(Lecture Notes in Civil Engineering 24) Fabio Bianconi, Marco Filippucci - Digital Wood Design - Innovative Techniques of Representation in Architectural Design-Springer International Publishing (2019

Lecture Notes in Civil Engineering
Fabio Bianconi
Marco Filippucci Editors
Digital
Wood
Design
Innovative Techniques of
Representation in Architectural Design
Volume 24
Series editors
Marco di Prisco, Politecnico di Milano, Milano, Italy
Sheng-Hong Chen, School of Water Resources and Hydropower Engineering,
Wuhan University, Wuhan, China
Ioannis Vayas, National Technical University of Athens, Zografou, Greece
Sanjay Kumar Shukla, School of Engineering, Edith Cowan University, Joondalup,
Perth, Australia
Giovanni Solari, University of Genoa, Genova, Italy
Anuj Sharma, Iowa State University, Ames, IA, USA
Nagesh Kumar, Department of Civil Engineering, Indian Institute of Science
Bangalore, Bengaluru, Karnataka, India
Chien Ming Wang, School of Civil Engineering, The University of Queensland,
St Lucia, QLD, Australia
Lecture Notes in Civil Engineering (LNCE) publishes the latest developments in
Civil Engineering - quickly, informally and in top quality. Though original research
reported in proceedings and post-proceedings represents the core of LNCE, edited
volumes of exceptionally high quality and interest may also be considered for
publication. Volumes published in LNCE embrace all aspects and subfields of, as
well as new challenges in, Civil Engineering. Topics in the series include:
– Construction and Structural Mechanics
– Building Materials
– Concrete, Steel and Timber Structures
– Geotechnical Engineering
– Earthquake Engineering
– Coastal Engineering
– Hydraulics, Hydrology and Water Resources Engineering
– Environmental Engineering and Sustainability
– Structural Health and Monitoring
– Surveying and Geographical Information Systems
– Heating, Ventilation and Air Conditioning (HVAC)
– Transportation and Traffic
– Risk Analysis
– Safety and Security
To submit a proposal or request further information, please contact the appropriate
Springer Editor:
– Mr. Pierpaolo Riva at pierpaolo.riva@springer.com (Europe and Americas);
– Ms. Swati Meherishi at swati.meherishi@springer.com (India);
– Ms. Li Shen at li.shen@springer.com (China);
– Dr. Loyola D’Silva at loyola.dsilva@springer.com (Southeast Asia and
Australia/NZ).
More information about this series at http://www.springer.com/series/15087

Fabio Bianconi Marco Filippucci
•
Editors
Digital Wood Design

Innovative Techniques of Representation
in Architectural Design
123
Editors
Fabio Bianconi Marco Filippucci
Department of Civil and Environmental Department of Civil and Environmental
Engineering Engineering
University of Perugia University of Perugia
Perugia, Italy Perugia, Italy
Contributions are published under double blind review mode
ISSN 2366-2557 ISSN 2366-2565 (electronic)

ISBN 978-3-030-03675-1 ISBN 978-3-030-03676-8 (eBook)
https://doi.org/10.1007/978-3-030-03676-8
Library of Congress Control Number: 2018963030
© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Contents
Part I Integrated Approach and Generative Design

WOOD, CAD AND AI: Digital Modelling as Place of Convergence
of Natural and Artificial Intelligent to Design Timber Architecture . . . . 3
Fabio Bianconi and Marco Filippucci
Beyond Form Definition: Material Informed Digital Fabrication
in Timber Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
David Correa, Oliver David Krieg and AnnaLisa Meyboom
New Workflows for Digital Timber . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Tom Svilans, Martin Tamke, Mette Ramsgaard Thomsen,
Jonas Runberger, Kai Strehlke and Martin Antemann
Negotiated Materialization: Design Approaches Integrating Wood
Heterogeneity Through Advanced Robotic Fabrication . . . . . . . . . . . . 135
Giulio Brugnaro, Angelo Figliola and Alexandre Dubor
Advanced Timber Concepts and the Design of Furniture
and Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Patrick Beale and Tristan Morgan
Lattice Shell Methodologies: Material Values, Digital Parameters . . . . 195
Mark Cabrinha, Dante Testolini and Ben Korman
Toward Mass Customized Architecture. Applying Principles of Mass
Customization While Designing Site-Specific, Customer-Inclusive
and Bespoke Timber Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
John Haddal Mork, Marcin Luczkowski, Bendik Manum
and Anders Rønnquist
Reciprocal Implications Between Design and Construction
Process of Timber Gridshell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Sergio Pone, Daniele Lancia and Davide Rando
v
vi Contents
Reinterpretation of Traditional Wood Structures with Digital

Design and Fabrication Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Philip F. Yuan and Hua Chai
Part II Centrality of Representation As a Synthesis Model

and Morphological Comprehension
Wood as an Essential Material in Architectural and Civil Engineering
Models from the Renaissance to the Architectural Avant-Garde . . . . . 285
José Antonio Franco Taboada
The Wooden Models of the Vatican Basilica by Antonio da Sangallo
and Michelangelo: Survey, Modelling and Interpretation . . . . . . . . . . . 321
Carlo Bianchini, Alfonso Ippolito and Luca J. Senatore
Representing with Wood: Carlo Lucangeli and the Model
of the Flavian Amphitheatre . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Graziano Mario Valenti, Cinzia Conti and Jessica Romor
The Teatro of Bologna as a Transformable Space: Drawing, Geometry
and Invention in the Study of the Wooden Model of the Theater
by Antonio Galli Bibiena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
Giuseppe Amoruso
Learning by Prototyping: Wood Design Course Experience . . . . . . . . . 395
Mariano Chernicoff
Parametric Kerf Bending: Manufacturing Double Curvature
Surfaces for Wooden Furniture Design . . . . . . . . . . . . . . . . . . . . . . . . 415
Mara Capone and Emanuela Lanzara
Digital Joinery for Hybrid Carpentry . . . . . . . . . . . . . . . . . . . . . . . . . 441
Shiran Magrisso and Amit Zoran
Geometric Patterns and the Art of Kundekari in Traditional
Turkish Woodworking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463
E. Sibel (Onat) Hattap
Digital Wood Trusses. Geometry and Parameters/Fabrication
and Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Andrea Giordano, Paolo Borin and Federico Panarotto
Digital Technologies for Cultural Heritage: 3D Representation
of Complex Wooden Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
Caterina Gottardi, Caterina Balletti, Simone Florian and Francesco Guerra
AHBIM for Wooden Built Heritage Conservation . . . . . . . . . . . . . . . . 533
Stefano Brusaporci, Alessandro Luigini, Starlight Vattano,
Pamela Maiezza and Alessandra Tata
Contents vii
Shape and Design of Ancestral Fishing Machines Between

Geometry and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 547
Domenico D’Uva
3D Digital Systems for the Documentation and Representation
of the Wooden Heritage Between Finland and Russia: Survey
Methods and Procedures for Detailed Analysis . . . . . . . . . . . . . . . . . . 565
Sara Porzilli and Stefano Bertocci
“Techos de Armaduras” in Cuban Colonial Architecture:
Cuban Coverings from XVII to XVIII Century . . . . . . . . . . . . . . . . . . 595
Cristiana Bartolomei, Caterina Morganti and Giorgia Predari
The Caribbean Bahareque: From Living Branches to the Wall . . . . . . 623
Massimo Leserri and Dayan-Ariadna Guzman-Bejarano
Part III Connection Between Nature Lessons and Material

Explorations
Computational Mechanical Modelling of Wood—From
Microstructural Characteristics Over Wood-Based Products
to Advanced Timber Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
Josef Füssl, Markus Lukacevic, Stefan Pillwein and Helmut Pottmann
Hygroscapes: Innovative Shape Shifting Façades . . . . . . . . . . . . . . . . . 675
Sherif Abdelmohsen, Sigrid Adriaenssens, Stefano Gabriele, Luigi Olivieri
and Rana El-Dabaa
Adaptive Wooden Architecture. Designing a Wood Composite
with Shape-Memory Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 703
Maryam Mansoori, Negar Kalantar, Terry Creasy and Zofia Rybkowski
Geometry-Induced System of Controlled Deformations. Application
in Self-organized Wooden Gridshell Structures . . . . . . . . . . . . . . . . . . 719
Efilena Baseta
Natural Complexity. An Introduction to Structural Design
with Tree Forks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743
Lukas Allner and Daniela Kroehnert
Evaluation of Wooden Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 761
Gülru Koca
Solid Wood and Wood Based Composites: The Challenge
of Sustainability Looking for a Short and Smart Supply Chain . . . . . . 783
M. Romagnoli, M. Fragiacomo, Antonio Brunori, M. Follesa
and G. Scarascia Mugnozza
viii Contents
Bamboo’s Bio-inspired Material Design Through Additive

Manufacturing Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
Francesco Buonamici, Yary Volpe, Rocco Furferi, Monica Carfagni,
Giovanni Signorini, Giacomo Goli, Lapo Governi
and Marco Fioravanti
Microtimber: The Development of a 3D Printed Composite Panel
Made from Waste Wood and Recycled Plastics . . . . . . . . . . . . . . . . . . 827
Sandra Karina Löschke, John Mai, Gwenaelle Proust
and Arianna Brambilla
Part IV Constructive Wisdom and Realization Challenges

Seeing a Tree as a Prerequisite to Timber Architecture . . . . . . . . . . . . 851
Rizal Muslimin
Constructing Correctly in Wood: New Insights into Timber
Technology Approaches Through Purist and Liberalist Schools
of Thought . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 871
Gabriel Tang and John Chilton
The Emergy of Digital Wood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 895
Chris Beorkrem
Housing Prototypes, Timber Tectonic Culture
and the Digital Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 911
Colabella Sofia and Gardiner Blair
Performative Architecture and Wooden Structures: Overview
on the Main Research Paths in Europe . . . . . . . . . . . . . . . . . . . . . . . . 937
Angelo Figliola and Alessandra Battisti
Adaptive Timber Towers. An Evolutionary Prototype for the 21st
Century Skyscraper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 971
Alessandro Buffi and Gian Maria Angelini
Knowledge-Based Design in Industrialised House Building:
A Case-Study for Prefabricated Timber Walls . . . . . . . . . . . . . . . . . . . 989
Graham Day, Eugenia Gasparri and Mathew Aitchison
(Re)construct with Wood, The Case Study of Amatrice’s
Prefabricated Bus Station Designed In BIM Environment . . . . . . . . . . 1017
Maria Grazia Cianci, Daniele Calisi, Francesca di Benedetto
and Matteo Molinari
Customizable Social Wooden Pavilions: A Workflow for the Energy,
Emergy and Perception Optimization in Perugia’s Parks . . . . . . . . . . . 1045
Marco Seccaroni and Giulia Pelliccia
Contents ix
Empathic Architecture: Digital Fabrication and Community

Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1063
Hiroto Kobayashi and Don O’Keefe
Non-orthogonal Light Timber Frame Design: Using Digital
Manufacturing Technologies to Facilitate Circular Economy
Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1087
Gerard Finch and Guy Marriage
Timber Plate Shell Structures: A Digital Resurgence
of Traditional Joining Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1117
Christopher Robeller
Computationally Derived Cross-Laminated Timber
Reinforcement and Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1135
Jefferson Ellinger, Chris Beorkrem and Calum Dodson
Beech Wood for Architectural Design: Three Studies Case
from an International Design Contest Terres de Hêtre® . . . . . . . . . . . 1151
Giuseppe Fallacara, Antonio Pantaleo and Giuseppe Scaltrito
Brise-Soleil House: Developing Software Tools to Enable
the Integration of Design and Fabrication in Timber . . . . . . . . . . . . . . 1183
Jonathan Nelson, Chris Knapp, Rory Spence and Joel Hutchines
Towards a 4.0 Mass Customized Wooden Housing
in the Mediterranean Area: The Ecodomus Project . . . . . . . . . . . . . . . 1201
Micaela Colella and Giuseppe Fallacara
HOUSE 1 Protostructure: Enhancement of Spatial Imagination
and Craftsmanship Between the Digital and the Analogical . . . . . . . . . 1229
Dieter Dietz, Dario Negueruela del Castillo, Agathe Mignon
and Julien Lafontaine Carboni
Part V Parametric Transfigurations and Morphological

Optimizations
KODAMA: A Polyhedron Sculpture in the Forest at Arte Sella . . . . . 1255
Kengo Kuma, Marco Imperadori, Marco Clozza, Toshiki Hirano,
Andrea Vanossi and Federica Brunone
Wooden Byobu. From Architectural Façade to Sculpture . . . . . . . . . . 1287
Atsushi Kitagawara, Marco Imperadori, Ryosuke Kuwabara,
Federica Brunone and Mayuko Matsukawa
Experimenting the Use of Wood in Contemporary Architecture:
Integrating Research into Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1317
Salvator-John A. Liotta
x Contents
Digital Design Thinking in Architectural Education Testing

Idea-Driven and Science-Driven Design Processes Towards
Researching Polymer/Wood Composite Structures . . . . . . . . . . . . . . . . 1355
Andrei Gheorghe
Digital Construction of Timber “Metabolism” . . . . . . . . . . . . . . . . . . . 1395
Yasushi Ikeda
Parametric Modeling of a Wooden Folding Structure . . . . . . . . . . . . . 1409
Tuğrul Yazar
Common-action Gardens: Performative and Parametric
Prototypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1421
Fulya Akipek
Algorithmic Craftsmanship For Bespoke Timber Architecture . . . . . . 1439
Sille Pihlak and Siim Tuksam
FracShell: From Fractal Surface to a Lattice Shell Structure . . . . . . . 1459
Iasef Md Rian
Developable Wooden Surfaces for Lightweight Architecture:
Bio-Dune Pavilion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1481
Andrés Martín-Pastor and Rodrigo García-Alvarado
SMALL IS MORE. Wooden Pavilion As a Path of Research . . . . . . . 1501
Marco Imperadori, Graziano Salvalai, Andrea Vanossi
and Federica Brunone
Introduction
Grandi invero sono le cose che in questo breve trattato io

propongo alla visione e alla contemplazione degli studiosi
della natura. Grandi, dico, sia per l’eccellenza della materia
per se stessa, sia per la novità loro non mai udita in tutti i
tempi trascorsi, sia anche per lo strumento, in virtù del quale
quelle cose medesime si sono rese manifeste al senso nostro.
In the present small treatise I set forth some matters of great
interest for all observers of natural phenomena to look at and
consider. They are of great interest, I think, both from their
intrinsic excellence, and from their absolute novelty, and also
on account of the instrument by the aid of which they have
been presented to my apprehension.
Galileo Galilei—Sidereus Nuncius (1610)
DWD
Digital Wood Design. Innovative Techniques of Representation in Architectural

Design intends to undertake the digital representation strategies that can change the
future of wooden architectures, through the combination of tradition and innova-
tion. Calling on anyone who intends to bring innovation and experience in wood
construction field, this volume is related to the advanced digital modeling with a
particular attention on the solutions associated with generative models and dynamic
value, inherent to the relation between knowing how to draw and how to build, a
relation that finds its foundation in geometry. Although wood is one of the oldest
construction materials, in recent years many innovations have affected the pro-
duction techniques and the design tools, and they have paved the way for new
formal, esthetic, and structural solutions, which can fill the application fields of this
material. Lightness and eco-compatibility of features make this material a valuable
alternative to materials that have characterized the recent architectural debate,
offering the ability to build with sustainable, fully renewable, and high-performance
materials in energy and structural aspects, capable of complying with architectural
solutions of great freedom. Thanks to the potential of computing, parametric design
xi
xii Introduction
and digital manufacturing are decisively contributing to renew a spinneret to look

forward in driving the future of construction.
The book is developed in the centrality of representation, starting from the lesson
of Marshall McLuhan about the correspondence between medium and message.
Representing corresponds to create a model, a condition marked by digital revo-
lution through the application of computational potentialities.
The new role of representation derived by digitalization is connected to the
techniques, but it is firstly a cultural question, an expression of the cultural passage
from a society of the image to a society of simulation. If maybe digital representation
is developed in correspondence of precision and replication necessity, if the com-
munication aspect promoted a valorization of tri-dimensional model, step by step, in
the predominance of morphological approach, the possibility to represent the
complexity of the reality has defined the role of digital design as fulcrum: On one
hand, it is becoming day after day the place where to converge different data, and on
the other hand, the medium to replicate the dynamical relation and the processes,
extending the question to time dimension. As in the digital era as in the prehistoric,
representation is the instrument useful to understand the world, to analyze and break
up the phenomena, and to prefigure future scenarios. Representation is a transdis-
ciplinary language, able to integrate specific technical language, by using the
abstraction of geometry to transform the analogy of the real in a simpler and
understandable discretized interpretation. It is so important to understand how these
tools are used to build the model, because in this interpretation it could be hidden a
reductive simplification or conceptual misunderstandings.
Wood represents a really contemporary material, rethought and transfigured by
design, always carrier of its natural essence, unable to forget the work of man and
the traditional lessons. Wood design is returning to have also in Europe a central
question, according also to its nature of renewable resource that can respond to the
energetic and climatic challenges. Digital studies are transforming wood production
and the workflows in the construction process, so that the distance between project
and realization is reducing more and more.
Digital wood design integrates different approaches, from the historical research
to the material analysis, the architectonical research of freedom. Different points of
view, the implicit need of interdisciplinary, support the explication of the hidden
relation is the basis of construction issues. But the particular approach of this
volume is connected to mark the centrality of the model and the value of digital
medium.
Drawing, model, and architecture are inseparable aspects of the same question.
Deepening the representative question helps to find the origin, the sense, and the
roots of our architectural aims.
Introduction xiii
Drawing as Model
Galileo Galilei is titled by Albert Einstein as the father of the modern science,
according to his introduction of the synthesis of model to the study of phenomena.
But Galileo found in the instrument, in his innovative tool, the fundamental help to
support his genius and his methodology. Medium is the message, as Marshall
McLuhan teaches us. This condition is really strong in our contemporary era,
characterized by a digital soul.
Digital wood design, a theme centered on innovation and digital tools, which
cannot forget the lesson of the history. The performance of instruments reduces
timing, which is dedicated to the reflections about essential questions and the aims of
architecture. We editors have the courage to promote this approach to digital wood
design not in virtue of the value of our innovative research, but because of the role of
Italian school in drawing studies, but also because the medium is useful to all. We
want to involve the tradition to help a critical vision of the future development.
The master of the past shows us contemporary lessons about the value of
drawing as model. In particular, Leon Battista Alberti, in 1485, starts his treatment
De re aedificatoria introducing the centrality of this field of research. It is necessary
to analyze the original Latin version because in some English texts it is often not
presented in a correct translation:
Tota res ædificatoria lineamentis et structura constituta est.
Lineamentorum tota vis et ratio consumitur, ut recta absolutaque habeatur via coaptandi
iungendique lineas et angulos, quibus ædificii facies comprehendatur atque concludatur.
Atqui est quidem, lineamenti munus et officium praecribere aedeficiis et partibus aedifi-
ciorum aptum locum et certum numerum dignumque modum et gratum ordinem, utiam tota
aedificii forma et figura ipsis in lineamentis conquiescat.
Neque habet lineamentum in se, ut materiam sequatur, sed est eiusmodi, ut eadem plurimis
in ædificiis esse lineamenta sentiamus, ubi una atque eadem in illis spectetur forma. Et
licebit integras formas præscribere animo et mente seclusa omni materia; quam rem
assequemur adnotando et præfiniendo angulos et lineas certa directione et connexione.
Hæc cum ita sint: erit ergo lineamentum certa constansque præscriptio concepta animo
facta lineis et angulis prectaque animo & ingenio erudito
Architecture is constituted by drawing and construction. The impact and the ratio of the
drawing lie in finding the correct and unique way of joining and fitting together those lines
and angles through which the building images are defined and understood. The function
and duty of the drawing consist then to assign to the buildings and to the parts that compose
them an appropriate place, an exact proportion, a convenient order, and a harmonious
arrangement, so that the whole form and figure of the building subsides in the drawing
itself. So, the drawing doesn’t have inside itself anything derived by matter; and that is why
we experience the same drawing in more buildings, where the visible form is just one. And
it will be possible to represent the integrated designs in our thought and imagination
entirely separate from matter, a condition that we understand by recording and presetting
angles and lines to defined directions and connections. This being so, the drawing will be a
defined and constant design, conceived in the mind, made by lines and angles, and carried
out by a person endowed with intelligence and culture.
xiv Introduction
The philological analysis of one of the most important architectonic treatises can
show us the full comprehension and anticipatory forecast that helps us to deeply
understand the sense of our research, especially in the digital age.
Lineamenta, drawing, modeling. The first reflection can be addressed to the
translation of the word “lineamenta”: Some English authors use the word “design”
or “lineaments,” but lineamanta is clearly connected to a system of lines, so as in
the first Italian translation the word used is “disegno,” now we can use the word
“drawing.” But inside this word, there is a deepening sense of these tools: Line, in
fact, is the first algorithm for primitive man; it defines the birth of culture, a magic
medium that transforms the void in form. Drawing is the system of lines, the
conscious use of this medium to understand the real. In this conception is forecast
our contemporary idea of generative modeling, composed by a link of simple
algorithm to define a digital path. Drawing is the modeling.
Architecture, drawing, modeling. Alberti shows us a series of considerations
about the value of drawing: “The architecture is constituted by the drawing and the
construction.” Drawing, as our digital models, is the constitution, the essence, and
the statute of architecture, which find its form in the construction, in the real
concretization. Alberti marks how architecture is an integrated process from the
virtual space of the idea to the defined form of the construction, these are two
structures composed together. But in Alberti’s theory, drawing and design are the
work of the architect, and construction is just a transcription of the idea. The model
is the space of generation, the place where it is possible to connect together the
elements. The main point of this sentence is the interpretation of a hybrid definition,
which is predominated by the virtual that defines the construction process.
Drawing code. About drawing, Alberti shows us as “tota vis et ratio,” its vitality,
its impact, its force, and its reason, but also the methods, “his way,” is not the role of a
simple tool or the results of a path, but it is an identification of relations and
instruction for architecture realization. In the words of the great theorist, drawing as
model is a net, defined by the connections of the designed forms multi-aspects.
Drawing is a system for the form-finding, aimed to have, after a path, after a hard
work in mind, the definition of elements and connections, a system that in virtuality
cannot be clearly verified. Drawing, in fact, is a scientific code, “correct and unique,”
absolute way of joining and fitting together lines and angles. Model and drawing are
both founded in the discretization of the elements: The digital logic is the expression
of man’s needs, but its goal is to return to the organicism of the architecture. These
relations are hidden in the form that expresses the complexity of architecture. Lines
and angles are the elements that configure the form of the building. In our digital
logic, it could be related respectively to the vector element and to the links between
them. Through these elements and these relations, through the construction of the
generative modeling, it is possible to obtain two different results: One is a defined
architecture, because the drawing is a scientific construction, and it is a two-way
function between the domain of the reality and the codomain of the ideas; the other,
marked by Alberti, shows a support process to understand, literally it takes together,
it transforms the multiplicity of perceptions and elements in a whole, the building
Introduction xv
images, not the form, but the results, the impact, the esthetic value, so that the value
of the drawing is found in the simulation function.
Drawing as language. Drawing is founded in a scientific language but, at the
same time, model is a cultural path. Alberti is a son of his time, and his words are full
of humanistic culture. His architectural theory finds a central reference in Vitruvius,
who, in De Architectura, writes that “proportion is a correspondence among mea-
sures of the members of an entire work, and of the whole to a certain part selected as
standard.” Drawing derives from measurement operations, in Latin mensura, from
metiri, which means to “distribute,” “divide.” Dividing up a space it involves
applying a rule, defining a relation; this is making architecture. From this, it results
the principles of symmetry. Without symmetry and proportion, there can be no
principles in design of any time; that is, if there is no precise relation between its
members, as in the case of those of a “well-shaped man” (Vitruvius, De Architectura,
III, 1, 1). He proposes “Six Principles of Design” as order (ordinatio), arrangement
(dispositio), proportion (eurythmia), symmetry (symmetria), propriety (decor), and
economy (distributio) (Vitruvius, De Architectura, I, 2, 1): Order consists in adapting
the right measure to single elements of the work; arrangement is their appropriate
collocation; eurythmy is the harmonic vision of the proportion between parts;
symmetry is the correspondence between each single part in the whole set; propriety
is the formal improvement; economy is the equilibrated administration of resources
and space (Vitruvius, De Architectura, II). Modeling as cultural language, as an
expression of a study of these relations, is not just a set of commands, it is not just a
procedure, but it is a path aimed to create architecture and its deeper esthetic value.
The virtuality of the drawing. Figuration recalls the concept of imaginary, the
interpretive action that gives a sense to the perceived form. This conceptualization,
that defines the real essence of modeling, is one of the central questions made by
Alberti, who marks in a different way, underlining the absence of materiality in this
differentiation between figure and form. Here is a reference in the value of vision,
so important in the construction of the model. Drawing is anyway an abstract
process, and in this, as a condition that defines the value itself of architecture, in its
connection with culture. Drawing defines itself in the virtual space of imagination,
in “animo et mente,” through the experience and feeling. There is a centrality of
virtual memory, so close to our digital approach, in the process of elaboration of
stimuli, in a centrality of the visual impact for the architectonic finalities.
Figure and form. Another central aspect introduced by Alberti is the distinction
between figure and form: The term “figure” derives from fingo, which means “to
give form,” with a mainly plastic connotation, so much that figolus is the potter. It is
distinguished from “form,” making clear etymologically the sense of “hold,” as it
testifies the correlation with the word “stop,” with which the linguistic root is in
common. If then “form” has a static sense of stabilizing and concretizing, “figure”
is tighter to dynamic, because form is defined as the arrest of a process. Figure is
also heuristic, abstracting from the visible matter through perception and through
reasoning form models, expression of the necessity of the eidetic-constructive
research for the comprehension of what is observed: It is in this context that is more
comprehensible because in geometry the figure is a set of points or lines or surfaces,
xvi Introduction
which present a whole construction. Both find their reason and their address “in
lineamentis conquiescat,” to stop itself, to find a quiet and a rest of mind process,
always dynamic and in exploration. In drawing, it is necessary to have choices, to
define logic, and to determinate univocal relations. Drawing is not an idea but the
concretization and the place of comprehension of the ideas, because, as written by
another great theorist of architecture, Giorgio Vasari, “il disegno è apparente
espressione e dichiarazione del concetto che si ha nell’animo, e di quello che altrui
si è nella mente immaginato e fabbricato nell’idea” (drawing is an apparent
expression and declaration of the concept that a person has in the soul, and of what
others have imagined in mind and built in the idea) (Vasari, Le vite, 1568, 11).
Drawing as generative modeling. Drawing is the set of rules and relations;
once it is frozen in the form, as a geometrical conceptualization, a triangle can
generate infinite forms, each one different from the others in terms of measure,
proportion, relations, etc. For this reason, “we experience the same drawing in more
buildings, where it is possible to define one and the same form”: Generative
drawing, the relation that defines modeling, is a reality with its own life. This
conceptualization is really close to our approach to digital design, in the mass
customization logic, in the replication of the design field of application. Centrality
of variation is another theme of the classicism, as Alberti remembers, in the same
book, when he writes: “Variety is always a most pleasing space, where distant
objects agree and conform with one another; but when it causes discord and dif-
ference between them, it is extremely disagreeable” (Alberti, 1485, I, IX). Ante
litteram, Alberti shows how parametric principles are the basis of architectural
theory, founded in the composition of this relations, showing how also in this field
difference is not diversity, in the unity of the model.
Integrated design. In the Latin text, Alberti defines the possibility of “integras
formas præscribere” that can be translated as “representing integrated designs.”
This interpretation is defined by the etymological sense of the word
“prae-scribere,” literally “writing before,” the signs made in a previous time,
connected to logic of “project,” throw forward, in the same sense of “re-present.”
“Integras formas” could be translated also as “defined solutions,” uniform and
complete forms, but the sense of this word is connected to the possibility of making
complex results. “Integras” is whole, full, derived by “integer,” in the sense of
making something complete. Integrated design is very close to our digital approach
where the distance between projects and construction is ever more reduced, also if
the project is separated from the material, and it finds his field of existence in the
virtuality of representation. In this context, for example, the actual condition of
wood design shows the breaking of the conditioning of material in form and
structure because wood itself is engineered. The real issue is the relation between
form, structure, and matter, which is understood in this text when the authors
connect the understanding process to the record and the presetting of geometrical
elements, interpreted as our concept of vectors and links. The integrated design is a
sort of second step; it is a complexification of the results when the model is stressed.
Introduction xvii
Hybridization as synthesis. To conclude his digression, Alberti proposes a

synthesis to describe all the aspects of drawing process: It is defined and constant
design, in this dialect between form and figure, between virtuality and reality,
marked also by the word constant, something that standing firm, invariable, and
uniform. Drawing is conceived in the mind, made by lines and angles, through
geometrical elements but also through their relations. In this sense, the model is
something hybrid, as asserted in the first sentence, connected to the construction, to
the interpretation of the complexity hidden in the architectural shapes, and to the
esthetic value of architecture. For this reason, drawing is reconnected to man, to his
author. In this humanism, the value of the drawing is not reduced to the tool, it is not
a simple machine, but it is “carried out by a person endowed with intelligence and
culture,” “animo & ingenio erudito,” a man with heart and mind, because drawing,
also in our digital world, always maintains its evocative purpose, using high-tech
solutions, in the deepen coordinates of architecture, which Alberti descript as “ad
vitam bene beateque agendam faciant,” building to have a better and happier life.
Starting from the origin and originality of the first architectural treatises, digital
wood design can find and redefine new models to innovate architecture.
The Structure of the Volume
The volume is composed of 61 chapters, written by 153 authors from 5 continents,

24 countries, and 69 centers of research (57 universities). These numbers witness
the international approach of the question, inedited integration of different schools
in the world.
In the value of integration and interdisciplinary, the richness of these variations
in the theme is organized in five parts that connect digital wood design to the
integrated approach and generative design, to the model synthesis and morpho-
logical comprehension, to nature lessons and material explorations, to constructive
wisdoms and realization challenges, and to parametric transfigurations and mor-
phological optimizations.
The first section concerns the integrated approach and the generative design
developed with different approaches by groups of research from all over the world.
The volume starts from our chapter (Bianconi, Filippucci) about the evolution of
digital representation in timber architecture, describing the activity of our group of
research in the balance between AI and NI and showing the researches applicative
developed founded in computational design. The section continues in the interna-
tional cooperation between Canada and Germany in the architectural language
through wooden prototype, in a critical approach to recent technical developments
in design, fabrication and material innovation and supported by a re-appropriation
of interdisciplinary teaching methods (Correa, Krieg, Meyboom). The theme of the
Danish Centre for IT and Architecture is new workflows for digital timber, which
explore how computation and a challenging of traditional material practice can
impact the use of timber in architectural design and fabrication (Svilans, Tamke,
xviii Introduction
Ramsgaard Thomsen, Runberger, Strehlke, Antemann). IAAC and at the Bartlett

School of Architecture show their collaboration in a series of project, case studies of
tolerance management in robotized fabrication with different kind of wood (Dubor,
Figliola, Brugnaro). The timber design across a variety of scales, from furniture to
structures, is the theme of the research activity developed by the Australian
Advanced Timber Concepts Research Centre (Beale, Morgan). The Californian
group of research describes Geodesic Lattice Shell Methodologies, hybrid approach
in modeling of these surfaces those combine an elegance of forms with the effi-
ciency of structure driven by the material constraints of straight lath members that
can be bent into shape (Cabrinha, Korman, Testolini). The NTNU Norwegian group
analyzes the mass customized architecture obtained through the application of
designing site-specific, customer-inclusive, and bespoke timber structures (Haddal
Mork, Luczkowski, Manum, Rønnquist). The gridshell.it research group linked to
Naples University describes reciprocal implications between design and construc-
tion process of timber gridshell, showing the development of a new information
technology design tool (Pone, Rando, Lancia). This part ends with the digital wood
researches of the Chinese Tongji University’s DDRC, centered in the relation
between traditional culture and innovation, a condition that involves also timber
architecture, in a digital design process based on the reinterpretation of traditional
wood structures that ends with fabrication technologies (Yuan, Chai).
The second part of the volume is based on the centrality of representation as a
synthesis model and morphological comprehension, deepening in lessons always
proposed by the cultural heritage and the value of drawing as knowledge instru-
ments. The section begins with an overview about the role of wood models from the
Renaissance to the architectural avant-garde as an alternative method to virtual
representations (Taboada). In connection here are reported three central case study
of the Wooden models of Vatican Basilica (Bianchini, Ippolito, Senatore), the
Coliseum (Valenti, Romor, Conti), and Bologna’s theater (Amoruso), paradigmat-
ical also if miniature architectures surveyed, represented, analyzed, and understood
by the support of the advanced digital tools. This part continues analyzing the use of
models and prototype in design practice (Chernicoff), the application of advanced
techniques as parametric kerf bending for double curvature surfaces for wooden
furniture design (Capone, Lanzara), and the integration of technique in digital
joinery for hybrid carpentry (Magrisso, Zoran). The connection between design and
wood is analyzed also in the study of cultural heritage, as in the case study of
geometric patterns in the traditional Turkish woodworking (Hattap). This part
continues reporting the research in architectural representation of wooden cultural
heritage, reporting at the beginning the Italian heritage documentation in the
application of HBIM in the Church of Eremitani at Padua (Giordano, Borin,
Panarotto), in Venetian complex wooden structures (Gottardi, Balletti, Guerra,
Florian), in South Tyrolean farm (Brusaporci, Luigini, Vattano, Maiezza, Tata), and
in the study of traditional fishing machines (D’Uva). The same Italian school is not
limited to local landscape, but it exports also to study Finnish and Russian heritage
(Bertocci, Porzilli), techos de armaduras in Cuban colonial (Morganti, Bartolomei,
Predari), and the Caribbean Bahareque in Colombia (Leserri, Guzman Bejarano).
Introduction xix
The third part describes the relation between nature lessons and material explo-
rations, in the value of digital design that involves also materiality. This part ana-
lyzes the mathematical interpretation inside the computational mechanical modeling
of wood, starting from microstructural characteristics over wood-based products to
advanced timber structures (Füssl, Lukacevic, Pillwein, Pottmann). The section
continues reporting research connected to the valorization of natural characteristic of
wood, as in the case of hygroscapes, understood as an Innovative Shape Shifting
Facades developed in an international cooperation between American University in
Cairo; Arab Academy for Science, Technology and Maritime Transport; Princeton
University; and the University of Roma Tre (Abdelmohsen, Adriaenssens, Gabriele,
Olivieri, El-Dabaa). In the same logic, the design of a Wood Composite with shape-
memory behavior is also connected to a responsive wooden architecture developed
in Texas (Mansoori, Kalantar, Creasy, Rybkowski). A similar approach is developed
in the exploitation of the physical response of wood, as in the innovation proposed in
a construction system to create scalable adaptive stiffness structures (Baseta) and in
the study of the conceptual joining linked to branch formations, (Allner, Kroehnert),
both the researches developed firstly in the University of Applied Arts of Vienna.
The study of the material increases in a detailed scale in the studies promoted in the
evaluation of mechanical properties of existing timber structures (Koca). Solid wood
and based wood composites and their potential of a short procurement chain of wood
represent the investigation of a national cooperation between Italian center of
research (Romagnoli, Scarascia Mugnozza, Fragiacomo, Follesa, Brunori). The
empirical research arrives to transfigure natural material, as reported in Bamboo’s
Bio-inspired Material Design Through Additive Manufacturing Technologies
developed in Florence University (Buonamici, Volpe, Furferi, Carfagni, Signorini,
Goli, Governi, Fioravanti). Also, on the other part of the world, in Australia, wood is
transformed into a 3D printed, gradient timber panel composed of forestry waste and
by-products (Löschke, Mai, Proust, Brambilla).
The fourth part concerns the constructive wisdom and realization challenges, by
analyzing the conceptualizations but also the paths to transform the research in
building. This section is opened by a shape grammar lessons derived from a tree saw as
a prerequisite to timber architecture (Muslimin). The second chapter presents a new
perspective for designers to work with timber in the twenty-first century, in the Nervi’s
request of “Constructing Correctly” in the values of the person places on material
understanding and constructional methods within specific context, analyzing also the
series of applications (Tang, Chilton). This part continues with an examination about
the emergent theme in digital wood design, exploring the ways in which researchers
have been exploring the use of technology to expand the use of wood while mini-
mizing the energetic impact of the construction (Beorkrem). Housing prototypes,
timber tectonic culture, and digital age are the theme of the following chapter, which
raises a series of discussion points centered on the role of timber-based products, in a
digitally enabled domestic construction industry (Colabella, Gardiner). Another two
overviews concern the main research paths in Europe, marking the research in
performance-based design, material culture and fabrication process (Figliola,
Battisti), and the wooden skyscrapers, connected to the design experience
xx Introduction
(Buffi, Angelini). This section continues analyzing the knowledge-based engineering

in timber construction, applied in industrialized building and data management by the
Australian group of research managing the complex and multidisciplinary nature of
design, fabrication, and installation (Day, Gasparri, Aitchison). The earthquake-proof
is one of the reasons for the application of this technology in territories afflicted by
earthquakes. The use of wood timber, designed in BIM environment, is descript in the
case study of Amatrice’s prefabricated bus station (Cianci, Calisi, Di Benedetto,
Molinari). The application of generative path is applied by the group of University of
Perugia (Seccaroni, Pelliccia) in the customizable social wooden pavilions, design
based on a workflow for the energy, emergy, and perception optimization in Perugia’s
parks. From Japan, an innovative methodology became the “Veneer House” project,
an empathic design founded in an agile technology applied in different solutions
(Kobayashi, O’Keefe). From New Zealand, it derived the proposal of an innovative
non-orthogonal assembly optimized light timber frame construction design, realized
by using digital manufacturing technologies (Finch, Marriage). From Germany,
another really important international case study regards the use of timber plate Shell
Structures, a historical technique of construction that becomes the inspiration for new
assembly methods according to the efficient realization that can now be achieved
through algorithmic geometry processing (Robeller). From USA, the proposal of a
computationally derived cross-laminated timber reinforcement and construction is
developed, a solution derived from the recent advances in computational analysis and
simulation (Ellinger, Beorkrem, Dodson). The application of the innovative tech-
nologies is applied also to the construction, as in the application of Beech wood for
architectural design, rewritten by a series of geometric and topological variations on
the theme of the wood-frame roof (Fallacara, Pantaleo, Scaltrito). Another really
interesting case study is developed in Papua Guinea, the Brise-Soleil House, a system
for seamlessly integrating design, fabrication, and assembly of a geometrically
complex timber veil (Nelson, Knapp, Spence, Hutchines). The successive case study
concerns the Italian Ecodomus project, a proposal that sees the mass customized
wooden housing in the Mediterranean area, a prototype created by an irregular
geometry made up of CLT panels and using digital fabrication technologies (Colella,
Fallacara). This section is concluded with the description of the HOUSE 1 as scaf-
folding and protostructure, experimentation of Wood Collaborative Design and
Construction in First Year Studio Teaching at EPFL (Dietz, Negueruela del Castillo,
Mignon, Lafontaine Carboni).
The last part regards the parametric transfigurations and morphological opti-
mizations, reporting firstly the case study of pavilions and installations from all over
the world, the spaces where it is possible to freely express and exalt digital wood
design. The section starts from the installation of KODOMA ideated by Kengo
Kuma in a context of international cooperation that sees the direction of Marco
Imperadori, installed in the Italian land art park Arte Sella, a polyhedron made only
by one wood section and connected without nails or screws (Kuma, Imperadori,
Hirano, Clozza, Vanossi, Brunone). In collaboration with the same Italian group,
Atsushi Kitagawara shares his installation Byobu, a reinterpretation of the really
famous EXPO Milano 2015 Japanese Pavilion, a naked structure defined by only
Introduction xxi
one element, conceived with studs, connected without any screws or nails, thanks to
the “compressive-tension” effect, whose concept interprets old Japanese
wooden-based techniques of constructions and even handcrafts of complex wooden
toys (Kitagawara, Imperadori, Kuwabara, Brunone, Matsukawa). The other chapter
reports at the same time wooden temporary architecture and constructions realized in
Japan, Italy, and France, developed through digital fabrication and controlled by
generative design, as an expression of the translation of the research into a real
high-quality architecture (Liotta). In the integration of idea-driven and
science-driven design process, a series of full expressive wooden pavilions witness
the combination between digital architectural design processes and digital design
fabrication technology, and it combines digital learning chain (Gheorghe). From
Japan, in a reinterpretation of metabolism in front of digital contamination, here is
reported the research experiences in digital wood design, showing its impact on
modularity, adaptability, compatibility, and transformability of architectural ele-
ments, in the paradigm shift of architectural design and technology between mass
production and smart production (Ikeda). Two experiences from Turkey show a
wooden folding structure, the A-Chord, a design family of non-standard wooden
structures, light-weight, portable (Yazar), and re-usable, and the common action
garden, expression of the integration between digital design and fabrication tech-
nologies with sustainable landscape issues (Akipek). From the Estonian Academy of
Arts, starting from local wooden house manufacturers, the architectonic research
shows in the installations the results of a young and vibrant digital design integrated
into a smart fabrication process, a way to return to a craftsmanship affordable for a
wider public (Tuksam, Pihlak). “FracShell” is a computational morphogenesis
project, based on the transformation of fractal geometry (Takagi–Landsberg’s fractal
surface) into a lattice shell structure, a demonstration of the easy manual construction
of a complex digital design taking the advantage of the rule-based shape but also
of the versatility of wood in transforming the complex digital design into its
real-scale physical structure (Rian). Within the broad context of teaching and
research, Bio-Dune Pavilion is the results of a project that involves a pavilion based
on developable wooden surfaces, parametric design, and digital fabrication, which
are presented in collaboration between Ibero-American universities; the project is
addressed to applications of wood to architecture through the CNC cutting of plates,
thanks to the properties of developable surface (Martín-Pastor, García-Alvarado).
The volume ends with the presentation of several shelters recently designed by the
Politecnico di Milano through high-level international cooperation, characterized all
by the smallness, a digital codesign that demonstrates the centrality of parametric
modeling and of the materiality, in the great potential of wood as a construction
material, according to the Vitruvian terms of structures, functions, and esthetics
(Imperadori, Salvalai, Vanossi, Brunone).
Part I
Integrated Approach
and Generative Design
WOOD, CAD AND AI: Digital Modelling
as Place of Convergence of Natural
and Artificial Intelligent to Design
Timber Architecture
Fabio Bianconi and Marco Filippucci
Abstract The contemporary development and digital culture in architecture, from

the idea to the realization, lead to a rewriting of the coordinates of the deep relation
between model and pre-figuration, especially in the timber structure field. Artifi-
cial intelligence opened new potentialities that rewrite the project paths through the
evaluation of computational design, with a model set as the place of simulation
and experimentation, in order to locate solutions for more and more high requests
made by architecture. Wood’s natural intelligence inspires artificial intelligence’s
principles, and it is projected as the new frontier of the research, in its possibility of
defying optimized solutions also in function of multiples objectives and parameters.
Wooden architecture design correlated to a history of tradition, which is established
on descriptive geometry, finds today multiple application fields for the research. In
this sense, representation supports the knowledge and the innovation, able to con-
tinue and express its operative aspect full of culture and, at the same time, its tecné
sense, which etymologically it is meant as art and technique. The present chapter
shows different ways to apply the contemporary principle of descriptive geometry
in digital wood design research, in a multidisciplinary and contaminated learning
environment. In all the illustrated cases, the generative design has a central role,
in an integration addressed to the need of optimization of architectural form, using
Genetic algorithms in order to analyze and to understand the relationship between
form, geometry, and construction.
Keywords Representation · Generative design · Artificial Intelligent
F. Bianconi · M. Filippucci (B)

Department of Civil and Environmental Engineering, University of Study of Perugia, Perugia,
Italy
e-mail: marco.filippucci@unipg.it
F. Bianconi
e-mail: fabio.bianconi@unipg.it
© Springer Nature Switzerland AG 2019 3

F. Bianconi and M. Filippucci (eds.), Digital Wood Design, Lecture Notes in Civil
Engineering 24, https://doi.org/10.1007/978-3-030-03676-8_1
4 F. Bianconi and M. Filippucci
1 Introduction
The new paradigm of digital tectonics is developing the coincidence between geomet-
ric representations of structuring and the program to modulate them (Oxman 2006,
pp. 229–266; Oxman and Oxman 2010b). In architectural representation, the model
allows to analyse and simulate according to the behaviours of structures and mate-
rials with the energetic performances of the analysed buildings (Brown and Mueller
2016). For example, the innovation of Building Information Model (Eastman 2011)
has to be considered firstly connected to the management of Big Data, to the pos-
sibility of replicating in virtual field the immaterial implications of the form. These
conditions tell us how CAD is orienting to Computational Design for Architecture,
and how now here it is changing the design process (Menges and Ahlquist 2011).
The theme of simulation (Baudrillard 1981) is the centre point of the actual devel-
opment of digital representation, which is connected to the theme of augmented real-
ity and the actual trend to replicate in a three-dimensional simulacrum the form that,
until now, was expressed in the planarity of the image.
Simulation is also related to the themes of prototyping (Chen and Sass 2017;
Larry Sass and Oxman 2006) manufacturing (Kolarevic 2004; Kolarevic and Klinger
2008) and fabrication (Austern et al. 2018; Corser 2010; Krieg et al. 2014; Sakamoto
and Ferré 2008; Sheil 2005), with a 3D printing (Correa et al. 2015; Le Duigou
et al. 2016), CAD/CAM system (Chaszar and Glymph 2010; Lawrence Sass 2012;
Lawrence Sass and Botha 2006) and robotics (Eversmann et al. 2017; Gramazio and
Kohler 2014; McGee and Ponce de León 2014; Menges 2012, 2013; Menges et al.
2017b), it developed according to the possibility of new tools to deconstruct the forms
for the constructive process. In this sense, if “architecture continually informs and is
informed by its modes of representation and construction, perhaps never more so than
now, when digital media and emerging technologies are rapidly expanding what we
conceived to be formally, spatially and materially possible” (Iwamoto 2009, p. 4). We
are living in a cultural revolution, because “fabrication is not a modelling technique,
but a revolution in the making of architecture” (Oxman and Oxman 2010a, p. 24).
Digital representation is ever more characterized by an hybridization between reality
and virtual, in the simulation that defines also the physical result of the architecture.
The digital representation is ever more characterized by a hybridization between
reality and virtual, in the simulation that defines also the physical result of the archi-
tecture.
Anyway, it is possible to replicate aspects connected not only to the morphologi-
cal aspects, or organized by the constructive workflow, it includes the form-finding
approach to functions like energy consumption or structural deformations also related
to the form (Adriaenssens et al. 2014; Menges 2012). The simulation firstly it opens
to interaction, as a new way to involve the user in the comprehension of architectural
implications in form and function (Greenough 1947).
This condition is strictly related to the revolution of parametric (Schumacher
2011). The extrapolation of logical connections in the construction of the form is
based on the individuation of relations and dependences (Jabi 2013), defining ranges
WOOD, CAD AND AI: Digital Modelling as Place … 5
Fig. 1 Digital wood: architectural survey and modelling of natural shapes of an olive tree
of the possible input data to determinate the options. The centrality of the model as
an organic system of relations, it defines the variation as a result, according to the
question that “difference is not diversity. Diversity is given, but difference is that by
which the given is given … Difference is not phenomenon but the noumenon closest
to the phenomenon” (Deleuze 1994, p. 222). For this reason, we think is more correct
to define this approach as a generative design, rather than parametric, in order to mark
the capacity of this modelling of discovering new forms.
In this sense, this methodology responds perfectly to the modern request, proposed
by the father of the descriptive geometry, Gaspard Monge, when he writes about its
objectives: “Le premier est de représenter avec exactitude, sur des dessins qui n’ont
que deux dimensions, les objets qui en ont trois, et qui sont susceptibles de définition
rigoureuse. … Le second objet de la géométrie descriptive est de déduire de la
description exacte des corps tout ce qui suit nécessairement de leurs formes et de
leurs positions respectives. Dans ce sens, c’est un moyen de rechercher la vérité;
elle offre des examples perpétuels du passage du connu à l’inconnu” (The first is
to represent with precision in drawings that have only two dimensions objects that
have three dimensions and that can be defined rigorously … The second objective
of the descriptive geometry is to deduce from the exact description of the bodies,
everything that is necessarily followed by their forms and their respective positions;
in this sense it is a means of seeking truth, as it offers perpetual examples of the
passage from what is known to what is unknown” (Monge 1789, p. 2) (Fig. 1).
The generative design, anyway is a two-dimensional drawing, defined by the
connections between simpler algorithms, it defines rigorously the form as any digital
path and it offers also the exact description of bodies, integrating also different
information. But the real question about generative essence is this “seeking truth”,
not as a path of passage from ignorance to wisdom, but from “what is known to what
is unknown”, a condition that operatively it is possible to test stressing the parameters
and their combinations.
In the construction of the system of relations and dependencies, the generative
design is a “representation of a representation”, a “square representation”, able to
amplify in an “exponential way” the passage form logic of model, linking geometry
and computer syntax (Filippucci 2012a). The deconstruction of elements, funda-
mental also for parameterization, together with the representation of connections
and structures it clarifies the mutual positions of algorithms. Digital revolution has
strengthened the requests of rigorous with its own syntactic alphabet, purely math-
ematical, able to procure synthetic elements and morphological patterns, a network
of nodes and connections that can be considered ideogrammatic morpheme of con-
temporary hypertext communication (Bianconi 2002, 2005). Generative modelling
becomes an instrument able to reinvigorate the union of the graphical representation
of space and digital space, a process that explains what a form is made of, and not
what kind of form it is. In representing a logical net, the parameterization of the ele-
ments is a consequence, an opportunity of digital descriptive text to regain possession
of its infinite potential and dynamic heuristic, analysis of a path, understanding sin-
gular elements of digital syntax. In this way, it is possible to have a “critical analysis
of digital representation’s performance”, as descriptive geometry discipline required
(Monge 1789).
Generative design, anyway, represents a spontaneous evolution of digital mod-
elling, aimed to make the most of its potential, valorising the capacity of computer
to simulate all the possible combination and to evaluate the solution for the best.
It represents a support and a response to the contemporary architectonical criteria
founded in the performances and defined by the organic system, which is oriented
to its formal definition (Hensel 2013; Oxman 2009; Turrin et al. 2011). Inside the
generative interpretation, here is in fact hidden an organic interpretation of the model
according to the “deep principles” of Nature (Petra Gruber and Jeronimidis 2012).
Natural Intelligence (NI), which is all the systems of control and not artifacts, but
rather is present in biology, is the font of inspiration for the architectonic research
(Goel et al. 2014; López et al. 2017; Vattam et al. 2007). Nature offers not forms
but processes to think about form (Oxman 2009) and it teaches how to create forms
(Barthel 1967; Bhushan 2009) in efficient structures (Knippers and Speck 2012;
Wester 2002) explaining how the roles of a design (Mattheck 1998; Mazzoleni 2013)
really is adaptive and optimized (Pawlyn 2011). NI defines the paradigms that could
be receipted not as “an ignorantly copy of the shapes” … but in the acknowledgement
“that biomimetics teaches that shape is the most important parameter of all” (Vincent
2009, p. 81), because at all levels it builds responsive and adaptive forms to conserve
material and energy resources through the use of modular components combined
with least-energy structural strategies (Pearce 1979).
The biological paradigm (Thompson 1917) becomes operative in digital repre-
sentation through the development of new technologies (Hensel et al. 2010). In this
centrality of biomimetic (Bar-Cohen 2016; Benyus 1997; Kuhlmann 2011; Lynn
1999; Myers 2012; Pedersen Zari 2015) recent developments focus on multi-scale
models and the interplay of mechanical phenomena at various hierarchical levels
(Knippers et al. 2016), as it is shown in some architectural experimentation (Bechert
et al. 2016; Krieg et al. 2015). Beginning with Ingo Rechenberg, the first applica-
tion of the Genetic Algorithms to the aerodynamic wing design (Rechenberg 1965),
the computer research began to develop algorithms inspired by natural evolution
(Fogel et al. 1966) in order to generate solutions to problems that were too difficult
to tackle with other analytical methods (Floreano and Mattiussi 2009, p. 1), founded
in the cycle between analysis, selection, coupling, and coalescence steps that are
repeated until reaching the solution (Rutten 2013). The logic of this approach is
to find “Solutions with better fitness values that are more likely to survive and to
join the next generation in a genetic algorithm” (Medaglia 2007); the application of
Darwinism’s two fundamental contributions, to “move in the direction of a science
of multiplicities: the substitution of populations for types, and the substitution of
rates or differential relations for degrees (Deleuze and Guattari 1987, p. 48). Evo-
lutionary Algorithms, applying the biological principles of mutation, selection, and
inheritance, reinterpret the parametric system that becomes the genome, the field of
alternatives is the same of the population, and the architect’s design goal becomes
the fitness criteria. It is possible to define the genetic algorithm as a series of math-
ematical operations that transform individual objects of a given population into a
subsequent new population, by selecting a certain percentage of objects according
to fitness criteria. Anyway, genetic algorithms are instruments to solve constraint
satisfaction problem, typically solved by using a form of search, pertinent to the
process of building (Fig. 2).
Genetic algorithms are the expression of the use of Artificial Intelligence in digital
representation; they are innovating this field by introducing a no-handed drawing,
with no direct representation of the form, no direct use of the fundamental algo-
rithms of representation, points and lines (Bianconi et al. 2017a; Filippucci 2015).
The development of evolutionary strategies reinforces the horizon of a generative
modelling (Renner and Ekárt 2003) by supporting the definition of unthinkable and
performative solutions (Menges 2009). In this simulation, it is possible to exploit the
computational capabilities of digital tools and to explore the possible combinations,
not only one parameter at a time, with the goal of improving a single aspect of a
building (Jones 2009a), but, more of them simultaneously.
The development of evolutionary strategies for construction issues (Kicinger et al.
2005) finds new life thanks to the support offered by AI: environment can be con-
sidered as Active Agent of the design (Hensel 2010) and in the range of solutions
and in all possible combinations AI chooses and draws the solution with the best
fitness. The proposed parametricism is not connected only to a variation of geo-
metrical parameters to change the final shape, but also it can be derived from an
input not directly connected to the form. In fact, it is possible to consider external
environment as a font of possible input data (Hensel and Menges 2008; Woodbury
2010), by integrating also those derivable from Big Data (Couldry and Powell 2014)
or from the internet of things (i.e. in domotic, a form that can derive from an external
signal—temperature—which drives a form—the opening of a louvre) to support the
Fig. 2 Digital design: didactic generative modelling of Toyo Ito’s snail shell-shaped timber struc-
ture inspired by NI
decision-making. The frontiers of the computational design see the organization of

the information and the management of big data, which can become the input data to
generate solutions. The volume of activity per minute on the Internet is something
like 347,222 tweets written, 51,000 apps downloaded from the Apple app store, 300 h
of videos posted on YouTube, 4,166,667 Facebook “Like”, 110,040 calls made by
Skype, 17,336,111 photos “liked” on Instagram… (Iafrate 2018, pp. 45–46). Arti-
ficial Intelligence will learn ever more quickly with the increasing amounts of data
to “consume”. Computer really aids design, the intelligent machine codesigns archi-
tecture, because it finds solution and draws form without a direct representation, and
also because in the acquisition of external (and big) data, AI also does self-learning,
remembering experiences of past actions (Fig. 3).
Fig. 3 Wood design: didactic modelling to materialize the space
As Darwin lesson about ex-aptation shows us, this process is not teleologic, so the
genetic algorithm shows us how it is possible to adapt a form to a changing context
(Woodbury 2010), by defining a new form of thinking design (Oxman and Oxman
2014). From another point of view, we can describe the same process by affirming
that an object described by parameters can be uploaded and it can evolve, so we can
say that AI redraws it (Vierlinger 2015), with results (Vierlinger 2013) that could
be valued also for their “artistic” value (Bittermann 2009; Pearson 2011) and their
role in space discovery, in a digital path anyway is open and flexible (Vierlinger and
Bollinger 2014).
Adaptation, responsivity, flexibility, evolution, are the keyword connected to NI,
that characterize the actual architectonic debate. Also, if NI opens to other hori-
zons and conceptualizations, in this context, it is possible to link this theme to the
relationship between design and construction, deepening, in particular, the wood as
a paradigmatic matter. A Morpho-Ecologies approach (Hensel and Menges 2006)
“commences from the unfolding of performative capacities inherent in material sys-
tems in relation to the specific environment they are embedded within, as well as an
intensively empirical mode based on physical and computational form-generation
and analysis methods” (Hensel and Menges 2008). This new approach is changing
the paradigms and the processes of architectonic research because structure, form,
and materials are not taken separated one from the other, but together in an organic
project where the digital process can simultaneously design and stress their charac-
teristics.
Artificial Intelligence appears now as the future of digital representation. The
smart culture needs to “do more with less”, it defines the need of a performative
result, and the intrinsic capacity of computation inside the digital process, with the
guaranty of achieving this goal in relation with the design (Grobman and Neuman
2013). Robots can adapt like animals (Cully et al. 2015) and through AI also drawing
is becoming every day more “complex, creative and surprising” (Lehman et al. 2018).
Someone is scared by this automation and the possibility of transhumanists, but
even if a drawing without hands is developing, the role of the designer is still the
center of modelling question; it is impossible to peerless our creativity and our
capacity of innovating, by disrupting our limits and ideas, the awareness, which
is the central question of the design process. At the basis, there is geometry and
its capacity of describing reality, by abstracting the phenomena in a mathematical
model. This because generative design changes the design thinking, by only marking
the abstraction process defined by a mathematical language and by an algorithmically
logic (Woodbury 2010).
In the abstraction of the virtual space, it is possible to forget the weight of materi-
ality, the concretism of the architectural question. In digital representation is possible
now to reproduce any form. In the real world, this is not useful. In the digital rep-
resentation, the time dimension can be unconsidered. In the real world, this is not
possible. It is really important not to fall in love with the model and the instrument.
Not totally at least, because in the real and concrete world there is what is concrete
and real, there is a natural intelligence.
In this context, wood represents a particular matter perfect for the application
of this approach. Wood is naturally connected to the smartness of biology (Ugolev
2014), the cause of its functional characteristic (Herzog et al. 2004), variable in qual-
ity. Priory renewable resource (Dangel 2016) with an aesthetic appeal, workability,
flexible, relatively light, versatility, low thermal conductivity, wood presents also
unwanted characteristic for its sizing limits and deformations, isotropism, hygro-
scopicity and degradation. Through its engineering aimed at homogenization of the
characteristics of this material every day more, wood represents a performing solu-
tion that integrates fabrication as a generative paradigm into the design process
(Gramazio et al. 2010, p. 111). For these conditions, wood represents one of the
most important field of application of parametric design (Chilton and Tang 2016;
Kaufmann and Nerdinger 2011; Menges et al. 2017a; Vierlinger 2015; Weinand
2016) where “non-standard timber structures can be efficiently aggregated from a
multitude of single timber members to foster highly versatile timber constructions”
(Willmann et al. 2017, p. 26). In hybridization and integration of digital wood design,
the innovative tools involve a transformation of paradigm and form, connections and
limits. Wood, CAD, and AI, all dynamic elements in evolution, those are transform-
ing our research, because as underline Achim Menges, “rethinking wood through a
computational perspective has only just begun” (Menges 2017, p. 108) (Fig. 4).
Fig. 4 Traditional parametric wood design: digital reconstruction of geometrical process for the
cutting of “trompe” ashlar in Philibert Delorme’s Le premiere Tome de l’Architecture” (1567)
2 Background
Defined these coordinates, it is possible to present how AI is changing our research

experiences, developed inside the Department of Civil and Environmental Engi-
neering at the University of Perugia and the International Research Laboratory on
Landscape. The themes of all the research are always linked to the field of study of
the science of representation, deepening the potentiality of digital instruments and
the construction of models.
This approach had deepened roots in history, in the research to apply geometry to
rationalize the real, and from the origin related to wood design. As implicitly reported
in the Laugier’s hut archetype (Laugier 1735), geometry represents a tecné to trace
signs in order to have constructive forms; a technical process that leads to developing
theoretical paths, which are defined by stereotomy traits. This science, particularly
asserted in France, was born from projective geometry’s principles (Evans 1995),
which evolve in the gothic (Sanabria 1989) until they reach a theoretic apex in XVIII
century with the definition of descriptive geometry by Gaspard Monge (Loria 1931).
Father of this technique could be Philibert de L’Orme, he who in his treaty call himself
as the creator of the method meant as a path “to unify the practice of geometrical
traits with Euclid’s theory”, with the aim of “reviewing Euclid” (De L’Orme 1568,
l. 3f.62). The passage from projective geometry to the Euclidean one is connected
then to Girard Desargues’ researches (Desarguer 1640), which are already connected
to stereotomic teaching (Saint Aubin 1994), this tries to connect prospective with
projective geometry (Field 1987, pp. 3–40). The value of the research, studied also
by Bosse (1643a), “it is not completely new and revolutionary in principles or in
mechanism: but in the drive to theory and in the abstraction from the particular
problem” (Trevisan 2011).
The wood stereotomy (Paris 2009) is connected to the descriptive geometry which
derived from Gaspard Monge’s revision, as testified it is by the nineteenth treaty of
Leroy (1857) or of Pillet (1887) or of the italian Peri (1884). Through the analysis
of the peculiarity of the material, the first researches were involved in supporting
carpenters in the cuts and in the realization of nodes and joints. Using the exact
instrument of the design, as rules and compass, in general the stereotomy is a proce-
dural strategy correlated to the constructive phases, following a space discretization
process, a simulation of the realization of the form in a geometry of the cut represen-
tation, the integration of the static analysis in the centrality of representation together
with the graphics analysis and the mechanic of rigid bodies. The growing complexity
imposes a structural strictness that leads to the transformation of stereotomy’s statute,
which moves from “art to science of the cut”, where intersection curves are treated
between section planes and geometric corps, and the projection of those curves on the
different planes (Fallacara 2007, p. 57): it is a historical evolution that finds a decod-
ing of already present notions, because Desargues (Bosse 1643b), La Hire (1596)
and Frézier (1737a, 1760) “didn’t know they knew descriptive geometry” (Trevisan
2011) but they had implicitly assimilated its prodromes (Salvatore 2012).
Maybe also the contemporary digital world design is not completely aware about
its knowledge on the rules of descriptive geometry, which is applied with the same aim
of setting in relation the design with the many performance exigencies of the construc-
tions, according to ante litteram mass customization logics (Andia and Spiegelhalter
2017; Paoletti 2018). Digital wood design is based on procedures that are so similar
to those used in the past, so much that there are many different reinterpretations
(Block et al. 2017; De Azambuja Varela and Sousa 2016; Fallacara 2006; Fernando
et al. 2015; Rippmann and Block 2011) surely influenced by the possibilities offered
by smart fabrication (Davis et al. 2012; Eastman et al. 2009; Li et al. 2008; Popov
et al. 2010).
In this cultural root, it is possible to find the reasons that describe the centrality
of design as a model and its transdisciplinary language. It is not a whole design, but
those are parametric procedures in an implicit way, able to respond to many different
solutions. Representative logic expressed, the same one that is the foundation of our
thought, it is revealed in its description capacity, to let know the project process.
In this relation, actually contemporary principles of the topological approach of the

generative design are present as prodromes, as the logic of similar elements with
“modifiers” that lead to the deformation of a “figure” (therefore to the creation of
the form) through the “repetition” of the creative act (Fallacara 2007).
This approach of representation is connected to the creation of models for the
interpretation of reality, a path applied “from the spun to the city”: drawing is always
the instrument to understand the complexity of the reality, to analyse and project it
and on these coordinates it is insert the writing action, which in the last years it lead
to analyse multiple and heterogeneous fields (Bianconi 2005; Bianconi et al. 2006,
2016a, b, 2017a, b, c, d, e, 2018a, b, c, d; Bianconi and Filippucci 2015, 2016a, b;
Filippucci 2010, b; Filippucci et al. 2016a, b, 2018, 2017). In the vastness of the
application it is possible to find out the centrality of representative research as an
instrument to know and innovate, able to continue and to express its operative aspect
and at the same time full of culture, the sense of its tecné, etimologically meant as
art and technique.
3 Methods and Materials
In our research we are testing different ways to apply the contemporary principle
of descriptive geometry to digital wood design research, almost contaminated in a
multidisciplinary learning environment. Moreover, the research reported shows also
the connection with local small enterprises and Administrations, a knowledge transfer
that is the strategy to survey the research. It is important also to mark how in Italy
the architectonic culture of wood building is really reborn after the predominance of
industrialized construction, even if in the last years these solutions have been finding
a vaster market.
In all these cases, generative design has a central role. Moreover, in this process,
here are always integrated different tools and software for multidisciplinary analysis.
This integration is address to the need of optimization of architectural form, using
Genetic algorithms to analyse and to understand the relation between form, geometry
and construction.
All the researches start redrawing themes. The transcription through the new
medium of generative design changes the messages, verifying the condition and
using more performing tools to explore the boundary condition.
In this chapter we describe our works proposing a path of five integrated step
in the valorisation of NI in wood design and the opposite engineering of the mate-
rial, the results of generative transcription, the horizon of form-finding and multi-
optimization in the use of AI, the realization of a pavilion.
4 Experimentations
4.1 Responsive Architecture: Study, Characterization

and Realization of “Unplywood” Panels for Passive
Ventilation Systems
This research is founded on the idea of indulging the natural characterises of the
wood, in a path that starts from phenomenal observation and it arrives to define a
digital model (Fig. 5). The research aim tries to answer to the needs of a reduced
energy consumption building, that has recently led to the development of expen-
sive and technological smart materials (Addington and Schodek 2005; Loonen et al.
2013). The exploitation of certain properties of natural materials can reduce the eco-
nomical and the environmental impact, which define the new technologies. Inside
the different characteristics, the study analyses the hygroscopic behaviour of wood,
always considered as a negative property of the material for its end-use, with the
aim of transforming this in a positive feature in the realization of a double-layered
composites which passively react to the variations in relative humidity (Holstov et al.
2015; Reichert et al. 2015). The bending reaction of “unplywood” is investigated and
applied to a passive ventilation system in indoor environments, where relative humid-
ity rapidly changes, and is used to parametrize the architectonical shapes (Figs. 6
and 7).
The research finds its inspiration in the previous works developed by the Institute
for Computational Design at the University of Stuttgart (Krieg et al. 2015; Menges
and Reichert 2012; Reichert et al. 2015; Wood et al. 2016, 2018), by Newcastle
University (Holstov et al. 2015, 2017) and by the ETH of Zurich (Rüggeberg and
Burgert 2015). The research is developed by a group from Department of Civil and
Environmental Engineering in University of Perugia, which includes the authors
and Giulia Pelliccia in the architectonic and representative theme, together with the
group of Giorgio Baldinelli and Antonella Rotili from CIRIAF center of the same
university to support the energy theme, and the group with Marco Fioravanti, Marco
Togni and Giacomo Goli from GESAAF center in Florence University to support the
wood research from. The research, started with a master thesis, is developing in a
knowledge transfer to the SMe ABITARE+ which is involved in timber construction
process.
The integration of different skills is linked to the increasing of building energy
efficiency, one of the most important goals to achieve both in existing and new build-
ings, thus, passive design strategies are becoming a strategic part of the contempo-
rary architecture projects. Wood is a perfect example of a natural smart material: its
hygroscopic behaviour makes it responsive to relative humidity variations and the
shape-memory effect (Ugolev 2014), it allows its return to the initial configuration
Fig. 5 “Unplywood” panels: empirical researches about the hygroscopic behaviour in displacement
of different panels in function of the time
that it had before humidity changes. The absorption of water results in the swelling
of wood; if, instead, the relative humidity decreases, wood will loss water and shrink
(Fig. 8).
The wood subjected to a variation in moisture content will change its dimensions
causing, as a result, aesthetic as well as functional problems. To take advantage of
natural hygroscopic properties of wood could be a better solution rather than prevent
them with expensive equipment, when the aim is that of designing responsive panels
to be used in building air conditioning.
Based on the example of the structure of the pine cones (Burgert and Fratzl
2009; Dawson et al. 1997; Reyssat and Mahadevan 2009; Song et al. 2015), whose
double-layered scales bend in reaction to humidity variations, an entirely wood-made
panel, called “unplywood”, is realized in this research. The two layers of the panel,
active and passive, must be cross-grained and the tangential direction of the active
Fig. 6 Modelling as result of observation of phenomena: curvature in function of the humidity in

climatic chamber tests
layer corresponds to the longitudinal one of the passive layers, which is almost non-
reactive to humidity (Rüggeberg and Burgert 2015; Vailati et al. 2018) in order to
have a bending of the “unplywood”.
The research stars from the study through experimental tests of some materi-
als and configurations, to evaluate the potential applicability of wood bilayers to
the indoor thermo-hygrometric comfort control in conference or meeting rooms,
or indoor vanes, where relative humidity rapidly increases and a humidity control
system is not present. Wood is used for the active layer beech (Fagus sylvatica L.),
while for the passive layer is used spruce (Picea abies Karst.) in some specimens
and European larch (Larix decidua Mill.) in others. A one component polyurethane
adhesive is used to glue the layers. Both active and passive layers are quarter cut
veneers oriented according to the wood anatomical directions. The specimens are
10 cm long, 10 cm wide and with a variable thickness, some of them have the grain
direction parallel to the edge (called “90 configuration”) and some others are parallel
Fig. 7 Wood NI for building energy efficiency: the exploitation of hygroscopic wood transforma-
tions for a natural indoor thermo-hygrometric comfort control
to the diagonal of the specimen (called “45 configuration”).

The experimental tests and the statistical analysis results lead to the construction
of the prototype of a modular panel for a false ceiling, simulated in a meeting room
and in a conference room. The humid air inside the room makes the false ceiling
open, as a reaction to the relative humidity difference between its faces. Then, air
flows in an interspace of 10 cm directly connected with the outside. The air is then
introduced in a space, wide not more than 50 cm, between the external wall and a
glass wall; for the greenhouse effect, the air inside that space is heated and it moves
upwards, until it reaches the opening to the outside at the upper floor level. The
application of this ventilation system requires a building of at least one floor above,
which must be dehumidified, in order to activate the stack effect. The external wall
where the glass is fixed can be totally opaque or transparent, or equipped with some
openable windows inwards (in this case, the indoor relative humidity control will be
lost).
The parametrization of the architectonical shapes allows to reproduce any plane
tessellation to be applicable on every dimension surface: whole surfaces can be cov-
ered with different types of tiles. The “unplywood” panels can be used to realize a
modulus for a false ceiling for the dehumidification of indoor environments. Through
Fig. 8 Digital panel design: parametric variation of the modular configurations for plane tessella-
tion
Grasshopper for Rhinoceros, these three typologies of plane tessellation are repro-
duced, realizing every tile with “unplywood” panels, whose movement, depending on
the maximum deflections reached in reaction to humidity variations, is parametrized
in some algorithms base.
Actually, this research represents just a first concretization of the potentialities
that “unplywood” shows in the bioclimatic architecture, with the possibility to be
applied to every indoor environment, thanks to the adaptability of their shapes and
dimensions (Fig. 9).
Fig. 9 Digital simulation: open and close configuration of the panelling in an hypothetic ceiling
of a meeting room
4.2 Innovations and Experimentations for Double-Curved

Timber Surfaces: Design and Characterization
of a Flexible and Engineered Solid Wood Panel
In contemporary architecture, the language used to draw forms is connected to the

possibility of expressing freely any shape. Geometrically, this condition is guaranteed
by the double-curvature surfaces and for this is fundamental the material, which
is transformed by innovative technologies to overcome the formal constraints for
the realization or by a discretization of the surface curvature through a tessellation
process. The topic of the research is the study of the double-curved timber surfaces,
in the limitation of the wood characteristic for its organic, heterogeneity, anisotropy
and personality. The real challenge is to make wood flexible in double-curvature and
to obtain a spatial bending, overcoming the limits imposed by nature of the material
itself (Fig. 10).
The research is developed by a group from Department of Civil and Environmental
Engineering in the University of Perugia that includes the authors and Simone Moroni
and Daniela Ripa for the architectonic and representative theme, and the support for
wood research from the group with Manuela Romagnoli from DIBAF of University
of Tuscia.
The study begins with the impossibility of creating a double curvature surface
using wooden planking, due to the not developability of this class of surface. The
path starts from the maps of the world and the necessity to cut double curvature
surface to develop it in a plane and little approximation, a technique applied in wood
design through an oriented cut (Grima et al. 2016).
Fig. 10 Flexible solid wooden planking: the discretization through the cut for the double curvature
of the panels
Also, if this kerfing technique is noted in the scientific literature (Menges 2011,
pp. 77–78; Zarrinmehr et al. 2017), the central question is the definition of engineered
solid wood panel as able, flexible and resistant to deform itself in a double-curvature
shape. For this reason, in a generalized vision of panelling (Eigensatz et al. 2010;
Konaković et al. 2016; Ohshima et al. 2013; Son et al. 2017), the research can be
considered as a modelling reverse of the design of a pattern of geodesics on a freeform
surface (Pottmann et al. 2010) to obtain a flexible developable surface (Solomon et al.
2012).
The study analyses the factors influencing the laser cutting of the wood (Barnekov
et al. 1986) and the proprieties of different wood species; it has developed more than
400 mechanical samples tests by using different wood species (beech, poplar, spruce,
lime), with the aim of characterizing their maximum bendability, their bending resis-
tance, and their load resistance. Furthermore, the trend of the stresses internal to the
samples is defined by the determination of the critical area of breakage. This provides
the ideal cutting frame dimensions which allow to reduce the internal stresses in the
weaker zones and to maintain the same bendability capacity, in order to achieve
a tenfold increase in curvature for carved panels with a quadrupled cross-section
(Figs. 11 and 12).
Through the parametric algorithm, it is possible to simulate the form and the
behaviours of these elements, and to modify, according to the performance requested
for the panel to have. The panel thickness, the cutting depth, and width, as well as
frame dimensions can be changed. For the analysis of the double-curved wooden
shapes, the surface cutting is first studied, then some wood folding techniques have
been analysed, through cuts, in order to make the material flexible. This condition
Fig. 11 Kerfing and resin: the construction of a developable panel and the transformation in a
structural element
represents just a first phase of the study, that test also the mechanical characteristics
of the panel adding different resins, considering the cut like a fist phase, to obtain
free form, that through this filling material arrives, in a second phase, to guarantee a
structural resistance (Fig. 13).
Empirical research is translated in the digital environment to define, in a different
way, the characteristic of the wood redrawn by the digital design. The experimentation
on the flexible panel allows to consider the application in design and to bring the
deformation capacity of the material to the limit.
Fig. 12 Mechanical tests: the definition of the cutting frame in 400 specimens in function of the
different wood species for the characterizing of maximum bendability, bending resistance, and load
resistance
In parallel to the empirical tests, the panels have been digitally reproduced and
simulated in digital environment to analyse them under a set of applied loads through
the software SAP. The behaviour of the panel has been simulated entering the Young’s
modulus values obtained from the real bending tests, in order to analyse the internal
stresses distribution both in the longitudinal and in the transversal direction to the
fibers and to more carefully evaluate the most solicited zones. The research shows as
the possible application of the flexible panel as structural element and for this reason
are tested different resins, to arrive at breaking strength of 1100 N.
Fig. 13 Digital analysis and design: internal stresses distribution in the structural simulation
4.3 Redrawing Descriptive and Projective Geometry:

Generative Design of Stereotomy and Japanese Joints
for CAD/CAM Application
Stereotomy means to cut solids, from Greek “stereós” and “témno”, geometrization
and generalization of architectural elements, and vice versa the application of pure
geometry to the architecture of “stereotomy policy”. This approach results extremely
congenial to the parametric tools today applied to many architectural projects, accord-
ing to a rule based on a design that today is evermore stronger in generative design.
The research, developed by the authors together with Filippo Bruno Palazzari,
starts from the transcription in generative environmental of the drawing published in
the eighteenth century Frézier publication (Frézier 1737b), where it is given special
importance to the quadric surfaces, by studying firstly the resulting curves from
reciprocal spatial relation, the intersection of two quadric surfaces (planar curves
or spatial curves). Rewriting these classes of surface in parametric language (Burry
2014) and linking them with the process of smart fabrication (Bidgoli and Cardoso-
Llach 2015; Duro-Royo and Oxman 2015; Rippmann and Block 2011; Svilans et al.
Fig. 14 Parametrizing gemetrical forms: three-dimensional development and parallel projections

of hump spirals
2018), it is possible to have parametrized solutions, but also “unknown” forms, also
interconnecting, contemning and transforming the classical path (Fig. 14).
It is possible to show this principle in the ideation of a staircase defined by stere-
ometric rules, reinterpreting the Vis de Saint Gilles (Pérouse de Montclos 1985;
Sanjurjo Alvarez 2010; Tamborero 2006). Beginning with the generative design of
the helicoids, the algorithm, which is created in Grasshopper for Rhinoceros, defines
a unique step in the application of classical rules, with variable solutions for its para-
metric nature in the possibilities of acting on decided upstream parameters, as the
diameter of the stairwell, height difference to be overcome, number of windings of
the helix and others. As it was in the past, the questions involved regard serializa-
tion, minimization of work waste, dry installation. Moreover, the characteristic of
the material is another boundary condition, as it was also for the stone cutting in
the past. To construct the stairwell by using wood, one of the most important aims
is the management of acute angles and maximize contact surfaces. To realize this
Fig. 15 Digital fabrication as contemporary carpentery: parametric transcription of the art of tra-
ditional Japanese wood Joinery
objective, it is necessary to design the form by using solids, which are characterized
by perpendicular surfaces, defining for every riser the normal surfaces to the planar
helicoid. In the experimentation developed, the helical staircases are digitally defined
by five principal surfaces, one planar helicoid, and four spatial helicoids, these define
the intrados and extrados of the staircase and its lateral surfaces. In the end, to design
the treads of the stairs, with Boolean operations, the extrados helicoid has to be cut
by parallel plans to Cartesian plane “xy” (Figs. 15 and 16).
The creation of optimized contact surfaces between the voussoirs of a vault
(whether it is a staircase or a “classic” vault), so important for wood stereotomy,
is realized by joints between the elements of a structure. To design the joint between
the steps, the open path guaranteed by the generative design is contaminated with
a really different culture of the Japanese carpentry (Brown 2014; Kuroishi 2015;
Fig. 16 Digital commingling in generative stereotomy: parametric design for the generalization of
helicoidal stairs and the standardization in the serial fabrication through the definition of an unique
step incastred by “Sumi Isuka Tsugi” oblique scarf join
Nakahara et al. 1995; Seike 1986; Zwerger 1997). Some joints are rewritten by gen-
erative design, in order to transform parametrically any solid. Inside the different
joints, it is proposed the application of Sumi Isuka Tsugi, which can redraw the serial
element. In digital design, by applying a morphing process, it is possible to create
stair characterized by the same form for all the steps (excluding the first and the last).
The digital model developed is useful for the structural analysis (applied with
Abaqus software) in order to define deformations and internal tensions, and also to
simulate the different behaviours, by using integrative structural support. The same
model is also useful for 3dprinting, tested by using the additive material, but also
useful for laser cutting. By using CAD-CAM technologies, it is then possible to skip
the last step, in fact, the designer himself can easily obtain the “Panneaux 2.0” or
rather he can obtain numeric codes to be supplied to CNC machines, that can easily
construct all of what the designer wants.
In this contamination of disciplines ad solutions, the purpose was reached by
letting the possibility of creating one reference step, provided of carpentry joints, into
Fig. 17 Generative path for the optimization of the results: computational design for the individu-
ation of the best combinations to minimze deformations and internal tensions
a simple parallelepiped volume, then with (simplistic) the use of topological criteria
(namely, geometric transformation without cutting, overlays or gluing) the reference
geometry has been morphing into the five basic surfaces of the helical staircase.
The research represents one of the new horizons opened by digital representation
and parametric tools in the renew of descriptive geometry. Rewriting through these
instruments, based on parametrization, the logic of a process, aimed to support the
seriality of applications, to guarantee a correspondence of intents from past scopes
to the present ones.
The integration of instruments and paths becomes an intercultural instrument
able to connect different cultures and different historical results in a contemporary
language, which useful to answer to the request of optimization. The historical lesson
shows its value and role for the construction of our future (Figs. 17 and 18).
Fig. 18 Generative trascritions: geometrical trasfromation of polyedra for the definition of tenseg-
rity structures
4.4 Wooden Tensegrity Structures and Morphological

Evolutions of Transpolyhedra
The construction of innovative structural solutions arises also from the graphic evo-
lution of simple forms. Digital representation, in the transcription of forms, helps
to understand the centrality of the medium to understand the form. It is the case of
tensegral structures, a study developed by the authors together with Margherita Stra-
maccia and Matteo Margutti (Bianconi et al. 2017g, 2018), where the geometrical
solution is originated by the movement of poles on polyhedric geometries, devel-
oped by generative paths. Without graphic support, it’s hard to understand what a
tensegrity is; it is definable as “a system in a state of stable self–balance including a
series of compressed components inside a continuum of tensed components” (Motro
2003, p. 19). Analysing this statement, it is possible to identify the “compressed
components” as the poles present inside the system, while the “tensed components”
are made up of a net of tie–rods that connect each element and create this continuum
(Lonardo 2011, p. 5).
The term “Tensegrity” was created in the early Fifties by Richard Buckminster
Fuller, as a contraction of the two words “Tensional” and “Integrity”. The main orig-
inal studies of these structures (Lalvani 1996) was firstly developed through different
approaches by Ioganson, the protagonist of the artistic research in Russian Construc-
tivism (Gough 1998), by Emmerich, who was interested in plane and spatial tessel-
lations (Emmerich 1988, 1996), by Snelson, who started from the polyhedral artistic
reinterpretation (Snelson 1996, 2012), and Fuller, the most famous (Fuller 1963;
Marks and Buckminster Fuller 1973; McHale 1964), who realized some buildings
using this class of structures (Buckminster Fuller 1961; Calladine 1978; Krausse and
Lichtenstein 2017; Marks and Buckminster Fuller 1973; Sadao 1996) considering
the structure as physical manifestation of his theories on synergetics (Buckminster
Fuller 1975). Tensegrity is linked to polyhedron (Pugh 1976a), and it can be inserted
in the theory of reciprocal structure (Baverel and Larsen 2011; Olivier Baverel and
Pugnale 2014; Gherardini and Leali 2017; Kohlhammer and Kotnik 2011; Popovic
Larsen 2003, 2008; Thönnissen 2014; Thönnissen and Werenfels 2011), which is
just known in history as Leonardo projects show (Di Carlo 2008; Pedretti 1988) but
their codification is recent, as testified by the sort of war in patent between the pro-
tagonists of last century (Buckminster Fuller, US3203144A 1961; Emmerich Patent
1377290, 1964; Snelson Patent US3169611A, 1960).
Tensegrity structures are analysed by several researches (Burkhardt 2008; René
Motro 2003; René Motro and Raducanu 2003; Pugh 1976b; Skelton et al. 2002),
those have shown how these structures present interesting peculiarities: for example
the extreme lightness accompanied by a surprising structural rigidity, a condition
that makes them ideal for the challenge of the research of the minimum mass design
(maximum resistance with minimum mass), therefore ideal for the achievement of
a more aware eco-friendliness, which is not just focused on mere energetic saving.
Other peculiarities that characterize these structures are the capacity of redistribution
of the loads on almost the totality of the structure, or once again, the recovery of its
own balance independently from the direction of the disturbances that solicit it, all
properties that are not possible to find in the actual structural configuration, which is
characterized by a superimposition of compression. Every stable three-dimensional
structure, composed by p-planks, with a polygon of p sides connected through cables
on the top and a polygon of p-sides, when connected through elements solicited by
traction as a base, it fits the definition of tensegrity prism (Skelton and de Oliveira
2009, p. 19). For these reasons, tensegrity is subject of theoretical research (Cefalo
and Mirats Tur 2010; Chandana et al. 2005; Connelly and Back 1998; Connelly and
Whiteley 1992; Fagerström 2009; Hanaor 1992; Masic et al. 2005, 2006; Mirats Tur
and Juan 2009; Murakami 2001; Oppenheim and Williams 1997; Tibert 2008; Tran
2002; Vassart and Motro 1999; Zhang et al. 2006) but it is also applied in different
field and applications (Abdelmohsen et al. 2016; Adriaenssens and Barnes 2001;
Bouzanjani et al. 2013; Bruce et al. 2014; Chandana et al. 2005; d’Estrée Sterk
2003; Dogra et al. 2015; Lenyra et al. 2006; Liapi 2004; Nestorovic 1987; Paronesso
and Passera 2004; Peña et al. 2010; Riether and Wit 2016; Schlaich 2004; Stephan
and Klimke 2004; Tachi 2012; Van Telgen et al. 2013; Tibert and Pellegrino 2011;
Wang 2004).
Our construction of this systems is based on the same geometric process used
for every type of polyhedron, whether it’s regular, semi-regular or Catalan. The gen-
erative path is structured by a set of rotations and translations, to which it follows
the connection between cables and poles, in order to provide the proper redistribu-
tion of the tensions. The morphological evolution of the polyhedra passes through the
intermediate polyhedric configurations, called “transpolyhedra”. The tensegral struc-
ture, according to the geometric process previously described, is totally inscribable
inside these solids and therefore, between them there are various geometric relations:
decomposing the polyhedron in regular figures is therefore possible to recognize the
number of poles of the tensegral structure, the number of cables, the geometries to
the knot and to have a map of the connections between the same. The creation of the
algorithm for the tensegrity structures can become from both the polyhedra and the
plane tessellations. In particular, the algorithm for the polyhedra realizes translations
and rotations of the corners of the examined polyhedron. It’s easy to grasp that the
number of corners and faces modifies the dimension of the algorithm generated. The
algorithm, due to how it was developed, allows us to vary in many cases the tenseg-
rity structure from both the dimensional point of view and the configuration one;
by doing so it will be possible to adapt the polyhedron with the respective tensegral
transpolyhedra to every planning necessity, whether it’s about a small object or about
architecture (Fig. 19).
According to the geometrical correlation between tensegral structures and trans-
polyhedra, it is possible to consider the generative design of the polyhedron as a
vehicle of representation of spatial definition. The same representation, in addition to
supporting the idealization process, becomes substantial for the analysis of the form.
It’s possible to know the number of poles, the number of cables, the dimensions of
the same and their spatial disposition, identifiable (though only partially), through
the inner corner in the plane figures of the relative transpolyhedra. This abstracted
research, founded in geometry and representation, needs to a have another important
step in order to be concretized, in the definition of a digital handcraft process. For
the characteristics of the material, wood represents in the history a particular field of
application for the design, and now it is possible to underline a new development of
digital woodcraft (Tamke and Thomsen 2009).
In this connection between geometry, representation, structures and form, it is
possible to develop a series of practical applications, also connected to digital wood
design. In particular, in this case, we tested the value of this structure in design object.
A little table, lamps and more objects in wood were created through the support of this
tools. Differently, from the usual condition of tensioning the cables, through wood
design, we created sleeves that allow the object to be tensioned, and then to be realized
through CNC technique. Also, in this case, the transcription of forms by generative
Fig. 19 Heuristic of the digital drawing: medium and message for the ideation and the design
modelling has created a series of solutions and applications, exploding the operative
concept of artisan 4.0 as a result of comprehension of a geometrical path: reconnect
with materiality and craftsmanship, at the same time, through standardization and
simulation, they take away the properties of unpredictability and emergence that
were inherent in processes that were harnessing materiality (Fig. 20).
Fig. 20 Design of wood furniture: from the idea to the realization of sleeves elements to obtain the
necessary tension of the tensegral structure
4.5 AI for Mass-Customized Housing. Multi-objective

Optimization as a Decision Support System
In the ideation process inside the architectural design, the performances required are
increasingly becoming a structural requirement of contemporary culture. Likewise,
the freedom in the formal expression, the flexibility requests and the multiple needs
Fig. 21 Variation and evolution: the genetic population for the analysis of different combinations
connected to the relation with the context are posed as favourable conditions for
the development of a generative approach as a process for a responsive design. The
research developed by the authors together with Alessandro Buffi, it aims to define a
series of mass-customized housing models, which can be adapted to different context
and conditions, providing different solutions as an interaction with the surrounding
environment. Through the use of optimization tools developed in the field of AI,
namely Genetic Algorithms (GA), the study deepens the possibility of using genera-
tive models and evolutionary principles in order to inform the customization process
and to promote efficient use of energy and materials in the early stage of design, with
the aim to building up a web-based user interface that allows customers to explore
the generative potential of a model (Fig. 21).
Result of a research agreement with the Italian start-up Abitare+ involved in timber
construction, the research proposes an integrated design and a production process for
CLT constructions, which is based on mass-customization experiences (Benros and
Duarte 2009; Bergin and Steinfeld 2012; Duarte 2005), multi-objective optimization
strategies (Aish and Woodbury 2005; Kolarevic and Malkawi 2005), and data-driven
design (Brown and Mueller 2017). In this process, natural inspired form-finding
strategies (Bergmann and Hildebrand 2015) are used to select optimal solutions from
a structural and environmental point of view, allowing the designer to visualize and
evaluate thousands of design options and variations (Self and Vercruysse 2017) of
the same product (Bianconi et al. 2017f) through the materialization of architecture’s
complexity question (Scheurer 2010).
The generative path is connected to the emerging context of Industry 4.0 (Paoletti
2018), in its contamination between design and digital fabrication (Corser 2010;
Richard 2005; Larry Sass 2006), aimed to promote timber houses (Turan 2009) by
redefining in a parametric way the traditional typology of the Italian single-family
house. The application of wood in architecture is a technique forgotten in the Italian
contemporary landscape, also because of the reconstructions needs due to the recent
earthquake, evolving from the concept of prefab to the theme of mass customization
(Anderson 2002; Dellaert and Stremersch 2005; Duray et al. 2000; Knaack et al.
Fig. 22 Housing and mass costumization: the generative path for the optimization of results
2012; Nahmens and Bindroo 2011; Page and Norman 2014; Pine and Slessor 1999;
Salvador et al. 2009; Willis and Woodward 2010; Zipkin 2001).
This model generates a wide variety of architectural solution, and each of them
differs from the others mainly for their orientation, size, type of ceiling, roof slope
and shape of the glazing elements. In this case, the definition of rule-based design
emerges from a study of local codes and CLT construction systems; indeed, while
defining the geometrical rules of the model, the researcher encoded constraints in such
a manner that each solution meets codes dimensioning and affordable fabrications
methods. As a result, each house in the series is unique in shape and size, even if it
shares with the others the same building system characterized by CLT panels and a
fixed number of manufacturing operations, in both the factory and the building side
(Fig. 22).
The integrated process proposed in this research is created entirely with the
Rhinoceros’s plug-in Grasshopper, introducing in each phase of the project different
add-ons for analysis, representation, and interoperability. The Grasshopper’s add-ons
Honeybee and Ladybug (Roudsari and Pak 2013) were used to inform the process
Fig. 23 Big data analysis for Data Driven Design: Pareto front (top right) and comparison between
energy consumption and heating loads (bottom left), cooling loads (bottom center) and lighting
loads (bottom right)
with climate data and advanced energy analysis, Octopus to perform multi-objective
optimization, TT Toolbox to record the optimization process, and Geometry Gym to
create IFC models for BIM interoperability. These tools along with the open-source
analysis software EnergyPlus, Open Studio, Radiance, and Daysim constitute the
toolset used to create the material contained in the web-based catalog. In a second
moment, web implementation and data visualization were realized with java scripting
starting from the open source project Design Explorer.
In this process, while some architectural characteristics are defined directly by
the researcher, other qualities are supposed to emerge from an optimization process
which encompasses environmental data and analysis. In this sense, Genetic Algo-
rithms (GA) are used as virtual prototyping tools to navigate the design space of
a generative model and optimize the architectural organism. In this research, the
optimization process is based on Grasshopper’s add-on Octopus (Bader and Zitzler
2011; Vierlinger and Zimmel 2015; Zitzler et al. 2001) that performs multi-objective
optimization and allows to combine more performances criteria, namely fitness, to
obtain multifunctional structures. The definition of the performance criteria started
with the study of a construction model through the definition of detailed solutions and
sizing of structural elements. In this phase, the construction cost was computed with
the company through the definition of a series of parametric costs for the elements
constituting the structural system and the envelope, while, energy performances were
evaluated through advanced energy analysis by estimating building energy consump-
tion, comfort, and daylighting (Fig. 23).
Building energy performance assessments are complex multi-criteria problems

(Asl et al. 2014) that can be effectively solved by genetic algorithms (Clune and
Lipson 2011; Jones 2009b). The goal of an environmental optimization is to ensure
a satisfactory comfort with the minimum use of energy, through the adaptation of
the architectural organism to its context and its inhabitants. Natural lighting then
becomes one of the major driving forces in this design process, which aims to rein-
force circadian rhythms and to reduce the use of electric lighting by introducing
daylight into space, and it results an effective reduction of energy consumption and
comfortable spaces. Acting with passive strategies since the initial stage of design,
it is usually the most effective solution characterized by the highest rate of return on
investment. Data-driven design can effectively support the designer in this process of
problem-solving, by comparing the performance of different design solutions. In par-
ticular, in this research, the goal is to inform the design process in order to optimize
solar gain and maximize the effectiveness of natural lighting and ventilation.
In this context, the combination of data visualization became an effective way
to enhance the decision-making process (Tsigkari et al. 2013) while design space
catalogs, which present a collection of different options for selection by a human
designer, have become a commonplace in architecture in the perspective of the design
democratisation (Kolarevic 2015). The aim is to create an open source design (Raja-
nen and Iivari 2015; Ratti and Claudel 2015; Weber 2005) as meta-project for adapt-
able and mass customized housing (Lawrence 2003). The user interface developed
in this research (https://www.algoritmi.abitarepiu.com) is based on Design Explorer,
an open source project realized by CORE Studio Thornton Tomasetti, that allows to
intuitively visualize and effectively navigate the design space of parametric models
developed in Grasshopper, Dynamo, and Catia. These tools can support the designer
in the complex problem-solving processes, through the combination of the designer’s
preferences with the great amount of information owned by modern construction
companies, thus filling the gap between technological advances and design prac-
tice. Furthermore, their usability and effectiveness will grow along with advances
in Building Information Modeling (BIM), performance simulations and parametric
design and hopefully, in the next future, a similar data-driven approach will help the
designer to deal with increasingly complex projects and achieve both performance
and aesthetic expression.
The approach developed in this research is a work in progress that is finding more
declination, as it shows, for example, the green pavilion project for Perugia Munic-
ipality (Bianconi et al. 2018f), a customized and responsive architecture (Hofman
et al. 2006; Huang 2008), unique design able to generate variable social pavilions
for more than 100 parks presented in the territory (Fig. 24).
Fig. 24 From AI to the digital fabrication: details of the BIM model for the timber structures cutting
4.6 The Ames Room Pavilion. The Reinterpretation

of the Classical by Generative Design
As a collateral result of the previous research path, our group has designed the Ames
room pavilion, a paradigmatic architectural example that shows the flexibility of
generative. The project, coordinated by the authors with the support of all the group
of research (Alessandro Buffi, Giulia Pelliccia, Marco Seccaroni, Elisa Bettollini,
Michela Meschini, Maria Pia Calabrò, Gianmaria Angelini, Lorenzo Ciculi), is a
spatial deformation that creates an illusory image, constructed starting from the
equivalence of perspective projections between homologous spaces, application of
“perception seen as transactions between the observer and the world… The physical
world is left as mysterious ‘forms’, serving as catalysts to evoke perceptions from
rich ‘assumptive’ common sense derived from past experience, which is the seeming
reality of our perceptions” (Gregory 1987, p. 277) (Fig. 25).
The study is attributed to the American psychologist and ophthalmologist of the
last century Adelbert Ames Jr., that describes these forms, variable in function of
monocular and binocular view (Ames and Ittelson 1952), as “… a whole “family”
of distorted rooms of vastly different shapes could be built, yet all of them, if our
reasoning is correct, should be seen as “normal” instead of some other way because
our past experience has made it a “better bet” to perceive level floors, upright walls,
rectangular windows, etc.” (Ittelson and Kilpatrick 1952, p. 48). As Ernst Gombrich
explains to us, “the illusion consists … in the conviction that there is only one way
Fig. 25 Ames room: the generation of a whole “family” of distorted shapes in function of the
correspondence of the image constraining the point of view
of interpreting the visual pattern in front of us. We are blind to other possible con-
figurations because we literally ‘cannot imagine’ these unlikely objects” (Gombrich
1960, pp. 210–211).
The morphological transformation starts from the visual pyramid that is created
in the vision of a canonically stereometric room: constraining the point of view and
starting from the same image, it is possible to find a family of deformed spaces by
varying the base of the pyramid. This creates a solid perspective defined by divergent
and inclined the horizontal and vertical surfaces, that deceives the observer in the
evaluation of spatiality, not included in its real form but as a function of the a priori
critical thought with which the image is correlated to space. The deformed space, that
appears regular, confuses the references and the evaluations on measures, making it
appear deformed in the judgment of the dimension of what is inside it, especially
in the extreme parts of the structure corresponding to the edges where objects and
people are positioned in the two extremes of the room they seem to be both bigger
and smaller at the same time. Besides, expanding the functioning of the process of
interpretation of the form related to vision and images, this expedient is also used in
cinema, for example, in Harry Potter or the Lord of the Rings, to contextually shoot
characters of really different height (Fig. 26).
Ames’ room is a classic experiment of visual perception (Ramachandran 1990),
developed in this context with the aim to analyse perception questions. In fact, as
demonstrated by Cornish, Ames’ room works well because “we instinctively regard
an object as extended in the plane at right angles to the line joining the object
to the eye” (Cornish 1935, p. 61). In this way, Ames’ room shows the invariance
hypothesis: “if two objects make images of the same size, the more distant object
must be larger, this is known as Size-distance invariance, therefore loses its status as
a powerful explanatory concept and becomes rather a description of results obtained
under conditions which have yet completely to be specified” (Kilpatrick and Ittelson
1953, p. 223). This experiment was so useful to analyse the effects of perspective
alterations on apparent size and distance scales (Vogel and Teghtsoonian 1972), on
distance and size judgments (Blessing et al. 1967; Epstein et al. 1961; Holway and
Boring 1941; McDonald and O’Hara 1964; Teghtsoonian 1965) and the importance
Fig. 26 The illusion in the Ames room: the great professor Franco Purini with the professor Roberto
de Rubertis
of object recognition in size constancy (Bolles and Bailey 1956, p. 222), all themes
that show how ambiguity in itself can never be perceived (Gombrich 1960) but that
illusions are more effective when they can count on certain inveterate expectations
and assumptions on the part of the observer (Gombrich 1960). In this way, the
camera becomes the topic of confrontation between the interpretation of perception
as unconscious inference (Gregory 1970) and the theory of Direct perception (de Wit
et al. 2015; Gehringer and Engel 1986; Gibson 1979; Runeson 1988) (Fig. 27).
The architectonic device developed was really firstly (Gregory 1994) guessed by
Hermann Helmholtz (Cz 1896), the principal reference for the Ames’ school (Pastore
1971), and it is connected to the artistic representation of the distorted room (Van De
Geer, De Natris) designed by artists to simulate three-dimensional spaces, as the case
of Samuel Van Hoogstraten (Brusati 1995) and of his amazing peepshow (i.e. van
Hoogstraten 1655–60), but also Brunelleschi perspective machines (Arnheim 1978,
1986), the tools to realize perspective (Moscati 2012) and its poetic (Elkins 1994):
in the correspondence between medium as message, Ames’ room concretizes in the
spatial the Leon Battista Alberti conceptualization of perspective as intersection of
the visual pyramid (Leon Battista Alberti, De pictura incipit, cod. II.IV,38, f.119v),
Fig. 27 The illusion in the Ames room: the great professor Roberto de Rubertis with the professor
Franco Purini
a geometric transcription of the optical concept of a visual cone (de Rubertis 2012),
the expression of its artificiality (Gioseffi 1957), “lo inganno de gl’occhi”, the eyes
deception (Accolti 1625). In this way, the Ames’ room is linked also to all the
architectures’ illusion, as the works of Andrea del Pozzo shows (Migliari 2000)
or Agostino Tassi (Negro 1996) with which they share the instrument (but not the
objectives): “on one hand Pozzo takes advantage of the eye’s rotation, pivoting in
the projection centre and resorting to other expedients, mitigating the effects of a
displacement of the observer; on the other hand Tassi disregards the code in order
to nullify the effects of the displacement. The perspectives painted by Tassi, thus,
behave like the Ames’ room, because it appeals to the viewer’s experience and to his
mental models, in order to hide the derogations imposed to the perspective rules and
the true shape of the space that these perspectives describe, if interpreted literally”
(Migliari and Romor 2015, p. 66). Reinterpreting the value of this approach for the
contemporary, Ames’ “perspective machine”, by using projections, is based on a
rigorous geometric path to create a wrong drawing (Koenderink et al. 2016), the
same theme of digital virtuality founded in a distort simulation (render-to-texture).
Fig. 28 Generative model: variation in the whole family of distorted rooms
A mechanical production of vision, Ames’ room for its genesis it is not an art-
work, but our proposal aims to demonstrate that it could become a design work.
Created by ABITARE+ and designed by the research group of the Department of
Civil and Environmental Engineering of the University of Perugia, coordinated by
the author, the research is an expression of the path of optimization of wooden houses
just described. Conceived for an exposition event, where many companies were par-
ticipating, Ames’ room is born inside an iconoclast strategy, that in our imagination
was (Latour and Weibel 2002) important also for the building market. The unusual
installation wants to respond to the homo game aesthetic (Pecchinenda 2010) and to
its role in representative projection in social synthetic universes (Castronova 2007):
the objective is to produce references to move around the net and to talk to the com-
pany, a result that appears reached as it is testified from Google trend data (surely
reductive), which show that in the last year the name of the company was searched
Fig. 29 Digital simulation: the analysis of the results and the study of the formal details
a lot after the event in which the house was installed. The success obtained with this
first manifestation leas to install the pavilion in front of the prestigious faculty of
Valle Giulia of Rome University “La Sapienza” and by Engineer pole of the Uni-
versity of Perugia; both were events that represent only a portion of the installation
series thought for the development program, a condition that granted the production
of thousands of images (Figs. 28 and 29).
The “vibration of appearances that is the genesis of things” (Merleau-Ponty 1962)
makes this space signed by divergent walls, dynamically conformed to the con-
temporary aesthetic, its form attracts people, simulacrum that transfigures the real
by killing its three-dimensionality, but at the same time it creates life in the image
(Debray 1994). The pavilion has its function in the substantial interactivity connected
to interpretation, which is intended as an opening need to a process of deconstruc-
tion and reconstruction (Neisser 1967) of what it is perceived. The design is then
reinforced, as an instrument to verify and to project, founded on geometry, that finds
in generative logics a new strength.
If conceptually Ames’ room summarizes a wealth of approaches inherent the rep-
resentation that ranges from descriptive geometry to projective through the value of
images and perception, its realization has been developed by exploiting the innova-
Fig. 30 BIM model: the digital fabrication of the parts and the simulation of the construction phases
tive logics of three-dimensional modelling and BIM approach. The paradigm of the
approach used for the design of optimized buildings, in the room is developed in our
first research as the value of generative modelling (Filippucci 2010c, 2012b), useful
to represent different families of spaces and to find the best solution to respond to
spatial limitations imposed by the space where it is set up. The generative process
dialogues and comes into contact with the logic of BIM, in the definition of details of
the constructive elements of the room, which however translates the techniques used
for contemporary constructions in wood. The process thus demonstrates the value of
research for industry 4.0 and the realization of the concepts of mass customization
that are exemplified here (Fig. 30).
For its realization, it was designed a timber platform frame, inscribed in a three-
dimensional space of 4 × 5 × 3.5 m, a solid wood structure with plywood sheathing
in Pinepanels Oriented Strand Board (OSB) and interior leaning in medium-density
fibreboard (MDF) panel and external leaning in black printed canvas. Our choice
was to create a box in the panel positioned in correspondence to the point of view
constrained, with the aim of discovering the deception only through the smartphone
screen. The effect is really understood when two people in the corners of the room
change their mutual position. The realization shows as restricted sight is not an
Fig. 31 The installation of the Ames’ room: the timber structure and the final realization
exact point, but an area, admitting an ample displacement of the observation point
(Migliari and Romor 2015), a condition that leads to an accommodation of perception
that becomes accustomed to this conformation (Mitchison and Westheimer 1984),
despite the clearest clues of stereoscopic vision (Pilewski and Martin 1991). To insert
deformed writings and logos in the deformed back wall in function of the different
events, our room doesn’t have deformed windows, but using quiconce drawn on the
ground, the camera pushes the people to assume that the images arise from objects of
the same size (Gogel 1969), also because under reduced conditions it is demonstrated
how objects are perceived at a default distance (specific distance tendency) (Gogel
1976), related to the resting state of vergence (Gogel and Tietz 1977).
Ames’ room becomes an interesting paradigm of our culture of the image and of
our mistrusted in them. As Karl Popper asserts, “if we are uncritical we shall always
find what we want: we shall look for, and find, confirmations, and we shall look away
from, and not see, whatever might be dangerous to our pet theories. In this way, it
is only too easy to obtain what appears to be overwhelming evidence in favour of
a theory which, if approached critically, would have been refuted” (Popper 1957,
p. 124) (Fig. 31).
5 Conclusions
“Centuries ago there lived–”A king!” my little readers will say immediately. No,
children, you are mistaken. Once upon a time there was a piece of wood. It was not
an expensive piece of wood. Far from it. Just a common block of firewood, one of
those thick, solid logs that are put on the fire in winter to make cold rooms cozy and
warm. I do not know how this really happened, yet the fact remains that one fine
day this piece of wood found itself in the shop of an old carpenter” (Carlo Collodi,
Pinocchio, 1883).
As in the children novel, also “a piece of wood”, a “common block of firewood”,
if animated by the creativity of an “old” but every innovative drawing, leads to the
discovery of something vital, that is inside the material itself. Wood, by meeting
human labour, is able to generate poetry. Geppetto is not just a simple character of a
fable, but is the touching symbol and metaphor of those who work with matter and
form, and from what appears inanimate he gives birth to something that projects and
redefines its own identity.
The (digital) technical capacities, material potentialities, vitality of the creativity,
all together characterize the experiences reported, they represent just a selection of
a series of research developed and in developing. In all these studies it is possible to
mark the centrality of form, finding the approach and the connection optimization
processes in the support offered by AI in representation tool, without forgetting
wood’s natural intelligence. The focus, for the designer, moves on the laws that rule
the organization of materials, architectures and environmental relationships.
The digital environment, the rewriting of descriptive geometry law, through the
rules and the logic of the generative design, supports the construction of the perfor-
mative architectural solutions. Digital representation does not appear to be a limit,
but the instrument to the Mongian “passage from what is known to what is unknown”.
AI aids designer in the morphogenesis process, also if is it represented by stereo-
metrics forms. All researches mark how the real question is the construction of the
model, founded in the value of interdisciplinarity, an open process aimed to under-
stand the weaving of themes included in the morphological results. The simulation
offered by digital processes is the means to define the ideated image, the real soul of
construction and of architecture.
References
Abdelmohsen S, Massoud P, Elshafei A (2016) Using tensegrity and folding to generate soft respon-
sive architectural skins. In: Proceedings of the 34th eCAADe conference of complexity & sim-
plicity. Oulu, pp 529–536
Accolti P (1625) Lo inganno de gl’occhi, prospettiva pratica de Pietro Accolti. Pietro Cecconcelli,
Florence
Addington M, Schodek DL (2005) Smart materials and new technologies: for the architecture and
design professions. Architectural, Oxford
Adriaenssens S, Barnes M (2001) Tensegrity spline beam and grid shell structures. Eng Struct
23(1):29–36. https://doi.org/10.1016/S0141-0296(00)00019-5
Adriaenssens S, Block P, Veenendaal D, Williams C (2014) Shell structures for architecture: form
finding and optimization. Routledge, New York
Aish R, Woodbury R (2005) Multi-level interaction in parametric design. In: Smart graphics.
Springer, Berlin, Heidelberg, pp 151–162. https://doi.org/10.1007/11536482_13
Ames A, Ittelson WH (1952) The Ames demonstrations in perception. Hafner Publishing, New
York
Anderson DM (2002) Build-to-order and mass customization: the ultimate supply chain manage-
ment and lean manufacturing strategy for low-cost on-demand production without forecasts or
inventory. CIM Press, Cambria
Andia A, Spiegelhalter T (eds) (2017) Post-parametric automation in design and construction.
Artech House, Boston
Arnheim R (1978) Brunelleschi’s Peepshow. Zeitschrift Für Kunstgeschichte 41(1): 57. https://doi.
org/10.2307/1481995
Arnheim R (1986) New essays on the psychology of art. University of California Press, Berkeley
Asl MR, Bergin M, Menter A, Yan W (2014) BIM-based parametric building energy perfor-
mance multi-objective optimization. In: Fusion—proceedings of the 32nd eCAADe conference.
eCAADe, Newcastle, pp 455–464
Austern G, Capeluto IG, Grobman YJ (2018) Rationalization methods in computer aided fabrication:
a critical review. Automation in Construction 90:281–293. https://doi.org/10.1016/J.AUTCON.
2017.12.027
Bader J, Zitzler E (2011) HypE: an algorithm for fast hypervolume-based many-objective optimiza-
tion. Evol Comput 19(1):45–76. https://doi.org/10.1162/EVCO_a_00009
Bar-Cohen Y (ed) (2016) Biomimetics nature-based innovation. CRC Press, Boca Raton
Barnekov VG, McMillin CW, Huber HA (1986) Factors influencing laser cutting of wood. For Prod
J 1(36):55–58
Barthel R (1967) Natural forms-architectural forms. In: Nerdinger W (ed) Frei Otto complete works.
Birkhäuser, Basel-Boston-Berlin, Architecture, pp 16–32
Baudrillard J (1981) Simulacres et simulation. Galilée, Paris
Baverel O, Larsen OP (2011) A review of woven structures with focus on reciprocal systems—Nex-
orades. Int J Space Struct 26(4):281–288. https://doi.org/10.1260/0266-3511.26.4.281
Baverel O, Pugnale A (2014) Reciprocal systems based on planar elements: morphology and design
explorations. Nexus Netw J 1(16):179–189
Bechert S, Knippers J, Krieg OD, Menges A, Schwinn T, Sonntag D (2016) Textile fabrication
techniques for timber shells elastic bending of custom-laminated veneer for segmented shell
construction systems. In: Adriaenssens S, Gramazio F, Kohler M, Menges A, Pauly M (eds)
Advances in architectural geometry 2016. Vdf Hochschulverlag AG an der ETH Zürich, Zürich,
pp 154–170
Benros D, Duarte JP (2009) An integrated system for providing mass customized housing. Autom
Constr 18(3):310–320. https://doi.org/10.1016/J.AUTCON.2008.09.006
Benyus JM (1997) Biomimicry: innovation inspired by nature. Harper Perennial, New York
Bergin M, Steinfeld K (2012) Housing agency system (HAS): multi-criteria satisfying and mass-
customization of homes. In: ACSA fall conference. OFFSITE, Philadelphia, pp 93–97
Bergmann E, Hildebrand S (2015) Form-finding, form-shaping, designing architecture. Mendrisio

Academy Press, Mendrisio
Bhushan B (2009) Biomimetics: lessons from nature–an overview. Philos Trans: Series A, Mathe-
matical, Physical, and Engineering Sciences 367(1893):1445–1486. https://doi.org/10.1098/rsta.
2009.0011
Bianconi F (2002) Tetraktis. Strumenti, luoghi, materia, rilievo. Digital Point, Perugia
Bianconi F (2005) Segni digitali. Morlacchi, Perugia
Bianconi F, Filippucci M (2015) The dams of Rio Grande’s basin (Amelia TR). In: Gambardella C
(ed) XIII Forum Internazionale Le Vie dei Mercanti. Heritage and technology mind knowledge
experience, vol 1. La scuola di Pitagora, Napoli, pp 1864–1875
Bianconi F, Filippucci M (2016a) Generative education: thinking by modeling/modeling by think-
ing. In: EGA. Congreso: XVI Congreso Internacional de Expresión Gráfica Arquitectónica “El
arquitecto, de la tradición al siglo XXI”, vol 1. Grupo Enlace Gráfico, pp 747–754
Bianconi F, Filippucci M (2016b) The parameterization of complex surfaces for engineering solu-
tions. In: Le ragioni del Disegno/the reasons of drawing, vol 1. Gangemi, pp 125–130
Bianconi F, Verducci P, Filippucci M (2006) Architetture dal Giappone: disegno, progetto e tecnica,
vol 1. Gangemi, Roma
Bianconi F, Filippucci M, Andreani S (2016a) Computational design and built environments. In:
3D printing: breakthroughs in research and practice, vol 1. IGI Global, Hershey (Pennsylvania,
USA), pp. 361–395
Bianconi F, Filippucci M, Verdecchia C (2016b) Body movement based architecture. In: Visual
computing and emerging geometrical design tools, vol 2. IGI Global, Hershey (Pennsylvania,
USA), pp 744–770. https://doi.org/10.4018/978-1-5225-0029-2.ch030
Bianconi F, Catalucci S, Filippucci M, Marsili R, Moretti M, Rossi G, Speranzini E (2017a) Com-
parison between two non-contact techniques for art digitalization. J Phys Conf Ser 882. https://
doi.org/10.1088/1742-6596/882/1/012005
Bianconi F, Filippucci M, Catalucci S (2017b) Line and Points. Critical analysis of evolution
of archaeological survey in forty years of experiences in Umbria. DISEGNARECON 10(19):
4–1–E4.20
Bianconi F, Filippucci M, Catalucci S (2017c) The identity landscape in the cataloging of scat-
tered assets in the area of Amelia. In: Putting tradition into practice: heritage, place and design.
Proceedings of 5th INTBAU international annual event. Springer, pp 984–993
Bianconi F, Filippucci M, Buffi A, Calabro’ MP (2017d) The value of image. The design of and
data streams from the perception by design. In: Proceedings 2017. International and interdisci-
plinary conference IMMAGINI? Image and imagination between representation, communication,
education and psychology. MDPI. https://doi.org/10.3390/proceedings1090933
Bianconi F, Filippucci M, Ciarapica A (2017e) Landscape, territory, knowledge. From Umbria
region’s atlas of objectives to the “Landscape Contracts” of Trasimeno Lake. In Crisis landscapes:
opportunities and weaknesses for a sustainable development. FrancoAngeli, Roma, pp 87–110
Bianconi F, Filippucci M, Clemente M, Salvati L (2017f) Green infrastructures and biodiverse urban
gardens for regenerating urban spaces. In: Gospodini A (ed) Book of abstracts of the international
conference on changing cities III spatial, design, landscape; socio-economic dimensions: 26–30
June 2017, Syros, Delos, Mykonos Islands, Greece. Grafima Publications, Thessaloniki, pp 42–42
Bianconi F, Filippucci M, Margutti M, Stramaccia M (2017g) Evoluzioni morfologiche di trans-
poliedri. Eloquenza delle immagini per generare strutture tensegrali Morphological evolutions of
transpolyhedra. Eloquence of the images to generate tensegrity structures XY(3):4–15
Bianconi F, Filippucci M, Margutti M, Stramaccia M (2018) Drawing Tensegrity, discover trans-
polyhedra. In: D’Uva D (ed) Analyzing form and morphogenesis in modern architectural contexts.
New York: IGI Global, pp 41–68 https://doi.org/10.4018/978-1-5225-3993-3.ch003
Bianconi F, Clemente M, Filippucci M, Salvati L (2018a) Re-sewing the urban periphery. A green
strategy for Fontivegge District in Perugia. TEMA 11: 107–118
Bianconi F, Filippucci M, Ciculi L (2018b) The form of music: experiments between cymatics and
engineering. In Nexus 2018 architecture and mathematics conference book. Kim Williams Book,
pp 233–244
Bianconi F, Filippucci M, Clemente M, Salvati L (2018c) Regenerating urban spaces under place-
specific social contexts: a brief commentary on green infrastructures for landscape conservation.
Int J Soc Sci 2:18–32. https://doi.org/10.20472/SS.2017.6.2.002
Bianconi F, Filippucci M, Margutti M, Stramaccia M (2018d) Drawing tensegrity, discover trans-
polyhedra. In D’Uva D (ed) Analyzing form and morphogenesis in modern architectural contexts.
IGI Global, New York, pp 41–68. https://doi.org/10.4018/978-1-5225-3993-3.ch003
Bianconi F, Filippucci M, Seccaroni M (2018e) Drawing architectonic choices. Representation
and optimization in design pathway. In: De-sign environment landscape city. Genova University
Press, Genova, p 47
Bianconi F, Filippucci M, Seccaroni M (2018f) Rappresentazione e variazione della forma architet-
tonica per l’ottimizzazione emergetica ed energetica. In: 18th CIRIAF national congress sustain-
able development, human health and environmental protection. Perugia, p Cod_018
Bidgoli A, Cardoso-Llach D (2015) Towards a motion grammar for robotic stereotomy. In: Ikeda
Y, Herr CM, Holzer D, Kaijima S, Kim MJ (eds) Emerging experience in past, present and future
of digital architecture. The Association for Computer-Aided Architectural Design Research in
Asia (CAADRIA), Hong Kong, pp 723–732
Bittermann MS (2009) Intelligent design objects (IDO): a cognitive approach for performance-based
design. www.boekenbent.com
Blessing WW, Landauer AA, Coltheart M (1967) The effect of false perspective cues on distance-
and size-judgments: an examination of the invariance hypothesis. Am J Psychol 80(2):250. https://
doi.org/10.2307/1420984
Block P, Rippmann M, Van Mele T, Escobedo D (2017) The Armadillo Vault: balancing computation
and traditional craft. In: Menges A, Sheil B, Glinn T, Skavara M (eds) Fabricate 2017: rethinking
design and construction. UCL Press, London, pp 286–293
Bolles RC, Bailey DE (1956) Importance of object recognition in size constancy. J Exp Psychol
51(3):222–225. https://doi.org/10.1037/h0048080
Bosse A (1643a) La pratique du trait à preuves, de Mr Desargues Lyonnois, pour la coupe des
pierres en l’architecture. Impr. Des-Hayes
Bosse A (1643b) La pratique du trait à preuves de Mr Desargues,… pour la coupe des pierres …
Impr. Des-Hayes
Bouzanjani BF, Leach N, Huang A, Fox M, Pomona CP (2013) Alloplastic architecture: the design
of an interactive tensegrity structure. In: Beesley P, Khan O, Stacey M (eds) Adaptive architecture:
ACADIA 2013. Riverside Architectural Press, Toronto, pp 129–136
Brown A (2014) The genius of Japanese carpentry: the secrets of a craft. Tuttle Publishing, Clarendon
Brown NC, Mueller CT (2016) Design for structural and energy performance of long span build-
ings using geometric multi-objective optimization. Energy Build 127:748–761. https://doi.org/
10.1016/J.ENBUILD.2016.05.090
Brown N, Mueller C (2017) Designing with data: moving beyond the design space catalog. Acadia
2017 Discipline + Distruption. MIT Press, Cambridge, pp 154–163
Bruce J, Caluwaerts K, Iscen A, Sabelhaus AP, SunSpiral V (2014) Design and evolution of a modu-
lar tensegrity robot platform. In: 2014 IEEE international conference on robotics and automation
(ICRA). IEEE, pp 3483–3489. https://doi.org/10.1109/ICRA.2014.6907361
Brusati C (1995) Artifice and illusion: the art and writing of Samuel van Hoogstraten. University
of Chicago Press, Chicago
Buckminster Fuller R (1961) Tensegrity. Portfolio Artnews Ann 4:112–127
Buckminster Fuller R (1975) Synergetics: explorations in the geometry of thinking. Macmillan
Publishing, Basingstoke
Burgert I, Fratzl P (2009) Actuation systems in plants as prototypes for bioinspired devices. Phi-
los Trans. Series A, Mathematical, Physical, and Engineering Sciences 367(1893):1541–1557.
https://doi.org/10.1098/rsta.2009.0003
Burkhardt RW (2008) A practical guide to tensegrity design. Cambridge

Burry M (2014) From descriptive geometry to smartgeometry: first steps towards digital architec-
ture. In: Peters B, Peters T (eds) Inside smartgeometry. Wiley, Chichester, West Sussex, United
Kingdom, pp 154–165. https://doi.org/10.1002/9781118653074.ch13
Calladine CR (1978) Buckminster Fuller’s “Tensegrity” structures and Clerk Maxwell’s rules for
the construction of stiff frames. Int J Solids Struct 14(2):161–172. https://doi.org/10.1016/0020-
7683(78)90052-5
Castronova E (2007) Universi sintetici: come le comunità online stanno cambiando la società e
l’economia. Mondadori, Milan
Cefalo M, Mirats Tur JM (2010) Real-time self-collision detection algorithms for tensegrity systems.
Int J Solids Struct 47(13):1711–1722. https://doi.org/10.1016/J.IJSOLSTR.2010.03.010
Chandana P, Lipson H, Cuevas FJV (2005) Evolutionary form-finding of tensegrity structures.
In Proceedings of the 2005 conference on genetic and evolutionary computation—GECCO’05.
ACM Press, New York, New York, USA, p 3. https://doi.org/10.1145/1068009.1068011
Chaszar A, Glymph J (2010) CAD/CAM in the business of architecture, engineering and con-
struction. In: Corser R (ed) Fabricating architecture: selected readings in digital design and
manufacturing. Princeton Architectural Press, Princeton, pp 86–93
Chen L, Sass L (2017) Generative computer-aided design: multi-modality large-scale direct phys-
ical production. Comput Aided Des Appl 14(1):83–94. https://doi.org/10.1080/16864360.2016.
1199758
Chilton JC, Tang G (2016) Timber gridshells: architecture, structure and craft. Routledge, London
Clune J, Lipson H (2011) Evolving three-dimensional objects with a generative encoding inspired by
developmental biology motivation and previous work. Eur Conf Artif Life. MIT Press, Cambridge,
pp 144–148
Connelly R, Back A (1998) Mathematics and tensegrity. Am Sci 86:142–151. https://doi.org/10.
2307/27856980
Connelly R, Whiteley W (1992) The stability of tensegrity frameworks. Int J Space Struct
7(2):153–163. https://doi.org/10.1177/026635119200700208
Cornish V (1935) Scenery and the sense of sight. University Press, Cambridge
Correa D, Papadopoulou A, Guberan C, Jhaveri N, Reichert S, Menges A, Tibbits S (2015) 3D-
printed wood: programming hygroscopic material transformations. 3D Printing Addit Manuf
2(3):106–116. https://doi.org/10.1089/3dp.2015.0022
Corser R (2010) Fabricating architecture: selected readings in digital design and manufacturing.
Princeton Architectural Press, Princeton
Couldry N, Powell A (2014) Big data from the bottom up. Big Data Soc 1(2). https://doi.org/10.
1177/2053951714539277
Cully A, Clune J, Tarapore D, Mouret J-B (2015) Robots that can adapt like animals. Nature
521:503–507. https://doi.org/10.1038/nature14422
Cz P (1896) Handbuch der physiologischen Optik. Monatshefte Für Mathematik Und Physik.
https://doi.org/10.1007/BF01708548
d’Estrée Sterk T (2003) Using actuated tensegrity structures to produce a responsive architecture.
In: Klinger KR (ed) Crossroads of digital discourse. Ball State University, pp 85–93
Dangel U (2016) Turning point in timber construction: a new economy. Birkhäuser, Basilea
Davis J, Edgar T, Porter J, Bernaden J, Sarli M (2012) Smart manufacturing, manufacturing intel-
ligence and demand-dynamic performance. Comput Chem Eng 47:145–156. https://doi.org/10.
1016/J.COMPCHEMENG.2012.06.037
Dawson C, Vincent JFV, Rocca A-M (1997) How pine cones open. Nature 390(6661):668. https://
doi.org/10.1038/37745
De Azambuja Varela P, Sousa JP (2016) Revising stereotomy through digital technology. In: Her-
neoja A, Österlund T, Markkanen P (eds) Complexity & simplicity—proceedings of the 34th
eCAADe conference. University of Oulu, Oulu, pp 427–434
De L’Orme P (1568) Le premier tome de l’architecture. Fédéric Morel, Paris
De La Hire P (1596) Traité de la coupe des pierres. Bibliothèque de l’Institut de France, Paris
de Rubertis R (2012) Piramide visiva. Retrieved 28 Aug 2018 from http://www.wikitecnica.com/

piramide-visiva/
de Wit MM, van der Kamp J, Withagen R (2015) Visual illusions and direct perception: elaborating
on Gibson’s insights. New Ideas Psychol 36:1–9. https://doi.org/10.1016/J.NEWIDEAPSYCH.
2014.07.001
Debray R (1994) Vie et mort de l’image une histoire du regard en Occident. Gallimard, Paris
Deleuze, G. (1994). Difference and repetition. Columbia University Press
Deleuze G, Guattari F (1987) A thousand plateaus: capitalism and schizophrenia. University of
Minnesota Press
Dellaert BGC, Stremersch S (2005) Marketing mass-customized products: striking a balance
between utility and complexity. J Mark Res 42(2):219–227. https://doi.org/10.1509/jmkr.42.2.
219.62293
Desarguer G (1640) Brouillon project d’exemples d’une manière universelle du sieur G.D.L.,
touchant la pratique du trait à preuve pour la coupe des pierres en architecture …. Melchor
Tavernier, Paris
Di Carlo B (2008) The wooden roofs of leonardo and new structural research. Nexus Netw J
10(1):27–38. https://doi.org/10.1007/s00004-007-0054-x
Dogra JK, Kaur B, Parsarokh A (2015) tensegri[city]. ELISAVA, Barcellona
Duarte JP (2005) A discursive grammar for customizing mass housing: the case of Siza’s houses at
Malagueira. Autom Constr 14(2):265–275. https://doi.org/10.1016/J.AUTCON.2004.07.013
Duray R, Ward PT, Milligan GW, Berry WL (2000) Approaches to mass customization: configu-
rations and empirical validation. J Oper Manag 18(6):605–625. https://doi.org/10.1016/S0272-
6963(00)00043-7
Duro-Royo J, Oxman N (2015) Towards fabrication information modeling (FIM): four case models
to derive designs informed by multi-scale trans-disciplinary data. MRS Proceedings, 1800. https://
doi.org/10.1557/opl.2015.647
Eastman CM (2011) BIM handbook: a guide to building information modeling for owners,
managers, designers, engineers and contractors. Wiley, Hoboken. Retrieved from https://
books.google.it/books?id=aCi7Ozwkoj0C&dq=Eastman,+C.+(2011+BIM+handbook:+a+
guide+to+building+information+modeling+for+owners,+managers,+designers,+engineers,+
and+contractors&hl=it&source=gbs_navlinks_s
Eastman C, Lee J, Jeong Y, Lee J (2009) Automatic rule-based checking of building designs. Autom
Constr 18(8):1011–1033. https://doi.org/10.1016/J.AUTCON.2009.07.002
Elkins J (1994) The poetics of perspective. Cornell University Press, London
Emmerich DG (1988) Structures tendues et autotendantes Monographies de géometrie constructive.
Editions de la La Villette, Paris
Emmerich DG (1996) Emmerich on self-tensioning structures. Int J Space Struct 11(1–2):29–36.
https://doi.org/10.1177/026635119601-205
Epstein W, Park J, Casey A (1961) The current status of the size-distance hypotheses. Psychol Bull
58(6):491–514. https://doi.org/10.1037/h0042260
Evans R (1995) The projective cast: architecture and its three geometries. MIT Press, Cambridge
Eversmann P, Gramazio F, Kohler M (2017) Robotic prefabrication of timber structures: towards
automated large-scale spatial assembly. Constr Rob 1(1–4):49–60. https://doi.org/10.1007/
s41693-017-0006-2
Eigensatz M, Kilian M, Schiftner A, Mitra NJ, Pottmann H, Pauly M, et al. (2010) Paneling archi-
tectural freeform surfaces. In: ACM transactions on graphics. Proceedings of ACM SIGGRAPH,
vol 29. ACM Press, New York, p 1. https://doi.org/10.1145/1833349.1778782
Fagerström G (2009) Dynamic relaxation of tensegrity structures. In: Proceedings of the 14th
international conference on computer aided architectural design research in Asia/Yunlin (Taiwan)
22–25 Apr 2009. CAADRIA, Taiwan, pp 553–562 (pp 553–562)
Fallacara G (2006) Digital stereotomy and topological transformations: reasoning about shape
building. In: Proceedings of second international congress construction history. Queen’s College
Cambridge, Cambridge, pp 1075–1092
Fallacara G (2007) Verso una progettazione stereotomica: nozioni di stereotomia, stereotomia dig-
itale e trasformazioni topologiche : ragionamenti intorno alla costruzione della forma. Aracne,
Roma
Fernando S, Saunders R, Weir S (2015). Surveying stereotomy: investigations in arches, vaults
and digital stone masonry. In: Architectural Research Centers Consortium (ed) ARCC 2015
conference—the future of architectural research. Perkins + Will, pp 82–89
Field JV (1987) Linear perspective and the projective geometry of Girard Desargues. Nuncius
2(2):3–40. https://doi.org/10.1163/182539187X00015
Filippucci M (2010a) Nuvole di pixel. La fotomodellazione con software liberi per il rilievo
d’architettura. DISEGNARE CON…, 3:50–63. https://doi.org/10.6092/issn.1828-5961/2081
Filippucci M (2010b) Virtual in virtual, discretization in discretization. Shape and perception in
parametric modelling for renewing descriptive geometry. In: Proceedings of ICGG 2010 14TH
international conference on geometry and graphics. Naomi Ando et al, pp 129–130
Filippucci M (2010c) Virtual in virtual, discretization in discretization. Shape and perception in
parametric modelling for renewing descriptive geometry. In: Ando N et al (ed) Proceedings of
ICGG 2010 14TH international conference on geometry and graphics. Kyoto, pp 129–130
Filippucci M (2012a) Dalla forma urbana all’immagine della città. Percezione e figurazione
all’origine dello spazio costruito. Sapienza Università di Roma
Filippucci M (2012b) Rappresentazione al quadrato. Il disegno generativo per il rinnovamento della
geometria descrittiva. In: Carlevalis L, De Carlo L, Migliari R (eds) Attualità della Geometria
Descrittiva Seminario nazionale sul rinnovamento della Geometria descrittiva, Roma dicembre
2009, marzo 2010. Gangemi, Rome
Filippucci M (2015) Primitive Urbane. Analisi interpretativa dei processi figurativi dell’immagine
della città. In: Novello G, Marotta A (eds). Gangemi edizioni, Torino, p 1219
Filippucci M, Bianconi F, Andreani S (2016a) Computational design and built environments: the
quest for an alternative role of the digital in architecture. 3D printing: breakthroughs in research
and practice. https://doi.org/10.4018/978-1-5225-1677-4.ch019
Filippucci M, Rinchi G, Brunori A, Nasini L, Regni L, Proietti P (2016b) Architectural modelling
of an olive tree. Generative tools for the scientific visualization of morphology and radiation
relationships. Ecol Inf 36. https://doi.org/10.1016/j.ecoinf.2016.09.004
Filippucci M, Bianconi F, Bettollini E, Meschini M, Seccaroni M (2017) Survey and representa-
tion for rural landscape. New tools for new strategies: the example of Campello Sul Clitunno.
In: Proceedings 2017. International and interdisciplinary conference IMMAGINI? Image and
imagination between representation, communication, education and psychology, vol 1. MDPI,
Bressanone, p 934. https://doi.org/10.3390/proceedings1090934
Filippucci M, Bianconi F, Bettollini E, Meschini M (2018) Visual perception analysis for landscape
evaluation. an experimental case, Campello Sul Clitunno. De_Sign Environment Landscape City,
vol p. Genova University Press, Genova, p 113
Floreano D, Mattiussi C (2009) Bio-inspired artificial intelligence: theories, methods, and technolo-
gies. Scalable Comput Pract Exp 10(4). https://doi.org/10.12694/scpe.v10i4.623
Fogel LJ, Owens AJ, Walsh MJ (1966) Artificial intelligence through simulated evolution. Wiley,
Oxford
Frézier AF (1737a) La théorie et la pratique de la coupe des pierres et des bois pour la construction
des voûtes et autres parties des bâtiments civils, militaires, ou Traité de stéréotomie, à l’usage de
l’architecture. Tome 3 /, par M. Frézier,… Paris: L.H. Guerin
Frézier AF (1737b) La théorie et la pratique de la coupe des pierres et des bois pour la construction
des voûtes et autres parties des bâtiments civils; militaires, ou Traité de stéréotomie, à l’usage de
l’architecture. Tome 3 /, par M. Frézier,… Paris: L.H. Guerin
Frézier AF (1760) Élémens de stéréotomie, à l’usage de l’architecture, pour la coupe des pierres.
Jombert, Paris
Fuller RB (1963) Ideas and integrities. Macmillan, New York
Gehringer WL, Engel E (1986) Effect of ecological viewing conditions on the Ames’ distorted
room illusion. J Exp Psychol Hum Percept Perform 12(2):181–185
Gherardini F, Leali F (2017) Reciprocal frames in temporary structures: an aesthetical and parametric
investigation. Nexus Netw J 19(3):741–762. https://doi.org/10.1007/s00004-017-0352-x
Gibson JJ (1979) The ecological approach to visual perception. Houghton Mifflin, Boston
Gioseffi D (1957) Perspectiva artificialis: Per la storia della prospettiva; spigolature e appunti.
Università Degli Studi di Trieste, Facoltà di Lettere e Filosofia, Trieste
Goel AK, McAdams DA, Stone RB (2014) Biologically inspired design. Springer, London. https://
doi.org/10.1007/978-1-4471-5248-4
Gogel WC (1969) The sensing of retinal size. Vision Res 9(9):1079–1094. https://doi.org/10.1016/
0042-6989(69)90049-2
Gogel WC (1976) An indirect method of measuring perceived distance from familiar size. Percept
Psychophys 20(6):419–429. https://doi.org/10.3758/BF03208276
Gogel WC, Tietz JD (1977) Eye fixation and attention as modifiers of perceived distance. Percept
Mot Skills 45(2):343–362. https://doi.org/10.2466/pms.1977.45.2.343
Gombrich EH (1960) Art and Illusion: a study in the psychology of pictorial representation. Phaïdon
Press, London
Gough M (1998) In the laboratory of constructivism: Karl Ioganson’s cold structures. October, 84,
90. https://doi.org/10.2307/779210
Gramazio F, Kohler M (2014) Made by robots: challenging architecture at the large scale AD. Wiley,
London
Gramazio F, Kohler N, Oesterle S (2010) Encoding material. In: Oxman R, Oxman R
(eds) The new structuralism: design, engineering and architectural technologies. Wiley, pp
108–115. Retrieved from https://books.google.it/books?id=035HAQAAIAAJ&q=The+new+
Structuralism:+Design,+Engineering+and+Architectural+Technologies+AD&dq=The+new+
Structuralism:+Design,+Engineering+and+Architectural+Technologies+AD&hl=it&sa=X&
ved=0ahUKEwjirozclondAhVMkiwKHQFfBbEQ6A
Greenough H (1947) Form and function: remarks on art, design, and architecture. University of
California Press, Berkeley
Gregory RL (1970) The intelligent eye. McGraw-Hill Book Company, New York
Gregory RL (1987) Analogue transactions with Adelbert Ames. Perception 16(3):277–282. https://
doi.org/10.1068/p160277
Gregory RL (1994) Even odder perceptions. Routledge, London
Grima JN, Mizzi L, Azzopardi KM, Gatt R (2016) Auxetic perforated mechanical metamate-
rials with randomly oriented Cuts. Adv Mater 28(2):385–389. https://doi.org/10.1002/adma.
201503653
Grobman YJ, Neuman E (eds) (2013) Performalism: form and performance in digital architecture.
Routledge, New York
Gruber P, Jeronimidis G (2012) Has biomimetics arrived in architecture? Bioinspiration Biomimet-
ics 7(1):010201. https://doi.org/10.1088/1748-3182/7/1/010201
Hanaor A (1992) Aspects of design of double-layer tensegrity domes. Int J Space Struct
7(2):101–113. https://doi.org/10.1177/026635119200700204
Hensel M (2010) Performance-oriented architecture: towards a biological paradigm for architec-
tural design and the built environment. FORMakademisk 3(1):36–56. https://doi.org/10.7577/
formakademisk.138
Hensel M (2013) Performance-oriented architecture: rethinking architectural design and the built
environment. Wiley, Chichester
Hensel M, Menges A (2006) Morpho-ecologies. Architectural Association, London
Hensel M, Menges A (2008) Inclusive performance: efficiency versus effectiveness towards a
morpho-ecological approach for design. Architectural Design 78(2):54–63. https://doi.org/10.
1002/ad.642
Hensel M, Menges A, Weinstock M (2010) Emergent technologies and design. Routledge, New
York
Herzog T, Natterer J, Schweitzer R, Volz M, Winter W (2004) Timber construction manual. DETAIL
- Birkhäuser, Basel
Hofman E, Halman JIM, Ion RA (2006) Variation in housing design: identifying customer prefer-
ences. Housing Stud 21(6):929–943. https://doi.org/10.1080/02673030600917842
Holstov A, Bridgens B, Farmer G (2015) Hygromorphic materials for sustainable responsive archi-
tecture. Constr Build Mater 98:570–582. https://doi.org/10.1016/J.CONBUILDMAT.2015.08.
136
Holstov A, Farmer G, Bridgens B (2017) Sustainable materialisation of responsive architecture.
Sustainability 9(3):435. https://doi.org/10.3390/su9030435
Holway AH, Boring EG (1941) Determinants of apparent visual size with distance variant. Am J
Psychol 54(1):21. https://doi.org/10.2307/1417790
Huang JC (2008) Participatory design for prefab house: using internet and query approach of
customizing prefabricated houses. VDM Verlag Dr. Müller, Saarbrücken
Iafrate F (2018) Artificial intelligence and big data: the birth of a new intelligence. Wiley, London
Ittelson WH, Kilpatrick FP (1952) Equivalent configurations and the monocular and binocular
distorted rooms. Human nature from the transactional point of view. Institute for Associated
Research, New York, pp 41–55
Iwamoto L (2009) Digital fabrications: architectural and material techniques. Princeton Architec-
tural Press, Princeton
Jabi W (2013) Parametric design for architecture. Laurence King, London
Jones NL (2009a) Architecture as a complex adaptive system. Faculty of the Graduate School of
Cornell University
Jones NL (2009b) Architecture as a complex adaptive system. Faculty of the Graduate School of
Cornell University
Kaufmann H, Nerdinger W (eds) (2011) Building with timber: paths into the future. Prestel Verlag,
Munich
Kicinger R, Arciszewski T, Jong K De (2005) Evolutionary computation and structural design: a
survey of the state-of-the-art. Comput Struct 83(23–24):1943–1978. https://doi.org/10.1016/J.
COMPSTRUC.2005.03.002
Kilpatrick FP, Ittelson WH (1953) The size-distance invariance hypothesis. Psychol Rev
60(4):223–231. https://doi.org/10.1037/h0060882
Knaack U, Chung-Klatte S, Hasselbach R (2012) Prefabricated systems: principles of construction.
Birkhäuser, Basel
Knippers J, Speck T (2012) Design and construction principles in nature and architecture. Bioin-
spiration Biomimetics 7(1). https://doi.org/10.1088/1748-3182/7/1/015002
Knippers J, Nickel KG, Speck T (2016) Biomimetic research for architecture and building con-
struction: biological design and integrative structures. Springer, Cham
Koenderink J, van Doorn A, Pepperell R, Pinna B (2016) On right and wrong drawings. Art Percept
4(1–2):1–38. https://doi.org/10.1163/22134913-00002043
Kohlhammer T, Kotnik T (2011) Systemic behaviour of plane reciprocal frame structures. Struct
Eng Int 21(1):80–86. https://doi.org/10.2749/101686611X12910257102596
Kolarevic B (2004) Architecture in the digital age: design and manufacturing. Taylor & Francis,
New York
Kolarevic B (2015) From mass customisation to design ‘democratisation’. Architectural Des
85(6):48–53. https://doi.org/10.1002/ad.1976
Kolarevic B, Klinger K (2008) Manufacturing material effects: rethinking design and making in
architecture. Routledge, New York
Kolarevic B, Malkawi A (2005) Peformative architecture. Routledge, London
Konaković M, Crane K, Deng B, Bouaziz S, Piker D, Pauly M (2016) Beyond developable. ACM
Trans Graph 35(4):1–11. https://doi.org/10.1145/2897824.2925944
Krausse J, Lichtenstein C (2017) Your private sky R. Buckminster Fuller: the art of design science.
Lars Müller, Zurich
Krieg O, Christian Z, Correa D, Menges A, Reichert S, Rinderspacher K, Schwinn T (2014)
HygroSkin: meteorosensitive pavilion. In: Gramazio F, Kohler M, Langenberg S (eds) FAB-
RICATE: negotiating design and making. UCL Press, Zürich, pp 272–279
Krieg OD, Schwinn T, Menges A, Li J-M, Knippers J, Schmitt A, Schwieger V (2015) Biomimetic
lightweight timber plate shells: computational integration of robotic fabrication, architectural
geometry and structural design. In Block P, Knippers J, Mitra NJ, Wang W (eds) Advances in
architectural geometry 2014. Springer International Publishing, Cham, pp 109–125. https://doi.
org/10.1007/978-3-319-11418-7_8
Kuhlmann D (2011) In: Gruber P, Bruckner D, Hellmich C, Schmiedmayer HB, Stachelberger H,
Gebeshuber IC (eds) Biomorphism in architecture: speculations on growth and form. Springer,
Berlin, pp 149–178. https://doi.org/10.1007/978-3-642-11934-7_8
Kuroishi I (2015) Mathematics of carpentry in historic Japanese architecture. Architecture and
mathematics from antiquity to the future. Springer International Publishing, Cham, pp 333–347.
https://doi.org/10.1007/978-3-319-00137-1_23
Lalvani H (1996) Origins of tensegrity: views of Emmerich, Fuller and Snelson. Int J Space Struct
11(1–2):27–55. https://doi.org/10.1177/026635119601-204
Latour B, Weibel P (2002) Iconoclash: beyond the image wars in science, religion, and art. Center
for Art and Media, Karlsruhe
Laugier M-A (1735) Essai sur l’architecture. chez Duchesne, Paris
Lawrence TT (2003) Chassis+Infill: a consumer-driven, open source building approach for adapt-
able, mass customized housing. Institute of Technology, Massachusetts
Le Duigou A, Castro M, Bevan R, Martin N (2016) 3D printing of wood fibre biocomposites:
from mechanical to actuation functionality. Mater Des 96:106–114. https://doi.org/10.1016/J.
MATDES.2016.02.018
Lehman J, Clune J, Misevic D, Adami C, Altenberg L, Beaulieu J, Yosinski J (2018) The surprising
creativity of digital evolution: a collection of anecdotes from the evolutionary computation and
artificial life research communities. Neural and Evolutionary Computing
Lenyra S, Campos Titotto M, Egmar T, Deifeld C, Marcelo De Oliveira Pauletti R (2006) The
monument to the Futile FormIi: conception, simulation and realization of a tensegrity membrane
sculpture. La Tensored Rlte
Leroy C-F-A (1857) Traité de stéréotomie, comprenant les applications de la géométrie descriptive
à la théorie des ombres, la perspective linéaire, la gnomonique, la coupe des pierres et la charp-
ente, avec un atlas… par C.-F.-A. Leroy,… 2e édition… par M. E. Martelet, (Gauthiers-), Paris.
Retrieved from http://catalogue.bnf.fr/ark:/12148/cb30799169v
Li H, Huang T, Kong CW, Guo HL, Baldwin A, Chan N, Wong J (2008) Integrating design and
construction through virtual prototyping. Autom Constr 17(8):915–922. https://doi.org/10.1016/
J.AUTCON.2008.02.016
Liapi K (2004) A computer based system for the design and fabrication of tensegrity structures. In:
Beesley P, Cheng NY-W, Williamson RS (eds) Fabrication examining the digital practice of archi-
tecture: proceedings of the 2004 AIA/ACADIA fabrication conference. ACADIA, Cambridge
Lonardo E (2011) Strutture Tensegrali. Retrieved from https://issuu.com/emiliolonardo/docs/
strutture_tensegrali_pubb
Loonen RCGM, Trčka M, Cóstola D, Hensen JLM (2013) Climate adaptive building shells: state-
of-the-art and future challenges. Renew Sustain Energy Rev 25:483–493. https://doi.org/10.1016/
J.RSER.2013.04.016
López M, Rubio R, Martín S, Croxford Ben (2017) How plants inspire façades. From plants to archi-
tecture: biomimetic principles for the development of adaptive architectural envelopes. Renew
Sustain Energy Rev 67:692–703. https://doi.org/10.1016/J.RSER.2016.09.018
Loria G (1931) Il passato e il presente delle principali teorie geometriche. Cedam, Padova
Lynn G (1999) Animate form. Princeton Architectural Press, New York
Marks RW, Buckminster Fuller R (1973) The Dymaxion world of Buckminster Fuller. Anchor
Books, New York
Masic M, Skelton RE, Gill PE (2005) Algebraic tensegrity form-finding. Int J Solids Struct
42(16–17):4833–4858. https://doi.org/10.1016/J.IJSOLSTR.2005.01.014
Masic M, Skelton RE, Gill PE (2006) Optimization of tensegrity structures. Int J Solids Struct
43(16):4687–4703. https://doi.org/10.1016/J.IJSOLSTR.2005.07.046
Mattheck C (1998) Design in nature: learning from trees. Springer, Berlin, Heidelberg
Mazzoleni I (2013) Architecture follows nature: biomimetic principles for innovative design. CRC
Press, New York
McDonald RP, O’Hara PT (1964) Size-distance invariance and perceptual constancy. Am J Psychol
77(2):276. https://doi.org/10.2307/1420135
McGee W, Ponce de León M (eds) (2014) Robotic fabrication in architecture, art and design 2014.
Springer Science & Business Media, Cham
McHale J (1964) R. Buckminster Fuller. Il Saggiatore, Milano
Medaglia AL (2007) An object-oriented framework for rapid genetic algorithm development. In:
Rennard J (ed) Handbook of research on nature-inspired computing for economics and manage-
ment. IGI Global, Hershey, pp 608–624. https://doi.org/10.4018/978-1-59140-984-7.ch040
Menges A (2009) Performative wood: integral computational design for timber constructions. In:
d’Estrée Sterk T, Loveridge R, Pancoast D (eds) Building a better tomorrow—proceedings of the
29th annual conference of the association for computer aided design in architecture. Association
for Computer-Aided Design in Architecture, Chicago, pp 66–74
Menges A (2011) Integrative design computation: integrating material behaviour and robotic man-
ufacturing processes in computational design for performative wood constructions. In: ACADIA
11: integration through computation. Association for Computer Aided Design in Architecture,
Banff, p 413
Menges A (2012) Material computation: higher integration in morphogenetic design. Architectural
Des 82(2):14–21. https://doi.org/10.1002/ad.1374
Menges A (2013) Morphospaces of robotic fabrication. In: Brell-Çokcan S, Braumann J (eds) Rob
| Arch 2012. Springer, Vienna, pp 28–47. https://doi.org/10.1007/978-3-7091-1465-0_3
Menges A (2017) Integrative design computation for advancing wood architecture. In: Menges A,
Schwinn T, Krieg OD (eds) advancing wood architecture. Routledge, London, pp 97–110
Menges A, Ahlquist S (2011) Computational design thinking. Wiley, London
Menges A, Reichert S (2012) material capacity: embedded responsiveness. Architectural Des
82(2):52–59. https://doi.org/10.1002/ad.1379
Menges A, Schwinn T, Krieg OD (eds) (2017a) Advancing wood architecture. Routledge, London
Menges A, Sheil B, Glynn R, Skavara M (2017b) Fabricate: rethinking design and construction.
UCL Press, London
Merleau-Ponty M (1962) Senso e non senso (1948). Il Saggiatore, Milan
Migliari R (2000) La rappresentazione e il controllo dello spazio: morte e trasfigurazione della
Geometria Descrittiva. Isegnare: Idee, Immagini XI(20–21):9–18
Migliari R, Romor J (2015) Perspective: theories and experiments on the veduta vincolata (restricted
sight). J Geom Graph 19(1):57–77
Mirats Tur JM, Juan SH (2009) Tensegrity frameworks: dynamic analysis review and open problems.
Mech Mach Theory 44(1):1–18. https://doi.org/10.1016/J.MECHMACHTHEORY.2008.06.008
Mitchison GJ, Westheimer G (1984) The perception of depth in simple figures. Vision Res
24(9):1063–1073. https://doi.org/10.1016/0042-6989(84)90084-1
Monge G (1789) Géométrie descriptive. Boudouin, Paris
Moscati A (2012) La prospettiva pratica. Gli strumenti per costruire la prospettiva. In: De Carlo
L, Migliari R, Carlevaris L (eds) Attualità della geometria descrittiva : seminario nazionale
sul rinnovamento della geometria descrittiva, Roma, dicembre 2009-marzo 2010 (Gangemi).
Gangemi, Rome. Retrieved from https://books.google.it/books?id=EskELgEACAAJ&dq=
attualitàdellageometriaescrittiva&hl=it&source=gbs_book_other_versions
Motro R (2003) Tensegrity: structural systems for the future. Kogan Page Science, London
Motro R, Raducanu V (2003) Tensegrity systems. Int J Space Struct 18(2):77–84. https://doi.org/
10.1260/026635103769518198
Murakami H (2001) Static and dynamic analyses of tensegrity structures. Part 1. Nonlinear equations
of motion. Int J Solids Struct 38(20):3599–3613. https://doi.org/10.1016/S0020-7683(00)00232-
8
Myers W (2012) Beyond Biomimicry. In: Myers W (ed) Bio design: nature, science, creativity.
Museum of Modern Art, London, p 288
Nahmens I, Bindroo V (2011) Is customization fruitful in industrialized homebuilding industry? J
Constr Eng Manag 137(12):1027–1035. https://doi.org/10.1061/(ASCE)CO.1943-7862.0000396
Nakahara Y, Sato H, Nii KP (1995) The complete Japanese joinery. Hartley & Marks, Vancouver
Negro A (1996) Il giardino dipinto del Cardinal Borghese: Paolo Bril e Guido Reni nel Palazzo
Rospigliosi Pallavicini a Roma. Àrgos, Rome
Neisser U (1967) Cognitive psychology. Appleton Century Crofts, New York
Nestorovic M (1987) Metallic integrally tensioned (Tensegrity) Cupola. In: Topping HV (ed) Pro-
ceedings of international conference on the design and construction of non-conventional struc-
tures. Civil-Comp Ltd., Edinburgh
Ohshima T, Igarashi T, Mitani J, Tanaka H (2013) Digital fabrication-volume 1-computation and
performance-eCAADe 31|. In: Proceedings of eCAADe 2013 education and research in computer
aided architectural design in Europe. eCAADe, Delft, pp 693–702
Oppenheim IJ, Williams WO (1997) Mechanics of tensegrity prisms. In: Proceedings of the 14th
international symposium on automation & robotics in construction. Pittsburgh
Oxman R (2006) Theory and design in the first digital age. Des Stud 27(3):229–265. https://doi.
org/10.1016/J.DESTUD.2005.11.002
Oxman R (2009) Performative design: a performance-based model of digital architectural design.
Environ Plan 36(6):1026–1037. https://doi.org/10.1068/b34149
Oxman R, Oxman R (2010a) Introduction. In: Oxman R, Oxman R (eds) The new structuralism:
design, engineering and architectural technologies. Wiley, pp 14–24
Oxman R, Oxman R (2010b) The new structuralism: design, engineering and architectural tech-
nologies. Wiley, New York
Oxman R, Oxman R (eds) (2014) Theories of the digital in architecture. Routledge, London
Page IC, Norman D (2014) Prefabrication and standardisation potential in buildings (SR 312).
Branz, Wellington
Paoletti I (2018) Informed architecture: computational strategies in architectural design. In: Hem-
merling M, Cocchiarella L (eds) Informed architecture: computational strategies in architectural
design. Springer, pp 77–88
Paris L (2009) Stereotomia del legno. In: Migliari R (ed) Geometria Descrittiva. CittàStudi, Torino,
pp 562–588
Paronesso A, Passera R (2004) The cloud of Yverdon. In: Motro R (ed) IASS symposium 2004.
IASS Secretariat, Madrid, pp 184–185
Pastore N (1971) Selective history of theories of visual perception 1650-1950. Oxford University
Press, Toronto
Pawlyn M (2011) Biomimicry in architecture. RIBA Publishing, London
Pearce P (1979) Structure in nature is a strategy for design. MIT Press, Cambridge
Pearson M (2011) Generative art: a practical guide using processing. Manning, Greenwich
Pecchinenda G (2010) Videogiochi e cultura della simulazione. La nascita dell’ «homo game».
Laterza, Roma - Bari
Pedersen Zari M (2015) Ecosystem processes for biomimetic architectural and urban design. Archi-
tectural Sci Rev 58(2):106–119. https://doi.org/10.1080/00038628.2014.968086
Pedretti C (1988) Leonardo architetto. Electa, Milano
Peña DM, Llorens I, Sastre R (2010) Application of the tensegrity principles on tensile textile
constructions. Int J Space Struct 25(1):57–67. https://doi.org/10.1260/0266-3511.25.1.57
Peri G (1884) Applicazioni della geometria descrittiva alle ombre, alla prospettiva lineare e aerea,
al taglio delle pietre e del legname (G. Belotti, ed.). Firenze
Pérouse de Montclos JM (1985) La vis de Saint-Gilles et l’escalier suspendu dans l’architecture
française du XVI siècle in L’escalier dans l’architecture de la Renaissance (Répertoire d’Art et
d’Archéologie, Ed.). Picard
Pilewski JL, Martin BA (1991) Effects of monocular versus binocular viewing in the ames distorted-
room illusion. Percept Mot Skills 72(1):306. https://doi.org/10.2466/pms.1991.72.1.306
Pillet J (1887) Traité de stéréotomie (C. Delagra). Paris

Pine BJ, Slessor C (1999) Mass customization: the new frontier in business competition. Harvard
Business School, Boston
Popov V, Juocevicius V, Migilinskas D, Ustinovichius L, Mikalauskas S (2010) The use of a vir-
tual building design and construction model for developing an effective project concept in 5D
environment. Autom Constr 19(3):357–367. https://doi.org/10.1016/J.AUTCON.2009.12.005
Popovic Larsen O (2003) Conceptual structural design: bridging the gap between architects and
engineers. Thomas Telford, London
Popovic Larsen O (2008) Reciprocal frame architecture. Architectural Press, Oxford
Popper KR (1957) The poverty of historicism. Beacon Press, Boston, pp 123–124
Pottmann H, Huang Q, Deng B, Schiftner A, Kilian M, Guibas L et al (2010) Geodesic patterns.
ACM Trans Graph 29(4):1. https://doi.org/10.1145/1778765.1778780
Pugh A (1976a) An introduction to tensegrity. University of California Press. Retrieved
from https://books.google.it/books?id=McEOfJu3NQAC&dq=A.+Pugh.+An+Introduction+to+
Tensegrity.+University+of+California+Press,+1976.&lr=&hl=it&source=gbs_navlinks_s
Pugh A (1976b) Polyhedra: a visual approach. University of California Press, Berkeley
Rajanen M, Iivari N (2015) Power, empowerment and open source usability. In: Proceedings of the
33rd annual ACM conference on human factors in computing systems—CHI ’15, pp 3413–3422.
https://doi.org/10.1145/2702123.2702441
Ramachandran VS (1990) Visual perception in people and machines. In: Blake A, Troscianko T
(eds) AI and the eye. Wiley, London
Ratti C, Claudel M (2015) Open source architecture. Thames & Hudson, New York
Rechenberg I (1965) Cybernetic solution path of an experimental problem. Royal AircraftEstab-
lishment, Ministry of Aviation, Farnborough Hants
Reichert S, Menges A, Correa D (2015) Meteorosensitive architecture: biomimetic building skins
based on materially embedded and hygroscopically enabled responsiveness. Comput Aided Des
60:50–69. https://doi.org/10.1016/J.CAD.2014.02.010
Renner G, Ekárt A (2003) Genetic algorithms in computer aided design. Comput Aided Des
35(8):709–726. https://doi.org/10.1016/S0010-4485(03)00003-4
Reyssat E, Mahadevan L (2009) Hygromorphs: from pine cones to biomimetic bilayers. J R Soc
Interface 6(39):951–957. https://doi.org/10.1098/rsif.2009.0184
Richard R-B (2005) Industrialised building systems: reproduction before automation and robotics.
Autom Constr 14(4):442–451. https://doi.org/10.1016/J.AUTCON.2004.09.009
Riether G, Wit AJ (2016) Underwood Pavilion. A parametric tensegrity structure. In: Adriaenssens
S, Gramazio F, Kohler M, Menges A, Pauly M (eds) Advances in architectural geometry 2016.
Vdf Hochschulverlag AG an der ETH Zürich, pp 114–203
Rippmann M, Block P (2011) Digital stereotomy: Voussoir geometry for freeform masonry-like
vaults informed by structural and fabrication constraints. In: Proceedings of the IABSE-IASS
symposium 2011. IABSE, Boston
Roudsari MS, Pak M (2013). Ladybug: a parametric environmental plugin for Grasshopper to help
designers create an environmentally-conscious design. In: Proceedings of the 13th international
IBPSA conference. Lyon, pp 3128–3135
Rüggeberg M, Burgert I (2015) Bio-inspired wooden actuators for large scale applications. PLoS
ONE 10(4):e0120718. https://doi.org/10.1371/journal.pone.0120718
Runeson S (1988) The distorted room illusion, equivalent configurations, and the specificity of
static optic arrays. J Exp Psychol Hum Percept Perform 14(2):295–304. https://doi.org/10.1037/
0096-1523.14.2.295
Rutten D (2013) Galapagos: on the logic and limitations of generic solvers. Architectural Des
83(2):132–135. https://doi.org/10.1002/ad.1568
Sadao S (1996) Fuller on tensegrity. Int J Space Struct 11(1–2):37–42. https://doi.org/10.1177/
026635119601-206
Saint Aubin JP (1994) es enjeux architecturaux de la didactique stéréotomique de Desargues. In:

Dhombres JG, Sakarovitch J (eds) Desargues en son temps. Librairie scientifique A. Blanchard,
Paris
Sakamoto T, Ferré A (2008) From control to design: parametric/algorithmic architecture. Actar-D,
New York
Salvador F, De Holan PM, Piller F (2009) Cracking the code of mass customization. MIT Sloan
Manag Rev 50(3):71–79
Salvatore M (2012) La stereotomia scientifica in Amédée François Frézier. Prodromi della geometria
descrittiva nella scienza del taglio delle pietre. Università degli Studi di Firenze
Sanabria SL (1989) From gothic to renaissance stereotomy: The design methods of Philibert de
l’Orme and Alonso de Vandelvira. Technol Cult 30(2):266. https://doi.org/10.2307/3105105
Sanjurjo Alvarez A (2010) La Vis-de-Saint-Gilles: analyse du modèle dans les traités de coupe des
pierres et de son influence sur les traités espagnols de l’âge moderne. In: Carvais R, Guillerme A,
Nègre J, Valérie Sakarovitch (eds) Édifice & Artifice. Histoires Constructives. Editions Picard,
Paris, pp 679–689
Sass L (2006) A Wood Frame grammar: a generative system for digital fabrication. Int J Architectural
Comput 4(1):51–67. https://doi.org/10.1260/147807706777008920
Sass L (2012) Direct building manufacturing of homes with digital fabrication. In: Gu N, Wan X
(eds) Computational design methods and technologies: applications in CAD, CAM, and CAE
education. IGI Global, New York
Sass L, Botha M (2006) The instant house: a model of design production with digital fabrication.
Int J Architectural Comput 4(4):109–123. https://doi.org/10.1260/147807706779399015
Sass L, Oxman R (2006) Materializing design: the implications of rapid prototyping in digital
design. Des Stud 27(3):325–355. https://doi.org/10.1016/J.DESTUD.2005.11.009
Scheurer F (2010) Materialising complexity. Architectural Des 80(4):86–93. https://doi.org/10.
1002/ad.1111
Schlaich M (2004) The Messeturm in rostock: a tensegrity tower. J Int Assoc Shell Spatial Struct
42(2):93–98
Schumacher P (2011) The autopoiesis of architecture: a new framework for architecture. Wiley,
West Sessex
Seike K (1986) The art of Japanese Joinery. Weatherhill, Tankosha
Self M, Vercruysse E (2017) Infinite variations, radical strategies. In: Menges A, Sheil B, Glynn R,
Skavara M (eds) Fabricate 2017 conference proceedings. UCL Press, London, pp 30–35
Sheil B (2005) Design through making: an introduction. Architectural Des 75(4):5–12. https://doi.
org/10.1002/ad.97
Skelton RE, de Oliveira MC (2009) Tensegrity systems. Springer Science & Business Media,
London
Skelton RE, Adhikari R, Pinaud JP, Chan W, Helton JW (2002) An introduction to the mechanics
of tensegrity structures. In: Proceedings of the 40th IEEE conference on decision and control
(Cat. No.01CH37228), vol 5, pp 4254–4259. IEEE, Orlando. https://doi.org/10.1109/CDC.2001.
980861
Snelson K (1996) Snelson on the tensegrity invention. Int J Space Struct 11(1–2):43–48. https://
doi.org/10.1177/026635119601-207
Snelson K (2012) The art of tensegrity. Int J Space Struct 27(2–3):71–80. https://doi.org/10.1260/
0266-3511.27.2-3.71
Solomon J, Vouga E, Wardetzky M, Grinspun E (2012) Flexible developable surfaces. Comput
Graph Forum 31(5):1567–1576. https://doi.org/10.1111/j.1467-8659.2012.03162.x
Son S, Fitriani H, Kim JT, Go S, Kim S (2017) Mathematical algorithms of patterns for free-
form panels. In: Proceedings of the 2nd world congress on civil, structural, and environmental
engineering (CSEE’17), vol 101, pp 2371–5294. CSENM,Barcelona. https://doi.org/10.11159/
icsenm17.101
Song K, Yeom E, Seo S-J, Kim K, Kim H, Lim J-H, Joon Lee S (2015) Journey of water in pine
cones. Sci Rep 5(1):9963. https://doi.org/10.1038/srep09963
Stephan S, Klimke H (2004) The making of a tensegrity tower. In: IASS 2004 symposium, inter-
national association for shell and spatial structures. Editions de l’Espérou, Montpellier
Svilans T, Poinet P, Tamke M, Ramsgaard Thomsen M (2018) A multi-scalar approach for the
modelling and fabrication of free-form glue-laminated timber structures. In: Humanizing digital
reality. Springer, Singapore, pp 247–257. https://doi.org/10.1007/978-981-10-6611-5_22
Tachi T (2012) Interactive freeform design of tensegrity. In: Hesselgren L, Sharma S, Wallner J,
Baldassini N, Bompas P, Raynaud J (eds) Proceedings of the advances in architectural geometry
conference. Springer, Wien, pp 259–268. https://doi.org/10.1007/978-3-7091-1251-9
Tamborero L (2006) The “Vis Saint-Gilles”, symbol of compromise between practice and science.
In: Proceedings of the second international congress on construction history. Construction History
Society, Cambridge, pp 3025–3040
Tamke M, Thomsen MR (2009) Digital wood craft. CAAD Futures
Teghtsoonian M (1965) The judgment of size. Am J Psychol 78(3):392. https://doi.org/10.2307/
1420573
Thompson DW (1917) On growth and form. Cambridge University Press, Cambridge
Thönnissen U (2014) A form-finding instrument for reciprocal structures. Nexus Netw J
16(1):89–107. https://doi.org/10.1007/s00004-014-0172-1
Thönnissen U, Werenfels N (2011) Reciprocal frames—teaching experiences. Int J Space Struct
26(4):369–371. https://doi.org/10.1260/0266-3511.26.4.369
Tibert G (2008) Advances in the optimization and form-finding of tensegrity structures. In: Proceed-
ings of the 6th international conference on computation of shell & spatial structures. Internet-First
University Press, Stockholm
Tibert AG, Pellegrino S (2011) Review of form-finding methods for tensegrity structures. Int J
Space Struct 26(3):241–255. https://doi.org/10.1260/0266-3511.26.3.241
Tran TM (2002) Reverse displacement analysis for tensegrity structures. University of Florida.
Retrieved from http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.15.2315&rep=rep1&
type=pdf
Trevisan C (2011) Per la storia della stereotomia: geometrie, metodi e construzioni. Aracne, Roma
Tsigkari M, Angelos C, Joyce SC, Davis A, Feng S, Aish F (2013) Integrated design in the simulation
process. In: Proceedings of the symposium on simulation for architecture & urban design. Society
for Computer Simulation International, San Diego
Turan M (2009) Reconstructing the balloon frame: a study in the history of architectonics. METU
J 2(26):175–209. https://doi.org/10.4305/METU.JFA.2009.2.10
Turrin M, von Buelow P, Stouffs R (2011) Design explorations of performance driven geome-
try in architectural design using parametric modeling and genetic algorithms. Adv Eng Inform
25(4):656–675. https://doi.org/10.1016/J.AEI.2011.07.009
Ugolev BN (2014) Wood as a natural smart material. Wood Sci Technol 48(3):553–568. https://doi.
org/10.1007/s00226-013-0611-2
Vailati C, Bachtiar E, Hass P, Burgert I, Rüggeberg M (2018) An autonomous shading system
based on coupled wood bilayer elements. Energy Build 158:1013–1022. https://doi.org/10.1016/
J.ENBUILD.2017.10.042
Van Telgen MV, Snijder H, Habraken APH, Beetz J (2013) Parametric design and calculation
of circular and elliptical tensegrity domes. In: IASS 2013 Wroclaw: ‘Beyond the Limits of
Man’—structural morphology. International Association for Shell and Spatial Structures (IASS),
Eindhoven, pp 1–8
Vassart N, Motro R (1999) Multiparametered formfinding method: application to tensegrity systems.
Int J Space Struct 14(2):147–154. https://doi.org/10.1260/0266351991494768
Vattam S, Helms ME, Goel AK (2007) Biologically-inspired innovation in engineering design: a
cognitive study, Atlanta. Retrieved from http://hdl.handle.net/1853/14346
Vierlinger R (2013) A framework for flexible search and optimization in parametric
design. In: Rethinking prototyping—proceedings of the design modelling symposium.
Berlin. Retrieved from https://www.researchgate.net/profile/Robert_Vierlinger/publication/
283073197_A_Framework_for_Flexible_Search_and_Optimization_in_Parametric_Design/
links/5628c83308ae04c2aeaeb6cb.pdf
Vierlinger R (2015) Towards ai drawing agents. In: Ramsgaard Thomsen M, Tamke M, Gengnagel
F, Faircloth C, Scheurer B (eds) Modelling behaviour. Springer International Publishing, Cham,
pp 357–369. https://doi.org/10.1007/978-3-319-24208-8_30
Vierlinger R, Bollinger K (2014) Accommodating change in parametric design. In: Proceedings of
ACADIA 2014. Association for Computer-Aided Design in Architecture, Los Angeles
Vierlinger R, Zimmel C (2015) Octopus. Retrieved 27 Aug 2018, from https://www.grasshopper3d.
com/group/octopus?groupUrl=octopus&id=2985220%3AGroup%3A742529&page=3
Vincent J (2009) Biomimetic patterns in architectural design. Architectural Des 79(6):74–81. https://
doi.org/10.1002/ad.982
Vogel JM, Teghtsoonian M (1972) The effects of perspective alterations on apparent size and
distance scales. Percept Psychophys 11(4):294–298. https://doi.org/10.3758/BF03210382
Wang B (2004) Free-standing tension structures: from tensegrity systems to cable-strut systems.
Spon Press
Weber S (2005) The success of open source. Harvard University Press, Cambridge
Weinand Y (2016) Advanced timber structures: architectural designs and digital dimensioning.
Birkhäuser, Basel
Wester T (2002) Nature teaching structures. Int J Space Struct 17(2–3):135–147. https://doi.org/10.
1260/026635102320321789
Wikipedia (2015) The poverty of historicism (2002nd ed.). Routledge, London
Willis D, Woodward T (2010) Diminishing difficulty: mass customisation and the digital production
of architecture. In Corser R (ed) Fabricating architecture: selected readings in digital design and
manufacturing. Princeton Architectural Press, pp 184–208
Willmann J, Gramazio F, Kohler M (2017) New paradigms of the automatic: robotic timber construc-
tion in architecture. In: Menges A, Schwinn T, Krieg OD (eds) Advancing wood architecture. A
computational approach. Routledge, London, pp 13–28. https://doi.org/10.4324/9781315678825-
11
Wood DM, Correa D, Krieg OD, Menges A (2016) Material computation—4D timber construction:
towards building-scale hygroscopic actuated, self-constructing timber surfaces. Int J Architectural
Comput 14(1):49–62. https://doi.org/10.1177/1478077115625522
Wood D, Vailati C, Menges A, Rüggeberg M (2018) Hygroscopically actuated wood elements for
weather responsive and self-forming building parts—facilitating upscaling and complex shape
changes. Constr Build Mater 165:782–791. https://doi.org/10.1016/J.CONBUILDMAT.2017.12.
134
Woodbury R (2010) Elements of parametric design. Routledge, London
Zarrinmehr S, Akleman E, Ettehad M, Kalantar N, Borhani A (2017) Kerfing with generalized
2D meander-patterns: conversion of planar rigid panels into locally-flexible panels with stiffness
control. In: Çagdas G, Özkar M, Gül LF, Gürer E (eds) Future Trajectories of computation in
design. Cenkler Matbaa, Istanbul, pp 276–293
Zhang L, Maurin B, Motro R (2006) Form-finding of nonregular tensegrity systems. J Struct Eng
132(9):1435–1440. https://doi.org/10.1061/(ASCE)0733-9445(2006)132:9(1435)
Zipkin P (2001) The limits of mass customization. MIT Sloan Manag Rev 42(3):81–87.
https://doi.org/ISSN: 1532-9194
Zitzler E, Laumanns M, Thiele L, Zitzler E (2001) SPEA2: Improving the strength pareto evolu-
tionary algorithm. Zurigo. https://doi.org/10.3929/ethz-a-004284029
Zwerger K (1997) Wood and wood joints: building traditions of Europe, Japan and China.
Birkhäuser, Basel
Beyond Form Definition: Material
Informed Digital Fabrication in Timber
Construction
David Correa, Oliver David Krieg and AnnaLisa Meyboom
Abstract This paper introduces a series of prototypes investigating a new

architectural language in wood that is driven by a critical approach to recent technical
developments in design, fabrication and material. Although wood is slowly being rec-
ognized as an advanced material for future construction due to its high performance
and sustainable nature, its differentiated and unpredictable material characteristics
have not only been progressively overlooked, but even been viewed as a negative
attribute. Wood’s varied dimensional range has been addressed through standardiza-
tion, its heterogeneous fiber structure ground and reconstituted into homogeneous
composites, and finally its complex aesthetic quality has even been caricaturized into
a skin-deep plastic-wood veneer texture. This paper seeks to extend research on the
implications of advanced robotic fabrication and its integration into design processes
that also integrate cross-disciplinary knowledge into architectural software. As inno-
vation in technology enables architects and engineers to engage with the complexities
of the material, the potential of wood is becoming accessible, leading to a new mate-
rial language. Through a series of full scale, robotically fabricated design prototypes,
the material performance of wood is investigated as a driver for form; its fabrication
and hygroscopic performance as a driver for assembly, and more importantly, the
entire design-to-fabrication-process as a method for investigation into innovation
and the structural and architectural potential of future wood.
Keywords Robotic fabrication · Computational design · Material computation

Elastic bending · Timber structures · Complex surfaces · Wood joints
D. Correa (B)
School of Architecture, University of Waterloo, Waterloo, Canada
e-mail: david.correa@uwaterloo.ca
O. D. Krieg
Institute for Computational Design and Construction, University of Stuttgart, Stuttgart, Germany
A. Meyboom
School of Architecture and Landscape Architecture, University of British Columbia,
Vancouver, Canada

62 D. Correa et al.
1 Introduction
Innovation in material results in innovation in architecture. Technology both supports

the development of new materials, and changes how we use traditional ones. In
parallel, these innovations bring about a change in how buildings are made, and in
many cases, how buildings are designed and conceived.
Developments in steel manufacturing, for example, revolutionized building con-
struction in the 19th Century—its strength increased significantly with new smelting
technologies and it, along with advances in elevator technology, freed the build-
ing from load bearing walls and allowed new forms such as high-rise buildings.
Reinforced concrete allowed for a similar revolution in the 20th century—it was
malleable, could take tension and allowed new forms to be built due to the casting
process. Designers such as Maillard, Nervi, Fisac, Corbusier, Eero Saarinen, and
Oscar Niemeyer took advantage of this material innovation and were foundational
in creating a new architectural language for the time, which engaged the properties
of the material and the innovations of the construction process—clean, curved, large
scale structural forms—a language recognized as representing contemporary design
in the mid-1900s.
Steel today is produced either as linear or sheet elements and requires a significant
amount of energy in the manufacturing process (Gordon 2003). Likewise, concrete
is very energy intensive (Gordon 2003) and very heavy—the weight of concrete
as dead load is often equivalent to the live load it can support. With the building
industry being one of the biggest sources of Green House Gas emissions (COM
2014), sustainable building materials and building techniques are most important for a
sustainable future. Especially in developing countries, where not only the production
of materials but also the production of energy itself poses an ecological challenge,
pollution by the construction industry has become a serious problem. As such, these
now common materials, while often giving yield to innovation in the past, face new
economic and environmental challenges that question their long-term feasibility.
If we were to put forward the criteria of the material of the future from the
viewpoint that we hold today, we would require it to be sustainable, lightweight,
structural, multifunctional, compelling, and receptive to parametrically driven forms
that are enabled by digital design. The material, answering to many of these criteria,
is wood. Wood can perform in tension as well as compression due to its natural fiber
structure. It can be employed as structure, finish or furniture and as a final attribute, it
is generally considered aesthetically pleasing because of its organic grain and depth
of tone and color. In its natural occurrence wood already exhibits high structural
strength, a positive carbon footprint (Kolb 2008), and very low embodied energy
(Alcorn 1996). Moreover, its local availability in moderate climate zones makes the
material particularly suitable for the development of more sustainable construction
methods (Krieg et al. 2015). It comes to no surprise that the general interest in wood
architecture has grown over the last decade.
In the long history of architecture, wood plays an almost continuously domi-
nant role as a building material. Buildings that still stand today showcase intricate
Beyond Form Definition … 63
construction systems made from posts and beams—the most famous of these being
Japanese and Chinese traditional temples, although this type of construction can be
observed in any civilization that had access to forests in their time. Structures made
from linear elements required a large amount of material knowledge and technical
skill to source and carve them, resulting in the emergence of extremely skilled crafts-
men and a culture of wood fabrication that persisted until the early 1900s. A shift from
local knowledge and local production to industrialization and mass production for a
global economy also brought a shift from value-added products to value-engineered
standardized components. This shift also precipitated a disciplinary change, from an
integrated design-to-fabrication model based on the mastery of a craft to the split
between design, engineering and manufacturing. In this new context, each step is sub-
sequently compartmentalized into highly specialized but also further disconnected
disciplinary domains; design is only concerned with form and function, engineering
with material and manufacturing, management with production and supply chain,
trades and operators with execution and assembly. Multiple hierarchies are concep-
tualized within this model, which limits the free movement of knowledge, expertise
and ultimately hinders innovation.
Research on innovation has pointed out that the building industry is on par with
other industries in terms of “incremental innovations”, which refers to minor changes
in the product. However, the same research also points out that the building indus-
try is a laggard adopter when it comes to “systemic innovations”, which refers to
product and process innovations that require multiple and inter-dependent industries
to change their processes (Taylor 2006, 16). Systemic innovation diffuses slowly in
project-based industries such as in today’s construction industry (Taylor 2006, 9),
as it requires multiple organizations acting together to implement the change; this is
known as an interorganizational network (Taylor and Levitt 2005, 9). The reason for
this is that “[w]hen organizational variety is high and the span of a systemic innova-
tion increases to impact two or more specialist firms, extra coordination is required
for inter-organizational knowledge to flow and accumulate” (Taylor and Levitt 2005,
15). Systemic innovations such as the one being discussed here—a major change in
the conceptual approach to wood design—would thus require not only an innovation
in the actual development of the technology but also a re-examination of the process
of design and building in order to avoid the friction inherent in the conventional
building industry of today.
Viewing this process in terms of the type of innovation—by use of Slaughter’s def-
initions of innovation—the types of innovations in these projects is radical in the form
of the robotic fabrication process. Particularly, the direct integration of design input
within the automation poses an entirely new model to the building industry. Slaughter
states that a radical innovation is one which “creates a new way of understanding a
phenomenon and formulating approaches through which to solve problems” and that
“All previous linkages and interactions may be irrelevant for a radical innovation, not
only with respect to the systems, but also the ties among organizations.” (Slaughter
1998, 227) Structural steel, she states, is considered a radical innovation because “a
whole new industry of steel manufacturing and fabrication emerged, as well as new
64 D. Correa et al.
components and systems linked to the new structural forms and systems.” (Slaughter
1998, 228).
In these terms, then, the radical innovation of steel also affected the various orga-
nizations which create systemic innovation, because as mentioned, the construction
industry is an interorganizational network. However, it was not considered systemic,
possibly because the innovation was outside the system and not derived within the
system. It is unclear whether the innovation described and anticipated by this paper
would be considered radical or systemic—it could be systemic because, if imple-
mented, it would likely be implemented by a partnership between architects, engi-
neers and fabricators. Still, the industrial robot applied to the wood fabrication indus-
try may be considered a radical innovation partly because of the implications of the
ability to change the workflow—taking the fabrication control back to the design
stage.
Winch has discussed the very slow innovation in the construction industry and
describes the problem as follows: “First, the systems integrator role is shared between
the principal architect/engineer and the principal contractor. Thus, construction typ-
ically has two separate systems integrators—one at the design stage and one at the
construction stage. Secondly, the fragmentation of the professional bodies in con-
struction has weakened their ability to act as honest brokers of innovations as they
typically threaten the interests of one or other amongst them.” (Winch 1998, 270)
“The development of separate and specialized design professions helped to bring
about divergent interests between engineers and architects and the organization of
design had no built-in mechanism to bring about innovation in the absence of exter-
nal stimuli from clients or materials’ producers.” (Gann 2000, 43) One method of
overcoming this inter-organizational friction is to adopt an organizational strategy of
integration which would reduce “organizational variety” and increase the “capability
to adopt and diffuse an innovation.” (Taylor and Levitt 2005, 9).
The process discussed in this paper operates under a “single system integrator”,
as per Winch, both design and construction are considered under the same umbrella
of execution, thereby reducing the systemic friction to innovation. By embedding
design, engineering and rapid prototyping within the manufacturing floor a more
horizontal exchange of information is made possible; an “Open-Source Workshop”,
where multiple disciplinary sources of expertise are engaged. It is only through this
type of “single integrator” framework that it may be possible to re-engage material
informed fabrication and assembly techniques. Professional and academic outcomes
of this type of approach are beginning to emerge more and more. Designers such
as Shigeru Ban (Sunny Hills Desert Shop, Triangle House, Aspen Art Museum,
Tamedia Headquarters in Zürich) and Kengo Kuma (Prosto Museum and Research
Centre, Daizaifu Temanga Shrine Starbucks interior) have experimented with this
integrated approach and have changed traditional wood fabrication into a much more
expressive modern architecture. For example, traditional joints through digital and
robotic fabrication, allow not only for the most intricate details, but also for the
possibility to precisely orient building elements. More comprehensively, this type
of systemic innovation across interorganizational networks has been demonstrated
by the group of firms Design to Production in collaboration with Blumer Lehmann,
and SJB Kempter Fitze (Antemann 2013) in order to realize complex projects by
SANAA, UN Studio, Zaha Hadid, Shigeru Ban, and Renzo Piano. This process of
decompartmentalization begins to open the door for a new and cross-sectionally
informed design language to emerge.
2 Patterns of Innovation in Materials and Their

Consequences
By looking at patterns of innovation in materials in the past, it is possible to strengthen

the argument that innovations in wood today have the potential to create a large change
in architectural design and, in fact, in the building industry generally. As Antoine
Picon has noted, “In the case of architecture and construction, as in many other
areas of material culture, innovation comes about at the intersection of technical and
social issues that need to be decoded.” (Picon 2010, 51) This section will look at
both what created the innovation in material and what changes therefore resulted in
building processes. The argument presented here is that the innovation in material is
concurrent and catalytic to other changes in the design and construction industries.
The prototypes described in Advanced Timber Concepts and the Design of Furniture
and Structures provide evidence of this shift, positioned at the intersection of tech-
nical, material and social developments. The projects offer an insight into a material
revolution on a larger scale rather than simply an innovation in material (Boake
2012).
2.1 Steel
The development and use of iron shows how the change in technology results in a
change in design and form. When iron was first used in construction, its building
elements and joints resembled those made from timber. Only once a better under-
standing of the differences in the material characteristics was established, joints and
building elements moved from an imitation of timber construction to a more appro-
priate technique for iron, such as bolting and riveting, or cast lugs and flanges (Peters
1996, 38). Secondly, “[t]he primacy of connection technology in iron changed the
way designers thought about construction.” Since iron has to be produced in a fac-
tory—it is a “process of assembling pre-fabricated components with prefabricated
connectors” (Peters 1996, 42). This was a change in thinking about how construc-
tion was carried out, as previously materials were manufactured or easily worked
into parts on site. This change was not instantaneous but took several decades to
occur (Peters 1996, 42). Compared to the initially used plain iron, these changes
became even more pronounced when wrought iron was developed, as it allowed a
symmetrical cross section with equal flanges because of its equal strength in tension
66 D. Correa et al.
and compression. As the knowledge of iron-related construction was developing,

engineers experienced the phenomenon of fatigue fracture in iron parts in mills. This
observation and the following efforts into the chemical makeup of iron led into the
development of the field of material science (Peters 1996, 44).
Steel further played a role in developing engineering as an academic field. Richard
Turner, in designing the Kew Palm House, developed a methodology which “demon-
strated that it was possible to fulfill contradictory criteria in a single detail by decou-
pling the problems, solving each on its own serially, and then reuniting the solutions
to form a component subset in a construction system” (Peters 1996, 220). Turner’s
“clear hierarchy of structural members and their relationships advanced technolog-
ical thought in building” (Peters 1996, 220). Like Turner, Gustav Eiffel segregated
issues to solve the more complex problems; whereas Turner detailed the solutions to
a specific building, Eiffel developed solutions which used an open system to build
any iron structure (Peters 1996, 266). The Eiffel Tower and the Galerie des Machines
at the 1889 Paris exhibition were demonstrations of this type of technological inno-
vation, defining structural systems still being used today.
While pre-designed and pre-engineered steel structures shape every city today,
both steel, and its predecessor iron, had to overcome significant uncertainties about
its application in structures and architecture at the outset. This uncertainty was not
only in how to approach its visual language but also in how to ascertain its structural
requirements for configuration. Steel was high strength and little was needed relative
to a stone or wood structure, but the question of its appearance and how to use it
visually was a conundrum as it did not fit in with the predominance of historicism
in architectural expression at the time (Rinke and Schwartz 2010b, 17; Picon 2010,
52). The public did not generally consider structural frames to be beautiful; iron
was “vulgar.” “Honesty” and “decency” seemed almost to be opposites. Even table
and piano legs were deemed indecent and had to be “decorously draped in cloth or
sheathed modestly in wooden casing” (Peters 1996, 275). The material uncertainty
and the attempt to define its strength with numbers as opposed to previous experience
led to an advancement in engineering analysis methodologies (Rinke and Schwartz
2010a, 67).
Rinke and Schwartz point out that although the developments in iron “did not
directly trigger the development of structural engineering and the process of sepa-
ration in the disciplines, it was at least a catalyst. […] the novel, homogeneous and
formable building material represented the promise of new foundational theories and
precise processes of calculation: a comprehensive degree of control and a redefini-
tion of the load-bearing structure. The increasing efforts to make the understanding
of the load-bearing structure as well as the new building material more scientific
had a radical influence on construction.” (Rinke and Schwartz 2010b, 19). Antoine
Picon has also commented on its radical influence on how we design: “the advent
of iron precipitated the abandonment of the Vitruvian tradition and the definitive
split between the professions of architect and engineer.” (Picon 2010, 51). What this
points to is that the innovation in material led to, or at least greatly accelerated, a
change in how we design.
2.2 Concrete
Compared to steel, concrete maintains the monolithic and massive characteristics of

stone. However, concrete had a more difficult time being accepted as an architectural
material when compared to steel. Although it was developed in the mid-19th century,
it was not until reinforcing steel technology developed and analysis of the hybrid
material was perfected, that it began to be widely experimented with in architecture.
The development of concrete as a modern building material involved material
innovation over the course of the 19th century in France by François Cointeraux,
Jean-Auguste LeBrun and finally François Coignet, who worked through mixes,
methods and aggregate types to develop the strength of the material. The material,
however, was not widely accepted—reluctance in uptake was due to uncertainty
about both its structural performance and its fire resistance. The structural history of
concrete throughout the 1800s is riddled with failures. These were in large part due
to inconsistencies in the quality of the material due to inconsistent workmanship and
a lack of quality assurance processes. The result of this slowed down uptake led to
the temporary end of the development in France for 20 years when Coignet went out
of business in 1872 (Collins 1959, Chapter 1: Beton). The enthusiasm was then taken
up in England and America for the next 20 years with the invention of formwork and
an early form of reinforcing. However, a few failures also led to reluctance to use
the material in these regions, enforced by the concern that the iron embedded in the
concrete would rust. This concern did not disappear until well into the 20th century.
Along with this resistance to use due to very practical reasons, however, was the
appearance of the material. In 1870 concrete was honeycombed and motley in color.
Victorians could not develop an affinity for the material and so the material was left
to the engineers, who developed it further for large scale civil works (Collins 1959,
Chapter 2: Concrete).
Iron reinforcing was the final innovation that propelled the use of concrete as we
know it today. In 1873 in America, Thaddeus Hyatt was looking for a solution to fire
proof iron, which had become known as more dangerous in a fire than heavy timber.
The innovation he patented was the use of a steel beam embedded in concrete. He
enlisted a retired railway engineer, Thomas Rickett, to help with the engineering
calculations required. They developed the modern calculation approach to the bend-
ing in a reinforced concrete beam. They tested beams to failure both under load and
under fire conditions and proved that steel and concrete expand at the same rate under
heat. Furthermore, they proved that the concrete protects the steel, leading the way
for use of reinforced concrete in hotels and many other structures. The results were
published in 1877. By 1905, there were standard textbooks on structural reinforced
concrete design. The standardization of both the material composition and the place-
ment of the reinforcing finally led to structural reliability, material use efficiency,
and its larger scale adoption, although full scale load testing was still required for
most projects (Elliott 1992, 183–84). The ability to engineer thin shell structures
in concrete advanced and by 1927, the Market Hall at Frankfurt am Main spanning
120 ft with 3–4-in.-thick concrete shells demonstrated this capacity. At this point in
68 D. Correa et al.
time, material efficiency was key, which made the use of vaulted structures desirable,
especially in World War II when building materials were scarce.
The innovation in concrete was originally meant as a cheaper alternative to stone.
The primary driver throughout its period of innovation was cost. Concrete did not
develop as quickly as steel in North America due to its relatively higher cost in
America than Europe. But once the material was recognized as reliable and fireproof
as well as inexpensive, its broad-based use increased. Its social acceptance, similar
to steel, lagged the material innovation.
3 Current Innovation Relating to Wood
As can be seen from the precedents of both concrete and steel, the innovation in a
material application generally has a series of both technical and functional challenges
that impact their wide scale adoption. These include change in societal expectations
as well as developments in design, construction, engineering and analysis. However,
these incremental developments in material and technologies can lead to more radical
or systemic innovations. In some instances, a radical change may occur at the inter-
section of social and technical needs, as it the case with the Market Hall in Frankfurt,
in others, it may open the door for a new design language in architecture. Wood is of
particular interest in this discussion as it is not newly invented, but instead, it is one
of the oldest building materials. At the current intersection of much larger cultural
and social challenges, the sustainable aspects as well as innovations relating to the
sourcing, processing, and use of the material have renewed the interest of consumers
and municipalities, as well as architects and engineers. Due to the long history of tra-
ditional wood construction it may be necessary to challenge pre-conceived notions;
a change in the language of wood architecture is arguably more difficult than the
invention of a new language for a new building material. That said, its accessibil-
ity as a resource combined with the economy of scale makes wood a very valuable
focus for continous innovation. It is therefore worth examining in more detail what
innovations in wood might be and how to take advantage of the material within the
framework of a new technology. Through this lens, the presented projects begin to
ask the question of what the future of building in wood may look like.
3.1 Innovation—Customized Timber Products
Sustainability concerns have brought a renewed interest in timber construction and

with it, the development of new timber products. Plywood, Laminated Veneer Lum-
ber, Glulam, and CLT have been developed to increase homogeneity and create
dimensional stability in otherwise anisotropic materials. Each product and each sub-
category of it, while exhibiting general enough characteristics to be used in very
different types of buildings, are made, and subsequently optimized, for a particular
structural purpose. Each has its grain oriented for a specific structural behavior or a
specific loading condition. In addition, many of these industrialized standardization
processes have been instrumental in positioning timber products as a prime mate-
rial for design and construction. While expanding the range of mass applications
for timber products is an essential dimension of industrial and academic research,
developing new value-added products and high-performance applications requires
a more inquisitive approach. Recent work into structural variability of timber com-
ponents (Self 2017), integrated Joinery (Robeller et al. 2017; Krieg and Menges
2013), custom-laminated doubly-curved components (Cheng et al. 2015), or envi-
ronmentally active shape-changing wood systems (Reichert et al. 2015) and actuators
(Ruggeberg 2015) present a new conceptual approach that is decidedly specific to
the performance characteristics of wood. These advances have only been possible
through a multidisciplinary approach to material research and the development of
specialized computational design-to-fabrication tools.
Moreover, a growing number of innovative architecture projects have started to
build on this material specificity in wood. For instance, the precise manipulation of
grain orientation for a specific architectural intent has been implemented in Shigeru
Ban’s curved elements for the Haesley Nine Bridges Golf Course Club House. The
structure is made from custom laminated veneers, manufactured by Blumer Lehman,
that have been structurally designed for specific architectural applications rather than
a generic structural loading condition. The ability to customize wood’s performance,
through material engineering and fabrication, for its structural requirements is one
of the reasons for its large potential in the 21st century.
3.2 Innovation—Digital Design
Today’s planning processes in building construction are well on their way to becom-
ing digitalized, but they are also fragmented into different disciplines (Kieranamp
and Timberlake 2004). On the one hand, digital planning processes may allow for
more complex and potentially performative buildings. On the other hand, they are
currently still characterized by a top-down design development where questions of
producibility and materiality are only answered at a late stage of the design process.
The current situation causes not only higher planning costs through changes in the
design, but it also requires more time and effort for manufacturers.
Leaving material performance and fabrication parameters for later stages perpet-
uates the disconnect in the design practice between form generation and materializa-
tion. This modus operandi leaves architects to design forms with limited fabrication
understanding while it simultaneously leaves builders and fabricators in the precar-
ious and costly position of fabricating structures without an understanding of the
intended architectural design performance. A common example of this disconnect is
the use of complex steel frames made of linear elements to build a framework for a
building skin, or covering, which manages to simulate the fluidity of the form in the
digital design such as the Burnham Pavilion in Chicago and the Riverside Museum
70 D. Correa et al.
in Glasgow by Zaha Hadid Architects. This however, is inefficient and illogical when
compared to the ability to simulate the natural bending capacity of the material itself
to represent the flow.
3.3 Innovation—Engineering Analysis
There is a general ambition that architectural design software may one day be able
to accurately simulate material acting structurally within its elastic range, and one
could imagine that this type of innovation, when available, would assist architects in
their pursuit to work with wood in innovative forms. However, at this point engineer-
ing software still has difficulties with the biological and anisotropic characteristics
of wood in plain members. It is still relatively difficult within mainstream engineer-
ing analysis software today to simulate wood with different strengths in different
directions due to grain orientations. Moreover, engineered wood can either make
the problem better or even worse: if the wood is layered in ways that create a more
isotropic material, this simplifies the material characteristics but, in some cases, mul-
tiple layers of laminate require the layering of already anisotropic layers of the wood,
creating an even more computationally complex material. These current limitations
render the implementation and testing of these models extremely labour intensive, or
their complexity makes them so computationally expensive that it becomes imprac-
tical to compute for most building construction applications.
In order to analyze many of the projects mentioned in the previous chapter, a
custom written staged finite element analysis is required that looks at each stage of
loading and how much of the elastic limit of the wood has been used prior to live
loading (Lienhard 2011a, b). This type of software is more common in mechanical
engineering to simulate staged loading but such a simulation is difficult to set up and
complete, especially with a material with varying grain structure and joints whose
behaviour are not conventionally known. In the experiments presented here, the scale
of the structures was small and estimates regarding behaviour were carried out so that
the benefits of the demonstration could be seen, with the knowledge that much more
analysis and sophistication in the engineering software is required before deployment
on a non-experimental application can take place.
3.4 Engaging with Material Characteristics
In addition to bending stiffness and its inherent anisotropic behavior, it is equally

important to note that wood’s material behavior and unique fibrous characteristics
come from its primary function as a naturally grown biological tissue.
The difference in cellular orientation across different species is critical to the
appearance and the structural characteristics in the transverse or radial plane of wood.
More generally, in the case of both softwood and hardwood species, the structure,
distribution and orientation of cells are the determining factors in the anisotropic
structural and hygroscopic characteristics of wood (Wagenführ 1999). Accessing
wood’s material complexity in a meaningful way requires a re-conceptualization of
architectural design that moves away from a single focus of form definition and
building performance, to a more bottom up approach of design formation. In the
latter, material behavior is an intrinsic component of building performance, from
embodied energy (Alcorn 1996), indoor moisture mediation (Simonson et al. 2001),
fire resistance, structural performance, to form generation.
The inter-dependence between materials and building performance is not new in
and of itself, but the quantitative understanding, albeit primitive, of their behavior
is. To develop predictive models for both building performance and the fabrication
of building components, a wider range of computational tools needs to be engaged,
where management of information becomes critical. Whether it is the acquisition
and catalog of data, simulation of structural behavior, simulation and control of
fabrication processes or the translation of data across several unrelated computational
platforms, each step requires its own set of information modeling tools. It is important
to note that many of these tools and methods are new while many of them have
been borrowed from initially disconnected research fields such as material science,
engineering and manufacturing. Therefore, the communication and translation of
different performance goals, data types and design criteria across multiple disciplines
are essential in the success of this research.
3.4.1 Elasticity
The uses of elastic bending for form finding in building construction is relatively
infrequent, as architects and engineers alike are generally trained to understand larger
deformations as potentially dangerous. Active bending for instance, describes curved
beams or surfaces whose geometry is defined by the elastic deformation of initially
straight elements (Gengnagel et al. 2012; Lienhard et al. 2011a, b). Active bending
has been employed in prototypical timber structures in the last years (Menges 2011;
Krieg and Menges 2013; Bechert et al. 2016) resulting in an innovative approach
to the material whose form embodies both the forces and the structural material
characteristics of timber itself. In bending active structures residual stress caused by
elastic deformation is used to act against external forces, such as load bearing, by
increasing their geometrical stiffness. Although integrating this material behavior
poses a challenge due to the complex interrelation of force, form, and fabrication,
it can provide unique formal and structural design opportunities that can facilitate
a much more effective use of the available material and thus fosters a material-
based architecture. One of the most advanced examples of elastically bent wooden
structures is the lattice shell of the Multihalle in Mannheim by Frei Otto, Carlfried
Mutschler and Ove Arup and Partners, completed in 1975 (Burkhardt 1978). Wooden
lattice shells are structures, which are initially constructed as a planar lattice of
timber laths and are later hoisted or lowered to form double-curved, form-active
surface structures. Here, the design approach has to incorporate both the flat, two-
72 D. Correa et al.
dimensional state, and the elastically bent, three-dimensional state of the grid shell.
Taking the material’s elasticity into account it was possible to fabricate the double-
curved grid shell from straight, planar, solid timber elements. The work outlined
by the projects presented by the authors continues this investigation—looking at
customizing material and its application to work together to create expressive and
material informed structural forms.
3.4.2 Hygroscopicity
Unlike other construction materials developed to meet specific manmade functional

requirements, wood has evolved as a highly efficient biological system to meet the
support, conduction and storage requirements of trees. All of its differentiated mate-
rial properties are themselves intrinsic to its physiological role within the tree as a liv-
ing organism and as such, they exhibit a great range of variation (Dinwoodie 2000).
Furthermore, within every geographic region, trees evolved into locally adapted
species and sub-species with differentiated material behavior in response to atmo-
spheric conditions, soil, seasonal changes and fauna. In other words, a bespoke
material for every ecosystem.
Considering this local adaptation can help in developing architectural and building
solutions that are more resistant to specific climatic conditions, it can foster sustain-
able sourcing of materials or support the re-conceptualization of existing wood con-
struction technologies in new contexts. However, it is also possible to augment some
of these biological characteristics to create new high-performance building systems
and components. One of the key biological characteristics of wood, as a vascular
tissue, is its dimensional change in relation to atmospheric conditions. The complex
cellular, polymeric composite characteristics of wood and its structural performance
implications are closely tied to the role of cellulose within the cell walls. The cel-
lular, axial differentiation of cellulose molecules within the fibrils and micro-fibrils
defines many of its structural properties, and it also establishes wood’s characteristic
hygroscopic behavior. This anisotropic expansion generates high pressures that can
greatly vary from specie to specie, from high theoretical values of up to 165 MPa
(Stamm 1964) to a more conservative 83 MPa (Tarkow 1958), depending on the
testing procedure, as explained by Rowell (1995). Extensive research on limiting or
controlling these dimensional changes has been conducted by the timber industry, but
very limited work has been done in integrating these expansion forces within design
applications. The presented work builds on the limited studies on wood movement
and joinery by Eckelman (1998) as well as the more recent development of shape
change bilayer wood actuators by Reichert et al. (2015), Holstov et al. (2015) and
Rüggeberg (2015); as well as the work on 3D printed hygroscopic wood actuators
by Correa et al (2015). Both wood joinery and actuators engage the dimensional
expansion of the material as a key performance characteristic, and therefore, as a
key driver of the design. Considerations of shape, fabrication and assembly are care-
fully considered in relation to the relative humidity (R.H.) of the environment as
well as the equilibrium moisture content (EMC) of the material itself. This integra-
tion seeks to precisely use and augment the hygroscopic expansion properties of the
material to employ its effects, for assembly or shape change, without external elec-
tronic mechanic sensors, controllers or actuators. The authors are building on this
research by using the intricate responsiveness of the material to aid in the assembly
and inter-locking fastening of timber elements in a larger building system.
3.5 Innovation—Fabrication
Digital fabrication tools like CNC machines have become well established in wood
fabrication in the last decades. Their main advantage lies in automating individual
aspects of the production and thus making standardized processes faster and more
efficient while at the same time exceeding the quality standards of manual fabrica-
tion. However, due to their complexity in programming and operating, most wood
processing machines have been equipped with closed, “black box” systems, with user
interfaces designed to simplify these processes by incorporating pre-defined mass
production fabrication workflows. This approach has been very successful in mak-
ing the technology accessible and therefore ubiquitous on every shop floor. However,
advanced users find themselves at odds with the overly constrained methods and the
limitations of a “black box” approach.
The industrial robot constitutes an opposite approach to these highly specialized
systems in both practicality and applicability. The multi-axis robot arm represents
a generic platform on which a multitude of tools, so-called effectors, and generic
Numeric Control platforms can be attached. Control platforms and effectors are
developed by multiple vendors, researchers or academic institutions, unrelated to
the robot manufacturer, but according to the tool or application. These NC tools
range from open ended multi-purpose applications to “black box” niche uses. It is
therefore only the effectors and the NC platform, or a combination of those, that will
ultimately define the variability and potential of the fabrication process. The potential
of industrial robots in wood construction ultimately derives from their extended
kinematic range and their adaptability to new fabrication processes, through custom
effectors and varied NC platforms, making the fabrication of more complex and
differentiated building parts possible. Most important yet, the ease in adaptability
of the industrial robot facilitates a shift in IP generation that does not rely solely on
the product but on the development of adaptable fabrication platforms. Nevertheless,
to really take advantage of the hardware’s flexibility the main challenge lies in the
development of a robust and open design-to-fabrication computational workflow
with a direct digital transfer of machine data from design to manufacturing.
One of the main foci of this research is therefore the demonstration of the implica-
tion of digital fabrication technology on how wood could be conceived and applied.
The 7-axis robot setup used in these experiments allows a very large range of options
for shape definition and assembly of wood components. This freedom combined
with the accuracy of the tool allow for tight fit joints at very precise angles. This
new-found fabrication ability in wood has been used here in three ways: to create
74 D. Correa et al.
a very precise joint capable of transmitting structural forces in predicable ways; to

create informed free-form geometries, which are precise and allow subtle changes in
curvature in response to natural material bending curvature; and to allow fabrication
and assembly instructions to be embedded within the component elements of the
installations themselves.
3.6 Innovation: Summary
The projects shown in the following chapters attempt to engage with innovation in
wood in the areas discussed above. We can see why innovation is halted across dis-
ciplinary borders and this research seeks to demonstrate an integrated approach that
facilitates innovation. Digital design does facilitate interdisciplinary collaboration.
However, when looking at such a radical change as the industrial robot in combination
with the evolving digital technologies, the entire conventional working methodology
has the potential to be disrupted. For example, the integration of accurate engineering
analysis and machine control into architectural software could produce unforeseen
disruption. In addition, the highly precise tolerances of the robot and the ability to
easily fabricate any shape or joint condition allows innovation. Further, more inno-
vation can be expected from the advanced understanding of material characteristics,
which can be integrated and taken advantage of in the design process, rather than
homogenized, approximated or ignored. All of these aspects can be combined into a
radical innovation in how we use and think about design in wood.
4 Methodology
The presented projects focus on the design and fabrication of two large scale, roboti-
cally fabricated, timber prototypes developed at the University of British Columbia,
Center for Advanced Wood Processing, in collaboration with the Institute of Com-
putational Design and Construction at the University of Stuttgart, and the School of
Architecture at the University of Waterloo, through two separate interdisciplinary
workshops. The workshops were led by the authors (architects and structural engi-
neers) in close collaboration with local wood researchers, industry participants, inter-
national faculty and students. Each prototype integrates a different set of material
performance goals but also different methodological strategies for both design intent
and assembly. Technically, the prototypes were developed using integrated ‘design-
to-fabrication’ computational and form finding tools, implemented using a 7-axis
robotic milling set-up.
Until now, traditional design processes in wood are typically clearly differentiating
between linear and planar elements in a hierarchical manner. In the case of the
innovations discussed here, the material takes on a design flexibility and a facility
with form. This form, however, is not imposed on the material but rather informed by
the material. The material characteristics, limitations and structural properties inform
the design, providing constraints but offering possibilities. The resulting techniques
and forms are presented here as initial steps towards a new language of design in
wood, which will better reflect the current design ethos.
Both material characteristics and fabrication possibilities were investigated in
parallel for the development of new construction systems. The presented prototypes
reveal both structural and architectural performance characteristics that are unique
to the presented integrative design approach. Their contribution here lies on the
reciprocities between prototyping robotic fabrication and material-based computa-
tional design. Both prototypes explore the concept of inherent assembly instructions.
Most of the building elements can be assembled without external instructions or plan
drawings because their assembly information is already inscribed in the connection
geometry. This method also ensures that the building elements are elastically bent
into their correct shape.
5 Results
The experimental installations created were full scale applications of the developed
technologies. Both were created with a custom design to fabrication computational
tool allowing for the direct generation of machine code from the design files. Both
prototypes were designed and built during workshops held at the University of British
Columbia (Figs. 1 and 2).
5.1 Prototype 1
The first prototype explored elastic bending of large scale timber sheets for double-
layered timber structures through a combined approach of material-oriented design,
traditional wood joints, and robotic fabrication techniques (Figs. 2, 3, 4, 5, 6, 7, 8,
9 and 21). Working with the hygroscopic and elastic capacities of wood, the project
uses elastically bent plywood sheets to form a finger jointed, segmented, double-
layered shell. Each joint is designed for self-aligning and self-locking through the
hygroscopic expansion of integrated dowels, as described in Sect. 3.4. The geometry,
joints and bending radius are parametrically defined within the custom design-to-
fabrication computation model. The direct associations, boundaries and other design
constraints are visually represented in the model. This visual access within a design
platform allows participants to develop multiple formal iterations while mediating
design intent with material and fabrication constraints (Fig. 4). The tool enables
participants, with different technical backgrounds and expertise, to provide design
input at multiple hierarchical levels, from global shell articulation through a double-
curved surface to precise multi axis control of the tooling path for the integrated 7
axis robot milling set up.
76 D. Correa et al.
Fig. 1 Timber Wave prototype, doubly curved timber assembly, at the University of British
Columbia Campus
Fig. 2 (Left): Photograph of the finished prototype made from a finger-jointed, double-layered,
elastically bent plywood construction
Fig. 3 (Right): Detailed photograph of the structure under construction. In the front, the extended
finger joints and shear dowels are visible
Fig. 4 (Left): Visualization of the computational process from the design model to the fabrication
data. The individual plywood strips get unrolled before the joint geometry and milling tool paths
get applied
While the elastic bending of timber sheets has been thoroughly explored by multi-
ple research institutions in the past years, their application in large scale construction
is prohibited by the relation between maximum bending radius, bending forces, and
78 D. Correa et al.
Fig. 5 (Right): Visualization of the double-layer construction, the finger joints, and the wooden
dowels that lock the connection in place
Fig. 6 (Left): Detail photogram and diagram for the hygroscopic induced locking mechanism of
the panel using wood dowels
material thickness. Generally speaking, timber construction elements usually have a

material thickness not suitable for elastic bending as a visible bending radius would
not be achievable during assembly. Instead, a multi-layered approach was investi-
gated in the first prototype, combining two layers of plywood sheets thin enough to be
bent manually, with two people on site, but resulting in enough material cross section
in aggregation to potentially carry higher loads in a larger structure. For this purpose,
both layers are interconnected with wooden shear dowels and spacers (Figs. 3, 5 and
Fig. 7 (Right): Photograph of the finished prototype for the double layer structure
Fig. 8 (Left): Robotic fabrication of the individual sheets. The fabrication data was directly
exported from the computational design model
80 D. Correa et al.
Fig. 9 (Right): Back side of the final prototype assembled on campus
6). This method allows for an additional space between the layers, allowing them
to act as a hollow sandwich panel. At the same time, the dowels are placed in such
a fashion that the two plywood layers can only be assembled when elastically bent
into their final shape.
The double-layered configuration allowed additional structural opportunities,
however, it also necessitated further development of curved finger joints for ply-
wood plates, building on previous research by the authors (Krieg and Menges 2013).
For this purpose, the finger joint was further developed to facilitate the assembly
sequence while maintain its structural capacity. Both plywood layers are connected
to their neighbors in consecutive fashion, allowing for finger joints to reach through
both layers of the neighboring elements. This type of multi-layered intersecting fin-
ger joint allows to rely only on wooden dowels to lock the connection after assembly
(Figs. 5 and 6). The necessary geometric information to generate the milling tool
path can be computed by the geometric relationship of two or more connected plates
in their final bent geometry, and their equivalent flattened, or unrolled shape. As the
connection angle and bending radius changes along the connection seam, the tool
path and vector are generated for each single joint and slot individually. Their param-
eters are informed by their position on the bent geometry and transferred to the flat
plate. Through the developed fabrication method, it is possible to cut curved finger
joint seams that specify the plywood sheet’s bending radius and geometry through
the individual joint’s shape and position. The resulting joints along the curved seam
only fit into their adjacent plate’s joints when the plywood sheet is elastically bent
into its correct shape (Figs. 4, 5, 6 and 7).
While the modulus of elasticity of the plywood sheets is used to define the global
form, the final assembly geometry is locked in place through the hygroscopic expan-
sion of the wooden dowels. Based on traditional wood joinery methods, as per
Sect. 3.4, the hygroscopic characteristics of the material are used to ensure a rigid
and permanent connection. In contrast to the plywood sheets that do not react to
changes in relative humidity due to their cross-laminated veneers, the wooden dow-
els consist of solid hardwood and therefore change dimensions depending on their
relative humidity. To allow for precise assembly, the diameter and moisture content
of the dowels is checked and equalized prior to fabrication—all dowels’ nominal
dimension and moisture content must be consistent. The dowels’ moisture is then
reduced to 4–6% EMC, or “dried” state, in a climate-controlled chamber resulting
in a 5–10% change in diameter. The perforations in the plywood sheets were milled
according to this smaller diameter, and therefore allowed the dowels to be inserted in
their “dried” state, and to become “locked” in place after adjusting to the surround-
ing atmospheric humidity (Fig. 6). The coupling of elastic bending and hygroscopic
expansion allows for the complete assembly of the self-standing structure without
the need for any metal fasteners or molds (Fig. 7).
The initial design of the prototype was conceived by understanding the relation-
ship between the locally elastically bent plywood strips and the connection angle
to their neighbors. In order to create a sufficiently expressive curvature, a thin ply-
wood was selected and the strategy of folding back and forth at the connections was
engaged to provide an additional structural depth to the shell form in order to increase
its structural stiffness while still maintaining the fluidity of the curved form, which
requires the elasticity of the material for its expression. This provided the additional
benefit of controlling the sheet size and maintaining a single curvature in the plywood
sheet elements.
5.2 Prototype 2
For the development of a second construction system and subsequent prototype

(Figs. 1, 10, 11, 12, 13, 14, 15, 16, 17, 18 and 19), the authors investigated the
relationship between assembly instructions and the geometric specificity of build-
ing parts for the construction of a doubly-curved building system. By encoding the
assembly instructions for a set of generic stock components into a different set of
specialized, CNC milled, building parts, traditional instructions such as plan draw-
ings become redundant. In addition, this strategy allows to fabricate and construct
elaborate timber structures partly from off-the-shelf building materials. By com-
bining two approaches—highly specific and prefabricated building components, and
standardized “stock” materials, the authors investigate the technical and architectural
reciprocities between mass production and bespoke fabrication. More specifically,
by coupling stock cedar planks with a robotically fabricated diagrid substructure, the
project uses geodesic methods to create a double-curved “timber wave” (Figs. 10,
14 and 15). The diagrid substructure can be assembled on site and is flexible but
82 D. Correa et al.
Fig. 10 (Left): Photograph of the finished prototype. The front face prominently shows the bent
cedar planks following the double-curved surface
self-supporting at first (Fig. 12). The intersection points of the diagrid are designed
to be structurally rigid only in combination with the horizontally arranged cedar
planks. These cedar planks are initially straight, and normally used as standard hori-
zonal cladding (Fig. 11). In this project, their functionality gets extended by fixing
them directly on the diagrid’s intersection points (Fig. 13). The diagrid is made
from two layers of CNC milled plywood strips, which, although initially planar, are
not straight but curved; the cedar planks, on the other hand, are initially planar and
straight. Both the diagrid and the cedar planks are elastically bent to achieve the
final structural assembly (Fig. 16). This additional constraint limits the design space
of double curved surfaces as the planks can only follow so-called geodesic lines.
Therefore, the planks become the defining design criteria.
In this project, hygroscopically activated wooden dowels are used to self-align and
fasten the diagrid components making metal fasteners mostly redundant. In this case,
the diagrid is double-layered but connects to thick plywood spacers at its intersection
points. In order to ensure a permanent connection between these elements, dried
wooden dowels are inserted during assembly. As the installation is located on site,
exposed to a higher R.H. environment, the dowels equalize their moisture content
and therefore expand to lock the layered connection.
Similar to the first project, the ‘design-to-fabrication’ approach and computational
tool enables the designer to engage with a complex technical problem through a
visual and design-oriented methodology. At the higher level, a double-curved surface
model provides global feedback on curvature and overall shape while subsequent
Fig. 11 (Right): Detailed photograph of the diagrid substructure. The diagrid is double-layered
and has thick plywood spacers at the intersection points, which also act as connectors to the cedar
planks
Fig. 12 (Left): Detailed photograph of the diagrid substructure during assembly

84 D. Correa et al.
Fig. 13 (Right): Detailed photograph of the cedar plank during assembly. Intersection point of
diagrid is half exposed awaiting fastening of the next plank
Fig. 14 (Left): Visualization of the diagrid elements and multi-layered assembly system. The
double-curved design surface is constructed with a double-layered diagrid and thicker spacers
at its intersection points
steps allow for parameter manipulation of plank overlap and joinery. Also, in this
project, the computational design tool pre-visualizes the result of the elastically bent
building elements but also recalculates their initially planar shape for fabrication.
This becomes especially evident when the design tool calculates the unrolled state of
the cedar planks, proving that they are initially straight and planar. A simulation of
the robotic milling process and export of robot control files are fully integrated in the
computational design process. Divided into two complimentary systems, the diagrid
Fig. 15 (Right): When unrolled from the design model, it becomes evident that the cedar planks
are indeed straight. They are only bent in one direction when connected to the diagrid
Fig. 16 (Left): The diagrid structure and the cedar planks during assembly
provides precise geometric definition through robotic fabrication and self-aligning

hygroscopic joinery, the planks are then easily aligned with the diagrid and fastened
through standard wood screws.
By using advanced timber fabrication techniques and taking full advantage of the
extended fabrication range of the multi-axis set up, large sections of plywood were
custom milled and assembled on-site into a unique one-to-one scale architectural pro-
86 D. Correa et al.
Fig. 17 (Right): From the back, the relation between planks and diagrid can be seen
Fig. 18 (Left): Diagram of fabrication components on the 7 axis robotic system
totype. Once assembled, the diagrid and the cedar planks form a stable, doubly-curved
building system. The prototype showcases distinctive wood fabrication possibilities
that integrate computational design, material characteristics, and digital fabrication
in a direct design to production paradigm, leading not only to innovation in timber
construction, but also for a re-interpretation of wood architecture.
Fig. 19 (Right): From the back detailed photograph of the diagrid substructure
Fig. 20 (Left): Hands-on robot control training during Robotic Fabrication workshop
88 D. Correa et al.
6 Discussion
Completed in week long workshops, the prototypes offer evidence that the unique
anisotropic and hygroscopic properties of wood have many unexplored design oppor-
tunities. The contribution of the prototypes presented here is their ability to position
wood as a driver of design intent and form. A contribution that is only possible
through a careful material understanding as well as a robust design to fabrication
workflow capable of integrating complex parameters and associative relations inher-
ent to fabrication, assembly and material performance.
The projects are case studies for both a manifestation of a design paradigm as well
as an assertion about the potential of integrated research into the potentials of wood,
and a re-emergence of applied research teaching methodologies. Using Picon’s words
(2010), “innovation comes about at the intersection of technical and social issues that
need to be decoded”. In here, the innovation is foregrounded primarily to be one of
material informed design, but in this discussion, it may be worth considering another
contingent dimension of the work. In an era of unprecedented connectivity and global
economies, disciplinary research remains a highly territorialized domain. The devel-
opment and execution of these projects as week-long workshops as opposed to term
long academic endeavors, is indicative of the larger challenges that lie ahead. While
interdisciplinary research is highly lauded as a pinnacle of modern research, insti-
tutional structures of disciplinary funding and professional teaching accreditation
Fig. 21 (Right): Manual assembly of elastically bent double shell prototype structure during
Robotic Fabrication workshop
struggle to identify frameworks that can support these types of initiatives. As pre-
sented in Sects. 2.1 and 2.2, technical developments or architectural design visions
in isolation are not enough to support meaningful building construction innovation.
Disciplinary fields can address individual challenges of quality control, structural per-
formance, system integration or visual expression. Meaningful development of new
technologies into commercial applications, however, only emerge at this intersection.
Interdisciplinary research at full scale, like the work presented here, is positioned to
be fertile ground for this type of exchange. The learning opportunities offered by full
scale prototyping projects are unique as they provide a defined scenario with real
testing conditions for design investigation, quality control, structural performance
testing, material research, project management and cost evaluation (Figs. 20 and 21).
An integrative design process, or design to fabrication process, is by definition and
inter-disciplinary design process. Computational, robotic, structural, material and
pedagogical tools and methods are positioned in support of each other both within
and outside of disciplinary boundaries.
7 Conclusion
Each prototype begins to demonstrate the potential of wood beyond current practices.
While mass industrialization capitalized on large-scale use of stock components,
demonstrated by stick frame construction, mass customization has focused on the
imposition of homogeneity on the varied anisotropic nature of the original material,
as can be seen with CLT, MDF, plywood, OSB and other CNC optimized materials.
Both technical and material innovations are not in themselves sufficient drivers of
meaningful change in architectural language or building practice. It is important to
remember that it was not only the development of reinforced concrete or steel that
brought about its implementation; meaningful innovation emerges at the intersec-
tion of needs and opportunities. The development of steel necessitated a new design
language and new specialized disciplines to truly have an impact in the built environ-
ment, while concrete required the development of methodologies for construction
standardization and disciplinary expertise before becoming an architectural staple.
It should be said that while technology may provide some answers to the practical
aspects of implementing materials in architecture, it cannot provide a road map for
research and on its own is unable to foster design interest or aesthetic performance.
The prototypes presented here do attempt to demonstrate the importance of a concep-
tual shift from representation and form definition to informed material investigation,
the re-purposing of robotic fabrication as method for informed material manipula-
tion and the role of teaching as shift from lecture to active interdisciplinary dialogue.
These investigations take small steps toward the larger research agenda of looking
at what might be possible in wood in the future, what are the architectural languages
and building types that these new possibilities can facilitate.
The projects are presented as a medium to begin a dialogue about methods of
construction: integration of material behavior in the design process, integration of
90 D. Correa et al.
computational tools and fabrication processes and most importantly, the knowledge
transfer to new architects and designers. The development, execution and teaching
methodologies necessary for these projects have inevitably generated larger ques-
tions about disciplinary roles, existing design paradigms and the potential of future
powerful design, analysis and fabrication tools—questions that are essential for the
future of the practice and the sustainable future of our cities.
Acknowledgements The workshops presented in this paper were organized and supported by
the Centre for Advanced Wood Processing, as well as the School of Architecture and Landscape
Architecture at the University of British Columbia. Funding was provided by the Forestry Innovation
Investment, Perkins+Will Vancouver, Perkins+Will Building Technology Lab, and the UBC SEEDS
Sustainability Program.
The authors would like to express their gratitude to Iain MacDonald, Jason Chiu and Jörn Dettmer
from the Centre for Advanced Wood Processing, as well as Dean Gregory from the UBC Campus
and Community Planning, and David Gill from the SEEDS Sustainability Program. Further, the
authors would like to thank all the participants of both workshops without whom the results would
not have been that successful.
References
Alcorn A (1996) Embodied energy coefficients of building materials. Centre for Building Perfor-
mance Research, Victoria University of Wellington
Antemann M (2013) Complex structures solutions in wood. Vancouver, BC, February
Bechert S, Knippers J, Krieg O, Menges A, Schwinn T, Sonntag D (2016) Textile fabrication
techniques for timber shells: elastic bending of custom-laminated veneer for segmented shell
construction systems. In: Adriaenssens S, Gramazio F, Kohler M, Menges A, Pauly M (eds)
Advances in architectural geometry 2016, vdf Hochschulverlag AG ETH Zurich, Zurich, pp
154–169. ISBN 978-3-7281-3778-4
Boake TM (2012) Understanding steel design: an architectural design manual. Basel, Birkhäuser
Burkhardt B (1978) IL 13: Multihalle Mannheim. Karl Krämer Verlag, Stuttgart
Cheng A, Gaudin T, Meyboom A, Neumann O, Tannert T (2015) Large scale wood surface struc-
tures. In: 3rd annual international conference on architecture and civil engineering (ACE 2015).
Singapore, pp 13–14
Collins P (1959) Concrete: the vision of a new architecture. Second Edition, 2004. McGill Queens
University Press, Montreal & Kingston
Communication from the Commission to the European Parliament, the Council, the European
Economic and Social Committee and the Committee of the Regions on Resource Efficiency
Opportunities in the Building Sector (2014). COM, European Commission
Correa D, Papadopoulou A, Guberan C, Jhaveri N, Reichert S, Menges A, Tibbits S (2015) 3D
printed wood: programming hygroscopic material transformation. 3D printing and additive man-
ufacturing. Mary Ann Liebert 2(3):106–116. https://doi.org/10.1089/3dp.2015.0022
Dinwoodie JM (2000) Timber: its nature and behaviour. E&FN Spon, London
Eckelman CA (1998) The shrinking and swelling of wood and its effect on furniture. Forestry and
natural resource 163. Purdue University, West Lafayette, I.N.
Elliott CD (1992) Technics and architecture: the development of materials and systems for building.
MIT Press, Cambridge, Massachusetts
Gann DM (2000) Building innovation. Thomas Telford, London
Gengnagel C, Alpermann H, Knippers J, Lienhard J (2012) Active bending, a review on structures

where bending is used as a self formation process. In: Proceedings of the international IASS
symposium, Seoul, Korea
Gordon JE (2003) Structures: or why things don’t fall down. Da Capo Press, Boston
Holstov A, Bridgens B, Farmer G (2015) Hygromorphic materials for sustainable responsive archi-
tecture. Constr Build Mater 98:570–582. https://doi.org/10.1016/j.conbuildmat.2015.08.136
Kieran S, Timberlake J (2004) Refabricating architecture. How manufacturing methodologies are
poised to transform building construction. McGraw-Hill, New York
Kolb J (2008) Systems in timber engineering: loadbearing structures and component layers.
Birkhäuser, Basel
Krieg O, Menges A (2013) Prototyping robotic production: development of elastically bent wood
plate morphologies with curved finger joint seams. In: Gengnagel C, Kilian A, Nembrini J,
Scheurer F (eds) Rethinking prototyping, proceedings of the design modelling symposium Berlin
2013, Verlag der Universität der Künste Berlin, pp 479–490. ISBN 978-3-89462-243-5
Krieg O, Schwinn T, Menges A (2015) Integrative design computation for local resource effec-
tiveness in architecture. In: Wang F, Prominski M (eds) Urbanization and locality: strengthening
identity and sustainability by site-specific planning and design. Springer Science and Business
Media, pp 123–143. ISBN 978-3-662-48492-0
Lienhard J, Fleischmann M, Menges A (2011) Computational design synthesis: embedding material
behaviour in generative computational processes. In: Proceedings of the 29th eCAADE confer-
ence, Ljubljana (Slovenia) 21–24 Sept 2011, pp 759-767. ISBN 978-9491207013
Lienhard J, Schleicher S, Knippers J (2011) Bending-active structures. In: Research Pavilion
ICD/ITKE, in symposium of the international association for shell and spatial structures 2011,
London, United Kingdom
Menges A (2011) Integrative design computation: integrating material behaviour and robotic manu-
facturing processes in computational design for performative wood constructions. In: Proceedings
of the 31st annual conference of the ACADIA, pp 72–81
Peters TF (1996) Building the 19th century. MIT Press, Cambridge, Massachusetts
Picon A (2010) The first steps of construction in iron: problems posed by the introduction of a new
construction material. In: Before steel. Niggli, Zürich
60:50–69
Rinke M, Schwartz J (2010) (a). The ambivalence of the revolution: an introduction. In: Before
steel. Niggli, Zürich. In: Before steel: the introduction of structural iron and its consequences.
Verlag Niggli, Zürich
Rinke M, Schwartz J (2010) (b) The ambivalence of the revolution: an introduction. In: Before
steel. Niggli, Zürich
Robeller C, Weinand Y, Helm V, Thoma A, Gramazio F, Kohler M (2017) Robotic integral attach-
ment. In: Menges A, Sheil B, Glynn R, Skavara M (eds) Fabricate 2017 conference proceedings.
UCL Press, London, pp 92–97
Rowell RM (1995) One way to keep wood from going this way and that. Am Rec XXXVI:12–16
Rüggeberg M, Burgert I (2015) Bio-inspired wooden actuators for large scale applications. Plos
One 10(4)
Self M, Vercruysse E (2017) Infinite variations, radical strategies. In: Menges A, Sheil B, Glynn R,
Skavara M (eds) Fabricate 2017 conference proceedings. UCL Press, London, pp 30–35
Simonson C, Salonvaara M, Ojanen T (2001) Improving indoor climate and comfort with wooden
structures. VTT Publications
Slaughter ES (1998) Models of construction innovation. J Constr Eng Manag 124(3):226–231
Stamm AJ (1964) Wood and cellulose science. Ronald Press Co., New York, NY
Tarkow H, Turner HD (1958) The swelling pressure of wood. For Prod J 8(7):193–197
92 D. Correa et al.
Taylor JE (2006) Three perspectives on innovation in interorganizational networks: systemic innova-

tion, boundary object change, and the alignment of innovations and networks. Stanford University,
Stanford
Taylor JE, Levitt RE (2005) Inter-organizational knowledge flow and innovation diffusion in project-
based industries. In: System sciences HICSS’05 proceedings of the 38th annual hawaii interna-
tional conference on system sciences. IEEE, pp 1–10
Wagenführ R (1999) Anatomie des Holzes. DRW Verlag, Leinfelden-Echterdingen
Winch G (1998) Zephyrs of creative destruction: understanding the management of innovation in
construction. Build Res Inf 26(5):268–279
New Workflows for Digital Timber
Tom Svilans, Martin Tamke, Mette Ramsgaard Thomsen, Jonas Runberger,

Kai Strehlke and Martin Antemann
Abstract This chapter describes a research thread at CITA which explores how
computation and a challenging of traditional material practice can impact the use of
timber in architectural design and fabrication. Several past research projects at CITA
have demonstrated the potential for streamlining the design-to-production process
using computational tools, and the value of working in concert with the inherent prop-
erties of wood. Current research continues this thread through a participation in the
Innochain research network (http://innochain.net/) and collaboration with industrial
partners White Arkitekter AB and Blumer-Lehmann AG. Through the embedding
of digital tools within established timber design a fabrication processes, new work-
flows are proposed which could lead to more intelligent design decisions, optimized
building components, and new timber morphologies.
Keywords Wood design · Complex timber structures

Parametric design and fabrication strategies · Optimization of wood architectures
Digital wood workflows
T. Svilans (B) · M. Tamke · M. R. Thomsen

Centre for IT and Architecture, Copenhagen, Denmark
e-mail: tsvi@kadk.dk
M. Tamke
e-mail: martin.tamke@kadk.dk
M. R. Thomsen
e-mail: mette.thomsen@kadk.dk
J. Runberger
White arkitekter AB, Stockholm, Sweden
e-mail: jonas.runberger@white.se
K. Strehlke
Blumer-Lehmann AG, Gossau, Switzerland
e-mail: kai.strehlke@bluimer-lehmann.ch
M. Antemann
Design-to-Production, Erlenbach/Zürich, Switzerland
e-mail: antemann@designtoproduction.ch

94 T. Svilans et al.
1 Introduction
1.1 Background
The design, fabrication, and use of free-form timber elements involves many steps:
design and CAD, CAE, joinery and detailing, supply and sourcing, fabrication, logis-
tics and assembly, maintenance and durability, and the component lifecycle. Such
design and fabrication workflows are typically linear and have little opportunity for
mutual informing or recursion between steps. This investigation focuses on how free-
form timber elements are designed, how they are fabricated, and how more knowledge
of material principles can be embedded into both domains through computational
processes and digital sensing tools. Glue-laminated timber beams—glulams—are
used as the starting point, due to their use in industry for creating large-scale free-
form building elements. The notion of the glulam blank—the raw laminated timber
piece that is subsequently machined into the final timber element—is central to the
investigation. Exploratory material probes question the typical categorization of glu-
lam blanks into straight, single-curved, or double-curved blanks by proposing new
types of blanks based on a variety of production and material principles.
New Workflows for Digital Timber 95
The work is undertaken in partnership with Blumer-Lehmann AG, a leading Swiss

contractor specialising in the development and realisation of free-form timber struc-
tures and envelopes,1 and Dsearch—a computational design development group
within White arkitekter AB,2 a multi-disciplinary Scandinavian architecture prac-
tice, which focuses on digital computation in architectural design and production
workflows.
1.2 CITA and Innochain
The continued development of digital technologies have brought about new

approaches and methods of working more fluidly between architecture and its con-
struction. Far from being merely an indulgence in the new formal possibilities made
available by digital modelling tools, the use of digital tools has allowed architects to
challenge existing practices and integrate a much wider breadth of knowledge into
the design and production of buildings (Kolarevic 2003). This type of challenge of
the status quo is taken up by the Innochain network—a research network comprised
of 15 researchers, at 6 different institutions across Europe, in collaboration with
approximately 14 international industrial partners.
Previous research at CITA has explored the relationship between computational
modelling, simulation, and timber. Lamella Flock (2010) examines how existing
methods can be extended and transformed using computational methods, using para-
metric timber modules in a large-scale agent-based simulation. Prefabrication and a
digital code-to-machine strategy enables the exploration of the geometric constraints
of the Zollinger system and customized wood joinery over a large-scale assembly of
simple, variated elements (Tamke et al. 2010). Dermoid (2011) investigates the use of
short timber elements and their active bending behaviour to create a hybrid structure
that combined aspects of reciprocal frames and elastic gridshells. The simulation of
bending and its effect on the overall structure through physics-based relaxation tools
seeks to connect the material reality of timber behaviour with the digital information
model (Tamke 2012). This interest in the connection between digital processes and
timber continues through participation in the Innochain network.
1.3 A New Case for Timber
The resurgence of timber as a large-scale construction material has come about

because of both social and technological reasons. Developments in adhesives and
the increase of precision in industrial processes have decreased the margin of error
typically attributed to crafted wood construction and made its structural analysis
1 http://www.blumer-lehmann.ch/en/.
2 http://whitearkitekter.com/white-research-lab/.
more robust and predictable (Radkau 2012). Engineered timber can be made stiffer
than natural wood and without the limits of scale usually imposed by the size of the
log. Even in its natural state, wood has long been championed as one of the most
fundamental materials in the development of modern civilization (Mumford 1934).
Some of its natural, inherent properties—such as heat insulation—outperform other
popular construction materials such as steel or concrete, and advances in material
sciences have also offered up new coatings and treatments to mitigate the suscepti-
bility of wood to decay and deterioration under the influence of moisture and harsh
environments. Its renewability as a natural resource and ability to store relatively
large amounts of carbon have also contributed to its favourable reception in a social
climate sensitive to environmental concerns and the effects of global climate change.
In terms of its use in fabrication and building, wood is easily machined and
processed compared with other materials—workable by hand or by machine, on- or
off-site, in garage workshops or large production halls. In relation to its volume, it is
quite a lightweight material, which positively impacts its ease of transportation and
handling. This versatility and mobility sets it apart from other large-scale building
materials which may require more specialized ingredients, tools, or environments.
Coupled with technological advances in computation, digital sensing, and
software-hardware interfaces, opportunities arise for the reconception of current tim-
ber workflows and how they might be challenged. In the face of mounting complexity
in the built environment and the specialization of the involved disciplines, the timber
renaissance provides an ideal case for devising new design and fabrication work-
flows which prioritize integration and communication between design, fabrication,
and material.
2 The Material Complexity of Timber
2.1 Diversity in Wood
Despite this promising outlook, wood still presents many challenges for its applica-
tion and analysis, borne out of its wide variance and individuality. As a multi-scalar
material,3 wood presents several distinct levels of detail, each with its own types of
parameters, behaviours, and properties, but each also interrelated and dependent on
the others. Most of the diversity in tree species, types of wood, and their resultant
behaviours can be attributed to small-scale variations in the cellular composition and
growth of trees, variations which result in much larger changes in structure and form
at the scale of the tree (Hoadley 2000).
The tracheid or wood fibre—the long cells that constitute the bulk of a tree—has
multiple roles throughout the tree—spatially and throughout its lifecycle. The gen-
erally parallel arrangement of these cells creates one of the most important constants
3 Similarto the way Frei Otto describes some materials as ‘[consisting] … of smaller elements
which are called “materials elements”’ in Otto (1992).
Fig. 1 Hankinson’s equation describes the relationship between grain angle and the compressive
strength of wood (Hankinson 1921). This has been proven to also be generally true for tensile
strength and other wood species
across all types of wood: its anisotropic character. This has fundamental implica-
tions for the growth and use of timber, as it means that different properties will
express themselves differently, depending on their relation to this cell orientation
(Fig. 1). From a structural perspective, this means that wood is stronger along the
fibres than across them. These cells and vessels serve as the tree’s primary structural
and transport mechanisms.
These particular mechanisms are, however, tailored and specialized locally by
higher-level factors. As the tree matures and thickens, the role of the older cells
changes to a less active one, becoming the more inert, structural core of the tree trunk
or branch as new cells form and grow along the outside ‘wet’ layer just beneath the
bark (Fig. 2). At junctions between the trunk and branches—moments of branching
or forking—the fibres intertwine and tangle at the areas of most splitting stress, thus
tying the splitting masses of the branches together more robustly. Prolonged exter-
nal forces acting upon the tree provoke a growth response in the form of reaction
wood—an allocation of more fibres and mass to the area under stress, in other words,
a localized strengthening—or functional grading—of the tree to ensure the structural
integrity of the whole. The strategic allocation of wood fibre over a tree’s life cycle
therefore demonstrates a gradual self-optimization, in response to its evolving func-
tional and structural demands.
Fig. 2 A photograph of the cell variation in the growth rings of a cross-section of Douglas Fir
The way in which these effects and growth behaviours are expressed, however, is
also dependent on the kind of tree in question. Deciduous trees allocate more material
on the side of the tree that experiences more compressive forces; coniferous trees
instead exhibit tension wood, which is an increased allocation of material on the side
that is under tension. Thus, the variety of the fibre—and its roles—is expanded by
the taxonomy of tree species. These in turn are shaped and individuated by a wide
variety of factors such as climate and more immediate, local environmental changes,
or even different spatial conditions (proximity to other trees, openness of the space
around it, etc.). Basic properties such as stiffness, density, susceptibility to swelling
or shrinking under changing moisture conditions, resistance to decay, and so on, are
particular not only to each species of wood, but also to the specific growth conditions
of each particular tree. What this goes to show is the ‘exceedingly regional character’
of wood (Radkau 2012), which contributes to the enormous parameter space or
taxonomy of wood properties and types, one which still challenges classification
efforts of wood (Hoadley 2000). However, the use of computational tools allows this
large parameter space to be leveraged by digital design processes, potentially leading
to better-informed and differentiated design solutions.
The material usage of wood—in construction or otherwise—further expands the
already vast space of possibility, bringing with it the multitude of processes, uses,
and infinite combinations thereof that come with contemporary timber processing
workflows.
2.2 Fibre in Construction
Because of this variance in properties and behaviours, the utility and application of
wood has also resulted in a huge diversity. In old times, even poor peasant house-
holds could have up to 27 different types or species of wood in various different
uses (Radkau 2012). While industrial wood production today has a narrower range
of source species, this is largely taken up by the multitude of ways in which wood
can be processed for further use, especially with the introduction of structural adhe-
sives (Fig. 3). Looking at the range of the most common timber products, certain
classifications become apparent: wood that is used in a solid form, such as lum-
ber, or individual lamellas in a glulam; versus wood that is used in bits and pieces,
such as OSB, OSL, or fibreboards; wood that focuses its anisotropic properties by
orienting its fibres in a common direction, such as lumber, LVL, or glulam beams;
versus wood that homogenizes these anisotropic properties through lamination and
randomization of fibre orientations, such as plywood, CLT, or OSB; and so on. Solid
timber retains many of the characteristics of the tree that it was sawn from before the
removal of knots and unwanted sections, whereas veneers and particle boards show
more homogenous, uniform properties—an abstraction or distancing from the raw
source material.
What this means for the design and construction of products or buildings in tim-
ber, is that different timber products have very specific allocations and distributions
of wood fibre, each suitable for a particular set of applications. This adds yet another
dimension to the complexity of wood. In addition to the taxonomy of trees and wood
structure in its raw state, there is an additional taxonomy of how it is deconstructed
and reassembled—in solid pieces, peels, chips, fibres. This creates another taxon-
omy of ‘industrial wood fibre’. This combination of reduction and aggregation—or
subtractive and additive manufacturing processes (Veltkamp 2007)—also suggests
that this process could become iterative, through cycles of machining and lamination.
The treatment and surface finishing of wood presents yet another additional field of
parameters.
To sum up, the integration of wood, its behaviours, and its properties requires a
focus on the principles that underlie them—in nature and in industry—in order to
avoid getting lost in this extremely broad territory. The use of wood in a particular
design context or situation requires an understanding of what the design is asking of
the wood, and how this vast parameter space can be navigated to arrive at an appropri-
Fig. 3 The production hall at Blumer-Lehmann AG and multi-axis timber mill for machining
large-scale glue-laminated components
ate solution. Computational methods—and the digital tools that embody them—of-
fer ways to augment architectural design and fabrication processes by managing and
harnessing this complexity, leading to new and better-informed timber designs.
3 Three Experiments
The following three experiments look at the consequences of grain orientation for
structure and durability. The first describes methods used to represent grain orien-
tation in digital models of free-form components as well as modelling approaches
that augment standard modelling tools with design feedback in the form of addi-
tional material and fabrication data. The second proposes a reshuffling of industrial
processes and a reconsideration of standard glue-laminated timber products, reach-
ing back to the constituent elements in industrial timber production to explore new
glulam morphologies and to create feedback loops between industrial processes. It
also proposes methods by which the fabrication of these new glulam types can be
improved by the integration of direct digital feedback in production. Finally, the third
experiment integrates the ideas and methods of the first two into a broader design
approach that is demonstrated through two architectural case studies.
3.1 Experiment 1: The Fibre
3.1.1 Fibre Orientation and Representation
Since the density, type, and orientation of wood cells are at the root of so many
of the material properties relevant to the design and construction of buildings, their
distribution throughout the wood member is an important point of focus. Wood grain
is generally understood as the overall orientation and arrangement of the long wood
cells, and variations thereof affect the use of wood in many different ways, from its
figure to its processing (Dinwoodie 2002).
Therefore, in a computational modelling approach, there must be a suitable way
in which to map the distribution of grain orientation across a digital model of a
timber component. This encoding is important for three reasons: grain orientation
must be represented in order to communicate differences between otherwise geo-
metrically identical timber components; the model must be tested in relation to this
grain orientation, which means that limits and areas where the grain orientation is
not optimal need to be identified; and the grain orientation needs to be determined
throughout the volume of the model so that it can be passed on to more in-depth
simulations, such as for form optimization or structural analysis. Most importantly,
because free-form elements and glue-laminated assemblies involve multiple timber
elements either bent or laminated in different orientations and arrangements, the
grain orientation throughout a component cannot be assumed to stay constant, and
so the model needs to accommodate a changing grain orientation as well.
3.1.2 Discretization Versus a Multi-scalar Approach
For describing the grain direction at a particular point within a timber component,
two approaches are explored. The first is a discretization of the component model
into triangular or tetrahedral elements, and the second is a multi-scalar approach
which links together the digital model of the architectural element with the digital
model of the timber assembly from which it is machined.
Meshing and Tetrahedralization
Discretization is a way to break down a model into individual elements that can be
examined and treated separately. Structures such as grids, voxels, or meshes allow
changing properties to be plotted across an area or throughout a volume. Each ele-
ment—a grid cell in a grid, a voxel, or a mesh face—may possess properties different
and potentially unrelated to those of its neighbours. Continuous changes in proper-
ties are approximated according to the resolution and distribution of elements within
the discretized model. Querying a particular point inside the model can therefore
return an interpolated value from the closest discrete elements to the queried point.
Fig. 4 FE simulation of a free-form beam. Arrows indicate the directions of the main orthotropic
directions—showing the longitudinal, radial, and tangential grain directions across the model
Fig. 5 A 2D discretization of Hetzer’s drawing, with grain direction displayed as color
This forms the basis for many simulation techniques, including FEA (Fig. 4) and the
design of functionally graded materials (FGM) (Nicholas 2011).
In the case of timber, an ideal discretization of a component model would use
each fibre or wood cell as a single element. For architectural scales, the amount
of data generated would be too dense and impractical. Instead, the resolution of
elements should respond to the rate of change of the wood properties in the component
model—grain orientation, in this case.
In this experiment, triangulations of 2D surfaces are the starting point in the
inquiry. This allows each triangle face to be assigned a unique grain orientation.
Thus, a free-form element can be simulated by varying the grain orientation prop-
Fig. 6 Otto Hetzer’s patent, 1905
erty across the mesh, and a laminated element can be simulated by creating zones
or groups of faces with similar grain orientations. This is tested on a drawing of
one of the first glulam prototypes, first described in 1905 in a patent by glulam pio-
neer Otto Hetzer (Fig. 6). Hetzer’s prototype describes how an actively bent wood
member can be laminated into a larger, straight member, effectively making the first
functionally-graded glulam prototype. The discretization of this model creates a finer
mesh resolution in areas close to the interfaces between the wood components and
in areas where the grain orientation was changing (Fig. 5). This technique is sub-
sequently expanded to three dimensions, using a tetrahedral discretization of solid
geometry or watertight meshes. To accomplish this, existing tetrahedralization tools
(Si 2015) is adapted and exposed in the design modelling environment.4
Although this technique allows the volumetric specification of grain orientation
at a configurable resolution and distribution through the component model, it is not
constrained enough to reflect the material reality of fabricating timber components.
Although the discretization of a volume into many elements opens the door to the
optimization of their individual properties in relation to the volume model and more
general explorations such as structural form-finding, it does not necessarily have
any links to the limits and methods of real material and fabrication. The component
can not be realistically fabricated from thousands of individual elements, therefore
groups of elements have to be used to describe individual pieces of timber, meaning
their specification should be linked somehow to a larger determiner of grain direc-
tion—meaning the grouping of elements must be related somehow to the real, larger
pieces of wood that constitute the real component.
4 In this case, Rhinoceros3d and the Grasshopper plug-in.

Multi-scalar Model
The second technique does away with the discretization of a component model and
instead replaces it with a secondary, linked model which provides information about
the grain orientation of the component as well as additional material and fabrication
parameters. This is conceptually similar to other research into Multi-Scalar Mod-
elling which investigates bidirectional data flows between models at different scales
and resolutions (Nicholas et al. 2016) and similar ways for defining grain directions
in digital models exist in commercial products.5 This multi-scalar approach intro-
duces the ‘blank model’, tailored to timber and glulam components with data about
the sizing and number of constituent lamellas, wood species and grading, overall
dimensions, and type classification—whether it is a straight glulam, single-curved,
or double-curved (Fig. 7). Sampling a particular point in the component model there-
fore returns the grain orientation as derived from the blank model, and not from a
collection of discrete elements. This method effectively links the component model
to a more materially detailed model that embeds and respects material and fabri-
cation parameters, and provides a more accurate description of grain orientation.
The advantage of keeping these two models linked but separate is that different glu-
lam types can be applied to the same component model, thus tailoring the material
outcome of the component to its particular structural or performative requirements.
This also allows the component model to be considered separately from the blank
model—introducing the fabrication data only when desired or needed, and keeping
within the multi-scalar mind-set of having separate models that communicate rather
than a single, centralized, and data-heavy model.
Fig. 7 The blank model (left), the component model (center), and an overlay of material data
derived from the blank model onto the component model (right)
5 For example, Scan&SolveTM is an FEA plug-in which allows a user to define a grain direction of
a timber component using reference geometry such as a curve or a surface, but does not go beyond
that.
3.1.3 Representation of Wood Grain Direction
Although the multi-scalar component-blank linking demonstrates an effective way

of embedding material properties into models of architectural components, a way of
representing and communicating this changing grain orientation is also required. This
is both in order to visually communicate the difference between two geometrically
identical components with differing grain orientations as well as to evaluate a com-
ponent against grain-dependent criteria. For example, for reasons of both structural
strength and durability, it becomes important to identify areas of the modelled com-
ponent where the wood is cut in such a way to reveal end-grain or otherwise destroy
the continuity of the wood fibres, or where exposed end-grain could potentially lead
to a premature deterioration of the component.
The use of the colour channels in digital images to describe additional spatial infor-
mation is probably most commonly seen in a computer graphics technique called
normal mapping.6 This technique encodes extra geometrical data—surface normal-
s—into the RGB channels of an image texture, which is then applied to a coarse
mesh model, thus exchanging a computationally-expensive higher polygon count
with a lightweight model with an extra texture map. Expressing multi-dimensional
data with colour and gradients is explored in other architectural research where this
is important to be visualized all at once (Heinrich and Ayres 2016). Mapping the
longitudinal vector of the grain orientation onto the colour channels therefore seems
a logical extension to grain representation (Fig. 8). This allows changes in grain
orientation in the component model to be quickly visualized as changes in colour,
making qualitative comparisons easy. There are two ways in which this colour change
can be mapped—as a global grain orientation indicator or as a relative comparison
against another vector or guide. The former maps the grain orientation vectors in
relation to a global frame of reference (Fig. 9). This means that, given an assembly
or structure of multiple components, similar colours show grain directions that are
pointing the same way. The latter assigns a colour based on the variation or relative
vector between the grain orientation and some local frame of reference—the local
space of the component, for example. Individual components can be evaluated based
on their grain deviation from their blank models, for example, yielding a kind of
‘grain orientation compliance map’ which highlights areas of the whole assembly
model where the component model begins to deviate from the ideal grain orientation.
The disadvantage of this mode of representation is that it only displays the surface
of the component model, revealing nothing about the interior organization of wood
grain. A further complication is in the mapping of colour to direction, because for the
purposes of timber modelling, a grain direction vector v is identical to its opposite −v.
This means that every grain orientation can be described by two colours on opposite
sides of the colour sphere illustrated above. A better colour mapping method is
therefore required.
6 https://en.wikipedia.org/wiki/Normal_mapping.
Fig. 8 A popular convention for representing directional information (3D vectors) as viewable
RGB colors. Three orthogonal views of the colour sphere created by these vectors
Fig. 9 Examples of using colour to represent grain direction in cross-laminated panels (left) and
in curved glulam members (right)
Fig. 10 A free-form element cut out of a straight glulam blank, represented as a wireframe around
the component (left). The fibre direction mapping shows a constant color (centre). The fibre cutting
analysis shows large areas of cutting past the angle limit (in red, right)
This color-coding based on a relative grain orientation thus makes it trivial to

identify end-grain conditions, simply by comparing a model surface to its underly-
ing grain colour representation. This can be easily done in a mesh-based way, by
analysing the face normals in relation to the fibre direction. The same component
model can thus be linked to different glulam blanks and the qualitative differences
seen at a glance (Figs. 10 and 11).
Fig. 11 A free-form element cut out of a closely matching free-form glulam blank (left). The
fibre direction mapping subsequently shows a changing color map (centre). Fibre cutting is greatly
minimized and continuity is kept intact except for the obvious end-pieces (in red, right)
3.1.4 Designing with Grain: Design Feedback
By using this multi-scalar linking of models and simple colour visualization of grain
orientation, embedded material and fabrication-based data becomes easy to com-
municate in a lightweight and accessible manner. To truly be useful as design tools,
however, these methods must be exposed in a design modelling environment, in a way
that is easy to interface with. This has led to the development of a software library,
which augments existing design modelling tools with ways of creating, manipulat-
ing, and analysing glulam blank models—a toolkit for glulam modelling (Fig. 12).
It is developed as an open-source library, allowing tools and techniques to be tested,
experimented with, and implemented for future use.7 A particular focus is placed
on arriving at a quick material specification from minimal data, while respecting
fabrication conventions and rules—such as the relationship between lamella size
and the maximum curvature of the blank.8 This provides fast, simulated feedback
to the designer about lamella composition, bending constraints, sizing parameters,
material usage, volume, and so on. This design-level feedback—or simply design
feedback—is therefore one of the methods by which an early-stage design phase
could be enriched and augmented with fabrication information (Fig. 13).
The crucial advantage that this notion of feedback presents is the ability to gauge
the consequences of early design decisions for later development stages, and estimate
what the impact on the fabrication process might be.
The toolkit goes further to incorporate the modelling of assemblies of multiple
glulam blanks, hinting at non-standard glulam geometries that are described in the
second experiment, as well as methods for developing a complete workpiece model
for fabrication. This workpiece model specifies joint details and associated machin-
ing strategies, and defines relationships between connected members to ensure that
corresponding joint data is kept intact and up-to-date.
Although these methods are based on their connection to material and produc-
tion constraints, these are not imposed on the designer in the sense of limiting the
7A more detailed description and the source code of this research toolkit can be found at http://
github.com/tsvilans/tas.
8 This particular rule is defined in (DIN EN 14080:2013-09 2013).
Fig. 12 Free-form glulam model, constructed from an input guide curve, key orientation frames,
and material data (left). Lamella sizing and composition of the glulam are embedded in the model,
and can be revealed as the centreline of each lamella in the model (right)
Fig. 13 A glulam beam model with an overlay of the blank model and intersecting beams. Connec-
tion detail geometries are generated parametrically, along with associated cutting strategies (left).
The same glulam beam model with an overlay of faces which exceed the fibre-cutting angle as well
as lamella bending states (right)
outcome of their use, or pushing the design towards a particular solution or formal
language defined by the tools. In this sense, they are quite passive, rather than actively
steering the design process. However, by providing an interface to these material and
fabrication properties, this toolkit provides a basis for the exploration and modelling
of new timber morphologies (Fig. 14).
3.2 Experiment 2: The Blank
The second experiment builds on this mapping and modelling of grain orientation
in glue-laminated timber assemblies by proposing new glulam types based on the
localized adaptation of grain direction and a reconsideration of industrial timber
processes.
Fig. 14 Tools for representation and analysis of new kinds of free-form glulam structures
3.2.1 Blanks Types and Their Implications
As previously seen, a tailored grain direction within a laminated free-form timber

component can lead to members that are more optimized for their particular role.
However, fabricating them remains a challenging and bespoke process. The usage of
standardized lumber and other timber construction materials—products of the mech-
anization of the timber industry—can be scaled to industrial proportions because of
the expedience and automation of their production. As a result, free-form timber
components are typically machined out of a few different types of glulam blanks:
straight blanks, single-curved blanks, and double-curved blanks. These three glu-
lam types display an increasing cost and production complexity—straight blanks are
simply produced by laminating a stack of straight lamellas in a simple press, curved
blanks require forming processes with modular presses, and double-curved blanks
require handling of a higher number of small lamellas, and more complex presses to
deliver an acceptable result within tolerances. This has a direct impact on cost and
risk for the design: higher curvatures and geometries that are more complex require
many more and much smaller lamellas that are actively bent into shape. Smaller
lamellas are increasingly more difficult to handle and organize, not to mention the
higher percentage of waste created in their production.
Bespoke billets and complex laminated pieces are commonly used in furniture
making and small-scale construction, however the reliance on special clamping appa-
ratus, jigs, and the labour-intensity of their fabrication prevents them from being used
more widely and at larger scales. For example, furniture-making very often employs
hot forming—or steam bending—to create curvilinear wood elements. This weakens
the wood, however, and cannot be easily scaled to industrial proportions. Instead, cold
Fig. 15 A map of common laminated construction products, organized by the type of constituent
industrialized timber, with the speculative blank prototypes loosely located in relation to them. In
addition to this categorization, some of the products typically form panels and others are geared
towards producing beams
forming—laminating actively bent lamellas—preserves the strength and integrity of

the wood while also introducing practical limitations in curvature and lamella size.
This means that, in order to operate within an industrial framework, these constraints
must be respected. Further, the taxonomy of industrialized wood—as described in the
introduction: fibres, strands, chips, veneer, and solid wood—can be used to describe
the timber products commonly seen on the market—glulam, crosslam, plywood,
OSB, etc. This describes a design space, within which different configurations of
these base elements could yield new types of glue-laminated assemblies (Fig. 15).
Therefore, the second experiment is divided into two parts: the first considers the
constraints of industrial lamination, common forming processes, and existing glue-
laminated products, and proposes four free-form glue-laminated prototypes which,
occupying the space between established timber products, attempt to bridge the
gap between optimally tailored grain distribution and fabrication realities (Table 1).
Second, the experiment further explores the application of direct feedback through
laser scanning and motion capture to the production process to determine how it can
mitigate some of the difficulties involved in producing bespoke, free-form timber
components.
Table 1 Material probe comparison with respect to process and end-grain exposure
New Workflows for Digital Timber
Voxel blank Finger-joint blank Crosslam joint blank Branching blank

Types of glulam involved Straight, double-curved Straight Straight or single-curved Double-curved
Process steps Lamination Lamination Multiple laminations Lamination
Machining Finger-jointing Machining (final) Machining (final)
Lamination Machining (final)
Machining (final)
End-grain exposure High (along sides) Medium–high Low Low
111
Fig. 16 Voxel blank prototype
3.2.2 Laminated Material Probes
Voxel Blank
The voxel blank addresses the difficulties and costs associated with creating double-
curved glulam blanks. Using simpler laminating techniques and a combination of
larger, straight lamellas and thinner, bent lamellas, this blank seeks to achieve a
stiffness similar to a double-curved blank while avoiding the use of complex presses
and springback. It is based on a similar principle as wooden I-joists, applied to a free-
form (double-curved) beam element. The core—or ‘web’—of the beam is aggregated
from short lengths of lumber, glued side-to-side and oriented perpendicular to the
main overall axis of the beam. It is ‘voxelized’ into a rough shape, making the
lamination process simple and making use of short members—offcuts, perhaps. This
voxelized volume is machined back to the double-curved surfaces of the beam. The
‘flanges’—made up of thinner lamellas—are subsequently laminated onto the core,
actively bent into shape (Fig. 16). This ensures a continuous grain along the double-
curved surfaces of the beam, where it experiences the largest bending stresses. The
main advantage of this ‘voxel blank’ is that the core is formed of simple, orthogonal
pieces of lumber, the double-curved form is then machined into it with the high
precision of a CNC router, and the actively bent lamination of the top chords or
flanges then uses the core as formwork, avoiding the need for precise presses and
lessening the effect of springback.
However, a major disadvantage of this blank is the large amount of exposed end-
grain along the flank of the member and the perpendicular glue interface between
the core and the actively bent flanges. Preliminary analysis using off-the-shelf FEA
Fig. 17 FE simulation of a simply supported voxel blank
Fig. 18 FE simulation of a simply supported straight blank
software shows qualitatively that, under simple loading conditions, its stiffness is not
quite as great as a double-curved blank, but larger than a straight blank (Figs. 17,
18, 19, and 20). This probe demonstrates the potential in iterative lamination and
machining processes, something that is avoided in current industry practice.
Finger-Joint Blank
The finger-joint blank also attempts to avoid the complex forming and small lamellas
associated with double-curved blanks. Instead, it considers the common process of
finger-joining and how it could be adapted to form free-form blanks (Fig. 21). Finger
joining is used throughout the timber industry, most commonly during the removal of
defects in lumber and lamellas in the glulam production process. The adhesives and
cutting pattern used result in a joint that is just as strong—if not stronger—than the
Fig. 19 Tension and compression forces (red and blue) are concentrated on the outer layers of the
voxel blank, where the fibres are continuous and follow the free-form geometry
Fig. 20 FE simulation of a simply supported double-curved blank
surrounding timber element. Although this is typically used simply to remove defects
and to extend individual lamellas to an arbitrary length, the piece-meal aggregation
of a larger member raises the interesting possibility to gradually build up a discretized
free-form member by modulated the orientation of the finger joiner, thus creating
a segmented free-form blank (Fig. 22). This lets the grain orientation approximate
the final free-form more closely than a straight blank, depending on the amount of
segmentation. A smooth, double-curved form can be subsequently machined out of
this finger-joint blank, smoothing out the inflection points around the finger joints.
This technique has been demonstrated in practice previously in the fabrication of a
free-form timber pavilion by Blumer-Lehmann AG and others.9 Again, one of the
main advantages of this process is the reduction of the amount of active bending
9 Z-Plus
Pavilion (2009) by Création Holz, designtoproduction, SJB Kempter Fitze, and Blumer-
Lehmann AG, in Gossau, Switzerland.
Fig. 21 Finger-joint blank prototype
Fig. 22 Individual components of the finger-joint blank
and curved lamellas required, and the usage of larger, straight lengths of timber to
achieve complex forms.
The main challenge for this process is the handling, pressing, and mounting of the
blank, as its segments need to be oriented individually and pressed into one another.
These production complications have prevented it from achieving wider industry use.
Crosslam-Joint Blank
The crosslam-joint blank uses the principle of cross-lamination—seen in products

like CLT and plywood—to explore multi-directional glulam blanks and integrated
joints. It employs cross-lamination as a way to incorporate joints or tight inflection
Fig. 23 Crosslam-joint blank prototype
points in a straight or curved glulam beam (Fig. 23). The main principle behind this
blank is the interleaving of standard lamellas at oblique or perpendicular angles,
much like the alternating layers in a CLT panel. Used at tight inflection points or
as a replacement for cross lap joints, it distributes the grain orientation throughout
the cross-section of the blank or joint—‘weaving’ the lamellas through each other
(Fig. 24). The advantage with this blank is that the grain orientation stays as consistent
as possible along the arms of the curving blank, and that it has multiple end-joints,
which has advantages for joining and assembly.
One important limit is that the faces of the joining pieces must be coplanar at the
joint, and their sections must be mutually perpendicular, owing to the rectangular
cross-section of the lamellas. This restricts the blank to either being an aggregation of
straight glulams or flat, single-curved glulams. The iterative layering and laminating
also requires more laminating cycles as well as precisely cut lamellas to minimize
gaps between the ends of the cross-cut lamellas.
Branching Blank
The branching blank follows a similar logic as the crosslam-joint blank and proposes a
multi-directional blank while keeping the grain orientation consistent along the lines
of the blank as much as possible. However, the multi-directionality is caused by a
splitting of a curved glulam into two halves, as opposed to an interleaving or crossing
as before. Instead, the branching blank employs a cross-laminated layer to act as a
structural bracing between its branching arms. This allows the grain orientation of the
blank to follow the curving arms continuously, while tying the seam together across
the grain. From a design perspective, the variable density of the cross-bracing layer
also hints at a module that is somewhere between a beam and a panel—a free-form
glulam/crosslam hybrid (Fig. 25).
Fig. 24 A typical cross lap joint breaks the continuity of lamellas over half the beam cross-section
(top). The crosslam-joint blank instead alternates this break in continuity between layers
Because it relies on forming two separate double-curved glulams, the fabrica-

tion process for this blank is particularly difficult. They can either be formed sep-
arately and joined afterwards—though then there is the question of integrating the
cross-laminated layer between the two arms—or the blank can be laminated all at
once—though this places constraints on the forming process and limits the amount
of accessible space around the blank within which to place clamps or formwork. A
series of blanks was created to test this process, with mixed results: several suffered
delamination because of insufficient consistency in clamping during lamination.
Fig. 25 Branching blank prototype
3.2.3 Feedback During Production
The move away from standardized, orthogonal timber elements and towards more
free-form, formally complex prototypes also presents challenges in their subsequent
processing and quality control. Despite being based in existing and well-established
industrial processes, the geometry and material specification involved necessitate a
new measure of direct feedback during fabrication. On a practical level, the unde-
sirable material behaviours that arise during the processing of free-form timber ele-
ments—such as springback and dimensional change—need to be detected and mea-
sured in order to mitigate their effect. On a deeper level, the continuous observation
of the material during production creates the potential for discerning patterns and
interpreting the accumulated data for future use—a kind of ‘encoded experience’ for
the production process—or, at the very least, a new input into the wider information
model of the whole design.
This is a key part of the workflow, as it forms the interface between the digital space
of modelling and material probes, and the space of material production. Scanning
and digital spatial capture are therefore employed as the most direct form of feedback
between the computational environment and the machine space. The problem of reg-
istration of free-form glulams before machining was confronted at Blumer-Lehmann
AG. Several projects dealt with non-orthogonal timber components at a large scale,
with no clear reference point or flat surface. This necessitated manual methods of
measuring corners and points, adjusting the machining model, and a slow process
of trial-and-error (Fig. 26). In addition to being the source of uncertainty and lost
time, this necessitated the use of jigs, templates, and other additional procedures to
ensure a suitable outcome. In response to this lack of direct feedback, a series of
Fig. 26 Free-form elements present a challenge for production
Table 2 Overview and comparison of the direct feedback methods employed and tested
Technique Spindle laser Rangefinder Motion capture LiDAR
Interface Machine tool Machine tool, Haptic, mounted Mounted device,
software device, software software
Point of use On the machine On the machine Handling, Arrival, machine
machine registration,
registration quality control
Immediacy Pre-programmed Pre-programmed Real-time Offline, requires
path, real-time path, real-time, processing
some software
processing
Type of feedback Visual Sparse data Sparse data High-resolution
points, visual points, haptic point cloud
scanning and motion-capture tests was conducted in the workshop, during an active
production of a free-form glulam project. Apart from speed and reliability of the
data, factors such as usability, robustness, and integration into the existing machin-
ing workflow were prioritized. Operating in a harsh industrial environment meant
that the equipment and techniques used needed to be replaceable and adjustable.
The direct feedback methods were implemented in order of complexity, from the
simplest, most rudimentary to more expensive and elaborate ones, involving more
specialist equipment (Table 2).
Fig. 27 Spindle-mounted laser pointer at Blumer-Lehmann AG
Spindle Laser
To begin with, the first method consisted of simply inserting a cheap laser pointer onto
the spindle of the production mill (Fig. 27). To account for the inevitable axial offset
in the laser, the spindle was set to spin at a slow speed. This traced a light circle onto
the material, from which a centre point could be inferred with reasonable accuracy.
Toolpaths from the free-form glulam workpiece were generated, which traced the
primary workpiece edges at a small distance offset. By observing the spinning laser
circle on the surface of the glulam blank, the machine operator could visually evaluate
the positioning of the material in the machine working volume. Tracing the edges
from multiple planes allowed a decent three-dimensional estimation of the location
of the workpiece model in relation to the material.
This first method was the simplest and easiest way to integrate a measure of
direct feedback into the production workflow, requiring very little extra work and no
specialist knowledge to implement, however it was also the least accurate and did
not have an interface to record any data or feed back into the software model. Any
change to the machining model still had to be done manually by the operator.
Rangefinder
The second method elaborated on the first by replacing the laser pointer with a
custom-built laser rangefinder device that communicated wirelessly with the oper-
ator’s workstation (Fig. 28). A custom housing was built out of waterjet cut and
welded steel, with an ABS cover to ensure wireless transparency. The axial offset of
Fig. 28 Custom-built laser rangefinder mounted in spindle
the laser measurement point was adjusted by calibration screws built into the device.
A point sample was recorded by moving the rangefinder to a specific pose and offset-
ting the known machine coordinate by the distance detected by the rangefinder. The
mounting of the device into the machine spindle meant that it used the same coor-
dinate system as the machine, preventing the need for any transformations between
coordinate systems—a potential source of error. A software interface was written to
receive, process, and visualize the real-time sensor data, and access the data through
the digital workpiece model. This provided a much more precise measurement of the
glulam blank with clear numerical output. The accuracy of the rangefinder enabled
measurements to within a tenth of a millimetre. Having a recorded set of data points
from the measuring process allowed the digital model of the workpiece to document
the fabrication process. The set of data points could be further used with an opti-
mization algorithm10 to propose an adjustment to the positioning and orientation of
the machining model to best fit the material on the machine.
While augmenting the previous method with a higher accuracy and by providing
a link to the digital fabrication model, this method was only usable when the blank
was already loaded onto the carriers for milling, meaning any time spent measuring
was time taken away from production and, in the event that the measurements were
not satisfactory, much time was lost unloading a defective blank.
10 A variation of the Metropolis-Hastings algorithm and a Simulated Annealing algorithm were
implemented in this case for quickness and ease of integration, though more robust implementations
of different registration algorithms are widely available, and have since been used. A good source
of these is the open-source Pointcloud Library (PCL).
Fig. 29 Real-time motion capture system in the Blumer-Lehmann AG production workshop
Motion Capture
The third method tested the use of real-time optical motion-tracking11 for measur-
ing precise points on the glulam blank and monitoring their movement (Fig. 29).
This involved setting up an array of cameras around a working volume and the use
of retroreflective markers and stickers. Computer vision techniques transform the
recorded bright ‘blobs’ of the reflectors into projected coordinates on each camera
frame, which are then subsequently transformed into 3d points using the relative
transformations between the cameras. By knowing the precise relationship between
cameras and their intrinsic parameters, points that are seen by multiple cameras
can be translated into spatial coordinates.12 The real-time nature of this system also
allowed for its use as a 3d measuring tape: a hand-held tracking marker could be
placed on points to be measured, and its precise spatial coordinates recorded in a
modelling environment, where a numerically accurate distance measurement can be
easily found. One advantage compared to the previous two methods was that this
system could be setup elsewhere in the workshop and, as such, its use did not inter-
fere with the production cycle of the mill—for example, measurement of incoming
blanks could be done well before they were loaded onto the carriers for machining,
saving time in the event of a defective blank. The amount of data collected with this
method was comparable to the last method, however the ability to hold a tracking
marker added a very haptic and intuitive dimension to its use.
11 NaturalPoint’s Optitrack system was used for this experiment.

12 A typical method for doing this is called Sparse Bundle Adjustment (SBA).
Using this system to track real-time behaviour of a blank during machining was
proposed but was not explored for logistical reasons. The tracking in real-time of the
movement and deformation of large-scale timber building components remains an
ambition and needs to be explored.
LiDAR
The last method tested during this time was the integration of large-scale LiDAR
scanning in the workshop (Fig. 30). As a 3d scanning technology, terrestrial LiDAR is
a highly accurate and high-resolution technique for creating large scanning datasets
in the form of point clouds. Compared to the previously described methods, this
creates a large measure of redundancy and extra information that can both offer
benefits and create complications for the wider workflow. More data allows a more
complete image of the scanned environment to be created, however the mass of
extraneous information often requires selective filtering, isolating, and reduction of
useful data, not to mention the larger strain it puts on data transmission and storage
requirements due to increased file sizes. At Blumer-Lehmann AG, this was used to
3d scan free-form glulam blanks both in the staging area of the workshop—upon
arrival, before machining—as well as during pre-machining—while being mounted
onto the machining carriers. Scanning during the staging phase was used to check the
geometric accuracy of the arriving blanks in order to catch errors early and ensure
the quality of the incoming product, before mounting and machining the erroneous
blank (Fig. 31). Scanning during the pre-machining phase was introduced as a way to
avoid the use of templates and alignment between the material placement and digital
model, as described in the previous methods. The usage of more point samples and
a larger point cloud meant that a more robust and complete fit could be estimated
between the point cloud and the model, at the cost of having to align the coordinate
system of the scanner with that of the mill—a crucial calibration step and potential
source of alignment error. This method also came with a much steeper learning curve,
requiring the usage of low-level software APIs and hardware interfaces.
3.2.4 Sensor Systems in Production: Direct Feedback
What these four tests and methods demonstrated was a form of direct feed-
back—manifested in different ways—that helps to close the gap between the digital
model and the material reality of free-form timber elements. In this way, it forms
a key step in the design-to-production workflow: knowledge of material behaviours
can be used to predict issues during production, but direct feedback during pro-
duction is crucial to understanding their effect and to contributing to an improved
future model. Further, in relation to the design and prototyping of complex new glu-
lam typologies, 3d scanning offers a robust way to efficiently process free-form or
non-rectilinear timber elements. This allows the timber production process to inter-
face with non-standard products, such as the material probes described in the first
Fig. 30 High-density point cloud of the Blumer-Lehmann AG production hall
Fig. 31 High resolution point clouds are used to compare the blank to the digital component model
part of the experiment. It also connects the material performance of timber during
the production process to its data or information model. This last possibility has
large implications for the information model: a sensor-laden timber production envi-
ronment could have a role to play in an integrated information model that spans
design, development, production, assembly, and occupancy—not just one of doc-
umenting the produced material and controlling quality, but as a live input into a
learning model. The feasibility of integrating new glue-laminated material proto-
types into industrial production requires adjustments and shifts in current industrial
Table 3 An overview of the two case studies and a comparison of the main design drivers and
aspects of the material concept
Grove Bridge
Major design Structure Function
drivers Fabrication Durability
Site constraints
Timber Experimental branching module Combination of solid timber,
components single-curved glulam, double-curved
glulam, and actively bent cladding
Material focus Material bending limits End-grain conditions
Assembly sequence Assembly sequence
Data management
timber workflows—ones which dispense with linearity and embrace feedback and
recursion as a regular part of the production process. The digital bridge created by
integrated digital sensors and feedback loops points to a way with which this could
be achieved.
3.3 Experiment 3: The Structure
3.3.1 Architectural Consequences and Early-Stage Design
The third experiment in this research attempts to connect the previous two experi-
ments to larger-scale processes of architectural design and development. At the scale
of a structure or building, different considerations come into play with regard to per-
formance and feedback. The integration of the two previous domains—the scale of
the fibre and the scale of the blank—raises questions of modelling, data organization,
and other performance metrics such as durability, weathering, and maintenance. It is
also an opportunity to explore the formal and tectonic potential of these new glulam
prototypes and this new mode of working with timber in a computational way. Two
case studies—one coming out of an academic setting and one in partnership with
White arkitekter—demonstrate the applications of the previous experiments to the
early-stage design and production of free-form timber buildings (Table 3).
Early stage design in practice is a condition with many unknown factors that
may include the timeline, the exact program, potential design strategy and concepts,
as well as material used. To address material performance in this contexts entails
working with hypothetical scenarios—assuming certain conditions while informing

early stage design with what is known from other situations—such as experiments
in practice.
3.3.2 Grove
Grove was a submission to the international Tallinn Architecture Biennale 2017 folly
competition (Fig. 32). This submission served as a test for ongoing development of
workflows and tools described in the previous two experiments, in combination with
notions of multi-scalar modelling and data management explored in another research
project at CITA (Poinet et al. 2016). Development of these workflows warranted a
demonstration within the wider context of an architectural design project, a site, and
a brief. The competition brief—to build a small timber pavilion in Tallinn, with a
budget and site considerations—provided an opportunity to test the research against
outside factors and reveal important shortcomings or aspects that require further
research and development. Applicants were invited to propose a pavilion that would
remain standing until the next Biennale, and which would be built in collaboration
with a local Estonian fabricator.
The Grove proposal explored the extension of the speculative blank types—or
material probes—described in the previous experiment into a building system. The
branching blank was considered as an architectural module and stimulant of a larger
Fig. 32 Digital rendering of Grove competition proposal

architectural form, and was developed in further detail for mass customization and
production. Architecturally, the pavilion proposal took the form of a set of inter-
connected catenary vaults, with multiple, branching legs, which spread out into the
surface of the vaults—or canopy of the pavilion. As the modules branched and prop-
agated, they thinned out, creating a density gradient from the thick leg supports to
the screen-like vault surface. This introduced a fluidity both of form as well as of
structural role: the relationship between beam, column, and panel was obfuscated.
Instead, the main leg supports dissolved into a light surface, and vice versa. The
intertwining of tree groves served as inspiration for this move—relating the leg sup-
ports to the trunks of trees, the vault surface to the enmeshed canopies of trees, and
the branching modules to the splits and branchings of the same. This type of tectonic
came directly out of the blank prototypes and the structural and formal possibilities
of free-form glulam.
Grove also presented a challenge for the integration of the different scales and
management of the relevant data. Building on the linked blank model and multi-
scalar approach in the first experiment, a similar multi-scalar approach was used to
simultaneously model the overall pavilion and to develop the individual blanks and
connection details for each of the 1000 modules (Svilans et al. 2018). At the level of
the module, the modelling toolkit (see Experiment 1) was expanded to support assem-
blies of multiple blanks and a workpiece model was implemented to enable tracking
relationships between modules and shared joint details. This permitted the overall
design to change and fluctuate, while keeping local relationships intact. For each
blank, these joint details and blank specifications formed the basis for the fabrication
model, which specified the sizing and amount of wood, the geometry of the nec-
essary mould for the blank, and the different machining surfaces. Organization and
grouping of the joint details and machining surfaces meant that a semi-autonomous
process could be used to generate toolpaths and machining strategies. This multi-
scalar management of the whole design-to-fabrication process allowed all fabrication
data to be accessed quickly and easily. This became important during discussions
with the local fabricators, as the daunting complexity of the proposal had to be made
friendlier. Demonstrating that the numerous components could be tracked and docu-
mented in detail—and that the proposal was based on material logic and fabrication
principles—helped gain the support of the jury and fabricators.
Ultimately, the sheer number of components and complexity of fabrication made
the proposal unsuitable for the budget and timeline of the competition, however the
confidence in the overall strategy and design framework still made it a worthwhile
effort for developing the research. Instead, it was awarded second place from among
16 international short-listed submissions.
3.3.3 The Bridge
The development of the Magelungen Park Bridge is part of the redevelopment of the
larger surrounding site by White arkitekter, including parkland as well as new resi-
dential development at a location south of Stockholm (Fig. 33). The initial program
Fig. 33 Early concept rendering of Magelungen Park Bridge
asked for a pedestrian bridge, assuming this would be a standard bridge in steel or
concrete. The bridge connects an elevated position at the entry of a park area, spans
a railroad, road and bikers’ lane, to connect nearby a bus stop and entry to the new
residential area while passing to a small forest of protected oak trees. The specific
site conditions, bridging two locations with a large difference in height while passing
through an area trees to be preserved, gave the bridge an organic form in plan (it had
to be almost looped in order to encompass the height difference while maintaining an
even slope of 5%). This in turn suggested an alternate strategy, where design could
play a bigger role, and an alternate material—timber.
At the time of writing, the Magelungen project is still at early stage design,
in anticipation of a revised planning document (Swedish “detaljplan”). The specific
conditions of the bridge demanded a much more refined development however, given
that aspects such as material choice, structure and dimensioning affects the possible
overall form of the bridge, which in turn affects bridgeheads, positioning, and in
particular, how the bridge can stay free of the protected oak trees of the site. This
required a series of design investigations where a structural principle had to be
proposed and assessed, and criteria such as weather protection of end grain as well
as clearance of snow had to be considered. This was conducted in stages, based on
previous knowledge in regards to limitations in stock material, as well as feedback
from structural engineers.
An initial frame-based construction with many columns later gave way to a diagrid
truss structure that could span over the road and railroad with fewer columns. The
over-arching organic form is emphasized with a rounded profile in cross section,
accomplished through repeated vertical panels formed in situ.
Fig. 34 Modelling of a bridge section, using the tools described in Experiment 1
The Magelungen Park Bridge project presents a different set of concerns than
Grove. Performance considerations such as durability, maintenance, assembly, and
so on in this case are stronger design drivers than concerns about data management
or developing a single material probe into a whole system. The extreme differences
between seasons in Stockholm present a harsh environment for exterior wood appli-
cations. The consequences of weather conditions such as standing water or snow
removal have to be addressed at the very beginning, to help steer the material solution.
In the case of wood, this sets two immediate guidelines for material development:
standing water must be avoided and end-grain conditions must be minimized. The
choice of wood species also becomes an important factor. Joints must take care with
contact surfaces, to prevent water seepage into the structure. Further, due to the two
high-traffic transportation routes running through the site, as well as the wish to min-
imize the impact on the adjacent forested area, the construction and assembly of the
bridge should be as quick and modular as possible. Finally, several structural layers
are present: the primary structure, secondary structure, timber cladding, handrails,
and the deck for pedestrians and cycling.
This presents an opportunity to develop bespoke timber components, which
address these criteria and integrate as many design parameters as possible into a pre-
fabricated solution. The glulam software toolkit described in Experiment 1 has been
used to develop an overall structural strategy using straight and curved glulams, how-
ever current development is looking at how these tools could be extended to propose
a more tailored solution (Figs. 34 and 35), akin to the material probes in Experiment
2. In particular, properties such as end-grain exposure become much more impor-
tant evaluation criteria for further probes, due to the focus on durability and water
shedding. Further probes are needed to explore these types of glue-laminated timber
assemblies that integrate multiple material criteria through the iterative aggregation
and machining of industrialized timber.
Fig. 35 Concept rendering of bridge section
4 Conclusion
There are several conclusions that come out of these three experiments in relation to
the digital design of timber buildings and the way in which they are produced today.
4.1 Digitally Designed Wood
The first is that a much wider space of possibility for timber structures is created by
both considering the principles of timber behaviour—such as its anisotropic char-
acter and multitude of properties—and reaching back beyond the arsenal of modern
timber products to the even more basic elements that constitute ‘industrialized tim-
ber’. It is through knowledge of how chips, fibres, veneers, lamellas, and solid wood
are transformed and aggregated into standard product forms that new material pos-
sibilities emerge. The combinations afforded by this new palette can lead to tailored
building elements and material forms that challenge notions of a standard sheet or
dimensioned product. The relationship between form and material can therefore be
relaxed by not having to pick between two products that might not quite fit the design
requirements, but rather having the opportunity to invent a more suitable third one
from the basic principles of industrialized timber.
This obviously has severe implications for the regulatory and communicative
processes governing the use and quality of timber products. Testing methods need
to be adapted, and simulation frameworks need to be expanded. This aspect needs
further thought about how designers and architects can begin to engage more openly
with material testing and standards, however material prototyping has always been
an integral part of the architectural design process.
4.2 Extended Digital Production
Secondly, the impact of this way of thinking on industrial producers and fabricators is
substantial, as it requires a rethinking of the typical supply and production processes
in glue-laminated timber. Consolidation or reconfiguration of these can be tricky, as
the barriers are often political, and not just technical. For example, moving glulam
production and component machining under the same roof makes a lot of practical
sense—communication would be immediate and adjustments could be implemented
right away—however supply arrangements between various independent producers
can prevent this from happening. On a technical level, a production apparatus that
can implement an iterative laminating and machining process—as explored in the
second experiment—needs to be constructed in such a way that the handling and
repositioning of multiple workpieces is minimized throughout their production cycle:
the work must not stop, but glue needs time to dry, but tolerances need to be kept.
The key in this case might be as simple as the difference between a static machining
portal over a moving workpiece, versus a moving portal over a static workpiece.
In the latter case, the track along which the machine runs could theoretically be
extended infinitely, with any number of workpieces fixed along its length, meaning a
mobile machine can engage with one workpiece while the glue sets on another. Used
in conjunction with pressing equipment along its length, and coordinated by highly
accurate laser scans and spatial capture markers, this kind of integrative production
workflow could begin to explore alternate laminated forms at an industrial pace and
scale (Fig. 36).
4.3 Digital Interfaces Across Disciplinary Silos
Finally, designers need access to timber knowledge and a measure of feedback

throughout the design process. As obvious as this might sound, the way to achieve
better buildings in timber has to begin with understanding what the basic building
blocks involved are and what the crucial criteria that the design needs to fulfil is in
Fig. 36 A production environment that integrates constant digital feedback and a connection to the
information model
the first place. That is to say, the material concept needs to be designed before even
beginning to address the architectural solution. On one hand, this knowledge transfer
can seem more pedagogical or organizational but, on the other, this material concept
also needs to find a concrete use as a tool during the design stages of a project, and
inform the user about the consequences of design decisions in approximate but fast
terms.
The computational modelling and analysis tools presented in this chapter attempt
to show how material models may be linked to architectural information models in a
non-obtrusive way. By taking a backseat role of illustrating the material consequences
of a model instead of imposing them, the feedback is provided transparently—the
design process is left untainted by the tools. This is contrast to the generative or
parametric literature which describes design processes where the designer sets some
settings and leaves everything else to the algorithm (Kolarevic 2003). In this case,
the feedback is there, but it is up to the designer to use it.
4.4 The Way Forward
To conclude, the link between the early stages of a design and the material knowl-
edge of the fabricator is one that must primarily begin with open communication.
To this end, architects have always employed prototypes to explore and demonstrate
the material embodiment of their ideas. Among both architects and fabricators, the
mockup is still the best way to test an architectural idea, short of building the project,
but it also represents a whole series of made decisions and a commitment to a specific
design path by the time it is commissioned. However, the maturing accessibility to
advanced tools of digital fabrication and new technology is driving material devel-
opment in architecture offices all around the world, bringing the tools of production
closer to the beginning of the design process. This is a positive step, as it allows
the possibility of the material concept developing with and feeding back into the
architectural concept.
Still, the tools by themselves are not enough: laser cutters and robotic arms in
design studios are a far cry from the mills of the timber industry, with all the cru-
cial background knowledge of supply, logistics, and execution, not to mention the
experience of material and process. But perhaps it is through more fluid and open
information models, partnerships, or research initiatives that straddle both worlds,
that the interface between design and fabrication could allow for the ‘digital contin-
uum’ (Kolarevic 2003) to become a deeper material continuum as well.
Acknowledgements This project was undertaken at the Centre for IT and Architecture, KADK, in
Copenhagen, Denmark as part of the Innochain Early Training Network. This project has received
funding from the European Union’s Horizon 2020 research and innovation programme under the
Marie Sklodowska-Curie grant agreement No 642877.
Special thanks to the industry partners and students of the CITAstudio Masters course.
References
DIN EN 14080:2013-09 (2013) Eurocode 5: design of timber structures

Dinwoodie J (2002) Timber: its nature and behaviour. CRC Press
Hankinson R (1921) Investigation of crushing strength of spruce at varying angles of grain. Air
Serv Inf Circ 3
Heinrich M, Ayres P (2016) For time-continuous optimisation: replacing automation with interactive
visualisation in multi-objective behavioural design. In: Living systems and micro-utopias: towards
continuous designing, proceedings of CAADRIA 2016. Melbourne, pp 373–382
Hoadley RB (2000) Understanding wood: a craftsman’s guide to wood technology. Taunton Press
Kolarevic B (2003) Architecture in the digital age: design and manufacturing. Taylor & Francis
Mumford L (1934) Technics and civilization. Routledge & Kegan Paul Ltd., London
Nicholas P (2011) Embedding designed deformation: towards the computational design of graded
material components. In: Ambience conference, pp 145–151
Nicholas P, Zwierzycki M, Ramsgaard Thomsen M (2016) Concepts and methodologies for multi-
scale modelling: a mesh-based approach for bi-directional information flows. In: Proceedings of
the 36th annual conference of the association for computer aided design in architecture (acadia),
pp 308–317
Otto F (1992) IL 23: form force mass 3—structure. Institute for Lightweight Structures
Poinet P, Nicholas P, Tamke M, Ramsgaard Thomsen M (2016) Multi-scalar modelling for free-
form timber structures. In: Proceedings of the IASS annual symposium, spatial structures in the
21st century
Radkau J (2012) Wood: a history. Polity
Si H (2015, Februrary) TetGen: a Delaunay-based quality tetrahedral mesh generator. ACM Trans
Math Soft
Svilans T, Poinet P, Tamke M, Ramsgaard Thomsen M (2018) Multi-scalar approach for the mod-
elling and fabrication of free-form glue-laminated timber structures. Humanizing digital reality.
Springer, Singapore, pp 247–257
Tamke M, Riiber J, Jungjohann H, Ramsgard Thomsen M (2010) Lamella flock. In: Advances in
architectural geometry 2010, pp 37–48
Tamke M, Hernández EL, Deleuran A, Gengnagel C, Burry M, Ramsgard Thomsen M (2012)
A new material practice: integrating design and material behavior. In: Proceedings of the 2012
symposium on simulation for architecture, p 3
Veltkamp M (2007) Free form structural design: schemes, systems and prototypes of structures for
irregular shaped buildings. IOS Press
Negotiated Materialization: Design
Approaches Integrating Wood
Heterogeneity Through Advanced
Robotic Fabrication
Giulio Brugnaro, Angelo Figliola and Alexandre Dubor
Abstract Whilst robots are predictable, repetitive, predefined and constant, natu-
ral materials present unpredictable complexity. Over the past few centuries, mate-
rials have been standardized to fit industrial processes, in an attempt to defy this
unpredictability. Thanks to new advances in sensing technologies and computational
design, today we have the opportunity to reintegrate the intrinsic properties of natural
materials in their full complexity. What is the potential of a synthesis between the
particularity of each specific material element—specific properties and parameter-
s—informing the fabrication process? Digital and Robotic Fabrication are based on
the use of flexible machines that open the possibility to mass-customize the produc-
tion process. Combined with sensors and computational analysis, they allow to work
with “soft systems”, both adaptable and continuously evolving, whose dynamism is
constantly fed by a flow of information. How can the designer integrate this uncer-
tainty and complexity in the design process? In this paper the authors specifically
discuss the management of structural and material tolerance inherent to large scale
construction and anisotropic materials, such as wood. A series of projects developed
and built at the Institute for Advanced Architecture of Catalonia and the Bartlett
School of Architecture are used as case studies to investigate tolerance management
in Digital Fabrication with different kinds of wood.
G. Brugnaro
The Bartlett School of Architecture, University College of London, London, UK
e-mail: g.brugnaro@ucl.ac.uk
A. Figliola
School of Architecture and Design, University of Camerino, Ascoli Piceno, Italy
e-mail: angelo.figliola@uniroma1.it
A. Figliola
Sapienza University of Rome, Rome, Italy
A. Dubor (B)
Institute for Advanced Architecture of Catalonia (IAAC), Barcelona, Spain
e-mail: alex@iaac.net

136 G. Brugnaro et al.
Keywords Wood design · Parametric design and fabrication strategies

CNC and woodworking technology · Complex wood structures
1 Introduction
Within the context of industrial production, robotic fabrication processes are devised
as short, constant and repetitive tasks, however, when handling natural, heteroge-
neous, materials, this requires to continuously adapt to the specificity of the mate-
rial’s specific properties and behaviours. Recent advances in sensing technologies
and computational processing allow to integrate such properties within our design
tools through different layers of complexity. Once materials can actively inform the
fabrication process, how can the synthesis of such information lead to novel design
opportunities? In the selection of projects presented in this chapter, the flexibility of
industrial robots provides the opportunity to explore the consequences of such new
material understanding towards the customisation of production processes.
1.1 Dynamic Blueprints and Cyber Physical Making
The integration of sensors and computational analysis allow to devise manufacturing

processes as “soft systems”, adaptable and continuously evolving frameworks whose
dynamism is supported by a constant flow of feedback information (Kwinter 1993).
The ability to dynamically act and react, presenting a certain quotient of intelligence,
make such systems sensitive to variations due to the interaction between the different
actors and components involved in the overall process, such as the anisotropy of the
materials used. This therefore progresses from a hard system of standard industrial
processes, to a soft one governed by an intelligent responsiveness: combining the
customization of tools, use of sensors and computational capabilities (Alemany and
Portel 2014; Vasey et al. 2015; Brugnaro et al. 2016). Such adaptive relationships
are defined by Achim Menges as “Cyber Physical Making” (2015) and described as
follows: “In behaviour-based making, data is continuously gathered and feedback to
the system, which means that new information is gained on the run and new insights
can be had; or in other words, design can evolve in the process of making”. According
to this definition, the role of computational design in processing real time data and
calibrating the digital model in accordance with the production sequence, is evident.
Hence, it is possible to define adaptive systems in relation to the variable parameters
of the manufacturing process, the material used and the interaction between the two
components that inform the digital model in real time. The described methodology
makes it possible to use the robot not as a simple performer of programmed actions,
returning to the definition of servo supplied by Capek (Capek et al. 2004), but as
a design agent able to materialize virtual and digital space through an adaptive and
responsive process (Menges 2015a, b). By extending the receptivity of the robot
Negotiated Materialization: Design Approaches Integrating … 137
through a series of sensors, it is possible to implement a system of feedback loop

relations between man, machine and material. Engaging with feedback loops sig-
nifies building relationships and designing a dynamic, non-linear system that can
guarantee adaptability to the local scale (Gramazio and Kohler 2014). Mario Carpo,
in conversation with Mathias Kohler during the 2014 International Fabricate Con-
ference, discusses the feedback loop process: “The topic you were talking about,
automatic feedback between the machine and the material it’s working with or an
intelligent machine that can interpret the resistance of the material, is the next step of
digital craftsmanship”. The machine becomes an extension of the artisan/designer,
who has perceived and analyzed the physical properties of a material through his
own hands (Carpo 2014). In this new context of production, design needs to negoti-
ate traditional (top-down) decisions of form and function with this new (bottom-up)
decision of materiality and making. A negotiation between digital design and digital
craftsmanship that requires dynamic blueprints rather than static ones. “Dynamic
Blueprints (Dubor and Diaz 2013)” offers a new tool for the designer to negotiate
the materialisation of the project, where uncertainty, inaccuracy and unpredictability
coming from the material and the fabrication process could be integrated in the design
within predefined boundaries and logics, in accordance with the overall design goals.
1.2 Digital Negotiations in Wood Constructions
How can designers integrate this material feedback in the design and making of
wood constructions? The technologies available today and the high level of integra-
tion between them allow to establish an intertwined logic among the various phases
of the design and construction process. The unpredictability and the complexity that
comes from the use of low-engineered materials such as wood and the fabrication set-
up, needs to be managed establishing a virtual connection between digital and phys-
ical model. This connection can be materialized in the different phases of the design
process using IoT protocol and various technologies such as 3D scanner, drones or
VR devices that helps to create a feedback-loop logic among digital and physical
models ensuring the materialization of the complex flow of information of the virtual
space. Being able to control and manage the behaviour of the materials enhance the
possibility to exploit the full potential of the computational design process without
referring to industrialized material and codified tectonics. This means designing in
accordance with the material and with machining tools defining a system of relations
able to self-organize according to dynamic external inputs. The intertwined logic
can have a relevance both in terms of the design process and shape generation and
in the assembly and production of technological components. The use of this design
methodology makes it possible to fabricate complex material systems for which it
is difficult to predict the result through digital simulation due to the non-linear rela-
tionship between material, manufacturing tool and external parameters (e.g. ambient
temperature and humidity) which makes it obsolete the use of digital manufacturing
paths due to their lack of adaptability to changes in the production process (Dubor
et al. 2016). Through the use of sensors, it is possible to adapt the parameters of the
production process with respect to real values and to approximate the result obtained
with respect to the digital model. The process experimented through the workshop
highlights two fundamental aspects with respect to the interactive process: the pos-
sibility of transferring complex and non-linear material systems from the digital and
virtual space to the physical one, and the result of which is not predictable through
digital simulation and adaptability to the tolerances of the production process. Over-
coming tolerances due to the relationship between material and production process
turn out to be a fundamental aspect in the materialization of complex morphologies
through robotic fabrication. The progressive accumulation of errors during produc-
tion process can be eliminated by programming a computational system based on
feedback loop logic, giving rise to adaptive assembly processes (Jeffers 2016; Dörfler
et al. 2016; Wu and Kilian 2016; Schwartz et al. 2016).
2 Tolerance Management in Digital Fabrication: Case

Studies
Following this thinking, the authors of this paper discuss specifically the management
of tolerance as “an allowable amount of variation of a specified quantity, especially
in the dimensions of a machine or part” inherent to large scale construction and
anisotropic material such as wood. A series of project developed and constructed
at IAAC are used as case studies of tolerance management in Digital Fabrication
with different kind of wood. The first project “Fusta Robotica” (2015) overcome
material and fabrication imprecision by integrating redundancy and tolerance in the
design and by using “green” wood that would later on dry, twist and lock the overall
shape into position. The Second project “Digital Urban Orchard” (2016) overcome
the challenge of imprecision by dividing the construction in two separate phases:
First production and construction of the raw wood structure; and in a second phase
3D scanning of the installed structure before manufacturing and assembly of the
additional layers of construction (skin, pipes, furniture, etc.). The Third Projects
“Robotic Sawmill” (2013) and “Digital Woodcraft” (2017) overcome the material
unpredictability of wood logs by 3D scanning each piece before fabrication, integrat-
ing its very specific shape in the design and robotic fabrication process. Finally, the
Fourth Project “Adaptive Robotic Carving” (2017), developed at the Bartlett School
of Architecture in London, focuses on the adaptive training of a robotic fabrication
system aimed toward the integration, within a design interface, of the instrumental
knowledge necessary to deal with timber properties and behaviors, through a com-
bination of sensor data and machine learning. The analysis of those four processes,
in addition to the definition of the limits and potentialities of the design approach
for each case, opens the debate on the role of technological experimentation in the
post-digital era, characterized by the process information that includes digital com-
putation, materials scanning and robotic fabrication. These case studies challenge
the conventional link between drawing and construction, looking at new dynamic
ways to design, towards “Dynamic Blue Print” capable to integrate the variety of
property that a material such as wood could offer to digital construction.
2.1 “Fusta Robotica” and “Digital Urban Orchard”
Fusta Robòtica Pavilion1 (Fig. 1) and Digital Urban Orchard2 (Fig. 2) have been
developed as part of the OTF 2015/2016, Open Thesis Fabrication, the Applied
Research program whose objective is to design and fabricate a full-scale functional
prototype to be implemented in urban public spaces. Material, functional, struc-
tural, and environmental criteria performance have been adopted to directly inform
and materialize the computational design generation through robot-assisted fabri-
cation and assembly processes. These two 1:1 scale wooden prototypes implement
the methodological application of the mass-customization paradigm in architecture
facing the issue of wood unpredictability in all the phases of the design process,
Fig. 1 IAAC, Fusta Ròbotica 2015. © Alexandre Dubor
1 https://iaac.net/educational-programmes/postgraduate-open-thesis-fabrication/past-editions/
fusta-robotica-otf-2015/.
2 https://iaac.net/educational-programmes/postgraduate-open-thesis-fabrication/past-editions/
digital-urban-orchard/.
Fig. 2 IAAC, Digital Urban Orchard 2016. © Andrea Quartara
from idea development to construction and assembly. They physically actualize two
different agenda and they propose two different methodological approach to digital
materiality in relation to wood properties.
2.2 Fusta Robotica: Design Against Material
At first, Fusta Robòtica Pavilion (Figliola and Dubor 2017), resulting from a collab-
oration between academic institution and industry, is a tectonic attempt to overcome
the limits of building with Catalan wood, highly deformable hence unsuitable for
construction. Fusta Robòtica has the virtue of being the first low-impact and zero-
kilometre Spanish project that has applied robotic technology to Catalan wood. The
Mediterranean pine timber solely used for palettes or biomass, due to its tendency
to warp while drying, have been used here. The material provided by the industry,
partner of the research project, is composed of wooden elements with a square cross-
section and dimensions of 38 mm × 38 mm × 2000 mm. Parallel to a formal and
tectonic exploration, the first phase of the project included an intense series of tests
concerning the technological and material system. In this specific case, the analogical
series of tests have been fundamentals to understand the behaviour of the material,
Fig. 3 Fusta Ròbotica: robotic fabrication process. © Andrea Quartara
exploring its mechanical and physical properties, and correctly inform the computa-
tional design process to ensure the feasibility of the project. Specifically, two areas of
investigation were outlined: the first concerned the possibility of combining different
digital fabrication strategies to design and materialize technological systems, the sec-
ond one concerned the study of connection joints between the various technological
units taking in consideration assembly logics. Through the experiments carried out
on the material in accordance with the digital manufacturing method and the tools
available (Fig. 3) (industrial robot KUKA KR 150-2 L110, circular saw and drill)
it was possible to define some criticalities with respect to which to make informed
design choices. Among these we can mention: the variation of the curvature follow-
ing the drying process; the need to preserve the resistant section of the material so as
not to damage the structural properties; the need to optimize the use of the material
to reduce the structural weight; the min and max dimension of the work space in the
robot room (1800 mm × 3600 mm). In relation to this, the design process has been
optimized through a series of informed choices:
– Fabrication with straight green wood, knowing the wood will deform after assem-
bly while drying;
– Redundant structure, composed of a multitude of small elements attached at least
in three point, as to force element to stay straight while drying and as to “lock” the
pieces in position- Nailed joints instead of milled joints to preserve the structural
section of the components;
Fig. 4 Fusta Ròbotica: computational workflow from design to construction. © Ji Won Jun
– Discretization of the surface through the production of sections with a constant

thickness whose dimensions can be traced back to 1800 mm × 3600 mm to be
assembled manually.
Through this methodological approach (Fig. 4), the design is informed about the
mechanical and structural properties of the material used as well as the tools involved
in all the phases. The bottom-up method adopted allows to guarantee the feasibility
and construction of the prototype and considers the material as an active agent and
not as an imposed element to be considered in the final phase of the design process.
Redundancy and tolerance are essential parameters both at (structural) design and
fabrication stages. Fusta Ròbotica has demonstrated how the tectonic configuration
can be the “tool” that helps to manage the tolerances of the low-engineered mate-
rial used for the construction. In this case, the overall shape overcame the material
properties in the design hierarchy.
2.3 “Digital Urban Orchard”: Design with Tolerances
Moving forward, Digital Urban Orchard expresses a more detailed functional pro-
gram: the interrelation of context, function, geometry and fabrication have been
directly embedded within the computational design (Figliola 2017). The soilless
cultivation system together with the setting of a new relational space can colonize
Barcelona’s rooftops, under the light of reducing the footprint of new built environ-
ments. The final shape has been selected out of a catalogue of options. The design of
the pavilion has highlighted a further aspect regarding the applicability of structural
wood systems in architectural practice: the control of formal generation through the
parametric process and the possibility of customization offered by the robotic fab-
rication (Fig. 5), allows to overcome and expand, the concept of structure to that
of an integrated system that manages to aggregate primary and secondary struc-
ture, systems and furnishings through specific morphological configurations. They
result from an optimization process informed by maximum stick-length, range of
possible angles to cut the stick edges as well as by different environmental analysis
that made able the designers to correctly mediate between design and digital pro-
Fig. 5 Digital Urban Orchard 2016: robotic fabrication process. © Andrea Quartara
Fig. 6 Digital Urban Orchard 2016: design with tolerances. Computational workflow and 3D scan-
ning process. © Angelo Figliola
cesses. Both constructions adopt the same structural principle: small size and square
section wooden elements have been cut, placed and nailed in position one-by-one to
create redundant structures. In particular, Digital Urban Orchard enriches the seam-
less design-to-fabrication process taking in account of manifold parameters, apart
from the solely geometry. Hence, the previous structural pattern, generated by the
sequence of main truss and structural stiffener, has been improved with hydroponic
system supports, skin supports, and furniture. This was made possible thanks to grip-
per technology applied for multitasking process based on external tailored effectors.
Fig. 7 Digital Urban Orchard 2016: scan of the overall wood structure allows to define the spatial
coordinates of the anchor points of the skin overcoming the mismatch between digital (x, y, z) and
physical (x1, y1, z1). © Angelo Figliola
Fig. 8 Digital Urban Orchard 2016: design with tolerances. From assembly to scanning and skin
assembly. © Angelo Figliola
The size of the robot-cell together with the custom-developed tools limits define some
design constraints but, at the same time, they trigger the implementation of a cost
and time-effective fabrication system. However, the pavilion present imprecisions
coming from two phases of the production process: the robotic assembly (including
a nailing process of the structural sections) and the in situ manual assembly where the
tolerance is constantly increasing. Due to the difference between the digital model
and the physical one, the major difficulties are in the integration between the main
structural system and subsystems such as technological devices and the silicone skin.
To overcome this problem and ensure the installation of the skin, a digital scanning
process was used after the completion of the pavilion (Fig. 6). The scan of the over-
all wood structure allows to define the spatial coordinates of the anchor points of
the skin overcoming the mismatch between digital (x, y, z) and physical (x1, y1, z1)
(Fig. 7). The new model obtained by the scanning process allows to update the digital
model and to inform the fabrication of the diagrid skin ensuring the correct installa-
tion (Fig. 8). These working prototypes lay the groundwork to further developments
scaling up the construction process, allowing a high-level of mass-customization and
cutting down the manufacture costs. Digital Urban Orchard introduces the theme of
communication between the different design phases to overcome the problems deriv-
ing from the large tolerances. The tolerances that accumulate during the various
phases of the project, from manufacturing to assembly of components, are managed
through a feedback loop process between the in situ assembly phase and the installa-
tion of the technological subsystems thanks to the 3D scanning of the overall shape.
The material properties overcame the overall design in the design hierarchy. Other
strategies can be applied to overcome this problem: interlocking connection systems
can be carved between adjacent sticks ensuring the overall structure alignment at
every stage of production; AR technologies to verify the correspondence between
digital and physical model during the robotic and the manual assembly process.
2.4 “Digital Woodcraft” and “Robotic Sawmill”: Design

and Fabrication Strategies Based on Natural Wood
Characteristics
One of the research lines explored at IAAC, concerns the use of natural and low
engineered material as an active agent in the design processes, thanks to the inno-
vative technologies of computation and digital fabrication that allow managing the
morphological and constructive complexity. The first experimentation in this regard
took place during the 2013 seminar, Robotic Sawmill, exploring the potential offered
by the robotic manufacture in the processing of wooden logs for the realization of a
prototype 1:1 scale of coverage for a shelter to be built near the Campus of Valldaura.
One of the peculiarities of the seminar is the use of a chainsaw, an end-effector that
recalls the analog fabrication/production processes of the trunks and of the structural
connections. Equipping the robotic arm with an analog tool means implementing
the potential offered by traditional techniques making it possible to materialize the
complexity of digital space derived from the lack of homogeneity of the material and
its tolerances (Fig. 9). The construction of the joints was preceded by the scanning of
the trunks and the development of a digital workflow to facilitate the digital recon-
struction of the physical component (Fig. 10). Parametric design is the instrument
by which the digital chain is synthesized, which includes data acquisition, the design
of joints and connections as well as the definition of the parameters necessary for
the manufacture and processing of components that integrate the specificities of the
scanned material (Fig. 11). The Digital WoodCraft project,3 a thesis developed within
the Master in Advanced Architecture of IAAC, MAA 2016/2017, represents an evo-
lution of the research line described above. Specifically, the thesis wants to explores
the tectonic and performative potentials deriving from the use of natural and non-
engineered material, such as tree trunks, in the design processes. Experimentation is
inspired by the observation of the natural world as opposed to the artificial/industrial
one: the morphological complexity that can be found in the natural forms of trees able
to guarantee excellent structural performance is standardized through an industrial-
ized production process aimed at eliminating imperfections and the inhomogeneities
that characterize the material for the production of the final product.
Some of the physical and mechanical properties that characterize the wood are
eliminated through the discretization and homogenization of the component with
consequent waste of energy and waste of material. In this regard, the research aims
to introduce and test a computational workflow that involves the use of innova-
tive technologies such as 3D photogrammetric scanning, parametric modeling and
robotic fabrication in order to collect, analyze and produce complex and customized
models and prototypes. The design process is based on the analysis and cataloging
of wooden logs in order to identify the physical and mechanical characteristics of the
components to be used as parameters of the design process (Fig. 12). Thanks to the
computational workflow, the geometric characteristics and the mechanical properties
3 http://www.iaacblog.com/programs/digital-woodcraft/.
Fig. 9 How to teach a robot to use a chainsaw?|Robotic Fabrication Workshop @ IaaC|2013. ©

Vicente Gasco
of the material become design parameters that contribute to the formation of the final
morphological solution. The direct connection between the log scan, the digital mesh
and the robotic manufacturing process made it possible to realize complex structural
joints and joints by controlling the tolerances derived from the use of natural and
non-engineered material. Node manufacturing was carried out using a subtractive
cnc milling process exploiting the potential offered by 3D photogrammetric scan-
ning in the positioning of the component within the work area and the calibration
of the tools used for the production process (Fig. 13). With the implementation of
traditional workflows, industrialized components and standard prefabrication can be
replaced and implemented by unique technological systems that cannot be repro-
duced in series, allowing a more efficient use of materials and exploiting the unique
structural and morphological characteristics of the trees and natural materials. With
the projects Digital Woodcraft and Robotic Sawmill another variable is introduced.
Instead of using industrialized and engineered material, the experimentation keeps
in consideration the natural material as the driver of the design choices.
148
Fig. 10 Scanning Log Geometry|Robotic Fabrication Workshop @ IaaC|2013. © Vicente Gasco

G. Brugnaro et al.
Fig. 11 Fabrication of 1:1 scale prototypes of timber joints with a chainsaw-equipped robot|Robotic
Fabrication Workshop @ IaaC|2013. © Vicente Gasco
Fig. 12 Digital WoodCraft|Design and fabrication strategies based on natural wood characteristics.
© Nikos Argyros
Fig. 13 Digital WoodCraft|Robotic fabrication process. © Nikos Argyros

3 Digital Craftsmanship
Within current design practices, the separation between the act of design and making
requires the standardization of fabrication tools and materials, significantly limiting
the range of design solutions available. Can we explore novel design opportunities
extending the range of manufacturing processes available to designers through the
integration of material and tool affordances as process drivers? If the fabrication
process is not completely predetermined in front of the computer screen, leaving
room for interpretation and exploration, the quality of the final outcome, because
only partially anticipated, is continuously at risk and significantly depends on the
capability of the fabrication framework to accommodate changes in terms of design,
tools and materials. The definition of “craftsmanship” provided by D. Pye in his book
“The Nature and Art of Workmanship” (Pye 1978), is particularly suited to this novel
fabrication paradigm (Kolarevic and Klinger 2008): “Craftsmanship…means simply
workmanship using any kind of technique or apparatus, in which the quality of the
result is not predetermined, but depends on the judgement, dexterity and care which
the maker exercises as he works. The essential idea is that the quality of the result is
continually at risk during the process of making.”. While the construction industry
is one of most adverse to changes and risks, within this notion of craftsmanship,
“risk” becomes a positive force that drives the making of an artefact. Digital fabrica-
tion tools could allow to embed this “risk” component through carefully articulated
design explorations and lead to the discovery of unanticipated formal configura-
tions or performance behaviours. For each task approached by a craftsman, there are
different evaluation criteria, for instance, aesthetic quality or cost efficiency, techni-
cal procedures and material consideration involved that “operate as positive forces
for action not determinants of outcome” (Keller and Keller 1993). The experience
knowledge of the craftsman, accumulated through years of practice, allow to address
these “dimensions” in order to create a new original plan of action. However, as dis-
cussed by Sharif and Gentry (2015), this preconception only initiates the task, while
the “the design concept evolves concurrent with the craftsman’s act of production
and the received feedback from the evaluation of material and objective conditions
of the work”.
3.1 “Adaptive Robotic Carving”: Digital Craftsmanship

and Machine Learning
The project “Adaptive Robotic Carving”, developed at the Bartlett School of Archi-
tecture as part of the InnoChain research network, focuses on the integration of
materials performance within design to manufacturing workflows combining sensor
data, machine learning and robotic manufacturing (Brugnaro and Hanna 2017). The
project examines manufacturing processes as moments of discovery, where the digital
model is constantly tuned as the fabrication progresses and designers are able to curate
Fig. 14 The encapsulation of instrumental knowledge through the training process is aimed towards
its integration within a design interface that allows to move back and forth between the prediction
of carved geometries and generation of robotic toolpaths. © Giulio Brugnaro
their own custom design to manufacturing workflow, integrating real-world mate-

rial behaviors to take better-informed design decisions, rather than working within
a standard CAM framework, which usually describes the workpiece as a block of a
generic non-material (Fure 2011). Focusing on subtractive manufacturing with tim-
ber, the technical core of the research questions whether is possible to use real-world
fabrication data, collected both by human experts and autonomous robotic sessions,
to achieve a more accurate geometrical prediction of non-standard tools operations
on a specific wood species, and, conversely, if it’s possible to reconstruct backwards
the robotic toolpath that generated a given carved geometry (Fig. 14). To achieve
this, the developed methods present a series of training procedures for a robotic fab-
rication system where the instrumental and material knowledge of skilled human
craftsman is captured, transferred, robotically augmented and finally integrated into
an interface that make this knowledge available to the designer. The sensor data,
collected in datasets, are used to feed a machine learning procedure to extract cor-
relations between the fabrication parameters and their material outcomes and use
such knowledge to inform future robotic fabrication tasks with a similar set of wood
species and carving tools (Fig. 15). In the specific, the robotic training methods are
structured around three main stages: (1) Recording, (2) Learning (3) Fabrication.
The goal of the recording stage is to collect sensor data and create a library of fabri-
cation datasets. During each recording session, a system of motion capture cameras
is used to track with high degree of precision the position and orientation of clusters
of reflective markers applied on the carving tool, reconstructing it in real-time in
the digital design environment. The combination of such methods with a force feed-
back sensor allows to record fabrication datasets simultaneously with the performing
of the carving operation. In addition, after the recording session, a digital scan of
each training board is generated through photogrammetric reconstruction. Each cut
is analysed through different parameters such as tool/workpiece angle, tool/grain
direction angle, force feedback, feed rate, target cut depth, target cut length and the
152
Fig. 15 The overall training process is articulated in three sequential stages (recording, learning, fabrication) focused around the transfer of instrumental
knowledge from the domain of human craftsmanship to robotic manufacturing. © Giulio Brugnaro
G. Brugnaro et al.
Fig. 16 During the fabrication stage, the robotic toolpath is optimized to match the original design
intention with tool affordances and material properties, such as the wood grain directionality. ©
Giulio Brugnaro
Fig. 17 The trained system could be used to predict combinations of multiple cuts to generate
specific patterns on the carved surfaces. © Giulio Brugnaro
material outcome that these generate measured in the length, width and depth of the
cut. Following the human demonstration, the carving dataset is consolidated through
autonomous robotic fabrication sessions aimed to collect more data to improve the
training performances through a finely interpolated collection of cuts generated by
sets of fabrication parameters within the operational range defined by the skilled
craftsman. In the next step, the aim of the learning stage is to extract meaningful
correlations in the recorded data, namely between the fabrication parameters and
their material outcomes, and use these to inform, in a new “unseen” configura-
tion, the robotic fabrication process. This implies the extraction of the instrumental
knowledge present in the dataset into a transferable and usable design tool through
a supervised machine learning process. The selected computational approach is an
Artificial Neural Network (ANN), which, loosely inspired by its biological equiva-
lent, could be described as a layered and interconnected network of “neurons” able
to “process information by their dynamic state response to external inputs” (Hecht-
Nielsen 1990). The network topology not only determines the performance of the
system but also needs to be configured in terms of inputs and outputs considering the
intended use of the trained network in a practical application during the fabrication
stage (Fig. 16). For instance, the learning objective could be structured to predict
the simulation of a subtractive operation from a user-defined toolpath and a series
of fabrication parameters, or conversely, generate a robotic toolpath out of a carved
geometry obtained from a 3D scanned element. While conventional digital Boolean
operations results insufficient in calculating the result of subtractive operations with
non-standard tools and heterogeneous materials, the trained network provides the
opportunity to explore design solutions through a more accurate simulation that takes
into consideration specific tool affordances and material properties such as the wood
grain direction (Fig. 17). Furthermore, the network prediction allows to optimize
individual fabrication parameters to increase the efficiency of the process measured
through the material removal volume or fabrication speed. For such reasons, the
project suggests the potential of machine learning strategies for the customization of
design to manufacturing workflows which can be flexibly trained according to dif-
ferent design intentions and material performances, through the extraction of human
tacit knowledge and its integration within a simulation framework.
4 Conclusions
The series of design strategies presented in this chapter address the issue of tolerance
control derived from the use of non-homogeneous material, such as timber, and how
its latent material agency can be exploited as an active driver for design processes,
instead of being conceived as a passive receiver of previously defined design choices.
The different approaches could be summarized as follows:
• Fusta Robòtica: Design in contraposition the material through specific morpho-
logical configurations that overcomes the material’s unpredictability. Analogue
testing is propaedeutic to understanding the material properties, and consequently
to extract rules with which to construct the digital model and overcome materials
imperfection.
• Digital Urban Orchard: Design negotiation connecting different phases of
the design process, utilizing communication protocols based on feedback loops
between design and construction. Through a scanning process and a digital recon-
struction of the physical prototype, it’s possible to overcome tolerances derived
from the material used and the robotic fabrication process. Based on the method-
ology described above, a perfect integration between technological systems is

ensured, taking into account the inhomogeneity of wood, and without forcing the
final shape in contraposition to the material properties.
Robotic Sawmill: Design in accordance with material’s heterogeneity, the material
becomes an active agent of the design process. The methodology is based on
the precise digital reconstruction of the physical component and its physical and
mechanical properties that allows to design taking into account the tolerances that
derive from the non-homogeneity of the material. Material properties as active
agent in the design process.
• Adaptive Robotic Carving: Integration of material and tool affordances as pro-
cess drivers through the training of a robotic fabrication system specifically tuned
to operate within a selected fabrication and material domain. The trained system,
based on sensor data collected during a series of recording sessions, has been used
to provide an accurate prediction/simulation of the carving operations and explore
multiple material solutions in terms of carved geometries and textures marks,
before moving to the actual fabrication stage. Such encapsulation of instrumen-
tal and material knowledge into a design interface allowed, not only an effective
geometric evaluation, but also provided the designer with an understanding of
the outcome generated by specific fabrication parameters and material properties,
such as the influence of the grain directionality in the actual carved shape.
In the light of the presented projects, the integration of material information within
advanced robotic fabrication opens new opportunities in the design and making
of wood construction. The need for tolerance and adaptation to the wood’s unpre-
dictability has led the design to fabrication to integrate more sensors and be more
flexible towards streams of feedback information. As shown in the presented exam-
ples, the digitized flow of information from the material and fabrication process
can inform the design before, during or after the construction process, leading to a
series of novel design approaches integrating tolerance and unpredictability at vari-
ous degrees.
In this Cyber Physical Making context, Dynamic Blueprints offers a new tool for
designer to negotiate the materialisation of the project, while the implementation
of machine learning strategies offers the possibility for robots to play a stronger,
active, role in the design and making process. With such technological advances, the
concept of craftsmanship has the opportunity to be re-integrated within the industrial
fabrication process, such as robotic and digital ones, towards a Digital Craftsmanship
capable of integrating wood properties in their full complexity (Fig. 18).
Fig. 18 Summary of design strategies proposed through the narrative of case studies. © Angelo
Figliola
Acknowledgements Special thanks to Dr. Mathilde Marengo for input.

Fusta Robòtica Pavilion is a research project of IAAC realized as part of the Open Thesis Fabrication
2015 developed with the generous sponsorship of Serradora Boix; in collaboration with Gremi de
Fusters, Tallfusta, Incafust, Mecakim, Decustik.
Digital Urban Orchard is a project of IAAC, realized as part of the Open Thesis Fabrication 2015,
developed with the generous sponsorship of Merefsa, and the collaboration of Windmill and Scanarq.
Both projects were led by Areti Markopoulou, Alexandre Dubor, Silvia Brandi; assisted by Djordje
Stanojevic; and developed by students: Andrea Quartara, Angelo Figliola, Monish Siripurapu, Ji
Won Jun, Josep Alcover Llubia, Yanna Haddad, Mohamad Mahdi Najafi, Fathimah Sujna Shakir
and Nada Shalaby.
Robotic Sawmill is a project of IAAC, realized as a workshop within the Master of Advanced Archi-
tecture 2012–13, led by Tom Pawlofsky, assisted by Alexandre Dubor, and developed by students:
Alexander Dolan, Kartik Ashok Gala, Robert Francisco Garita Garita, Vincent Huyghe, Stefanos
Levidis, Iker Luna, Stuart Maggs, Dirce Medina Patatuchi, Pedro Moraes, Boleslaw Musierowicz,
Urte Naujekaite, Amir Reza Saheb, Dori Sadan, Ahmed Selim, Jin Shihui, Anand Singh, Sofoklis
Giannakopoulos, Georgios Soutos, Angeliki Terezaki, Ali Yerdel, Maria Kuptsova, Vicente Gasco
Gomez, Inder Prakash Singh Shergill.
Digital Woodcraft is a project of IAAC, realized as part of the Master of Advanced Architecture
Individual Thesis, in 2016–17, led by Marcos Cruz, with the input of Raimund Krenmueller and
developed by student: Nikolaos Argyros.
The “Adaptive Robotic Carving” project is part of ongoing Ph.D. research conducted by Giulio
Brugnaro, supervised by Prof. Bob Sheil and Dr. Sean Hanna, at the Bartlett School of Architec-
ture, University College of London, within the framework of the “InnoChain Training Network,”
supported by the European Union’s Horizon 2020 research and innovation program under the Marie
Sklodowska-Curie grant agreement No. 642877.
References
Alemany M, Portel J (2014) Soft tolerance: an approach for additive construction on site. AD Archit
Des J (214):122–127
Brugnaro G, Hanna S (2017) Adaptive robotic training methods for subtractive manufacturing.
In: ACADIA 2017: disciplines & disruption, proceedings of the 37th annual conference of the
association for computer aided design in architecture (ACADIA). Cambridge, MA 2–4 Nov 2017,
pp 164–169
Brugnaro G, Baharlou E, Vasey L, Menges A (2016) Robotic softness: an adaptive robotic fabrication
process for woven structures. In: “ACADIA//2016: Posthuman frontiers: data, designers, and
cognitive machines, proceedings of the 36th annual conference of the association for computer
aided design in architecture (ACADIA), pp 154–163
Capek K, Novak-Jones C, Klima I (2004) R.U.R. Penguin Classics
Carpo M (2014) Mario Carpo in conversation with Matthias Kohler. In: Gramazio F, Kohler M,
Langenberg S (eds) Fabricate: negotiating design & making. Gta-Verl, Zurich, pp 12–21
Dörfler K, Sandy T, Giftthaler M, Gramazio F, Kohler M, Buchli J (2016) Mobile robotic brickwork.
In: Reinhardt D, Saunders P, Burry J (eds) Robotic fabrication in architecture, art and design 2016.
Springer International Publishing, pp 205–218
Dubor A, Diaz GB (2013) Magnetic Architecture. In: Brell-Çokcan S, Braumann J (eds) Rob|Arch
2012. Springer, Vienna
Dubor A, Camprodom G, Bello Diaz G, Reinhardt D, Saunders R, Dunn K, Niemelä M, Horlyck
S, Alarcon-Licona S, Wozniak-O’Connor D, Watt R (2016) Sensors and workflow evolutions:
developing a framework for instant robotic toolpath revision. In: Reinhardt D, Saunders P, Burry J
(eds) Robotic fabrication in architecture, art and design 2016. Springer International Publishing,
pp 411–426
Figliola A (2017) Post-industrial robotics: esplorazione di architetture informate nell’era post-
digitale. Techne. https://doi.org/10.13128/techne-xxxxx
Figliola A, Dubor A (2017) Fusta Robòtica: generic tools for complex structure through the perfor-
mances. In: Favargiotti S, Staniscia S (eds) Monograph.Research REDS 03. Flowing Knowledge,
Monograph.Research 03. List, Trento, pp 228–232
Fure A (2011) Digital materiallurgy: on the productive force of deep codes and vital matter. In:
ACADIA 11: integration through computation, proceedings of the 31st annual conference of the
association for computer aided design in architecture (ACADIA), Banff (Alberta), pp 90–97
Gramazio F, Kohler M (2014) Made by robots: challenging architecture at a larger scale. Wiley,
London
Hecht-Nielsen R (1990) Neurocomputing. Addison-Wesley Pub. Co, Reading, Mass
Jeffers M (2016) Autonomous robotic assembly with variable material properties. In: Reinhardt
D, Saunders P, Burry J (eds) Robotic fabrication in architecture, art and design 2016. Springer
International Publishing, pp 49–62
Keller CM, Keller JD (1993) Thinking and acting with iron. In: Chaiklin S, Lave J (eds) Under-
standing practice: perspectives on activity and context. Cambridge University Press, Cambridge
Kolarevic B, Klinger K (2008) Manufacturing material effects: rethinking design and making in
architecture. Routledge, New York
Kwinter S (1993) Soft system. In: Culture lab. Princenton Architectural Press, pp 206–227
Menges A (2015a) Material synthesis: fusing the physical and the computational. Architectural
Design Profile, Wiley, London
Menges A (2015b) The new cyber-physical making in architecture. In: AD Archit Des J (237):40–47
Pye D (1978) The nature and art of workmanship. Cambridge University Press, Cambridge
Schwartz T, Andraos S, Nelson J, Knapp C, Arnold B (2016) Towards on-site collaborative robotics.
In: Reinhardt D, Saunders P, Burry J (eds) Robotic fabrication in architecture, art and design 2016.
Springer International Publishing, pp 388–398
Sharif S, Gentry R (2015) Design cognition shift from craftsman to digital maker. In: Emerging
experience in past, present and future of digital architecture, proceedings of the 20th Interna-
tional Conference of the Association for Computer-Aided Architectural Design Research in Asia
(CAADRIA 2015). pp 683–692
Vasey L, Baharlou E, Dörstelmann M, Koslowski V, Prado M, Schieber G, Menges A, Knippers J
(2015) Behavioral design and adaptive robotic fabrication of a fiber composite compression shell
with pneumatic formwork. In: Combs L, Perry C (eds) Computational ecologies: design in the
anthropocene, proceedings of the 35th annual conference of the association for computer aided
design in architecture (ACADIA). University of Cincinnati, Cincinnati OH, pp 297–309
Wu K, Kilian A (2016) Developing architectural geometry through robotic assembly and material
sensing. In: Reinhardt D, Saunders P, Burry J (eds) Robotic fabrication in architecture, art and
design 2016. Springer International Publishing, pp 241–251
Advanced Timber Concepts
and the Design of Furniture
and Structures
Patrick Beale and Tristan Morgan
Abstract Since 2006, the Advanced Timber Concepts Research Centre (ATC) has
pursued the development of innovative approaches to designing with timber across a
variety of scales. Through teaching-based research and applied research within a pro-
fessional context, the research group has engaged in cutting-edge practice uniquely
relevant to Australian timber culture, whilst engaging with digital wood design in a
global context. Projects undertaken since 2006 include a ten-year project running an
intensive furniture design course at the University that focused on re-imagining the
role of low-value timber within fine-furniture making, employing traditional tech-
niques along-side cutting edge digital fabrication and computer-aided-manufacturing
(CAM) processes. Concurrently, a pair of prototypical flat-pack timber houses emerg-
ing from teaching programmes were designed, developed, and built, alongside several
other prototypical timber structures and assembly systems initiated within academic
teaching and developed by the research centre and the professional studio—ATC
Studio. These include a third prototype house, an experimental permanent large-
span timber shell structure, a continuing five-installment advanced computational
design and manufacturing design-and-build teaching program, and a currently under-
construction sports hall and community centre. Through these projects, we have
engaged with the use of timber as a vibrant material that possesses active proper-
ties that can be employed productively within the design process. Further, advanced
digital processes such as computational design, simulation, analysis, form-finding,
evolutionary solving, and CNC manufacturing, uncover the opportunity to explore
the nature of the relationship between designer and material by way of the digital.
We increasingly understand that the design of computational processes leads to a
condition of ‘dual-authorship’, where the designer is both responsible for the pro-
cess and the computed result. Add to that the active role of a natural material such
P. Beale (B)
The University of Western Australia and Advanced Timber Design Research Centre, Perth,
Australia
e-mail: Patrick.beale@uwa.edu.au
T. Morgan
The University of Western Australia and Cox Architecture, Perth, Australia
e-mail: Tristan.morgan@uwa.edu.au

160 P. Beale and T. Morgan
as timber, and we now encounter a third author—the forest—which may begin to

approach the achievement of a truly holistic, sustainable architecture that considers
its environmental impacts in an active and performative manner from the outset.
Keywords Fabrication · Prefabrication · Furniture · Digital design

Computational design · Material properties
1 Introduction
Since 2006, the Advanced Timber Concepts Research Centre (ATC) has pursued the
development of innovative approaches to designing with timber across a variety of
scales. The Research Centre was funded initially with a brief to ‘re-value’ the native
hardwood resources of Western Australia in the wake of the state governments re-
structuring of the native forestry industry which resulted in significant reductions in
the volume of timber that could be harvested each year.
Through teaching-based research and applied research within a professional con-
text, the research group has engaged in cutting-edge practice uniquely relevant to
Australian timber culture and in particular, West Australian timber culture, whilst
engaging with digital wood design in a global context. Projects undertaken since 2006
include a ten-year project running an intensive furniture design course that focused on
re-imagining the role of low-value timber within fine-furniture making, employing
traditional techniques along-side cutting edge digital fabrication and computer-aided-
manufacturing (CAM) processes. Discoveries made during this program of research
included the establishment by empirical methodologies of the strength properties
of young and mature re-growth native timbers. This was significant because what
emerged was that the properties of the re-growth forests were quite different to those
from the old growth timbers on which all of the available data was based.
It should be noted that commercial logging in the West Australian native forest did
not get started until the late 1880s (Carron 1985). One hundred years later in the 1980s
the last serious testing and evaluation of native timber properties was systematically
undertaken. The research group concentrated on applications in furniture as this
represented the highest value adding of all timber products. The strength properties
of the native hardwoods were such that we were able to work successfully with very
small sections of timber, and with low value ‘thinnings’ rather than with mature
timber but this in turn demanded a precision of manufacturing that was ill suited to
manual practice but very well suited to digital fabrication.
Concurrently, with this hardwood based research, a pair of prototypical flat-pack
timber houses which emerged from studio based design teaching programs were
designed, developed, and built, alongside several other prototypical timber struc-
tures and assembly systems initiated within academic teaching and developed by
the Research Centre and the professional studio—ATC Studio. These include a third
prototype house, an experimental permanent large-span timber shell structure, a con-
tinuing four-installment advanced computational design and manufacturing design-
Advanced Timber Concepts and the Design of Furniture … 161
and-build teaching programme, and a 3700 m2 sports hall and community centre that
is currently under construction.
The requirement for ‘flat-packing’ emerged here from demand from the rural and
remote areas of this very thinly populated state. Western Australia has a popula-
tion now of about 2.5 million, the majority of whom (2.2 million) live in the Perth
metropolitan area. The result is that regional towns and centers are very small and
have difficulty supporting building companies of any size, and most buildings are
therefore either supplied from or are built using labour from the Perth metro area.
This makes for very expensive work with a premium of up to 40% over metropoli-
tan costs quite common. Obviously, this inhibits development and maintenance and
so the housing and building stock are in poor condition. Providing buildings where
the majority of the time/labour costs are incurred in controlled factory conditions
rather than on site in conventional building is very advantageous. The challenge in
these projects was establishing an appropriate technology for the context, rather than
deploying the most effective production processes available. It became clear very
quickly that many previous projects had failed because they had not taken into con-
sideration the context and lifespan of the building, or the ‘life cycle costs’ of a given
remote project.
Through these projects, we have engaged with the use of timber as a vibrant
material that possesses active properties that can be employed productively within
the design process. In each case, the context in which the design is to perform or to
be located, is key to developing appropriate responses to program or to demand. Fur-
ther, advanced digital processes such as computational design, simulation, analysis,
form-finding, evolutionary solving, and CNC manufacturing, offer the opportunity
to explore the nature of the relationship between designer and material by way of
the digital. We increasingly understand that the design of computational processes
leads to a condition of ‘dual-authorship’, where the designer is both responsible for
the process and the computed result. Here we must add that the ‘designer’ is both the
architect and the structural engineer, working as equal and symbiotic partners in the
process. Add to that the active participation of a natural material such as timber, and
we now encounter a third family of authors—the forest and its raw and engineered
products. With this array of authors—man, machine and nature, we may begin to
approach the achievement of a truly holistic, sustainable architecture that considers
its human engagement in context and its environmental impacts in an active and
performative manner from the outset.
The following chapters serve as a review of more than a decade’s worth of research
and practice over multiple scales and contexts. We present an unfinished and con-
tinuing history of Advanced Timber Concepts as research centre and commercial
practice, which has advanced the knowledge of endemic timber within a developing
industry through applied and experimental research as pedagogical and commercial
practice. What emerges from this review is the rise of the role of digital practice
in the understanding and delivery of wood in architecture, where digital design and
fabrication methods have supplanted traditional crafting in the search for greater
precision, complexity, and material efficiency.
2 The Architecture of Furniture
The ten years of workshop/studio teaching program was based on a project the
studio undertook professionally to develop an all timber chair using West Australian
Hardwoods—Jarrah (Eucalyptus marginata) and Karri (Eucalyptus diversicolor), two
of the more dense of the Eucalypt species and quite different in their structural and
material properties. The profiles of the two trees indicates the fundamental difference
in their structural characteristics. Both timbers are very efficient in compression
which is derived from their density (around 1200 kg/m3 ) but whereas Jarrah is short
fibred and quite brittle in bending, Jarrah has long stringy fibers and is much more
elastic, making it quite suitable for structures—in this case chairs—that are subject to
eccentric loads. The structural behavior of small scale (furniture) timber structures
are poorly understood, and the data available on material properties of regrowth
Australian hardwoods is sparse.
We worked with both timber species with the common aim to produce a chair that
weighed no more than 3 kg but that was robust enough to meet stringent commercial
testing standards. The parameters we took from Gio Ponte’s Superleggera1 chair
(and Cane Seat) and with some hubris, attempted to emulate his work using timber
that was several times denser than European Ash (density 700 kg/m2 ). To approach
the lightest possible results for the work. We needed to keep the member size to
a minimum, and so with a mixture of data extrapolated from what was available,
and some preliminary empirical testing we established a member size of between 20
and 24 mm. for each species. Joints were to be kept to a minimum and a number of
different configurations were explored and tested on site and at the Australian Furni-
ture Research and Development Institute in Tasmania, the only facility in Australia
certified to test furniture to Australian Standards. The outcome of the testing and
evaluation process was to use 22 mm × 22 mm square members in the Jarrah version
of the chair, and 20 mm × 20 mm square in the Karri version. In this configuration
the Karri chair achieved a rating of level 3 (out of 5) commercial rating while the
Jarrah chair achieved only a domestic rating (2) at the larger section size.
The Shimmer Chair (2009) was the result, shown in Udine, London (100% Design,
where it was runner up to best in show) and Valencia in that year (Fig. 1). The chairs
use a compound haunched double Mortice and Tenon joint at each of the critical
junctions, and spline joints elsewhere (the leading edge ‘barrel’ and the leg/back-
seat junction). The surface texture of the seat battens was milled to a depth of 1 mm
using the Masterwood 327 CNC machine we had available. Each of the joints and
each member was prepared on this machine from our 3d Studio files. While these
joints could be made with hand skills and patience, we were aiming at a chair for
mass production and so the emphasis’s was always on machine production rather
than hand preparation. While a flatbed CNC machine is far from ideal for preparing
end jointed members, it is possible with the inclusion of specially fabricated jigs. A
robotic arm or 5 axis CNC machine would have been more convenient.
1 Superleggera Chair. Designed 1957 Gio Ponte. Produced 1965 by Cassina S p A. Weight 1.7 kgs.
Fig. 1 Shimmer chair and table, 2008
The premise for the studio teaching program was founded in this base data. One
criteria that was established early on in the program was to use one timber species
(Karri) and to limit the section size to 20 mm × 20 mm or 400 mm2 of any member
of the structure. The fabrication of joints at this scale in these timbers became very
awkward without sophisticated joinery skills and so we developed a digital compo-
nent manufacturing processes during the program drawn from models generated in
3D studio and later in the Rhinoceros software (McNeel 2018), a process that soon
carried over to both the architectural projects and the later generative fabrication unit.
Fig. 2 Student chair, 2012. Georgia Taylor-Berry
The Architecture of Furniture Unit explored and took to its limits, the potential to
design and make very lightweight chairs using timbers that heretofore had been used
for both outdoor and indoor furniture but with a very ‘agricultural’ aesthetic. Heavy
frames, weighty pieces. The chairs became not only a means to directly engage with
the materiality of making things with timber, but also to explore structural systems,
comfort limits, and form in a pure sense (Fig. 2).
The carryover from the prototyping (each student had to design, prototype and
make their chair) in the area of three dimensional modeling as part of the design pro-
cess, rather than merely an illustrative form, was very useful when students returned
to projects at other scales. This also allowed the studio to explore and develop alter-
native ways of setting up modeling systems to interface with production systems.
3 Flat-Pack Housing
The conditions that lead to the need for a flat pack housing system in Australia are
somewhat different to those that drive similar systems in other parts of the world that
are more densely populated. The driver in our projects was the very remoteness of
many rural communities here, where remote means anywhere beyond the peri-urban
zone which for Perth has a limit of about 2 h’ drive or 150 km. At this distance it
is the cost of imported labour (from the metro area) that raises costs and so speed
of erection on site is paramount. The second most important element is the use of
an appropriate technology. The market is not large enough to have a sophisticated
conventional prefabricated building industry such as is found in the USA where
there is a long tradition of relatively high quality factory made or prefabricated
buildings. In Australia this does not extend beyond the provision of single person or
dormitory accommodation for the numerous remote mining operations in the State.
These buildings, affectionately known as Dongas, are minimal in design content and
rely upon a format that can be delivered on a flatbed truck fully fitted and ready to
use. They are environmentally poor performers as energy to run 24/7 cooling units
is readily available in such a high value industry. Each one will comprise 5 or 6
single rooms with no communal space and therefore there is no effective ‘after life’
or secondary application. The life span of a typical donga is 5–7 years after which
they are scrapped.
Through a variety of simulations and client group feedback sessions we arrived at
a panel format that has two variations in dimension, and three variations in cladding
formats to deal with straight run perimeter walls, corners and internal wall condi-
tions. The panels have a more or less conventional frame structure and a lightweight
core cladding over which different external finishes can be applied (typically corru-
gated steel, timber weatherboards or engineered timber panels such as plywoods).
The panels are made up on a jig with CNC operated screwing and nailing, but are
also simple enough to be able to be assembled by hand in small volumes. This sim-
plicity also keeps the overall technology of maintenance within generally understood
building practices and can be performed with standard skills and hand tools. This
two level technology allows for rapid assembly and if necessary, stockpiling as well
as for low key or semi-skilled erection and modification or repair and maintenance
over time. The high tech or digital production of the fundamental components allows
for low tech/low skill intervention at a later date without jeopardizing the structural
or performance criteria of the building. Also, by limiting the panel size and type, the
configuration in plan and section, while flexible, does not allow for infinite varia-
Fig. 3 Perenjori 2 flatpack housing, 2013
tions. It also means that doors, windows and corridor widths remain at optimum sizes
rather than getting reduced to minimum or sub minimum dimensions. In terms of the
‘partnership’ of the parties involved referred to above, this project involves the user
and the erector in equal partnership with the architect or designer in putting together
these projects. Designing also becomes an assembly operation, albeit initially carried
out in virtual space. In the real life conditions, options are limited, but the system is
variable enough to be able to adapt to different site and orientation configurations to
optimize environmental conditions and amenity at any given site (Figs. 3 and 4).
4 Generative Fabrications
The intimate understanding of native Western Australian hardwoods such as Karri,

coupled with early experimentation with the then emerging computational design
software combination of Rhinoceros 3D and Grasshopper, prompted ATC researchers
to posit whether Karri was capable of being exploited within the typology of long-
span structures, initially as post-formed gridshells in the Weald and Download mold,
and later using short lengths with complex joints. In early 2011, ATC initiated a
series of pedagogical experiments to explore the relationship between timber mate-
Fig. 4 Perenjori 2 flatpack housing, 2013
rials, emerging digital design practice, and digital fabrication technology. The first
of these examined the long-length structural bending performance of green karri
within the context of a pavilion design project to be assembled on the ground of
the University. Findings from this research were developed into a Masters level unit
called “Generative Fabrication”, which examined developments in computational
practice and digital fabrication and pre-fabrication technologies within the context
of a design-driven pavilion project undertaken by students. The scope of this unit
changed year-to-year, though the core focus remains the relationship between natural
material and digital design technologies.
4.1 Gridshell Experiment
The Gridshell Experiment, as the first pedagogical experiment in utilizing karri in

long-span structures was dubbed, took place over a single semester in 2011. Students
were invited to research the use of timber as a structural material in expressive
architectural formations through the production of scale models. From these models
were then derived structural principles, which were translated into parametric models
using the Rhino/Grasshopper combination. These parametric models were parlayed
Fig. 5 Students assemble full-size prototype of green karri grid-shell, 2011
into prototypical timber gridshell joints, which were then built and tested at full-scale
using green karri. Several student teams proposed pavilion designs, resulting in an
array of options from which a single project was chosen to be assembled for display
at the Faculty of Architecture’s end-of-semester exhibition.
While the bending performance of the green karri in long span application was
recorded and acknowledged by the project teams, at this time the digital models
produced merely approximated the timber’s performance. Manual adjustment of
rudimentary parametric models was used to attain shell-type forms for the pavilion
proposals, but the models possessed no real material information, and thus were not
tested for performance or viability. It was only during the 1:1 assembly period (Figs. 5
and 6), under extreme time pressures, that the material performance assumptions were
tested, with non-optimal results. While the bending performance of green karri, when
considered within a near-homogenous material density and quality, was well-known
to the teaching and learning group by this point, imperfections along the length
of timber in terms of grain and knotting was not accounted for. Under moderate
bending stresses, the karri buckled and splintered. This performative material affect
accelerated as the green timber dried out during the Australian autumn.
The ATC teaching team concluded that for such research to continue, methods of
embedding natural material performance within computational models needed to be
employed or developed. This must necessarily include the ability to model, computa-
tionally, material properties and behaviours, structural performance and optimization,
and the automation of fabrication processes using computer-numerically-controlled
manufacturing practices. Similar to the way that Bernard Cache’s Objectile practice
Fig. 6 Example prototype grid-shell connection, 2011
embeds the machining data directly within the design model; future experimental
research into structural applications of timber products must also minimize transla-
tion events to ensure the 1:1 transmission of design information to fabrication and
actualization.
4.2 Research Pavilions
Following on from ‘The Gridshell Experiment”, a permanent Masters level unit at the
University was established. This new unit, beginning in 2012 and named ‘Generative
Fabrication’, saw four experimental pavilion-scale projects designed and prototyped
at 1:1 scale up until 2017. Each pedagogical research project explored the relationship
between computational design tools and material properties via translation events.
Through this teaching programme we uncovered the immense pedagogical value in
prototyping at 1:1 scale and exposing students to the impacts of translations from dig-
ital to actual in practice. The importance of understanding material behaviours—and
limitations—was essential in enabling the fabrication and assembly of projects, while
exposure of students to theories of digital design was an essential in situating the
use of computational design approaches within a long and continuing history of
architectural representation and execution.
Fig. 7 Partially assembled pavilion, 2012
2012—Plywood Shell
The first instalment of Generative Fabrication challenged students to examine the
opportunities inherent in using a single material for structure and fixings—marine
plywood, a single software combination—Rhino/Grasshopper, and a single manu-
facturing process—a Masterwood Project 327 CNC machining center to design a
small experimental structure.
Several options were proposed, and students elected to develop for fabrication and
assembly a structure that utilized plywood as structural panels connected through
plywood discs that acted as universal nodes (Fig. 7). The benefit of these discs was
that they were simple to machine and could meet any requirements. In fact, this
methodology was an in-practice example of mass-customisation made possible only
through digital design to fabrication processes using then-emerging computational
design practices. The plywood disc connectors allowed the structural plywood panels
to be planar, yet take on a non-standard form. The form itself was inscribed in the disc
connectors thanks to the unique slotted connections generated automatically within
the generative Grasshopper script (Fig. 8).
The project engaged with non-standard modes of production, bearing some resem-
blance to Objectile’s Fast-Wood prototypes of 2004 (Cache et al. 2007). Whilst the
pedagogical aims of this first instalment of Generative Fabrication were limited to
Fig. 8 Fabrication of circular nodes, 2012
modes of production and computation, it is impossible to avoid entering into think-

ing of design through the frame of digital representation, data-driven form, and
the philosophy of objectile and the non-standard (Cache 2011). The Plywood Shell
pavilion, as constructed, represents just one single actualization of a process that
can easily generate thousands of variations. Unwittingly, and unavoidably, entering
into the space between computational design and fabrication forces one to encounter
philosophies of form and formation.
2013—Sculptural Narratives
The following year, we introduced to the unit a focus on karri to parallel the research
being undertaken in the Architecture of Furniture unit. Where the furniture unit
would explore the material at the human scale, Generative Fabrication would instead
explore karri as a structural material in an experimental pavilion. In this instalment
of the unit, students were immediately introduced to the neo-materialist philosoph-
ical framework of Manuel DeLanda, wherein the material itself was understood as
an active contributor to form through morphogenetic processes. The relationship
between designer, material, and machine was also highlighted within the contin-
uing history of digital design theory and practice that has developed since Ivan
Sutherland’s Sketchpad in 1963, although Nicholas Negropontes’ later Architecture
Machine project was more significant in understanding the role of the machine in
design thinking and, eventually, fabrication.
The limitations of the material were quickly understood, particularly as the avail-
able material was limited to the same used in the furniture unit—boards of 19 mm ×
Fig. 9 3d model view of half-lap joint, 2013
120 mm × 1200 to 1500 mm long. This influenced the scale of the proposed struc-
tures, which were limited to a sculptural scale rather than a habitable pavilion scale.
The project selected for development, fabrication, and assembly, proposed a novel
connection method of unique half-lap joints that could be fabricated using the com-
bination of a mitre saw and Masterwood Project 323 (Figs. 9 and 10). The mitre
saw was used to cut the timber boards to length and correct angle, which were then
placed in a pre-cut jig for milling of the half-laps and pre-drilling of fixing screw
holes. The assembly required several plywood jigs and some careful balancing and
maneuvering due to the weight of the karri. The finished structure was sculptural in
nature, yet invited interest due to interplay of light and shade and precision of the
joints (Fig. 11).
The critical findings of this experiment, however, were in the interplay between
precision digital crafting and the essential intervention of manual processes using
common wood-working tools. To fabricate this structure required an intimate under-
standing of the behaviours and characteristics of the solid timber material, which
were embedded within the computational model through machining tolerances and
toolpaths. The design-to-fabrication process exposed students to the impacts of mate-
rial characteristics on formation of structural and architectural objects, where usually
the actualization of design is through inert homogenous materials such as the melted
and squashed filament employed in FDM printing currently so popular. Crucially,
these synthetic and composite materials bear no relation to actual material for con-
struction and therefore, we believe, have limited value in any discussion of a material
specific practice such as wood or ‘timber’.
Fig. 10 Karri members in jig after machining, 2013
2015—Material Experiments
The third installment of Generative Fabrications proposed as an open-ended mate-
rial investigation, where the only limitations were budget and fabrication. Only lim-
ited access to industrial CNC machining technology was available—students were
encouraged to explore methods of making using manual tooling and reduced geo-
metric complexity. Three distinct projects were undertaken: an architectural screen
wall using karri strips in bending-active configuration; a shading structure explor-
ing the possibility of using plywood to emulate textile behaviours; a traditional
research pavilion employing plywood as a structural material. Each of these projects
investigated the behaviours and characteristics of the timber material employed. For
instance, the bending-active screen translated observed behavior of thin karri strips
into a dynamic digital model using the Kangaroo physics simulation plug-in within
the Rhino/Grasshopper software combination. The dynamic model, when manipu-
lated, enabled the students to identify an aesthetically appropriate outcome (Fig. 12).
Translation to fabrication involved the manual location of fixing points along the tim-
ber, which controlled the bending behavior of the karri once fixed in-place. We found
that the simple coordinates locating fixing points enabled them to embed complex
material behavior and formation in a highly efficient and compressed manner. This
was critical to the success of the project as the students chose to utilize traditional
hand-tools rather than a direct data-to-fabrication process employing CAD/CAM
technology.
Fig. 11 Generative fabrication sculpture, 2013
In achieving a textile form effect, the second prototype employed computational

form-finding using a similar software combination to the first prototype. A simulated
mesh achieved the stretched-membrane behaviour, which was then fixed in time to
leave a static form. This static form was subdivided into planar surfaces, which were
CNC machined and then zip-tied together. The ‘shell’, once suspended, fell into
shape, appearing as a light fabric despite weighing many times more (Fig. 13). We
found that formal goals are just one small part of the computational design process,
where simple constraints such as self-weight of material plays a large part in the
structural effectiveness of a form, which is not adequately represented within current
common form-finding techniques without significant customization.
The final prototype pursued the more traditional form of a research pavilion,
though it did so with a focus on simplicity of connections and fixings with a min-
imal material palette. Though the student team were able to complete only a small
portion of the designed structure, the resulting prototype demonstrated the structural
performance of plywood in compression as part of a vaulted typology with simple
pine shims and wood screws achieving the continually changing section-to-section
geometric relationships (Fig. 14).
2016—High-Tide Hub
In 2016, we were approached by a local Western Australian architect to assist with
the concept design and development of a small pavilion structure to be fabricated,
Fig. 12 Active-bending karri strip screen, 2015
assembled, and utilized during the inaugural event of an international arts festival.
The architects had little experience with projects of this scale; having developed a
portfolio of work comprised mostly of multi-residential complexes and mixed use
developments. They provided an early conceptual sketch of the pavilion in the form
of a timber ring made of cut plywood sheets, and sought further development from
the “Generative Fabrication” unit. The students examined several connection and
jointing methodologies utilizing the Rhino/Grasshopper computational design soft-
ware combination before exploring opportunities for application and development
Fig. 13 Textile wood, 2015
within the constraints of an exhibition and event space for the arts festival which
would be held across a two-week period.
The three projects developed by the students varied in their approach to structural
performance and fabrication, but shared material performance and assembly thinking.
The most challenging aspect of the “High-Tide Hub” project for each group was the
circular nature of the pavilion, which required attention to the relationship between
connections and geometry. While within a fixed radius circular pavilion a singular
type of connection joint is viable, but where the geometry becomes non-standard,
clever jointing methodologies needed to be developed. The three projects each dealt
with this issue in a unique manner.
The first proposal chose to explore “moiré” effects through layering a standardized
hexagonal frame around the pavilions radius. Each frame was infilled with threads; a
different orientation for each frame created subtle differences in appearance (Fig. 15).
The second proposal also employed a single curvature, but explored the threshold
between inside and outside by using vertical rectangular posts with a modulating
density indicating accessible points around the perimeter. Inside, a meandering wall
organized the space into distinct zones of activity—bar, gallery, performance, and
seating (Fig. 16). The final proposal explored complex geometry through the manip-
ulation of the circular pavilion into a non-standard swooping form constructed of
plywood ribs. Each rib, however, was identical and merely rotated as it swept along
Fig. 14 Pyramid unfolded structure, 2015
the perimeter rail curve. This simple gesture also ensured that a single family of joints
were required to be fabricated, as the rotation was linear in nature and identical from
one frame to the next (Fig. 17).
The pedagogically driven projects undertaken within the Gridshell Experiment
and Generative Fabrication units since 2011 have allowed focused interrogations
of the relationship between designer, material, and machine to occur. While the
students have grappled with developing digital computational design skills within
a rigourous fabrication-focused unit, as conveners we have been able to develop a
concurrent development of understanding of emerging approaches to designing with
natural materials within increasingly complex design and delivery frameworks. For
instance, the ability to embed datasets pertaining to material, structural performance,
fabrication, and assembly within a single digital model through complex computation
calculations and data-driven design approaches enables us to build virtual models of
any architectural structure that can be reliably assessed for viability as if they were
already actualized. This understanding has transferred from the pedagogical through
that of commercial architectural practice, where we continue to challenge the status-
quo in Australia with novel approaches to the design and delivery of timber projects.
The next section deals specifically with a selection of these projects, which have
been undertaken in parallel with the pedagogical projects, and should be viewed as
the natural extension and development of outcomes from teaching practice.
Fig. 15 Moiré pavilion, 2017
5 Complex Timber Structures
Over the course of the past seven years ATC Research and the commercial entity, ATC
Studio have carried out two complex timber projects, with a third currently in planning
stages. The first of the projects, the South Hedland Town Centre Performance Space
evolved from a brief for a shade structure in the revitalized town square of this
northwestern township through which much of Western Australia’s Iron ore travels
on its way to processing plants in China and elsewhere. The project offered the
opportunity to explore the gridshell concept at a number of different scales and with
both native hardwoods and engineered timber products. A remote site 1500 km from
Perth in an area where very aggressive wood devouring ‘Mastotermes darweniensis’
are endemic added to issues of detailing and overall performance of the structure.
Fig. 16 Circular pavilion, 2017
This was our first engagement with large a scale design modeling and manufactur-
ing process and we were lucky in many ways that the building had a very simple pro-
gram whereby the more complex ‘typical’ elements of the building (change rooms,
machine rooms, stores, bathrooms) were assembled in a stand-alone building that
acts as a backdrop to the timber structure. The result was that there was only the light-
ing consultant and the membrane manufacturer to coordinate within the process. Of
these two, the membrane manufacturer used similar processes to manufacture their
product and indeed worked directly from the same model that we constructed for the
timber fabrication.
A very different story emerged when we undertook the design of a large recreation
and sports facility in another regional township, this time only 150 kms from the
metropolitan center of Perth. The complex comprises two large buildings, a sports
hall and a function centre and cultural hall, each quite different in program and scope
of design manufacturing, and a third structure, an extensive verandah that provides
Fig. 17 Twisted pavilion, 2017
the context for the other two buildings and hosts many smaller functional areas such
as change rooms and clubrooms. In this project most of the disciplines in the building
industry have been involved and the single outstanding characteristic among them
has been their unwillingness to understand the precision required in designing for
manufacture and importantly the precision of the engineered timber product as it is
shipped to site for assembly. Importantly for us the three different building ‘types’
that make up the complex have allowed exploration of three different structural and
assembly systems, again each appropriate to its task functionally, aesthetically and
structurally. The complex uses a New Zealand produced LVL product for all structural
applications and a native plantation hardwood (Eucalyptus muellerana) for decking,
cladding and interior floors and linings.
The third of these major projects is for a forest observation and visitors tower
which is currently in concept design phase. This project offers the opportunity to
mix the structural elements between hardwood and engineered timbers. As the tower
is primarily for educational (school and tourist) purposes and is situated in an area of
conserved ‘demonstration’ Jarrah forest, it is important to maintain reference to the
native timber and how it can be used in tall structures. The lengths of structural timber
produced from the Jarrah forest are not significant and therefore complex jointing
systems are required to create a tall structure that appropriately demonstrates and
exploits the capabilities and scope of the material and the design/fabrication system.
The tower will have five platforms at levels that correspond with the change in the
forest ecosystems and will reach a height of 50 m. The historic context for the project
are the towers that were constructed as fire watch towers in the earlier days of logging
in the South West forests most of which were based on a tapered truss format.
5.1 South Hedland Shade Structure
Inspired by European gridshell developments we began looking at assemblies of

short length components for ease of transport and assembly but moved away from
this after introducing timber fabrication engineers and building contractors to the
design team and a variety of shell forms had been explored. Timber construction in
Australia’s West has long been eclipsed by the ‘brick’ industry although the state has
a significant history of timber building. To execute large scale projects in timber it
is necessary to do a lot of ground work in preparation of individuals, client groups,
as well as of the contractors who will actually be implementing the work. At ATC
we prefer to work with teams who are involved with the project from the outset to
completion, although this is often difficult when state procurement protocols have
to be followed. This project too is also in a remote location some 1500 km from
Perth and over five thousand by road form Melbourne where the only manufacturing
expertise and equipment was located at the time.2
We were keen to use a Pinasta pine based LVL that is manufactured in Western
Australia and is now the only LVL plant in Australia.3 It is a robust material for
which the mechanical properties are known and yet its application is seldom seen
as a featured or externally visible material. The form of the shell was explored
extensively in Rhinoceros/Grasshopper formats and in the typically collaborative
design process, evolved from a homogenous shell to a ribbed shell with continuous
spanning members braced equally with transverse short members (Fig. 18). The final
project is a rigid shell structure that was entirely factory built on a Hundegger K2
carpentry centre by Timberbuilt Solutions in Melbourne—the only one of its kind in
Australia at the time—from the Rhino models generated by ATC.
The shell’s form was adjusted to attenuate the cyclonic storms that are common in
this latitude and to allow for a membrane covering that was only perimeter fixed and
also to permit the maximum pre-assembly in the Timberbuilt factory before shipping.
The structure spans in excess of 16 m and rises to over 6.0 m at the apex (Fig. 19). It
2 Timberbuilt Solutions Pty Ltd. Hallam, Victoria.

3 Wesbeam PtyLtd: Neerabup, Western Australia.
Fig. 18 Virtual model in Rhino3D, 2010
rests on two plinths that begin 1.6 m off the ground, rising to approximately 1.8 m in
part to make the structure more termite proof. The plinths were also developed in form
by ATC and further manufactured by Timberbuilt Solutions. The whole assembly is
symmetrical about the centre axis and thus each element is repeated twice. The arcs
were able to be preassembled in Melbourne and the complete structure loaded onto
a single flatbed trailer. The assembly took three days to erect off the truck in Port
Hedland, with a further two days to install the perimeter fixed PTFE membrane. The
membrane was manufactured in Japan by the Taiyo Membrane Corporation and was
designed and supplied through MakMax, Brisbane.
The complexity of the form of this structure due to the non-Euclidian geometry
would have made it almost impossibly complicated to make without sophisticated
digital modelling. Knowing the mechanical properties of the predictable engineered
LVL material made the structural work challenging and rewarding rather than impos-
sibly difficult. The design process employed the Rhino/Grasshopper software com-
bination, enabling a computational model to be developed that linked the material
properties with form and performance. This ensured that design, engineering, and
fabrication teams could work collaboratively despite the large distances involved,
as the single source of digital information—the computational model—was easily
translated as simple datasets from one party to the other. The consultant teams worked
within a productive feedback loop where material and structural modifications were
easily assimilated into the design model under full control of the ATC design team.
Fig. 19 South Hedland shade structure, 2011
Moreover, the fabrication process was handled without any projective drawings, as
dataset transfer was simple and lightweight, exhibiting no loss of information as it
moved from design to fabrication model.
The erection process was facilitated by the fact that so much of the (virtual) assem-
bly and (fabrication) processes were modeled accurately in the design and engineer-
ing team’s offices, rather than being worked out empirically on site. The timber
superstructure was shipped complete with all fixing plates, screws and bolts where
appropriate. Each element is referenced so that as it is unpacked from the flatbed
truck, it can be fitted into its exact designated position in the assembly (Figs. 20 and
21). This kind of precision is not extraordinary in the world of engineered timber
fabrication and ensures that the onsite assembly runs as smoothly as possible, pro-
vided that the on-site set up for the elements is correct. The bases of the structural
arcs on this project were spliced onto the steel members, which in turn were welded
onto the plinth base plates one by one, as they were assembled. Any error on the part
of the site preparation on a project like this would mean that the assembly would not
work without significant modifications being made to accommodate the inaccuracies
or over generous tolerances (Figs. 22 and 23).
Fig. 20 Shade structure assembly, 2011
5.2 Pingelly Cultural and Recreation Centre
This project was characterised by extreme financial constraints and was challenged
by the arcane procurement processes of local government regulations in Australia.
ATC was called into re-conceptualize a project that had been costed at $12 million
for a client budget of $4.5 million. A rationalized prefabricated timber structure was
proposed that eliminated duplicated facilities and offered similar accommodation for
$4.75 million, a budget that eventually grew to 8.5 million over the course of the
two year design and documentation period. What is notable in this cost creep is that
the increased costs went almost entirely in more elaborate fit outs and servicing with
only a small increase in the originally proposed footprint. The primary structural
elements changed little except for the substitution of an elaborate air-conditioning
plant which required additional support and bracing, in place of a passively ventilated
system.
There are three structural typologies used in the project, which is entirely con-
structed on a structural timber deck that floats between 700 and 1200 mm above
existing grade on the site. The Sports Hall has the largest span at 28.5 ms with a clear
height internally of 9.1 m. A portal frame system was adopted for this element. The
Function Centre, which programmatically is more complicated than the great hall,
uses a combination of portal systems and truss frame systems, while the extensive
Fig. 21 Shade structure assembly, 2011
verandah structure, which is part open and part enclosed acting as an umbrella roof to
smaller functional areas like clubrooms and change rooms, uses a post and structural
roof cassette system to allow for a free volume below the roof.
The Sports Hall portal frames are LVL box beam construction (columns and
rafters) with 6.5 m longitudinal bays. Upper and lower wall cassette frames were
designed for ground assembly to be lifted into position between each bay. The time
allowed for erection of the structure of the Sports Hall was halved once from the
contractors estimate by the design teams’ estimate of time required, and was halved
again on site in the actual time taken to assemble and erect the components (Figs. 24
and 25).
The regular geometry of the Function Centre grid was compromised during the
latter stages of the design to accommodate client demands and a level of servicing
that had not been part of the original concept. Consultants appointed directly by
the client later in the process did not appreciate the need to maintain discipline
in the structural configuration if economy of costs in design and manufacturing
effort was to be minimized. A common thread throughout this project has been the
unwillingness of other disciplines to appreciate and work to the level of precision
that the architectural and engineering design team had worked to in the concept-
design/fabrication process. Surprise and consternation has been a characteristic of site
meetings as typical tolerances and lack of detail have proved inadequate for this type
Fig. 22 Shade structure at night, 2011
of building production. It has emerged that the need to maintain precision throughout
the design, fabrication and erection/installation process is better appreciated by the
discerning builder than by some consultant teams. With more exposure to this kind
of design and production of buildings no doubt this will change but it does represent
Fig. 23 Shade structure detail, 2011
the need for a shift in culture as much as a shift in understanding of a new type of
building technology.
The verandah structures are attached to the larger buildings but also stand inde-
pendent of them in parts of the complex (Fig. 26). The same regular grid system is
applied throughout and the cassette module for the roofing bays remains constant
wherever the rectilinear geometry of the buildings permits. The exception is at the
entry where the cassettes are opened up to accommodate a fan shaped main entry
canopy. The structure of the cassettes however remains the same. Like the floor decks,
Fig. 24 Pingelly recreation and cultural centre (PRaCC) assembly, 2017
these cassettes are all designed as co-planar assemblies, minimizing structural depth
and optimizing overall structural performance (Fig. 27).
The complex uses 350 m3 of plantation grown hardwood (Eucalyptus muellerana)
for the decking, cladding, interior floors and wall linings. Where the timber is exposed
to weather heavy duty 35 mm thick boards are used while in protected areas a 19 mm
board is used. Interior floors are either 19 mm or 12 mm depending on the subfloor
and the function, and wall linings are similarly 12 mm.
The mix of engineered timbers in the structural work and native hardwoods in
the claddings and decking’s makes the most appropriate use of the different charac-
teristics of these timber products: less uniformity is required in areas of deck and
cladding than in structural applications where the precision and predictability of the
engineered timber is required. Our design research processes however indicate that
the aim of building with timber should be to optimize the structural properties of the
forest product with the least amount of processing possible. This should be possible
with better monitoring of timber growth patterns in different environmental condi-
tions: different nutrition, insolation, exposure to rain, soil types and depth, gradient,
elevation and exposure all play a part. Until this becomes an industry standard, the
engineered product, however wasteful it is in the processing, is the material with
which we must work.
Fig. 25 PRaCC interior, 2018
Fig. 26 PRaCC verandah, 2018

Fig. 27 PRaCC function centre under assembly, 2018
5.3 Forestry Tower
The forestry tower project currently under study in the studio for an education and
orientation center poses a number of interesting problems and opportunities. As a
largely open structure by nature, issues of durability, safety and accessibility are
inherent in the design process. The project comprises a 45 m tall timber tower with
viewing platforms every 5 m in elevation corresponding to the growth stages of the
Jarrah Forest. It is currently under study as a two part structure with a conventional
rectilinear engineered timber core containing the required servicing including lifts
and escape stairs, and a curvilinear lattice enclosing structure in native hardwood
(Figs. 28 and 29).
The design issues center around form finding and optimization for material use
and for both manufacture and on site assembly. These criteria will determine the size
of components and thus weight, ease of assembly and assembly method. The design
process thus encompasses spatial, material, engineering and assembly design within
the studio environment. This amounts to a comprehensive ‘end to end’ modeling
process.
Fig. 28 Forest tower study shell, 2017. Sami Joer
Fig. 29 Forest tower study hardwood shell, 2017. Damien Lim

The aim of the project is as a celebration of timber and the digital design and pro-
duction processes, showcasing the properties of the native hardwoods and their poten-
tial as structural elements when formed through the complex design and machining
processes that CNC carpentry centres makes possible. At the same time both the
processed products of the forest—the engineered timber core—and the view to the
growth systems of the native forest complete the picture of the timber production
systems in our environment.
6 Designer-Material Relationships and Conclusion
One of the fundamental aims of the studio has been to establish a base of knowledge
about the properties of the local WA hardwoods and their suitability for use in a
number of different structural conditions. Essential to a process of rapid design
and prototyping/manufacture, the material properties and mechanical behaviors are
necessary data fields for the efficient design in timber using digital modelling and
digital driven manufacturing. Our experience with the use of engineered timbers is a
testament to this and the conceptual link between the designer (architect/engineer),
the fabricator and the resource has been thrown into sharp focus. This has been done
through our learning and understanding of the details of the mechanical properties
or characteristics of different timbers, in particular, the available native hardwoods
of the West Australian forest. But in fact any forest is the same in the terms we wish
to speak of. When once the forest or the natural resource is considered as one of the
active players in the design and production cycle then the need to better understand
the characteristics of that player becomes very important. At first this was played out
in the Shimmer Chair project at ATC, where we had no reliable data on the properties
of the regrowth timbers we were using and had to establish the basic performance
parameters ourselves by carrying out our own testing. This need for reliable data
and performance characteristics became important again through a number of the
Genfab projects and through some of the projects currently under study.
By describing the forest as an ‘active participant’ in the design system and process,
we are not suggesting that the forest is ‘personified’ any more than any other material
or important contributor to the design process is. Rather we want to emphasise that
it is no longer enough to be content with the knowledge that timber is an anisotropic
material and we just have to live with it! We need to know more about the forest, the
trees in it, the soils, the micro climates and other factors that pertain to the growth
and development of the tree, so that we can make better use of the ‘anisotropicality’
of any given log. The current system of using reductive blanket categorizations (F
grades and MGP grades in Australia) devalues the tree by not acknowledging the
unique characteristics of a given log—something that every sawmiller knows and
acknowledges but all too often remains unheard—and which in turn devalues the
forest making investment outcomes of the harvest ‘unreliable’. In turn, the banking
sector then is reluctant to invest funds in either research, forest development or forest
‘husbandry’, and forest maintenance. We need to know the ‘terroir’ of our forests
as well as the viticulturist knows his vineyards and vines. We need to acknowledge
and be able to work with the subtle differences in trees as the wine maker knows and
refines the subtle differences of the grapes he has tended and harvested. The perceived
‘problem’ that the industry has to deal with could emerge as the materials greatest
asset, but until we invest in and develop the knowledge of our forests the problem
will continue and one of the planets greatest resources will remain undervalued.
Ultimately we would like to be in command of enough data about the growth and
maturity of a given tree or timber, in a given part of the forest, to be able to reliably
predict the most appropriate applications for that timber. We believe that over time
this would fundamentally change the economics of forestry and investment in the
downstream forest industries. While we acknowledge that the Australian example
is not characteristic of many of the more developed forestry industries, the peculiar
mix in the resource, and its precarious state of survival or annihilation in the face of
ignorance, it remains an important example and laboratory for the forest industries
worldwide.
At the same time, we have encountered philosophies of design modelling, and
material representations within digital space in particular, which have provided a crit-
ical frame within which to view the research that we have undertaken. Of significant
influence on our work has been understanding of translative events between data (the
virtual) and object (the actual); a continuum within which immanent matter—em-
bedded in the digital—presents an opportunity for architectural structural forms to
be informed by the irregularities and peculiarities of wood. An immanent theory of
matter establishes that form emerges from within, independent of the human mind
(De Landa 1999), characterized by processes of translation from one form to another
as part of a larger dynamic system of feedback loops and the collision between
extensive structures and intensive difference (De Landa 2002).
We reject the notion of ‘wood’ as a singular homogenous category of mate-
rial, instead acknowledging the near-infinite variability that exists within each tree,
let alone the forest. Computational design approaches enable the designer to engage
with natural material as a dynamic system. It is quite easy to see how the use of a
visual scripting engine such as Grasshopper, through the use of nodes and wires,
reflects the series of phase transitions and translation events that are characteristic of
a neo-materialist view of matter. The Grasshopper script itself represents a narrative
of a dynamical model, wherein embedded material characteristics play an active role
in the genesis of form, in cooperation with the designer.
The precision of the digital model is translated directly into the engineered timber
product with minimal tolerances that can be held true for very large structures in a
way that can be more efficient and reliable than in other materials typically employed
in prefabrication processes. When this means of production is used for hardwood
timber or for the forest product generally, then known calculated tolerances can be
built into that model system. Clearly, data sets directly relevant to the material being
used, and a clear understanding of the theoretical understanding of the relationship
and process of idea, data, and object will better inform the architectural product. This
emphasis on the recognition of the context and importance of the equal and symbi-
otic partners—the ‘family of authors’—in the process designer (architect/engineer),
forest, (timber as raw and engineered products) and machine (digital model, digital
fabrication) should make possible a truly holistic approach to sustainable practice
that can recognize and actively be accountable for the complexities of working in
the human and natural environment.
Acknowledgements Members of the ATC, 2006—Present: Gary Marinko (Director), Tobias Beale,
Daniel Bubnich, Robyn Diggins, Gwinyai Dzinotyiweyi, Mannchee Mishelle, Tara Moore, Shannon
Stanwell, Domenic Trimboli, Harry Reynoldson, Ian Weir.
Narrogin Joinery: Stan Samulkiewicz, Michael Samulkiewicz.
Gridshell Experiment and GenFab 2012 Students.
GenFab 2013 Students: Tobias Beale, Henry Beeck, Abby Chittick, Ian Dodson, Eliza Langham,
Chia Lee, Mark McKenna, Michael Warr, Steven Wong.
GenFab 2015 Students: Maeda Bahremand, Thomas Ffoulkes, Luke Geiles, Daniele Giuffre,
Mitchell Hender, Rui Hou, Geok Sing Khor, Kah Wai Leong, Xunyi Luo, Ashley Mutton, Luke
Myers, Veng Fei Ong, James Palmer, Simon Sawyer, Rui Tan, Jacob Turner, Zachary Wood, Hao
Zhang.
GenFab 2017 Students: Jack Bradshaw, Hazem Halasa, Mark La Riviere, Jovin Lim, Shin Low,
David Smith, Michele Stambulich, Shamisa Tahery Darestany, Trevor Wong.
References
Cache B et al (2007) Objectiles: Patrick Beauce + Bernard Cache: Fast Wood: a brouillon Project.
Springer, New York, p 2
Cache B (2011) Projectiles. Architectural Association, London, p 20
Carron LT (1985) A history of forestry in Australia. ANU Press, Canberra, pp 137–139
De Landa M (1999) Deleuze, diagrams, and the open-ended becoming of the wold. In: Grosz EA
(ed) Becomings: explorations in time, memory, and futures. Cornell University Press, Ithaca, pp
29–41
De Landa M (2002) Philosophies of design: the case of modelling software. In: Verb processing.
ACTAR Press, Madrid, p 1
McNeel R (2018) Rhinoceros 3D. McNeel North America, Seattle, Washington
Lattice Shell Methodologies: Material
Values, Digital Parameters
Mark Cabrinha, Dante Testolini and Ben Korman
Abstract Lattice shells combine an elegance of form with the efficiency of struc-
ture driven by the material constraints of straight lath members that can be bent into
shape. While formally expressive, the form is the result of an explicit methodol-
ogy combining form-finding, material constraints, and construction logistics. As the
boundary curve establishes the constraints of the system, it is the boundary curve,
not the surface, which gives the designer discretion over form. From this boundary
constraint, the form is developed through the forces applied in the form-finding pro-
cess such as a vertical load vector (pushing) and/or surface relaxation (stretching or
equalizing). Although these values can be adjusted by the designer, they are only
meaningful when calibrated by material constraints. Through physical testing real-
time material feedback can be embedded into the parametric system. In combination
with form-finding, the use of geodesics constrains fits lath members to the compound
curved shell such that it can be constructed from straight lath members. As an elegant
response to how material can inform form, by integrating these processes into a para-
metric workflow, further attention can be applied to other design criteria including
spatial development and environmental response while maintaining the elegance and
structural economy of shell structures.
Keywords Lattice shells · Gridshells · Form-finding · Geodesics

Digital fabrication
1 Introduction
The tension between form and material is neither a new one, nor is it necessarily a
digital matter. As early as 1934, Henri Focillon suggests that matter imposes its own
M. Cabrinha (B)
California Polytechnic State University, San Luis Obispo, USA
e-mail: mcabrinh@calpoly.edu
D. Testolini · B. Korman
d2bdesign, Cayucos, CA, USA

196 M. Cabrinha et al.
form upon form (Focillon 1934). With today’s digital fabrication tools, material is
all too often seen as a homogenous substrate with the outputs of digital fabrication
tools essentially creating physical representations of digitally derived form. Useful
and at times inspiring, yet clearly this matter does not impose its own form upon
form. Said another way, there is not an inherent material feedback loop in the digital
design process. Contemporary digital fabrication tools are not simply about material
output of preconceived digital form, but can be used as a means to extend the material
capacity, complexity, and variation of a material system.
The potentials for a bottom-up, material centric design approach have strong
philosophical, conceptual, and even ethical values. Manual DeLanda, like Focillon,
has been critical of the philosophical indifference to material that is often preva-
lent today. This Platonic indifference suggests that ideas are simply actualized by
the addition of matter—a view that matter is an inert receptacle of external ideas.
Rather, the view DeLanda takes up is that material is an active participant in the
genesis of form. He identifies two central aspects of material qualities: their capac-
ity and tendency (DeLanda 2004). To oversimplify this case in the wood spline, its
capacity is its strength in bending and its tendencies develop from its axis in bend-
ing (bi-axial vs. uni-axial). While these capacities and tendencies can be viewed as
constraints, constraints are not simply limitations, but become productive by shap-
ing design parameters. Viewing constraints as productive can have a powerful effect
on design concepts. After all, attentiveness to materials directs attention beyond the
symbolic and representational nature of design and instead prioritizes the environ-
mental, kinesthetic and haptic experience of architecture. Consequently, the primary
visual interface of digital media may be balanced by the material influence through
taking the wood spline quite literally. Through this, there is a connection between
material values and digital parameters. Suggesting that values and parameters are
not, in the end, synonymous, provides an opportunity to productively connect the
two rather than conflate them as one and the same thing.
The formal flexibility in today’s computational tools, such as NURBS surface
modeling, places emphasis on “form” without consideration for material or structure.
As Frei Otto has remarked, “the cry for form comes from those who have no idea
of the origin and extent of the infinity of forms” (Nerdinger 2005). As exemplified
by the expressive forms of Otto and his Institute for Lightweight Structures, “form”
is not independent or isolated from real-world criteria, but develops from the very
constraints of the situation. In the case of lattice shell structures, material constraints
become operable design criteria.
Ironically, the origins of NURBS based tools in use by current design practices
developed from the ability of wood to bend—the physical spline. In the 1960’s,
mathematician-engineers such as Pierre Beziér abstracted this physical spline, includ-
ing the material resistance of the wood and the geometric constraints of weighted
ducks, into what we now know of as NURBS. While many may argue that the free-
form fascination in contemporary architecture is a consequence of NURBS based
software, an overlooked historical fact is that material constraints and parametric
flexibility were the foundations of the system (Bezier 1998). The power of today’s
computational tools enables the designer to couple particular constraints of the sit-
Lattice Shell Methodologies: Material Values, Digital Parameters 197
uation with the flexibility to adaptively shape these constraints through parametric
control. Consequently, the issue is not the flexibility of the tools, but the very mis-
conceived notion of “free-form”.
The distinction between “free-form” and “form-finding” are often conflated
between architects and engineers. Historian of structural engineering David Billing-
ton has noted the distinction between scale, use, and form that has separated architec-
ture from engineering: while architecture is focused on creating space, the structure
is designed to control force, rather than space (Billington 1983). Replacing this free-
form fascination with pure-form equilibrium shells is problematic for the reasons
Billington identifies: it shifts the focus from space to the optimization of forces;
from the work of the architect, to the engineer.
The motivation to develop a parametric methodology for lattice shells is to identify
a design approach through which material constraints, form-finding, and construction
logistics can be integrated. By integrating these processes into a parametric workflow,
further attention can be applied to other design criteria including spatial development
and environmental response while maintaining the elegance and structural economy
of shell structures.
2 Lattice Shell Precedents
While work on lattice shell structures occurred much earlier, such as the steel lattice
structures of Russian engineer Vladimir Shukhov in the 1890’s, Frei Otto along with
a young Ted Happold, then at Ove Arup, pioneered the use of timber lattice structures
exemplified by the Mannheim Multihalle gridshell in 1975. Although relatively few
lattice shell structures have been built since then, there has been a recent resurgence of
gridshell structures most notably Shigeru Ban’s Japanese Pavilion (2000), the Weald
and Downland Gridshell (2002), and the Savill Building (2006). These examples also
demonstrate the limited formal morphology in their similar use of a doubly curved
hourglass shape. This similarity in shape is driven as much by structural necessity, as it
is from their construction approach. These structures are fabricated from a flat matt of
straight laths, or paper tubes in Ban’s pavilion, and then slowly raised and/or lowered
into shape, often with the use of complex telescoping and articulating scaffolding.
Helsinki University of Technology Wood Studio’s “timber bubble” at the Helsinki
Zoo (2003) is unique in its capacity to support its floors, as well as in its construction
technique assembling the wood laths in a piece-wise approach, instead of a flat matt
(although in this case this also required steam bending the members into place).
Thomas Herzog’s Expodach, also for Expo 2000 in Hannover as was Ban’s Japanese
Pavilion, was constructed from repeatable prefabricated lamella lattice shells each
formed over an immense mold. Toyo Ito’s media center in Gifu (2015), provides an
alternate lattice shell construction creating its long-span structure through relatively
thin lattice members that are glued together on site as if a gigantic glulam lattice
structure.
Shigeru Ban has pushed the limits of timber gridshell design with his formally
exuberant design for the Centre Pompidou-Metz (2010), in France, as well as the ele-
gant Nine Bridges Country Club (2009) in South Korea. Unlike the previous bending
active structures, these structures employ heavy timber sections digitally fabricated
through the complex parametric workflow of design-to-production, a design consul-
tancy that linked the complex design intent with the Swiss timber fabricators Holzbau
Amann (Pompidou Metz) and Blumer-Lehmann (Nine Bridges). Despite the com-
plexity of design, engineering, and fabrication these two structures exhibit, in the end
they do not demonstrate Otto’s “lightweight principle” at the core of bending-active
lattice shell structures. It is interesting to note that the parametric sophistication of
design-to-production is brought in as an essential part of the fabrication logistics
toward the end of the design process, rather than earlier in the conceptual design
process. When I interviewed Fabien Scheurer, founder of design-to-production, and
coincidentally when he was just finishing his role on Nine Bridges, he had this to
say:
In most of the cases where this fabrication [is] actually done, we’re hired by either the
fabricator or by the general contractor and not by the architect. This is mainly a problem of
the process itself. There’s a straight cut between the design phase and the building phase,
which means that there’s no information coming from the backend of the process up the
chain and informing the design process, which is a shame I have to say. The quality of the
outcome can be much higher if the form, the shape, the design and the materiality match in
a way. (Cabrinha 2010)
3 Parametric Process as Manipulable System from Design

to Construction
The biggest challenge in gridshells is not simply the geometry and its representa-
tion, but the separation between material intuition, design process and its digital
representation, and construction. In reflecting on the design and construction of the
Weald and Downland Gridshell, architect Richard Harris notes the collaborative
nature of the design team, including not only the client, architects and engineers, but
timber framers and scaffolding contractors that were central to the “fluidity of think-
ing” which the primary material—wood—suggested (Johnson 2006). This dynamic
process between material, design and construction is as true of the concrete shell
builders. After all, Pier Luigi Nervi was known not only as a great engineer, but as
a builder connecting intuition, analysis and construction that defines Nervi’s work
(Poretti 2010). Similarly, Felix Candela remarked “the only way to be an artist in this
difficult specialty of building is to be your own contractor” (Billington 1983). The
Korlandia gridshell, which presents the proof-of-concept of the process presented
here, is no different. As the outcome of an intensive if obsessive design-build process,
the construction of this lattice shell combines material intuition and analysis, engi-
neering, geometric precision, and experimentation from early conceptual design in
2012 to the completion of its construction in 2018. Rather than the mythical ideal of a
single parametric process in complete control of all variables, innovation and impro-
visation proceeded at moments of human discretion from the form-finding process
to the construction of full-scale templates, from the precision of digital fabrication
to the tacit knowledge of experienced craftsmen. This relationship between design,
intuition, and precision through full-scale experimentation is a critical component of
design innovation that is as old as the profession itself.
The Gothic cathedrals can be seen as a working laboratory between the construc-
tion site and the lodge—between tacit and explicit knowledge. Templates, geometry,
and a set of simple tools from compasses, straightedges and string enabled incredible
innovation. The full-scale templates themselves became the primary currency, the
medium of exchange between design as conceived and constructed. Rather than a
single point of control, the aggregated knowledge of skilled specialists, experimen-
tation and practice, through string, geometry and templates has been described as a
manipulable system (Turnbull 1993). This system begins with the ability of natural
phenomena to generate data—as simple as giving form to raw stone from templates
and importantly the ability to manipulate this system from design through construc-
tion. When viewing parametric design as such a manipulable system, material values
must be given as explicit design parameters. As exemplified in the construction of
the Korlandia gridshell, rather than a mythical ideal of total parametric control, as
a manipulable system in continuous play from design to construction ad hoc local
adjustments and decisions were able to be made throughout the construction process
enabling human discretion at critical points throughout the process from design to
construction. In the attempt to outline an explicit working methodology for lattice
shell structures as a manipulable parametric system, we must not forget the goal
of this geometry is to extend the possibilities of our material imagination through
the execution of built work. In the end, as with all shell builders, material intuition,
geometry, and construction must be interrelated.
4 Lattice Shell Principles
The very definition of lattice shells combines form, material, and construction logic.
In the thorough description of the Multihalle Mannheim in IL 13, Ted Happold and
Ian Liddel define the lattice shell as:
The term lattice shell is used in this paper to describe a doubly curved surface formed from
a lattice of timber laths bolted together at uniform spacing in two directions. (Burkhardt
1978)
Surface curvature in shell design has utility—uniting form and structure. Form
finding is a critical component in developing shell designs, such as the hanging chain
models Otto employed. While there are numerous form-finding approaches for shell
structures (see Bechtold 2008), in general the surface can be conceived as a flexible
membrane (soap film, hanging chains, catenary network) but it is the boundary curve
that constrains the membrane to take its shape under a given load. Consequently, in
a form-finding workflow it is the boundary curve, not the surface, which gives the
designer primary discretion over form. As will be described later, the boundary curve
warrants much further attention in both initial design concepts as well as fabrication.
In the approach identified by Happold and Liddel, the uniform spacing of timber
laths connects both to the form-finding process (analogous to hanging-chain) as well
as to the flat matt construction technique in which the gridshell is either slowly
raised and/or lowered into its shell form. While elegant in this approach, the flat
matt technique limits the formal morphology of the shell, often restricting the shape
to straight vaults (such as Ban’s Japanese Pavilion and the Weald and Downland
Gridshell). Other approaches are possible, such as the aforementioned HUT’s Timber
Bubble, or in a hybrid approach, such as the Savill Building, which was constructed
from a two-layer matt lowered onto the perimeter beam, with the next two layers
placed piece by piece over this surface (Harris et al. 2008). Otto’s Mannheim Pavilion
was unique as it was constructed from a series of flat matts that were pieced together
in their final free-form appearance.
If the formal morphology is limited by uniform spacing, can a wider formal
morphology be achieved by using non-uniform lath spacing? Non-uniform spacing
also means the flat matt approach will not work, as overlapping individual laths
can only be pinned when they are in precisely the desired bent shape. Employing the
geometric precision in today’s digital modeling tools, each precisely placed hole acts
as a registration system pinning each member at only their desired location in its final
form, which in theory, also means that the complex telescoping scaffolding used in the
flat matt approach is not required. Connecting form-finding with construction logic is
now met with material logic: this is only made practically possible by understanding
the difference between elastic and plastic bending of lath members.
Traditional approaches to wood lattice structures require continuous lengths of
straight grained green solid timber, such as white oak or larch. These members are
slowly bent into their desired shape over a matter of weeks allowing the wood fibers to
relax into the shape, and therefore a green wood member is desired to allow for plastic
bending, subtly deforming the member into shape over time. While plastic bending
results in the subtle deformation of material through stretching, elastic bending is
what we normally experience in holding a piece of wood allowing it to spring back
into its original shape. For the non-uniform spacing proposed here, employing elastic
bending is essential to fit each individual lath into place, as the time required for plastic
bending for each individual member would be far too slow, and steam bending is
both time consuming as well as weakens the strength of the wood. At the same time,
the use of elastic bending makes possible the use of engineered lumber, such as
laminated veneer lumber (LVL), which is precisely graded for strength, and can be
manufactured in long lengths.
Taking this alternative approach of non-uniform grid spacing opens up a wider
typology of lattice shell shapes than could be created from the flat matt approach,
while providing an alternate construction technique that can minimize the complexity
of scaffolding required, and through elastic bending enables the use of engineered
lumber. But this approach leaves one last question: how to predict exactly how a
straight lath segment will lay across a doubly curved surface?
This is made possible through geodesics. Geodesics are mathematically defined

as the shortest distance between two points on a surface. Because a straight line
is the shortest distance between two points, it follows that when a geodesic curve
on a surface is flattened or “unrolled” it will likewise be straight. This is a purely
mathematical centerline without consideration of material properties or cross section,
as the width of material moves from this centerline twisting is possible. However,
based on the small lath cross section relative to the limited degree of curvature
possible in elastic bending, twisting is negligible. On the other hand, the material
limit of elastic bending must be introduced (or analyzed) bringing this material
constraint back into the system.
5 Parametric Approach
This parametric approach to lattice shells has developed through an incremental

multi-year development of bending-active structures through teaching, research, and
consultancy. This parametric approach provides a flexible framework that can capture
the combination of a bottom-up approach driven through material constraints, econ-
omy, and structure, while maintaining a certain degree of top-down formal expres-
sion. The intention of this parametric methodology is not to automate lattice shell
structures, but rather to identify the precise points at which human discretion is
needed.
This approach employs geodesics fit to a single NURBS surface, and therefore,
the surface morphology is likewise constrained to the basic constraints of NURBS
surfaces. While the tendency of novice users is to simply loft everything and join
multiple surfaces into a formal representation, structural continuity and continuity of
surface logic go hand in hand. If surfaces are joined, each surface would then need
a ridge beam at this joint to carry this load to the ground, while each single surface
in this assemblage would follow the same process identified here. For example, a
multi-surface approach could be developed similar to the Expodach (see Herzog
2000).
6 Lath Pattern: Projected, Applied, and Geodesic
From a given NURBS surface, a pattern can be projected onto the surface (typically
from the ground up), a pattern can be applied stretching this pattern proportionally
across the surface in reference to the density and shape of the UV curves, or a pattern
can be mathematically derived from a network of geodesic curves. If lath surfaces
are developed from projected or applied patterns, they will not be straight. While
these arc-shaped unrolled surface profiles could be cut from standard board sections
on a CNC router, not only would this be materially inefficient, but in fact cutting
Fig. 1 In following the

shortest path between two
points, geodesics often fail to
yield useful results
suggesting a necessary
relationship between surface
and geodesic
across the grain would lead to material splitting when stressed in its curved position
on the surface leading to failure.
Only a geodesic pattern is able to be unrolled into straight laths, and yet developing
a lattice pattern from a network of geodesic curves presents a unique challenge.
As the geodesic curve is a mathematical solution to find the shortest path, it does
not necessarily yield useful results (Fig. 1). While mathematically accurate, these
“failures” need to be resolved through the reciprocity between form-finding and
pattern logic.
7 Previous Work
Previous work (Cabrinha 2008) found that by employing surface relaxation to a

given surface, a geodesic network can be evenly distributed and is therefore able to
be developed from straight laths (Fig. 2). Note the boundary curves for Figs. 1 and
2 are identical, but the surface of course is not. While successful as a first step, this
solution neither accounted for the constraints of material to bend into this position,
nor if the relaxed surface was structurally viable. To evaluate the structural stability
of this relaxed surface, a preliminary analysis by Buro Happold, Los Angeles found
that a combination of surface relaxation with an upward load vector was necessary.
Perhaps not surprising in hindsight, this upward load vector resulted in a catenary
cross section.
Fig. 2 By employing
surface relaxation to the
surface in Fig. 1, an even
distribution of the geodesic
pattern can be achieved
suggesting the relationship
between form-finding and
geodesics
8 Parametric Lattice Shell Development
Parametric processes require an explicit methodology typically reduced to geometry

alone. The approach taken here incorporates material constraints as geometric criteria
through empirical bending tests.
8.1 Parametric Wood Seminar
In Winter 2011, my parametric wood seminar made two developments from this
previous work. First, a parametric solution was achieved that developed a four-layer
geodesic lattice network over a given surface, while unrolling each unique individual
member for rapid fabrication of laser-cut models. This parametric solution enabled
a matrix to be developed from a family of solutions including changes in warp
and weft lath densities, as well as variations of different lath cross sections (Fig. 3).
Second, through empirical testing of minimum bending radii of physical lath samples,
minimum curvature circles were measured so that a simple parametric model would
be limited by this minimum curvature (Fig. 4). These values were then tested at
full-scale through a rain-screen application using 19 mm × 88 mm pine laths. From
this work, the remaining thread was to introduce form-finding into this parametric
workflow.
Fig. 3 The parametric wood seminar developed a parametric solution for the development of
geodesics on a surface that were automatically unrolled for fabrication
Fig. 4 Empirical testing of lath members to encode the bending radius into the parametric definition
8.2 Smart Geometry Gridshell
The Smart Geometry Gridshell introduced form-finding into the previously devel-
oped parametric workflow and tested this process through a full-scale installation.
Fig. 5 The smart geometry gridshell was a proof-of-concept of this approach resulting in a formal
type that could not have been constructed through the flat matt technique
This served as a proof-of-concept of the non-uniform spacing approach resulting in

a formal type that could not have been developed through the flat matt technique
(Fig. 5).
Along with streamlining the previous work of the Parametric Wood Seminar,
form finding was integrated into this workflow through Kangaroo, a physics-based
plug into Grasshopper. The boundary condition was established from a catenary
control rig, from which the surface was form-found through a network of springs
in Kangaroo which balanced the equalization of spring lengths (surface relaxation)
with an upwards load vector (inverse hanging-chain model). Curvature analysis was
employed to make sure that the individual 38 mm × 6 mm pine laths could be bent
into the shape proposed (Fig. 6).
Rather than unrolling individual laths for full-scale construction, perhaps the most
interesting development in the Smart Geometry Gridshell is that the only output
required for full-scale fabrication was the lath length and the spacing between holes
supplied through output to a spreadsheet compiled in a single A3 fabrication table for
each of the four layers. A single plan drawing was all that was needed for assembly by
organizing the laths around the central ring. After the first attempt to raise the entire
surface into place proved impossible, the structure took shape easily when lifted from
smaller catenary arch sections. With the first two layers assembled, the surface was
highly unstable until the additional two layers were put into place piece by piece.
Consequently, in the future a simple frame scaffold at precise surface targets would
enable a much quicker, precise, and more logical assembly process. Furthermore,
while the surface was developed through a precise boundary curve, the complexity
Fig. 6 Minimum radius of bending was analysed relative to bending tests on physical lath samples
to make sure the geodesic curvature was within the material constraints
of fabricating this boundary curve in the time and material we had added to the
complexity of assembly, and in the end meant the structure was not stable.
The Smart Geometry Gridshell provided a proof-of-concept that a parametric loop
between form-finding, material constraints, and geodesic lattice network with a sim-
ple fabrication process and (somewhat) simple assembly process could be achieved.
However, as a working methodology for full-scale structures, this approach needed
to be tested beyond installations with flexible thin pine laths, to structural lath mem-
bers such as the 35 × 50 mm laths of the Weald and Downland Gridshell to the 50 ×
80 mm laths of the Savill Building.
9 Korlandia Gridshell
The Korlandia Gridshell, in California, represents the full implementation of the

methodologies presented here through a 50 (15.25 m) × 140 (42 m) lattice shell
constructed from 2 (38 mm) × 4 (89 mm) Laminated Veneer Lumber (LVL) mem-
bers. Design development of the gridshell began in August 2012, with the gridshell
completion in May 2018. With the gridshell now complete, these construction real-
ities confirmed this methodology and provide insights into future implementations.
The jobsite of the Korlandia gridshell is not only unique due to its 40-acre bluff-
top views of the Pacific Ocean, but due to the vision, interest, and fortitude of the
design-builder/owner, Ben Korman, who set out to develop this gridshell for his own
residence. As a design-builder of several innovative houses on the central coast of
California, developing an innovative lattice shell for the first time involved risk and
uncertainty.
We set out on a project that we believed would work, but required constant inno-
vation, persistence, and dedication. The connection to design ambition, construction
and innovation was enabled through digital tools, but only realized through exper-
imentation in full scale construction. Constructed previously on the property is an
innovative wood tensile structure used as the workshop testing out full-scale jigs
and templates with a 3-axis MultiCam CNC router, the development of a custom
site-mounted 12 (30 cm) × 60 (18.25 m) 3-axis CNC router to fabricate the LVL
laths, and most recently the acquisition of a used Kuka KR2 to implement numerous
tasks. In this way, this workshop is akin to a working laboratory, just as the lodge
was the site of experimentation in Gothic cathedral building (Turnbull 1993).
9.1 Design Process
The initial schematic design of the house developed by Ben Korman and designer
Humberto Norman included the rough approximation of an ovoid roof outline for the
gridshell. As they researched gridshells, they came across my previously published
work (Cabrinha 2008). I had heard of Ben’s previous work, but had yet to meet
Ben or Humberto, though we were only 20 miles apart. While it is common in
a conventional design process to proceed from floor plan to roof, the governing
geometry of the boundary curve in shell design means that you start with the roof
and work down from these parameters. A guiding design philosophy of this project
was that as more design parameters were incorporated into the project, through
these constraints a beautiful result would be resolved. The introduction of working
parametrically, through Grasshopper, was crucial to this philosophy which extended
from design through construction as new parameters became known.
With this parametric shift in design development, the parametric model developed
to include a feedback loop between the programmatic needs and formal interests
established in schematic design, form-finding with quick export for structural anal-
ysis feedback, and real-time visualization of the bending capacity for the 38 mm ×
89 mm Laminated Veneer Lumber (LVL) members. To give the owner control of the
shape generation, the parametric set-out geometry was rationalization from the given
plan geometry, a control surface to manipulate the undulating boundary curve, and
a series of catenary arches fit between column lines to form the structural shell sur-
face (Fig. 7). Rather than simply pure form-finding, this provided an idiosyncratic
approach with just enough form-finding logic while giving the owner individual
control of the boundary curve, and therefore the roof shape. While form finding
was initially developed in Kangaroo, simply developing the form through paramet-
ric catenary curves proved to be computationally more efficient, enabling a quicker
feedback loop in this iterative approach. In either approach, structural and material
criteria need to be incorporated into the system as the parametric force values are
not calibrated to real world criteria. While the structural engineers for the residence
analyzed several iterations of the surface using deflection analysis in RISA, in the
Control Surface:
Plan Curves are projected onto Control
Surface to create boundary curve
Variable:
Height of Catenary Curves
Inputs:
2 Plan Curves
5 Column Lines
Fig. 7 The parametric set-out geometry included only two plan input curves, 5 column lines, and
a control surface to manipulate the undulating boundary curve. From these inputs, the boundary
curve was formed, and projected up to the control surface, with 3 catenary curves developed with
the column lines, from which a form-found surface was created
end, it was an understanding of the material constraints that provided the balance
between economy, aesthetics, and structure.
9.2 Material Constraints and Analysis
Custom milled green timber was beyond the budget for this project, and the code
requirement for high-strength timber members pointed the way to engineered lumber.
Nominal 2 (38 mm) × 4 (89 mm) LVL members were readily available in 60
(18.25 m) lengths as they are typically used as chord members in wooden trusses.
Green lumber in typical lattice shell construction allows the wood cells to stretch,
resulting in plastic bending. The use of engineered lumber, however, would restrict us
to only elastic bending due to the laminated nature of the material. At the same time,
an elastic approach to bending was more appropriate for the non-uniform lath-by-
lath piecewise assembly process developed here. Based on the manufacturer’s load
values of the LVL, the engineers were confident the material had sufficient strength.
However, we did not know if we could actually bend the laths into the curvature
necessary. In consult with the structural engineers, we conducted 3-point bending
tests with 20 (6 m) long members to identify the minimum radius in bending of
these members. Based on the current design at that point, we needed 24 (609 mm)
of displacement at the center-point of the member, which established our target.
Using a come-a-long, nine members were bent at ½ (12 mm) increments until
they failed, noting and logging audible sounds of cracking along the way. One mem-
Failure
Design
25.75”
215/8”
24”
20’-0”
3”
r. 297
r. 27
Bending Failure
30"
28"
26"
24" Design Target
22"
Failure
20"
Audible Cracking
18" Start
16"
14"
12"
LVL1 LVL2 LVL3 LVL4 LVL5 LVL6 LVL7 LVL8 LVL9
1.5” x 3.5” Laminated Veneer Lumber (LVL), 2400 PSI
Fig. 8 LVL curvature tests to identify geometrical bending constraints prior to material failure
ber failed prematurely before the target, but the rest of the members made it past the
target curvature before failing (Fig. 8). Based on these displacements, the radius of
curvature fit through these displacements was drawn so this minimum radius could
be analyzed in the parametric model forming a real-time material feedback loop.
During this final adjustment to allow for allowable bending, the parametric model
also provided a particular insight into future construction logic. Based on the roof
shape at that time, many of the lath members were beyond the 60 (18.25 m) length
limit of each individual lath requiring the development of a connection joint that
would have to be determined and tested, or best avoided all together. Through a real-
time quantitative assessment of all the lath lengths made possible by the parametric
model, it was found that with a 6 (1.8 m) offset in the boundary curve, subtly reducing
the width of the plan, combined with a slight angle of rotation in the lath grid,
every lath would fit within the standard 60 (18.25 m) lengths. With this parametric
approach, the bending capacity of the members constrained the shape, and the length
of the members established the extent of the shell.
In the end, the material constraints drove the successful outcome of the shell.
Previous to the bending tests, the structural analysis indicated many areas in the
shell with unacceptable deflection, but with the material feedback loop added to the
parametric form finding model, the “right” shape resulted. I distinctly remember the
engineer’s response: “I don’t know what you did, but this model is working great.”
Fig. 9 The compound

curved boundary curved was
rationalized into 13 radial
arcs. As each member is a
single arc, it lies within a
single plane as indicated by
each of the 13 great circles
9.3 Design for Fabrication
At this stage of design, emphasis was placed on the form of the shell for structural
analysis, while the residence was redesigned for the optimized shell boundary curve,
and submitted for building permits. However, the downstream fabrication logistics of
the double curved 12 (32 cm) steel boundary curve and the associated connection of
the lath members to the boundary curve (edge beam) were not introduced until after
design and permitting. With this parametric approach, it was surprisingly easy to
rationalize the doubly curved boundary condition into 13 radial arcs fit to a tolerance
of 3 from the initial boundary curve as required by the engineers (Fig. 9). The more
complex detail was connecting the laths and their associated load into the edge beam,
which in turn effected the lath layout.
In the early design phase, two approaches were identified to develop the lath end
points: projecting a uniform plan grid to intersect with the boundary curve, or to
subdivide the boundary curve allowing lath members to intersect at equally spaced
points along the edge beam. At this early stage, it was decided to follow the projected
approach as this gave a desired appearance of a more uniform grid spacing to allow
for greater optimization of plywood sheathing panels. When it came to fabrication
planning, however, the steel connections of these projected lath members meant there
was no logic in there spacing nor how the members came together at the edge beam.
As these connections would also be exposed, at the end of design and permitting it
was decided to evenly subdivide the boundary curve, thus changing the lath patterns
across the entire surface, and resulting in each lath member coming together at the
edge beam in a series of equally spaced points. While the surface did not change,
this lath spacing developed more evenly across the surface with the opposing warp
Fig. 10 Each of the 152 unique lath connections were welded in their orientation off-site located
using centerlines and a tape wrap template. After erection on site, and with the rotation of the 13
singly curved edge beam sections, the continuity of the compound curve edge beam can be seen,
and the angle the surface will take
and weft of lath members connecting at a V-node. As this now followed a logical
approach to the lath termination, a parametric approach to developing each of the 152
steel connections could be developed, each unique in their geometry but following
the same logic (Fig. 10).
9.4 Steel Fabrication and Erection
Just as the boundary curve is the governing geometry for the design of shells, the edge
beam controls the construction logic and assembly of the shell. In the Smart Geometry
gridshell, given its 4-day design to assembly time and budgetary constraints, there
was no edge beam to work from. The installation was only manageable due to the
small size, flexibility of the thin laths, and a crew of people. What it did illustrate was
not only the crucial role of the boundary curve in shell design, but the construction
logistics provided by the edge beam. For the Korlandia gridshell, the steel fabricators
received the 12 HSS bent to the rationalized radii, and individually labeled water-jet
cut pieces for each of the 152 lath connectors. Considering that each of the singly
curved members would be rotated in 3d space to construct the doubly curved edge
beam, the trick was how to accurately locate and weld each lath connector to the edge
beam at the precise angle required. The fabricators were able to identify a level line
for each member on a flat plane using a “curve-o-matic” level, through the digital
model a series of full scale tape-wraps were made to locate the center point of each
lath connector plate from this level line. A full-scale protractor-like jig was created
to precisely scribe the ends of each of the 13 pieces of the edge beam, along with
alignment tabs to aid in the positioning of the rotated HSS sections during erection.
The edge beam is supported by a series of chevron braces that resist gravity, lateral,
and torsional loads of the shell on the beam, with these chevron braces resolved to
the foundation by 18 cast pin joints (Fig. 11). While there were challenges in the
steel fabrication and erection, primarily on account of the ovoidal deformation of the
circular cross sections of the pipe common to radial bending, in the end, it was the
craftsmanship of the fabricators with the use of an old if unique bubble level, which
connected the curved edge beam back into the digital model, such that accurate full
scale templates and jigs from the on-site workshop could be used to position each
connector and alignment tabs for the steel erection (Fig. 12).
9.5 Lath Fabrication and Assembly
The construction of previous lattice shells, from the Multihalle to the Savill, develop
the shell from a flat mat of laths with nodes at equal spacing. The approach developed
here uses the digital model to determine the hole spacing, which requires a piecewise
approach to the assembly of the shell. For the construction of the Korlandia gridshell,
a custom site-built 3-axis 12 (30 cm) wide by 60 (18.25 m) long CNC router, coined
the “lath-o-matic,” was developed to drill each individual hole accurately from the
digital model and cut the uniquely angled ends of the lath as they terminated at the
edge beam. The machine itself was constructed from parts purchased on-line, using
Mach3 control software and a Gecko G540 controller with stepper motors. The G-
code was developed directly from a custom Grasshopper script to fabricate just over
28,750 LF (8763 m) of lath members (Fig. 13). The ability to cut each lath member
quickly and precisely, meant that a small 3-person team was able to fabricate and
assemble the entire gridshell. Without the need of any complex telescoping scaffold-
ing, the team was able to assemble the 4-layers of the gridshell in about 3 weeks per
layer, using a skytrack boom to lift the laths at their centerpoint, with another person
at each end to pull the lath member down and attach it to the lath connector. After the
first two layers were installed, and therefore the shell was closed, lath members were
loaded from above, and slid down the previous layer. Based on reports from previous
gridshell construction, particularly the Weald and Downland Gridshell (Harris et al.
2003), we anticipated about 10% failure rate for laths. Although our sample size was
small from our initial material testing, the fact 1 member of 9 tested failed to meet
maximum bending desired, we were confident we were within around 10% break-
age. In the end, 24 lath members broke during assembly, which is about 14% of the
total number of laths. New lath members could be quickly fabricated with the on-site
custom CNC, and the broken laths were reused for blocking. With experience, the
Fig. 11 18 cast steel tapered pins, from Cast Connex, resolve the chevron bracing to the foundation
team also began to anticipate breakage in high curvature areas by successive clamp-
ing and strapping to minimize breakage. As the first lattice shell constructed from
LVL engineered lumber in this piecewise approach, the results from the Korlandia
gridshell are very promising for future shells constructed in this method (Figs. 14
and 15).
While each layer of the gridshell was constructed in about 3 weeks, it was the
shear blocking between the layers that required much more time and effort than
anticipated. The shear blocking is used to transmit loads between each lath layer
and the additional blocking between layers 2 and 3 increased the effective depth of
Fig. 12 Final edge beam segment MN being erected to close the boundary curve. Note the tabs at
end of each beam segment to aid in erection and alignment. After each segment of the edge beams
are connected with full penetration welds, the tabs were grinded off
the shell reducing deflection to allowable limits. Each piece of blocking was unique,
parametrically derived from file to fabrication with the on-site CNC and grouped
in labeled blocking packages. While the prefabrication of each blocking package
made the operation go smoothly and precisely, it took about 6–8 weeks for each
layer of shear blocking. This was no doubt the most tedious aspect of the lattice shell
assembly, but with each successive layer the shear blocking began to give interesting
visual effects as well. Based on stresses in the shell, three different block types
were used according to a graphic schedule provided by the engineers. At times, this
created a smooth gradient from solid blocking to more open blocking (Fig. 16), while
in other areas the solid blocking in high stress areas created a blotchy effect. While
in the end extremely minor, in the future incorporating the blocking pattern into the
parametric model is an additional material constraint which can drive both structural
and aesthetic performance.
Fig. 13 Custom 12 × 60 3-axis CNC was built on-site to fabricate over 28,000 lineal feet of lath
Fig. 14 Complete assembly of the LVL lattice shell prior to sheathing

Fig. 15 Detail of complete assembly of the LVL lattice shell
9.6 Scanning and Verification
Through the direct file to fabrication approach, the lath members also indicated areas
in which the edge beam was out of tolerance with the digital model as they did not
precisely fit. This was not discovered until reaching the mid-way point of layers 1 and
2. A surveyor was hired using a traditional transit to verify points on the connector
plate, but due to a miscommunication in prism offsets, it was only discovered after
additional fabrication from this survey that in fact the survey was inaccurate! At this
point, a LiDAR scan was completed using a Faro x330 compiled from 10 scans,
Fig. 16 From specific angles, the three different levels of shear blocking provide a smooth gradient
from closed blocking to more open blocking
coordinating this point cloud with the parametric model in Grasshopper using the
plug-in Volvox. Despite a long delay with a lot of trial and error, the LiDAR scan
revealed that the edge beam was off in some areas as much as ¾ of an inch, but on
average closer to only ¼ . A clear lesson learned here is to LiDAR scan the edge beam
directly after erection, such that the parametric model can be updated if need be prior
to proceeding with lath fabrication. Through this integrated parametric approach, and
with the on-site lath-o-matic CNC, the investment in the LiDAR scanning would have
eliminated construction delays.
9.7 Sheathing
The lattice shell is sheathed in ½ (12 mm) OSX plywood. The original projected
lath spacing was designed such that the sheathing spacing followed the centerlines
of the laths, but in the end it was far more economical to run the sheathing across
the entire shell using h-clips to connect sheathing. This became possible for two
reasons. From the underside of the shell, the h-clips are visible and include a slight
gap between panels which is visually not desireable. However, the spacing between
the laths will be covered with LED edge lit arcylic/wood veneer panels to provide
lighting and a finished appearance to the underside of the shell. From a construction
logic, there are 374 uniquely shaped sheathing panels which follow the curvature of
the roof cut directly from the parametic model on the MutliCam CNC sheet router
(Fig. 17). A custom airbrush attachment was afixed to the router to locate each of
the 16,000 screws as well as label each sheathing panel, and thereby did not need to
align the edges of the sheathing panels with lath members. At this point, a used Kuka
KR2 was added to the toolset of the workshop, and introduced as a palette loader
34
33
32
31
30
29
28
27
26 NESTED W/ 26-2
25
24
23
22
21
20
19
NESTED W/ 19-2
NESTED W/ 19-2
18
17
16
15
14
13
12
11
10
09
08
07
06
05
04
03
02
01
00
Fig. 17 Each of the 374 sheathing panels were fabricated directly from the parametric script includ-
ing a custom airbrush attachment to locate each of the 16,000 screw locations into the LVL laths
covered by the sheathing
Fig. 18 The finished gridshell with sheathing
for the nearly 400 sheets of sheathing loaded onto the multicam CNC router. From
file to fabrication and installation, the sheathing was complete in less than 4 weeks
(Fig. 18).
10 Conclusion
The successful completion of the Korlandia Gridshell provides a full-scale demon-

stration that this piece-wise approach to an engineered lumber LVL lattice shell
works. Along with the focus on this parametric methodology combining form-
finding, material constraints, and construction logic, emphasis should also be placed
on the ingenuity, craft knowledge, and innovation throughout the construction process
from the full-scale jigs and templates in the steel fabrication process, to the develop-
ment of a site-built CNC to fabricate the 60 long laths, and the sheer determination to
see the construction of the lattice shell through the 10 months it took to assemble the
gridshell after steel erection. With a clear span of 50 (15.25 m) × 140 (42.5 m) with
a shell thickness of only 9 (23 cm), the resulting shape combines the play between
formal expression, structural economy, and material constraints resolved elegantly
into the edge beam and its associated lath connections. When combined with the
formal elegance of the Smart Geometry Gridshell, the opportunity for future lattice
shell structures that were simply inconceivable through previous form-finding and
mat construction logics can now be explored.
References
Bezier P (1998) A view of the CAD/CAM development period. IEEE Ann Hist Comput 20(2):37–40
Bechtold M (2008) Innovative surface structures: technologies and applications. Taylor Francis
Billington D (1983) The tower and the bridge: the new art of structural engineering. Princeton
University Press
Burkhardt B (1978) IL 13: Multihalle Mannheim. Institute for Lightweight Structures, Stuttgart
Cabrinha M (2008) Gridshell tectonics: material values digital parameters. silicone + skin. In: Pro-
ceedings of the association for computer-aided design in architecture (ACADIA), Minneapolis,
Michigan, pp 118–125
Cabrinha M (2010) (In)Forming: the affordances of digital fabrication in architectural education.
Dissertation, Rensselaer Polytechnic Institute. Troy, New York
DeLanda M (2004) Material complexity. In: Leach N, Turnbull D, Williams C (eds) Digital tectonics.
Wiley-Academy Press, London
Focillon H (1934/1992) The life of forms in art. Zone Books, New York
Harris R, Romer J, Kelly O, Johson S (2003) Design and construction of the downland gridshell.
Build Res Inf 6(31):427–454
Harris R, Haskins S, Roynon J (2008) The Savill garden gridshell: design and construction. Struct
Eng 86(17):27–34
Herzog T (2000) Expodach: roof structure at the world exhibition, Hanover 2000. Prestel
Johnson S (2006) The architecture ensemble. In: Architectural design, pp 96–99
Nerdinger W (2005) Frei Otto complete works: lightweight construction natural design. Birkhäuser
Basel
Poretti S (2010) Pier Luigic Nervi, an Italian Builder. In: Olmo C, Chiorino C (2010) Pier Luigi
Nervi: architecture as challenge. Silvana Editoriale, Milan
Turnbull D (1993) The ad hoc collective work of building gothic cathedrals with templates, string,
and geometry. Sci Technol Human Values 18(3):315–340
Toward Mass Customized Architecture.
Applying Principles of Mass
Customization While Designing
Site-Specific, Customer-Inclusive
and Bespoke Timber Structures
John Haddal Mork, Marcin Luczkowski, Bendik Manum

and Anders Rønnquist
Abstract Mass customization is established in many industries, but are not yet inte-
grated in architecture and the building industry. This article presents a parametric
timber toolkit under development. A flexible toolkit for parametrically designed tim-
ber structures, a toolkit that simplifies and substantiates a continuous digital workflow
from global shape to digital fabrication and assembly—a toolkit that requires para-
metric thinking, not only parametric modelling skills. The toolkit proposes solutions
to four recurrent workflow-related challenges that limit efficiency and quality while
designing timber structures. A series of built case projects are used to exemplify
and explain the toolkit. An important finding discussed in the end of the article is
that parametric modelling, and partly the toolkit, changes our conception of what is
considered a similar structure.
Keywords Parametric design and fabrication strategies

CNC and woodworking technology · Parametric timber engineering
J. H. Mork (B) · M. Luczkowski · B. Manum · A. Rønnquist

Norwegian University of Science and Technology, Trondheim, Norway
e-mail: John.h.mork@ntnu.no
M. Luczkowski
e-mail: marcin.luczkowski@ntnu.no
B. Manum
e-mail: bendik.manum@ntnu.no
A. Rønnquist
e-mail: anders.ronnquist@ntnu.no

222 J. H. Mork et al.
1 Introduction
Is it possible to integrate the benefits of mass customization (MC) into architec-

ture? How can architects apply the principles of MC while designing site-specific,
customer-inclusive and bespoke timber structures? Mass customization, a term intro-
duced by Stanley Davis in 1987 (Davis 1997), was later redefined by Andreas M.
Kaplan and Michael Haenlein as follows: “Visionary traditional MC in a strategy
that creates value by some form of company-customer interaction at the design stage
of the operations level to create customized products, following a hybrid strategy
combining cost leadership and differentiation” (Kaplan and Haenlein 2006).
A challenge that distinguishes architecture from other professions is the large
design space that has to be offered. Naval architects always make variations of a
ship with a hull, an engine, and a variety of equipment. Designers in the footwear
industry always deal with feet of different sizes and customers with different uses
and styles. In contrast, architectural design projects range from working on a bridge
to a high-rise to small furniture. Johanna Daaboul stated that MC can be offered
either via product variability or process variability (Daaboul et al. 2011). Offering
both, a wide range of process variability and product variability, is a core service of
being an architect and makes mass customization especially complex. First, every
project site is unique and includes the functional requirements, locally available
materials, climate and building culture of a given location to be unique. Second, the
user and the user’s needs and functional demands are unique. Adapting projects to
their contexts is what makes architecture, architecture. Furthermore, architecture is
a highly subjective task, partly a poetic process from a vision to a built structure, and
mass customization must not remove such an important quality.
The use of Computer Numerical Controlled (Rapp and Johnson 1980) (CNC)
Machines has become widespread in the building industry (BI). CNC-machines are
used for a large range of manufacturing operations, such as cutting, drilling and
milling, and are digitally controlled. One of the biggest benefits of using such tools
is that the production process is automated and uses the same amount of time man-
ufacturing a set of unique components, such as a set of identical components. How-
ever, automated manufacturing of unique building components does not implicate
automated, manufacturing ready computer drawings. In contrast, unique building
components require unique drawings and demands instead of an automated strategy
to be economically sustainable.
Computer-Aided Design (Burns 1986) (CAD) has existed since the 80s, but in
reality, it is more or less a digital, augmented variant of manual drawing. Later,
Building Information Modeling (Kymmell 2007) (BIM) was introduced. From dumb
geometry, 3D geometry suddenly represented a building component—a building
component aware of where it was, what function it had, optional data about cost,
manufacturer, etc. According to Jared Banks, BIM enabled designers to spend more
time designing and less time documenting and coordinating (Banks 2016).
Toward Mass Customized Architecture. Applying Principles … 223
Digital parametric workflows, which widely appeared a decade ago, represent a

dramatic new way of applying digital power while designing our built environment.
Instead of drawing geometry, parametric workflows let the user define the design as
a series of decisions, systems and relationships (Krymsky 2015). Hence, the digital
design process has the potential of becoming more flexible than existing, partly
standardized, BIM processes. Robert Aish described it as the computation era and
argued that the objective of design computation is to overcome many of the limitations
of BIM (Aish 2013). We see a computational approach and parametric thinking as
the next, decisive step to reach mass-customized architecture and BI.
Parametricism (Schumacher 2008) is a buzzword that has been around for some
years and is often associated with organic, prestigious buildings made by architects
such as Zaha Hadid or Frank Gehry. However, parametric workflows have greater
potential than creating signature architecture and are thoroughly described in Archi-
tectural Design’s special edition, “Parametricism 2.0”. Patrik Schumacher claims
following: “In order to reverse the current marginalisation of Parametricism, it is nec-
essary to relaunch it in a self-critical redirection as Parametricism 2.0. Parametricism
is architecture’s answer to contemporary, computationally empowered civilisation,
and is the only architectural style that can take full advantage of the computational
revolution that now drives all domains of society” (Schumacher 2016).
Great examples of such methodology can be seen in the works of Shigeru Ban
and Achim Menges. Shigeru Ban’s famous Nine Bridges Country Club grid shell is
realized with the help of Design To Production (consultant company) parameteriza-
tion and automated manufacturing (Scheurer 2013). Another outstanding example
is Achim Menges, who applies parametric tools throughout the entire process while
making his pavilions and structures. Nevertheless, the mentioned projects are more
customized than those that are mass customized. Large budgets and partly research-
funded projects make such projects happen.
What is mass customization in architecture? Parametric workflows are extremely
efficient when established, but establishing such workflows in each project may not
be economically justifiable. Hence, the goal must be to create a continuous paramet-
ric workflow from sketch to fabrication that is applicable to a wide range of projects.
Through designing and building a series of timber structures according to an estab-
lished parametric approach, we identified four recurrent workflow-related challenges
that limit efficiency and quality while designing timber structures. These are also the
challenges that often cancel a continuous workflow from sketch to fabrication and
are the motivation to create the parametric timber toolkit:
(1) Parametric Detailing: Finding a robust way to sort and identify geometric data
that are time-consuming in each individual project. Among other reasons, data
is sorted to be able to design the different details in a structure.
(2) Tectonic Architecture: Processing timber components that are subtractive,
meaning that the material is subtracted from a stock. In contrast, digital design is
largely additive, e.g., extrusions, revolves, sweeps, etc. Thus, a recurrent design
challenge focuses on how a component is physically manufactured.
(3) Manufacturing Descriptions: Concerning timber building design, architects

and structural engineers tend to output geometry that the manufacturer has to
redraw. Geometry that does not include information about the manufacturing
process. The result is redundant work and increases the chances of modeling
errors.
(4) Structural Analysis: Since the timber is orthotropic material and the most
demanding phase in structural analysis is the connection design, the faster the
structural analysis is implemented, the greater the chance of achieving a feasible
project. Similarly, in manufacturing conversion, a challenge is to automatically
convert geometry modeled by an architect to structural analysis.
“First build your tool” is the title of the abovementioned Aish’s article (Aish 2013).
The objective of the ongoing research project has been to develop a flexible toolkit for
parametrically designed timber structures, a toolkit that simplifies and substantiates
a continuous digital workflow from global shape to digital fabrication and assembly.
A toolkit that requires parametric thinking, not only parametric modeling skills.
The toolkit is based upon the process of designing timber structures, but many
of the principles are general and apply to other materials. This article thoroughly
describes the system and principles of using elements and nodes as a basis for most
types of parametrically designed architectural structures. It describes how archi-
tects, engineers and manufacturers considers an element, and a shared solution is
proposed. Furthermore, the article describes the algorithm that sorts details into
detailing groups. A series of built case projects are used to exemplify and explain
the toolkit. An important finding discussed in the end of the article is that parametric
modeling, and partly the toolkit, changes our conception of what is considered a
similar structure. A structure can be similar in parametric description but can vary
much in function, form, scale and all other attributes. What appears as visually and
radically different structures from an architectural point of view is very similar from
an algorithmic point of view.
2 The Computational Workflow
2.1 Software Platform
The toolkit is implemented in Grasshopper 3D (McNeel 2015), an add on to

Rhinoceros 3D (McNeel 2009). The visual programming software was introduced
in late 2007 and accelerated the architectural software revolution (Krymsky 2015).
Currently, the toolkit is an in-house beta-version, but is planned to be released as
an open-source plugin early 2019. Using C# and Visual Studio, the toolkit is now
developed as a proper plugin. Earlier development, as described in the case-studies,
have been a combination of Grasshopper and Iron Python scripting. The current
toolkit is based on C# classes, which corresponds to the detailing groups described in
Fig. 1 Rhinoceros to the left, BTL-viewer to the right
Fig. 2 The grasshopper script required to create the geometry illustrated in Fig. 1
Chap. 4.1. The export from Rhinoceros is based on Building Transfer Language
(BTL), and will be described in Chap. 4.4
Figures 1 and 2 shows a simple example of the concept described in this paper.
A few bars are connected in a node. If the bars are shorter than 925 mm, they are
trimmed with an angle. The left side of Fig. 1 shows the result in Rhinoceros. The
right side of the figure shows the manufacturing ready BTL-export. Figure 2 shows
the script in Grasshopper.
2.2 Workflow
To use parametric modelling as an efficient method, it is crucial to ensure a continuous

flow from overall geometry, via detailing and structural analysis to manufacturing
output. Furthermore, it is crucial to sustain a workflow that suits both architects,
engineers and manufacturers. This chapter explains briefly the chosen computational
workflow. The following chapters explains in-depth how the introduced challenges
are solved and implemented.
The components and workflow are described in Fig. 3, and the procedure is as
follows:
Overall geometry
(1) Centerline-geometry is generated with the help of Rhinoceros or conventional

Grasshopper-components.
(2) The centerline-geometry is fed into one or multiple element components. Here,
the centerline is materialized.
(3) In the element components, the cross-section, cross-section orientation and
material are defined.
(4) The property description tools describe each detailing group. Here, the node-
properties are also assigned.
(5) The loads are defined.
(6) The loads, the elements and the descriptions of the detailing groups are attached
to the assembler. The assembler generates relations between the nodes and
elements, generates the details and assigns the details to its detailing groups.
This step relates to challenge 1 and are thoroughly described in Chap. 4.
Structural analysis
(7) Finite element analysis is performed using Karamba (Preisinger and Heim-
rath 2014). However, the pre-processing and post-processing of the analysis is
included in the toolkit. To integrate the structural analysis and the architectural
design, the structural FE objects are being made simultaneously with definition
of the geometrical elements and its components (material, cross section). More
over the results from FE analysis are automatically send for post-processing. The
components are checking the elements and joints according to the EC5 criteria
and are informing user (designer) about utilization ratio and which combination
of failure state (compression, tension, bending, combinations) is crucial.
Detailing
(8) The respective detailing groups are extracted by a component. Here, the user
can extract each member of the detail.
(9) Properties from the members of the details are further extracted.
(10) Subtractive tools detail the timber structure. This step relates to challenge 2
and 3. The Subtractive tools are described in Chap. 5.
Output
(11) The outputs from the subtractive operations can be used for visual inspection,
FEA or BTL-export.
Fig. 3 The computational workflow
3 Establishment of a Shared Parametric Approach

for Architects, Structural Engineers and Manufacturers
An established method of modeling digital structures is to start with the center-line

geometry, modeling the theoretical center-line of any building component. Columns,
beams, chords and bars are described by two points—the start and end points that
construct a line. If two lines share the same points, the lines are connected neighbors.
The shared point will then be a connection node. With only points, lines and geometric
properties such as cross section and materials, a majority of the different kinds of
structure can be described in architecture or engineering.
As in many other applications, the parametric timber toolkit is based on a system of
curves and points. Curves can represent any building component with one dominant
geometrical direction. In this article, these components will be described as elements.
In buildings, elements are mostly attached by a physical connector, but digitally, they
can be described as single points where elements start, end or intersect (hereby called
nodes). See Fig. 4.
As a general description of a topology, the overall shape of the structure, the
description of using curves and points, works very well. However, when architects,
engineers and manufacturers are going into detailed design, analysis and manufacture
of the structure, the described system is not optimal. The following is a generalized
description of how geometry generation methods are preferred according to each of
the building-process participant.
Fig. 4 The article and toolkit’s definition of elements and nodes
Fig. 5 How architects, structural engineers and manufactures define an element
3.1 Architect’s Conception of an Element
What you see, is what you get. For an average architect, it is enough to consider
an element as a building component. The physical component ordered from the
manufacturer is what is being considered as one element. The example in Fig. 5
shows two columns and an arched beam. Depending on the detail-level, further
detailing of the node and even the refinement of the elements might be required.
Regardless, the architect can consider the illustrated structure as three elements.
3.2 Structural Engineer’s Conception of an Element
The structural engineers view the objects from two perspectives, physical and numer-
ical. A numerical model always aims to simulate the physical behavior of the structure
or at least it part (here called element). Currently, the numerical models are mostly
built using the finite element method (FEM).
In general, (the physical model) one element is understood by the structural engi-
neer as an object with material continuity that allows continuous stress distribution.
The engineering simplification of the objects to the beam element is very intuitive
and allows operating on force and deformation for estimating the element capacity.
Two rules are essential:
(1) Commonly, the beams are represented by the linear finite element; for repre-
senting curves, we divide it into a sufficient number of smaller finite elements.
(2) The connection/continuity between elements is described by the nodes, and the
finite elements have to begin or end in the nodes.
With the rules in mind, we see that the architectural model on the left side (Fig. 5)
does not fulfil the requirements for a structural analysis. First, the arch must be
segmented into a poly line and fragmented into linear elements. Second, the beam
must be divided where the columns are attached. That is, for a structural engineer,
the structure consists of seven elements.
3.3 Manufacturer’s Conceptions of an Element
The architectural conception of an element is also often sufficient for a manufacturer.

One element is one building component. However, sometimes, an element consists
of multiple sub-elements joined together as one element. There are many reasons
contributing to this phenomenon. For glulam-manufacturers, the reasons to split the
element can be due to size limitations or being able to mill a complex detail.
For such purposes, the element must be divided into multiple elements, but not
likely in the same manner as for the structural engineer.
If the structure in the example were detailed and slotted in plates and dowels as
nodes, the manufacturer would likely prefer to construct each element from at least
three sub-elements. In this way, space for the plates could be easily milled, thereby
being glued together as one element. The result is seen in Fig. 5.
3.4 A Solution that Integrates the Three Conceptions
How is this problem solved? How does one make a model that satisfy these three
approaches? Owing to object-oriented programming (Goldberg and Robson 1983),
a class-structure is developed to contain the three ways of describing an element.
The master element is similar to the physical component delivered to the building
site. However, a subclass is also storing one or more structural elements within
the main element. Similarly, one or more manufacturing sub element is stored in
the main element. The system is illustrated in Fig. 6. Note that the structural sub
element is subdivided transversal while the manufacturing sub element is subdivided
longitudinal.
Fig. 6 Object-oriented programming gathers architects, structural engineers and manufacturers’

conceptions of an element in one master element
In addition, geometric representations of the element and a series of other data

are generated and stored in the element object. First, geometric data such as length,
height, width and local planes and vectors are stored. Second, cross-sectional data,
material properties, manufacturing data and other metadata are stored in the ele-
ment object, making it trivial to extract relevant data when detailing or analyzing a
structure.
4 Rethinking Parametric Detailing: Introducing

Property-Based Detailing Groups
The beauty of parametric modeling is the ability to create a limited amount of para-
metric details that become valid for numerous instances in a given structure. The
challenge is to create a robust sorting algorithm that updates any geometry correctly.
The first level of complexity is to make the sorting work for a given topology; a more
complex sorting is to make it work for any kind of topology, allowing the topology
of a structure to change fundamentally, but the details are updated and distributed
correctly. Through the research project, a general and flexible sorting algorithm for
parametric modeling has been developed. The method is surprisingly simple.
However, to describe the concept properly, let us start from the beginning. What
is a detail? When detailing a node, one is also detailing the ends of the connected
elements. Hence, a detail of the node influences the node, and due to, for example,
holes in the element, the detail changes the element size. Mutually, when detailing
an element, the connected nodes may be affected. Thus, the system defines a detail
as a coupled system of elements and nodes, that is, either a node and its elements or
an element and its nodes. The principle is shown in Fig. 7.
Every type of structure, both parametrically and conventionally designed, contains
a set of details. How the column meets the ground, how the members in a grid shell
are connected, and how the top-chord is connected to the two bars are examples
of details. However, there is a big difference between parametric and conventional
detailing:
Parametric detailing principle
One parametric recipe for a detail generates specific geometry for all instances of the
details in the structure. A good parametric recipe of a detail is able to handle a wide
range of variations. The more general the recipe is, the bigger algorithm has to be.
One model—several possible detail instances
Fig. 7 Node-focused details and element-focused details. A node has elements and an element has
nodes
Conventional detailing principle

If there are no variations, then one detail applies to all. However, if there are variations
within the same detailing principle, one either has to draw several variations or
describe only one in detail. The rest depends on the builders to replicate. Since each
detail needs a new model in this approach, the time of design and the probability of
making a mistake by the designer increase.
One model—one detail instance.
4.1 Detailing Groups
How a detail is crafted is decided by a series of considerations. Architectural, struc-

tural and manufacturing considerations are taken into account, but all of the other
disciplines included in a building project influences a design. Are the details of the
structure load-bearing? Is the detail geometrically planar? Is the connection hinged
or fixed? Will the connectors be visible? Do the details need fire-protection? What
Fig. 8 “Guess who?” Property descriptions filter the different groups

Fig. 9 Similar to “Guess who”, the detailing groups can be described by property descriptions
class of materials will be applied? How will the components be assembled on the
construction site? The possible considerations are almost endless.
Figure 9 shows two structures. An experienced human can intuitively point out
what is similarly detailed. The challenging part is to precisely describe the logics in
a language that a computer understands. The method used in proposed algorithm is
surprisingly similar to the children’s game called “Guess who?” The purpose of the
game is to use as few possible descriptive questions to ask of whom the competitor
is thinking. It is not allowed to ask topology-based questions, such as “is it a person
in the bottom-left corner?” See Fig. 8.
A similar approach is suitable to identify details in a structure. Figure 9 shows
two structures. By using property descriptions, the same descriptions can be used on
two different structures. In Fig. 9, all details that contains two elements becomes a
foundation node.
The key advantage of the described sorting principle is not to use descriptions
relative to the topology and space. Rather, descriptions of local parameters, only
influencing each individual detail, are used to identify the details. Hence, a change
in the topology does not disturb the sorting system.
Fig. 10 Diagram of the sorting algorithm. The algorithm checks each detail and determines if it
fits one or several detailing groups
The sorting principle is implemented in the toolkit by a series of components

regarding property description. Detailing groups are defined by telling whether a
property description is true or whether a property description is false. When the
detailing groups are defined, each detail is analyzed to determine of which detailing
groups they belong. A diagram of the system is described in Fig. 10. Table 1 presents
a list of property descriptions that have been or will be implemented in the toolkit.
The result from the described sorting system is that one may end up with details
present in either zero or multiple detailing groups. If that is not the intention of
the specific project, then the detailing groups must be further described. However,
there are cases in which details connected to zero or multiple detailing groups are
relevant. First, being connected to zero detailing groups implies that no detailing is
required, the stock does not need any refinement. Second, there might be sub-details
that overlap in some details. An example is the chord connections of the Follo Bridge
shown in Fig. 11. The upper chord and lower chord are primarily similar, but the
bottom chord node detail includes a suspended connection to the secondary structure.
Hence, the connections in the top chord only belong to the Chord Node Detail, while
the connection in bottom chord also belongs to the Suspension Detail.
Table 1 Property descriptions that have been or will be implemented in the toolkit
Name Description Option
Element length Defining Min/max length Valid for one or all
Element angle Max/min angle between elements Valid for one or all
Element Defining min/max parameter where Valid for one or all
parameter the node is connected to the element
Element amount Min/max amount of elements
Node position Checks if node is inside defined
bounding box
Element names Checks if detailing group contains Including at least or exactly defined
defined names elements
Element Defines min/max angle deviation Valid for one or all
direction from defined vector
Normal force Defines min/max normal force value Valid for one or all
Moment force Defines min/max moment Force value
Fig. 11 The bottom chord detail belongs to the two detailing groups
4.2 Node Properties
The properties of the elements can be created simultaneously by defining the ele-
ments. Assigning properties to the nodes is a slightly more complicated process
because the node geometry is a consequence of the intersection of the elements. The
features of the node can be changed, added or limited according to changes in the
global geometry. The system assumes that all nodes within one node-based detail-
ing group have shared properties. Hence, the desired properties are connected to the
property description component.
Generating a sufficient node plane is crucial for efficient parametric detailing. As
previously explained, the node is generated from the end points or the element inter-
sections and is by default just a point with no normal direction. However, when detail-
ing a spatial object around this point, the orientation of a local coordinate-system, or
a local node plane, must be clearly identified relative to the global coordinate-system
of the overall structure. How a node plane is oriented is highly dependent on the
type of structure and how the node is detailed, but this orientation is valid for all
parametric detailing procedures; a consistent and logical node plane generation is
crucial for efficient detailing.
What is a consistent and logic generation of a node plane? The concept can be
well-explained in a truss bridge example. By default, a node plane is an XY-plane
(Fig. 12a) but is most likely not suitable for the nodes in a bridge. If the bridge is
straight and parallel to a global axis, an XZ-plane or a YZ-plane is usable, but it is
not flexible. A more consistent rule would be to say that the plane is planar to the
truss plane (Fig. 12b). Then, the structure can be oriented in any direction. However,
the logics of the planes are still not consistent. The normal direction may be flipped,
and the X-alignment may not be defined. By stating that the normal direction of
the plane always faces from or to the secondary structure and the X-axis is parallel
to the bottom chord, a fully consistent and logical plane-generation is established
(Fig. 12c).
These rules apply well for a truss bridge, but the logic does not necessarily apply
for all other structures. For example, a grid shell structure is easier to detail if the
normal axis of the node plane is parallel to the normal surface. Thus, the toolkit
should supply various node plane alignment components.
5 Digital Subtractive Tools
There are many purposes of creating a 3D CAD-model. The way a model is made
influences the model output possibilities. A sketched volume model is relevant to
understand the space to be created but most likely is not adequate for fabrication
purposes.
Most of the big manufacturers in the timber industry use CNC tools, such as Hun-
degger Speedcutter (Hundegger 2001). These tools have two shared characteristics:
they are controlled by numerical data, and they subtract material. These features
must be taken into account when designing timber structures digitally. Since man-
ufacturing timber is highly specialized and tool-dependent, a design must often be
redrawn by the manufacturer. Redundant work is done, but it also increases the
chances of human modeling errors. To solve this challenge, the modeling purposes
and geometrical outputs of the architects, structural engineers and manufacturers
Fig. 12 Different strategies of node-plane generation. Figure 9c shows a consistent and logical
plane-generation
were investigated. The timber toolkit aims to create an effective solution that fits
all three professions. The following is a brief description of these three professions’
objective for 3D-detailing timber structures. The description focuses on the part that
influences all three professions, namely, the load-bearing system.
5.1 Architect
While some architects leave the load-bearing system to structural engineers, other
architects base the architectural expression on how a structure is made. They use
architecture to communicate how the forces in a building work and how a building
has been crafted. To design such architectural expressions, it is suitable to make
a detailed 3D-model, including the bolts and brackets. Indeed, the architect does
not dictate the dimensioning but can hold a great influence on how the structural
load-bearing system and detailing are composed. Thus, the architectural geometrical
output is primarily visual. Secondary output such as cost, manufacturer, and volume
may be generated.
5.2 Structural Engineer
The timber structures due to natural imperfections in the material (orthotropic, not
linear grain angle, rods.) are very sensitive to imprecise detailing. Even the best calcu-
lations and perfectly chosen static schemes can be ruined because of bad explanations
of the joint fabrication/production (the detail).
The issue that is characteristic of timber engineering is finding the connection
stiffness. Most of the connections are designed to be rotation-free hinges. Finding
the real characteristic of the joint behavior is left either to experimental studies or
sophisticated experiments.
Two computational approaches are described—one which is simple and fast, and
one which is advanced and more CPU-demanding. The simple detail analysis is made
by applying analytical equations (e.g., from Eurocode 5). These equations can check
a whole structure in real time. The advanced analysis is made based on sub-modeling.
Critical details from global analysis are sent to local analysis. Finite element analysis
is performed with volume-objects and more precise material descriptions, and it
includes eventual steel connectors such as dowels, bolts and screws.
At first glance, the geometry needed for advanced analysis is very similar to
an architectural model. However, the major difference is that how the mesh of the
analysis model is built up is critical to make the analysis work.
5.3 Manufacturer
In regard to digital fabrication, the manufacturer’s output is preferably an assembly-

ready building component, but the input is a primitive stock. Thus, one of the roles of a
high-tech manufacturer is to create instructions for a CNC machine. Two-dimensional
cuts performed by lasers or mills are possible to import from a standard CAD-
model, but more complex operations have to be defined as virtual timber processing
operations.
5.4 Finding a Shared Detailing Solution
The three described methods and geometric outputs look surprisingly similar, but
the difference is how the geometry is being made. There are some existing tools
that are bridging the gap between the disciplines, but assuming a user-controlled
parametric workflow, the options are limited. Cad Work and HSB CAD are tools
that are tailored for timber structures and connect architecture and manufacturing.
Further, Woodpecker, an innovative tool developed by Lignocam and Design to
production, enables the export from Grasshoper 3D to any CAM-software. This
export occurs through an open-source format called Building Transfer Language.
However, Woodpecker is not a design tool; rather, it is more of a geometry conversion
tool (Stehling et al. 2017).
The solution developed in this toolkit aims to bridge the gap between all mentioned
disciplines while modelling parametrically. The concept is to allow the user to refine
timber components exclusively based on subtracting material from a digital stock.
Fig. 13 Different outputs from one subtractive operation
Table 2 The table shows which elements of the toolkit have been used to design and build the
structures
Project Script language Detailing Subtractive FEA BTL-export
groups tools
Orkla Bridges Iron Python × × × ×
Water Ramp Iron Python × × ×
Log House GH-components ×
Book-shelf GH-components ×
Fig. 14 Detailing groups in Follo Bridge. The same rules were applicable while designing Evjen
Bridge
Fig. 15 Subtractive operations to achieve wanted detail
Thus, the designer is forced into thinking about physical processes while designing
virtually. When the architect applies a subtractive operation to a stock, three outputs
are automatically generated. A 3D preview geometry, an FEA-ready mesh and a
BTL-description. (In the current version of the toolkit, the FEA mesh is not yet
implemented). The concept is illustrated in Fig. 13.
6 Case Studies
A series of case projects have been built to test the capability and flexibility of the
toolkit. The following sub-chapters shows how the toolkit was used to parametrically
design two timber bridges, a freestyle water ramp, a log house and a shelf. The
structures are visually and fundamentally different but surprisingly similar when
applying the ideas of elements and nodes. The toolkit have been developed while
designing the case-studies. Mock-ups of the toolkit have been developed in both
Grasshopper-scripts, Iron-python and C#. Table 2 shows what elements of the toolkit
have been used designing the structures (Figs. 14, 15 and 16).
6.1 Orkla Bridges
The bridges are designed for a pathway in a park and cross two small rivers. The
distance between the bridges are approximately 1 km; hence, it was natural to create
two bridges that had a similar architectural expression. However, the boundary con-
ditions at the sites were different. The first bridge, Follo Bridge (shown in Fig. 17)
had a span of 10 m and on height-differences but had to be arched to allow the poten-
tial ice drift through. The second bridge, Evjen Bridge (shown in Fig. 16) had a span
of 15 m and had a relatively large height difference. Furthermore, the bridge needed
Fig. 16 Evjen Bridge. Photo Arnfinn Sæthre

Fig. 17 Follo Bridge. Photo Arnfinn Sæthre
Fig. 18 The sauna (mostly L-nodes) on the left side and the water ramp with secondary connection
and primary extension nodes
curved platforms on both sides to connect to the path and fulfill the requirements of
a maximum 1:20 rise.
The different boundary conditions gave a nice opportunity to build a paramet-
ric model with a large enough design-space to be adaptable for both bridges. The
architectural expression is based on a well-established system of dowels and slotted-
in-plates as a connection system (Mork and Luczkowski 2017). Due to relatively
heavy loads, a secondary steel structure had to be used, but the rest of the structure is
glulam-based. While the smallest bridge has a classic arched shape, the larger bridge
has a more organic appearance with a doubled curved railing.
Using the principle of detailing groups as outlined, only four different detailing
groups had to be developed in the small bridge: the bottom chord, the top chord, the
foundation points and the zero-force node. Figure 14 shows the rules applied.
The bottom chord node is the most complex detailing group. The structural bars
are connected to the bottom chord, and the secondary structure is suspended from
the metal-plates. Due to the scale of the bridge, the suspension connections were
integrated inside the chord, making a more compact detail. As previously shown,
the basis for the detailing is stocks that are not processed. Subtractive tools are used
to design the timber elements. In addition, steel components were designed using
conventional parametric tools in the grasshopper environment. In the bottom chord
detailing group, these processes were performed (Fig. 15).
• Cutting the bars parallel and with an offset half the height + 30 mm from the
tangent direction of the bottom chord center curve.
• A grid of holes for the dowels. The amount of dowels was calculated based on the
force in the bar/chord, and the size of the grid was calculated based on the required
edge distances.
• A pocket milling to make space for the metal-plates. Both the plate that connected
the bars to the chord and the plate for the suspended secondary structure.
6.2 Freestyle Water Ramp
A freestyle water ramp is a ski jump that ends in a lake and is used for summer training.
This structure was a simpler design than those of the two bridges. Due to low loads,
a beam-column structure was feasible. However, the principle of using dowels and
slotted-in metal plates applied in this structure was used as well. Surprisingly, the
water ramp had more details to develop than those of a bridge. The detailing groups
can be seen in Figs. 19, and 18 shows the ramp under construction.
6.3 Log House Sauna
Until now, the examples have shown rules based on element names. However, there
are many cases in which element names cannot be used. Often, the elements are
nameless or a set of similar elements. In addition, previous examples have described
spacious structures. Is it possible for the principle to apply a more homogeneous
structure, containing only one type of element?
An algorithm was developed to be able to model the wall’s surfaces. The algorithm
then splits the structure into layers with a center-line geometry. What is challenging is
Fig. 19 Relatively high amount of detailing groups. Note how the yellow detailing group can be
used to trim building components
to generate the actual length of each individual element. Some elements are extended,
and some elements are trimmed to become an overlapping system.
Looking at the system, there are four kinds of connection types, i.e., four types of
detailing groups—the X-node, the T-node, the L-node and the I-node. These are easy
for a human to determine, but what rules can be applied in the toolkit to distinguish
the nodes? An X-detail contains only two laths, but that is the case for the L and
T-detail, as well. In this case, information regarding where the node is relative to
the lath was used to distinguish the details. Descriptions of the detailing groups are
shown in Fig. 20 and the final result is shown in Fig. 18.
Fig. 20 Detailing groups applying the positioning of the node relative to the element to distinguish
the details
With such rules, all the different types are separated, but there is one more condi-
tion needed to generate a log house, namely, the shift in which element is extended.
Thus, one more condition to all the detailing groups is if the group is on an even or
odd layer. If the T-node is on an odd layer, the element connected to the end will be
extended, and the opposite will happen if the node is on an even layer. This principle
is shown in the bottom-right of Fig. 20.
Fig. 21 Same property descriptions as the log-house, but different design of the details
6.4 Plywood Shelf
Finally, the principle has been tested on furniture: a plywood shelf prefabricated using
a 3-axis CNC-mill. Both visually and in function, a shelf is something extremely
different from a truss, a ski jump and a log-house; there are different scales, different
appearances, and different functions. However, algorithmically, a shelf is identical
to one layer of the log-house principle. The same type of detailing group applies to
the X, T, L and I-node. Figure 21 shows the nodes and Fig. 22 shows the detailing
of the X-node.
7 Discussion and Conclusion
This article presented a flexible approach designing structures parametrically. The

aim was to investigate how to apply the benefits of MC while designing site-specific,
customer-inclusive and bespoke timber structures. The main finding is that when
abstracting any structure to elements and nodes, most of the structures can be designed
through a similar set of rules.
Architects are trained to design unique projects. Reusing elements from a com-
pleted project almost feels like cheating. Such mindset stands in contrast to the idea of
mass customization. However, learning from the described study and case-projects,
one conclusion is particularly prominent: To be able to achieve Mass Customiza-
tion in architecture, we must standardize projects. Thanks to parametric modelling,
we do not have to standardize the physical result. Rather, we can standardize the
Fig. 22 Detailing the X-node. The parts are slotted to simplify assembly
workflow, the process and the interfaces between the project stakeholders. This paper
has exemplified a standardized, still flexible, building system in the age of parametric
architecture.
A structure can be similar in a parametric description, but can vary much in
function, form, scale and all other attributes. What appears as visually and radically
different structures from an architectural point of view is very similar from an algo-
rithmic point of view. The log-house and the shelf exemplify this issue very well.
In contrast, two structures that are functionally similar are fundamentally different
from an algorithmic perspective. With such a view, a truss bridge is more similar to
a free-form grid shell than a beam bridge. A tendency in described case-projects, is
that one parametric workflow outputs a much larger design space than in ordinary,
manually drawn CAD-projects. By parametric thinking, we are able to shift from
standardized products to standardized, still scalable, parametric building systems.
The toolkit made in fully parametric design environment allows the designer
(architect or engineer) to observe and control the design in a holistic way. From
digital conceptual draft to the production codes, every important decision can be
programmed, followed in real time and changed according to the need of hour.
The toolkit does not introduce such methodology, but simplifies and substantiates a
continuous digital workflow from global shape to digital fabrication and assembly.
This by offering a series of tools that reduces technical complexity of parametri-
cally designing timber structures. Especially the detailing groups based on property
descriptions and the digital subtractive tools have drastically reduced the time needed
to set up a parametric model of a project.
According to Aish [9] in computation design: “The designer is no longer directly
modelling the building: instead he develops a graph or script whose execution
generates the model”. The timber toolkit presented here allows the designer to go out
from strict BIM design, where everything have to be described “a priori”, to described
by Aish definition called computation design. The variety of presented cases done
with one approach and one toolkit is a conclusive evidence of applicability of it in
building design and industry.
One can definitely state that the Orkla Bridge project fulfils the idea of mass
customization. The first bridge, Follo Bridge were extremely time-demanding, but
designing the Evjen Bridge was much more efficient. Updating the shape of the bridge
required only 60–90 s, including a simple FEA, regenerating all details and not the
least outputting BTL-instruction. However, these bridges were relatively similar.
How about the other structures? The already-designed bridges contributed to a more
efficient detailing of the water ramp. The slotted-in metal plates for the detailing
algorithm were reused when modeling the foundations for the columns. They are
different shapes but based on the same knowledge-based engineering (KBE)-rules.
The flexibility of a building system is often dependent of the hours spent develop-
ing the system. A flexible and highly reusable parametric building system is lightly
not to pay off if thinking one project a head. However, if developing the building sys-
tem is considered as an investment in a series of future projects, we strongly believe
that such mindset is economically sustainable. Further, to achieve Mass customiza-
tion in architecture, we must continue shifting our mindset to better understand the
possibilities of applying parametric tools and computation.
At last, does the timber toolkit ensure site-specific, customer-inclusive and
bespoke timber structures? The author’s answer is, luckily, that the toolkit does
not make great architecture. At best, the toolkit reduces the amount of work spent on
redundant work. However, the hours earned can be spent making better architecture.
Architecture is still made by architects; tools are just tools simplifying the creative
process.
Acknowledgements This research was partially supported by Innovation Norway. We thank our
collaborators from Nikken Sekkei Digital Design Lab who provided insights and expertise of devel-
oping the toolkit’s concept and technical solutions. Further, we thank Moelven Limtre who provided
important manufacturing knowledge.
References
Aish R (2013) First build your tools. Inside smart geometry: expanding the architectural possibilities
of computational design. pp 36–49
Banks J (2016) Why BIM is still bankrupting your firm. Shoegnome 9:12
Burns WE (1986) Computer-Aided Design (CAD). Ind Educ 75(6):10–11
Daaboul J, Cunha CD, Bernard A, Laroche F (2011) Design for mass customization: product variety
versus process variety. CIRP Ann 60(1):169–174. https://doi.org/10.1016/j.cirp.2011.03.093
Davis SM (1997) Future perfect. Basic Books, New York
Goldberg A, Robson D (1983) Smalltalk- 80: the language and its implementation. Addison-Wesley
Longman Publishing Co., Inc., Boston
Hundegger H (2001) Woodworking machine
Kaplan AM, Haenlein M (2006) Toward a parsimonious definition of traditional and electronic
mass customization. J Prod Innov Manag 23(2):168–182
Krymsky, Y. (2015, December 11) The Architecture Software Revolution: From One Size Fits All
to DIY. Retrieved January 18, 2017, from http://www.archdaily.com/778619/the-architecture-
software-revolution-from-one-size-fits-all-to-diy
Kymmell W (2007) Building information modeling: planning and managing construction projects
with 4D CAD and simulations (McGraw-Hill Construction Series): planning and managing con-
struction projects with 4D CAD and simulations. McGraw Hill Professional, New York
McNeel R (2009) Rhinoceros 3D. Retrieved January 15
McNeel R (2015) Grasshopper 3D
Mork JH, Luczkowski M (2017) One algorithm, two timber bridges built in Orkdal, Norway. In
Forum Wood Building/Nordic Trondheim 17. Faculty of Architecture and Design, Trondheim
Preisinger C, Heimrath M (2014) Karamba—A Toolkit for Parametric Structural
Design. Structural Engineering International 24(2):217–221. https://doi.org/10.2749/
101686614X13830790993483
Rapp EJ, Johnson GA (1980) Computer numerical control machine tool. Google Patents. Retrieved
from https://www.google.com/patents/US4199814
Scheurer F (2013) Digital craftsmanship: from thinking to modeling to building. Digital Workflows
in Architecture. Birkhäuser, Basel, pp 110–129
Schumacher P (2008) Parametricism as style-Parametricist manifesto. 11th Architecture Biennale,
Venice. pp 17–20
Schumacher P, (2016) Parametricism 2.0: gearing up to impact the global built environment. Archi-
tectural Des 86(2):10. https://doi.org/10.1002/ad.2018
Stehling H, Scheurer F, Roulier J (2017) Bridging the gap from cad to cam: concepts, caveats
and a new grasshopper plug-in. In: Fabricate 2014: negotiating design & making (DGO-Digital
original, pp 52–59). UCL Press. Retrieved from http://www.jstor.org/stable/j.ctt1tp3c5w.10
Reciprocal Implications Between Design
and Construction Process of Timber
Gridshell
Sergio Pone, Daniele Lancia and Davide Rando
Abstract This contribution focuses on the latest acquisitions of the Gridshell.it

research group on active-bent timber gridshells. In particular, the main topic is related
to the development of a new information technology design tool (GFFT 2.0) devel-
oped by Davide Rando for his Master’s thesis in Architecture in January 2018 under
the supervision of Sergio Pone and Daniele Lancia; this design tool is a new version
of the previous one developed in 2012 by Daniele Lancia and Andrea Fiore for their
Master’s Degree thesis called GFFT (Gridshell Form Finding Tool) (Pone et al. in
Beyond the limits of man, Wroclaw, Poland, 2013). We will also describe the two
latest projects to which we have committed ourselves since 2016 up to today. We will
also illustrate how the creation of the new IT tool was based on previous experiences
that have influenced subsequent projects (Pone 2017).

Active bending in timber structures · Complex timber structures
Wood complex surface
S. Pone (B) · D. Rando

Department of Architecture (DiArc), University of Naples Federico II, Naples, Italy
e-mail: pone@unina.it
D. Rando
e-mail: randavides@gmail.com
D. Lancia
Gridshell.it, Naples, NA, Italy
e-mail: lanc.daniele@gmail.com

252 S. Pone et al.
1 Introduction
The creation of the GFFT 2.0 tool highlights Gridshell.it’s1 determination to change
the resistant section within the same structure; a decision that starts from the assump-
tion that an active bent timber gridshell can refer to the structural behavior of thin
vaults in reinforced concrete, mainly resistant to compression, but that can also have
mixed behavior in which all compressed zones are joined by parts subjected to bend-
ing and compression (Pone 2012).
This study therefore starts from the intention to map a generic surface, due to the
different levels of stress that it will be subjected to, and as a results propose different
resistances. These resistances will vary according to the inertia of the section and
other possible factors, in order to build gridshells that are no longer homogenous in
all of their parts.
Our latest IT tool, focuses on varying only the inertia of the resistant section and
questions the assumption according to which in the design of wooden gridshells, the
systematic use of funicular forms is the only way possible to interpret this structural
typology.
Of much greater allure is the theory that wants to use the timber gridshell as a
structure capable of building a variety of different forms, regardless of their struc-
tural behavior. These forms will respond well to the stresses of pure compression,
conforming to its essence of “shell”, as well as showing its ability to withstand even
bending stresses, as its “grid” nature suggest.
The in-depth examination and analysis of the theory of the variability of the cross-
sections would make the active bent timber gridshell, structures adaptable to many
interesting forms and would naturally make them compatible with the construction
of free-form buildings due to the following:
– the transformation of its parts by crossed rods into real crossed reticular beams
– localized condensation of the structural mesh2
– the thickening or lightening of the cross-section.
1 The research unit Gridshell.it is coordinated by Sergio Pone and composed by Sofia Colabella,
Bianca Parenti and Daniele Lancia; in the past it was of use to the assistance of Bernardino D’Amico
and Andrea Fiore.
2 Structural thickening has already been used by Edward Cullinan and Buro Happold in the con-
struction of the Weald & Downland Living Museum gridshell in Singleton, Chichester, Sussex built
in 2002.
Reciprocal Implications Between Design and Construction … 253
2 Gridshell Form Finding Tool 2.0
2.1 The Bent Timber Gridshell
The research unit Gridshell.it decided years ago to elaborate a chain of IT tools in
order to design gridshells from any free-form surface; the chain (Fig. 1) starts with
GridMaker (Pone et al. 2016) a tool built with the Rhinoceros® Grasshopper plug-in
and with a part specifically written in the PythonTM programming language.
GridMaker uses computational morphogenesis to identify a constant pitch grid
belonging to a given surface. The morphogenetic part of the tool aims to find the
geometric conditions that allow a minimizing of the curvature of the rods. The Soft-
ware input is a generic surface while the output is a flat-grid that once deformed is
capable of generating the given surface (Toussaint 2007 and Kuijvenhoven 2009).
The flat-grid and the ground connection lines of the surface, which are GridMaker
output, become the input of the Gridshell Form Finding Tool (GFFT).
GFFT uses parametric design softwares, in particular Grasshopper for
Rhinoceros®, and the add-on Kangaroo® (Piker 2013), which is a particle-springs
system simulator that is needed to simulate the forming process of the gridshell
through dynamic relaxation. The tool assigns to some of the nodes located on the
perimeter of the flat-grid a displacement in order to make their position coincide
with the ground connection lines, which belong to the initial surface, and calculate
the consequential deformation of the grid.
This GFFT application gives back a coherent simulation if the forming action
of the gridshell is limited to moving the knots destined to remain on the ground
(Colabella et al. 2015).
At this point it is possible to make a comparison between the Gridmaker input
surface and the gridshell obtained with GFFT. If the two structures are quite similar,
the process stops at this point, if instead they show differences that are too relevant
the two shells become the input for the third software: InFormer (Pone et al. 2017) .
Fig. 1 The tool chain developed by the RU Gridshell.it

254 S. Pone et al.
InFormer, as well as the previous tools, uses Rhino-Grasshopper environment and

the Karamba® Large Deformation module to the finite elements with parts written
in Python by the research unit and the MOGA (multi objective genetic algorithm)
of the software Octopus. The two gridshells will be the starting geometry, but the
former gridshell, which originates with GridMaker, will be the arrival one.
The software works in order to make both gridshells coincide, with the com-
bination of the nodes and with the construction of vectors from the nodes of the
latter gridshell to the corresponding nodes of the former grid. The tool extends the
direction of the vector until it meets the ground plane of the structure and determines
whether the displacement of the node, that we are analyzing, should be carried out by
pushing or pulling, for example if a strut or a tie-rod must be applied or not. Among
these numerous possibilities, InFormer identifies the minimum number of struts and
tie rods that must be applied to make the mesh generated by GFFT to coincide (or
almost) with the desired one, designed by GridMaker.
2.2 The Chain Last Link
The latest addition to the gridshell design tools is the second version of GFFT (GFFT
2.0) born after the recent publication of Kangaroo 2.0, which optimizes performances
of the first version and that is integrated with a completely new module. The new
GFFT 2.0 (Fig. 2) module uses Karamba (Preisinger and Moritz 2014) and the FEA
(Finite Elements Analysis) method to optimize the cross sections of the single rods.3
This module proposes to adapt the shell resistance to the different deformation
conditions that are derived from the structural analysis. The input that the tool uses
is the single-line structure obtained from the application of the tools chain described
Fig. 2 GFFT 2.0 flowchart
3 On the theme of variation of the cross section, of wooden post-formed gridshell laths, the tool
presented in this paper is recognized as a useful simplification to the project carried out by Bernardino
D’Amico (D’Amico et al. 2015a, b).
previously, in which the springs of the main mesh have a rigidity calibrated on the
wood chosen for the simulation, while the diagonal rods are deactivated to allow
simulation of the forming phase; when the structure has reached its final shape, it
is converted into a system with rods and nodes (Fig. 3a, b) to which the operator
attributes a uniform cross section, that is compatible with the minimum curvature
radii detected automatically (Fig. 3c). At this point the Finite Elements Analysis
(Fig. 3d) reads the bending moment in the nodes as well as at the central point of
the micro-modules. If the values are acceptable, which generally happens when the
structure is stressed by simple compression and behaves like a shell in all its parts,
the process stops at this point and goes to the structural verification conducted with
the traditional methods.
If, vice versa, in a mesh area the Finite Elements Analysis reads moment values
that are exceeding the limit (Fig. 3e), the tool automatically applies an algorithm that
calculates the height of the section necessary to bring the values back into the field
of the acceptable one, and automatically modifies the cross section by adjusting its
dimensions (Fig. 3f, g).
This procedure gives rise to a new structure in which the rods of the grid have
a variable section which can then be analyzed again to rectify the heights of the
resistant sections in a more refined manner (Fig. 3h).
In the areas of the structure where the moment is particularly strong, the section
will be considered full, which means that the shear blocks completely saturate the
space between the two layers of structural rods. As mentioned in the introduction,
if the application of the structural algorithm were to propose an excessive height of
the cross section, nothing would prevent treating these mesh parts, as an intersection
of reticular beams oriented in the direction orthogonal to the plane of the surface.
This hypothesis has not yet been fully developed but it does not seem to lack the
requisites of feasibility and a certain charm.
Fig. 3 a The starting geometry, b A load to each node, c Cross section for each lines, d Finite
Element Analysis, e Bending moments values, f New cross sections assigned, g The two compared
analysis, h New bending moments values
256 S. Pone et al.
3 The First Experimental Applications of the Tool
The new tool has been developed and tested in some of Gridshell.it’s latest projects,
carried out for the first time in Italy, whose theme of the construction of active bent
wooden gridshells were built as a part of a public building. This is the case of two
projects connected directly or indirectly to the themes of the design of archaeological
sites: the Coverage of the Entrance to the Archaeological Park of San Vincenzo al
Volturno in Castel San Vincenzo4 (Is) in 2016, and the Coverage of the ruins of the
Neolithic settlement of the Parco dei Campi Diomedei in Foggia5 in 2017.
3.1 The Covering of the Entrance to the Archaeological Park

of San Vincenzo al Volturno
The Gridshell project (Fig. 4) is integrated into a small building to be constructed

with a cross-laminated timber panel that designs the entrance to the park with a ticket
office along with annex staff services, a small room for lessons and an open space,
covered with the gridshell, as a starting point for guided tours of the archaeological
complex. The gridshell has a modular structural step of 600 × 600 mm made of a
double layer of 25 × 60 mm larch rods and is partly supported by the same building
and partly by the edge of a height gap created to access the floor baseline of the
building (Fig. 5). The gridshell provides a structural edge made of single metal tubes
or assembled to form a reticular beam, which lies on the surface plane. This expedient
presents a series of positive aspects among which the ability to better contain the
stresses oriented on the plane of the surface.
These aspects may be seen when coming from exceptional and dissymmetric
load conditions, that may be connected for example to the snow, or to the viscous
deformation phenomena to which wood can meet during the operating phase.
The second technological theme dealt with was the laying of the stiffening ele-
ments, in addition to assigning the task of preventing further unwanted forming
kinematisms, a task traditionally entrusted to a system of diagonal rods. This func-
tion has been accomplished by a double plating of wooden slates that allows the
inner surface to be made with the same material of the structural grid and therefore
4 The project was commissioned by the Archaeological, Fine Arts and Landscape Heritage Depart-
ment of Molise, and its long implementation process of the executive designs has now reached
approval.
The design group, formed by Gridshell.it, was directed by architect Bianca Parenti for the
architectural project and engineer Oreste Mammana for structural analysis.
5 The structure was commissioned by the Municipality of Foggia and it is part of the project for
the construction of the large urban park of the Campi Diomedei. The project is the result of the
collaboration of Gridshell.it with Asdea Srl, coordinated by architect Sergio Pone as well as by
engineer Guido Camata. In addition the executive project was approved and the building site was
opened, however in the process of this publication, the construction of the structure has not yet
begun.
Fig. 4 San Vincenzo gridshell rendering, 2016
Fig. 5 San Vincenzo entrance to the Building, 2016
showing a homogeneous image to the visitors of the Park, similar to what was already
obtained in the Trio Gridshell of Lecce of 2009 (Figs. 6 and 7) (Chilton and Tang
2016).
The construction of this gridshell is inspired to the constructive strategy developed
by Buro Happold for the construction of the Gridshell of the Savill Garden Building
in Windsor Great Park, Berkshire (Harris et al. 2008). This solution was experimented
258 S. Pone et al.
Fig. 6 Trio Gridshell, Lecce 2009
Fig. 7 Trio Gridshell’s wooden slates, Lecce 2009

Fig. 8 Portici, 2012. The

shear blocks are applied on
the grid before its
deformation
by our group in the small gridshell of Portici of 2012 (Figs. 8 and 9), which plans to
form the grid in place only with two layers of rods already equipped with the relative
shear blocks for the laying of the third and fourth layers that takes place after forming
the gridshell.
The choice of this different organization of the construction process allows us to
simplify the forming phase, due to the fact that the grid with only two layers is much
more docile.
In addition the technological problems connected to the translatory movement,
which the nodes need to have during the formation phase, disappear.
Nonetheless, we have decided not to give up on a common feature of the 12
structures built till today: the macro-modules decomposition of the grid and the
systematic renunciation of glued joints. The macro-modules are assembled in the
workshop, they are prepared by creating holes in order to join the modules allowing
them to be close and complete with one of the two sets of shear blocks (assembling
both would make it impossible to pantograph the modules and consequently com-
plicate the transportation of the said modules). The dry assembly using self-tapping
screws and bolts allows the simple substitution of damaged parts (or possibly dam-
aged in the forming phase of the structure). Furthermore, the exclusion of glued
260 S. Pone et al.
Fig. 9 Portici, 2012. In this picture the stiffening elements and the second layer of the grid are
placed once the structure is formed
joints greatly simplifies the waste recycling in the event of a future demolition phase
as well as maximizes the natural eco-compatibility of wood.
3.2 The Coverage of the Remains of the Neolithic Settlement

of the Campi Diomedei Park in Foggia
Foggia’s gridshell will cover an area of more than 600 m2 (Fig. 10) and it will
represent the third largest structure in Europe (after the forefather of Mannheim and
the gridshell of Savill Garden). The shape of the structure is very simple and it is
inspired by the shape of the two large dunes of land located near the archaeological
site to be covered, according to the intentions of the authors of the architectural
project, awarded with the first prize of the architectural project contest.
In contrast to the previous one, the Foggian gridshell rest on the ground completely
but is also edged by a system of beams made of single metal tubes.
As far as the case of the ground attachments are concerned, they are assembled
to form a reticular beam contained on the surface plane for the edges that draw the
large access arches to the covered area.
Fig. 10 Project for the Neolithic settlement of the Campi Diomedei Park in Foggia in 2017
Fig. 11 The introduction of the metal edge, is a solution used for the gridshell of Campi Diomedei
Park in Foggia in 2017
While in the case of S. Vincenzo the metal structure was incorporated in the
slatted roofing system of the gridshell, in this case the trusses and the tubular of the
connections to the ground remain in direct contact with the wooden parts, with a
900 × 900 mm grid composed of larch wood sticks 50 × 80 mm (Fig. 11).
The stiffening system to be executed when the structure is formed, consists of a
skin of OSB panels (Fig. 12), thin but sufficiently rigid on their surface, covered with
a PVC waterproofing sheath with a fabric anchored to the surface with a network of
cables.
The main critical point that emerged due to this technological choice was that of
being in a position to increase the horizontal dimension of the resistant section to
allow the installation of two rows of joints, whether screws or bolts. This necessity
262 S. Pone et al.
Fig. 12 Design drawings for the assembling phases used for the project of the gridshell of Campi
Diomedei, Foggia 2017
has reduced the possibility of bending the structural rods on the surface plane, at
the same time conditioning the geometric-physical freedom of the forming phase
and therefore indirectly influencing the freedom of expression of the architectural
project.
The executive phase of the Foggia project was inspired by the “Savill method”
(Kelly et al. 2001); as a result the expectation in the forming phase of the grid was
with only two layers while the laying of a third and fourth layer occurred after the
structure had been formed.
3.3 The Idea of the New Tool and Its First Application
It was precisely the adherence to the Savill model, carried out with the project of San
Vincenzo, to suggest the idea that with the two-stage assembly of the layers of sticks
one could, with a certain simplicity, operate on the thickness of the shear-block to
vary the inertia of the cross-section between different part of the same structure.
If we bare in mind the shear-block composed by overlapping layers, it is immedi-
ately possible to vary the number without immediately incurring problems of stability
of the section, that is stabilized at each micro-module from the intersection with the
orthogonal layer.
It was this precise constructive intuition that pushed towards the creation of an
adequate system for the design of possible variations: hence GFFT 2.0 was born.
In the case of Foggia, immediately following that of San Vincenzo, there was a
first experimental use of the newborn tool. Unfortunately, the influence of a design
nature have forced the shell to choose a rather simple shape, characterized by a single
direction of curvature, with only a slight deformation of the rods on the surface.
The synclastic surface subjected to the previously described chain of tools, gener-
ated a gridshell that did not present particularly significant stress differences among
the various zones. In fact, the variation of the section proposed by the software was
at most one layer (50 mm) between the compressed zones and those subjected to
combined compression and bending.
For this reason, the constant thickness which makes it possible to simplify some
phases of the construction process, has been considered preferable. Experimental
software has shown a greater incidence capacity in structures characterized by more
complex surfaces deriving from the harmonic composition of synchlastic parts and
anticlastic parts. For the time being, the greatest effectiveness has been tested only
on models that have not yet reached the executive level but the first results justify
a cautious optimism on the possibility that GFFT 2.0 can contribute to the effective
design of active bent timber gridshell and not only of a theoretical nature.
4 Conclusions
The choices on the project tools and on the technologies that have guided these two
most recent projects, have often been characterized by the novelty constituted by
the unprecedented public nature of the clients and therefore often imply a slight
“heaviness” of the construction processes in addition to a relative complexity of the
executive techniques. Nevertheless, we believe that we cannot fail to betray our his-
torical attitude inspired by the motto “complex design for easy construction”, whose
main objective has always aimed at naturally accepting the increasing complexity
of the design process by resisting, passively or actively, against the temptation of
raising the technological level of the construction.
264 S. Pone et al.
References
Chilton J, Tang G (2016) Timber gridshells: architecture structure and craft. Taylor & Francis,
London
Colabella S, Lancia D, Repola L, Memmolo R, Pone S (2015) A monitoring system for wooden post-
formed gridshells. In: Proceedings of the international association for shell and spatial structures
(IASS) symposium 2015, Amsterdam Future Visions, 17—20 August 2015, Amsterdam, The
Netherlands
D’Amico B (2015) Timber Grid-shell Structures: form-finding, analysis and optimisation. Ph.D.
Thesis, Edinburgh Napier University
D’Amico B, Kermani A, Zhang H, Pugnale A, Colabella S, Pone S (2015) Timber gridshells:
numerical simulation, design and construction of a full scale structure. Structures. http://dx.doi.
org/10.1016/j.istruc.2015.05.002
D’Amico B, Kermani A, Zhang H, Shepherd P, Williams CJK (2015) Optimization of cross-section
of actively bent grid shells with strength and geometric compatibility constraints. Comput Struct.
http://doi.org/10.1016/j.compstruc.2015.04.0068
Harris R, Haskins S, Ronyon J (2008) The Savill Garden gridshell: design and construction. Struct
Eng 86(17)
Kelly O, Harris R, Dickson M (2001) The Construction of the downland gridshell. Struct Eng 79(17)
Kuijvenhoven M (2009) A design method for timber grid shells. MSc Thesis Delft University of
Technology
Piker D (2013) Kangaroo: form finding with computational physics. Archi Design 83:(2)
Pone S (2012) Gridshell. I gusci a graticcio tra innovazione e tradizione, Alinea, Firenze
Pone S (2017) Digital tools and experimentations for structures realised with the active bending.
Techne—J Technol For Archit Environ 13:306–312. https://doi.org/10.13128/Techne-19748
Pone S, Colabella S, D’Amico B, Fiore A, Lancia D, Parenti B (2013) Timber post formed grid
shell: digital form finding/drawing and building tool. In: Obr˛ebski JB, Tarczewski R (eds) Beyond
the Limits of Man, Wroclaw, Poland, proceeding of IASS Symposium 2013
Pone S, Mirra G, Pignatelli E (2017) InFormer: designing forming actions in post-formed gridshells
by means of MOGAs. In: Bögle A, Grohmann M (eds) Interfaces: architecture-engineering-
science, Hamburg, Germany, proceeding of IASS symposium 2017
Pone S, Mirra G, Pignatelli E, Colabella S, Lancia D (2016) Specialised algorithms for different
project stages in a post formed timber gridshell design. In: Proceeding of ICSA symposium,
Guimares, Portugal
Preisinger C, Moritz H (2014) Karamba—a toolkit for parametric structural design. Struct Eng Int
Assoc Bridge Struct Eng 24(2):217–221
Toussaint MH (2007) A design tool for timber gridshells. Ph.D. Thesis, University of Technology,
Delft, Nederland
Reinterpretation of Traditional Wood
Structures with Digital Design
and Fabrication Technologies
Philip F. Yuan and Hua Chai
Abstract While digital design and fabrication technology has strongly promoted the
innovation of wooden architecture from different aspects, they have also provided
us with a new perspective to re-examine traditional wood structures. In the digital
wood researches of DDRC, Tongji University, traditional culture has often played an
important role. Both the culture of Chinese traditional wood architecture and west-
ern structure systems have become the source of inspiration. This paper presents
two research projects that reinterpret traditional wood structures through experimen-
tal construction of large scale prototypes. “REVERSE RAFTER” demonstrates the
possibility of simulating and reinterpreting Chinese traditional wood tectonics with
structural performance based-design and digital fabrication technology, while “Digi-
talFUTURE 2017 gridshell” showing an approach for extend gridshell system of the
west into more complex application. By putting digital technologies under the con-
text of wood culture, reinterpretation of traditional wood structures has the potential
to continuously provide inspiration and resources for innovative wood architecture
practice.
Keywords Reinterpretation · Traditional wood structure · Structural performance

Digital fabrication
1 Introduction
In the past ten years, almost unprecedented technological development has strongly
promoted the innovation of wooden architecture from different aspects. While
wood has long been the most tempting building material with the development
of material technologies, only in recent years has wood architecture been freed from
P. F. Yuan (B) · H. Chai

College of Architecture and Urban Planning, Tongji University, Shanghai, China
e-mail: philipyuan007@tongji.edu.cn
H. Chai
e-mail: chaihua@tongji.edu.cn

266 P. F. Yuan and H. Chai
the limitations of frame and truss structure, which should be attributed to the emerg-
ing and fast development of digital design and fabrication technology (Menges et al.
2016). Digital design methods, especially performance-based design approaches, has
vastly expanded the possibilities of wood structure. At the same time, the growing
trend of research in digital fabrication has provided a strong backing for the appli-
cation of digital wood architecture, accelerating the practice of innovative wood
architecture.
Not only digital design and fabrication technologies has brought about innova-
tion in wood construction, it also provides us with a new perspective to re-examine
traditional wood structures. Digital design methods enable architects to reinterpret
exiting structural types with parametric thinking, meeting complex needs of modern
life without sacrificing their structural performance. At the same time, while tradi-
tional wood tectonics are the accumulation of the experience and understanding of
craftsmen through thousands of years, digital simulation technologies has made it
possible to excavate the implicit knowledge inherent in tradition wood architectures,
which could be simulated in digital environment, and reinterpreted into innovative
wood design. While innovation from scratch is difficult, the excavation and rein-
terpretation of traditional wood structures has provided a feasible solution for the
development of digital wood architecture. Some studies have shown great potential
for this approach. The researches on folding wood structures carried out by IBOIS
has investigated the potential of folding structure in wood industry through a series
of digital modeling and prototyping experiments (Buri 2010). Bending-active struc-
tures has also been widely researched as an extension of strained gridshell structure
put forwarded by Frei Otto (Lienhard 2012). A variety of traditional wood joints
have also been reinterpreted and developed with digital fabrication technology like
CNC milling. As wood construction has a long history in both the East and West,
there is a wealth of potential out there to be further explored.
In the digital wood researches of DDRC, Tongji University, tradition has often
played an important role. Both the culture of Chinese traditional wood architec-
ture and western structure systems have become the source of inspiration for digital
wood research. This paper presents two projects in different culture context that show
different reinterpretation approaches mentioned above. The first is a project called
“REVERSE RAFTER” which explores the possibility of simulating and reinterpret-
ing Chinese traditional wood tectonics with structural performance based-design
and digital fabrication technology (Yuan and Chai 2015). The second one, “digital-
FUTURE 2017 gridshell”, tried to extend gridshell system of the west into more
complex application (Yuan et al. 2018).
Reinterpretation of Traditional Wood Structures with Digital … 267
2 REVERSE RAFTER—Reinterpretation of Traditional

Chinese Wood Tectonics
Reinterpretation of wood tectonics plays an important role in contemporary practices

and researches of Chinese architectural culture. Among all the digital technologies,
structural performance-based design has provide a solution for architects to deal with
traditional tectonics scientifically, integrating technology with culture. Through the
comprehensive analysis of form and performance, the qualitative knowledge can be
transformed into quantitative calculation, and the tacit knowledge can be transformed
into explicit knowledge, which could become an important reference for modern
wood practice.
“Reverse rafter” is a structural installation from “DigitalFUTURE Shanghai
2014”. Taking “eaves rafter” as research prototype, this project tried to reveal the
underlying structural principles of “eaves rafter” by employing topology optimiza-
tion as research method, based on which a reciprocal structure installation was design
and fabricated.
2.1 Prototype Research
Big roof with deep overhangs is the most prominent feature of traditional Chinese
architecture, which presents both aesthetic expression and reasonable force flow.
This project took “eaves rafter” as prototype, which plays the most important role in
supporting the overhanging structure.
Traditional roofs mainly bear loads like self-weight and snow loads by rafters with
two ends laid on purlins. The overhangs at the eaves put forward a higher structural
performance requirement. Supported by eaves purlin, purlin on hypostyle and eaves
tiebeam, eaves rafter cantilevers out obliquely to support the overhanging part of the
roof or flying eaves (see Fig. 1).
The book “Qingshi Yingzao Zeli”, which is comprised of the design principles,
manufacturing standards of Qing Dynasty extracted from “Qing Gongbu Gongcheng
Zuofa Zeli”, provides the overhanging size of eaves rafter, the horizontal distance
from eaves tiebeam to the eaves edge as 14doukou (a unit of measurement of Chinese
traditional wood architecture). The horizontal distance from eaves tiebeam to eaves
purlin is defined as 6 doukou, while that between eaves purlin and purlin on hypostyle
is 24doukou (Liang 1981). As the book does not give a clear explanation, it is not sure
whether the overhanging size was dictated out of structural consideration. Chinese
scholars have given the overhanging size a structural interpretation through structural
calculation and bending moment diagram (Jiang et al. 2011). By simplifying the
eaves rafter into a statically indeterminate oblique beam and balancing the positive
and negative peak moments, the research come out with a result consistent with the
description of “Qingshi Yingzao Zeli “(see Fig. 2).
Fig. 1 The location of eaves rafter
Fig. 2 Bending moment diagram of of eaves rafter
Structure calculations verified the rationality in the force distribution of eaves

rafter, but did not reveal the underlying structural principles. In contrast digital anal-
ysis software allows for more dynamic and interactive analysis and optimization. In
this project, the physical environment for eaves rafter was simulated in grasshop-
per plug-in “Millipede”. At the beginning, a rafter model with material information,
load condition, and external constraints was created. During the simulation, shape
changes were carefully observed with the adjustment of the overhanging ratio. Sim-
ulations showed that, with the same slope of rod, the reduce of the overhanging ratio
would cause the deformation between support points B, C to increase; When raising
the overhanging ratio, the cross-section near point A need to be strengthen to resist
bending moment action. Therefore, the appropriate overhanging ratio should be the
balancing result between the deformation of BC and section size of point A. After
several attempts, we found out that when the length proportion of AD and AC was
within the 0.46–0.48 range, the appropriate overhanging ratio could be obtained (see
Fig. 3).
Further studies showed that, when the position of support point B was changed,
the balanced proportion between AC and AD was almost unaffected, maintaining at
0.46–0.48 range. This results are basically consistent with the description in “Qing-
shi Yingzao Zeli”. It can be concluded that a most reasonable value exists in the
overhanging size of multi-span continuous beam. In the practice, the overhanging
Fig. 3 structural simulation: Simplified rafter model for simulation (up left); Equilibrium state
when AD/AC equals 0.2(up right); Equilibrium state when AD/AC equals 0.8(down left); Equilib-
rium state when AD/AC equals 0.47(down right)
size of multi-span continuous beam should be determined according to the specific

conditions, thus saving material, reducing project cost. Compared with the traditional
structure calculation method, the advantages of “Millipede” lies in its dynamic and
interactive interface, however, as diagrams, the analysis result of “Millipede” requires
further explanations to be reasonable for applications.
Analysis above has fully proved that traditional tectonics accumulated through
thousands of years contain a wealth of structural information. Structural analysis
nowadays can help to extract the underlying principles. At the same time, the con-
sistency between analysis results and traditional standards also demonstrates the
effectiveness and potential of digital structural tools.
2.2 Structural Performance Simulation and Optimization
On the basis of the prototype analysis, the project further studied the role and sig-
nificance of structural performance-based tools in structural design.
For further research, this project first accomplished a structural installation design
based on structural principles analyzed above. Taking advantages of topological opti-
mization function of “Millipede” in shape optimization and material size optimiza-
tion, this project then simulated and optimized the form and cross-section of the
structural installation.
Reciprocal structure was introduced in design phase. In reciprocal structure, each

rod overlaps with two adjacent rods, forming overhangs at both ends. Similar to the
force condition of eaves rafters, rods bear mainly bending moment action. Reciprocal
structure often served as supporting structure of bridges or pavilions in ancient China.
As horizontal extension is the major feature, reciprocal structure is conventionally
used to build a planar continuum. In contrast, reciprocal structure developed vertically
in this project.
Structure performance can be obtained by combining the structural principles with
certain algorithms. The basic unit was derived from the reciprocal unit with three
rods overlapping each other with reasonable overhanging proportion (see Fig. 4).
While the upper unit was generated based on the lower one in vertical development,
the unfolding degree of units increases layer by layer, rendering the overall structure
an umbrella form. Only a few parameters such as the radius of the first layer were
needed to generate the overall form and ensure each individual element to behave
adequately.
Since each unit undertook the weight of all the upper ones, the deformation degree
varies with height. Therefore cross-sections vary to respond to different force condi-
tions of different layers. “Millipede” FE analysis assumes totally fixed connections
between elements, which show completely different performance with overlapping
connections. Therefore, the simulation model was built with each rod simplified into
its central axis with auxiliary line connecting the axis at the joints. In this way, the
force flow in simulation model was made closer to real situation. After building
the simulation model, the project simulated the structural performance—axial force,
hear force, bending force, and resistance to deformation with Millipede” (see Fig. 5).
The built-in topology optimization function of “Millipede” showed the appropriate
cross-sections after optimization, which provided references for design adjustment.
Sectional sizes after optimization varied nonlinearly following changes of stress state.
Taking into account the accessibility and economy of materials, the project observed
the optimization results and classified cross-sections into several types, which were
then feedback to simulation. After repeating the analysis loop several times, the
Fig. 4 The generate process of a unit based on appropriate proportion

Fig. 5 The structural performance simulation—axial force (1), shear force (2), bend-ing force (3),
and resistance to deformation (4) with “Millipede”
cross-sections of the rods were finally divided into four categories: 90 mm × 40 mm,
70 mm × 30 mm, 55 mm × 20 mm, 40 mm × 20 mm, which set an end to the opti-
mization process. At this stage, the topology optimization function of “Millipede”
effectively ensured the design and optimization of wood tectonics.
2.3 Digital Fabrication
Dgital fabrication technology provides critical support for the realization of the
comprehensive structural performance-based design. In this project, although the
structural unit of each layer followed the same logic, the rod length, cross-sectional
shape, overlapping position and inclination angle varies from each rod, resulting
in extremely complex junctions. Deviations accumulation resulted from traditional
manual fabrication process would inevitably lead to the loss of control of the over-
all form. The capacity of 5 axis CNC’s provides suitable response to this demand,
preserving the intention of design initiatives.
This structure comprises of 63 rods in 21 different types, with lengths varying from
1 to 3 m. Before fabrication processes, each rod was labeled in sequence in both digital
and physical platforms. As each piece was unique, this was essential for assembling.
The localization and milling process of junctions was mainly performed by the 5
axis CNC, while construction on site was achieved manually (see Fig. 6). Similar
to Chinese traditional mortise-tenon structure, the entire structure was assembled
without any connecting members or reinforcement components (see Fig. 7).
Fig. 6 Label system and digital fabrication
Fig. 7 The final installation with mortise-tenon joints

Fig. 8 Reverse rafter installation
2.4 Evaculation
The final structure presents an umbrella form of 6 m high, force flow passing spirally
from top to bottom along the rods. The bottom radius of the structure is 0.5 m, while
the top radius extends up to 3 m. The triangle bench placed at the bottom not only
serves as resting space, but also acts as structural components to counter weight
and resist the overturning force to maintain the overall stability (see Fig. 8).This
project combined structural performance-based design with traditional wood tecton-
ics, endowing the structural installation with both cultural significance and structural
aesthetics.
3 DigitalFUTURE 2017 Gridshell—Reinterpretation

of Strained Gridshell
Brought forward by Frei Otto, strained gridshell is known as one of the most efficient
structure to cover large spans with lightweight materials, which is constructed from
an equal mesh net of continuous timber laths bent into the planned shape (Lienhard
et al. 2012). Strained gridshell, as opposed to the ‘unstrained’ gridshell which is

stress-free in the initial curved state (Adriaenssens et al. 2014), gains its strength and
stiffness through the bending action in its double curvature configuration (Pone et al.
2013).
Nevertheless, strained gridshell structure has been seldom used since the suc-
cessful construction of the Multihalle Mannheim in the 1970s. This situation can be
attributed to the difficulties of several aspects, among which design method is the
most important one. When trying to extend the range of gridshell to organic shapes
or complex geometries, finding an appropriate grid pattern that is compatible with
the material properties, is the most crucial issue. Several attempts have been made
to apply the principles of strained gridshells to free forms, such as the Smart Geom-
etry 2012 gridshell and the Ongreening Pavilion. All experiments explored complex
gridshells at the cost of requiring a more complex fabrication and assembly. But
the inconvenience of a complex fabrication process was well balanced by digital
fabrication technologies.
In this context, the DigitalFUTURE 2017 Gridshell aimed to explore another
solution for the design and fabrication of complex strained gridshell structure, which
took full advantages of digital fabrication technology to enable much flexible form
and pattern design. DigitalFUTURE 2017 Gridshell mainly focused on the design
and fabrication method of Enneper surface-shape gridshells.
3.1 Grid Pattern Generation and Optimization
The arrangement of grid patterns on the designed surfaces is a key issue concerning
the design of gridshell structure. Instead of using a flat grid at the first place, the
form-finding method proposed in this research is to generate a quadrilateral grid
pattern directly on the designed complex surface. The grid pattern will be optimized
afterward to meet the limitation of material performance by simulating the bending
action in the computer.
The initial form was a parameter-defined Enneper surface, which was subse-
quently went through some simple operation to meet the design requirements, such
as vertically scaled and a central opening. There are many ways to create a quadrilat-
eral grid on the surface, such as isocurves, principal curvature lines, asymptotic lines,
geodesic lines, principal stress lines. This research tried to use principal curvature
lines as the initial grid pattern for the Enneper surface-shape structure. The Princi-
pal curvature lines were generated on the surface using Grasshopper-based plugin
Millipede. Firstly the principal curvature of the designed surface was calculated in
grasshopper so as to define the parameterization field in Millipede. Then a network of
curves aligned to the field were extracted with the Reparameterization Outputs com-
ponents of Millipede. The density of the output curves can be controlled by manually
adjusting the scaling factor, which allowed designer to choose the most appropri-
ate curves density. After reconstructing, the resulting grid pattern was a relatively
uniform, orthogonal grid (see Fig. 9).
Fig. 9 The generation of the principal curvature lines on the designed surface
As the resulting curves may have problems like excessive curvatures, they could
hardly be used directly as the grid pattern for gridshells and built with timber. Further
measures were taken to optimize the curves. The most important objective of opti-
mization was to reduce the curvature of the curves. Therefore, a hypothesis was put
forward that, the curvature and complexity can be reduced effectively if an appro-
priate bending force was applied to the grid in computer.
The Live Physics engine of Kangaroo played a significant role in this stage. Each
continuous curve was considered as one timber lath in the strained gridshell. Starting
with the unstrained curves, a certain amount of bending force was applied to the
continuous curves in kangaroo. At the same time, a few other constraints were also
applied to the curves to keep the curves on the surface throughout the optimization.
Therefore the curves were allowed to move freely on the surface under bending
action. Once the simulation was started, the curvature tended to decrease with the
increase of the bending force. The interaction between curves will cause the whole
grid to slide on the surface, so that the overall shape will be optimized simultaneously
(see Fig. 10).
Since the optimization criterion was that the curvature of each curve need to be
smaller than the maximum curvature allowed by the material, the maximum curvature
of the materials that will be used to construct the large scale prototype - plywood
plate with a thickness of 5 mm—was roughly measured by bending several sample
pieces of 100 mm × 1000 mm. The ratio between the height of the arch and the initial
length was used to quantify the material’s ability of bending, which is was about 0.34
in this case. This ratio was reduced from 0.416 to 0.151 after optimization. Therefore
this optimization greatly reduced the overall curvature (see Fig. 11), thus effectively
reduce the difficulties of assembly process.
Fig. 10 The optimization process of the grid pattern
Fig. 11 The comparision of laths curvature before and after optimization
3.2 Structural Optimization
Since this method intended to integrate gridshells and digital fabrication techniques,
each component in this structural system does not necessarily need to have consis-
tent cross-section. Therefore, after the detailed gridshell system was developed, this
research further optimized the cross-section of the structure to improve utilization
efficiency of materials.
Following the structural system of traditional gridshells, this structure consists of
three parts: rigid edge beam, continuous laths and rotatable joints. The edge beam
took rigid Glued-Laminated Timber as material to resist bending action of the shell.
Using 5 mm plywood plate laths as the material, the laths employed the double-layer
Fig. 12 The double layer laths system and joints design
Fig. 13 Structural simulation and optimization process with Karamba
system developed by Frei Otto. As for the joints between laths, instead of using
tradition joints, this research invented a mortise and tenon joint system following the
same principle, which was much convenient to install and remove, and thus more
suitable for this experimental project (see Fig. 12).
The gridshell was optimized afterwards based on FEA (Finite Element Analysis).
An analysis model was built in Karamba (see Fig. 13). Taking digital model and
material properties of different components as input, both the cross-sections of laths
and edge beam were optimized according to the material utilization efficiency. As
the bending action could hardly be simulated directly in Karamba, bending action
was considered to as pretension stress applied to the laths beforehand.
Fig. 14 The Gridshell system after optimization: Edge beam (left); Laths (middle); the Gridshell
Structure (right)
Besides gradient graphs, the simulation results were also output in the form of
parameter values correspondence with the geometric components one-to-one. The
efficiency of material utilization was used as the parameter to determine the cross-
sections. Larger cross-sections were used where material utilization efficiency was
high and smaller cross-sections where efficiency was low. The initial cross section of
edge beam was 80 mm × 100 mm, while the width of laths was 70 mm. After the opti-
mization, the cross-sectional dimensions varied with material utilization efficiency,
which of edge beam varied between 60 mm × 90 mm and 90 mm × 120 mm, and
that of laths varied between 40 mm and 90 mm. The optimization process enhanced
the structure performance of gridshell by creating a direct connection between form
and inner forces (see Fig. 14).
3.3 Robotic Fabrication and Assembly
The resulting complexities of the gridshell were fully addressed in the fabrication
and erection process of a large scale gridshell structure. Different digital fabrication
methods were explored in the fabrication of different parts.
The laths were straightened onto a plane with all the information retained, such
as the location of joints and the width of the laths. Then the straight laths were
subdivided into segments which can easily fit into the plywood boards of 1220 mm ×
2440 mm, and milled with CNC (see Fig. 15). The milling of 38 plywood boards
were completed in 2 days.
The fabrication of spatial curved edge beam which was made of GLT (Glue
Laminated Timber) was the main challenge of this stage. Currently, while plane
curved GLT could be easily produced in the factory, there were still no efficient
solutions for spatial curved GLT production. Robotic timber wire-cutting technique,
whose capacity was tested in the previous research project (Yuan and Chai 2017),was
employed in this project to deal with the spatial curved edge beam.
Fig. 15 The laths were straightened on to the plywood boards and milled by CNC
Band-saws are employed in wood industry in many areas from primary log con-
version to furniture manufacturing. Compared with circular saws, band-saws have
advantages such as lower kerf waste and noise levels (Ulsoy et al. 1978). More impor-
tantly, as the width of the band-saw blade is much smaller, it’s possible to cut ruled
surfaces with a band saw by continuously rotating (either material or blade) while
cutting. While in conventional band-saw cutting process workers need complex jigs
to stabilize the cutting path when feeding wood through a band-saw, robots can make
this process flexible and precise.
A modified 14-inch band saw was mounted on a robust Kuka R2700 robot, to
cut the beam segments out from prefabricated simple ones. Some spare parts of the
band-saw were removed to reduce weight and increase the cutting range. With a
Motor of 1500 W, the saw is possible to cut materials with a cross-section of 35 and
30 cm.
The edge beam was divided into 12 segments according to the machine range (see
Fig. 16). As the beams were space curved, it was more complicate to find the form of
raw beams with minimal volume. Genetic Algorithm was employed in this stage to
find the smallest volume that can accommodate the target beams by rotating the beams
on XZ and YZ planes to changing the projecting angles onto XY plane. The beams
were oriented onto a horizontal plane to extract the outer contour. After offsetting the
contour 1 cm outwardly to make tolerance, the maximum width of the contour was
taken as the cross-sectional width of the raw beam. The raw beams were produced
in the factory with one of the most commonly used materials in the construction
of wood structures. The production process was guided by CNC templates which
was used to inform the curvature of timber beams. In spite of this, the accuracy of
the beams was difficult to be guaranteed in the manual production process, 1 cm
more tolerance was given in this stage. As the structure is centrosymmetric, the 12
segments actually had only two different shapes, and all the segments shared the
same form of raw beam.
Fig. 16 The edge beams were cutted from planner beams with robotic bandsaw
Fig. 17 Robotic cutting simulation in Kuka Prc
As there were substantial spatial torsions in most surfaces that need to be cut, the
tables for fixing the material were not feasible in this case. Triangle wooden frames,
which is more flexible to be customized for each beam, were developed as supports.
The raw beam was firstly fixed to the supports with woodworking clips, and the part
to be cut was cantilevering outside them. After measuring the exact position of the
beam, the model was generated in Rhino and Grasshopper, and then robot tool paths
were designed and simulated with Kuka Prc (see Fig. 17). In this way, each beam
only needed to be turn around once during the cutting process. Each segment was
cut from the raw one through 6 cuts (see Fig. 18). During the fabrication process,
the speed of the robot was continuously changing according to the thickness to be
cut. The speed of the cutting process was about 3–6 m per hour. The entire cutting
process took about 10 days in total. The perfect matching between segments proved
the accuracy of this technique.
As the edge beams and laths are all fully informed, the on-site assembly process
of the large scale structure was quite simple and efficient. First of all 12 edge beam
segments were connected with bolts and erected in place; then laths were assembled
one by one in way similar to weaving. The entire assembly work was done in a joint
effort of one designer and four workers in less than 20 h. The final pavilion shows the
feasibility and efficiency of robotic band-saw cutting in more complicate structures.
Fig. 18 Robotic cutting process
Fig. 19 Digital future 2017 gridshell
3.4 Results and Discussion
With a height up to 6 m, the large scale structure demonstrates the feasibility and
efficiency of the design and fabrication methods proposed in this paper (see Fig. 19).
Although some advantages of gridshells are lost in this method, the most important
merit of gridshells–the form-resistant mechanism imparted by the bending of the tim-
bers is retained. Not only the complexities can be balanced by the novel robotic tech-
nology, more importantly, this method demonstrates its potential to extend strained
gridshell into structures of complex geometries.
4 Conclusion
This paper presents two research projects that reinterpret traditional wood structures
through experimental construction of large scale prototypes. Form-finding, struc-
tural optimization, digital fabrication and assembly process are described in detail to
investigate different possibilities of reinterpretation. The final outcomes shows the
constant negotiation between design expression and fabrication constraints, between
tradition and state-of–the-art technology. By putting digital technologies under the
context of wood culture, reinterpretation of traditional wood structures has the poten-
tial to continuously provide inspiration and resources for modern wood architecture
practice.
Acknowledgements This research is funded by National Natural Science Foundation of China

(Grant No.51578378), National Key R&D Program of China (Grant No.2016YFC0702104), Sino-
German Center (Grant No.GZ1162), and Shanghai Science and Technology Committee (Grant
No.16dz1206502, Grant No.16dz2250500, and Grant No.17dz1203405).
Part of the text and images presented in this paper were modified from two previously published
papers in the proceeding of CAADRIA 2015 and CAADRIA 2018. See:
Yuan and Chai (2015).
Yuan et al. (2018).
References
Adriaenssens S, Block P, Veenendaal D et al (2014) Shell structures for architecture: form finding
and optimization. Routledge, London and New York
Buri HU (2010) Origami: folded plate structures. https://doi.org/10.5075/epfl-thesis-4714
Pone S, Colabella S, D’Amico, B et al (2013) Timber post formed grid shell: digital form find-
ing/drawing and building tool. In: IASS 2013: Beyond the limits of man
Jiang Y, Mao LT, Cao XL (2011) The structural analysis of overhanging length of ancient building
eave-rafter. In: 17th Beijing society of theoretical and applied mechanics congress proceedings.
2011, pp 572–573
Liang SC (1981) Qingshi Yingzao Zeli. China Architecture & Building Press, p 117
Lienhard J, Gengnagel C, Knippers J et al (2012a) Active bending, a review on structures where
bending is used as a self-formation process. Int J Space Struct 28(2–3):187–196
Lienhard J et al (2012b) Active bending, a review on structures where bending is used as a self-
formation process. Int J Space Struct 28(2–3):187–196
Menges A, Schwinn T, Krieg OD et al (2016) Advancing wood architecture: a computational
approach. Routledge
Ulsoy AG, Mote CD, Szymni R (1978) Principal developments in band saw vibration and stability
research. Holz als Rohund Werkstoff 36(7):273–280
Yuan PF, Chai H (2015) Reverse Rafter: Structural performance simulation based on wood tectonics.
CAADRIA 2015:693–702
Yuan PF, Chai H (2017) Robotic wood tectonics. Fabricate. Stuttgart, pp 44–49
Yuan, PF, Chai, H and Jin, JX (2018) Digital form-finding and fabrication of strained gridshells
with complex geometries. In: Proceedings of the 23rd CAADRIA conference, Beijing, China,
2018, pp. 267–276
Part II
Centrality of Representation As a Synthesis
Model and Morphological Comprehension
Wood as an Essential Material
in Architectural and Civil Engineering
Models from the Renaissance
to the Architectural Avant-Garde
José Antonio Franco Taboada
Abstract The mock-ups, as important project instruments, have accompanied the

work of the technician from ancient times. In this process, wood has always been an
essential material due mainly to its abundance and properties, particularly its flexibil-
ity, strength and even some malleability. This work aims to provide an overview of the
history of wooden models through some notable examples of the use of this material
in architectural and engineering mock-ups, analysing their purpose from antiquity to
the present day, as well as their role today as an alternative or complement to virtual
models.
Theme or topic: History of representation in wood design.
Keywords Models · Wood · Architecture · Engineering
1 Introduction
“A history of architectural models as an integral part of the design process has yet to be
written”.
(Millon 1994, p. 72)
These words by historian and architect Henry A. Millon, one of the world’s leading
experts on architectural models, were written in 1994 for the catalogue of the most
complete and comprehensive exhibition ever organized on Renaissance models. A
quarter of a century later and a few months after his death, they are still in full force.
Like the laureate professor Millon, many historians and architects have pointed
this out. Among them was John Winton-Ely, who said, in connection with the
many studies and publications on recent architectural drawings—referring to the
1960s—that the intrinsic qualities and meaning of the models had not been suffi-
ciently recognised. Specifically, this refers to the models in which the skills of the
J. A. Franco Taboada (B)

Higher Technical University College of Architecture, University of La Coruña, Galicia, Spain
e-mail: jafranco@udc.es

286 J. A. Franco Taboada
carpenter and the architect are combined, that is, the wooden models (Wilton-Ely
1968). This chapter focuses on this speciality which, given the unmanageable nature
of the study of scale models and in particular wooden ones, as well as the limitations
of the work, has focused on those projects which, in my opinion, best represent their
evolution.
Since ancient times, models have been made for a wide variety of purposes, rang-
ing from religious to construction (Franco 2017). In relation to the latter, they have
become an essential tool in the process of conception and implementing the ideas of
the architect or engineer. Although many materials have been used throughout his-
tory, we could establish with some certainty that wood prevailed in the Renaissance,
mainly due to its nobility and durability.
The 19th century brought with it a paradigm shift in the construction of scale
models, brought about by the new architecture of iron, which led to the combina-
tion of traditional materials with others that were newer at the time, such as metals.
Since then, the raw material of the models has evolved over time, giving way to new
materials, some as novel as the plastics used in 3D printers. Wood has enabled many
generations of architects—including great Renaissance geniuses such as Brunelleschi
and Michelangelo, and later, their spiritual heirs such as Sir Christopher Wren and
Filippo Juvara—to realise their great dreams in order to convince important institu-
tional clients, but also to understand their own ideas and perfect them, in light of the
models.
The field of construction encompasses both the great architectural creations and
those of civil engineering, because how could we classify Leonardo da Vinci in his
multifaceted construction activity? Was he an architect, an engineer, or both? It is
believed that this artist made the scale models of his works mainly in wood—although
some say that he also used marzipan for this purpose1 —or at least this is how all those
who reproduce his creations today interpret them, who do not conceive them in any
other way than essentially in wood. In this material, magnificent models were created
that reflected important works of engineering or military architecture such as bridges,
lighthouses or dykes, which were increasingly daring, and large fortifications of a
tectonic nature, which sometimes and in their ambition, described entire cities such
as Madrid, in order to understand their functioning and plan their future development.
In this sense, if in an architectural model the scale was fundamental, in an engi-
neering model it became essential, reaching the smallest details that could represent
the viability or lack thereof of the future construction. The wooden dome model
of the Grain Market in Paris, completed in 1783 according to the assembly system
developed by the Renaissance architect Philibert de l’Orme, consisted of more than
13,000 pieces of the same material.
In the field of civil construction, they were essential elements of the Universal
Exhibitions and other important exhibitions to explain convincingly to all visitors
1 Several authors have included in their books references to the work Leonardo’s Kitchen Notebooks,
written by the couple of historians Shelagh and Jonathan Routh with the intention of entertaining
their readers, giving for certain some of the “supposed facts” that appear there, such as attributing to
Leonardo the realization of architectural models with marzipan, possibly inspired by Vasari, which
speaks of edible models.
Wood as an Essential Material in Architectural and Civil … 287
the technical achievements and the historical heritage of the participating countries.
These models later became part of museums such as the Musée des Plans-Reliefs in
Paris or the model galleries of the technical education centres of the time.
They played a major role as an element of teaching in both the colleges of architec-
ture and engineering, and still do, despite the unfortunately relatively few examples
that remain. Some are authentic monuments or, to put it more modestly, assets of
cultural interest that reflect the historical memory of achievements that in many cases
have disappeared, remaining as silent but very expressive witnesses of the genius of
their creators.
The role of wooden models in archaeology should also be considered. A field in
which they have proved particularly useful as a tool for the restoration of buildings
already built or for the recreation of those that have already disappeared. In the latter
case they have even become true objects of desire for galleries and museums that,
like MoMA, already have professionals specialized in the restoration of architectural
models of historical and cultural interest.
An interest that is understood above all because sometimes the only thing left of a
construction—which may not even have been made—is its model. And in some cases,
when not even the scale representation is preserved and only the graphic testimony
of the latter remains, in the form of photographs, these images force us to consider
the relationship between the camera and the model, the visual interpretation of the
former by the latter.
2 Wood as a Scale Building Material
The world of wood as a material used to create architectural models is much more
extensive than it might seem. In fact, on the one hand we have wood that comes
directly from the tree and, on the other, products derived from it subject to varying
degrees of processing.
Among the former, in turn, we have a variety of colours, textures, and grains,
ranging from those easiest to work with, such as balsa wood—usually used in the
form of elongated boards of various thicknesses-and linden (Fig. 1), followed by
pine and then going all the way up to the hardest and therefore more resistant, such
as maple, rosewood, teak or mahogany.
Among the various wood-derivative products we must consider plywood and
particleboard, especially of the medium density type, DM or MDF, which can be
water-repellent, made with fibres or shavings generally of pine wood, which has had
its lignin removed and is agglomerated with synthetic resins. These water-repellent
boards, the most suitable for models where durability is desirable, can be divided into
three groups, from least to highest strength: particleboard (MDF and OSB), phenolic
plywood and compact panels (HPL phenolic wood fibre). Strictly speaking, paper and
cardboard, both made of cellulose fibres, should also be considered wood derivatives,
and therefore models made with these raw materials, which can be painted or pre-
printed, should be considered to be wood.
Fig. 1 White-painted balsa wood models of Le Corbusier buildings. Agglomeration contour lines.
Le Corbusier 25 + 1. Models to aid the understanding of architecture. Higher Technical University
College of Architecture (A Coruña). Photographs by the Author (2018): a Houses for artisans, 1924;
b Convent of Sainte Marie de La Tourette, 1956–60
Throughout history, the choice of the type of wood has been the result of a wide
range of factors, from the most elementary such as the ease with which it can be
obtained—that is, based on reasons of opportunity and economy of means—to the
most complex or sophisticated such as the search for beauty through its semantic
character, which evolved in line with technique and fashion, often influenced by the
discovery of new foreign, and therefore exotic, species.
In the Renaissance and over the following centuries, the choice of wood variety
depended to a large extent on the phase of the project in which the model was to be
used and the importance placed to it for the future. Thus, although most of the models
Sir John Soane produced were made of wood, half were made of pine and originally
painted. However, when a greater degree of detail and durability was needed, or they
were required to impress the customer, they were made of mahogany.
This tradition continues today, with the exception that the choice is no longer
limited to a simple disjunctive between the most common woods and the noble ones,
since it includes a wide variety of derivative products that can also be combined
with multiple non-wood materials, from whose dialectic new forms of architectural
expression are born, such as the Tatlin tower or the different versions of Gehry’s Bil-
bao Guggenheim, models that have contributed significantly to enriching the silent
language of architecture. The latter is perhaps the best example of the use of scale
models as successive representations of the final form, in which wood and its deriva-
tive, paper, are the ideal vehicles to materialise an idea. However, they are not the only
ones, since depending on the phase of the project, they are combined with materials
such as plaster or iron depending on the elements to be shown and the meanings to
be conveyed.
3 The Models of Antiquity and the Dark Centuries
The models of ancient Mesopotamia and Egypt, known as “houses of the soul” or
Spirit Houses, have long been analysed and valued for their supposed votive intention,
although it seems logical to suppose that the architects of the time used similar devices
to convince their clients, generally religious or civil authorities, of the benefits of their
proposals. Other examples such as the Assyrian reliefs of Khorsabad show tributaries
carrying the model of a fortress, around 742–705 BC (Esparza 2006, p. 59). For
Gentil, “The mere votive consideration that is usually given to them does not detract
from their significance; whether they represented an ideal building or one which had
already been constructed, nothing prevented them from having been used—even at
a later date—as the basis for a real building”. In fact, “it is more than possible that
their use was, simply, much more common than that of flat representation” (Gentil
1998, pp. 16 and 18). We can also assume, precisely because they have not made
it to our days, that many of them would possibly be made of wood, given that the
relative durability and flammability of this material would make it very difficult to
survive over so many centuries.
Similar models had to exist in Greece, ancient Rome and its predecessor Etruria,
as well as other Celts and the ancient Mediterranean, not to mention those of the pre-
Columbian peoples or the Far East. In Greece, we know from the famous Constitution
of Athens, attributed to Aristotle, that the city’s Council of Five Hundred required
architects to present their projects in the form of models, which should obviously be
made to scale (Adam 1997, p. 31) and presumably also in wood.
There are not many examples of models from the Middle Ages either, but in view
of the importance of the scale model since the First Renaissance, it seems logical to
assume that, at least in the Late Middle Ages, scale models were built prior to the
construction of works of a certain religious or civil importance, especially taking into
account the technical and material difficulties to produce accurate and convincing
drawings.
In this sense, although with a clearly commemorative purpose, the relief in Ulm
Cathedral is surprising. It depicts former mayor Lutz Kraft and his wife literally
putting the responsibility for the construction of the new church on the shoulders of
architect Heinrich Parler (Fig. 2a) in the form of a polychrome stone model, like the
rest of the relief, but the original of which was probably made of wood.
4 The Golden Age of Wooden Models: The Renaissance
Leon Battista Alberti in the Second Book of his work De re aedificatoria, dedicated
to La Materia, shows that in the 15th century there was already an “old” custom in
Italy of making models or that, in any case, they should be made in order to build
properly. And although he raises the possibility of using other materials, first of all
he speaks of “slats”, that is, of wood: “I will therefore always approve of the old
Fig. 2 a Detail of the relief of Ulm cathedral. Photograph: Rufus46 via Wikimedia Commons (CC
BY-SA 3.0); b Brunelleschi’s wooden model of the dome and sections of the apse of Florence
Cathedral. Photograph: Piero Baguzzi, © Museo dell’Opera di Santa Maria del Fiore (Millon 1994,
p. 23); c Antonio da Sangallo the Younger’s unbuilt project for St Peter’s, Fabbrica di San Pietro,
Vatican (Alimonti 2018)
custom of those who build well, that we should not only (think) with writing and
painting, but also with models, and making examples on a tablet, or whatever else,
of all the work and each of the measures of all the parts, as if they were the most
exercised time and again, and again, and let them be examined first to begin with
something else that requires expense and care. But in making the models you will
be offered an opportunity to look beautifully at and consider the site of the region,
and the roundness of the space, and the number and order of the parts, and the facie
of the walls, and the firmness of the ceilings…”.2 In other words, for Alberti, the
model is clearly a project instrument, not only for the building itself, but also for its
relationship with the environment. And he adds that the “shaved (so to speak) models
that are embellished with soft and delicate touches of paint” are only made by the
ambitious ones who seek to move away from the right to examine them and “analyse
the parts”. He concludes that the models must be “naked and simple, in which the
inventor’s creativity shines through instead of the hand of the creator” (Alberti 1582,
p. 34).
For Henry A. Millon, “The model, then, for Alberti, was not a vehicle to present
an idea to a client, but a means to study and realize an idea. For Brunelleschi and later
for Michelangelo, on the contrary, the model was apparently the representation of
an idea already formed in the mind to serve as a guide for workmen in construction”
(Millon 1994, p. 24).
The Renaissance models continued the tradition of ancient Greece, in the sense
that they were compulsory in competitions as patrons demanded them when inviting
several architects. As Gentil affirms, for Renaissance architects, their main design
instrument was not the perspective, as has normally been believed, but the model
generally made of wood which, given its flammability—accidental or provoked—and
2 The first Spanish translation of Alberti’s work, De re aedificatoria, was published in 1582 under
the name of Los diez libros de Architectura de Leon Baptista Alberto and included the following
year in the index of forbidden books of the Inquisition.
its difficult conservation, have led to very few examples, relatively speaking, reaching
our times (Gentil 1998, pp. 23–42). Thus, for Filippo Brunelleschi, the fundamental
figure of the first Renaissance, the wooden model of the dome and the sections of
the apse of the Cathedral of Florence, had to be essential for its definition3 (Fig. 2b).
The lantern that finally crowned the dome, the subject of a later competition that
Brunelleschi also won, does not appear on his model.
The great exhibition Italian Renaissance Architecture. From Brunelleschi to
Michelangelo, held in 1994 at Palazzo Grassi in Venice, brought together and showed
the public an extraordinary collection of materials on the important Italian 15th and
16th centuries. As Paolo Viti, director of cultural activities at Palazzo Grassi said,
“This ambitious exhibition to present an image of the Renaissance will be impossible
to repeat, since the models and numerous works will not be on loan from Italian and
foreign museums again” (Samaniego 1994).
Viti mentions the scale models first, as they undoubtedly constituted the essen-
tial or, at least, the most spectacular part of the exhibition. As Fernando Samaniego
pointed out, “The first visitors who entered yesterday from the same place praised
the artefact that represented the greatest structure in the history of architecture”
(Samaniego 1994). He was referring to the model of Antonio da Sangallo the Younger,
at a scale of 1:30, which was the star of the exhibition (Fig. 2c). Its wooden construc-
tion by Antonio Labacco lasted seven years, from 1539 to 1546, and was used by the
workers. It is the largest surviving model of the Italian Renaissance, few models of
which are preserved, despite the fact that hundreds or even thousands were possibly
built. It is an authentic model that can be visited, which shows the interior and the
exterior and which, as we know, Michelangelo did not like. He made his own wooden
model of the drum and the dome.
Up to thirty wooden models were exhibited at the exhibition. For the author of
the montage, Mario Bellini, “An exhibition of architecture has a quality of power-
lessness, since buildings cannot be brought in”, but the models are “an entrance into
an illusion machine, so that the spectator can get closer to the truth of architecture”
(Samaniego 1994). Millon, publisher of the exhibition catalogue, mentions several
wooden models of which payment documents have been preserved but which have
now disappeared: by Francesco di Giorgio, Bramante, Leonardo or Sangallo him-
self. For St. Peter’s there were models of Bramante, Raphael, Peruzzi, Sangallo the
Younger, Michelangelo, Giacomo della Porta and Maderno, of which only that of
Sangallo and two of Michelangelo (Fig. 3), one modified by della Porta, remain
(Millon 1994, p. 35).
Another extraordinary example of a wooden model that has been preserved is that
of the Cathedral of Pavia. Although the work has been attributed to Bramante—who
really seems to have intervened in a later phase, the initial project of 1487 corre-
sponds to the engineer, architect and sculptor of Pavia Cristoforo Rocchi, carried
3 The wooden model was fundamental for the design of the dome, but for its construction
Brunelleschi used the Cupulina discovered in 2012, which he kept secret from his competitors.
With its almost three meters in diameter, it was used to test the famous structural placement of the
herringbone bricks that saved him a huge scaffolding.
Fig. 3 Wooden model of the dome of St. Peter’s in which Michelangelo, Giacomo della Porta and
Luigi Vanvitelli took part. Fabbrica di San Pietro, Vatican. Photographs: Antonio Solazzi (Millon
1994, pp. 44–45)
out in collaboration with his fellow countryman Giovanni Antonio Amadeo. It is

not surprising, in view of the model, that it was known above all for its mastery in
woodworking (Fig. 4).
Looking at the large model, its similarity with the building that was finally built
stands out, possibly due to the subsequent intervention on it and on the model by
Giovan Pietro Fugazza himself, who succeeded Rocchi in the direction of the works.
5 Virtuosity in Model Making During the Baroque Period
The tradition of making models, mostly of wood, was maintained throughout the
17th and early 18th centuries, coinciding with the period that we now stylistically
define as the Baroque.
Just as the Italian Renaissance Architecture exhibition commissioned by Henry
Millon brought together for the first time the major models of that period, the same
researcher organised in Turin in 1999 the exhibition entitled The Triumph of the
Baroque: Architecture in Europe 1600–1750, with 80 original wooden models from
the 17th and 18th centuries, although there were no Spanish or Portuguese models
due to the lack of response from both countries (Millon 1999).
Fig. 4 Woden Model of Pavia Cathedral attributed to Cristoforo Rocchi. Fabbriceria della Catte-
drale, Pavia. Photograph: Carlo Dell’Orto via Wikimedia Commons (CC BY-SA 4.0)
Spain was asked for models of buildings with such importance as the basilica of
El Pilar in Zaragoza, the work of Ventura Rodríguez, or the façade of the city hall
of Salamanca (Fig. 5a), made of cedar wood, the work of architects Juan García
Berruguilla and Andrés García de Quiñones, although initially designed by Alberto
Churriguera. The latter is revealing when showing the two towers projected on the
sides of the façade, which were not built due to weight problems, and would have
given it a religious air and possibly interrupted the harmonious crowning line of the
square, from which the façade of the town hall stands out only just enough to highlight
its importance. And possibly the belfry of the model inspired the construction, almost
a century later, of the one that now exists in the centre of the building (Fig. 5b).
An extraordinary wooden model in the Madrid History Museum is the model
of the main façade of the Madrid Alcázar or fortress, dated by the museum itself
between 1630 and 1646, although some researchers place it around 16224 (Fig. 5c,
d). It is the work of the Roman architect and painter Giovanni Battista Crescenzi,
who after working for Pope Paul V, was called to Madrid by Felipe III, to create
and decorate the pantheon of Kings (to express gratitude, Felipe IV appointed him
marquis of the tower and knight of the Order of Santiago).
4 Among others, Juan Luis Blanco Mozo, lecturer at the Department of History and Art Theory of
the Autónoma University of Madrid, in the conference entitled “Idea and Project. Giovanni Battista
Crescenzi and the model of the Alcázar de Madrid”, given during the conferences organised by the
National Archaeological Museum of Spain on 22 February 2018 (the conference is recorded and
can be consulted online at https://www.youtube.com/watch?v=Oi4Rkf4rWmo).
Fig. 5 a Wooden model of the Salamanca City Hall (Estella 2005, p. 71); b The Salamanca City
Hall in the Main Square. Photograph by the Author (2009); c Scale model of the main facade of the
Alcázar of Madrid. Madrid History Museum. Photograph by the Author (2018); d View from the
west of the model. Photograph by the Author (2018)
He was a friend of another great painter and architect, Diego Velázquez, a facet
of the latter generally unknown but which has been studied in detail by Antonio
Bonet Correa (Bonet 1960), with whom he shared the name of “Tracist” or inventive
architect. Crescenzi is credited with consolidating this model of architect in the court
of Madrid to the detriment of practical architects or “master builders”: “More and
more frequently, clients (including the king) preferred to entrust the blueprints of
their buildings to painters or sculptors skilled in artistic or descriptive drawing and
versed in geometry and perspective, which enabled them to produce proposals of
remarkable beauty, audacity and suggestion, thanks to the incorporation of touches
related to their main area of expertise and their extraordinary capacity to represent
imagined space” (Blasco 2016, p. 298). The two “Tracists” did not get along well
with the great “practical architect” Juan Gómez de Mora, to whom works of the
former two have been attributed.5
The Alcázar is an “incomplete” model in the sense that, since it is essentially
intended to create a new façade for the large building that was already the Alcázar, it
does not represent the entire building; that is to say, it consists of the façade and the
5 Juan Gómez de Mora (1586–1648) is credited with works in the Alcázar, the Plaza Mayor, the Prison
of the Court of Madrid—currently the Palace of Santa Cruz, originally attributed to Crescenzi—and
the Casa de la Villa, among others.
Fig. 6 Wooden model of Juan de Villanueva’s project. Prado Museum, Madrid. Photograph by the
author (2018)
first centreline of the building, without a rear part. The primitive medieval fortress
was reformed by the Trastámara dynasty and then by the monarchs of the house of
Austria, with successive architects such as Luís de Vega, Alonso de Covarrubias, Juan
Bautista de Toledo and Francisco de Mora, succeeded by his nephew, the controver-
sial Juan Gómez de Mora, who apparently executed the final project. The scope of
Crescenzi’s intervention remains in the air, although we can assume that if he made
the model he would be the author of the project that was finally executed. The com-
plex disappeared after the fire that devoured it on Christmas in 1734, although there
are still many images that testify that the façade was built according to this model,
which has miraculously been preserved for almost 400 years. It is interesting to note
that between Marcelli’s “3D” axonometric plans of 1622 and those of Teixeira of
1656, the final façade of the Alcázar appears in the second, but not in the first.
Another extraordinary example of a scale model prior to an architectural per-
formance is Juan de Villanueva’s second project for the Cabinet of Natural History
and the Academy of Sciences of Madrid, then the Royal Museum of Paintings and
now the Prado Museum. It is a conceptual and at the same time prepositive model
that the architect presented to King Carlos II in 1787 and whose interest is even
greater because the corresponding plans have been lost. The project finally built
would vary little, some details, especially the central space for the conference room
which in the model was rectangular in shape and was built topped by a semicircular
apse (Fig. 6). Perhaps the most important variation is the elimination of the elevated
central body, which is not known what it would be destined for, possibly because
Villanueva himself considered that it broke the horizontal harmony of the whole.
It also shows what its rear façade would have been like without the subsequent
additions by more than twenty architects, the most recent and important of which
was the extension by Rafael Moneo, winner of the Pritzker Prize for Architecture in
1996, who communicated it underground—in fact with a low building with a green
roof—with the cloister of the former monastery of Jerónimos.6
The model was built in 1787 by an unknown craftsman with lemon, box,
mahogany, walnut and pine wood, which were assembled in great harmony, as can
be seen in the photograph, made in 2018. It measures no less than 362 cm long by
94 cm wide and 66 cm high.7
As for Italy, the Correr Museum of Venice, in the heart of St. Mark’s Square,
exhibits two unique 18th century wooden architectural models in the Sala delle
Quattro Porte, the importance of which is increased not only because they are among
the few that are preserved, but because they correspond to civil architecture and cer-
tify that at that time expert craftsmen worked in Venice. According to the Museum:
“In such demanding works, the conclusion of which always required several years,
the presence of a perfect on-site scale model that finished in every structural and
decorative detail both inside and outside (for this reason they could always be dis-
assembled and observed in every detail) was essential. It was a concrete reference
and even more invaluable than the paper designs; they were therefore essential and
a very tangible reference for the vast array of operators called upon to contribute to
the construction in a unified and organic manner (bricklayers, stonemasons, wood
and iron carpenters, decorators, etc.)”.8
The most complete model is the Large wooden model for Palazzo Venier dei leoni,
attributed to Domenico Rizzi, made with several types of wood and measuring 95 ×
118 × 153 cm. It represents a grandiose palace with great definition in all its refined
details. It is in fact a second version of the first commission by the Venier brothers
to Lorenzo Boschetti, with the same general scheme but a less refined architectural
language. Work was interrupted when they had just completed the first floor and
were finished off in what became the house-museum of Peggy Guggenheim, today
the museum of the same name (Fig. 7a, b).
The other unique architectural model is the Wooden model for a large Venetian
palazzo (Fig. 7c,d), which according to the museum itself could date back to the
late 17th or early 18th century. Designed to look out over the Grand Canal, it is a
perfect example of a design proposal that was never built. Its façade plays adequately
with its three storeys of height, distributed in turn into three vertical bodies, the
central one divided into five arcades per floor and the two lateral ones with two
slender rectangular windows per floor. According to the Museum: “Incomplete (note,
however, an astonishing Baroque staircase with multiple ramps), perhaps because of
its fragmentary condition, the model has a particularly evocative charm”.9
6 The successive modifications and enlargements of the museum can be consulted at: https://www.
museodelprado.es/museo/ampliacion-jeronimos.
7 https://www.museodelprado.es/coleccion/obra-de-arte/maqueta-o-modelo-del-tercer-proyecto-
de-juan-de/af672159-8cbd-4714-a4c0-6fb71e676044?searchid=8e8d02ca-3e32-9d30-13cc-
a5a72b6824c0.
8 http://correr.visitmuve.it/en/eventi-en/archivio-eventi-en/wooden-architectural-scale-models-
from-the-18th-century-presentation/2016/10/18211/louis-vuitton-wooden-architectural-models-
restoration/.
9 Idem.
Fig. 7 a, b Large wooden model for Palazzo Venier dei leoni; c, d Wooden model for a large
Venetian palazzo. Both at the Correr Museum, Venice. Photographs by the Author (2017)
In the Palazzo Madama in Turin there is another extraordinary wooden model,

in this case of the palace where Vittorio Amedeo II wanted to transform the castle
of Rivoli, so that it competed with the European greats, although the works were
stopped in 1720. The project carried out by Filippo Juvara in the first quarter of the
18th century served both for Carlo María Ugliengo to make the model and for the
six great panoramic views of what the castle would look like once it was converted
into a palace. Something like detailed infographics combined with a 3D printed
model of the present day. The perspectives were entrusted to the main specialists of
the time. The great panoramic view by Giovanni Paolo Pannini, entitled Castello di
Rivoli ripreso da levante, is shown in the same room where the model is exhibited
(Fig. 8a).
In England arose what is perhaps the best example of architectural model with a
triple purpose: serve as a design tool, as a way to convince the client and as a guide
for the builders of the building. It is the wooden model that the English architect Sir
Fig. 8 a Wooden model of the Castello di Rivoli with a Paolo Pannini’s oil painting on the left
wall with a panoramic view of the building; b Other view of the model. Palazzo Madama, Turin.
Photographs by the Author (2015)
Christopher Wren made in 1669 for the London Cathedral (Fig. 9), after the great fire
that devoured the city in September 1666 and destroyed the former church of Saint
Paul. As Wilton-Ely points out, “Wren brought to architectural design a scientific
cast of mind that found its expression in such experimental and explanatory aids as
models. The Great Model, made by William and Richard Cleer with twelve joiners
and a team of specialized craftsmen over a period of nearly ten months in 1673–74,
represents the unfettered imagination of the designer. Despite the subsequent aban-
donment of this design, mainly due to the mental climate of post-Civil War England,
something of its ideal grandeur was retained in the executed building” (Wilton-Ely
1976, p. 186).
Fig. 9 Sir Christopher Wren’s “Great Model” for St. Paul’s Cathedral in London. Photograph:
© 2017 The Chapter of St Paul’s Cathedral https://www.stpauls.co.uk/SM4/Mutable/Uploads/
generic_image/Great_Model.jpg
6 The Triumph of Drawing Over Models: Beaux Arts

Versus Industrialization
In 19th century architecture we have two fundamental milestones. On the one hand,
the strengthening of what was called Beaux Arts style or French academism, of the
Académie royale dÁrchitecture and its successor, the Académie des Beaux-Arts.
The projects were presented in large format plans and with techniques that hardly
allowed corrections, for which the contestants to the Grand Prix of the institution had
to master technical drawing, flat and descriptive geometry, with special emphasis on
the accuracy of shadows, which obviously led—at least among the most ambitious
architects—to the progressive abandonment of the models in favour of the beautiful
drawings of the academy.
This was not the case with Sir John Soane, who considered that the models were
not only necessary for the correct execution of the buildings, but also to obtain the
approval of the projects, as managers did not hesitate to introduce modifications
after seeing them (Wilton-Ely 1969). He also maintained that he always made scale
models and that the few times he did not, he feared that the building had suffered the
consequences, whether in terms of solidity, convenience or perhaps both (Wilton-Ely
Fig. 10 Model Room at Sir John Soane’s Museum, London. Photograph: Darley (2016)
1969). In his house he had a room dedicated to his own wooden models and to those
he bought outside, usually in other materials such as plaster, created by the famous
French model makers Jean Pierre and François Fouquet (Fig. 10). This room has
been restored in a recent renovation of the museum that bears his name: “The Model
Room, as recreated, faithfully follows a watercolour by C. J. Richardson of around
1834/35” (Darley 2016).
Even before the industrial revolution itself, a great technological level had been
reached, as shown by the model for the expansion of the hydraulic saw of the
Havana arsenal, made by military engineer Francisco Suárez Calderín around 1757
in mahogany wood. The expansion, which was twice the size of the existing 1748
installation, was never completed. This is an obvious example of the fact that it was
carried out as a preliminary test to ensure it worked properly (Fig. 11a, b).
On the other hand, the Industrial Revolution led to the increasing use of iron
in architecture, especially in engineering. In both cases, scale models, which often
incorporated iron like their corresponding constructions, were still considered fun-
damental. A notable example of this is the model of the elevated deposit of Santa
Engracia in Madrid (Fig. 11c), which was crafted after its construction in 1912
although the project dated from 1907, in other words, at the beginning of the 20th
century. It is made of painted pine wood, painted iron plate and patent leather card-
board and was built for the 1929 Ibero-American Exhibition in Seville (Lopera 2017,
p. 180).
To represent the model of the Santa Engracia tank—today converted into an
exhibition hall for the Canal de Isabel II, the company responsible for the full water
Fig. 11 a, b Model for the extension of the hydraulic saw of the arsenal of Havana, Francisco
Suárez Calderín, ca. 1757. Photographs by the Author (2017); c View of the 3D model created by
the author using photographic scanning of the model of the deposit of Santa Engracia in Madrid
(2018)
cycle in Madrid—I made a 3D photographic scan of the model, 2.20 m high, made
to a scale of 1:20. That is to say, the image that is published corresponds to one of
the multiple points of view of a 3D digital model made from the real model, in a
virtual-real dialogue that could be studied in more depth.
Another curious model from this period is the scaffolding system for placing
columns designed by Spanish engineer Agustín de Betancourt for St. Isaac’s Cathe-
dral in St. Petersburg, from which a wooden model that was designed to guide the
workers (Fig. 12) is preserved. The architect of the cathedral, Auguste de Mont-
ferrand, published in 1845 in Paris a luxurious album, Église Cathédrale de Saint-
Isaac…, with 49 lithographs documenting the development of the construction when
it had not yet been completed.
In the Madrid History Museum there is a large wooden model of the Teatro Real,
made by Custodio Teodoro Moreno around 1835. In the same way as the Renaissance
and Baroque models that have already been studied, the main characteristics of the
project are simplified, including the rear arcaded galleries and the upper gallery of
the façade, which was finally not built (Fig. 13).
The theatre was built in front of the Royal Palace, although due to economic diffi-
culties it was not inaugurated until 1850, with Donizetti’s La Favorita. Its architects
were, in order, Antonio López Aguado, Teodoro Custodio Moreno and Francisco
Cabezuelo, which explains the differences between the project initially conceived
represented in the model and the one finally inaugurated twenty years later, which is
reflected in the 1860 engraving of Fig. 13.
Fig. 12 a Scaffolding model for St. Isaac’s Cathedral (Navascués and Revuelta 2017, p. 273); b
Engraving by A. Montferrand, with the scaffolding in situ (Navascués and Revuelta 2017, p. 272);
c Model made of old pine and boxwood by Luís Cónsul Tedó in 1995 for the exhibition Betancourt,
the beginnings of modern engineering in Europe, Madrid. Author’s photograph (2017)
Fig. 13 a Model of Teatro Real by Custodio Teodoro Moreno. Author’s photograph (2018); b The
Teatro Real in 1860 (Amador and Rosell 1864, p. 436)
In this museum, and also made of wood, we have the model of the now disappeared
bullring of the Puerta de Alcalá in Madrid, made between 1843 and 1846 by retired
lieutenant colonel D. Juan de la Mata Aguilera (Fig. 14a, b). This bullring operated
between 1749 and 1874 and replaced the one built 12 years earlier—the first round
bullring—taking advantage of the existing space (López 1985, p. 115). Until then,
the bullrings were square or rectangular in shape and even irregular, since bullfights
were held in public squares in the cities, such as the Plaza Mayor in Madrid, sharing
space with other public activities.
In this case, the importance of the model also lies in the fact that this square was
visited by characters such as Francisco de Goya and inspired his works such as La
Tauromaquia, along with those of artists such as Gustave Doré or Edouard Manet, as
well as Richard Ford, Alfred de Musset, Théofile Gautier or Edmondo De Amicis,
among others. It inspired all of them with their artistic or literary creations thanks to
its architecture, evoked by the vivid reconstruction of the model, in which hundreds
of different figurines dressed in period costumes recreate the environment that these
characters lived in, although the perception of character and the acceptance of it have
changed radically in these times.
Fig. 14 a, b Wood scale model of the demolished bullring in the Puerta de Alcalá, 1846. History
Museum of Madrid. Author’s photographs (2018); c Alfred Guesdon’s View Taken from the Top of
the Bullring, ca. 1855 (Sáenz-López and Aragón 2013, pp. 9–10)
It is interesting to compare the shape of the square as it appears in the Madrid

Model of 1830 (Fig. 22) and the aerial view of Madrid of 1854 by the French architect
and lithographer Alfred Guesdon (Fig. 14c). In both of them there is a closed area
that occupies the equivalent of three of the arcades that occupy the two upper floors of
the building, while the rest are open. This enclosure with a higher roof overhanging
the square was probably modified in the reformation of the square in 1833, which
would explain the differences in detail between the two representations.
7 The Anthropocene Models: Two Iconic Examples
For some scientists, the 20th century marked the beginning of a new geological
era on Earth, the Anthropocene,10 a neologism coined by Nobel Prize winner Paul
Crutzen in the hope that “the term ‘Anthropocene’ will be a warning to the world”
(Kolbert 2011). This period is characterised essentially by the use of materials such as
aluminium, cement and plastics, but also by radioisotopes from atomic bombs since
the middle of the century. Obviously, these materials, especially plastics, in many
cases replaced wood and its derivatives, including paper and cardboard. However,
many others harmoniously combined them or simply continued to use wood, often
with a more conceptual approach.
7.1 The Tatlin Tower
There has been much speculation about the materials used by Tatlin in the orig-
inal model of his Monument to the Third International for St. Petersburg—then
Petrograd—in 1919, since it has disappeared, although there are several later ones
(Fig. 15a). This is possibly due to the description of the writer, critic and revolution-
ary Viktor Shklovsky, who in Knight’s Move refers to the utopian monument, not
the model, as “The monument is made of iron, glass and revolution” (Shklovsky and
Sheldon 2005, p 70).
Fig. 15 a Vladimir Tatlin’s monument to the third international or Tatlin’s Tower (unbuilt) with the
architect. Photograph: public domain, via Wikimedia Commons; b Vladimir Tatlin with his assis-
tants building the model. Exposition Russian Dada 1914–1924, Museo Reina Sofía. Reproduction
of the original photograph of an unknown creator taken by the author (2018); c Circle of Tobias
Verhaecht, The Tower of Babel. Photograph: Public domain via Wikimedia Commons
10 TheInternational Stratigraphy Commission, belonging to the International Union of Geological

Sciences, is studying the incorporation of the term Anthropocene within the geological time scale.
However, a photograph belonging to the Catherine Cooke Collection of the Uni-

versity of Cambridge shows several of Tatlin’s assistants working on the model,
which is made mainly out of wood joined by metal plates. In fact, one of the workers
is connecting11 with Tatlin’s help two parts with a hammer (Fig. 15b).
Nick Dunn believes that the construction of the enormous helical model of a
building with three rotating bodies that would be 400 m high and was destined to
be the tallest in Europe—surpassing even the Eiffel Tower, without any previous
definitive plans, allowed Tatlin to explore possibilities as he built it (Dunn 2010). Its
ambitious height and shape were obviously inspired by the Flemish paintings on the
tower of Babel. Although it is usually compared to the one made by Pieter Brueghel
the Elder in 1563, it is in fact much more similar to the one attributed to the Circle
of Tobias Verhaecht (Fig. 15c).
The symbolic importance of this model of an unconstructed building is clear, to
the extent that it was considered by Bruno Zevi, who called it the “flying machine”,
the most famous work of the period. Kenneth Frampton, Leonardo Benevolo and
William Curtis, among others, also considered it key, although analysing it from
very different points of view. But its importance lies in the fact that the model of
a building not erected and perhaps impossible to build in that period because of
its crazy costs and futuristic technology, has influenced the theory and practice of
architecture far beyond the countless most important buildings built at that time and,
of course, all the rest from constructivism.
Unlike the Renaissance models, key to their approval by the competent author-
ities and the construction of the corresponding projects, Tatlin’s model was able to
“do without” their practical realisation—likely impossible on the other hand, as has
already been pointed out—in order to become an autonomous piece of “architec-
ture”, independent of its existence as a building constructed in the history of this
field.
7.2 Adolf Loos’ House for Joséphine Baker
It is doubtful whether Loos ever met Joséphine Baker personally. What is certain
is that he admired her and that he must have at least attended her performances in
Paris or coincided at some party with her. According to Ines Weizman, director of
the Bauhaus-Institute of History and Theory of Architecture and Planning of Weimar
(Germany): “Baker—who at another earlier occasion had taught Loos how to dance
the Charleston—probably talked to many more guests of the cabaret that night, while
Loos in a mixture of anger, trying to show-off and most certainly in admiration of the
Black Venus of Paris, conceived a love letter in architectural form”(Weizman 2017,
p. 35).
The fact that the house was not built and that it was only known through the plans
made by Loos and, above all, through its model, did not prevent it from having a
11 De izquierda a derecha, Sofía Dymshits-Tolstaia, Tatlin; Tevel Shapiro e Iósif Meerzon, 1920.
Fig. 16 a Josephine Baker’s house © To the descendants of Ludwig Münz (Photograph: Albertina,
Vienna); b Huis van Roosmalen’s building facing the river Schelde in Antwerp (Photograph: Dirk
Verwoerd via Mimoa.eu); c EM2N building for social housing in Paris. Photograph: © EM2N via
designboom.com
long and somewhat morbid journey, essentially due to the originality of its model
(Fig. 16a). This is how “tributes” have been built, as clear and acknowledged as the
building facing the river Schelde in Antwerp, Holland, by Huis van Roosmalen in
1986 (Fig. 16b).12 Or the really elegant interpretation of the Swiss studio EM2N for
a social housing building in Paris (2012), in which the black and white marble strips
have been reinterpreted in a clearly modern language (Fig. 16c).13 It has even raised
its influence in the Newfoundland Shopping Centre built by Aldo Rossi in Olbia,
Italy, in 1997.14 However, I believe that, as Gravagnuolo has already pointed out,
both projects are based on Tuscan and Minoan architectures.
It is interesting to note that the design of the interior spaces of the house is even
more attractive than the model of the project he made for the dancer. According
to Albert Fuster, “In a letter from Kurt Unger (who collaborated with Loos on the
project) to Ludwig Münz of 23 July 1935, written in response to a request for clarifi-
cation from Münz, he explains how the pool should have large windows open to the
corridors, to “see people swimming and diving into the crystal clear water, flooded
[with light] from above: a ‘magazine’ underwater, so to speak” (Fuster 2006, p. 111).
This allows us to appreciate the differences between the approaches when design-
ing an essentially volumetric model like this one, in which what stands out is the
exterior aspect of the building and other more conceptual factors, as we will see later
with the example of Rafael Moneo, in which the essential thing is the description
of the interior spaces. In this case, the model that went down in history and had the
greatest influence on architecture was this one and not one of the interior spaces,
which apparently was never made.
Ines Weizman has argued in a revealing article on the copyright of architec-
tural works, which in Germany expire 75 years after the author’s death, that Loos’
12 https://www.mimoa.eu/projects/Belgium/Antwerp/Huis%20van%20Roosmalen/.
13 https://www.designboom.com/architecture/em2n-draws-from-adolf-loos-for-parsian-social-
housing/.
14 http://www.sardegnacultura.it/j/v/253?s=19398&v=2&c=2488&c1=2130&t=1.
house could become “a new project with almost no manipulation by the architect”.
Consequently, and in collaboration with Andreas Thiele, in 2008 she promoted a
reconstruction of it for the city of Ordos, Inner Mongolia (China): “I proposed to
celebrate the making public of the copy-rights over Loos’ oeuvre by building a fac-
simile of House Baker—the house he designed in 1928 in Paris for the legendary
singer Josephine Baker, but never realized” (Weizman 2017, p. 33). The master plan
was developed by Herzog & de Meuron, of Basel, and the artist Ai Wei Wei’s Fake
studio in Beijing.
8 Models as Design Instruments in Contemporary

Architecture: Some Pritzker Prizes
One of the most outstanding architects in the use of models is Rafael Moneo, who
uses the models not only to present his projects to contests or clients, like Luís de
Villanueva or Sir Christopher Wren and his Renaissance precursors did, but also
as a working tool. He used this system for one of his most important works, the
National Museum of Roman Art in Mérida (Fig. 17a), in which he made use of
simple agglomerated panels to help define what would later become the majestic and
evocative spaces of the building, the work that brought him international fame.
In the model for the parish church on the banks of the Loiola river, also in San
Sebastian (2001–2013), the wood not only helps define the construction but, carefully
treated, also defines the undulating shapes of its surroundings, which thus acquires
the prominence desired by the architect (Fig. 17b). In the model of L’Illa Diagonal,
the gigantic “horizontal skyscraper” that he designed and built for Barcelona in
collaboration with Manuel de Sola-Morales between 1987 and 1994, he humanises
its gigantic dimensions with, once again, the warmth of wood (Fig. 17c).
However, he does not hesitate to use more noble woods and combine them with
new materials such as methacrylate to show the interior and exterior of buildings
simultaneously, as he did with the model of the Los Angeles Cathedral (Fig. 17d). And
in the Kursaal in San Sebastian, he opens the building’s cold, translucent envelope
to show its warm, wooden entrails (Fig. 17e).
A well-known case of an architect using successive models as an integral part of the
design process is that of Frank Gehry, another Pritzker Prize winner for architecture.
Perhaps the building that made him most famous was the Guggenheim Museum in
Bilbao, for which he made up to six major models of the exterior volume during the
long development process in the 1990s.
The first, from 1991 to a scale of 1:500, was presented to the competition and was
made of lime wood and paper; the second, to a scale of 1:200, was made of plaster and
metal and was presented to the Guggenheim Foundation and the Basque government;
the third, also made of wood and metal to a scale of 1:200, was completed in 1993.
The fourth, to a scale of 1:200, was made of wood and paper and was shown at the
beginning of the work on 20 October 1993. The fifth, much larger one, was made
Fig. 17 Rafael Moneo’s models from the exhibition Rafael Moneo. A Theoretical Reflection from
the Professional Practice. Archive materials (1961–2013). Author’s photographs (2013): a Con-
ceptual model for the National Museum of Roman Art in Mérida; b Parish church on the banks of
the Loiola river; c “Horizontal skyscraper” L’Illa Diagonal in Barcelona; d Los Angeles Cathedral;
e Kursaal in San Sebastian
almost simultaneously to a scale of 1:100 in lime wood and paper. The sixth, of the
same size, was carried out in 1994 as a test and used industrial foam and wood (Salles
2006).
As we can see, all but the second used wood and secondarily paper. For Gehry it
is clear that the models are a way of working with successive approximations to the
final form, in which wood and its derivative paper are the ideal materials. But they
are also simultaneously a way to compete and to present their work to their clients
without falling into the temptation of mannerist finishes, simply representing that
which is essential.
When another Pritzker Architecture Prize winner, Álvaro Siza, collaborated with
Gehry on a project that was not built—a new library for the Art Center College of
Design in Pasadena (California), they worked with wooden models. In the Fig. 18a
we see both discussing a model and in the Fig. 18b, the existing building with its
Fig. 18 a, b Álvaro Siza and Frank Gehry working together on the Art Center College of Design
enlargement (unbuilt) and the model of The Art Center and its enlargement on the right. Unpublished
photographs are courtesy of M. Blanco Longueira (2002); c, d Two views of the Richard Meier
Model Museum, in which wood models in natural finish stand out. Photographs: © Richard Meier
& Partners Architects
rectangular shapes and the new library with its mainly curvaceous shapes, that are
perfectly distinguishable.15
The importance that some great architects have given to models is also clear with
the example of Richard Meier, winner of the Pritzker Prize for Architecture in 1984,
who in 2014 inaugurated his own museum under the name Richard Meier Model
Museum, in Jersey City (Fig. 18c, d). It exhibits a large number of models made by
his professional studio in an area of about 1400 m2 . Although the museum houses
all kinds of architectural material, sculptures, paintings and books from more than
five decades of activity, as its name suggests, it emphasizes scale models. Despite
the fact that they are made with many materials, those made entirely or substantially
out of wood stand out due to their number and importance.
15 The 21 February 2002 the President of the institution announced that «Frank Gehry, Alvaro Siza,
HsinMing Fung and Craig Hodgetts have been chosen to plan a series of “significant structures”»
(http://www2.artcenter.edu/news/wire/02-02-21.html, last access: 03/15/2018).
9 Some Unique Uses: From Palladio to “Plans-Reliefs”
9.1 Scenography as a Distorted Model of Reality
We are aware of the perhaps fundamental problem of classical scenography: to place

in an reduced space essentially in terms of its depth the proscenium, a plausible
three-dimensional representation of the reality in which the play is performed. For
example, the city of Thebes in Oedipus Rex, Sophocles’ great tragedy. Much has
and will continue to be written about the importance of the theatre, today a World
Heritage Site, designed by Andrea Palladio and in which architects of the stature of
Vincenzo Scamozzi collaborated, especially after Palladio’s death. As we know, it is
in fact a Roman theatre, like the ones the author had studied and drawn, but for the
first time it was not only closed but also roofed, turning it into an enclosed area, thus
establishing the model of those that would be built thereafter.
From our point of view, it would have been very important to have the wooden
model he had previously made to convince the members of the Olympic Society he
had created and the possible patrons of the city. But unfortunately it has disappeared,
although instead he left us a much more important wooden model, the proscenium
of the theatre itself. Although it was built for a specific site, it was not dismantled
and became one of the main elements of the site and one of its main architectural
attractions (Fig. 19).
In fact, it is not a model of an existing building, nor one that could be built, as it is
an example of what has been called architecture in perspective or false perspective
that precedes the works of Bramante, such as the false choir of Santa Maria presso
San Satiro of Milan (1480–82) or also of Borromini, such as the Palazzo Spada
(1652–53). The fact that no model of Palladio has been preserved underlines the
importance of this work.
Wooden models have been used in modern theatrical scenography as a relatively
simple way of checking the suitability of the sets they represent. In Fig. 20a, b, we can
Fig. 19 Andrea Palladio y Vincenzo Scamozzi. Proscenium of the Olympic Theatre of Vicenza.
Photograph: Didier Descouens via Wikimedia Commons (CC BY-SA 4.0)
Fig. 20 a, b Wooden model for The Return of Ulysses; c, d Wooden model for Lulu. Both from
the William Kentridge exhibition. Enough and more than enough. Museo Reina Sofía. Author’s
photographs (2018)
see the set for the 1998 opera The Return of Ulysses, designed by William Kentridge,
the South African artist who won the Princess of Asturias Award for the Arts in
2017, in collaboration with the Handspring Puppet Company. The scale model for
the opera Lulu (2013–2014), from the same creator in collaboration with Sabine
Theunissen, can be seen below, which in addition to wood uses other materials such
as agglomerates, paper, photocopies and cardboard (Fig. 20c, d).
9.2 “Plans-Reliefs” or City Models
The scale models of fortified cities and castles, also known as “relief plans” and
“bas-reliefs”, date back to the 16th century and were used for military tactical pur-
poses, but also for the intellectual interest of their rulers in getting to know the main
cities of European countries. As Marino Viganò points out, “The 18th century collec-
tions of the French, Spanish and Neapolitan Bourbons «private» models with public
use—reveal the proper meaning of this court collecting passion, between military
Fig. 21 a Wooden model of Cádiz ordered by King Carlos III; b Detail of the cathedral. Pho-
tographs: Cádizpedia (CC BY-SA 3.0)
usefulness and symbolism” (Viganò 2007, p. 219). King Carlos III of Spain, who
ruled the Duchies of Parma and Plasencia and later the kingdoms of Naples and
Sicily, ordered the construction of models of all the cities of military importance and
Spanish fortifications.
The one in Cádiz was the only one that was completed between 1777 and 1779.
It was made from a very recent plan for the time, from 1749, drawn by Ignacio Sala.
Its author was an infantry lieutenant colonel and engineer, Alfonso Ximénez who,
with the help of cabinetmakers, made it in four parts for later assembly (Fig. 21a).
It is made of Flanders pine wood, finished in carved cedar and with the streets in
acana wood (Manilkara albescens), native of Cuba and from which latex is extracted,
with ebony balconies. However, the cathedral, still in the planning stage at the time,
was made of ivory (Fig. 21b), as well as other important buildings of the time. As
is logical given its purpose, the military buildings stand out, from the walls to the
Castle of La Villa, now disappeared. Given that many parts of the complex have been
lost, it is a notable example of the importance of the conservation of these works
for the knowledge of the evolution of cities that would otherwise be very difficult
to know. Thanks to the restoration done by Pena López16 in the second half of the
20th century, we can admire it today as a key piece of urban iconography (Martínez
1999), probably as it was almost 250 years ago.
Artillery Lieutenant Colonel León Gil de Palacio took almost twenty-three months
to make the model known as the Madrid Model, commissioned by King Fernando VII
of Spain and completed in 1830. It is located in the History Museum of Madrid and
is composed of ten irregular blocks of poplar wood with a surface area of 18.34 m2 .
It details the central area of the city in three dimensions, with a perimeter of 12.5 km,
defined by the wall that was not demolished until 1869 to develop the new urban
expansion area (Fig. 22).
16 Master of the carpentry workshop of the School of Beaux-Arts of Cadiz.

Fig. 22 Model of Madrid by León Gil de Palacio, 1830. History Museum of Madrid. Author’s pho-
tographs (2018): a The Royal Palace; b San Francisco el Grande and its convent, now disappeared;
c The Plaza Mayor; d The Bullring of the Puerta de Alcalá
For the details, in addition to poplar wood, cork, sawdust, paper and cardboard
painted with glass and lead were used. It is extraordinarily precise when picking up
the altimetry, the details, the apple yards and even the clotheslines on the banks of the
Manzanares River. It is made to a scale of 1:816 and measures 5.20 × 3.50 m. Much
of its historical importance lies in the fact that it represents the city in 3D before
the controversial confiscation of Mendizábal in 1835–1836, when a large number of
buildings belonging to the Church, such as monasteries and convents, disappeared
in a short time.
The basic plans of the model—which can be seen in the same room - are usually
considered, among other older ones, the Geometric Plan of Madrid from 1785,
made with burin and etching by Tomás López de Vargas Machuca to a scale of
approximately 1:5500 and, closer in time, the Plan of Madrid divided into 10 barracks,
of 1812, published by the geographer Juan Lopez and also made with burin and
etching to a scale of approximately 1:7800.17
In the same room there is another maple wood model, this time contemporary,
erected from the lithograph of the facsimile plane of 1881 also exhibited, enti-
tled MANTVA CARPETANORVM. SIVE MATRITVM VRBS REGIA, known as
Topographia de la Villa de Madrid, described by D. Pedro Teixeira, of 1656 and
17 http://bibliotecadigital.rah.es/dgbrah/es/consulta/registro.cmd?id=12892.
representing the city at the time of Felipe IV to a scale of approximately 1:1600.18

The map of the Portuguese cartographer engraved in Antwerp is made in a mili-
tary perspective,19 which allowed for a sufficiently reliable construction of the 3D
model, which would have been impossible without this map. Madrid was defined by
the same fence that appears in the great model of Leon Gil de Palacio and you can
see the old medieval roads, converted into the main streets of the new capital of the
Spanish empire, established in 1561.
An important precursor of this last plan is the one known as La Villa de Madrid
Corte de los Reyes Católicos de España.20 Although it has no data, its scale has been
calculated to approximately 1:1600, in other words, the same as the previous one,
and because of the buildings that appear or are missing, it is believed to have been
built about 1622, possibly by the draftsman and illuminator Antonio Marcelli. It was
widely distributed and possibly used by Pedro Teixeira for his own one.
City models in large-scale continue to be made in urban planning studies, firstly
to visualise together the real state of a city or of a major new development, and
secondly, to enable citizens to better understand the environment in which they live
thanks to their public exhibition.
There are large scale models of cities—authentic urban scenographies—made
entirely or largely out of wood in cities like Moscow, which has one in the Royal
Radisson Hotel, or the so-called The Panorama of the City of New York, built for the
1964 World’s Fair—located in one of its old pavilions and in constant updating—and
many more, but perhaps the one that stands out for its purpose is the Model of
Havana, made by the model maker and topographer Orlando Martorell in solid
wood, apparently cedar, with plywood to a scale of 1:1000. It was carried out with
a clearly urban purpose as a way of controlling its growth, for which it includes the
most recent works and even those under construction. Although it should remain as
it is, as a witness of an era, given that if the political regime were to change, the city
would quickly become unrecognizable.
9.3 Photography as Trompe L’oeil
The photography of scale models is a speciality of great interest in architecture

because it allows multiple points of view to be generated, both in linear perspective
and in an axonometric approach, which facilitates the spatial understanding of the
building.
18 http://www.bibliotecavirtualmadrid.org/bvmadrid_publicacion/i18n/consulta/registro.cmd?id=
725.
19 According to the digital library of the Community of Madrid, cavalier perspective, which is wrong.
20 http://www.bne.es/es/Micrositios/Exposiciones/BNE300/resources/img/300anos_160_01_gr.
jpg.
Fig. 23 a, b Canonical view of the axonometric model of the House X and another view. Pho-
tographs: © 2015 Eisenman Architects; c Lithography with axonometry in colour, Maison Partic-
ulière of Theo van Doesburg, Netherlands Institute for Cultural Heritage
Peter Eisenman took to the limit an axonometric vision of an architectural model

with his well-known House X, a project from 1975 which was not built,21 in which
photography can only be taken from a single point of view (Fig. 23a). The result,
drawn up by the architect himself, uses mainly wood and its derivatives, such as
different types of cardboard, and can be compared with the image of the 1925 Maison
Particulière project by Theo van Doesburg and Cornelius van Eesteren (Fig. 23c),
since it is clear that Eisenman intended the same effect.
But Eisenman’s model is an exceptional case or rather, an academic investigation.
More common is the making of models in which the interior space, for example, an
auditorium or a place of worship, is as important or even more so than the exterior. In
these cases a model is made in which part of the façades and the roof can be removed,
or it is directly built openly, as in the examples of Rafael Moneo for the cathedral of
Los Angeles or the Kursaal of San Sebastian, seen above.
9.4 The Model of the Great Hall of the Elbphilharmonie
More and more models are being used for different or complementary purposes, as
in the case of the scale model of the Grand Hall or main auditorium of the Elbphil-
harmonie in Hamburg, by the architects Herzog & de Meuron. This large 1:10 scale
model was made of wood and has more than 10,000 pieces with an approximate
weight of 4.5 tons. It was carried out by Japanese acoustics expert Yasuhisa Toyota
to study the acoustics of the main hall.
Although this was the main motivation of the model, it was later found to be of
great use during the long years of construction of the building: it was the centre
of attraction of the Elbphilharmonie Pavillion, located near the building and which
21 Project for Arnold Aronoff in Bloomfield Hills, Michigan (https://eisenmanarchitects.com/House-
X-1975).
Fig. 24 a The Elbphilharmonie Pavillion with the building under construction in the background;
b Partial view from its centre of the 1:10 scale model of the Grand Hall. Author’s photographs
(2016)
helped explain it to visitors (Fig. 24a). Located on the third floor, the view was
accessed through a hole in the centre of the layout room: from this standpoint the
viewer obtained a 360° view, as a virtual-real panorama (Fig. 24b): “Looking at the
model, one can get a good spatial impression of the concert hall” (Kämpermann
2014, p. 15).
10 New Technologies in the Manufacture of Wooden

Models
The difficulties inherent in the traditional and therefore artisan manufacturing of

wooden models have been practically eliminated with the new 3D manufacturing
processes using CNC machines, i.e. Computerised numerical control, which use
microprocessors to operate them and can be of two types: milling machines and
laser cutting machines.
Both systems actually work in 2D and cut or print plates in the same way as the
old cork contours and cut with knives. Now, a simple collection of 2D cut-out pieces
can create a three-dimensional illusion as in Rafael Moneo’s model for the Diestre
Transformer Factory, 1965–1967 in Zaragoza, Spain (Fig. 25a).
It is also possible to make three-dimensional meshes from pieces cut out in 2D,
such as the model of German architect Jürgen Mayer for the Metropol Parasol build-
ing (2005–2011), better known as “the mushroom”, in the Plaza de la Encarnación
in Seville, for the new municipal market (Fig. 25b, c).
This is a particularly significant example, since the very need to assemble the
elementary pieces of the model could have led to the creation of the building using
the same technique. If so, the resulting work can be understood as a gigantic, almost
literal extension of the scale model.
Obviously, the models are just one example of what this technology can accom-
plish, as furniture and even complex architectural elements for interiors and exteriors
Fig. 25 a Model for Rafael Moneo’s Diestre transformer factory. Photograph by the author (2014);
b, c Metropol Parasol building by Jürgen Mayer. Author’s photographs (2016); d, e Sculptural
element for the entrance hall to the ARUP Offices in Madrid.
Photographs: © Controlmad
can also be manufactured, as will surely be discussed in other chapters of this mono-
graphic work. Here I will simply highlight a curious example: that of the sculptural
element designed for the entrance hall of the ARUP offices in Madrid (Fig. 25d, e),
which could be said to be a 1:1 scale model. Made from a standard 25 millimetres
pine plywood board, it was designed in parametric 3D CAD and manufactured with
a three-axis CNC milling machine.22 The result, which at a glance might appear to
be moulded or carved, is actually a kind of ‘ground’ made up of successive layers
finished with ecological protective wax.
It is also possible to use 3D printers to make wood models, thanks to the new
materials being experimented with, such as the LayWoo-D3 filament, which is made
out of 40% recycled wood, and the rest are binding polymers. The result can be
sanded or even painted much more easily than other materials such as plastics.
11 Conclusions
The wood models reached their maximum development in the Renaissance, in which
they basically served three purposes: firstly, to serve as a design tool for the architect,
that is to say, a personal conceptual use; secondly, to be presented to the competitions
and to convince the corresponding patrons or clients, as in the case of Brunelleschi
for the Florence Duomo; and thirdly, to serve as a guide for the workers of the work.
In the centuries that followed, these aims did not change much, if perhaps the
former was accentuated, as Sir Christopher Wren and Sir John Soane stated. In
modern times the best examples in this sense are Álvaro Siza and Frank Gehry,
among many others.
In addition, the wooden models were not only used and are used to represent build-
ings, but have also been widely used for urban representations and, more specifically,
cities, where they either fix in time data of great archaeological value that would be
unknown today or allow the control of the development of these cities, promoting
understanding between those who live in them or those who visit them.
Finally, new technologies allow new approaches and utilities, making it easier to
make them by hand-even with 3D printers-or substantially minimizing them. Golvin
and Vergnieux affirmed, in relation to the constructions of the past, that the restored
image constitutes an important and indispensable stage between the scientific analysis
of the documentation and the production of the research instrument which is the
electronic mockup of the ancient buildings (Golvin and Vergnieux 1997, p. 40).
In this sense, I would like to suggest that the new technologies could be used,
opening new lines of research, to generate three-dimensional digital models of the
old models that currently exist, either by means of 3D laser scanners or photographic
scans with programs such as Photomodeler, Photoscan or ReCap Photo. As an exam-
ple, the photographic scan of the model of the Santa Engracia tank (Sect. 6, Fig. 11c),
chosen for its special characteristics and manufacturing difficulty, has been carried
out specifically for this chapter. All this work could lead to a European project for
22 Sculptural element for the entrance hall of the ARUP Offices, Madrid. Design and production:
Controlmad. Sergio Alonso del Campo and Diego García Cuevas. Contributors: Víctor Armas and
Aitor Jaurrieta. The software used was Rhinoceros, Rhinonest and Grasshopper. See http://www.
tectonica.es/espacio-tectonica/pdf/controlmad.pdf.
the creation of a universal and interactive database of the great models that have laid
the foundations of architecture and engineering as we know them today.
This would make it possible for researchers and future architecture professionals
to get to know multiple points of view and not only those of some photographs of
the original model, deepening in their study. In addition, it would allow printing
on the increasingly advanced 3D printers already available at most universities and
would somehow restore the original meaning of the corresponding models. Likewise,
the mere existence of these archives would in some way guarantee the preservation
for the future of models that already form part of the history of architecture and
engineering.
References
Adam JP (1997) Dibujos y maquetas: la concepción arquitectónica antigua. In: Azara P (ed) Las
casas del alma. Maquetas arquitectónicas de la antigüedad (5.500 a.C./300 d.C). Diputación de
Barcelona, Barcelona, pp 25–33
Alberti LB (1582) Los diez libros de Architectura de Leon Baptista Alberto. Traducidos de Latin
en Romance. Alonso Gomez impressor de su. Magestad, Madrid
Alimonti A (2018) La più grande chiesa del mondo. Abitare (article online: http://www.abitare.it/it/
habitat/patrimonio-storico/2018/03/25/san-pietro-storia-chiesa-piu-grande-mondo/. Last Access
15 Apr 2018
Amador J, Rosell C (1864) Historia de la Villa y Corte de Madrid, vol 4. Establecimiento tipográfico
de M, López de la Hoya, Madrid
Blasco B (2016) Invención, traza y proyecto. El proceso arquitectónico en las Obras Reales de la
Casa de Austria. Artigrama 31:279–306
Bonet A (1960) Velázquez, arquitecto y decorador. Archivo Español de Arte 130(33):215–250
Darley G (2016) Sir John Soane’s private apartments are a public treasure. Apollo (arti-
cle online: https://www.apollo-magazine.com/sir-john-soanes-private-apartments-are-a-public-
treasure/. Last Access 20 Apr 2018
Dunn N (2010) Architectural Modelmaking. Laurence King Publishing Limited, London. Spanish
edition: Dunn N (2010) Maquetas de arquitectura. Medios, tipos, aplicación (trans: Melús C).
Blume, Barcelona
Esparza V (2006) Maquetas en el mundo antiguo: entrevista a Pedro Azara. DC Revista de crítica
arquitectónica 15–16:55–62
Estella A (ed) (2005) 250 años de la Plaza. La plaza Mayor de Salamanca, vol vol., Caja Duero,
Salamanca, p 3
Franco JA (2017) The search for three-dimensionality in professional practice. From the patesi
Gudea to the Elbphilharmonie in Hamburg. In: Cardone V (ed) Territories and Frontiers of
Representation. Roma Gangemi Editore International, Roma, pp 1551–1557
Fuster A (2006) Casa para Joséphine Baker: Adolf Loos (1927). DC. Revista de crítica arquitectónica
15–16:110–111
Gentil JM (1998) Traza y Modelo en el Renacimiento. Instituto Universitario de Ciencias de la
Construcción—Universidad de Sevilla, Sevilla
Golvin JC, Vergnieux R (1997) Primer análisis para la elaboración de una maqueta electrónica del
santuario del gran templo de Atón en Amarna. In: Azara P (ed) Las casas del alma. Maquetas
arquitectónicas de la antigüedad (5.500 a.C./300 d.C). Diputación de Barcelona, Barcelona, pp
35–40
Kämpermann M, Hotes K (eds) (2014) Elbphilharmonie Hamburg. HamburgMusik gGmbH and
Elbphilharmonie und Laeiszhalle Betriebsgesellschaft, Hamburg
Kolbert E (2011) Enter the anthropocene—age of man. Nat Geograp 219(3):60–85

Lopera A (2017) Maqueta del depósito elevado nº 1 del Canal de Isabel II (Madrid). In: Navascués
P Revuelta B (2017), pp 180–185
López F (1985) Plazas de toros de Madrid y otros lugares donde se corrieron. Avapiés, Madrid
Millon HA (1994) Models in reanissance architecture. In: Millon HA (ed) Italian renaissance
architecture. From Brunelleschi to Michelangelo, Thames and Hudson, London, pp 19–74
Millon HA (1999) The triumph of the baroque: architecture in Europe 1600–1750. Thames and
Hudson, London
Navascués P, Revuelta B (2017) Maquetas y modelos históricos. Ingeniería y Construcción, Fun-
dación Juanelo Turriano, Madrid
Sáenz-López S, Aragón S (2013) 101 Obras Maestras: ciencia y arte en los museos y bibliotecas
de Madrid. FECYT, Madrid
Salles R (2006) Museo Guggenheim Bilbao: Frank O. Gehry (1999–2006). DC. Revista de crítica
Samaniego F (1994) Toda la arquitectura del Renacimiento italiano, en una exposición única
en Venecia. El País (article online: https://elpais.com/diario/1994/03/31/cultura/765064801_
850215.html. Last Access 17 Apr 2018
Shklovsky V, Sheldon R (2005) Knight’s move. Dalkey Archive Press, Normal
Viganò M (2007) Colecciones de modelos de plazas fuertes de los Borbones de Francia, España y
Nápoles en el siglo XVIII. BSAA Arte 72–73:219–243
Weizman I (2017) Thing Rights. An architectural re-enactment of the Josephine Baker
House by Adolf Loos (1928). ARQ (Santiago) 95:30–39. http://dx.doi.org/10.4067/S0717-
69962017000100030
Wilton-Ely J (1968) The Architectural Model: English Baroque. Apollo 88:250–259. Wilton-Ely J
(2006) La maqueta arquitectónica: Barroco inglés (trans: Caballero AJ). DC. Revista de crítica
Wilton-Ely J (1969) The architectural models of Sir John Soane: a Catalogue Raisonné. Architectural
History 12:5–38. Wilton-Ely J (2006) Los modelos arquitectónicos de Sir John Soane: un catálogo
(trans: Roca M). DC. Revista de crítica arquitectónica 15–16: 41–51
Wilton-Ely J (1976) The rise of the professional architect in England. In: Kostof S (ed) The Architect:
Chapters in the History of the Profession. Oxford University Press, New York, pp 180–208
The Wooden Models of the Vatican
Basilica by Antonio da Sangallo
and Michelangelo: Survey, Modelling
and Interpretation
Carlo Bianchini, Alfonso Ippolito and Luca J. Senatore
Abstract The role that models have played in the development of Renaissance
architectural thought is well known but the massive usage of digital modelling and
visualization tools often prevents us from fully appreciating their original role in
the conception and communication of architecture. Nowadays infact, bits and pix-
els have almost completely replaced wood, chalk or plaster. However, either tangible
(material) and intangible (digital) models still unquestionably represent an extremely
effective design tool. More than drawings, in fact, they show a strong ability to crys-
talize in an immediate and evident form the architect’s idea. Under this perspective,
material and digital models are not as distant form each other as it seems and actu-
ally live a sort of parallel life. They share in fact the modelling, namely that process
of abstraction establishing a biunivocal correspondence between the object and its
representation. This is why representations have been so important for the making of
architecture: exploiting their being models, architects do use them to journey across
space (and time occasionally) in order to understand the essence of architectural con-
cepts, designs and even buildings. In this framework, this paper summarizes some
results of the research developed by our group about Sangallo’s and Michelangelo’s
maquettes for the New St. Peter’s Basilica that have been surveyed in 2005/2006
and investigated since then. Our objective is to provide an insight about the work
and a discussion of some major issues encountered while experimenting the virtual
modelling of those historic models.
Keywords Material model · Antonio da Sangallo · Michelangelo

St. Peter Basilica · 3D modelling
C. Bianchini (B) · A. Ippolito · L. J. Senatore

Department of History, Representation and Restoration of Architecture,
e-mail: carlo.bianchini@uniroma1.it
A. Ippolito
e-mail: alfonso.ippolito@uniroma1.it
L. J. Senatore
e-mail: luca.senatore@uniroma1.it

322 C. Bianchini et al.
1 Introduction
The role that models have played in the development of Renaissance architectural
thought is well known (Briggs 1929; Ribichini 2007; Benedetti 1986, 1994, 2009).
Nowadays though, we have only a very limited perception of their original role in the
conception and communication of architecture as we currently use digital modeling
and visualization tools instead of wood, chalk or plaster.
Moreover, we can reasonably imagine that in the Renaissance context the material
models were to produce on the public the same effect that many of the most advanced
digital representations do on us.
We can find clues of this symbolic and communicative value in many of the
treatises and commentators of the time (i.e. Vasari’s words about the Sangallo’s
Model for the New St. Peter’s Basilica).1
However, renaissance material models unquestionably represent also an extremely
effective design tool. More than drawings, in fact, they show a strong ability to
crystalize in a material form, immediate and evident, the architect’s idea (Docci
2007; Silvan 1994).
The commitment lavished in the last years of his life by Antonio da Sangallo
the Young around St. Peter’s model appears under this light perfectly clear and
reasonable. While officially this extraordinary wooden model intended to present to
Pope Paolo III the project Sangallo had conceived for the new St. Peter Basilica, in
1 […] Ma tutto quello che Antonio fece di giovamento e d’utilità al mondo è nulla a paragone del
modello della venerandissima e stupendissima fabbrica di San Pietro di Roma. La quale, essendo
stata a principio ordinata da Bramante, egli con ordine nuovo e modo straordinario l’aggrandì e
riordinò, dandole proporzionata composizione e decoro, così nel tutto come ne’ membri, come si
può vedere nel modello fatto per mano d’Antonio d’Abaco suo creato, di legname, et interamente
finito; il quale modello, che diede ad Antonio nome grandissimo, con la pianta di tutto l’edifizio
sono stati dopo la morte d’Antonio Sangallo messi in istampa dal detto Antonio d’Abaco, il quale
ha voluto perciò mostrare quanta fusse la virtù del Sangallo, e che si conosca da ogni uomo il parere
di quell’architetto, essendo stati dati nuovi ordini in contrario da Michelagnolo Buonarroti; per la
quale riordinazione sono poi nate molte contese, come si dirà a suo luogo. Pareva a Michelagnolo,
et a molti altri ancora che hanno veduto il modello del Sangallo e quello che da lui fu messo in
opera, che il componimento d’Antonio venisse troppo sminuzzato dai risalti e dai membri, che
sono piccoli, sì come anco sono le colonne, archi sopra archi e cornici sopra cornici. Oltre ciò pare
che non piaccia che i due campanili che vi faceva, le quattro tribune piccole e la cupola maggiore,
avessino quel finimento o vero ghirlanda di colonne molte e piccole; e parimente non piacevano
molto, e non piacciono, quelle tante aguglie che vi sono per finimento, parendo che in ciò detto
modello immiti più la maniera et opera tedesca, che l’antica e buona che oggi osservano gl’architetti
migliori. Finito dall’Abaco tutti i detti modelli, poco dopo la morte d’Antonio, si trovò che detto
modello di San Pietro costò (quanto apartiene solamente all’opere de’ legnaiuoli e legname) scudi
quattromilacentoottantaquattro. Nel che fare Antonio Abaco, che n’ebbe cura, si portò molto bene,
essendo molto intendente delle cose d’architettura, come ne dimostra il suo libro stampato delle
cose di Roma, che è bellissimo. Il qual modello, che si truova oggi in San Piero nella cappella
maggiore, è lungo palmi trentacinque e largo 26 et alto palmi venti e mezzo. Onde sarebbe venuta
l’opera, secondo questo modello, lunga palmi 1040, cioè canne 104, e larga palmi 360, che sono
canne 36, perciò che secondo la misura de’ muratori, la canna che corre a Roma è dieci palmi
(Vasari 1568).
The Wooden Models of the Vatican Basilica … 323
Fig. 1 Antonio da
Sangallo’s wooden model for
the New St. Peter’s Basilica,
Main Façade (before
restoration)
the designer’s mind it had to serve as a univocal document of his project in order to
shield it against future heretical interpretations or changes.
The object itself (Figs. 1, 2 and 3), manufactured during the fourth and fifth
decades of XVI Century, is actually a small building: both for the time and money
spent on it (more than seven years and nearly 5000 scudi—enough to build a real
church) as well as for its dimensions. The model is 7.36 m long, 6.02 m wide and
4.68 m high at dome summit; the bell towers are instead 4.56 m high; the overall
scale is about 1:30.
The story, as we all know, has taken a different, unpredictable twist with
the appointment of Michelangelo as Architetto della Fabbrica at Sangallo’s
death. Michelangelo, in fact, abandoned almost completely Sangallo’s idea and
actually restored much of the initial Bramante’s concept (Wittkower and Brunetti
1964; Tessari and Bruschi 1996; Ciulich 1983; Argan 2015). For several reasons,
however, Michelangelo as well found himself in a condition similar to that of his
predecessor. On the one hand, he could not afford to neglect completely Sangallo’s
model (and fabric of course), both as a material object and as a virtual representa-
tion of his predecessor’s project. On the other hand, getting himself on in years and
probably realizing that he would not be able to see the Great Basilica and its Dome
finished, he was pressed by the same need to firm his idea as much as possible and
to pass it on untouched (and untouchable in his hope…) to his successors.
Fig. 2 Antonio da
Sangallo’s wooden model for
the New St. Peter’s Basilica,
detail of the main dome
These urgencies, along with many others of course, led to the manufacturing of at
least two models of Michelangelo’s St. Peter: one embracing the entire basilica and
another regarding only the dome. The first, lost, probably built just before receiving
the appointment from Pope Paul III with the famous Motu-proprio of January 1st
1547, represented the Fabbrica in a scale close to 1:90. We have an idea of this
maquette from the painting Michelangelo and Paolo IV by Passignano where the
architect illustrates his project to the Pope (Fig. 4). The other, initiated probably in
1558, is still conserved in an octagon of the Vatican Basilica almost facing the huge
Sangallo’s wooden model (Figs. 5, 6 and 7).
Both of them have been studied since 2005 (Guidi and Bianchini 2007) by our
research group and this paper provides an insight about this work and a discussion
of some major issues encountered while experimenting virtual modelling on those
historic models.
Fig. 3 Antonio da Sangallo’s wooden model for the New St. Peter’s Basilica, Side view with
general dimensions
2 Tangible Versus Intangible
The process of 3D modelling has been radically changing since the appearance of
digital tools (Centofanti et al. 2014; Bianchini et al. 2017). However, tangible and
intangible models are not so distant form each other as it seems. They share in fact
the modelling, that is to say that process of abstraction that establishes a biunivocal
correspondence between the object and its representation. This is why representations
have been (and still are) so important for the “making” of architecture: exploiting their
being models, architects do use them to journey across space (and time occasionally)
in order to understand the essence of architectural concepts, designs and buildings.
Moreover, material and virtual models seem to live a sort of parallel life.
In a digital environment in fact, models are three-dimensional exactly as in the
real world: any modelling software creates in fact an entire 3D virtual space that,
initially empty, is systematically populated by elements that together build the virtual
model (Figs. 7 and 8).
Hence, the digital environment created by 3D modelling software provides an
actual spatial framework for all the following constructions. Therefore, our monitor
becomes the interface between two parallel universes: the real one on this side, the
virtual instead on the other. However, the screen represents also the ultimate limit of
our exploration, the border that we can never cross and that causes our interaction
with all virtual entities (navigation, modelling, and manipulation) to occur only using
special, digital tools (Fig. 9).
Fig. 4 Domenico Cresti

detto il Passignano,
Michelangelo and Paolo IV
In this framework, modelling means essentially constructing all the elements that
together do compose the final object. Geometric construction, obviously, as any entity
that is or may be created is geometric. Nevertheless, a construction that is quite alike
to the processes typically adopted in a building yard or, in our case, by a team of
artisans manufacturing a wooden artefact.
3 Sangallo’s and Michelangelo’s Models: Capturing, Study

and Reconstruction
The research developed by our group on both wooden models has entailed many
different activities (Bianchini 2007; Bianchini and Ippolito 2009; Senatore 2009).
First, a systematic documentation and survey campaign aiming at collecting an as
much as possible complete set of data and information. Then, an analytical study
Fig. 5 Michelangelo’s
wooden model of the Dome
of the New St. Peter’s
Basilica
about the geometric and morphological features of both objects oriented toward the
construction of 3D models as virtual replicas of the corresponding originals.
In order to achieve the above-mentioned objectives, the entire process has entailed
research and investigation of archival material, historical drawings, previous recon-
structions, etc.; a systematic and detailed photographic campaign; an integrated 3D
survey (total station, 3D laser scanner).
The final 3D models are actually the result of the interleaved elaboration of all
these data.
Before starting with the acquisition phase, an accurate planning of activities has led
to define accurately the most effective procedure vis-à-vis the research objectives.
In this context, given the peculiarity of the artefacts, we decided to use multiple,
integrated capturing techniques.
Yet another element has heavily influenced the development of our work: the
location of the models. Stored in the octagon of San Girolamo (one of the few
originally built by Sangallo himself), relevant portions of both models are too close to
Basilica, details of the
external tambour and dome
the walls for an easy surveying. In particular, only two sides of Sangallo’s maquette2
are comfortably accessible (even if the main façade cannot be captured entirely from
any point), while the model of Michelangelo is currently positioned with the entire
external façade close to a wall.
Materialized on the spot the network of reference points, we used the following
equipment during the data acquisition phase:
(1) Leica HDS 3000 3D laser scanner (operating uncertainty +5 mm.);
(2) Leica TCR 1201 R300 total station equipped with a laser distance meter;
(3) Minolta 3D Vi-9i triangulation laser scanner (operating uncertainty of less than
1 mm.);
(4) Canon EOS digital camera.
2 Such a physical model, that is presently hosted in one of St. Peter’s Octagons (one of the few
originally constructed by Sangallo himself), has therefore a size which lays midway between a
small object and a building. Even if its present placing (by now practically permanent) suits at best
preservation needs, it has deeply affected our scanning campaign: the vaulted space of the octagon
is in fact almost completely occupied by the model so that, except for the two access to the room,
there is on average only one meter of space available around the object.
Basilica, detail of the dome
crowning
Three were the stations used for the construction of a topographic polygonal while
the scanning has entailed seven different shooting positions, six at ground level and
one on a mobile scaffold (for the upper part of the dome). Two the sample spacing
used: a general one 5 × 5 mm and a denser one with a 2 × 2 mm sampling (Figs. 10,
11, 12 and 13).
Finally, we captured those details that requested a higher accuracy (capitals, dec-
orations, etc.) using the Minolta Vi-9i scanner.
Before proceeding with the 3D modelling phase, some 2D information was
extracted from the general numeric model, that is to say from the comprehensive
point cloud.
We started generating some horizontal and radial sections, both on the 2 × 2 mm
and the 5 × 5 mm models.
For Michelangelo’s maquette, the considered horizontal profiles lay on three
planes that cut the dome in the area without windows; from the interpolation of
these sections we could reconstruct the directrix circumference of the dome. The
profiles traced on the 5 × 5 model are also located near the previous ones. In this
case tough, the curves embrace the entire model (half dome). These curves build a
Fig. 8 The digital reconstruction of Antonio da Sangallo’s design for the New St. Peter’s Basilica,
the general wireframe model
set of three ideally concentric circumferences positioned at different heights. The

vertical line that best interpolates their centres was therefore assumed as the vertical
axis of the dome and thus used as support for the beam of radial planes sectioning
the rib vaults (Figs. 14 and 15).
These geometric results as well as a direct re-elaboration of original point clouds
actually fed the real 3D modelling phase aiming at the construction of surfaces
(Fig. 16).
The elaborations, still in progress for the Michelangelo’s model and that we
present here only as prototypes, have been instead the core of the work developed on
Sangallo’s maquette. In this case in fact, our approach has been much more modular,
something we could now define as sort of BIM-ready one. Since the very first acqui-
sitions, in fact, we started the 3D modelling project aiming at developing prototypes
of those modular elements creating a compositional and decorative blueprint of the
external and internal surfaces.
This turned out to be very important because it allowed us to: establish the basic
geometries of each architectural element; check their proportions (by comparing
them with Sangallo’s drawings and with those of the most important treatise writers
of the Renaissance); establish the level of detail needed for the specifics of the final
Fig. 9 The digital reconstruction of Michelangelo’s design for the Dome of the New St. Peter’s
Basilica, 3D model prototypes
model we wanted to build; optimise the surfaces created by the modelling software
and, finally, estimate the overall size of the 3D model by establishing a procedure
that allowed us to manage either the whole model or each of its parts.
We solved the prickly problem of cataloguing and management of all this infor-
mation by defining a database for each element of the model: these included also two-
dimensional plans, elevations and sections as well as a three-dimensional NURBS
object (Figs. 17 and 18).
It would be redundant in the context of this paper to discuss more in detail the
whole Sangallo’s project: for that we invite to refer to Docci (2007), Bianchini (2007)
and Ribichini (2007).
Nevertheless, we must underline once more how the 3D model represented the
final product of a very careful analytic process and, actually, our best interpretation
of the artefact at that time. Interpretation that besides was choosing to provide a
reading of Sangallo’s maquette in terms of design concept (directly referring to the
Fig. 10 Michelangelo’s wooden model of the Dome of the New St. Peter’s Basilica, the 3D cap-
turing process
compositional, formal and decorative framework the author had prefigured for the
New St. Peter) instead of considering the wooden model as an object in itself (that
is to say with its own shape, technology, matter) (Figs. 19 and 20).
However, over and above all this, that project achieved a result that, not immedi-
ately evident to us, on the contrary we presently consider one of the most valuable:
show to the entire community of scholars, even to the most unwilling towards the
digital tools, the potential impact of 3D modelling for its ability to provide a virtual
environment in which explore, verify and display any reconstructive hypothesis.
On this very basis, some years after our “first” release, a new version of the
3D model of Sangallo’s maquette was elaborated (Senatore 2009) in order to take
into account some new grounded conjectures that allowed to better understand the
refined illumination concept designed for the inner Basilica by Antonio da Sangallo
(Benedetti 2009) (Figs. 21 and 22).
Fig. 11 Michelangelo’s wooden model of the Dome of the New St. Peter’s Basilica, texturized
views (RGB from camera) of the captured numeric model (point cloud)
Basilica, texturized views
(RGB from reflectance) of
the captured numeric model
(point cloud)
Fig. 13 Antonio da Sangallo’s wooden model for the New St. Peter’s Basilica, texturized views
(RGB from camera) of the captured numeric model (point cloud)
Fig. 14 Michelangelo’s wooden model of the Dome of the New St. Peter’s Basilica, 2D investiga-
tion of radial geometry
Fig. 15 Michelangelo’s wooden model of the Dome of the New St. Peter’s Basilica, horizontal and
vertical, radial 2D sections
Fig. 16 Michelangelo’s wooden model of the Dome of the New St. Peter’s Basilica, from the point
cloud to the NURBS surfaces
a sample of the 2D analysis of architectural basic elements
the modular re-building of the model
transversal section and main façade elevation as 2D elaboration form the 3D general model
texturized rendered view of part of the ground level
Fig. 21 The revised digital reconstruction of Antonio da Sangallo’s design for the New St. Peter’s
Basilica, texturized rendered elevation of the building side (in red the new elements)
Fig. 22 The revised digital reconstruction of Antonio da Sangallo’s design for the New St. Peter’s
Basilica, 2D transversal section (in red the new elements)
4 Conclusions
Renaissance models undoubtedly represent a tremendous tool for architectural inves-

tigation as well as tangible deliverable of designers’ program. Nowadays they are
indeed extraordinary masterpieces of artisans’ ability and skill and, for us, pieces of
cultural heritage to preserve and pass on to next generations.
Nevertheless, they are indeed also 3D models, not far in terms of meaning and
impact from the virtual ones we experiments every day.
The study presented has enlightened these different aspects and the peculiarities of
their documentation, surveying and, finally, digital interpretation and reconstruction.
Acknowledgements We would like to thank the Veneranda Fabbrica di San Pietro: first its respon-
sible, Card. Angelo Comastri, for his high support and then Luca Virgilio and Pietro Zander for
their constant help and advice.
Furthermore, we must underline that the Sangallo’s research has been part of a wider activity
commissioned by the Veneranda Fabbrica in 2005 for the making of a multimedia product to be
shown during the Petros Eni Exhibition. The Principal Investigator of this research was Prof. Luca
Ribichini.
The activities focusing on Michelangelo’s model, instead, were part of an investigation project
coordinated by prof. Paolo Rocchi that aimed at studying the Great Vatican Basilica from different,
complementary points of view.
Finally, all the equipment used has been provided by the Department of History, Representation
and Restoration of Architecture, Sapienza—Università di Roma.
References
Argan GC (2015) Michelangelo. Giunti

Benedetti S (1986) Il modello per il San Pietro in Vaticano di Antonio da Sangallo il Giovane.
Centro di Studi per la storia dell’architettura
Benedetti S (1994) Oltre l’antico e il gotico: il profilo della cupola vaticana di Antonio da Sangallo
il giovane
Benedetti S (2009) Antonio da Sangallo il Giovane: Il grande modello per il San Pietro in Vaticano.
Gangemi
Bianchini C (2007) From reality to virtuality (and back): the wooden model by Antonio da Sangallo
for the new St. Peter’s in Vatican. Disegnare idee immagini 34:36–49
Bianchini C, Ippolito A (2009) Il Modello di Michelangelo per la Grande Cupola Vaticana: doc-
umentazione e prime analisi. In: Rocchi P (ed) Nuove ricerche sulla Gran Cupola del Tempio
Vaticano, pp 278–289
Bianchini C, Viscogliosi A, Aglietti A (2017) Innovative digital heuristic approaches in architectural
historical research. In: 2017 21st international conference on information visualisation (IV). IEEE,
pp 444–449
Briggs MS (1929) Architectural models-I. Burlington Mag Connoisseurs 54(313):174–183
Centofanti M, Brusaporci S, Lucchese V (2014) Architectural heritage and 3D models. In: Com-
putational modeling of objects presented in images. Springer, Cham, pp. 31–49
Ciulich LB (1983) Nuovi documenti su Michelangelo architetto maggiore di San Pietro. Rinasci-
mento 23:173
Docci M (2007) La basilica vaticana non costruita. L’opera di Antonio da Sangallo. Disegnare Idee
Immagini, 34:24–35
Guidi G, Bianchini C (2007) TOF laser scanner characterization for low-range applications. In:
Videometrics IX, vol 6491, p 649109
Ribichini L (2007) The fortunes and history of Models. Disegnare idee immagini 34:50–61
Senatore LJ (2009) Analisi studio e confronto tra il modello virtuale della Basilica di San Pietro e
l’iconografia del Labacco. In: Benedetti S (ed) Antonio da Sangallo il Giovane: Il grande modello
per il San Pietro in Vaticano. Gangemi
Silvan PL (1994) San Pietro: Antonio da Sangallo, Antonio Labacco; un progetto e un modello;
storia e restauro; Santa Maria del Fiore; quattro modelli per il tamburo della cupola; restauro.
Bompiani
Tessari C, Bruschi A (1996) San Pietro che non c’è: da Bramante a Sangallo il Giovane, vol 93. Art
Books Intl Ltd.
Vasari G (1568) Vita d’Antonio da Sangallo architettore fiorentino. In Le vite de’ più eccellenti
architetti, pittori, et scultori italiani, da Cimabue insino a’ tempi nostri. Parte terza, primo volume
Wittkower R, Brunetti E (1964) La cupola di San Pietro di Michelangelo. Sansoni
Representing with Wood: Carlo
Lucangeli and the Model of the Flavian
Amphitheatre
Graziano Mario Valenti, Cinzia Conti and Jessica Romor
Abstract The wooden model of the Flavian Amphitheater built by Carlo Lucangeli
between 1790 and 1812 is among the largest and most detailed ever made on the
subject. Unlike the coeval models, normally made of cork and retracting the building
to a state of ruin, the Lucangeli wooden Coliseum, divided into seventy distinct
sections, is an interpretation of the ideal form of the monument, extremely detailed
and expressive thanks to the use of different wood essences and accessory materials,
such as metal and putty. This essay aims to illustrate the peculiarities of this model and
to define the importance of the contribution of Carlo Lucangeli in the vast historical
panorama of the representation of the monument.
Keywords Colosseum · Construction systems of wooden model

Integrated survey · Architectural morphology · Architectural analysis
1 Carlo Lucangeli and the Modeling Tradition
During the eighteenth century, artists and architects developed a great interest in
ancient monuments, which were studied and represented through drawings and phys-
ical models that reproduced, at a certain scale, the contemporary appearance or the
ideal configuration. Carlo Lucangeli, architect and cabinetmaker who lived between
1747 and 1812, was among the protagonists, together with Antonio Chichi and
Augusto Rosa (1738–1784) in Rome and Giovanni Altieri (1767–1790) in Naples,
G. M. Valenti · J. Romor (B)

Department of History, Design and Restoration Architecture, Sapienza University of Rome,
Rome, Italy
e-mail: jessica.romor@uniroma1.it
G. M. Valenti
e-mail: grazianomario.valenti@uniroma1.it
C. Conti
Rome Special Superintendence for Archaeology, Fine Arts and Landscape, Rome, Italy
e-mail: cinzia.conti@beniculturali.it

344 G. M. Valenti et al.
Fig. 1 Antonio Chichi, Colosseum, polychrome cork, XVIII–XIX century, Victoria Museum, Mel-
bourne. © R. Gillespir
of this tradition born as part of the scenery of the eighteenth century Neapolitan
cribs. In fact, during this period the felloplastica develops (from the Greek ‘phellos’,
that is ‘cork’: the art of making cork models), a technique that had its progenitor in
Augusto Rosa, collaborator of Gian Battista Piranesi. This art was intended repro-
duction of the ancient ruined buildings to be used, in principle, to the decoration of
the cribs, then affects the passionate collectors of ancient ruins. Cork was among
the favorite materials at that time for the production of ruined architectural models,
since its workability allowed, with a sculptural approach, to imitate the deteriorated
appearance of the masonry by removing material (Fig. 1) (Conti 2001, 2010; Conti
and Schwartz 2015).
The story that links Carlo Lucangeli to the Flavian Amphitheater is long and
charming. He devoted himself for a long time to the study of the monument, which
had even had the good fortune to observe first from a very unusual perspective. In
1788, according to local chronicles and literary and artistic sources (Fig. 2), Lucangeli
was participating in a flight demonstration of a balloon designed by Vincenzo Lunardi
(1754–1806), near the Mausoleum of Augustus. After several vain attempts to take off
that had disappointed the waiting crowd, the vehicle, following some modifications
made at the same time by Lunardi, unexpectedly took off the flight, lifting only
Lucangeli, who was unintentionally on board. For the first time in history, a man
flew over the city of Rome, and could enjoy an extraordinary, unusual overview of
the Colosseum, in all its magnificence (Conti 2010). The experience, which ended
after about a quarter of an hour with a fortuitous landing, particularly impressed the
architect, who dedicated the last twenty-two years of his life, from 1790 to 1812, to
the study of the Colosseum, for which he carried out excavation campaigns that led to
exceptional discoveries in the unexplored area of the hypogea, such as the discovery
of columns belonging to the portico in summa cavea and the passage of Commodus
(Conti 2001).
Lucangeli created several architectural models—among which we also remem-
ber the wooden model of St. Peter’s Basilica preserved in the Museum of St. Peters-
Representing with Wood: Carlo Lucangeli and the Model … 345
Fig. 2 Engraving of the

flight of the aerostatic
balloon built by Vincenzo
Lunardi in Rome in 1788
burg—and at least two of these concern the Flavian Amphitheater (Conti 2010; Conti
and Schwartz 2015; Schingo 2001). The first, in cork, dating back to 1792/1805,
reproduces the appearance of the monument as it must have appeared at the end of
the eighteenth century. It is kept at the École nationale supérieure des Beaux-Arts in
Paris but, given its current unavailability to the exhibition and consultation, it is pos-
sible to observe it only indirectly, through archive photographs (Fig. 3). This model is
particularly important, not only for its artistic merit and its continuity with the coeval
tradition, but above all because it faithfully reproduces the image of the Colosseum
in a very precise period and constitutes a precious benchmark for the evaluation of
the interpretative content of the second model, which we are going to analyze. This
other wooden model, built between 1790 and 1812, proposes instead an ideal recon-
struction of the monument and constitutes, for the wealth of information content, a
precious testimony in the history of representation and interpretation of its original
aspect (Fig. 4). The model, built on the 1:60 scale, is not only the largest ever on
the theme up until then, but boasts an exceptional fidelity in proportions and dimen-
sions compared to the royal monument, unlike what happened for traditional coeval
cork models, not particularly rigorous from the dimensional point of view. To certify
the quality of the work, in a booklet published in 1824 containing the description
of the model written by Lucangeli (Lucangeli and Dalbono 1813; Lucangeli 1851),
are some of the most authoritative academics of San Luca—Pietro Bianchi, Antonio
Canova, Vincenzo Camuccini, Giuseppe Camporesi, Giuseppe Antonio Guattani,
Raffaele Stern, Giuseppe Valadier, Filippo Aurelio Visconti—, G. Guillaume Le
Thières, Director of the French Academy in Rome at the time, and Laurent Re, pro-
fessor of archeology at the University of Rome. In fact, they exalt the originality,
the precision, the high stylistic quality of the model compared to the real monument
and praise the scientific reliability of the reconstruction, linked to the discoveries and
investigations conducted by the same author through the archaeological excavations
(AA.VV 1985; Conti 2001).
The present study deals with this wooden model, to which it was possible to access,
thanks to the Special Superintendence Archeology Fine Arts and Landscape of Rome,
for the operations of survey and study of the 70 parts that constitute it, during its
preparation for the exhibition “Colosseum—An Icon” at the Flavian Amphitheater,
from 8 March 2017 to 25 March 2018 (Conti 2017).
2 The Survey of a Wooden Model: Purposes

and Methodologies
2.1 Purposes
Carlo Lucangeli’s wooden model is today considered one of the most important and
surprising examples of architectural modeling. Despite the great credit attributed to
it by the academics of San Luca in its first years of life, the model has experienced
in the following decades moments of lesser fortune that have endangered its conser-
vation and concealed its precious value. Relocated numerous times, it has long been
mistaken for a model of the Thirties, because of the painting that made it similar
to the models of the Museum of the Roman Civilization of the Universal Exposi-
tion of Rome, painted with white tempera by the will of Benito Mussolini (Conti and
Schwartz 2015). Subsequently, the origins of the model were rediscovered and it was
understood that the painting was actually the result of a gray-colored painting carried
out in 1949 which, with the restoration of 1996, began to be removed, revealing the
original value of the delicate travertine paintings, of the decorations that imitate the
arrangement of stones and bricks, and the use of metal parts in cast tin, all attributable
to Lucangeli (Fig. 5).
The high artistic and scientific value of the model, combined with its need for
conservation and protection, given its fragility that makes it difficult to handling, has
therefore necessitated an accurate survey of the same, never done previously. This
Fig. 3 Photographs of the cork model of the Colosseum by Carlo Lucangeli conserved at the École
des Beaux-Arts in Paris
survey was therefore conducted with a dual purpose: the first consists in recording the
current state of the model, “photographed” as accurately as possible with respect to
the tools offered today by technology; the second, in the possibility offered to scholars
to observe remotely each part of the model, to establish, for example, methods and
criteria for conservative intervention or to foresee the movement and the exhibition
configurations indirectly.
Fig. 4 Photographs of the assembled model of the Colosseum by Lucangeli: © Zeno Colantoni.
Subdivision in plan of the 70 pieces constituting the wooden model and example of a sequence of
contiguous pieces
2.2 Methodologies and Results
The survey project first of all required a reflection on the techniques and methods of
intervention, being the survey of a model on a reduced scale not unlike that one of
the original monument. Various technologies have been adopted in the acquisition
phase, aimed at the production of Ultra High Res images, IMB models and laser
scans that document each of the 70 pieces constituting the wooden model.
In a small window of time, coinciding with the preparation of the model for its
subsequent public exposure, we have modulated survey and technology objectives,
opting for a massively redundant, rapid acquisition of data, to recover later procedural
errors, and necessarily referring to a second, much longer and more complex phase
of selection, normalization and representation of data and, therefore, of study of the
model (Valenti et al. 2017).
Fig. 5 Front of piece A1 and detail of the portion in which we can see the partial removal of 1949
painting
2.2.1 Gigapixel Images—The Survey of the Model as It Appears
The first phase of the survey consisted in the acquisition of high resolution photo-
graphic images of each piece of the model. Ring by ring, we photographed each
piece capturing for each of its sides a series of images, taken from the same point
of view with the help of a panoramic head and adequate lighting, then recomposed
through a planar projection with an appropriate stitching software. In this way we
have obtained for each element a series of photographic images able to reproduce the
main fronts at a scale of about 2:1 (Fig. 6), considering a print resolution of 300 dpi.
This choice allows us to clearly appreciate the more minute parts of the architectural
details and the consistency, quality and degree of conservation of materials, tints and
annotations reported by Lucangeli himself. For each ring, we have therefore com-
posed an abacus of analogous representations of the different components (Fig. 7),
Fig. 6 Three different level of details of the gigapixel image of A10 front. The percentage refers
to the real dimension of the model
graphically placing in relation to each other the representations of each piece and
the ones of contiguous pieces, so as to facilitate reading and identification of the dif-
ferent parts of the structure within the single piece and understand the connections
established in the model, in imitation of the real object, between the various spaces
that characterize it. This abacus was subsequently very useful in the analysis phase of
the model, as we will see in the part of the present study dedicated to its composition
and decomposition.
Fig. 7 Abacus of the 70 pieces composing the model. For each piece, we have here five main sides
2.2.2 Image Based Modeling—The Survey of the Model as It Is
The previous phase of survey allowed us to produce materials aimed at a thorough per-
ceptive control of the various components of the model, representing it as it appears.
The second phase instead wanted to introduce a different methodology of investi-
gation, Image Base Modeling, aimed at the reproduction of its three-dimensional
consistency. We carried out close-up photo shots for each of the seventy pieces,
using small and high-resolution digital cameras (12Mpx), which produced for each
piece, depending on the size, a number of images from 200 to around 800. These
photographs, appropriately selected, were then used to obtain, through a software
of photogrammetry, clouds of points suitable for describing in a discrete way the
various acquired surfaces, which on average were each represented by a cloud of
300 million points (Fig. 8).
The main problem in this second survey, as it would be for the real monument, was
the lighting project, aimed at reducing the shaded areas—necessarily present, given
the complex spatial articulation of the pieces and their small size—that otherwise
would have negatively affected quantity and quality of points in each cloud. In order
the camera did not interfere with the light sources necessary to illuminate the whole
piece, we decided to create two separate photographic sets, with relative dedicated
lighting system, for each component, postponing orientation and composition in a
single cloud to the next three-dimensional restitution phase. From this operation, we
have thus obtained the three-dimensional models of the seventy portions constituting
the wooden model.
Fig. 8 The process of IBM, from the capture of the photographs, to the generation of the digital
model. In the figure below, detail of the point cloud generated by IMB of piece C1
2.2.3 Laser Scans—The Model and Its Measurement
The third phase of the survey was therefore aimed at acquiring the dimensional data
relating to the seventy pieces of the model, in order to be able to size and measure
the models produced in the previous phase. For the scans, we have chosen to use two
types of scanners, short range and mid-range, the first with the aim of comparing
the quality of the surfaces acquired through Imaging Based Modeling, the second to
quickly obtain useful measurements to scale the acquired models.
2.2.4 Direct Survey—An “Interview” to the Model
After the two weeks dedicated to the instrumental survey, we turned off the tech-
nology and dedicated the days before the reassembly and subsequent exposure to
a careful and scrupulous observation of the model. Armed with a notebook and a
caliber, we dedicated ourselves to a synchronic analysis of the object, starting an
intimate and exciting dialogue with the maquette that revealed its charm and its high
communicative and cognitive value.
3 Composition and Decomposition: The Wooden Model

as an Instrument of Knowledge and Communication
3.1 Composition and Knowledge
The model—reproducing the monument in 1:60 scale, for a total size of 318 cm by
261 cm and 82.5 cm height—is divided into six rings, each of which is decomposed
into a variable number of pieces according to radial sections (Fig. 4). This decompo-
sition, which we will discuss in detail in the next part, favoured a thorough autoptic
examination in which it was possible to record the characteristics of the various parts,
from the supporting structure to the decorations.
We have first reasoned on the logic of selection and approximation of the various
parts of the structure with respect to the actual monument, classifying the types
of reproduced elements (walls, pillars, arches, vaults, and stairs) and the way of
representing them three-dimensionally, also in relation to the subdivision of the
model. Later, we investigated the characteristics of the decorative apparatus (capitals,
cornices, bas-reliefs, parapets, supports of the velarium, etc.), for which the use of
wood was complemented by the use of putty and metal.
3.1.1 What It Is Represented
Let’s analyze below the parts of the monument that are represented, following the
compositional logic of the Lucangeli model, ring by ring, from the outside to the
inside.
Ring A
This ring, divided into 16 portions symmetrical to the building’s axes (Fig. 9),
describes the first ambulacrum, at the three different orders, and the porch in summa
cavea, thus also representing the exterior elevation of the building. Each portion
contains a fixed number of arches, five, and is laterally sectioned at the pillars. The
third-order ambulacrum has an intermediate passageway, unlike the other two that
are full-height.
Fig. 9 Gigapixel images composition of a piece belonging to the ring A
The structure is made up of pillars that support the round arches of the eighteen
fornices, at three levels, and the part of the attic on the last level, in which there
are forty windows—one every two arches—overlooking the portico. Other forty
windows, smaller, squared and arranged alternately with respect to the first, are
instead obtained at the intermediate passageway.
The first two levels are covered by simple annular vaults, the third level and the
intermediate passageway by annular vaults with cylindrical lunettes in correspon-
dence with the openings of the arches; where we find the stairs of the intermediate
passageway, the annular vault is interrupted by sections of flat ceiling. The porch
roof is instead constituted by a flat ceiling supported by a system of beams and joists
that connect the perimeter wall of the attic to the columns, present in the next ring.
On the outer wall, we can see the 240 beams built to support the velarium, of which
there is no trace, at whose ends there are rings to which were fixed the pulleys that
allowed, through ropes, to stretch the cloth.
Regarding circulation and connection between the various spaces, represented in
the first ring and the next, we record the presence of four different types of stairs
(Fig. 10). The first, which only partially appear, consist of the last steps of the
sixteen stairs that come from the intermediate passageway of the second order of the
second ambulacrum, represented in the next ring B. Then, there are the 24 stairs of
the intermediate passageway of the third order, consisting of two opposing ramps
confluent in a second one that leads to the portico. Then we find 4 staircases, leaning
against the wall of the attic, which lead from the portico to the roof. Finally, on the
roof there are 6 small stairs with two symmetrical ramps leading from the roof to
the top of the perimeter wall. Parapets and metal gratings on the floors complete the
description of the structure.
The minute description of the architectural orders is one of the peculiar charac-
teristics of this ring. On the façade, from the bottom to the top, the Doric, then Ionic,
and two orders of Corinthian adorn the four levels. The first three orders are char-
acterized by semi-columns leaning against the pillars, while the attic is decorated
with pilasters, elements that we find also in correspondence of the pillars of the first
two levels, towards the inside. The pieces A1 and A9, which contain respectively
in the middle North and South arches, are enriched by a portal that emphasizes the
entrances. East and West entrances, belonging to the pieces A5 and A13, are instead
characterized by a greater width of the homonymous fornix. An exclusive feature of
these pieces is the presence, on the second and third levels, of the two rows of statues
placed on pedestals in correspondence of the arches.
Ring B
This ring, divided into 16 portions symmetrical to the building’s axes (Fig. 11),
describes the second ambulacrum, at the three different orders, and contains the
columns of the portico and the top part of the cavea. Each portion contains a variable
number of arches, from four and a half to five and a half, and is laterally sectioned
at the pillars or at the key of the arches. The first level houses the full-height ambu-
lacrum, the second and third levels have lower ambulacra (like the one on the third
level of the first ring) and are respectively surmounted by another intermediate pas-
sageway and by the top portion of the cavea.
The structure, also in this case, is made up of pillars supporting round arches, on
three levels, and to which the eighty columns of the portico correspond to the top. The
columns are surmounted by an entablature on which are fixed metal elements always
intended for the operation of the velarium. The first level is surmounted by a simple
annular vault, while the second and third level and the intermediate passageway by
annular vaults with cylindrical lunettes at the arches and stairs.
Fig. 10 Examples of the different type of stairs belonging to ring A
Regarding the circulation and the connection between the various rooms, we
register here the presence of two different types of stairs (Fig. 12). The first ones are
consist of the sixteen scales that lead from the intermediate passageway of the second
order of second ambulacrum to the third level of the external ambulacrum, described
previously. Then, there are the fourteen scales, much more articulated, which start
from the second level with three ramps, one placed in the first ambulacrum—and then
protruding from the B ring—and two, opposite, in the second ambulacrum. These
ones lead, with two distinct ramps, to the intermediate passageway of the third level
of first ring and to the cavea. Also in this case, parapets and metal gratings on the
floors complete the description of the structure.
This ring does not have a particularly developed decorative apparatus. The
columns of the portico are characterized by the Corinthian style and we find again
Fig. 11 Gigapixel images

composition of a piece
belonging to the ring B
the pilasters leaning against the pillars on the first level, on both sides, and on the
second level, only towards the outside.
Ring C
This ring, divided into 11 parts (Fig. 13), contains the portion of radial septa which
delimit the wedges, included between the second and third ambulacrum and in axis
with the pilasters placed on the previous rings. The pieces are symmetrical with
respect to the two main axes of the plan, except for C5 and C6, which are distinct,
but which together are still symmetrical to C7 and C1, and therefore analogous to
Fig. 12 Examples of the different type of stairs belonging to ring B
C11. Each portion contains a different number of wedges, comprised between four
and eight and a half, and is laterally sectioned in correspondence with a septum or
with the axis of a wedge.
The structure consists of vertical partitions which include the connecting stairs
between the various ambulacra and the cavea and which support the most substantial
part of the cavea itself. Each piece is delimited by a wall that leads to the second
ambulacrum and emerges at the top, at the level of third order, to delimit the view of
the ambulacrum to the arena through a series of doors and windows. The flights of
stairs are covered with barrel vaults.
The circulation between contiguous spaces is rather articulated and complex,
especially for the depth reserved for the pieces of this ring. We can however distin-
guish four types of stairs according to their function. There are stairs that connect the
first and second level of the second ambulacrum; others that connect the first level
of the third ambulacrum to the second level of the second ambulacrum; others that
connect the passageway placed inside the ring C just below the cavea to the inter-
mediate passageway of the second level; and then, there are the stairs that connect
the ambulacra and passageways to the cavea, through vomitoria.
Particular attention is paid in this ring to the modelling of the decorative apparatus
at the entrances put on its two axes. In fact, rich decorations characterize the walls
and the barrel vaults of the triads of wedges used to enter the main axes (Fig. 14):
Fig. 13 Gigapixel images composition of a piece belonging to the ring C
in particular, the positioning of the section plane at the North-South axis makes it
easier to observe the decorations placed on the sides of couples of pieces C11-C1
and C6-C7.
Ring D
This ring, also divided into 11 portions (Fig. 15), represents the third ambulacrum
and the overlying cavea portion. The pieces are symmetrical with respect to the two
main axes of the plant, including D11, which, despite being a single piece, is still
symmetrical to the pairs of pieces D1-D2 and D7-D8, and therefore similar to the
pair D5-D6. Each portion contains a different number of arcs, between four and a
half and ten and a half, and is laterally sectioned at a radial wall or at the axis of a
wedge.
Fig. 14 Examples of stucco decorations belonging to ring C
The structure consists of an external wall on which there are arches in correspon-
dence with the wedges of the previous ring. On the upper part of the wall, some
windows overlook the stairways of the previous ring. The ambulacrum is covered by
an annular vault, which supports a portion of the cavea steps.
Regarding the circulation and the connection between the various spaces, we
record the presence of the vomitoria that come from the intermediate passageway of
the previous ring, at the level of the first order. Also in this case, as in A and B rings,
parapets and metal gratings on the steps of the cavea complete the description of the
structure.
In this ring there are no details relating to the decorative apparatus.
Fig. 15 Gigapixel images composition of a piece belonging to the ring D
Ring E
This ring, divided into 8 portions (Fig. 16), includes the innermost part of the wedges
and the cavea. Each portion contains a variable number of wedges, from two to fifteen,
and is sectioned laterally at a radial septum or at the axis of a wedge. The division of
the pieces is not as regular as in the other rings, even if some symmetry relationships
remain.
The structure consists of radial septa that host single-level rooms and stairways,
covered by barrel vaults that support the last steps of the cavea. In this ring, there are
the stairs that connect the third ambulacrum to the cavea, while the passing wedges
allow the transit from the third to the fourth ambulacrum.
Fig. 16 Gigapixel images composition of a piece belonging to the ring E
As in C, we find some decorations in correspondence of the wedges that identify

the main entrances.
Ring F
The last ring is subdivided, like the previous one, into 8 portions (Fig. 17), and
contains the fourth ambulacrum and the areas near the arena. The ring does not
present a homogeneous subdivision, but the various portions are symmetrical with
respect to the main axes of the building.
The structure consists of a wall that separates the last ambulacrum from the arena,
covered by an annular vault, and supports the lower passageway of the cavea. In this
ring, we can see the stairs that lead from the fourth ambulacrum to the base of
the cavea and the openings towards the arena at four entrances. The piece F8 also
represents the so-called “passage of Commodus”, discovered in those years thanks to
Fig. 17 Gigapixel images composition of a piece belonging to the ring F
Fig. 18 Examples of stucco decorations belonging to the piece F8
the excavation campaign funded by Lucangeli. Parapets and metal gratings, draining
on the floors or closing the windows, complete the description of the structure.
In this ring, in the F4 and F8 pieces, there are the decorations on the internal walls
and on the vaults of the North-South axis. Piece F8 is particularly rich in decorations,
which—even if not directly visible—underline the importance of the presence of the
passage of Commodus (Fig. 18).
3.1.2 How It Is Represented
Let us come to the considerations on the construction of the model, describing on the
one hand the used materials—including the tree species used for the wooden parts and
the accessory materials (such as putty and metal) for the most minute details—and the
similarity of the chromatic and decorative aspects compared to the original reference,
and on the other the fixing techniques of the various components—both within the
single piece, and between contiguous pieces during assembly of the whole model.
Materials
The model is mainly made of wood, different types of wood, a feature that, as we have
already said, distinguishes it from contemporary cork models and allows the author
greater expressiveness in the representation of the different types of elements, from
the structural level to the purely decorative one. It represents a complex and refined
work of cabinet-making, technically well-designed: in fact, it not only demonstrates
resistance to the action of time without presenting significant deformations, but the
constructive choices adopted in the model reflect the actual construction system,
which makes the model not only aesthetically, but also substantially conform to the
original. Naturally, we can note constructive simplifications, given the reduction in
scale, as in the case of pillars, which are not tapered towards the inside.
The structure and the main architectural details are entirely made of wood
(Fig. 19a). Depending on the type of element represented, Lucangeli uses differ-
ent essences in relation to their workability. The whole structure is made of poplar
wood, solid and easy to work, present in the form of bands, little bands, strips and
blocks assembled together to form the supporting structure, imitating the logic of
real architecture (Conti 2001). For smaller elements such as moldings and columns,
Lucangeli uses beech wood, hard, compact, and therefore particularly suitable for
lathe and carving. Bases and shafts of the columns are in fact worked on the lathe,
while the capitals are made assembly tiny wooden tablets, blocks and extruded stucco
in paste, to allow a direct design of the decoration. The moldings are carved on juxta-
positions of panels, while lintels and frames were carved with a gouge. The elements
that represent the arena, made later by the son-in-law Paolo Dal Bono, are made of
elm wood (Conti 2001).
The more minute decorations, consisting of bas-reliefs that are located at the
entrance axes, are made of stucco, molded separately and glued after to the model
(Fig. 19b).
Many details of the ancient building were then represented with metal elements,
in imitation of the original: railings, gates, grilles for the discharge of water, were
made of a material resulting from the melting of tin and lead (Fig. 19c). The railings
have been fixed with iron nails and the grids with glue. The external archivolts were
also made of lead and tin casting and obtained on an external matrix. Other particular
elements are the supports of the velarium, made of wood and covered with a metal
sheet. A metal ring is fixed on the top of each one, which was used to slide the
ropes necessary to stretch the surface of the velarium thanks to pulleys. The working
pulleys of the velarium, visible on the external side of the pieces B (Fig. 11), under
Fig. 19 Examples of the different materials used to build several parts of the model
the wooden supports fixed to the trabeation, are made by assembling minute pieces
of metal and bone and clearly call to mind the tradition of Genoese naval models
(Conti 2010).
Finally, we observe the chromatic characteristics of the model, clearly visible in

correspondence of the pieces for which, during the restoration of 1996, the layer of
grey painting dating back to 1949 was removed. The original colors of the model,
spread with fluid and light brushstrokes on the wood prepared with a cooked lin-
seed oil, indicate the different parts of the Colosseum, depending on the material
that characterizes them in reality: there are therefore different colors to represent
travertine, brick, cement conglomerate and peperino marble. On the outside, super-
imposed on the ivory painting that imitates the travertine, there are also graphic signs
that evoke the marble blocks, as well as in section indicate the arrangement of the
bricks (Fig. 19d) (Conti 2001).
Assembly
Regarding the construction techniques, we first consider the question of fixing the
elements within a single piece. The structure is made, as we said, with pieces of
wood of various types, cut according to verse, and held together with cervona or
rabbit glue or small wooden dowels and metal nails (Conti 2001).
By way of example, let us analyze for example the sections of the A ring. They
are composed of two main portions of wood, held together by a wooden butterfly
joint at the level of the entablature of the second order (Fig. 20a). Vaults and floors
are then fixed to these using glue and nails (Fig. 20b, c). The stairs are built with
steps made separately and fixed on the climbing structure (Fig. 20d). In the pieces
of ring B, the main structure is divided into two parts at the height of the third level
floor. Unlike the solution adopted for the previous ring, reciprocal joints resulting
from the articulation of the section plane ensure greater rigidity of the element at
the connection between the two parts (Fig. 20e). The steps of the cavea are built to
imitate the real structure, modeled one by one and characterized by a sloping rear cut
that follows the slope of the cavea and allows it support to the underlying structures
(Fig. 20f). The rampant vaults made to cover the stairs placed between the radial
septa are made with the same technique of the annular vaults.
Let us now look at how is structured the assembly of the different sections to
obtain the closed configuration of the model (Fig. 21). Each section is fixed to the
others with wooden nails and screwed clamps. For the assembly of the model, there
are currently some long screws through the ambulacra, but they do not seem to be
original. Let us analyze in detail fixing systems and criteria.
The A elements are fixed to the ground with a pair of screws present at the left
and right margins of the base, on the outer side (Fig. 21a). They are also connected
to each other by means of metal plates placed on the horizontal plane at the top, to
avoid being seen on the façade (Fig. 21b). The discontinuity between one element
and the other is hidden by the overlap of the architectural order, which is entirely
realized on only one of the sides common to two elements (Fig. 21c).
Fig. 20 Assembly techniques of the parts within a single piece
The elements B, tall and slender, would be folded up very easily, and for this
reason they are anchored to the C elements by means of passing screws with visible
heads, placed in correspondence of the holes where we can see the flared entrance
for the screw head (Fig. 21d). The elements are connected to each other with metal
plates arranged vertically at the ends of the element on two levels (Fig. 21e). They
are also fixed together at the top with dowels to give continuity to the trabeation of
the portico in summa cavea (Fig. 21f).
Fig. 21 Assembly techniques adopted for the assembly of the 70 pieces, ring by ring
The elements of ring C, joined in sequence by means of metal clamps placed on

the outer side about one third and two thirds of the height and fixed with screws,
constitute the annular band which, having greater thickness and weight than the
others, gives stability to the whole model. We can see on the outside the screw holes
coming from the elements B (Fig. 21g).
Fig. 22 A sequence of contiguous pieces that shows the relation between the different spaces
represented
Fig. 23 Some drawings depicting the Colosseum by Luis Joseph Duc, 19th Century
The ring elements D, E and F are connected to each other by means of metal
plates placed vertically near the upper lateral ends of the elements (Fig. 21h–j).
Several elements E also have screws on the steps of the cavea to fix on the elements
F (Fig. 21k). All the elements F, in turn, have holes to accommodate fixing screws on
the lower level in order to anchor themselves drawing an oval profile on the support
surface where the model will be assembled (Fig. 21l).
Fig. 24 Some drawings depicting the ideal aspect of the Colosseum by Antoine Desgodets (a) and
Carlo Fontana (b)
3.2 Decomposition and Communication
As we have seen and as written by the illustrious academics of San Luca above, the
wooden model of Carlo Lucangeli can be considered the first architectural model of
the story that deals with the theme of ideal reconstruction of Flavian Amphitheater
in such a systematic and passionate way, both for the precision in the definition of
the spaces and their dimensions, both in the philological rigor of the interpretations,
due to the deep knowledge of the site (Lucangeli and Dalbono 1813; Lucangeli
1851). This model, however, has another important record: to tell, with a narrative
logic normally proper of technical architectural drawing, the structural morphology
of the monument and the complex relationships between the various spaces that
characterize it.
The wooden model of Lucangeli is characterized by a constructive logic that

not only manifests the will, common to its era, to perceptively evoke the outward
appearance of the monument in its original composition, but above all the intention
to describe its structure and the internal articulation (Fig. 22) (Valenti et al. 2017).
The first subdivision, in rings, allows to highlight the internal and external eleva-
tions of each portion, showing in particular the development of the ambulacra, the
consequentiality of the covered spaces, from the first ambulacrum to the arena, and
their relationship with the cavea. In fact, the section surfaces are always placed in
such a way as to completely unveil the ambulatories and to observe and catalogue
the various types of staircases that connect the different rooms together.
The second subdivision, the radial one that follows the course of the wedges,
instead highlights the sections of the various rings, thus showing an information
content comparable to that of traditional section drawings. Lucangeli uses many
different section planes and almost never aligned with each other from one ring to
the other. This allow to observe the typologies of vaults represented and to understand
the functioning of the stairs leading from a ring to the other, especially at the deep
C ring, which contains various types of stairs between the radial walls that delimit
the wedges. The only part of the model that admits the same section plane, on the
two opposite sides, for the rings B, C, D, E, is that relative to the North-South axis,
where in fact the author maintains a unity to emphasize the rich decorative apparatus
represented there.
Through these two levels of subdivision, the model becomes an atlas of three-
dimensional drawings, to be flipping through piece by piece, ring by ring, to fully
understand the composition of the monument. A similar effect also come from the
precious drawings of the Colosseum that a few years later will be realized by Luis
Joseph Duc (Fig. 23), who certainly saw Lucangeli’s model and took inspiration
from its subdivisions for the creation of his own drawings (AA.VV 1985; Conti
2001, 2010; Conti and Schwartz 2015).
4 Conclusion
As we have seen, the wooden model of the Flavian Amphitheater by Carlo Lucan-
geli plays a role of primary importance in the history of the representation of the
monument. It is in fact the expression of the most advanced and in-depth knowledge
gathered up to that moment on the Colosseum and on the interpretation of its ideal
form.
Lucangeli personally dedicated himself, for years, to the survey and study of the
monument, with such passion to generate a model that is able to communicate with
perfect synthesis the essence of the building, thanks to his brilliant logic of con-
struction and decomposition and the use of different types of wood, that allow him
to diversify hierarchically the various parts in relation to their workability. “Flip-
ping through” the model, as if it were a “three-dimensional album” of detailed and
expressive architectural drawings, we can immediately understands the typology
and succession of the various spaces characterizing the building and easily grasp
the proportional relationships between them. The observation of the different types
of stairs and their distribution moves the viewer to virtually explore the monument,
to discover the paths that connect the different spaces, both inside and outside. The
hierarchy of spaces becomes evident where the eye pauses on the richest decorations,
highlighted by appropriately placed sections.
Moreover, the comparison with the previous graphic sources, in particular with
the famous reconstructions by Fontana (1725) and Antoine Desgodets (Cellauro and
Richaud 2008; Desgodets 1972) (Fig. 24), shows how the interpretation of the ideal
form by Carlo Lucangeli from one hand drew on this tradition, from the other pro-
posed also different and unpublished solutions, where important discoveries—like
the columns of the portico in summa cavea—allowed new reconstructive hypotheses.
This is the starting point for a new research path aimed at analyzing and comparing
the wooden model with the real model, whose image at the time of Lucangeli is
recorded in the cork version we mentioned. In fact, we believe that the comparison
with the real monument, made possible by the integration of data deriving from
the different surveys carried out over time, including that performed in the year
2000 under the direction of Mario Docci and the scientific responsibility of Riccardo
Migliari (AA.VV 1999), may further to return to this masterpiece of architectural
representation the fame and prestige that had accompanied his birth and that for so
many years have remained dormant.
References
AA.VV (1985) Roma Antiqua. “Envois” degli architetti francesi (1788–1924). L’area archeologica
centrale, catalogo della mostra (Roma, 29 marzo-27 maggio 1985; Parigi, 7 maggio-13 luglio
1986). Roma/Parigi
AA.VV (1999) Il Colosseo. In: Docci M (ed) Disegnare Idee Immagini. Gangemi, Roma
Cellauro L, Richaud G (2008) Antoine Desgodets. Gli edifici antichi di Roma. De Luca Editori
d’Arte, Roma
Conti C (2001) Il modello ligneo dell’Anfiteatro Flavio di Carlo Lucangeli: osservazioni nel corso
del restauro. In: La Regina A (ed) Sangue e arena. Electa, Milano
Conti C (2010) Rilievo, restauro, restituzione: il modello in legno del Colosseo di Carlo Lucangeli.
In Fagiolo M, Tabarrini M (eds) Giuseppe Piermarini tra barocco e neoclassico: Roma, Napoli,
Caserta, Foligno. Perugia
Conti C, Schwartz E (2015) Les maquettes du Colisée par Carlo Lucangeli. In: Amsellem G (ed) La
maquette: un outil au service du projet architectural, actes du colloque qui s’est tenu les 20–21
mai 2011 à la Cité de l’architecture & du patrimoine. Edizioni Cendres, Parigi
Conti C (2017) Il modello di Carlo Lucangeli: un Colosseo di legno. In (a cura di) Prosperetti,
F., Rea, R., Romano Gosetti di Sturmeck, S., Santangeli Valenziani, R.: Colosseo. Un’icona,
catalogo della mostra (Roma 8 marzo 2017-7 gennaio 2018). Mondadori Electa
Desgodets A (1972) Les édifices antiques de Rome. Collegium graphicum, Portland
Fontana C (1725) L’ Anfiteatro Flavio descritto e delineato dal cavaliere Carlo Fontana. Isaco
Vaillant, Haia
Lucangeli C, Dalbono P (1813, I ed.; 1815, II ed; 1824, III ed.) Modèle qui offre la restauration
du Colisée de Rome, tel qu’il était originairement; De la soixantième partie de la réalité, mesure
linéaire. Fait par Carles Lucangeli, célèbre Architecte mécanicien. Roma
Lucangeli C (1851) L’anfiteatro Flavio detto il Colosseo della grandezza di palmi 2449 di circon-
ferenza ridotto alla circonferenza di pal. 40: misura lineare dall’uno al sessanta per opera di Carlo
Lucangeli romano. Tip. di C. Puccinelli, Roma
Schingo G (2001) I modelli del Colosseo. In La Regina A (ed) Sangue e arena. Electa, Milano
Valenti GM, Conti C, Romor J (2017) L’Anfiteatro Flavio nell’interpretazione lignea di Carlo Lucan-
geli: un modello conoscitivo, 39° Convegno dei Docenti delle Discipline della Rappresentazione,
XIV Congresso UID, Napoli, 2017. Gangemi Editore, Napoli
The Teatro of Bologna as a Transformable
Space: Drawing, Geometry and Invention
in the Study of the Wooden Model
of the Theater by Antonio Galli Bibiena
Giuseppe Amoruso
Abstract The teatro of Bologna, designed by Antonio Galli Bibiena, was opened
in 1763 in the town of Father Giovanni Maria Martini, a leading musician consulted
also by Wolfgang Amadeus Mozart, and Farinelli, celebrated as one of the greatest
singers in the history of opera. Bibiena proposed an innovative solution for the cavea
shape, a multiple curved profile with a bell-shaped morphology. The paper describes
the architectural features of the theater and its innovation starting from the survey
of the wooden model exhibited at the Museo Internazionale della Musica and also
the unique wooden mechanism under the parterre stalls that allows to transform and
improve the functionality of the space.
Keywords Teatro Comunale di Bologna · Antonio Galli Bibiena · Wooden model

Wood machine · Architectural morphology · Italian theater design
1 The Galli-Bibiena Family and the Tradition of Theatrical

Scenery in Bologna
In 2006 Bologna was nominated UNESCO Creative City of Music celebrating its rich
musical tradition and the continuous importance of initiatives in the field of music
practice and studies. Music for Bologna, in continuity with a glorious past, is tradi-
tion, creativity and identity: Father Martini, Farinelli, Rossini, Donizetti, Respighi
and many others as well as the presence of leading institutions such as the Munic-
ipal Theater, the International Museum and Music Library, the Conservatory, the
Department of Arts at the University of Bologna, the Philharmonic Academy.
Bologna has also a long tradition about scenography, perspective and illusory
frescoes and theater design; was the town of the Galli-Bibiena family, a dynasty born
under Giovanni Maria Galli who was from Bibbiena (Florence). It was a family of
G. Amoruso (B)
Polytechnic University of Milan, Milan, Italy
e-mail: giuseppe.amoruso@polimi.it

376 G. Amoruso
artists, designers, architects who dedicated themselves to architectural work and set
design for the theatre: among them Ferdinando, the author of L’architettura civile
preparata su la geometria e ridotta alle prospettive (1711) and his son Antonio,
designer of the Teatro comunale in Bologna, the Teatro Scientifico in Mantova (1769)
and the Teatro Fraschini in Pavia (1771–1777) and many other theaters.
In Bologna, in those years, a school of scenography and applied arts was founded;
it was addressing the study of perspective and the building of ephemeral architectures,
architectural perspectives and illusory paintings both in public buildings than in
the most important private palaces and theaters. The Galli-Bibiena family, known
in many European courts for the design of stages and theaters, is considered the
most important reference in the field. Using the highly ornate style of late baroque
sculpture and architecture, the members of the Galli-Bibiena family designed a series
of theatrical scenery and other masterworks.
In 1756 the Bolognese Senate, thanks also to the strong commitment of Cardinal
Lambertini, commissioned Antonio Galli Bibiena to design a new theater. The Teatro
Comunale di Bologna was opened in 1763 also to compensate for the loss of the
Malvezzi theater, which burned a few years earlier.
During those years the city had a transformation and also an unusual renewal
through the administration of the Cardinal Legato, a change in the image of the
townscape without affecting the urban pattern. The city changed its appearance, its
way to represent its feature, a new scene as it was for the system of porticoes that
gradually became no longer built in wood but rebuilt as brick columns; the town
was decorated with street scenography perspectives and also courtyards, gardens
and staircase were the subject of painters and architects that painted their walls and
ceilings. (AA VV 1998)
The architecture of the interior became more and more sophisticated, painted
and intangible, interiors were improved with new architectures of deception through
illusory frescoes and decorations. The renowned school of scenography and the
Accademia Clementina (currently the Accademia di Belle Arti) were destined to
leave their mark all over Europe. The skill of the set designer moved on the delicate
balance between the true and the false using the perspective illusion, paraphrase of
the truth. What was happening was pure theatrical representation of a new social life,
where the actors were the ladies and the knights and the audience was the people
who discovered themselves as spectators within the contemporary society (Figs. 1
and 2).
2 The Teatro of Bologna as a Transformable Space
The teatro Comunale di Bologna was erected on the area occupied in the past by
the magnificent Domus Area of the Bentivoglio family who ruled the city till 1506;
along an almost unbuilt road dedicated to San Donato, Sante Bentivoglio designed
the site and conceived the architectural feature for his magnificence by building his
main residence, a loggia with twelve arches, the family chapel from the renovation of
The Teatro of Bologna as a Transformable Space: Drawing … 377
Fig. 1 A miniature of 1722 by Leonardo Sconzani that shows the teatro del Pubblico o della Sala,
built inside the Palazzo del Podestà towards the end of the 16th century and which remained active,
even after its fire, until 1768, when it was demolished (Bologna, Archivio di Stato, Insigna degli
Anziani Consoli, vol. XIII, c. 125.)
the ancient oratory of Santa Cecilia, a portico with 36 columns and the new Palazzo
degli Stanardi (Ceccarelli 1998).
In Bologna, governed by the papal state, it happened that the Pope Julius II resolved
the end of the Bentivoglio family, driving away the family; the building that had
become a real court was destroyed and it remained unbuilt and ruined for more than
two centuries. Thus the image of that place, which was the scenic representation
of personal pomp and wealth, became the indelible mark of punishment, of public
drama, a place occupied by ruins as a result of a spiritual power that put an end
to the richness; but the site still remained a place of representation, the “guasto”
(damaged), so called by the Bolognese, celebrated itself according to a scene that
brought impressions of its memory.
In the second half of the eighteenth century Bologna was a great construction
site, a new culture of modernity was spreading and senatorial families represented
themselves in their rooms, hosting parties of increasingly lavish and richly decorated
installations.
The shows were made outdoors in the civic main space, Piazza Maggiore, but also
it was a common opinion that the town needed a place for the celebration of events.
378 G. Amoruso
Fig. 2 The wooden theater settled in Piazza Maggiore for the Festa della Porchetta of 1627 (it was
celebrated every year on August 24th); these celebrations are considered as a public representation
and also exhibition of the bolognese society of those times, namely the “Bononia Felix” according
to several authors, and the birth of the tradition to build temporary scenographic installations in
Bologna (Bologna, Archivio di Stato, Insigna degli Anziani Consoli, vol. V, cc. 102–03.)
Antonio Galli Bibiena, known to numerous European courts for the design of
scenery and theaters, was a theatrical engineer and set designer as well as having
worked at the court of Vienna as “First Imperial Architect”.
He proposed an innovative form for the cavea, a multiple curved profile with a
bell-shaped imprint; he also introduced new constructive solutions such as the use
of load-bearing masonry (instead of the usual wooden frame), innovations that cost
him a heavy denigration campaign but which instead proved to be very effective also
for example for acoustic performance (Lenzi 1975) (Figs. 3, 4, 5 and 6).
For ten years Antonio Galli Bibiena and his project were the object of a campaign
of denigration and satire; Bibiena, to defend himself, prepared a series of drawings
to support his thesis and presented publicly also a wooden model to describe the
solutions for the cavea. The oppositions addressed the controversial issue of the “new
shape” of the theater and the use of masonry for the main structures and architectural
frame (Garbero Zorzi 1997)
The bell-shape cavea designed by Bibiena turned out to be a non-functional solu-
tion according to the expected acoustic behavior of the room and in this regard the
opponents wrote: “What is the basis of such a refined invention, it is easy to see the
similitude, that is, or analogy, who imagined having to find between the sound made
by the bell, and the figure of the bell itself that makes it” (Algarotti 1763, p. 77).
History and modern science will prove to be unfounded accusations and the blamed
bell-shape cavea will then be reused also in the nineteenth century as in the case of
Carlo Rossi’s Aleksandrinsky Theater in St. Petersburg.
Also, the choice of the masonry for the construction of the upstairs galleries
in place of the traditional wood appeared in contrast with the characteristics of
euphonicity of the theater, but also this time the opponents were wrong. The new
social function of the theater, assumed as an organism, required a new position within
the townscape and an innovative morphology for the architectural structure that was
thought of greater monumental value with an improved capacity and new facilities.
Antonio Galli Bibiena defended himself almost to the point of exhaustion from the
accusations justifying the solution of the masonry with the need to have solid, durable
and fire-resistant structures, the main danger of the wooden theaters. Unfortunately,
the criticism and material obstacles to the project caused the incompleteness of the
main facade, which remained altered in the composition until the 30s of the last
century (Amoruso 2009)
Regarding the bell shape cavea, the designer assured that it was “of complete liking
to the musicians and the spectators” and to reinforce this conviction he designed later
the Teatro Scientifico of Mantova according to the same shape (Figs. 7, 8 and 9).
380 G. Amoruso
Fig. 3 The design solutions used by Antonio Galli Bibiena: load-bearing structure and stages
built in masonry and the innovative bell solution for the cavea (Bologna, Biblioteca Comunale
dell’Archiginnasio, Gabinetto Stampe)
Fig. 4 Plan and section of the Teatro Comunale published in the collection “Pianta e spaccato del
Nuovo Teatro di Bologna” by Lelio dalla Volpe, Bologna 1771 (Bologna, Biblioteca Comunale
dell’Archiginnasio)
3 Modelling the Theatre, a Wooden Tool for Design

Visualization: 3D Reconstruction and Analysis Workflow
The study of the wooden model and of its constructive characteristics is one of
the subject of this research which has foreseen a data collection phase on site at
the Museo Internazionale della Musica in Bologna where the model is exhibited.
The case study presents the process of automatic 3D reconstruction from a set of
photographs: the survey is based on the accurate structure-from-motion technique
and photogrammetric process in order to study the design features and the shape
parameters of the cavea. The process allows to create a 3D data set from a selection
of photographs and, at the end of the process, to build an accurate model for the
geometric and architectural analysis (Figs. 10, 11, 12 and 13).
The use of such algorithms avoids the use of any in-scene markers that is required
by some photogrammetric applications, often impractical, like a not accessible site
or not allowed when a non-contact approach is mandatory. The latest technology is
targeting automated approaches as much as possible and the improvement of stereo
382 G. Amoruso
Fig. 5 Plan and section of the Teatro Comunale published in the collection “Pianta e spaccato del
Nuovo Teatro di Bologna” by Lelio dalla Volpe, Bologna 1771 (Bologna, Biblioteca Comunale
dell’Archiginnasio)
matching techniques is one of the most investigated research topic of the last years
in photogrammetry and computer vision.
The survey and 3D image-based reconstruction starts from the images, correct
for lens, and computes feature points between them; it’s a scale invariant feature
transform process. The algorithm computes camera positions and other intrinsic
camera parameters, creating the sparse 3D point cloud, called “bundle adjustment”.
A step further is to create the dense point cloud and, according to a set of geometric
constraints like the number of polygons, it is formed the mesh from the dense point
cloud.
Fig. 6 Section of the cavea and of the proscenium of the Teatro Comunale published in the collec-
tion “Pianta e spaccato del Nuovo Teatro di Bologna” by Lelio dalla Volpe, Bologna 1771 (Bologna,
Biblioteca Comunale dell’Archiginnasio)
When the mesh surface is ready it is possible to re-project images to derive texture
segments; it is also optionally possible to simplify and optimize the mesh through a
quadratic edge collapse decimation and fill holes too.
The final step is to export the 3D model for the post processing step and then
to extract a set of orthogonal views and further illustrations like exploded views or
video animation.
The survey of the wooden model was done with Agisoft PhotoScan, a commercial
software, developed by Agisoft LLC company, that follows a multi-image approach.
After the images orientation, the software refines calibration camera parameters (the
geometry of the images sequence allows to estimate a set of interior orientation
parameters for each camera, whether these are not previously assigned); in a second
step, it proceeds to the DSM generation. The image acquisition phase has been par-
ticularly complex because the model is exhibited in a museum room and is also kept
inside a transparent box, only partially openable. In the image-based 3D reconstruc-
tion process, image quality is closely linked to the accuracy of the data subsequently
384 G. Amoruso
Fig. 7 Interior of the Teatro Scientifico di Mantova designed by Antonio Galli Bibiena for the
Accademia dei Timidi, 1767
extracted from the structure-from-motion algorithm and the first step is to import
and select a good set of pictures.
Tie points are detected on two (or more) different images and they are actually
matches between SIFT points (named “key points”). Those points can be filtered,
using different methods (Gradual selection tools in PhotoScan) in order to converge
to a more precise relative orientation. On picture overview panel, after the camera
alignment step, it is possible to validate the process, blue points are valid tie points
used to get this 3D relative orientation, whereas grey points are disabled tie points,
disabled because of their imprecision. They are the fundamental link between differ-
ent images to get 3D relative positioning. The software performs with a high level
of automation, being very intuitive, and this is the reason of its diffusion among
practitioners and the user can interact to improve the points cloud density and the
3D modelling quality.
The depth map computation is based on the overlap of image pairs, merging all
the results in a single, final, 3D reconstruction. In fact, a multi-baseline matching
extension is more robust with regard to occlusions detection and wrong matches,
realizing the fusion of disparity information given by all the match images and
producing smoother results.
Fig. 8 The wooden model of the Nuovo Teatro Pubblico by Antonio Galli Bibiena which was
crafted in 1756 by Giovanni Battista Martorelli and Antonio Gambarini (Bologna, Museo inter-
nazionale e biblioteca della musica di Bologna)
The analysis of the model confirms the choices of Antonio Galli Bibiena and high-
lights the transformations that the theatrical space has undergone: the reconstruction
of the vault of the stalls, the suppression of the cornice that ran to the top of the hall,
the restoration of the proscenium and the rectification of the performance of the para-
pets of the boxes. In the years 1853–54, under the direction of Carlo Parmeggiani,
the shelves of the third order were modified, the cornice of the gallery, the vault of
the stalls and again the proscenium.
In 1907, together with some modifications concerning the halls and the decoration,
the theater was equipped with electric lighting. The fire hit the theater in 1931 when a
fire destroyed the curtain and the entire stage. The last “architectural” chapter on the
theater was written in 1980–81 when complex restoration work gave new life to the
main wooden structures (threatened by an army of woodworms) and allowed a radical
retrofitting that, for example, led to the constraint of the stalls at the surrounding
cavea.
386 G. Amoruso
Fig. 9 The wooden model of the Nuovo Teatro Pubblico by Antonio Galli Bibiena which was
crafted in 1756 by Giovanni Battista Martorelli and Antonio Gambarini (Bologna, Museo inter-
nazionale e biblioteca della musica di Bologna)
4 The Teatro of Bologna as Transformable Space:

Illustrating the Wooden Mechanism
The extraordinary musical instrument that is the Teatro Comunale of Bologna pre-
serves inside a wooden mechanism placed under the parterre stalls.
This apparatus was designed by Filippo Ferrari in the early 1800s and came
into operation in 1820 on the occasion of radical technical adaptation interventions
to which the theater was subjected in competition with the new Teatro del Corso
opened in Bologna in 1805.
The mechanical apparatus allowed the raising of the stalls at the level of the stage
and arose from the need to achieve continuity of the walkable plane between cavea
and stage, very useful in occasion of special events, such as carnival; in this way the
theater “got” the raised roof and easier equipment for many years desired.
If there are copious bibliographic sources about the Teatro Comunale that allow
us to reconstruct the lively events that have accompanied it construction and its
evolution, not so much for the apparatus object of this study which is completely
unknown in its morphology and little studied. (Amoruso 2006)
The presence is reported in the inventory written after the works of the early
1800s of the new mechanical apparatus that it has be considered as a “mechanism”
according to the definitions used in the scientific literature.
Fig. 10 Automatic 3D reconstruction from a set of photographs. The wooden model is recorded
with a camera and the picture alignment is processed with a structure-from-motion algorithm
(3D reconstruction and illustrations by Andrea Manti under the scientific supervision of Giuseppe
Amoruso)
In this regard it should be noted that “machine” means, more in general, a device
that has the dual function of transmitting movement and to transmit force. Therefore
the parts that compose the machine are built and entrenched by their possibility of
movement following the application of certain forces.
It is more legitimate to speak of “mechanism” since the object of the research
examines the parts from a prevalent geometric and cinematic point of view ignoring
the considerations inherent in the transformation of forces and the transformation of
energy.
The wooden apparatus object of this research can be described as a mechanism
consisting of lifting machines and flexible parts. Among the lifting machines they
are to be mentioned the capstan, the levers, the wheels and the pulleys, while as a
flexible element, in this case, the machine uses a hemp rope.
Flexible parts are mechanical organs that have the ability to react only to tensile
stresses and for this reason they are largely used in lifting and transport machines:
in this category we find ropes, chains and straps.
The mechanism inside the theater is equipped with ropes made of natural fiber
(hemp) and simple ropes and stranded ropes are recognizable as a consequence of
388 G. Amoruso
Amoruso)
Amoruso)
the solicitations applied in the different organs. The mechanism is still perfectly
preserved today but has been unused since the beginning of 1900 (Figs. 14 and 15).
Ferdinando Galli Bibiena, father of Antonio and known for his numerous pub-
lications on scenography and perspective, wrote in the Direzioni Della Prospettiva
Amoruso)
Teorica Corrispondenti a quelle dell’Architettura (Bologna 1732) that “La Mecanica

è Arte, per la quale si cerca di superare le cose maggiori con le minori, e col mezzo
di poca forza movere cose di molto peso.” (“Mechanics is Art, for which one tries
to overcome the greater things with the minors, and with the medium of little force
to move things of great weight. TbA”). This concept has been transferred in the con-
struction of the wooden machine that is located under the auditorium of the theater;
390 G. Amoruso
Fig. 14 The parterre stalls without the floor plan during the 1981 restorations; the underlying lifting
mechanism is visible from the cavea, Bologna, Archivio Storico del Teatro Comunale
due to its importance and thanks also to the state of conservation, part of the research
was dedicated to the graphic documentation of the lifting mechanism of the platea;
this apparatus was used during special events to obtain a single practicable horizontal
level that connected the stage to the main parterre space.
It is a complex wooden mechanism, still intact even if unused, consisting of
a central and big winch and an idler wheel which, by means of hemp ties, move
a system of four levers; the vertical translation movement is transmitted to two
transversal trusses on which the entire floor is supported.
The general working model of the apparatus and the different configurations that
the individual elements take at rest and in action are described by volumetric models;
an axonometric plan presents the overall position of all the parts of the mechanism
in relation to the space that contains them while an axonometric view from below
Fig. 15 The parterre stalls without the floor plan during the 1981 restorations; the underlying lifting
mechanism is visible from the cavea, Bologna, Archivio Storico del Teatro Comunale
describes the interaction between the platea floor, the trusses that support the floor
and the lifting levers.
Representation of the individual elements is given by graphic drawings and
3D modelling after the metric survey which illustrate constructive and functional
issues through orthogonal projections and axonometric exploded views (Figs. 16, 17
and 18).
392 G. Amoruso
Fig. 16 Geometric survey and 3D modeling of the platform lifting mechanism (3D modelling and
illustrations by Giuseppe Amoruso)
5 Conclusions
Vitruvius, in his famous treatise, entrusted to the architect the instruments of the
Ichnographia, of the Orthographia and of the Scaenographia. Nowadays, the prac-
tice of representing cultural heritage is still traditionally entrenched to tools and
procedures for the visualization in orthogonal projection of the main building or
space floors and/or planes. But nevertheless the contemporary scholar is entrusted
with new digital tools that allow to deepen his analysis, to develop the complex rela-
tionships between the parts within a 3D environment, to read, to discriminate and to
accurately describe the geometries of heritage; research in the design sector requires
to correctly perform the complex task of “reducing” the geometries, according to
Ferdinando Galli Bibiena, stratified by the long historical evolution of the artefact,
carrying out a process of analysis that develops the three-dimensional data made
available by digital instruments. For these objectives, representation techniques and
the 3D reconstruction methodologies have carried out a refinement process with
an improved efficiency aimed at optimizing the relation among the drawing sheet
(today the computer screen) and the different projection systems. Today it is possible
to foster representations that sometimes are far from the real vision of the object or
are augmenting its physical appearance with extra layers and information but which
give grounds for the possibility of producing correct graphic-proportional outputs
and illustrations.
394 G. Amoruso
Acknowledgements This work is part of a wider research on digital heritage and also on the
theater of Bologna that was the subject of the PhD thesis defended by Giuseppe Amoruso in
2001 at the University of Ancona (nowadays Università Politecnica delle Marche), “Teatro come
spazio trasformabile: rilievo di una macchina di legno”, supervisor Roberto Mingucci (University
of Bologna). Author acknowledges Andrea Manti for the wooden model survey 3D reconstruction
and illustration and the Teatro Comunale for the support.
References
AA VV (1998) Il teatro per la città, Bologna

Algarotti F (1763) Saggio sopra l’opera in musica, Livorno
Amoruso G (2006) Una macchina di legno nel Teatro Comunale di Bologna in Rivista Digitale
On line. DISEGNARE CON… vol 0, Dipartimento di Architettura e Pianificazione Territoriale,
Università di Bologna
Amoruso G (2009) Il Teatro Comunale di Bologna di Antonio Galli Bibiena. Disegni e rilievi per una
facciata incompiuta in IKHNOS 2009 Annale di Analisi grafica e Storia della Rappresentazione,
Università degli Studi di Catania—Facoltà di Architettura, Siracusa
Ceccarelli F (1998) Il Teatro Comunale nella strada San Donato, in A.A.V.V., Il teatro per la città,
Bologna
Garbero Zorzi E (1997) La vicenda toscana della “nuova forma” dei teatri di Antonio Galli Bibiena,
in I Galli Bibiena. Una dinastia di architetti e scenografi, a cura di Deanna Lenzi. Atti del Convegno
(Bibbiena, 26–27 maggio 1995), Accademia Galli Bibiena, Bibbiena, pp 83–98
Lenzi D (1975) Riedizione anastatica con note introduttive della Pianta e Spaccato del Nuovo Teatro
di Bologna offerto al Nobil’Uomo ed eccelso Sig. Senatore Conte Girolamo Legnani da Lorenzo
Capponi, In Bologna Per Lelio dalla Volpe…MDCCLXXI (con alcune varianti rispetto all’ed.di
Venezia del 1764), D. Atesa Editrice, Bologna
Pianta e spaccato del Nuovo Teatro di Bologna fatta in occasione dell’apertura di esso il 14 Maggio
1763… In Bologna MDCCLXIII. Nella stamperia del Longhi
Pianta e spaccato del Nuovo Teatro di Bologna al Nobil Uomo, ed Eccelso Sig. Senatore Co.
Girolamo Legnani Ferri…da Lorenzo Capponi, Venezia MDCCLXIV (con cinque incisioni)
Learning by Prototyping: Wood Design
Course Experience
Mariano Chernicoff
Abstract Since 2012, a new teaching and learning ethos has been embedded in the
prototype design workshop at Politecnico di Milano. This is based on rediscovering
the past design traditions in the light of the new present and future technologies. This
education experience has been applied in the Wood Design course of the Design
School, a teaching course inspired and with roots within the traditional Arts and
Crafts academies, but also integrating new tools and skills for design/production of
wood products. One of the key features of this course is that students are expected
to express their own ideas through usable prototypes, making products and furni-
ture with different styles but always with a high technical and finishing quality. This
teaching approach can be associated with Bruno Munari’s motto, when he evokes
Confucius saying “I hear and I forget. I see and I remember. I do and I understand.”
For these reasons MAKING is the backbone of this course. The focus of the chapter
is around the great potential of this way of teaching, through direct constructive activ-
ities, to understand how such a practical approach about representation, design and
knowledge transmission can become a solid basis education for future generations
of designers.
Keywords Wood design · Prototype · Arts and craft · Educational innovation

Practical approach · Knowledge transmission
1 Introduction
The School of Design in Milan has a long tradition of professional design practice,
which grew out of the material culture of craftsmanship and industry in Lombardy
at the beginning of the last century and boomed after the war and in the fifties
with an original, authoritative voice. Industrial Design oversees the designing and
enhancement of industrial products, integrating design disciplines (rooted in art and
M. Chernicoff (B)
Design Department, Polytechnic University of Milan, Milan, Italy
e-mail: mariano.chernicoff@polimi.it

396 M. Chernicoff
architecture) with technological (engineering and information technology based) and

management (economics linked) disciplines. For over fifty years industrial design
has been part of Universities and Scientific/Technological Institutions throughout
the world: first in the industrialized European countries, in North America and in
Japan; later in emerging and recently developed countries such as India, China,
Brazil, Argentina, Korea and Taiwan. In 1993 Milan, with its Politecnico, became the
birthplace of the Degree Course in Industrial Design, the first of its kind in Italy. The
Politecnico is a place of avant-garde training, a meeting point of different cultures;
it links architectural and artistic studies, with their creativity and focus on form, to
technical, scientific and engineering studies. In addition, this university is noteworthy
for its longstanding dialogue with the great entrepreneurial management tradition of
small and medium enterprises which enrich the territory of Lombardy and of Italy
in general. The history of the Degree Course in Industrial Design—Design Faculty
since June 2000 and now the Design School—is characterized by its constant desire
to innovative study, experimental development lines that respond to the real market
needs of contemporary society.1 The Design School in the Politecnico di Milano is
today the largest international university for the training of product, communication,
interior and fashion designers, both by number of students and of teaching staff. It
is active in the newly built Milan campus. The following figures give some idea of
its extent: approx. 5000 students, over 450 teaching staff and a further 400 or so
research and teaching assistants working in various disciplines. Degree programs
are organized at two levels: 3 year first level degree (Bachelor or B.Sc.) and a further
2 year second level degree “Laurea Magistrale” (Master or M.Sc.).
In this chapter I would like to share with you a very unusual didactical experience
in the university, far from the aseptic classrooms and the excess of virtuality of our
times. I will show you a special place where students can transform solid wood and
wood derivatives through manual tools (e.g. Japanese saws) or complex machinery
as traditional and new CNC drills. Since 2012, a new teaching and learning ethos
has been embedded in the prototype design workshop at Politecnico di Milano. This
is based on rediscovering the past design traditions in the light of the new present
and future technologies. This education experience has been applied in the Wood
Design course of the Design School (Fig. 1), a teaching course inspired and with
roots within the traditional Arts and Crafts academies, but also integrating new tools
and skills for design/production of wood products.
In the course, which is 50 h long, students can learn by ‘getting their hands dirty’
and at the same time be guided by the intelligence of their own hands. Over the past
five editions, more than 300 Italian and foreign design students took this course. One
of the key features of this course is that students are expected to express their own
ideas trough usable prototypes, making products and furniture with different styles
but always with a high technical and finishing quality. Given these rules, a crucial role
is played by the environment where the course takes place: the Design Department
of Politecnico di Milano workshop, equipped with machines, tools and qualified
technicians able to support students in giving shape to their ideas. This represents a
1 The School, http://www.design.polimi.it/en/the-school, [22/02/2018].

Learning by Prototyping: Wood Design Course Experience 397
Fig. 1 Wood design course

logo
special moment in which the students become makers of their own projects and can
check functionality, dimensions and aesthetic character.
I often say that in this course “we play or pretend to be carpenters to become more
competent designers”. We try to resolve a basilar problem: for the new generation
of students, without experience, it is not easy to decide HOW to join two pieces
of wood in a proper and durable way! This teaching approach can be associated
with Bruno Munari’s motto, when he evokes Confucius saying “I hear and I forget.
I see and I remember. I do and I understand.” For these reasons MAKING is the
backbone of this course. This teaching experience proposes basic exercises to acquire
knowledge about the material: building a jigsaw puzzle with different types of wood
or a kitchen cutting board. And we develop more complex projects: for example
games for children, lamps, traditional furniture, projects realized with waste materials
and many others.
2 Before You Run Learn to Walk
We are living a special technological development phase: the massive spread of 3D

printers and CNC machines makes it seem that everything can be prototyped or
produced very quickly, much too quickly I would dare to say! Obviously, models
and prototypes are very important tools to check the projects but these have to be
used with conscience and awareness about the motivations for which they are made
and above all having developed a minimum degree of design ability and knowledge
of materials (Hannessey and Papanek 1977). Would you give a powerful motorbike
to a person who cannot ride a bike? It is not just a question of the rational use
398 M. Chernicoff
Fig. 2 Example of Wood

puzzle_realized by Anastasia
Di Lorenzo, Gaia Migliore,
Cristina Pepe, Camilla
Pilotto
of resources and materials from the ecological point of view. Without manual and
analogical experiences in the practice of the project the risk and the greatest damage
would be to not dedicate the necessary time for basic and sensory training when
managing projects using such a special material as wood (Munari 2001).
In this sense, the course includes preparatory exercises: a wooden puzzle (Fig. 2),
a kitchen cutting board (Fig. 3) and the composition and creation of a samples-
collection of different types of wood and materials derived from wood such as ply-
wood panels and other semi-finished materials (Fig. 4). These exercises are made
exclusively with manual tools and traditional carpentry machinery; by cutting or
smoothing the wood by hand you learn in a direct and unforgettable way the main
characteristics of the essence: hardness, fragility, compactness of the fiber, aesthetic
characteristics of the grain. Using traditional band saws or circular saws, planes,
drills and other power tools it is possible to grasp the principles of machining and
to deal with the main problems. Only at the end of these preparatory exercises there
is the possibility, only if needed by the type of processing, to use new technologies
such as CNC machines with 3 or 5 axes.
In Italy the Matthew Crawford book entitled “Shop Class as Soulcraft: An Inquiry
Into the Value of Work” has been translated as “Manual labor as medicine of the soul”
and this book brings alive an experience that was once quite common but now seems
to be receding from society: the experience of making and fixing things with our
hands. Those of us who sit in an office often feel a lack of connection to the material
world, a sense of loss, and find it is difficult to say exactly what we do all day. For
anyone this risk exists: to take too theoretical paths against their own inclinations
and natural bents. The book seeks to restore the honour of the manual trades as a life
worth choosing. On both economic and psychological grounds, Crawford questions
the educational imperative of turning everyone into a “knowledge worker,” based
on a misguided separation of thinking from doing, the work of the hand from that
of the mind. Crawford shows us how such a partition, which began a century ago
with the assembly line, degrades work for those on both sides. But Crawford offers
good news as well: the manual trades are very different from the assembly line, and
Fig. 3 Example of cutting board_realized by Sonora Del Barba, Marta Fancellu, Alessandra Fuma-
galli Romario, Enea Gianfranco Colombo
Fig. 4 Samples wood exercise presentation
from dumbed-down white collar work as well. They require careful thinking and
are punctuated by moments of genuine pleasure. Based on his own experience as an
electrician and mechanic, Crawford makes a case for the intrinsic satisfactions and
cognitive challenges of manual work (Munari 2001).
With the same beliefs as Crawford, but in the area of wood design and with a 5-
year experience in the wood design course, we can begin to analyze the first results,
under three different lenses: (1) case studies of commendable projects, (2) collective
400 M. Chernicoff
experiences in the course, and (3) students who have undertaken the profession
independently or in companies in the sector.
In the next sections I will highlight some aspects for each strand. Through these
projects I will tell you about experiences and personal stories that merit to be shared,
including technical drawings, photographs and insights.
3 “Contact”, When the Best Design Found the Digital

Fabrication
Contact is a sensory puzzle. A game for blind people or to play blindfolded with the
mask provided inside the packaging. It is composed by 16 square blocks of 4 × 4 cm
solid wood and variable height that can be flanked just in a single way to rebuild the
soft geometry surface (Fig. 5). This project uses the sensorial potential of wood in
the best way, putting the relationship between the hands and the material at the center
of the game. We know that in the absence of the sense of sight all the other senses
develop greater sensibility and capacity. This happens when you try to cover the soft
surface letting yourself be guided through the hands and the mental recomposition
of the geometries inside the brain. The pleasantness of the contact with the well-
finished and polished beech wood completes this interesting proposal designed by
the students Luca Macrì, Clara Marcolin, Eleonora Poletti, Angela Razza, Enrico
Stefano Tirelli (Fig. 6).
The technical peculiarities for its realization can be summarized in the following
steps:
1. it starts from a single raw beech wood table big enough to obtain the whole puzzle;
therefore reconstituting the puzzle surface you also recompose the movement of
the wood grain;
2. division of the raw table with the band saw in four strips;
3. planing the wood strips with a thickness wood planer machine, 60 × 250 mm
thickness 50 mm;
4. parting the strips of wood with circular saw to obtain the 16 blocks 60 × 60 ×
50 mm;
5. fixing in the saddle the blocks maintaining the original order of the wood grain
to recompose the original table;
6. soft surface on the top of the blocks by the 5-axis CNC drill (a 3-axis machine
would have been sufficient using a semi-circle tool);
7. hand finishing, progressive processing with grit sandpaper 150, 250 and 400.
As you can see from the images, besides the realization of the prototype, the project
has been developed to the finishing details. The pack containing the 16 blocks of wood
works as a venue on which to play the game, made with great attention from the point
of view of graphics and “design for all” principles: in fact both the packaging and the
instruction book contain the Braille translation, and a mask to blind normal-sighted
players is included in the game.
Fig. 5 “Contact” sensorial wood puzzle, concept and book project
Fig. 6 “Contact” sensorial wood puzzle, prototype mix photo

402 M. Chernicoff
The points to be emphasized regarding this project are on the one hand the intrinsic
poetics of the game and on the other the use of machinery in the most intelligent
and appropriate way. We can see in the process of woodworking of this project the
mix between traditional and new technologies both in the design with 3D modeling
programs and in the production phase.
4 “Add a Seat to the Table” and “from the Tree

to the Design” Projects: Build Together, Learn Together
In these years of experience there have been two collective activities that I believe
brought together two important concepts: the transmission of knowledge from the
master to the apprentice in a direct way during the wood working, and the collective
construction as a team building exercise.
The first one, titled “Add a seat to the table” took place during the 2014 Milan
Design Week on the Bovisa campus of the Politecnico di Milano, headquarter of
the School and Department of Design. In this job-performance-event-construction,
teachers, students and technicians with guests and visitors designed and built a very
special piece of furniture in 5 days: a 20 m long table in the gardens of the university
campus (Fig. 7).
In this self-construction workshop everyone was allowed to become maker and
designer using panels and materials of different nature and format, donated by partner
companies of the initiative. This activity was an integral and complementary part of
the Design × Designers exhibition, developed over 1500 m2 in the exhibition area,
where it was possible to see posters, models and prototypes of all the courses and
masters of the Politecnico di Milano Education design system.
“The Table”, an icon of aggregation, integration and consumption of food, had in
itself part of the topics on which the future EXPO 2015, the universal exhibition held
in Milan, would have been based on. For this special furniture performance instal-
lation we built using 2.5 m modules with different solid wood and wood derivatives
and techniques but uniform in height and width of the top.
Technical information, organization and steps:
1. The project purpose was to build this table in just four days, so it was neces-
sary to obtain the material for the activity beforehand. We therefore contacted
some partner companies of the course willing to donate the material, and subse-
quently found ourselves designing and building with: plywood panels in okumè
and Sapelli mahogany from the Bellotti spa company; scraps of solid wood of
canaletto walnut, oak, iroko and teak kindly donated by Riva 1920; structural
blocks and cross-lam panels used for the construction of wooden houses sup-
plied by Binderholz and Mozzone; OSB panels (oriented strand board) made of
poplar wood produced in Italy by the Bonzano group.
2. The place chosen for the construction and installation was the garden in front
of the building that houses the main lectures of the School of Design. This is
Fig. 7 “Add a seat to the

table” project, collective
activity mix photo
construction during the
Milano Design Week
called “The Oval”, and it is a usual meeting point for students. For the occasion,
25 m2 of gazebo housed our nomadic laboratory with the necessary equipment:
jigsaws, drills, screwdrivers, clamps and manual tools.
3. The goal was to build 5 m of table a day, two modules of 2.5 m with the respective
seats. In this way, the installation would progressively increase to the desired
length in the four days of activity. Every day started with a brief project, the
selection of the materials to be used and of the work phases, organizing the job
404 M. Chernicoff
Fig. 8 “Add a seat to the table”, 20 m long table realized with the students. Credits photo Matteo
Bergamini
in a workgroups and involving guests and visitors. Every evening, the modules
had to be realized and ready for use. It was a very beautiful challenge!
4. The constructional technical features of each module reflected the design inter-
pretation and the potential or intrinsic limits of each material. Specifically they
were made as follows: one module with “dry joints” (without glues or screws)
where a band of denim fabric held the table-top system in tension; two modules
with 4 cm thick OSB tops and 10 cm thick cross-lam legs (very heavy duty mod-
ules!); two tables entirely made of plywood panels and shaped legs of refined
geometry. One module, of particular interest, was made with solid wood strips
that came from production waste as too thin or from the sapwood (the less noble
part of the trunk): nothing is destroyed, only its form is changed!
5. The temporary installation, held from Tuesday to Friday of the Milan Design
Week, hosted the final event on Saturday evening. It then remained available to the
students for the following three months, offering them a functional meeting point
for lunch or relaxing breaks during the Italian spring-summer season (Fig. 8). At
the end of the cycle the modules of the tables were donated to the “Cultivating”
solidarity group that manages the urban vegetable garden of the Bovisa district,
once again trying to give a second life to the available resources. After four years
it is still possible to see some table or bench that has survived the atmospheric
agents and the intensive use by the students.
More than a hundred people contributed differently to create a single meeting
place with 80 seats during those days. It was an educational experience for the
Fig. 9 Milano city

Christmas tree 2017–2018
students/makers and a great cohesion activity among the participants: professors,

technical staff, students and visitors; an interesting team building exercise applied!
The second project of design and collective construction worthy of insights is
called “From the tree to the design”, and is born from the collaboration between
the Politecnico di Milano and Sky Academy, the TV channel of the Sky TV multi-
media group that supports initiatives to promote knowledge and training. The project
launched in September 2017, and still underway, will lead to the transformation of
the Christmas tree located in Piazza del Duomo in Milan in December 2017 (Fig. 9)
in a new wood installation designed and built for an urban area of the city: the Sky
square in Milan (zone n. 4, a peripheral area of Milan).
To reuse in a creative way the wood of the Christmas tree the backbone of the
initiative is based on two intensive courses: the first one is a design workshop where
the students have to design furniture and installations for the city using the 30 mt
long spruce used in Piazza Duomo; the second one is a constructive workshop where
materially the 5 m3 of solid wood are transformed inside the wood laboratories of
the Design Department.
The 5-day design workshop held in February 2018 aimed to provide students with
a training and methodological approach that can enable future designers to intervene
406 M. Chernicoff
in the complexity of design processes. An intuitive and practical professional pro-

file, able to range from the brief’s presentation to the definition of the concept, and
concretely interact from the first sketch to the design of a complex object and sub-
sequently to its realization, up to its positioning in the urban space of the city. The
methodological principle applied to this course of study is “Learning by designing”.
The constructive workshop, scheduled for May–June 2018, aims to materialize
the prototype chosen by the jury downstream of the design workshop. The training
proposal of the course, mainly practical, is set according to the different stages of
wood processing up to the finished product ready for installation. Organized in work
groups, which will be defined every day on the basis of calendar activities and with
the support of technical staff, students will cut, smooth and assemble the different
pieces of the project. Some woodworking of complex geometry will be realized with
new generation instruments as CNC machines with 3–5 axes. The entire piece will
be assembled at the Bovisa Campus of Politecnico di Milano and will be ready by
the end of June. The methodological education applied to this workshop is “Learning
by doing”.
The overall project is not yet complete and for the moment it is possible to share
with you the winning project and three of the 23 proposals chosen by the jury (Fig. 10).
The group composed by the students Luca Mussi and Luca Polcari has elaborated
the most convincing proposal called “2nd chance”, which is now under construction
at the Politecnico Laboratories (Fig. 11). The idea is to repopulate Piazza Santa
Giulia offering people a multi-purpose recreational space free of schemes, which can
involve people and leave them the opportunity to express themselves: this is the main
concept of the project. The installation in the square is suitable for sports and games,
with particular attention to the alternation of geometric shapes that lend themselves
to multiple interpretations and uses. Technically the trunk of the tree is processed
in a single type of raw table from which all the components of the installation are
obtained. In this way, machining, waste and assembly operations are reduced, making
simplicity the key element of 2nd Chance. The mainly horizontal development of
the project proposes, arranging the various modules according to a dynamic lay out,
an enhancement to the square’s visual impact. Furthermore, after checking the half-
light of the central corridor of the square during the night, a treatment based on
phosphorescent paint on the sides of the modules could be considered to increase the
visibility of the installation, creating suggestive optical effects. With the involvement
of exceptional partners such as Sky and the City of Milan, it is possible to imagine
sporting and entertainment events that will expand and support the vast panorama of
Milan’s events.
This educational project was carried out by the Politecnico di Milano in col-
laboration with the professors Davide Bruno, Emmanuele Villani and the designer
Giacomo Moor. In this case the professionalizing value of the experience for the
students should be highlighted. The multiple constraints: limited available material,
location of the project and timing are comparable to those of real work. As a training
center, preparing future designers for the profession is one of the keys to doing one’s
duty well.
Fig. 10 2nd chance project, winner of the workshop to transform the Christmas tree
5 Students Wooden Start-Up Companies
In continuity with the conclusions of the previous paragraph, in this section I will
present you relevant experiences of delicate passage between the world of study
and that of work. Case studies of students who completed the degree course have
408 M. Chernicoff
Fig. 11 Students Luca Mussi e Luca Polcari with the 2nd chance prototype during the wood
working workshop
transformed the experiences gained within the Wood Design course in opportunities
to start a profession or create new businesses founding start-ups in the wood industry.
This first case study summarizes the new trends in the market and the innovative
paths of the design profession: limited edition series, product customization, integra-
tion between traditional woodworking with new production technologies and online
sales. In October 2017 three ex-students evolved the kitchen cutting board exercise
in a real line of products and founded the QUID brand (Fig. 12).
This company started from the observation that in today’s fast-paced world we
often do not become attached any more to the possessed objects because they are
soon ephemeral and obsolete. Instead they, with their projects, propose to recover
and enhance that desire to own an object, to preserve it and make it become part
of their lives. The intention is to create unique products able to enclose memories
and stories to tell: “Wood Desires”. To do this, the strategy is to combine the typical
innovation of Made in Italy design with the know-how of craftsmanship, so they
chose to use wood, “because only those who know how to wait will see that our
products are alive, change and enrich with details and new colors, always remaining
up to date as the world of today imposes.”2
Laura Guerrini, Giacomo Guaragna, Simone De Stefani are the QUID founders.
Their personal design history starts from knowing each other during university studies
in Milan where a strong friendship developed, allowing them to grow and improve
both as persons and as designers and to give life to this project (Fig. 13). Their
goal is to give life to products that can transmit emotions to those who own them,
establishing a special bond between person and object; therefore they decided to
2 https://www.quidwood.com, [7/04/2018].
Fig. 12 QUID website. Wooden start-up founded by three wood design course ex-students
redesign and ennoble common objects, rediscovering the value of craft processes,
and accompanying the m with a continuous research and experimentation.
The second case study is similar to the previous one and in this case the company
name is called MIDA—More Than Gold, wooden jewelry-. In 2013, within the Wood
design course, an enthusiastic group of students worked with small scraps of precious
or particular woods and developed a series of rings and fashion accessories. From
this experience Simona Gorini and Giuditta Sartori founded a small company that
now produces and markets them.
To conclude this point, focusing on the profiles of former students of the Design
School of Politecnico di Milano and professional wood design careers is worth talking
about Giacomo Moor. He is now an established international designer who has
focused research and manual labor on wood since his university careers. Born in
Milan in 1981, Giacomo Moor begins to work in a carpenter’s workshop at an early
age. During his years at the University, under the guidance of professor and architect
Beppe Finessi, he started to bring together rigorous planning with his fascination for
410 M. Chernicoff
Fig. 13 Laura Guerrini, Giacomo Guarangna, Simone De Stefani. QUID founders
the beauty and versatility of woodworking. In university he completed his studies

with a wood design degree thesis entitled “Valuable Imperfections. Cabinetmaking,
from sculpture to design”.
Giacomo now has a very special workspace with an integrated design studio and
wood laboratory (Fig. 14). His lab has become an articulated and multidisciplinary
facility, a place of exchange and growth. Alongside an additional team of designers
and carpenters, Moor exploits expertise, innovation and research as well as tradi-
tional working methods and utmost craftsmanship to fulfil the needs of a variety of
clients. Always striving to improve his woodworking techniques, Moor combines
the precision in design with the hands-on experience of an artisan, articulating his
work in three main areas: Limited, Industrial and Bespoke.3
Moor executes and refines ideas: he designs products for companies, creates lim-
ited editions for design galleries and produces one-of-a-kind pieces for private clients,
always supervising the entire creative process (Fig. 15). In 2016 he is nominated Best
Young Designer 2016 by Elle Decor Italia at EDIDA and is granted the Best Young
Designer prize from Salone del Mobile di Milano Award; in 2017 he wins the Wall-
paper Award.
3 https://www.giacomoor.com, [13/05/2018].
Fig. 14 Wallpaper’s article about Giacomo Moor
Fig. 15 Outline project, Giacomo Moor for Acerbis company

412 M. Chernicoff
Fig. 16 PoliMi student using the band saw during the course
6 Conclusion
The methodology and practice of the project are at the heart of the designer’s profes-
sion and consequently of teaching and learning within the School of Design of the
Milan Polytechnic (Sacco and Tagliabue 2003). Physical modeling and prototyping
are the aspects I have always dealt with; the selection made for this writing arises
from my daily experience, and I consider these tools fundamental and indispensable
tools, that allow to shape, verify and communicate ideas and intuitions.
I am sure that these experiences would have been impossible to achieve in the
absence of the appropriate context for this activity: the instrumental laboratories
of the Design Department (Fig. 16). The system of laboratories for teaching and
services was established in 2002 on the Politecnico di Milano’s Bovisa Campus. It
occupies an area of around 10,000 m2 of the building which is home to the School
of Design’s teaching activities. In addition to supporting these activities, for over ten
years it has provided services for businesses, associations, bodies and professional
organisations in various areas ranging from communication design to product design,
from interiors to construction and from textiles to fashion. A special mantion goes
to the Paolo Padova Exhibit Lab, a space devoted to planning, implementation and
experimentation in the world of interior design divided in home, retail, exhibition,
work and transport spaces. Its activities in all spheres adopt a concrete experimental
assessment approach to project implementation. Building spaces on a natural scale
and prototyping décor and layout elements—these are the characteristics which mark
out the lab’s work processes. Spaces, equipment and technical staff know-how tar-
get support and experimental teaching methods, planning and implementation of
structures and systems for exhibitions and events, experimenting with and using new
materials in the furniture and interior design sectors.
We have all seen where Renzo Piano works, his atelier, his workbench and, if I am not so
naïve as to think that today, in the first person, he is concretely engaged in this workstation,
I observe that the presence is not secondary in his study of the subject and of the tools to
modify its shape is a significant fact. (Padova 2003)
As professor, this kind of didactical approach through direct constructive activities

I hope can help to understand how such a practical approach about representation,
design and knowledge transmission can become a solid basis education for future
generations of designers.
Acknowledgements Livio Riceputi, assistant of the wood design course. Mattia Favalli, Fabio
Manfredini, Camillo Sangiorgio and Matteo Piccoli. Special thanks to all the students, teaching is
the most interesting way to learn together!
References
Crawford M (2009) Shop class as soulcraft: an inquiry into the value of work. Penguin Press, New
York
Hannessey J, Papanek V (1977) Nomadic furniture (arredamento mobile). Longanesi & Co., Milano
Munari B (2001) Wood working (lavori in legno). Orsa maggiore edizioni, Milano
Padova P (2003) Hands intelligence (Intelligenza delle mani). Università di Trento Sociology
School. Degree thesis
Sacco E, Tagliabue S (2003) Do it yourself. Politecnico di Milano Design School. Degree thesis
Parametric Kerf Bending:
Manufacturing Double Curvature
Surfaces for Wooden Furniture Design
Mara Capone and Emanuela Lanzara
Abstract Geometric understanding allows us to solve many problems in manufac-

turing complex surfaces and geometric rules can lead the project from theory to
practice (Pottmann in Architectural geometry. Kindle edn, 2015). Digital fabrication
technologies are evolving and becoming more and more widespread. Knowledge of
fabrication methods available and parametric design tools, based on geometry, are
changing the designer way to think. Advances in this field promote experimental use
of new materials but also innovative use of traditional materials, such as wood. Our
research in the field of non-developable surfaces fabrication move from “paneling”
to “kerfing”, or “kerf bending”. This subtractive manufacturing technique consists
in transforming a rigid material in a flexible one and the problem to solve is how
to cut the flat shape to obtain the 3D design surface. This is a quite simple question
to solve for a developable surface, but it is a very complex problems for a double
curvature surface manufacturing. The aims of our research are: analyzing the rela-
tionship between the curvature of surfaces and kerfing techniques; testing different
ways to optimize the design of cutting patterns to bend a rigid materials; testing
different ways to manufacture 3D surface from 2D panel. We are manufacturing
some prototypes that are made up of pieces of developable helicoid and hyperbolic
paraboloid.
Keywords Parametric kerfing · Flexible wood · Digital fabrication

Wooden furniture design · Wooden complex surface
M. Capone (B) · E. Lanzara

University of Naples Federico II (UNINA), Naples, Italy
e-mail: mara.capone@unina.it
E. Lanzara
e-mail: emanuela.lanzaral@unina.it

416 M. Capone and E. Lanzara
1 Introduction
Geometric understanding allows us to solve many problems in manufacturing com-

plex surfaces and geometric rules can lead the project from theory to practice
(Pottmann 2015). Digital fabrication technologies are evolving and becoming more
and more widespread. Knowledge of fabrication methods available and parametric
design tools, based on geometry, are changing the designer way to think. Advances in
this field promotes experimental use of new materials but also innovative use of tradi-
tional materials, such as wood. Our research in the field of non-developable surfaces
fabrication move from “paneling” to “kerfing”, or “kerf bending”. This subtractive
manufacturing technique consists in transforming a rigid material in a flexible one.
The aims of our research are: analyzing the relationship between curvature of sur-
faces and kerfing techniques; testing different ways to optimize the design of cutting
patterns to bend a rigid material; testing different ways to manufacture 3D surface
from 2D panel. To do that we have defined these main Research Questions:
– How do you make wood more flexible?
– How do you manufacture a double curvature surface from a flat panel?
– How do you choose the best pattern for manufacturing your wooden furniture
design?
2 Theoretical Background—Geometry
Knowledge of geometric properties of surfaces is crucial for resolution of many

manufacturing problems (Capone 2012). Developability is an important intrinsic
property of a surface, because it allows its manufacture starting from a flat “strip” of
a “flexible” and “non-deformable” material.
Our research, in the field of wooden curved surfaces fabrication, move from
developable surface manufacturing and from “paneling” to “kerfing”. This technique
consists in transforming a rigid material in a flexible one and the problem to solve is
how to cut the flat shape to obtain the design surface. This is a quite simple question
to solve for a developable surface but it is a very complex problem to address for
double curvature surface manufacturing.
In our research, we can summarise the methodological approach for wooden
curved object manufacturing that moves from theory to practice:
– theoretical framework;
– kerf bending patterns classification;
– analysis of different approaches (patterns uniform distribution vs. patterns opti-
mized distribution);
– tests for single curvature surfaces manufacturing (k 0 cone and developable
tangential);
– tests for doubly curvature surfaces manufacturing (k < 0 hyperbolic paraboloid);
– analysis of results and formulations of new hypotheses.
Parametric Kerf Bending: Manufacturing Double Curvature … 417
2.1 Differential Classification of Surfaces
Differential classification of surfaces, introduced by Leonhard Euler (1707–1783)

and subsequently used by Monge, allows us to group surfaces according to cur-
vature definition, which will be precisely defined by Carl Friedrich Gauss (Gauss
1902), in four categories: surfaces with zero curvature, surfaces with positive curva-
ture, surfaces with negative curvature and surfaces with variable curvature.
The curvature of a curve in P is k, where k 1\r and r is the radius of the osculating
circle of the curve. We can define the main sections of a surface, the sections obtained
with planes passing through the normal to the surface in P, with minimum and
maximum curvature.
As we know, Gaussian curvature is the product of the two main curvatures, so
it can be positive, negative or zero: it is positive when the osculating circles of the
main sections are on the same side of the tangent plane, negative when they are on
opposite sides, zero when one of the two main sections is a straight line.
The surfaces with zero curvature are specific ruled surfaces, also called devel-
opables (Fig. 1).
Fig. 1 Differential classification of surfaces (Lanzara 2015)

2.2 Developable Surfaces
A developable surface is a surface for which every generatrix intersects the generatrix
infinitely close, then:
– when they intersect on curve c (edge of regression), we have a tangent developable;
– when they intersect in a point V (the edge of regression is a point), we have a conic
surface;
– when they intersect at infinity, we have a cylindrical surface.
For this reason, we can group developable surfaces in three groups: conical sur-
faces, cylindrical surfaces and tangential surfaces. As we know, you can obtain a
tangent developable using the tangents to a generic curve c, the directrix c is called
the edge of regression (Migliari 2009, pp. 213–218).
We can easily generate cylindrical or conical surfaces using a 3D modeling soft-
ware, extruding the curve in one direction or to a point. It’s more complex generating
a tangential developable.
In Descriptive Geometry a tangential developable is generated by motion of tan-
gent line on a spatial curve (the edge of regression). Using algorithmic modeling we
have developed a definition to construct developable ruled surfaces using a generic
spatial curve. This spatial curve can be imported by Rhino or parameterized in rela-
tion to specific needs.
Dividing the assigned spatial curve (the edge of regression) in n parts, our algo-
rithmic definition allows to generate the surface by constructing n lines (generatrix
of the surface) passing through the n points and tangent to it. It is possible to obtain
infinite developable surfaces by modifying the length of the generatrix and the edge
of regression. This surface may be cut to define the edge which is otherwise auto-
matically generated as a function of the generatrix length (Fig. 2).
A tangential developable specializes if the edge of regression is a cylindrical
helix: the surface generated by motion of a tangent line to a cylindrical helix is a
developable helicoid.
Fig. 2 Tangential developable and developable helicoid

The case of the developable helicoid is the simplest, in fact, if the edge of regression
is a cylindrical helix, in order to generate the surface we can construct the tangent at
a point P and then make it move along the helix.
In this case, generative modeling is powerful tool, useful not only for reiterating
procedures but for verifying theories. In fact, a tangential developable can be unrolled
with some unavoidable approximations, as two consecutive generatrices intersect
each others on the edge of regression only in an infinitesimal neighborhood, with n
tending to infinity.
We realized manufacturing tests using pieces of developable helicoid with the aim
of building a wooden lamp.
2.3 From 3D to 2D Shape to Cut
One of the main questions to manufacture a 3D surface from a planar panel is to find
the 2D shape to cut.
If we have to fabricate a developable surface it is always possible but, if we have to
fabricate a complex surface, we have to decide how to determinate its approximate
develop. It’s very easy to find conical or cylindrical surfaces development using
traditional methods or digital tools, on the contrary it is very difficult to find the
unrolled shape of tangential developable surfaces. Using 3D modeling, there is a
command that is able to automatically unroll both conical and cylindrical surfaces.
The method for finding the unrolled shape of a tangential developable is more
complex, in this case differential geometry application is crucial.
Monge uses the principles of differential calculus to study developables sur-
faces properties (Migliari 2009, pp. 106–108). Each developable surface can be flat-
tened onto a plane without distortion and, in a limited region, without overlapping.
The unrolled shape of the surface generated by the infinite tangents to a space
curve is obtained by considering n generatrices and flattening onto plane the surfaces
included between two consecutive generatrices. Two consecutive tangents, t1 t2 , can
be considered coplanar if they are very close, then: if we rotate t2 t3 around t2 , and
we repeat for the following tangents, we can find the unrolled surface.
The unrolled surface depends on the edge of regression.
Using algorithmic modeling first, we have done a tools that allows to develop a
developable helicoid and then we have done another tool that allows to develop any
developable tangential.
We know that helixes on developable helicoid turn into concentric circles, for
this reason the edge of regression will turn into a circle whose radius depends on R,
the radius of curvature of the helix. To draw the unrolled shape of the developable
helicoid, it is enough to fix the length of the assigned generatrix (for example AB)
on the helix development and drawing a concentric circle with radius OB (Fig. 3).
Starting from this theoretical framework and using a generative algorithm in
Grasshopper, we have develop a method that allows to find the unrolled shape of
any developable tangential.
Fig. 3 Tangential developable and developable helicoid: finding developed surfaces
If we divide the edge of regression into n parts and we consider n tangents (gener-
atrices of ruled surface) we have that two consecutive tangents intersect on the edge
of regression. This is true only in a small, infinitesimal neighborhood. In fact, if we
divide the edge of regression into n parts and we consider two successive tangents, t1
and t2 , from points 1 and 2, we define the non-flat quadrilateral A12B. If we extend
B2, it intersects the A1 in the point 2* (Fig. 3). Therefore, we can construct the flat
triangular face A2*B. In the same way, we extend D3, we have the B3*D triangle,
thus the surface can be divided and approximated in the triangles faces: A2*B, B3*D,
D3*E, D4*E… It is true when n go to infinity, so the point 2 goes to point 1, 3 to
2 …. We can unroll the surface composed by n triangular faces. The approximation
of the unrolled surface obviously depends on n. We can evaluate it comparing the
metric values of the 3D surface, the length of the edges and the area, with the unrolled
shapes that we have constructed (Fig. 3).
It may happen that the surface configuration is such that portions of unrolled
surface overlap with the others, in these cases, it is necessary to divide the design
surface into parts in order to manufacturing it.
2.4 Non Developable Surfaces
There are two different approaches to define approximate develop of non developable
surface: first we can design the surface and then we can find the best flat shape to
manufacture it, or we can design 3D surface morphing a planar surface.
We can divide and approximate the 3D surface in order to cover a complex shape
by developable strips, which can be unfolded to the plane in an isometric way, without
stretching or tearing (Pottmann 2015).
In this way we can fabricate 3D complex shape using materials that can be bent
in one direction or rigid material with no possibility of being bent at all. In our
research we have considered materials and manufacturing techniques that can solve
approximation problems allowing non isometric transformation of the panel thanks
to “cuts”, kerfing, or “overlapping”, bending.
The kerfing technique, in addition to make flexible the panel, can also make it
deformable. Consequently, it is possible to manufacture double curvature surfaces
using this technique.
Our tests to unroll the developable helicoid and to unroll a tangential developable is
the starting point of our research in progress about manufacturing of non-developable
ruled surfaces. Our tests regard non-developable surfaces and in particular we are
dealing with problem of identifying the approximate unrolled surface according to
the different ways of cutting the panel. We can find approximate unrolled shape of
non-developable surfaces useful for certain applications.
We have studied non-developable surfaces and in particular the case of hyperbolic
paraboloid, to highlight some of the problems and to define some possible approaches
to transform a non-developable surface into a flat surface that, with better approxi-
mation, is able to preserve the main features of 3D surface.
One of the main research goal is to highlight, through the applications, how these
approaches can influence the figurative outcome and the manufacturing process.
The hyperbolic paraboloid is a ruled surface that may be generated by motion of
a straight line that is parallel to a fixed plane. It is a non developable surface because
two consecutive generatrices are always skew lines and Gaussian curvature is always
negative.
There are several tools that allow you to automatically obtain the approximate
unrolled shape of a non-developable surface: using Rhino the command “smash”
and the command “squish”.
The critical analysis of the results obtained using 3D modeling software is part
of our experimentation.
Using the “smash” command we can automatically generate an approximate
unrolled shape for a double curved surface, but using this flat shape we can recon-
struct the real 3D shape only if we use a deformable material. The “squish” command
uses a different algorithm, performs the smoothing of meshes or 3D nurbs surfaces,
modifying the starting area, allowing the display and control of the local compres-
sion and stretching zones. Applying the “smash” and “squish” commands to a piece
of hyperbolic paraboloid, used in our tests, we obtained different shape (Fig. 4). We
have done these observations based on results: the area changes compared to the
real one and the generatrices of one of the two groups deform themselves. It follows
that it will necessarily be breakings and/or overlaps to transform the flatten shape
into 3D designed shape. In fact, if the generatrix AD becomes curve, it turns into
the curved edge A*D*, this must be deformable, therefore the cuts must be made to
allow the curve A*D* to assume the configuration straight of the designed shape.
Similarly, if the generatrix AB is deformed, it will be necessary to allow that the
curve A*B* is able to transform into the straight segment AB (Fig. 4). We tested
different methodologies to simulate deformation according to kerfs made, using gen-
Fig. 4 Non developable surface: approximate develop of hyperbolic paraboloid for wooden man-
ufacture
erative modeling. Our goal is to identify processes and to develop tools, to define the
approximate flatten shape of a double curvature based on the knowledge of geometric
properties.
3 Theoretical Background—How to Make Wood Flexible
The kerfing technique consists in subtracting material in some points of the panel in
order to improve its flexibility. The maximum radius of curvature, that the panel can
reach, depends on the material, on the panel thickness, on the type of kerf and on the
different distribution ways on the panel in relation to the curvature of the design
surface.
The main goal of our research was to analyze and to classify the different kerf
pattern geometries in relation to the nature of the surfaces and especially to the
Gaussian curvature. We have used 2 mm MDF, 3, 6 and 120 mm plywood to make
our tests.
There are three different traditional ways to bend wood: bending wood with the
steam box method, bending wood with the lamination method, based on using of
molds, and bending wood with the kerf-cutting method, based on subtractive manu-
facturing technique. We are interested in the latter case.
Focuses on creating digital models for kerf bending, a traditional method for
creating efficient and flexible forms out of rigid materials, we have defined three
different ways to remove material to make wood more flexible (Fig. 5):
cutting only on one side of the panel;
cutting on both sides;
cutting through thickness of the panel.
Fig. 5 Kerfing: how to make wood flexible. Different way to cut
3.1 Cutting Only on One Side of the Panel
The traditional kerf bending technique consists in a series of cuts on one side of the
rigid panel to transform it in a flexible material that can be bent and it looks like a
continuous surface on the other side. The distance between the kerfs and their depth
will determine panel flexibility and the radius of the bend. We can find the geometric
spacing of the cuts so when bent, the inside edges of the cuts join to create the curve
(Fig. 6). The depth of the cuts and remaining wood thickness for the bend depends
on the different kind of wood, material behaviors, and panel thickness.
To bend wood with minimum loss of strength, right kerf depth and spacing could
be determined using tests. To avoid stress concentration at the end of the kerfed length
Fig. 6 Cutting only on one side of the panel: regular kerfing versus parametric kerfing
caused by use of constant kerf depth we can vary kerf depth gradually in relation to
the stress distribution.
Using constant kerf it is possible manufacturing only single curvature surfaces.
The evolution of this technique consists in changing the shape and the distribution
of the cuts according to the curvature.
The cut shape depends on fabrication technique, you can cut using different kind
of saw or CNC machine (saw, milling or laser cutter). In the first case you can use
a radial arm saw, table saw, band saw and an indexing system to keep the spacing
uniform. The more uniform the spacing is, the more uniform the curve will be. On
the contrary, if you use laser cutter and parametric tools, you can obtain different
distribution of the cuts and cuts shape in relation to design form (Menges et al. 2016).
The research project, designed by Brad Crane, Andrew McGee, Marshall Prado
and Yang Zhao (Performative Wood Studio, Prof. Achim Menges, Harvard GSD
2010), developed an integrated computational design tool and robotic manufacturing
process that allows programming the bending and twisting behavior of tensioned
wood elements through specific kerfing (Fig. 6).
Starting to study of regular kerfing and applications in boat construction, furniture
making and other fields, technique for fabricating wooden parts bent in one direction,
this project explored how the computer controlled variation of kerf depth, length,
frequency and orientation allowed for achieving more elaborate bending and warping
figures (Menges 2011).
3.2 Cutting on Both Sides
Advanced methods are developed by traditional kerf bending: cutting on both sides
of the board. We are analyzing the main two different approaches: uniform pattern
and research of optimized kerfing.
In the first case, such as Dukta Duna incision types, the double sided incision grid
makes wooden panel able to bent and twisted. The incisions cross over and span both
sides meaning it can be bent into 3D shapes: it is possible manufacturing a double
curvature surface.
Instead of creating basic cuts there are some experimentations about optimized
solutions research through parametric tools. Most of them are looking at creating
more precise cuts related to the curvature of design form. Very interesting case
study, applications of traditional kerfing method using parametric tools, is recent
experiments by LMN Architects. They are “considering a technique somewhere in
between a kerf bend and a series of miter joints, which they call a mitered kerf. With
this technique, a triangle is cut (on a laser cutter or a router) out of the material. The
triangle’s shape is determined by the curvature of the generating form so that the
faces within each kerf cut become flush when the object is bent into place. By using
a kerf pattern to create the circle, they can save about 33% of our material vs. laser
cutting the circle directly. The material savings increase significantly when the tool
is used for more complex and volumetric forms. LMN are using Grasshopper and
Fig. 7 Cutting through thickness of the panel
Hoopsnake to generate the cutting templates for these forms, and the digital process
works like the fabrication process in reverse. The definition starts with a complete
digital model from which contoured polylines are created for the cut” (Fig. 7).
3.3 Cutting Through Thickness of the Panel
One of the main goals of our research is to classify the different kerf bending tech-
niques in relation to geometry and curvature. We have tested some of them to try
to achieve different results, most of which were failed attempts at making a doubly
curved surface.
Following a gradual study of optimizing the management and fabrication of curved
surfaces from 2D planar materials, we have chosen a case study: the hyperbolic
paraboloid that is a doubly ruled surface with negative curvature.
The challenge is to find a pattern which enabled bending in two directions and
research of optimized solution using parametric tools. We are testing some different
techniques to manufacture the same surface.
We have grouped kerf bending into four, according to the type of cut: spirals,
fringes, zigzags and slit (Muñoz et al. 2011).
The experimentation is part of a broader research that has involved the study of
different kind of “cutting through kerf” in relation to the Gaussian curvature of the
designed surfaces.
Our tests show that there are some problems to manufacture double curvature
using many of the analyzed patterns.
We have grouped the “cutting through kerf” in four main groups (Muñoz et al.
2011): spiral kerf, fringe kerf, zig zag kerf and slit kerf (Fig. 7).
Fig. 8 Different approaches for manufacturing double curvature surface: regular kerfing or para-
metric kerfing
1. spiral kerf: it is single or double kerf from the center to the border, it is continuous
and it makes the panel flexible in orthogonal direction of the plane. 3D shape
depends on the cut line (Fig. 8).
2. fringe kerf: a series of kerfs, distributed on the surface according to a geometric
rule, come up to the edge and break it. Flexibility depends on kerfs length and
geometric rule.
3. zig zag kerf: a series of kerfs that start from the edge of the panel, they are
staggered. Flexibility achieved depends on frequency, overlapping of the kerf
and on panel thickness. The distance between kerf and the kerf shape is very
important for furniture design. Flexibility can be multidirectional.
4. “slit” kerf: they are narrow kerfs that follow a lattice order or pattern. Generally,
the plan becomes flexible along the direction of the kerfs and this flexibility
can be controlled by providing sectors not affected by kerfs or with different
densities. The empty space is very important in this case, because otherwise the
contact of the edges is a limit.
First we studied the geometry of the kerf and then we studied the different ways of
the pattern distribution on the panel, identifying two different approaches: uniform
distribution or optimized distribution (Fig. 9).
The studies carried out show that the maximum radius of curvature, that a panel
of a given thickness can reach, depends on the type of cut, but also on the pattern
distribution on the surface.
Therefore, we have identified some fundamental variables and, through the use of
parametric representation tools, we have fabricated a series of prototypes using a slit
pattern uniformly distributed on the panel in order to define the type of cut according
to the maximum curvature of the designed surface.
The kerfing technique increases the flexibility of the panel by subtracting material
and modifying its resistance.
Fig. 9 Cocoon kerfing experiment by IaaC, Institute for Advanced Architecture of Catalonia. Make
wood flexible in more than one direction
The goal of the most innovative research in this field is to define processes able to
find optimized solutions: “cut only where it is needed” to get the designed surface.
We have analyzed some fundamental case studies to understand the different ways
of kerfs pattern distribution to optimize surfaces fabrication with variable curvature.
4 Experimentation Process—Different Approaches
The experiments we have performed are built upon analysis of four case studies that
we have selected to understand different approaches: LMN Architects experiments
for furniture, Cocoon kerfing experiment by IaaC, Institute for Advanced Architec-
ture of Catalonia, The Wooden Waves by Mamou-Mani Architects\BuroHappold
Engineering and Gregg Fleishman’s works.
Cocoon is a kerfing experiment, the patterns design are the result of various tests
to create a multi-directional bending surface to fabricate a lamp.
The pattern design starts from the geometry cut shape that is overlapped on inter-
section points of grid layout. The cocoon pattern is used to fabricate a double positive
curvature surface (Fig. 9).
The Wooden Waves by Mamou-Mani Architects and Buro Happold Engineering,
the Golden Winner 2016—American Architecture Prize, is an architectural installa-
tion suspended, not structural. It is a result of testing in lattice-hinge design, makes
use of flat, stock plywood.
The free form is generated using a parametric slit kerf pattern, in this case the
laser cut lines alters the global proprieties of plywood sheet making it locally more
flexible. More than a hundred prototypes were tested to inform the digital model and
master the curvature of the final piece. In this case the surface curvature is variable.
Fig. 10 Gregg Fleishman’s design object
Gregg Fleishman has been working on folding plywood furniture design for 1975.
His works are often obtained by a single panel of wood using a combination of zig
zag kerf and slit cuts. The cut shape is very important for Gregg Fleishman’s design
object and strength of material too.
Fleishman’s chairs are made of laminated birch, 20 mm thickness, manufacturing
using CNC fabrication technique (Fig. 10). The 3D surfaces are single curvature
surfaces and there is no torsion stress.
4.1 Prototypes
We tested some pattern using parametric tools and defining assumptions and vari-
ables in relation to kerf technology. With regard to this question we have considered
the actual machines used for kerfing. In particular, we are analyzing the difference
between laser cut and CNC milling, and how they may influence the design process.
We have designed our experiment using 3 mm fir plywood and 2 mm MDF. The
goal is to know what is the maximum bending limit of a material, according to
kerf patterns geometry and how to manufacture same 3D shape using material with
different thickness.
In our case study manufacturing techniques are used as a design tool. Our aim is
to test this methodological path to explore bending behavior of the material using
kerfing technique.
The results of our tests show us that it isn’t the subtraction operation itself that
increases the bending ability of the surface, but it is lay out pattern distribution
that might increase the flexibility. In that way the planar surface begin to perform
anisotropic behavior (Güzelci 2016).
We have tested two patterns and their variations to study the relations between
patterns geometry and material property and some different kind of kerf pattern,
especially slit kerf.
The shape is the start point of our research and not a casual result of trials patterns,
among many tests to put on the plane one pattern or mixed patterns to manufacture
an assigned doubly curvature surface.
To do that we have produced several prototypes.
We have considered two different approaches:
– pattern distribution on surface using a grid layout on straight line generatrices;
– pattern distribution based on curvature and geometric properties of 3D surface.
In this paper we show some case studies that we have fabricate using ruled sur-
faces:
lamp prototype 1—cone;
lamp prototype 2—developable surface composed by parts of developable helicoid;
lamp prototype 3—non-developable ruled surface composed by parts of hyperbolic
paraboloid;
stool prototype—non-developable ruled surface composed by parts of hyperbolic
paraboloid.
4.2 Pattern Design Process
Pattern design process usually starts from grid layout which has a predefined size. A
single geometry is overlapped on intersection points of grid layout as a one or two
dimensional array organization (Güzelci 2016). Variables like size of pattern units,
distance and direction between pattern elements, thickness of the material can be
modified to explore the strength of the material.
There are many kerf bending patterns, we used straight cuts (slit kerf) for designing
our lamps prototypes. Lattice hinge is the most common and the most reliable. We
have done our tests for ruled surfaces manufacturing using this pattern because we
have distributed it on 3D surface using the straight generatrices.
The radius of the bend depends on the length of the cuts, the distance between
them and the thickness of the material (Porterfield 2014). We have done many proto-
types tests using 2 mm an 3 mm plywood and to design the final pieces which forms
usable prototypes. We are studying how to optimized the design process using math-
ematical modeling that let us to evaluate errors and reduce the cost of the project.
When you cut a 2D surface using a set of lattice cuts, we improve its ability to
warp: allowing expansion and compression perpendicular to the line of the cuts.
These abilities (expansion and compression) allow the distortion takes place. The
connections became spring because the two ends are able to move relative to each
other, for this reason the connected system might work as a hinge system and then
it is called “kerf bending lattice hinges”. The spring connection operates in torsion
only, then it is relatively straight forward to calculate the maximum torsional stress
of that region. This torsional behavior depends on material, thickness and kerf size.
If the stress in each spring connection is kept below to the torsional yield stress of
the material, then the material will always be operating elastically. This means that
the deformation will not be permanent; so once any force is removed, the hinge will
return to its original, flat shape. This also gives much better fatigue properties than
if the yield stress is exceeded, meaning that the joint could be bent and unbent many
times without risk of material fracture. It is possible calculate maximum torsional
stress using a formulas or using tests prototypes (Fenner 2011).
4.3 Relation Between Pattern and Curvature: Tests

Prototypes
Using computational design we parametrized some of the analyzed patterns to study

the relationship between pattern size and curvature, especially for double curvature.
We have done samples using 3 mm mdf and 6 mm plywood (Fig. 11).
Our tests show that we have to look at issues relating to manufacturing. As we can
see, the mechanical stresses of milling machines can cause breakage and it is very
important to evaluate the pattern size in relation to thickness of the panel. We also
experimented with laser cutting as it turned out to be more suitable for our pattern
tests (Fig. 12). We fabricated a sample, using laser cutter, with the same pattern that
we have manufactured with CNC milling machine (Fig. 13).
Fig. 11 Uniform pattern using parametric tools

Fig. 12 Samples for double curvature. Fir plywood 6 mm_cnc milling machine
Fig. 13 Samples for double curvature. MDF 3 mm_laser cutter
5 Design Results
Lamp prototype 1—cone

To analyze the different manufacturing ways of a conical surface we have studied the
prototype of a lamp composed by pieces of cone delimited by a continuous border
obtained as intersection between a circular cone and an ellipsoid. The distribution
of the pattern on the unrolled surface has presented some problems related to the
presence of a single edge.
To evaluate the different possibilities of pattern distribution on the project surface
we have parametrized three matrices: pin, slit and Zig Zag.
We have parametrized the matrices so as to modify both the geometry of the cut
and the overlapping of the cuts, on which the flexibility of the panel depends.
To distribute the pattern on the surface and on its development, as previously illus-
trated, we used the “morph 2D” component present in “paneling tools” for Grasshop-
per (Fig. 14).
Obtaining the result that we want, it was necessary to construct a grid using the
straight generatrices according to the following procedure:
– extract the isocurves from the surface part;
– consider only the straight generatrices;
Fig. 14 Samples for developable surfaces. Fir plywood 6 mm_cnc milling machine
– construct a grid dividing the straight generatrices into n parts;

– distribute the matrix on the grid using “morph 2d”.
We have obtained the pattern distributed on the unrolled surface linking it to 3D
surface (Fig. 15).
Lamp prototype 2—developable helicoid
To distribute the cutting pattern on the 3D surface, we have created a grid using
the straight generatrices which we have divided into m parts (variable) and allow us
to modify the grid.
We used Paneling Tools for Grasshopper to generate patterns on curved surfaces.
The Paneling Tools “morph 2D” tool allows you to morph a 2D curve onto a 3D
surface. It does this by dividing a surface with a grid. Then it stretches a pattern into
the frames created by the grid.
Using algorithms we have distributed the pattern on the surface and on the unrolled
surface. In this way we have manufactured the pieces (Fig. 16).
We are working how to turn our tests into an usable prototype and then into a design
product. We are designing how to assemble a complex 3D object from 2D pieces, in
a simple way. First of all, we are studying the relationship between geometry, kerfs
and shadows. We know that the developable helicoid sections by planes orthogonal
to cylindrical helix axis are the involutes of the cylinder that contains the helix,
therefore we can obtain these shadows up and down of the lamp, on the roof and
on the floor (Fig. 16). We are addressing some problems linked to manufacturing
technologies.
Fig. 15 Lamp prototype 1—cone
Fig. 16 Lamp prototype 2—developable surface composed by parts of developable helicoid

Fig. 17 Lamp prototype 3—non-developable ruled surface composed by parts of hyperbolic

paraboloid
Lamp prototype 3—hyperbolic paraboloid

A hyperbolic paraboloid is a doubly ruled surface generated by motion of a straight
generatrix on two straight directrices.
The directrices are parallel to the same plane, the first plan-director, and also the
generatrices are parallel to the same plan, the second plan-director. The main feature
of hyperbolic paraboloid is to have two groups of straight generatrices. It is negative
double curved surface and it is very complicated to perform. We have to need to
increase its ability to bent in two directions, thus we are studying how cut operations
can augment bending performance of the planar surfaces.
We have two problems to solve: first, we have to find the 2D shape to cut and then
we have to decide the kerf bending pattern to use and the best distribution in relation
to curvature.
The hyperbolic paraboloid it isn’t developable but, using the “smash” command in
Rhino, we can automatically generate two approximate unrolled shapes for a double
curved surface. We have used one of these to manufacture our lamp prototype.
We have created a grid using one of generatrices groups and we have distributed
slit kerf pattern using Paneling Tools (Figs. 17 and 18).
Our lamp prototype is composed by four pieces of hyperbolic paraboloid and two
elements to connect them.
It is very important designing how to connect each peaces and how to shape them.
It is a very complex work transform the 2D approximate develop into 3D design
shape in the case of doubly curved surface.
If you have to bend only in one direction you have only to make flexible the panel.
In fact, when you shape a single curvature object you have to force the piece only
Fig. 18 How to find approximate develop for hyperbolic paraboloid
in one direction, instead, if you have to bend a negative double curvature object you
have to bend in two opposite direction, thus there are stresses that tend to return 3D
shape into 2D shape.
Therefore, designing how to connect each piece it is a very hard work.
5.1 An Hyperbolic Paraboloid to Design a Stool
Kerfing technique is usually used to manufacture objects that are not needed a good
structural performance. An early example for structural application comes from
Gregg Fleishman’s furniture. Fleishman use the zig zag cut and he distributes it
following the principal direction of curvature to manufacture chairs obtained by a
single panel of wood. This is example of optimization of the final shape. As you can
see, there is a strong relationship between the thickness of the material, the shape
of the cut and the final shape that you want to obtain. All Fleishman’s furnitures are
developable surfaces.
Starting from this study case we are addressing with manufacturing of double
curvature structural object using kerfing technique: a stool made with hyperbolic
paraboloid pieces.
As we know, the first problem to solve is how to make planar a non developable sur-
face. In this case structural performance led the search for the approximate develop.
As we saw earlier for the fabrication of our lamp prototype, the best way to fabricate
the hyperbolic paraboloid piece, as we have designed it, is to use the approximate
develop obtained using isocurve u and to distribute pattern using a grid composed
on u generatrices. This solution can’t be used for our designed stool.
Fig. 19 Stool 1:2 prototypes
Fig. 20 Stool 1:1 prototypes. Stress problems
We have fabricate many prototypes 1:2 prototypes an two 1:1 protypes to find the
best approximate 2D shape and the best way to cut to fabricate our design 3D shape
(Figs. 19 and 20).
We have analyzed the main structural problems caused by the torsonial stress that
are caused by stress accumulation in one point.
We are working to find different solution to manufacture this doubly curvature
surface with structural behavior.
6 Digital Fabrication: Problems and Solutions
Kerfing is the act of cutting a series of kerfs (cuts) in a piece of wood in close
proximity, so the wood can be curved. Kerf is defined as the width of material that
is removed by a cutting process. It was originally used to describe how much wood
was removed by a saw, because the teeth on a saw are bent to the side, so that they
remove more material than the width of the saw blade itself, preventing the blade
from getting stuck in the wood. When talking about CNC shape cutting with typical
cutting processes, kerf is the width of material that the process removes as it cuts
through the plate. Over the years some people use the word “kerf” generally referring
to cut for bending but now it should only be used when talking about the actual cut
width. Each cutting process removes a different amount of material, or kerf. Water
jet and laser, remove a smaller amount of kerf, which is one of the reasons they can
be more precise.
In our research we have address the relationship between “kerf” of the laser cutter
and CNC milling machines and the design of our objects.
The technologies involved in our tests are laser cutter and CNC milling machines.
One of the main research aims is to evaluate which problems could be into the
production process and to compare results.
We have tested CNC milling machine to make our stool prototype and we have
used laser cut to produce lamps prototype. The choice usually depends on the thick-
ness of the panel and on the object size.
Our tests show that the main problems linked to CNC milling machine are the
drill bit size that we have to take into account from start of design process and work
piece fastening to prevent breakage during milling operations.
When we use laser cutting technology we have to evaluate the right distance
between cuts to prevent burns.
As a matter of fact that any areas in our design where cut lines come closer than
0.5 mm together could burn away entirely (Fig. 21). Any details narrower than 1 mm
are likely to be very fragile and in some cases can cause the material to warp whilst
cutting. As a benchmark, we think that minimum cut widths be no smaller than the
corresponding thickness of the material. We can go smaller but this can make our
pieces very fragile. We have corrected our first prototypes of developable helicoid
using pattern distribution on our grid in order to respect this rule (Fig. 21).
Fig. 21 Laser cutter: problems and solutions
7 Conclusions and Further Research
We have presented methods for manufacture design objects using kerfing technique.
Our approach is based on geometric knowledge of ruled surface and pattern distribu-
tion using a grid on straight generatrices. As we have seen, in the cases of developable
surface that we have manufactured, developable helicoid, pattern density automati-
cally achieve the surface curvature.
It isn’t true for doubly ruled surfaces. We are studying a technique for applying the
pattern based on surface curvature, measuring principle curvature in each direction,
and then morph a grid along two axis rather than just one. Our method is illustrated by
means of different examples, some of which correspond to existing work in kerfing.
We have shown how to morph a grid along one axis and how to apply the pattern
density to achieve the surface curvature.
According to advances in kerf bending techniques our aim is to tests geometric
pattern generation process for specific needs (pattern for needs\customized) in field
of wooden furniture design.
We defined the surface geometry based on 2D array organizations and grid layout
for pattern design. This grid layout itself became a limitation along the experiments.
In the further studies the manipulations of the grid organization might enrich the
geometric pattern generation process, thus the bending behavior (Güzelci 2016).
Another issue concerns the kind of design object we want to build. We are not
interesting in objects that can be bent in place permanently but in objects that can
back flat: flexible objects, complex forms, eco-packaging and easy transportation.
Least but not least we are addressing the issue of structural strength in wooden
furniture design, in our future researches we are going to analyze the patterns in rela-
tion to structural problems. The aim is search of geometric manufacturing solutions
for complex shapes using a parametric model not only as a representational tool, but
as a tool for real-time shape exploration.
We think that representation and applied geometry play a key role in this design
process: advances in computational design, simulation tools and digital fabrication
techniques allow us to experiment with analogue and digital tools, informing the
creative process by establishing feedback between digital media and physical proto-
types.
References
Capone M (2012) Geometria per l’architettura. Edizioni Giannini, Napoli

Fenner P (2011) https://www.defproc.co.uk/blog/2011/laser-cut-lattice-living-hinges/
Gauss KF (1902) General investigations of curved surfaces of 1827 and 1825. The Princeton Uni-
versity Library
Güzelci O et al (2016) Enhancing flexibility of 2D planar materials by applying cut patterns for
hands on study models. In: SIGraDi, XX Congreso de la Sociedad Iberoamericana de Gráfica
Digital, Argentina, 9–11 November 2016
Lanzara E (2015) Paneling complex surfaces. Razionalizzazione di superfici complesse per
l’industrializzazione. Phd thesis. fedoa unina
LMN Architecs. https://lmnarchitects.com/tech-studio/fabrication/kerf-bending/
Menges A (2011) Integrative Design Computation integrating material behavior and robotic manu-
facturing processes in computational design for performative wood constructions. In: Proceedings
of the 31st annual conference of the association for computer aided design in architecture (ACA-
DIA)
Menges A et al (2016) Wood architecture: a computational approach. Taylor & Francis Ltd., Oxford
Migliari R (2009) Geometria descrittiva. CittàStudi, Roma
Muñoz P et al (2011) Fabricación digital y morfología: la flexibilidad en la generacion de la forma.
Ediciones de la forma, Argentina
Mamou-Mani Architects. https://mamou-mani.com/project/wooden-waves/
Porterfield A (2014) Curved laser bent wood. http://www.instructables.com/id/Curved-laser-bent-
wood/
Pottmann H et al (2015) Architectural geometry, Kindle edn
Digital Joinery for Hybrid Carpentry
Shiran Magrisso and Amit Zoran
Abstract The craft of carpentry relies on joinery: connections between pieces of

wood to create multipart structures. In traditional woodworking, joints are limited to
the manual chisel skills of the craftsperson, or to capabilities of the machines, which
favourite 90° or 180° angle joints with no more than two elements. We contribute
an interactive design process in which joints are generated digitally to allow for
unrestricted beam connectors, then produced from Nylon-12 using selective laser
sintering (SLS) 3D printing. We present our Generative Joinery Design Tool and
demonstrate our system on a selection of stools. This chapter exemplifies the potential
of Digital Joinery to enhance carpentry by incorporating a hybrid and interactive level
of design sophistication and affordances that are very hard to achieve with traditional
skills and tools.
Keywords Joinery · 3D printing · Hybrid · Digital fabrication

Computer-aided design
1 Introduction
Good joinery… is difficult to design and even more difficult to execute. It should be thought
of as an investment, an unseen morality.
George Nakashima (1981)
The craft of designing and implementing wood joints has a long-standing and
important role in all traditional woodworking practices (Nakashima 1981; Selke
1977). Joints are the elements that transform lumber into a practical artifact. They are
S. Magrisso · A. Zoran (B)

Bezalel Academy of Arts and Design, Jerusalem, Israel
e-mail: zoran@cs.huji.ac.il
S. Magrisso
e-mail: shirandm@gmail.com
A. Zoran
The Hebrew University of Jerusalem, Jerusalem, Israel

442 S. Magrisso and A. Zoran
fundamental to wooden artifacts and are probably the most technologically advanced
elements in woodworking.
The complexity of the joinery craft applies severe constraints on carpentry, lim-
iting the design possibilities of wooden artifacts. It is difficult to master high-end
joinery craft, as it requires free-hand chisel techniques and the design of non-trivial
joints. Thus, common joinery restricts design possibilities to flat or right angles
between pieces of wood, rarely connecting more than two pieces in the same joint.
In our work, we are motivated to liberate contemporary woodworkers from this
limitation and enable new design affordances using generative design and additive
manufacturing (AM) of plastic joints.
Digital Joinery contributes a new type of design freedom and construction affor-
dances to furniture making.1 For instance, a maker who wishes to design a complex
joint posits the 3D lumber plan virtually and selects all of the surfaces that need to
be connected. A parametric design procedure helps in generating a Voronoi diagram
skeleton, which acts as a bridge between the wooden surfaces.2 Our tool allows for
parametric control over the characteristics of the joint, such as density, thickness, and
style. After structurally evaluating the design, the user 3D prints the joints in Nylon-
12 through selective laser sintering (SLS). In addition, we contribute new designs for
3D-printed anchors, which are virtually added to the generated joint before printing.
These anchors, which can be assembled without special woodworking or glue, are
the locking mechanism between the joint and the wood.
In the following section, we review related work, then discuss wood joinery. Then
we turn to detailing the technical aspects of Digital Joinery, including the anchors,
our Generative Joinery Design Tool (GJDT), and a design workflow. We present
a design case featuring three stools. Before concluding, we discuss feedback we
received from makers regarding the future potential of Digital Joinery.
2 Related Work
Digital Joinery brings new design affordances to carpentry, like other hybrid design
projects that mix traditional craft with digital practice (Cheatle and Jackson 2015).
Hybrid design is a booming research territory in human-computer interaction (HCI),
featuring new hardware (Zoran et al. 2014) and software tools (Efrat et al. 2016),
along with aesthetics and creative procedures (Devendorf and Roykai 2015; Jacobs
and Buechley 2013; Jacobs and Zoran 2015; Tsaknaki et al. 2017; Zoran 2015). In
general, the hybrid practice aims at extending the creative spectrum of makers, going
beyond the secure boundaries of autonomous fabrication (Zoran 2016).
1 This chapter is a modified version of a research paper under the same title, published in ACM CHI
2018, Montreal, Canada.

2 Voronoi diagrams partition the 3D space into regions and are common in natural structures as well
as digital design.
Digital Joinery for Hybrid Carpentry 443
Today, designers rely on hybrid practice in diverse ways (Efrat et al. 2016; Jones
et al. 2016; Laarman n.d.; Mizrahi et al. 2016; Rosner and Ryokai 2010). For exam-
ple, Zoran (2013) demonstrates a process where plastic structures are designed dig-
itally and 3D printed, then manually reinforced with organic woven fibers. Hybrid
reAssemblage features an alternative process, where broken ceramic elements are
restored by 3D scanning and printing to create a new type of aesthetic (Zoran and
Buechley 2012).
With respect to craft, digital design can produce complex patterns that are hard to
make manually (Reas and McWilliams 2010). These patterns, such as the Voronoi
diagram that we use in our work, can resemble organic structures (Haeckel) and are
common in many computer graphics (CG) projects. For example, Lu et al. introduce
a tool that utilizes the Voronoi diagram to compute irregular volume tessellations
(2014). Martínez et al. enable 3D printing of flexible structure by Voronoi foams
(2016). A newer work discusses a foam-like printed material that adapts to uneven
load scenarios (Martínez et al. 2017). In addition, the idea of modifying 3D forms
to achieve a new aesthetic is also common in CG, where 3D computer-aided design
(CAD) objects are virtually modified to allow for surface decoration (Chen et al.
2016; Zehnder et al. 2016).
Recently, several projects have explored the application of parametric and gener-
ative design to both joinery design and furniture making. Zheng et al. have developed
a parametric joint-generation tool for 2D laser-cut assemblies (Zheng et al. 2017).
TrussFab is an integrated end-to-end system that allows users to fabricate large-scale
structures (Kovacs et al. 2017). Yao et al. present a tool for designing furniture joints
with user-controlled graphics (2017). SketchChair is an application that allows novice
users to control the entire process of making their own chairs (Saul et al. 2011). Fu
et al. present a computational solution to support the design of a network of joints
that form a globally-interlocking furniture assembly (2015). Meanwhile, Garg et al.
present software to support the interactive design of reconfigurable furniture, featur-
ing tools that resolve infeasible configurations (2016). Yet these projects have grown
out of an academic landscape, while many joinery developments are happening in
design studios and professional woodworking workshops.
3 Wood Joinery: From Traditional to Digital
In The Joint Book, Terrie Noll reviews joint techniques and orientations (2009).
Achieving a joint between pieces on two different planes is very difficult and therefore
rare. Several beam and board orientations are common: parallel and I orientation
(boards joined by their edge or end-to-end to increase width or length); L and T
orientation (90° connectors); and crossed or angle orientation (modification of one
of the orientations to change the joining angle in one axis to anything other than 90°
or 180°).
There are numerous techniques to implement the joint itself. One can use glue,
nails, screws, or other fasteners to attach one piece of lumber to another. Dovetail
joints are interlocking joints with great mechanical strength. They are constructed
with an angled male part shaped like a dovetail that fits into a similarly shaped female
socket. Mortise-and-tenon joints are also very common, having a tongue (tenon) that
fits into a hole (mortise). In many traditional practices, such as Japanese carpentry
(Seike 1977), skilled craftpersons master dry wood (no glue) joint techniques that are
reversible and feature several complex wooden parts, enabling sophisticated locking
mechanisms.
Recently, as digital fabrication devices have became more common, a growing
number of projects have featured CAD joints. Jochen Gros has presented 50 digital
wood joints, designed using CAD and fabricated by a computer-numeric control
(CNC) milling machine (n.d.). In addition, several projects explore the use of 3D
printing in fabricating plastic joints, mostly to allow for modularity in their con-
struction (Bernier 2013; Brink 2015; Gellért 2015; Oh 2015; Studio Minale-Maeda
2014).
4 Digital Joinery for Hybrid Carpentry
Working in the same realm, and implementing 3D-printed plastic joints for wood-
working, we aim at adding an interactive stage to the joinery design process. Our
work allows users to create unique and specific solutions to complex conditions. Our
digital joints comprise two parts: the joint body and anchors, both printed in Nylon-
12 [(CH2 )11 C(O)NH]n (PA12), a durable, low-cost material with good mechanical
properties, which enables production of hollow and latticed structures.
The joint body bridges all of the wood surfaces that need to be connected, and
is generated with our software. In order to save material costs and weight, a sparse
Voronoi skeleton makes up the body, which also contributes a distinguished style.
The anchors are set manually, as the designer chooses the type of anchors she or he
would like to use and adds them virtually to the joint body prior to printing.
Our Generative Joinery Design Tool allows makers to specify unique joints for
custom design furniture. Using our portfolio of anchors, the 3D-printed joints can be
assembled easily with no need to master manual chisel work or have special CNC
machines. We now review our custom anchors prior to discussing GJDT (All of these
files can be downloaded from the project website3 ).
5 Digital Joinery Anchors
Anchors enable easy assembly of strong and stable furniture. The anchor is an element
of mechanical dry connection (without adhesive). Only manual tools are necessary
to prepare the wooden parts to receive the anchor, as opposed to only using CNC.
3 http://www.amitz.co/DigitalJoints.html.
Fig. 1 Digital Joinery and five development stages of a hybrid stool; a an early sketch of a stool
design relying on digital joinery; b using the generative joinery design tool, the software generates
digital joints; c a rendering of a complete stool with digital joints; d a 3D-printed (by selective laser
sintering) Nylon-12 joint with its anchors; e a closeup photograph of a dyed and assembled Voronoi
diagram skeleton joint, connecting four wooden beams; f a photograph of the finished stool (by
Daniel Shechter)
We conceptualize different methods to implement the anchors, and present below

the final four anchor designs we use in our work. CAD models of these anchors are
available from the project website; Fig. 1 illustrates the design process.
Anchor 1 (Native) relies on friction in a 6 mm-diameter cylindrical structure 16 mm
in length with one-directional 1 mm serration (see Fig. 2c). To ensure anchor effi-
ciency, seven different models with varying angles and quantities of teeth were eval-
uated. All of the models underwent a pulling test to see if they would move out of
place under stress (see Fig. 2d). The tests showed that the quantity of teeth is less
significant than tooth angle. We observed that a decisive 25° angle provides maxi-
mum durability when an anchor is pulled. Optimal results were received from model
20/25, which could withstand a maximum pulling force of 29.25 kg.
Anchor 2 (Wedge) relies on Anchor 1 (serrated structure), with a wedge that locks
the anchor when the joint reaches its final location (see Fig. 2d, e). Unlike Anchor 1,
which resists insertion (as friction makes it extremely difficult to achieve an optimal
tight connection), Anchor B can be inserted smoothly and easily. In our proposed
structure, Anchor 2 can withstand a maximum pulling force of 65.62 kg, double the
amount of Anchor 1. One limitation of this anchor is that because it requires access
to the joint from both sides of the wood, it is not suitable for every situation.
Anchor 3 (Nails) relies on Anchor 2, but instead of using a 3D-printed wedge, we
used steel nails inside the anchor (see Fig. 2d, e). When the anchor is pushed into
the wood, the nails remain inside the anchor’s body. The nails reinforce the plastic
anchor with a metal structure, allowing it to withstand a maximum pulling force of
101.83 kg.
Anchor 4 (Jumbo) is a jumbo anchor (see Fig. 2d, e). Unlike the previous models,
this anchor uses the latticed body of the joint to set in a 2.5 mm screw that holds
a tenon. As the screw is tightened, the tenon retracts the anchor, which is pushed
forcefully against the wood. This joint reinforces the anchor with a metal structure
and can withstand a maximum pulling force of 35.98 kg. It should be noted that this
Fig. 2 Digital Joinery

anchors; a initial sketches for
anchor concepts, and (a1) the
selected solution; b physical
stress-strain evaluation and c
the results as a function of
teeth angle/quantity of a
cylindrical anchor; d
improved designs with
anchors utilizing different
locking mechanisms (nail,
wedge, jumbo and flat
20–25); e the collection of
different anchors we used
(Photograph by Daniel
Shechter)
anchor never failed a test, as the anchor never collapsed, although the nylon body
began to stretch.
In the Design Case: Stool Collection section we present a set of stools that demon-
strate the use of all the anchors described here. In real-world applications, we extend
the length of the anchors up to 25 mm if the thickness of the wood allows. Before
examining this work, we outline the main interactive and technical principles of our
GJDT tool to design joints, to help others re-implement or incorporate our work in
new developments.
6 Generative Joinery Design Tool (GJDT)
Using Grasshopper, a parametric design plugin for Rhino, we developed GJDT to

allow users to customize joints. Where the design consists of several joints, the
designer will need to process each of them separately. The generative design proce-
dure is as follows (see Figs. 3 and 4):
1. In Rhino, the user specifies the lumber position in 3D space, or imports a design
made elsewhere. The user can use as many separate wooden parts as he or she
wishes, as long as the surfaces that need to be connected are flat and rectangular
(see Fig. 3a).
Fig. 3 The interactive procedure and workflow of our Generative Joinery Design Tool (GJDT),
a specifying lumber position in Rhino; b selecting surface to be joined in GJDT; c defining the
position and orientation of the supporting contour curves, (c1) and controlling their fillet angles; d
applying a loft function between the surfaces to create a watertight solid object; e populating this
object with random points, or (e1) using attractor points; f generating a 3D Voronoi diagram with
or without isocurves; g creating a solid 3D Voronoi skeleton in the arm thickness defined by the
user; and h outputting the design for external structural validation tests
Fig. 4 The Grasshopper interface of GJDT. GJDT’s main components are: a 3D object import;
b surface selection; c generating supporting curves; d lofting between surfaces to create a solid
watertight joint body; e population of the object with random points; f generating a 3D Voronoi
diagram; g creating a smoothed, solid skeleton body. Initial settings are in the center of the sliders
range for ease of use
2. In GJDT, the user selects the wooden elements and their surfaces S i (where i is
the surface index) that need to be joined (see Figs. 3b and 4a, b).
3. GJDT generates contour curves C i for S, and a set of curves C *i that offset C *i by
distance d (the user’s input parameter) in the corresponding normal direction
ni , such as that C *i d × ni + C i (see Fig. 3c). The user has parametric control
over C Tota
i
l
= {C i , C *i } corners, filleting, and the angles of C *i (see Fig. 3c1 ).
LoftSet
4. GJDT lofts each set of C l {C Total
i , C Total
j } to create a solid object
LoftSet
(0 ≤ j i < number of surfaces), such that SOl Loft(C l , control_param).
Lofting can be challenging, and sometimes Rhino’s manual loft function oper-
ates better than the function in Grasshopper, allowing more control options.
Therefore, if the lofts do not satisfy the user, they can be redone manually (see
Fig. 3d). In both cases. the user can choose the type of loft function to use
(Loose; Normal; Straight sections; Tight; Uniform).
5. GJDT applies boolean union U on all SOl . Sometime boolean operations in
Rhino and Grasshopper fail due to numeric conflicts; this is solved by re-scaling
each SOl by different frictional factor.
6. GJDT populates U with points inside and on the boundary surfaces. The user
controls the number of points to populate. The surface bases are populated with
a dense grid of points to stabilize the connection to the wood (see Figs. 3e and
4e). The user can implement an attractor to create an uneven distribution of the
points, in order to change the local resolution of the joint skeleton (see Fig. 3e1 ).
Fig. 5 Five different photos of printed joints with varying finishing colors and joints’ body visual
style. These joints demonstrate a small part of the stylistic variety potential of GJDT
Fig. 6 a A CAD model of a stool, and different Digital Joinery alternatives for joints (b–d). The
user can consider structural analysis evaluation (here done in the Scan&Solve plugin for Rhino),
cost, weight, and aesthetics: b a three-surface joint and its danger level simulation; c a three-surface
joint and a displacement simulation; d a two-surface joint and a displacement simulation. Cost
calculated from www.shapeways.com website in May 2017
7. GJDT generates 3D Voronoi curve skeletons based on the population points.

GJDT creates Voronoi 3D solid objects, produces curves based on the corners
of these objects, and deletes duplicate curves. The user can decide whether to
add the original isocurves of U to the skeleton (see Figs. 3f and 4f).
8. GJDT generates a volumetric skeleton based on U curve structure. The skeleton
SK bones radius is set by the user (see Figs. 3g and 4g).
9. GJDT applies Laplacian smoothing on SK s LaplacSmooth (SK,iterations),
when iterations is set by the user (see Figs. 3h and 4g).
10. The joint SK s is finished. The user can evaluate it aesthetically and structurally
using an external finite element method (FEM) tool (we used the Scan&Solve
Rhino plugin), and review alternatives (see below and Fig. 6). A future GJDT
version may include internal (built-in) topological optimization abilities.
11. In this stage the joint body is ready. The user now adds the anchors she or he
wishes to use to the design and generates a STL mesh file for print.
7 The Hybrid Design Process and Workflow
The hybrid design workflow is subject to personal design preference and habits. We
can only suggest a recommended workflow we used and executed, as demonstrated
in the following section and later on three stools.
As in many design processes, the maker starts by sketching concepts and exploring
alternative solutions to the project requirements (see Fig. 7). The maker is free from
traditional carpentry restrictions and can envision designs going beyond traditional
shapes. For example, multiple-beam joints and unusual angles are now possible, and
new types of organic aesthetics (as influenced by the Voronoi diagram style) may
suggest decorations and new patterns influenced by the 3D-printed work. The maker
Fig. 7 Concept designs, sketches, and investigations. a Initial sketches of concept designs for hybrid
stools that require Digital Joinery for construction, and b several 3D-printed miniature stools to
evaluate design ideas
can select the preferred anchor for the specific application. A color treatment and
material bridge may be needed to aesthetically link the printed plastic with wood.
See Fig. 5 for varying types of printed and assembled joints, and Fig. 8 for examples
of the impact of the joint on the overall stool design.
When the maker is satisfied with the sketch, she or he models it in Rhino, largely
by positioning wooden beams and boards in the 3D CAD space. Then, joint-by-
joint, the designer creates digital joints using GJDT. For each solution, the designer
can consider alternatives, taking into account varying parameters such as structural
analysis results (FEM is not yet an integrated part of GJDT), printed price and weight,
and the aesthetic style of the joint. Anchors are then added to the design, and the
work is sent to print.
Meanwhile, the maker manually produces the wooden parts. Upon receiving the
printed joints, she or he assembles the work and finishes it using any accessible,
useful technique and process. One of the main advantages of Digital Joinery is that
the construction can be dry (no glue) and does not require any special CNC work
and/or chisel skills. This makes the furniture assembly reversible and easy for non-
professionals to accomplish.
8 Design Case: Stool Collection
To personally evaluate Digital Joinery and demonstrate its capabilities, we designed

and built a collection of hybrid stools. The first model constitutes our initial attempt
to fully complete the design and construction of an object that merges woodcraft
with digital practice
T1 model: The first stool is a 120° radial symmetry stool (see Fig. 8a, b). It consists
of a central joint joining three sections of the top together, and three sets of two joints
connecting the three legs to the top. During the production of the joints, we learned
that connections made at multi-faceted junctions have multiple advantages over those
made at connection points where one face is joined to another.
T1 is based entirely on Anchor 1. In the process of creating this stool, we discov-
ered many design principles that helped in the following designs. The joints were
fastened together with friction, thus the degree of precision needed between the holes
in the wood and the printed anchor parts was extremely high. Precision to the tenth
millimeter is required to enable the anchor to enter without damaging the 3D-printed
plastic and to lock the part into the wood. The printed parts suffer from imprecision
due to production accuracy and the placement of each anchor in a different vector on
the printing tray. Additionally, as the printed material absorbs moisture after printing,
its dimensions change. These factors all make the connection points slightly loose,
which causes a sense of instability when a user sits on the stool.
Visually, the meeting point between the wood and the printed Nylon-12 creates an
aesthetic unease. The visual challenge raises several design-related concerns about
how technologies are sacrificed for each other; hence, this was a main consideration
for the subsequent developments.
Fig. 8 The two first hybrid

stools in our collection. a T1
model with no load and b
with load, and c T1 model
with off-white dyed anchors.
Photographs (a and c) by
Daniel Shechter
T2 Model: In this second model, different types of joints were generated (see
Fig. 8c). Here our design purpose was two-fold. First, we aim at exploring varied
aesthetic possibilities received from GJDT. In addition, this stool allows for exam-
ination of the joints’ maximum mechanical ability to bear a load, compared with
FEMs (see Fig. 6).
The range of styles resulting from GJDT enabled a wide variety of possibilities,
as the user-controlled parameters in GJDT result in joints with rich aesthetic charac-
teristics (see also Fig. 5). In addition to stool T1, here we considered for the first time
how manual woodwork could welcome the plastic joints with better visual integra-
tion. We colored the joints using a special paint solution, and explored how manual
carving can reflect some of the joint patterns in the wooden elements (in Fig. 5d, for
example, note the decorative pattern in the wooden beam below the joint).
The use of Anchors 2, 3 and 4 allowed for easy and quick construction. The
required precision of the joint placement was made possible with zero tolerance in
the connection points. Given that the form chosen for the joints bearing the seat
was one that has lattice strips with a thickness of 0.8 mm (see Fig. 6d3 ), which in
the preliminary analyses revealed extensive areas of failure, the stool top becomes
springy and soft. This contributes an additional quality in our Digital Joinery: the
ability to use FEM to predict and design the dynamic behavior of the joints.
T3 model: The last model was inspired by the aesthetics of the Voronoi diagram
implementing the joints. This model demonstrates a stool design with a structural
style unique to the technology. It reveals the design potential that can be achieved
using our new process, which allowed the creation of a complex geometric structure
that cannot be achieved using traditional technology. Digital Joinery allowed relative
ease in the design, production, and assembly of the resulting structure, which is both
stable and unique.
In this model, most of the joints were locked using Anchor 4. A latticed structure
resulting from Voronoi diagrams (see Fig. 9) enables joint precision using screws
that are almost invisible, yet provides an excellent degree of structural strength (see
Fig. 10). We used CAD work to create stylish 3D printed parts in the backrest and
the center of the seat (such as in Fig. 10b), colored to compliment the aesthetic of
the wood. This creates an additional visual link between the wooden parts and the
3D-printed parts. A thin white ring was painted on the wooden beams near the edge
that meets the joints to stylize these contact points.
9 Evaluation
One can easily see how our work evolved from the first stool to the last, and how
we refined our hybrid craft skills while we worked on the three stools. The design
evolution we experienced hints at the full aesthetic potential of hybrid furniture with
Digital Joinery. Only in the last design did we really start to free ourselves from
traditional carpentry constraints, limitations, and habits, and to fully explore the
Fig. 9 Eight sketches of 3D Voronoi-diagram inspired stools, an early stage in designing stool T3
design freedom we now have at hand. Thus, we see the stools collections as an initial
step towards exploring hybrid carpentry in depth.
To gain some perspective on how professional designers and carpenters would
value Digital Joinery, we engaged in meetings with four carpenters and three design-
ers. However, our study participants did not have a great deal of time to invest. The
design and fabrication of each of our stools took two to three weeks from sketch to
completed furniture; this is not a process we can compete in short evaluation meet-
ings. Moreover, not all of the participants are skilled in CAD. Hence, we used these
evaluation meetings to learn more about design potential rather than GJDT usability,
focusing on concept design rather than production of real furniture.
We had a single meeting of approximately one hour with each of the participants,
in their studios or workshops. In each meeting, we presented our work, the capabilities
of Digital Joinery and our portfolio of stools. We asked the carpenters and designers
to share any feedback they had. Furthermore, we asked them to draw some ideas for
future objects they would design using this technology.
Overall, we received positive reactions. None of the professionals we met with
failed to understand or appreciate the contribution of Digital Joinery. Yet, as we were
aware that these responses might be biased or incomplete, we put more emphasis
on the drawings and visions participants expressed about how they might implement
these new design and construction capabilities in their own work. Figure 11 shows a
collection of drawings we made based on the rough sketches we collected from the
designers and carpenters who examined our work.
Roughly speaking, it seemed that the carpenters, who in their daily practice largely
produce designs made by others (such as architects or designers), value Digital
Fig. 10 a The last (T3)

stool, b a decorative,
3D-printed part and (c–d)
digital joints. Photograph a
by Daniel Shechter
Fig. 11 Four drawings based on sketches made during the evaluation stage by external carpenters or
designers. Hybrid Joints are in yellow. a a coffee table uses Digital Joinery in a manner that resembles
our stools; b a toy tent allowing for many shapes enabled by Digital Joinery; c a traditional chair
design whose most demanding craft section is replaced with a 3D-printed joint; and d a table with
a central structural joint that becomes the most important design element of the work
Joinery as a tool for efficiency or for labor-saving. Figure 11c, for example, shows
how one carpenter envisions a printed joint replacing part of a chair he makes, thus
freeing him from the most demanding part of the job. This perspective was repeated
in the open discussions, as the carpenters we met value automation and computers
as labor-saving technologies, not necessarily as design or innovation tools.
On the other hand, the designers we met, who are used to computational design
in their daily practice and do not manually operate machines or master woodcraft,
envision new design affordances enabled by Digital Joinery (see Fig. 11a, b, d). As
this group of designers is familiar with 3D printing and CAD, they were excited to
consider new possibilities for design agency. For example, one designer imagined the
use of Digital Joinery in fabric toys, and another considered pieces of furniture with
a large, central joint that would serve as a structural base for all elements, holding
the furniture together.
Our overall impression was that while the carpenters we met with value the tra-
ditional craft, they were unable to demonstrate design creativity related to CAD
practice, as they are simply unfamiliar with CAD capabilities and design potential.
In other words, they were a bit conservative compared to the group of designers.
On the other hand, designers easily integrate Digital Joinery into their conceptual
work. Although our pool of participants was too small to draw further conclusions
about the division of skills and design intent, the concept of hybrid practice seems
easy for makers who are already familiar with computational design to accept.
Nevertheless, CAD designers are not necessarily part of the community of mak-
ers who work in woodshops. This means that our motivation to liberate traditional
carpentry from some of its technical limitations will not develop fully with CAD
practitioners. Thus, we argue that our innovation will be a perfect device for a third
type of makers, hybrid makers, who explore a wide spectrum of traditions and tech-
nologies, and use manual and physical craft together with digital fabrication and
CAD. This community of makers is still small, but increasing (see Zoran 2016 for a
wider discussion on hybrid craft).
10 Conclusions and Future Work
In this paper we presented Digital Joinery, a woodworking design paradigm aiming

to liberate carpentry from traditional construction limitations. We contribute a new
joinery design software tool that supports new types of connectors, to let multiple
pieces of lumber meet at unconventional angles, manifesting a new type of aesthetics
and structure for furniture. We demonstrated our tool in a collection of three stools,
each of which illustrates a different aspect of new joints for stool making.
Our work aims at reinforcing the bridge that has already started to pave the way
for makers, designers, and researchers to merge computational practices with craft
(Rosner and Ryokai 2010; Tsaknaki et al. 2017). In doing so, we continue the lines
of prior work (Zoran and Buechley 2012) aiming to keep computational design open
and allow for some sort of manual freedom to encourage makers to explore the full
potential of the hybrid medium.
In addition to the technical work presented here, we discussed our work with pro-
fessional carpenters and designers. The designers seemed a bit more open-minded
about the affordances of such technology, envisioning a new type of human-computer
interaction via the furniture design process. Yet, as our work aims to assist craftper-
sons as well as designers, we believe that hybrid design tools may be most fruitful in
the hands of hybrid makers. Otherwise, a broader introduction is needed to expose
traditional craftpersons to digital design and fabrication prior to training them to use
Digital Joinery.
Considering future work, digital joinery can be extended to serve as a connect-
ing agent between other ready-made artifacts (not just wood). Moreover, we would
like to extend the deployment of Digital Joinery to a diverse community of mak-
ers wishing for hybrid design prosperity. We envision future craftpersons equipped
with manual and digital machines, mastering handwork and computational work.
While this community of professional hybrid makers already exists, more tools and
technology for hybrid practice are required, and our work aims at this need.
Finally, and on a slightly different track, we would like to learn how Digital Joinery
could personalize mass furniture manufacturing. For instance, given a warehouse
with a huge amount of lumber in a finite set of dimensions and types, how can
consumers interact with parametric and generative tools to customize their furniture?
How can we optimize the use for this collection of raw material, yet maximize the
design freedom of the end users?
For example, let us assume a warehouse includes 4 cm × 4 cm wooden beams
in lengths of 80 cm and 50 cm. When a customer requests a chair that will support
a load of 100 kg, or a table in a specific dimension, the software will be able to
generate an optimal or semi-optimal (using a minimal amount of wood) solution
enabled by 3D-printed joints and an optimizing design algorithm. In addition, the
user can specify a finished style and other details.
This vision require extensive research on (1) finite-set catalog selection mecha-
nisms integrated into generative design software; and (2) extended generative design
procedures that optimize full furniture architecture for production, considering wood
costs, 3D printing costs, and the system’s abilities. This will be the topic of our future
research, and hopefully will interest other researchers in the HCI and CG communi-
ties.
Acknowledgements We would like to thank all of the people who supported and helped in finishing
this work, including the Hybrid Lab group at HUJI; the external advisors in Bezalel Academy of
Arts and Design, Jerusalem; Moran Mizrahi; Raanan Fattal; Chia. L. Evers; Tal Levi; Lior Uri; Oz
Banderman; Hai Phalach; Nimrod Shamia, Ronen Dor; and Alon Greber and others who helped by
allowing the use of their technical facilities. The work was partially supported by the Israeli Science
Foundation (ISF).
References
Bernier SN (2013) Draisienne: hack of IKEA’s Frosta stool. In: Coroflot. http://www.coroflot.com/
samuelbernier/Hack-of-IKEAs-Frosta-stool. Accessed 27 May 2018
Brink N (2015) Christian sjöström develops link modular furniture system. In: design-
boom|architecture and design magazine. https://www.designboom.com/design/link-furniture-
christian-sjostrom-03-24-2015/. Accessed 27 May 2018
Cheatle A, Jackson SJ (2015) Digital entanglements: craft, computation and collaboration in fine
art furniture production. In: Proceedings of the 18th ACM conference on computer supported
cooperative work and social computing. ACM, New York, pp 958–968
Chen W, Zhang X, Xin S et al (2016) Synthesis of filigrees for digital fabrication. ACM Trans
Graph 35(98):1–98. https://doi.org/10.1145/2897824.2925911
Devendorf L, Ryokai K (2015) Being the machine: reconfiguring agency and control in hybrid
fabrication. In: Proceedings of the 33rd annual ACM conference on human factors in computing
systems. ACM, New York, pp 2477–2486
Efrat TA, Mizrahi M, Zoran A (2016) The hybrid bricolage: bridging parametric design with craft
through algorithmic modularity. In: Proceedings of the 2016 CHI conference on human factors
in computing systems. ACM, New York, pp 5984–5995
Fu C-W, Song P, Yan X et al (2015) Computational interlocking furniture assembly. ACM Trans
Graph 34:1–91. https://doi.org/10.1145/2766892
Garg A, Jacobson A, Grinspun E (2016) Computational design of reconfigurables. ACM Trans
Graph 35:1–90. https://doi.org/10.1145/2897824.2925900
Gellért O (2015) Print to build, 3d printed joint collection. In: Behance. https://www.behance.net/
gallery/27812109/Print-To-Build-3D-printed-joint-collection. Accessed 27 May 2018
Gros J 50 Digital Wood Joints. In: Winterdienst. http://winterdienst.info/50-digital-wood-joints-by-
jochen-gros/. Accessed 27 May 2018
Haeckel E (1998) Art forms in nature: the prints of Ernst Haeckel. Prestel
Jacobs J, Buechley L (2013) Codeable objects: computational design and digital fabrication for
novice programmers. In: Proceedings of the SIGCHI conference on human factors in computing
Jacobs J, Zoran A (2015) Hybrid practice in the Kalahari: design collaboration through digital tools
and hunter-gatherer craft. In: Proceedings of the 33rd annual ACM conference on human factors
in computing systems. ACM, New York, pp 619–628
Jones MD, Seppi K, Olsen DR (2016) What you sculpt is what you get: modeling physical interactive
devices with clay and 3D printed widgets. In: Proceedings of the 2016 CHI conference on human
factors in computing systems. ACM, New York, pp 876–886
Kovacs R, Seufert A, Wall L et al (2017) TrussFab: fabricating sturdy large-scale structures on
desktop 3D printers. In: Proceedings of the 2017 CHI conference on human factors in computing
Laarman J (2018) Joris Laarman Lab. In: Joris Laarman Lab. http://www.jorislaarman.com/.
Accessed 27 May 2018
Lu L, Sharf A, Zhao H et al (2014) Build-to-last: strength to weight 3D printed objects. ACM Trans
Graph 33:1–97. https://doi.org/10.1145/2601097.2601168
Martínez J, Dumas J, Lefebvre S (2016) Procedural voronoi foams for additive manufacturing.
ACM Trans Graph 35:1–44. https://doi.org/10.1145/2897824.2925922
Martínez J, Song H, Dumas J, Lefebvre S (2017) Orthotropic K-nearest foams for additive manu-
facturing. ACM Trans Graph 36:1–121. https://doi.org/10.1145/3072959.3073638
Mizrahi M, Golan A, Mizrahi AB et al (2016) Digital gastronomy: methods and recipes for hybrid
cooking. In: Proceedings of the 29th annual symposium on user interface software and technology.
ACM, New York, pp 541–552
Nakashima G (1981) The soul of a tree. In: Adamson G (ed) (2009) The Craft reader. Berg Publishers,
Oxford, New York, pp 219–225
Noll T (2009) The joint book: the complete guide to wood joinery. Chartwell Books, London
Oh S (2015) Plumb modular furniture. In: Sehun Oh Design. http://www.sehunoh.com/plumb.html.

Reas C, McWilliams C (2010) Lust: form and code. Princeton Architectural Press, Princeton
Rosner DK, Ryokai K (2010) Spyn: augmenting the creative and communicative potential of craft.
In: Proceedings of the SIGCHI conference on human factors in computing systems. ACM, New
York, pp 2407–2416
Saul G, Lau M, Mitani J, Igarashi T (2011) SketchChair: an all-in-one chair design system for end
users. In: Proceedings of the fifth international conference on tangible, embedded, and embodied
interaction. ACM, New York, pp 73–80
Seike K (1977) The art of japanese joinery, 1st edn. Weatherhill/Tankosha, New York
Studio Minale-Maeda (2014) Keystones. http://www.minale-maeda.com/KEYSTONES. Accessed
27 May 2018
Tsaknaki V, Fernaeus Y, Rapp E, Solsona Belenguer J (2017) Articulating challenges of hybrid
crafting for the case of interactive silversmith practice. In: Proceedings of the 2017 conference
on designing interactive systems. ACM, New York, pp 1187–1200
Yao J, Kaufman DM, Gingold Y, Agrawala M (2017) Interactive design and stability analysis of
decorative joinery for furniture. ACM Trans Graph 36. https://doi.org/10.1145/3054740
Zehnder J, Coros S, Thomaszewski B (2016) Designing structurally-sound ornamental curve net-
works. ACM Trans Graph 35:1–99. https://doi.org/10.1145/2897824.2925888
Zheng C, Do EY-L, Budd J (2017) Joinery: parametric joint generation for laser cut assemblies.
In: Proceedings of the 2017 ACM SIGCHI conference on creativity and cognition. ACM, New
York, pp 63–74
Zoran A (2013) Hybrid basketry: interweaving digital practice within contemporary craft. ACM
SIGGRAPH 2013 Art Gallery. ACM, New York, pp 324–331
Zoran A (2015) Hybrid craft: showcase of physical and digital integration of design and craft skills.
ACM SIGGRAPH Art Gallery. ACM, New York, NY, USA, pp 384–398
Zoran A (2016) A manifest for digital imperfection. XRDS 22:22–27. https://doi.org/10.1145/
2893491
Zoran A, Buechley L (2012) Hybrid reassemblage: an exploration of craft, digital fabrication and
artifact uniqueness. Leonardo 46:4–10
Zoran A, Shilkrot R, Nanyakkara S, Paradiso J (2014) The hybrid artisans: a case study in smart
tools. ACM Trans Comput-Hum Interact 21:1–15. https://doi.org/10.1145/2617570
Geometric Patterns and the Art
of Kundekari in Traditional Turkish
Woodworking
E. Sibel (Onat) Hattap
Abstract One of the most important areas contributing in the wealth of Turkish
art is the woodworking that displays fine aesthetic spirit in the best way. Once the
Turks arrived in the Anatolia, they sustained their traditions of art remained from the
Central Asia in field of woodworking as the case of every field. It is possible to deter-
mine the most beautiful and richest examples of this in the art of Anatolian Seljuk
state. Anatolian Seljukians has not only utilized wood as a construction material in
architecture but also formed some other handcrafts made of wood such as mimbar,
rostrum, lectern, the Koran case cover, drawers, symbolic coffin and other fine art
works. It is seen that Anatolian Seljukians and Ottomants mostly use the techniques
of carving (embossing), kundekari (build-up, interlock), latticed carving and dyeing.
Being made by combining several components together by means of lathes without
using nails and glue, kündekâri technique is an expensive and very demanding tech-
nique that prevents breakage, cracking and distortion of wood due to external effects
such as moisture and heat. The contribution of kundekari techniques in traditional
woodworking shall be investigated in the study.
Keywords Wood and geometry in history and traditional · Woodwork

Kundekari · Geometric patterns
1 Introduction
Woodworking art has been developing with considerably successful workmanship

without being overshadowed by wonderful stonemasonry since the Seljukian period.
It has been used in religious monumental structures, gates, mimbar (pulpitum), cab-
inet doors. Simplicity has been preferred especially on exterior facade while indoor
areas have been enriched and woodworking art has started to be used in greater num-
ber of areas during the Ottoman Empire era. Kundekari technique has been used in
E. S. (Onat) Hattap (B)

Mimar Sinan Fine Arts University, Istanbul, Turkey
e-mail: sibel.hattap@msgsu.edu.tr; sibelhattap@gmail.com

464 E. S. (Onat) Hattap
various areas like gates, window lids, mimbar (pulpitum), preaching desk. Depend-
ing on significance of cultural asset, some transition from simple, quinary, nine-fold
use to geometric patter varieties has occurred and the works have been enriched with
some metal ornamentations like pearl, ivory, tortoise shell, brass. Previously used as
a constructional element in architecture, timber has taken a turn towards art by the
various ornamental techniques used. Kundekari technique has been used especially
in the ceilings, doors, windows and closet doors has become one of the indispensable
techniques in the Ottoman.
The components used in the kundekari technique that has a very significant posi-
tion in woodworking do not include any fitting such as adhesive or nail, all compo-
nents are brought together with the meshing technique and requires good command
of geometry knowledge. The use of nails or glue is not preferred. Attention is taken
to ensure that wood’s water is on the opposite direction in every piece standing side
by side in a wood ornament made by use of kündekâri technique. By this way, any
separation or rotation thay may arise from drying of woods does not occur in time
regardless of how large the sizes in a plate made with this technique. A kündekâri
plate is made up a number of small interlocking pieces, a wooden carrier frame is
constructed in general on the back side to ensure hold itself and for this reason front
and back sides are different.
Being a commonly used woodworking technique by Anatolian Seljukians, kün-
dekâri was mostly used to manufacture door wings and particularly mosque mimbars.
The essential pattern is composed of a fabric made of the lines extending by intersect-
ing each other and the polygons, diamonds and asterisk shapes formed between these
lines in different forms in the kündekâri technique the earliest examples of which
can be seen in Egypt, Aleppo and Anatolia in 12th century. Inside of these inter-
faces is filled with curved branch patterns made with the carving technique. Each of
polygon and asterisk shapes is made by being processed from a separate wood piece
individually and diligently, broken lines among them are made of separate lathes. A
hollow slot is made along with a pattern in a way opposing to each other on each of
these pieces and a bulge is made to interlock the slot of the other pattern that pattern
corresponds to. Kündekâri technique is achieved by means of interlocking of slots
and bulges made in a way that the patterns correspond to each other. Lamp-dowel
interlocking application is an indispensible character of the kündekâri technique.
Indent technique has been mostly used during the Ottoman period and pearl, ivory
and tortoiseshell have been used as the indent material. The panel on side stern-
boards processed with the kündekâri technique is composed of octagon asterisks on
the circles centers of which cut vertically and are tangent to each other and pentagon
asterisks around them and hexagon pieces around them. On the other hand the bottom
of the minaret balcony consists of various Rumi ornaments on asterisks, hexagons
and octagons on the areas where the lines of geometric pattern that includes decagon
asterisks on the center are intersecting.
Geometric Patterns and the Art of Kundekari in Traditional … 465
2 General Characteristics of Kundekari Technique
2.1 Real Kundekari
Real kundekari is constituted by inner filling pieces and lathes which surround these
pieces. The components used in the kundekari technique that has a very significant
position in woodworking do not include any fitting such as adhesive or nail, all
components are brought together with the meshing technique and requires good
command of geometry knowledge. Upper part of these lathes is ornamented with
various profiles and the edges are grooved in order to ensure placement of inner
filling pieces. The trimming table that is also called as ornament or decoration is
created with combination of backbone system and inner filling pieces and with the
outer locking frame that surrounds this. By means of very thin joints occurred among
the components existing on this table, separations do not occur in shrinkages in
case wooden material contracts by loosing moisture and any disruption does not
occur on the trimming table by the action occurring in the grooved combination. For
that reason, real kundekari is not only an ornamentation art but also a method of
combination (Özdemir 1999).
In kundekari that is implemented by means that the edges have been carved as
negative and positive, in other words male and female, cut separately in the shape
of polygon and asterisk, the pieces ornamented with rumi and palmet embossments
and wooden girders are meshed to each other, decoration composition is based on
a geometric scheme. It is implemented with a variety of geometric patterns like
asterisk which symbolizes infinity, octagon, diamonds. The composition is colored
by placing small wooden plaques in different kinds and colors between each other
(http://ismek.ist/blog/icerik.aspx?p=496).
The foundation of the technique is based on combination of small timber pieces
in a way grains, so the curvature directions are opposite to each other with grooves
and cavities-rims. By this way, the problems like volume changes and distortions
which occur as a result of distortion mechanisms like water-moisture mostly affecting
wood and leading it to activate, namely activation of wood is prevented (http://www.
semerkandvakfi.org/699-2/).
Narlama, backbone lathes which form the backbone system combine with mor-
tise and gain resistance on the system by extending along one axis. If the existence
of narlama lathe along the kundekari trimming table that is created by alternating
geometric figures is to disturb continuity of pattern, inner filling pieces and back-
bone lathes are apparent on the front surface while narlama components which are
continuing as a whole and connected to outer locking frame are used on the lower
surface. By this way, ornamentation is provided on the front surface and construction
is provided on the lower surface.
Inner filling pieces are the wooden pieces having various geometric shapes with
a thickness about 16–20 mm on the edges of which lamps or grooves with 3–4 mm
depth and 5–6 mm width are opened. These pieces may undergo some processes
such as chamfering, carving and indent to improve ornamentation of the work. The
Fig. 1 Edirne Selimiye

Mosque Geometric pattern
kundekari door (Hattap
2015)
inner filling piece is combined as interlaced with their correspondences located on the
lathes which form the backbone system. Adhesives like glue etc. is not used on joining
points, and they are made smaller than the hollows remained between backbone lathes
in order to release superficial expansion that may occur by humidification of wooden
material.
The outer locking frame surrounds all elements of the kundekari system and makes
it a stable structure. The lathes of this frame are wider than backbone lathes, the inner
parts have the same detail with one edge of the backbone lath. On the other hand,
the outer edge is manufactured in compliance with the system according to work
characteristic (Kürklü 2011).
Mostly used in door flaps and side surfaces of pulpits, this technique is the most
original and the hardest. The first examples of this technique in Islamic art come from
Egypt, Aleppo and Anatolia in the XII. Century and these examples found seem to
be concurrent in all three centers. Kundekari is a technique that requires expert
workmanship. It fastens octagons, rhombuses, asterisk and other such geometric
shapes by a framing technique. These wooden shapes are attached to each other by the
usage of wooden laths grooved in order to produce interlocking elements. Adhesives
or mechanical fasteners are not used. The interlocked compositions of the pieces are
secured by a wooden carcass behind the surfaces, since a possible problem is the
drying and detaching of the elements. The pieces, which are usually geometric shapes,
are decorated with arabesk (arabesque), rumi (zoomorphic ornaments), botanical
ornaments and flower motifs, either etched or embossed (Ersoy 1993) (Fig. 1).
In religious works produced in 15th century in Bursa, in door and window shutters
and pulpits in cities such as Edirne and Manisa it is possible to see early Ottoman
works where real kundekari has been used.
The geometric compositions seen in Anatolian Turkish Art are usually applied to
structural timber. Materials preferred for these kinds of applications are oak, ebony,
walnut, boxwood, lime, apple, pear, cedar and rose trees. Woodworking, brought
to Anatolia by the Seljuks has reached its peak in the 12th and 13th centuries. In
the Clans period of the Seljuk Dynasty it has been possible to distinguish this art
according to its geographic location, however in the Ottoman period, this art has
become enriched by new materials introduced and complex compositions produced
(Mülayim 1982).
2.2 False Kundekari
Applied mostly in pulpit side plates, door and window shutters, this technique dif-
ferentiates in itself in ways such as nailing and embossing, or nailing and gluing.
Different wood blocks are brought together in this technique. These blocks are again
in the geometric forms such as octagons, hexagons, stars and rhombus shapes. These
blocks are usually decorated with embossed botanical motifs. The laths between the
wooden blocks are fixed in place by nails. In this cruder and less expertise requiring
technique crevices occur between the blocks after the drying out of the wood. In the
nailed and embossed kündetori technique, laths, which outline the geometric design,
are nailed between protruding surfaces. In nailed and glued kündekarı\laths are nailed
on top of the wooden block surfaces in order to create the geometric scheme (Ersoy
1993). Grooving depth does not go over 1 cm in any of the examples. However
this technique has produced examples, which imitate the kundekari technique with
deeper grooves such as the pulpit and window shutters of Divriği Ulu Camii (1241)
(Mülayim 1982).
During the 14th century kundekari becomes enriched by the support of inlay
technique. Ottoman woodworking has produced astounding works with the use of
Seljuk techniques such as carving and framing, and also with the use of materials
such as mother-of-pearl, tortoise shell, ivory, bone and even gold and silver. Produced
by the hammering in or gluing of the material into the receptacle hole opened in the
wood, this technique produces geometric surfaces, which can be distinguished by
color and material (Mülayim 1982).
Overall among the techniques, Kundekâri technique is commonly used on side
backboards and doors of mimbars, carving technique is used door and window sashes,
in various places like lecterns, mimbar, mihrab (altar), desks and sarcophaguses;
bevel cutting technique is used on side backboards of mimbars; framing technique

is used on mimbar guard rails; hemstitch technique is used in crown parts of mimbar
doors and lecterns, Mashrabiye style framing technique is used on lecterns, on-
wood painting technique is applied on column heading, consoles and beams (Soysal
Demirci 2016).
3 Geometric Motif Use in Kundekari Technique
The geometric ornamentation art has gained its current form by being synthesized
with Turkish-Islam culture with pre and post Islam period.
Particularly we can see many works of art performed in field of geometric motif
ornaments during the Anatolian Sejukian era. On the other hand, the symbolic mean-
ing at the origin of such motifs lost and they were let turn into a decoration material
in time during the Ottoman period.
Geometric motifs are formed by combination of many geometric shapes like dots,
lines, squares, rectangular, triangles, circles and asterisk and symbolize the infinity
of the universe in terms of meaning. Geometric compositions arise from combination
of geometric shapes with broken, curved and straight lines. Since drawing human
and animal paintings and carving sculptures is not permissible in the Islam, the artists
have become skilled in geometric motifs (Figs. 2 and 3).
Fig. 2 Geometric patterns used in Kundekari technique (Yalçın 1997)

Fig. 3 Kundekari example with asterisk in Istanbul Topkapı palace window lid (Yalçın 1997)
Even though geometric motifs are used in every branch of ornamentation, they
are particularly used widely in wooden materials. The kundekari technique that does
not require any joining agent but just produced with interlocks is the most beautiful
example of geometric ornamentation. These motifs are used also in grid carvings
made of wood. Geometric motifs may be divided into two groups as polygon and
circular motifs and border and interlocking motifs. Polygon motifs include triangle,
Turkish triangle, asterisk, multiaxial motifs, circles, arches and interlocks within cir-
cular motifs, polygon, asterisk, zigzag, shuttle, whirligig, meander, interlocks within
border motifs, meshes and guilloches in interlock motifs. These motifs may be both
used as plain and as dual and triple forms (Megep 2008) (Figs. 4 and 5).
4 Kundekari Examples
Bursa Grand Mosque mimbar is the most monumental one among the examples
belonging to Seljukian and Principalities period. In this work of art that also includes
a perfect implementation of real kundekari technique, wide expansions of geometric
compositions were able to be placed on surfaces since the pieces were made smaller
and mimbar size was made larger. Large and small embossments some of which
are made with kundekari technique have gained a considerably moving appearance
Fig. 4 Istanbul Suleymaniye

Mosque post restoration
2018)
particularly on the east front of the mimbar that views the mihrab side. Both door
wings and doorjambs of the mimbar are also made by kundekari technique. The use
of kundekari technique on doorjambs is an implementation that is not familiar for
the Anatolia (Figs. 6 and 7).
The reduction of size in geometric pieces starting by the mimbar gate in Manisa
Grand Mosque has been seen in fewer examples until mid 15th century, but it has
started to become common beginning from door wings of Edirne Üç Şerefeli Mosque
(M. 1447); by this way the kundekari technique which had been rarely preferred in
door and window gates in the past has started to become widespread beginning from
the mid century and this technique has become dominant almost all examples by the
16th century. In parallel with this, plant ornamentations on geometric pieces have
become plain and carving depth reduced; inlaying and tarsi technique has increas-
ingly continued; gradually all pieces have started to be evaluated with the inlaying
technique (Figs. 3 and 8).
Once marble mimbars have started to become widespread after mid 15th century
in the Ottoman period, traditional wooden mimbars which may be considered as mon-
umental have started to disappear gradually, so the use of real kundekari technique
on mimbars passed into oblivion; conversely it has found different implementation
Fig. 5 Istanbul Kılıç Ali

Paşa Mosque Kundekari
door (Orbeyi 2005)
areas such as preaching desks, cabinet doors etc. in addition to door and window
wings (Fig. 9) (Bozer 2002).
The most beautiful examples of kundekari are found in Alhamra Palace located
in Granada city of Spain in addition to Turkish woodwork art. In this palace the use
of kundekari technique is also apparent in ceilings which is rare in Turkey (Figs. 10
and 11).

2018)
5 Kundekari Technique in Restorations
The building elements such as doors, windows, preaching desks made in kundekari
technique considered among restored cultural assets are today mostly displaced and
ensured to be protected against deterioration mechanisms. It is observed that wooden
building elements become porous and their sections have weakened particularly by
the influence of insects and worms which prefer wooden material because of their
foodstuff preference (Tarim and Yardimli 2017). Once the damage analysis of the
current situation is made, biological, chemical and mechanic deteriorations occurred
on the structure are eliminated. There are some methods for this. For example, micron
scale granules obtained from nutshell are applied by means of spraying method, so
the surface is corroded and cleaned without disturbing the details for the patina
layer formed on the surface. Some methods are applied to prolong life of wood and
make it resistant against heat and moisture by destroying some living organisms
which are formed on some buildings elements like door, window lids and result
in biological deterioration, deficient components are completed. While doing this
completion process, the same material is tried to be used by considering which tree
has been used for making the wooden material in the original kundekari work.
Fig. 7 Istanbul Rüstem Paşa Mosque Kundekari door (Hattap 2015)
Fig. 8 Istanbul Yeni Valide Mosque Kundekari door (Holat 2014)


kundekari window lid
(Hattap 2018)
The use of impregnated substances which are water-soluble has also considerably
increased in recent years. Impregnation is used to prolong life of material in wooden
kundekari doors and preferred in outdoor areas. Odor does not cause a problem
in impregnated wooden material, surface processes may be applied on wood after
impregnation. Durability of the materials can be ensuredby means of these impreg-
nated substances (Koca et al. 2013). Rüstem Paşa Mosque (1561) of the Architect
Sinan located in Istanbul indicates design richness with its outstanding landscaping
that has been shaped as a result of some necessities, there is also a bazaar on the ground
floor because of the land value (Kalle et al. 2015) Deficient pieces are being completed
in kundekari doors disassembled in Rustem Pasa Mosque that is currently restored
(Figs. 12 and 13). Wooden door and window lids made of kundekari workmanship
have also been disassembled and placed in an assembly created in the mosque gar-
den in Suleymaniye Mosque (1550–1557) located in Istanbul that has been recently
restored and mentioned as the masterpiece by the Architect Sinan; additionally they
are ensured to be protected from biological degradation mechanisms by performing
a chemical application for insects and worms inside the assembly, remaining pro-
cesses have also been completed and the door and window lids reassembled on their
Fig. 10 Kundekari door in

Granada Alhamra Palace
(Hattap 2017)
location. Life of wood material has been prolonged and its sustainability has been
ensured by means of protection and repair methods implemen ted (Figs. 6 and 9).
Surfaces of works are being lubricated and polished in order for protection of
works processed with Kundekâri technique against weather conditions and give them
a live appearance. They are occasionally colored with natural paints. For this purpose,
mostly linseed oil, rarely olive oil or bitter almond oil and turpentine (by being mixed
in equal ratio with pure pine naphtha oil) are applied for this purpose. Or some
vegetable oils like teak oil are used and shellac polish is applied to give a bright
appearance (Soysal Demirci 2016) (Fig. 14).
6 Conclusion
In our modern day, the number of master artisans can be counted with the fingers
of one hand, and they have a hard time finding young craftsmen to teach this art to.
In the restoration works performed, unfortunately, kiindekari does not carry the soul
of older works, and its color, the texture of wood, the varnish produce a much more
false effect which is also due to the lack of ageing and life prints on the material. In
reality it is very hard to extend the lifespan of this art, which has come through the
Fig. 11 Kundekari ceiling

in Granada Alhamra Palace
(Hattap 2017)
Seljuks to the Ottomans, especially considering that the elements are wooden. The
examples, which have arrived to this day from the past, have to be restored, if not
they are sentenced to perish: which has happened to many examples until now. Due
to these reasons kundekari needs to be cared for. There is need for the tutoring of
kiindekari technique, and establishing coordination between universities and work-
shops. The craftsmen who can do Kundekari (kundekar) should be researched and
benefited from these persons in field of education, new craftsmen who could continue
this wood workmanship should be trained and continuity should be ensured by this
means. Suleymaniye Mosque and Soup-Kitchen Construction belonging to Ömer
Lütfi Barkan published in 1979 (1550–1557). The books preserved from publication
in the construction include registered the names of craftsmen who constructed the
products such as doors, windows, cabinet wings, lectern, the cities they are from,
the number of pieces they produced, and the amounts they have been paid (Barişta
2009). Such details indicate that craftsmen of that period were significant and high
in number, and greater efforts were taken to transfer wood art and ensure that to be
sustainable. If the fact that art and culture is universal is considered, the need for
the conservation of this technique as a part of world heritage in cooperation with
international universities arises.
Fig. 12 Istanbul Rüstem Paşa Mosque Kundekari door at repair stage (Hattap 2018)
Fig. 13 Istanbul Rüstem

Paşa Mosque quinary
Kundekari door at repair
stage (Hattap 2018)
Fig. 14 Wooden interlock fixing in Kundekari door at repair stage (Yalçın 1997)
References
Barışta ÖH (2009) Osmanlı İmparatorluğu Dönemi Istanbul Cami ve Türbelerinden Ağaç İşleri.
Ankara, Turkey, p 42
Bozer R (2002) Ortaçağ Anadolu Türk Ahşap Sanatında Kundekari Tekniği, Türkler, Cilt 7., Ankara,
Turkey, pp 911–916
Ersoy A (1993) Osmanlı Ağaç İşçiliği XV. Century Ottoman Woodworking, Marmara Üniversitesi
Yayın No (Marmara University Publication number): 509, Atatürk Eğitim Fakültesi Yayın No
(Atatürk Education Faculty Publication Number): 14, Istanbul-Turkey
Kalle S, Hattap ES, Orbeyi N (2015) Mimari Bir Sınırlama Örneği Tahtakale Rüstem Paşa Camisi.
9. International Sinan Symposium (161), Edirne, pp 199–204
Koca G, As N, Aroğlu N (2013) Ahşap Dış Cephe Kaplama Elemanları, 7. Ulusal Çatı ve Cephe
Sempozyumu, Istanbul, Turkey, pp 15–23
Kürklü G (2011) Geleneksel Türk Ahşap Sanatı Kundekari ve Günümüz Teknolojisine Sahip Atölye
Ortamında Yapılabilirliği, Afyon Kocatepe Üniversitesi Fen Bilimleri Dergisi, Afyon, Turkey, pp
13–20
Megep (2008) Mesleki Eğitim Ve Öğretim Sisteminingüçlendirilmesi Projesi İnşaat Teknolojisi
Geometrik Motifler, Ankara, Turkey, p 1
Mülayim S (1982) Anadolu Geometrik Selçuklu Çağı’ [Geometric Decorations in Anatotian Turkish
Architecture, Seljuk Period], Kültür ve Turizm Bakanlığı Yayınları (Publication of Ministry of
Culture and Tourism): 503 Sanat Eserleri Dizisi (Art Works Series):1, Ankara-Turkey
Özdemir F (1999) Kündekâri Tekniğinin Dekorasyonda. Uygulaması, 1. Ulusal Mobilya Kongresi,
Ankara, Turkey, pp 427–434
Soysal Demirci H (2016) Yozgatta Yaşayan Kundekari Ustasi Yakup Kılıç, Bozok Sempozyumu,
cilt: 3, Yozgat, Turkey, pp 592–601
Tarım A, Yardımlı S (2017) Title of paper deformations appeared in wooden frame residence
construction elements in Istanbul historical peninsula zone X. In: International Sinan symposium,
27–28 April 2017, neighbourhood, proceeding book, Edirne, Turkey, pp 291–302. ISBN: 978-
975-374-208-5
Digital Wood Trusses. Geometry
and Parameters/Fabrication
and Monitoring
Andrea Giordano, Paolo Borin and Federico Panarotto
Abstract This paper aims to demonstrate how to translate features and properties
of existing wooden structures about function, performance and aesthetic into BIM
objects. This purpose has a double effect: from an operative perspective, the mod-
elling of existing wooden structures is necessary for any structural restoration project,
while culturally for the preservation of construction techniques and their transferabil-
ity in the current design theory. This refers in particular to wooden trusses as com-
pound objects, such as to comply with a single structural and configurative purpose
within the built space. The study proves the feasibility of describing the relationship
among geometry, material features, construction techniques and simulation of trusses
within a BIM environment. From a methodological point of view, starting from a
point cloud, the process involves the modelling of beams and trusses, the construc-
tion of a library of nodes with parametric geometry drawn from nineteenth-century
treatises (also called “wood stereotomy”), the information exchange from the BIM
object to structural simulation environment. Moreover, there is the development of
a parametric relation structure, which enables the creation of cutting parameters of
the truss’s structural components. The case study from which the system is obtained
is the covering system of the Church of the Eremitani, which hosts a so-called revo-
lutionary covering system by a fourteenth-century monk, Giovanni degli Eremitani.
Keywords Wood stereotomy · BIM · IFC · Parameters · Interoperability
A. Giordano (B) · P. Borin · F. Panarotto

University of Padova, Padua, Italy
e-mail: andrea.giordano@unipd.it
P. Borin
e-mail: paolo.borin@unipd.it
F. Panarotto
e-mail: federico.panarotto@unipd.it

480 A. Giordano et al.
1 Introduction1
There is an analogy between the meaning of algorithm and the concept of ‘repre-
sentation’ of architecture. Let’s try to consider the aspects that make representation
in general—and design in particular—an ‘operation’ of problem-solving, where the
latter is meant as the set of criteria and processes in line with specific standards
for the analysis, the definition and the concrete explanation of a problem. These
three actions combine design and representation with the concept of algorithm: the
algorithm turns out to be the set of rules and laws to be followed in calculation and
problem-solving operations, in particular when using the computer.
Likewise, if we add the meaning of the term ‘parametric’ to ‘design’, we imme-
diately see how the latter is suited to the operation of parameterization: thus, design
is also that set of values and parameters that aim at describing, as well as critically
comparing, rules.
Finally, there is a further meaning that can be suitably added nowadays to repre-
sentation, therefore intended as an operation of ‘simulation’, not in terms of mere
imitation or copy, but accepting the meaning of use of a model to ‘virtually repeat’
certain conditions for purposes of study, research, experimentation, training.
Therefore, nowadays we refer to a new algorithmic and parametric way of rep-
resenting architecture, aimed at simulating reality, whether existing or in progress.
This is the aim of this study, which puts in relation parameters and algorithms to
represent reality, in this case constructive.
This paper is part of a broader study that aims at enhancing those ancient construc-
tion techniques strongly linked to the use of descriptive geometry, in the light of the
new processes of architectural production. The advantage of this type of approach
lies in the fact that geometry is constituted as the common element that allows this
passage. If in the past this geometric complexity was defined and conveyed by spe-
cialized treatises, nowadays it is used through modelling. However, it is to be stated
that the role of geometry has never been linked exclusively to configurational or
aesthetic aspects. Architectural systems made the geometric, performative and pro-
ductive requests strongly interdependent. Evans was among the first to find a link
between architectural aspects and scientific development in the field of descriptive
geometry (Evans 2000). The result is a relationship between architectural needs,
complex paper or tin models for stone cutting for the construction of, for example,
bizarre trompe, and technological advancement, towards the Mongian horizon of
orthogonal projections. This paper investigates the possibility of replicating this pro-
cess, extending the study on one hand of historical treatises, on the other of the need
to safeguard the built heritage to understand if and which links there are with the
temporary. As recently reported by Antoine Picon, the digital fabrication movement
in the architectural field is going through a nostalgic phase: it is no coincidence that
the work of one of its greatest exponents, Philippe Block, identifies the origin of his
studies both in Gothic cathedrals and in nineteenth-century studies about the theory
1 Andrea Giordano and Paolo Borin wrote Chap. 1; Paolo Borin and Federico Panarotto wrote
Chaps. 2, 3 and 5; Federico Panarotto wrote Chap. 4.

Digital Wood Trusses. Geometry and Parameters/Fabrication … 481
of construction. This study shows how the Building Information Modelling is nowa-
days the best tool to give continuity to the use of the best construction methods. This
continuity is certainly useful on the maintenance level, but, on a subtler level, it has
a procedural nature and deals with geometrical knowledge, integration of different
and distant multidisciplinary fields, relation of design qualities, suitable use of the
available materials. In a scientific production by now based on the terms optioneer-
ing and optimization, this study differs by evaluating the continuity of these terms
in time. The culture established as OPENBIM also imposes a correct exchange of
information, which is addressed for the structural part through the IFC standard. The
use of the standard not only allows to show the hierarchy of information between
main objects (the cover) and secondary objects (the beams that compose it), but it also
determines the relationships between the parameters of the main objects and those
at the base of the manufacture, corners of cutting, at the same time representatives
of the structural design constraints.
Section 2 identifies a reference literary review, Sect. 3 describes the methodology
used, Sect. 4 shows the meaning of the research for the case study of the Church of
the Eremitani in Padua, where the spatial configuration is subjected to a series of
wooden vaults.
2 Literary Review
This article is based on three pillars in literature: the BIM parametric design and return
methods, the digital production of building components, and the structural analysis
of these components and its sub-components. In a recent publication about this topic,
an interview with Manfred Grohmann and Jan Knippers confirmed this domain
of applications and research in support of wooden constructions: “The advance of
computational design, simulation and fabrication has significantly expanded the pos-
sibilities of wooden structures and related engineering approaches in recent years”
(Krieg 2016). In a taxonomy of the problems related to the new wooden design, there
are some critical issues that persist in complete wooden buildings. “[…] the complex-
ity in wooden processing, where both subtractive and additive fabrication methods
are necessary, made it more challenging to achieve a high degree of automation”
(Krieg 2016). An example of solving of these problems can be found in the activity
of the studio Foster and Partners, that uses digital fabrication for structural covering
components inspired by the mechanical functioning of reticular beams (Maddock
et al. 2017). The main difficulty for wooden constructions is the approach according
to which “truss structures are still designed in such a way that by using complex steel
connectors, the geometry of wooden beams is kept simple” (Krieg 2016). There is no
doubt that the link between the project and its mass realization is based on a digital
modelling, simulation and manufacturing continuum.2 This digital continuum is
2 Referto the example of Nelson 1996 for a study about some differences between the connections
within wooden trusses.
defined in literature as ‘computational design’, and it questions “knowledge-based

planning methods that encapsulate the knowledge of a specific domain, such as
Building Information Modelling or Computer Aided Manufacturing” towards
behaviour-based approaches using an agent-based modelling and simulation method
(ABMS), which are however monothematic, as they cannot nowadays fully describe
the complexity of the building in its multiple disciplines (Schwinn 2016).
There are not many contributions that show the benefits of using a BIM model for
prefabrication. Hamid and others developed a BIM approach for the construction of
wall components. Although it is clear that often formalized knowledge is embedded
into a fabricator-specific application, they conclude that the results achieved deal
with a multitude of domains, ‘reducing the need for design reviews, obviating the
need for shop drawings, automating part list calculation and hardware enumeration
activities […], part of the process planning and part programming required for the
CNC machining of the cabinet parts’3 (Nelson 1996).
Currently, with the increasing use of BIM applications for the construction indus-
try, the role of component manufacturers has generally increased, as they provide
designers with digital objects that represent their actual components in the manu-
facturing sector. These BIM objects (also known as models of construction objects)
can:
• help designers to select the most appropriate component in line with the design
aim;
• facilitate the provision of the product;
• improve the workflow between design and production disciplines;
• guarantee a high level of personalization.
Despite these advantages, commonly available BIM objects fall into specific com-
ponent production categories: MTS (Made-to-Stock) components and MTO (Made-
to-Order) components. In this sense, they delimit the designer to a fixed set from
the manufacturer’s catalogs. This is rarely the case with building components, as
some require a high degree of customization to meet the specific requirements of a
project. Thus, these components align with another category: the ETO components
(Engineered-to-Order) (Hamid et al. 2018).
ETO components generally involve a more complex work process than their MTS
or MTO counterparts. By using BIM, the production of these customized components
could be facilitated by what literature indicates to be a superimposable relationship
between designers and manufacturing professionals in which there is a bidirectional
dependence on the information shared in the workflow between them (Eastman et al.
2011). It is thus clear that this production approach fits very well the need for occa-
sional replacements of wooden elements, because in the ETO process the procedure
of engineering only begins when the order arrives and guarantees another degree of
customization.
3 Inthis regard it can be mentioned the plugin Autodesk Revit, Wood Framing Wall+ http://www.
aga-cad.com/products/bim-solutions/wood-framing-wall.
There are many initiatives in this sense, such as studies on the development of
processes aimed at improving the ability to customize the building elements through
techniques of parametric and generative design. Among the most interesting ones,
it is noteworthy the work of Monizza et al. (2017), who directed their efforts in the
creation of a specific algorithm for GLT (Glued-Laminated Timber) products aiming
at reducing the use of unnecessary raw materials and facilitating the production
process of complex products, like those with a double curvature. The study shows
how, with the increase of prototypes and the general growth of procedural reliability,
these design techniques can be effective to push the AEC (Architecture, Engineering
and Construction) sector towards mass customization techniques.
New computational opportunities, which allow a direct production of CAM
(Computer-Aided Manufacturing) files for CNC machines (Computer Numerical
Control) and customized products directly based on the design needs, encourage a
stronger control of the constraints and production possibilities.
The mutual relationship between software progress and the use of computer
numerical control (CNC) or robotic arms in construction has opened new inves-
tigative frontiers that push towards high-performance systems and components with
reasonable timing and costs. The main advantage derives from the possibility of
producing building components through unique pieces that otherwise could not be
economically sustainable to be produced using traditional manufacturing techniques
and processes (Paoletti 2017).
By the way, one of the characteristics of wood, unlike other materials, is the pre-
disposition to be cut. Nowadays, cutting methods can be expressed with processes
called “advanced production techniques”, in this case made using CNC systems
(Computer Numerical Control). To date, the practice that is used allows the realiza-
tion of complex cutting procedures using CNC machines that receive data translated
by a CAM process and which, in turn, interpret values derived from a 3D CAD model.
In addition to the increasing use of CAD/CAM technology in the engineering and
construction industry (AEC), a parallel but less developed field of work involves the
use of BIM applications in these types of process (Hauschild and Karzel 2011). It
emerges from a recent report that a process that relates the BIM methodology and
CNC production has the potential to achieve what is called model-based prefabri-
cation, where the fabrication of components is directly related to the information
that comes from the designer’s BIM model (Jones and Bernstein 2014). This direct
fabrication allows to relate the geometries of a BIM model, linked by parameters, to
latest-generation robotic arms that allow the creation of multi-axial cuts of consid-
erable complexity. The 6-axis industrial robot, for example, allows the fabrication
not only of 2D curved surfaces, but also of high-quality 3D geometries through the
continuous rotation of the blades, which can guarantee a higher resolution compared
to traditional milling (Yuan and Chai 2017).
The use of BIM objects to convey knowledge of buildings is already known (Oreni
et al. 2013; Fai et al. 2013). In this case, the geometric and parametric production
of the objects represents a process of reverse engineering of this same process,
serving as much for cognitive reasons related to the restoration and protection of
the components, as for didactic purposes. This passage allows a direct connection
Fig. 1 Mathurin Jousse,

L’art de charpenterie”, 1702
with the treaties of wooden stereotomy that make it possible to identify the geometric
and technological relationships between the matching elements of a truss. The main
application of descriptive geometry is found in nineteenth-century treatises, with
particular interest in wood-wood connections in which operations of intersections of
solids and rotations of planes are recalled. Although the first stone stereotomy treaties
also contained references to the cutting of wood, a subtractive technique useful to
the construction of the support ribs of the vaults, only some of them integrate the
discourse to describe the connections between the covering elements. These include
the treatise by Mathurin Jousse, (L’art de charpenterie”, 1702) (Fig. 1) written in
collaboration with Gabriel Philippe de La Hire, or the one by Nicolas Fourneau
(L’art du trait de charpenterie, 1767) (Fig. 2), and the nineteenth-century treatises
by Amand-Rose Émy (Traité de l’art de la charpenterie, 1870) (Fig. 3), Jules Pillet
Fig. 2 Nicolas Fourneau, L’art du trait de charpenterie, 1767, Planches 5 and 15
(Traité de stéréotomie (charpente et coupe des pierres, 1887) (Fig. 4) and Charles-
François-Antoine Leroy (Traité de stéréotomie: comprenant les applications de la
géométrie descriptive, 1877) (Fig. 5).
3 Methodology
This chapter describes the particular methodology used in the BIM parametric mod-
elling of the roof of the Church of the Eremitani, which will then be concretely
presented in Sect. 5. The process ideated for this type of work refers to the specific
context on which the work is carried out, the built cultural heritage. The text deals
with the description of the techniques used to manage, process and share the geomet-
ric information of the structure that supports the roof, composed of wooden trusses.
The development of BIM models for the historic construction industry exceeds the
practice usually used for new buildings in some cases. This type of procedure often
forces the operator to face and solve specific problems. Normally this procedure turns
into the design and creation of new objects and methods that do not fall within the
Fig. 3 Amand-Rose Émy, Traité de l’art de la charpenterie, 1870, Planches 41 and 68
standards of traditional modelling software. The consequences of this are the devel-
opment of processes that enable to improve the semantic content of BIM objects and,
more generally, the methods used to create them.
Fig. 4 Jules Pillet, Traité de stéréotomie (charpente et coupe des pierres), 1887, p. 38, Pic. 95
The ‘Eremitani project’4 to which we refer involved several people who dealt
with different issues. Starting from an accurate survey thanks to the most advanced
techniques, we aim at describing the methodological approach used for the realization
of the BIM model of the ecclesiastical complex and, above all, the characteristics
identified for a correct sharing of the information involved. One of the most important
aspects of this work is the quick and optimal circulation of the information obtained.
The interoperability of a BIM model is guaranteed by the IFC standard, which allows
the sharing of data between different software and, above all, allows the digital model
to be relevant throughout its lifecycle and the legacy data not to become obsolete
and un-usable. This aspect is more important for a historical building than for the
common building records.
In Sect. 3.1 there is a brief introduction to the issues related to wood stereotomy,
with particular attention to the functioning of the dry assembly of wooden structures
4 FSEproject carried out between 2014 and 2015, thanks to P.O.R. VENETO F.S.E. 2007–2013,
Human Potential Development Area, titled: “Methodologies for the acquisition, processing and
communication of data related to cultural heritage and architectural and technological interventions
aiming at the conservation and improvement of the use for cultural tourism”. DICEA, University of
Padua, resp. Luigi Stendardo, co-directors Andrea Giordano, Agostino De Rosa, Stefano Zaggia.
Fig. 5 Charles-François-Antoine Leroy (Traité de stéréotomie: comprenant les applications de la

géométrie descriptive, 1877, Planche 68
and the characteristics of the material. A work of this kind can only start from an
extensive study of the wooden construction techniques described in the treatises of
stereotomy. In Sect. 3.2 there is an evaluation of the operations to be performed to
generate a correct process of information sharing through the IFC.
3.1 Stereotomy and Woodworking
Stereotomy of wood is defined as the set of operations aimed at modifying the pris-
matic or parallelepiped shape of the seasoned wooden boards, suitably arranged to
allow their use in construction industry and furniture (Paris 2009). The stereotomy
of wood is therefore the fundamental discipline for woodworking, i.e. for the con-
struction of direct or indirect supporting structures. The study of wood stereotomy
was widely dealt with in some important nineteenth-century scientific treatises that
refer to the discipline of post-mongian descriptive geometry, such as the treatises
on stereotomy by Leroy (1870) or by Pillet (1887). In the first treatise, the topic is
developed in the chapter dedicated to the charpente (carpentry), together with other
topics on the applications of descriptive geometry, such as the theory of shadows,
the linear perspective, the gnomonic and the cutting of stone. The second treatise is
dedicated exclusively to stereotomy and is divided into two parts: the first deals with
the charpente and the second one with the ‘coupe des pierres’, i.e. the cutting of stone
(Paris 2009). The differences between wood and stone are obviously substantial: the
Fig. 6 a Simple truss, b simple truss with king post, c Palladian, d composite truss, e composite
truss with webs (retrieved from: Koenig et al. 1995)
geometric operations designed and applied to stone cases are not adaptable to the
wood ones, and vice versa. On the one hand, stone can be sculpted indifferently in
the three dimensions, keeping the characteristic of strong resistance to compression
constant; on the other hand, the cutting of wood is strictly influenced by the nature
of the material itself. In wooden works there is the predominance of one dimension
compared to the other two, so that in wooden structures, in addition to the normal
compression stresses, push-up stresses are often produced too. Moreover, being com-
posed of fibres, wood has different resistance capacities depending on the direction of
the forces and of the fibres themselves. Finally, in wooden structures, unlike in stone
ones, it is possible to use some pure traction elements, the so-called ‘tie-rods’. Cuts
made for unions can weaken their strength, so they must be designed appropriately.
The aforementioned treaties describe in detail the unions between the various
components of a wooden structure. Wooden unions are often composed of specific
cuts of wood, which create the optimal contact surface between two or more ele-
ments. These unions are often made through “Tenon” and “Mortise”, which allow a
connection between the various structural components in the most solid way and are
often used in wooden supporting structures.
Among wooden structures, the truss is certainly one of the most important and
used. The complexity of a truss can vary depending on the type used, from a simple
one—with the classic elements, such as bottom chord, top chords and king post—to
the most complex ones such as the Palladian, the composite and so on (Fig. 6). It is
known that trusses build the supporting elements of vast coverings: the typological
choice of a truss is functional to the characteristics of the space to be covered, defined
by particular shapes and dimensions and by the quality of the structural supports to the
wooden elements. A truss needs that every element that composes it must withstand
certain stresses, generally traction stresses (the king post and especially the bottom
chord), compression (the truss webs) and buckling (the top chords). To do this, the
structural elements must be correctly connected among each other. These junctions,
often complex, are realized through particular cuts of the wood described in the
treaties about wooden stereotomy.
3.2 The BIM Model: Parameterization and Interoperability
A BIM model contains a lot of information available to the numerous people involved
in the building process. For specific operations, different programmes compatible
with the BIM process are used. In this specific case, there is the need to express the
information contained in the file of the model to transfer it to the next simulation
phase.
Interoperability, i.e. the ability to exchange information between the various pro-
grammes, especially between BIM modelling platforms and complementary pro-
grammes related to specific disciplines, is guaranteed by live-links (plug-ins that
allow a direct exchange of information and developed ad hoc) or by an exchange
format shared with all software based on BIM technology. It is an IFC (Industry
Foundation Classes), managed by buildingSMART’s,5 an open format, interopera-
ble between the various operators. BuildingSMART’s purpose is to allow the sharing
of information throughout the lifecycle of any built environment asset between all the
participants, regardless of which software application they are using.6 The machine
readable, good quality data should be available for use throughout the design, pro-
curement, construction, maintenance and operation phases. For each asset the inten-
tion is that two versions exist—the real and the digital.
BuildingSMART’s technical core is based around a common data schema (model)
called IFC, that makes it possible to hold and exchange relevant data between different
software applications. It relies on the ISO-STEP EXPRESS language and concepts
for its definition, with a few minor restrictions on the EXPRESS language.7
In the modelling environment, the trusses were made thanks to a parametric truss
Item. As a matter of fact, a truss is the simplest processing of the reticular beam
concept and uses the triangle as the main isostatic element. The truss is composed of a
series of sub-elements called beams, and each one of them responds to characteristics
of the type to which they belong.
In the IFC format the truss can be defined by IfcElementAssembly8 (Fig. 7). The
IfcElementAssembly is an aggregate being composed by other elements and acting
as an assembly using the objectified relationship IfcRelAggregates, referring to it by
its inverse attribute SELF\IfcObjectDefinition.IsDecomposedBy. Components of an
5 BuildingSMART is the worldwide authority driving the transformation of the built asset economy
through creation and adoption of open, international standards.

6 https://www.buildingsmart.org.
7 Ibidem.
8 The IfcElementAssembly represents complex element assemblies aggregated from several ele-
ments, such as discrete elements, building elements, or other elements. The IfcElementAssembly
should have (and in most implementation scenarios it is mandatory) a relationship for its hier-
archical containment in the spatial structure of the project. The IfcElementAssembly is placed
within the project spatial hierarchy using the objectified relationship IfcRelContainedInSpa-
tialStructure, referring to it by its inverse attribute SELF\IfcElement.ContainedInStructure.
Subtypes of IfcSpatialStructureElement are valid spatial containers, with IfcBuildingStorey
being the default container. (http://www.buildingsmarttech.org/ifc/IFC2x4/rc2/html/schema/
ifcproductextension/lexical/ifcelementassembly.htm).
Fig. 7 Spatial containment and element aggregation relationships for IfcElementAssembly

(http://www.buildingsmarttech.org/ifc/IFC2x4/rc2/html/schema/ifcproductextension/lexical/
ifcelementassembly.htm)
assembly are described by instances of subtypes of IfcElement. In the case of the

truss, the IfcElements are represented by IfcBeam, a structural member designed to
carry loads between or beyond points of support, usually narrow in relation to its
length and horizontal, or nearly so. In the case of the simple truss, these elements
are the bottom chord, the top chords and the king post; in the case of more complex
trusses, queen posts, truss webs and other wooden elements must be added to the
elements already mentioned. Each element has its own characteristics:
• Type Use Definition: the IfcBeamType (if present) may establish the common
type name, usage (or predefined) type, common material layer set, common set of
properties and common shape representations;
• Property Set Use Definition: the property sets relating to the IfcBeam are defined
by the IfcPropertySet and attached by the IfcRelDefinesByProperties relationship;
• Material Use Definition: The material of the IfcBeam is defined by the IfcMaterial
or IfcMaterialList and attached by the IfcRelAssociatesMaterial.RelatingMaterial;
• Quantity Use Definition: The quantities relating to the IfcBeam are defined by the
IfcElementQuantity and attached by the IfcRelDefinesByProperties relationship;
• Containment Use Definition: The IfcBeam, as any subtype of IfcBuildingElement,
may participate in two different containment relationships. The first (and in most
implementation scenarios mandatory) relationship is the hierarchical spatial con-
tainment, the second (optional) relationship is the aggregation within an element
assembly;
• Geometry Use Definitions: The geometric representation of IfcBeam is given
by the IfcProductDefinitionShape, allowing multiple geometric representations.
Local Placement, Geometric Representations, SweptSolid Representation, Clip-
ping Representation and MappedRepresentation are included.
The IFC model contains every possible property that the individual objects and
the whole model have, but, in the analysis phase, we must be able to exchange only
useful information. For a correct structural analysis, for example, we need only infor-
mation regarding the analytical model and the acting loads, such as the geometry
of the structure, the nodes, the constraints, the material used containing the rela-
tive performance properties. In the case of a structural analysis, there is a specific
data-schema, called Structural Analysis Domain, which is intended to describe the
analytical model, the connections between elements and the forces applied through
entities and types. Finally, the analysis results can be included within the file, whether
they are expressed in forces or tension or movement. Speaking of the structural anal-
ysis, the IfcBeam elements that form the aggregate have a different behavior, and they
are then connected by nodes that guarantee the connection. The model is composed of
absolute entities such as IfcStructuralItem, that describes the objects of the analytical
model, where multiple items are grouped into an IfcSystem; IfcStructuralActivity,
in which loads are applied to the structural elements (described by IfcStructural-
Load), i.e. actions that generate certain reactions. The actions are then combined
(IfcStructuralLoadGroup); IfcStructuralAnalysisModel, the analysis model consists
of IfcSystem and load combinations, the result is the IfcStructuralResultsGroup.
The description of the objects of the analytical model (IfcStructuralItem) of the
IfcBeam elements of the truss are described by the ifcStructuralMember sub-entity:
IfcStructuralCurveMember, which is able to describe one-dimensional elements of
the analytical model (IfcBeam, IfcColumn). Two aspects can be specified: the Profile
of the element determined within a standardized list using IfcProfileDef and the
extrusion trajectory of the profile chosen with IfcProductDefinitionShape through
a typological description (IfcTopologyRapresentation) expressed by a well-defined
and parametrized curve described from the entity IfcEdge.
The possible types that can be assigned are:
• RIGID_JOINED_MEMBER: a member with capacity to carry transverse and axial
loads, i.e. top chords, truss webs, king post and queen posts. Its actual joints may
be rigid or pinned. It is typically used in rigid frames.
• PIN_JOINED_MEMBER: a member with the capacity to carry only axial loads,
i.e. a link. It is typically used in trusses.
• CABLE: A tension member which can carry transverse loads only under large
deflection.
• TENSION_MEMBER: a member without compressional stiffness.
• COMPRESSION_MEMBER: a member without tensional stiffness.
• USERDEFINED: a specially defined member.
• NOTDEFINED: a member without further categorization.
To define internal or external constraints to the structural element, the abstract
super-entity IfcStructuralConnection is used, which represents the schematization
of a structural connection. Depending on their geometry, there are three types of
connections: IfcStructuralPointConnection, IfcStructuralCurveConnection and Ifc-
StructuralSurfaceConnection. The IfcStructuralPointConnection is used especially
for the nodes of the truss. It is an entity that describes a structural node or a punctual
constraint. It is located within a local coordinate system thanks to the ConditionCo-

ordinateSystem attribute; these coordinates are then referred to the global coordinate
system using IfcVertexPoint and the underlying IfcCartesianPoint. IfcTopologyRep-
resentation is used instead to describe the node from a typological point of view.
4 Case Study
The famous roof of the Church of the Eremitani in Padua, defined as ‘ship’s hull’,
is a wooden vault that recalls the keel of a ship (Fig. 8). That particular structure is
attributed to Fra’ Giovanni degli Eremitani and was built in the early 14th century.
Giovanni degli Eremitani was already involved in numerous works, among which
the roof of Palazzo della Ragione in Padua. After great success, he dedicated himself
to the roof of the Church of the Eremitani. The complex articulation of the wooden
ceiling, idealized and realized by the monk, is intended as domain figurative and
architectural elements, able to characterize the internal space, adjusting and varying
the monotone path and, therefore, able to build themselves not as a furniture, but as
essential elements for the construction. The focal moment of Fra’ Giovanni’s inven-
tion is placed at 3/4 of the nave, where the roof changes suddenly its rhythms and
design: the series of transversal vaults is interrupted and linked with the cylindri-
cal segments in a longitudinal sense (Carpeggiani 1975). Unfortunately, in 1944 the
church was bombarded, and part of the original architecture was destroyed. One of
the most affected areas was the roof, which suffered numerous damages. The current
building is a result of a focused work of reconstruction and recondition carried out
in the years following the bombing by the superintendent Ferdinando Forlati. The
roof of the church, in the absence of any sign that denounces its almost complete
reconstruction, is not recognizable as a contemporary work, but rather as a faithful
reproduction of the fourteenth-century hull roof. The dates engraved on the stone
elements redone and the scratching work are the only truly real testimonies, as said
by Forlati, of the reconstruction. However, they give an account of the historical
vicissitudes suffered by the monument in an implicit way, and it is instead clear that
the modern elements, as the superintendent indicated, are ‘made invisible’ where
possible (Bordignon 2012). It is therefore difficult to identify which elements of the
current wooden roof are original and which ones were made during the reconstruc-
tion. The elements of the trusses seem to belong to different eras, even though it is
difficult to identify the actual remakes.
4.1 Structure of the Roof of the Church of the Eremitani
The roof has a supporting structure composed of 20 trusses (Fig. 9), which support the
trilobate-arch vault and the roof above. It covers the long and narrow rectangular nave,
measuring 66.78 × 17.64 m, and which ends in the three apses, one big and with the
Fig. 8 View of the nave of the Church of the Eremitani
altar and two lateral smaller ones. Each truss, whose weight is about 82.000 kg, rests
on the south and north perimeter walls and partly on the original stone barbicans,
or on barbicans in reinforced concrete, if they were rebuilt. The larch vault is an
imposing structure whose total weight is 115.920 kg, while the roof composed of
purlins, rafters, slabs and roof tiles weights 139.944 kg.9 Table 1 shows the areas and
masses involved in the analysis of the loads. The roof has a uniform weight over the
entire area, while, as previously mentioned, the wooden vault changes configuration
at about 3/4 of the nave. The table shows also how the second part of the vault, about
1/4 of the aisle, has a greater unit weight than the first 3/4. It follows that the first
15 trusses (starting from the façade) have a lower operating weight than the last 4
trusses, towards the apses. Truss number 16 instead must be mentioned separately,
because it is located exactly where the vault changes its structure.
The single truss (Fig. 10) has a complex shape and is one of the cases named
‘compounds’ in literature, but with substantial differences. The geometric and static
procedures that adapt the truss to the environment in which it is installed are reused:
in this case, there are not only geometric parameters, but there is also the need to
define the geometric spatial relationship of a decorative element, the wooden vault,
which relates to the environment itself. In this type of truss, the vertical element of the
upper triangular body is suppressed, and the two lateral top chords are divided into
9 Thedata were calculated using the objects of the BIM model, with a specific wood weight of
700 kg/m3 .
Fig. 9 3D model of the Church of the Eremitani with its vaults in evidence
two parts, the lower of the two having a greater inclination than the upper one. This
specific composition is characterised by the disproportion of the height of the vault
with respect to the width of the nave. This setback was solved by the introduction of a
beam which supports the purlins in the lower part, having the same slope as the upper
top chord, thus guaranteeing a slope continuity at the roof pitch (Fig. 11). Without
this expedient, the connection between the top chord and the bottom chord would
have occurred about 1.70 m from the external surface of the wall. This particularity
guarantees also a greater resistance to buckling of the two top chords.
The first four trusses, towards the apse, where the vault changes its appearance,
have two oblique beams. The bottom chord is composed of three elements. The
central one is thinner (22 × 28 cm) and is the part of the bottom chord also visible
from the nave inserted into both sides, towards the side walls, in elements of greater
Table 1 Loads analysis in horizontal projection

Specific weight of wood 700.00 kg/m3
Nave Length 68.78 m
Width 17.64 m
Area 1213.28 m2
“1/4” Nave Length 16.64 m
Width 17.64 m
Area “1/4” 293.53 m2
“3/4” Nave Length 52.14 m
Width 17.64 m
Area “3/4” 919.75 m2
Vault Vault mass 115,920.00 kg
“1/4” vault mass 32,980.00 kg
“3/4” vault mass 82,940.00 kg
Unit weight “1/4” vault 112.36 kg/m
Unit weight “3/4” vault 90.18 kg/m
Roof Roof mass 139,944.00 kg
Unit weight 115.00 kg/m
Truss Medium-distance trusses 3.13 m
Competence area 1 truss 55.21 m2
Unit weight on the first 15 trusses 11,328.46 kg/m
Unit weight on 16° trusses 11,940.78 kg/m
Unit weight on the last 4 trusses 12,553.09 kg/m
section (28 × 34 cm) that incorporate it. Depending on the case, the connection is
guaranteed by bolted or nailed elements, also using metal clamps or bars (Fig. 12).
These elements, which the bottom chord consists of on the sides, have an underly-
ing reinforcement consisting of a sub-bottom chord with a metallic connection that
increases the total section and ensures an increase in the performances compared to
the bending stress induced by the load of the vault (Fig. 13). Another very interesting
aspect of the truss of the Church of the Eremitani is the particular structure of the two
queen posts (Galliani 2001). Each of these elements is composed of the assembly of
two identical but mirrored portions, shaped in the lower part so that they incorporate
the bottom chord and the sub-bottom chord (Fig. 14).
Fig. 10 Axonometric exploded view of the vault
Fig. 11 Detail 1
Fig. 12 Detail 2
Fig. 13 Detail 3
Fig. 14 Detail 5
4.2 The BIM Model: Relationship Between Parameters
It was possible to realize the 3D model of the church with a BIM software10 thanks to
two digital evaluations. The first one was possible thanks to the LIDAR technique,11
the second one thanks to photogrammetry. The evaluation survey, which lasted about
three months, involved also the portion of the church that is located between the
extrados of the wooden ceiling and the intrados of the roof. This enabled to obtain
precise geometric data and provided digital cloud evaluations, imported into the 3D
modelling environment.
The irregular shape of the Church determines a variation of the width of the
nave in plan (Fig. 15), and this translates into a different inclination of the trusses’
top chords in precise points. Table 2 shows the variation of width of the nave to
the tax quota of the trusses.12 The results of the survey indicate that the maximum
difference in width is 35 cm and the average width of the nave is 17.64 m. This
variation in width is determined above all by the out-of-centre lead of the north wall
in its central part caused by the war. Besides concepts deriving from constructive and
manufacturing operations of that period, which limit its precision, each truss differs
from the others, and therefore the cuts of the wooden elements differ too. Historical
structures may have intrinsic architectural defects or undergo plastic deformations
10 Reference model in virtual modeling environment, Autodesk Revit® .

11 Made through Leica® ScanStation C10.
12 The tax quota of the trusses is about 17.80 m from the floor of the nave.
Fig. 15 Variation in width of the nave
over time, both in the planimetric imprint and in their vertical sections. This leads to
a lack of uniformity, typical of the contemporary construction industry. Therefore,
we cannot disregard a different approach from the methodological practice of BIM
and, in particular, from a parametric modelling of the elements.
As already mentioned, the structure consists of 20 trusses, all with a different tax
light. Each truss then consists of 18 beam elements and 11 nodes13 (Fig. 16). These
data enabled to derive the information needed to reconstruct the entire typological
set of connection between the elements.
The flow of data elaborated by the process14 used during the modelling phase
allows a connection structure of the parameters, exemplified by the graph in (Fig. 17).
The process aims at highlighting the hierarchy of the data: the parameters that define
the geometry of the nave influence those parameters that define every single cut of
the elements that the trusses are made of. In the designed structure, these data are
linked in such a way that they can be compiled, in several cases, by other parameters
that come from a hierarchically superior element, with a substantial increase in the
system productivity. In the same way, the exchange of parameters can be inverse or
among the same category of elements. This system allows not only a rapid increase
in the modelling phase, but also a fundamental semantic structuring for their use.
13 Thestudy of the nodes was carried out through current photographs, in situ observations and
through LIDAR digital survey.
14 Made through a VPL (Visual Programming Language) Autodesk Dynamo® .
Table 2 Width of the nave to the tax quota of the trusses
Upstream of the system, we thus find the geometric data that define the nave. The
transition from the truss components (IfcElementAssembly in the IFC structure),
from the beam elements (IfcBeam) and from the cuts (defined by IfcPropertySet),
with their related parameters, enables to eventually define the morphology of each
single connection. Each node is the meeting point between different beam elements
and is composed of the cutting parameters that define its shape. Among these cutting
parameters of elements that have a contact relationship in a specific node, there is a
bi-directional transfer of data (automatic filling of the fields). This system has also
the advantage of making cross-checks between the data, therefore improving the
adhesion of the connection among the elements.
The parameters of the truss element are mainly linear (ridge height), but, if nec-
essary, they can identify angular data (slope angle of the roof). The parameters of
the beam element are linear too, and they define the length, the width and the height
in section. The length of a beam element is defined by the corresponding parameter
of the truss element, which in turn defines the extrusion linear depth using IfcEx-
trudedAreaSolid. IfcRectangleProfileDef.XDim and IfcRectangleProfileDef.YDim
define the rectangular section that must be extruded. The cutting parameters referred
to the beam elements are 60. For this treatise linear parameters were preferred, and
among these, two are radial and belong to the elements B15 and B18. The geometries
of the identified cuts are not too complex and can often be traced back to essential
geometric shapes. It is therefore possible to deduce them through simple trigono-
metric operations.
One of the fundamental aspects is the link between the parameters. Let’s consider
a generic example: the dimensional datum Tn_Bn_Ln of whichever beam element is
Fig. 16 Elements and nodes that the truss consists of with parameters
Fig. 17 Connection structure of data
compiled by the value of the Tn_Bn parameter of the corresponding truss element.
Now, let’s consider a specific case referring to the first truss: if the parameters T1_L13,
T1_L14, T1_B14_C2_L1, T1_B17_C2_L1 are added to the parameter T1_L, and
the parameters T1_B15 and T1_B17 are subtracted to the parameter T1_L, the result
is T1_B16_L, the length of the B16 element, which is also equal to the value T1_L7.
Fig. 18 Scheme of decomposition and cutting of beam elements
The cutting parameters are often determined by the inclination of the beam element
of the truss to which they refer (Fig. 18). All the geometries of the cuts and of their
related parameters are shown in the following images (Figs. 19, 20, 21 and 22).
5 Conclusions
Considering the specificity, this research shows a methodological approach rather

than a universal solution for the improvement and the use of BIM models applied to
the design and manufacture of structural wooden components. The universality of a
solution is questioned when the contexts are framed by specific skills and business
processes, as well as technologies related to systems, machinery or software.
This work can be useful to provide indications of the correct management of
data concerning the cultural heritage. The BIM model is an essential database of
information and improves facility management processes (Eastman et al. 2011). The
methodology described can be a useful tool for investigation and knowledge available
to superintendents and administrators of construction works. Historic buildings such
as churches are often barely monitored, especially in parts difficult to reach, such as
roofs. There are often cases of collapse of these works, due to the deterioration of the
structural elements and to the consequent loss of load-bearing capacity. Therefore, it
is important to emphasize the maintenance criteria, in some cases also replacing the
damaged elements. In the case of a highly historical and culturally valuable building,
Fig. 19 Cutting parameters

of elements 2 and 8
Fig. 20 Cutting parameters of elements 9 and 13

Fig. 21 Cutting parameters of elements 11 and 18

Fig. 22 Cutting parameters of elements 12, 18 and 16

the chosen option is the restoration, while the replacement is only used when strictly
necessary, for example in cases of degradation due to active agents. The proposed
work describes the approaches used to create a structure within the modelling soft-
ware, with the aim of obtaining the geometric parameters of the connection nodes
of a truss. These data can be used to create the single element through advanced
cutting techniques, as described in Sect. 3.2. In this way, it is expressed the need to
introduce concepts of digital fabrication, mass customization and ETO manufactur-
ing processes in the operations actuated in relation to the historical context/cultural
heritage. This can bring these processes closer to the concepts of lean constructions
and lean manufacturing.
The use of the IFC format allowed the sharing of the information identified, main-
taining the structure created. The parameters that characterize the geometry of the
elements’ cuts are defined by the Geometry Use Definitions of the ifcBeam elements
in IFC. These data can be communicated through successive implementations to
CNC machines that cut wood. The study also provides a tool for the exportation of
structural analysis programmes and indications on the possible type of connection
between the elements to increase the reliability of this type of analysis.
The construction industry lags in the process of exploiting the potential offered
by computational design. The current materials and production techniques are not
flexible enough to support highly specialized works. Within this rapid development
where production capacities are constantly evolving, these cutting-edge methods
suggest promising results in a long-term perspective (Naboni and Paoletti 2015).
References
Bordignon G (2012) “Le pietre parleranno” distruzione e ricostruzione postbellica della chiesa degli
Eremitani a Padova tra storia e propaganda. In: Belli G, Centanni M (eds) Opus incertum vol
6–7: Costruzioni e ricostruzioni dell’identità italiana, vol 6–7. Polistampa, pp 156–175
Carpeggiani P (1975) Gli Eremintai. In: Bellinati C, Puppi L (eds) Basiliche e chiese, vol 1. Neri
Pozza Editore, Vicenza, pp 217–234
Eastman C, Teicholz P, Sacks R, Liston K (2011) BIM handbook: a guide to building information
modeling for owners, managers, designers, engineers and contractors, 2nd edn. Wiley, Hoboken,
New Jersey
Evans R (2000) Drawn stone. In: Evans R (ed) The projective cast. The MIT press, Cambridge, pp
179–239
Fai S, Filippi M, Paliaga S (2013) Parametric modelling (BIM) for the documentation of vernacular
construction methods: a BIM model for the commissariat building, Ottawa, Canada. ISPRS—Ann
Photogramm Remote Sens Spat Inf Sci II-5/W1:115–120
Galliani GV (2001) Dizionario degli elementi costruttivi. UTET, Turin, pp 183–188
Hamid M, Tolba O, El Antably A (2018) BIM semantics for digital fabrication: a knowledge-based
approach. Autom Constr 91:62–82
Hauschild M, Karzel R (2011) Digital processes (detail practice). Birkhäuser GmbH, Basel
Jones SA, Bernstein HM (2014) The business value of BIM for construction in major global markets:
how contractors around the world are driving innovation with building information modeling.
SmartMarket Report. McGraw-Hill Construction
Koenig GK, Furiozzi B, Brunetti F (1995) Tecnologia delle costruzioni 2. LE MONNIER, Firenze
Krieg OD (2016) Engineerin perspectives. In: Menges A, Schwinn T, Krieg OD (eds) Advancing
wood architecture. Routledge, A computational approach, pp 199–209
Leroy CFA (1870) Traitè de Stéréotomie comprenant les applications de la géométrie descriptive a
la théorie des ombres, la perspective linéaire, la gnomonique, la coupe des pierres et la charpente.
Gauthiers-Villars, Paris
Maddock R, De Kestelier X, Ridsdill Smith R, Haylock D (2017) Maggie’s at the Robert Parfett
Building, Manchester. In: Menges A, Sheil B, Glynn R, Skavara M (eds) Fabricate rethinking
design and construction. UCL PRESS, London, pp 74–81
Monizza GP, Rauch E, Matt DT (2017) Parametric and generative design techniques for mass-
customization in building industry: a case study for glued-laminated timber. Procedia CIRP
60:392–397
Naboni R, Paoletti I (2015) Advanced customization in architectural design and construction.
Springer, New York
Nelson LH (1996) Early wooden truss connections versus wood shrinkage: from mortise-and-tenon
joints to bolted connections. APT Bull 27(1/2):11–19
Oreni D, Brumana R, Georgopoulos A, Cuca B (2013) HBIM for conservation and management
of built heritage: towards a library of vaults and wooden bean floors. ISPRS—Ann Photogramm
Remote Sens Spat Inf Sci II-5/W1:5–221
Paoletti I (2017) Mass customization with additive manufacturing: new perspectives for multi
performative building components in architecture. Procedia Eng. Elsevier Ltd, Amsterdam, pp
1150–1159
Paris L (2009) Stereotomia del legno. In: Migliari R (ed) Geometria descrittiva vol 2. CittàStudi,
pp 562–588
Pillet J (1887) Traité de Stéréotomie (charpente et coupe des pierres), Ch. Delagrave, Paris
Schwinn T (2016) Landesgartenschau exhibition hall. In: Menges A, Schwinn T, Krieg OD (eds)
Advancing wood architecture. Routledge, A computational approach, pp 111–124
Yuan PF, Chai H (2017) Robotic wood tectonics. In: Menges A, Sheil B, Glynn R, Skavara M (eds)
Fabricate rethinking design and construction. UCL PRESS, London, pp 44–49
https://www.buildingsmart.org
Digital Technologies for Cultural
Heritage: 3D Representation of Complex
Wooden Structures
Caterina Gottardi, Caterina Balletti, Simone Florian and Francesco Guerra
Abstract In line with the growing demand of digital documentation in the field
of Cultural Heritage, nowadays survey technologies allow an immediate reading
of a whole system directly in a 3D environment. The Photogrammetry Laboratory
at the Iuav University of Venice had often dealt with surveying and documenting
complex wooden structures. In this work, we will present the methodology used to
acquire two different types of structures with an in-depth analysis of two Venetian
study cases: the SS. Giovanni e Paolo’s wooden dome and the Magazzini del Sale’s
trusses, starting from a previous work in the Ducal Palace. The purpose of this
research was to analyse the shape and the geometry of these structures: because
of their complexity, a laser scanning survey, with the support of more traditional
methods, seemed to be the best way to analyse them and to obtain the information
directly in a 3D environment. The final aim was not only the creation of 3D models,
useful for studying the spatial complexity, but also the identification of an operational
procedure for such particular constructive techniques. This paper analyses the issues
concerning the survey processes and it illustrates the solutions chosen to overcome
them.
Keywords 3D modelling · Complex architecture · Survey · Laser scanning

Cultural heritage
C. Gottardi (B) · C. Balletti · S. Florian · F. Guerra

Photogrammetric Laboratory, Laboratory System, Iuav University of Venice, 191 Santa Croce,
30135 Venice, Italy
e-mail: cgottardi@iuav.it
C. Balletti
e-mail: balletti@iuav.it
S. Florian
e-mail: simone.florian@live.it
F. Guerra
e-mail: guerra2@iuav.it

512 C. Gottardi et al.
1 Introduction
In the last few years, thanks to the continuous development of new digital tools,
both hardware and software, the metric survey has had a significant push forward
in the field of 3D digitization of Architectural Heritage, also due to the lower prices
of the instruments and faster times of acquisition (Tucci et al. 2016; Balletti et al.
2014b, 2015; Vacca et al. 2012). In fact, digital acquisition of Cultural Heritage
allows to study some aspects that are otherwise difficult to understand in a detailed
way; visualizing an architecture or a complex structure directly in a three-dimensional
environment can allow analysing from different points of view those elements which,
sometimes, are not easily accessible on site. The in-depth study of the shape and
geometry of an object is, moreover, a basis for the analysis of its state of conservation
and for an eventual planning of a restoration intervention.
Up to now, the 3D model seems to be the best choice for representations of con-
voluted structures, especially for what concerns spreading the knowledge even to
people not strictly specialized in the sector. Particularly in museum contexts, appli-
cations of virtual or augmented reality—and the reproduction of objects with 3D
printing techniques—are becoming increasingly widespread for the enhancement
and dissemination of researches’ results. In addition, a model, more than easy to
read and understand, can be useful as a basis for structural analysis or as a virtual
database in which to catalogue all the information collected. The project of BIM
(Building Information Model) for the representation of Architecture and, in partic-
ular, of Cultural Heritage and its history (HBIM—Historical Building Information
Model) will allow to easily catalogue a large amount of information and it will make
the model useful to various users with different purposes.
The model has to be not only accurate, but also easily understandable in each
of its different elements and this could be possible only thanks to a precise survey
method. Moreover, the model has to be thought not only as an end in itself, but to
be measured and analysed in the future, adding new information, if necessary (Fassi
et al. 2011a; b; Fregonese et al. 2009). In fact, we have not to forget the purpose that
necessarily represents the model itself: sometimes, the traditional 2D orthogonal
projections, possibly correlated by technical sheets for the catalogue of elements,
still fully meet the needs of architects and restorers, especially in a workplace, where
not always there is enough time to create detailed 3D models. For this reason, in the
work presented here two types of final representations, a three-dimensional model
and 2D drawings are presented; both products, although different, derive from a
metric survey of which the instrumental precision and accuracy of the acquired data
are known.
Specifically, we will present the methodology used to acquire two different types
of structures, the wooden domes and the trusses, explaining two peculiar Venetian
study cases. We will both describe the acquisition and processing of the SS. Giovanni
e Paolo’s wooden dome and, for the trusses’ analysis, the case of the Magazzini del
Sale. The experience acquired during the previous survey (2002–2003) of the wooden
trusses of Ducal Palace (Sect. 2) allowed a more rapid planning of the procedure to
Digital Technologies for Cultural Heritage: 3D Representation … 513
carry out in further works but, above all, allowed to apply a methodology of metric
data acquisition partly successfully tested. In 2012, we planned the survey of the
wooden dome of SS. Giovanni e Paolo with the aim of analysing the shape and
its spatial complexity. Because of the complexity of the structure, a laser scanning
survey, with the support of more traditional methods, such as direct and topographic
techniques, seemed to be the best way to analyse these elements and to obtain the
information directly in a 3D environment.
In the following paragraphs we will—first of all—briefly describe the history of
this articulated structure very common in the Venetian architectural tradition of the
XIV and XV century (Sects. 3 and 3.1), then we will illustrate the procedure imple-
mented in order to acquire (Sect. 3.2) and process (Sect. 3.3) the metric information.
After the data elaboration it was possible to create a 3D model and some digital rep-
resentations of the entire structure (Sects. 3.4 and 3.5). In the fourth section, we will
then explain the work carried out at the Magazzini del Sale in Venice. After a brief
historical introduction (Sect. 4, 4.1), we will describe the methods and techniques
used for data acquisition (Sect. 4.2) and processing (Sect. 4.3). This most recent
work is slightly different from the previous one, because the lighting conditions have
also allowed the images acquisition for the photogrammetric models: in this way, we
were able to join together both the photogrammetric and laser scanning point clouds
in order to obtain a complete metric data. At the end of the elaboration process, the
17 trusses were appropriately catalogued for restoration analysis in order to plan
potential conservative interventions (Sect. 4.4).
The purpose of this work, as highlighted also in the conclusions (Sect. 5), is not
only the creation of 3D models or 2D representations, useful for studying the spatial
complexity, but also the identification of an operational workflow that can be applied
in surveys of complex wooden structures.
2 The Metric Acquisition in the Ducal Palace
In 2002, the Photogrammetry Laboratory started a survey campaign of the wooden

system of the Sala del Maggior Consiglio of the Ducal Palace in Venice (Balletti
et al. 2003). The structure has the double task of holding up the frescoed ceiling
of the Sala and, at the same time, of supporting the roof covering. The Sala del
Maggior Consiglio is the main room of the Ducal Palace, once location of the highest
magistracy of the Serenissima; the huge ceiling contains 35 paintings on canvas,
some of which are made by Tintoretto, Veronese, Palma il Giovane and Bassano.
The articulated decorative structure, together with the gilded wooden frames, is
supported by a dense network of wooden trusses measuring 24.5 × 52 m and about
7 m high. The large size and low visibility, due to the lack of light, were two of
the main reasons that led to the choice of a laser scanning survey. Moreover, the
elements’ irregularity of the 22 trusses could not be studied in detail with only a
topographic or direct survey: the small planking level, only in the central portion of
the trusses, made some areas inaccessible by the operators.
Because of the described issues, it was decided to acquire the metric data with one
of the first laser scanners used by the Laboratory, the LMS-Z 210 Riegl 3D, which
had an accuracy of ± 2.5 cm. A total of eight scans were performed from different
positions to obtain the entire geometry of the structure. The scans were oriented
using reflecting targets, whose coordinates were calculated thanks to a topographic
acquisition. In those years, laser scanning tools applied to Cultural Heritage and
Architecture were still a novelty. For this reason, instrumental precision was relatively
low for a detailed survey and the data processing phase was very laborious: the errors
of the scans orientation was from 5 mm to 2.4 cm. Subsequently, from the point
cloud, 220 sections were extracted that allowed 2D representations of the trusses,
with the relative structural joints. Finally, also a three-dimensional model was created
to facilitate the understanding of such a complex structure as the ceiling of the Sala
del Maggior Consiglio.
The work briefly described here was useful for dealing with an even more compli-
cated survey about ten years later. In fact, despite the technological and instrumental
development, the methodological procedure is basically the same: in order to obtain a
precise and accurate metric data for an architectural scale, it is necessary to integrate
different survey techniques together.
3 Surveying a Wooden Dome
This work came to life while studying and observing the wooden domes, that are
one of the most enchanting and singular examples of the Venetian architecture. The
dome of SS. Giovanni e Paolo had not yet been studied in its entirety and it was easy
to access to carry out the surveys and the geometrical analysis necessary for this
study (Balletti et al. 2013, 2014a).
Sometimes it can be hard to read through big architectural structures as many
elements are added throughout the centuries during the various restorations. That’s
the reason why a 3D model was needed in order to represent and clearly read a
complex wooden system such as the one of the venetian domes. Only in this way it
was possible to read the entire structure.
Data acquisition through laser scanner is now a very renowned method in the Cul-
tural Heritage field as well as in Architecture in general. The laser scanning technique
allows the analysis of the object in a 3D environment, also from the georeferenced
point cloud. The result is a virtual model of the single dome’s elements, regardless of
whether it is wooden, metallic or wall. To achieve the end result, there were different
phases to go through:
• on-site inspections of the Basilica and internal part of the dome;
• on-site data acquisition that was achieved in two days;
• data elaboration and creation of the 3D model.
To be able to analyse the wall intrados’ curvature of the dome, it was mandatory
to scan the internal section of the Basilica. This survey has been carried out together
with an historical research. The aim was to acquire as much information and drawings
as possible about the architectural development as the documentation (both graphic
and bibliographic) was indeed quite scarce. Thanks to this survey we were able to
obtain an in depth knowledge of the structure and therefore to read the XIX-century
drawings and historical representations in a more critical way than when we first
started.
The aim of the research we carried out was to define a methodology for the creation
of accurate digital representations, going from a high resolution laser scanning survey,
to the analysis of the shape and the spatial geometry of a convoluted wooden structure.
3.1 The Dome of SS. Giovanni e Paolo
The SS. Giovanni e Paolo Basilica (Fig. 1) overlooks the northern part of the lagoon
in Venice’s city centre. The Basilica has always been used as a Pantheon for Venetian
Doges. The funeral monuments inside are a mere key to understand the evolution of
the Republic of Venice. In fact, depending on how the doge chose to be remembered
to posterity through the sculptural art, we can read the power, the charisma, the
cultural influences and the characteristics that he had in life. Even if there is much
information on the evolution of this Basilica, unfortunately there are not as many
materials concerning the evolution and the dating of the dome and the overlapping
wooden structure. We don’t know the exact date but probably the dome was built
around the mid of the XV century (Merotto Ghedini 2000). What we do know is that
the dome follows the architectural tradition found in the five domes of St. Mark’s
Basilica: at first glance one might think that the reason for this complex structure was
to protect the dome from the elements. However, there were lots of less articulated
ways to protect it from the rain: therefore, the only reason must have been that of
elevation. The SS. Giovanni e Paolo dome, together with other domes of the same
period, presents differences to some of the system’s wooden elements when compared
to St. Mark’s Basilica. Nevertheless, the architectural model is pretty much the same:
wooden structures that start from the tambour of the wall domes which rise on the
lagoon landscape, looking for volumetric elevation if compared to the surrounding
buildings.
Fig. 1 The dome of SS. Giovanni e Paolo in Venice

This dome represents the first attempt to change the constructive process. In fact,
for the first time in the Venetian tradition, the wall dome has an oculus in the central
part: with the current wooden structure it is impossible for light to shine inside the
Basilica; certainly a different kind of structure must have been thought and planned.
Actually, we do not know the reason of this attempt, but what we do know is that the
oculus was closed with a wooden plank and the dome was built with a much more
traditional shape. It was only with the construction of San Giorgio’s dome by Palladio
in the second half of the XVI century that tradition underwent a simplification process,
which influencing all wooden structures thereafter (Piana 2009).
As far as the architectural evolution of SS. Giovanni e Paolo’s dome is concerned,
the documents and sources date to the XIX century. Substantial changes were carried
out around the middle of the century resulting in a radical change of the wooden
structure. In addition to some important but routine maintenance works, such as
the reconstruction of the wooden roof and of the external lead structure, a new
complex system of wooden and metal elements was added on. This completely
changed the overall wooden architecture of the dome. The most dramatic changes
occurred between 1826–28 and 1850–53. During the first period, the lower crossed
timbers in the central part of the structure were renovated and a new horizontal
curved structure made of oak wood was added, supported by a new system of 12
larch elements. Moreover, in the fifties, because of the fragile conditions and decay
of the structure, a new substantial restoration intervention was planned: during this
time, another three horizontal curved structures were renovated and a new elements’
system to support them was created. It was also decided to add a new system of metal
tie beams useful to join the horizontal element together. Finally, also the structure
of the lantern was radically modified. In the following years until today, the only
changes were made to preserve the structure and the wooden system was not altered.
3.2 On-site Data Acquisition
During the first on-site inspections we immediately realized the main difficulties that
would be encountered in the following survey campaigns: the reduced visibility due
to the absence of lighting, a very small planking level and a high number of wooden
elements with irregular shapes were some of the main reasons that led us to the
decision of using a laser scanner for the acquisition of metric information.
The dome has a diameter of about 14.5 m and a height from the planking level to
the roof approximately of 10 m: these become 23 m if we also consider the lantern and
the cross on the top. The planking level, where the timber elements of the structure
are based, has a width of 60 cm: as we can therefore imagine, walking inside this
environment was particularly complicated. Furthermore, there wasn’t any artificial
illumination and the natural light was coming only from the upper part where a little
trapdoor allowed people to go outside.
With the laser scanning method it was possible to overcome all the described
issues: contrary to the passive sensors (topography and photogrammetry), the laser
scanner uses an encoded light (active sensor) that is almost indifferent to the envi-
ronment’s light conditions in which we operate (Guidi et al. 2010). In this way it was
possible to acquire the metric information even without light. Furthermore, the laser
scanner allowed to acquire elements even if not directly accessible by the operators.
In fact, inside the wooden structure there is only a single staircase that leads from
the base of the dome to the external top: there are no other easy ways of reaching the
distant beams. Moreover, the reduced dimensions and weight of the laser scanner we
used—definitely handier than the one used in the Ducal Palace—allowed us to take
it up to the dome with no problems.
The structure has an octagonal scheme, but with the changes made in the XIX
century, the observer will perceive a different division, highlighted by the 12 systems
of wooden elements. We decided to put the laser scanner in each of these spans; the
other scans were carried out from the intermediate and top landing of the staircase
that takes people to the top of the dome and from the stairs that go outside, acquiring
also a great part of the lantern. For all the scans we used the phase based laser scanner
Faro Focus 3D S120 (CAM2 Faro Focus 3D): this laser scanner is particularly suited
for fast data acquisition (up to 976,000 points per second) maintaining a high level
of precision (±2 mm in a range from 0.6 to 120 m). Specifically, the scans have been
acquired with angular increments of 0.035°, in order to have a point every 6 mm at
a distance of 10 m. However, the high number of scans and their very close distance
to the analysed object have allowed to obtain a very dense point cloud for a total
of more than 800 millions points; the main problem of such a huge quantity of data
acquisition is the very difficult management in the processing phase.
The scans’ elaboration software recognises easily some signalization systems.
For this reason, high-contrast and spherical targets with a diameter of 12 cm were
positioned on the timber elements (Fig. 2). The acquisition of the scans was carried
out together with a topographic support with the Leica TCR 1103 total station for
a better control in the alignment and registration phases and especially in order to
guarantee the verticality of the Z axis.
Data acquisition was carried out in two days: this was possible thanks to the very
fast instrument’s acquisition times, that allow to optimize the phase of the on-site
data measurements. During the first day we materialized the topographic network
and we performed some of the laser scanning acquisition: we carried out the scans
from the intermediate and the top landing and three scans on the planking level;
Fig. 2 The wooden structure and some targets used for the survey
during the second day we finished the nine scans at the basement. A total of 20 scans
and 860 million points were acquired inside the wooden dome.
Finally, to analyse the curvature of the intrados, it was necessary to acquire also
two scans from inside the Basilica. In this case the targets were positioned in the
apsidal part and in the last two columns of the aisle before the transept. These last
scans were performed with a closer angular increment setting in order to obtain a
higher number of point of the intrados with only two acquisitions. This procedure is
not always used in survey operations because, although the final point cloud is really
dense, the acquisition and processing times are greatly increased.
3.3 Data Processing
After data acquisition, the first step was the alignment of the 20 scans through a semi-
automatic recognition of the targets positioned on the timber elements: in this way,
all the scans are recorded together and georeferenced in the topographic reference
system. In order to perform this operation, a six-parameters geometrical transforma-
tion is applied, that is a spatial rototranslation. At the end of the process, the average
precision obtained in the scans alignment was about 4 mm.
The Faro Focus 3D laser scanner has also an integrated sensor with a 70-megapixel
resolution in order to “colour” the scans with the RGB values of the images acquired.
In this case, the point cloud of the wooden system was exported only with the intensity
value (Fig. 3): the RGB information had not been used due to the weak lighting
conditions illustrated before. Furthermore, radiometric values were not essential in
order to study the geometry of the structure and the wooden material’s response to the
laser scanner did not cause any particular problems in terms of the intensity value. The
same processing work was done with the two scans inside the Basilica. However,
in this case the light conditions made it possible to realise a better photographic
acquisition than the one done inside the dome: for this reason, it was possible to
acquire the scans also with the radiometric values.
At the end of the process we obtained a very dense and detailed point cloud: in the
wooden dome, a total of 863,665,301 points were scanned, processed and registered.
The point cloud was then edited: all the unnecessary data recorded by the laser
Fig. 3 The recognition of the targets using Faro Scene

Fig. 4 The point cloud of the wooden structure and the intrados of the dome
scanner was deleted, like a few parts of the roof that could be seen through the holes
at the base of the dome. The final elaborated cloud had a total of 861,912,073 points.
For the alignment and the registration of the scans we used the Faro Scene software
(CAM2 Scene), whereas for a better management and for the point cloud editing we
used Pointools (Bentley Pointools): with this software it is possible to visualise in
a rapid way a huge number of points, to take some distance measurements and to
extract points’ coordinates. Moreover, it is also possible to create 3D visualizations of
the point cloud (Fig. 4): in fact, also some flythrough animations have been prepared
to show the survey’s first results.
The thickness of the wall dome was taken directly from the nineteenth-century
historical sections found in the archives; in fact, to connect the scans of the intrados
with those of the extrados we should have performed a topographic survey along
the steep and narrow path of the spiral staircase that, from the apse of the Basilica,
leads to the wooden dome. Since the knowledge of the masonry section was not
the main purpose of this study, it was therefore decided to simplify the topographic
acquisition.
The final elaboration allowed a representation for at least a 1:50 scale, as previ-
ously set during the survey’s project phase. However, the traditional drawings would
have allowed only in part to understand the wooden system in its spatial complexity
and to analyse the process carried out during the technical construction phase: for
this reason, we decided to choose a 3D digital representation. Moreover, once the
accurate 3D model is created, it is possible to extract the useful 2D information and
views directly and automatically from the 3D environment.
Once the data processing phase was completed, the point cloud was ready to be
used as a basis for the three-dimensional elaborations of wooden elements.
3.4 3D Digital Representations
From a communicative point of view, the virtual reconstruction of a complex his-

torical architecture is an important tool (Guidi and Russo 2011): it is easier than
traditional 2D drawings to analyse geometrical information and it is fundamental in
the spreading of knowledge of the constructive technique to a non-expert audience.
Fig. 5 Generation of solids from laser data sections
After the point cloud’s processing, we moved on to the 3D modelling: the entire
model was created using AutoCAD (Autodesk AutoCAD). The workspace choice
doesn’t allow any automation during the modelling process and it guarantees the
coherence of the model with the surveyed data. In fact, thanks to a Pointools for
AutoCAD plug-in, it was possible to visualize and work directly on the point cloud to
extract sections and profiles (Fig. 5). Each vertical and oblique element was modelled
with solid or loft surfaces from the extraction of profiles derived from horizontal
sections of the cloud at different heights: this modelling method allows keeping a
good adherence to the data surveyed. The horizontal elements, geometrically simpler,
were then created with the extrusion of solids from vertical profiles. For the curved
objects, such as the wall dome, sections were extracted radially with respect to
a common centre. With the more complex elements we made a high number of
sections and we specified the path or the guides for a better control on the surface
or solid shape, as for the curved roof elements. Thanks to the Boolean operations,
all the wooden system’s joints were also made. These conditions, together with the
possibility of using a high number of profiles directly from the point cloud, guarantee
the geometrical coherence with the real data.
All data for the graphical part came from processing the 20 scans of the wooden
dome and the two scans from inside the Basilica. Moreover, Aerial Lidar data was
used only for the modelling of the external cross. This was acquired for a bigger
research of Venice and of SS. Giovanni e Paolo insula; in this way we were able
to read some of the measures between the summit points, despite being sparse, and
getting the size of the cross on the lantern.
The modelling process of the wood, metal and masonry elements was carried out
without applying a general simplification of geometry. Although the regularisation
of the wooden structure would have allowed faster times of processing, we decided
not to do any rationalization of the shapes, in order to analyse the real geometry and
to allow future applications of the model in the restoration and structural analysis
fields, if required. For this reason, every single element was modelled according to
the geometry and the real position.
The final result was obtained after several months of data elaboration and mod-
elling in AutoCAD. The 3D model visualization and the video-animations were made
by 3D Studio Max (Autodesk 3D Studio Max); in particular, the latter are crucial to
understand the constructive process of this Venetian architectural technique (Fig. 6).
Fig. 6 Images of the 3D model obtained
In fact, thanks to the 3D digital representations and to the subsequent animations

that can be achieved, it is possible to clearly grasp the spatial distribution of all the
elements that compose the dome of SS. Giovanni e Paolo. In particular, three video-
animations were created to show the complexity of this great structure: one of the
entire system and two of different details. In the first case every single element was
virtually disassembled, from the external lead structure to the wall dome, with the
intention of showing the hierarchical structure and to make people to easily under-
stand the constructive technique. The second video animation focused on one of the
12 main wooden elements added in the 1850s that support the entire dome: in this
case, with a much more detailed view, the aim was to underline that every element
is different from the other and also that every joint of the wooden system was cre-
ated. Finally, only thanks to the methodology applied in this research, we created an
animation of the staircase that takes people up to the top, to show that every single
element, functional or structural, was modelled according to real geometry.
3.5 The Purpose of the 3D Model
The laser scanning survey and digital modelling used in this case study allowed us
to understand the geometry and the spatial complexity of one of the most singular
techniques of the Venetian architectural tradition. Moreover, this method allowed the
comparison of the actual survey with existing historical representations.
There are only a few examples of 3D models of wooden domes that can be used
as a basis for comparison because of the complexity and constructive distinctiveness
of this structure, except for the analysis conducted on two of the five domes of St.
Mark’s Basilica, the Pentecoste and Profeti’s (Fregonese and Taffurelli 2009). In fact,
more often than not, surveys and structural analyses are carried out on the intrados
of the wall domes and vaults (Tucci et al. 2011). Moreover, even though it is known
that the general organization of the wooden beams of the Venetian domes derives
from St. Mark’s domes, each one is unique, with a different element composition
and its collocation in space.
Thanks to the 3D model it is possible to obtain an immediate reading of the shape
and dimensions of the structure and thanks to the virtual navigation it is possible to
Fig. 7 Digital representations. On the right: a comparison between the historical section and the
actual one obtained with the laser scanning survey
grasp the whole architectural system, that would otherwise be impossible to recognize
so clearly in the real environment. In addition, having created these animations it
is possible to spread the knowledge of this technique amongst those who are not
specialized in this particular sector. Analysing the shape of this structure is essential
not only to understand the complicated wooden system, but also in view of other
analysis and considerations, such as the state of condition and decay or its historical
architectural evolution.
The model obtained with this method gave us the possibility to compare a section
of the actual state of the dome with a drawing of the 1850s (Fig. 7). It has been
discovered that the XIX-century drawing had some simplification of the elements:
the position of some wooden beams does not reflect reality and also the geometry of
some elements (such as the curvature of the roof and of the wall dome), was drawn
with a regular shape, whereas the 3D model tells us that they cannot be classified as a
known geometry. Naturally, this type of geometric simplification was necessarily due
to the tools available in 1850s; thanks to the continuous technological development,
today it is possible to study and analyse with a deeper approach and increasing
accuracy the Architectural Heritage.
4 Surveying Wooden Trusses
The second part of the work concerns the metric study of the sixth building of
the Magazzini del Sale in Venice, the Magazzino 184 (Fig. 8). The building is a
rectangular space 40 × 9 m long and 10 m high, with only a single entrance along the
Fondamenta dei Saloni. Inside, the space is described by buttresses on walls defining
regular spans. This structure, once used as a salt warehouse from the Republic of
Venice, has given the opportunity to analyse another typical construction technique:
the geometric survey of the building and, in particular, of the wooden trusses was
performed with the aim of giving back to the city an area of such historical and
cultural interest as well as a unique architectural charm. The restoration interventions
Fig. 8 The Magazzini del Sale in Venice
already done in the first three Magazzini under the supervision of the Soprintendenza
of Venice testify the effective willingness of citizens to actively take advantages of
these spaces intended to be a pole of art and culture.
To provide a starting point on which to develop a conservative intervention, an
accurate and precise metric survey was therefore necessary. Thanks to the integra-
tion of various acquisition techniques, such as laser scanning and photogrammetric
survey, it was possible to produce drawings and orthophotos useful for the definition
of technical sheets about the materials’ state of condition. The metric analysis of this
architecture was therefore necessary in order to know how to conserve and preserve
the structure. Once again, the survey is an integral part of the knowing process, the
first essential phase of a restoration project.
Particular attention was given to the wooden trusses system of the Magazzino
184: moreover, the 2D drawings and the sheets produced can allow a possible data
implementation at different times. Furthermore, the integrated survey techniques
allowed the creation of a 3D model where the geometrical complexity of a specific
truss did not permit a clear reading of the system only through traditional drawings.
4.1 The History of the Magazzini del Sale
About the history of saltworks in Venice there is a rich documentation dating back to
the 1000 A.D. (Vitale d’Alberton 2006; Hocquet 1991; Hocquet 1990). Already in the
XII century—the period of maximum salt production in the lagoon—many salt pans
and wooden warehouse were concentrated in Dorsoduro, the same neighbourhood
where the monumental Magazzini dell’Emporio dei Sali or di San Gregorio would
have been built (Fazzini 2006). Towards the end of the century, Venice adopted a real
control policy on the trade of this property: the history of salt is therefore connected to
the history of that architecture built to contain and store in total security the so-called
“white gold” of the Venetians.
Between the end of the fourteenth century and the beginning of the fifteenth cen-
tury, the nine Magazzini were built: today, they represent a monumental complex
architecture under protection in the Dogana area (Berlucchi and Mazzetto 2006;
Dell’Aquila 2006). The Magazzini di San Gregorio were initially all of the same
length and perfectly aligned one to each other: furthermore, they were built perpen-
dicular to the Giudecca canal to facilitate the unloading and storage of salt.
During the centuries many changes were done to these structures, in particular to
the external wall, the length of the buildings and the roofs. The Magazzini had their
slow and gradual decline towards the end of the eighteenth century with the French
domination, who suppressed the Magistrato al Sal and, leaving the city, emptied the
structure of their precious content. The current appearance dates back to the second
Austrian domination when, between 1820 and 1830, the Magazzini underwent rad-
ical restorations: the reconstruction of the façade, the creation of internal buttresses
supporting the walls and the roofs renovations were only some of the interventions
performed in those years. In 1974 the Magazzini were used by the Biennale di Venezia
for temporary exhibitions: it was the first time that the structure had been opened to
the public and it was immediately evident the vocation of these spaces for a cultural
use. This opens the phase that will lead the area of the Punta della Dogana to become
a central point for art and cultural activities; in 1978 the Magazzini 184 and 185 were
entrusted to the Reale Società Canottieri Bucintoro and, still today, they are used as
a depot of ancient and modern boats.
4.2 Integrated Techniques for Data Acquisition
The survey of the Magazzini del Sale has been divided into several phases using
different techniques: at first, a topographic network was carried out to allow the geo-
referencing of photogrammetric models and laser scanning data. The materialization
of a semi-permanent system with cornerstones on the ground, as in this case, has the
advantage of being able to implement metric acquisition in different temporal phases
and to place instruments in the same position.
The topographic survey was carried out with the Leica TCR 1103 total station in
two separate phases: during the first acquisition (July 2017), three cornerstones were
materialized: two on the Fondamenta and one inside the Magazzino 184. During the
second campaign (October 2017), four more points were added to allow a topographic
support in the lateral Magazzini (183 and 185). In this way, it was also possible to
deduce the wall thickness. At the same time, a detailed network was also created
to obtain the coordinates of some points necessary for the georeferencing of the
laser scanning data and for the absolute orientation of photogrammetric models; A4
format targets were used with a black and white chessboard, easily recognizable by
the data processing software. The observations were adjusted with the least squares
method implemented in the StarNet software (MicroSurvey StarNet). At the end of
the procedure, a standard deviation on the coordinates of about ±2 mm were obtained.
Also in this case, a laser scanning survey with the phase based laser scanner CAM2
Faro Focus 3D S120 was performed. As previously said, there are many advantages
of using this instrument, as for example the practicality, or the battery life, to the
possibility of working even in the absence of lighting conditions by recording the
material’s response to the laser due to the intensity of the return signal. As already
Fig. 9 The instruments used for data acquisition
known, this laser scanner is particularly suited for fast data acquisition maintaining
a high level of precision. The scans have been acquired in order to have a point every
6 mm at a distance of 10 m: in this way it is possible to guarantee a final data accuracy
suitable for an architectural and detailed scale.
For this work, we performed a scan for each span defined by the buttresses on
walls. During the first survey, eight scans were carried out, two outside and six inside
the Magazzino 184 while, in the second campaign, 15 scans were acquired in the
two lateral Magazzini. Inside the building, the lighting conditions varied a lot due to
the unique window above the entrance door, causing an extremely overexposed area
and a darker one towards the end of the Magazzino. For this reason, in the redrawing
phase the scans near the entrance were used only with the intensity values, while all
the other ones were used with the radiometric values acquired by the laser scanner
photo sensor.
Due to the geometry of the building and the last complex spatial truss, the points
obtained with the laser scanner were very sparse in the final part of the Magazzino:
in fact, the high number of beams caused many areas with no data acquired. In
order to overcome this issue, it was finally decided to also use the technique of
digital photogrammetry to implement the metric data (Barazzetti and Scaioni 2008;
Bezoari et al. 1999; Fangi 1997).
The photogrammetric survey was carried out using two digital cameras: the Nikon
D800 with a 20 mm focal length for the walls and the Sony RX100 for the external
façade and the final spatial truss (Fig. 9). For the acquisition of the façade, the
compact Sony RX100 camera was used together with a carbon-fibre telescopic pole
that allows to reach heights up to 12 m. In fact, the reduced size of the Fondamenta
did not allow to keep the needed distance from the object in order to guarantee the
acquisition of the entire height in a single photogram. In this way, instead, it was
possible to acquire two strips with an optical axis perpendicular to the façade at
different heights and with a good overlap between them. To increase the accuracy of
the photogrammetric model, it was also decided to acquire some photograms with a
45° optical axis at regular intervals.
Both the cameras were used inside the Magazzino. For the survey of the walls
the photograms were acquired with the Nikon D800: a parallel acquisition scheme
with respect to the object was planned with a particular attention to images overlap.
Because of the different areas of exposure, a system of spotlights on a bogie was also
used in order to be positioned near the wall surface from time to time.
For the acquisition of the final spatial truss, necessary to complete the metric
description of the wooden structure of the Magazzino, the Sony RX100 was used
together with the telescopic pole. In this case, given the complexity of the structure,
we did not apply a typical photogrammetric acquisition geometry, but the photograms
were taken by placing the pole in every interstitial zone of the beams to obtain a high
number of images. At the end of the process, a total of 118 photos of the external
façade, 88 of the internal walls and 123 of the wooden system were acquired.
4.3 Data Elaboration
The acquired data was then processed with different software. The laser scanning
information was oriented and georeferenced in the topographic system using the
Faro Scene software (Fig. 10): at the end of the procedure, the precision obtained in
the scans alignment was about 4 mm.
The photogrammetric process was more difficult: all the images were elaborated
in the Photoscan Professional software (Agisoft Photoscan), which allows to obtain
a three-dimensional model starting from photographs. The software is based on the
recognition of the same point in two or more images: in this way, it is possible to
calculate the camera position at the acquisition time and to reconstruct the geometry
of the object. The photogrammetric restitution, that corresponds to the creation of
the dense cloud in Photoscan, allows to obtain an almost continuous description of
the object. Subsequently, it is also possible to generate a mesh model, applying the
Delaunay triangulation algorithm (Fig. 11).
Fig. 10 The recognition of the targets using Faro Scene and the point cloud in Recap
Fig. 11 The photogrammetric model. From the left: the point cloud, the mesh and the texturized
model
We finally generated the photogrammetric models of the external façade, of the

internal walls and, of course, of the spatial truss. In particular, in this last case, photos
of the wooden structure (123) and some of the walls (12) were elaborated together
in order to identify some targets for the absolute orientation of the model: in this
way it was possible to georeference also the truss in the same reference system of
the Magazzini del Sale’s survey.
We applied a free-network approach in the bundle adjustment (Nex and
Remondino 2014): only at the end of the process, after the creation of the model,
a Helmert transformation—a rototranslation with a scale variation—was applied to
georeference the model in the topographical reference system. The RMSE (Root
Mean Square Error) on the four identified GCPs was obtained almost equal to 1 mm.
The photogrammetric and laser scanning point clouds have been visualized
together in the Recap software (Autodesk Recap) and subsequently imported in
AutoCAD for the further elaborations. The differences between the two clouds were
lower than the tolerance required for the 1:50 scale of representation: in this way, it
was possible to obtain a complete metric data accurate for the architectural repre-
sentation chosen.
4.4 Representations and Final Considerations
The two-dimensional representations have been elaborated in AutoCAD that allows to

directly import the point cloud for an accurate redrawing of the surveyed object. This
procedure, similar to the one applied in the case of the dome, allows to represent
the analysed elements without simplification. In this way, by defining horizontal
and vertical section planes, it was possible to produce drawings for the 1:50 scale,
including the plan of the trusses and two longitudinal and transversal sections in
order to study the entire wooden system (Fig. 12).
Subsequently, monographic technical sheets were performed and implemented
for each single wooden truss of the roof (Fig. 13). This one is composed by 17
trusses parallel to the façade and placed one to each other with a distance of about
2 m: it was the result of an intervention that involved the whole Emporio dei Sali
in the twenties of the XIX century. With the new trusses arrangement, a connection
Fig. 12 Generation of 2D drawings and 3D model from laser data sections

Fig. 13 An example of a technical data sheet
with the lateral Magazzini was created due to some shelves that, crossing the wall,
offered a common plan for the trusses support.
During the survey it was possible to verify this connection and some hypothe-
ses were formulated on the current deterioration phenomena. The in-depth study
had the aim of documenting the state of condition of the trusses: this was possible
through an analysis of the single elements that compose them and the joints that
guarantee the correct reticular function. For each beam, three sections were done
perpendicular to the element in order to represent the real geometry, any warping
or section variations. This was also possible thanks to the integration of the laser
scanning and photogrammetric point cloud, which ensured an accurate elaboration
of the individual component.
The sheets also present a qualitative description of the elements according to the
wood defects (number and size of the joints, ring shake, shrinkage crack) deduced
from the images acquired with the help of the telescopic pole: it was then possible to
define the timber quality according to some parameters established in the legislation
for the visual classification of wood.
The photograms have also brought out traces of the axe work, particularly visible
on the horizontal elements; moreover, they gave the possibility to identify the pres-
ence of xylophage organisms, especially in correspondence of substituted elements.
All this information is essential for defining and containing possible interventions,
such as the reinforcement of joints to improve the building behaviour to horizontal
actions or localized biocide treatments.
Only for the case of the final spatial truss, to allow a clear reading also to users not
necessarily specialized in architectural construction techniques, we decided to model
this part of the structure in its three dimensions (Fig. 14). The applied procedure is
very similar to the one used to create the 3D model of the dome: thanks to the
possibility to import the point cloud directly in the CAD environment, each element
was modelled with solid or loft surfaces from the extraction of profiles: as already
said, this modelling method allows keeping a good adherence to the data surveyed.
This work on the trusses is an example of how, not always, a three-dimensional
model must necessarily be created: each representation should be elaborated accord-
ing to the final purpose. Sometimes, however, there is a tendency to represent every-
thing in three dimensions, greatly increasing data processing times: in this case,
only the spatial truss needed a three-dimensional representation, given the construc-
Fig. 14 Digital elaborations obtained. On the left: a 2D drawing; on the right: the 3D model of the
spatial truss
tive complexity that distinguished it from the other ones. The traditional 2D draw-
ings and the monographic technical sheets provided an accurate description of the
wooden structure of the Magazzini del Sale in Venice; furthermore, the integration
of laser scanning and photogrammetric point cloud allows, already as itself, a three-
dimensional representation, though discrete, that can be used to understand surveyed
elements.
5 Conclusions
The methodology described in this research is intended as an application procedure

for an in-depth study of convoluted wooden structures. The structural and conserva-
tive analyses of an architecture cannot be able to leave the shape knowledge of the
object out of consideration: for this reason, the geometrical survey must be at the
beginning of the cognitive path and it also must guarantee the necessary accuracy for
further studies. As an essential step for the knowledge process, the survey can allow
the preservation of Cultural Heritage and its historical and material enhancement.
Up to now, the continuous development of both hardware and software tools
allows to acquire data and to produce elaborations useful for the complete knowledge
of an object in a relatively short time. However, we always need to pay particular
attention to the interpretation of representations, critically analysing the acquired
information, above all for what concerns the accuracy and precision obtainable with
the used survey techniques.
As we have already described, survey’s results can be different, from the three-
dimensional model to the generation of orthophotos and to the more traditional 2D
orthogonal projections. Sometimes the 3D model, as for the dome of SS. Giovanni
e Paolo, becomes a necessary and unavoidable product for the knowledge of the
structure: in this case, the two-dimensional drawings would not have sufficiently
described the complexity of the dome. Moreover, the model was created not only
as an instrument for the geometrical and spatial knowledge of the dome, but also in
order to be useful for further researches: the survey is firstly an analysis of the shapes
and then it can be used as a support for other investigations in the architectural field
(Bertolini Cestari et al. 2013).
On the other hand, as for the most of the Magazzini del Sale’s trusses, traditional
orthogonal projections can be a suitable product for the representation of the wooden
system and to support further analyses. In this work, the 3D model was needed only
in the case of the final spatial truss to describe the complexity of the timber elements,
focusing the modelling operations only on a specific part of the structure otherwise
difficult to study in its entirety.
What is common in the case studies described is the starting point of the procedure,
that is a metrically accurate survey for the architectonic scale: through the processing
of the acquired information, we obtain a three-dimensional point cloud able to provide
a detailed description of the elements and to represent them with their deformations
and irregularities. The survey allows an easy comprehension of the wooden elements
of these complex constructive systems and it is the basis for the creation of a 3D or
2D database where it is possible to georeference all the collected information. Every
single element can be catalogued, for example, according to its structural function,
type of wood, conservation status and renovation interventions carried out in the past.
Combining and comparing data from different sources requires greater integration
of highly varied professional skills: up to now, a multidisciplinary research, which
connects different expertise, seems to be the appropriate procedure for enhancing the
Cultural Heritage and allowing possible restoration interventions for the preservation
of Architectural Heritage of our country. Furthermore, Venice has always had an
important role for commercial and cultural exchanges: the enhancement of its history
of architecture, relatively forgotten or, at first sight, hidden, must necessarily be part
of a project in order to restore the structures and to spread the knowledge of this
urban Heritage that belongs to everyone.
At the end of this work, we would like to present this methodology as an opera-
tional workflow for the survey of Venetian historical wooden roofs and, in particular,
for wooden domes.
Acknowledgements The authors would like to thank Francesco Gerbaudi and Matteo Berto for
their co-operation in the researches, respectively, of the wooden trusses of Ducal Palace and of SS.
Giovanni e Paolo’s dome.
References
Balletti C, Guerra F, Gottardi C, Scocca V (2015) 3D integrated methodologies for the documen-
tation and the virtual reconstruction of an archaeological site. ISPRS Arch 40:215–222. https://
doi.org/10.5194/isprsarchives-xl-5-w4-215-2015
Balletti C, Berto M, Gottardi C, Guerra F (2014a) 3D technologies for the digital documentation
of an ancient wooden structure. IJHDE 3(1):19–32
Balletti C, Brussa N, Gottardi C, Guerra F (2014b) The documentation and reintegration of a lost
past. ISPRS Ann II(5):49–55
Balletti C, Berto M, Gottardi C, Guerra F (2013) Ancient structures and new technologies: survey
and digital representation of the wooden dome of SS. Giovanni e Paolo in Venice. ISPRS Annals
II-5/W1:25–30
Balletti C, Guerra F, Gerbaudi F (2003) The survey of the wooden structure of the roof of Palazzo
Ducale in Venice. ISPRS Archives XXXIV-5/W12:49–53
Barazzetti L, Scaioni M (2008) SIFT e Geometria multi-immagine per la modellazione tridimen-
sionale automatica da sequenze di immagini. In: Proceedings of 12th ASITA National Conference,
L’Aquila, pp 269–274, 21–24 oct 2008
Berlucchi N, Mazzetto M (2006) L’intervento di restauro delle coperture lignee. In: Cattani M,
Berlucchi N (eds) I magazzini del sale a Venezia. Indagini storiche e diagnostiche per un intervento
di restauro conservativo, Marsilio, Venezia, pp 124–143
Bertolini Cestari C, Chiabrando F, Invernizzi S, Marzi T, Spanò A (2013) Terrestrial laser scanning
and settled techniques: a support to detect pathologies and safety conditions of timber structures.
Adv Mater Res 778:350–357
Bezoari G, Monti C, Selvini A (1999) La fotogrammetria per l’architettura. Officine Grafiche
Liguori, Napoli
Dell’Aquila G (2006) I magazzini del sale detti “saloni”. In: Cattani M, Berlucchi N (eds) I magazzini
del sale a Venezia. Indagini storiche e diagnostiche per un intervento di restauro conservativo,
Marsilio, Venezia, pp 41–54
Fangi G (1997) Note di fotogrammetria. CLUA, Ancona
Fassi F, Achille C, Fregonese L (2011a) Surveying and modelling the main spire of Milan cathedral
using multiple data sources. Photogram Rec 26(136):462–487
Fassi F, Achille C, Gaudio F, Fregonese L (2011b) Integrated Strategies for the modeling of very
large, complex architecture. ISPRS Arch XXXVIII-5/W16:105–112
Fazzini G (2006) Il sale di Venezia. In: ArcheoVenezia, anno XVI, n. 1–2, pp 1–4
Fregonese L, Scaioni M, Taffurelli L (2009) Generation of a spatial information system for archi-
tecture with laser scanning data. ISPRS Arch XXXVIII-3/W8:87–92
Fregonese L, Taffurelli L (2009) 3D model for the documentation of cultural heritage: the wooden
domes of St. Mark’s Basilica in Venice. ISPRS Arch XXXVIII-5/W1:1–6
Guidi G, Russo M (2011) Diachronic 3D reconstruction for lost Cultural Heritage. ISPRS Arch
XXXVIII-5/W16:371–376
Guidi G, Russo M, Beraldin JA (2010) Acquisizione 3D e modellazione poligonale. McGraw-Hill
Companies, Milano
Hocquet JC (1991) Chioggia capitale del sale nel Medioevo. Il Leggio Libreria Editrice, Sottomarina
(VE)
Hocquet JC (1990) Il sale e la fortuna di Venezia. Jouvence Società Editoriale, Roma
Merotto Ghedini M (2000) Santi Giovanni e Paolo. In: Veneto Istituto, Valcanover F, Wolters W (eds)
L’architettura gotica veneziana. Istituto Veneto di Scienze, Lettere ed Arti, Venice, pp 115–122
Nex F, Remondino F (2014) UAV for 3D mapping applications: a review. Appl Geomat 6:1. https://
doi.org/10.1007/s12518-013-0120-x
Piana M (2009) San Giorgio e le cupole lignee lagunari. Annali di Architettura 21:79–90
Tucci G, Bonora V, Fiorini L, Conti A (2016) The Florence baptistery: 3-D survey as a knowledge
tool for historical and structural investigations. ISPRS Arch 41:977–984. https://doi.org/10.5194/
isprs-archives-xli-b5-977-2016
Tucci G, Nobile A, Riemma M (2011) The Basilica della Madonna dell’Umiltà in Pistoia: survey,
analysis and documentation. In: Proceedings of XXIII International CIPA Symposium, Prague,
pp 1–8
Vacca G, Deidda M, Dessi A, Marras M (2012) Laser scanner survey to cultural heritage conservation
and restoration. ISPRS Arch XXXIX-B5:589–594
Vitale d’Alberton R (2006) Dalle saline alle salere. Una storia di sale e non solo. In: I magazzini
del sale a Venezia. Indagini storiche e diagnostiche per un intervento di restauro conservativo,
Cattani M and Berlucchi N eds., Marsilio, Venezia, pp 17–40
Websites
Agisoft Photoscan: http://www.agisoft.com/

Autodesk 3D Studio Max: https://www.autodesk.it/products/3ds-max/overview
Autodesk AutoCAD: https://www.autodesk.it/products/autocad/overview
Autodesk Recap: https://www.autodesk.com/products/recap/overview
Bentley Pointools: https://www.bentley.com/en/products/product-line/reality-modeling-software/
bentley-pointools
CAM2 Faro Focus 3D: https://www.faro.com/it-it/prodotti/costruzione-bim-cim/cam2-focus/
CAM2 Scene: https://www.faro.com/it-it/prodotti/design-del-prodotto/faro-scene/
MicroSurvey StarNet: http://www.microsurvey.com/products/starnet/
AHBIM for Wooden Built Heritage
Conservation
Stefano Brusaporci, Alessandro Luigini, Starlight Vattano, Pamela Maiezza
and Alessandra Tata
Abstract The chapter investigates the potential offered by the Building Information
Modeling for the conservation of the wooden architectural heritage. Based on the
architectural survey, the BIM modeling of a South Tyrolean farm was carried out,
with particular attention to the aspects related to its management and maintenance.
Inside the database connected to the model, information about the building has been
included, which can also be used on site through the use of mobile devices. The
sharing of such information within a single platform avoids risks connected to the
failure or incorrect transmission of data, allowing the optimization of the manage-
ment, maintenance and conservation of the architectural asset.
Keywords BIM · Built heritage · Parametric design · Conservation

Facility management
S. Brusaporci (B) · P. Maiezza · A. Tata

DICEAA—Department of Civil, Construction-Architectural and Environmental Engineering,
University of L’Aquila, L’Aquila, Italy
e-mail: stefano.brusaporci@univaq.it
P. Maiezza
e-mail: pamela.maiezza@graduate.univaq.it
A. Tata
e-mail: alessandra.tata@virgilio.it
A. Luigini · S. Vattano
Free University of Bolzano, Bolzano, Italy
e-mail: alessandro.luigini@unibz.it
S. Vattano
e-mail: starlight.vattano@unibz.it

534 S. Brusaporci et al.
1 Introduction
Building Information Modeling (BIM) is an efficient methodology aiming to opti-

mize the building process in all its phases. The BIM model, considered in its dual role
as a three-dimensional representation of the building and an information database,
is configured as a system for the digital management of the entire building process,
with important repercussions in economic and timing terms. The option to insert,
manage and display complex and heterogeneous data within a single environment
avoids problems related to information redundancy or errors in its transmission and
sharing. The advantages of using BIM platforms are therefore not limited to the
design phase but include the management and maintenance of the building. In fact,
the greater reliability of information, favored by the reduced risk of data loss or errors
in the transition from the construction phase to the management one, allows a more
efficient maintenance process.
In addition, the possibility of accessing the information content of the model
remotely, through virtual or augmented-reality applications, means that the collabo-
rative approach of BIM can be extended to the construction site. So all the profes-
sional figures involved in the building process, including the technicians working
on site, interact through the mediation of a shared database—consisting of the BIM
model—in which it is possible to view and add data interactively and in real time.
These advantages of the use of the BIM methodology assume a specific value in
the case of application to the architectural heritage. The specific features of architec-
tural assets, such as the uniqueness of a building and the large amount of information
to be managed for a complete knowledge of it, mean that the BIM methodological
approach, designed for new buildings, requires specific reflections for use on his-
torical buildings (Brusaporci et al. 2018a; Bianchini et al. 2016). The modeling of
complex historical geometries in a context characterized by standardization, and the
articulation of the information system for the description of historical buildings are
the main themes of research on Historical Building Information Modeling (HBIM).
A further requirement is the appropriate structuring of the information through spe-
cific parameters, in anticipation of the management and conservation of an historic
building through BIM.
This chapter aims to test the opportunities created by the BIM process for Facility
Management of historical buildings, and more specifically wooden buildings.
The contribution proposes a case study of the Aschner hof farm, located in Renon
(Bolzano). The BIM procedure applied to wooden constructions was trialed through
BIM modeling of a historical South Tyrolean building, representative of a more
widespread construction tradition in the Dolomites, testing the limits and potentiality
of this process in the case of architectural heritage conservation.
AHBIM for Wooden Built Heritage Conservation 535
2 AH-BIM for the Architectural Heritage
The BIM process allows integration of all design components (architectural, struc-
tural, plant, economic, etc.) in a single “environment” consisting of a model, favor-
ing interoperability among the various parties involved in the building process
(Brusaporci and Maiezza 2016).
The possibility of managing all information concerning the entire lifecycle of the
building (design, construction, maintenance, management, decommissioning) within
a single platform makes BIM of great interest both for the new buildings for which it
was developed and for existing ones (HBIM) for which it is still under consideration
(Bianchini et al. 2017).
Documentation and conservation activities produce a large quantity of data, rang-
ing from technical-constructional characteristics to information of a historical or
formal nature. This heterogeneous body of information (images, photos, drawings,
archival documents, etc.) needs to be filed and managed appropriately in order to
guarantee ease of use by all the players involved in the conservation process of the
architectural heritage (scholars, architects, restorers, etc.) (Apollonio et al. 2017).
It is this need for a shared database and full collaboration between the different
professional figures that has motivated those who are interested in the conservation
of architectural heritage to focus on Building Information Modeling (AHBIM—Ar-
chitectural Heritage Building Information Modeling).
BIM, with its object-oriented semantic organization, fosters a deep knowledge of
the architectural asset through the creation of a shared information system, consisting
of the database model itself (Garagnani 2013). As well as geometric information, BIM
includes data concerning structure, materials, technological system, deterioration,
damage and the different historical phases of the building, reconstructed from the
survey of the existing building and from archival documents (Fig. 1) (Dore and
Murphy 2013; Brusaporci et al. 2018b).
This system sets the conditions for a new approach to the conservation of the build-
ing, in which the project of renovation, maintenance and enhancement of the building
is the result of an interactive and real-time collaboration between the various sub-
jects involved. Within the AHBIM model, intended both as a graphic representation
and as a database, both historical and geometric-dimensional features of historical
architecture can be recorded. Furthermore, the conservation status of a building and
its components can be reported in the database model through the mapping of spe-
cific wood-degradation phenomena. According to the logic of BIM, such forms of
degradation can be easily calculated and also supported by other information such as
the causes of the phenomenon or the type of intervention required (Sampietro et al.
2018). To this end, it is necessary to create specific, shared parameters for including
all information relating to the Facility Management of the building. The different
construction phases of the building can also be included within the same model,
documenting its historical transformations up to the present day, interventions over
time and maintenance activities carried out.
Fig. 1 Aschner hof farm: architectural survey and rendered views
The availability of such heterogeneous information in a single platform promotes

the sharing of knowledge, avoiding the risks inherent in the non- or incorrect
transmission of data. Furthermore, this shared knowledge is offered as a support for
the development of renovation projects and maintenance plans.
Access to this information system is also possible on site through the use of
virtual reality and cloud-based platforms. These applications allow visualization and
utilization of the complex information system that substantiates the BIM model
directly on site, with important repercussions on the monitoring, management and
conservation of buildings.
3 Survey and Analysis
Following a survey phase to acquire the measurements necessary for digital elabora-
tion of the farm building investigated in this study, the canonical drawings of plans,
elevations and sections relating to the two-dimensional representation of the build-
ing were elaborated (Fig. 2). Subsequently, realization of the digital model provided
graphic information regarding the structural, technological and constructional char-
acteristics of the architectural object, highlighting some of the functional, formal and
compositional elements of the Alpine building tradition.
The digital drawings elaborated, including three plans, one longitudinal section
of the farm and four elevations, provide an insight into the internal division of the
building and its layout in relation to agricultural functions. Its architectural typology
Fig. 2 Plan of the first, second and third level of the farm and texturized elevations
belongs to the category of farm known as Haufenhof , meaning groups of buildings.

Unlike the other three categories of farm—two separate buildings (Paarhof), a single
long building (Einhof ) and a single building with two flanked roofs (Zwiehof )—the
Haufenhof has a main building used as a stable and barn and three other separate
buildings: one used as a house (Wohnhaus), one for machinery and another for the
storage of wood (Maschinenraum and Holzhütte).
The plan has a rectangular shape (17.80 × 18.16 m) which determines the compact
volume of the building, divided into two large longitudinal spaces on the ground floor
with a recess in the façade corresponding to a balcony on the west side, a single space
accessed via two wooden staircases (one inside and one outside) on the first floor,
smaller than the level below to accommodate the balcony, and a still smaller attic
space.
The morphological characteristics of the site favored a spatial configuration sep-
arating the internal environments, accessed from three different levels. They corre-
spond to the three levels of the architectural body and are also recognizable by the
differentiation of materials used: the ground floor has a structure in masonry; the first
floor is built of wood and cladding with corner partitions in masonry, and the third
is entirely in wood.
The farm building, on three floors, is characterized by four entrances at a height
of −3.10 m corresponding to the ground floor, located on the west and east sides,
two by two and arranged in axis, and a further entrance to the space used for milk
production. The main entrance of the farm is located on the first floor, with direct
access to the areas used for storage and distribution of forage. Another entrance
is located near the outside balcony, connected to the ground floor via an external
wooden staircase, on the opposite side to the woodshed. The second floor, with the
Fig. 3 Elevations 1-1 and 2-2 of the farm and axonometric render with the exposed roof structure
function of threshing area, has an entrance at a height of +4.26 m, reached via an

external ramp that fills in the height difference.
Drawings of the elevations reveal the layout of the internal spaces in relation
to their functions. The different materials in elevation, along with the presence of
openings or the closure of entire sides, in fact indicate a recognizable constructional
approach in traditional South Tyrolean and, more broadly, Dolomitic architecture.
The part of the building in masonry corresponds to the spaces used for livestock—for
adult cattle, the granary, forage for pigs, storage of bedding and milk production—ar-
ranged around a large central area where hay and sawdust could be stored. The two
levels above built of wood served for the storage and distribution of forage in the
case of the first floor, and for threshing in the case of the uppermost level (Figs. 3
and 4).
As regards the roofing system, the trusses mainly used for this type of construction
are distinguished as Pfettendach and Sparrendach: the former is a structure in which
wooden pillars support the horizontal elements placed on them and with a supporting
function for the joists above, while the latter is an independent structure of sloping
beams to support the load (Fig. 5).
The roof, a variant of the so-called Walmdach (Marino and Pietramellara 1998)
with a wooden structure, is characterized by a system composed of a rafter, a collar tie
and a suspension of the collar tie. Corresponding to the corner solutions, the structure
is characterized by beams that protrude with a progressively greater extension, sup-
porting both the balcony and the roof, a specific construction process of the so-called
“bunk” system. Until the fourteenth century, the type of trussed roofing with a ridge
Fig. 4 Elevations 3-3 and 4-4 and texturized views of the farm with the wooden structure of the
roof
Fig. 5 Section and rendered views of the farm. In the axonometric exploded view, the components
of the roofing system are highlighted
beam, together with the rafter roof structure, represented the construction technique
traditionally adopted in south-western Germany (Panzini et al. 2016).
The entire architectural body is completely visible from the south-facing side,
corresponding to the balcony on the first floor. The masonry construction system of
the ground floor has a wooden superstructure corresponding to the two levels above,
with a mixed system of planks and columns (ständerbohlenbau) and a cross-shaped
frame, better known as bundwerk, with a wooden paneling envelope.
This system of reinforced frames, a typical static-structural element of South
Tyrolean farms, has also become, over time, an architectural and decorative charac-
teristic of the carpentry of this area (Marino and Pietramellara 1998).
The connection between the two structural systems is achieved with wooden
breakers that follow the planimetric profile of the building. Compared to the other
two levels, the ground floor is characterized by a greater number of openings almost
entirely on the west, east and south elevations, while the first floor is completely
covered in wood, apart from the corner sections in masonry. The only openings on
the first floor are in correspondence with the external balcony, two doors on the same
axis.
The top floor is characterized by a large arched opening on the southern facade,
corresponding to the central axis of the plan, on the opposite side with respect to the
entrance.
The paneling is characterized by a series of perforations that alternate with the
wooden cladding system, defining a geometric treatment of the building envelope
and characterizing the architectural type of the farm.
The digital drawings brought to light several typical characteristics of the South
Tyrolean farm, from the identification of the inside-outside connections, the internal
paths produced by the dislocation of the rooms in plan, the division on the basis of
agricultural functions and the spatial configuration as affected by the morphological
complexity of the site.
In this graphic study, the elaboration of two-dimensional representations and the
development of the digital model made for a greater understanding of the spatial
relationships defining the architectural organism. Exploded and axonometric views
activate processes of decomposition and isolation for structural, constructional and
compositional elements that trace the relationship among the parts and clarify a series
of issues concerning the methodological qualities of the architectural typology being
studied.
4 Parametric Modeling
Based on the semantic decomposition deriving from the survey of the farm, the BIM
model was created using Autodesk’s Revit 2016 software.
The advantage of the use of BIM for this case study is the possibility of creating
a model elaborated through almost exclusive use of parametric system families.
This creates the opportunity to use all the parameters already set within the current
Fig. 6 Screenshots of the Revit Architecture software: plan, elevation and three-dimensional view
of the model
platforms, ensuring full recognition of all information entered (structural function,

quantities, physical and mechanical characteristics, etc.).
Regarding the elevations, the two distinct wall typologies, created within the same
walls family, were divided into two different types based on the stratigraphy. The
ground floor wall is the masonry wall, divided into external plaster, masonry and
internal plaster; the wall above is the wooden wall, created as a single layer made of
the wood itself.
The roofing was also created, in its separate parts, from the perspective of semantic
modeling. So the actual roofing consisting of a system roof with stratigraphy in
planks and tiles was created; the main beams and trusses already present in the
software library within the Structural Frame family were loaded into the model,
entered and adapted through simple modification of geometric parameters; and lastly,
the secondary beams were created through the use of the “beam system” option within
the program.
The wooden superstructure, consisting of a mixed system with planks and columns
and a crossed frame, is the only part it was not considered necessary to elaborate using
system families. This superstructure could be realized by three different methods:
creation of an adaptive curtain wall, modeling with system families, local modeling.
All the three methods are correct and possible, the choice of one methodology over
another depends only on the purpose of the model itself (Fig. 6).
The first hypothesis was rejected because the complexity of this superstructure and
the scant possibility of re-use in other models did not justify parametric modeling,
which would be useful only for this model or for identical models with the same
dimensions. As regards semantic modeling, it is possible to subdivide the structure
into pillars, beams and braces, but this complex modeling is justified where the aim
is to obtain a structural model.
In our case, however, the goal being to test the validity of the use of BIM for
wooden constructions and their management and maintenance, local modeling was
preferred. Within this method, the distinct elements were separated in order to assign
different characteristics to them, including those related to deterioration.
As well as a reduction in modeling time, the in-place modeling of these elements
has the advantage of allowing all necessary information to be entered in view of
future monitoring of the building.
Through the creation of shared parameters, it was possible to create new fields
for the insertion of information within the model. Specifically, in order to use the
model for management and maintenance, shared parameters were created regarding
Degradation, Interventions to remedy degradation, the last maintenance work carried
out, planned future maintenance and a field for the insertion of images to monitor
the degradation, its spread and speed of development.
For easy management and use of the model, annual phases were created to facil-
itate planning future maintenance and entering information. The use of such phases
effectively creates a maintenance history, in which information can be updated at
each stage and archived so there is no loss of data and information. Another advan-
tage of their use is the lightness of the model: though the model is complete, all the
information is not always visible but can be recalled by changing the phase in the
view.
5 Knowledge Sharing: Virtual Reality and Cloud-Based

Platforms
The sharing of information among all the professional figures involved in the building
process is the foundation of the BIM approach (Barazzetti and Banfi 2017; Osello
et al. 2018). The opportunity for the various professionals to operate within a single
digital environment reduces the risk of errors due to the loss of information, allowing
the optimization of the entire building process (Fig. 7).
Accessibility to the information content of the BIM model is further enhanced by
current technologies for mobile devices (Williams et al. 2015). Virtual reality applica-
tions and cloud-based platforms, which can also be used on tablets and smartphones,
allow data to be accessed and shared remotely and on site, with important advantages
in terms of management and maintenance.
In particular, cloud-based platforms such as Autodesk’s A360 and BIM 360 allow,
through the sharing of a centralized model, visualization of the information content
associated with the model and the possibility to add data on site. This creates a digital
space in which to interact remotely and in real time.
At the same time, virtual-reality applications which maintain the information
foundation of BIM are also now available: as well as model navigation, applications
such as Revit Live and Stingray by Autodesk also allow the display of database
Fig. 7 Screenshots of the Revit Architecture software: model of the farm and visualization of the
documents inserted in the BIM database
contents, with the sole exception of the parameters added specifically for the case
study.
These virtual reality applications together with cloud-based platforms are useful
tools for the management and maintenance of historical buildings, further strength-
ening the overall aim of the BIM methodology to include the entire life cycle of
the building. The use of these technologies, in fact, facilitates collaboration and
interaction between all the operators involved, including technical staff operating on
site.
In the light of these considerations, the creation of the BIM model of the farm also
took into account the specific needs dictated by the use of BIM as a tool for the Facility
management of the building (Semeraro et al. 2017). In addition to the description
of the architectural elements constituting the farm, new parameters were added to
meet requirements related to the management and maintenance of the building. This
information, used on site through cloud platforms or virtual reality, facilitates and
optimizes the operations of monitoring the building and carrying out repairs.
In particular, the operator will be able to directly access, on site, information
on the latest maintenance work carried out, scheduled maintenance operations, the
degradation of a building component, planned interventions for its removal, etc.
Via the cloud-based platforms, notes and comments can be entered directly into the
digital environment, communicating in real time with other professional figures.
Documents attached to the model and images illustrating an element, useful for
comparing the previous situation with the current one, can be viewed using Revit Lite,
a lighter version of Autodesk’s BIM software. This tool, which runs on any Windows
mobile device, allows the user to access all information stored in the database, display
all linked contents (archival documents, images, etc.), and modify and integrate the
model with additional data.
Finally, management of the temporal phases of the model, set during the modeling
of the farm, allows continuous updating of information without losing previous data.
In this way, a history of damage and maintenance interventions is created, useful
both for the management of the building and for a deeper knowledge of it. In addi-
tion, setting of annual phases makes the model easier to use remotely because it is
significantly lighter when unnecessary data are not displayed.
The use of current technologies to access information on site means that technical
staff responsible for the maintenance of the building can assess the state of conserva-
tion of the building, monitor any phenomena of deterioration and damage, and carry
out repairs more easily and efficiently.
The possibility of accessing the information content of the BIM model from
mobile devices on site, therefore, further enhances the potential of the BIM process
as a tool for the management and conservation of historical buildings.
6 Conclusion
Building Information Modelling is a tool of great interest for the conservation of

wooden architectural heritage such as the Aschner hof farm.
The conservation and enhancement of historical buildings are based on a deep
knowledge of the building that requires management of a large quantity of extremely
heterogeneous data: archival documents, photos, historical phases, construction
equipment, information on degradation and damage, maintenance and restoration,
etc. (Fig. 8). This diversified information needs to be properly managed and shared
among the various professional figures working on heritage, so that problems due to
information redundancy or transmission errors can be avoided.
Building Information Modeling, with its capability for storing and managing
data related to transformations and the use of buildings over time, certainly offers
interesting opportunities for conservation of architectural heritage, including wooden
structures.
Furthermore, technologies enabling the use of models and their databases
remotely, from mobile devices, allow information to be accessed and shared on site,
making maintenance and management of the building easier. The ability to visualize
the information content of the model, compare the conservation status recorded in the
database with the current observable situation, insert annotations and contents, etc.,
is an important and significant advantage for optimization of the process, confirming
the crucial role that BIM can now play not only in the design of the new buildings
but also for management and conservation of the wooden architectural heritage.
Fig. 8 Rendered exploded view of the farm
Acknowledgements Although the chapter was conceived unitedly, Brusaporci and Luigini are
the authors of paragraphs 1 and 6; Vattano is the author of paragraph 3; Maiezza is the author of
paragraphs 2 and 5, Tata is the author of paragraph 4.
References
Apollonio FI, Gaiani M, Sun Z (2017) A reality integrated bim for architectural heritage conser-
vation. In: Ippolito A (ed) Handbook of research on emerging technologies for architectural and
archaeological heritage. IGI Global, Hershey, PA, pp 31–65
Barazzetti L, Banfi F (2017) Historic BIM for mobile VR/AR applications. In: Ioannides M,
Magnenat-Thalmann N, Papagiannakis G (eds) Mixed reality and gamification for cultural her-
itage. Springer, Cham
Bianchini C, Inglese C, Ippolito A (2016) Il contributo della Rappresentazione nel building infor-
mation modeling (BIM) per la gestione del costruito. DISEGNARECON, 9(16):10.1–10.9
Bianchini C, Inglese C, Ippolito A, Maiorino D, Senatore LJ (2017) Building information modeling

(BIM): great misunderstanding or potential opportunities for the design disciplines? In: Ippolito
A, Cigola M (eds), Handbook of research on emerging technologies for digital preservation and
information modeling. IGI Global, Hershey, PA, pp 67–90
Brusaporci S, Maiezza P (2016) Re-loading bim: between spatial and database information modeling
for architectural heritage documentation. In: AA.VV. Dibujar, Construir, Sonar. Investigaciones
en torno a la expresion grafica aplicada a la edificacion. Tirant Lo Blanch, Valencia, pp 835–847
Brusaporci S, Maiezza P, Tata A (2018) AHBIM come sistema di rappresentazione complesso dei
beni architettonici. In: Marcos CL; Gutiérrez P JJ; Gresa JD (eds) De trazos, huellas e improntas.
arquitectura, ideación, representación y diffusion. Byprint Percom SL, Valencia, pp 485–494
Brusaporci S, Maiezza P, Tata A (2018b) A framework for architectural heritage HBIM semantiza-
tion and development. Int Arch Photogramm Remote Sens Spatial Inf Sci, XLII-2:179–184
Dore C, Murphy M (2013) Semi-automatic modelling of building façades with shape grammars
using historic building information modelling. In: Int Arch Photogramm Remote Sens Spatial Inf
Sci, Volume XL-5/W1
Garagnani S (2013) Building information modeling a real world knowledge. A methodological
approach to accurate semantic documentation for the built environment. In: Digital heritage
international congress (Digital Heritage), pp 489–496
Marino L, Pietramellara C (eds) (1998) Tecniche edili tradizionali: contributi per la conoscenza e
la conservazione del patrimonio archeologico. Alinea Editrice, Firenze
Osello A, Lucibello G, Morgagni F (2018) HBIM and virtual tools: a new chance to preserve
architectural heritage. Buildings. MDPI, Basel
Panzini N et al (2016) Tecniche tradizionali e artigianato nella Scuola di Stoccarda. Collana di
Architettura del Politecnico di Bari. Gangemi Editore, Bari
Sampietro F, Masciopinto ML, Cincotta I WJ, Bergonzoni G (2018) Computational design nel
processo HBIM. Dn. Build Inf Model Data Semant 2/2018, pp 86–96
Semeraro F, Osello A, Rapetti N (2017) Un approccio basato sul BIM per la gestione e manutenzione
del patrimonio edilizio. DN Building Information Modeling, Data & Semantics
Williams G, Gheisari M, Po-Jui C, Irizarry J (2015) BIM2MAR: an efficient BIM translation to
mobile augmented reality applications. J Manage Eng, 31(1)
Shape and Design of Ancestral Fishing
Machines Between Geometry
and Technology
Domenico D’Uva
Abstract The choice of studying trabucchi as bearer of knowledge is justified

through the aim of clarifying the expertise they hand down and make them available,
after they have been interpreted and organized. For this purpose, the technological
analysis of the studied material has been critically conducted, considering historical,
political and cultural issues they are connected with. It will be cleared how structural
conception of trabucchi is strictly connected with the aim why they have been built,
and any technical choice has been motivated by practical needs, which can be under-
stood through a deep knowledge of the places, of the possibilities and the economic
environment where trabucchi were born and evolved. By analyzing the logic with
which its parts have been assembled (as a base for any constructive choice) it will be
pointed out the extraordinariness of the structural scheme, which reached excellent
results by exploiting the available resources with great intelligence. The case study
is the Punta Turchino trabucco, through its specific environment, it will be analyzed
a preservation/reconstruction design.
Keywords Trabucchi · Construction · Fishing machine · Adriatic sea · Ancestral
1 Introduction
A study regarding trabucchi is a choice which can firstly arouse a sense of amaze-
ment, (D’Annunzio 1894) because the knowledge of these machines is currently very
limited.
The first question which is worth to be asked is “what is a trabucco”? It is essen-
tially a wooden machine adapted to fishing used along center-south Adriatic sea
shore, in the regions of Abruzzo, Molise and Puglia (La Regina 1980). It exploits a
squared net, uplifted by a winch and supported by poles. Trabucco is a fishing struc-
ture which can be located off shore, even more than 100 m far from mainland, thanks
D. D’Uva (B)
Department of Architecture and Urban Studies, Polytechnic University of Milan, Milan, Italy
e-mail: doduva@gmail.com

548 D. D’Uva
to a system of poles anchored to the seabed. At the end of a connection boardwalk,

a platform is located, standing on four poles growing from water; long wooden joint
rods are added to the main structure toward the sea, for sustaining the fishing net.
The second question in need of a clarification is whether trabucchi are different
from any other human constructions because of their building inner logics. Isostatic
or hyperstatic constructions represent more than 90% of the architectural available
heritage. Trabucchi are instead labile constructions, built following different criteria
which gained immense advantages from lability to face several needs.
The choice of studying trabucchi as bearer of knowledge is justified through the
aim of clarifying the expertise they hand down and make them available, after they
have been interpreted and organized. For this purpose, the technological analysis of
the studied material has been critically conducted, considering historical, political
and cultural issues they are connected with.
The work will analyze in detail which typological component distinguish a tra-
bucco from any other manufacture and how these elements react with environment
(Misino and Trasi 1995). The aim of this work is the definition of a strategy and
issues for preservation to be applied to the case study of Trabucco di Punta Turchino
(Fig. 1), located in San Vito Chietino, in Abruzzi region of Italy.
Fig. 1 Trabucco Punta del Turchino, 2003. Courtesy of arch. Carla D’Uva
Shape and Design of Ancestral Fishing Machines Between Geometry … 549
2 Constructive Elements Technology
Trabucco, as it is meant in this work, constitutes a general category of manufactures

which embrace the following characteristic features. The elements which define a
trabucco as defined in (Barone and Marino 1999) are:
1. Platform
2. Cabin
3. Winch
4. Flagpoles
5. Load bearing poles
6. Net
7. Catwalk
2.1 Platform
Platform is the part of trabucco located in the farthest position from land, where
all fishing activities are carried through: immersion and lifting of the net, waiting,
collection of catch and sharing among fishermen. On the platform there is a winch,
and a cabin for the storage of tools, which initially also served as a shelter for the
fisherman in long waits; it is placed so that the network is visible from inside. The
supporting structure of the platform is made up of wooden pillars, linked with various
framework to the seabed, in the rock or on the rocks. Sometimes it happens that the
platform is placed directly on the rock without the need for an artificially constructed
load-bearing structure. In the case of pillars embedded in the rock, on the other hand,
they are made up of a single piece of wood, made integral with the foundation by
means of wooden ‘wedges’ and concrete castings. If the pillars are partly submerged
by water, they are formed by two elements: an upper one of wood and a lower one
of iron, in direct contact with the water. If the pillars are partly submerged by water,
they are formed by two elements: an upper one of wood and a lower one of iron, in
direct contact with the water. The iron beams, which are anchored to the bottom) are
frequently made by a piece of rail, coming from tracks of nearby railway, recycled
for this use. The anchoring of the foundation depends on the type of actual seabed:
if it is sandy, then the beam is inserted at a greater depth (from 80 to 1.5 m), to avoid
a subsequent subsidence. With the presence of rocks, instead, a hole is made into
the rock to place the base of the beam; everything is initially blocked with wooden
wedges, then with cement. The union between timber and iron parts is instead assured
by iron or steel bolts; in the past they were also used nails or even ropes. This frame
of bracing beams is joined at the top by a system totally made up of wooden logs.
In some cases, beams can form simple trilithic systems, in others they can create
squares or rectangular-mesh systems by means of two orthogonal frames. In more
recent trabucchi it is possible to find trunks placed diagonally to form cross bracing,
placed on the horizontal plane. In many cases, side reinforcements made of timber
550 D. D’Uva
are added to this structure. These consolidations are formed by a pillar and by at least
two beams placed according to a triangular shape, with the vertex towards the pillar
and the base towards the platform. On this supporting structure, secondary beams are
placed or smaller beams with a square cross-section, which support a deck of boards
(‘scurette’ of 20 mm or ‘mezzanelle’ of 15 mm). The size of the platform, which in
antiquity was always 25 m2 , today reaches 40 m2 , and includes both uncovered and
covered parts.
2.2 Cabin
The covered part of trabucco can sometimes occupy the entire surface of the platform;
in this case the vertical walls are present only on three sides, while the part facing
the sea is left open to allow all operations to maneuver. In the Abruzzese-Molise
typology the covered part is reduced to the cabin, usually 3 × 3 m, a shelter formed
by wooden planks, in the vertical elements, with flat or two-pitched roofing. This
is used by the fishermen as a shelter in long waits or in case of adverse weather
conditions and as a deposit of tools or other stuffs. In the reconstructed trabucchi
in recent times, the cabin (Fig. 4) can also perform other functions, becoming in
some cases a real home, equipped with kitchen and other comforts. In the past wood
and straw were the only materials used for the construction of the cabin, today, they
are used also various types of plastic materials, especially for roofing. In trabucchi
located in harbors, the tendency to abandon traditional materials is even more evident
and the cabin sometimes takes on the appearance of a colorful collage of sheets and
plastics.
2.3 Winch
It is the tool used to pull the ropes linked to the net and effectuate all of the fishing
maneuvers (Fig. 4). It is usually made of a wooden trunk, about 2 m high, with
variable diameter. Its rotational movement is made possible by an upper and a lower
linchpins, which were made of wood, afterward replaced by iron bars in modern
trabucchi. The lower end is hinged to the base of the platform by means of wooden
boards or iron plates, hooked to the pavement; the upper end is usually anchored to
a wooden crosspiece, thanks to iron stirrups or, in the case of newer winches with
steel elements of different shapes and typologies. The rotation of the central part of
the winch is obtained by pushing the horizontal elements (two or four) connected to
it, at a height of about one meter and twenty, so that the angler can perform, with
the minimum effort, the maneuver to rewind the ropes. The horizontal rods, which
previously were exclusively made of wood, today are made up of steel bars and the
rotation of the winch, in some cases, is no longer carried out manually, but using
the electricity. In oldest winches, they were nailed to it some thick wooden boards
placed in a circular crown for increasing the upper section of the trunk. In this way,
the point where the horizontal arms are interlocked had a smaller section, while the
upper part, where the ropes are collected, had a larger diameter so that a limited
number of turns could give rise to the winding of many meters of ropes. In some
types of trabucco, only one winch is not sufficient for lifting the network, which can
span up to 70 m2 . In this case we find two winches, placed on the same axis in the
longitudinal direction, each of which is intended to collect half of the ropes, to lift
the net by dividing the load in two.
2.4 Flagpoles
There are four flagpoles in a trabucco (Fig. 2), two of which are longer and more
resistant, called antennas, and two shorter and more external ones called “antenninas”
(small antennas). The antennas are formed by overlapping and connected perches,
up to a maximum of 40 m of overhang. The wooden perches (ranging from 6 to
8 m in length) range from a minimum of one to a maximum of six elements. The
junctions are made by means of bolts and wire but in the past only hemp ropes
could be used, as the trabucchi were much smaller and the antennas much shorter.
Junction between two perches is obtained by overlapping the parts, for a length
of about 80 cm, in various directions: laterally, above or below, according to the
constructional traditions developed in the place where the trabucco was installed.
The antennas are also connected to the load-bearing piles, in the section closest to
the platform, by means of bolts or iron stirrups. The antenninas, shorter than the
antennas and often formed by a single perch, usually start from the supporting poles,
at a height higher than the latter. They bear a lower load than the antennas and are
oriented in a direction to form an obtuse angle between them.
2.5 Load Bearing Poles
Antennas and antenninas are supported, by means of tie rods, by the respective
uprights, present in number from two to four. These poles (Fig. 3) are directly sticked
into the seabed or in the rock and reach a height of 40 m. They have a proportion ratio
with the antennas because their stamina relies upon them. Between the uprights and
the flagpoles, there are strands of iron, generally three per pole, which support the
entire cantilevered structure. We can mainly distinguish two types of tie rod anchors:
one, which is used more frequently in the Gargano type, the other in the remaining
cases. The first type exploits the presence of the rocky ridge where wooden poles are
placed, in a position diametrically opposite to the antennas, with respect to the center
constituted by the upright. The wire, tied to the antennas, passes through a hole made
at the end of the upright, wraps around it and is pulled from the opposite side, up
to the anchoring post. In the Abruzzese-Molisano typology, the wooden post (called
552 D. D’Uva
codittone) that acts as anchor, is instead fixed in the seabed (with the iron-wood
system used for all foundations). So its position depends on the arrangement of the
rocks compared to the trabucchi and is not specified a priori. The same codittoni
also support iron wires that have the task of bracing the structure and that are kept in
different places, according to the possibilities offered by the nature of the place. The
bracing devices are not stretched (as opposed to the tie rods) but only if necessary,
when the structure is stressed (during the hoisting of the network, in the event of
storms, strong winds and other weathering). Some iron wires, bracing or pulled, are
also connected to the uprights of the walkway that is the element of union of the
platform with the ground.
Fig. 2 Survey of trabucco, flagpoles marked in red, 2003. Courtesy of arch. Carla D’Uva
Fig. 3 Survey of trabucco, bearing poles marked in red, 2003. Courtesy of arch. Carla D’Uva
2.6 Net
It has a quadrangular shape and usually measures a minimum span of 15 × 15 m, in

the Abruzzo-Molisani trabucchi, up to dimensions of 30 × 25 m, in the Gargano ones.
In ancient times, the network was woven by hand, by the families of fishermen, with
cotton and hemp threads along with other vegetable fibers. Today, mostly artificial
fibers such as nylon, plastic or metals are used, made by industrial manufacture. The
nets used for fishing are usually made of cotton and nylon, have very tight meshes
that thicken towards the center or ‘bottom point’, where the meshes reduce up to be
almost closed in the central part of about 1 m2 . The net perimeter is formed by a thick
and resistant cord called “file” in local dialect to which are connected the four main
strings, called “vendle”, which support the corners of the net. There are also smaller
and less resistant strings, “filacciuoli”, (small ropes) which start from the perimeter
file, and which help to lift the net. The main strings pull the net upwards, when they
are recalled by the winch, passing through the pulleys (once in wood, today in iron
or steel) placed at the ends of antennas and antennas and, in some cases, even in the
middle. The number of pulleys used depends on the size of the net and on the number
of ropes used to lift it.
2.7 Catwalk
It is only present in the Abruzzese-Molise typology and it is a strongly characterizing

element. It is the main connection element with land (Fig. 4); its length reaches 40 m
in some cases and its trend is never perfectly regular nor exactly perpendicular to the
coastline. It is in fact based on the layout of the rocks, describing always different
paths with some sudden changes of direction. Its struts usually form trilithic systems,
by means of wooden crossbars, and are anchored to the bottom in the same way as
the bearing piles of the platform. These portals are also braced in different ways:
with simple diagonal elements, with a St. Andrew’s cross scheme or with horizontal
reinforcements. Sometimes the trilithic structure is replaced by systems of crossed
diagonals or by types of supports obtained by the ingenious union of several parts.
All these systems have been created to adapt, in the best possible way, the product to
the environment in which it is placed. The resulting structures do not follow therefore
codified rules, but there is a research for the most suitable method, to ensure that
nature accepts them and not discards them, assuming always new and original forms.
The catwalk therefore rests on this succession, (more or less regular) of portals and
supports, through the use of main and secondary crosspieces, placed orthogonally.
The boards forming the walking surface are nailed directly onto the crosspieces,
placed perpendicular to the short side of the walkway. Usually catwalks are made of
fairly regular boards of width from 15 to 20 cm and variable length. The handrail can
consist of wooden posts, ropes or wires, which are connected to the uprights from the
foundation and up to about a meter and a half, compared to the base of the catwalk In
554 D. D’Uva
Fig. 4 Cabin, Winch, Catwalk, 2016. Google photos
some cases the voids formed between the handrail, the uprights and the base boards
are closed by nets of cotton, hemp or nylon, in other cases they are left totally open.
The ground connection of the catwalk is obtained in a way that is always different
depending on the case and the conformation of the territory. Sometimes it is facilitated
by ladders and wooden ramps, in other cases it is simply leaning against the rock,
exploiting the natural slope of the land. Despite having well-defined technological
features, trabucco does not present itself as a standardized product and the elements
that are repeated slavishly in all contexts, but as a set of structures, always original,
characterized by the same purpose. Timber, the main trabucchi construction material,
deserves a separate mention, thanks to which it has been possible, since the most
ancient times, to obtain constantly evolving results, achieving its full potential.
3 Analysis of Materials Decay
Before starting with preservation project, it is necessary to understand the main causes
which brings the decay of trabucco, its causes and its remedies. Decay of trabucco
may be considered from different points of view: material decay with which trabucco
has been built, static decay of construction elements which cannot be evaluated with
ordinary criteria, the decay owed to artificial overlapping parts, the decay caused
by the loss of primary function as ancient fishing machine and eventually the decay
caused by a state of neglect and abandon. With this work it will be analyzed the
material decay by which main component of trabucco are built, timber and iron.
Starting with the material that make 90% of the structure, the timber, it is possible to
characterize some among main causes of the main causes for its decay, specifically
located in the applied environment. For this purpose it is useful to remind that the
most ancient trabucchi had tree trunks which were erected directly in the sand, as
they were offered by natural environment. Frequently these logs were used with the
barks and they were not treated with anti-parasitic or other protective substances,
except for poles once used for phone lines. This approach increased the probability
for timber structures to be attached by insect and fungi, but originally it was not
a problem because supporting poles were replaced for other causes as collapses or
subsidence. Currently, on the other hand, the most ancient part of trabucchi, being
reinforced with new techniques are able to resist for many years and to host a wide
variety of insect and parasites. A certain number of studies dealt with this issue,
therefore results may be summed up as follows. Decay agents of timber have been
splitted into biotic and abiotic factors; among biotic factors are some fungi as caries,
soft caries, blooming, insects and clam. Among abiotic factors are light, humidity,
temperature variations and whether in general. Caries are fungi which cause alteration
or decomposition of vegetal fabric given by basidiomycetes. Other fungi belonging to
the type of deep chromogenic are tricoderma lignorum, which establish themselves in
intrados of trabucco flat surfaces, remaining permanently in a humidity state because
of the lack of direct sunlight. Timber structures are attacked mainly by insect called
cerambycid, belonging to the family of pest beetles. Cerambycids are categorized by
the wood they colonize, therefore it is possible to notice the presence of cerambyx
cerdo (oak tree cerambycid) and hespherophanes cinerens (in robinia). Other families
of insects are spreader in wet parts are anobiids. These are small beetles belonging to
bostrychus family with dark and uniform livery, whose larvae live in wood, excavating
meandering galleries and causing severe damages. One among the most dangerous
woodworm belongs to this family, the xestobium ufovillosum, commonly known
as the death-clock and anobium punctatum, known as Saint Paschal clock. It is
commonly found in the fissure created by wood contraction, for example in chestnut
wood, which shows an alveolar aspect in some of its parts, up to the detachment
of sapwood. Another paramount biotic decay factor is sea water, that causes decay
to metallic structures, and generates phenomenon of marine aerosol. However, the
most important among humidity-related biotic decay factors is humidity. In fact
timber whose trabucco is made is in a permanent humidity condition, higher than
20%, since its construction until it is dismantled. In biological risk classification,
issued on 21/12/1998 by CEE guideline 89/106 “durability of timber and wooden
derived material”, timber used in trabucchi are valued as risk class 4 and 5. In the
specific environment which characterizes Abruzzo coastal area, the wide diffusion of
robinia on escarpment allowed the use of this essence as a main fabrication material
for the construction of trabucchi. Climatic conditions and humidity remains constant
during the several phases of locating, cut and implementing, therefore contributing to
considerably improve timber performances. Other abiotic decay factors for timber are
light and temperature variations, which influence stamina in great measure because
of the alternance of fiber expansion/contraction, damaging the material after a series
of continuous cycles. Light is a paramount factor for timber decay, because sun
causes chromatic alteration, changing colors toward darker or paler, usually without
chemical variations. It may occur in addition, that the concurrent action of freshwater
and brackish water contributes in melting the lignin, which is the most abundant
wood component after cellulose. This factor has the effect of damaging the median
lamella, which is the primary membrane of the cells of a fabric, with the result of
a chromatic alteration toward grey and a substantial decay of structural stamina.
This loss is substantial, especially in the wood produced during spring season. It is
556 D. D’Uva
therefore possible to frame decay agents into biotic and abiotic that cause weakening
and eventually destruction mainly in timber, while damaging is limited regarding
metallic part of trabucchi. Decay can be brought to extreme consequences due to
the destruction of the manufacture for scarce maintenance of structures, bad design
which can, together with adverse weather condition bring to the collapse of the
structure.
4 Phases for a Methodic and Coherent Preservation
Structural conception of trabucchi is strictly connected with the aim why they have
been built, and any technical choice has been motivated by practical needs, which
can be understood through a deep knowledge of the places, of the possibilities and
the economic environment where trabucchi were born and evolved (Gasparri 1995).
By analyzing the logic with which its parts have been assembled (as a base for any
constructive choice) it will be pointed out the extraordinariness of the structural
scheme, which reached excellent results by exploiting the available resources with
great intelligence.
The case study is the Punta Turchino trabucco, through its specific environment,
it will be analyzed a reconstruction design (Zecca 1993). Despite the lack of law
references about trabucchi, it seems appropriate to set general guidelines to be applied
to analyzed manufactures.
Once decay has been analyzed, it is paramount to differentiate elements which
belong to the continuous process of ordinary maintenance from added parts outside
of the original structural conception. For this purpose the survey must be carried
out with critical methodology, standing as a tool for active analysis of the structure,
answering to objectives and aims acquired before. The phase after data collection is
the element discretization, by choosing working elements among all and selecting
damaged parts that must be removed. The choice criteria are based on the constructive
logic with which trabucco has been built. All the accretions must be removed. For
this purpose, as it is impossible to remove all the added parts without modifying the
perception of the manufacture, it would be a smart methodology the prohibition of
the reconstruction of non-suitable pieces, when environmental agents provoke their
decay.
The choice to finalize the preservation to the integration of cultural heritage is due
and inevitable, because of the uniqueness of the manufacture and its original function.
It is possible, in this way to save the essence itself through the construction, without
applying modifications which would deteriorate it.
It is an extreme measure, although effective to hand down to future generation
a testimony of the past in the best possible conditions. Meaning trabucco as a liv-
ing museum appears to be a sign of civilization, aimed to spread the diffusion of
knowledge of which there is a witness.
It seemed suitable to set an array of general rules which can be applied for
preservation and restoration of a wide range of trabucchi (De Gotzen and Laner
1988). Each proposal should be specifically aimed to restoring the constructive

essence without leaving aside the renovation of damaged parts. Unluckily, due to
a lacking normative framework and diverse ownership purposes, the interventions
made on trabucchi were disproportionate and inordinate, without stringent rules on
preservation to comply with. The enormous waste of energies and material along
with the loss of original function demonstrated a complete lack of knowledge of the
target of the preservation, denoting as their only aim the pursuit of personal interests.
It has been demonstrated how modern structures have been designed with an hyper
static structural approach, (Mombelli 2000) with a large use of crossed bracing
beams and modern commercial timber grid structure, wastefully refined with few in
common with constructive simplicity and essentiality which characterized ancient
trabucchi. For these reasons it is necessary to highlight specific guidelines which
allow designers to effectuate aware restoration interventions, aimed to magnify the
peculiarity of manufactures that retain original features.
Proposal for an aware restoration starts from a punctual survey of all of the struc-
tural and non-structural elements which make trabucco, with the purpose of verify-
ing the actual coherence with the identifies constructive logic. For this purpose it is
possible to start the survey task from historical photos which represent paramount
elements for understanding the several interventions occurred through time. The
difficulty of this method is in separating elements “outside” from the original struc-
tural conception from elements which belongs to the continuous ordinary process of
maintenance of trabucco. Following these criteria survey must be carried out with
a critical approach in a way to be as a tool of active analysis of the structure, by
answering to aim and purposes defined ex-ante.
Following this very first phase of essential information collecting, there is data
analysis, whose result is a discretization of the elements which can be restored
or replaced and the outside elements which damages the correct operations of the
manufacture and must be removed. Criterions which justify division, evaluation
and selection are based on learning the constructive logic with which trabucco has
been conceived, as it was before explained. Parts which aim is coming back in their
original working position undergo a decay analysis of their constituent materials, for
understanding their replacement with new elements which are equivalent to the fore-
most, following the same requirement and locating with the same constructive logic.
It should be stressed that in this very case that material replacement is not a fake, due
to the features of trabucchi, but it is part of the continuous regeneration process of
the structure, meant as a ceaseless building site. In fact, as it has been several times
highlighted, a continuous process of maintenance can ensure the preservation of
the constructive essence of the trabucco; the addition of structural stiffening modify
trabucco static behavior, causing a damage to preservation. All of the unnecessary
additions go against essentiality and economic logic that it is worth to be preserved.
The most delicate issue concerns cases where the functioning of trabucco has been
completely transformed due to a series of non-suitable modifications made in pre-
vious time. This option indeed concerns about 90% of the analyzed cases, therefore
it is suitable to propose a solution coherent with the premises and the target set for
this work. It is a standing that can embrace facts and environment, without omitting
558 D. D’Uva
the knowledge acquired with interpretation of original structural conception. Being

aware of the impossibility to remove all of the fake accretions that make trabucchi
completely different from their ancient counterparts (not only structurally), this
work propose a prohibition of reconstruction non suitable parts when external agents
causes their decay. Therefore an aware proposal would be a further prohibition to
transform actual manufacture in dimensions, material and constructive elements.
5 Preservation and Reconstruction Activity
The aforementioned detailed method for preservation of trabucco di Punta

Turchino has been proposed on the base on the survey phase, which was conducted
in 2003, when the manufacture was in an average state of conservation. The survey
phase was not limited to the manufacture itself, but it embraced the environment
around it, in fact also variations of sea level and underwater rocks were surveyed in
order to have a full comprehension of the structure. In this very case it was funda-
mental because of the future events that occurred. In fact the enduring activity of
weathering and the lack of maintenance has weakened the timber and the connections
with railings so that few years after the survey activity the trabucco was in a severe
state of decay. The continuous activity of maintenance is the base for preserving life
of trabucchi, as it was defined a continuous construction site (Laner and Trabucchi
2012). Instead the system continued in its decay, as it was possible to point out in
Fig. 5 Progression of trabucco collapse, 2013, 2014. Courtesy of Domenico Gargarella. www.
domgarga.com
Fig. 6 Trabucchi location in Abruzzo, Molise and Puglia. Map from Bing
Fig. 7 Trabucco from the shore, 2018. Courtesy of arch. Carla D’Uva
2013, that the platform started collapsing. The decay continued through all the sum-
mer 2014, when the night of 26th July 2014 a severe storm destroyed what remained
of the ancient structure (Fig. 5). The following year a preservation design was laid
out; in June 2016 the reconstructed trabucco was inaugurated (Figs. 6 and 7).
It is worth to underline how the survey and knowledge process in destruction
phases (Laner 1983) is paramount to preserve and rebuild the characteristic elements
of the original trabucco. Therefore the subject of this work has been under all of the
phases of its life, going through decay, collapsing and reconstruction. The analysis of
the life of this manufacture would make easier to understand the issues that similar
560 D. D’Uva
structures, as trabucchi located in different locations can face. The first and most
important consideration to be expressed is that the chosen manufacture has under-
gone a correct preservation process, because it has maintained original functions and
feature elements has been rebuilt with correct form and material. This is absolutely
uncommon because most of the trabucchi in this area, which were undergoing a
severe decay process were restored and transformed in seaside (or more properly on-
the-sea) restaurants, vanishing all the knowledge for preservation process explained
before. Trabucco del Turchino, instead was rebuilt following the original design and
functions, restoring all of the original elements with few modifications.
This can be verifying by the overlapping of original drawings of 2003 and new
construction on 2014. Although very similar in most of the details, two trabucchi
differ in small aspects (Fig. 8).
The comparison between the different surveys was carried out by overlapping the
platform and cabin (as shown in Fig. 9), because from a photographic survey held
after the collapse, it was clear how the old railing foundations were preserved. In the
current phase of this research it was impossible to increase precision because the 2003
survey was conducted with direct measures of the manufacture and a geo-localization
system was not used. Location all of the already defined features of trabucchi has
been roughly maintained and material have been replaced following the knowledge
criteria. The catwalk reconstruction is slightly different from the original; the position
is lightly different, it has been flattened and a safer handrail has been added. The only
objectionable modification is a general increase in the number of bracing beams. The
overlapping between the two construction provided even more results, as the angle
among the antennas has been widen, but their length remained the same.
6 Conclusion
The result of this work can be summed up in the recognition of the importance of
survey (Fig. 10) and understanding of trabucchi, to preserve structure and environ-
mental features of the Abruzzi and Molise coasts. The case study represent a virtuous
example because it is a symbol of trabucchi, a paramount manufacture, which was
widely surveyed and studied. There is a large number of original trabucchi that have
few (or no) studies; these manufacture are worth to be preserved as a cultural heritage
of the ancient life of fishermen in these regions of Italy. An accurate survey of the
manufacture and a continuous work of maintenance are the only possible method for
achieving preservation or allowing reconstruction of this ancient fishing machines.
Fig. 8 Reconstruction design of the trabucco, plan and elevation, 2014. Drawings from Comune
di San Vito Chietino
562 D. D’Uva
Fig. 9 Overlapping of 2003 survey and 2014 reconstruction. 2003 survey marked in red. Courtesy
of arch. Carla D’Uva
Fig. 10 3d model of destroyed trabucco, 2003 survey

References
Barone P, Marino L (1999) I Trabocchi, macchine da pesca della costa adriatica Cierre Grafica,
Verona
D’Annunzio G (1894) Trionfo della Morte. A. Mondadori Editore, Milano
De Gotzen S, Laner F (1988) La chiglia rovesciata, Ricerche di tecnologia dell’architettura, Franco
Angeli, Milano
Gasparri G (1995) La costa magica dei vecchi trabocchi. Una corsa dalla marina di S. Vito a
Fossacesia lungo l’Adriatico, tra le dannunziane “macchine” per la pesca, in Bell’Italia n°114,
Milano
La Regina A (1980) Molise, una regione piccola e di poche risorse, BNL, Electa editrice
Laner F (1983) Sicurezza e crolli. Cluva Università, Venezia
Laner F, Trabucchi (2012) Artorigenerazione come chiave di letture, in Geocentro Magazine, 20,
Fondazione Geometri Italiani
Laner F (1994) Inferenze tecnico-progettuali di originarie concezioni strutturali, in Dialoghi di
restauro n°3, Trento
Misino P, Trasi N (1995) Resistenti leggerezze, Capanni da Pesca sulla costa abruzzese, col.
“Ossimoro”, Pescara
Mombelli R (2000), Trabocchi. Anonymous timber architecture in Italy, Detail, 1/2000
Zecca G (1993) Il ragno pescatore, in D’Abruzzo, edizioni Menabò
3D Digital Systems
for the Documentation
and Representation of the Wooden
Heritage Between Finland and Russia:
Survey Methods and Procedures
for Detailed Analysis
Sara Porzilli and Stefano Bertocci
Abstract This contribution presents research experiences carried out on historic

wooden heritage located between Northern Finland and Russia. Many cultures
all over the world hold their own testimony related to historic wooden archi-
tecture (Iori 2005), different in their constructive techniques and compositional
results but at the same time full of common features. For this reason we can
assume that wooden architecture represents one of the oldest building systems
adopted all over the world by people who developed specific approaches in build-
ing with timber material while respecting local tradition (Pryce 2005). This vast
and unique wooden world heritage strongly needs today to be surveyed and
documented with a deep level of detail and accuracy by using the most accu-
rate survey methods and digital systems. The contribution describes three spe-
cific case studies located in different geographical areas: The Unesco World Her-
itage Site of the Pogost Complex on Kizhi Island and rural settlements in Kare-
lia (Russia), The Wooden Farm House of Lamminaho in Vaala region (Finland)
and the residential wooden district of Raksila in Oulu (Finland). The research
experiences presented have the aim to offer a wide viewpoint on how survey opera-
tions and analysis should be performed and which are the main strategies and proce-
dures useful for obtaining specific results. From 2012 until now European Funding
has supported researches on Wooden Architecture with a Seventh Framework Pro-
gramme, Marie Curie Actions People (years 2012–2014) and from 2017 until 2019
with a Marie S. Curie Individual Fellowship held by Post Doctoral Researcher Sara
Porzilli, actually based in Oulu, Finland.
S. Porzilli (B)
Department of History of Architecture and Restoration Studies,
University of Oulu, Oulu, Finland
e-mail: sara.porzilli@oulu.fi
S. Bertocci
Department of Architecture DIDA, University of Florence, Florence, Italy
e-mail: stefano.bertocci@unifi.it

566 S. Porzilli and S. Bertocci
Keywords Wooden architecture · Timber structures survey

Parametric timber engineering · Laser scanning · Photogrammetry
1 Introduction
Wooden Heritage represents one of the oldest example of architecture developed

by different cultures from all over the world, who developed specific features with
unique compositional solutions (Zwerger 2012). In Northern Europe as well as in
Russia, wooden buildings existed already during the early 17th century (Anker and
Snitt 1997). They were built in a tight connection with the surrounded environment
creating intimate relationships with the landscape. Materials belonged to the near
forests, decorative motif and details took inspirations by the Nature and its elements
(Fig. 1). In many occasions, craftsmen and artisans, direct authors of these unique
heritage, were “simple” skilled experts of the place, inhabitants of the surrounded
areas, people of whom we often have not today any precise information. Today,
because of negligence, sudden events like fire, or even because of absence of accurate
plans of preservation most of our world wooden heritage is already reduced, part of
it is dramatically abandoned or totally vanished without any documentary testimony
(Fig. 2).
However, even if historic wooden buildings and timber structures were not consid-
ered enough important in the past, evidently due to a greater preference and attitude
for the more ancient stone and brick architecture, part of the wooden heritage still
survived until now and it is still present in many regions (Hansen Hans 1969).
Fig. 1 Historic wooden architecture from Russia and Finland has many examples of constructive
and decorative details that seem to explicitly emulate the forms of Nature. Perhaps also, for this
reason, wooden buildings and landscape appear harmoniously integrated creating a unique cultural
landscape in which architecture and northern forests live together
3D Digital Systems for the Documentation and Representation … 567
Fig. 2 The historic wooden settlement of Jylhämä in the region of Vaala, Finland. This place it used
to be a former outdoor museum along the Oulujoki River. After years it was sold to privates that
are today the owners. Buildings are in a serious state of abandonment. The vegetation has already
started to surround the architecture. Timber structures are dramatically compromised. Now there
are not any documentary actions of intervention for repairing this heritage, which is destined to
disappear if specific restoration activities will not start soon
From the first approach on this typology of heritage, it appeared clear that the
documentation of historic wooden architecture constitute today a precious testimony
not only for the understanding of specific construction techniques used in the past but
also for a sensible dissemination of a wider documentation related to social customs
and traditions existed in many different cultures.
Because for these reasons the research is proposing a specific study approach in
which multidisciplinary methodologies of documentation together with specific and
updated technical approaches are deepened in order to propose new digital methods
for the documentation of wooden heritage. The research approach intends to high-
light not only the architectural value of the cases studied but also wants to increase
and improve their cultural, symbolic and environmental importance. For this rea-
son during years of researches, the activities have included not only singular timber
monuments but also examples located in evolved contexts, places where the origi-
nal environment and townscape changed and where general functions lived sensitive
modifications.
Two main objectives settle this research: the necessity to improve and develop
updated digital survey techniques applied specifically on wooden heritage and timber
structures; the volunteer to produce detailed inventory analysis with the main scope
to increase the visibility and sensibilities for authorities, inhabitants and responsible
of these heritage in order to active specific protection programmes.
The two goals appear even more connected if we consider that the process of
knowledge through accurate analysis is in itself the first step towards a deeper aware-
ness and education for the preservation of the architectural heritage for future gen-
erations.
In addition to these aspects, it is also evident that digital survey methods and pro-
cedures are nowadays increasingly requested because for the accuracy that can be
achieved and for the level of detail that can be represented in 2D and 3D dimension.
Guidelines and methods for preservation need today to be always updated and imple-
mented, registering changes by updating the data. Evidently, this process is meant
to be entirely digital and will not avoid the utilization of the newest technologies by
using specific software of detection and representation as for instance BIM systems.
The contribution explains the strategies, methods and procedures used by show-
ing three different case studies. Naturally the knowledge of the state of the places,
the survey of the dimensional and material aspects, constitutes in these works the
first fundamental step for the conservation. Through analysis and survey operations,
the aim is to demonstrate that the value of an architecture consists not only in its
monumentality and in its richness of details and decorations, but mainly in the iden-
tification of its tangible and intangible traits. The identification of the value of an
architectural object (as well as for a group of buildings) in the history and develop-
ment of the place, can represent the only real and authentic testimony of a specific
part of the history of our culture.
1.1 The Support of European Funding Schemes
Wooden Architecture represents one of the main research topic carried out on inter-
national scale by Departments DIDA from Florence and DICAr from Pavia with the
support of important institutions and universities from Finland and Russia. In 2007,
Prof. S. Bertocci1 and S. Parrinello2 organized the First International Conference
on Wooden Architecture, defining a first important network of relations between
academy, business companies and singular experts (Bertocci and Parrinello 2007).
In 2009, a Second International Conference improved definitely the common con-
sciousness about the belief that today this vast and unique wooden heritage needs to
be documented and preserved (Bertocci and Parrinello 2009a). Expeditions to Kare-
1 Prof. Arch. Stefano Bertocci is Full Professor at the Department of Architecture DIDA of the
University of Florence, Italy.
2 Prof. Arch. Sandro Parrinello is Associate Professor at the Department of Civil Engineering and
Architecture of the University of Pavia, Italy.

lia and Northern Finland started during the same years, with the aim of survey and
study interesting examples of traditional wooden villages and specific buildings. The
approach was already at that time aimed to increase at the same time documentation
of inedited places while deepening research methods and scientific approaches.
From 2012 until 2014, the first European Funding (Seventh Framework Pro-
gramme, Marie Curie Actions People) supported the activities. “Wooden Architec-
ture” European Project3 allowed early researchers Ph.D. Sara Porzilli and Ph.D. F.
Picchio coordinated by same professors of the university of Florence and Pavia,
to perform intensive and massive survey expeditions to Karelia (Russian side). This
opportunity allowed the arrangement not only of trips and of investigations to Karelia,
but also encouraged educational activities with the organization of summer schools in
Karelia, involving international students, Ph.D. students and professors for working
all together to this topic (Fig. 3).
During the massive survey analysis several wooden heritage, villages and mon-
uments were documented. The research team implemented technical survey proce-
dures and methods of representation for the documentation of these contexts. This
contribution intends to present and explains part of the results obtained.
More recently, from September 2016, Ph.D. Sara Porzilli applied and won a Marie
S. Curie Individual Fellowship4 by presenting a research project entitled: “Preserv-
ing Wooden Heritage. Methods for monitoring wooden structures: 3D laser scanner
survey and application of BIM systems on point cloud models”. From June 2017
European Funding is supporting her researches, which are related to wooden her-
itage and the application of the latest digital methods for its preservation. Part of
Porzilli’s research includes dissemination activities and secondments to expert part-
ners chosen from the beginning in order to increase technical skills and cooperate
with a wider group of different skilled experts. An experience in which she is rep-
resenting the key-figure and link between Academy and Business environment by
using the general formula of the “teaching-and-training” method. The study is now
3 Complete data of the European Project: Seventh Frame Programme Marie Curie Actions People
International Research Staff Exchange Scheme. Full title: “Wooden Architecture. Traditional Kare-
lian Timber Architecture and Landscape”. Identification number: 269185. Departments involved:
Department of Architecture DIDA of the University of Firenze, Department of Civil Engineering
and Architecture of the University of Pavia, Department of History of Architecture and Restoration
Studies of the University of Oulu (Finland), Engineering Faculty of the Petrozavodsk State Uni-
versity (Fed. Russia). Scientific Responsible Prof. S. Bertocci from the Department of Architecture
DIDA, Prof. S. Parrinello from the Department of Civil Engineering and Architecture in Pavia.
Thanks to this research opportunity of international collaboration, participants have undertaken a
series of missions to the Karelia regions and carried out research experiences.
4 Complete data of the Marie Skłodowska Curie Project: European Commission, Horizon 2020—Re-
search and Innovation Framework Programme. Marie S. Curie Actions—Individual Fellowship.

Call: H2020-MSCA-IF-2016. Scientific Panel: ST-ENG. Duration of the Project: 24 months. Title
of the Project: Preserving Wooden Heritage. Methods for monitoring wooden structures: 3D laser
scanner survey and application of BIM systems on point cloud models”. Beneficiary: University
of Oulu (Finland). Researcher enrolled Ph.D. Arch. Sara Porzilli. Supervisor of the research: Prof.
Anna-Maija Ylimaula. The Marie Skłodowska Curie Grant Agreement protects project results.
Mitta Oy Company, official partner of the project based in Oulu, has provided the equipment used.
Fig. 3 Within the European Project “Wooden Architecture” dated between 2012 and 2014 many
students from Italy, Finland, Russia and Israel have taken part in summer schools and survey
activities carried out in Karelia. Main goals of the researcher were the documentation of historic
testimonies of wooden settlements and architecture by using all different level of analysis, survey
and representation. Observe, draw, sketch-fixing notes helped the recognition of the place in order
to proceed with the elaboration of more technical and detailed material
mainly concentrated in developing digital methods of representation, deepening the

relation between laser scanning survey of wooden structures and BIM systems.
2 The Case Studies Between Russia and Finland
The three case studies chosen are located in different areas, characterized by specific
and unique features related to their own cultural context, constructive techniques and
functions. Even if these cases are noticeably different from each other in terms of
the nature of the object investigated and location, they have all contributed to the
increasing of knowledge deepening a “technical consciousness” on how to conduct
detailed surveys on wooden heritage. In this process of documentation both environ-
ment, landscape, architecture, structures have been investigated and treated as equal
parts of a same process, aimed at the elaboration of a complete recognition of the
site.
First case study5 is the World Unesco Heritage Site of the Pogost Complex on
Kizhi Island with detailed investigations of other rural settlements located in Karelia,
Russia (Fig. 4). These surveys generated a precious contribution in the understanding
of how to handle a wide amount of metrical data. The representation of the wooden
structures in 2D and 3D for technological assessment analysis and the elaboration of
highly detailed technical drawings have represented the main goal of this work. The
complexity of the wooden structure of the Church of Transfiguration and the pres-
ence of a dense metal support structure transformed this survey into an interesting
challenging experience.6 The researches on the wooden settlements have produced
interesting experiences on how to develop digital census analysis, elaboration of
descriptive atlases and experimentation of 3D photo modelling results from the gen-
eral scale of the entire village until the description of a singular detail.7
The second case study is the historic Lamminaho Farm House located in Vaala
region, Finland (Fig. 5). The place preserves and documents the original features
of a traditional historic Finnish wooden complex dating between the 18th and 19th
century. The place was initially built as a private farmhouse and it was then used often
by anglers because for the strategic position along the Oulu River. For almost two
centuries, Lamminaho was a private property. In 1992 the entire place was donated
5 Researches carried out on the Pogost Complex on Kizhi Island were developed during year 2011.
The elaboration of results, analysis and technical drawings showed belong to Sara Porzilli’s Master
Thesis in Architecture, defended in July 2011. Title of the Thesis: “The Pogost Complex on Kizhi
Island. Laser scanner survey for the architectonical analysis of the structure of the Church of
Transfiguration”. The analysis on the rural settlements in Karelia were developed within the First
European Project “Wooden Architecture”, which has involved researchers, professors and students
thanks to the organization of international summer schools and events.
6 Responsible and Coordinator for the survey operations and post production supervision is Prof.
S. Parrinello, from the University of Pavia (Italy) and Prof. S. Bertocci from the University of
Florence (Italy). Surveyors involved in the laser scanner survey of the Pogost Complex are Sara
Porzilli and Aurora Sorini, both authors of the post production phase and of all the results obtained.
7 The studies on Karelian villages were developed during the First European Project “Wooden
Architecture” thanks to the organization of two main Summer Schools, which have contributed
with important insights the progress of the research. For a detailed study consult Porzilli (2015),
Rilevare l’architettura in legno. Protocolli metodologici per la documentazione delle architetture
tradizionali lignee: i casi studio dei villaggi careliani in Russia. FUP—Firenze University Press,
Firenze.
Fig. 4 General views and representations of the wooden settlements and architecture analyzed in
Karelia. From the upper part: the hand drawn map of the historic village of Korza located in the
Vedlozero region. Preliminary sketches of the Unesco World Heritage Site of the Pogost Complex.
Point cloud result of the laser scanning survey of Yamka wooden village with its authentic windmill.
General images of the survey activities carried out on Kizhi Island, located in the archipelago of
the Onega Lake
to the National Board of Antiquities8 who started a documentation which have had
a precious updating during the survey activities shown in this contribution. Lam-
minaho represents a unique, well-preserved example of wooden heritage, a place
where restoration and repair interventions have been undertaken by respecting the
so-called genius loci (Norberg-Schulz 1992), by using traditional techniques, which
have preserved the authenticity, without compromising the spirit of the place. Today
experts and responsible of the place9 are promoting the idea of using this place as
an open-air museum. For this reason, the National Board of Antiquities and Senate
Properties10 have commissioned an important survey including not only the archi-
tecture but also the surrounded environment. The survey of the architecture jointly
with the documentation of its environment gains a complete framework of the actual
situation of the place. The main aim has been to set up a documentary corpus for
supporting all the different technical actions. This research project has contributed
of increasing important practical strategies for carrying out laser scanning survey
of wide areas taking measurements of both the environment and the architecture
by implementing the metric database with additional survey phases along a certain
period of investigation.
The third case concerns the laser scanning survey and analysis of the residential
wooden district of Raksila in Oulu, Finland (Fig. 6). This residential area was planned
during the early century of the XX century as a regular district, characterized by
wooden buildings aligned along the main streets. The image of this place did not
changed during the years and today it almost appears in its authentic configuration.
Among this, still there are not specific guidelines of interventions and technical
recommendations for the preservation of its authenticity. In some parts it is possible
to identify inaccurate interventions and actions on restored buildings with the addition
of volumes, use of non-original materials and organization of the private yard not
totally conformed with the original plan of the place. For this main reason, the survey
of Raksila has involved the architecture and its main environment as two essential
8 National Board of Antiquities (NBA or Museovirasto in Finnish language) is responsible, together
with other authorities and the museum field, for protecting Finnish environments with cultural
historic value, archaeological culture heritage and architectural heritage. It operates under the Min-
istry of Education and Culture in Finland and it has the responsibility of Finnish properties even
abroad. The Mission statement is to recognize and nurture the ever-changing culture. “The Board
ensure the maintenance and accessibility of cultural heritage, produce and mediate new information
about cultural heritage together with its partners, and participate in social value discussion on the
significance of culture, the state of the cultural heritage, and the direction in which it should be
developed”. More detailed information are explained on the official webpage: http://www.nba.fi/
en/index.
9 Museovirasto—National Board of Antiquities (NBA) which has the main office based in Helsinki
(Finland) represents the first responsible of Lamminaho Farm House. Senatti Properties is the second
partner, which has financed the activities jointly with NBA the research activities carried out by
Ph.D. Sara Porzilli.
10 Senate Properties is a Company working in the field of work environment and specialist partner
of the Finnish government. The company has the main responsibility to manage the government’s
property assets and their efficient use. Official webpage: https://www.senaatti.fi/en.
Fig. 5 The historic village of Lamminaho has represented a pilot project for the “Preserving
Wooden Heritage” European Project. Starting from laser scanning survey activities it has been
developed a detailed scientific report with the representation and description of all the buildings
preserved in this area. A sort of technical dossier available for technicians and experts, responsible
for the preservation of this Heritage
Fig. 6 The study case of the historic district of Raksila in Oulu, Finland. Laser scanning survey,
analysis and documentation of the street facades with census activity of the architecture and its
environment. Analysis and results included in the report Özlem Özer-Kemppainen (edited by Özer-
Kemppainen 2017)
parts of the same research topic. The laser scanner survey of Raksila is part of Ph.D.
Sara Porzilli’s ongoing research project “Preserving Wooden Heritage”.
3 Research Approach and Criterions
All the research cases presented here arose from the idea that accurate surveying
operations nowadays constitute the fundamental basis for designing any kind
of architectural project and critical analysis. Innovative methods for surveying
architecture and its environment allow technicians to acquire exact knowledge of
the status of the object studied and give exact information in order to produce
effective intervention strategies. True metrical information and data are fundamental
for understanding the formation and development of an architectural entity, village,
or city, as well as for planning conservation and restoration projects starting from
the present state of a building. In order to get the maximum information from a
survey campaign, an integrated11 research project requires detailed organisation of
the different activities carried out combined with an indicative time line.
In order to obtain certain and specific results from each type of research activity,
it is always fundamental to approach the study with a careful planning of all the theo-
retical and practical phases involved. During the preliminary phase, it is important to
define the activities involved in the process, elaborate operational schemes on which
to affix notes in order to identify all the different typologies of data obtained and to
understand the achievable results. “Input-Output” schemes, tables and lists help the
researcher in the organization of the works and can give important insights in the
elaboration of cross checks and transversal implementations (Table 1).
Second fundamental aspect is the identification of the different levels of studies
according with the aim of the research. Different activities relate in fact to a different
scale of investigation and should be handled by using a process, which from the
general aspects includes progressively the details and the particular items (Fig. 7).
From the general → To the particular

Analysis of available material Goal setting (elaboration of the Structuring of activities with
main purposes of the research) an “input-output” method.
Each activity generates results
(=output) which should
represent the “input” of the
next phase
Analysis of the context Survey of the architecture Analysis of the details (detail
(territorial scale) (architectural scale) scale)
Identity and authenticity of the Analysis of the architectural Documentation of the
place values particular elements
11 The term “integrated” is specifically intended for those research projects where different types of
survey activities are carried out concurrently.

Table 1 Survey activities planning: instruments, input, results, possible crossing checks and
outputs
The necessity to go through accurate analysis is today supported by the sensible

development of digital technologies and methods thanks to with it is possible to link
for instance census activities with surveys, or point clouds with mesh models, and
again 3D models with GIS databases.
The interdisciplinary and the connection of cross-thematises support the identi-
fication of the new interesting insights and features of an entire place as well as of
a singular architecture. These features are again responsible of the identification of
intrinsic and extrinsic aspects related to the style, the authenticity, the identity, the
narrativity and reversibility, all elements, which transcend evidently above Time and
Place being responsible of the real comprehension of the object analysed.
Investigations and survey activities follow some main criterions:
• Preliminary recognition of the area, by concerning also archive investigations and
historic analysis available;
• Definition of objectives, final results and limits;
Fig. 7 From the general recognition of the place through the analysis of the singular details of a
certain building. The understanding of the investigated object require the practicing of a constant
zoom-in zoom-out analysis in order to include the detail in the over-all point of view and keep the
general aspects while detecting a particular element
• Identification and organization of the documentary and survey activities by pro-

ducing time schedules, Gantt charts and table of contents.
This type of organization appear even more necessary if we consider that different
level of analysis produce different scale of values and conclusive considerations
which contribute to the elaboration of a multidisciplinary research, that consider the
object of investigation in its whole material and immaterial configuration.
Within the field of the analysis for documentation, preservation and restoration
purposes, survey activities define the preliminary key-points, thanks to with it is
possible to increase the consciousness of the place and deepen the understanding of
the architectural structure. Each case study has represented in fact the testing ground
for the elaboration of useful principles both in theory and in practice.
4 Digital Survey Methods and Procedures

for the Documentation of Wooden Heritage
According with the objectives defined during the preliminary phase, the activities
carried out started with a first recognition of the available documentation and a second
step more concentrated in collecting updated information and metrical data on site.
A third step regards the post production activity. The first two phases include:
1. Preliminary and propaedeutic activities for supporting the survey analysis and
investigations:
• Archival researches with general understanding of the available material;
• Meetings with people and responsible of the architectural theme investigated,
or in charge with the place;
2. Practical survey activities:
• Laser scanner survey, from which obtain a complete point cloud both of the
architecture and its environment.
• Topographic survey, in order to georeferenced the point cloud and be able to
allocate different parts of point cloud in a mutual geometrically and georef-
erenced correct position. The topographic survey helps also in the reduction
of the error during the registration phase of wide areas (for example for the
case study of the Pogost Complex and the Karelian villages and for the case
of Lamminaho Farm House in Finland);
• Direct survey, by using simple tools, for measuring details and parts which the
other survey operations should be not able to surveyed with enough metrical
accuracy;
• Photo Documentation for general observations, for the photo-mapping of
facades, sections, floor plans;
• Photo modelling reconstructions, by using the possibility to reproduce from
2D photos (obtained through specific and accurate photo documentation) 3D
models with the so-called structure-from-motion process;
• Census activities, for the creation of descriptive inventories and technical
atlases of the buildings investigated;
• Additional activities: landscape analysis, studies related to environmental and
cultural aspects by using interviews and external supports from other special-
ized technicians.
Archival documentation, made by historic pictures, drawings and representations
of the Pogost Complex and of the whole island of Kizhi, gave to this case study
important insight for the understanding of the original situation both of the architec-
ture and of the landscape. During the last decades, the Church of Transfiguration has
been in fact under careful restoration and consolidation activities. The original inner
parts, characterized by the presence of an ancient wooden iconostasis surmounted
by a timber ceiling, which reproduced the vault of the sky, were disassembled and
Fig. 8 Scheme related to the architectural composition of the Church of the Transfiguration on
Kizhi Island. Thanks to the historic previous documentation it has been possible to understand and
know about the original composition of the building. Which where the main elements and how they
were located in the structure
stored in order to build a metallic structure for restoration operations. Neverthe-

less, the architectural parts supporting these structures, remained in the church as an
integral part of the main wooden structure. During the survey and analysis, it was
fundamental to reconstruct and understand deeply these parts and how they were
originally built in order to be able to carry out an accurate survey and a coherent
redrawing of the real state of the timber structures (Fig. 8).
In addition to this aspect, previous old drawings have allowed the correct under-
standing of architectural details as for example the wooden onion domes which
characterize the roof structure of the Church (Fig. 9). In other occasion the support
of preliminary drawings, although indicative, have supported the organization of the
survey.
During the survey activities on the karelian villages in Russia, a preliminary
archival documentation was on the contrary almost inexistent. Only for few cases
it has been possible to approach the research starting from some available data and
technical documentation, mainly linked to the initial representation of the settlement.
This type of information was clearly useful for undertaking comparison and dating
of buildings, identifying the origin and development of settlements, documenting the
lost and destroyed buildings and defining the additions (Fig. 10).
Both for the survey of an architecture as well as for an entire settlement the laser
scanning survey should be always linked to a topographic survey or to a GPS data set
of points. Wide and complex survey operations can produce in fact geometrical errors
Fig. 9 Constructive details related to common architectural elements present in the Karelian
Wooden Architecture. Below the structure of a traditional timber roof, below the composition
of the onion domes, one of the main architectural features present in the orthodox Russian churches
Fig. 10 Comparison and analysis of different maps related to the evolution of the Karelian settle-
ment of Korza
during the registration process.12 A topographic recognition as well as georeferenced

information can help the surveyor in decreasing sensibly the amount of error in order
to obtain a result with a high metrical accuracy. In addition to this, it can be also
12 The term “registration” represents a technical way to indicate the operation of connections of the
singular scanwords made during the re-processing of data after the laser scanning operations on
field.
Fig. 11 In the organization of wide digital survey, it could be useful to use different levels of
resolution for performing the scan words defining specific polygonals. In this image Scans 1, 2, 3,
4 have “super high” resolution and they create a close polygonal loop. On this basis, the other scans
registered in smaller groups, can be linked and find a correct mutual position. By registering each
group of scans with the main polygonal of super high scan words, each group will be definitely also
correctly linked with the other groups of scans
useful to conduct ultra-high resolution scans for connecting different areas of the
same survey. These connecting-scans have the main role to acquire enough metrical
information of the environment and architecture and be the key-connections in-
between different groups of scans (Fig. 11).
For all the case studies a precise organisation of the scan positions was defined
in advance in order to set the practical activities and quantify the amount of work
necessary in the field. This procedure is recommended because it gives a real and
practical understanding of the work and obligates the surveyor to elaborate in advance
all the necessary work strategies useful for carrying out the research and obtaining
the results. This type of method allow also a recognition of the estimated time of the
activities of field. The organisation of the scan positions is generally determined by
specific factors, for example:
• The presence of obstructions which obligate the surveyor to move the instruments
into different positions in order to avoid the possibility of missing data;
• The necessity to create the right connections between external areas and inner
parts;
• The necessity to survey the object investigated with the highest resolution, avoid-
ing, when possible, shadows and holes in the point clouds (for example with the
case of the Church of Transfiguration).
For the Lamminaho project two different types of laser scanner and software were
used and tested (the first campaign was performed in 2015 with a Leica GeoSystem
and the second in 2017 was performed with a Zoller + Fröhlich laser scanner). This
possibility offered interesting comparisons of the data acquired and better under-
standing of different procedures (Fig. 12). During the second campaign, a massive
project was developed in addition to the first one, by making more than 90 scans in
order to survey all the external and inner parts in detail. These scans were registered
with Z + F Laser Control Software giving the possibility to visualize all the scans
with a top view mode directly on the map generated and recognise each singular
scan position and navigate within it (with a double click on the blue sphere).
Same laser scanner has been used also for the documentation of Raksila district in
Oulu. In this case the object of the analysis were concentrated to the documentation
of the townscape of this particular example of historic wooden district. The survey
operation has been carried out elaborating the scans along all the main roads of the
district. The massive presence of the tree crowns obligated the definition of the scan
positions systematically during the activities. The map illustrates in fact that the
positions are organized along a zig-zag open polygonals (Fig. 13).
For the registration processes, two methods were adopted:
– use of targets: they are specific points which give the possibility to elaborate
geometrical links between different scan words during the registration operations.
The target needs to visible in all the scans involved in the process of registration.
The laser ScanStation 2 from Leica Geosystems uses generically targets which
have reflector parts, this aspect helps the recognition of the exact point from the
software used. With ZF lasers is possible to use black and white paper targets,
metallic orientable targets on tripods, or sphere targets.
– cloud-to-cloud alignment system: this method can be used only when wide surfaces
are acquired by different scan positions. The software used for the registration gives
the possibility to make a geometrical pre-alignment of two scans just by moving
and overlap a scan to another one. When the pre-alignment is done the software
can proceed with the accurate final alignment recognizing the same surface in both
the different scans. This method needs to be done with attention and only in those
situations where the point clouds have enough areas for overlapping each scan to
another one.
The two methods can be used concurrently in order to increase the level of accuracy
and for reducing the geometrical error.
Fig. 12 Representation of the scan positions for the laser scanning survey of the wooden Farm
House of Lamminaho in Finland. In red color is the first survey campaign (2015). In yellow color
the second campaign (2016) and in blue color the third campaign (2017)
4.1 Post Production Phase and Out-Puts Obtained

from the Data Acquired
Analysis and survey activities produce a large volume of updated documentation and
information, characterized by the different typologies of data:
• Metric databases, point clouds;
• Vector bidimensional drawings obtained from the elaboration of the point clouds, in
particular: general plans, environmental sections, vertical and horizontal sections,
detailed technical drawings for specific analysis and observations;
• Three-dimensional models obtained through the elaboration of the point clouds.
Fig. 13 Scan positions of the laser scanning survey of the wooden district of Raksila in Oulu,
Finland. The main aim of the survey has been the study of the streets facades of this historic place
and the recognition of the present townscape
For each case study presented detailed 2D CAD drawings were produced, using
different metric scales according to the purposes planned, as well as 3D models
and/or simulations and photomaps to represent the architecture in its real aspect.
It can be useful to summarize in detail the main purpose of each project and the
technical materials produced:
– Pogost Complex Project: Analysis and understanding of the architectural structure
of the Church of the Transfiguration. Documentation of the island and its rural
settlements for landscape analysis and for general dissemination useful for touristic
purposes too. Material produced is environmental sections in 1:50 scale, floor
plans, longitudinal and transversal sections with a 1:5 scale of resolution, 3D model
of the first octagonal basement and different tests of modelling straight from the
point cloud information. Photomaps of each CAD drawing (plans, sections and
facades) completed the documentation using photo documentation built up during
the survey campaign (Fig. 14).
– Lamminaho Project: Documentation of the entire area in terms of both architectural
and environmental aspects for preservation and restoration activities. Materials
Fig. 14 Different phases of the reprocessing data for the digital survey of the Pogost Complex on
Kizhi Island, Russia. For this projects, the post production has involved accurate elaboration of
technical drawings for the representation and comprehension of traditional timber technologies and
architectural aspects
produced: elaboration of environmental sections (metric scale 1:50) for an updated

recognition of the landscape; updated measures of the river bank area; analysis of
the relations between buildings and open areas. For each building produced: floor
Fig. 15 The case study of the Farm House in Lamminaho in Vaala region, Finland. The place
is under a careful plan of intervention and analysis for the restoration of the architecture and its
landscape. Digital survey, technical drawings by using different scales and photo documentation
have been produced in order to help experts and technicians in this important process of conservation
plans, technical drawings of the facades, longitudinal and/or transversal sections

in 1:50 scale. Each façade has a real photomap elaborated in 1:20 metric scale
(Fig. 15).
– Raksila Project: Documentation and survey of all the streets of the district. Analy-
sis of the architecture and its environment for the elaboration of updated environ-
mental sections (metric scale 1:50). Census and inventories about buildings with
detection of the main characteristics related to the main structures, architectural
elements and details, state of conservation with recognition of possible damages.
Fig. 16 Digital survey and preliminary results of the wooden district of Raksila in Oulu, Finland.
The work is still under processing and it will constitute a new updated digital documentation of this
historic area in Raksila. This case study is part of “Preserving Wooden Heritage” European Pro-ject
held by Post Doc. Researcher Sara Porzilli assisted by stud. Francesca Messeri
3D model of the whole area in order to link the information collected in the census
analysis with the singular building represented in 3D. Each environmental section
has a vector CAD drawing and photomaps of the buildings elaborated thanks to a
detailed photo documentation. The resolution of photomaps is 1:20 (Fig. 16).
In the post-production phase it is important to check each ScanWorld and clean
them one by one from different types of noise and useless element scanned, for
example objects in the field, furniture, presence of people passing in front the laser.
In addition to this, it is possible to separate the point cloud into layers in order to have
a more cleaned organization of the data within the point cloud file. The elaboration
of the sections has been done with Cyclone software from Leica Geosystem. The
surveyor create and defines the cut-planes necessary for making the sections, then
it is possible to elaborate the orthoimages and export them in a .bmp file. These
images are loaded into a CAD file and moved in their specific reference positions
thanks to the coordinates assigned by the software while exporting the orthoimage
(the reference positions and all the necessary geometrical information are listed in the
text file generated at the same time with the orthoimage). When all the orthoimages
are loaded into the CAD file it is possible to start drawing. For the 2D drawings a
technical layer list is set and defined in AutoCAD according to the main rules for the
representation of the architecture and environment (Bertocci and Bini 2012). All the
material produced needs to be archived with specific organization in order to give to
different operators and technicians the possibility to handle the quantity of data and
navigate within the database without compliance. For this reason, laboratories, units
and departments should define and agree specific storage systems for the organization
of their research work.
4.2 Final Results: Digital Documentations

and Representations
Alongside the technical-operational aspect, this experience has also involved a the-
oretical and an academic approach for the advancement of the new integrated digital
survey systems and for increasing the 2D and 3D post-production methods in order
to obtain the newest and most up-to-date procedures for systematic analysis.
It is interesting to notice that even if each case study had a specific output and
purpose, all started from accurate survey operations. The challenging aspect of this
method and procedure is being able to cover many necessities even in different scien-
tific areas while approaching the subject with these specific activities and conscious-
ness. Landscape analysis, study of architecture in its history, engineering approaches
and experimentations, dissemination needs, sociological analysis and historical and
archival studies can start from survey recognition on site and find a deep help from
the scientific sector of the survey and representation. It is certainly possible to assume
that a laser scanning survey may be necessary for not only the main reason for which
it can be requested but also crucial for future reasons that may arise later on. The
research experiences illustrated in this contribution have definitely underlined and
confirmed that today our world heritage should have as a rule metric documenta-
tion performed through a laser scanner survey, in order to generate documentation
available for the next generation.
In the specific sector of Wooden Heritage digital survey methods are giving impor-
tant insights and positive results. Because for the structure and nature of the material
it should be in fact extremely difficult to achieve accurate results from survey oper-
ations without considering the contribution of a digital survey systems.
In all the three cases showed, there has been an updating of the state of the art of
the subject investigated. The responsible of the place received not only the metrical
database of the investigated area but also the additional elaborations.
For the Pogost Complex the attention has been put on the representation of the
timber structures achieving a full description of the technologies used and construc-
Fig. 17 Post production results of the case study of the Church of the Transfiguration. In order to
understand the wooden structure nine floor plans and four sections were produced by elaborating
orthophotos by using the photo documentation elaborated on field and detailed technical drawings
tive techniques adopted. Detailed and updated drawings help the understanding of
how this magnificent building was built and how it was planned (Fig. 17).
For the Lamminaho Project the laser scanning survey has been the starting point
for the elaboration of an updated documentation, represented in form of catalogue of
the entire village. Each building has a full description both in the inner and exterior
parts. Point cloud, cad drawings, photomaps of the facades, sections and detailed
photo documentation are from now available for the responsible of the restoration
project.
More recently, the case of the wooden district of Raksila has demonstrate that even
well preserved urban areas needs in any case to be checked and controlled activating
strategies of documentation. Survey information and accuracy in the representation
of the data have represented two main aspects of the same purpose. The census elab-
orated for the buildings in Raksila has produced an important descriptive inventory
in which the main information are catalogued in digital form. This means that it will
be easily updatable in the future collecting the history of each building and support-
ing the actions of future interventions. Digital instruments and documentation will
definitely allow technicians to switch from a static method to a new dynamic and
more efficient scientific approach.
5 Conclusion
According to the state of the art related on wooden heritage and methods for its
documentation, researches carried out have increased the knowledge and the digital
archive related on this topic along with the “Wooden Architecture” and “Preserving
Wooden Heritage” European Projects. Both projects started out from the evident and
urgent necessity to keep and preserve wooden architecture by developing systematic
analysis including 2D/3D representations for diagnostic analysis and cataloguing all
the elements with census activities. According to the preface of the Conservation
of Historic Timber Structures manual (Larsen and Marstein 2016): “There are no
standard technical solutions which can be applied universally. Our experience is that
repair approaches must be geared towards the specific cultural, architectural and
environmental challenges in the country or region where the historic timber structure
is located. With this background, one of our most important tasks […] has been the
development of the Principles for the Preservation of Historic Timber Structures”.
The research experience is having the important challenge and purpose of
operating within the ICOMOS Principles for supporting technicians and operators
involved in different types of activities: restoration, documentation, re-assessment
and re-designing necessities, re-use, accessibility projects and preservation needs.
Alongside the technical and operational aspect, the experiences illustrated here have
also involved a theoretical and academic approach for the advancement of the state
of the art related to the new integrated digital survey systems for improving 2D and
3D post-production methods. All the analysis presented have been done in respect
of the Research-Theory-Practice triangular scientific approach.
This research proves that laser scanner survey represents nowadays a fundamental
documentary base that any heritage site should have and request. From a laser scanner,
in fact, many different types of project can be formulated with intentions ranging
from the simple updating of cartography and technical drawings, up to restoration
and consolidation projects. In addition, even projects related to more educational
aims for dissemination purposes such as touristic info should find a powerful benefit
from digital surveys. The results of these research activities have highlighted that the
development of intervention strategies for the preservation of cultural heritage must
today be based on updated documentation. It is evident that the careful acquisition of
data has a fundamental role in validating each decision in any sector of detection. The
importance of the documentation becomes even higher considering conservation in
its widest sense, thinking about “physical characteristics” and “immaterial intrinsic
elements”, in consideration always of the principle of the “minimum intervention
and maximum retention of materials” (Larsen and Marstein 2016). Today the main
urgent needs are related to the ascertainment that:
– Nontraditional materials are altering and undermining the structural image of
wooden heritage;
– A renewed necessity for skilled carpenters needs to be resolved in order to keep
specific knowledge alive and available for future generations;
– Traditional materials, crafts and craft techniques must be disseminated and spread
by using training programs and specific studies addressing the sustainability of
traditional materials and craftsmanship13 ;
– Related to practical strategies with laser scanners and monitoring assessments, it
would be useful to leave in place the surveyed documentary targets in order to
obtain the same metrical joint system throughout the years. Periodical surveys can
aid the technical and deep understanding of timber structures in order to check their
state of maintenance and operate with fast, safe coherent actions when needed.
The loss of knowledge of this wide wooden heritage, the disappearance of the
traditional cultural identity, the memory of these places and the loss of knowledge
of craft techniques are seriously compromising the conservation of wooden tradi-
tional architecture. The elaboration of new typologies of analysis and intervention
strategies for wooden heritage represents a strong, highly urgent necessity. Compa-
nies, academic units, and state entities should start to invest in this sector, asking for
updated documentation and financing courses for carpenters and craft technicians in
order to keep this precious part of our history of architecture alive and available for
the next generations.14
Credits
Stefano Bertocci is author of paragraphs “1. Introduction” and “1.1 The support of
European Funding Schemes”, “5. Conclusion”.
Sara Porzilli is author of paragraphs “2. The case studies between Russia and
Finland”, “3. Research approach and criterions”, “4. Digital survey methods and pro-
cedures for the documentation of Wooden Heritage”, “4.1 Post production phase and
out-puts obtained from the data acquired”, “4.2 Final results: digital documentations
and representations”, “5. Conclusion”.
Copyright
The Marie Skłodowska Curie Grant Agreement of the Researcher Ph.D. Arch. Sara
Porzilli, PostDoctoral Fellow at the University of Oulu (Finland) protect project
results of Lamminaho and Raksila surveys and results located in Finland. These two
projects have been financed by European Commission | Horizon 2020—Research and
Innovation Framework Programme MARIE S. ACTIONS—INDIVIDUAL FEL-
LOWSHIP. Call: H2020-MSCA-IF-2016. Project title of the Research: “Preserving
Wooden Heritage. Methods for monitoring wooden structures: 3D laser scanner sur-
vey and application of BIM systems on point cloud models”. Mitta Oy Company
13 These principles are already underlined and treated in the ICOMOS International Wood Com-
mittee’s Principles. This contribution aims to underline and confirm these urgent necessities with
practical examples.
14 The same principles and considerations are published in S. Porzilli, “Laser Scanning Surveys of
Wooden Heritage and Timber Roof Structures and their Post-production. Methods for Detailed 2D
and 3D Representations”, in M. Huttunen, L. Eerikäinen, L. Laine, P. Saarinen, P. Savolainen (a
cura di) “Ruotsin suurvalta-ajan vesikattorakenteet suomessa”. Grano Oy editore, Helsinki, 2018.
pp. 196–211. ISBN 978-952-7239-38-4.
based in Oulu with a “Survey Lab.” has provided the equipment used for performing
the laser scanning survey. External collaborator for the case study of Raksila is stud.
Francesca Messeri, who is developing her master thesis in Architecture collaborating
within the “Prewoodenheritage” European Project.
Some images and information are already published in S. Porzilli, “Laser Scanning
Surveys of Wooden Heritage and Timber Roof Structures and their Post-production.
Methods for Detailed 2D and 3D Representations”, in M. Huttunen, L. Eerikäinen,
L. Laine, P. Saarinen, P. Savolainen (a cura di) “Ruotsin suurvalta-ajan vesikattorak-
enteet suomessa”. Grano Oy editore, Helsinki, 2018. pp. 196–211. ISBN 978-952-
7239-38-4.
References
AA.VV (1994) The language of wood. Wood in Finnish sculpture, design and architecture. F.G.
Lonnberg, Helsinki (Finlandia)
Anker L, Snitt I (1997) Our Nordic Heritage. World heritage sites in the Nordic countries. Kom
Forlag Vagevein, Norvegia
Bertocci S, Bini M (2012) Manuale di rilievo architettonico e urbano. Città Studi edizioni, Torino
Bertocci S, Parrinello S (a cura di) (2007) Wooden Architecture in Karelia. A collaboration pro-
gramme for the preservation of the traditional Karelian timber architecture, Edifir, Firenze
Bertocci S, Parrinello S (a cura di) (2009a) Wooden Architecture in Karelia II. Timber architecture
as a phenomenon of National Culture, Edifir, Firenze
Bertocci S, Parrinello S (a cura di) (2011) Carelia. Segni, immagini, momenti, OOO Sezam-print,
San Pietroburgo, Russia
Hansen Hans J (1969) Architetture in legno. Vallecchi Editore, Firenze
Iori I (2005) Elementi di persistenza e di tradizione nelle forme del costruire. In: Bertozzi P, Ghini
A, Guardigli L (a cura di), Le forme della tradizione in architettura. Esperienze a confronto,
FrancoAngeli, Milano
Larsen KE, Marstein N (2016) Conservation of historic timber structures. An ecological approach,
Oslo
Norberg-Schulz C (1992) Genius loci. Paesaggio ambiente, architettura, Mondadori Electa Milano.
Ed. orig. Genius Loci: towards a phenomenology of architecture. Edinburgh University Press,
Edimburgh
Özer-Kemppainen Ö (ed) (2017) Historiallisen Kaupunkimiljöön Suojeluatlas: Oulun Raksila”
(English translation: Historical urban environment protection class: Raksila in Oulu) Oulun
Yliopisto: Oulu
Porzilli S (2015) Rilevare l’architettura in legno. Protocolli metodologici per la documentazione
delle architetture tradizionali lignee: i casi studio dei villaggi careliani in Russia. FUP—Firenze
University Press, Firenze. ISBN 978-88-6453-354-4 (print). ISBN 978-88-6453-355-1 (online)
Pryce W (2005) Architettura del legno. Una storia mondiale. Bolis Edizioni, Bergamo
Soikkeli A (2000a) Management of the European Wooden Building Heritage, University of Oulu,
Department of Architecture, Oulu, Finland
Soikkeli A (2000b) Restoration of Old and Modern Wooden Buildings, University of Oulu, Depart-
ment of Architecture, Oulu, Finland
Zwerger K (2012) Wood and Wood Joints. Building Traditions in Europe, Japan and China.
Birkhauser Basel, Germania
“Techos de Armaduras” in Cuban
Colonial Architecture: Cuban Coverings
from XVII to XVIII Century
Cristiana Bartolomei, Caterina Morganti and Giorgia Predari
Abstract The main research investigation deals with wooden roofs in the context
of colonial architecture in Cuba, focusing especially on civil and religious buildings
from XVII to XVIII centuries. The roofing technique treated in the present research
comes from the Cubans span—roofs called “armadura de maderas” and they origi-
nate from Spanish sloping roofs called “armadura de pares”, typical of the mudéjar
style. Thanks to the colonialization process, the experience and the cultural back-
ground in building them came directly from Spain. Like Spanish roofs, Cuban ones
hide a powerful but visibly light structural whole provided with little elements of
small size. Together with Canary Islands, Cuba recalls its incredible alikeness with
constructive solutions used in Castilla and Andalusia. It can be noticed that both in
Canary Islands and Cuban roofs there are great similitudes as well as great differ-
ences: for example, in the Canary Islands wooden decorations are often very rich,
while in Cuba is more likely to be found simpler decorative items. What is empha-
sized in the Cuban wooden roof is instead the structure, that is always manifest
independently by the type of its decorations.
Keywords Wood and geometry in the history · Roofs · Mudéjar · Cuba · Havana
1 Introduction
The peculiarities of every world society develop along the years and they get enriched
by the direct contact and the fusion with other communities. Cuba is not an exception,
at the contrary: its actual population originated thanks to this process. After the
C. Bartolomei · C. Morganti (B) · G. Predari

Department of Architecture, Alma Mater Studiorum University of Bologna, Bologna, Italy
e-mail: caterina.morganti4@unibo.it
C. Bartolomei
e-mail: cristiana.bartolomei@unibo.it
G. Predari
e-mail: giorgia.predari@unibo.it
596 C. Bartolomei et al.
extermination of the aborigines, due to the hardest conditions to which they were
continuously subjected during the colonization, Spanish people started introducing
into the island African slaves in order to use them for the heaviest jobs in sugar
industry, and some thousands of Chinese peasants for the cultivation of fields.
European colonizers successively came, especially from England, France and
Italy, more than Arabs and Jews, making the island the melting pot heart of the
process, favored by a privileged geographical position and both by the importance
of its harbor and the relevant role of its square in the whole American continent.
As well as the strong impact in terms of social identity, such fusion led to an essen-
tial influence for what is concerning arts and architecture during the colonial age:
mudéjar constructive tradition, here exported by Spanish colonizers, had a singular
evolution on the island thanks to this cohabitation, assuming its own personality. Old
Havana preserves one of the richest heritage of coverings, the so called armadura de
madera, inspired by mudéjar style in a typical Latin American city (Venegas Fornias
2003); all of these examples are characterized by a good aesthetic and manufacturing
quality but, at the same time, they suffer from a quick deterioration (AA VV 2009a).
Restoring works managed to conserve some of the roofs in good conditions, most
of which making part of historical-cultural edifices, even though those works didn’t
even consider modest edifices, where abandon and progressive substitution repre-
sent a big risk for the material witness of this heritage (AA VV 2009b). To face this
study means to undertake deeply Havana’s identity with its authenticity features and
exceptions often hidden or neglected (AA VV 2011).
2 Coverings in Mudéjar Architecture
Mudéjar architecture derivates from the fusion of Islamic and Christian culture after
seven centuries of Arab colonialism in Spanish peninsula. Spanish adjectives mudéjar
(it comes from arab mudajjan that is to say “tamed”) generally refers to those muslims
who perseverated into the Iberian territory after Christian Reconcquista, which ended
in 1492 with the fall of Granada Sultanate. Mudéjar art is therefore a typical Iberian
artistic phenomenon, generated from a specific historical situation and due to Islamic
elements persisting in civil architecture or Christian ones in religious architecture
field, thanks to cultural roots ingrained in this territory associating Muslim art with
the prevailing European style. Important exponents were both Muslim artists moved
to Spain or Christian ones attracted by Andalusian solutions, especially found in the
big constructions in the caliphate of Cordova during the X sec, in the almohadi in
Sevilla and in the nasridi of Granada between XII and XV centuries (Aranda Bernal
2011).
Mudéjar architecture was spreaded and developed thanks to many circumstances;
the economic factor is definitely essential, because artesans realized their construc-
tion with simple layouts and fronts and with wooden roofs, so with cheap and easily
available materials like wood, rocks, ceramic and plaster (Chateloin 2000). Mudéjar
architecture influenced civil and ecclesiastic buildings, intensively developing in a
“Techos de Armaduras” in Cuban Colonial Architecture … 597
characteristic way among fluvial lowlands in Iberian peninsula, where good quality
rocks were rare and clay was awesome, especially in Ebro valley, in Aragon, in some
Duero and Tago basin spots, close to Sevilla. Elements originated from Mudéjar style
in structures and decorations can be found from the end of XII century, in Castilla-
Leon country (the heart is Toledo), Andalusia (especially Cordova and Sevilla) and
Aragon.
Here the interpretation of mudéjar architecture developed in a peculiar way start-
ing from XIII century, when Christian or Muslim masterworkers perfectly cohabi-
tated and they built several churches following the Gothic Catalan typical scheme;
the cathedral of Saragozza, built in the XV century, must be mentioned, together
with Teruel and Tarazona’s one in XVI century, closing the great age of Aragonian
mudéjar art (Chateloin 2003) (Fig. 1).
For what is concerning architectonic decorations, mudéjar arts inspires itself to
Islamic ornamental repertoire turning around three essential themes: arab alphabet,
stylized vegetal elements and geometrical entanglement strongly based on math
offered wide creative possibilities.
This repertoire could perfectly adapt to any architectonic surface or others, and
it could be easily realized by means of the already mentioned materials. Islamic
and Gothic style join themselves in vegetal stylized elements, entanglements and
epigraphic motifs.
Fig. 1 Coverage of the Sala Justicia in the Alcazar in Seville

These features are especially appreciated for what is concerning wooden coverings
which are rich, various and genially set, in the decoration of surfaces with slightly
connected bricks, in the application of majolicas and other ceramic elements, in
vaulted roofs and domes of churches using plaster and chalk, in the wide usage of
mudéjar style decorated woods to build eaves, shelves, balconies, doors and shutters
(Cuza Pèrez and Rivery Carreras 2015).
Despite the greatness of possible references to this amazing form of art, specific
attention has to be paid to the realization process of one of the building’s most impor-
tant element: the roof. Within mudéjar architecture, armadura is the structure formed
by a conjunction of wooden elements put together and used to constitute the pitched
roofs of residencial or ecclesiastic buildings. The structural scheme employed for
roofings has been perfectioned over the time, reaching its maximum growth between
XIV and XVI centuries (García Santana 2009); this led to a high standardization
despite a high level of creativity was maintained during this phase, and for this rea-
son ornaments became essentials in this kind of art (Campuzano and Vallejo 2003).
The simplest form of wooden carpentry for roofings is the so calle armadura de
par e hilera, a technical solution derived from visigothic spanish building tradition,
with formal and constructive roots very close to european carpentry tradition. It is
formed by couples of reclined wooden beams very close to each other1 (pares) defin-
ing the geometry of the roof; beams are upperly connected thanks to an interposed
element, a simple board joining them (hilera). In more complexes forms, an addic-
tional intermediate element is inserted to connect the couples of beams, the nudillo,
which gives the typical A shape.
Usually positioned along the beams at 1/3 from the top, nudillo is subjected
to compressive stress; from a constructive point of view, it prevents an excessive
deformation and reclined beams instability. The connection between par and nudillo
is made through beams shaping and subsequent nailing with carvings in the wooden
part of reclined beams and giving a fork-shaped to the head of the horizontal element.
From a formal point of view the sequence of those elements then plugged with a
wooden board, internally give the roof a trapezoidal section and generates a horizontal
level (almizade or hameruelo) which is an ideal position to realize a decorative
repertoire, often made from wooden carvings: trapeze, triangles, squares and stars
are combined into space for an incredible decorative richness. Such structures allow
to build double pitched roofs, and when roofings become more articulated (like
for example with four pitches, or pavillon type, or even hexagonal or octagonal)
more wooden elements are added to define the conjunctions between pitches on the
costruction corners. When a single reclined beam generates the union between two
pitches, the covering is called de par y nudillo con limas bordon, when there are two
similar and coupled beams they are called limas moamares.
The roof structure, made by triangular elements generated from pares (with or
without nudillo), is headed on a border continuous beam defined estribo, working
like a ribbon and receiving horizontal thrusts from them. The estribo is made by big
longitudinal beams, squared and hidden, on which the extremities of pares were fixed
1 Pares’ interaxle spacing is generally equal to the double of their own thickness.
after being specifically shaped to fit. Due to its strictly structural nature, the estribo is
the most stimulated part of the whole roof and sometimes the most damaged because
it takes high oblique loads.
In order to make it crushproof under the stresses given by the wooden reclined
beams, it is implemented with transversal distanciated beams, single or coupled
(called tirantes) which absorb horizontal thrusts generated from the system. When
coupled, two beams are connected in the middle and in the extremities from transver-
sal elements, the ligaduras. In good quality roofs, the estribo is arranged right above
the tie-beams, inside proper slots into the wooden section.
Tie-beams are always in a small number and they present a bigger section than
other elements, as they must substain their own weight more than absorbing the tensile
stresses coming from pares and estribo, sometimes on wide spans. For this reason,
they are substained by wooden shelves (called canes) which help the distribution
of stresses; their free extremities are decorated with geometrical, zoomorphic and
human head shapes. For big spans, ties are coupled with elements diagonally disposed
at 45° along the wall corners, called cuadrales; their function is to connect the
orthogonally positioned estribos to each other.
The connection of the roof with perimetral walls is made through a further continu-
ous wooden ring, formed by boards or wooden beams, called solera; such connection
element between the wall and the roof works as a wall plate, fulfilling the fundamen-
tal function of receiving and uniformly transmitting vertical loads. Furthermore, it
works as a barrier preventing moisture. The installation of solera occurs on wooden
pieces transversally leaning on the wall (called nudillos themselves) functioning as
a laying flat.
The support system of the armadura is hidden by a series of boards called arro-
cabe, in its turn divided in two types: aliceres and tocaduras, generally organized
on two levels and differently decorated. Finally, the upper closure is made with a
wooden board generally covered in mug, straw and tiles.
A peculiarity that must be mentioned about this system is that it uniformly reacts
like a rigid unit thanks to its triangle shaped conformation and to the creation of
unions and constraints ensuring system continuity. The loads distribution occurs in
a uniform way along the perimeter, creating a sort of “lid” put on the upper part of
the building which avoids irregularities in the distribution of stresses of the support
structure, often made in low mechanic resistance materials like raw bricks and mug,
characterizing hispano-muslim architecture, especially mudéjar. Furthermore, the
elevate interaxle spacing among ties is a good compromise between the necessity of
absorbing tensile stresses and leaving free space to decorate the roof.
3 Comparison Between European and Mudéjar Roofings
In Europe wood was one of the first material employed in constructions, together with
stones. It was the only one which could be used interchangeably with compression,
tensile and bending stresses. In relation to woodlands spreading, wood undertook
several functions as the unique material in the building construction or as a functional

element just for coverings and floors.
Central-European prehistoric constructions were constituted by a series of wooden
poles fixed on the ground with horizontal beams substaining the roof on them. Until
the Dark Age, in Europe, there was not a clear geographical distinction in the dif-
fusion of the various types of wooden buildings; from that time on, at the contrary,
techniques started to be differentiated into wooden framework constructions, with
vertical bearing elements constituted by poles (called fachwerk in Germany, pan de
bois o colombage in France, half timber work in UK, bindingsverk and korsvirke in
Norway and Sweden) and entirely wooden constructions with horizontally positioned
tree trunks composing walls (this technique is called blockbau).
A lattice works technique, called wattle and daub, became typical of Central
Europe and England, regions where, despite not rich of conifers woodlands, defor-
estation happened more than in other regions like Scandinavia. It thereas became the
more spreaded along the whole Middle Age and continued evolving over time and
becoming during 1800s the base for the definition of ballon frame system, set up by
wooden board-made frameworks in standardized dimension nailed together.
In Mediterranean basin area, despite the lively carpentry tradition, the employment
of wood for buildings seems to have been generally limited to roofings during Middle
Age. The main reason for this process is certainly spotted in minor availability
of wood with respect to Northern Europe and the subsequent necessity to employ
different materials where permitted.
For what is concerning vertical elements, mainly subjected to compression
stresses, walls made of stones and bricks perfectly fulfill this need. In the attempt
to reproduce the same kind of roofing techniques employed for wooden framework
solutions with rafters departing from the perimeter wall to unite on the top, the prob-
lem about the horizontal thrusts generated by this static system comes up. In contrast
to what happens with wooden frame solutions, where the high slope of the roof tends
to reduce the thrusts and material elasticity-together with the settle of a supporting
plate for roofing graft-avoids excessive structural deformations, structural elements
realized with unelastic material can’t contain the horizontal thrusts generated from
roofing.
The easier solution is to arrange an element dedicated to this function between the
walls, assuming the role of tie-beam and avoiding deformation and overturning of
supports. This is the birth, in Roman times, of “capriata”, a flat reticular truss undoing
horizontal thrusts thanks to its triangular shape, where the horizontal element (the
tie-beam) suppresses the stresses of reclined ones (called rafters) (Munafò 2002).
The truss works on the principle of triangle distribution: the cohesion among its
elements is fundamental, with rafters and tie-beam properly shaped to be fixed with
each other and strenghtened by several iron brackets joining them.
The construction conception of “capriata” reached its climax around IV century
a.c and it was massively employed during Early Christian age as a roofing technique
for the first Christian churches. During Romanesque age, despite the preference
toward the usage of vaults for monumental architecture, the truss equally dissemi-
nated around medieval Europe, giving birth to several variations. In Mediterranean
regions the most essential style was perfectioned, and in Italy the usage of truss con-
tinued to be employed for big monumental churches. In continental Europe, more
complex and hyperstatic trusses were used, while in Northern Europe tridimensional
systems with very pitched roofs and useable roof-spaces spreaded widely. Used as
the basic element for pitched roofs, the “capriata” is constituted by a flat reticular
truss vertically positioned; it works only on the specific plane where it’s arranged,
vertically transferring to the supports the stresses coming from the covering. Other
elements are responsible for the superficial distribution of the roofing in space, gen-
erally horizontal purlings laying on the trusses. As it transmits only vertical actions,
in the construction tradition it is leaned directly on the walls without any connection
joining them between each other in proximity of the truss impost.
The substantial difference between truss coverings and mudéjar ones is precisely
in their static conception: the first is flat, and the second is spatial. In the spatial
solution generated by the armaduras, the containing effect opposing to the horizontal
thrusts is however given to tie-beams, even though they’re not singularly distributed
in any pares couple, but they are arranged in specific positions. The estribo provides
the connection and collaboration between elements subjected to tensile ad bending
stresses. It is a kind of perimetral beam which solves the main function of receiving
pares horizontal thrusts and nullifies them through tie-beams holding. A mudéjar
covering can’t work out if it is spatially discontinuous and if it doesn’t develop itself
along four or more sides, for the necessity of closing the estribo upper ring; at the
contrary, two developing sides are enough for reticular truss as they are necessary to
lay down the trusses.
Certainly there is no direct connection between the birth of mudéjar carpentry
and the wooden truss2 one, but they can represent two different solutions to the same
problem related to the containment of horizontal thrusts coming from pitches. Despite
the great employment of truss solution in Italy, some examples of roofings showing
some analogies with the already described spatial system can be found. They are
all placed in a single territory which is the actual Veneto region, and they belong
to the kind of roofings called “ship’s hull”, existing in fine ecclesiastic buildings in
Venice, Padova, Verona and Treviso3 (Piana 2000). The peculiar solution adopted for
roof structures allowed the construction of reclined wooden ceilings, whose shape
recalls the profile of an overturned vessel, even though it doesn’t have any structural
value: it’s a wooden covering made of boards fixed on ribs and substained by the
bearing warping that’s often different from traditional truss solution. The overall
structural conception is different and, in particular, it is for Santa Caterina Church
in Mazzorbo, where the overall system is solved with ties-beams more distanciated
than rafters designing the pitched roof. The large spacing of the tie-beams solve the
exigence of composition: tie-beams are distanciated to avoid an excessive density in
2 Gómez-Moreno, Arts and Architecture’s historian, considered the origin of mudéjar carpentry to
be found in Almohadi’s culture, as a possible interpretation of wooden roman carpentry.
3 In Venice we can find steamlined ceilings in Santa Caterina di Mazzorbo, San Giacomo dall’Orio
and Santo Stefano Churches; in Padua in Eremitani’s Church, reconstructed after 1944 bombings.
In Treviso, we can mention San Niccolò Church and in Verona San Zeno and San Fermo Churches.
internal spaces and to show the ceilings. This is a roof solution employed in Italian
territory resembling to the spatial conception typical of mudéjar architecture.
Therefore, there is no a clear link with the european context; so mudéjar art
experts thought there could be a derivation by the islamic culture, even though it’s
not actually proven because in those territories the most used model is the truss one.
What was generally realized with wood was the internal hull of the domes, and the
positioning of the decorative repertoire in mudéjar art certainly derives from this
practise, reaching a whole covering of the internal ceiling surface during its mature
phase.
Cuban armaduras show some peculiar features which are not found in any other
country, clearly connected with mudéjar roofs thanks to the privileged relationship
between Cuba and Andalusia and the mastery that spanish artesans brought on the
island when they moved.
The whole South America presents wooden roofs with similar characteristic,
even though the most similar to Andalusian ones are still Cuba and Canary Islands
ones. Most of the experts believe that Cuban armaduras directly originated from
Spain and then they developed independently untill creating their own constructive
tradition. Other academics underline the influence that could come from Canaries,
which still conserve relevant armaduras examples from XVI century on, because a
lot of Canarian carpenters emigrated to Cuba in XVIII century.
Despite the fact that Cuban and Canarian armaduras are very similar, it must be
underlined how Canarian decorations are generally more abundant; one of the most
relavant examples is the principal Church in El Salvador, founded in Santa Cruz de
la Palma in 1500, which is spectaculary covered both along the aisles and chapels,
entirely painted and decored with geometrical 8 and 12 sides wheels.
As in Andalusia, here the limas moamares solution is frequently used, totally
absent in Cuba. Such difference is essential if we consider that this implies great
discrepancies in the structural and assembling solution of the roof, because the latter
allows to build independent pitches, in contrast with the lima bordon solution.
4 Main Features of Mudéjar Architecture in Cuba
Back to the XV century, as the Spanish conquerors anchored on its shores, Cuba was
an unscarred island whose territories were inhabited by the aboriginal natives called
Aruaco. The most developed aboriginal tribes used to employ simple and rural archi-
tecture strategies to build their villages, only by assembling tree trunks to work up stilt
houses and hut. In 1513, the cities of Asuncion and San Salvador started their raise
as the very first cities in Cuba to be constructed, along with the Santísima Trinidad,
Sancti Spiritus e San Cristóbal de La Habana in 1514, and Santa María de Puerto
Príncipe e Santiago de Cuba in 1515. These settlements were inhabited by groups
of white colonists which forced the aboriginals to submit to the rigid free-labour
working system, known as encomienda. In XVI century, agriculture and livestock
breeding started to develop as minor working sectors, despite mines extraction was
the leading sector of the island economy.
In the next century, La Habana began to witness to the construction of several
fortresses, such as La Fuerza, El Morro and La Punta, along with the castle of San
Pedro de la Roca in Santiago de Cuba and Jagua castle in Cienfuegos, designed by
the most popular master builders and engineers of all Europe (García Santana 2012).
At that time, the sugar and tobacco production underwent an economic boom that
was expected to meet the foreign market growing demands; at the same time, the
strong development of the wood manufacture industry came along to supply wooden
material to assemble furniture, building and ships.
Thanks to an efficient trade network, Cuba could be directly connected to Europe
and, obviously, to Sevilla where the mudéjar style was originally born; and it is
through this trading route that millions of migrants carrying their culture and tradi-
tions moved form Spain and settled down in Cuba. The port of Sevilla was the main
boarding point to sail to America and, at that time, the city hosted the Chamber of
Commerce. Engineers, artisans, chisel sculptors and construction materials departed
form Sevilla to be sent to Cuba and employed there. During XVI century, trading
ships departing from all around America used to stop to the port of La Habana and
then, they continued their sailing towards Sevilla, rewarding the Cuban town as the
most important communication knot between Europe and the New World (Roig De
Leuchsering 1963).
An efficient trading route and an enormously productive wood industry were two
of the main reasons why the mudéjar architecture started to be implemented, thanks
also to the architecture knowledge that the colonists heirs, who were likely to be
artisans themselves, learnt and brought over; in fact, they were able to reproduce the
widely popular European constructive schemes on the Cuban territory. This building
design could be imposed with no oppositions due to the lack of a competitive building
market and due to the fact that the island had been inhabited by aboriginal natives
and their huts, ever since.
From XVI to XVII centuries, a slow and constant cultural development paved the
way to more complex architectural solutions, one of those is the construction of the
very first masonry houses which usually had a square base and a patio in their centre,
as to follow the Muslim tradition (Weiss 2002); they were two floors high and the
ground floor was used for warehouses and shops. The construction materials were
really simple such as earth, mud and bricks; the wall were usually worked up with
raw earth alternated with bricks and stone pillars.
The armadura de pares was the widest used construction solution from the very
beginning of the colonist period up to its end in XIX century, and it served to create
roofs for both religious and civil buildings; with their pitched roofs covered by the
typical creole tiles and their façades, this buildings offered a real manifestation of the
peculiar aspects of a colonial town, while from the inside they offered a real glimpse
to the perfection of their wooden ceilings (Prieto Herrera 2006). These armaduras de
pares were meant to define both the urban asset and the architectural spaces and they
represent the first aesthetic ambitions of Cuban architecture and the development of
the carpentry work (De Las Cuevas Toraya 2001).
It is possible that the armadura method had been already employed during the
XVI century although there are no material evidences that can confirm it. Nowadays,
the oldest existing roofs can be considered the one on top of religious buildings,
which are less likely to undergo architectural modifications than residential build-
ings; in fact, frequent modifications in their height and base were the cause of the
removal or substitution of the original roofs. This constructive tradition continued to
be employed up to XIX century, until the island deforestation and a new neoclassic
aesthetic trend marked its end. Since the second half of XVIII century, the Baroque
style had been in use as the major construction style for official and religious building
until its replacement with the eclectic style-composed of an abundancy of heteroge-
nous shapes-in XX century (Prat Puig 1947). The mudéjar style was then labelled as
a lower class architectural tradition.
5 Declination in the Constructive System of Cuban

Colonial Architecture
In traditional Cuban architecture, characterized by residential and religious buildings

dating back the colonial age (XVII–XVIII century), the term armadura de madera
represent a particular construction technique employed in pitched roofs with entirely
wooden structure, composed by different load bering elements working unitedly to
transmit them to vertical structures (AA VV 1998).
It is clear that the birth of armaduras de madera is directly connected with mudéjar
architecture. In Cuban roofings, construction techniques are very close with the
already described Spanish ones, with pitched roofs par de hilera and par y nudillo
made, with squared or polygonal floor plan variations. Par y nudillo is the most
diffused solution in Cuba, built on squared or rectangular floor plans even though
some perfect constructions with a polygonal base are found in this area.4 The roof
is generally organized in four pitches, lima bordon solution is the most used in this
case, easier and purely functional. Tie-beams can be singular or coupled, on more or
less wide shelves, with or without cuadrales in the corners of the area to be covered
(Weiss 1978).
The most employed type of woods in Cuba were the cedar for boardings, acacia or
jucaro for structural elements; the connections among elements were achieved with
different kinds of iron nails and with wooden chocks in some particular cases. The
type of material of the bearing walls, its quality and width are strongly connected
with the dimension and the kind of armadura. In Old Havana the most ancient walls
are made of mud, rarely in irregulary shaped stones; just from the XIX century, bricks
appeared to be used as a construction material (Sánchez et al. 2010).
Cuban armaduras hide a complex structural assemblage, just like the Spanish
ones; only the pairs of inclined beams very close one to each other and with a small
section are left in sight and they become the protagonists of the Cuban system.
4 With regard to this, we recall Santo Domingo Church roofings in Guanabacoa.

The size of rooms to be covered, conditioned by the specific building function

and, for what is concerning residential edifices, by the number of inhabitants, defines
the dimension of construction elements to be employed in the process. Generally,
span goes from 3.50 to 6 m; exceptionally spans reach 10 m in some of the main
rooms located in luxurious abodes and in ecclesiastic buildings. The dimensions of
pares are about 12 × 15 cm to cover 6 m span, but if the dimension reaches 10 m
then the dimensions rise to 14 × 17 cm. The dimensions of nudillo are similar.
Matajuntas are a further element which is always visible; they are wooden 5–7 cm
slats hiding connections between boards constituting the roof frameworks; more
than for this functional motivation, their presence generates a spreaded connection
between every pares triangle with adjacent ones, favouring the collaboration in the
plan of the roof and reducing instability phenomena.
The main feature of the armadura is the existence of high horizontal thrusts,
which must be opposed to reach the proper duration and efficiency of the roofing.
These horizontal thrusts are bigger when covered span increases and when the pitch
decrease. The roof pitch goes from 50 to 70% and this allows to reduce horizontal
thrusts generated by coupled beams. On the other side, a very high pitch could trigger
the shifting of covering tiles, requiring a proper vinculation for them.
On top of the wall, the solera is always present; it is a board having 40 cm of
width and 5 cm of thickness covering the whole perimeter of the internal walls and
functioning as a wall plate to uniformely distribute armadura’s loads on it. It is often
protruding for 5 cm from the wall line, and it is there graven with a gridded molding.
Departing from it, boardings composing the arrocabe are useful to hide the structural
system; every level of the arrocabe level is surmounted by a frame called tocadura,
and on top of it we find an enclosing table shaped on the counterform required by
the section of the pares.
Estribo is another crucial element in the system, a beam having a great section
where the extremities of the pares, properly shaped, are fixed. It works as a perimetral
beam and it presents a 30 × 20 cm sections in residential buildings and a 35 × 30 cm
section in ecclesiastic ones, because its function is to limit the significant horizontal
thrusts generated by the pitched roof. To reduce the deformations of the estribo, it is
blocked by transversal tie-beams and on the corner, by cuadrales, both subjected by
tensile stresses. These two elements have about 15 × 20 cm section, or even more.
The estribo is hidden by a series of boards composing a framework, usually decorated
and called arrocabe.
For what is concerning luxurious roofings it is very common to find coupled tie-
beams. The origin of this choice is not clear: it could derive from a kind of precaution,
to avoid the possible structural collapse of one of them, or the habit to use the laceria
as a systematic decorative element, compelling to duplicate them.
When tie-beams are singular, they are closer. In order to grant the best system
functionality, the construction practice fixed the sides of an octagon inscribed in the
room as the interaxle spacing between tie-beams, so to define also the dimension
and position of cuadrales, when existing: in this way, all the estribos works at the
same conditions. Naturally this rule can’t be strictly applied every time, because
it often happens that the dimension of the ambient that has to be covered doesn’t
Fig. 2 Optimal ties’ distribution
allow to divide the sides lenghts in a way that would match with the side of the
octagon inscribed in the short side; but they are still useful as a reference to adapt
the distribution of ties-beams (Fig. 2).
A peculiarity of Cuban armaduras, never found in Spanish mudéjar roofings, is
that shelves aren’t just in correspondance of tie-beams and cuadrales, but they are
always in a greater number and they are symmetrically distributed along the wall
perimeter, in different positions, such as in the corners. The reason of this kind of
solution is unclear, but it may essentially have a building meaning. The assemblage
of the roof needs that the estribo is inserted on the heads of the tie-beams which,
as already described, are located above the shelves. For very long walls, this is the
only position where the estribo can be interrupted, to let another wooden element
pass through from one side to the other without losing stability and continuity. The
positioning of a big number of shelves, even without tie-beams on them, could
descend from the necessity to generate further break points of the estribo, maybe
due to the impossibility to find adequately long trusses in the proper section. The
finality of the free tie-beams shelves, included those located on the corners, would be
to allow the usage of quite short pieces of wood beams for the estribo, guaranteeing
their stability and the impossibility to trigger deformations or torsional movements.
A great number of roofs comes up with a real profusion of shelves, joining together
the essential bearing function with the decorative one. The building located in Tacòn
4 is an extraordinary example of what described above, with very closely arranged
shelves and one of the richest ornaments among Cuban armaduras.
Some differences in the organization of armaduras can be pointed out in relation
to their construction period and to the prosperity of the building. The main difference
is the presence or not of shelves substaining ties-beams, and consequently the number
of layers composing the arrocabe.
The simplest kind of armadura has no shelves and the arrocabe is on one level;
tie-beams are positioned directly on the solera and above them the estribo receives
the pares. The weight of the roof itself stabilizes the tie-beams, and loads are directly
distributed on the walls to the solera through them. This kind of solution has been used
for humble houses but also in ancient quality homes, like Obispo’s house 117-119,
probably the most ancient in La Habana dating back 1594 (Menéndez García 2007).
It is two floors high, with an internal yard, and it has adobe bricks walls (a mixture of
ground and calcareous stones) joined and reinforced with lime mortar and wooden
trellis.5 Its armadura de par y nudillo has four pitches with lima bordon, rectangular
harneruelo, a central single tie-beam and cuadrales located on the corners of the
room. This is a very simple solution, with the tie-beams laying directly on the solera,
which is not visible, and the estribo positioned above it and covered by a simple
undecorated wooden frame. The only decorations are fine lines carved on the lower
part of pares and matajuntas.
Inside the same building, there are also different arrangements, probably because
following intervention changed some rooms during the life of the building. The
roofing represented below belongs to same type, that is to say with no shelves, but
it has more decorations on the ligaduras of the coupled central ties-beams and on
the insertion of the pitches with the horizontal plan, where there is a decorative stars
textures. It is clearly elegant and restrained, but it proves that this was probably one
of the main rooms of the building (Fig. 3).
In a further type of solution, shelves are presents, but the arrocabe is arranged on
one single level, like the previous one; tie-beams and shelves are located at the same
level, on the solera. Those kinds of solutions are the most spreaded in bourgeois
houses, more or less decorated; among the simplest, the House in Tejadillo road n. 9
and the House of Asia must be mentioned, both with armadura de par e hilera, with
coupled tie-beams and ligaduras without decorations (Franco Rubio 2009). Among
the richest, there is the armadura de par e hilera with coupled tie-beams decorated by
geometrically adorned ligaduras in Wilfredo Lam’s main lounge, and the armadura
de par y nudillo in Al Medina’s restaurant, both painted in greenish shades.
Still in Obispo 117-119 Home, such type of roof is present, with one-level arro-
cabe and strangely without tie-beams, probably thanks to the limited dimension of
the room, which doesn’t generate much horizontal thrusts on perimetral walls.
About the more articulated solutions, many shelves, both singular or double,
appear with two or more levels of the arrocabe; this is typical in recent luxury
homes and ecclesiastic buildings. The solera supports the shelves and tie-beams and
cuadrales lay on them. Shelves are positioned in correspondance of ties-beams, as
it always occurs in Cuban roofings, in the middle of the walls, on corners, always
maintaining the roofing simmetry. Since 1626, casa de Pedroso is one of the oldest
5 The edifice was built during XVI century’s second half on the account of Don Antón Recio,
General Prosecutor and Major of Municipal Council in La Habana. Despite changings over time,
the house maintains its original structure, reason why it is considered the oldest house in La Habana.
Originally it was one floor high, then in XVI century another level was added with masonry walls
and a red roof tiles.
Fig. 3 Casa de la calle Obispo No. 117-119
buildings with such system, in Baratillo road 101.6 During more than three centuries,
the building suffered heavy changings, losing the majority of its original elements.
The decorations of the ceiling are lost, but the elegant armadura de par e hilera,
completely left on sight, is still admirable with its coupled tie-beams connected by
simple unadorned ligaduras. The only existing decoration is a mild scallop on the
boardings of the arrocabe. Inside the same building, in other rooms there are roofings
belonging to the cathegory with tie-beams and shelves on the same level, to testify a
probable previous building phase, less transformed if compared to the other (Fig. 4).
The residence located in calle de Amargura 65 is more recent; this road is one of
the oldest urbanised area in La Habana. The earliest documentation on the building
go back to 1731, when the construction of a colonial one-floor house is certified. In
the middle of XVIII century Miguel Ciriaco de Arango bought it, and made it bigger
and higher, transforming it in a two floors house. The existing par e hilera roof is
tracked back to middle XIX century. As time went by, the construction was even more
modified and splitted into several apartments. The Oficina del Historiador decided
6 The original access stood on old San Pedro street; only in the second half of XVII century it moved
on Baratillo road, where the trilobal arch of the entrance was built.
Fig. 4 Casa de la familia Pedroso
to start a restoration plan in 2001 due to its advanced degradation status, preventing
the loss of this precious testimony of colonial architecture (Chateloin 2008) (Fig. 5).
The roofs of Santa Clara Monastery (Arduengo García and Cruz Pérez 2012) are
strongly impressive thanks to the big dimension of the rooms. It was the first female
monastery built in the city between 1638 and 16437 ; the armaduras de par y nudillo
have coupled tie-beams, singular shelves and elegantly decorated arrocabe with jags
and circular carvings; some particular details are unique for this roofing, such as
some stiffening elements below the harneruelo constituted by horizontal wooden
pieces of beams, thick as the pares and arranged to connect between each other two
adjacent couples of pares, creating a sort of reticular truss increasing the rigidity and
bonding among reclined elements (Fig. 6).
7 The monastery has been strongly transformed over time, extended and elevated by adding a second
floor. It was abandoned in 1922 because of its decay. It was re-opened for public events celebrations
after its restoration in 1923. Nowadays we find two yards out of three completely renewed: one
is hosting the Centro Nacional de Conservación, Restauración y Museología (CENCREM), the
organization in charge of restoring different manufactures coming from all over Cuba; the other
containing a small independent building managed like an economic hostel in mudéjar style.
Fig. 5 Casa de la calle Amargura No. 65
6 Mudéjar Decorations in Cuba
Mudéjar architectural decoration is widely known as one of the most beautiful in the
world art, particularly for its abundance of details, the large number of ornamental
elements, the remarkable polychrome pattern and the geometric lines and shapes
which cover the entire building. In terms of decorative features, mudéjar architecture
is known for its different distinctive elements that involve bricks and rocks carv-
ing, which are sculptured in fine arabesques, for the enamel decoration of ceramics
employed to create frames and pavement, for the chalk manufacturing for cupolas
and geometric pattern paintings (Weiss 1972).
In Cuba, the mudéjar style focuses more on aspects related to construction rather
than decorative ones, marking its difference with the Spanish tradition; in fact, mudé-
jar roofs in Cuba are far less decorated than Spanish ones dating back to the XVI
century whose decoration tends to completely hide the building structures, whereas
the Cuban case aims to maintain the structure visible, simple and functional (Bauer
Arnold 1999).
Fig. 6 Convento de Santa Clara
Cuban armaduras have their own identity thanks to their clear structural language
and its strong relation with their decoration; the simplicity awarded to the structure
is given by the importance of the arrocabe, the presence of single or double shelves
located below the cuadrales, the moulding clearness involving the inferior part of
the elements and the moderate use of lacerias.
Geometric shapes are the most typical and ancient decoration of the armaduras
that are visible on the buildings of La Habana Vieja, although some ornamental
templates may appear with flower or figurative shapes. Other decorations that may
appear quite often consist of simple carvings, used to draw single or double lines
along the hedge of the inferior surface of the pares, tirantes, cuadrales and nudillos,
geometrical shapes, i.e. broken continuous line or concentric circle that may be
also discontinuous, parallel or crossed, dots and more rarely sgraffito decorations
(esgrafiado).
Other elements that are more likely to be decorated are shelves and ligaduras; the
latter usually shows lacerias decorations, twines and tapered polygons between the
tie-beams.
Religious buildings usually present several decorations in the harneruelo that may
show paired nudillos as the ones visible in Casa Obispo 117-119. Painting decoration
is not that frequent; however, it is possible to observe some painted ornaments in the
octagonal roof of Santo Domingo’s church in Guanabacoa.
What enhances and characterizes the decoration of these interiors is certainly
their bright chromatism consisting of a range of blue, green, caramel or beige; these
colours have been employed in later stages, replacing the traditional use of wax layers
and flax oil.
Abundantly decorated ceilings are certainly the one existing in residential build-
ings, usually owned by upper class families, or in religious ones; depending on their
size, the ornaments might be executed in par y hilera or par y nudillo, even though the
biggest buildings are generally more likely to be covered by par y nudillo decorations.
On squared structure, the armadura might have up to eight pitches which make
it similar to a wooden cupola, as it is possible to observe in Casa de Tacòn 4. The
latter is definitely the most widely decorated building of La Habana Vieja, though the
simple blue chromatism of the whole site. A plethora of shelves which curiously hang
down the tie-beams of the main hall just for decorative purposes, matches perfectly
with the tight geometric carves pattern executed in the wooden beams (Figs. 7, 8, 9
and 10).
7 Survey and Structure Form Motion of Cuban Roofs
Survey and representation are referred as documentation, measurement, analysis

and restitution activities; these are fundamental instruments to understand on the one
hand, the traditional Cuban wooden roofing procedure with its constitutive elements,
and on the other, the knowledgeable methodological base that is necessary to manage
and enhance the potential value of these buildings (Bartolomei and Ippolito 2014).
To fulfil the objectives coming from analysis, knowledge and preservation of any
remarkable architectural construction, the study of all aspects of any single element
that composes it is mandatory; thus, the high-priority goal of this research is to
elaborate a useful procedure that can be used in both the study and the documentation
of this type of buildings (Bianchini et al. 2014).
To represent such peculiar objects, it is necessary to be deeply aware of the con-
struction techniques employed during their construction in a definite historical time
and, more specifically, what peculiar construction techniques were employed in Cuba
in the same historical time. After a historical, typological and construction based
analysis, it has been necessary to arrange measurement ad hoc to elaborate a proper
documentation model that could be analysed ex-post and to confirm the hypothesis
that had been previously formulated.
The investigations have been conducted mainly to establish both formal and
dimensional configuration of the roofs. The formal configuration is the specific
knowledge of size and geometry of the pieces composing wooden element and it
is crucial to understand its functioning. To be more specific, this research zoomed
Fig. 7 Casa de la calle Tacón No. 4
a major focus on some buildings in the core of La Habana Vieja: Casa de Obispo
117 and 119, the Archaeological Museum located in Calle Tacón, the Monastery of
Santa Clara and the Museum of Colonial Arts.
The structure from motion is estimated to have been the most adequate procedure
and the best able to fulfil the purposes that the documentation model required (De
Luca 2010). Digital photogrammetrical instruments, projective geometry laws and
new mathematic algorithm such as collinearity equations and some epipolar geom-
etry principles, allow to generate metrically accurate texturized 3D models starting
from two images, whose shooting conditions are well-known (De Luca 2011). The
ability of retrieving three-dimensional data like metrical and colorimetric coordinates
thanks to pictures, is specifically linked to the automatic-individuation software abil-
ity that recognizes the common traits of model and pictures by moving from different
shooting centre.
The evolution of elaboration system and shooting cameras, the development of
modern feature extraction algorithms and the high-speed picture analysis system,
guarantee now a fast and automatic features extraction followed by their unambigu-
Fig. 8 Example of assembly par y nudillo
ous classification. It is still necessary to refer to a wide number of frames to create

3D models of surfaces where blind spots are reduced to the minimum.
The camera used is a Reflex, because generally, this technology has an advanced
image detector system among which the biggest is a full-frame, one that measures
24 × 36 mm and it is similar to the image detector of Reflex film cameras. In regular
Reflex cameras, the APS-C reduced frame is widely used and it is available in different
Fig. 9 Detail of the match between nudillos and pares
versions that changes from company to company, specifically documentation cameras

by Canon measure 22.2 × 14.8, measure 23.6 × 15.7 mm for the ones of other brands.
Two aspects that must be considered while choosing the instrument for structure
from motion are the control shooting and the lens of the camera; Reflex cameras are
the best option on this regards, provided that the control shooting can be used both
manually and automatically and the lens is easily replaceable and adaptable to the
different documentation conditions.
In the present case study, the employed camera is a Digital Reflex Canon EOS-
600D fully equipped with CMOS1 image detector of APSC size, DIGIC 4–14 bit
processor, 5184 × 3456 pixel of maximum resolution countable as 18mpx, EF-S
18-55 IS II lens, SD, SDHC e SDXC card support and picture detector that simulta-
neously captures the image in both jpeg and raw formats. There are two types of lens,
specifically a fixed focus and a mobile focus one (i.e. zoom); in Reflex camera is pos-
sible to assemble together both lenses at the same time. Digital cameras allow to use
both optical zoom and digital zoom but, it is necessary to acquaint the correct func-
tioning of any component of these cameras to issue valid documentation images,
such as the exposition depending on the shutter opening and speed, white-colour
balancing depending on the scene brightness and the detector sensitivity.
Structure from motion made compulsory the choice of the most adequate type
of pictures to use during the procedure, in fact the most important characteristic of
a structure from motion picture must be the correct line sharpening because fuzzy
Fig. 10 Different types of armadura de par y nudillo
images reduce considerably the possibility to spot the common traits which are crucial
to create the 3D model (Gaiani 2015). In case of fuzzy pictures, the selection mistake
usually overcomes 10px. Picture sharpening generally depends on two main factors
which are the focus and the shutter speed; a longer shutter speed time, caused by a
simple hand movement, may be the cause of fuzzy images. In case of architecture-
driven images, it all depends on the ability of the photographer and, by all means,
it is highly recommended the use of tripods and the activation of shooting timer or
auto shoot to avoid any vibrations that may interfere during this process.
The filling speed for free hands shootings has to be properly settled for a value
which is superior to the contrary of the value of focal lenses. For example, if we have
a 50 mm focal lense the suggested filling speed is major or equal to 1/60 of a second.
Another essential characteristic is the contrast, because the post-processing phase
of the images allow to recollect more details on well balanced pictures. Infact over-
exposed areas or shades doesn’t allow the proper identification of every point. The
contrast level depends on several factors: environmental lightnings conditions, bal-
ance between filling speed and diaphragm opening. Lightning conditions is another
critical point that we found during the project, it is necessary to efficiently reduce the
problem of shades which could damage the good extraction of the textures and colors
of the material. The camera was settled with closed diaphragm to have more field
depth, low ISO to reduce noise, consequently calculating the exposition and manu-
ally settling the focus to infinite. The pictures were shooted by putting the camera
on the ground, being careful to position the gripping points on three parallel tracks
and diagonally orienting the camera to take as much as possible every level of the
object.
Furthermore, for what is concerning this specific case, it was necessary to carry
out the shooting thanks to the help of scaffoldings so that the operator could grasp
every point of the element articulated on more levels, some of which dominants. This
occurred because we dealed with architectonical objects located in a high position.
The definition of those levels helps the determination of the number of shoots to be
taken more than their orientation and position.
Every dominant level must have at least two pictures taken in an oblique position
with respect to itself. Additionally to it, to grasp the texture of a surface it is proper that
the photographic collimation axes is the more perpendicular as possible to the surface
that has to be mapped. Depending on the type of object that must be restituted and on
the aims of the representation, different techniques can be chosen or combined. The
methodologies here employed for the shooting were the one with paraller axes, where
the optic axes is always oriented towards a specific direction, and the panoramic one,
where portions of shooting of spherical pictures is carried out from separate spatial
positions, executed by carrying the camera around its own focal point. This allowed
to achieve images with a good overlapping standard, more than 30%, which had very
homogeneous taking-condition.
Agisoft Photoscan is the program used to process images; it consented to generate
a polygonal model geo-referenced and texture in HD. Through feature extraction
and matching operations, that is to say the recognition of Tie Points (TPs) which
are common to more images, and their gathering thanks to SIFT (Scale-Invariant
Feature Trasform) algorithm we achieved a threedimensional model composed by
Tie Points, useful for the orientation (sparse point cloud) to whom followed the dense
matching: during this phase the algorithm allows to generate a points-made model of
the interested object called dense point cloud, once analyzed orientated photograms
Fig. 11 Structure from motion of the coverings of Casa de la calle Obispo No. 117-119
with an established field partition (Styliani 2016). Then the scaling and referencing
of the model was done by means of measurements or known points recollected on
field, with mesh creation and projection of HD oriented pictures on the achieved
mesh model (Figs. 11 and 12).
The creation of a kind of model which is measurable and navigable is clearly a
great advantage not only to document and monitorate such precious buildings, but
Fig. 12 Structure from motion of the coverings of Casa de la calle Tacón No. 4
also to reconstruct their state in case of natural disasters like typhoons, so common
in Cuba.
With its application methodologies scientifically consolidated, direct survey con-
stitutes the most adequate tool to satisfy such finalities. The choice to employ image-
based techniques during the architectonic survey of cuban wooden roofings led
to several advantages, speeding up the time taken to recollect data and giving an
unprecedented documentation on it. Other surveys of such buildings are nowadays
non-existing and some existing drawings about them are by now old and unreliable.
Survey operations on cuban roofings allowed us to implement a real informatic
database for every building, containing qualitative data, more than very highly
detailed parametric and dimensional factors. When comparing all the recollected
elements, it was necessary the ability to read and evaluate them, quite apart from
the potential of employed tools. Surveyed objects were interpreted and what was till
then only carefully “registered” was graphically translated.
The contribution deriving from knowledge, awareness and from the type of criti-
cal and methodological approach pursued during the survey, always requires proper
interpretations and graphic transpositions. It is necessary to overcome numerical data
and “read” the parts connecting structures, individuate differences, point out possi-
ble discrepancies, find connections with analogue buildings. This kind of activities
defined a wide documentary corpus, which embraced different typologies of analysis
and developed interesting considerations on different topics.
8 Conclusions
The art of Cuban wooden carpentry, directly derived from the culture of the Spanish
Mudéjar style, can be defined as uniquely characterized by the process it comes
from, as the result of construction techniques accrued along centuries in Europe and
so advanced that it had already solved all possible static problems.
Skillful masters of carpentry brought this building solution on the island, and they
knew how to mix it with local traditions that were pretty simple at the time, achieving
a kind of product that almost exclusively maintained its functional motivation and
noticeably simplitying its aesthetic form.
The elegance of decorations, often made in simple lines and geometrical engrav-
ings, stands in opposition with the always clear structural rationality.
The only exceptions is represented by some luxurious building in Old Havana with
cultural-historical value which hold a richer decorations setup, and for this reason
has already been restored (Fernàndez Santalices 2000). The historical center of La
Habana maintains one of the richest heritage for what is concerning colonial roofs,
most of them suffering a strong degradation because they belong to humble edifices
where dereliction and progressive substitution put their survival at risk.
While wishing for an expansion of the conservation of historical construction
techniques, survey and representation become essential tools to understand the whole
roofing system and its singular components, also becoming the methodological base
to know how to better manage and valorize those products. Focusing on some specific
historically valuable buildings, structure from motion is recognized as the proper
method towards every necessity that the model had to resolve, that is to say the
determination of a dimensional and formal configuration of the roof, more than the
geometry of pieces to understand their operation.
The choice to use image-based techniques during the architectonic survey of
Cuban wooden roofings allowed to generate a polygonal model which is also geo-
referenced and HD texturized, starting from pictures taken on field, speeding up
data acquisition times with respect to traditional survey, and contributing to give
an unique documentation on roofings theirselves, implemented with an informatic
database containing qualitative data and detailed dimensional and parametric values.
Despite the support given by employed tools, the competence of the researcher
in reading and interpreting data is essential; a proper graphic translation is only
possible with critical and methodological knowledge achieved by a specific historical,
typological and construction analysis of the available documentation carried out at
the beginning.
References
AA VV (1998) Guìa de arquitectura. Ed. consejo editorial, La Habana

Aa VV (2009a) Manejo y gestiòn de centros històricos. Ed. Boloña, La Habana
Aa VV (2009b) Regulaciones Urbanisticas Ciudad de La Habana - La Habana Vieja - Centro
Historico. Ed. Boloña, La Habana
Aa VV (2011) La plaza Vieja de La Habana - proceso de recuperaciòn. Ed. Consejerìa de Obras
Pùblicas y Vivienda, Sevilla
Aranda Bernal A (2011) El Orígen de la Casa de Pilatos de Sevilla. 1483–1505. Atrio. Revista de
Historia del Arte 17:133–172
Arduengo García DA, Cruz Pérez A (2012) Una nueva lectura del antiguo convento de Santa Clara
de Asís desde la arqueología y la historiografía. Arquitectura y Urbanismo 33(1):77–90
Bartolomei C, Ippolito A (2014) Low-cost cataloging methodologies for architecture: the system
of gates in Bologna. In: Ioannides M, Magnenat Thalmann N, Fink E, Yen A, Quak E, Way E
(eds) EuroMed 2014 digital heritage: progress in cultural heritage, documentation, preservation
and protection. Multi-Science Publishing Co., United Kingdom, pp 249–257
Bauer Arnold J (1999) La Cultura material. In: Carmagnani M, Hernádez Chávez A, Romano R
(eds) Para Una Historia de América: Las estructuras. Fondo de Cultura Económica, Mexico City,
pp 404–497
Bianchini C, Borgogni F, Ippolito A, Senatore LJ (2014) The surveying and representation process
applied to archaeology: a quest for invariants in a highly variable context. In: Di Giamberardino P,
Iacoviello D, Jorge RN, Tavares JR (eds) Computational modeling of objects presented in images
fundamentals, methods and applications. Springer International Publishing, Berlin
Campuzano L, Vallejo C (2003) Yo con mi viveza. Textos de conquistadoras, monjas, brujas, poetas
y otras mujeres de la colonia. Fondo editorial de Casa de las Américas, La Habana
Chateloin F (2000) Los techos de armadura cubanos. Arquitectura y Urbanismo XXI(4):23–32
Chateloin F (2003) Techos de madera en las iglesias habanaeras. Cimientos la revista de información
científico-técnica 7:57–60
Chateloin F (2008) El Centro Histórico ¿Concepto o criterio en desarrollo? Arquitectura y Urban-
ismo XXIX(2):10–18
Cuza Pèrez A, Rivery Carreras R (2015) Maderas de La Habana colonial. Anales del Museo de
Amèrica XXII:95–109
De Las Cuevas Toraya J (2001) 500 años de construcciones en Cuba. Ed. Chavìn, La Habana
De Luca L (2010) An iconography-based modeling approach for the spatio-temporal analysis of
architectural heritage. In: Shape modeling international conference, Washington, pp 78–89
De Luca L (2011) La fotomodellazione architettonica. Rilievo, modellazione, rappresentazione di
edifici a partire da fotografie. Ed. Dario Flaccovio, Palermo
Fernàndez Santalices M (2000) Calles de La Habana. Agualarga Editores, Madrid
Franco Rubio G (2009) La vivienda en el antiguo régimen: de espacio habitable a espacio social.
Revista de Historia Moderna 35:63–103
Gaiani M (2015) I portici di Bologna. Architettura, Modelli 3D e ricerche tecnologiche. Bononia
University Press, Bologna
García Santana A (2009) Urbanismo y arquitectura de la Habana Vieja - siglos XVI al XVIII. Ed.
Boloña, La Habana
García Santana A (2012) Treinta Maravillas del patrimonio Arquitectónico Cubano. Ediciones
Polymita, Guatemala
Menéndez García M (2007) La Casa Habanera. Tipología de la arquitectura Doméstica en el Centro
Histórico. Ediciones Boloña, La Habana
Munafò P (2002) Le capriate lignee antiche per i tetti a bassa pendenza. Evoluzione - dissesti -
tecniche di intervento. Alinea, Firenze
Piana M (2000) La carpenteria lignea veneziana nei secoli XIV e XV. In: Istituto veneto si scienze,
lettere ed arti (ed) L’architettura gotica veneziana, Venezia, p 73
Prat Puig F (1947) El pre barroco en Cuba. Una escuela criolla de arquitectura morisca. La Habana
Publisher, La Habana
Prieto Herrera O (2006) Arquitectura Habitacional Cubana y Camagüeyana. Síntesis del desar-
rollo arquitectónico de la vivienda colonial cubana. Caracterización tipológica de Camagüey.
CECONS, Camagüey
Roig De Leuchsering E (1963) La Habana Apuntes históricos. Editorial Consejo Nacional de
Cultura, La Habana
Sánchez F, De Julián JJ, Ordóñez A (2010) Tipologías constructivas en una ciudad patrimonio de
la humanidad. Revista de la Construcción 9(1):89–97
Styliani V (2016) Multi-scale 3D modelling of damaged cultural sites: use cases and image-based
workflows. In: Ioannides M, Fink E, Moropoulou A, Hagedorn-Saupe M, Fresa A, Liestøl G,
Rajcic V, Grussenmeyer P (eds) Digital heritage. Progress in cultural heritage: documentation,
preservation, and protection. Springer International Publisher, Cyprus, pp 50–62
Venegas Fornias C (2003) Plazas de intramuro. Ed. Mayra Fernàndez Peròn, La Habana
Weiss JE (1972) La arquitectura colonial cubana. Ediciones de Arte y Sociedad, La Habana
Weiss JE (1978) Techos coloniales cubanos. Editorial Arte y Literatura, La Habana
Weiss JE (2002) La Arquitectura colonial cubana. Siglos XVI al XIX. Instituto Cubano del
Libro/Junta de Andalucía, Sevilla
The Caribbean Bahareque: From Living
Branches to the Wall
Massimo Leserri and Dayan-Ariadna Guzman-Bejarano
Abstract This research project begins with a two-hour journey in the Cordoba
department, (Colombian Caribbean Region), two hours of endless pathway and
abrupt braking dividing the city of Monteria from Covenas town; namely the inland
from the sea, the green from the blue. A trip that has allowed landscape observation
presenting itself as an endless extension where nature mixes with man and cattle
presence. The project begins with landscape observation where inherency of fences
of live branches nailed in the ground is a constant. In such a landscape, what is
built seems to be obtained using wood, by means of self-construction and character-
ized by an emergency visualization of consensual stability. Long series of lined up
vertical boards, for example, describe horizontal lines which probably pursued the
need to define them as private fields and properties. Approaching the towns seems
to intensify poles’ deformation, but it can be just a sensation, and it hypothesizes a
new idea of curtain wall appearing urbanized when definitions begin to look alike.
The city is approaching, or it is a campaign that slowly hits in the lure of moder-
nity. Irregular enclosures begin to harden, to acquire new rules and orders. Heights
appear regularized. Even parcels will appear permanently blocked in a wall. The
research has been developed through the study of bahareque, an ancient technique
of Caribbean construction to build walls of wooden sticks and a final layer of humid
ground. Nonetheless, the study’s objective is to analyse the historical genesis of
vernacular architectures where sticks become wall. The wall is born, but hypothet-
ically the branches’ braided system has not completely disappeared (Fig. 1). The
wall continues to delimit borders and architecture as a branches’ set stuck in the
ground. It has neither lost its nature nor origin. They are signs of the landscape and,
at the same time, the most evident expression and symbol of knowing how to do and
live through a territory’s materials. The study ends with an analysis of the genesis
and use of living branches in fields such as architecture where their extraordinary
construction techniques represent an indelible knowledge.
M. Leserri (B) · D.-A. Guzman-Bejarano

Pontifical Bolivarian University, Montería, Colombia
e-mail: massimo.leserri@upb.edu.co
D.-A. Guzman-Bejarano
e-mail: dayan.guzmanb@upb.edu.co

624 M. Leserri and D.-A. Guzman-Bejarano
Fig. 1 Bahareque in San Antero (Colombia). Source Massimo Leserri, 2017
Keywords Colombia · Bahareque · Architecture · Caribe · Vernacular
1 Introduction
Department of Cordoba (Colombia). Landscape, a mixture of exuberant nature of

tropical vegetation and livestock. Elements appearing to derive from nature are pos-
sible to observe. The result of self-building characterized by an emergency portrait of
consented stability. Long series of aligned vertical planks describe horizontal lines
in landscape and probably pursued the need to define them as private fields and
properties.
Each series is shaped by branches nailed to the ground, and each element singu-
larity is confirmed by the adjacent one according to a comparison evaluation made
by the invited observer. Then, each element is vertical enough to keep balance and
at the same time maintain its natural essence.1
It seems that the walls´ architecture follows tectonics according to Semper cate-
gories, which in, there is a different relationship between architecture and nature; or
even better , the ground. Thus, this vernacular architecture works by reading through
1 Leserri et al. (2018).

The Caribbean Bahareque: From Living Branches to the Wall 625
tension regarding another architecture that works for compression; and therefore, it
appears more plastic and inspired by the ground. The basis is that a wooden con-
struction is always a light, serial work and with several levels of transparency and
permeability.2
2 From Living Branches to the Wall
Fences distinguished from living nature share in the landscape through a different
order. They are also a set of finite or pseudo-finite elements (branches, trunks) not
objects of processing. The tectonics reconfirms it is the art of assembly or connec-
tion not only of construction pieces, but objects as well. Message is that matter’s
architecture aims at the space to be continuous with nature, because that, through a
human intention, it becomes material. This construction has an image of a mobile,
nomadic structure arising from the spatial need to link where it is based, as well as
other performances resulting of a work to achieve a space where it can be safely
covered.3
Nonetheless, by observing the region´s landscape, the relationship between wall
and fence is also evident because stakes are closed without losing their characteristics,
determining a wall of living branches that seems the result from the transformation
of an authentically discontinuous form in an apparently continuous one (Fig. 2). It is
a mobile wall, also tectonic, built with thickening stakes where emptiness is reduced
to the minimum possible. But this is a wall because it definitively closes the space
where senses appear; the soul and the genesis with sensitive and spiritual components.
Movement is contrasted with life in a continuous space full of living nature, indeed.
Used materials come from nature adapted through ideas and intuitions. Each idea
is possible by presence of concrete matter as well as author’s experience. Now the
fence becomes a wall when its transparency and permeability lessen; inviting us to
reflect not only on the wall’s concept of the continuum, but also on the vernacular
architecture relating to thought where matter is a whole in the idea (Fig. 3).
As in a wall, all equal parts come together as those found in landscape and are
integrated into its thickness. Thus, the vernacular architecture retains its immobile,
static, but not necessarily permanent character. Pieces within it are not identified
in the whole, namely, they are not part of a predefined or prefigured mechanism.
Space is born from the inner idea and not from partial operations with assembled
components. The created space, in this sense, is discontinuous with nature in terms
that the architectural idea and space are achieved through the critical use of matter
and nature; so much, that matter becomes material.
Over time, the house is closed by a wall providing the idea of being a “protector”
of primitive cane. Semper found a clear example for his fourth elements theory of
architecture in the Universal Exhibition from 1851; in his painting of an Indian hut in
2 Ugo (1990).
3 Ugo (1990).
Fig. 2 Living Branches in the Cordoba Departments (Colombia). Source Massimo Leserri, 2017
Trinidad (elevation and floor) with the four identifiable elements. In this Caribbean
cabin all the elements of ancient architecture appear in their most original and unadul-
terated form: the home as a centre, the bench surrounded by a framework of poles
as a terrace, the roof supported by columns, and mats as enclosure of space or wall.
This modern Caribbean cabin was a good example of the link of primitive domes-
tic construction and confirmed the writings of Vitruvius on derivation of the Greek
temple from a wooden structure.4
… to nail a stake, not to mention a whole row of poles, therefore, is considered more and
more as a metaphor and an act of a generic construction work. It is also an act of taking
land possession: each stake driven implies a circular area around, if nothing else than the
shadow that projects, which direction is always parallel to the projection of the upright
structure. A pole planting is a primary gesture - the ability to guide, to know the proper
body’s perpendicularity with respect to the ground. That is a condition of our being Homo
sapiens put on “erectus” and; before taking charge and possession of the first cut stakes, men
can have learned the technique in living trees nailed to the ground…5
Regarding the trees, they find the ground, the most requested point. In addition,
to support high tensions, a strong bond is needed, and trees have a dense root system
playing this role, as a trunk’s anchor to the ground in an indissoluble way. For some
scholars, wood was the first building material used by man since prehistoric times,
and its peculiarities (easy availability, good physical and mechanical characteristics,
easy workability and, in some cases, the possibility of use, even without any pro-
cessing type) have made it a raw material mostly in construction especially among
4 Aparicio Guisado (2006).

5 Rykwert (2010).
Fig. 3 Image idea on “From the Living Branches to the Wall of Bahereque”. Source Massimo
Leserri, 2018
primitive society. Examples are old wooden buildings dating from Palaeolithic times
and covering structures like buried curtains, constructions which in vertical walls and
flat roofs were made with tree trunks aligned, glued and covered with ground.6 The
study and use of nature made possible to understand its mechanical language. Some
of these laws are stronger, again, and not only to provide a characteristic common
to all elements (standardization), but also to rank (locate) length and diameter of
6 Tomei (2010).
branches, which gradually decrease and go upwards; or the design of the lymphatic
system that brings life through the body.7 All that observation suggests knowledge
for construction and its success.
Good wood has volume that decreases in proportion as it loses temper; unlike
ropes, drying achieves more in length and lowers thickness. With drying, more weight
is lost, and more volume decreases than from the centre, and therefore, large trunks
drying lose their weight, but their volume. The first architect and builder must know
how to get these and other trees and know their destiny8 “… It is necessary to cut
trees in autumn, and throughout winter, because in this period they recover from the
roots, the vigor and firmness that in spring and summer had lost. They will be cut
without the moon; because not to bribe, the forests at that time are not followed by
moths or woodworm. They should only cut half of the bone marrow, and thus leave
them, as they dry up… Vitruvius, to that place, gives good learned instructions, who
other men, copiously wrote9 …”
Nature was domesticated, understood, submitted to taxonomic organizations that
reasonably explained what had previously been conceived as an ungraspable and
threatening mystery. In fact, the traditional system to build a skeleton or structural
part is based on wooden horcones in the countryside as in the urban centre, resulted
from a tree of appropriate size. All over the world, the choice of materials and con-
sequent typologies given life to homes, do not have an arbitrary character. It is also
about results of a profound evolution process. “… This was concretized through
experimentation with multiple possibilities, that later, the indigenous people under-
stood as the necessary elements for the houses to acquire the multiple benefits they
enjoy - including those of a bioclimatic and sustainable nature, so fashionable in
our days10 …”
2.1 The Bahareque
Structure made of domesticated and unprocessed logs regarding to a modern idea of

uniformity is presented, in fact, with a still natural morphological portrait. The use
of wood that, together with combination of cane and clay, is the main material used
in this Caribbean region vernacular architecture (Fig. 4). There are cases of urban
houses having a structural system formed by pillars and wooden beams, derived
from use of trees which selection of records seems to occur in accordance with the
recognition of the cross sectional dimensional conditions.
Each structural component (trunk or branch) still appears because of a simple cross
section, without use of additional longitudinal cuts or work planned for elimination of
7 Fontana (2012).
8 AAVV (1785).
9 Palladio (1570).
10 Sanchez Suarez (2017).
Fig. 4 An example of the bahareque in Monteria (Colombia). Source Watercolor by Victor Julio
Silva De la Ossa, 2007
Fig. 5 Some examples of bahareque in Cereté (Colombia). Source José David Puche Berrocal
(Colombia) 2013
geometric perception irregularities such as the continuous spiral pattern.11 However,

house location can indicate different compositional typologies between urban or rural
solutions. The four macro-construction phases are, first, to raise the main structure,
then place the splice, then build the perimeter wall and finally place the elements door
and windows. A constant is true, the wall theme made through bahareque technique
(Fig. 5).
11 Leserri et al. (2017).

Fig. 6 Structural of the vernacular architecture of the Bahareque. Source Massimo Leserri 2018
Often the type of housing of one floor and therefore, the simpler one, shows
horcones (vertical elements) representing the first structural work to which, then, the
table is assembled; namely, a structural ring of braces (longitudinal) and seatings
(transversal) making up the system rising on the horcones (Fig. 6). Above the table,
a construction of the cover roof set of canes develops. In fact, the cover roof is
composed of a ridgepole (or caballetera), scissors, set of canes, limestones, rods and
caiman.12
Experience suggests that horcones must be made with wood resisting humidity,
since they are going to be buried, and the loads for them to resist, are great, (the whole
cover weight, winds and earthquakes). They must have a sufficient diameter: (about
15 cm). In sawn wood, a 4 × 4 “(10 × 10 cm) dimension is used.” “… When the wall
is a stood stick, it does not matter horcones are not very straight, since they will be
left out of it. When it comes to grid or trabilla walls, where horcones are included,
their length depends on the height that has been foreseen and depth at which they
are to be buried …” When the terrain is a slope, it is likely to require horcones of
different lengths. Guaiacwood is the preferred wood.
To carry out this activity, a minimum number of workers with a construction
foreman and at least two assistants are needed; due to the horcones’ weights, making
them difficult to manipulate. Layout is carried out by the foreman, as everything that
has a significant reason, such as the horcones’ leaking and leveling; tasks that, due
to complexity, require skilled workers. Nonetheless, this activity is done quickly.
Instruments such as machete, set square, meter, plumb-line (lead), barretón or pala-
draga, shovel, pisón, thread (lienza), level and saw are all needed to place horcones
inside holes, making co-locate the center of these with the crossing of lienzas. The
head or upper part should be plumbed with the foot or lower horcon’s part (a). This is
also true when using round wood, which is not completely straight. Fill the gaps with
selected soil, not organic and preferably sandy. Tamp down around, taking care that
the horcon does not lose verticality with bangs. Level the horcones to the height at
which they are going to be left, generally between 2 and 3 m from the finished floor
and cut them with a saw (c). To protect them from humidity and animals attacking
them, the part to be buried is spread with burning oil, gas (petroleum), tar, or with
12 Angel and Sánchez (1990).

some immunizing product (a). It is also a tradition to slightly burn the tip or pour
quicklime into the hole (b). Immunizers should be applied regularly because they
lose their effect over time.
Today, if there is no good wood, stakes can be made with heart wood or reinforced
concrete piles for the part buried to which the available wood horcones are tied with
wire, nails or bolts, (e). Bolted connections work much better but require a “barbiquí ”
(tool builders usually do not have in the region). The wood used for seatings and
braces must have a minimum of 3 (8 cm) in diameter. The most even and straight
one is chosen for seatings and braces and for scissors and limestones. Portaletes
come from the sticks cutting. Several woods and guadua are used for seatings and
braces. For scissors and ridgepole, mangroves and guadua are mainly used. For the
framework, guadua and some cans and branches of some trees are used. Some trees
tend to be replaced today by more commercial materials such as wire, bolts and nails.
The cover assembling work requires an experienced foreman because of the many
details that must be considered. To assemble the butt is the most complicated part,
which must be done very carefully because it has many details.13
Different systems are used in walls construction, depending essentially on local
resources. Fence is made up by a complementary structure allowing union between
main structure and filling. This structure which in materials such as wood and reeds
are used, linked to the main structure, must allow “assembly” made from filling of
ground and fibers, as well as union with the other elements. The complementary
structure is generally composed, of a “frame” serving as support for the “skeleton”,
the set of pieces intended to support the filling.
In the stand cane system, horcones are not included in the fence, which axis is
approximately 20 cm away from the horcones’, in such a way that they are located
towards the fence interior. The placement of the fence itself, consists of the vertical
placement, an ally of the other canes or cans covering the spaces between stands. It
is carried out as follows: (1) Chop canes and/or woods with machete on the twisted
side (in the sunken part or belly) to straighten them and improve mud adherence. (2)
Place canes one by one interspersing head and foot, for the fence’s verticality. (3) Tie
the ribbons with wire or bejuco; or nail them. Generally, this work is carried out by
two men with simple or double-chain locks. These are placed, one inside the fence
and the other outside, so that while one executes the ties at the top, the other makes
them at the bottom. Depending on the thickness of the canes or cans, they can be tied
together.
Woods used for the stands are Willow, Carbonero tree, Dividivi, and Vivaseca.
When they are placed directly on the ground, their diameter varies between 3 and
2 . It is technically originating a repetition of the waterproof fence with autonomous
living branches. Structurally, when wood is not locally available, bamboo and man-
grove are used, and one or two layers are placed depending on the desired thickness.

One of the recommendations is that the palm veins, the cans or the list of sawn
wood should be of 2 by 2.5 . The stand should be placed on a stake. Other woods
used in the stands are: the heart of Mora, Mataraton, Coquillo, Solera Oak, Pijino
and Carreta.14
When the final fence is done, its vertical and/or horizontal element forms a kind
of “fabric.”15 Nature, in fact, offers a significant production of plants with which
various fabrics are created, but also knots and moorage for construction. Every cre-
ation made by hand, which is an architecture or artisan element is the product of
researched nature. The same construction seems by a verse, a large fabric for the
home. The bahareque is finally, the technique where the living matter intentionally
placed becomes material, and the wall defines the life of the architectural space.
It is really the moment to definitively close the interior space to the outside and,
the vertical elements, the canes, are placed. The branches of the analogue near the
countryside, one, and then the other. But the arrangement can vary depending on
the fabric that will be carried out (single or double). Chance does not exist, because
the empirical technique suggests that the woods to be used must be thin branches of
hardwoods that do not thicken.
These are long lasting and available materials, since they rapidly reproduce in the
fields and are a constant part of the Colombian Caribbean landscape. The place of
construction coincides with the same where they can be taken directly from the trees
allowing easy access and economy; and often the same owner also becomes a builder
and architect. The walls can be, generally non-structural, but in the case where they
are formed with a framework of guadua and/or wood composed of lower and upper
sill. That is, structural parts and inclined elements (at about 45°) between them.
Guadua is a material which best quality is achieved in season plants, (more than
four years old). Immunized Guadua does not mean protection against other envi-
ronmental effects, so, the wood cannot be exposed to sun or water in any part of
the building, because ultraviolet rays produce cracking, discoloration and loss of
brightness. Plus, moisture changes can cause decay (Fig. 7).
After the fabric’s placement, the smear should be applied immediately so that
“horizontal” elements of the “fabric” are not “drained”. This is done by pasting both
sides. Traditionally repelling uses local sandy loam and “dung” (cagajon or dung) of
cattle both in the layer of the “embrute” or “smear” and as in the finish repelle (Fig. 8).
Nowadays, cement use has been introduced into the mixture of the compound for the
final finishing. To carry out the repelling, the mixture to be used must be prepared,
the kneading must be done and finally the smear and repelle must be applied. At the
end, Bahareque, Bareque or Bajareque can be defined as a construction system from
woven sticks or reeds and mud.16 This technique has been used since ancient times
for the construction of housing from indigenous peoples of America, widely used by
Amerindians, mainly in Colombia and Venezuela.

Fig. 7 Making the last layer of the bahareque. Source José David Puche Berrocal (Colombia) 2013
Fig. 8 Making the last layer (repelle) of the Bahareque. Source José David Puche Berrocal (Colom-
bia) 2013
The wall is not only a continuous division that allows to close an interior space,
but it can also be a dialogue with the outside and allow the passage of light, air and
landscape that are realities that become elements of the whole, of the architecture’s
continuous.
3 Conclusion
We must ask ourselves if this live architecture, important both in Colombia and in
the rest of the world, demands a reflection on how to articulate an immediate and
universal sense of beauty. Something endowed with simplicity and intensity, which
is never insulting or arrogant. The problem is, perhaps, that man separating himself
from the hostile nature, through the conquest of a secure space at the same time,
loses the continuity of the mystical relationship that links him to nature, to the world,
losing with his own house, “the measured and excellent environment” of nature and
begins to feel the act of separation as a violent act against him.
But perhaps this fruitful journey is not only a work understood in the direction to
comprehend the nature of the beauty of this Caribbean scenery, how to discover and
confirm, perhaps, the universal need of man to make architecture through essentially
the concepts of fences. The meaning of the architecture is perhaps proper in the
archetype of the enclosure, which controlled space, separated an internal and an
external space through a delimitation. Man builds and manifests his existence through
the creation of enclosures.17
This study developed through drawing and anthropological analysis forces us
to continue researching the great heritage of this extraordinary region of the world
where understanding constructive technique as a cultural fact seems consistent with
the arguments outlined above on landscape and living.18
References
Angel OC, Sánchez GC (1990) El bahareque en la región del caribe Modulo 7 La técnica. Sección
de Publicaciones. SENA General Dirección on Bogotá, pp 18–25
Aparicio Guisado JM (2006) El Muro. Biblioteca Nueva, Madrid, p 205
Fontana L (2012) Lezione africane, per un’architettura Materica, sociale, Organica e Giovale, Cre-
ative Commons CC BY-NC 3.0. Ropi (Ethiopia), p 44
Herrera Castro S, Hernandez Spinola F (2017) la Arena in Naturaleza. In el Habitar 01, Tradición
constructivas en maderas y fibras naturales. UNAM Mexico, pp 36–37
Leserri M, Olmos Lorduy JP, Castillo Ayazo HE, Castillo Sarmiento MA (2017) The survey of ver-
nacular architecture: casa Grandeth. In: Amoruso G (ed) Putting tradition into practice: heritage,
place and design. INTBAU Lecture Notes in Civil Engineering, vol 3. Springer, Basel (Suiza),
pp 441–448
Leserri M, Chaverra Suarez M, Rossi G, Guzman Bejarano DA (2018) From an Enclosure to the
Corraleja. An Analysis of the Genesis of an Ephemeral and Vernacular Colombian Architecture.
In: Brusaporci S, Amoruso G (eds) Buildings Journal v.8 fascicle. n/a pp. 1–14, MDPI, Basel
(Switzerland). https://doi.org/10.3390/buildings803004
Milizia F (1785) Principi di Architettura civile, Tomo terzo. Anno MDCCLXXXV, Remondini di
Venezia, Bassano, p 86
Oliva G (2012) L’esperienza del paesaggio in Riflessioni, Collana hortusbook. Nuova Cultura,
Roma, p 67
17 Oliva (2012).
18 Herrera Castro and Hernandez Spinola (2017).
Palladio A (1570) Dei Legnami. In: I libro dell’architettura, Cap. II, I Qvattro libri dell’architettura,
Venezia
Real Academia Española. http://www.dle.rae.es/?id=4mknvqT
Rykwert J (2010) La colonna danzante. Sull’ordine dell’architettura, Libri Scheiwiller Ore Cultura.
Milano, p, 88
Sanchez Suarez A (2017) De los Angeles Vizcarra de los reyes M, A la Sombra de la Selva. In el
Habitar 01 Tradiccion constructivas en maderas y fibras naturales, UNAM, Mexico, p 139
Tomei M (2010) Architettura. Schemi riassuntivi, quadri di approfondimento, Istituto Geografico
De Agostini. Novara, p 256
Ugo V (1990) Laugier e la dimensione teorica dell’architettura. Dedalo, Bari, pp 198–199
Part III
Connection Between Nature Lessons
and Material Explorations
Computational Mechanical Modelling
of Wood—From Microstructural
Characteristics Over Wood-Based
Products to Advanced Timber Structures
Josef Füssl, Markus Lukacevic, Stefan Pillwein and Helmut Pottmann
Abstract Wood as structural bearing material is often encountered with skepticism

and, therefore, it is not used as extensively as its very good material properties would
suggest. Beside building physics and construction reasons, the main cause of this
skepticism is its quite complex material behavior, which is the reason that design
concepts for wood have so far not achieved a desirable prediction accuracy. Thus, for
the prediction of effective mechanical properties of wood, advanced computational
tools are required, which are able to predict as well as consider multidimensional
strength information at different scales of observation. Within this chapter, three
computational methods are presented: an extended finite element approach able to
describe strong strain-softening and, thus, reproduce brittle failure modes accurately;
a numerical limit analysis approach, exclusively describing ductile failure; and an
elastic limit approach based on continuum micromechanics. Based on illustrative
results, the performance of these methods is shown and discussed. Furthermore, a
finite-element-based design procedure for an elastically-deformed wooden structure
is outlined, showing how advanced mechanical information of the base material could
be exploited within digital design of complex timber structures in future. Finally, geo-
metric design concepts applicable within digital wood design are discussed, giving
insights into possible future developments.
Keywords Computational mechanics · Mechanical modelling

Structural analysis · Wooden microstructure · Wood-based products
J. Füssl (B) · M. Lukacevic · S. Pillwein

Institute for Mechanics of Materials and Structures, TU Wien, Karlsplatz 13, Vienna, Austria
e-mail: josef.fuessl@tuwien.ac.at
H. Pottmann
Computer, Electrical and Mathematical Science and Engineering Division, King Abdullah
University of Science and Technology, Thuwal, Saudi Arabia

640 J. Füssl et al.
1 Introduction
Engineering structures, built by and for the use of people, have to meet certain
requirements with regard to structural stability, load-bearing capacity, and service-
ability. To check in advance whether a planned structure will fulfil these requirements,
its mechanical behavior needs to be described by mathematical models and assessed
by design concepts derived from it. This is referred to as structural analysis, which
has its roots in the 17th century, if one ignores the very first insights from Archimedes
in the 3rd century BC and the great achievements of the ancient Romans. In the 17th
century, Galileo Galilei, Robert Hooke, and Isaac Newton laid the foundation for
structural analysis by outlining the science of the strength of materials, the motion of
objects, and providing an understanding of the elasticity of materials. With the works
of Leonhard Euler and Johann Bernoulli, formulating beam theory and the theory of
virtual work, the mathematical analysis of structures was finally established in the
18th century.
Since then, structural analysis has developed very fast and experienced a new
boost in the late 20th and early 21st centuries through the development of computer
technology, allowing mechanical field equations to be solved numerically to a great
extent. Finite element methods and related concepts have become significant tools
for structural analysis and design, and experience a broad application in almost all
areas of engineering. Reliable predictions of the mechanical behavior of complex 3D
structures allow constructions with sophisticated design to become larger and larger
in an economically and efficient way.
However, most of the structural analysis performed by civil engineers is still heav-
ily based on linear elastic material behavior and non-linear effects are considered
only very simplified. This has three main reasons: First, considering non-linearities
leads to a sharp increase in the computational effort, secondly, appropriate numerical
methods are still under development and often not sufficiently stable for practical
applications, and, third, the non-linear material properties are rarely available to the
necessary extent for 3D calculations. The reason for the latter lies in the fact that nec-
essary material input parameters for structural analysis are mainly obtained through
experiments. These experimental programs include selected loading situations, like
uniaxial tensile, compression, or bending tests, respectively. A comprehensive 3D
information of the material behavior cannot be gained therefrom. This applies in par-
ticular to building materials with anisotropic behavior and several clearly different
failure mechanisms depending on the loading situation, like it is the case for wood.
Thus, the mathematical modelling of building materials has become more and
more important in recent decades, providing structural analysis and computational
design with better input and, simultaneously, providing the basis for optimization and
further development of building materials themselves. Mathematical models were
therefore extended to lower length scales of building materials, making it possible
to include microstructural characteristics, like the volume, shape, and orientation of
material phases, into mechanical design approaches. By describing the mechanical
behavior at each scale of observation and linking this behavior through the different
Computational Mechanical Modelling of Wood—From Microstructural … 641
scales of observation up to the macroscale, effective (macroscopic) properties of

building materials can be determined as a function of their composition. For different
compositions of similar material systems it is then straightforward to determine
the corresponding material properties and, with a sufficiently validated multiscale
model at hand, no additional experiments are needed. A more profound mechanical
understanding of building materials can be gained and optimization procedures are
easier and more efficient to implement. However, this shift of “natural variability” into
the design process on the computer also results in more complex mathematical models
and requires computational methods able to describe the mechanical processes taking
place at different scales of observation of the considered building materials.
For some building materials, such as e.g. cementitious materials and steel alloys,
quite advanced multiscale concepts already exist and computational design tools
benefit from the material understanding gained from them. However, for the very
traditional building material wood the research on multiscale description of the
mechanical behavior is only at the beginning. This represents the main motivation
for this contribution, proposing advanced computational modelling concepts, for the
microstructure of wood over wood-based products to advanced timber structures.
1.1 Motivation
Wood is a naturally grown material with excellent mechanical properties. Comparing

mass density to representative strength value ratios of different common building
materials, defect-free wood is a factor of three better than steel, considering tensile
strength, and about ten times better than concrete, even with respect to compressive
strength. Nevertheless, wood as structural bearing material is often encountered with
skepticism and, therefore, it is not used as extensively as its very good material
properties would suggest. Besides building physics and construction reasons, the
main cause of this skepticism is the quite complex material behavior of wood, the
huge variety due to hundreds of species, and its strong inhomogeneity, resulting from
branches and growth irregularities. Until now this hasn’t been a crucial problem,
because, traditionally, wood as a structural building material has mainly been used
in rural areas for one- or two-storey residential buildings or simple halls and stables.
Due to the relatively small dimensions of such buildings, the low complexity, and
the fact that each structural element only appears in a small number, it has not
been necessary to exploit the full mechanical potential of wood. Simple design rules
combined with practical experience and considerably oversized components have
together met all requirements.
In recent years, this situation has changed dramatically, as becomes obvious by
looking at all the high-rise buildings made of wood mentioned, e.g. in Füssl et al.
(2017). The excellent mechanical and physical properties of wood, combined with
the general trend of growing environmental awareness, have put timber structures
into the focus of private as well as public building developers—not just to realize
small buildings, but to use wooden building elements for highly sophisticated engi-
neering structures. The design of geometrically complex building structures also

requires advanced knowledge about the underlying material itself. Suddenly, struc-
tural design, which has historically been concerned mostly with beam-like structures,
and thus 1D design concepts, has to deal with complex 3D stress and strain fields
arising within, and in connections between, structural elements. These stress distri-
butions are often responsible for the initiation of cracks or plastified material zones,
and therefore have to be taken into account accurately. This is only possible when a
detailed characterization of the multidimensional strength behavior of the considered
structural element is available. This is actually never the case, because, as mentioned
before, the enormous experimental program that would be required is hardly feasible.
This applies in particular to timber structures, where the very complex mechanical
behavior of the material wood often hinders the realization of challenging and excit-
ing design ideas.
1.2 Structure and Objectives
Derived from the previously described evolution of structural analysis and inspired
by the reasonable trend of linking the macroscopic mechanical behavior of building
materials to their microstructural characteristics, and in that way providing compu-
tational design tools with reliable 3D mechanical behavior, the presented scientific
work in this chapter describes building blocks for a holistic computational modelling
approach for wooden structures as displayed in Fig. 1. This concept can be subdi-
vided into five areas, where each of them is addressed in this chapter, devoted to the
following objectives:
1. Determination of the orthotropic clear wood material properties with respect to
the principal material directions by means of a multiscale model (described in
Sect. 3), where at each level the elastic behavior, failure stress states (strength
properties), as well as failure modes are determined and the information obtained
is transferred—and serves as input—to the next higher scale of observation. Three
scales of observation are schematically illustrated in Fig. 1 (top), showing the
basic characteristic (cellulose fibers embedded in a polymer network) of the
wood cell wall material, the periodically arranged honeycomb cells of the early-
and latewood layers, and the laminated structure at the clear wood scale. For
the so-called upscaling of mechanical properties, three different computational
methods are used depending on the respective (micro)structure, such as methods
from the field of continuum micromechanics as well as extended finite element
approaches. All three computational methods are described and discussed in
detail in Sect. 2. Finally, comprehensive information about the 3D mechanical
behavior of clear wood is obtained, reflected in a so-called multisurface failure
criterion with respect to the strength behavior of wood (proposed in Sect. 4.1).
2. Revealing performance-influencing morphological characteristics in wooden
boards (addressed in Sect. 4.2), like knots and fiber deviations, and generating
Fig. 1 Basic concept for a holistic computational modelling approach for wooden structures
3D digital models of wood and wood-based products, which can further be used
as input to 3D numerical simulations at the wood-product scale (Sect. 4.3). They
allow for a realistic description of mechanical processes within wood, by taking
into account brittle as well as ductile failure modes. Together with the material
property information obtained through the multiscale model, also stiffness and
strength profiles for wooden boards can be obtained (as exemplarily shown in
Fig. 1), serving as input to deterministic simulations as well as to stochastic

concepts for wood-based products.
3. The determination of the effective behavior of wood products taking into account
the inherent material property variation of wood and wooden boards. Therefore,
different stochastic approaches can be applied, as briefly outlined in Sect. 4.
4. The application of computational methods to guide a design process for complex
timber structures, such as the one shown in Fig. 1 (bottom right). Therefore, focus
is laid on a structure built by elastically-deformed lamellas. The design process
to obtain the arrangement of the geometrically compatible lamellas is outlined
in Sect. 5.
5. Finally, promising geometric design concepts which could be used in future in
timber engineering are proposed, on condition that sufficient information about
the mechanical behavior of wood is available and can be used as a physical
foundation for these methods (Sect. 6).
2 Computational Modelling Tools
The failure behavior of wood is quite complex, being strongly dependent on the
type of loading and the level of consideration (observation scale). The majority of
the models in the scientific literature focus on the wooden board scale (see Fig. 1),
describing the failure behavior of clear wood and the influence of knots. At this scale,
under pressure-dominated stress states, wood behaves in a very ductile manner and
strain hardening may even occur. Under tensile- and shear-dominated stress states, a
rather brittle failure behavior can be observed, but with a fracture toughness which
can be very different dependent on the wooden structure considered. A brief overview
of some existing methods for predicting/modelling structural behavior at the wooden
board level is given in the following.
The first group of approaches avoids any direct description of failure mecha-
nisms, and instead uses mean stress concepts, see Masuda (1988) and Landelius
(1989), where averaged stresses over a small finite volume are assumed to indicate
failure. These volumes can be adjusted to typical features of wood, such as structural
characteristics of wood fibers, see Aicher et al. (2002). In Serrano and Gustafs-
son (2006), Sjödin and Serrano (2008), and Sjödin et al. (2008), this approach was
applied in combination with findings of linear elastic fracture mechanics. The suit-
ability of different volume definitions, over which the stresses are averaged, and
also the effectiveness of various failure criteria, was compared in Guindos (2011).
A physically-based structural failure criterion was presented in Lukacevic and Füssl
(2014) and Lukacevic et al. (2014), where it was assumed that in wooden boards with
knots, global failure can be related to a stress transfer mechanism, which is identi-
fiable by evaluating averaged stress fields in the vicinity of knots. Models based on
mean stress concepts are mostly dependent on empirical parameters, and the true
failure mechanisms cannot be mapped. This can be overcome by directly modelling
failure processes.
(a) (b)
Fig. 2 Characteristics and qualitative representation of three different computational methods:

a effective/macroscopic stress-strain response and b microscopic stress-strain relationships (Füssl
et al. 2017)
The most sophisticated approaches for this are based on multi-surface yield/failure
criteria, as described in Mackenzie-Helnwein et al. (2003), Lukacevic and Füssl
(2016) and Lukacevic et al. (2017). In the latter, orthotropic yield surfaces are defined
to describe the onset of ductile failure, whereas orthotropic crack surfaces indicate
stress states under which cracks are to be expected. Cracking is normally modelled
with cohesive elements, including an anisotropic traction separation law for wood.
Applications of this model show promising results regarding the estimation of load-
carrying behavior, see Schmidt and Kaliske (2007, 2009).
These approaches work well for a uniform fiber orientation and if the location of
the crack path is known in advance. However, due to knots and the fiber deviations
around them complex stress states usually prevail, and such an approach becomes
impossible to apply. Thus, to be able to assign reliable failure surfaces and cor-
responding crack directions to clear wood at the wooden board scale, multiscale
approaches should be considered. For such an approach, failure mechanisms at dif-
ferent observation scales need to be analyzed by appropriate computational methods.
Three different promising methods are briefly proposed and discussed in the follow-
ing sections. Furthermore, the basic characteristics of each method are described and
also qualitatively shown in Fig. 2.
The type of structural response that can be described by each method, in terms
of relationships between the effective stress tensor and the effective strain tensor
E, is illustrated in Fig. 2a, where Ceff denotes the effective elastic stiffness tensor.
Figure 2b shows the corresponding relationships between microscopic stresses σ (xi )
and microscopic strains ε(xi ), for a material point i of the considered structure. The
values given therein denote strength and stiffness values of the wood microstructure
(see Fig. 1, top), which will be explained in detail in Sect. 3.
The most obvious way to deal with this complex problem, with respect to mor-
phological as well as material behavior aspects, is to apply the conventional 2D or
3D finite element method. This has been done by the authors recently, see Lukacevic
et al. (2015a, b). Therein, XFEM was applied successfully to representative wood
unit cells, delivering effective stress states which cause the initiation of a crack,
and simultaneously giving the corresponding effective crack direction. Neverthe-
less, even if XFEM, especially in combination with multi-surface plasticity criteria,
is a powerful tool allowing for a very realistic description of complex failure mecha-
nisms, its high computational cost and the overall complexity of the approach make
its use within a multiscale approach rather tedious.
For this reason, additional focus is given to the numerical LA approach and its
applicability to microstructures of wood. This method exclusively concentrates on
the time instant of failure, and delivers lower and upper bounds for the ultimate
strength of the considered material structure. Compared with conventional finite
element approaches, where the complete load history has to be considered and, in
order to predict the correct failure mechanisms, proper regularization techniques
must be used, the LA approach is much more stable and efficient. Moreover, it leads
to rigorous bounds on the material strength, and thus gives a reliable error measure
for the prediction. The completely different nature of this so-called direct method,
compared with XFEM, becomes obvious in Fig. 2.
Of course, the efficiency of LA results from the strict assumptions on which
this method is based. These include an associated plastic flow rule and ideal plastic
(ductile) behavior of the considered material. Although these two assumptions are
not entirely correct for wood, the first is generally made due to a lack of information
about the non-associativity, while the second does not exclude the possibility that
good strength predictions can also be obtained for strain-softening structures. Wood
certainly has the ductile potential to ‘activate’ the strength of many points along a
potential crack surface before brittle failure occurs. It therefore seems that LA could
be an appropriate method for determining effective strengths, at least for pressure-
and shear-dominated loading states.
The EL approach has been chosen as third method. It is based on continuum
micromechanical homogenization schemes, as presented in Zaoui (2002) and Suquet
(1997), characterized by linearly averaging stress and strain fields of material phases
to obtain the effective stiffness tensor of the corresponding material system. Higher-
order strain averages, in combination with an appropriate failure criterion for the
failure-pertinent material phase, can then be used to predict effective strengths. These
strengths refer to elastic limit states in case of ductile failure, but represent ultimate
strengths in the case of brittle failure. For this reason, the EL approach ideally com-
plements the LA approach, and thus an effective multiscale approach delivering
comprehensive failure information for clear wood seems to be possible.
The adopted formulations of the three different computational methods, which
are subsequently applied to the wood microstructures, are presented in the following
subsections.
2.1 Extended Finite Element Method (XFEM)
The XFEM, first introduced in Belytschko and Black (1999) and based on the par-
tition of unity method proposed in Melenk and Babuska (1996), allows, through
targeted enrichment of the approximation possibilities of the displacement field, the
description of singularities and weak or strong discontinuities of the considered field
quantity. A field u(x), in our case a displacement field, can thereby be approximated
by
⎡ ⎤
NN
NG
u(x) ≈ Nu,i (x)⎣qu,i + G j (x)qenr,i j ⎦ (1)
i1 j1
where NN is the number of nodes of the finite element mesh, Nu,i represents the
conventional
N N element interpolation functions obeying the partition-of-unity prop-
erty, i1 Ni 1, with NNE as the number of nodes of an element, qu,i is the
displacement vector, and qenr,ij contains the additional degrees of freedom of node i,
representing the amplitude of the jth extension function G j (x). NG is the number of
extension functions. In this chapter, the basic XFEM capability of a commercial finite
element package was used, wherein the so-called level set method, see Osher and
Sethian (1988), Moes et al. (2002) and Gravouil et al. (2002), is applied to describe
the location and geometry of discrete moving cracks. Two level set functions, which
are two almost-orthogonal signed distance functions, are sufficient to describe arbi-
trary material interfaces and cracks by using conventional finite element interpolation
functions. Commonly, within the XFEM, singularities caused by crack tips are mod-
elled with extension functions derived in the framework of linear elastic fracture
mechanics, see Belytschko and Black (1999). For elements that are completely cut
by a crack, the corresponding nodes are enriched by sign or Heaviside functions. In
this chapter, for describing cracks (strong discontinuities) only the modified Heavi-
side function, see Moes et al. (2002), is used. This element-by-element propagation
of a crack has been successfully used in Song et al. (2006) for several simulations of
crack propagation, with only minor mesh dependency. They also noted that, although
the crack speed might be overestimated by the element-by-element propagation of
a crack, its propagation path is usually in reasonable agreement with experiments.
After crack initiation, perfect brittle damage evolution is assumed, meaning that the
transferable stresses drop almost immediately to zero (the displacement from crack
initiation to stress-free crack surfaces is set to 10−5 mm, see Fig. 2b). To permit
subsequent stress transfer over crack surfaces, a node-to-surface contact formulation
with finite sliding is used for fully cracked elements.
2.2 Limit Analysis (LA)
As opposed to conventional finite element methods, XFEM included, LA is a ‘direct

method’, focusing only on the moment of impending plastic collapse. All deforma-
tion up to this point is assumed to be of the same order of magnitude as the elastic
deformation, so that the influence of geometry changes on the collapse load can be
neglected. In the state of impending plastic collapse, all elastic strain rates in the
considered body vanish and it behaves like a rigid–plastic rather than an elastic—
plastic body. For this reason, only the strength behavior of each material phase is
required, and the need to specify deformation properties (which are often difficult
to determine) can be avoided. This argument was first presented by Drucker et al.
(1951), together with the first complete formulations of both the lower bound and
upper bound theorems of limit analysis in Drucker et al. (1952), which can be stated
informally as follows, according to Lubliner (1990): “The loads that are in equilib-
rium with a stress field that nowhere violates the yield criterion do not exceed the
collapse loads, while the loads that do positive work on a kinematically admissible
velocity field at a rate equal to the total plastic dissipation are at least equal to the
collapse loads.” The required key assumption behind these two theorems is the prin-
ciple of maximum plastic dissipation, which requires a perfectly plastic material with
a convex yield surface and associated plastic flow (normality of incremental plastic
strain rates to the yield surface). If the loads resulting from the two theorems are
equal to each other, then the true collapse load has been determined. If not, rigorous
lower and upper bounds on the exact solution are provided, giving an immediate
error indication for the result. Design approaches based on limit analysis represent
very efficient methods with a reliability that is often sufficient for many civil and
mechanical engineering problems such as structural design and safety assessment,
and geotechnical capacity calculations.
An intuitive statement of the upper bound theorem could read: The considered
material structure will collapse under a given loading if an admissible velocity field
can be found forming a failure mechanism dissipating less power than the rate of
work of the given loading. Thereby, a velocity field is considered to be admissible
when it obeys, in every point of the material structure, the associated flow rule. The
finite element discretization for the upper bound problem uses triangular elements
with quadratic shape functions to approximate the velocity field, as proposed in
Makrodimopoulos and Martin (2005a, b), resulting in a linear variation of strain rate
within the element. If the vertices of the element are used to enforce the flow rule
and evaluate the plastic dissipation, the solutions obtained are strict upper bounds on
the exact collapse load, as stated in Makrodimopoulos and Martin (2005a, b, 2007).
An intuitive statement of the lower bound theorem could read: A body will not
collapse under a certain loading if an admissible stress field can be found that is in
equilibrium with a loading greater than the considered loading. Thereby, a stress field
is considered to be admissible when it does not violate the yield criterion at any point
of the material structure. To obtain a discretized version of the lower bound problem,
the stress field is approximated by a piecewise linear variation using triangular finite
elements. Each element has its own set of three internal stress evaluation nodes,
thus multiple nodes from adjacent elements may share the same coordinates. At the
resulting stress discontinuities between elements, only the continuity of normal and
shear stress components needs to be enforced; a jump in the stress component parallel
to the discontinuity is allowed.
Finally, the discretized forms of the lower and upper bound optimization problems
can be solved efficiently by an interior point method. In this work, the algorithm
outlined in Andersen et al. (2003), which has been implemented in the optimization
software MOSEK (2006), was used.
2.3 Elastic Limit (EL) Approach
Unlike the XFEM and LA approaches described before, the EL approach used within
this work is solely based on linking the effective stiffness properties of the con-
sidered (loaded) material with the stiffness properties of the underlying material
phases. Therefore, depending on the observation scale, either continuum microme-
chanical schemes or analytical solutions for unit cell structures, see Suquet (1987),
are employed. Finally, effective stress states of the loaded material structure can
be directly related to microscopic strains in an arbitrary material phase. Under
the assumption that failure of a certain material phase is responsible for failure of
the whole material structure, critical strain states in such a phase can be identified
and, subsequently, effective failure stress states of the whole material structure are
obtained.
Continuum micromechanics.
Within the framework of continuum micromechanics, a material is seen as a micro-
heterogeneous body filling a macro-homogeneous representative volume element
(RVE) at different scales of observation. The microstructure within such an RVE
can be described by so-called material phases, representing quasi- homogeneous
subdomains with known physical characteristics such as volume fractions, charac-
teristic shapes, and mechanical properties. The size of the inhomogeneities defined
by the material phases has to be significantly smaller than the characteristic length
of the RVE, and the size of the RVE again has to be smaller than the characteristic
dimension of the structure built up by the material, or the RVE representing the next
higher scale of observation, respectively. In such a way, multiscale homogenization
schemes are set up.
In continuum micromechanics, the effective stiffness Ceff of an RVE can be linked
to its microstructure according to Zaoui (2002) and Benveniste (1987), based on the
solutions in Eshelby (1957) and Laws (1977), as follows (Hofstetter et al. 2008):

N
Ce f f f i ci : Ai with
i1
⎧ ⎫−1
⎨N
⎬
−1 −1
Ai I + Pi0 : (ci − C0 ) : f j I + P0j : (c j − C0 ) (2)
⎩ ⎭
j1
where ci and f i denote the elastic stiffness and the volume fraction of phase i of
total N material phases, respectively, Ai is the concentration tensor of phase i and I
is the fourth-order identity tensor. Ceff relates effective/macroscopic RVE-related
stresses to corresponding strains E, while ci relates (average) phase stresses
σ i (1/Vi ) Vi σ (x)d V to (average) phase strains εi (1/Vi ) Vi ε(x)d V . x is
the location vector for positions inside the RVE, σ (x) and ε(x) are the stress and
strain fields inside the material phases, and V i is the volume of phase domain i inside
the considered RVE. The fourth-order tensor Pi0 accounts for the characteristic shape
of phase i in a matrix with stiffness C0 . The choice of this stiffness describes the inter-
actions between phases. If C0 describes one of the phase stiffnesses, a material with a
contiguous matrix with inclusions is represented and Eq. (2) renders the Mori-Tanaka
scheme as given in Benveniste (1987), while for C0 Ce f f , a dispersed arrangement
of the phases is represented and Eq. (2) reflects the self-consistent scheme.
Unit cell homogenization.
If the material microstructure is characterized by the spatial variation of physical

quantities which can be represented by a combination of local fluctuations at the
level of the elementary cell and a spatial repetition of this elementary cell, a unit cell
homogenization scheme may be employed, also applicable to LA, as can be found
in Füssl et al. (2008). The honeycomb structure inherent to wood (see Fig. 1, top)
represents such a material microstructure, which can be built up/approximated using
identical basic repetitive units. Moreover, the length of these cells is two orders of
magnitude larger than their cross-sectional dimensions, justifying the assumption of
an infinite extension of the unit cells in the longitudinal direction (i.e., conditions of
plane strain). Thus, the unit cell homogenization methods proposed in Suquet (1987)
and Böhm (2004) can be applied. Thereby, the unit cell is subjected to periodic,
symmetry, or antisymmetry displacement boundary conditions such that the spatial
averages of the corresponding microscopic strains ε(x) are equal to the macroscopic
strains E related to the cellular material. Linking these macroscopic strains to the
spatial average of the periodic microstresses they induce, i.e., to the macroscopic
stresses , yields the homogenized effective stiffness Ceff of the cellular material.
The effective stiffness expressions, as functions of the unit cell characteristics, and
the original application of this method to wood can be found in Hofstetter et al.
(2006). In contrast to continuum micromechanics, potential plate-like bending and
shear deformations of the cell walls, building up the material structure, can be taken
into account with this homogenization method.
Elastic limit estimates.
Both of the described methods, continuum micromechanics as well as unit cell

homogenization, can be used, either alone or in combination, to determine the effec-
tive stiffness Ceff of a material structure as a function of its microstructural charac-
teristics. Let us now assume that the intensity of shear loading of the weakest phase
is responsible for failure (an elastic limit state) of our considered material struc-
ture. Then, we would need an appropriate stress or strain measure for this material
phase and a corresponding failure criterion. In Suquet (1997) such a strain measure
is proposed, in fact, quadratic strain averages (second-order moments) are assumed
to appropriately represent (failure-causing) local strain peaks in a material phase.
Through the described homogenization scheme, such strain measures can be linked
to the macroscopic stresses , and in combination with a failure criterion for the
weakest phase, effective elastic limit stress states and, subsequently, an effective
failure criterion can be obtained.
3 Multiscale Model for Wood
The most common wood species used in engineering constructions is spruce, thus, it
is a very well investigated material with a large body of experimental data about its
microstructural characteristics and the properties of its constituents. For the results
presented in this chapter, the composition of spruce wood was taken from Eber-
hardsteiner et al. (2005), in which a comprehensive micromechanical model for the
elasticity of wood is presented, and many references to experimental data for the
wooden microstructure are given. Those data and similar models were also used in
further works, e.g., in Hofstetter et al. (2006), presenting an improved microme-
chanical model for the elasticity of wood, in Hofstetter et al. (2008), proposing an
extension to obtain elastic limit states of wood, and in Bader et al. (2011), including
a poromechanical approach in the micromechanical model.
Based on the underlying data of these models, appropriate observation scales have
been defined in the present work, tailored for predicting the effective strength of clear
wood based on its microstructural characteristics. The highest observation scale in
this multiscale approach, illustrated in Fig. 1 on top, represents the clear wood scale,
consisting of earlywood and latewood layers with a typical thickness between 0.3 and
3.5 mm. Considering the stem cross-section of a tree, the earlywood and latewood
layers form a pattern of concentric circles around the pith, known as growth rings.
T indicates the tangential direction and R the radial direction with respect to these
growth rings, while L shows the longitudinal direction (stem direction).
At the next scale of observation in Fig. 1, the annual ring scale, the earlywood and
latewood layers consist of periodically arranged honeycomb cells, and thus they can
be appropriately represented by the shown unit cells. For the earlywood and latewood
layers, two different unit cells have been defined. In reality, the cell wall consists of
five different layers (often referred to as the middle lamella, P, S1, S2, and S3 layers).
The volumes of the P, S1 and S3 layers are very small compared with the S2 layer
and the middle lamella. For this reason, only two layers are distinguished in the unit
cell for the numerical approaches: the middle lamella (dark grey) and a homogenized
layer (light grey) representing all S layers and the P layer. Each cell wall layer is itself
made up of a non-cellulosic matrix (termed polymer network) in which cellulosic
fibers are embedded, extending spirally at an angle MFA (micro fibril angle) towards
the cell axis. The volume contents of these different material phases, at the different
observation scales, as well as all parameters defining the geometries of the unit cells
can be found in Füssl et al. (2017).
4 Mechanical Behavior of Wood
To obtain the effective elastic behavior at the cell wall scale (cf. Fig. 1, top left side),
Eq. (2) can be applied, with C0 as the isotropic stiffness of the polymer network,
and c1 as the anisotropic stiffness of the cellulosic fibers, which were both calculated
according to the model in Bader et al. (2011). The cellulosic fibers are modelled
as cylindrical inclusions, defined through the concentration tensor Pc0f , which can
be found in Zaoui (2002). The resulting effective cell wall layer stiffnesses define
the elastic behavior of the wood unit cells at the annual ring scale, and thus can
serve as input to XFEM simulations as well as to the unit cell homogenization
within the EL approach. The elegance of this continuum micromechanical approach
is that it efficiently relates the elastic cell wall layer behavior to the behavior of
its microstructural building blocks, like amorphous cellulose, hemicellulose, lignin,
water, crystalline cellulose, and so on. The mechanical properties of these phases are
the same for different wood species, and thus the input parameters for the presented
multiscale concept can easily be adapted for different species of wood. No elastic
material properties are required for the LA approaches.
The strength behavior assigned to the wood cell structures is explained in the
following. In the course of the EL approach in Bader et al. (2011), lignin is assumed
to be the weakest phase in the wood material structure, failing in shear once a certain
threshold is achieved, see Zimmermann et al. (1994). Thus, an appropriate failure
criterion reads
d ∧ d d
Fω (εω ) Flignin (εlignin ) (2μlignin εlignin )2 − (τlignin )2 (3)
where τlignin denotes the lignin shear strength, assumed to be 14.3 MPa according
to Bader et al. (2010) and Lukacevic et al. (2015a, b). Since it can be assumed that
failure of the middle lamella is primarily caused by failure of its main constituent
lignin, the shear strength of lignin (and the Von Mises-type failure criterion) was also
assigned to the middle lamella of the unit cells for the numerical approaches.
In contrast to the EL approach, for the numerical approaches the strength behavior
of the explicitly modelled S-layer also needs to be defined. This layer resembles a
unidirectional fiber-reinforced composite, a material for which several failure criteria
have been developed. Following Puck’s failure criterion, see Puck and Schuermann
(1998), which identifies fiber failure and inter-fiber failure in a unidirectional com-
posite, the following three criteria have been assigned to the S-layers, indicating
crack initiation within the XFEM approach in Lukacevic et al. (2015a, b),
σL
f sl,1 (σ ) ≤ , (4)
σ Lsl
σ
f sl,2 (σ ) ≤ , and (5)
σsl
2 2
τ⊥ τL
f sl,3 (σ ) ≤ + . (6)
τ⊥
sl⊥
τL
sl
Equation (4) describes Mode I fiber failure in L-direction, whereas Eqs. (5) and (6)
both describe inter-fiber failure, leading to a fracture plane with its normal pointing
toward the local -direction (the symbols and ⊥ respectively denote the local
parallel and perpendicular directions in the ‘numerical’ unit cell). The corresponding
tensile and shear strength values were chosen in the same way as in Lukacevic et al.
(2015a, b).
For strength definition in the LA approach, two different strategies were followed.
On one hand, the in-plane failure criteria as used in the XFEM approach, Eqs. (5)–(6),
were assigned to the S-layers, and on the other hand, a single 2D anisotropic Tsai-Wu
yield criterion, see Tsai and Wu (1971), was used, reading
σ2 σ2 τ⊥
2
f sl,T W (σ ) ⊥2 + 2 + 2 ≤ 1. (7)
σ⊥sl σsl τ⊥
sl
Both sets of failure criteria gave almost identical results, but due to the slightly
lower numerical effort of the Tsai-Wu criterion, it is used for all LA simulations
presented in this chapter. All the criteria from Eqs. (5) to (7) can be cast in a gen-
eral quadratic form, and thus can be expressed as second-order cone constraints, as
required for the LA optimization problems.
Finally, each of the three presented methods delivers predictions of effective
strength for two different scales of observation: the annual ring scale and the clear
wood scale. Selected results from all these computations will be presented in the
following section.
4.1 Effective Strengths at Annual Ring and Clear Wood Scale
In Fig. 3 selected failure modes obtained with the upper bound LA approach (Fig.
(a1) and (a3)) and the XFEM approach (Fig. (a2) and (a4)) for wood unit cells as well
as selected XFEM results at the clear wood scale (Fig. (b1) and (b2)) are shown. The
LA method seems to capture the basic failure characteristics quite well, see also Li
et al. (2018a), and shows qualitatively similar failure behavior to that delivered by the
XFEM approach. At this point, it is thus shown that the application of LA approaches
to wood might be reasonable, and taking their great efficiency into account, even very
promising. All strength parameters used as input to these calculations as well as the
exact geometric definitions of the considered unit cells can be found in Füssl et al.
(2017).
By varying the effective (macroscopic) loading situation applied to the unit cell
under consideration, either the honeycomb unit cell at the annual ring scale or the
layered unit cell at the clear wood scale, effective failure surfaces for those unit
cells can be obtained. Within the XFEM approach, loading is applied through a
prescribed displacement field, fulfilling periodicity, at the boundary of the relevant
unit cell. The maximum attainable reaction force at the boundary, for a certain load
situation, defines the corresponding effective failure stress state. The resulting failure
stress states at the clear wood scale, in the radial-tangential stress plane, are shown
in Fig. 3c. There, the colors indicate the type of failure (the yellow/red colormap
denotes brittle and the green/blue one ductile failure) as well as the magnitude of
the in-plane shear component τ RT (red and blue ends of the colormaps account for
high in-plane shear stresses), and the size of the points correlates to the magnitude
of longitudinal shear (the bigger the points the higher the sum of τ L R and τ L T ). This
6D point cloud of failure stress states can be approximated by means of 8 Tsai-Wu
failure surfaces, as shown in Lukacevic et al. (2015a, b), delivering comprehensive
information about when and how clear wood fails.
Such effective failure stress states can also be obtained, in a more efficient way,
with the proposed LA approach and through elastic limit estimates. Within the upper
bound LA approach, an effective strain rate tensor is varied and effective failure
stress states are obtained by taking the mean value of the microscopic stresses of
the optimal solution. The lower bound LA approach leads to different failure stress
states through variation of the effective traction at the boundary. Finally, using the
EL approach, effective elastic limit stress states can easily be obtained as described
in the last paragraph of Sect. 2.3. In Fig. 4 the resulting effective failure surfaces are
compared to each other and to experimental results, obtained through bi-axial testing
of clear wood in Eberhardsteiner (2002), in the longitudinal (L)–radial (R) plane.
A very satisfactory result is that all methods, without any empirical calibration, are
able to predict the correct magnitude and orientation of the experimentally-obtained
effective strength region. The LA results almost perfectly form an outer envelope
to this region. Since no significant hardening effects were encountered during the
experiments, linear-elastic perfectly-plastic failure should lead to the highest effective
strength, which exactly corresponds to the underlying assumptions of LA. By means
of the LA results, the interpretation of the experimental results can now be extended.
As it is stated in Denton and Morley (2000): ‘A structure does not need to exhibit
perfect plasticity for the theoretical plastic collapse load based on the peak yield
stress of each component to be approached closely. Rather, it is necessary that, at
the point when a collapse mechanism forms under a particular loading, all those
regions within the structure which are undergoing straining lie very close to the
(a1) (a2) (b1)
(a3) (a4) (b2)
(c)
Fig. 3 Selected failure modes for wood unit cells obtained with the LA approach (a1 and a3) and
the XFEM approach (a2 and a4), as well as selected failure modes at the clear wood scale obtained
with the XFEM approach (b1 and b2). c Resulting failure stress states at the clear wood scale in
the radial (R)–tangential (T) plane
experiments elastic limit estimates XFEM limit analysis
Fig. 4 Failure stress states of clear wood in the longitudinal (L)–radial (R) plane. Comparison of
experimental results (Eberhardsteiner 2002) and effective failure surfaces obtained with the three
different methods presented
peak yield stress which they can achieve.’ This statement may help explain the large
fluctuation of the experimental strength values in tension-dominated regions. Even
if the structural response of wood is known to be quite brittle under tensile loading,
it seems that the microstructure of wood has a high capacity to redistribute stresses
before ultimate failure. For this reason, the failure loads can (and do) reach values
that would be expected for purely ductile failure. This supports the previously made
assumption that wood has the ductile potential to ‘activate’ the strength of many
points along a potential crack surface before brittle failure occurs. To what extent
this potential can be exploited depends on how homogeneous and defect-free the
considered specimen is. The natural variation of such defects may then introduce the
experimentally-observed scatter of effective strengths.
Interestingly, the EL estimates deliver a very good inner envelope of the exper-
imental results, and justify their name at this scale of observation. Thus, the EL
approach has the potential to be an ideally complementary method to the LA
approach. And as Fig. 4 demonstrates, together they could be highly effective tools
for predicting the effective strength of wood under multidimensional stress states.
Based on the results obtained, the three computational methods proposed can be
generally evaluated as follows:
– The XFEM approach is by far the most extensive with respect to computational
effort and algorithmic implementation. This is not surprising, since this method
includes the largest description flexibility, and thus has the potential to reproduce
the processes in nature most realistically. It will not, however, be available for
quick engineering estimates or for the build-up of material databases for wood
species or products in the near future.
– Compared to XFEM, with the LA approach basic characteristics of failure modes
and the overall strength behavior could be captured correctly, both qualitatively and
quantitatively. Along with the observed computational efficiency and numerical
stability, these results render the LA approach as a very promising tool for this
kind of application in the future. A new upper bound approach, including adaptive
velocity discontinuities, can be found in Li et al. (2018b).
– The continuum micromechanical basis of the EL approach makes it unbeatable
with respect to efficiency and the elegance with which material properties of dif-
ferent phases can be linked across several scales of observation. However, when
it comes to strength estimates, stress and strain averages over material phases
are perhaps not able to capture all failure-triggering processes. Nevertheless, the
potential of this method is undisputed and due to its efficiency, it may become
invaluable for rapid strength predictions within engineering design processes.
Having a mechanically sound characterization of clear wood at hand, the next
step is to implement this information into design tools for wood-based products.
Therefore, it is necessary to identify the 3D morphology of wooden boards, building
up many commonly-used wood-based products. How this can be done is explained
in the next section.
4.2 Morphology Identification at Wooden Board Scale
In contrast to other building materials wood is a highly inhomogeneous material

due to the natural growth process, and thus, exhibits large variations of mechanical
properties in different directions, see Steffen et al. (1997). The reasons are mani-
fold, ranging from a ‘simple’ variation in density up to changing microfibril angles
at the wood cell wall scale, all of them influencing the mechanical properties of
wood, see Persson (2000) and Kandler et al. (2015b, 2018). However, one of the
most prominent sources for the variation of mechanical properties are knots and the
resulting fiber deviations around them, as discussed in Johansson (2003), wherein it
is also concluded that failure in wooden structural elements is mainly observed in
the vicinity of knots. For this reason, 3D morphological information about knots and
fiber distributions are an absolute necessity for a reliable performance prediction of
wooden elements.
To obtain such morphological information, conventional wooden boards, rep-
resenting the main building element of many different wood-based products and
structural elements, are standardly sent through laser scanning devices to get a rough
estimate about the inhomogeneity of a board. Non-standard and currently in scien-
tific focus is exploiting the tracheid effect, which describes the transformation of
an originally circular laser dot to an elliptical shape when hitting the surface of a
softwood sample, see Nyström (2003). Since light spreads further in longitudinal
direction than in lateral direction, the major axis of the ellipse will be oriented paral-
lel to the fiber direction, giving access to 2D fiber angle information on all surfaces
of a wooden board. Such a fiber angle distribution for one side of a wooden board is
exemplarily illustrated in Fig. 5a. In addition, photographs of the cross sections on
both ends of the board allow for an estimation of pith locations. The fully automated
approach described in Kandler et al. (2016) can be used to obtain a geometrical
description of the knot morphology. Therein, the fiber angle measurements are used
to estimate knot areas on the board’s surfaces. This information in combination with
the pith location is used to fit rotationally symmetric cones, representing knots, into
the board (see Fig. 5f). Thereby, the reconstruction error is minimized by employing
a so-called simulated-annealing optimization scheme, which reduces the difference
between actual knot areas and knot areas obtained from intersecting the reconstructed
cones with the board. In Fig. 5b all recognized knots within the exemplary board
are displayed, whereas in Fig. 5c only the significant knot groups according to a
modified version of the criterion are shown. Here, significance is defined by using a
size measure for the knots. Knots with a visible knot area above the 70%-quantile are
considered to be large knots and, thus, to have a significant influence on the board’s
mechanical behavior. In addition, the mutual distance between adjacent large knots is
checked and those with longitudinal distances below 200 mm are grouped together.
All other (small) knots are now disregarded unless they lie in close vicinity of larger
knots, defined as a maximum distance of 100 mm. The reason for this approach is that
small knots are considered to be of small mechanical significance if by themselves,
but they might have an influence on sections, where interacting multiple knots lead
to variations in fiber deviations and, thus, influences on failure mechanisms.
In a next step, the three-dimensional fiber angle course within the wooden board
can be computed based on a numerical model proposed in Foley (2003). Therein, the
fiber angles in the radial-tangential plane are assumed to follow a course analogous
to the flow of a laminar fluid around an obstacle. Fiber angles in longitudinal-radial
planes are estimated through polynomials previously fitted to year ring distributions.
By means of this 3D fiber angel course in combination with the effective stiffness
tensor for clear wood, obtained with the multiscale model proposed above, longi-
tudinal stiffness profiles of wooden boards can be computed from cross-sectional
averages. One such longitudinal stiffness distribution is shown in Fig. 5d, showing
distinct stiffness drops in regions where knots are present. Alternatively, effective
board stiffnesses can be calculated section-wise by means of 3D finite element mod-
els, exploiting the morphological information obtained before.
To get access to strength distributions, as shown in Fig. 5e, the board is divided
into so-called knot sections, with a maximum length of 300 mm and interjacent clear
wood sections without knots. Each knot section is then subjected to a set of indicat-
ing properties (IPs), proposed in Lukacevic et al. (2015a, b). The IPs are empirically
derived quantities, relating morphological characteristics, like knot areas, volumes,
or interfaces areas, to effective strength values of knot sections. For the clear wood
sections, the longitudinal tensile strength is computed following the approach in
Olsson and Oscarsson (2014), which is based on scaling experimentally obtained
strength values according to density-dependent results obtained by a micromechan-
ical model for the elastic behavior of wood as proposed in Sect. 2.3. The resulting
strength profiles for the exemplary wooden board are displayed in Fig. 5e. Such
profiles of alternating defect-free sections and knot sections are referred to as weak
zone models.
Fig. 5 a Wood fiber distribution on a conventional wooden board obtained from laser scanning, b
reconstructed knots by means of the algorithm in Kandler et al. (2016) f, c significant knot groups
triggering failure behavior, d longitudinal stiffness distributions of the board, and e different strength
profiles of the board obtained by applying indicating properties (IPs) according to Lukacevic et al.
(2015a, b)
4.3 Effective Strengths at Wood-Product Scale
By combining the 3D information about the mechanical behavior of clear wood

obtained in Sect. 4.1 with the 3D morphological information about wooden boards
obtained in Sect. 4.2, it becomes possible to build up realistic digital models of
glued-laminated timber (GLT) as well as cross-laminated (CLT) products. Through
application of finite-element-based methods, like the XFEM and LA approach briefly
presented in Sects. 2.1 and 2.2, respectively, accurate predictions of the effective
mechanical behavior of such products can be obtained. Thereby, highly non-linear
effects, like brittle cracking, as well as plastic deformation in pressure-dominated
regions, as described in Hochreiner et al. (2014a, b), can be described quite realis-
tically. As illustrative example, the failure mode of a unit cell, representing a cross-
laminated timber element under pure bending load, is shown in Fig. 6. It can be seen
(a) (b)
Fig. 6 a Accurate description of cracking around knots by means of an XFEM approach, simu-
lating the mechanical behavior of a CLT unit cell under bending load. b Related CLT element and
visualization of the crosswise arranged (glued) wooden boards
Fig. 7 CLT plate under bending load investigated by numerical LA, considering longitudinal
strength variations (values in MPa) within wooden boards (a, c). The resulting deformation (b)
and the failure mode (plastic regions) across the plate in exploded view (d, e)
that by the XFEM approach complex failure mechanisms around knots, characterized
by discrete cracks, can be described appropriately.
Beside such unit cell calculations, where the 3D structure of a wood-based product
is resolved in detail, the longitudinal stiffness and strength profiles shown in Fig. 5
can be used to consider the material variability within whole building elements. In
Fig. 7a, c the modelled longitudinal strength variation within a CLT structural element
is shown, where the section-wise different strengths are obtained through analysing
140 individual boards of a certain strength class (C24) as described in Sect. 4.2.
The single boards were then divided into several knot and clear-wood sections
with assigned strength parameters determined by the board’s strength profile infor-
mation. Through a random board selection process, an arbitrary number of different

CLT plates can be assembled using wooden boards with different strength profiles,
and their failure behaviour under bending was predicted using the numerical LA
approaches described in Sect. 2.2. An exemplary result of these calculations is shown
in Fig. 7b, and, in order to get a complete image of the failure pattern across the whole
plate, exploded views showing the top and bottom of each layer are given in Fig. 7d, e.
The advantage of this approach is the high computational efficiency of numerical
LA, making it possible to run hundreds of simulations and, in this way, studying the
influence of section-wise varying strength values using a strength profile database.
4.4 Stochastic Aspects
The proposed stiffness and strength profiles also give access to the application of
stochastic concepts and metaheuristic optimization algorithms to wooden structural
elements, as schematically shown in Fig. 1 (bottom left).
Thereby, common stochastic approaches consist of two parts: (i) The develop-
ment of an appropriate random process model based on previously acquired data
(realizations), in our case stiffness or strength profiles, and (ii) the implementation
into a stochastic FE model where the mechanical and stochastic problem is coupled.
The discretization of the generated random process can be done either with a spa-
tial discretization method or using a serial expansion. These discretization methods
can be implemented into different ‘closed’ stochastic FE formulations, e.g. (i) the
perturbation method, where the stochastic system matrix and the response vector
are expressed as Taylor series expansions, and (ii) a spectral approach, where the
stochastic part of the system matrix is written as a sum of certain ‘basis functions’.
The application of these methods to a GLT element, in Füssl et al. (2016), Kandler
et al. (2015a), and Kandler and Füssl (2017), has shown that both methods are able to
capture important effects, such as lamination effects (see Fig. 8), and deliver appro-
priate effective stochastic information, similar to the Monte-Carlo simulation, but
with much less computational effort.
Another utilization of stiffness and strength profiles represents the optimization of
the assembly of wooden boards within wood-based products by using metaheuristic
optimization algorithms. Based on the fundamental properties of them described in
Blum and Roli (2003), the most suitable class of metaheuristic algorithms for the
present combinatorial problem seems to be genetic algorithms, see Goldberg (1989),
which are also a subclass of evolutionary algorithms. The starting point of every
genetic algorithm is a population of individuals, in our case, for example, represented
by a generated sample of GLT beams. In the following, this sample of GLT beams
will adapt and improve over successive generations through concepts of natural
selection, crossover, and mutation. During the selection phase, strong individuals
(GLT beams) from the current population are selected to form the next generation.
The crossover phase generates the offspring based on the parent population and
the mutation phase assures that the genetic diversity is maintained. Finally, a near
Fig. 8 Longitudinal stiffness probability density functions (PDFs) for two classes of wooden boards
and the corresponding PDFs for 4-layered GLT elements, showing much smaller coefficient of
variations (COV), and thus, reflecting the so-called lamination effect
optimal population will be achieved, meaning that the assembly of wooden boards
with respect to the desired performance characteristics (e.g., the maximum deflection
or effective bending strength) is optimal, and thus, the raw material is ideally utilized.
In Pech (2017) and Pech et al. (2018) it could already be shown that by means of
such algorithms the performance of a GLT beam sample regarding its maximum
deflection can be improved up to 30%.
5 Elastically-Deformed Wooden Structures
Based on sound mechanical models of how wooden structural elements behave under
3D stress states, it now makes sense to move up to the next higher scale of observa-
tion—the structural scale. By again applying computational methods at this scale, the
whole information presented before can theoretically be implemented and exploited
in modern design procedures. This hasn’t been achieved so far, however, in the fol-
lowing a finite-element-based concept for investigating elastically-deformed wooden
structures is proposed, showing in which way a digital design procedure can be linked
to a powerful prediction tool for the mechanical behavior.
Elastically deformed structures offer great possibilities for free-form architecture.
The initial stage of design usually involves creating a design surface, defined by the
requirements the structure has to fulfill. Using an appropriate geometric concept,
suitable deformed structural members can be found and arranged on it. Thereby, the
deformed shape of these members can be reached by bending and twisting them.
Describing the behavior of a structure designed this way is difficult, since it can be
assumed that the desired form of the structure does not correspond to a state of equi-
librium. Thus, to evaluate the quality of a design, computing its closest equilibrium
position delivers important information.
The interaction of multiple deformed members can result in unexpected structural
behavior. Since all members can be seen as springs trying to obtain a state of lower
energy, their spatial configuration belonging to an equilibrium position is unknown.
This also applies for their intersection points. For this reason, the initially-designed
form of the structure may not be viable. In general, motions of intersection points
lead to undesired deviations from the designed form and possibly also to buckling of
members or parts of the structure. Analyzing members individually is an improper
simplification because of their interconnectedness. Finding an equilibrium position
for such a structure is an iterative process, where the initially designed form can
serve as a starting point.
In the following, a procedure of how this analysis can be done efficiently and with
high geometric precision is proposed. The main steps of this analysis are implement-
ing the deformation process, defining the interactions of the members and carrying
out a relaxation process.
5.1 Design Procedure
By using finite element analysis, this problem is approached with a sound mechan-
ical method. Thereby, a set of points is defined on every un-deformed member and
then moved (displacement-controlled) to the position of the desired deformed mem-
ber, using multiple steps. As soon as all members are successfully deformed con-
tact between them is initiated. Subsequently, the applied displacement conditions to
deform the members are removed, giving the structure the possibility to find a state
of equilibrium.
To carry out this procedure, accurate information describing the deformation pro-
cess is required. To generate this data, the CAD Software Rhinoceros and the embed-
ded editor Grasshopper were used. The visual programming language of Grasshopper
offers designers as well as engineers an easy-to-use programming environment. Since
the amount of data generated to describe the deformation can become very large, the
usage of a pre-processing module within a finite element software becomes inef-
ficient or impossible. Within the proposed concept the pre-processing module of
Abaqus and Grasshopper is used.
To generate the necessary data for simulation, an input file is created using the
pre-processing module. It contains material properties, finite-element mesh infor-
mation of members, the assembled model, and interaction definitions. Subsequently,
information to carry out the deformation as well as the relaxation process is added.
Since Grasshopper can process text, this information can be written directly in text
form. This whole process is completely automated. Finally, the modified input file
can be sent to the solver.
Fig. 9 Left: design surface and prototype structure (supports clamped and highlighted in black),
right: deformation path of the upper edge of a member and deformation vectors for a second member
Elastically deformed wooden structures can be constructed in multiple ways:

Assembled in a plane and deformed globally or deformed individually and then
assembled. The presented method was developed for the latter case. In the following
section an example will be presented. Nevertheless, the method can be applied to
structures constructed in the former case as well.
5.2 Deformation Data
The structure in Fig. 9 is based on a design surface and geodesics, describing the
centerlines of the deformed members. Designed this way, each member experiences
uniaxial bending and torsion. The deformed members can be created by constructing
tangential surface strips to the design surface for each geodesic and assigning a
thickness to them. The maximum applicable thickness depends on the mechanical
properties of the material, curvature and torsion in each point of the centerline. For
the non-deformed member a straight strip with the same size is created.
There is a unique deformation path for every point, mapping it from the straight
strip to the tangent strip without stretching or compressing the strip. These spatial
curves can be constructed as trajectories of the rolling straight strip on the tangent
strip, ignoring any further motion of points once they have reached the tangent strip.
Therefore, these curves can be considered as involutes of the tangent strip. The set of
points used to enforce the deformation on the member should be distributed evenly
across the strip. It is reasonable to use intersections of the observed geodesic with
other geodesics of the network and some additional points, depending on the total
quantity of torsion and curvature of the member (see the two highlighted members
in Fig. 9). The discretization of the deformation paths plays an important role. Using
chords results in undesired shortenings of the member rolling along the discrete path.
This may lead to buckling. Therefore, it is strongly recommended to use a sufficient
number of segments tangent to the deformation path.
The sets of deformation vectors now have to be assigned to the right points. To
make this possible, special attention has to be put on the meshing process during
pre-processing, since chosen points need to be vertices of the member’s mesh. Algo-
rithmically comparing the vertex positions with the desired point positions, the vertex
index can be read from the input file. Subsequently, the necessary information linking
the deformation vectors to the desired point positions can be generated and written
in the modified input file.
5.3 Interaction and Relaxation
The interaction of the structural members in Fig. 9 is modelled by using the options
available in Abaqus for surface interaction. Separation after initial contact is not
allowed, which makes the transfer of forces trying to separate two members possi-
ble. For models with many layers, modelling the contact with pin joints is another
option. After having deformed each member and brought them to contact, the bound-
ary conditions introduced to perform the deformation process have to be suspended,
causing the structure to relax. To ensure the stability of the simulation during relax-
ation, it should be carried out using multiple steps.
A deformation analysis for the structure of Fig. 9 is shown in Fig. 10. The initial
design is highlighted in grey, the simulated deformation is highlighted in green. In
some parts of the structure, significant deviations from the designed form can be
observed. On the right upper side a whole region of the structure has buckled during
the relaxation process of the members. The structure’s unsymmetrical behavior can
be explained by necessary offsets using multiple layers of members.
Carrying out deformation analysis for an elastically deformed structure provides
valuable information to evaluate an initial design. Undesired deviations from the
initial form and tendencies of buckling can be determined. This enables the designer
to adapt the initial design, e.g. by adding more members or changing properties of
the design surface.
Finally, the arising 3D stress states within the structure during the deformation
process can be assessed by means of the multi-surface criteria presented in Sect. 4.1,
allowing for the identification of critical regions where brittle cracking or high plastic
deformation possibly takes place. Furthermore, in a subsequent step the structural
behavior under different live loads could be investigated realistically, since stresses
due to the deformation process are already taken into account.
Fig. 10 Left: comparison of initial design (grey) and relaxed structure (green), right: stress distri-
bution of von Mises stresses (N/mm2 )
6 Geometric Design Concepts Applicable to Wood
The trend towards free form structures in contemporary architecture and the numer-
ous challenges in realizing such structures led to the development of an interdisci-
plinary research area, called Architectural Geometry. For an overview of this field,
we refer the reader to Pottmann et al. (2015). A relatively large number of contribu-
tions to this area is applicable to wood. We focus here only on a few topics, which
have been motivated by the use of wood. We first continue the subject of the previous
chapter and deal with bent straight strips and then address freeform structures from
flat panels and straight beams.
6.1 Geodesic Strips
A straight strip of material whose deformation behaviour is strongly dominated by

bending rather than stretching can assume shapes which we call geodesic strips in
the following. Such a geodesic strip is a piece of a developable surface and its central
curve c is a geodesic (shortest path) on that surface. The strip surface is part of the
so-called rectifying developable surface of c (see e.g. Pottmann et al. 2007). Hence,
if one attaches such a strip in a tangential way to a surface, the strip follows a shortest
path on that surface. This fact has motivated research on the optimized arrangement of
Fig. 11 Left: design surface segmented into patches which can be covered by geodesic strips.
Middle: wooden gridshell from geodesic strips which are tangential to a surface and arranged in
a tri-hex pattern. Right: Wooden gridshell from geodesic strips orthogonal to a reference surface
(image courtesy E. Schling)
geodesics on surfaces. The applications include panelling and supporting structures

or gridshells.
For panelling, a major optimization criterion has been to find nearly parallel
families of geodesics on a given design surface. If that surface is sufficiently far away
from a developable one, it is usually necessary to segment the surface into patches
along which the geodesics are nearly parallel (see Fig. 11, left). Essential tools
are from differential geometry (geodesic vector fields, Jacobi fields) and numerical
optimization, see Pottmann et al. (2010). A built example for such an arrangement of
geodesic strips is the wooden roof in the entrance hall of the Burj Khalifa by Gehry
Technologies.
To obtain gridshells, one may arrange the strips tangentially along a surface, such
as in Figs. 9, 10 and 11 middle. From a geometric perspective, this leads to the problem
of designing specific patterns of geodesics on a surface. For example, in Fig. 11 we
show a tri-hex pattern from geodesics obtained as level sets of functions defined on
the surface (geopattern). The existence of special patterns from geodesics limits the
possible shapes of the reference surface. For built structures, see e.g. Chilton and
Tang (2017), Naicu et al. (2014) and Pirazzi and Weinand (2006).
In the construction of wooden gridshells, one may arrange the geodesic strips
normal to a reference surface, enabling an elastic assembly via their weak axis and
a local transfer of loads via their strong axis. An example for this approach is seen
in recent work by Schling and Barthel (2017) and (2018), and Fig. 11, right. If a
geodesic strip is orthogonal to a reference surface S, it follows an asymptotic curve
on S, and therefore the surface must have negative Gaussian curvature. If in addition
the strips meet at a right angle (as in Fig. 11), S has to be a minimal surface. For
generalizations of these concepts, see Schling et al. (2018).
6.2 Freeform Structures from Flat Panels and Straight

Beams
Let us assume that we want to realize a freeform surface by flat wooden plates of
constant thickness w. If the appearance of both sides of such a faceted freeform
structure matters (as in Fig. 12, top, left and central), the plates have to be arranged
at a mesh M with planar faces and an exact face offset M w . This means that offsetting
each face plane of M by a constant distance w should result in the face planes of a
mesh M w with the same combinatorics as M. This is only possible if each vertex v of
M is conical, i.e., all face planes meeting at v are tangent to a sphere (equivalently,
tangent to a right circular cone). A vertex v of valence 3 is always conical and
thus patterns with only vertices of valence 3 pose no problems when materializing
them by plates of constant thickness (Fig. 11, top). Meshes from planar quads and
with exact face offsets, so-called conical meshes, are known as discrete versions
of principal curvature parameterizations of smooth surfaces, see Liu et al. (2006),
Pottmann et al. (2007, 2015). Hence, the edges of a conical mesh follow roughly
the principal curvature directions of a smooth underlying surface and the freedom in
designing such a mesh for a given reference surface is limited. A triangle mesh with
an exact face offset is only possible if the reference surface is close to a sphere. More
general meshes with planar faces and the combinatorics of semi-regular patterns have
been studied by Jiang et al. (2015). There, the appearance of faces with more than
4 vertices depends on the sign of Gaussian curvature K. For positive K, the patterns
appear like the corresponding semi-regular patterns in the plane, while significant
shape changes occur in negatively curved regions (see Fig. 12, top). These properties
hold for piecewise flat panelizations where kink angles at edges are so small that they
are good approximations of the underlying reference surface. However, there are
architectural timber structures which exhibit high kink angles by design. Prominent
examples are inspired by Origami. These timber folded plate structures combine the
structural advantages of timber panels with the efficiency of folded plates. For a
detailed discussion, we refer to Stitic and Weinand (2015).
Built freeform structures in which the panels do not transfer loads, need a support
structure. In case of planar panels, the support structure would typically follow the
edges of the base mesh M and consist of prismatic beams. Such a node configuration
can be complicated. It is an advantage to have a torsion-free node, where the central
planes of beams pass through a straight line, called the node-axis (Fig. 12, bottom,
left). However, when working with wood, it may be better to resolve the nodes into
cells. An example is provided by Fig. 12, bottom, where the nodes of valence 3
in the original structure become triangular cells. They are stable and easier to be
manufactured.
Fig. 12 Top, left and central: flat panels following a hexagonal pattern have been used for the
Landesgartenschau Exhibition Hall in Schwäbisch Gmünd, Germany, see Krieg et al. (2014) (images
courtesy A. Menges). Top, right: a more general pattern from flat panels on a freeform surface, see
Jiang et al. (2015). Bottom: a packing of spheres on a surface, see Schiftner et al. (2009), as basis
of a hexagonal support structure with nodes of valence 3. At the Kreod Pavilion (by Pavilion
Architecture, London; Architect: Chun Qing Li), the nodes have been resolved into triangular cells
(images courtesy Evolute)
7 Conclusion
A mechanical modeling approach for wood, depicting a comprehensive strategy

from microstructural length scales, defining the raw material, to the final wood-
based product, was proposed. Focus was thereby laid on the identification of the 3D
morphology of wooden boards, the determination of material properties at different
scales of observation by means of three different computational methods, stochastic
modeling of wood products as well as on a design concept for elastically-deformed
timber structures.
Important motivation of this work has been to get insights into the potential
of multisurface concepts which are able to provide detailed information about the
3D strength behavior of clear wood. Information which cannot be achieved with
identification experiments only, but which is required for the definition of realistic
orthotropic failure envelopes at the clear wood scale, independent on their mathe-
matical description. By relating strength behavior to its microstructural origin, where
failure processes are much more ‘universal’ with respect to different wood species,
the number of strength properties to be identified for an accurate description of
wood should be greatly reduced. Finally, reliable 2D and 3D strength information
for wooden boards as well as wood-based products could be obtained in an efficient
way, which would subsequently help to exploit their full potential in modern engi-
neering structures, as well as serve as strong basis for computational design concepts
for complex timber structures.
Finally, making wood predictable through computational methods would pave
the way to smart algorithmic-driven manufacturing processes following the vision
of Industry 4.0, targeted computer-based development of new wooden structural

elements, and integration of wood analysis into general concepts of computational
design. This will prepare timber industry for the future challenge of our society to effi-
ciently design, produce, and build sustainable infrastructure with all the possibilities
provided by the fast-developing areas of digitalization and cognitive computing.
References
Aicher S, Gustafsson PJ, Haller P, Petersson H (2002) Fracture mechanics models for strength
analysis of timber beams with a hole or a notch. A report of Rilem TC-133
Andersen ED, Roos C, Terlaky T (2003) On implementing a primal-dual interior-point method for
conic quadratic optimization. Math Program Ser B 95:249–277
Bader T, Hofstetter K, Hellmich C, Eberhardsteiner J (2010) Poromechanical scale transitions of
failure stresses in wood: from the lignin to the spruce level. ZAMM - Zeitschrift für angewandte
Mathematik und Mechanik 90:750–767
Bader T, Hofstetter K, Hellmich C, Eberhardsteiner J (2011) The poroelastic role of water in cell
walls of the hierarchical composite softwood. Acta Mech 217:75–100
Belytschko T, Black T (1999) Elastic crack growth in finite elements with minimal remeshing. Int
J Numer Meth Eng 45(5):601–620
Benveniste Y (1987) A new approach to the application of Mori-Tanaka’s theory in composite
materials. Mech Mater 6:147–157
Blum Ch, Roli A (2003) Metaheuristics in combinatorial optimization: overview and conceptual
comparison. ACM Comput Surv (CSUR) 35(3):268–308
Böhm H (2004) A short introduction to continuum micro-mechanics. In: Böhm H (ed) Mechanics
of microstructured materials. CISM Lecture notes 464. Springer, Wien/New York, pp 1–40
Chilton J, Tang G (2017) Timber gridshells: architecture, structure and craft. Routledge, New York
Denton SR, Morley CT (2000) Limit analysis and strain-softening structures. Int J Mech Sci
42:503–522
Drucker DC, Greenberg HJ, Prager W (1951) The safety factor of an elastic-plastic body in plane
strain. J Appl Mech 18:371–378
Drucker DC, Prager W, Greenberg HJ (1952) Extended limit design theorems for continuous media.
Q Appl Math 9:381–389
Eberhardsteiner J (2002) Mechanisches Verhalten von Fichtenholz: Experimentelle Bestimmung
der biaxialen Festigkeit- seigenschaften, Mechanical Behavior of Spruce Wood: Experimental
Determination of Biaxial Strength (Properties), in German. Springer, Wien, New York
Eberhardsteiner J, Hofstetter K, Hellmich Ch (2005) Development and experimental validation of a
continuum micromechanics model for the elasticity of wood. Eur J Mech A/Solids 24:1030–1053
Eshelby JD (1957) The determination of the elastic field of an ellipsoidal inclusion, and related
problems. Proc R Soc Lond Ser A 241:376–396
Foley Ch (2003) Modeling the effects of knots in structural timber. ISSN: 0349-4969
Füssl J, Lackner R, Eberhardsteiner J, Mang HA (2008) Failure modes and effective strength of two-
phase materials determined by means of numerical limit analysis. Acta Mech 195(1–4):185–202
Füssl J, Kandler G, Eberhardsteiner J (2016) Application of stochastic finite element approaches
to wood-based products. Arch Appl Mech 86(1–2):89–110. https://doi.org/10.1007/s00419-015-
1112-6
Füssl J, Li M, Lukacevic M, Eberhardsteiner J, Martin CM (2017) Comparison of unit cell-based
computational methods for predicting the strength of wood. Eng Struct 141:427–443. https://doi.
org/10.1016/j.engstruct.2017.03.005
Goldberg DE (1989) Genetic algorithms in search, optimization, and machine learning, 1989.
Addison-Wesley, Reading
Gravouil A, Moes N, Belytschko T (2002) Non-planar 3d crack growth by the extended finite
element and level sets-part ii: level set update. Int J Numer Meth Eng 53(11):2569–2586
Guindos P (2011) Three-dimensional finite element models to simulate the behavior of wood with
presence of knots, applying the flow-grain analogy and validation with close range photogramme-
try. PhD thesis, University of Santiago de Compostela, Department of Agroforestry Engineering
Hochreiner G, Füssl J, Eberhardsteiner J (2014a) Cross-laminated timber plates subjected to con-
centrated loading: experimental identification of failure mechanisms. Strain 50(1):68–81
Hochreiner G, Füssl J, Serrano E, Eberhardsteiner J (2014b) Influence of wooden board strength
class on the performance of cross-laminated timber plates investigated by means of full-field
deformation measurements. Strain 50(2):161–173
Hofstetter K, Hellmich Ch, Eberhardsteiner J (2006) Micromechanical modeling of solid-type and
plate-type deformation patterns within softwood materials. A review and an improved approach.
Holzforschung 61:343–351
Hofstetter K, Hellmich Ch, Eberhardsteiner J, Mang HA (2008) Micromechanical estimates for
elastic limit states in wood, revealing nanostructural failure mechanisms. Mech Adv Mater Struct
15:474–484
Jiang C, Tang C, Vaxman A, Wonka P, Pottmann H (2015) Polyhedral patterns. ACM Trans
Graph 34(6):#172, 1–12
Johansson CJ (2003) Grading of timber with respect to mechanical properties. In: Timber engineer-
ing, pp 23–43. ISBN: 0-470-84469-8
Kandler G, Füssl J (2017) A probabilistic approach for the linear behaviour of glued laminated
timber. Eng Struct 148:673–685
Kandler G, Füssl J, Eberhardsteiner J (2015a) Stochastic finite element approaches for wood-based
products: theoretical framework and review of methods. Wood Sci Technol 49(5):1055–1097
Kandler G, Füssl J, Serrano E, Eberhardsteiner J (2015b) Effective stiffness prediction of GLT beams
based on stiffness distributions of individual lamellas. Wood Sci Technol 49(6):1101–1121
Kandler G, Lukacevic M, Füssl J (2016) An algorithm for the geometric reconstruction of knots
within timber boards based on fibre angle measurements. Constr Build Mater 124:945–960
Kandler G, Lukacevic M, Füssl J (2018) Experimental study on glued laminated timber beams with
well-known knot morphology. Eur J Wood Wood Prod 1–18
Krieg O, Schwinn T, Menges A, Li JM, Knippers J, Schmitt A, Schwieger V (2014) Biomimetic
lightweight timber plate shells: computational integration of robotic fabrication, architectural
geometry and structural design. In: Block P et al (eds) Advances in architectural geometry 2014.
Springer, pp 109–125
Landelius J (1989) Finit area metoden. en bra metod för beräkning av uppfläkningsbrott? Rep
No TVSM 5043:66 (in Swedish)
Laws N (1977) The determination of stress and strain concentrations at an ellipsoidal inclusion in
an anisotropic material. J Elast 7(1):91–97
Li M, Füssl J, Lukacevic M, Eberhardsteiner J, Martin CM (2018a) Strength predictions of clear
wood at multiple scales using numerical limit analysis approaches. Comput Struct 196:200–216
Li M, Füssl J, Lukacevic M, Eberhardsteiner J, Martin CM (2018b) A numerical upper bound
formulation with sensibly-arranged velocity discontinuities and orthotropic material strength
behaviour. J Theor Appl Mech 56
Liu Y, Pottmann H, Wallner J, Yang YL, Wang W (2006) Geometric modelling with conical meshes
and developable surfaces. ACM Trans Graph 25(3):681–689
Lubliner J (1990) Plasticity theory, Revised Pdf Edition 2006. University of California Berkeley,
Previously published by Pearson Education, Inc.
Lukacevic M, Füssl J (2014) Numerical simulation tool for wooden boards with a physically based
approach to identify structural failure. Eur J Wood Wood Prod 72(4):497–508. https://doi.org/10.
1007/s00107-014-0803-y
Lukacevic M, Füssl J (2016) Application of a multisurface discrete crack model for clear wood
taking into account the inherent microstructural characteristics of wood cells. Holzforschung
70(9):845–853
Lukacevic M, Füssl J, Griessner M, Eberhardsteiner J (2014) Performance assessment of a numerical

simulation tool for wooden boards with knots by means of full-field deformation measurements.
Strain 50(4):301–317. https://doi.org/10.1111/str.12093
Lukacevic M, Füssl J, Lampert R (2015a) Failure mechanisms of clear wood identified at wood cell
level by an approach based on the extended finite element method. Eng Fract Mech 144:158–175.
https://doi.org/10.1016/j.engfracmech.2015.06.066
Lukacevic M, Füssl J, Eberhardsteiner J (2015b) Discussion of common and new indicating prop-
erties for the strength grading of wooden boards. Wood Sci Technol 49(3):551–576
Lukacevic M, Lederer W, Füssl J (2017) A microstructure-based multisurface failure criterion for the
description of brittle and ductile failure mechanisms of clear-wood. Eng Fract Mech 176:83–99
Mackenzie-Helnwein P, Eberhardsteiner J, Mang HA (2003) A multi-surface plasticity model for
clear wood and its application to the finite element analysis of structural details. Comput Mech
31:204–218
Makrodimopoulos A, Martin CM (2005a) Limit analysis using large-scale socp optimization. In:
Proceedings of 13th national conference of UK association for computational mechanics in engi-
neering, Sheffield, pp 21–24
Makrodimopoulos A, Martin CM (2005b) A novel formulation of upper bound limit analysis as
a second-order cone programming problem. In: Proceedings of 8th international conference on
computational plasticity, Barcelona, pp 1083–1086
Makrodimopoulos A, Martin CM (2007) Upper bound limit analysis using simplex strain elements
and second-order cone programming. Int J Numer Anal Meth Geomech 31(6):835–865
Masuda M (1988) Theoretical consideration on fracture criteria of wood—proposal of finite small
area theory. In: Proceedings of the 1988 international conference on timber engineering, Seattle,
vol 2, pp 584–595
Melenk J, Babuska I (1996) The partition of unity finite element method: basic theory and applica-
tions. Comput Methods Appl Mech Eng 139(1):289–314
Moes N, Gravouil A, Belytschko T (2002) Non-planar 3d crack growth by the extended finite
element and level sets-part i: mechanical model. Int J Numer Meth Eng 53(11):2549–2568
MOSEK ApS (2006) The MOSEK optimization tools version 4.0 (revision 35). User’s Manual and
Reference, available from http://www.mosek.com
Naicu D, Harris R, Williams C (2014) Timber gridshells: design methods and their application to
a temporary pavilion. In: World conference on timber engineering, Vienna
Nyström J (2003) Automatic measurement of fiber orientation in softwoods by using the tracheid
effect. Comput Electron Agric 41(1):91–99
Olsson A, Oscarsson J (2014) Three dimensional fibre orientation models for wood based on laser
scanning utilizing the tracheid effect. In: WCTE 2014, World conference on timber engineering,
Quebec City, Canada, August 10–14
Osher S, Sethian JA (1988) Fronts propagating with curvature-dependent speed: algorithms based
on Hamilton-Jacobi formulations. J Comput Phys 79(1):12–49
Pech S (2017) Metamodel assisted optimisation of glued laminated timber systems by reordering
laminations using metaheuristic algorithms. Master thesis, TU, Wien
Pech S, Kandler G, Lukacevic M, Füssl J (2018) Metamodel assisted optimization of glued laminated
timber systems by reordering wooden lamellas using metaheuristic algorithms. Comput Struct
(submitted)
Persson K (2000) Micromechanical modelling of wood and fibre properties, vol 1013. Division of
Structural Mechanics, Lund Institute of Technology. ISBN: 91-7874-094-0
Pirazzi C, Weinand Y (2006) Geodesic lines on free-form surfaces—optimized grids for timber rib
shells. In: World conference on timber engineering, vol 7
Pottmann H, Asperl A, Hofer M, Kilian A (2007) Architectural geometry. Bentley Institute Press
Pottmann H, Huang Q, Deng B, Schiftner A, Kilian M, Guibas L, Wallner J (2010) Geodesic
patterns. ACM Trans Graph 29(4):#43, 1–10
Pottmann H, Eigensatz M, Vaxman A, Wallner J (2015) Architectural geometry. Comput
Graph 47:145–164
Puck A, Schuermann H (1998) Failure analysis of frp laminates by means of physically based
phenomenological models. Compos Sci Technol 58(7):1045–1067
Schiftner A, Höbinger M, Wallner J, Pottmann H (2009) Packing circles and spheres on surfaces.
ACM Trans Graph 28(5):#139, 1–8
Schling E, Barthel R (2017) Experimental studies on the construction of doubly curved structures.
Detail Struct 01:52–56
Schling E, Kilian M, Wang H, Schikore D, Pottmann H (2018) Design and construction of curved
support structures with repetitive parameters. In: Advances in architectural geometry (submitted)
Schmidt J, Kaliske M (2007) Simulation of cracks in wood using a coupled material model for
interface elements. Holzforschung 61(4):382–389
Schmidt J, Kaliske M (2009) Models for numerical failure analysis of wooden structures. Eng Struct
31(2):571–579
Serrano E, Gustafsson J (2006) Fracture mechanics in timber engineering—strength analyses of
components and joints. Mater Struct 40:87–96
Sjödin J, Serrano E (2008) A numerical study of methods to predict the capacity of multiple steel-
timber dowel joints. Holz als Roh- und Werkstoff 66(6):447–454
Sjödin J, Serrano E, Enquist B (2008) An experimental and numerical study of the effect of friction
in single dowel joints. Holz als Roh- und Werkstoff 66(5):363–372
Song JH, Areias PMA, Belytschko T (2006) A method for dynamic crack and shear band propagation
with phantom nodes. Int J Numer Meth Eng 67(6):868–893
Steffen A, Johansson CJ, Wormuth EW (1997) Study of the relationship between flatwise and edge-
wise modull of elasticity of sawn timber as a means to improve mechanical strength grading
technology. Holz als Roh- und Werkstoff 55(2–4):245–253
Stitic A, Weinand Y (2015) Timber folded plate structures—topological and structural considera-
tions. Int J Space Struct 30:169–176
Suquet P (1987) Elements of homogenization for inelastic solid mechanics. In: Sanchez-Palencia
E, Zaoui A (eds) Homogenization techniques for composite media. Lecture notes in physics 272.
Springer, Wien, New York, pp 193–278
Suquet P (1997) Continuum micromechanics. Springer, Wien, New York
Tsai SW, Wu EM (1971) A general theory of strength for anisotropic materials. J Compos Mater
5(1):5880
Zaoui A (2002) Continuum micromechanics: survey. ASCE J Eng Mech 128(8):808–816
Zimmermann T, Sell J, Eckstein D (1994) Rasterelektronenmikroskopische Untersuchungen an
Zugbruchflächen von Fichtenholz. Holz als Roh- und Werkstoff 52(4):223–229
Hygroscapes: Innovative Shape Shifting
Façades
Sherif Abdelmohsen, Sigrid Adriaenssens, Stefano Gabriele, Luigi Olivieri

and Rana El-Dabaa
Abstract This chapter focuses on the testing and design of shape shifting façade
prototypes that are programmed to passively sense stimuli and respond in a con-
trolled setting based on the hygroscopic properties of wood. Wood is introduced in
this context as a low-tech smart material with a naturally soft responsive mecha-
nism that offers a substitute for mechanical actuators. First, a set of physical experi-
ments were conducted to deduce the design parameters that affect wood morphology,
behavior and response time upon changes in humidity levels and moisture content,
including dimensional ratio, grain orientation, material thickness, type of wood, and
lamination. We then report on the process and outcome of a workshop held at the
American University in Cairo, with the main challenge of regulating the morphology
and hygroscopic behavior of wood to work as an actuator with specifically desired
motion for adaptive building façade prototypes. Based on the observations and analy-
sis of concepts and mechanisms, we discuss shape shifting grammars as a framework
for devising adaptive façade prototypes from a generative design perspective, where
specific combinations of motion parameters are used to induce semantic rules and
customized commands for the overall behavior of shape shifting mechanisms.
S. Abdelmohsen (B)
American University in Cairo, New Cairo, Egypt
e-mail: sherifmorad@aucegypt.edu
S. Abdelmohsen
Ain Shams University, Cairo, Egypt
S. Adriaenssens
Princeton University, Princeton, NJ, USA
e-mail: sadriaen@princeton.edu
S. Gabriele · L. Olivieri
Roma Tre University, Rome, Italy
e-mail: stefano.gabriele@uniroma3.it
L. Olivieri
e-mail: lolivieri89@gmail.com
R. El-Dabaa
Arab Academy for Science, Technology and Maritime Transport, Cairo, Egypt
e-mail: rana.bahaa@aast.edu
676 S. Abdelmohsen et al.
Keywords Flexible wood · Shape shifting facades · Hygroscopy

Generative design · Programmable matter
1 Introduction
Designing for adaptive and responsive facades requires a controlled loop of sen-
sors, processors and actuators. This technology relies on active complex systems
with a multitude of rigid devices and hard mechanistic operations. Advances in
computation, material science, and fabrication technologies have revolutionized the
development of soft adaptive building skins and brought about the application of
programmable materials in shape shifting facades. The ability to program a single
sheet of material to seamlessly sense and respond has highly impacted the design
of actuation and motion in adaptive systems. In natural materials such as wood,
this mechanism is more complex yet promising. Hygroscopic properties of wood
add to this complexity, where the specific behavior and direction of motion can be
controlled. This is due to the orthotropic properties of wood, where its mechanical
properties vary when measured longitudinally, radially or along the grain. This mech-
anism introduces a novel way to enhance responsiveness and interactivity, allowing
for a passive alternative to mechanical actuators in responding to specific stimuli
including climatic conditions.
In this chapter, the hygroscopic behavior of wood is utilized as a low-tech smart
material with a naturally responsive mechanism that exhibits different shrinkage and
expansion values when exposed to different humidity levels. Wood is known for its
low environmental impact, low embodied energy, carbon impact and its ability to
change volume passively under certain stimuli. This chapter focuses on the design
and application of a material-based system that is programmed to passively sense
stimuli and respond in a controlled setting. A set of physical experiments were con-
ducted to deduce the design parameters that affect wood behavior. These parameters
included type of wood, moisture content, thickness, grain orientation, sample dimen-
sional ratio, and lamination. These parameters were seen to collectively control the
morphing behavior and response time of wood to changes in humidity. Material fab-
rication and design parameters were studied to encode the direction and deflection
value of the programmable wood system.
This chapter illustrates the effect of each design parameter on the responsive
behavior of wood, with primary focus on dissecting each parameter independently
to deduce a traceable benchmark for the impact of each on wood response. The
experiments included testing the response behavior of different types of wood such
as softwood and hardwood. The response speed of wood to differences in moisture
content was also presented using a moisture content meter, digital calipers, and
humidity and temperature sensors. Response speed was calculated using outputs
such as sample deflection value, difference in length, and moisture content over
time.
Hygroscapes: Innovative Shape Shifting Façades 677
Material thickness was another parameter that was seen to affect the response
speed and behavior of wood to differences in humidity levels. Several thicknesses
were tested to evaluate wood behavior in terms of speed response and deflection,
ranging from veneer (0.5 mm) to thicknesses of 1.8 mm. Another significant param-
eter was grain orientation, which differs according to the direction and location from
which the sample is extracted; namely tangential, longitudinal and radial orienta-
tions. Each of these directions differs in terms of their structure and direction of
pores, thus affecting the ability of the sample to hold different amount of water,
and consequently the shrinkage and expansion values. Three different dimensional
ratios were tested to identify the response of wood. Dimensional ratio was shown to
directly affect sample deflection value, which varied from bending to a full rolling
up of the sample.
All these parameters were shown to affect the morphing behavior of wood. For a
more controlled and accurate responsive behavior, a combination of these parameters
is necessary to achieve the desired response. This was further demonstrated in the
lamination process, where unlimited variations were conducted to test all parameters
in conjunction, relying on the concept of using passive and active layers. Lamination
typically uses the difference in shrinkage value for each layer to account for difference
in wood response to humidity, where the active layer is responsible for motion, and
the passive layer accounts for resistive behavior.
Experiments were conducted on Fir (softwood) and Beech (hardwood). The struc-
ture of hardwood and softwood differs in the ability of handling water. Hardwood has
pores while softwood does not have pores, thus affecting their mechanical properties
in terms of change in volume. The combination of hardwood and softwood is usually
used in the lamination process, where hardwood acts as the active layer and softwood
as the passive layer. The implementation of this concept of passive and active layers
to wood allows for a passive ‘programmable’ system that potentially replaces highly
mechanistic digitally controlled systems.
Lamination does not rely solely on the previous parameters for each layer, but
also on the state in which lamination took place. The programming process relies
significantly on the moisture content of wood during the lamination process. The
response behavior has a reversible property, where it can be programmed to begin
or end with the desired motion or shape; either curved or flat. If the sample was
laminated in a wet state, its initial state is curved, and when exposed to an increase in
humidity it flattens. If the lamination however was initially applied in a dry state, the
initial state is flat and when exposed to an increase in humidity it curls. This leads
to unlimited variations in terms of achieving a specific desired morphology while
designing an adaptive façade. Exhausting the endless possibilities resulting from
these variations was necessary to inform the experiments and workshops outlined in
this chapter.
After our preliminary experiments, we report on the process and outcome of a
workshop conducted in July 2017 in the Department of Architecture at the American
University in Cairo, and held as a collaboration between Princeton University, the
American University in Cairo and Rome Tre University. The challenge in this work-
shop was controlling and regulating the behavior and response of wood to work as
an actuator with specifically desired motion for an adaptive building façade. Partici-
pants in the workshop were presented with state-of-the-art technology of hygroscopic
responsiveness of wood. The hygroscopic design parameters of wood were utilized in
the workshops to design and fabricate responsive facades that respond to differences
in humidity levels.
Thirty undergraduate students of architecture, in semesters 7 and 8 of their study
plan, participated in the workshop. They were divided into five groups and were asked
to utilize the hygroscopic behavior of wood to design adaptive kinetic façade pro-
totypes. Participants were encouraged to set a series of experiments and personally
investigate the design parameters of wood to develop an actuation concept and mech-
anism. Physical and digital methods were used by the participants to test and explore
their programmable material approach. Participants were able to produce a variety
of responsive façades using proposals of multi-layer wood consisting of softwood
and hardwood, veneer, and different thicknesses with different grain orientations.
Finally, this chapter explores the challenges and solutions for using hygroscopic
behavior of wood as programmable material that allows for a passive responsive
material with low energy consumption and carbon foot print. We conclude with a
description of the observed shape shifting mechanisms and identify a framework
for utilizing shape shifting grammars in our further research to demonstrate a con-
trolled and regulated kinetic motion of adaptive façades from a generative design
perspective.
2 Relevant Work
The state of the art in hygroscopic design generally comprises three domains of
research: the physical and numerical testing of the mechanical properties of wood
and its deformation, controlling and programming the behavior of wood, and the use
of hygroscopic properties in architectural applications. This chapter focuses on the
architectural applications of wood and the ability to control and program its response.
Recent work has included testing the ability to program materials through their
mechanical properties in relation to their deformation and shrinking values, and
developing the concept of programmable matter, such as the work of the Institute
for Computational Design (ICD) at the University of Stuttgart (Dierichs and Menges
2016), and the Self-Assembly Lab (SAL) at MIT (Correa et al. 2015). Different fab-
rication methods and activation techniques have been developed in that area includ-
ing 3D printing of materials or using a combination of laminated layers such as
printed nylon with wood. This was seen to increase the ability to program the exact
properties of wood through printing them in a way similar to the original fiber ori-
entation (Correa et al. 2015). Other methods used lamination of specific segments
of the wood samples that act as hygroscopic programmed actuators (Rüggeberg and
Burgert 2015). This involves the combination of different hygroscopic design param-
eters such as grain orientation, dimensional ratio, and wood type, in addition to the
type and nature of the lamination process.
Utilizing the hygroscopic properties of wood in architectural applications was

applied by means of shifting to large scale fabrication and self-assembly techniques.
For example, the ability to fabricate joints in 3D and program their initial state condi-
tion in relation to the overall structure allowed for self-actuated wooden membranes
in relation to the difference in humidity levels (Wood et al. 2016). The ‘Hygroscopic
Aggregate Structure’ workshop introduced by the ICD lab (Dierichs and Menges
2016) demonstrated techniques by which participants could achieve robotic con-
struction with granular materials. This was achieved by programming the material
and substituting mechanical techniques. The granular materials, acting as embedded
wooden actuators, were stimulated by changes in humidity levels, and expanded to
reach approximately 500% increase in volume, with a reversible behavior capability.
The Hygroskin project (Krieg et al. 2014) used the embedded properties of wood
and multi-layer lamination to passively respond to changes in humidity levels, with-
out the need for any complex mechanical actuators. The responsive pavilion’s façade
relied on the hygroscopic properties of wood as a response system; the façade opens
in dry conditions and closes as humidity increases. The work by Holstov et al. (2015b)
used hygroscopic properties to program operable facades and passively regulate the
amount of daylight, and reduce both heat gain and glare in space. The responsive
façade prototype was preprogramed and controlled to promote shading by closing
openings in dry conditions with clear sky condition, while opening under humid and
overcast sky conditions.
3 Hygroscopic Properties of Wood: Preliminary

Experiments
Controlling the behavior of wood can be done through different design parameters
such as moisture content, material thickness, grain orientation, type of wood, and
laminating multiple layers together. Each parameter affects the response behavior
of wood, represented in the deflection of the material sample, shrinkage value and
response time as humidity levels change. The response of wood to humidity is a
reversible action, where wood returns to its initial state upon drying. The initial state
of wood can be controlled and programmed in either a flat or curved state through
controlling the level of moisture content in the lamination process. We present below
a set of preliminary experiments conducted on different samples of wood, discussing
three of the main parameters affecting the response of wood to changes in humidity
levels: dimensional ratio, grain orientation, and lamination.
3.1 Dimensional Ratio
Depending upon the specific dimensional ratio, wood samples will behave more like
a beam, plate or shell, upon an increase in moisture content. Beamlike elements have
a width and a thickness that is small in comparison to their length. As a result, the
deformation in the width direction is negligible and the dimensionality of the digital
model can be reduced to one dimension. A wood sample that has a small thickness
but a larger width and length, will behave as a plate when it is programmed flat in
its initial state. In contrast, a wood sample with similar dimensions that is initially
programmed as curved in its initial state, will behave like a curved thin shell system.
An experiment was conducted to test the effect of dimensional ratio on the
response of wood. Three tangential beech veneer samples were used with differ-
ent dimensional ratios: 1:1, 1:2 and 1:3. The curvature of samples varied according
to their length. It was shown that maximum deflection was achieved in the sample
with dimensional ratio 1:3, while the least curvature was in the ratio 1:1, as shown
in Fig. 1. These results were significant in terms of selecting lengths and propor-
tions of wood samples in an adaptive façade system, where the expected deflection
and response speed vary for samples with different dimensional ratios, allowing for
a clear prediction of the expected type of motion, as well as allowing for a wider
toolset of elements and patterns with varying configurations and motion responses
within the designed façade.
3.2 Grain Orientation
The response of wood to humidity varies according to its grain orientation, whether
in the longitudinal, radial or tangential direction. Figure 2 shows preliminary exper-
iments conducted on three samples of a single layer of wood veneer with three
different grain orientations. The orientation of the sample at the bottom is at 90°
to that of the sample in the middle, while the orientation of the sample at the top
is at 45° to the middle sample. While applying the same amount of moisture to all
three samples, the resulting motion occurs in a direction perpendicular to the grain
orientation, causing a variation in the morphology of the final state of all three sam-
ples. These results were significant in terms of expecting the type and direction of
motion when designing an adaptive façade system, especially in the design of façade
apertures and louvers.
Another experiment examined the response of five beech veneer samples with
different dimensional ratios (width to length 1:3, 1:2 and 1:1) and different grain
orientations (tangential and longitudinal), as shown in Fig. 3. To validate the results,
the experiment was conducted on three similar samples for each case, and the
average reading was taken. Samples were sprayed by water from one side, and the
Fig. 1 Maximum deflection value in three tangential beech veneer samples with different dimen-
sional ratios
deflection value and difference in length (sample shrinkage) were measured based
on specific time intervals at room temperature 21 °C and relative humidity ranging
from 47 to 50%.
It was shown that cases 1 and 4 (tangential grain orientation) were the fastest
in response, as they both completely curled around themselves in approximately
one minute. The uncurling process varied based on dimensional ratios, where case
1 (1:3) took approximately 50 min and case 4 (1:2) took approximately 20 min.
The difference in their dimensional ratios was found to affect the response time
and deflection value, while the grain orientation was found to affect the movement
behavior.
The response in cases 2 and 3 (longitudinal grain orientation) was slower than
the tangential orientation cases. Case 2 (1:3) took approximately 3 min to reach its
highest deflection value (28.08 mm) and 14 min to return to its initial flat state. Case
4 (1:2) took approximately 2 min to reach its highest maximum deflection value
(25.75 mm) and 13 min to return to its initial flat state. It was observed that the
Fig. 2 Behavior of wood with different grain orientations
Fig. 3 The response of five beech veneer samples with different dimensional ratios and grain
orientations
dimensional ratios did not have a significant effect in longitudinal grain orientation
cases, as the movement was taking place on the fixed length side. Case 5 (with a
square proportion of 1:1) reached a complete curl in 2 min and returned to its initial
flat state after 11 min, with an average highest deflection value of 25.65 mm.
3.3 Lamination
In order to explore the reaction of multi-material properties to moisture, a set of exper-

iments was conducted by implementing a lamination process with polyurethane glue.
The behavior of wood bilayer generally relies on the difference in expansion coef-
ficients between active and passive layers. Two types of veneer (beech and spruce)
were laminated with different grain directions, to perform a natural curvature or
twisting by adding and removing moisture. The initial experiment used one layer
of beech (1 mm—90° grain direction) and one layer of spruce (0.8 mm—0° grain
direction). The samples were first completely dried in an oven and then laminated
with polyurethane glue (type D4). The samples were then moved into a humidified
chamber to control the quantity of moisture correlated with the curvature deflection.
The result shown in Fig. 4 produces a visible curvature of the sample after only
40 min at 85% relative humidity and temperature of 18 °C.
The response of wood bilayer samples to humidity basically relies on the differ-
ence in the microstructure cells of each of the two layers. Each layer has its own
swelling and shrinkage value when exposed to moisture. The active layer has larger
swelling and shrinkage values than the passive layer (Holstov et al. 2015a). This
behavior produces a working mechanism by creating internal forces that bend the
element without the use of external forces.
A reversible fabrication process was developed during the preliminary experi-
ments in order to achieve the same result with two different conditions, using both
single-layered and laminated veneer samples. The samples could bend or twist when
the moisture was added or removed, due to the 45° grain orientation of the sample in
both cases. As shown in Fig. 5, the result can change according to the thickness and
the layers of used materials. In the first case on the left, the sample used a single layer
Fig. 4 Laminated wood samples (Beech and Spruce)

Fig. 5 Programming wood veneer samples through lamination to induce a twisting motion resulting
in a helical morphology
of paperback veneer, and the result after 45 min was the illustrated helical shape.
In the second case on the right, the veneer was laminated, and resulted in a similar
twisting into a helical shape. This was caused by the two different coefficients of
expansion and contraction, where the material could be programmed to achieve a
specific desired type of motion and rotation. Manipulating other parameters such as
grain orientation, type of wood, material thickness, or dimensional ratio could yield
different results, thus regulating the final morphology to the desired type of kinetic
response.
4 Shape Shifting Prototypes
4.1 Method
To explore the different effects of dimensional ratio, grain orientation, wood type,
thickness, and lamination, a workshop was held at the Department of Architecture
in the American University in Cairo in July 2017, in collaboration with Princeton
University and Rome Tre University. Thirty undergraduate students of architecture
(at semester 7 and semester 8 of their study plan) participated in the workshop. They
were divided into five groups, and the aim was to utilize the hygroscopic behavior of
wood to design adaptive kinetic façade prototypes. The challenge in the workshop
lied in the ability to control wood behavior to simulate a specific motion to be applied
as an adaptive façade technique.
The workshop was divided into three integrated components: lectures and hands-
on sessions that illustrate the main concepts to students by tutors from the partic-
ipating institutions, experimentation by the students for preliminary samples and

fabrication of shape shifting prototypes, and student presentations to conclude pro-
cess and ideas and to exhibit their fabricated prototypes. The preliminary sessions
demonstrated the concept of programmable materials to students, their architectural
applications, the mechanical properties of materials and specifically wood, in addition
to an introduction to parametric and numerical simulation techniques for designing
and fabricating the shape shifting prototypes.
This was followed by generating scripts using Rhino and Grasshopper software
for the manifestation of the students’ ideas, then conducting experiments on different
types of veneer and laminated samples of wood to measure the deflection, shrinking
factor and radius of curvature for each sample upon applying moisture. Students
were provided with 1.8 mm beech (hardwood) veneer samples, and 0.8 mm fir (soft-
wood) veneer samples. The tools provided to conduct the required measurements
were digital calipers, weight gauges and scales. Polyurethane glue was used for the
lamination process as a water resistance-based glue.
The outcome of the workshop was five shape shifting façade prototypes (70 cm ×
70 cm × 20 cm) that exhibit variations but mainly involved three basic concepts. The
first concept was designing wooden vertical louvers that respond to the variation of
humidity levels through a twisting motion, thus allowing for openings in between
louvers while in motion. The second concept focused on programming the hinge
between different wood segments to act as a passive actuator that can substitute
mechanical actuators. The programmable hinge was able to lift a veneer louver. The
first and second concepts implemented their work using laminated sample of wood,
while the third concept relied on using a single layer of wood. Three student groups
worked on the third concept which featured the idea of modular panels with opening
and closing apertures. The three prototypes used either a double façade concept or
relied on the difference in response rate of the beech and fir veneer.
4.2 Prototype 1
The movement mechanism in this prototype is close to the motion of lamellas in

the façade of the One Ocean, Thematic Pavilion EXPO 2012 in Yeosu, South Korea
(Knippers et al. 2012). The façade of that precedent project was made of glass-
fiber-reinforced polymer (GFRP) with 9 mm thickness. The motion gave a gradient
motion of the lamellas’ percentages of opening by means of using the same radius
of lamellas with variation in length. The longer the length the larger the curvature
it allows for in the overall pattern. This resulted in a variation in the percentage
of openings. Mechanical actuators were used to give a sliding motion that pushes
the material causing the bending mechanism. To achieve a similar gradient effect
with louver motion, the motion of wood was programmed in relation to change in
humidity levels. The design of the shape shifting façade prototype relied on using
laminated layers with different types and grain orientations to achieve the twisting
motion mechanism (Fig. 6).
Fig. 6 Achieving the twisting mechanism of vertical louvers by controlling the laminated properties
of wood and using different grain orientations in the active and passive layers
Fig. 7 Testing the response time of single and bilayer beech and fir veneer samples with 45° grain
orientation
The challenge lied in designing vertical louvers that twisted in a gradient pattern.
This required a series of physical experiments to understand and test the response
behavior of single and laminated samples in relation to the response time of each
louver to achieve the necessary gradient design (Fig. 7).
The laminated samples consisted of beech as the active layer and fir as the passive
layer. All the designed louvers comprised a main active layer consisting of beech, at
a 45° grain orientation. The passive layer was then used to regulate the direction and
response speed of the twisting motion. In the middle of the prototype, the students
designed a single layer of beech with 45° grain orientation. The passive layers were
then gradually used to the outer sides, including different orientations to achieve
the required pattern variation, including tangential grain orientation (0° grain orien-
tation), 25° grain orientation, longitudinal (90° grain orientation), and −45° grain
orientation, as shown in Fig. 8. Louvers are fixed from the upper point and free from
the lower point. The form and proportions of the “I” shaped louvers are utilized to fit
Fig. 8 The different grain orientation and lamination patterns used for the twisting prototype
in a rectangular groove, thus enabling a smoothly twisting motion, as shown in Fig. 9.

The ability to control a gradient twisting motion is a resultant of combining different
grain orientations through the lamination process. Programming each vertical louver
twisting degree affects the view angle between the louvers. The wider viewing angle
is in the center of the façade, while the angle decreases towards the right and left
edges where the two layers of wood have opposite grain directions. The final shape
shifting façade prototype is shown in Fig. 10.
Fig. 9 Fixation connections of Prototype 1
Fig. 10 The final design of Prototype 1
4.3 Prototype 2
The mechanism of this prototype utilized the hygroscopic properties of wood in

developing an adaptive hinge that potentially could substitute mechanical actuators.
Students developed different iterations and experiments to understand and test the
behavior of wooden hinges thus reaching an optimum hinge prototype with regulated
motion. Figure 11 demonstrates the testing conducted for response time of single and
laminated hinge actuators.
Fig. 11 Testing the response time of single and bilayer sample hinges to act as passive actuators
Fig. 12 Using programmable wood to regulate the motion and rotation of hinge actuator
Different types of wood were studied. The development of the prototype relied
mainly on the idea of laminated preprogrammed wood as a hinge, with a curved initial
state. The hinge was fabricated using two wet laminated beech and fir veneer samples
with thickness 1.8 and 0.5 mm respectively with a tangential grain orientation. Due
to the use of the wet lamination method, the hinge was preprogrammed in a curved
state, as shown in Fig. 12. When the level of humidity increases, the hinge deflection
value decreases towards a slightly flat state.
In their final prototype (Fig. 13), the students fabricated two laminated layers of
tangential beech and fir in a wet lamination condition. The wet lamination process
allowed for programming the initial state of the veneer sample to be curved, then
to be flattened in the final state when as humidity increases. The vertical louvers
are fixed with the hinge using polyurethane glue. This prototype utilizes the folding
mechanism of vertical louvers to act as horizontal shading devices. The range of
motion is related to the response of the hinge to the level of humidity. The folding
angle ranges from 0° to 90°.
4.4 Prototype 3
The third prototype relied on achieving different percentages of kinetic façade open-
ings. This tends to rely typically on mechanical actuators that use different types of
motion such as sliding, rotation and folding. The main idea relied on programming
the response of wood to changes in humidity according to type, response time and
thickness. Students studied previous examples of kinetic facades achieving different
aperture openings and dynamic opening patterns by means of mechanical actuators
(Abdelkader and Park 2017). The prototype was designed as a double façade, with
each skin layer consisting of a single layer of wood. The front side consisted of beech
with thickness 1.8 mm, while the back side consisted of fir with thickness 0.5 mm,
both with 45° grain orientation. Different physical experiments were conducted on
the two types of wood to understand and record the different response time and wood
behavior upon increase in humidity levels (Fig. 14).
The architectural design of this façade prototype divides the pattern in two lay-
ers, resulting in dynamic shading throughout the day and different viewing angles
(Fig. 15). It was designed such that the skins vary in terms of their response time,
where the beech layer consumes more time to respond due to its large thickness and
slow response time.
Fig. 14 Experiments to understand the effect of wood type, thickness and grain orientation on
response time and behavior to humidity
Fig. 15 Double layer skin that allows for dynamic shading and different viewing angles
The fixation of the two facade layers is different, as the orientation of louvers is
either parallel or perpendicular to the façade skeleton. Louvers with perpendicular
orientation are fixed using tongue and groove connections with the façade skeleton,
while the parallel louvers are fixed by gluing them to the façade structure using
polyurethane glue, as shown in Fig. 16.
Fig. 17 Dynamic effect and variations of aperture percentages in the double skin façade prototype
This dynamic effect and the discrepancies in aperture percentages were augmented
by means of the tilting of the second layer skin and the difference in wood response
between the first and second layers (Fig. 17), leading to varying patterns within
each modular panel and across the overall prototype. The final prototype is shown
in Fig. 18.
4.5 Prototype 4
The fourth group designed a double skin façade prototype consisting of fir veneer
on the front side and beech veneer on the back side, all implemented using 45° grain
orientation. The layers were aligned to produce different solid and void ratios and
consequently different aperture sizes and opening and closing mechanisms (Fig. 19).
Two different structures are fabricated for each façade layer. Both types of louvers
are fixed to the façade structure using polyurethane glue. The beech façade is fixed
from the center of the louvers, while the fir façade is fixed from the louver edges, as
shown in Fig. 20.
The dynamic effect and discrepancies in aperture percentages were augmented
by means of the differences in response of the overlapping patterns for the two skins
(Fig. 21). The final prototype is shown in Fig. 22.
4.6 Prototype 5
This prototype was composed of nine modular panels with triangulated patterns,
each differently responding to changing levels of humidity. Single layers of fir and
beech were used with different combinations of grain orientations. Each modular
panel is composed of two overlapping triangles. The most important factor while
designing this prototype was the continuous evaluation of the response speed and
time of each triangular panel to avoid collision between the different modular panels.
Fig. 19 Variations in aperture sizes and mechanisms in the double skin façade prototype
Several trials and experiments were conducted on beech and fir triangular samples
with different grain orientations and thicknesses to evaluate their response behavior
and time (Fig. 23). These movable panels are glued to a frame with nine square panels
behind them using polyurethane glue.
It was observed that beech with a tangential cut exhibited the fastest response.
This impacted the design of the modular panels, where the beech layer was used as
the outermost layer in six of the modular panels. The second fastest was the fir veneer
with tangential grain orientation which was used as the outermost layer in three of the
modular panels. The beech and fir veneer triangular modules with longitudinal grain
orientations exhibited the slowest response and were therefore used in the innermost
layer. The final prototype is shown in Fig. 24. A wooden frame with nine square
panels is fabricated and the louvers are glued to this structure by polyurethane glue.
Fig. 20 From left to right: fixation of the beech layer from the center of the prototype, and fixation
of the fir layer from the louver edges and the two-layer structures with attached louvers
Fig. 21 Testing the response time of single veneer samples
5 Discussion
5.1 Shape Shifting Mechanisms
The results of the five prototypes demonstrated a number of architectural features

pertaining to shape shifting façades, as shown in Table 1. These features include
the specific challenge of the façade skin (e.g. direction and range of motion, speed
response control, aperture percentage, etc.), the type of motion (e.g. rotation, sliding,
twisting, etc.), and the lamination parameters, including the number of layers (single,
double or multi-layered), lamination condition (wet or dry), and the nature of the
active and passive layers (including parameters of grain orientation and type of
wood). For each prototype, a different arrangement and combination of parameters
was used to induce a specific type of shape shifting mechanism, including fully
responsive envelope, vertical louvers, responsive hinges, etc.
In Prototype 1, the main shape shifting application was twisting vertical louvers,
similar in principle to the One Ocean, Thematic Pavilion EXPO 2012 façade in
Yeosu, South Korea. This was done using a double layer of beech and fir using dry
Fig. 23 Testing the response time and behavior of single and bilayer samples with different grain
orientations
lamination. The challenge in this prototype was threefold: (1) controlling the direction
of motion for each louver (with often varying directions and degrees of twisting, to
induce a dynamic and non-repetitive façade prototype pattern), (2) controlling speed
response per louver, and (3) controlling the percentage of the aperture, which is
indirectly impacted by the dynamic twisting. The strategy therefore was to use beech
(hardwood) as the active layer at 45° to induce the initial twisting motion, and fir
(softwood) as the passive layer with different grain orientations per louver to regulate
the response speed of each louver and induce a dynamic and free effect for each of
the elements of the shape shifting prototype.
In Prototype 2, the main shape shifting application was folding louvers, which
relied on the principle of responsive hinges as an actuator to induce the holistic louver
motion based on a combination of rotation and sliding. This was done using a double
layer of beech and fir using wet lamination. The challenge in this prototype was
twofold: (1) controlling the range of motion for each louver hinge, with varying angles
to induce a dynamic and non-repetitive façade prototype pattern, and (2) controlling
speed response per louver. The strategy therefore was to use beech (hardwood) as the
active layer at a tangential grain orientation, and fir (softwood) as the passive layer
at a tangential grain orientation to regulate the motion of the beech active layer in
the louver hinge and consequently the angle of rotation of the louver.
In Prototype 3, the main shape shifting application was a responsive double skin
façade prototype consisting of a set of rotating modular panels with closing and
opening apertures, similar in principle to the rotating iris of the Arab World Institute
in Paris and the Al Bahar Towers responsive façade. This was done using a single
layer of both beech and fir per skin with perpendicular grain orientation on the
center of each of the modular polygons. No lamination was applied. The challenge
in this prototype was threefold: (1) controlling the direction of motion for each of
the modular panels, (2) controlling speed response per panel, and (3) controlling
the percentage of aperture closing and opening. The strategy therefore was to use
both beech (hardwood) and fir (softwood) to induce varying percentages of apertures
rotating at different speeds based on type of wood and specific grain orientations.
Table 1 Description of the challenges, types of motion, lamination conditions, and architectural
applications for the shape shifting prototypes
Prototype 1 Prototype 2 Prototype 3 Prototype 4 Prototype 5
Number of Double layer Double layer Single layer Single layer Single layer
layers
Lamination Dry Wet – – –
condition lamination lamination
Active layer Beech at 45° Tangential Beech/fir with Beech/fir with Beech/fir with
grain Beech perpendicular perpendicular different grain
orientation grain grain orientation to
orientation on orientation on control speed
panel center panel center of response
Passive layer Fir with Tangential Fir – – –
different grain to regulate the
orientations to motion of
regulate beech wood
twisting
motion
Challenge Control speed Control speed Control speed Control speed Control speed
of response + of response + of response + of response + of response +
direction of control range direction of direction of direction of
motion/louver + of hinge motion + motion motion
percentage of motion percentage of
aperture aperture
Type of Twisting Folding Rotation Rotation Rotation
motion (rotation +
sliding)
Architectural Vertical Responsive Responsive Responsive Responsive
applications louvers hinges double façade double façade façade
In Prototype 4, the main shape shifting application was a responsive double skin
panel with varying aperture openings. This was done using a single layer of both
beech and fir per skin with perpendicular grain orientation on the center of each of
the sub-components of the panel. No lamination was applied. The challenge in this
prototype was twofold: (1) controlling the direction of motion for each of the panel
sub-components, and (2) controlling speed response per sub-component. The strategy
therefore was to use both beech (hardwood) and fir (softwood) to induce varying
rotational motions for each of the sub-components moving at different speeds based
on type of wood and specific grain orientations, and consequently a dynamically
changing double skin aperture.
In Prototype 5, the main shape shifting application was a responsive façade proto-
type consisting of a set of rotating modular panels with closing and opening apertures.
This was done using multiple overlapping layers of both beech and fir per modular
panel with different grain orientations on the center of each of the modular polygons.
No lamination was applied. The challenge in this prototype was twofold: (1) control-
ling the direction of motion for each of the panel layers, and (2) controlling speed
response per panel layer. The strategy therefore was to use both beech (hardwood)
and fir (softwood) to induce varying rotational motions for each of the panel layers
moving at different speeds based on type of wood and specific grain orientations, and
consequently a responsive façade prototype consisting of a set of panels that open
and close with varying percentages.
5.2 Shape Shifting Grammars
By observing the material behavior and mechanism in all five prototypes, it is con-
cluded that each prototype exhibits a specific type of motion (e.g. bending, folding,
twisting, rotation, etc.) based on the particular combination of parameters, including
lamination, type of wood, material thickness, dimensional ratio, grain orientation,
etc. In order to arrive at those mechanisms and types of motion from a generative
design perspective, it is necessary to identify a general framework for shape shift-
ing. We base this framework on the concept of shape grammars, which are a set of
transformation rules applied on an initial shape to generate a new shape, whereby
a computational method identifies a rule that describes two parameters: the shape
that is replaced and the method it is replaced with (Knight 2000). More complex and
dynamically transforming 3D shapes require other types of grammars such as motion
grammars (Bidgoli and Cardoso-Llach 2015), leading to methods of investigating
generative designs of shapes in continuous motion, as the combination of different
motions devise the rules that produce composite motions.
As demonstrated by Moloney (2011), kinetic motion is typically defined by means
of three spatial transformations: translation, rotation, and scaling. Based on motion
grammar logic, any motion is defined through context-free language by means of
a set of elements, including sensor readings, output tokens, tokenizing function,
non-terminals, semantic rules, productions, and commands (Dantam and Stilman
2012). Based on the findings of this study, we devise a shape shifting grammar
based on this logic. According to this framework, the motion grammar elements are
understood as follows: Sensor readings refer to readings of wood moisture content,
temperature or humidity, in addition to basic data of the wood samples including
length and deflection upon response to humidity. Output tokens refer to the changes
in the property of wood in terms of sample length, deflection, radius of curvature,
and duration of response time. Tokenizing functions refer to the primary driver for
motion in wood, which is moisture content. Non-terminals refer to the parameters
that affect the type and outcome of motion, including type of wood, grain orientation,
material thickness, and lamination layers. Semantic rules refer to the rules directly
linked to the expansion and shrinking mechanical properties of wood and its ability
to absorb and handle moisture. Every group of semantic rules and events constitutes
a series of productions. Commands refer to the overall expected action of wood, as
a higher-level action, e.g. bending, flattening, twisting, rotation, etc.
Based on this shape shifting grammar, in a generative design context, the command
for bending a wood veneer sample involves the following semantic rules: an increase
in moisture content, leading to an increase in deflection and radius of curvature, and
a decrease in sample length due to the shrinking property of the material. These
are typically affected by non-terminals including type of wood, lamination, grain
orientation, and material thickness. A single active layer of beech veneer for example
would exhibit larger bending, or a higher increase in deflection, radius of curvature
and shrinkage of length than a laminated sample with active beech and passive
fir layers. Also, twisting the same sample would involve the same semantic rules
but with variation in grain orientation in the active layer for example to induce a
rotational output. Similar to prototype 1, the larger the difference in grain orientation
between the active and passive layers, the larger the rotational movement of the
vertical louvers. In addition, the larger the material thickness of either the active
or passive layers, the more the impact on speed and response time. The command
for flattening that same sample involves the following semantic rules: a decrease in
moisture content, leading to a decrease in deflection and radius of curvature, and an
increase in sample length due to the expansion property of the material.
Our future work will build on this shape shifting grammar logic to devise specific
and well-defined rules for generative design, with the ultimate objective of program-
ming the shape shifting mechanisms of adaptive façades based on the hygroscopic
properties of wood, whereby a general framework is developed to acquire sensor read-
ings of samples to a high degree of precision, and regulating specific combinations
of non-terminals to induce controlled semantic rules and consequently customized
commands for the overall behavior and motion of the shape shifting mechanism.
Using this shape shifting grammar logic can be used in studying the effect of these
responsive facades on energy performance in buildings in different climatic zones.
6 Conclusion
This chapter explored the challenges and solutions for using hygroscopic behavior
of wood as programmable material that allows for a passive responsive material with
low energy consumption and carbon footprint. Based on the conducted experiments
to observe and understand the hygroscopic properties of wood and its morphing
behavior in relation to the main affecting parameters, we discuss the main findings
of a workshop held to design prototypes of shape shifting façades. Thirty under-
graduate students of architecture at the American University in Cairo investigated
the design parameters of wood to develop actuation concepts and mechanisms. They
produced a variety of responsive façades using proposals that incorporated single
and laminated softwood and hardwood veneer, with different thicknesses, dimen-
sional ratios and grain orientations. Based on this programmable material approach,
we concluded with a description of the observed shape shifting mechanisms and
identify a framework for utilizing shape shifting grammars in our further research
to demonstrate a controlled and regulated kinetic motion of adaptive façades from a
generative design perspective.
Acknowledgements The authors are grateful for the support of the Bartlett’s Fund for Science
and Engineering Research Collaboration provided to the American University in Cairo, Princeton
University, and University Roma Tre. They would also like to thank Veronica Boyce, Aly Ibrahim,
Passaint Massoud and Tasbeh Mokbel for their support in the workshop setup, and students of the
Department of Architecture at the American University in Cairo: Maia Abdelalim, Salma Abdelfat-
tah, Maram Abou El Enein, Yara Adly, Dina Afifi, Rehab Ahmed, Omar Assem, Moshira Ayad, Laila
Azmi, Nouran Badawy, Mariam Barsoum, Rami El Hosseiny, Aly El Sinbawi, Omar Elbayoumi,
Reem Gamil, Maryam Gazzaz, Mennatallah Helal, Nour Hussein, Raghda Ismail, Zeina Lasheen,
Lina Mahgoub, Noura Mohamed, Kareem Rabie, Sandra Saad, Atheer Salama, Yara Saleh, Lama
Saleh, Reem Sayed and Malak Seoudi for their active participation during the workshops.
References
Abdelkader R, Park J-H (2017) The evolving transformation of Mashrabiya as a traditional Middle
Eastern architecture element. Int J Civ Environ Eng 17:15–20
Bidgoli A, Cardoso-Llach D (2015) Towards a motion grammar for robotic stereotomy. In: Pro-
ceedings of the 20th international conference of the association for computer aided architectural
design research in Asia (CAADRIA), Hong Kong, 2015
Correa D, Papadopoulou A, Guberan C, Jhaveri N, Reichert S, Menges A, Skylar T (2015)
3D-printed wood: programming hygroscopic material transformations. 3D Print Addit Manuf
2(3):106–116. https://doi.org/10.1089/3dp.2015.0022
Dantam N, Stilman M (2012) The motion grammar: linguistic perception, planning, and control.
In: Robotics: science and systems VII. Cambridge, Mass., MIT Press
Dierichs K, Menges A (2016) Towards an aggregate architecture: designed granular systems as
programmable matter in architecture. Granular Matter 18:25. https://doi.org/10.1007/s10035-
016-0631-3
Holstov A, Bridgens B, Farmer G (2015a) Hygromorphic materials for sustainable responsive
architecture. Constr Build Mater 98:570–582. https://doi.org/10.1016/j.conbuildmat.2015.08.136
Holstov A, Morris P, Farmer G, Bridgens B (2015b) Towards sustainable adaptive building skins with
embedded hygromorphic responsiveness. In: Oliver E (ed) Advanced building skins: proceedings
of the international conference on building envelope design and technology (ICBEDT), Graz,
Austria, 2015
Knight T (2000) Introduction to shape grammars. MIT Press
Knippers J, Scheible F, Oppe M, Jungjohann H (2012) Bio-inspired kinetic GFRP-façade for the
Thematic Pavilion of the EXPO 2012 in Yeosu. In: IASS—APCS symposium 2012: from spatial
structures to space structures, Seoul, Korea, 2015
Krieg O, Christian Z, Correa D, Menges A, Reichert S, Rinderspacher K, Schwinn T (2014)
HygroSkin: meteorosensitive Pavilion. In: Gramazio F, Kohler M, Langenberg S (eds) Fabri-
cate: negotiating design and making. Zürich, Austria, pp 272–279
Moloney J (2011) Designing kinetics for architectural facades: state change. Routledge, Abingdon,
Oxon; New York
Rüggeberg M, Burgert I (2015) Bio-inspired wooden actuators for large scale applications. PLoS
ONE 10(4):e0120718. https://doi.org/10.1371/journal.pone.0120718
Wood DM, Correa D, Krieg OD, Menges A (2016) Material computation—4D timber construc-
tion: towards building-scale hygroscopic actuated, self-constructing timber surfaces. Int J Archit
Comput 14:49–62. https://doi.org/10.1117/1478077115625522
Adaptive Wooden Architecture.
Designing a Wood Composite
with Shape-Memory Behavior
Maryam Mansoori, Negar Kalantar, Terry Creasy and Zofia Rybkowski
Abstract Wood is a sustainable and attractive material with a venerable history of

use in architectural construction and carpentry. It also has a promising and inno-
vative future in architecture and design. The aim of our research is to reintroduce
wood as a responsive and transformable material for use in novel adaptive archi-
tectural design. By combining a refurbished technique of wood-cutting known as
kerfing with the use of a shape-memory polymer resin, we have created wood-based
surfaces that can turn into precise curvilinear forms without incurring damage, and
then self-transform to their original shape in response to environmental stimuli. We
developed a temperature-based responsive polymer and a flexible, diamond-shaped
kerfing pattern in our prototype testing and were able to achieve the desired results.
This method enabled us to design and control the material and its behavior by taking
advantage of the micro-scale resin polymer’s effects, combined with wood’s specif-
ically cut geometry. In addition to demonstrating the possibilities of shape memory
behavior for wood-based architecture, this prototype offers a practical technique that
can be used by designers to create flexible and inexpensive wood-based fabrications
on the required scale with compact storage and transportation alignments.
Keywords Adaptive architecture · Wood design · Shape memory behavior

Curvilinear surfaces · Responsive wooden surfaces
Smart Materials in Architecture
M. Mansoori (B) · N. Kalantar · T. Creasy · Z. Rybkowski

Architecture Division, California College of the Arts, San Francisco, CA, USA
e-mail: maryammansoori@tamu.edu
N. Kalantar
e-mail: kalantar@cca.edu
T. Creasy
e-mail: tcreasy@tamu.edu
Z. Rybkowski
e-mail: zrybkowski@tamu.edu

704 M. Mansoori et al.
1 Introduction: A New Understanding of Wood

as an Architectural Material
Environmentally sustainable and aesthetically pleasant, wood is one of the most tra-
ditional of construction materials. It also has a promising and innovative future in
architecture and design. Wood’s rich legacy ranges back ten thousand years, from the
use of timber in homes to today’s 3D-printed and robot-assembled wooden furniture.
With the current rise of free-form architecture and a growing interest in sustainable
materials, wood has once again become an important focus of contemporary design
research. Advanced fabrication tools are steadily eliminating many of the manu-
facturing limitations that previously prevented wood from being used in curved
and geometrically complex designs. This has led to a renewed interest in analyz-
ing and predicting wood’s behavior in relation to innovative architectural geometry
(Pottmann et al. 2008).
At the same time, other developments in materials science have introduced a new
generation of “smart” or responsive materials with advanced performance capabili-
ties. Smart materials can undergo changes in physical qualities such as form, color,
and flexibility in response to environmental stimuli such as moisture, heat, and light.
The use of responsive materials in combination with computational design and digital
fabrication has created a powerful shift towards adaptive designs, in which structures
may change their shape or rigidity according to the user’s needs (López et al. 2017).
One of many examples of such innovation is the “fluid assembly chair” that was pro-
totyped at the MIT Self-assembly Lab. In this design, several independent elements
come together to automatically form a chair when submerged in water (Tibbits et al.
2014).
To create such possibilities, designers must work in collaboration with special-
ists in materials science, fabrication, and mechanical engineering, and eventually
become fluent in these disciplines. By becoming knowledgeable about materials
development, designers will be able to take an active role in the process of creating
new materials to meet specific design purposes.
Since design culture has begun to synergistically integrate geometry, materials,
and the environment to provide intelligent and active solutions to design problems
(Oxman and Rosenberg 2007), we must ask what role traditional materials can play
in an era of adaptive and intelligent architecture. Wood, an excellent example of
a commonly used material in architecture, is often viewed as a static architectural
component. In reality, however, it has demonstrated a substantial potential to be
both responsive and transformable. Wood’s environmental responsiveness has been
examined through two main lenses. One is the material’s natural capacity to react
to humidity through reversible shape-changing (i.e., warping) (López et al. 2017).
Wood absorbs and releases moisture in response to the environment, which can result
in dimensional changes and bending.
Shape changes in wood’s bilayer structures in response to humidity are caused
by the material’s tendency to expand when wet. Wood’s outer and inner layers swell
when exposed to humidity and shrink when dried, but the inner layer’s change
Adaptive Wooden Architecture. Designing a Wood Composite … 705
is significantly smaller. Bilayer wood material bends because of the difference in

expansion in the two layers. Timoshenko (1925) first analyzed bending from ther-
mal expansion in bimetal and determined how contrasts in thickness and stiffness
defined and limited the resulting curvature. Bending by swelling in wood has been
explored (Timoshenko 1925) by various researchers, and perhaps most notably by B.
N. Ugolev in the late 1970s (Ugolev 1998). More recently, A. Menges and colleagues
reintroduced humidity-based responsive wooden structures as an architectural tech-
nique through a rigorous series of research studies, prototypes, and adaptive pavilion
designs (Menges 2012).
The more recent direction by which researchers have explored wood’s responsive-
ness is using 3D-printing technologies to create specially formulated wood–plastic
composite materials. This allows materials designers to determine grain structures
and other properties of the wood composite, so that the material’s shape changes can
be precisely controlled. Such approaches rely on wood’s tendency to expand when
wet, but incorporate a greater level of control of the material’s microscale proper-
ties to produce the desired macroscale mechanical performance, without requiring
additional electronic or mechanical stabilizers (Correa et al. 2015).
Previous efforts at adapting wood as a contemporary responsive material have led
to novel results, particularly with regards to specific weather-responsive elements.
Researchers have also achieved large-scale surfaces 1 m in length by joining multiple
designed wooden pieces (Wood et al. 2018) and increasing the moisture change rate
by milling grooves for faster actuation (Vailati et al. 2017). However, for producing
responsive wooden material for different applications, these methods remain lim-
ited in terms of the response time, range of curvature, and ability to continuously
change shape over time without incurring damage. One important aspect of these
methods is using wood’s natural and inherent capacities; however, this means they
invariably rely on the stimulus of humidity, which can significantly decrease the
range of practical applications (especially as compared to other responsive materi-
als). The 3D-printed composite materials also require a very complicated analysis
and technological fabrication process to be effective.
In order to propose a responsive wood-based material for large-scale architectural
purposes, our research has focused on initiating an alternative method of develop-
ing responsive wooden composites. The research goal was to develop a method that
offered greater precision, more extensive transformation abilities, and faster transfor-
mation speeds compared to the previous work done in this area. The manufacturing
process we used was simpler than that employed in many previous applications. Our
solution involves using a traditional relief-cutting method called kerfing to create
flexible wooden sheets, and then bonding these wooden sheets to an appropriate
responsive resin polymer. The kerfed wood allowed us to obtain desirable curvatures
while providing stability and attractiveness, and the responsive polymer permitted
the material to quickly reversibly transform in response to a temperature stimulus.
Fig. 1 From a flat surface to a 3-dimensional curved surface: a original flat surface, b zero-curve
bent surface, c positive double-curved surface, d negative double-curved surface or “saddle-like”
surface. e Zero curvatures can be made by folding a plane’s surface, while positive or negative
curvatures (i.e., c, d, e) cannot transform into 2-dimensional planes without some degree of surface
distortion. Images created by the authors
2 Curved Surfaces and Kerfing Techniques
Contemporary architecture tends to prominently feature curved surfaces, both as a

means of providing sturdy and aesthetically pleasing structural forms and to make
efficient use of space. These surfaces can be simple single curvatures such as cylin-
drical vaults, double curvatures such as domes, or complex combinations of single
and double curvatures, which are becoming more common in contemporary free-
form architecture. Gaussian curvature measurements are often used to describe these
curved surfaces. The Gaussian curvature (K) of a surface at any given point is the
product of the principle curvatures in each direction: K (κ1) · (κ2) (Branko 2009).
This means that both flat and single-curved surfaces possess zero Gaussian curva-
tures, since in those structures the curvature in at least one direction is zero (Fig. 1).
Surfaces with Gaussian curvatures of zero are easier to produce, because they
can be transformed or “developed” to the plane without any distortion. For example,
lifting up two opposite sides of a flat sheet of paper creates a simple single curvature
(K 0) with no distortion, stretching, or tearing of the material. In contrast, double-
curved surfaces such as domes (K 0) cannot be formed out of flat sheets without
stretching or tearing the material (Kalantar et al. 2016).
Since wood is not ordinarily a flexible material, conventional building techniques
that use wood are generally limited to flat or light single-curved surfaces that are most
Fig. 2 Kerfing allows the rigid material to bend through controlled compression/expansions per-
pendicular to the cut lines. Images created by the authors
frequently formed by bending wooden sheets until they warp into place. Techniques
to produce more complex surfaces involve joining a large number of individual
single-curved wooden strips to obtain the desired overall curvature, as is the case in
“grid shell” construction techniques (Pottmann et al. 2008).
One way to create highly flexible wood surface is the relief-cutting process known
as kerfing. The flexibility obtained from kerfing is produced by removing parts of
the material’s surface through a series of cuts (or kerfs). Removing material via this
technique allows for in-plane expansions and compressions perpendicular to the cut
lines (Fig. 2). In traditional fabrication, kerfing is done with a saw; it is a well-
established carpentry technique for producing a variety of gently curved geometries
in products ranging from tambour clocks to arched windows.
With the advent of CNC fabrication machines, designers can now make extremely
complex and precise patterns of cuts in wooden sheets. The new possibilities these
machines make available are what allowed us to create a wide range of bending abil-
ities that would not have been feasible through traditional kerfing (Fig. 3). Thus, the
kerfed wooden surfaces can have more complex single and double curved geometries
and even locally differentiated flexibility controlled by the designed cutting pattern
(Zarrinmehr et al. 2017).
For our purposes, kerfing provided a variety of desirable shape-changing potentials
in wooden surfaces; at the same time, the removal of portions of the wood created
appropriate spaces to fill with the temperature-responsive resin polymer material that
provided responsive properties. In our research prototype, we used a laser cutting
machine to apply a diamond lattice pattern of cuts (drawn in the Rhino software
platform) to an ordinary sheet of plywood (see Fig. 4). This relatively simple and
open design allowed for maximum bending of the sheet along one axis, and minimum
Fig. 3 Wooden curvilinear surfaces generated by kerfing flat sheets: a positive double-curved
surface, b negative double-curved surface, and c single-curved surface. Projects were completed
in the studio of Prof. Alireza Borhani at Texas A&M University by Chandler Fleming (a), Rotana
Hok (b), and Joseph Verd (c)
bending along the perpendicular axis. We labelled these as the “flexible axis” and
“stiff axis,” respectively (see Fig. 5).
To create our composite, we focused on single curved surfaces; the designed
pattern allowed us to achieve the maximum curvature along its flexible axis. This
pattern limits the possibility to achieve positive double-curved surfaces, however,
creating negative double-curved surfaces by bending diagonal axes was possible
(Fig. 6).
3 Developing and Testing the Responsive Polymer Resin
While the kerfed wood had the ability to transform from one stable shape to another,
it lacked the inherent impetus to do so. To add a responsive quality to the material, we
needed to integrate an advanced polymer that would adhere strongly to the wood and
provide a “smart” response to environmental stimuli. Rigorous efforts in materials
science over the past two decades have been focused on developing such materials,
so we had ample options to test and develop. We selected a shape-memory polymer
(SMP) with the ability to respond to temperature. In contrast to alternatives such
as shape-memory alloys, polymers are moldable, transparent, inexpensive to work
with, and have a greater bonding capability (Wei et al. 1998).
SMPs typically begin as liquid mixtures that are then solidified through the appli-
cation of heat. This means that they are very easy to shape; the process is as simple
is pouring the mixture into a mold or applying it with a brush prior to hardening. The
specific properties of the polymer, including the extent of its temperature response
and transformation, can be customized during the manufacturing process. Many of
these polymers are also transparent, which has the valuable advantage of highlighting
the aesthetics of the wood component. Perhaps most importantly, when creating an
SMP for use in a wood composite, an epoxy resin can be used as the base material
for the responsive polymer. Epoxies are widely employed as adhesives and coating
Fig. 4 Diamond lattice cutting pattern offered high flexibility and spaces appropriate for filling
with a temperature-responsive polymer material that provided responsive behavior
materials in wood design and manufacturing, and they have a significant bonding
affinity with kerfed wood.
To create a shape-memory polymer suitable for our purposes, we used an epoxy
(Epon 826) with a curing agent (Jeffamine D-230) and added neopentyl glycol digly-
col ether (NGDE). When combined and chemically cross-linked by heating, these
materials demonstrated a strong, temperature-mediated shape-memory effect (Xie
and Rousseau 2009). Temperature-responsive SMPs remain in a rigid state as long as
they are below their “glass transition temperature” (Tg). When the temperature rises
above Tg, they become flexible and can be reshaped through the application of force.
They retain the new shape after cooling. However, if the temperature rises above Tg
and no external force is present, the material will rapidly self-transform back into
its original manufactured shape. Thus, if a flat sheet of SMP is manufactured, it can
be heated and pressed into, for example, a dome shape, and then cooled in order to
Fig. 5 Flexible axis versus stiff axis
Fig. 6 Curved surfaces formed out of kerfed wood
retain that shape. Reheating the material in the absence of force will lead it to rapidly
revert back to its original flat position (see Figs. 7 and 8).
As discussed above, the shape-memory effect in our resin-based polymer was
the result of the SMP components and their reactions. The Jeffamine D-230 bonded
the components and hardened the polymer materials to create the SMP structure.
The NGDE diluted the polymer and strengthened it against cracking and brittle-
ness. The NGDE segments in the SMP became flexible and bendable when the
temperature was above the Tg, while the Epon 826 created the frozen segments.
The self-transformation to the original shape was the result of a reorientation of the
polyurethane molecular chains that encouraged crystallization; this could have been
repeated nearly indefinitely (Wei et al. 1998).
Before creating the wood composite, we first tested a small sample of the polymer
in isolation to confirm that we were using the right proportions and the resulting
Fig. 7 The general temperature response pattern of an SMP (on left); The SMP shape-changing
cycle (on right). Images created by the authors (adapted from Yang et al. 2016; Beites 2013)
Fig. 8 Coefficient of thermal expansion (CTE) versus temperature for: a deformation by loading
and b recovery without loading. The composite exhibited a significant ability to deform between
40 and 60 °C as well as at higher temperatures. Images created by the authors (adapted from Huang
et al. 2016)
polymer responded to temperature changes in the desired fashion (Fig. 9). In our
composite material, we wanted to achieve a Tg of approximately 40–60 °C. This
range allowed the material to remain firmly rigid at room temperature, while not
requiring overly excessive heating to reach the flexible state. To achieve the desired
Tg, we mixed 73.6 g of Epon 826, 46.5 g of Jeffamine D230, and 43.7 g of NGDE.
Fig. 9 Initial test of SMP responsiveness: a the deformed shape of a fabricated polymer sample
after applying heat and pressure, and b, c, d the shape recovery process at 2, 6, and 8 s, respectively
The fabrication process involved four simple steps: (1) mixing the appropriate
amounts of the three substances in a plastic cup and stirring vigorously with a wooden
stick for about 10 s; (2) removing any bubbles by placing the mixture in a heat-
resistant container, heating to 80 °C, and vacuuming to expel all of the air trapped in
the mixture for 40 min; (3) placing the mixture in a silicone rubber mold and heating
it in an oven at 100 °C for 90 min to cure; and (4) cooling the material. Then, the
SMP was complete (Xie and Rousseau 2009).
4 Fabrication of the Wood Composite
After successfully testing our SMP, we created the composite material by coating and
laminating the flat kerfed wooden surface with the temperature-responsive polymer
and heating it (see Fig. 10). We followed the process as explained above, but first
placed the wooden surface in the mold after carefully brushing the wood with the
liquid SMP mixture. We placed the materials in the silicone rubber mold and heated
them at 100 °C for 90 min. Then, the composite was cooled at room temperature to
obtain the desired composite (see Fig. 11).
The resulting composite obtained its responsiveness from the shape-memory poly-
mer and its shape-changing direction and much of its structural shape from the kerfed
wooden framework. The individual materials did not dissolve or merge completely
in the composite, but rather acted together as one. The two components could still be
physically identified, and the transparent nature of the polymer helped to highlight
the aesthetic qualities of the wood.
In our prototype tests, the combined SMP—wood laminate performed extremely
well. Unlike the isolated kerfed wood component, the composite material remained
fully rigid at room temperature. When the environment temperature was raised to
above 40 °C, however, the material became flexible and could be readily pressed into
any position allowed by the kerfed wood’s range of motion (see Fig. 11). Pressure
was applied to hold the material in a curved shape. Then, the heat source was removed
and the composite allowed to cool. After 20 s at room temperature, the composite
material cooled sufficiently to become rigid again and retain the new curved form
Fig. 10 SMP creation process
Fig. 11 Performances of wood-SMP composite at room temperature (left image: rigid composite)
and 40 °C (right image: flexible composite)
(see Fig. 12). When once again heated to above Tg and without applying any force,
the material “remembered” its original flat shape and returned to it within 48s. This
cycle could have been continuously repeated (see Fig. 13).
Fig. 12 Process of surface bending by heating with external force, then cooling to obtain a rigid
surface
Fig. 13 Process of shape self-recovery by heating without an external force
5 Conclusion
Our research demonstrates a new and unique way of combining refurbished carpentry
techniques with “smart” polymers to create innovative wood-based surfaces. We
developed a novel method for producing responsiveness and promoting reversible
shape-changing characteristics in a wooden surface by laminating an epoxy resin-
based polymer onto kerfed wood. The outcome is a wooden surface that can take on
attractive and complex geometries and is capable of changing shape in response to
the environmental stimulus of heat (see Fig. 14).
The composite surface can be transformed into a variety of curvilinear shapes
through the application of heat and pressure, followed by cooling. Upon being heated
again in the absence of external pressure, the composite reverts to its original shape
within a few seconds. This material offers designers the opportunity to reconcep-
Fig. 14 The responsive wood-polymer composite can be used to transform to a variety of stable
curvilinear surfaces, such as flat surfaces, single-curved surfaces and saddle-like negative double-
curved surfaces
tualize the use of wood in contemporary architecture. In addition to providing new

conceptual and aesthetic opportunities for design, there are also practical benefits in
terms of shipping. These wooden surfaces can be returned to a flat shape as often as
desired, facilitating storage and transportation.
In future projects, we plan to continue studying the composite’s performance
by applying different relief cutting patterns that will encourage alternative curva-
tures and degrees of flexibility. In addition, ongoing work in this area should focus
on improving the wood–composite interface, and analyzing the recovery times and
temperature ranges for specific shape deformations. While there is literature that sep-
arately discusses the properties of wood and SMP, the mechanical characteristics and
structural properties of this new composite as a whole (which behaves differently
from either of its two components) needs further investigation. The inherent brit-
tleness of polymers, for example, often limits their application, and so additional
work should determine if there are similar concerns with the structural properties of
this new composite. This additional research will help us to examine the applicabil-
ity of the proposed technique on a broader architectural scale, which is one of the
main goals of this ongoing research. Therefore, the relationship between strain on
the composite surface and failure/deformation at various temperatures will benefit
future work in this area.
Acknowledgements This project was funded by the National Science Foundation’s EAGER Award
#1548243 and titled “Interaction of Smart Materials for Transparent, Self-regulating Building
Skins.” A previous description of this ongoing project was published in the CAADRIA 2018 con-
ference proceeding (Mansoori et al. 2018).
References
Beites S (2013) Morphological behavior of shape memory polymers toward a deployable, adaptive
architecture. In: Association for computer aided design in architecture (ACADIA) 2013: adaptive
architecture, University of Waterloo Cambridge, Ontario, Canada, pp 121–127, 24–26 October
2013
Branko K (2009) Digital production. In: Branko K (ed) Architecture in the digital age: design and
manufacturing, 2nd edn. Taylor and Francis, New York, pp 44–48
Correa D, Papadopoulou A, Guberan C, Jhaveri N, Reichert S, Menges A, Tibbits S (2015)
3D-printed wood: programming hygroscopic material transformations. 3d Print Addit Manuf
2(3):106–116. https://doi.org/10.1089/3dp.2015.0022
Huang WM, Zhao Y, Wang CC, Ding Z, Purnawali H, Tang C, Zhang JL (2016) Thermo/chemo-
responsive shape memory effect in polymers: a sketch of working mechanisms, fundamentals
and optimization. J Polym Res 19:9952. https://doi.org/10.1007/s10965-012-9952-z
Kalantar N, Borhani A, Akleman E (2016) Nip and tuck: a simple approach to fabricate double
curved surfaces with 2d cutting. In: eCAADe 2016, 1: 344–337, Aulikki Herneoja Toni Österlund
Piia Markkanen (ed) Oulu School of Architecture, Finland, 22–26 August 2016
López M, Rubio R, Martín S, Croxford B (2017) How plants inspire façades. from plans to archi-
tecture: Biomimetic principles for the development of adaptive architectural envelopes. Renew
Sustain Energy Rev 67(C):692–703. https://doi.org/10.1016/j.rser.2016.09.018
Mansoori M, Kalantar K, Creasy T, Rybkowski Z (2018) Toward adaptive architectural skins design-
ing temperature-responsive curvilinear surfaces. In: CAADRIA 2018, Osaka T, Huang W, Patrick
C, Crolla Chinese K, Alhadidi S (ed) Tsinghua University, Beijing, China, 17–19 May 2018
Menges A (2012) Material computation: higher integration in morphogenetic design. Archit Design
82(2):14–21
Oxman N, Rosenberg JL (2007) Material based design computation material-based design compu-
tation: an inquiry into digital simulation of physical material properties as design generators. Int
J Archit Comput 1(5):25–44
Pottmann H, Schiftner A, Wallner J (2008) Geometry of architectural freeform structures. Int Math
Nachr 209(2):15–28
Tibbits S, Zuniga B, McKnelly C, Papadopoulou A (2014) Fluid-assembly chair. [online] Available
at: https://selfassemblylab.mit.edu/fluid-assembly-chair/. Accessed 21 Mar 2017
Timoshenko S (1925) Analysis of bi-metal thermostats. J Opt Soc Am 11:233–255. https://doi.org/
10.1364/JOSA.11.000233.
Ugolev BN (1998) Effect of freezing wood deformations at complex force and heat actions. In: 2nd
international symposium on wood rheology, Rydzina
Vailati C, Hass P, Burgert I, Rüggeberg M (2017) Upscaling of wood bilayers: design principles for
controlling shape change and increasing moisture change rate. Mater Struct 50(250). https://doi.
org/10.1617/s11527-017-1117-4
Wei ZG, Sandstrom R, Miyazak S (1998) Review Shape memory materials and hybrid composites
for smart systems: part I Shape-memory hybrid composites. J Mater Sci 33(15):3763–3783
Wood D, Vailati C, Menges A (2018) Hygroscopically actuated wood elements for weather respon-
sive and self-forming building parts: facilitating upscaling and complex shape changes. Constr
Build Mater 165:785–787. https://doi.org/10.1016/j.conbuildmat.2017.12.134
Xie T, Rousseau IA (2009) Facile tailoring of thermal transition temperatures of epoxy shape
memory polymers. Polymer 50(8):1852–1856
Yang Y, Chen Y, Wei Y, Li Y (2016) Novel design and three-dimensional printing of variable
stiffness robotic grippers. ASME J Mech Robot 8(6):061010–061010-15. https://doi.org/10.1115/
1.4033728
Zarrinmehr S, Ettehad M, Akleman E, Kalantar N, Borhani A (2017) Kerfing with generalized
2D meander-patterns conversion of planar rigid panels into locally-flexible panels with stiffness
control. In: CAAD Futures 2017, Istanbul, Turkey, 12–14 July 2017
Geometry-Induced System of Controlled
Deformations. Application
in Self-organized Wooden Gridshell
Structures
Efilena Baseta
Abstract This chapter presents a novel construction system which offers an efficient
materialization method for double-curved surfaces. This results in an active-bending
system of controlled deformations. The latter system embeds its construction manual
into the geometry of its components, thus it can be used as a self-formation process.
The two presented gridshell prototypes are composed of geometry-induced, variable
stiffness elements. The latter elements are able to form programmed shapes passively
when gravitational loads are applied. Each element consists of multiple layers and
a slip zone among them. The slip allows the element to be flexible when flat and
increasingly stiffer when its curvature increases. The presented system eliminates
the need for electromechanical equipment since it relies on material properties and
geometrical configurations. Wood, as a flexible and strong material, has been used for
the prototypes. The fabrication of the timber laths has been done via CNC industrial
milling processes. The scalability of the system shows potential for applications in
large-scale transformable structures. The comparison between the predefined digital
design and the resulting geometry of the physical prototypes is reviewed here. The
aim is to inform the design and fabrication process with the extracted performance
data and thus, optimize the system’s behaviour.
Keywords Variable-stiffness · Active bending · Self-organization processes

Flexible wood · Wood joint systems
1 Introduction
The paradigm shift from mass production to mass customization and the technolog-
ical burst in the second half of the 20th century changed radically the predominant
design thinking. However, unlike shipbuilding, automotive and aircraft industries,
E. Baseta (B)
Institute of Architecture, University of Applied Arts Vienna, Vienna, Austria
e-mail: efilena@noumena.io
URL: http://www.dieangewandte.at/

720 E. Baseta
construction industry was not able to effectively adjust to the new reality and exploit
the full potential of the new technologies that the digital era offered. Thus, the con-
struction industry remains labor intensive and unsustainable (Frazer 1995).
The urgency of that shift was not only due to changes in economic models of
production but also due to social and environmental changes. After the mid of 20th
century, our society underwent rapid transformations, which derived by the faster
means of transport and communication that enabled tremendous flux of information
and globalization. According to Castells, space became fluid (space of flows) and
gained dynamicity, influenced by the perpetual transformations of the modern society
(Castells 1992). This instability is the cause why architects and engineers started
looking for inspiration into nature since the latter is an evolving system that can
adapt to ever-changing conditions, and thus survive. As a result, adaptive structures
appear as a key to the aforementioned problematic, allowing a system to react in
real-time to specific inputs deriving from its surroundings (e.g. weather data, use of
space and number of users). The direct repercussion of that is the sustainability of
the built environment.
In this framework, observing living organisms, scientists found that they are
able to form hierarchically organized structures through genetically controlled self-
organization, which allows their survival (Knippers et al. 2016). The process of
self-organization is activated each time that the equilibrium of their ecosystem col-
lapses (Weinstock 2004). The principle of self-organization has concerned several
researchers, due to the fact that design complexity can emerge through simple ele-
ments and rules (Frazer 1995). By reverse engineering, applying hierarchical princi-
ples in man-made structures, self-organized systems can emerge. Consequently, new
design processes which confront complexity can be defined.
2 Background
2.1 Transformable Structures
Transformable (form-adaptive) structures have been increasingly concerning the

architecture community for the last years. Besides prototypes made in academic con-
text, industrially driven projects have been developed with scales varying from man-
sized products (prosthetics and robotics) and façade components (kinetic facades
and adaptive photovoltaics) to self-supported structures (airfoils and structural parts)
(Galloway et al. 2013; Barozzi et al. 2016; Jayathissa et al. 2016; Kuder et al. 2016;
Senatore et al. 2017).
The aforementioned projects employ various means to achieve shape-adaptation,
such as shape/phase-changing materials, pneumatic and electromechanical systems
as well as geometry-induced, variable stiffness structures. Most applications relying
on smart materials are still constrained to small scales or short life cycles, due to low
strength and high fatigue respectively. On the contrary, electromechanical systems
Geometry-Induced System of Controlled Deformations … 721
and pneumatics allow large-scale and long-term shape-adaptation. However, they

become unsustainable since they add complexity and weight to the designs. In addi-
tion, they increase maintenance and energy costs. On the contrary, self-organized
structures, which exploit material properties and hierarchical geometrical configura-
tions, have been proven adequate for cost-efficient adaptivity in large-scale systems.
2.2 Self-formation Processes in Bending-Active Structures
Self-organization processes applied in contemporary architecture appeared in the

beginning of the 20th century. An outstanding precedent is the hanging-chain models
of Colonia Güell church and Sagrada Familia by Antoni Gaudi (Burry et al. 2005).
At Gaudi’s hanging-chain models, equilibrium states were reached by unevenly
distributed gravitational loads along uniform members (strings) (Huerta 2006).
This approach complies with D’Arcy Thompson’s understanding of form as it is a
system which organizes itself in the presences of both internal and external forces
(Thompson 1961).
Another pioneer who embedded self-organization driven by natural laws in design
processes was Frei Otto. Otto used the equilibrium of natural forces as his only design
criterion. An example of that was the prototype for Mannheim Multihall. However,
passing from small-scale form-finding prototypes to large-scale constructions many
difficulties are found. During the erection of the Mannheim Multihall extra shear
blocks had to be added manually between the layers of the gridshell in order to
increase its stiffness and stabilize it. In addition, many laths broke. Despite the fact
that Mannheim Multihall is one of the best examples of large-scale gridshells, its
complicated construction process led to a final shape deviating from the original
(Liddell 2015).
In recent times, active bending is being used as a self-formation process. Active
bending is an efficient form-giving process which exploits elastic deformation to
shape initially straight linear elements or planar surfaces into curved configurations.
It has been widely used in vernacular architecture since it is a simple construction
method for deployable lightweight structures such as yurts (Lienhard et al. 2013a).
The last 50 years, active bending has been widely used for the construction of large-
span gridshell structures. This is due to economic reasons and more specifically
due to: (a) volume reduction of transportation, (b) a single production process for
elements with different radii of curvatures and (c) structural performance and adapt-
ability (Gengnagel et al. 2013).
Nevertheless, the disadvantage of the aforementioned form-giving process is the
initial stress caused to the elements by bending. This has been apparent during the
erection process of large span gridshells where many breakages occurred. In order to
reduce that, for a specific material, either the cross-sectional height should decrease
or the radius of curvature should increase. This is a conflict due to the fact that
smaller cross-sectional height reduces the stiffness of a structure and bigger radius
of curvature limits the design possibilities of curved structures. An optimized active
722 E. Baseta
bending structure would allow small radii of curvature and obtain high stiffness. The
aforementioned conflict results in scaling limitations of active bending structures,
which also depends on the strength of the used material (Lienhard and Knippers 2013;
Gengnagel et al. 2013). According to recent studies, the most efficient materials for
active bending are wood, GFRP and NFRP (Kotelnikova-Weiler et al. 2013).
To overcome the scaling limitation, researchers developed hybrid structures.
The latter structures reach an equilibrium state by equating the compressive forces
induced by tensile fabrics and the tensile forces induced by elastically bent rods (Lien-
hard et al. 2013b). In hybrid structures, elements such as textiles or cables are added
to stiffen the bent elements. Thus, the cross-sectional height can be reduced as well as
the radius of curvature. A precedent of this construction method is the MoD-shelter,
which consists of membrane restrained arches which span 5 m (Gengnagel et al.
2013). Another precedent is the M1 which consists of actively bent rods restrained
by a membrane (Lienhard et al. 2013b). Other approaches to overcome the low stiff-
ness of bending active structures have been conducted at the StretchPLAY prototype.
For the latter prototype, laminated beams were created by bundling together three
GFRP rods with a knitted sleeve. In order to remove the shear between the three rods,
and thus increase the overall stiffness, the laminated beams have been impregnated
with epoxy resin once the structure has been shaped in its curved shape (Ahlquist
2015).
A further advantage of the bending active systems is their flexibility. The lat-
ter property allows the creation of compliant mechanisms. Thus, bending active
systems can find applications in elastic kinetic and shape adaptable structures. How-
ever, precedents of the latter structures remain in components scale supported and
controlled by electromechanical equipment. Recent examples are the kinetic louvers
Flectofin and Flectofold which fluctuate from closed to open configuration (Lienhard
et al. 2013a; Körner et al. 2016). Besides that, shape-adaptive photovoltaic textiles
have been made for the Softhouse (Lienhard and Knippers 2015). Experimental pro-
totypes of self-supported, bending active, transformable structures have been realized
recently and rely on biaxial braided systems. In the latter systems, the stiffness of
the structure fluctuates as the applied loads and support conditions change. Despite
the fact that the prototypes show potential for applications in large-scale, they have
still limited freestanding capabilities due to the slenderness of their parts (Sparrman
et al. 2017).
3 Inspiration from Microscale and Research Objectives
3.1 Self-actuating Mechanisms of Natural and Synthetic

Composites
Passive, transformable structures in microscale (here 0.01–0.1 m) have been obtained

through custom, synthetic, self-actuated composites such as Natto-Cells biofilm, pho-
toresponsive polymer springs, 3D-Printed Wood and 3D printed hydrogel composites

(Yao et al. 2015; Iamsaard et al. 2016; Correa et al. 2015; Sydney Gladman et al.
2016). The aforementioned smart materials can adapt to external stimuli finding
applications in medicine, soft robotics, aeronautics, design field and self-assembly
processes (Lendlein and Langer 2002; Lan et al. 2011; Jenkins and Landis 1995;
Farahi 2016; Farahi et al. 2013; Raviv et al. 2014; Wood et al. 2016). On the other
hand, self-actuated mechanisms exist also in nature. The wheat awn, the pine cone,
the ice plant and the awn of the Pelargonium are some of the plants whose movements
and structures have been studied by scientists (Burgert and Fratzl 2009; Guiducci
et al. 2016; Jung et al. 2014).
Examining the self-actuated behaviors of natural and synthetic composites, a com-
mon principle has been noticed: A hierarchical structure, consisting of two rigidly
connected parts (bi-layer principle) with different properties, defines the deforma-
tion. Each bi-layer consists of one active part, which responds to the stimulus (e.g.
humidity), and one resistive, which regulates the deformation. In some cases, the
differentiation of properties of the parts relies on the orientation of microfibrils
(anisotropic materials) across the thickness of the part. In other cases, it relies on
the differentiation of the material itself. Conclusively, the programmed deformation
depends on the relation between the parts. This complies with the systemic design
approach described by Christopher Alexander, who states that every object/system is
a hierarchy of components and its behavior is accomplished through the interaction
among parts. The knowledge of the components has to be complemented with the
knowledge of how the components interact (Alexander 1964).
3.2 Wood as a Case Study Material
Wood, as a natural fibrous cellulose-based composite, demonstrates self-actuated

behaviors in microscale. More specifically, it is characterized by a hygroscopic behav-
ior, an ability of a material to absorb and release moisture from the surrounding atmo-
sphere. An increase of the water in the cell walls expands the distance between the
micro-fibrils of the cell tissue resulting in a significant dimensional change and vice
versa. Thus, when combined, in a bi-layer system, with another non hygroscopic
part, it can create bending or twisting forms due to humidity increase. Moreover,
wood, as an anisotropic material, its fibers’ orientation defines its flexibility and
consequently its deformation. This complex property allows wood to compete with
high-tech contemporary, passive, shape-shifting materials (Reichert et al. 2015). The
hygroscopic property of wood has found applications in experimental projects in the
field of architecture. The latter projects focus either on adaptive envelope skins or
adaptive photovoltaics. However, no self-supported, transformable structure has been
developed, due to the difficulty of scaling up the material properties (Reichert et al.
2015; Holstov et al. 2015; Rüggeberg and Burgert 2015).
724 E. Baseta
Fig. 1 Bi-layer systems of controlled deformation made of wood; a, b hygroscopic system in

microscale where the deformation is defined by the fibers orientation of the resistive layer; c, d
scalable, variable stiffness system where the deformation is defined by the connecting geometry
between the layers
From the above derives a question: How principles of self-organized material sys-
tems in microscale can be scaled-up and find applications in transformable structures
in the construction industry?
Considering that wood is a renewable material that is widely used in construction
industry, it has been selected as a case study material, in an attempt to transfer its
self-actuating bending behavior in microscale (0.1 m) to a self-organized material
system in macroscale (10 m). Applying the aforementioned, bi-layer principle of
self-actuated composites in mesoscalar (1 m) active bending elements, it has been
observed that controlled deformation of the system can be achieved by regulating
the relative movement (slip) between the layers. In this case, the stimuli are forces
that are induced externally (Fig. 1). In the following sections, a novel construction
system based on the aforementioned observation is described.
3.3 Research Framework
As mentioned above, the hierarchy and relation of the parts of self-actuating sys-
tems can be the driving force for controlled deformation. The transfer of the bio-
logical principles of morphogenesis and evolution to architecture can lead to an
inseparable relation between material and structure, and thus to the creation of self-
supporting systems (Frazer 1995). Thus, this research aims to create self-supported
transformable structures relying on geometrical and material explorations rather than
mechanical parts and electronics. As such, passive actuators can emerge, minimiz-
ing the energy and the maintenance needed compared to conventional shape-adaptive
systems.
As mentioned in Sect. 2, structures that employ geometry-induced, variable stiff-
ness principles have been proven advantageous for scalable cost-efficient, shape-
adaptive applications. As a result, the proposed system intends to facilitate the real-
ization of self-supported, shape-adaptable, bending-active structures by developing
passive variable stiffness elements. This offers alternative ways to achieve scalable
active bending structures with a more efficient stiffness-radius relation, eliminating
the need of additional elements (e.g. membranes, cables, shear blocks) and stiffening
processes after assembly (e.g. impregnation of laminated beams).
The result is an active-bending system of controlled deflections. The latter system
embeds its construction manual into the geometry of its components, thus it can
be used as a self-formation process. When this process is reversible, transformable
structures can be achieved. Considering that the evolutionary character of nature
enables it to adapt, and thus withstand to the ever-changing conditions, responsive
architecture can lead to a sustainable built environment.
4 Geometry-Induced Variable Stiffness Elements
4.1 System’s Principle of Operation
The proposed construction system aims to create shape-adaptable elements, which

shift from flat and flexible, to deformed and stiff configurations, and vice versa, due
to the change of the applied loads. The operating principle of the system relies on
controlled active-bending linear elements. Controlled displacement is achieved by
the juxtaposition of multiple layers (minimum two) and insertion of shear blocks
among them. By embedding rectangular shear blocks in the layers, notched linear
elements are created. The presence of transversal gaps between the shear blocks
of two consecutive layers, along the longitudinal axis, allows slip among the layers.
Through dimensional and geometrical variations of the gaps, the slip can be regulated,
and thus variable stiffness can be obtained. One locking point among the layers, where
no slip occurs, is fundamental in order to regulate the deformation (Fig. 2).
To further clarify the behaviour of the system, a linear element consisting of two
layers of equal cross section height is considered. Each layer expands at its upper
surface and contracts at its lower. As a result, the relative displacement Δl of two
initially adjacent points (one on the bottom of the upper layer and one on the top of
the lower layer) between the flat and the elastically bent state, is doubled compared
to the displacement of a point on a single layer (Δl Δll + Δlu ). This is because
there is positive and negative stain at the lower (l) and upper (u) layer respectively
(Fig. 3) (Baseta et al. 2018).
By creating gaps of lengths Δl between the shear blocks of a notched double-
layered element, the latter obtains its maximum stiffness as soon as the aforemen-
726 E. Baseta
Fig. 2 Cantilevering double-layered beam with 2 states: (1) Flat state with no loads, (2) Deformed
state induced by force F where beam (a) includes shear blocks and thus its deformation is controlled,
while beam (b) has no shear blocks and thus it deforms freely (Baseta et al. 2018)
Fig. 3 Displacement of adjacent points of a double-layered beam due to constant moments (Baseta
et al. 2018)
tioned bent state 2 is reached. Thus, the system exploits the flexibility of the small
cross section heights h of its individual layers to allow shape-adaptation, while also
preventing failure due to its final statically enhanced cross section h . The latter occurs
when the gaps between the shear blocks are closed and shear resistance between the
layers is activated. Consequently, the bending stiffness I2b of the double-layered
3
beam without shear blocks would be twice the bending stiffness of its parts I bh 12
(Fig. 2). However, the bending stiffness of the deformed, notched, double-layered
beam I2a can theoretically reach 8 times the bending stiffness I of its parts. This is
summarized in mathematical terms as follows:
Fig. 4 a Tabled scarf joint detail of the element; b double-layered variable stiffness element. State
(a) shows the element straight and stiff when its top layer faces upwards. When the element is
rotated 180° around its longitudinal axis (top layer faces downwards) it becomes flexible and starts
to deform until it reaches its predefined bent state (b) where it reaches again its maximum stiffness
(Baseta and Bollinger 2018)
3
bh 3 bh 3 bh 3 h 2h b(2h)3 bh
I2b + 2I and I2a ⇒ I2a 8 8I
12 12 12 12 12
(1)
When the aforementioned linear elements are used in gridshell configurations,

they can create surfaces which are able to shift from flat to doubly-curved and vice
versa. Thus, the proposed system leads to a novel construction system which facili-
tates controlled deformations (Baseta et al. 2018).
4.2 Design Process and Geometrical Parameters
In order to design and fabricate the aforementioned system, a curve representing the
predefined deformation should be given. On top of that, the boundary conditions
(loads and supports) and material properties of the proposed structure should also be
known. The cross section height of each layer depends on the maximum curvature
of the given curve and the number of the layers on the load case scenario.
Regarding the joinery detail (shear blocks) between two consecutive layers, this
research focuses on repetitive, tabled scarf geometries. The proposed geometry
blocks upward deformations (Fig. 4). Although, variations of the shear blocks geom-
etry can lead to different behaviours and thus applications of the system. Generally,
the length of each ‘tooth’ defines the resolution of the deformed curve and its height
defines the magnitude of the exerted shear forces. Finally, the calculation of the
effective cross section height of the notched linear elements and the length of their
gaps are vital to describe the system and therefore are explained in the following
sections (Baseta et al. 2018).
728 E. Baseta
Fig. 5 a System with constant moment for testing the beams A and B; b longitudinal section and
graphs of u and u for notched beam; c longitudinal section and graphs of u and u for rectangular
beam (Baseta et al. 2018)
4.2.1 Cross Section Height Calculation
In order to specify the maximum elastic bending of each notched layer, the effective
cross section height h3 of the latter needs to be calculated. Prior to that, the effective,
at bending, lengths of l1 , l 2 that correspond to the variable cross section heights h1 ,
h2 must be calculated (Fig. 5). For this reason, it has been assumed that 80% of the
tooth length l2 is effective. Further FEM analysis is needed to verify this assumption.
Given that, the effective lengths l1 , l2 of the notched layer are calculated as follows:
8 2
l2 l2 , l1 l1 + l2 and l l1 + l2 l1 + l2 (2)
10 10
Subsequently, to calculate h3 , a simply supported beam AB with 2 forces F at C,
D at a distance e from the supports is considered. Both the notched beam A and a
rectangular beam B are tested (Fig. 5). Assuming that the moment M between C and
D is constant (M F e) and the displacement of beam A and beam B are equal, i.e.
ua ub u, the first step to calculate h3 is to take the double integral of curvature
u to find displacement u:
M M M l2
u ,u l + c1 , u + c1l + c2 , (3)
EI EI EI 2
where E is Young’s modulus, I is the second moment of inertia and c1 , c2 are inte-
gration constants.
For beam A, the second moment of inertia is denoted I 1 for length l 1 and I 2 for
length l2 , whereas for beam B I 3 for length l. Drawing the graph of u with respect
to l for beams A and B, deflection u is calculated as the area between the graph and
the x axis:
l22 l1 l2 l 2 (l + l )2

ua ub ⇒ + + 1 1 2 (4)
2I2 l2 2I1 2I3
Solving (3) with respect to I 3 the effective cross section height h3 is:

l2 bh 33 3 12I3
I3 ⇒ h3 with h 1 < h 3 < h 2 (5)
2u a 12 b
4.2.2 Gap Length Calculation
Following the previous analysis to find the effective cross section height h3 , the
length of the gaps along the given curve can be calculated. Considering the radius
of curvature r along the curve, as well as the boundary conditions of the proposed
system, the position of the locking point can be defined. Subsequently, for a pre-
defined deformation, the lengths of the gaps (gap) along the longitudinal axis are
calculated according to the strain values of adjacent points between two consecutive
layers (Figs. 2 and 3). The strain ε for each point is:
Δl h/2
ε ⇒ Δl εl, ε ± , (6)
l r
where r is the radius of curvature and h is the cross section height of a single layer.
Integrating the strain along the length of the upper (u) and lower (l) layer, starting
from the locking point of 0 deformation (t 0) and ending at the end of the linear
element (t 1), the gap length is calculated as follows (Fig. 6):
t0 t0
h/2 h/2
Δlu εldt, ε − and Δll εldt, ε + (7)
r r
t1 t1
Δl Δll − Δlu ⇒ gap gapl − gapu (8)
5 Physical Experiments
Physical experiment has been chosen as a method to evaluate the potential use of the
aforementioned geometry-induced, variable stiffness elements for the self-formation
of gridshells. Thus, two experimental prototypes have been built. The design and
digital fabrication process of the latter prototypes are reviewed in Sects. 5.1 and 5.2
(Baseta and Bollinger 2018).
730 E. Baseta
Fig. 6 Graph of gap lengths in respect to the length of the predefined curve (a) with constant
curvature u and locking point at its left end. The resulting notched beam is illustrated in (b)
(Baseta et al. 2018)
Fig. 7 a Suspended gridshell designed by form finding process. The colors and numbers represent
the curvature values along the laths; b plan view with numbered elements (Baseta and Bollinger
2018)
5.1 Suspended Gridshell
The first prototype refers to a suspended double curved gridshell (6 × 3 m footprint).

The design has been realized by digital form-finding processes. More specifically
gravitational loads have been applied to an initially flat mesh. The equilibrium state
of the specific mesh and loads was frozen and rationalized into a distorted (30°)
rectangular grid that consists of planar curves with various curvatures (Fig. 7). The
primary structure consists of 12 laths (a–l) with single signed curvature and the 8
laths (1–8) of the secondary structure have double signed curvature with up to two
inflection points.
Fig. 8 a, b Robotic fabrication and resulting timber lath layers (Baseta and Bollinger 2018)
Subsequently, the aforementioned planar curves were given to an algorithm (see

Sect. 4.2.2) which calculates the appropriate gap lengths along the linear elements.
This leads to the design of flat double-layered notched laths, which are deformed
to the predefined shape when loads are induced. For the single signed curves, the
locking point between the layers was inserted at one of their ends. For the curves with
one inflection point, the locking point was placed where the sign of the curvature
changes. Finally, the curves with two inflection points, of the secondary structure,
were left flexible in their middle part to deform while being supported by the primary
structure.
The cross section dimensions of the double-layered laths were 15 × 20 mm (5 mm
gap height) which was due to the minimum radius of curvature of the design (r
1.4 m). The laths were made of white ash and due to the slenderness of their parts and
the fabrication precision needed, robotic CNC milling was selected as a fabrication
method. With a 6 mm milling bit and a pneumatic clamping mechanism mounted on
the wall, each linear element (max 3 m long) was fabricated in 7 min (Fig. 8). This
results to a fabrication time of approximately 4 h. 1 mm tolerance was given to the
notched geometries and small fabrication imprecisions (0.5 mm) were created by the
vibration of the robot when cantilevering. The laths that were longer than 3 m were
glued via diagonal shifters. Finally, the two layers of each lath were joined together
with zip ties.
For the assembly of the structure, the elements were placed on a DOKA frame
so that their both ends are free to slide. The primary structure was placed first and
then the secondary. Subsequently, loose zip ties were put around the nodes of the
gridshell as a quick joint that allows sliding between the primary and the secondary
structure. When equal loads of 5 N were hanged from each node, the grid formed
a double-curved shape (Fig. 9). The resulting shape relies on the variable internal
geometry of each double-layered lath which defines its maximum deformation and
732 E. Baseta
Fig. 9 Suspended gridshell; a dead loads; b gravitational loads applied at all the nodes (Baseta and
Bollinger 2018)
consequently the global deformation of the gridshell. The installation process was
uncomplicated and quick (Baseta and Bollinger 2018).
5.2 Cantilevering Gridshell
The second prototype refers to a cantilevering gridshell structure (5 × 4 m footprint).

It consists of five (1–5) double-layered beams in the primary direction and four (a–d)
in the secondary. Spruce GL24 was chosen for the materialization of the beams as
it is broadly used in the construction industry. In addition, the cross section of each
double-layered beam was decided to be 60 × 60 mm due to fabrication constraints.
The ‘tooth’ height was 1/3 of the beam height and its length (180 mm) was constrained
to the width of the clamping mechanism of the CNC machine. Each beam of the
primary structure has been designed to have constant curvature. However, its five
beams exhibit different curvatures, decreasing gradually from beam 1 (r 7.6 m) to
beam 5 (r 60 m). The primary structure defines the deformation of the gridshell,
while the secondary, perpendicular to the first one, follows the curvature of the latter
(Fig. 10).
One of the aims of this prototype was the industrial fabrication of the notched
beam. Thus, Hundegger K3 was chosen as an industrial CNC milling machine opti-
mized for rapidly machining joinery details of long, flat timber beams (Fig. 11a).
Considering that standard beams used in the construction industry have few joints
and larger cross sections than the layers of the proposed variable stiffness beams,
the fabrication of the latter beams was challenging. The first fabrication test was
conducted for the single layers of the 5 m long beams (Spruce C24 40 × 60 mm)
which were machined from the bottom. Due to their small cross section, the clamp-
ing mechanism of the K3 was unable to hold them still. On top of that, their lightness
allowed high vibrations during the machining of the ends of the beams, which led to
Fig. 10 Cantilevering gridshell; a perspective view showing the radius of curvature of the beams
and the moments (green) of the primary structure; b plan view with numbered elements (Baseta
and Bollinger 2018)
Fig. 11 a Hundegger K3 at Blumer Lehmann AG; b–d evolution of milling quality (Baseta and
Bollinger 2018)
breakages. As a result, the milled pieces had rough finishing and imprecisions of the
scale of centimeters (Fig. 11b, c).
Due to the unsatisfactory results of the abovementioned fabrication test, a second
one was conducted. The intention was to increase the weight of the milled beams
and thus minimize the vibrations. The two layers of each double beam were milled
back to back out of a single glue-laminated beam (Spruce GL 24) of bigger cross
section 60 × 100 mm. The ‘teeth’ of the two layers were aligned in order to offer a
strong grip to the pneumatic clamping mechanism of K3. Thus, the latter could tightly
grasp the beams from their full cross section width (100 mm) and slide them along
the rail of K3, where they were milled simultaneously at both sides. The fabrication
ran without problems and the resulting beams had a good finishing (Fig. 11d). The
1.5 mm tolerance that was given to the fabricated pieces proved adequate for the
given machine. The fabrication time of each double-layered beam took approximately
19 min. Consequently, the milling of all the elements lasted 3 h. Finally, the beams
were sliced along their longitudinal axis with an electric saw so as to separate the
upper from the lower layer. The layers were joined together with zip ties.
During the assembly, the cantilevering beams were mounted on a fixed base and
subsequently the secondary beams were placed on top. The installation was simple
734 E. Baseta
Fig. 12 Cantilevering gridshell; a dead loads; b gravitational loads applied at the 5 cantilevering
tips (Baseta and Bollinger 2018)
and completed by two people in a couple of hours. When loads of 50 N were added
at the ends of the primary structure, the gridshell formed a double curved surface
(Fig. 12). Despite the fact that all the beams of the primary structure had the same
cross section and the same loading, beam 5 had short gaps between its layers and thus
allowed small slip and thus small deflection (stiffest beam). On the contrary, beam
1 had longer gaps and thus allowed bigger slip and larger deflection (most flexible
beam) (Baseta and Bollinger 2018).
6 Results from Physical Experiments
The aforementioned prototypes were the first experiments to prove that the geometry-
induced, variable stiffness beams can be used for the self-formation of gridshells.
From Sect. 5 it has been evident that there is a qualitative correspondence between the
predefined form and the resulting geometry (Figs. 9 and 12). However, the detailed
comparison between the initially designed curves and their physical form has been
considered as a crucial step to evaluate the performance of the system. The quantita-
tive results are described in Sects. 6.1 and 6.2 for the suspended and the cantilevered
gridshell respectively.
6.1 Geometrical Evaluation of Suspended Gridshell
The physical form of the suspended gridshell was extracted from a point cloud col-
lected by a laser 3D scanner. Juxtaposing the point cloud over the digital design
showed that the physical form has a maximum deviation of 34 cm in the middle
of the structure (Fig. 13a). Additional small deviations of maximum 100 mm have
been noticed in xy plane (Fig. 13b). A comparison between the physical and the dig-
Fig. 13 Comparison between the 3D scan of the suspended gridshell (brown) and the digitally
designed form (grey/black) (Baseta and Bollinger 2018); a perspective view; b top view
Fig. 14 a, b Comparison between the curvature of digital (orange) and physical (black) elements
of the suspended gridshell (Baseta and Bollinger 2018)
ital form of every single double-layered lath shows their exact deviation (Fig. 14).
As mentioned before, larger deviations have been noticed at the middle laths of the
primary structure (e–h) and at the laths with two inflection points of the secondary
structure (3–5). Despite the fact that the laths with two inflection points showed dou-
ble signed curvature in their physical form, the magnitude of the resulting curvature
was decreased and mainly induced by the primary structure.
For comparison purposes, the aforementioned gridshell has been structurally ana-
lyzed as it consisted of single solid rectangular laths (h 8.2 mm) with equal bending
stiffness as the flat double-layered notched laths. Considering the same boundary con-
ditions, the deflection of the hypothetical structure resembles, as expected, a catenary
form with the maximum deflection at the middle. Comparing the latter form with the
form of the physical prototype, it is evident that the double-layered notched beams
indicate a different bending behavior than the solid beams (Fig. 15) (Baseta and
Bollinger 2018).
736 E. Baseta
Fig. 15 Colored laths: deformation of a hypothetical gridshell which consists of rectangular single
laths (8.2 × 20 mm) with 5 N point loads at every node, black lines: physical form of the prototype,
grey mesh: predefined digital shape (Baseta and Bollinger 2018)
Fig. 16 Comparison between the curvature of digital (orange) and physical elements of the can-
tilevering gridshell (Baseta and Bollinger 2018)
6.2 Geometrical Evaluation of Cantilevering Gridshell
The comparison between the physical and the digital form of the cantilevering grid-
shell focuses more on the shape of the primary structure. While beam 1 shows a
smaller maximum deflection than the predefined curve, the beams 3, 4 and 5 show
bigger deflections than initially planned (Fig. 16). Examining the radii of curvature of
the primary beams (1–5), it has been noticed that whereas the predefined curves have
constant curvature, the resulting curves have decreasing curvature towards their tip.
This curvature is similar to the curvature induced to a standard cantilevering beam
by a point load at its tip (maximum curvature at the support to 0 curvature at the tip).
For verifying the controlled stiffness of the notched beams the cantilevering grid-
shell has been structurally analyzed as it consisted of single solid laths (h 32 mm)
with equal bending stiffness as the flat double-layered notched laths. Considering the
same boundary conditions, the analysis showed that the deflection of the hypothetical
gridshell is larger than the physical result of the experiment (Fig. 17). This proves that
as the gaps between the shear blocks decrease, the stiffness of the elements increases.
Thus, the element is able to block before it reaches the maximum deformation of its
layers. In order to verify that quantitatively, a stiffness test was conducted. The aim
Fig. 17 Colored laths: deformation of a hypothetical gridshell which consists of rectangular single
laths (32 × 60 mm), with 50 N point loads at every cantilevering tip, black lines: physical form of
the prototype, grey mesh: predefined digital shape (Baseta and Bollinger 2018)
Fig. 18 Deformation of beam 2 by increasing forces
of the experiment was to compare the structural behavior of double-layered beams

with the one of solid beams.
More specifically, the deflection of beam 2 was measured under different loads
(Figs. 18 and 19) in order to obtain its second moment of inertia I, related to its
Fl 3
bending stiffness. Solving the formula of maximum deflection d y 3EI of a can-
tilevered beam with respect to its second moment of inertia I, the latter can be found.
Taking into account that on top of the deformation in the y axis d y there is a consid-
erable deformation in the x axis d x as well, the aforementioned formula translates as
follows:
Fl 3 Fl 3 ll0 −dx F(l0 − dx )3

dy ⇒I ⇒ I , (9)
3E I 3Ed y 3Ed y
where l0 is the initial length of the beam and F the applied force at the tip of the
cantilever.
Figure 19 shows the results from the aforementioned test, where the notched
doubly-layered beam exhibits increasing stiffness as its curvature increases. At small
curvature (F = 14.7 N), the stiffness of the notched beam is close to the stiffness of
the equivalent double-layered beam without shear blocks (Fig. 19b) as expected. As
the beam reaches its predefined deformation (Fig. 19, blue triangle) it has almost
achieved its maximum stiffness. Due to fabrication tolerances, the maximum bending
738 E. Baseta
Fig. 19 Bending stiffness graph with respect to force of the notched beam 2: the triangle represents
the stiffness at the predefined deformation while the circle represents the maximum stiffness (Baseta
et al. 2018)
Fig. 20 Strength test of beam 1. Juxtaposition of the physical prototype and the digital structural
analysis. The colors and the lines indicate the distribution of the principal stresses
stiffness is achieved at slightly larger curvature (Fig. 19, blue circle). However, the
maximum bending stiffness of the beam does not reach 8 times the stiffness of its
single layer I (Fig. 19a) as expected (Eq. 1). Instead, it reaches stiffness 3.6 I (Fig. 19c)
(Baseta et al. 2018). The aforementioned test needs to be repeated in various settings
and with different parameters in order to identify the reason for its underperformance,
and subsequently improve it.
Finally, a strength test was conducted for the beam 1 in order to see the weak points
of the double-layered beam. The first crack has been noticed at the thin part of the top
layer, close to the support when 350 N where exerted at the tip of the cantilevering
beam. Small cracks, parallel to the long axis of the beam, were induced at the bottom
layer at the point that the cross section height jumps from h1 to h2 when bigger force
was exerted. Digital structural analysis of the beam verifies its week points (Fig. 20).
The same strength test was conducted for beam 5 and no breakage was noticed with
F = 530 N. This verifies the differentiation of the maximum stiffness of beams with
different gap lengths.
6.3 General Evaluation and Discussion
The qualitative and quantitative comparison between the physical and the digital form
of the aforementioned prototypes brings valuable insights for the further development
of the presented self-formation process. Both prototypes showed similar deficiencies
which are summarized below.
Firstly, inaccuracies in the resulting curvature of the individual elements have
been noticed. This has been partly evoked by the fabrication tolerances since 1 mm
increase of the gap length results in deviations of several centimeters (e.g. 130 mm
for the elements of the suspended prototype). As a result, further fabrication tests
need to be conducted to achieve the precision of 1/10 of a millimeter. An additional
reason for the curvature inaccuracy was the elasticity of the notched joinery detail.
This led to a reduced, than expected, resulting stiffness of the elements and thus
bigger deflections. Therefore, Finite-Element-Analysis (FEA) for strengthening the
joinery detail should be conducted.
Secondly, the difficulty of sliding between the primary and the secondary structure
led to a constrained deformation of the system. The increased friction at the nodes of
the gridshell was caused by the transversal torsion induced during the deformation.
This is particularly visible in the plan view of the suspended structure, where the
elements do not remain planar. Thus, further tests with industrial or custom joinery
details of the nodes should be conducted.
7 Conclusion
In conclusion, by inserting variable stiffness zones among thin layers of elastic mate-
rial, controlled deformation can be achieved by load differentiation. It has been
proven, through physical experiments, that the presented double-layered beams with
embedded shear blocks increase their stiffness as their deflection increases. Thus,
they can be used for passive self-formation of gridshell structures. When the self-
formation is reversible, the presented system can find applications in transformable
structures. In addition, the joinery details between the double-layered beams act like
an embedded construction manual. As a result, digitally designed double curved sur-
faces can be easily and rapidly constructed. Specific applications which exploit the
full potential of the system are yet to be defined. Regarding the material used, this
research focuses on wood, since it is a natural fibrous composite with high flexibility
and strength which can be easily machined. In addition, the ability to fabricate the
variable stiffness elements with industrial CNC machines, relatively cost-effectively,
shows great potential in scaling-up the system. However, further physical experi-
ments in larger scale need to be conducted to prove its structural performance. In
addition, the fatigue after multiple cycles of transformations is a topic that should
be investigated. Alternatively, materials with high strength to stiffness ratio can be
used, such as fiber reinforced polymers (FRPs) and carbon fibers. The two examined
740 E. Baseta
experimental prototypes proved the qualitative correspondence between the prede-

fined digital and the resulting physical form. Nevertheless, quantitative tolerances
need to be explored. Therefore, additional fabrication tests, as well as FEA of the
joinery detail and digital simulations of the performance of the elements are currently
under development.
Acknowledgements This project has received funding from the European Union’s Horizon 2020
research and innovation program under the Marie Sklodowska-Curie grant agreement No 642877.
The fabrication and installation of the cantilevered prototype would not have been possible without
the support from the Blumer Lehmann AG (Gossau, Switzerland) team and its leaders Kai Strehlke
and Martin Antemann.
The design, fabrication and installation of the suspended prototype was part of the course ‘Digital
Design and Full Scale Fabrication 17’ in the University of Applied Arts Vienna-Institute of Archi-
tecture led by Andrei Gheorge. Philipp Hornung representing the Angewandte Robotic Lab and the
Wood technology laboratory led the robotic fabrication. Students of the course: Adrian Herk, Afshin
Koupaei, Aleksandra Belitskaja, Alex Ahmad, Alexandra Moisi, Andrej Strieženec, Anna Tuzova,
Ben James, Charlotte Krause, David Rüßkamp, Jan Kováříček, Jelinek Johanna, Jonghoon Kim,
Julian Heinen, Kaspar Ehrhardt, Leonie Eitzenberger, Ludmila Janigova, Madeleine Malle, Michael
Tingen, Minho Hong, Polina Korochkova, Rudolf Neumerkel, Sadi Özdemir, Shaun McCallum,
Toms Kampars, Zarina Belousova.
References
Alexander C (1964) Notes on the synthesis of form. Harvard University Press, Cambridge, MA
Alquist S (2015) Social sensory architectures: articulating textile hybrid structures for multi-sensory
responsiveness and collaborative play. In: Proceedings of computational ecologies: design in the
anthropocene—ACADIA 2015, Cincinnati, USA
Barozzi M, Lienhard J, Zanelli A, Monticelli C (2016) The sustainability of adaptive envelopes:
developments of kinetic architecture. In: International symposium on novel structural skins:
improving sustainability and efficiency through new structural textile materials and designs. Proc
Eng 155:275–284
Baseta E, Bollinger K (2018) Construction system for reversible self-formation of gridshells: cor-
respondence between physical and digital form. In: Proceedings of recalibration on imprecision
and infidelity—ACADIA 2018, Mexico City, Mexico, pp 366–375
Baseta E, Preisinger C, Antemann M, Strehlke K, Bollinger K (2018) Geometry-induced variable
stiffness structures. In: Proceedings of the IASS symposium 2018 creativity in structural design,
MIT, Boston, USA
Burgert I, Fratzl P (2009) Actuation systems in plants as prototypes for bioinspired devices. Philos
Trans R Soc 367(1893):1541–1557
Burry J, Felicetti P, Tang J, Burry M, Xie M (2005) Dynamical structural modeling: a collaborative
design exploration. Int J Arch Comput 3(1):27–42
Castells M (1992) The informational city: economic restructuring and urban development. Wiley-
Blackwell, Cambridge, MA
Correa D, Papadopoulou A, Guberan C, Jhaveri N, Reichert S, Menges A, Tibbits S (2015) 3D-
printed wood: programming hygroscopic material transformations. J 3D Print Addit Manuf
2(3):106–116
Farahi B (2016) Caress of the gaze: a gaze actuated 3D printed body architecture. In: Proceedings
of posthuman frontiers—ACADIA 2016, Michigan, pp 352–361
Farahi B, Leach N, Huang A, Fox M (2013) Alloplastic architecture: the design of an interac-
tive tensegrity structure. In: Proceedings of adaptive architecture—ACADIA 2013, Cambridge,
Ontario, Canada, pp 129–136
Frazer J (1995) An evolutionary architecture. Architectural Association, London
Galloway KC, Clark JE, Koditschek DE (2013) Variable stiffness legs for robust, efficient, and
stable dynamic running. J Mech Robot 5(1):011009
Gengnagel C, Alpermann H, Lafuente E (2013) Active bending in hybrid structures. In: For-
m—rule|Rule—form 2013. Innsbruck University Press, Innsbruck
Guiducci L, Razghandi K, Bertinetti L, Turcaud S, Ruggeberg M, Weaver JC, Fratzl P, Burgert I,
Dunlop JWC (2016) Honeycomb actuators inspired by the unfolding of ice plant seed capsules.
PLoS ONE 11(11)
Holstov A, Morris P, Farmer G, Bridgens B (2015) Towards sustainable adaptive building skins
with embedded hygromorphic responsiveness. In: Proceedings of building envelope design and
technology—advanced building skins, Graz, Austria, pp 57–67
Huerta S (2006) Structural design in the work of Gaudi. Arch Sci Rev 49(4):324–339
Iamsaard S, Villemin E, Lancia F, Aβhof SJ, Fletcher SP, Katsonis N (2016) Preparation of
biomimetic photoresponsive polymer springs. Nat Protoc 11:1788–1797
Jayathissa P, Caranovic S, Begle M, Svetozarevic B, Hofer J, Nagy Z, Schlueter A (2016) Structural
and architectural integration of adaptive photovoltaic modules. In: Proceedings of the advanced
building skins, Bern, Switzerland, p C6-3
Jenkins PP, Landis GA (1995) A rotating arm using shape-memory alloy. In: 9th aerospace mech-
anisms symposium. NASA Johnson Space Center, USA, pp 167–171
Jung W, Kim W, Kim HY (2014) Self-burial mechanics of hygroscopically responsive awns. Integr
Comp Biol 54(6):1034–1042
Knippers J, Nickel K, Speck T (2016) Biomimetic research for architecture and building construc-
tion: biological design and integrative structures. Springer, Switzerland
Körner A, Madar A, Saffarian S, Knippers J (2016) Bio-inspired kinetic curved-line folding for
architectural applications. In: Proceedings of posthuman frontiers: data, designers, and cognitive
machines—ACADIA 2016, Ann Arbor, pp 270–279
Kotelnikova-Weiler N, Douthe C, Lafuente E, Baverel O, Gengnagel C, Caron J (2013) Materials
for actively-bent structures. Int J Space Struct 28(3–4):229–240
Kuder II, Fasel U, Ermanni P, Arrieta AF (2016) Concurrent design of a morphing aerofoil with
variable stiffness bi-stable laminates. Smart Mater Struct 25(11):115001
Lan X, Zhang R, Liu Y, Leng J (2011) Fiber reinforced shape-memory polymer com-
posite and its application in deployable hinge in space. In: Proceedings of the 52nd
AIAA/ASME/ASCE/AHS/ASC structures, structural dynamics and materials conference, Den-
ver, Colorado
Lendlein A, Langer R (2002) Biodegradable, elastic shape-memory polymers for potential biomed-
ical applications. Science 296(5573):1673–1676
Liddell I (2015) Frei Otto and the development of gridshells. Case Stud Struct Eng 4:39–49
Lienhard J, Knippers J (2013) Considerations on the scaling of bending-active structures. Int J Space
Struct 28(3–4):137–148
Lienhard J, Knippers J (2015) Bending-active textile hybrids. J Int Assoc Shell Spat Struct 56:37–48
Lienhard J, Alpermann H, Gengnagel C, Knippers J (2013a) Active bending, a review on structures
where bending is used as a self-formation process. Int J Space Struct 28(3–4):187–196
Lienhard J, Alquist S, Menges A, Knippers J (2013b) Extending the functional and formal vocab-
ulary of tensile membrane structures through the interaction with bending-active elements. Re]
Thinking lightweight structures, Proceedings of Tensinet symposium, Istanbul
Raviv D, Zhao W, McKnelly C, Papadopoulou A, Kadambi A, Shi B, Hirsch S, Dikovsky D, Zyracki
M, Olguin C, Raskar R, Tibbits S (2014) Active printed materials for complex self-evolving
deformations. Sci Rep 4(7422)
742 E. Baseta

60:50–69
Rüggeberg M, Burgert I (2015) Bio-Inspired wooden actuators for large scale applications. PLoS
ONE 10(4):1–16
Senatore G, Duffour P, Winslow P, Wise C (2017) Shape control and whole-life energy assessment
of an ‘infinitely stiff’ prototype adaptive structure. Smart Mater Struct 27(1):015022
Sparrman B, Matthews C, Kernizan S, Chadwick A, Thomas N, Laucks J, Tibbits S (2017) Large-
scale lightweight transformable structures. In: Proceedings of 37th annual conference of the
association for computer aided design in architecture: disciplines + disruption, USA, pp 572–581
Sydney Gladman A, Matsumoto EA, Nuzzo RG, Mahadevan L, Lewis JA (2016) Biomimetic 4D
printing. Nat Mater 15:413–418
Thompson D (1961) On growth and form. Cambridge University Press, Cambridge
Weinstock M (2004) Morphogenesis and the mathematics of emergence. In: Hensel M, Menges
A (eds) Emergence: morphogenetic design strategies, vol 74, No. 3, Architectural design. Wiley
Academy, London
Wood DM, Correa D, Krieg OD, Menges A (2016) Material computation-4D timber construction:
towards building-scale hygroscopic actuated, self-constructing timber surfaces. J Arch Comput
14(1):49–62
Yao L, Ou J, Cheng C, Steiner H, Wang W, Wang G, Ishii H (2015) BioLogic: natto cells as
nanoactuators for shape changing interfaces. In: Proceedings of 33rd annual ACM conference on
human factors in computing systems, Seoul, pp 1–10
Natural Complexity. An Introduction
to Structural Design with Tree Forks
Lukas Allner and Daniela Kroehnert
Abstract Amongst construction materials wood offers a unique potential for deriv-
ing complex spatial geometries from its grown fibrous structure. The inherent mate-
rial properties such as grain direction, can be used as generative parameters for
innovative timber structures. With the possibilities of 3D-scanning and Computer
Tomography the anatomy of wood can be exploited as a design driver for spatial struc-
tures. This chapter introduces a design concept of utilizing natural forked branches
as components in structural frameworks.
Keywords 3D scan · Naturally grown form · Complex wood structures

Spatial frameworks · Discrete element aggregation
1 Introduction
Trees naturally form intricate configurations of roots, branches, leaves, etc. that
correspond to multiple forces and conditions. As a result, wood is a dynamic inho-
mogeneous material with varying densities, directional grain and a complex geom-
etry. Usually trunks are cut into rectilinear elements and the more irregular parts
of the tree are chipped and processed further into timber products. The resulting
standardized components have the highest possible homogeneous quality in order to
assemble structures according to the logic of repetition of identical parts, mainly in
the form of grids and other regular organizational principles. However “the resulting
This Article is based on a paper presented at the IASS Conference 2018 in Boston: L. Allner, D.
Kroehnert (2018) Conceptual Joining: Branch Formations, Proceedings of IASS 2018: Creativity
in Structural Design.
L. Allner (B) · D. Kroehnert (B)

University of Applied Arts Vienna, Vienna, Austria
e-mail: lukas.allner@uni-ak.ac.at
D. Kroehnert
e-mail: daniela.kroehnert@uni-ak.ac.at

744 L. Allner and D. Kroehnert
timber-products are showing substantial deficiencies in comparison to the undam-

aged biological system ‘Tree’” (Müller et al. 2014). This practice of homogenizations
is nevertheless today’s standard, since the components can be optimally manufac-
tured and controlled by industrial machines and because it corresponds to a design
method in which a spatial scheme is first developed with a virtual (digital) abstract
geometry and then adapted to a building material.
Advanced technologies and methods are generally driven by digital abstraction.
The physical translation commonly aims to be as close as possible to the mathematical
pure geometry as an ultimately virtual “ideal”. Complex geometries and organic
freeform shapes, e.g. beams of the Centre Pompidou Metz (Ban 2010), are often
realized as elements that are CNC-cut from solid blocks with a wasteful material
utilization.
We believe wood offers a potential for an aesthetic of complexity by its heteroge-
neous material properties and its grown form, but also for novel structural models.
With the possibilities of 3D-scanning and Computer Tomography, the anatomy of
wood can be activated and exploited as a design driver for spatial structures closely
related to the material logic. In fact, we intend to develop conceptually organic struc-
tures that are directly derived from the natural properties of the grown material and
which, despite forming artificial compositions, are nevertheless closer to reality than
to virtuality.
This article introduces ‘Branch Formations’ an investigation in the context of the
research project “Conceptual Joining” at the University of Applied Arts in Vienna,
funded by the Austrian Science Fund (FWF). The aim of this project is to explore the
relationship between material and space and more specifically with an intention to
translate the grown complexity of wood into architectural experience (Fig. 1). Branch
Formations investigates the use of grown forked branches as structural components
in spatial frameworks to take advantage of these naturally optimized nodes.
A reference is the “Wood Chip Barn” project of the Architectural Association
from 2016 (Self and Vercruysse 2016) where natural crotches of beech were used to
compose the primary structure of a roof. In “Branch Formations” the potentials of
grown form are exploited independent from a specific structural function. With this
agenda this project navigates in between the uniqueness of a spatial configuration
affected by the singular characteristic of each branch element and the similarities
within different types of trees. The use of forked branches as naturally optimized
structural nodes allows us to design filigree structures with minimal cross-sections.
This advantage of maximum structural performance at a minimum weight has already
been employed for tools in pre-industrial times (Spindler 1995). Some trees were
even cultivated with a designed geometry (Agote Aiziprua 2009).
Today with computer aided technologies we have the capability to look at wood
differently, to liberate its nature and invent solutions beyond industrial standards.
Natural Complexity. An Introduction to Structural … 745
Fig. 1 Concept
2 Aims
The research presented in this chapter attempts to understand the complex and irregu-
lar parts of trees as a potential. The wood industry considers these as disadvantageous
because branch nodes disrupt the desired homogeneity and structural performance
of linear timber profiles. With an alternative concept of following the inherent logic
of wood elements and utilizing them accordingly we will be able to activate a hidden
potential and eventually transform the presumed disadvantage into a benefit. This
would increase the value of non-standard tree parts—they no longer serve only as
a source material for chip- and fiber-boards or fuel—and activate their potential for
construction. Our research represents a quest to develop rulesets for assembly as well
as design methods for the use of these parts in structural architectural schemes. The
present text describes an ongoing research and is partially a report of undertaken
surveys but it also includes speculations on possible outcomes and future investiga-
tions.
The dialogue and tension between control (design intention) and coincidence
(natural singularity) is to be negotiated constantly.
3 Digital Nature
3D Scanning Technology and computational tools allow us to cope with the com-
plexity of naturally grown unique elements such as branch nodes and consider a
“non-standard approach” (Schindler et al. 2013) to structural design.
3.1 Data Capturing
As an approach to designing structures using natural tree parts we choose a method

of simplification in a digital model. The geometrical complexity of the individual
pieces is reduced to a minimum maintaining the basic topological principles in an
axis model which serves as reference for parameterization. The basis of this method
of working is a digitized catalogue of 3D scanned branch nodes. Tools such as
Microsoft™ Kinect, Artec Eva and photogrammetric software capture geometry as
textured 3D envelopes. For the studies presented here this method is providing the
basic data in the form of 3D mesh geometry and a sufficient amount of information to
approximate the digitized elements’ structural performance. Computer Tomographic
Technology (e.g. Microtec CT Log) additionally registers internal structures, offering
more advanced data sets for future investigations.
Evaluating branch forms in various tree species suggests the use of hard wood
branch nodes as architectural and structural components. Deciduous trees have
branches of larger diameters in relation to the trunk than those of conifers and they
form crowns which are usually not yet considered for construction. Crown parts rep-
resent a considerable proportion of the overall tree and because of the slower growth
rate of hard wood concepts for activating this unused potential are most relevant.
Additionally hard wood will become increasingly relevant to the wood industry in
middle Europe due to climate change, that requires the consideration of wood species
other than those that have prevailed so far. The most commonly used spruce is more
sensitive to increasingly warm weather with longer dry periods than other trees and
stocks are expected to shrink in the coming decades. Various types of deciduous trees
on the other hand are more robust and thus represent a valid and perhaps necessary
alternative (Schüler et al. 2013). Various studies suggest that in the future, trees of
that category such as beech will become ever more important for the timber indus-
try in Europe (Hanewinkel 2010). Utilizing complex tree parts for structural design
additionally offers the potential to consider a greater variety of wood types; species
that offer good structural material properties but have been so far disregarded due to
their complex shape, including trees with shorter trunks or shrubs.
For the present studies we are working with branches of Hornbeam (Carpinus
betulus) which is characterised by a high rigidity and a tendency to extensive branch-
ing. Therefore it was the favoured tree type used already by the Romans to grow
natural ramparts (parts of the Limes). In middle Europe it was used in combination
with beech and oak until the late 1800 for growing “Landwehr” (Lüster 1913). The
wood of this tree has also been used for tool grips and mechanical parts in preindus-
trial times but does not play a relevant role in contemporary industry mainly due to
its complex shape and the relatively small diameter of its trunk (Grabner 2017).
In an empirical analysis we observed in a sample set of 154 Y-branches an average
crotch angle of 42.7°, ranging from a minimum of 16.8° to a maximum of 62.2°, of
which 87% are within a range of 30°–55° (Fig. 2). The set contains elements that are
unequal but very similar. The same range of angles occurs across all generations and
Fig. 2 Crotch angle range of hornbeam
Fig. 3 Comparison: oak–hornbeam
scales of branches within the sample tree crown which allows for tests in models of
small scale to approximate geometric properties of full scale configurations.
In other tree types such as Oak the average crotch angle as well as minima and
maxima are different, resulting in a diverging crown architecture (Fig. 3). There is
very little research or scientific data on typical crotch angles and statistical distribu-
tion for various tree species. We assume, that angles are varying only in a limited
range within one species under similar growth conditions. Across species on the
other hand there is a large variety of angles which will affect the geometrical and
structural configurations formed by the natural branch nodes.
3.2 Abstraction Principles and Control Geometry
A given set of forked branches is 3D scanned and the digitized data simplified into
approximated axis models. Each branch can be represented by 4 points and 3 lines.
Fig. 4 From 3D scan to axis model (version a)
Fig. 5 Additional abstraction principles (version b + c)
The node-element represented in modules with three line segments is used as a

‘Y’—typology (Fig. 4).
The axis models are used as components in a parametric model. They also serve
as reference geometry to eventually reposition the original scan models into the
assembled configuration and thus allow for designing specific joints and modification
of the physical parts further along in the process of our research.
There are different possible approximations/translations of branch nodes into Y-
models (Fig. 5):
(a) Nodes with variable branch lengths. Each branch is approximated as one axis
line.
(b) Nodes with branches in varying length. End segments are approximated as an
axis lines, interpolated within branch node.
(c) An inscribed circle around the center point trims branches to same lengths.
Branches are represented as approximated axis lines.
The axes in version (c) are nearly planar, whereas version (a) and (b) mostly result
in 3-dimensional models. All of these approximations result in similar simplified
digital typology and are conceptually interchangeable in the digital assembly model.
But the difference in their relationship to the respective physical origin offers different
options of a physical translation of the digital assembly in the subsequent steps.
Fig. 6 Branch-forks connection principle
4 Aggregation Principles
To stay close to the principle of the tree with its layout of continuous (bifurcating)
fibers we set up an assembly concept in which branch nodes are joined in extension
of the wood grain to form a potentially endless continuum (Fig. 6). We aim to develop
spatial frameworks for architectural purpose made from wooden parts exclusively,
but without being yet tied to a specific use. To allow for a maximum flexibility and
range of possible applications we conceptualize configurations of Y-elements with
their open ends joined to form closed cells to provide general structural integrity and
bracing. This offers possibilities for multidirectional orientation and load bearing
capacity by maintaining the objective of fiber and grain continuity. The emerging
configurations assemble the Y-elements into bifurcating space loops. In a first series
of studies we examine possible geometrical formations and methods of control and
generative logic.
4.1 Discrete Element Aggregation
Based on findings from intuitive physical and digital models (Fig. 7) we decided to
distinguish three basic principles for aggregations: Repetitive patterns (a), Iterative
aggregation/self organisation (b) and Population of a designed structure
(a) Repetitive patterns:
Offering the highest level of control and predictability this principle was our starting
point and is mainly in focus of our investigation up until this point in the project.
A computational process is set up for aggregating repetitive and regular structures
based on mathematical-geometrical principles. Depending on the defined aggregation
rule set 2 and 3-dimensional cellular patterns can be formed using the branches
as edges. From all possible cell geometries a Tetrahedron is the most compact 3
Fig. 7 Intuitive studies
Fig. 8 Tetrahedron module from Y-elements
dimensional form. The tetrahedron is the platonic solid with the least amount of
faces. Thus it also represents the most compact spatial cell with the highest stiffness
formed by y-shaped branch nodes (Fig. 8). As a result aggregations of tetrahedrons
(Fig. 9) are the most stable regular spatial frameworks for the chosen connection
concept of fiber continuity.
Rotation and length relating to the branch axis represent sufficient levels of free-
dom necessary to combine elements of any angle into a Tetrahedron cell when con-
sidering a set with crotch angles ranging from 17° to 62°.
Forming platonic solids or other 3 dimensional cells with more faces and edges
would allow for a larger flexibility of possible combinations (Fig. 10). But the
increased flexibility also results in forms with lower stiffness. In a different set with
a larger angle range combinations of extremely differently angled elements would
eventually fail to form a continuous framework.
With the Wasp Plugin (Rossi 2017) for the 3D modeling environment
Rhinoceros/Grasshopper (McNeel, Rutten), regular frameworks of tetrahedrons can
be achieved by duplicating and aggregating one Y-branch element applying the
respective ruleset. Current design practice in working with discrete element aggre-
gation is mainly working with a limited amount of different digital entities within the
aggregation processes. In this project however we are confronted with a large amount
of unequal but similar objects. To cope with this challenge we introduce an operation
Fig. 9 Tetrahedron framework
Fig. 10 Tetrahedron and octahedron framework
in which each instance of the identical duplicates is randomly replaced by one (sim-
plified) representation of a “real” specimen from the digitized catalogue, each with
a different crotch/reflex angle. The resulting discontinuous pattern is modified with
the employment of a relaxation algorithm using Kangaroo for Grasshopper (Piker
2017). The algorithmic definition is programmed with constrained angles within
each Y-element, variable length/distance between modules and a goal set to align
adjacent branch axis. The simulation results in a slightly modified irregular pattern
of identical topology but with adjusted edge-lengths and an angle variance stemming
from the scanned catalogue (Fig. 11).
In a first step a random positioning of elements is used as means of handling
formations composed of sets of unequal elements. It is one instance of a design
setup in a process of negotiating levels of control and design intention on one hand
and coincidence introduced by the variance of natural elements on the other hand.
Fig. 11 Tetrahedron framework: regular–mismatching–optimized
Fig. 12 Tetrahedron frameworks from different sets of angles
In subsequent steps we aim at introducing means of control according to local and

global design goals and conditions.
Sets with varying ranges of angles representing branches from different tree
species would result in very different appearances (Fig. 12) and structural behavior
(when disregarding material specificity). This shows a direct relationship of specific
typical growth form related to a species and the resulting spatial formation. In these
structures the distinct material property of wood represented in geometrical form
becomes spatial and thus architectural experience. The palette of perception of wood
structures is expanded beyond the sensing of visual and haptical quality of architec-
tural surfaces. This spatial material specificity opens a spectrum of possibilities for
non standard forms and building typologies directly related to the specific material
used to build them. The close relationship of material and form not only allows for
spatial perception but possibly offers models for an optimal structural use of a certain
type of wood.
The material does not have to fit into a preconfigured structural (standardized)
system but instead can be used to its full potential. The assembly of structural com-
ponents is partially derived from the material geometry and structural performance
in a process of form finding that aims at exploring structures that are natural to the
material.
The aggregations of irregular elements in a regular topology allows for a first
grasping of the essence of branch formations. These also provide the basis for struc-
tural analysis models and speculations of usability. Pattern-based assemblies are
strong structuring devices for spatial organizations. The pattern principles imply an
order as a guiding logic for a possible architectural scheme. The dominant role of
the pattern in the design process ensures a specific architecture closely related to the
logic of the branch formation but the strict order also comes with some limitations.
Fig. 13 Study model: iterative assembly
(b) Iterative aggregation/self organisation:
In a procedural assembly (Fig. 13) parts are put together in iterative steps. Instead
of a global logic controlling the formation local parameters define the aggregation
in every step. There is no fixed or predefined framework that is put together instead
the form of the aggregation is found in the process of making.
Self organized structures can be adapted to local conditions by maintaining an
overall systematic coherence dependent of the respective ruleset. The resulting irreg-
ular structures represent individual solutions to actual problems and conditions as
unique formations, a quality very similar to the structure of a tree. Its growth is
dependent on the characteristics of its species but it finds a unique form in relation
to its individual growing conditions.
Aggregations based on self-organizational principles with their inherent irregu-
larity are more difficult to control and put an additional level of complexity to branch
formations. But those offer various possibilities for the addressing of specific prob-
lems, whereas the regular topology of principle (a) is more restrictive and fixed to
its pattern logic. Future investigations in this project will address the development
of concepts and algorithms to design and control iterative aggregations.
(c) Population of a designed structure:
The nodes of a predefined framework are being replaced by branch elements with
crotch or reflex angles matching the respective angles between axes of structural
members. Following principle (c) a conventionally designed structural framework
composed of generic axes could be modified by replacing structural nodes by forked
branches with matching or similar angles. Altered geometrical configurations and
the related effect on structural behavior can be analyzed. This concept is currently
not in focus.
5 Joining
Considering those formations as physical structures the joining of elements becomes

a key problem to address.
In a complex framework typically linear members are connected at the axis nodes.
Especially with a high number of members joined, details at the nodes require a high
level of precision and technical as well as manufacturing efforts. In such frameworks
standardized (simple) members are connected with complex joints.
This project is working with an inverse situation: the non-standard (complex)
members represented by branch nodes are connected with simple joints. With the
chosen concept of grain continuity all joints are instances of extension joints. The
principle of the Y-shaped element offers bifurcating directions which allows for a
high level of spatial complexity with only one type of joint.
Structural members are joined at the least complex location within the axis model
and only two elements are joined in each location. This way the complexity and
specificity of the element is compensated by the geometric simplicity of the joints.
The complexity of the non-standard members, which is a product of the tree’s growth
pattern as a representation of the material property of wood, reflects in spatial com-
plexity and variation.
5.1 Joining Principles
To translate the digital aggregates into structures assembled by physical elements we

consider two possible concepts:
(a) branch elements joined with additional bridging elements (see Fig. 14, V1 &
V3)
Adding elements of variable length in between branch nodes allows for flexible
aggregations. In each joint two degrees of freedom apply for the combination of
elements along the same axis: axial rotation and translation provided by the extension
element. Although allowing almost any combination of crotch angles, the insertion
of an intermediate bridging element weakens the system structurally by introducing
additional joints.
(b) branch elements directly joined (see Fig. 14, V2)
This concept requires a precise selecting and matching of the available specimen.
With the limitation of branch lengths also a translation along the connection axis
as a degree of freedom is reduced. Although geometric limitations apply, this join-
ing principle provides maximum stability of a structural system due to a minimum
number of joints. The goal in this project is to achieve formations assembled with
elements joined without extensions.
Fig. 14 Tetrahedron assembly variations in relation to abstraction models (see Sect. 3.2)
5.2 Joints
The joining of branch nodes in the context of this project needs to address the fol-
lowing problems:
Stability:
Within the complex frameworks with differently oriented structural members and
loads acting in various directions, the joints need to have the capacity to compensate
a set of different forces. Possibly a variation of joints in some instances is needed.
Tolerance:
Given the spatial complexity in the system some tolerance is required to be able to
join branch elements into closed cells without applying tension.
Degrees of freedom for assembly:
Combining elements to close a loop requires sufficient degrees of freedom regard-
ing the direction in which parts are put together. Possibly different joint types for
“standard” and “closing” joints are needed.
With the intention of developing joints without metal fasteners or glue we chose a
lap joint as a basis for a study (Fig. 15). Subsequently the potential of more complex
joint types such as the Japanese "Kanawa Tsugi" will be investigated. This joint is
typically used in vertical orientation to extend beams that receive bending loads. It
has the benefit of being also relatively stable under application of other loads. This
way it is possible to solve all instances with one kind of joint.
Other joint types and techniques such as other traditional extension joints, steel-
and 3D printed joints are studied additionally. Especially current developments in 3D
Fig. 15 Joint study
printing technology support speculations on a joining technique that would literally

extend wood fibers of the connected elements. Perhaps concepts can be developed
for interpolating grain patterns by 3D printed strands.
6 Structural Evaluation
One of the key interests in Branch Formations is the exploration of the spatial and
structural qualities inherent in wood as a dynamic material. In this context we are
approaching structural design not as a means of solving a given problem but instead
we are working with an experimental reversed method: The structural purpose and
usability of spatial frameworks are found in an investigative process and developed
on the basis of geometric and structural qualities provided by the formations that are
defined by principles derived from the grown logic of the material. Various aggrega-
tions of branch node elements are tested with regard to ideal structural performance
under influence of different load cases. Potential functions are assigned accordingly
and the structural system adapted in a feedback loop.
At the current state of the project mainly instances of pattern based aggregations
(see also paragraph 4.1.a) are used as structural samples. A bounding box is populated
with a theoretically endless framework of elements and evaluated in a structural
simulation with Karamba3d for Grasshopper (Preisinger 2013). Differently oriented
aggregations of the same pattern are tested, each with gravitational loads and supports
set at the bottom ends show the maximum deformation in each instance within the
same load scenario (Fig. 16). The alternative aggregation samples are evaluated with
the goal to find an optimized orientation of the structural members. The open ends
at the borders of the bounding volume appear as the weakest areas. Within the grid,
in areas where all elements are interlinked the system shows a much higher stability.
Fig. 16 Structural analysis: deformation of different orientations of tetrahedron grid
Further studies will investigate to which extent and in which cases irregular sys-
tems are structurally superior to regular systems. In a feedback with the development
of joints a more in depth analysis of the distributions of forces and moments across
the structural model will eventually inform an optimized positioning of joints along
the axes in which branch ends are connected. Complementary to a global optimiza-
tion a structural analysis and simulation could also influence the morphogenesis of
the joints themselves. Joining techniques are studied in correspondence with struc-
tural behavior and performance. Besides the analysis of the aggregations as abstract
form we additionally conduct an investigation in case studies. With this method the
potential of branch aggregations as structural systems are explored in architectural
scenarios that relate to a projected usability. The integration of a new geometry is
also a first step towards a speculation of scale and use in an architectural sense.
In a first instance the aggregation showing the lowest deformation in the initial
testing is used for a case study of a simple vertical architectural scheme (Fig. 17).
In this scenario the aggregated branch elements serve as a structure supporting hor-
izontal planes. Structural simulation shows that the introduction of floors stabilizes
the system and partially solves the problem of open ends at the periphery. A cross-
section optimization algorithm distributes a preset range of variable widths for each
member across the entire model. This information will be used in a next step to
position the specimen from the scanned set with matching cross-section according
to local structural utilization to form a model with an ideal global structural perfor-
mance. The introduction of irregular parts would also transform the regular grid into
an irregular model. Different to the previous studies in which a random combination
of elements resulted in a transformation of the regular grid (Sect. 4.1, Figs. 11 and
12) in this study specific local (structural) parameters define the distribution of the
varying branch elements and result in an optimized irregularity (Fig. 18).
7 Speculation of Use
In a similar way as we are working on developing structural models we also intend

to find adequate use and architectural function related to the spatial properties and
features in the process. Usability and program are not initially set out as a goal to be
achieved. With this agenda we start with the exploration of the potential of spatial
Fig. 17 Regular tetrahedron-tower load scenarios (gravity and dead weight)
Fig. 18 Tower case study, close up view
configurations and develop a possible functionality in a process of negotiating geom-

etry and use. We aim at finding out more about the relationship between architecture
and the material wood but also at finding novel modes of experience, usability and
architectural typology, a new materialist wood architecture across multiple scales
from furniture to high-rise building.
8 Conclusion and Outlook
A key quality and challenge of the proposed method is the tension between control and
coincidence. Further research will investigate structural performance and potential
function of various aggregations as well as methods of structural optimization. With
the aim of arriving at full scale prototypes there is an ongoing development of joint
details for optimal load transmission.
The researchers will further explore the relationship of the angle spectrums occur-
ring in various tree species and geometric and structural possibilities and limitations.
“Branch Formations” represents the exploration of a distinct conceptual agenda,
but could be seen as one case study in the broader context of a new understanding
of architectural and structural design using wood in its naturally grown form. With
constantly improving computation and technology paired with a growing awareness
for renewable resources wood does not need to be processed into a homogeneous
material but can be addressed and utilized in its full complexity.
Acknowledgements This work is part of the PEEK research project Conceptual Joining (AR 395-
G24) funded by the Austrian Science Fund (FWF). Christoph Kaltenbrunner and our fellow team
members contributed with expertise and feedback. Philipp Reinsberg developed the initial joint
study. Matthew Tam of Bollinger + Grohmann engineers contributed with an initial Karamba defi-
nition. Sandra Tuider provided the branch samples. The Department of Geometry of the Angewandte
made available an Artec EVA 3D Scanner.
References
Agote Aiziprua X (2009) ct timber suppliers for shipyards. In: Bertain#23 ‘from tree to ship’, p 109
Departamento de Cultura y Euskera
Ban S (Architects) (2010) Centre Pompidou. Metz, France
Grabner M (2017) WerkHolz. Kessel, Remangen-Oberwinter
Hanewinkel M (2010) Baumarteneignung Fichte und Buche bei Klimawandel. http://www.
waldwissen.net
Lüster G (1913) Das Rheingauer Gebück. In: Naturdenkmäler in Nassau 2. Bechtold, Wiesbaden
Müller U, Gindl-Altmutter W, Konnerth J, Kaserer D, Keckes J (2014) Stamm-Astanbindung – eine
biologisch optimierte Struktur mit hoher mechanischer Leistungsfähigkeit. 10. GraHFT’14
Piker D (2017) www.food4rhino.com/app/kangaroo-physics
Preisinger C (2013) Linking structure and parametric geometry. Arch Des 83:110–113. https://doi.
org/10.1002/ad.1564
Rossi A (2017) www.food4rhino.com/app/wasp
Schindler C, Tamke M, Tabatabai A, Bereuter M (2013) Serial branches. In: Proceedings of eCAADe
31-computation and performance, vol 1: Material studies
Schüler S, Grabner M, Karanitsch-Ackerl S, Fluch S, Jandl R, Geburek T, Konrad H (2013) Fichte
- fit für den Klimawandel? BFW-Praxisinformation 31:10–12
Self M, Vercruysse E (2016) Wood chip barn. Design + Make. The Architectural Association London
Spindler K (1995) cf wooden handle of the cooper axe used by the neolithic hunter Otzi 5.000 years
ago. In: Der Mann aus dem Eis, Goldmann
Evaluation of Wooden Structures
Gülru Koca
Abstract In order to preserve the architectural heritage and sustainability of cities,

the accurate evaluation of the mechanical properties of existing buildings is crucial.
While inorganic building materials such as natural stones can be evaluated more eas-
ily, it is difficult to accurately assess the mechanical properties of wood. Mistaken
evaluations of structural wooden members may lead to large-scale replacements in
the maintenance and restoration of buildings. The techniques used in the evalua-
tion of wood are; destructive, semi-destructive and non-destructive tests. Although
destructive tests give accurate information about the mechanical properties of wood,
they are not preferred in the evaluation of the existing structures because they cause
the loss of structural integrity. The semi-destructive and non-destructive methods are
being widely used for the last decades in the evaluation of structural wooden mem-
bers. As these techniques do not give harm to the structural members, they allow the
in situ evaluation of wooden structures. While semi-destructive tests are carried out
with the extraction of a small piece without influencing the mechanical properties of
wood, non-destructive techniques are carried out with the help of small devices in
order to detect the interior defect and deteriorations. In this study, it is aimed to give
information about some of the most used semi-destructive and non-destructive test
methods.
Keywords Wooden structures · Sustainability · Ecology

Non-destructive techniques · Semi-destructive techniques
1 Introduction
Timber, which is the only renewable source on earth, has been one of the most used
materials in construction from prehistoric times to present. It has been used in building
structures, from simple shelters produced by trial and error methods to computer-
G. Koca (B)
Department of Interior Architecture, Faculty of Fine Arts, Işık University, Istanbul, Turkey
e-mail: gulru.koca@isikun.edu.tr

762 G. Koca
aided engineering design spatial constructions. The most important reasons in the
preference of timber are; it can easily be obtained, it has good strength to weight
ratio, it is renewable material well appreciated by the customers and it has fine ther-
mal insulation properties. The use of timber has never disappeared even after a slight
decline following the Industrial Revolution, when modern building materials have
emerged. Following the Industrial Revolution, which took place in the late 18th cen-
tury, timber framed building systems were replaced by reinforced concrete and steel
building systems. The use of these materials with higher strength properties makes
it possible to produce high structures. However, when they are evaluated according
to their ecological properties, reinforced concrete and steel are more harmful com-
pared to timber. The embodied energy and carbon emissions during the manufacture
of reinforced concrete and steel are very high. Also a great amount of waste is left in
the air, water and soil during the production process of both materials (Yeang 2006).
The problems which were caused by these materials were neglected in the post-war
period, but the environmental problems started to attract attention recently. Nowa-
days important problems such as air pollution, non-renewable resource consumption
and global warming are emerging due to high carbon emissions (Yeang 2006; Fenger
2009; Jim 2014). It is necessary to take some precautions to solve these problems.
One of the most frequently used solution for this purpose; is to increase the use of
ecological materials.
Ecological material can be defined as a material with low embodied energy and
waste. When timber is evaluated according to these properties it can be accepted as
an ecological building material and its usage should be supported. It is necessary
to build structures with ecological building materials and to properly evaluate their
mechanical properties. Mistaken evaluations can cause large number of replacements
and this is economically and socially unsustainable due to increased natural resource
consumption and costs (Morales Conde et al. 2014a). The evaluation of the mechan-
ical properties of timber is more difficult than the other building materials because
of its organic structure. Timber’s properties change depending on the geographical
region, the climate, the orientation, and the properties of soil and water where the
tree grows. The mechanical properties of timber is evaluated by; destructive, semi-
destructive and non-destructive methods carried out together with visual inspection
(Feio 2006; Sousa 2013).
The primary technique used in the evaluation of timber structures is visual inspec-
tion. Although this method does not require any special tool and depend on the deter-
mination of the defects and decay signs, it requires a well assessed experience by a
specialist (Feio 2006). The actual strength of timber could be measured by destructive
tests which are applied according to standards and destroy the timbers. However, this
type of approach is usually not acceptable in existing buildings as it causes the loss of
structural integrity (Calderoni et al. 2010). In recent years semi-destructive and non-
destructive tests are used in the evaluation of timber structures. Non-destructive and
semi-destructive tests are methods that enable to obtain information about internal
structure and level of deterioration without harming the material (Ross and Pellerin
1994). Non-destructive tests are applied with special tools without giving damage
to the specimen and are used to predict various properties that can give information
Evaluation of Wooden Structures 763
about the mechanical properties of the material. The non-destructive tests are based
on the relationship between the measured parameter and the mechanical parame-
ter. Here the relationship is usually obtained by correlations. Semi-destructive tests
are usually carried out by the extraction of small specimen from the timber without
influencing the strength properties and the results are measured precisely as in the
destructive tests. The non-destructive and semi-destructive tests are usually used to
provide information about a single feature of the material, so it is more efficient
to use a combination of visual inspection and several non-destructive and/or semi-
destructive tests in the evaluation of a structure. It is even more convenient to establish
a test procedure before application (Sousa 2013). By using the data obtained from
destructive, semi-destructive and non-destructive tests, it may also be possible to pro-
duce simulation programs in order to evaluate the behavior of the structural timber
members. In this study, some important non-destructive and semi-destructive test
methods which is used in the evaluation of timber structures, their efficiency and
reliability are given.
2 Using Wood for Construction
Wood has been one of the most used natural resources from the earliest known years
of human beings to obtain tools, food and shelter. The areas of use and methods of
production have changed over time as needs have grown, but the first production
methods have been the infrastructure of today’s modern systems. The material is
characterized by a remarkable strength to weight ratio (Kelley et al. 2000; Asıf
2009).
The oldest known examples of wooden constructions are tent-like shelters built
with wooden branches and twigs. In the Paleolithic and Mesolithic period the
branches were buried into the ground, bent over to the center and bound together to
form the frame. The wall and roof was usually produced from interwoven branches.
Terra Amata (Fig. 1) which is an open air Paleolithic Period archeological site, dated
to about 400.000 BC consists of this kind of shelter ruins (de Lumley 1969).
Neolithic Period is the period in which human beings began cultivating plants and
settle down. As this period is the beginning of the early civilization, the investigation
of constructions and materials also starts from this moment. The circular shelters of
Terra Amata continued to be used in the Neolithic period, however the form turned
into a rectangular plan in the last years of this period. The use of stone and adobe
increased after the Neolithic Period, but the use of wood has never been abandoned
(Davey 1961; Kelley et al. 2000).
Wood also has been used widely in ancient civilizations. Egypt, Babylonian,
Greek and Roman civilizations used wood in different parts of the structure; and
even the carpenters were among the leading people of the society at that period.
Wood has especially been used in the production of roof and furniture in Egypt,
while it was used mainly in the production of roofs and roof supports in Greek and
Roman civilizations. Even after the transition of construction materials from wood
764 G. Koca
Fig. 1 Terra Amata, Nice (URL 1)
to stone in Greek Architecture, some temples continued to have wooden structural

members. It has been suggested that the columns of The Temple of Hera at Olympia
was originally made of wood, however the columns were later replaced with stone
in order to increase the service life of the temple (Derry and Williams 1960; Kelley
et al. 2000).
Wood has mainly been used as a structural material before the Industrial Revo-
lution. During this period, wooden constructions have been produced according to
the culture and local traditions of the people, by using available wood species of
the surrounding area. The material has been used in different parts of the building
from wooden frame construction systems to furniture. Early wood buildings were
produced with solid walls. This type of construction was produced by using split
trunks vertically edge to edge at the beginning and named as “stave construction”.
A modern structure produced with the stave construction system can also be seen in
Fig. 2.
Later on “log-cabin” or “log construction” systems have started to be used with
logs placed horizontally on top of each other. A modern log-cabin construction system
structure can be seen in Fig. 3. Although these building types were common in many
places around the world, solid walls consumed a great amount of wood and were
also sensitive to biological deterioration factors. Therefore the section dimensions of
wood decreased and the spaces between the studs were filled with different materials
such as, wattle and daub or bricks in the later periods (Kelley et al. 2000). There
are some important settlements which were built with the mentioned technique and
is still being used, such as Cumalıkızık, Turkey. Cumalıkızık is a village in Bursa
which was built around 700 years ago. The buildings were constructed in post-and-
pane structural system and the upper floors were built with mud-brick wooden frame
technique (Tarım and Hattap 2016).
Fig. 2 Stave cabin construction (URL 2)
Fig. 3 Log cabin construction (URL 3)
As a result of the technological developments following the Industrial Revolution,

the use of reinforced concrete and steel has increased due to their high strength values.
Nevertheless, wood continued to be used in the production of low-rise buildings,
interior finishings and furniture production. Nowadays, wood composites are being
used in constructions to prevent the consumption of natural resources. Due to the
high dimensional stability and chance to obtain elements of large dimension, wood
766 G. Koca
Fig. 4 Initial embodied

energy and carbon emissions
of different construction
materials (URL 4)
composites can be used in large span timber constructions and higher buildings
(Kelley et al. 2000; Astrubali et al. 2017).
3 Importance of Evaluating Buildings in Terms of Ecology

and Sustainability
There are significant environmental problems that are caused by modern building
materials such as; the consumption of non-renewable natural resources, the use of
fossil fuels in the production processes and the excessive amount of waste produc-
tion. One of the most important environmental problem is the air pollution caused
by the emission of greenhouse gases occurred from the production and use of build-
ing materials (Fig. 4). The air pollutant particles not only cause significant health
problems, but also harm the nature and buildings. Historical buildings which have
been used without any damage for many years have been severely damaged due to
air pollution in a short time after the Industrial Revolution (Watt et al. 2009).
In recent years the preservation of the architectural heritage has gained importance
within the scope of sustainable development. In order to provide sustainability in the
construction industry; buildings with long service life which does not increase land
and resource consumption and does not increase air pollution are supported (Berke
and Conroy 2000; Yeang 2006). Therefore, it is crucial to properly evaluate the
residual structural properties of the buildings in order to increase the service life of
the constructions.
During the evaluation of current situation of the buildings, the behavior of the
materials vs environmental factors becomes important. All building materials are
subjected to physical, chemical and biological degradation factors (Addleson and
Rice 1991). Fungi and insects are responsible from the bio deterioration of wooden
members and there are many types of fungi and insects which can live in different
environmental conditions. They may grow inside the wood or on the wood surface and
sometimes they can live undetectable for many years (Bodig and Jayne 1982). The
activity of biodegradation agents depend on the natural durability of the wood species
and by the thermo-hygroscopic features of the surrounding environment. These types
of microorganisms grow depending by the environmental conditions (Illston and
Domone 2001). Wooden structures, especially those of high historical value, are
highly susceptible to deterioration since they are directly exposed to environmental
factors. Therefore it is crucial to protect wooden structures and to properly evaluate

the mechanical properties of them.
4 Methods Used in the Evaluation of Structural Timber

Members
In the evaluation of wooden structures first of all the difference between the historic
structures and other existing structures needs to be determined. Although many of
the evaluation methods are common for different type of structures, a great care has
to be taken in historical structures. Usually the most conservative approach is by far
to be preferred (Cruz et al. 2015).
In the evaluation of wooden structures a protocol should be established before
the intervention and the inspection has to be carried out according to this protocol.
Recently some standards have started to be used for this purpose in order to avoid
mistaken evaluations. The main standards used are ISO 13822 (2016), BS EN 1995-
1-1 (2004) + A2 (2014), UNI 11119 (2004) and SIA 269/5 (2011) (Deuse 2017). A
comprehensive study is also being carried out for a possible European Standard as
discussed with CEN/TC 346 (Conservation of Cultural Heritage). During the in situ
evaluation of wooden structures the main aim is to obtain adequate information by
having minimal impact on the structure. The procedure carried out in the evalua-
tion of wooden structures are; desk survey, in situ visual inspection, measurements
(dimensions of structural members, sign of damages), structural analysis (high stress
points, critical zones) and a preliminary report (Cruz et al. 2015).
Desk survey is carried out to obtain knowledge about the history of the struc-
ture, and if the structure has a historical value the previous interventions also have
to be investigated and stated. Visual inspection is carried out to determine the cur-
rent situation of the structure in order to plan the next stages of the intervention.
Measurements include the dimensions of the structural members, signs of damage
and decays to locate the problematic parts of the structure. Structural analysis has
to identify the areas of high stress. After these stages, a preliminary report has to be
prepared. The preliminary report also has to include the types of wood species, the
moisture contents of the wooden members and the parameters of the surrounding
environment (Cruz et al. 2015). According to the results of the preliminary report the
critical structural members can be identified and then should be evaluated by visual
inspection, semi-destructive and non-destructive methods.
4.1 Species Identification
A large number of wood species is being used in the construction industry and
the identification of wood species helps to determine the original durability and
768 G. Koca
mechanical strength of the wood. The identification may be carried out by using
different techniques and combining them, observing macroscopic and microscopic
characters. Macroscopic identification is the first stage and is carried out by the
observation of wood’s structural properties, color, growth rings, porosity, etc. with
the naked eye or with a hand lens. The characters of annual rings, early wood/late
wood proportions and resin canals are useful elements in the identification of tree
species. Before the identification the structural member has to be cleaned. During
the evaluation the unknown specimen is compared with standard properties of the
tree species defined in a scientific wood species atlas (Macchioni 2010; Riggio et al.
2014).
Microscopic identification is characterization of wood by using a microscope. The
macroscopic features of wood can sometimes be hidden by the surface properties
of the structural member and therefore the macroscopic evaluation is prevented. In
such cases, microscopic evaluation is essential to identify the wood species. In order
to identify the wood species with a microscope, small specimens has to be extracted
from the structural member. A thin specimen is prepared along three principal direc-
tions. Then the specimen is evaluated under a microscope in order to identify the
characteristic anatomical features of the species using an atlas of wood anatomy.
4.2 Estimation of Moisture Content
Wood, like many natural materials is hygroscopic and it absorbs moisture from the
surrounding environment. Moisture content is defined as the proportion of water to
the proportion of wood material and it influences many properties of wood such as
elasticity and strength parameters. While the mechanical properties, including brit-
tleness, of wood tend to increase at lower moisture contents, biological degradation
starts in high moisture contents which strongly effect the mechanical properties of
wood. In existing structures the moisture content of wooden members can change
easily as the wood absorbs and desorbs moisture from the surrounding atmosphere.
It can also change within the structural member daily or seasonally (Bodig and Jayne
1982; Glass and Zelinka 2010).
The most accurate way of determining the moisture content of wood is the oven-
dry method, however it is time consuming and cannot be carried out in existing
buildings. Therefore some other indirect methods which can determine the moisture
content rapidly are being used in the estimation of moisture content in situ. The
methods which is commonly used in situ are; capacitance and resistance methods.
Capacitance method (pin-less meter)
The capacitance method use the dielectric properties of wood with the help of alter-
nating current. Decrease of the frequency signal indicates the degree of moisture
content during the measurement. This method is suitable in the evaluation of mois-
ture content of structural wooden members for moisture contents between 7 and 30%.
And any type of bipolar preservative, flame retardant, chemical or surface treatment
Fig. 5 Capacitive (pin-less)

moisture meter (URL 5)
may affect the accuracy of the measurement and requires special calibration as men-
tioned in EN 13183-3 (2005).
In this method the moisture content is measured with surface electrodes (Fig. 5).
The electrodes used for this type of moisture meters are mostly surface contact
electrodes, however in rough surfaces different type of rods which can adjust the
irregularities are used (Kasal and Lear 2010).
The device has to be calibrated for density before the measurement and also has
to be applicable for the thickness of the specimen. The electric field which is created
from the electrodes penetrates to a depth of 20 mm and an average thickness of
40 mm can be tested with this type of moisture meter. The measurement area has to
be free from bark, knots, resin pockets or checks.
Resistance method (pin meter)
The resistance method uses the relation between moisture content and the direct
current conductance of wood. As water is a good and wood is a poor conductor of elec-
tric; when the moisture content increases, the resistance to electric current decreases
(Kasal and Lear 2010). The method is usually suitable in temperature between 5 and
60 °C and can measure moisture content between 7 and 30%. A calibration should
be done according to the examined wood species before the measurement and some
types of preservatives, flame retardants, chemical treatments can affect the accuracy
of the measurement as mentioned in EN 13183-2 (2002).
The device consists of a portable resistance type moisture meter with different
type of electrodes. Especially pin type electrodes give reliable readings as they are
insulated into the wood and permit local moisture content estimation as can be seen
770 G. Koca
Fig. 6 Resistance type moisture meter
in Fig. 6. Electrodes should be placed into one face of the piece in the direction of
the grain and the measuring spot has to be free from defects (Riggio et al. 2014).
4.3 Semi-destructive and Non-destructive Techniques
The test methods used in the evaluation of wooden members are classified in two dif-
ferent groups; according to the type of parameter obtained which takes into account
the measurement extent and to the damage level. The classification made according
to the type of parameter obtained which takes into account the measurement extent
is divided into local and global test methods. Local test methods determine the prop-
erties of small parts of the member and the properties of full member is considered
by correlations or empirical knowledge. Semi-destructive test methods are types of
local test methods as they provide information only about the extracted specimens.
However, global test methods measure parameters of relatively large areas or even
the entire section of the member. Both non-destructive and destructive methods are
global test methods (Baldassino et al. 1996; Sousa 2013).
The test methods are classified into three groups according to the damage level;
non-destructive, semi-destructive and destructive methods. The test is classified as
a non-destructive test if it does not cause any damage on the member. If a minor
damage is occurred, then the test is classified as a semi-destructive test. However, if
the member is destroyed after the measurement, the test is classified as a destructive
test (Sousa 2013).
Destructive tests don’t use estimations, they accurately determine the mechanical
properties of the structural wooden members. The specimens are tested in the labora-
Table 1 Some of the most preferred testing methods used in the evaluation of timber
Evaluation technique Test method
Non-destructive Visual inspection
Acoustic methods (sonic and ultrasonic stress waves)
Digital radioscopy
Semi-destructive Drilling resistance test
Needle penetration test
Pin pushing test
Screw resistance test
Extraction and testing of small non-standard specimens
tory to the limit values as mentioned in the standards and the mechanical properties
are determined. However, as destructive tests destroy the wooden member during
the measurement it is incompatible in the evaluation of existing buildings. In the
technical evaluation of existing buildings these types of tests can be used to support
the non-destructive ones.
Semi-destructive and non-destructive tests are more and more used in the most
recent years in order to evaluate mechanical properties of structural wooden members.
The semi-destructive tests are carried out by the extraction of a small specimen from
the wooden member without reducing the strength of the specimen. Semi-destructive
tests also allow a more reliable estimate of mechanical properties of the member
(Tannert et al. 2014). Non-destructive tests are carried out with the help of simple
devices to obtain information about the internal structure and deterioration level of the
specimen. Non-destructive tests are based on the relationship between the measured
parameter and a mechanical parameter (Riggio et al. 2014). A brief list of some tests
according to the classification by the damage level can be seen in Table 1.
Semi-destructive and non-destructive test methods give information only about a
single parameter of the member. Therefore it is more convenient to establish a test
procedure, combining different semi-destructive and non-destructive tests, in order
to increase the reliability of the evaluation.
4.3.1 Non-destructive Techniques
Non-destructive tests are carried out with the help of simple devices to obtain infor-
mation about the internal structure and deterioration level of the specimen. Some of
the most used non-destructive tests are as follows;
• Visual inspection
It is the primary method used in the evaluation of the current situation of wooden
structures. Visual inspection is based on the inspection of the structural members
without the need for a special tool. The technique is suitable for the application of
wooden structures, specifically for solid wood and of limited use when applied to
772 G. Koca
glue-laminated wood. The inspection should be started with the members that are
more susceptible to defects. The geometry, joint types, support points also should
be measured in order to identify deformations due to loads. Important problems
such as mechanical damage and the presence and extent of wood defects should
be determined at this stage and the effect of these problems has to be interpreted.
If a historical structure is being evaluated, the previous interventions also has
to be reported. During the inspection, the structure should be examined as close
as possible to determine the problematic parts. Thus, a general comment can be
made on the structure and the next interventions can be planned which includes the
evaluation of the members by semi-destructive and/or non-destructive tests (UNI
11119 2004; Riggio et al. 2014).
The equipment which are used in semi-destructive and non-destructive tests often
require calibration according to different tree species, therefore the tree species
needs to be determined after the visual inspection (UNI 11118 2004). At this stage,
the moisture content of the member must also be determined. Due to wood decay
occurs at high wood moisture content (generally over 20%) the moisture content of
the critical deterioration points has to be determined (EN 13183-1 2002; EN 13183-
3 2005). After the determination of the tree species and the moisture content, the
dimensions of the structural members has to be measured, the problematic points
has to be marked and a report has to be prepared. Although visual inspection is the
most important method, the accuracy depends a lot by the experience of the expert.
Therefore it is recommended to support visual inspection by other non-destructive
techniques (Cruz et al. 2015; Feio and Machado 2015)
• Acoustic methods (Sonic and ultrasonic stress wave)
Stress waves are being used for a long time in the non-destructive evaluation of
wooden members and are usually used as a predictor of inner voids, deteriorations
and the mechanical properties of the specimen. The application is based on the
transmission of sound waves along the wooden member (Kasal et al. 2010).
The application is carried out by measuring the time of flight of the stress wave
to determine the wave transmission velocity. The method require knowledge of
material density and moisture content. In this technique the stress wave can be
produced in sonic or ultrasonic sound frequencies. In sonic stress waves (usually
named as “stress waves”) the impact is created on the specimen with frequencies
in the audible range. In ultrasonic stress waves the impact is created in inaudible
range, above 20 kHz. Ultrasonic stress wave method is mostly used in homoge-
neous and non-porous materials. Therefore it is usually less effective in wooden
materials due to porous and heterogeneous structure. Different application types
of the test can be seen in Fig. 7.
The stress wave is produced in transverse or longitudinal direction with a hammer
in sonic stress wave technique. The timer is started when the stress wave is initiated
at the first accelerometer and stopped when it reaches the second accelerometer.
The ultrasonic stress wave technique is similar to the sonic stress wave method
but uses higher frequencies. The ultrasound is usually created by a piezoelectric
material which convert electrical current into waves. The testing frequencies are
Fig. 7 Types of different

acoustic experimental setup
(URL 6)
typically in the range of 20–500 kHz which is relatively low again because of the
heterogeneous structure of wooden materials (Feio 2006; Kasal et al. 2010).
The setup of the application is usually the direct transmission of the wave, but
this can sometimes be inapplicable in situ. In this case indirect application can be
done. The time of flight (TF) is measured with the equation;
V L/TF
where
V velocity
L length
If the application is used to estimate the dynamic modulus of elasticity, density
must also be measured. By using velocity and density (ρ) of the specimen the
dynamic modulus of elasticity can be determined with the equation;
MoED V2 × ρ
The velocity of the stress wave depends on the elastic properties of the material.
If the signal is deviated despite the heterogeneity, anisotropy or inner voids the
transmission time increases. Therefore longer transmission times generally implies
defects or deteriorations in these techniques (Fig. 8). Also, as sonic stress waves
have longer wavelengths they are not sensitive to small defects, ultrasonic stress
waves have a great ability in detecting small defects (Feio 2006; Kasal et al. 2010;
Senalik et al. 2015).
• Digital radioscopy
This method can also be named as the advanced visual inspection method. Digital
radioscopy is based on the use of X rays in the evaluation of wooden members.
By using the penetration property of X rays to different materials the image of the
member can be obtained and the internal defects such as insect attacks, construction
defects, splits and cracks can be observed. X rays with short wavelengths and
energy, which can easily penetrate wooden members, are being used to create
a digital image of the structural member. In order to evaluate wooden members
portable X-ray systems are used with different imaging systems.
774 G. Koca
Fig. 8 Stress wave propagation in sound and decayed wood (URL 7)
The most important factor while using X ray is its influence on health. Basically,
a portable digital radioscope is not radioactive and becomes unhealthy when it
starts to produce pulses. If the energy of the system is low and the user is far away
from the source (usually stated as 3 m), it can be mentioned that it is safe. This
requirement can sometimes prevents using the method in narrow spaces (Riggio
et al. 2014; Anthony 2004).
4.3.2 Semi-destructive Techniques
A small specimen is extracted from the wooden member without influencing the
strength properties and then tested in semi-destructive tests. Due to the variability
within the wooden member the number of measurements will greatly affect the reli-
ability of the evaluation. Therefore usually a large number of specimens is required
in order to obtain a reliable result. Some of the most used semi-destructive tests are
as follows;
• Drilling resistance test (Resistograph method)
Drilling resistance test is used to evaluate the density profile, mechanical proper-
ties, the current state and level of deterioration of the specimen. The test is carried
out by micro-drilling of wood at a constant speed and measuring the required
energy during the process. The density profile of the specimen along the depth is
obtained depending on the change in the resistance of the material. The drilling
holes need to be small enough, which is stated between 1.5 and 3 mm, to identify
the differences in density and to detect the existing deteriorations of the specimen
(Feio 2006; Tannert et al. 2014).
This type of damage does not have a weakening affect and does not change the
mechanical properties of the specimen. After the test is carried out with an appro-
priate number of drillings, the density profile which is a graphical display of the
results is obtained. The graph shows the deteriorations and internal defects of the
specimen through the penetration depth as seen in Fig. 9.
Although this method is efficient in the evaluation of the structural timber members
there are some important factors which effects the application such as; accessibility
Fig. 9 Drill resistance equipment and a graph obtained
and damage to the specimen. Accessibility is an important factor which effects the
application of the method and the length of the test device sometimes limits its
applicability in narrow spaces. The drilling hole is also sometimes not allowed due
to the historical importance of the member. Also this technique can be sometimes
time consuming as it often requires multiple drills on the member (Morales Conde
et al. 2014b).
• Needle penetration test (pylodin)
Needle penetration test works similarly to the drilling resistance test. As density is
directly related to the mechanical properties and hardness of wood, the main aim
of the test is to evaluate the density of the specimen. The method is mainly based
on the penetration of a pin into the specimen and the evaluation of the current state
of wooden members by using the correlation between the depth of penetration and
the density (Feio 2006; Branco et al. 2010).
Needle penetration test is usually named as “Pilodyn” which is a brand name
widely used for this purpose. There are several types of device; some working
with increasing strike forces and some allowing repeating shots while estimating
the density of the specimen. The most popular model 6 J gives a single shot to the
specimen and drives the attached striker into the wood. The penetration depth can
be read from the device (Fig. 10) and wood’s resistance to the penetration of the
striker can be evaluated (Feio 2006).
Although it is a simple and rapid technique there are some weaknesses of the
test such as; it gives local information about the specimen and it can evaluate
only the outer 30–40 mm part of the wood member due to the length of the
striker. Also the penetration depth is effected from factors such as; species, density,
moisture content and defects of wood. However, it can be mentioned that the needle
penetration test can predict the mean outerwood density values much easily and
with a reasonable accuracy than the equivalent techniques such as core drilling
(Feio 2006; Watt et al. 1996).
• Pin pushing test
Pin pushing test is a mechanical resistance measurement technique which differs
from the other penetration tests. In this technique the pin is not pushed into the
wood dynamically, it is gradually pressed down. In this way, the applied force can
776 G. Koca
Fig. 10 Needle penetration test equipment and application (URL 8)
be continuously monitored and recorded. The method has also the opportunity of
evaluating the entire cross-section of the specimen up to a depth of 120 mm as can
be seen in Fig. 11 (Kloiber et al. 2011).
The test apparatus can be mounted to the member to be tested and fixed to the
specimen with a strap. After the apparatus is fixed to the specimen, the pin is
pressed into the wood. The force is measured and recorded during the application
and is transferred to a PC through a wireless connection (Tippner et al. 2011).
During the pin pushing process the device monitors and records the applied force
to obtain a graph. The graph shows the force applied through the depth to which the
pin was pushed. The peaks in the graph shows the resistant parts of the specimen
and the lower values show the lower resistant parts (Kloiber et al. 2015).
• Screw resistance test
Screw withdrawal test is developed to evaluate the biological degradation of
wooden members. The method is mainly based on the insertion and extraction
of the screw probes and the measurement of maximum pull out loads as a with-
drawal resistance. According to previous researches it has been mentioned that
MOE and MOR of wood are correlated to screw resistance of wood. Although
density is a good indicator to evaluate the degradation of wood, MOE and MOR
are better indicators of the mechanical properties of structural timber members
(Yamaguchi 2010).
There are different types of tools used for the procedure which include a load cell
and measure the withdrawal resistance of wood during the application. Probes has
to be as long as the specimen depth and small probe diameters should be preferred
in order to prevent damages on the specimen. After obtaining the withdrawal
resistances, the average of multiple withdrawal resistances should be evaluated
and the results of sound new wood and specimen has to be compared. The method
gives local information about the structural member and therefore in order to
Fig. 11 Pin penetration test

(URL 9)
obtain a global information it has to be applied to different parts of the specimen

(Winandy et al. 1998; Tannert et al. 2014).
• Extraction and testing of small non-standard specimens
A semi destructive technique based on the testing of small non-destructive spec-
imens taken from structural timber members was developed by Kasal (2003) in
order to evaluate wood. In this technique two types of tests were carried out; micro
tension test and radial core test.
Micro-tension test is used in the evaluation of bending strength of the specimens,
which is an important loading type in constructions. A small triangular shaped
specimen is extracted from the member according to the test procedure. The
extracted specimen is later attached to small wooden blocks in order to avoid
problems due to clamping. After the specimen is extracted from the member and
attached to a wooden block, it is placed into a testing machine and the tensile
strength and modulus of elasticity (MoE) in tension parallel to grain is obtained
(Kasal 2004; Tannert et al. 2014).
778 G. Koca
Fig. 12 Radial core extraction test (URL 10)
Radial core test is carried out to obtain the compression strength and the modulus
of elasticity (MOE) of the member. Small cylindrical test specimens, which are
in radial direction, are extracted from the structural member, fixed to the testing
device and then tested. The device extracts small specimens with a diameter of
5 mm from the member with a core drill which can be seen in Fig. 12. The local
information which is obtained from the extracted specimen is later used to evaluate
an overall strength of the specimen (Kasal 2004; Tannert et al. 2014).
Although the extraction of small specimens leads to a reliable estimation of the
mechanical properties of wood as the results are obtained directly with destructive
tests, it must be mentioned that these methods are time consuming than the other
semi destructive tests.
5 Conclusion
Non-destructive and semi-destructive test methods are widely being used together
with visual inspection for the last few years in the evaluation of wooden structures.
Visual inspection is still accepted as the primary technique used in the evaluation
of wooden members. However, visual inspection depends on the knowledge of the
expert and it is better to be supported by other non-destructive and semi-destructive
methods. Non-destructive test methods are based on the relationship between the
measured parameter and a mechanical parameter. Therefore non-destructive tests
depend on the correlations obtained between the parameters. The above mentioned
non-destructive test methods have medium to high correlations with the mechanical
parameters of wood. Semi-destructive tests can evaluate the accurate mechanical
properties of wood, however a large number of small specimens has to be extracted

from the member in order to obtain a reliable result.
Although non-destructive and semi-destructive tests are usually effective in the
evaluation of mechanical properties there are still some deficiencies and problems in
the applications. Recently some standards have started to be used in the evaluation
of wooden structures, but most of them include different evaluation techniques. For
example; Eurocode 5 (EN 1995) is mainly based on the construction procedures of
new wooden structures and UNI 11119 is mainly carried out by the visual inspection
of the wooden members. There is not a main standard which includes all of the
mentioned non-destructive and semi-destructive methods and describes an evaluation
procedure for the wooden structures. The experts usually use different techniques
in the assessment procedure and therefore the combination and the results of the
methods change. However, a comprehensive study is being carried out for a possible
European Standard.
Also, there are many types of wood species used in the building industry and
because of the growing habitat, all of them have different properties. Due to the high
variability between wood species the coefficient of correlations also change and
there must be more researches about the most used species in order to obtain reli-
able results. However, using the non-destructive and semi-destructive tests combined
together with visual inspection in the evaluation of wooden structures can enable the
sustainable development of the cities. With the help of the data obtained from non-
destructive and semi-destructive tests, it may also be possible to produce simulation
programs in order to evaluate the behavior of the structural timber members easily.
References
Addleson L, Rice C (1991) Performance of materials in buildings: a study of the principles and
agencies of change. Butterworth & Heinemann, Oxford
Anthony RW (2004) Condition assessment of timber using resistance drilling and digital radioscopy.
APT Bull J Preservation Technol 35(4):21–26
Asıf M (2009) Sustainability of construction materials. In: Khatib J (ed) Sustainability of timber,
wood and bamboo in construction. Elsevier, Amsterdam, pp 31–54
Astrubali F, Ferracuti B, Lombardi L, Guattari C, Evangelisti L, Grazieschi G (2017) A review
of structures, thermo-physical, acoustical and environmental properties of wooden materials for
building applications. Build Environ 114:307–322
Baldassino N, Piazza M, Zanon P (1996) In situ evaluation of the mechanical properties of timber
structural elements. In: Proceedings of international symposium on non-destructive testing of
wood, pp 369–377
Berke PR, Conroy MM (2000) Are we planning for sustainable development? An evaluation of 30
comprehensive plans. J Am Plann Assoc 66(1):21–33
Bodig J, Jayne BA (1982) Mechanics of wood and wood composites. Van Nostrand Reinhold, New
York
Branco JM, Piazza M, Cruz PJS (2010) Structural analysis of two King-post timber trusses: non-
destructive evaluation and load-carrying tests. Constr Build Mater 24:371–383
BS EN 1995-1-1 (2004) + A2 (2014) Eurocode 5: Design of timber structures General. Common
rules and rules for buildings. British Standards Institute (BSI), UK
780 G. Koca
Calderoni C, Matteis G, Zamperini E (2010) Experimental correlations between destructive and

non-destructive tests on ancient timber elements. Eng Struct 2(32):442–448
Cruz H, Yeomans D, Tsakanika E, Macchioni N, Jorissen A, Touza M, Mannucci M, Lourenço
PB (2015) Guidelines for on-site assessment of historic timber structures. Int J Archit Herit
9:277–289
Davey N (1961) A history of building materials. Phoenix House, London
De Lumley H (1969) A paleolithic camp at nice. Sci Am 220(5):42–51
Derry TK, Williams TI (1960) A short history of technology: from the earliest times to AD 1900.
The Clarendon Press, UK
Deuse M (2017) Non-destructive diagnostics for timber structures in historic buildings: investigation
methods and testing tools. Master thesis, University of Cantabria, Spain
EN 13183-1 (2002) Moisture content of a piece of sawn timber. Determination by oven dry method.
European Committee for Standardization (CEN), Brussels
EN 13183-2 (2002) Moisture content of a piece of sawn timber: part 2: estimation by electrical
resistance method. European Committee for Standardization (CEN), Brussels
EN 13183-3 (2005) Moisture content of a piece of sawn timber: part 3: estimation by capacitance
method. European Committee for Standardization (CEN), Brussels
Feio A (2006) Inspection and diagnosis of historical timber structures: NDT correlations and struc-
tural behaviour. PhD thesis, University of Minho, Portugal
Feio A, Machado JS (2015) In-situ assessment of timber structural members: combining infor-
mation from visual strength grading and NDT/SDT methods—a review. Constr Build Mater
101:1157–1165
Fenger J (2009) Air pollution in the last 50 years—from local to global. Atmos Environ 43(1):13–22
Glass SV, Zelinka SL (2010) Moisture relations and physical properties of wood. In: Ross RJ (ed)
Wood handbook wood as an engineering material. FPL, Madison
Illston J, Domone P (2001) Construction materials: their nature and behavior. Taylor & Francis,
New York
ISO 13822 (2016) Bases for design of structures—Assessment of existing structures. International
Organization for Standardization (ISO), Geneva
Jim CY (2014) Air-conditioning energy consumption due to green roofs with different building
thermal insulation. Appl Energy 128:49–59
Kasal B (2003) Semi-destructive method for in-situ evaluation of compressive strength of wood
structural members. For Prod J 53:55–58
Kasal B (2004) Advances in in-situ evaluation of timber structures. Prog Struct Eng Mater 6:94–103
Kasal B, Lear G (2010) Moisture measurement. In: Kasal B, Tannert T (eds) In situ assessment of
structural timber. RILEM State of the Art Report. Springer, Nederland, pp 99–104
Kasal B, Lear G, Tannert T (2010) Stress waves. In: Kasal B, Tannert T (eds) In situ assessment of
Kelley SJ, Loferski JR, Salenikovich AJ, Stern EG (2000) Wood structures: a global forum on the
treatment, conservation and repair of cultural heritage. ASTM, USA
Kloiber M, Tippner J, Drdachy M (2011) In: Proceedings of international conference on structural
health assessment of timber structures, Portugal, Lisbon, June 2011
Kloiber M, Drdachy M, Tippner J, Hrivnak J (2015) Conventional compressive strength parallel to
the grain and mechanical resistance of wood against pin penetration and micro drilling established
by in-situ semi-destructive devices. Mater Struct 48(10):3217–3229
Macchioni N (2010) Species identification. In: Kasal B, Tannert T (eds) In situ assessment of
Morales Conde MJ, Rodriguez Lińan C, Rubio de Hita P (2014a) Use of ultrasound as a non-
destructive evaluation technique for sustainable interventions on wooden structures. Build Envi-
ron 82:247–257
Morales Conde MJ, Rodriguez Lińan C, Saporiti-Machado J (2014b) Predicting the density of
structural timber in service. The combine use of wood cores and drill resistance data. Materiales
de Construccíon 64(315):1–11
Riggio M, Anthony RW, Augelli F, Kasal B, Lechner T, Muller W, Tannert T (2014) In situ assess-
ment of structural timber using non-destructive techniques. Mater Struct 4(47):749–766
Ross RJ, Pellerin RF (1994) Nondestructive testing for assessing wood members in structures.
USDA. FPL-GTR-70, Forest Product Laboratory, Madison
Senalik A, Schueneman G, Ross RJ (2015) Ultrasonic-based non-destructive evaluation methods
for wood. In: Ross RJ (ed) Non-destructive evaluation of wood, 2nd edn. FPL, Madison, pp 21–51
SIA 269/5 (2011) Existing structures—timber structures. Swiss Society of Engineers and Architects,
Zurich
Sousa H (2013) Methodology for safety evaluation of existing timber elements. PhD thesis, Uni-
versity of Minho, Portugal
Tannert T, Anthony RW, Kasal B, Kloiber M, Piazza M, Riggio M, Rinn F, Widmann R, Yamaguchi
N (2014) In situ assessment of structural timber using semi-destructive techniques. Mater Struct
5(47):767–785
Tarım A, Hattap S (2016) In: Proceedings of adobe 2016, international conference on cultural
landscape: rebuilding after decay, Istanbul Aydın University, Istanbul, 17–18 December 2016
Tippner J, Kloiber M, Hrivnak J (2011) In: Proceedings of 17th international nondestructive testing
and evaluation of wood symposium, Sopron, 14–16 September 2011
UNI 11118 (2004) Cultural heritage: wooden artifacts: criteria for the identification of the wood
species. Ente Nazionale Italiano di Unificazione, Milan
UNI 11119 (2004) Cultural heritage: wooden artifacts: load bearing structures: on-site inspections
for the diagnosis of timber members. Ente Nazionale Italiano di Unificazione, Milan
URL 1. http://www.thesubversivearchaeologist.com/2014/05/. Access date: 18 July 2018
URL 2. http://countryliving.ws/wp-content/uploads/2015/11/vertical-101.jpg. Access date: 18 July
2018
URL 3. http://www.onlyinyourstate.com/idaho/cozy-cabins-in-idaho/. Access date: 18 July 2018
URL 4. https://www.researchgate.net/publication/265051465_Quantifying_life_cycle_energy_
and_carbon_footprints_of_china%27s_residential_small_district/figures?lo=1. Access date: 18
July 2018
URL 5. https://www.fibreglast.com/product/Pinless-Moisture-Meter/Instrumentation. Access date:
18 July 2018
URL 6. https://www.researchgate.net/publication/225334752_Mechanical_properties_assessment_of_
Cunninghamia_lanceolata_plantation_wood_with_three_acoustic-based_nondestructive_methods/
figures?lo=1. Access date: 18 July 2018
URL 7. https://opus.lib.uts.edu.au/bitstream/10453/35427/4/Recommendation%20of%20RILEM%
20TC%20215-AST%20-%20Stress%20Waves%20-%20v6.pdf1. Access date: 12 July 2018
URL 8. https://stavba.tzb-info.cz/drevene-konstrukce/9669-metody-pro-zjistovani-spolehlivosti-
historickych-drevenych-konstrukci. Access date: 25 June 2018
URL 9. https://ars.els-cdn.com/content/image/1-s2.0-S0950061815006923-gr22.jpg. Access date:
24 June 2018
URL 10. https://www.researchgate.net/publication/257896291_In_situ_assessment_of_structural_
timber_using_semi-destructive_techniques/figures. Access date: 18 July 2018
Watt M, Garnett B, Walker J (1996) The use of the Pilodyn for assessing outerwood density in New
Zealand radiate pine. For Prod J 46(11/12):101–105
Watt J, Tidblad J, Kucera V, Hamilton R (2009) The effects of air pollution on cultural heritage.
Springer, New York
Winandy JE, Lebow PK, Nelson W (1998) Predicting bending strength of fire-retardant-treated
plywood screw-withdrawal test. FPL-RP-568, Forest Product Laboratory, Madison
Yamaguchi N (2010) Screw resistance. In: In situ assessment of structural timber. RILEM State of
Art Reports, vol 7. Springer, Dordrecht, pp 81–86
Yeang K (2006) Eco-design: a manual for ecological design. Wiley, Chichester
Solid Wood and Wood Based Composites:
The Challenge of Sustainability Looking
for a Short and Smart Supply Chain
M. Romagnoli, M. Fragiacomo, Antonio Brunori, M. Follesa

and G. Scarascia Mugnozza
Abstract The paper takes into account the most important wood based products
used in architecture, structural engineering and design. The amount of roundwood,
sawnwood and wood panel production is analysed and forest certification is reported
as possible tool to ensure a sustainable forest management, fighting illegal logging
and deforestation. A smart review of the most used wood-based products is performed
together with the chance to activate a modern short supply chain. The state of art
of the wood species actually used is considered together with most important actual
challenges addressed to obtain sustainable wood-based products referring to eco-
friendly process of gluing and increasing durability. The possibility to use short chain
species for structural purposes is reported looking to Italian case studies (chestnut,
eucalyptus and pines). A lot of successful prototypes and products have been obtained
using short supply species, but research still is needed on the most type or products
especially if they are to be used for structural purposes.
Keywords CLT · Panels · Gluing · Forest certification · Forest production
M. Romagnoli · G. Scarascia Mugnozza (B)

Department of Innovation in Biotechnology Agrifood and Forestry (DIBAF),
University of Tuscia, Viterbo, Italy
e-mail: gscaras@unitus.it
M. Romagnoli
e-mail: mroma@unitus.it
M. Fragiacomo
Department of Civil, Construction-Architecture and Environmental Engineering (DICEAA),
University of L’Aquila, L’Aquila, Italy
e-mail: massimo.fragiacomo@univaq.it
A. Brunori
PEFC, Perugia, Italy
e-mail: info@pefc.it
M. Follesa
dedaLEGNO, Firenze, Italy
e-mail: follesa@dedalegno.com

784 M. Romagnoli et al.
1 Introduction
In recent years, there has been a renewed interest in wood products, due to a growing
awareness of wood’s role as a carbon pool, as well as in its aesthetic value, and to
the technological performances especially in terms of high mechanical properties to
specific weight ratio. The interest in the material is expected to increase especially
due to the role of urban’s settlements in the near future. By 2050 world population is
estimated to be 9.8 billion, while income per capita is expected to triple. Nowadays
50% of the world’s population is located in urban areas and the prediction for 2050
is an increase in the percentage that will reach 75%, in Italy it was 69.12% in 2016.
The impact on the building industry cannot be neglected due to the more strict
requirements of new construction and/or the requalification of the oldest ones. Rea-
sonably, wood will be one of the most preferred materials also looking to the envi-
ronmental perspective. It is well known for its role as carbon stock (1 m3 of wood
means 1 ton of CO2 stored in the woody tissues), less known is the estimate of carbon
footprint reduction up to 50% considering the substitution effect on concrete or steel.
Wood in urban context respond to the customer’s needs differently from wood
products in rural environment. Urban customers are quite more demanding especially
in the design of the final artefacts resulting in the request of wood based composites
which can be transformed in daring shapes, using a reduced size of the elementary
component, and which can reach ad used in larger sizes. According to this figure
there will be an increasing demand of wood in the future which must be faced
with a vision and approach undoubtedly based on the science of sustainability. This
concept takes into account two main challenges, the first is related to wood supply,
while the second is linked to the technology of wood based products. From the latter
perspective, the possibility to obtain sustainable composite products must take into
account two important elements: bonding and durability; both processes respond
more and more to the need and to the wish to apply the products in severe use
conditions which are represented by high humidity and therefore a high risk of
degradation due to biotic attack. On one hand, wood is well appreciated because it is
a natural material and renewable, on the other hand there is the requirement to have
a product long lasting in time and with suitable mechanical performances, easy to
manage and to conserve, like other materials. For that reason, one of the major goals
in wood technology research, in order to make wood products more competitive
and attractive, is not only to develop a large variety of shapes and forms but also to
improve the technological performances related to the biodegradability of the woody
material. According to this general picture, looking at the building sector as the main
driver of wood economy, the main topics which need to be taken into account may
be summarized as:
– Forest supply and the role of short forest—wood supply chain
– Solid wood and wood base composites: a smart overview
– Forest management certification and chain of custody certification
– Sustainable wood durability—wood gluing
– Wood for structural application
Solid Wood and Wood Based Composites … 785
2 Forest Supply and the Role of Short Forest-Wood Supply

Chain
At global scale, forests cover is composed by 2969 Mha (mega hectares or mil-
lion hectares) of public forests, 433 Mha of individually-owned forests, 116 Mha
of community-owned forests and 559 Mha of a combination of community- and
individually-owned forests (FAO 2018). Total land surface area of the Earth is
14,900 Mha; forests cover 30.6% of this surface. In many areas of the world,
forest cover is still declining except in temperate regions, a fundamental issue of
the environmental sustainability concept. The five countries with the largest forest
areas in the world are Russian Federation, Brazil, Canada, United States of Amer-
ica and China. Jointly they account for more than half of the global forest sur-
face with the Russian Federation which is by far the most forested country of the
world. The European forest resources (EU-28) are extensive as they cover approx-
imately 182 million ha (over 43% of the EU territory), of which 130 Mha can be
sustainably managed and used for timber supply (EC-Agricultural and Rural Devel-
opment 2018). European forests account for 5% of the total world forest surface,
while the European most forested country is Sweden (30.5 Mha), and the small-
est is Malta (300 ha) (https://ec.europa.eu/eurostat/statistics-explained/index.php/
Forests,_forestry_and_logging, 2015). The largest part of the forests in the world
(80%) has been reported to be commercially exploitable (or already exploited). The
percentage is 63.5% in Africa, 81% in Europe and 85.5% in Asia. Looking at the
Mediterranean region, at country level, wood volume of commercial value ranges
from 2% of Tunisia to 100% of Morocco. In France and Slovenia wood in “com-
mercial” forest represents more than 90% of total wood volume of the countries
(Matteucci et al. 2011).
In the FAO-State of the World’s Forest (2018) the Global Index of Wood pro-
duction from 1970 related to roundwood and sawnwood + panels, shows a slightly
increasing trend, furthermore roundwood production continued in step with produc-
tion of sawnwood and panels between 1970 and 1999, after this period sawnwood
and panels production increased in much more substantial way. This trend can be
observed also considering the different product categories (Table 1); in Figs. 1, 2, 3,
4, 5, 6, 7 and 8 the production trends are also reported referring to Europe and to all
the different products, sawnwood, fiber based and particle based products, mixing
coniferous and not coniferous species (FAO-STAT 2016). It is quite evident as there
has been a big crisis of the whole wood sector in the decade 1990–2000, with a new
surge after this period which is more evident in plywood and veneers sheets, while
fiberboards and fiber based panels have reached the pre-crisis levels; less evident is
the recovery of the particle board panels except the oriented strandboard (OSB) which
shows a continuous trend of production from 1970 up to now. At European level the
most important producer of roundwood are Sweden, Finland, Germany, France and
also Poland (cf. Table 2). The most important sawnwood production is from Germany
(Table 3) which has also increased over time, while quite relevant and more stable
in time are also the sawnwood productions from Finland and Austria (Table 3).
Table 1 An overview of world-wide production of solid wood and wood-based products

Product Absolute Percentage (%)
production
2016 2015 2000 1980
Roundwood million m3 3737 1 8 19
Wood fuel million m3 1863 0 5 11
Industrial million m3 1874 3 11 30
roundwood
Wood pellets million t 29 6
Sawnwood million m3 468 3 21 11
Wood-based million m3 416 4 123 310
panels
Veneer and million m3 174 3 161 296
plywood
Particle- million m3 242 5 102 321
board, OSB
and
fibreboard
Wood pulp million t 180 2 5 43
Other fibre million t 12 −7 −19 70
pulp
Recovered million t 230 1 60 354
paper
Paper and million t 409 0 26 142
paperboard
Absolute value referred to 2016 and percentage of variation compared to the years 2015, 2000 and
1980 (FAOSTAT 2017)
In 2015 about 420,000 enterprises were active in wood-based industries across the
EU-28 and they represented 20% of manufacturing enterprises across the European
Union (https://ec.europa.eu/eurostat/, 2018). The economic importance of the wood-
based industries in the EU-28 as measured by gross value added was equivalent to
139 billion e or 7.3% of the overall manufacturing gross income in 2015 (Table 4).
Noteworthy, there has been a decrease in number of enterprises, Gross value added at
factor and the number of persons employed. Within the EU-28’s wood-based indus-
try, the highest share was recorded for pulp, paper and paper products manufacturing
(32.9% or 46 billion e), while the other three sectors had nearly equal shares—print-
ing and service activities related to printing and the manufacture of furniture each
amounted to 21–22% of the gross value added of wood based industries, while the
manufacturing of wood and wood products made up 24%.
In Italy wood industry turnover amounts at 41 billion e with 80,000 enterprises
and 320,000 employees, while the pulp and paper industry has an annual turnover
of 22 billion e, over 3800 companies and more than 20,000 employees. However, it
should be noted that 80% of the wood used by Italian manufacturers is imported from
Fig. 1 Europe production of sawlogs and veneer logs, coniferous (Source Eurostat)
other countries. Therefore, one of the primary objectives of forest policy in Italy is
to increase the degree of self-supply in this sector (Italian Strategy of Bioeconomy
2015). The necessity to resort to wood import is quite evident if we compare the
production of wood based products by European countries (see Figs. 1–8) with the
total amount of roundwood production in Europe; also, this is a quite recurrent
scenario even in countries which have a remarkable wood production. In 2016 the
total imports of the EU wood sector were estimated as 31.9 billion e, of which
around 23 billion e are from EU countries representing of 73.4%, while 26.6% are
from non-EU countries. In Europe the first importing country is Germany, Italy is
the fourth importer (Federlegno Report 2016). China is one of the most important
wood products exporting country to Europe.
However, the large role of wood trade does not always fit with the concept of
sustainability and it raises questions regarding legality and sustainability in the use
of resources. As for legality a crucial aspect would be forest certification, but a more
efficient and sustainable use of forest domestic resources is an issue which must be
also taken into account.
The chance to have a greater utilization of domestic resources must be explored
taking into account the expansion of forest land in many countries, particularly in
Europe and also in the Mediterranean region.
Fig. 2 Europe production sheets based panels: veneer sheets (Source Eurostat)
As for the latter, it is interesting to note that the increasing trend of forest cover,
or at least its stabilization, is widespread in almost all the Mediterranean countries.
Total forest cover increased from 69.13 Mha in 1990 to 76 Mha in 2000, reach-
ing 79.15 Mha in 2005. Most of the “new forest surface” of 10 Mha is located in
Europe (84%), 10% is in Africa and only 6% in Asia. The percentage increment has
been similar in Europe and Africa (around 1% per year) and lower in Asia (0.4%)
(Matteucci et al. 2011).
Focusing on Italy, the overall country’s annual increment of wood volume is
about 35 million m3 in forests and 1 million m3 in woody plantations (INFC 2007);
that is a much greater amount than the rate of wood logged every year which is
approximately 10 mm3 year−1 (MCPFE 2007), 60% of which is used as firewood.
The remaining 40% is mainly used in the furniture sector and, in a small percentage,
in the construction sector.
Therefore, the forest sector in Italy surely constitutes a valuable resource for the
environment, for tourism and landscape, it reduces landslide and erosion hazards,
but the potential use of wood as structural and industry material cannot be neglected.
In this context using wood may create a clear benefit in terms of local economies
and employment, particularly for local and mountain communities, as well as an
incentive to increase forested areas and therefore safeguard the local environment
for future generations (Paletto et al. 2017).
Fig. 3 Europe production sheets based panels: plywood (Source Eurostat)
However, until now most of the wood used for timber construction in Italy is
imported from Central and Northern European countries, especially Austria and
Germany, but in the recent years an important role is played also by Croatia with
high economic and environmental costs due to transportation.
Further opportunities for a wise use of forest resources derive from the biodi-
versity concept applied to wood science and technology. There is the chance to use
a larger variety of forest tree species as compared to those presently used. There
are 50,000 tree species on the planet, but only few of them are characterized for
their wood, particularly looking at the industrial potential utilization. For example,
in French Guyana, where there are no less than 1600 tropical species, only 250
have been characterized for their mechanic properties, and 150 for their resistance
to fungi. In France, where there are about one hundred species, only oak is well
documented (Cailloce 2016). In Italy most of wood research activity was conducted
on conifers, although most of Central-Southern Italy’s forests are composed of hard-
woods species.
However, the opportunity to use a wider range of tree species should not neglect the
deforestation issue which is still a dramatic problem especially in tropical countries
where there was a net forest loss of 7 Mha per year in 2000–2010 period while
net forest losses in 2010–2015 (most of which were natural forests) were partially
offset by a combination of natural expansion, often on abandoned agricultural land
Fig. 4 Europe production of particle based panels: particleboards (Source Eurostat)
(2.2 Mha per year), and the establishment of planted forests (3.1 Mha per year).
Also, the issue of planted forests with very high productivity is an important aspect
to reduce the pressure on natural forests around the world. However, governance
of the forest sector at local and global scale is a key issue to be developed further
in the future as well as the coordination of forest policy at international level and
geographical sustainability of forest management.
3 Forest Certification
The fulfilment of world demand on timber and assurance of forest conservation are
not conflicting issues, if forest management is implemented in a sustainable way.
Sustainable forest management (SFM) is the management of forests according to the
principles of sustainable development. Sustainable forest management has to keep
the balance between three main pillars: ecological, economic and socio-cultural. Suc-
cessfully achieving sustainable forest management will provide integrated benefits
to all, ranging from safeguarding local livelihoods to protecting the biodiversity and
ecosystems provided by forests, reducing rural poverty and mitigating some of the
effects of climate change.
Fig. 5 Europe production of particle based panels: OSB (Source Eurostat)
A definition of SFM was developed by the Ministerial Conference on the Protec-

tion of Forests in Europe (FOREST EUROPE), and has since been adopted by the
Food and Agriculture Organization (FAO). It defines sustainable forest management
as: “The stewardship and use of forests and forest lands in a way, and at a rate, that
maintains their biodiversity, productivity, regeneration capacity, vitality and their
potential to fulfil, now and in the future, relevant ecological, economic and social
functions, at local, national, and global levels, and that does not cause damage to
other ecosystems.”
In simpler terms, the concept can be described as the attainment of balance—bal-
ance between society’s increasing demands for forest products and benefits, and the
preservation of forest health and diversity. This balance is critical to the survival of
forests, and to the prosperity of forest-dependent communities.
A number of sets of criteria and indicators have since been developed to evaluate
the achievement of SFM at the global, regional, country and management unit level.
Criteria and indicators are tools which can be used to conceptualise, evaluate
and implement sustainable forest management. Criteria define and characterize the
essential elements, as well as a set of conditions or processes, by which sustainable
forest management may be assessed. Periodically measured indicators reveal the
direction of change with respect to each criterion.
Fig. 6 Europe production of fiber based panels: medium density fiber board
Criteria and indicators of sustainable forest management are widely used and many
countries produce national reports that assess their progress toward sustainable forest
management. Principles, Criteria and Indicators also form the basis of third-party
forest certification systems such as PEFCTM and FSC®.
3.1 Forest Certification, a Tool for Local and Sustainable

Forest Supply Chain
Growing environmental awareness and consumer demand for more socially responsi-
ble businesses helped third-party forest certification emerge in the 1990s as a credible
tool for communicating the environmental and social performance of forest opera-
tions. In details, forest certification systems work throughout the entire forest supply
chain to promote good practice in the forest and to ensure that timber and non-timber
forest products are produced with respect for the highest ecological, social and eth-
ical standards. Due to eco-labelling, different agreement to the specific certification
standard, customers and consumers are able to identify products from sustainably
managed forests. Therefore, forest certification offers assurances about the sustain-
Fig. 7 Europe production of fiber based panels: other fiberboard
ability of forest based products to customers throughout the supply chain, from forest
to the final user.
In addition, forest certification can be considered as a tool for fighting the illegal
logging and the illegal timber trade, introducing into the market timber from legal ori-
gin: according to the European Parliament, about the 20% of the total wood imported
into Europe is illegal, and intervention on the already existing European normative
is needed. From 2013, the European Union Timber Regulation (EUTR) has been a
useful tool to combat the phenomenon, but it appears that this is not enough. Forest
certification verifies that forests are well-managed—as defined by a particular stan-
dard—and chain-of-custody certification tracks wood and paper products from the
legal and sustainable forest through processing to the point of sale.
According to FAO, there are more than fifty national certification standards world-
wide, addressing the diversity of forest types and tenures. Globally, there are only two
main international certification programs, namely Programme for the Endorsement
of Forest Certification (PEFCTM ) and Forest Stewardship Council (FSC® ).
In July 2018, PEFC was the world’s largest forest certification system, with
308 Mha (two-thirds of the total global certified area certified to its sustainability
benchmarks), while FSC certified forests were around 192 Mha. In total, 431.4 Mha
of forests were certified to either PEFC or FSC, of which 71.1 Mha (or 16.5%) were
double certified.
Fig. 8 Europe production of fiber based panels: hardboard
4 Solid Wood and Wood Base Composites: A Smart

Overview
Currently, the term composite is being used to describe any wood material adhesive-
bonded together. This product mix ranges from fiberboard to laminated beams and
components. In Fig. 9 there is the logical basis for classifying wood composites
according to the size and shape of the raw material. The wood species used for
wood based composite sometimes are related to their availability in the territory, and
by the highest know how of the industries. The final end-use will distinguish if the
product is for structural or not structural purposes. According to the used raw material
(veneers, flakes, chips, powder) some remarks can be focused on the traditional and
more recent wood-based composites.
Among the solid wood composites, besides glulam, an important slice of the
market has been occupied by the innovative CLT panels, which represent a new
opportunity for the use of locally-grown timber in the construction sector. CLT
panels are structural elements which can then be used for walls, floors and
roofs of buildings, prefabricated by gluing cross layers of small timber boards,
with a high strength and stiffness and good overall dimensional stability. Due
to the system effect and the large size of the panels, the presence of defects
is less critical compared to other wood-based products such as sawn timber
Table 2 Roundwood production of European Countries and the most important producers (Eurostat
2016 modified)
Country Total Fuelwood Roundwood for
industrial purposes
1000 m3 under bark
Bulgaria 6410 2298 3481
Czech Republic (1) 16,163 2336 13,827
Denmark 3483 2015 1468
Germany 52,194 9413 42,789
Estonia 9735 3161 6574
Ireland 3050 316 2734
Greece 1217 894 323
Spain 17,848 4523 13,325
France 51,131 25,859 25,272
Croatia 5165 1768 3397
Italy (1) 5052 3004 2048
Latvia 12,651 1300 11,351
Lithuania 6747 2085 4662
Hungary 5586 2363 2950
Netherlands 2271 1397 874
Austria 16,763 4590 12,173
Poland 42,137 5295 36,481
Portugal 11,985 1090 10,895
Romania 15,117 5164 9953
Slovenia 5381 1272 4110
Slovakia 9267 515 8752
Finland 61,434 7107 54,327
Sweden 80,959 7214 73,745
Liechtenstein 8 4 4
Norway 12,028 1724 10,304
Switzerland 4263 1591 2672
Turkey (2) 22,835 4300 18,535
(1) Data referred to 2015 (2) data referred to 2014. Source Eurostat (online data codes for_remov)
and glue-laminated timber, therefore low-quality timber can be used. CLT is there-
fore particularly suited for regions where, due to geographical and climatic reasons,
the size and quality of the trees may be lower compared to Northern Europe.
Plywood is by far one of the oldest wood-composite composed by veneers. The
two general classes of plywood are (a) construction and industrial plywood and (b)
hardwood and decorative plywood, the two classes are covered by separate tech-
nical standards. Construction and industrial plywood has traditionally been made
Table 3 Sawnwood production 2000–2016 (1000 m3 )

Country 2000 2005 2010 2012 2013 2014 2015 2016
EU-28 100,706 108,706 100,815 98,948 99,691 103,705 104,094 105,994
Bulgaria 312 569 554 698 803 838 938
Czech Republic 4106 4003 4744 4259 4037 3861 4150
Denmark 364 196 448 358 358 358
Germany 16,340 21,931 22,059 21,081 21,459 21,772 21,490 22,200
Estonia 1436 2063 1771 1491 1558 1554 177 200
Spain 3760 3660 2038 1971 1047 2245 1691 1737
France 10,536 9715 8316 8067 7901 7697 7626 7742
Croatia 642 624 677 651 1192 1294 1488 1434
Italy 1630 1590 1200 1370 1360 1430 1470 –
Latvia 3900 4227 3150 3316 3367 36,757 3479 3903
Lithuania 1300 14,445 1272 1150 1120 1345 1248 1406
Hungary 291 215 133 302 109 121 488 518
Netherlands 389 279 231 190 211 228 185 184
Austria 10,390 11,074 9603 8952 8534 8460 8731 9216
Poland 4262 3360 4220 4249 4321 4725 4835 4911
Portugal 1427 1010 1045 1097 854 1035 1156 1239
Romania 3396 4321 4323 5500 5532 6019 5936 5449
Slovenia 439 527 760 660 660 700 725 730
Slovakia 1265 2621 2576 1560 1430 1750 1600 1580
Finland 13,420 12,269 9473 9440 10,440 10,920 10,640 11,420
Sweden 16,176 17,600 16,750 16,492 16,074 17,500 18,174
UK 2622 2780 3101 3409 3581 3764 3493 3671
Norway 2280 2326 2118 2289 2206 2407 2444 2533
Switzerland 1625 1591 1457 1135 1044 1140 1122
Turkey 5528 6445 6243 6682 6405 6635
Brazil 21,300 23,557 17,452 15,167 15,397 15,397
Canada 50,465 60,187 38,667 40,564 42,813 43,351
China 6675 17,960 37,231 55,740 63,040 68,440
India 7900 14,789 6889
Indonesia 6500 4330 4169
Russia 20,000 23,913 28,870 32,230 33,500 33,900
USA 91,076 97,020 60,013 67,474 71,115 74,803
Source Eurostat (online data code for_swpan)
Table 4 Main indicators for wood-based industries, EU-28, 2005 and 2015 (Eurostat, modified)
Activity Number of Gross value added cost Number —1000
enterprises—1000 (billion EUR)
2005 2015 2005 2015 2005 2015
Manufacturing 2183 (e) 2097 1668 (e) 1900 (du) 32,032 (e) 29,986
Wood-based 470 420 153 139 3959 3325
industries
Manufacture of 188 170 36 (e) 33 1105 976
wood and wood
products
Manufacture of 20 19 (de) 40 46 730 (e) 644
pulp, paper and
paper products
Printing and 133 110 (du) 41 30 978 706
service activities
related to
printing
Manufacture of 130 (eu) 120 (e) 36 (e) 29 1147 (e) 1000 (u)
furniture
In 2005 it is EU 27 countries. e estimated, u low reliability
Fig. 9 The logical basic wood element for composites (Marra 1979 from Youngquist 1990)
from softwoods, however, the current standard lists a large number of hardwoods as
qualifying for use (Youngquist 1999). At the same time, the technical standard in
use for hardwood and decorative plywood cover certain decorative softwood species
for non-construction use. The species suitable for plywood are classified according
also to their strength (Youngquist 1999) among the conifers the most widespread
are cedar, American cypress, Douglas-fir, fir, hemlock, pines, larch, spruce with a
large variety of Pinus species (mainly American pines). Among hardwoods poplar,
birch, maple, are very represented. Eucalyptus is suitable and tropical species like
lauan and meranti, and rosewood are used mainly as finishing layer. Okoumé must
be cited because in the last times it is quite used for plywood due to the excellent
gluing, workability and finishing, and it is suitable and used for maritime panels;
the species is harvested in Gabon, in Equatorial Guinea and in Congo. Plywood

production process is quite in evolution due to the chance to use special bonding
for external uses, and the possibility to use more species than in the past. Beech
for LVL is just an example but also some conifers like maritime pine could have
a more abundant use because the species could be much more exploited especially
in Mediterranean regions to obtain plywood; after all this lasts species has proved
to be suitable for plywood products since the 1970s. The use of poplar cannot be
considered so obvious at least in some countries, in Italy the last CGA (Censimento
Generale dell’Agricoltura) has shown from 2000 to 2010 a strong reduction in the
surface which decreased from 126,697 to 101,628 ha.
Particle based wood products have the big advantage that can be realised for
almost any type of wood also from less known tropical species (Karlinasari et al.
2012). The most used species are fir and spruce, even pruning could find a reasonable
use for this production (Nasser et al. 2014). The raw material could be also derived
also from forest and wood industries residues, buildings demolition, and agricultural
crops (Nishimura 2015), as with the residues of wheat and rice crops due to their
lighter bulk density. Particle based products need adhesives for which the most used
are urea phenol and melamine formaldehyde; therefore, the main issue for this type
of products relates to VOC emission due also to the high amount of adhesive which
need to be used. Particle based wood products are still the most important wood
composite which also offer a chance to use recycled wood.
OSB is a particle-based product and it has been developed as an alternative to
plywood in order to overcome the declining availability of large diameter logs. The
main advantage of OSB compared to plywood is that it can be produced from small
diameter logs and it can be made from both hardwood (poplar and aspen) and soft-
wood (pine, spruce, douglas fir) timber; however, juvenile wood could affect the
quality of OSB.
Oriented strands lumber (OSL) and laminated strands lumber (LSL) offer a chance
to use wood of low quality by optimizing the strand thickness, strand length, slen-
derness ratio and specific surface area, in the future they could be a valid product to
be exploited in short value chains.
In theory each lignocellulosic material could be used for fiber board panel, but it
has been shown that fiber characters as length and diameter greatly affect the final
quality of the panel (Dae Park and Kim 2001) and it depends also on the desired final
density. Hardboard such as MDF preferably should be made with conifers due to their
longer fibers; possibly, conifer fibers are mixed with hardwoods. Also recycled paper
can be used in fiber based panels. Among the most used species there are radiata pines,
even mixed with eucalypts wood, but also poplar, robinia and black pine are suitable
(Alpar et al. 2010); in Italy there was an interesting example of a fiberboard produced
with chestnut mixed with conifers fibers (Fig. 10). Interesting results are also with
fibers of non-wood material like bamboo and rice straw (Hiziroglu et al. 2007). Agri-
cultural residues and crops like hemp, wheat straw, rice straw, jute, coconut fiber,
oil palm, bagasse and bamboo agro-based oil palm biomasses (including the trunk,
frond, empty fruit bunch, and palm) could increasingly be considered in the future for
particle and fiber-based panels especially for non-structural purposes. Fiber-boards
Fig. 10 MDF produced using also local species like chestnut
have the large advantage to be utilisable also for shaping complicated and curved pro-
files by machines, the only one big disadvantage being the large amount of adhesives
to be used.
Fiber-based products have a higher degree of moldability and tree bark can be
considered also as raw material. It contains much shorter cells and more heteroge-
neous chemicals than wood, with a potential use as for example tannins. Tannins
have an interesting utilization in bark-board production which is a special particle
board (Yemele et al. 2008) and as filler material in other biocomposites, due to their
hydrophobic characteristics that make wood less prone to biotic attack; however,
there could be problems with compatibility with the other wood particles.
Moreover, wood plastic composite (WPC) and natural fiber composite (NFC)
occupy an important role among innovative materials, as with them it is quite easy to
obtain daring and novel shapes in the artefacts; in fact, especially in automotive sector
they have been extensively used. The raw material for WPC is mainly wood flour
coming from conifers even if some hardwoods like beech have proved to be suitable.
As for the natural fiber composites, they can be derived from many agricultural
wastes like wheat straw, olive stones, hazelnut shells, coconut shells, cork bark,
rice husk, corn stalks, corn cobs, sugarcane bagasse, chicken feathers, cottonseed
(Spear et al. 2015). The list of woody and herbaceous crops which could be used for
these composite products is increasing and the compatibility with plastic like HDPE,
LDPE, PVC, PP, ABS, PS, and PLA sometimes needs a compatibilizer and this is a
huge aspect for improving environmental sustainability.
The potential for innovative industrial products and processes is directly linked to
the 3D-printing technology, in order to implement final products designed to better
address the biomimetic principles; those products could be based on wood and non-
wood mixed flour, using also natural binders even based on lignin (Keplinger et al.
2018). Fusion deposition modelling, based on wood flour mixture with bio-polymers
is also a quite novel technology (Le Duigou et al. 2016), as well as new functional
materials obtained from densified cellulose, after extracting lignin and densification
of wood (Graichen et al. 2017), or micro and nano-cellulose compounds and lignin
extracted from forest and agricultural wastes. The application of circular, green econ-
omy concept should allow a high resource availability, but the sustainability of the
production process and the compatibility with other chemical compounds are aspects
which must be carefully considered.
5 Improving Performances: Gluing and Wood Treatments
Increasing wood durability and the use of eco-friendly gluing are major challenges
for the wood industry which could take advantage of the short value chain approach.
In general, since a large group of classical biocides are banned due to environmental
and health concerns, and in response to negative public opinion on existing bio-
cides and exploitation of tropical timber, research strategies are changing. Focus is
on improving the durability and prolonging the service life of wood (Humar et al.
2017). Recent models clearly indicate that the service life of wood in aboveground
applications is a function of inherent durability, as a result of the presence of biocides
and/or biologically active extractives, and of water exclusion efficacy (Meyer-Veltrup
et al. 2017) of wood species. Industry in Europe is therefore looking for less utilized
tree species whose wood properties may exhibit good water exclusion efficacy, due
either to the anatomical structure or to the chemical composition. One of the over-
looked forest tree species is Pinus pinea (De Angelis et al. 2018), which showed
excellent results related to its natural durability and the chance to be treated by the
most common and less toxic biocides.
There is a lot of research on chemical compounds which can improve wood dura-
bility (for a complete review cf. Schultz et al. 2008), but from the industrial point
of view the most used active compounds, alone ore mixed, are mainly Cu-based,
Boron-based and different organic compounds; most of those chemicals and active
substances are derived from agricultural practices. Some of those chemical treatments
are used also in biocomposites (i.e. Boron-based), but lately wood modification is
one of the most important challenges to be applied also in wood based composites.
Among the techniques of wood modification, heat treatment is considered an eco-
friendly process, aimed at upgrading low-quality wood (Romagnoli et al. 2015a, b,
2018), and it has proved to be a fine reasonable perspective both in plywood and in
OSB (Okino et al. 2007). The overall extent of modification and the final result of
the process is quite variable while the state parameters of the process (temperature,
pressure and application time interval) determine the final result. Applying heat treat-
ment to composite products is highly problematic, because heat treatment reduces
bondability; however, researches on cross-laminated timber have been carried out and
thermally modified wood as decorative face-layer of plywood and laminated flooring
are already in the market. Heat treated glulam for windows is a quite promising prod-
uct and no large effects seem to occur on OSB panels (Spear 2015); test on poplar
plywood seems to produce interesting results (Goli et al. 2014), and heat treated
wood powder as byproduct seems to increase some performances of WPC (Kallakas
et al. 2016) with some criticisms as for the mechanical strength. Acetylated modified
wood may be glued depending by the condition of wood assortment and the type of
adhesive (Frihart et al. 2017).
Gluing is the other very big challenge towards sustainability of wood-based prod-
ucts and references report a lot of applications to replace or at least to reduce the fossil
based, not eco-friendly adhesives, in primis urea formaldehyde. Just to mention a
not exhaustive list of the possible natural compounds which have been tested, there
are reports on the application of tannin based adhesives mainly on particleboards
(Spina et al. 2013); lignin adhesives coupled to synthetic resin and with not synthetic
resins (Mittal and Sharma 1992); protein-based adhesives; carbohydrate adhesives;
adhesives from unsaturated natural oils; cottonseed meal (Liu et al. 2018); cellulose
nanofibrils (Jiang et al. 2018). A very original perspective belongs to welding wood
where bonding is obtained without adhesives, but simply using high speed vibra-
tion which makes intercellular wood lignin like a thermoplastic-type binder (Pizzi
2016), In conclusion, a lot of efforts are still needed to make the use of natural
glue widely applicable in many biocomposite especially considering the possible
structural applications.
6 Possibility of Using Short Chain Species for Structural

Purposes. Mechanical Characterization, Modelling
and Italian Case Studies
The possibility of developing a short procurement chain of wood for structural use
was demonstrated from the results of different research projects recently conducted in
different Italian regions. In 2009, a project was carried out in Tuscany, the first Italian
region for forested area with more than 1 million ha of forests. The project started
with the selection of the wood species best suited for structural purposes (locally-
grown Douglas fir trees), then continued with the mechanical characterization of
timber boards, and with the derivation of rules for visual grading of Douglas fir. In
2010, thanks to some European and regional grants, a pilot plant was implemented
in Tuscany for the production of CLT structural panels made with locally grown
Douglas Fir timber. Finally, in 2011 a two-storey public building was constructed in
Florence using the same panels (Berti et al. 2009 ).
A similar project was completed in 2013 in Sardinia, a region with approximately

600,000 ha of forest. Again visual and machine grading rules for local Maritime
Pine wood were derived, and CLT panels made of maritime pine were produced and
tested to destruction, leading to a full mechanical characterization (Fragiacomo et al.
2015; Concu et al. 2016).
Again in the 2013–2015 period a project on the glulam for structural purposes
of chestnut was conducted (Romagnoli et al. 2015a, b; Brunetti et al. 2015), and
a prototype of chestnut glulam was built in the Latium region. Other important
projects have been carried out in central-southern Italian by ALForLab (http://www.
alforlab.it/) and the national project PRIN 2015 where species like Pinus laricio as
well as beech and chestnut were used to develop short supply chain in the wood-
biomass sector for traceability, certification and carbon sequestration, as well as to
build structural components as CLT, glulam and bio-based varnishes by lignin (Zikeli
et al. 2018).
6.1 Technological Potentials and Technical Regulations

for Some Type of Products, Especially for Structural
Purposes
When used for structural purposes, timber shows additional advantages over con-
crete, including: (i) high strength-to-specific weight ratio, which results in lightweight
structures, reduced impacts from earthquakes, and reduced imposed load on founda-
tion; (ii) rapidity of construction: the entire construction process is ‘dry’ and thus does
not require propping and formwork as there is no concrete pouring on the building
site; and (iii) possibility of extensive prefabrication: most of the construction work
can be moved from the building site to fabrication plants where it is carried out in
controlled factory conditions, with extensive use of mechanisation, improved quality
control, and reduction of hazardous work operations on the building site.
There is a long tradition of using timber in Europe, even though in the previ-
ous century the use of timber was limited to small houses and low-rise residential
buildings mostly in Northern and Central Europe. In the new millennium, however,
enormous progresses have been made in timber engineering. New wood-based prod-
ucts such as cross-laminated timber (CLT) and laminated veneer lumber (LVL) have
been invented and have markedly improved the material behaviour by reducing the
influence of knots, grain deviation and other defects on the strength, reducing the
anisotropy, and limiting the shrinkage/swelling processes (Schickhofer et al. 2010).
With these new products used to produce mass timber panels, new frontiers were
quickly reached in civil engineering. Smith and Frangi (2008) reported that it is pos-
sible to construct multi-storey timber buildings up to 10 floors with excellent fire
protection and resistance to lateral loads. The growing diffusion of timber structural
systems in Italy is also due to the improvements gained in the automation process
and performance of Computer Numerical Control (CNC) machinery, as well as in
the advance reached in the glueing process of wood-based products, which led to the
development of new wood-based products like the cross-laminated timber (CLT). The
introduction of new types of mechanical fasteners, especially self-drilling screws,
have greatly enhanced the possibility of prefabrication of structural components and
made the construction process easier and faster.
At the same time, important research projects on multi-storey timber buildings
have recently been completed. The Sofie project undertaken in Italy (Ceccotti and
Follesa 2006) has explored the feasibility of residential buildings made of CLT pan-
els. Particular emphasis was placed on the seismic and fire resistance, which were
investigated via full-scale shaking table tests of 3- and 7-storey buildings (Ceccotti
et al. 2013) and a fire tests of a 3-storey building. Several cyclic (Gavric et al. 2015)
and fire (Fragiacomo et al. 2013) tests of wall and floor panel components were
also carried out, demonstrating the excellent seismic and fire performance of these
buildings. The SERIES international project (2010–2013) focused on the seismic
performance of multi-storey CLT, Log House and Light Frame buildings (Tomasi
and Piazza 2013).
All these examples, together with the great speed of construction due to an almost
completely dry construction process, can explain why in Italy the use of timber struc-
tural systems is steadily increasing especially for the construction of new residential
and school buildings with respect to traditional construction systems such as rein-
forced concrete, masonry and steel structures. This resulted in the construction of
several multi-storey buildings ranging from 4 to 9 storeys in different Italian regions
with different levels of seismicity. The tallest buildings were erected in Milano, where
four 9-storeys CLT residential buildings were built in 2012.
According to Centro Studi Federlegno Arredo Eventi SpA (2015), in 2015 Italy
was the 4th European Country for the construction of residential timber buildings
with a share of 7% of the whole residential building market.
Considering the outcomes of the above-mentioned research projects and expe-
riences, the development of a short supply chain of wood for structural purposes
should therefore include the following steps:
1. Selection and sampling of local species suitable for structural purposes from
local forests and/or plantations.
2. Conducting first preliminary grading in order to identify the most suitable trees
on selected standing trees, based on measurements taken at different heights.
3. Harvesting and assortment of the logs based on size and external defects.
4. Cutting of the logs in a local sawmill, selecting different board sizes for the
visual and machine grading process and for the production of CLT panels and
glued-laminated (GLT) beams.
5. First machine grading of boards using non-destructive measurements of the
Modulus of Elasticity with acoustic wave velocities measurements.
6. Drying process of the boards.
7. Non-destructive measurements of their physical properties (density, humidity,
defects etc.) in order to calibrate the machine grading.
8. Measurements of the size and position of the main visual defects such as knot
diameters and positions, annual rings thickness, grain deviations.
9. Failure tests of the boards in order to measure their strength and correlate these
values to the size and position of visual defects measured in 8.
10. Derivation of visual and machine grading rules based on results of steps 7–9.
11. Selection of boards for the production of CLT panels and GLT beams, based on
the grading rules derived in 10.
12. Production of CLT and GLT test specimens.
13. Failure tests of CLT panels and GLT beams according to the specification of
the European Standard for CLT EN 16351 and for GLT EN 14080.
14. Derivation of the mechanical properties of the CLT panels and GLT beams.
15. Analytical simulations of the mechanical behavior of CLT panels and GLT
beams with the most common analytical methods (Modified Gamma Method
and Shear Analogy Method for CLT, transformed section method for GLT).
16. FE modelling of CLT panels and GLT beams and numerical simulations of the
mechanical behavior of CLT panels and GLT beams.
17. Comparison of the results of step 15 and 16 with the results of step 13 and
calibration of analytical formulas in step 15 for use in structural design of CLT
and GLT made of locally-sourced timber.
And possibly,
18. Construction of a pilot plant for the production of CLT and GLT.
19. Construction of a building with CLT panels and GLT beams produced with
locally grown timber.
20. Physical and structural tests on the building.
It should be pointed out however that a strong and firm support by the local public
authorities and national government is crucial in order to establish a short supply
chain of timber that can be economically competitive with the strong and more
organized international producers. Within the framework of European regulations,
local and national authorities should promote the use of locally-grown timber, with
tangible benefits for the users complying with those guidelines.
7 Conclusion
It is quite evident that wood-composites is a field of research and knowledge trans-

fer in rapid evolution. Even if some composites can now be considered as already
mature products, i.e. the first rough plywood can be traced up to the Egyptian times,
they do not only still hold much interest for wood industry and technology but could
be approached with a new perspective, for example concerning the wood species to
be employed for their production. Short supply chain and exploitation of different
species is one of the most important potential development in order to decrease trans-
port effect on carbon footprint of the biocomposite products, consequently decreas-
ing also the amount of wood import and the risk of deforestation while reducing
the pressure on the most utilized tree species in composite like spruce, in Europe.
Biocomposites are the products of the future, because they are a one of the ground-
work of the circular economy approach to material application. From a technical
point of view, wood-based composites or in general bio-based composites, could be
dealt with at a finer level of fiber, particle, flake, or as veneer. Properties of such
materials can be changed by combining, reorganizing, or stratifying the fibers and
particles. Both wood-based composites and natural fiber composites have much to
offer also to the design sector for their high content of renewable fibers and poten-
tially of polymers. Designers can use semi-structural elements to utilize agricultural
residues where timber resources are scarce or are under strong competition with the
traditional use in construction, furniture and panel products application.
Acknowledgements Grant: MIUR PRIN 2015. Prot. 2015YW8JWA. La filiera corta nel settore
biomasse-legno: approvvigionamento, tracciabilità, certificazione e sequestro di Carbonio. Inno-
vazioni per la bioedilizia e l’efficienza energetica (led by Giuseppe Scarascia Mugnozza). The
research work was partially supported also by the “Departments of Excellence-2018” Program
(Dipartimenti di Eccellenza) of the Italian Ministry of Education, University and Research, DIBAF-
Department of University of Tuscia, Project “Landscape 4.0—food, wellbeing and environment”
project of excellence MIUR “Landscape 4.0” DIBAF.
References
Alpar T, Fàczàn T, Ràcz I, Kàtoli G (2010) MDF/HDF production from plantation wood species.
Drvna Industrija 61(3):183–191-
Arwinfar F, Hosseinhashemi SK, Latibari AJ, Lashgari A, Ayrilmis N (2016) Mechanical proper-
ties and morphology of wood plastic composites produced with thermally treated beech wood.
BioResources 11:1
Berti S, Brunetti M, Capone P, Ciapini E, Fedrigo C, Follesa M, Lauriola MP, Lavisci P, Macchioni
N, Palanga G, Palanti S, Pizzo B, Terranova M, Vasta S, Vignoli A (2009) Linee guida per l’edilizia
in legno in Toscana – Coordinamento Editoriale: Maurizio Follesa e Marco Pio Lauriola, Regione
Toscana, Giunta -Regionale, Direzione Generale della Presidenza, 2009 (in Italian)
Brunetti M, Silvestri A, Nocetti M, Burato P, Portoghesi L, Carbone F (2015) Travi lamellari in
castagno. Innovazione di prodotto nella filiera del legno per uso strutturale. Sherwood 215:31–35
Cailloce L (2016) Wood material for the future? CNRS News, 08 Aug 2016
Ceccotti A, Follesa M (2006) Seismic behaviour of multi-storey X-lam buildings. In: Proceedings of
426 COST E29 international workshop on earthquake engineering on timber structures, Coimbra,
Portugal, pp 81–95
Ceccotti A, Sandhaas C, Okabe M, Yasumura M, Minowa C, Kawai N (2013) SOFIE project—3D
shaking table test on a seven-storey full-scale cross-laminated building. Earthquake Eng Struct
Dyn 42(13):2003–2021
Centro Studi Federlegno Arredo Eventi SpA. Rapporto case ed edifici in legno 2015. Assolegno-
Federlegno Arredo, MADE Expo, proHolz; 2015
Concu G, De Nicolo B, Fragiacomo M, Trulli N, Valdes M (2016) Grading of maritime pine from
Sardinia (Italy) for use in cross laminated timber. In: Construction materials—proceedings of the
institutions of civil engineers (in press). https://doi.org/10.1680/jcoma.16.00043
De Angelis M, Romagnoli M, Viljem V., Poljanšek I, Oven P, Thaler N, Lesar B, Kržišnik D, Humar
M (2018) Chemical composition and performance of Italian Stone pine (Pinus pinea L.) wood
against fungal decay and water. Ind Crops Prod 117:187–196. https://doi.org/10.1016/j.indcrop.
2018.03.016
Dae Park B, Kim YS (2001) Effect of fiber characteristic on medium density fiberboard (MDF)
performances. J Korean Wood Sci Technol 29(3):27–35
EN 16351 (2015) Timber structures—cross laminated timber—requirements. CEN; 2015
EN 14080 (2013) Timber structures. Glued laminated timber and glued solid timber. Requirements.
CEN; 2013
FAO (2018) State of world forests 2018, Rome
Frihart R, Brandon R, Beecher JF, Ibach RE (2017) Adhesives for achieving durable bonds with
acetylated wood. Polymers 9:12
Fragiacomo M, Riu R, Scotti R (2015) Can structural timber foster short procurement chains within
Mediterranean forests? A research case in Sardinia. South-East Eur For J 6(1):11 pp. http://dx.
doi.org/10.15177/seefor.15-09
Fragiacomo M, Menis A, Clemente I, Bochicchio G, Ceccotti A (2013) Fire resistance of cross-
laminated timber panels loaded out-of-plane. ASCE J Struct Eng 139(12):11 pp, 04013018.
https://doi.org/10.1061/(asce)st.1943-541x.0000787
Gavric I, Fragiacomo M, Ceccotti A (2015) Cyclic behavior of cross-laminated timber (CLT) wall
systems: experimental tests and analytical prediction models. ASCE J Struct Eng 141(11):14 pp,
04015034. https://doi.org/10.1061/(asce)st.1943-541x.0001246
Goli G, Cremonini C, Negro F, Zanuttini R, Fioravanti M (2014) Physical-mechanical properties
and bonding quality of heat treated poplar (I-214 clone) and ceiba plywood. Iforest 8:687–692
Kallakas H, Martin M, Goljandin D, Poltimae T, Krumme A, Kers J (2016) Mechanical and physical
properties of thermally modified wood flour reinforced polypropylene composites. Agron Res
14:994–1003
Karlinasari L, Hermawan D, Maddu A, Sudo Hadi Y (2012) Development of particleboard from
tropical fast-grown species for acoustic panel. J Trop For Sci 24(1):64–69
Keplinger T, Frey M, Burgert I (2018) Versatile strategies for the development of wood-based
functional materials. In: Proceedings of SPIE—the international society for optical engineering,
article 1059313
Hiziroglu S, Bauchongkol P, Fueangvivat V, Soontonbura W, Jarusombuti S (2007) Selected prop-
erties of medium density fiberboard (MDF) panels made from bamboo and rice straw. For Prod
J 57(6):46–50
Humar M, Kržišnik D, Lesar B, Ugovšek A, Rep G, Šubic C, Thaler N, Ugovšek A, Zupančič
K, Žlahtič M (2017) Thermal modification of wax-impregnated wood to enhance its physical,
mechanical, and biological properties. Holzforschung 71:57–64
INFC. Inventario Nazionale delle Foreste e dei serbatoi forestali di Carbonio. I caratteri quantitativi
- MIPAAF Corpo Forestale dello Stato, CRA-MPF, Trento, 2007 (in Italian)
Jiang W, Tomppo L, Pakarin T, Sirvio JA, Liimatainen H, Haapala A (2018) Effect of cellulose
nanofibrils on the bond strength of polyvinyl acetate and starch adhesives for wood. BioResources
13(2):2283–2292
Matteucci G, Cammarano M, Dezi S, Mancini M, Scarascia-Mugnozza G, Magnani F (2011)
Climate change impacts on forests and forest products in the Mediterranean area. In: Nararra
A, Tubiana L (eds) Regional assessment of climate change in the Mediterranean-synthesis and
the assessment of adaptation measures. Springer, Netherlands
Meyer-Veltrup L, Brischke C, Alfredsen G, Humar M, Flaete PO, Isaksson T, Larsson Brelid P,
Westin M, Jermer J (2017) The combined effect of wetting ability and durability on outdoor
performance of wood: development and verification of a new prediction approach. Wood Sci
Technol 51(3):615–637
Mittal, Sharma (1992) Studies on lignin-based adhesives for plywood panels. Polym Int 29(1):7–8
Le Duigou A, Castro M, Bevan R, Martin N (2016) 3D printing of wood fiber biocomposites: from
mechanical to actuation functionality. Mater Des 96:106–114
MCPFE (2007) State of Europe’s forests 2007. MCPFE Liaison Unit, Warsaw
Nasser RA, Al-Mefarrej HA, Abdel-Aal MA, Alshahrani TS (2014) Effects of tree species and
wood particle size on the properties of cement-bonded particleboard manufacturing from tree
prunings. J Environ Biol. 35(5):961–971
Nishimura T (2015) Chipboard oriented strand board (OSB) and structural composite lumber. In:
Wood composites, Woodhead Publishing series in Composites science and engineering, vol 54,
pp 103–119
Okino E, Teixeira D, Del Menezzi C (2007) Post-thermal treatment od oriented strandboard (OSB)
made from cypress (Cupressus glauca Lam.). Maderas Ciencia y tecnologia 9(3):199–210
Pizzi A (2016) Natural adhesives binders and matrices for wood and fiber composites. Chemistry
and technology. In: Lignocellulosic, fiber and wood handbook renewable materials for today’s
environment, pp 277–303
Romagnoli M, Cavalli D, Pernarella R, Zanuttini R, Togni M (2015a) Physical and mechanical
characteristics of poor-quality wood after heat treatment. IForest 8:884–891 (online) (22 May
2015). http://www.sisef.it/iforest/contents/?id=ifor1229-007
Romagnoli M, Cecchini M, Carbone F, Portoghesi L, Mattioli W, Moroni S, Brunetti M, Nocetti M
(2015b) Innovazione di processo e di prodotto nella filiera del legno di castagno per uso strutturale:
lamellare e innalzamento delle classi di resistenza meccanica del legno massiccio. PSR Regione
Lazio Assessorato Agricoltura Finanziata con Reg. CE N.1698/05 PSR LAZIO 2007/2013 D.G.R.
76 DEL 18/02/2014 Misura 124 – Cooperazione per lo sviluppo di nuovi prodotti, processi e
tecnologie, nel settore agricolo, alimentare forestale – PROGETTO N°: 8475921170 – Provved-
imento di Concessione R.L. n. 57/124/10
Romagnoli M, Vinciguerra V, Silvestri A (2018) Heat treatment effect on lignin and carbohydrate in
Corsican pine earlywood and latewood studied by Py-GC-MS technique. J Wood Chem Technol
38:57–70 (Scopus cit. 0, WOS cit. 0). https://doi.org/10.1080/02773813.2017.1372479
Schickhofer G, Bogensperger T, Moosbrugger T (2010) BSPhandbuch – Holz-Massivbauweise in
Brettsperrholz. Technischen Universität Graz, Austria
Schultz P, Militz H, Freeman MH, Goodell B, Darrel N (2008) Development of commercial preser-
vatives, efficacy environmental and health issue. ACS Symposium Series 982
Smith I, Frangi A (2008) Overview of design issues for tall timber buildings. Struct Eng Int
18(2):141–147
Spear MJ, Edee A, Carus M (2015) Wood polymer composites. In: Ansell M (ed) Wood Composites,
Woodhead Publishing series in Composites and Engineering, vol 54, pp 195–240
Tomasi R, Piazza M (2013) Investigation of seismic performance of multi-storey timber build-
ings within the framework of the SERIES Project. In: International conference on structure and
architecture, Guimaraes [ICSA2013], Portugal
Yemele MC, Blanchet P, Cloutier A, Koubaa A (2008) Effects of bark content and particle geometry
on the physical and mechanical properties of particleboard made from black spruce and trembling
aspen bark. For Prod J 58(11):48–56
Youngquist J (1999) Wood-based composites and panel products. Forest Product Laboratory
Bamboo’s Bio-inspired Material Design
Through Additive Manufacturing
Technologies
Francesco Buonamici, Yary Volpe, Rocco Furferi, Monica Carfagni,

Giovanni Signorini, Giacomo Goli, Lapo Governi and Marco Fioravanti
Abstract Bamboo is one of the longest-used organic raw materials in the tropics
for a large number of different purposes in the daily lives of human beings. Because
of its excellent physical-mechanical properties, in many parts of the world it is
widely used as a structural material, especially for the construction of scaffolding
and the construction of buildings. The bamboo can be modelled as a composite
material, consisting of a of a parenchyma cells matrix, similar to a foam, reinforced
by bundles of fibers associated with vessels. The present work aimed to explore the
possibility to design a 3D printed biomimetic composite material able of keeping
advantages from the bamboo morphological structure. Samples made of PVA, ABS
and PVA + ABS were manufactured using Fused Deposition Modelling and tested
under compression and bending conditions. The behavior under compression has
shown to depend mainly on the material used while in bending the structure has
shown important effects leading the sample made of PVA + ABS to have the same
performances of much expensive pure ABS.
Keywords Additive manufacturing · Bamboo · Structure optimization
1 Introduction to Bamboo Anatomy and Mechanics
Some common features characterize every solution adopted by nature for designing
structural elements within living organisms: the optimization in the use of constitu-
tive materials, the lowest possible amount of energy used in their processing, and
mostly, the continuing assessment of the solution adopted and its relevant modifi-
cation through a process that normally took very long time. Being part of living
F. Buonamici · Y. Volpe · R. Furferi (B) · M. Carfagni · L. Governi

Department of Industrial Engineering (DIEF), University of Florence, Florence, Italy
e-mail: rocco.furferi@unifi.it
G. Signorini · G. Goli · M. Fioravanti
Department of Agricultural, Food and Forestry Systems (GESAAF),
University of Florence, Florence, Italy

810 F. Buonamici et al.
organisms, these structures can sometimes benefit of self-repairing mechanisms that

greatly increase their efficiency in service. Accordingly, many of them can be used
as source of inspiration for building new materials or structures. This is particularly
true for the natural Bamboo, whose utilization for designing structures is well estab-
lished. In fact, bamboo is one of the longest-used organic raw materials for a large
number of different purposes in the daily lives of human beings (Grosser and Liese
1971). Because of its excellent physical-mechanical properties, in many parts of the
world it is widely used as a structural material, especially for the construction of
scaffolding and the construction of buildings (Mannan et al. 2017).
The Bamboo are monocotyledonous, perennial, ubiquitous and evergreen herba-
ceous plants originating mainly from tropical and subtropical regions. They grows in
the temperate regions of Japan, China, Chile and the United States. They are natives
of all continents except Europe and the highest concentration of species occurs in
Asia, in the Indo-Burmese region, which is also considered their area of origin. Taxo-
nomically, the Bamboo belong to the family of Poaceae (Graminaceae), subfamily of
the Bambusoideae, that is articulated into 75 genera and about 1250 species (Mannan
et al. 2017).
Used as a construction material, bamboo has a significant economic impact;
according to the 2010 FAO forest resource assessment, there are 31.4 million hectares
of bamboo worldwide, 60% of which concentrated in the rapidly developing coun-
tries of China, India and Brazil (Dixon and Gibson 2014).
Bamboo, unlike trees, have only a primary growth (the stem developed only
in length), not followed by the secondary one, in thickness, as they do not show
the cambium (the meristematic tissue that produces the wooden tissue). The final
diameter of the stem will therefore be already determined at the initial stage of
the growth phase of the shoot, which translates into a thin and hollow stem inside
(Wang et al. 2014). This type of stem is called culm. The bamboo culm is roughly a
cylindrical shell, more or less thick wall, with large central cavities inside separated
by periodic transverse partitions, called diaphragms, whose purpose seems to be to
prevent the ovalization of the cross section when the culm it flexes. The structure of the
culm wall represents the most interesting aspect of Bamboo bio-mimetic mechanics.
It consists of a matrix of parenchymatous cells in which numerous vascular bundles
are immersed.
The shape, size, arrangement and number of vascular bundles determines the
transverse aspect of the culma wall according to Liese (1998). The presence of fibers
complete the anatomical structure of the culm, constituting the sclerenchymatous
tissue, which is the main element of mechanical support, and stiffening of the bam-
boo culm (see Fig. 1). The sclerenchymatous tissue comes in the form of sheaths
externally wrapping the vessels, the protoxylem and the phloem of each vascular
bundle. In some species, moreover, the fibres are also present as isolated beams.
In Bamboo, due to the lack of the vascular cambium and of a secondary growth in
thickness of the culms, which would allow geometric adaptations to the mechanical
loads, there is a need to optimize the structure at the level of the tissues (Wang et al.
2012). The plants, in fact, cannot use adaptive growth in the same way as a tree does;
in fact, they modify the geometry of the stem thus increasing its moment of inertia to
Bamboo’s Bio-inspired Material Design Through Additive … 811
Fig. 1 Transversal cross-section of bamboo with parenchyma cells vascular bundles
cope with the bending stresses caused, for example, by wind loads. Therefore the culm
tissues have an organization that is able to maximize the mechanical strengths. In this
perspective, the mechanical stability of the culms depends on the correct insertion of
the rigid sheaths of fibers of the vascular bundles in the soft and compressible matrix
of parenchyma, consisting of thin-walled cells that do not contribute to the stiffness
of the culms (Kanzawa et al. 2011).
In particular, the difference in frequency and shape of the vascular bundles in the
culm, especially in the transversal direction, makes the difference. In fact, within the
culm, the distribution and the shape of the vascular bundles is not uniform. In the
peripheral area of the culm wall the beams are smaller and more numerous, while in
the inner part of the wall they are larger and sparse. This marked radial gradient, with
densely packed bundles in the periphery of the culm, is due to the fact that the area of
the culm where the bending stresses are higher is the most external one (Amada et al.
1997). The greater quantity of vascular bundles in the external wall of the culm is
therefore a perfect structural adaptation towards a high rigidity and flexural strength
of the bamboo culm.
The fibers are long and tapered cells at the ends, with a thick wall and reduced
cellular lumen. In bamboo the ratio between fiber length and width varies between
150:1 and 250:1. Generally, bamboo fibers are much longer than those of wood (Liese
1992).
Bamboo fibers have a very thick wall, characterized by a particular multilayer
structure. In the bamboo culms the fibers remain alive and are modified during the
entire development of the plant. As the culm develops, new layers of secondary cell
wall are deposed within the support fibers, resulting in increased cell wall thickness
and cellular luminal reduction (Gritsch et al. 2004; Lin et al. 2002; Murphy and Alvin
1997).
The resulting thick walls are poly-laminar and this lamellar consists in the alterna-
tion of thick and thin layers. At the ultrastructural level, these layers have a different
orientation of the cellulose microfibrils (that can be considered as the elementary
anisotropic bricks used for building the cell wall). In the thick lamellae the cellu-
lose fibrils are oriented in axial direction, almost parallel to the longitudinal axis
of the fiber, with a very modest microfibrillary angle (MFA) (2°–20° with slight
gradual increase of the microfibrillar angle going from the median lamella to the
cellular lumen). The thin lamellae show an orientation of the fibrils in a direction
mostly transverse, almost perpendicular to the fiber axis, with a constant microfibril-
lary angle of 85°–90° (lower the MFA higher the longitudinal strength) as stated by
Parameswaran and Liese in 1975, 1996 and confirmed in 1980. Differences between
the lamellae of the bamboo fibers are also due to the degree of lignification: the thin
lamellae show a higher lignin and xylan content. The thick bamboo fibers have a
circular section and are joined together by a very strongly lignified median lamella,
which gives them greater ability to withstand shear stresses.
The poly-laminar wall structure is present above all in the fibers on the periphery of
the culm, conferring excellent flexural strength properties especially useful because
the bamboo culms must be able to sustain a great flexural deformation due to wind
loads. This deformation produces very strong stresses in the compressed half of
the culm. In this regard, Koch and Lybeer (2005) suggested that the poly-lamellar
wall structure of bamboo fibers can withstand considerable stresses due to axial
compression, thanks above all to frequent changes in the microfibrillar corners in the
multilamellar cell wall.
In addition, the cellulose fibrils of thick, axially oriented lamellas, act maximizing
the longitudinal elastic modulus of the fibers. Such an optimized structure may be
the result of a winning plant strategy to compensate for the lack of secondary growth
in stem thickness, and therefore the lack of opportunity to increase their moment of
inertia. In fact, bamboo moves the maximum resistances on high thresholds through
the maximization of the elastic modulus of the culms. Within the fibrous sheaths of
the vascular bundles, the bamboo develops a particular gradient of rigidity through a
differential thickening of the cell wall. Indeed, in the fibrous sheaths of the vascular
bundles of immature culms, it is easy to see that the innermost fibers first complete
their thickening of the walls, while those at the periphery are in a transitional state
made visible by the large cellular lumen. The reasons for this stiffness gradient, and
therefore the thickness of the fiber wall, can be interpreted as follows:
(1) Maximum rigidity is required to the internal fibers, in contact with the conductive
fabric, from the thin walls, in order to strengthen it and to prevent the vessels
from sticking because of the negative pressures exerted by the rising of the
water.
(2) A gradual reduction of the stiffness is necessary going from the internal fibers
towards the external fibers that are connected to the soft surrounding parenchy-
mal tissue.
Consequently, the bamboo can be modeled as a composite material, transversely
isotropic, consisting of a base matrix of parenchyma (from now on named “matrix”),
similar to a foam, reinforced by bundles of fibers associated with vascular bundles
(from now on designed with the term “fibers”). In this complex system also the
parenchyma has its importance in determining the physical-mechanical properties
of the bamboo, helping to transfer the loads between the fibers in shear stress. The
density determines largely the mechanical properties of the culm. Within the culm it
increases as the fiber content increases and the thickness of the cell wall increases.
In the bamboo culms there is, therefore, a pronounced gradient of radial density,
with a denser tissue towards the outside of the culm. This radial density gradient
has a profound effect on the mechanical properties of the bamboo: the axial tensile
modulus Young varies from about 5 to 25 GPa and the axial tensile strength varies
from about 100 to 800 MPa, for samples taken, respectively, from inner and outer
part of the culm.
Accordingly, the main aim of the present work is to design, by using Additive
Manufacturing (AM) technology, a biomimetic material capable of keeping advan-
tages from a morphological biomimetic structure inspired by bamboo natural one.
Bio-inspired structures can be considered a complicated structural material whose
design learns from nature’s design principles. As nature far exceed the capability of
traditional design and fabrication technologies, AM proves to be the most suitable
solution for designing such complex shapes, especially when topologically opti-
mized structures characterized by high lightness and mechanical strength have to
be reproduced. Therefore, by designing an internal optimized structure (both matrix
and fibers distribution) it will be possible to manufacture products with optimal
mechanical properties. In detail, this work demonstrates the feasibility of creating a
bamboo-like synthetic structure by using Fused Deposition Modelling (FDM) able
to combine two different materials within the same part.
2 Background
Additive manufacturing has been largely applied in the study and development of
innovative materials and biomimetic materials, specifically. The potential for fabri-
cating complex geometries on macroscopic, mesoscopic and microscopic levels as
well as the possibility of integrating different materials and produce parts character-
ized by locally-controlled mechanical properties represent massive strengths for the
study of advanced materials.
The first structure that has been studied is, arguably, the trabecular bone. Titanium
alloys have been used extensively for the fabrication of prosthesis and scaffolds.
Indeed, bone-like porous structures ease grafting and increase implant solidity. Sev-
eral works have been published on this specific topic (Zhang et al. 2018; Fantini and
Curto 2018). The study and reproduction of wood structures, on the other hand, is
still an open topic. For instance, a study that considers the shape of banana leaves is
presented in Ahlquist et al. (2015).
Considering the use of FDM technology, the range of materials still represent a
perceivable limit to the development of advanced materials (Gu et al. 2016). Never-
theless, several applications dealing with the fabrication of plastic composites have
been presented in literature (Shofner et al. 2003; Ning et al. 2015; Zhong et al. 2001).
This study provides new insights into the use of FDM to fabricate composite
materials inspired by wood. Indeed, the deposition mechanism used by FDM could
potentially be well-suited to simulate wood fibers. On a general level, mechanical
properties of FDM materials could be similar to wood ones. De Vries (2010) proposed
several digital models reproducing the bamboo structure; different fibers concentra-
tions and diameters are examined. Authors validated the model using FE analyses
and compared with bamboo wood mechanical properties. Finally, a specimen of the
chosen configuration has been produced via PolyJet and tested in a 4-point bend-
ing test. While results are promising, further studies are required to test different
technologies and materials.
A thorough review of latest advancement in bio-mimetic AM materials is pre-
sented by Studart in 2016; the author points out that, while the design flexibility
introduced by AM is beneficial to access a wide range of properties and function-
alities, it has also complicated the identification of optimal shapes. The study of
biomimetic materials answer to the need of a lean strategy to identify valid geome-
tries and structures.
3 Design of Biomimetic Bamboo-Like Material:

Framework
The fabrication of specimens of a composite plastic material capable of replicat-

ing the internal structure of the bamboo wood is a task that can be addressed using
different tools both for the CAD modeling and the manufacturing phases. The pro-
posed approach relies on Additive Manufacturing techniques for the fabrication; as
mentioned above, this paper aims at demonstrating the possibility of fabricating a
plastic composite material, bio-inspired by bamboo wood, using the Fused Deposi-
tion Modeling (FDM) technology. This technology would allow, in the future, the
generation of smart materials and complex components, with an internal structure
tailored and optimized for a specific application.
The study needs to take into account the following simplifications:
(1) This study aims at reproducing only the cross-sectional composite structure of
bamboo, and not its longitudinal structure. Specifically, only the inter-nodal part
of the wood is considered. Accordingly, the subject of this study is limited to
the study of composite microstructure observed in the bamboo culm wall.
(2) Only a part of the entire culm cross-section is investigated. In other words, the
whole cross-section of the culm (whose geometry resembles a circular crown) is
analyzed to produce a model. Moreover, to ease the production of the prototype
using AM a square section was preferred to a circular one. Such a section is
also most suitable to perform compression and bending tests according to ISO
3787 and ISO 3133 standards. The composition and distribution of fibers in
the model are determined simulating the same distribution that can be observed
moving from the external surface of the culm to the internal one (see Fig. 2).
Consequently, the reproduced structure can be considered representative of the
original structure under a small set of specific load conditions even if its geom-
etry is not characterized by a circular section.
(3) As previously mentioned, the shape of the fibers of the bamboo change in shape
in the cross-section, as they are typically more circular near the central lacuna
Fig. 2 Cross-sectional
images of bamboo structure
and more elliptic moving towards the external surface, as it can be seen in Fig. 2.
For the sake of simplicity, the cross-sectional shape of the fibers is modelled as
perfectly circular.
(4) The dimension of the fibers and the whole composition (i.e. density and distri-
bution of the fibers) needs to be adjusted according to the limits imposed by the
fabrication process.
(5) Generated specimens are tested by means of both a compression test and a 3-
point bending test. Accordingly, specimens’ geometries are modeled following
the indications prescribed by the ASTM D790 standards.
3.1 Modeling
CAD modeling has been performed within Solidworks, a well-known 3D modeler;

the goal of this phase is to build digital models of the specimens (i.e. STL files),
which are required to generate the G-CODE that contains the instructions for the
FDM machine.
The two specimens, required for the compression and the 3-point bending tests,
are rectangular prisms with a square section, with a side of 20 mm. The compression
test specimen is characterized by a length of 40 mm (overall dimensions: 20 × 20 ×
40 mm); the specimen for bending test is 400 mm long, with overall dimensions
(20 × 20 × 400 mm). These dimensions have been determined in order to make a
compression test compliant with ISO 3787 and a bending test compliant with ISO
3133.
Both parts have been generated modeling fibers and matrix of the composite
material as separate entities to allow manufacturing via additive technologies of the
two parts. Moreover, the specimens are composed of two separate entities, which can
be modeled by performing an extrusion of a single 2D sketch (i.e. the cross-section
Fig. 3 Measurement of
fibers
of the specimen). Accordingly, the entire modeling procedure is essentially reduced

to the definition of a valid 2D sketch representing a correct distribution of fiber and
matrix materials, faithfully inspired by the bamboo structure.
Accordingly, the definition of a significant cross-section of the specimen started
with an accurate analysis of the distribution of the fibers in bamboo wood. Specif-
ically, considering a cross-section of the culm and moving from the inner to the
outer surface, fibers density typically spans from 20 to 60%. This distribution allows
bamboo to withstand significant bending loads, as those caused by the wind. Inter-
estingly, the fibers have different dimensions along the culm radial direction. In the
external area, where fiber concentration is higher, their dimension is smaller. More-
over, their shape tends to an ellipse, with the major axis aligned with the culm radius.
For sake of simplicity, this last aspect is neglected during this study, and it will be
later introduced in future analyses.
In order to faithfully reproduce the characteristics of distribution/dimension of
the wood, a measurement campaign has been performed using an image of the cross-
section of a bamboo taken with a microscope. The image shows a sector of the culm,
with both the inner and outer wall visible.
In order to measure the diameter of the fibers, the area represented in the picture
has been subdivided in 5 horizontal regions of equal height (Fig. 3). For each area,
all the fibers have been measured in diameter in order to extract a mean dimension
(radius) describing the fibers. The results of the measurement are summarized in
Table 1.
Evidently, a scaling operation is required in order to maintain the correct pro-
portion between the cross-section of the specimen and the original section that has
Table 1 Measurement of fibers

Region Mean diameter (mm)
1 0.49
2 0.39
3 0.35
4 0.29
5 0.21
Table 2 Fiber parameters. Sketch A (standard concentration of fibers). Sketch B (lower fibers
concentration)
Region Mean diameter (mm) Sketch A—fibers Sketch B—fibers
concentration (%) concentration (%)
1 1.17 20 10
2 0.93 30 15
3 0.83 40 20
4 0.66 50 25
5 0.50 60 30
been taken as reference. Moreover, a scaling operation is likewise needed to allow

the fabrication of the smallest fibers, as the level of detail introduced in the digital
model needs to be reproducible by using the available FDM machine.
Values of Table 1 have been scaled by the ratio evaluated comparing the dimen-
sion of the actual bamboo wood (i.e. radial distance between external and internal
wall) and the dimension of the specimen (i.e. side of the square). Subsequently, the
minimum fiber diameter has been set equal to 0.5 mm (minimum diameter accepted
for correct fabrication). All other values are adjusted accordingly to maintain the
correct proportionality within the specimen (see Table 2).
Two sketches, represented in Fig. 4, have been modelled with different fiber
concentrations, partially relying on the results discussed by De Vries in 2010, where
circular/hollow specimens are designed, for determining a valid topology. Sketch
A is characterized by a fiber distribution comparable with bamboo wood. However,
after a visual inspection, sketch A has been judged as characterized by a general
fiber/area ratio too high (47% of the section area). Sketch B, on the other hand,
presents a fiber/area ratio of 24%.
Moreover, preliminary fabrication tests showed that specimens defined using
sketch A entail also severe fabrication difficulties. As a result, only the sketch B
has been used to generate the actual specimens that have been tested in this study.
Future studies will investigate into the different mechanical responses that can be
obtained varying the distribution and concentration of the reinforcement material.
Sketch B has been used to model the two specimens (compression and 3-point
bending) by means of an extrusion. The digital models can be seen in Fig. 5.
Fig. 4 Sketches of the cross-section of the specimens. Grey: matrix material; white: circular fibers
of various diameters. a Sketch A—characterized by a 47% fiber concentration; b Sketch B—char-
acterized by a 24% concentration
Fig. 5 Digital models of the compression (green) and 3-point bending (orange) specimens
3.2 Fabrication
FDM is an additive process where a thermoplastic filament is forced through a hot

nozzle and deposited in a semi-fused state on the workpiece, layer-by-layer. Usually
characterized by a typical gantry architecture, FDM machines move the extruder
relying on a horizontal crossbeam, moved by step motors, allowing the generation
of 2D horizontal paths; the platform, on the other hand, can be moved vertically to
generate each layer of the object. Once that is extruded, the plastic material imme-
diately cools, solidifying where has been deposited and providing a solid support
for subsequent layers. The FDM machine used for fabrication is a S2 produced by
Gimax 3D. The machine is characterized by a print volume of 400 × 400 × 400 mm,
heated bed and two parallel extruders.
Possible materials that have been considered to be used to print both the matrix
and the fibers of the composite are reported in Table 3, together with most significant
mechanical properties. It is important to note that sensible differences in the values
characterizing a specific filament can be found depending on the producer of the
filament, environmental conditions, integrity of the filament. Moreover, fabrication
parameters deeply affect the mechanical properties of the final printed part.
Consequently, Table 3 reports a range of possible values extracted referring to
different literature sources i.e.:
– Carbon Fiber Carbonium 3D Printing Filament—Treed Filaments—http://
treedfilaments.com/3d-printing-filaments/carbon-fiber-carbonium-2/ (accessed
May 30, 2018).
– Filamenti stampanti 3d vendita filamenti stampanti 3d filamento abs
n.d. https://www.filoprint.it/abs-diam-175-3-mm/2-abs-nero-o-175-mm-plastink.
html (accessed May 30, 2018).
– PLA Technical Data Sheets | Best PLA Filament Brand n.d. https://www.innofil3d.
com/material-data/pla-technical-data/ (accessed May 30, 2018).
– PLA, Renewable Thermoplastic | Stratasys n.d. http://www.stratasys.com/PLA
(accessed May 30, 2018).
Table 3 Mechanical properties of various FDM materials

Filament Ultimate tensile Tensile modulus Density (g/cm3 ) Cost (e/kg)
strength (MPa) (MPa)
PLA 20–53 2500–3600 1.26 20–25
ABS 33–60 1850–2200 1.01–1.04 30
Carbonium 110 8000 1.07 120–150
(polyamide
reinforced with
carbon fibers)
Ninjaflex 26 – 1.19 70
Semiflex
Table 4 Mechanical properties of the bamboo wood

Ultimate tensile Tensile modulus Density (g/cm3 )
strength (MPa) (MPa)
Bamboo—fiber 610 46,000 1.16
Bamboo—matrix 50 2000 0.67
Table 5 Recommended filament printing settings

Filament Extrusion Bed temperature Print speed Minimum
temperature (°C) (°C) (mm/s) diameter extruder
ABS 235 85 40–110 0.2
PLA 190 0 30–150 0.2
Ninjaflex 225–235 40 10–20 0.2
Semiflex
Carbonium 230 100 60 0.5
– ABSplus n.d. http://www.stratasys.com/materials/search/absplus (accessed May

30, 2018).
– PLA—Italian | Eumakers n.d. http://www.eumakers.com/it/filamenti-per-
stampanti-3d/pla (accessed May 30, 2018). [27–32].
The choice of the materials has been made trying to maintain the same relations
that can be observed within the bamboo wood in terms of mechanical properties
between fibers/matrix materials. Table 4 reports the data measured on the bamboo
wood that have been taken as reference.
By comparing Tables 3 and 4 it is evident that no FDM material is capable of
replicating the mechanical properties measured on the fibers of the bamboo wood.
Accordingly, the choice of materials will be made trying to simulate, to the best of
the possibilities guaranteed by available materials, the mechanical properties of the
wood.
It is important to note that the choice of materials needs to consider also the
manufacturability of the entire composite; a dual extrusion of materials that require
different settings of the machine may cause problems of cohesion at the interfaces
of the two components and even compromise the solidity of the entire specimen.
Accordingly, preliminary tests are required to verify the compatibility of different
materials for a combined extrusion.
Standard settings, recommended by filament producers, have been used for fab-
rication. Most important parameters are reported in Table 5 for each filament. The
slicing and G-CODE generation operations have been carried out within Cura, an
Ultimaker® slicing free software.
Cura allows the management of two extruders within the same print with different
print settings. In order to minimize the anisotropy effects introduced by the FDM
process (Buonamici et al. 2018), the deposition path of the extruder has been set for
the fibers as parallel to their principal direction (see Fig. 6).
Fig. 6 Fabrication preview

generated by Cura. The
deposition path for the fibers
(green) is defined by
rectilinear segments parallel
to the fibers principal
direction
This is necessary to maximize the performances of the specimen as well as the

similarity between the mechanical response of the specimen and that of the bamboo
wood.
Preliminary tests have been carried out with different combination of materials
to evaluate their compatibility within a single print. Ninjaflex® by Ninjatek showed
severe difficulties when combined with all the other materials (PLA, ABS, Carbo-
nium), mostly due to the low rigidity and the particular printing settings that requires.
Moreover, its mechanical properties could have made it suitable for use as matrix
material, but its high cost prevents this option in a real application.
Similarly, the use of the carbon-reinforced filament resulted positive only for the
fabrication of the compression specimens, that have limited dimensions. A specimen
for the 3-point bending test made with a combination of PLA and Carbonium was not
fabricated mostly due to the problems caused by thermal shrinkages. Such problems
could not be removed by adjusting different process settings. Some examples of the
results obtained in this phase are represented in Fig. 7.
Summing up, only the specimens composed of PLA + ABS were fabricated with a
satisfying quality level (considering adhesion of the layers and structural integrity of
the whole specimen) and have been put through the physical tests. The next section
will describe the results obtained.
Fig. 7 Examples of the specimens built for compression tests (top) and 3-point bending test (bot-
tom). The materials used are: PLA (red), ABS (grey), Carbonium (Black)
3.3 Experimental Tests
In order to provide a baseline for comparison, each test has been performed also
on two specimens of identical external geometry but fabricated in a single material
(PLA or ABS).
Test 1: Compression
Compression test has been performed according to ISO 3787 and using the setup
depicted in Fig. 8a, by compressing the specimen between two metallic plates and
imposing a perfectly vertical force that is measured by a load cell.
Results obtained are summarized in Table 6. All the tests have showed a classical
behavior of the specimen, highlighted by the presence of a clear shear band (see, for
instance, Fig. 8b).
Table 6 Compression test results

Specimen Max compression tension (MPa)
100% PLA 76
100% ABS 95
26% ABS + 74% PLA 80
Fig. 8 a Compression test setup. Specimen of the PLA/ABS composite material. b Compression
test of the ABS specimen. A shear bend, causing the failure of the specimen is clearly visible
Table 7 3-Point bending test Specimen Max flexural tension (MPa)

results (max flexural tension)
100% PLA 68
100% ABS 119
26% ABS + 74% PLA 105
Although minimal, there is an increase in the max stress sustained by the composite
specimen w.r.t. the mono-PLA specimen. This is an important result because it proves
the validity of the modelled structure (i.e. the topology established by the sketch
presented in Sect. 3.1) to assure a cohesive structure that is not prone to instability
failures. Considering the numerical value obtained, a max compression tension of
80 MPa is coherent with the composition of the specimen—i.e. with the amount of
ABS introduced in the PLA matrix. Considering that the fibers are oriented parallel
w.r.t. the load direction, their distribution within the cross-section does not contribute
significantly to the overall mechanical behavior of the specimen. In other words, any
distribution of 26% ABS would show the same result, provided that the solidity of
the specimen is guaranteed.
Test 2: 3-Point Bending
Bending test has been executed in compliance with ISO 3133 using a 3-point fixture
that is showed in Fig. 9. Results obtained throughout the tests are summarized in
Table 7. The test performed on the composite materials showed good results, that
prove the validity of the tested fiber distribution. The measured max flexural tension
resulted really close to the one measured for the mono-ABS specimen.
Figure 10 shows the fracture section observed in the composite specimen during
the bending test. By considering the image, it can be observed the distribution of
Fig. 9 3-point bending test performed on the mono-ABS specimen
Fig. 10 Section of the

fracture observed in the
bending test. PLA matrix
(red) and ABS fibers (grey)
the fibers (in grey) closely resembles the one designed in the digital model. A per-
fect cohesion between the two different materials has been observed throughout the
section, proving also the validity of the parameters chosen for the fabrication.
4 Conclusions
This study has proved the feasibility of producing biomimetic AM-generated com-
posite materials. The topology investigated is inspired by the analysis of the bamboo
cross-sectional structure, which is responsible for impressive mechanical perfor-
mances of the bamboo wood. Results obtained prove the feasibility of the entire
framework, as well as the validity of the choices made in all the phases (CAD mod-
elling, fabrication) even if for a limited amount of samples because of the technical
limits encountered in manufacturing them. This preliminary investigation lays the
basis for future in-depth studies on this topic.
It is important to note that fabrication problems cannot be overlooked, as they
represent a major aspect to be dealt within the whole framework. Specifically, con-
sidering FDM technology, the dual extrusion introduces problems that need to be
considered both in the CAD modelling phase and in the fabrication. For instance,
extrusion temperatures that differ significantly may cause adhesion problems at the
interface between different materials. Future tests will introduce a porosity in the
matrix material. This could limit the amount of material used for the matrix, which
is responsible for the shape and cohesion of the composite, but not primarily for the
structural behavior. Therefore, lighter and economic composite materials could be
further studied and tested.
References
ABSplus (n.d.) http://www.stratasys.com/materials/search/absplus. Accessed 30 May 2018

Ahlquist S, Kampowski T, Oliyan Torghabehi O, Menges A, Speck T (2015) Development of
a digital framework for the computation of complex material and morphological behavior of
biological and technological systems. Comput Aided Des 60:84–104. https://doi.org/10.1016/J.
CAD.2014.01.013
Amada S, Ichikawa Y, Munekata T, Nagase Y, Shimizu H (1997) Fiber texture and mechanical graded
structure of bamboo. Compos B Eng 28:13–20. https://doi.org/10.1016/S1359-8368(96)00020-0
Buonamici F, Carfagni M, Furferi R, Governi L, Saccardi M, Volpe Y (2018) Optimizing fabrication
outcome in low-cost FDM machines. Part 1—metrics. Manuf Technol 18(3)
Carbon Fiber Carbonium 3D Printing Filament—Treed Filaments (n.d.) http://treedfilaments.com/
3d-printing-filaments/carbon-fiber-carbonium-2/. Accessed 30 May 2018
De Vries DVWM (2010) Biomimetic design based on bamboo
Dixon PG, Gibson LJ (2014) The structure and mechanics of Moso bamboo material. J R Soc
Interface 11:20140321. https://doi.org/10.1098/rsif.2014.0321
Fantini M, Curto M (2018) Interactive design and manufacturing of a Voronoi-based biomimetic
bone scaffold for morphological characterization. Int J Interact Des Manuf (IJIDeM) 12:585–596.
https://doi.org/10.1007/s12008-017-0416-x
Filamenti stampanti 3d vendita filamenti stampanti 3d filamento abs (n.d.) https://www.filoprint.it/
abs-diam-175-3-mm/2-abs-nero-o-175-mm-plastink.html. Accessed 30 May 2018
Gritsch CS, Kleist G, Murphy RJ (2004) Developmental changes in cell wall structure of phloem
fibres of the bamboo Dendrocalamus asper. Ann Bot 94:497–505. https://doi.org/10.1093/aob/
mch169
Grosser D, Liese W (1971) On the anatomy of Asian bamboos, with special reference to their
vascular bundles. Wood Sci Technol 5:290–312. https://doi.org/10.1007/BF00365061
Gu G, Su I, Sharma S, Voros JL, Qin Z, Buehler MJ (2016) Three-dimensional-printing of bio-
inspired composites. J Biomech Eng 138:021006. https://doi.org/10.1115/1.4032423
ISO 3787 (1976) Wood—test methods—determination of ultimate stress in compression parallel
to grain
ISO 3133 (1975) Wood—determination of ultimate strength in static bending
Kanzawa E, Aoyagi S, Nakano T (2011) Vascular bundle shape in cross-section and relaxation
properties of Moso bamboo (Phyllostachys pubescens). Mater Sci Eng, C 2011(31):1050–1054.
https://doi.org/10.1016/j.msec.2011.03.004
Koch G, Lybeer B (2005) Lignin distribution in the tropical bamboo species Gigantochloa Levis.
IAWA J 26:443–456. https://doi.org/10.1163/22941932-90000126
Liese W (1992) The structure of bamboo in relation to its properties and utilization. In: International
symposium on industrial use of bamboo, Beijing: Intern. Trop. Timber Organization, Chinese
Academy of Forestry
Liese W (1998) The anatomy of bamboo culms. International Network for Bamboo and Rattan
Lin J, He X, Hu Y, Kuang T, Ceulemans R (2002) Lignification and lignin heterogeneity for various
age classes of bamboo (Phyllostachys pubescens) stems. Physiol Plant 114:296–302
Mannan S, Knox J, Basu S (2017) Correlations between axial stiffness and microstructure of a
species of bamboo. R Soc Open Sci 4:160412. https://doi.org/10.1098/rsos.160412
Murphy RJ, Alvin KL (1997) Fibre maturation in the bamboo Gigantochloa scortechinii. IAWA J
18:147–156. https://doi.org/10.1163/22941932-90001476
NinjaTek|High Performance 3D Printing Materials (n.d.) https://ninjatek.com/. Accessed 30 May
2018
Ning F, Cong W, Qiu J, Wei J, Wang S (2015) Additive manufacturing of carbon fiber reinforced
thermoplastic composites using fused deposition modeling. Compos B Eng 80:369–378. https://
doi.org/10.1016/j.compositesb.2015.06.013
Parameswaran N, Liese W (1975) On the polylamellate structure of parenchyma wall in Phyl-
lostachys edulis Riv. IAWA Bull Int Assoc Wood Anat
Parameswaran N, Liese W (1980) Ultrastructural aspects of bamboo cells. Cellulose Chem Technol
Parameswaran N, Liese W (1996) On the fine structure of bamboo fibres. Wood Sci Technol
10:231–246. https://doi.org/10.1007/bf00350830
PLA—Italian|Eumakers (n.d.) http://www.eumakers.com/it/filamenti-per-stampanti-3d/pla.
PLA, Renewable Thermoplastic|Stratasys (n.d.) http://www.stratasys.com/PLA. Accessed 30 May
2018
PLA Technical Data Sheets|Best PLA Filament Brand (n.d.) https://www.innofil3d.com/material-
data/pla-technical-data/. Accessed 30 May 2018
Shofner ML, Lozano K, Rodríguez-Macías FJ, Barrera EV (2003) Nanofiber-reinforced polymers
prepared by fused deposition modeling. J Appl Polym Sci 89:3081–3090. https://doi.org/10.1002/
app.12496
Studart AR (2016) Additive manufacturing of biologically-inspired materials. Chem Soc Rev
45:359–376. https://doi.org/10.1039/C5CS00836K
Ultimaker Cura 3D Printing Software (n.d.) https://ultimaker.com/en/products/ultimaker-cura-
software. Accessed 1 Mar 2018
Wang X, Ren H, Zhang B, Fei B, Burgert I (2012) Cell wall structure and formation of maturing
fibres of moso bamboo (Phyllostachys pubescens) increase buckling resistance. J R Soc Interface
9:988–996. https://doi.org/10.1098/rsif.2011.0462
Wang X, Keplinger T, Gierlinger N, Burgert I (2014) Plant material features responsible for bam-
boo’s excellent mechanical performance: a comparison of tensile properties of bamboo and spruce
at the tissue, fibre and cell wall levels. Ann Bot 114:1627–1635. https://doi.org/10.1093/aob/
mcu180
Zhang B, Pei X, Zhou C, Fan Y, Jiang Q, Ronca A et al (2018) The biomimetic design and 3D
printing of customized mechanical properties porous Ti6Al4 V scaffold for load-bearing bone
reconstruction. Mater Des 152:30–39. https://doi.org/10.1016/J.MATDES.2018.04.065
Zhong W, Li F, Zhang Z, Song L, Li Z (2001) Short fiber reinforced composites for fused deposition
modeling. Mater Sci Eng, A 301:125–130. https://doi.org/10.1016/S0921-5093(00)01810-4
Microtimber: The Development of a 3D
Printed Composite Panel Made
from Waste Wood and Recycled Plastics
Sandra Karina Löschke, John Mai, Gwenaelle Proust

and Arianna Brambilla
Abstract This chapters reports research conducted in the context of the

multi-disciplinary research project at the University of Sydney—Microtimber:
Development of a 3D printed, gradient timber panel composed of forestry waste and
by-products (2015–2019). Funded by Forest and Wood Products Australia (FWPA),
the research intends to valorise forest and plastic waste by combining saw dust
with recycled acrylonitrile butadiene styrene (ABS) to develop an environmentally
sustainable composite material suitable for 3D printing, using a fused filament fab-
rication process. First, the research explores the mechanical performance and print-
ability of wood-plastic composites and variations in their respective compositions
and second, it develops new 3D printing processes that achieve material and aes-
thetic gradients through the optimisation of printing parameters and development of
printing algorithms. The aim is to achieve a fluidly variable gradient material that rep-
resents a new design paradigm in architecture and replaces traditional architectural
systems that rely on the mechanical layering of different elements such as structure,
rain screen, insulation, lining etc. Preliminary testing showed that from a perspective
of environmental sustainability, the unproblematic recycling of these Microtimber
specimen promises to close the loop between the material sourcing stages and the
end of life management of Life Cycle Assessment (LCA).
Keywords 3D printing · Wood · Waste · Architecture · Aesthetics
1 Introduction
3D printing is becoming increasingly popular and its use in architectural construction

is growing fast. Unlike current fabrication technologies that mechanically cut, fix or
glue different standardized elements for serial production, 3D printing chemically
fuses different composites at a microlevel and allows for customized variation of
S. K. Löschke (B) · J. Mai · G. Proust · A. Brambilla

Faculty of Architecture, Design and Planning, The University of Sydney, Sydney, Australia
e-mail: sandra.loschke@sydney.edu.au

828 S. K. Löschke et al.
the material. To illustrate this point in more concrete terms: In timber stud wall
construction, for example, we have a layering of discrete materials that are screwed,
nailed, bolted or glued together. Typically, this is a layer of plaster board, timber studs,
insulation, vapour barrier, battens and rainscreen. In contrast, 3D printing allows
us to replace layered construction with a homogeneous composite material. The
material’s composition can be changed to respond to performance requirements such
as structural strength, weather resistance and surface appearance. It meets different
performance criteria through chemical variation rather than mechanical assembly
and allows for customized panels that use materials economically.
To date, this innovative potential has been overlooked, and 3D-printing often
simply imitates traditional fabrication processes. The Chinese company Winsun, for
example, 3D prints entire multi-storey concrete houses. However, the fabrication
process requires the mechanical integration of reinforcement bars, mimicking tra-
ditional concrete wall construction. The Microtimber project is less interested in
mimicking existing construction methods and instead asks what 3D printing can do
that other fabrication processes cannot achieve.
The research focuses on identifying and testing new composites that can be made
from waste and by-products and combines natural resources such as saw dust from
the forest industry as well as ‘non-natural’ resources such as recycled plastics. The
utilisation of waste for 3D printing could have important benefits for the timber,
construction and waste industries and could deliver sustainable alternatives to raw
material-intensive building products that contribute to landfill. For architects, the
opportunity to design directly through material composition offers new possibili-
ties of aesthetic expression that are currently restricted to the skilful manipulation
of ready-made building materials. The 3D printing of waste-based composites rep-
resents an entirely new design paradigm that forces a rethinking of architectural
aesthetics (fluid variation), practices (co-design with engineers, material scientists
and other specialists) and environmental thinking (waste as resource).
2 Developing the 3D Printing Technology for a New

Gradient Timber-like Aesthetic
This section reports the development of 3D printing processes that permit the fab-
rication of an aesthetically- and materially-graded Microtimber panel, using com-
mercially available filaments made from wood-plastic composites. In particular, it
explores speed, temperature and printing angles to achieve an appearance that aims
to be more timber-like and attractive than existing wood-plastic composite prod-
ucts. The study evaluates the material properties of these gradient timber prototypes
in terms of printability and tests their mechanical performance in terms of tensile
strength.
Microtimber: The Development of a 3D … 829
2.1 Aesthetic Perceptions: Solid Wood Versus Wood-Plastic

Composites
During the past two decades, the number of timber substitutes for structural and
architectural applications has risen steadily. These new materials seek to deliver tech-
nically advanced solutions that improve the structural performance and durability of
natural wood. Because they are mostly fabricated from wood waste- or by-products,
they are not in direct competition with solid wood but represent a new subset of prod-
ucts that permit the more efficient use of timber resources. Wood has a very high
appeal when compared to other building materials like brick, concrete, steel or plastic
(TDA 2015). But the appeal of wood does not translate to wood composites and it is
useful to gain a better understanding of how wood composite materials are perceived
by consumers in comparison to natural wood, and how the respective associations and
sensory reactions of consumers can inform the development of new composites—in
this instance, the development of a Microtimber product (Pritchard 2004). Research
has shown a clear consumer preference for natural wood over composite products
(Jonsson et al. 2008). Consumers used predominantly perceptual attributes, vision
and tactility, to describe their responses to different wood materials, identifying an
underlying quality that can be described as ‘wood-likeness.’ This was associated
with qualities such as natural, soft and living, and stood in contrast to the qualities
that were associated with wood-plastic composites, which were seen as artificial and
hard. These perceptions, might explain why many wood-plastic composite products
do not enjoy the same popularity as solid wood although in many instances, they are
more sustainable, durable, affordable and offer a better structural performance. The
aesthetic appearance of solid wood is determined by its microscopic wood features,
namely the colouring and structural features that are easily visible to the eye. These
are the result of biological processes and environmental influences such temperature
and moisture levels, and lend wood characteristic features such as rings, knots and
resin or bark inclusions (Kolb 2008). As a result, every piece of wood is unique
but different species are recognisable by the particular combination, prominence
and articulation of these features. To make wood composite materials more ‘wood-
like’ then, a solid understanding of these underlying aesthetics and how they can be
translated into fabrication methods is of key importance.
2.2 “Timber-like”: Developing Advanced 3D Printing

Technologies to Generate a New Wood Aesthetic
3D printing offers new possibilities to address the shortcomings of more established

wood composite products such as laminated boards and macro-extruded wood-plastic
composites. Wood laminates mimic the appearance of solid wood through a textured
surface that is applied to MDF or other composite boards that by themselves have a
dull and unappealing appearance. They are extensively used in household and office
furniture. Macro-extruded wood-plastic composites attempt to imitate the texture

of wood through physical indentations that also act as a non-slip surface for appli-
cations such as decking boards, floor boards and hand railings. However, standard
industrialised fabrication processes rely on the repetition of a standard pattern which
is a visual indicator of the non-authenticity of the material.
The potential for the use of 3D printing as a method for simulating the features
of timber is still in its early stages. While 3D printing has been used in construction,
often post-processing of the product is needed to smoothen out the characteristic
horizontal striations that result from the layer-by-layer application of the material
during the 3D printing process. In addition, there has been limited exploration of
custom-designed aesthetics in 3D printing.
Many 3D printed wood-plastic composite products display an aesthetically unap-
pealing uniformity that lacks variation in surface texture and colour and negatively
impacts on their marketability as viable alternatives to natural wood. To address
these shortcomings, the research first, explores new 3D printing processes that
eliminate the need for post-processing and as a result, improve the cost-efficiency
as well as the embodied energy of 3D printed wood-plastic composites, and second,
develops an innovative and attractive aesthetic for these products that can be
generated by varying the parameters of the additive production process (speed,
temperature, printing angle, nozzle size etc.) and/or the material composition. By
fusing and fluidly varying structure and surface, it becomes possible to create 3D
printed wood-plastic composite panels without any post-processing.
2.3 Manipulation of Print Parameters to Define Wood Grain

and Colour
Timber generally contains light and dark striations in the form of growth rings because
of the seasonal variations that occur during the life of a tree. Depending on the species
and specific climatic conditions, a tree’s striations may be distinct, have a gradient
variation, or a blend of both (Hafner 2015). 3D printer functions can be manipulated
to recreate a wood grain on a 3D printed object when using a suitable wood-plastic
filament. For this research project, a commercial wood-plastic composite filament
known as Laywoo-D3 (developed by CC-Products, Köln, Germany) was explored
because it displayed large visual variation between light and dark tones when exposed
to different print parameters (Jovanovic 2015). These parameters include the move-
ment speed of the print head and the temperature of the nozzle.
The velocity of the tool path can be manipulated by changing the speed of the
print head and bed. Distinct light and dark tones can be produced, as the speed can be
changed with almost instant effect. Slower print speeds allow for the wood particles
to char for longer, hence producing darker tones, while faster print speeds reduce
charring and hence produce lighter tones. Using Laywoo-D3, contrasting striations
were printed by changing print speeds: a print speed of 10 mm/s produces dark tones
while one of 60 mm/s produces light tones.
Striations can also be produced by changing the nozzle temperature. In order to
produce a contrasting pattern through temperature variation, Laywoo-D3 was printed
at a temperature of 180 °C to produce the lighter tones, while a temperature of 230 °C
was used to produce darker tones.
In contrast to the instant changes in speed, changes in temperature require more
time and produce a more gradual change in tone, due to the longer transition time.
The time taken to reach the desired temperature can be programmed to vary and
depends on the type of 3D printer. For example, samples created using the Leapfrog
Creatr Dual Extruder (developed by Leapfrog 3D Printers, Alphen aan den Rijn,
The Netherlands) of dimensions 80 mm × 40 mm with a 0.5 mm nozzle at 15 mm/s
printing speed showed a gradient over a width of 5 mm as it takes 60 s to cool down
from 230 to 180 °C (Mai et al. 2016).
Both the temperature and speed variation have their benefits and drawbacks. The
range of values within which temperature and speed can be varied is limited and
exceeding the limits may adversely affect the print in quality. For example, extremely
high temperatures can burn the wood and severely degrade material properties, while
excessive speeds can reduce the contact between the material coming out of the nozzle
and the already deposited material on the print bed and produce a low-quality surface
while also affecting mechanical performance.
To ensure that striations that resulted from various degrees of charring did not
impact the structural integrity of the material, tests to analyse the tensile strength were
conducted according to standard ASTM D638-14 Standard Test Method for Tensile
Properties of Plastics. From this standard, ASTM D638-14 Type 1 tensile specimens
were printed at temperatures between 180 and 230 °C at 10 °C intervals (Mai et al.
2017). Three specimens were printed at each of the eight different temperatures.
Over this wide range of temperatures, the tensile strength of the specimen was not
significantly altered, as shown in Fig. 1.
2.4 Algorithms to Generate Timber-like Textures

for a Microtimber
The principles for creating a wood-like texture were established by adapting an algo-
rithm that provides a more realistic appearance to computer-generated surfaces—Per-
lin noise. This algorithm was created by Ken Perlin in 1983 to provide a natural
texture to an otherwise plain graphic design (Shiffman 2012). Rather than being rep-
resented as a completely random sequence of numbers, Perlin noise smoothens out
the randomness by incorporating a pseudorandom sequence that only varies slightly
between adjacent values.
This algorithm was adapted to develop a realistic wood grain that exhibits the
seemingly random variations in wood grain width both ‘along the grain’ and ‘across
14
12
Peak Tensile Stress (MPa) 10
0
180 190 200 210 220 230 240 250
Temperature (°C)
Fig. 1 Peak tensile stress of Laywoo-D3 tensile coupons printed at different temperatures. (8)
Fig. 2 Evenly spaced lines (left) is modified to have varying width (right) by application of a
one-dimensional Perlin noise map. (7)
the grain’ textures. Trees experience grain variations due to varying climatic con-
ditions, localised shading or competition from neighbouring objects (Fritts 2012).
These realistic variations can be approximated with the Perlin noise algorithm.
The Perlin noise algorithm can be visualised as a black and white map that exhibits
a pseudo-random change in tone across the map. The change in tone allows for
the manipulation of certain wood grain characteristics, such as spacing density and
curvature.
As shown in Fig. 2, the one-dimensional Perlin noise map has darker regions that
produce grains with closer spacing, while lighter regions produce grains with wider
spacing. The variation in the spacing can be varied according to the wood species or
the desired visual appearance.
Expanding to a two-dimensional Perlin noise map, Fig. 3 shows a shift in the wood
grain in the y-direction depending on the tone of the map at a specific point. A lighter
region shifted the line up in the positive y-direction while a darker region shifted the
line down in the negative y-direction. Like the spacing density, the resultant curvature
from the shifting of lines can be varied according to the wood species or the desired
visual appearance.
To achieve the desired texture, a combination of temperature change, speed change
and tool path manipulation was developed. The grain is produced using a tool path
that is predominantly in the x-direction, with subsequent pathways shifted in the
Fig. 3 Linear lines (left) is modified to have varying curvature (right) by application of a two-
dimensional Perlin noise map. (7)
Fig. 4 A combination of
temperature change, speed
change and tool path
manipulation (7)
y-direction. Depending on the nozzle size, overlapping of the print path can increase
the darkness of that region, as the heated nozzle passes over a previously printed
surface. The results are shown in Fig. 4.
2.5 Flexibility to Manipulate 3D Surface
Traditionally, 3D printing using fused filament fabrication technology applies flat,

layer-by-layer deposition. This is the norm for slicer software that translates a 3D
drawing created with standard 3D modelling software into a G-code (a generic name
for a control language) that can be read and printed by the 3D printer. To overcome
this flatness, it is possible to customise the G-code instructions that are sent to the 3D
printer to achieve a wood-like curved surface. This can eliminate the stepped surface
that would otherwise be produced on a curved surface.
The curved layer printing provides an opportunity to imprint a physical texture
onto the surface, which can be varied depending on the printing orientation. 3D
printers normally have a print head with a fixed vertical alignment. When printing on
a curved surface, the width of the nozzle can scrape on the existing printed material
while printing the new layer, partially remelting it to produce a roughly textured
surface (Fig. 5). This may be desirable to create a non-slip surface or achieve a more
natural texture of wood for interior applications.
Fig. 5 Developed prototype

sample showing rough
surface features with a more
natural appearance
In order to achieve a smoother curved surface, it is necessary to align the print head
orientation such that it is always perpendicular to the print surface to avoid excessive
nozzle interference. Since conventional 3D printers are unable to rotate the print head,
the six-axis KUKA KR 6 R-900 robot (developed by KUKA, Augsburg, Germany)
was used to achieve the desired smooth curved surface (Fig. 6). In addition to the
flexibility of movement, a six-axis robot has the additional advantage of higher print
speeds and the potential for upscaling to industry-size production. By attaching the
extruder and hot end combination as an end-effector on the robot, the functions of a
3D printer can be replicated by controlling the movements through the KUKA robot
software and synchronising with the Arduino firmware to control the temperature
and extrusion settings.
The design of conventional 3D printers usually incorporates slow z-axis move-
ments compared to x- and y-axis velocities. This increases the opportunities for
printing more complex surfaces that involve significant variation in height at differ-
ent points on the surface. The six-axis robot has the additional advantage of providing
more freedom to create the form, compared to desktop 3D printers that have a cross-
bar or gantry which limits their range.
Upscaling of the print can be achieved by increasing the effective print area and/or
the extrusion rate. The larger reach of the robot arm effectively increases the available
printing area compared to that of conventional desktop 3D printers. The design of the
print head can also be modified to increase the flow rate, for example by incorporating
a larger nozzle or an extended melt zone. Both the increase in print area and extrusion
rate can allow for larger elements such as façade panels to be produced without the
constraint of size limitations and slow speeds.
Fig. 6 a + b: Developed prototype sample showing smooth surface (left) produced using KUKA
robot (right)
3 Experimenting with Material Composition—New

Wood-Plastic Composites
To date, research into the material composition of wood-plastic composites has

largely focused on pressed, extruded or moulded products rather than on 3D printed
elements. To develop a Microtimber material, that addresses this gap in knowl-
edge, a series of experiments were conducted that created and tested various wood-
plastic composites suitable for 3D printing, using a fused filament fabrication process.
In particular, the following was investigated: (a) varying the ratio of wood/plastic
in the composite; and (b) using different timber species for the timber compo-
nent—Australian Hardwoods: Red Gum (Eucalyptus camaldulensis), Grey Box
(Eucalyptus microcarpa), Tasmanian Oak (Eucalyptus delegatensis/Eucalyptus reg-
nans/Eucalyptus obliqua) and Softwood: Pine (Pinus radiata). A series of tests eval-
uated the material properties of the composites in terms of tensile strength and print-
ability. The strength of the wood species in their natural state did not translate to
their performance as part of a wood-plastic composite. It emerged that in compos-
ites, the strength inherent to the natural fibrous structure of wood no longer comes
into play and aspects such as chemical composition and micro-particle size and shape
exert a more decisive influence on the properties and performance of the material.
Our testing matrix therefore included particle sizes in addition to the range of wood
species and wood-plastic ratios. In a second stage, the impact of recycling on the
performance of the selected wood-plastic composites was investigated by recycling
our own printed material, re-extruding the samples and printing new coupons for
testing.
3.1 Rationale for the Selection of Wood Species and Their

Relevant Properties
For the series of experiments, a range of wood species was selected that are com-
monly used in Australian construction: three Australian Hardwoods (Grey Box, Red
Gum and Tasmanian Oak) and a Softwood (Pine). The Australian Standard AS2878
classifies the ultimate strength of natural seasoned timbers in eight groups, SD1 rep-
resenting the strongest and SD8 the weakest. According to the standard, Grey Box
is classified as SD2, Red Gum as SD5 and Tasmanian Oak as SD2. With a density
of 1100 kg/m3 , Grey Box is considerably heavier than Red Gum and Tasman Oak
which have respective air-dry weights of 900 and 740 kg/m3 . A detailed study for
Red Gum and Grey Box has been published earlier (Huang 2017). In contrast to the
Australian hardwoods, Pine is softwood species with and average air-dry density of
535 kg/m3 and has been classified as strength group SD6.
During the process of creating wood flour, different wood species have produced
material with different wood particle sizes and particle shapes, and these were iden-
tified as factors that influence the tensile strength and density of the respective 3D
printed samples. Measurements of the particle diameter, aspect ratio, circularity and
elongation of the wood particles in Grey Box showed that it has a greater average
particle size and also a wider distribution of particle diameters when compared to Red
Gum, for example (Huang 2017). As a result of these differences, the ‘3D printabil-
ity’ of the individual wood species also varied and the parameters for the extrusion
of filaments and the 3D printing process had to be fine-tuned to obtain homogenous
materials with optimal mechanical and aesthetic properties.
3.2 3D Printability and Particle Size
For 3D printing, wood-plastic filaments with a 2.8–3.0 mm diameter were produced.

First 3D test prints showed that larger wood particles resulted in a congestion of
material in the printing head and to ensure a smooth and constant printing process,
the maximum wood particle size had to be set at 212 µm diameter in order to
avoid clogging. To control the maximum particle size, the wood powders were dried,
sieved and separated according to three different particle size ranges before being
mixed with the ABS and maleic anhydride, which was used as a binding agent.
The three ranges of particle sizes were tested for Australian hardwood powders to
investigate their impact on the composites’ printability and properties. Wood particles
in range 1 had a diameter size of 150–212 µm; range 2 comprised particles with
Fig. 7 Dimensions of
coupons for tensile testing
diameter sizes of 90–150 µm; and range 3, diameters below 90 µm. Using a Noztek
Pro filament extruder (developed by Noztek, West Sussex, England) with a pre-heater
band and a 2.75 mm extrusion head, the filaments were extruded at a temperature
range of 150–160 °C. From these, printed tensile specimens were printed, using a
LULZBOT 3D printer (developed by Aleph Objects, Colorado, USA) with a 0.8 mm
printing head, heated between 200 and 250 °C and moving at a printing speed of
15 mm/s. Figure 7 shows the dimensions of the tested tensile specimens adapted
from the standard ASTM D638. For the Red Gum, Grey Box and Tasmanian Oak
the printing temperature was fixed at 250 °C. The Pine specimen were printed at
200–240 °C. As a benchmark, pure ABS specimen were printed to evaluate the
effects resulting from the inclusion of wood particles.
3.3 Mechanical Performance of the Wood-Plastic Composites
Tensile tests on the printed specimens were performed using an MTS Criterion load-
ing frame with a constant deformation rate of 0.5 mm/min. At least three tensile
specimens for each wood species were tested and loaded in tension until failure.
The loads and displacements were recorded during the tests to plot the stress-strain
curves from which the tensile strengths were extracted. Figure 8 shows the data for
the single-extruded filaments.
Fig. 8 Tensile strength test

data for: pure plastic (ABS)
and single-extruded
wood-plastic composites
made with Red Gum, Grey
Box, Tasmanian Oak and
Pine woods
The tensile strength test data show that the addition of wood particles to ABS
produces a composite material with a lower strength than pure ABS, which can be
explained by the hydrophilic nature of wood which causes improper bonding with the
plastic (Panthapulakkal and Sain 2007). Therefore, the higher the wood content of
a wood-plastic composite, the lower the strength of the material. Additional testing
of a Red Gum-plastic specimen with a weight ratio of 29:70:1 (wood: ABS:binding
agent) confirmed the negative impact of wood presence on the final strength of the
material. In fact, the tensile strength was measured at 11.3 MPa, which represents a
35% loss in strength for a 10% increase in wood content.
The tensile strength varied also across the different wood species. Red Gum has
the highest strength followed by Pine and Grey Box. Tasmanian Oak produces the
weakest composite in this study. Analysis of the fracture surface of the different
specimen using scanning electron microscopy has shown that when compared to
the Red Gum specimen, Grey Box and Tasmanian Oak specimen presented a large
amount of porosity, which might explain the lower strength of these composites
(Huang 2017). However, this requires further investigation.
The Pine wood-plastic specimens were printed at temperatures ranging from 200
to 240 °C and no relation between the printing temperature and the strength of the
composite was observed. This aligns with the results presented in Sect. 2.3 of this
chapter, where tests of specimens produced with commercial wood-plastic composite
filament showed that no relation between the tensile strength and the 3D printing
temperature exists (Mai et al. 2017).
Figure 9 shows the effect of varying the particle sizes of the wood powder on
the tensile strength of the wood-plastic composite fabricated using Grey Box. The
graph shows that decreasing the particle sizes slightly increases the strength of the
composite. The same test was conducted for Red Gum, but the trend was not as
obvious as the one observed for Grey Box (Huang 2017).
Fig. 9 Tensile strength test

data of wood-plastic
composites made from Grey
Box powder sieved to
different particle size
distributions
3.4 Recycling and Re-Extrusion
Red Gum, Grey Box and Pine wood filaments can be recycled and re-extruded.
Having been fabricated at a temperature of 150–160 °C, the material was broken
into small pieces and placed in the extruder to fabricate a new filament, this time at
a lower temperature of 115 °C.
In all but one case, the specimens produced from the re-extruded material dis-
played higher tensile strengths than the specimen produced from the original single-
extruded filaments. The only exception occurred in the Red Gum-plastic composite
that used wood particle sizes between 90–150 µm, where the result was reversed.
Importantly, it could be observed that the re-extrusion process homogenises the
composite material and could explain the decrease in the standard deviation of the
measured mechanical properties (Huang 2017).
3.5 Future Research into Wood-Plastic Composites for 3D

Printing
The results presented here show that the wood provenance, particle size and process-
ing parameters affect the tensile strength of 3D printed wood-plastic composites.
Other research (Panthapulakkal and Sain 2007) has shown that binding between
wood particles and polymer does not happen naturally and therefore the careful
selection of a binding agent is essential to achieve cohesion between wood and plas-
tic to improve the composite material’s strength. There is great potential for future
research into the role of binding agents both in terms of investigating the quantity
of agent used but also in testing different types of binding agent that might respond
to the chemical compositions of various woods. It is not difficult to imagine that
the high resin content of Australian hardwoods, for example, might influence the
selection of binder and that an investigation of alternative binders could improve the
tensile strength of the wood-plastic composites tested in this research project.
Whilst this research focused on tensile strength to date, other properties, such
as durability, thermal performance and aesthetic perceptions will be investigated to
determine the useability of these particular wood-plastic composites in architectural
applications.
Additionally, Pine-plastic composites were recycled after having been printed
once and the preliminary results show that there is no change in the tensile strength
measured on these new fabricated specimens.
4 Evaluation of Microtimber in Terms of Building Life

Cycle Assessment (LCA)
This section evaluates the anticipated opportunities and limitations of Microtimber

as a 3D-printed wood-plastic composite in comparison with commonly used com-
posite wood products in terms of: raw material extraction and materials processing,
manufacture, distribution, use, repair and maintenance, and disposal/ recycling.
4.1 Life Cycle Assessment of Buildings: General Framework
Buildings have a considerable impact on the environment and are among the greatest
contributors to global carbon emissions and energy use. In 2014, the construction sec-
tor was directly and indirectly responsible for 18% of global greenhouse gas (GHG)
emissions and one of the largest consumers of materials (IPCC 2014). Improving
the efficiency of material usage and deploying recycled or waste materials, such as
waste saw dust and recycled ABS plastics, can therefore significantly help to reduce
the GHG emissions below current baselines scenarios.
One of the tools to assess energy consumption and environmental impact of mate-
rials and products is life cycle assessment (LCA). Although LCA has been used as
a method to evaluate the environmental impacts of a product for some time, its use
in the construction sector is relatively new, with the first application dated to around
2003 (Ortiz et al. 2009; SCBI 2006). LCA is the systematic analysis of the environ-
mental impacts of a product or a process across all stages of its life cycle, from raw
material extraction to manufacturing, use, maintenance and finally disposal.
Among the different approaches available, the framework for life cycle assessment
published by the International Organization for Standardization (ISO) is the most
developed and has been progressively reviewed since its inception in 2006, resulting
in an entire series of new international standards focused on LCA (ISO 1404:2006;
ISO 1404:2006). These standards introduce a clear framework for life cycle analysis,
with a step by step methodology based on four main stages:
1. The first stage defines the goal and scope of the LCA. Its main aim is to establish
clear boundaries, assumptions, metrics and evaluation criteria for the successive
stages.
2. The Life Cycle Inventory (LCI) database is set up in the second stage, when all
the data about the impacts of products and processes are collected and calculation
procedures are defined.
3. Life Cycle Impact Assessment (LCIA). It is the actual calculation phase that
aims to understand the extent and significance of impacts of a product system
4. Data interpretation. It deals with the interpretation of the results of stages two
and three based on the criteria set up in stage one and seeks to reach conclusions
and recommendations.
The importance of stage one is vital, as it sets the rules and the assumptions for the
following assessments and determines the success and validity of the entire process.
During the first stage the boundaries of the analysis are defined, namely the phases
and the processes that will be analysed.
Specific standards for the building sector have been introduced to guide the LCA
practitioners in the first stage (EN 15643-2:2011; EN 15978:2011; EN 15804:2012).
Standard EN 15978:2011 Sustainability of construction works. Assessment of envi-
ronmental performance of buildings. Calculation method defines the building’s life
cycle through different sequential stages, as shown in Table 1.
A comprehensive LCA accounts for the impacts of materials and processes during
all stages of the building life cycle as outlined in Table 1. The assessment is not a
simple aggregation of individual assessments into a single overall score. Instead
it is an iterative process that evaluates impacts of building materials, processes,
building services, equipment, operational uses, retrofit and disposal, and the complex
relations between these (Cabeza et al. 2014). As new information becomes available
during the assessment process, trade-offs are required and the scope might need to
be reconsidered and modified as a result.
Of particular difficulty is the assessment of impacts related to the end of life stages
when a product leaves the life cycle process. The amount of data available for these
stages is very limited in the building sector. Common end of life options are recycling,
reprocessing, reuse, or disposal; but the LCA evaluation considers processes and
strategies that would come into effect several years after the calculation (Blengini
and Di Carlo 2010) as building life span is considered to be between 50 and 70 years.
This issue was already detected in 2001, when the European Commission declared
that the end of life management should be a priority (European Commission 2001).
The EU Construction & Demolition Waste Management protocol (September 2016)
states that waste from construction and demolition is the largest source of waste in
the European Union, accounting for a third of all waste production (ISO 1402). The
aim of the protocol is to transition Europe towards a circular economy by closing the
loop of product lifecycles and to achieve this a concrete target of recycling 70% of
all Construction & Demolition waste by 2020 has been set.
As a wood-plastic composite material, Microtimber, comes in exactly at this point,
as its potential lies in closing the loop. At the “Product Stage” it uses recycled/waste
842
Table 1 Building life cycles, defined by EN 15978:2011

Building life cycle
Product stage Construction Use stage End of life
process
A1 A2 A3 A4 A5 B1 B2 B3 B4 B5 B6 B7 C1 C2 C3 C4
Benefits and loads beyond the system boundaries
Raw Transport Manufacture Transport Installation Use Maintenance Repair Replacement Refurbishment Operational Operational Deconstruction Transport Waste Disposal
material process energy use water use processing
supply
D: reuse, recovery, recycling − potential
S. K. Löschke et al.
materials as a ‘raw material’ and the “End of Life” Stage it can be recycled by being
powdered and reprinted without the loss of mechanical performance as preliminary
testing has shown.
4.2 Expected Environmental Performance of Microtimber
At the present stage, the Microtimber wood-plastic composites are in the early stages
of development and for this reason, a definitive life cycle assessment is not possible.
However, it is important to consider the likely environmental performance of the
product at an early stage so that material composition and fabrication processes can
be optimised to target an overall reduction of the material’s carbon footprint.
Due to Microtimber’s use of small participles and reliance of chemical bonding
action, the closest set of comparable products are timber particle boards and reconsti-
tuted timber panels. They can help to provide an informed estimate for the expected
environmental performance of Microtimber.
In order to have reliable and consistent data for input, it is important to refer to a life
cycle inventory source that is officially recognized and can be applied in accordance
with relevant standards—for example, Environmental Product Declarations (EPDs)
(19, 20). An EPD is a document that certifies the environmental data of products,
based on the LCA method described by the ISO 14025 (Passer et al. 2015). This
type of document is essential for a reliable building LCA, as because it represents
a comprehensive database, one of the benefits of an EPD is that it provides the
standardized documentation needed for certification as it aligns with the first step of
LCA, namely goal and scope definition. EPDs clearly state the phases considered in
the assessment, the boundaries, the assumptions and the data sources, which together
constitute the first essential step of LCA.
Table 2 shows the two building products used for comparison with Microtim-
ber—particle board and reconstituted timber panels. The data and LCI database vary
with the geographical location (different processes, cost of primary energy, trans-
portation of the products etc.), therefore an Australian database1 to contextualize the
analysis to specific criteria was used.
Particle board uses a large percentage of recycled timber as raw material, however
its short lifespan and the low strength make it largely unsuitable for external or
structural use in buildings. The reconstituted timber panel seems to be more similar
to the Microtimber wood-plastic composite in its potential use. However, although
it is deemed to have high recycling potential, it is estimated that only 7.5% of all
panels are redirected from landfill. Assuming that Microtimber would have similar
environmental process to the products used in this comparison, it is possible to
identify the end of life management and the material sourcing as the building life
cycle stages, in which significant improvement could be achieved.
1 https://epd-australasia.com/?s=plastic&post_type=epd
Table 2 Environmental impacts of two building materials similar to Microtimber. Data are taken
from the Australian EPD database (see footnote 1). In this analysis only the global warming potential
(GWP) was considered, defined as the kg of carbon dioxide emitted by the process, the total use
of non-renewable primary energy sources (PENRT) and the total use of renewable energy sources
(PERT). Impacts are defined per kg of final product
Comparartive analysis
Product name Timber particle board Reconstituted timber Microtimber
18 mm panel (wood-plastic wood-plastic
composite) composites
Components Softwood 84.1% + urea Wood powder 52.5% + Wood powder: up to
formaldehyde 5.9% + polyvinyl chloride 28% + acrylonitrile
melamine formaldehyde 38.2% + calcium butadiene styrene
0.8% + paraffine wax carbonate 5.8% + (ABS): 81% + maleic
0.5% + lamination pigments 2.5% anhydride: 1%
paper 1.3% +
ammonium sulphate
0.1% + water 7.3%
Recycled content Variable—depending on 20% of the feedstock 100% from waste:
of wood availability mix comes from either by-product of timber
internal recycling or industry (saw dust) o
offcuts other recycled sources
such as furniture etc.
Life cycle A1 + A2: materials A1 + A2: Raw materials A1 + A2: Wood powder
process transported by truck are delivered by truck, delivered by truck from
from nearby producer, no data available on nearby producer,
usually sawmill proximity usually sawmill.
A3: wood particles are A3: materials are mixed Recycled ABS pellets
chipped (or flaked) and and extruded as from nearby supplier.
then mixed with resin granules, granules are A3: Wood powder is
and wax, the mixture is mixed with crap and ground and sieved to
pressed and cut, drying extruded (extrusion desired particle size and
process, trimming and process cooled with mixed with ABS and
sanding of the surface, water), aging process of maleic anhydride; the
lamination of the board 48 h, additional surface matrix is extruded under
A4: packaging—not treatment and embossed heat to form filament
better defined of wood grain and then 3D printed.
B: short life span A4: packaging in A4: not defined
C: the board can be cardboard boxes and B: varies depending on
recycled but not define shipped by sea application and
data available on the B: long life span composition
percentage C: part of the product is C: the product is 100%
shipped again to the recyclable, it can be
factory for being reground and reprinted.
recycled, the rest is
landfilled
(continued)
Table 2 (continued)
Pros and Cons EPD of the type of EPD of a specific EPD not available
material, data averaged product, data available
form available sources are refined
A + factories close to − factory in China, − still a non-industrial
sawmill and primary resulting in higher process, high impacts
wood suppliers transportation cost from production
resulting in lower + possibility to increase + use of waste materials
environmental impacts the content of recycled + use of variable matrix
+ ideally, high wood
percentage of recycled
wood form the nearby
B − short life span + long life span + variable life span and
(evaluated around 5 (evaluated > 10 yrs.), low impacts from
yrs.), resulting in higher resulting in lower cost maintenance
impacts for replacement for replacement
C-D − difficult to reuse as it + possibility to recycle + product completely
damages easily part of the product recyclable
− recycling process is − recycling needs + product can be re-used
difficult transportation back to as it is highly durable
the factory site (China)
LCA phases A, C, D A, C, D
Impacts for GWP: −11.1 kgCO2 eq GWP: 1.44 kgCO2 eq The environmental
phase A1-A3 PENRT: 126 MJ PENRT: 32.3 MJ impacts of Microtimber
PERT: 243 MJ PERT: 9.68 MJ are expected to fit
in-between these two
products
Impacts for Not considered GWP: 0.22 kgCO2 eq
phase A4 PENRT: 1.86 MJ
PERT: 0.036 MJ
Impacts for Not considered GWP: 0.01 kgCO2 eq
phase C1 PENRT: 0.28 MJ
PERT: 3.12e−3 MJ
Impacts for Not considered
phase C2
Impacts for GWP: 19.9 kgCO2 eq
phase C3 PENRT: −16.6 MJ
PERT: −193 MJ
Impacts for GWP: 2.09 kgCO2 eq
phase C4 PENRT: 15.4 MJ
PERT: 0.956 MJ
(continued)
Table 2 (continued)
Impacts for 1. scenario 100% 1. real scenario: 7.5%
phase D landfilled recycled (sent to
GWP: −0.0747 production factory),
kgCO2 eq 92.5% landfilled
PENRT: −0.852 MJ GWP: −0.072 kgCO2 eq
PERT: −0.0957 MJ PENRT: −1.93 MJ
2. scenario 100% PERT: −0.603 MJ
recycling: board is
shredded into woodchip
and re-use for new
particle board
GWP: −2.55 kgCO2 eq
PENRT: −31.2 MJ
PERT: −45.1 MJ
The biggest potential for improvement that Microtimber could offer concerns the
raw materials sourcing at stage A. The wood powder used is a waste- or by-product of
other production processes, mainly saw mills, and this means that zero virgin wood
is used in the manufacture. The other component—ABS plastics—is a recycled
material and also uses zero natural resources. Materials made with recovered wood
and plastics have positive effects on LCA, if compared to non-wood products or
products obtained from virgin wood. Direct effects can be achieved in relation to the
reduced environmental impacts of phase A, and indirect effects include the reduction
of land use, as less timber is harvested, and land can be used differently as well as
the reduction of land use for landfill (Sathre and Gustavsoon 2006). On the other
hand, Microtimber has high production impacts, as the process it is not optimized and
industrialized. The latter is an area for future research when the material is developed
towards specific industry applications.
At the end of their life, timber products can be burnt to produce energy, re-used
as solid material when possible (e.g. solid timber frames), reprocessed in new timber
products, pulped to form paper products or landfilled. Although landfilling is pro-
hibited in many parts of Europe (Dodoo et al. 2009), it is still the standard scenario
for the Australian construction sector, indicating that the waste hierarchy of timber is
still underdeveloped and requires improved management and innovation. Microtim-
ber offers the opportunity to enhance the life cycle management of timber composite
products both in phase A and phases C-D. Ultimately, it offers opportunities for
closing the loop of product lifecycles and advancing a circular Australian economy,
akin to the European model.
5 Conclusion
Research into the development of 3D printable wood-plastic composites made from

waste materials is still in its infancy. The possibilities that such products can open up
for improving design options, construction processes and environmental sustainabil-
ity are evident. They need to be developed towards industrial applications and require
the expertise of multiple disciplines including architecture, structural engineering,
bio-engineering and commitment from industry. The Microtimber research project
provides important directions for future investigations.
Acknowledgements This research was conducted in the context of the multidisciplinary research
project Microtimber: Development of a 3D printed, gradient timber panel composed of forestry
waste and by-products (2015–2019). Funded by Forest and Wood Products Australia (FWPA).
The authors would like to thank students and staff who contributed to this research project: Yerong
Huang, Jordan Girdis, Yicheng Todd Zhou, Eduardo De Oliveira Barata, and Pamela Kahwajy. We
also thank Susana Alarcon Licona from the Faculty of Architecture, Design and Planning’s DMaF
lab for her excellent support in robotic fabrication.
References
Blengini GA, Di Carlo T (2010) The changing role of life cycle phases, subsystems and materials
in the LCA of low energy buildings. Energy Build 42:869–880
Cabeza LF et al (2014) Life cycle assessment (LCA) and life cycle energy analysis (LCEA) of
buildings and the building sector: a review. Renew Sustain Energy Rev 29:394–416
Dodoo A, Gustavsson L, Sathre R (2009) Carbon implications of end-of-life management of building
materials. Resour Conserv Recycl 53:276–286
EN 15643-2:2011 Sustainability of construction works. Assessment of buildings. Part 2 Framework
for the assessment of environmental performance
EN 15804:2012 Sustainability for construction works. Environmental product declarations. Core
rules for the product category of construction products
EN 15978:2011 Sustainability of construction works. Assessment of the environmental performance
of buildings. Calculation method
European Commission. Task Group 3: construction and demolition waste. Final report. Brus-
sels; 2001. Accessed on 27 Apr 2018 at: https://ec.europa.eu/docsroom/documents/20509/
attachments/1/translations/en/renditions/native
Fritts H (2012) Tree rings and climate. Elsevier Science
Hafner P et al. (2015) Variations in environmental signals in tree-ring indices in trees with different
growth potential. PLoS ONE 10(11)
Huang Y, Unpublished thesis (2017) Relevant parts and additional details have been published in
Huang Y, Girdis J, Dong A, Proust G, Löschke S (2017) Designing Material Performance: Inves-
tigating the use of Australian hardwoods in 3D printed wood-plastic composites. Proceedings
of the 5th Annual International Conference on Architecture and Civil Engineering (ACE 2017),
Singapore, 8–9 May 2017, pp 339–343
IPCC (2014) Climate Change 2014: synthesis report. Contribution of working groups i, ii and iii to
the fifth assessment report of the intergovernmental panel on climate change [Core writing team,
R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, p 102
ISO 14025:2006 Environmental labels and declarations—type iii environmental declarations—prin-
ciples and procedure
ISO14040:2006. Environmental management—life cycle assessment—principles and framework.

CEN, Brussels
ISO 14040:2006. Environmental management—life cycle assessment—requirements and guide-
lines. CEN, Brussels
Jonsson O et al (2008) Consumer perceptions and preferences in solid wood, wood-based panels,
and composites: a repertory grid study. Wood Fiber Sci 40(4):663–678
Jovanovic I (2015) 3d printing using wood composite materials. Machine Design 7(1):23–30
Kolb J (2008) Systems in Timber engineering: loadbearing structures and component layers.
Birkhäuser, Basel, pp 32–33
Mai J, Thistleton C, Löschke S, Proust G, Dong A (2016) Towards a new techno-aesthetic paradigm:
Experiments with pattern, texture and colour in 3D-printed wood-plastic composites. In: Bourell
D (ed) Proceedings of 27th Annual International Solid Freeform Fabrication Symposium—An
Additive Manufacturing Conference, Austin, August 2016
Mai J, Girdis J, Proust G, Dong A, Löschke S (2017) Shades of wood: The effects of temperature
variation on the appearance and physical and mechanical properties of 3d printed wood-plastic
composites. In: Anderson M, Anderson P (eds) 5th Annual International Conference on Archi-
tecture and Civil Engineering, Global Science and Technology Forum, Singapore, May 2017,
246–250
Ortiz O, Castells S, Sonnemann G (2009) Sustainability in the construction industry: a review of
recent development based on LCA. Constr Build Mater 1:28–39
Panthapulakkal S, Sain M (2007) Agro-residue reinforced high-density polyethylene composites:
fiber characterization and analysis of composite properties. Composites Part A: Appl Sci Manuf
38(6):1445–1454
Passer A et al (2015) Environmental product declarations entering the building sector: critical
reflections based on 5–10 years experience in different European countries. Int J Life Cycle
Assess 20:1199–1212
Pritchard G (2004) Two technologies merge: wood plastic composites. Reinf Plast 48:26–29
Sathre R, Gustavsoon L (2006) Energy and carbon balances of wood cascade chains. Resour Conserv
Recycl 47(4):332–355
Scbi U (2006) Sustainable building and construction initiative: information note. DTIE, Paris,
France
Shiffman D (2012) The nature of code: simulating natural systems with processing. The magic
book project, and Perlin K (July 1985). An image synthesizer. SIGGRAPH Computer Graph-
ics 19 (0097–8930): 287–296. https://doi.org/10.1145/325165.325247
Submission by TDA (Timber Development Association of New South Wales) to the Forest Industry
Advisory Council (June 2015): 1–10 (1). Accessed on 27 Apr 2018 at: http://www.agriculture.
gov.au/SiteCollectionDocuments/forestry/australias-forest-policies/fiac/submissions/timber-
development-association-nsw.pdf
Part IV
Constructive Wisdom
and Realization Challenges
Seeing a Tree as a Prerequisite to Timber
Architecture
Rizal Muslimin
Abstract This chapter discusses the way in which a tree’s geometry is visually
perceived and constrained in architectural design. The discussion is centered at the
intersection of visual perception and geometry to explore the design opportunities
in a tree’s complex geometry. The first part outlines the representational framework
and the role of visual perception in seeing tree geometry. The second part repositions
timber architecture typologies using the framework discussed in the first part. The
third part illustrates alternative approaches for embedding tree geometry in timber
designs. This representational platform aligns well with the emerging movement on
organic timber architecture and is timely due to the recent advancements in timber
manufacturing and technology.
Keywords Tree · Visual perception · Embedding · Wood and geometry

Computational design
1 Introduction
Despite the diversity of trees’ geometry in nature, the initial references for designing
timber architecture are mainly fixated on engineered wood shapes. Timber architec-
ture, which has gained attention through tall buildings and wide-span structures in
the last few years, has been driven by a specific way of perceiving the geometry of a
tree that needs to be cut into boards or lumbers. Projects such as Treet Timber Tower
in Norway and Brock Commons in Canada demonstrate engineered woods’ capacity,
e.g. glulam and cross-laminated timber (Abrahamsen and Malo 2014; Poirier et al.
2016). The woods appear in standard geometry to suit various architectural shapes
and functions. In other cases, non-standard shapes in timber architecture are often
achieved by modifying wooden material with standard geometry through subtractive,
deformative, and additive techniques. This can be seen in the Centre Pompidou-Metz
R. Muslimin (B)
School of Architecture, Design and Planning, The University of Sydney, Sydney, Australia
e-mail: rizal.muslimin@sydney.edu.au

852 R. Muslimin
Fig. 1 Embedded design in tree geometry from the 18th-century ship’s futtocks (left) and chocks
(right) (after J. Adams 2013)
in France by Shigeru Ban, where the glulam beams were bent and milled to form a
woven roof structure (Brell-Cokcan et al. 2009; Scheurer 2010).
Nevertheless, effort to embrace the tree’s geometry in its natural form is evident in
the past. Ship hull curvature in 19th century Northern Europe by Sir Robert Pessing,
for instance, corresponds to the original tree’s curvature, with the futtocks cut and
segmented to fit the hull curvature (Adams 2013) (Fig. 1). Appreciation of a tree’s
natural geometry can also be seen in the modification of a tree in its milieu, for
example in the 17th century Willow Cathedral in Scotland by Sir James Hall, where
the willow’s elastic branches were deformed into a small Gothic cathedral (Reeve
2012).
More recently, Hall’s vision has been revived with modern technology, as can be
seen in the living plant tower by the Baubotanik Research Group in Stuttgart, where
the tree growth was modified to allow the branches to become an integral part of
the building’s structure (Ludwig et al. 2012). A similar approach by Robert Pessing
was achieved through digital fabrication in Hooke Park Woodchip Barn in the UK
by Architectural Association (AA) students, where the main trusses’ arch beam was
assembled by 25 forked beech branches, each in its natural form (Self 2016; Self and
Vercruysse 2017). The branches were 3D scanned to arrange the trusses and cut by
a robotic arm for more accurate joints.
Similar approaches have been demonstrated in other AA Hooke Park campus
projects, as well as in projects at the Zurich University of Applied Sciences, Switzer-
land and The Royal Danish Academy of Fine Arts, Denmark (Schindler et al. 2013,
2014). These projects achieved non-standard shapes with minimum compromise of
the natural geometry. They show what can be achieved from a tree’s natural geometry
in comparison to mass-produced timbers.
Seeing a Tree as a Prerequisite to Timber Architecture 853
While the use of digital technology offers accuracy and efficiency in timber fab-
rication, most computational research on wooden design is focused on structure and
production and less so on the methodology used to generate tree-driven designs.
This chapter argues that to better synthesize wooden design in timber architecture,
a representational framework is required to assist architects in perceiving the design
opportunities in a tree’s complex geometry. To benefit from such geometry, we need
to understand the way in which a tree might be visually perceived.
This chapter addresses this argument in three sections. Section two introduces a
theoretical framework and the role of visual perception in perceiving a tree’s geome-
try. Pertinent theories from perceptual studies and computational vision are discussed
and the methods used to represent a tree’s geometry are outlined. Section three ana-
lyzes timber architecture typology based on the way in which our visual perception of
the tree’s geometry constrains the design and vice versa. This analysis compares the
differences between the ways trees are commonly perceived and the alternative way
of perceiving trees that is not necessarily related to the standard building typology.
Section four provides examples of embedding a tree’s geometry in a house design to
synthesize alternative timber designs.
2 Theoretical Frameworks: The Algebra of Shapes

and Gestalt Laws
To navigate our analysis on perceiving a tree’s complex geometry, this section begin
by providing a framework from computational design, i.e. the algebra of shapes from
Shape Grammar theory, accompanied by a discussion on a classical theory in visual
perception, i.e. Gestalt laws to explain why we see a tree the way we normally see it.
This discussion is followed by some examples from the digital and manufacturing
process to show how visual fixation may continue to impact the way we use the
technology on timber design.
2.1 The Algebra of Shape Uij
The algebra of shape U ij maps an object based on its shape’s dimension (denoted
with i) and the space dimension where the shape is represented (denoted with j)
(Stiny 1993). For example, representing a shape in algebra U 12 means perceiving an
object as a one-dimensional shape in a two-dimensional space. The space dimension
confines and constrains the shape’s transformation. A cylindrical object in real space,
for instance, can be mapped into different algebras, such as U 33 , U 22 , U 23 , U 12 , and
U 13 (Fig. 2). If perceived in algebra U 33 , the cylinder can be transformed into a
three-dimensional shape, such as a conical, spherical, polyhedral, Euclidian, or non-
Euclidian shape.
854 R. Muslimin
Fig. 2 The algebra of

shapes and example mapping
of a cylindrical object
Two-dimensional shapes can also be perceived from a cylinder, e.g. a circle,

rectangle, or ellipse, by slicing the cylinder along a two-dimensional plane in algebra
U 22 . The two-dimensional cutting plane constrains the space for modifying the shapes
from the cylinder. If the cutting plane is perceived in a three-dimensional space, the
circle, ellipse, and rectangular shapes can be warped into shapes in algebra of U 23 ,
e.g. a dome, barrel-vault, or hyperboloid. In three-dimensional space, these two-
dimensional shapes can be further multiplied to occupy the original cylinder.
Similarly, lower dimensional shapes, such as lines and curves, can be perceived
in multiple ways. They can be retrieved from the circle’s, ellipse’s, and rectangle’s
edges in algebra U 12 . A line and curve can also be perceived from the warped shape’s
boundary, the cylinder’s axial line, or any curves that can be confined within a cylinder
in algebra U 13 . The zero-dimensional shape in algebra U oj will be discussed in the
later Sect. (2.3) as it represents symbolic data (digital bits) in the algebra of shapes.
The algebra of shapes framework proposes that an object might be temporarily
perceived into a different shape and space dimension in the mind, regardless of its
actual dimension in algebra U 33 . A paper, for instance, is a three-dimensional shape
(U 33 ) because it has a thickness, yet one can temporarily perceive it in algebra U 22 as a
planar surface, where the thickness does not contribute to the transformation process
(e.g. tiling), or in algebra U 23 , where the transformation requires three-dimensional
space (e.g. folding). Hence, developing an awareness of how an object is perceived
and represented is pivotal as it might limit or expand the opportunities to transform
it during the design and making process (Muslimin 2014).
Within this framework, the number of shapes that can be perceived from an object
is enormous. When seeing a tree’s geometry, one can shift back and forth across
algebras to see different shapes, i.e. as an assemblage of points, lines, surface, and
solid or a combination of them. Such perceptual diversity can be seen in sculpture
and painting, where a tree is perceived and modified in many ways.
Artists and sculptors take this liberty by embracing their tools and techniques
to project different ways of seeing a tree. By projecting their three-dimensional tip
brush in various ways onto the two-dimensional canvas, a painter can generate a new
kind of tree in algebras U 12 and U 22 (for illustration, see Piet Mondrian’s different
depictions of a tree in Red Tree, 1908–1910; The Flowering Apple Tree, 1912 and The
Grey Tree, 1912 (Wallace 1968)). A wooden sculptor materializes their perception
of the tree as a one-, two-, or three-dimensional shape onto the three-dimensional
trunk (e.g. with chiseling, sculpting, and engraving).
In architecture, however, a tree is perceived in a more specific way. While several
architectural experiments in the introduction highlight the different ways of per-
ceiving a tree (e.g. the willow cathedral, the baubotanik tower, and the barn), the
derivation of a tree’s geometry in mainstream timber architecture is mostly fixated
on cuboid shapes, as seen in lumber, plywood, and timber. This suggests that there
are other drivers that might influence this fixation, either internally within the human
perceptual system or externally in digital technology and the manufacturing process.
The following interprets these factors via Gestalt law and the way this operates in
computational vision, design algorithms, and the timber manufacturing process.
2.2 Gestalt Redux
According to the Gestalt laws, the human’s visual perceptual system tends to recog-
nize a ‘whole’ from the ‘parts’ (Koffka 1922; Wertheimer 1923). The law of Pragnanz
outlines several ways in which the parts can be grouped and perceived as a whole,
such as whether the parts are in a symmetrical arrangement, close to each other,
tend to form a continuous shape, or look similar to one another (Wertheimer 1938).
These principles imply that shapes in the lower algebra (U i ) will be grouped and
recognized as a whole in the higher algebra (U i+1 ) (Muslimin 2012). For example, a
row of points will be grouped into a line, a group of lines with similar directions that
are close together will be perceived as a face, and faces that suggest an enclosure
will be recognized as a solid.
Classical Gestalt laws also argue that in the case where different groupings are
possible, our mind will pick the “good” shape, for example, the one that is symmetri-
cal, has an enclosure, or is stable regarding its gravitational direction (Koffka 1922).
While vaguely defined, the consideration of what a “good” shape is might be related
to one’s cognitive capacity to find the simplest shape to be processed and memorized.
For example, a tree trunk could be recognized as a line rather than intricate branches
of cylinders because of the way it is perceived from a certain distance.
When looking closely, the tree’s three-dimensional appearance is more observ-
able, but when looking from far away, one could lose this three-dimensional quality
and perceive the trunk as a curve or line (U 13 ) (Fig. 3). From a much further distance
856 R. Muslimin
Fig. 3 Perceiving a “good” shape from a tree. A simpler geometry of a Eucalyptus tree model
is perceived in algebra U 13 , where the most straight and continuous lines are recognized, before
projected back into real space U 33
away, the trunk might become less salient and instead one might distinguish the
upper shape’s outline of a tree. This process is recursive across scales, for instance,
perceiving a group of branches based on simple geometries, e.g. a ‘Y’ shape, is more
manageable than memorizing the geometry of each branch, e.g. a meandering shape.
The order of recognizing a “good” shape and reproducing it might be different.
For example, when asked to draw a tree, some might start by drawing a straight
line first for the trunk and then the branches, followed by the leaves. Others might
start with the shape of the upper part of the tree (e.g. a cone or triangle for a pine
tree) and then draw the branches and the trunk. An artist might consider the tree’s
intricate geometry as the salient shape that uniquely characterizes the tree (e.g. the
meandering branches, the outline of the leaves, or the bark texture). Yet, when it
comes to visualizing it, the artist might start drawing the basic shape of the tree first
(e.g. the trunk line and upper part) before elaborating on the intricate geometries.
Fig. 4 Sensing a tree digitally. A tree is represented as a points field in U03 before being visualized
into its digital twin in higher algebra. While the tree can be perceived in a higher dimension, the
computation process, such as for structural analysis in finite element modeling, continues operating
in algebra U0j
Therefore, Gestalt laws do not always work consistently as the order of “good”
shape recognition and reconstruction could go back and forth between the ‘whole’ and
the ‘parts’ across algebras. It also does not explain, for example, the duality of seeing
a tree, both as a line and plane, or that the role of visual perception in drawing might
be independent from creative seeing. Indeed, Gestalt laws are generally considered
obsolete and lacking mathematical rigor.
Nevertheless, Gestalt principles such as “visual illusions” and “perceptual switch-
ing” remain influential to current vision and cognitive science studies. Some
researchers improved Gestalt principles by testing the hypothesis with advanced
technology (Wagemans et al. 2012a, b), others by challenging its premises. For
example, computational vision researchers have criticized Gestalt psychology for its
lack of mathematical evidence and subsequently conducted an in-depth study on the
systemization of the parts-to-whole grouping processes (Marr et al. 2010). Marr, a
pioneer in computational vision, did not negate the Gestalt influences in computer
vision, however.
For our purpose, navigating Gestalt principles in relation to the algebra of shapes
helps to investigate the perceptual issues of perceiving a tree’s geometry across
disciplines in architectural timbers. The following section illustrates the magnitude
of Gestalt principles in digital and manufacturing technology.
2.3 A “Good” Shape for a Digital Tree
In computational vision, a tree is digitally represented as a set of points captured via

sensors and reconstructed with computer vision algorithms (Fig. 4). For example,
in LIDAR technology, a tree’s exterior is first sensed as a set of points in space
(U 03 ) that are then reconstructed along a linear arrangement U 01 (to form a line) and
planar U 02 (to form a face) (Rosette et al. 2011; Tittmann et al. 2011). They are then
grouped together as a whole to allow the user to ‘see’ the digital twin of a tree (as an
assemblage of line U 13 and mesh U 23 ).
858 R. Muslimin
In object recognition algorithms, a two-dimensional image of a tree is represented

as a two-dimensional point field (U 02 ) imbued with graphical properties (e.g. bright-
ness). The algorithm will then detect the edges, brightness, and directionality of
each pixel and deduce a group of shapes to be recognized as a tree (Marr et al. 2010;
Strahler et al. 2008). Similar logic is applied when a tree’s internal properties are
sensed using CT scan technology, for instance, to analyze the cambium lines through
a layer of two-dimensional slices of faces (U 23 ) comprised of the voxel (U 02 ) (Born
et al. 2016).
Thus, through the computer, we see a tree how we want to rather than how the
computer ‘sees’ it. This sensing technology shows that the digital recognition mech-
anism is agnostic as to which algebra a tree must be perceived through, except that
it must start from point(s)—U 0j . The higher dimensional appearance is constructed
computationally using a human-defined algorithm to perceive a “good” shape of a
tree that is more manageable for the user’s mind and the way in which they use the
computer, e.g. by embedding vectors with magnitude and direction to analyze the
trunk’s skeleton structures (U 13 ) and using a Cartesian plane to analyze the trunk’s
sectional profile (U 23 ) (Raumonen et al. 2013; Mattheck 1998).
We can then safely say that a “good” shape for both the computer and the user
is the discrete shape as this is countable and therefore computer-friendly. It is also
atomic, which makes it easier for the user to perceive it as a whole. For example,
to understand the plant’s growing mechanism, L-system and Venation algorithms
represent a plant as discrete lines and points, respectively (Fig. 5). While these logics
are applied for different scales (one for tree branches, another for leaf veins), they
show how different the ways of perceiving a plant across algebra could lead to a
different logic and outcome.
In L-systems, the tree trunk and branches are simplified into their axial lines (U 1 )
which is then incorporated with additive transformation to generate a variety of tree
typologies (Prusinkiewicz and Lindenmayer 1996). Starting with a line as an initial
shape for the trunk, L-system algorithms add new lines that represent the branches,
then reproduces itself recursively into different scales and orientations to form a tree.
In a leaf venation logic, however, the vein line of a leaf grows based on the point
distributions (U 0 ) (Runions et al. 2005). Unlike the L-system, where the tree topology
is predefined, there is no line to begin with in Runions’ venation logic. Instead, it
starts with a set of point fields (representing the hormone-auxin source) from which
a root (also represented by a point) starts to grow into vein lines by computing its
proximity to the neighboring points (Fig. 5).
With lines (U 1 ), the L-system logic is relatively easier to perceive and compute
compared to the Venation logic, which is driven by point position (U 0 ) and an iterative
trigonometric operation. Yet, the venation logic’s closer analogy to the biological
mechanism (i.e. how the vein reaches the food) and its reliance on computer power
to process the random point locations iteratively (i.e. where the next food exists)
result in a more organic and realistic output. Nevertheless, the lines and points are
treated as discrete shapes in both algorithms and are therefore compatible for bit
calculation in a computer.
Fig. 5 L-system and venation logic representing a tree growth. The L-system focuses on a line
composition and predefines how the lines will be composed with a set of additive rules (top), while
the venation logic relies on the algorithm’s capacity to develop its own line composition by sensing
the neighboring points, which represents the food source (bottom)
Discrete representation is the backbone of digital computation and can help rec-
ognize the tree in ways that might not align with the Gestalt principles or the way in
which humans typically recognize a tree. A dataset of discrete shapes for a computer
is open to interpretation, for example, a computer might consider the fork, the branch,
or the bark as the discrete shape and can assemble them in different structures or
hierarchically (Fig. 6). The computer’s inexperienced view of ‘seeing’ the tree can
produce different ways of looking at a tree to our overly matured eyes.
2.4 A “Good” Shape for the Timber Industry
Seeing a tree’s geometry back and forth across the algebra of shape becomes more
challenging in the wood manufacturing process due to the physical constraints and the
mechanization process of the timber industry. A long tradition in the timber industry
has crystalized a consensus on perceiving a tree into its cylindrical form (i.e. tree logs)
860 R. Muslimin
Fig. 6 Tree-data and data-tree: exploring parts of the tree through different types of discrete shape
data and the dataset’s structure (shown in graph) to allow various interpretations of a tree’s geometry
and outputs material into a family of cuboid shapes (e.g. timber panels, lumbers, and
plywood) (Ramage et al. 2017). This consensus governs the manufacturing phases
(e.g. debarking, cutting off, and peeling) and the decision-making process to plant a
type of tree that is manufacturing-friendly. Through slicing, cutting, and peeling, the
cuboid is imposed onto the initial perception of the log’s cylindrical shape and even-
tually replaces our initial cognition of a tree’s geometry when the parts are assembled
and bundled into lumber, planks, plywood, CLT, glulam, or board (Maloney 1996).
This manufactured cuboid timber sets the perception on perceiving the wooden
materials within linear (U 1 ) and planar algebra (U 2 ) upon which architects base their
designs. This process does not suggest that a manufacturer’s perception is limited to
cuboid shapes, but rather the fixation on cuboid shapes causes them to process the tree
in certain ways. The woodworking machinery can shape the timber in many ways,
e.g. as cylinders, spirals, lines, circles, etc., with which one can form many shapes
(Ernest 1989). Some of this machinery is the same as is used in other industries, such
as car and ship building, which demands more shape variety. Yet, once the priority for
a particular geometry is set, the tools are constrained to allow the intended geometry
to emerge from the tree.
3 Perceiving a Tree in Architectural Timber
The previous sections discuss the way a tree’s geometry is perceived through digi-
tal sensing, computational algorithms, and the timber manufacturing process. They
highlight the key role of visual perception in constraining the geometry from the
material point of view. This section extends this discussion from an architectural
design viewpoint by outlining the factors that might contribute to similar perceptual
prioritization.
To accommodate the different level of complexity in perceiving architectural
shapes, the algebra of shapes framework can be extended by including maximum
or minimum shapes as additional parameters for perceiving an object (Stiny 2006).
Maximum shapes refer to shapes that we perceive as they occupy their full size,
such as the maximum length of a line in algebra U 1 , the maximum area of a surface
in algebra U 2 , and the maximum volume of a solid in algebra U 3 . Therefore, the
number of maximum shapes of an object in each algebra is finite. A maximum shape
of a cube, for example, consists of one solid, six faces, and 12 lines as these are the
shapes that occupy their maximum dimensions, whereas a cylinder has one solid,
one face, and two curves.
Minimum shapes refer to all shapes that one can perceive from an object that has
less than or equal to its maximum size. Hence, the number of minimum shapes within
an object is infinite. For instance, the amount of line segments that can occupy a line
is endless as there are many ways to split the line based on any splitting points that
can be embedded along the line. Not to mention the number of overlapping lines that
can be perceived from such a line. Hence, the minimum shape operates in the same
shape (i) and space (j) in algebra U ij . In the overlapping line example it is U 11 as we
862 R. Muslimin
perceive other possible one-dimensional lines on an existing line. The same goes for
the other algebras, for example, when perceiving smaller faces in a surface in U 22 or
chunks of a solid object in U 33 .
This framework holds that we not only perceive an object across different algebra,
but also in-between the maximum and minimum shapes in each algebra. Taking an
analogy from an architectural design process, an architect might start with a one-
line drawing when designing a floor plan. The first few lines could be perceived
as a maximum shape where one line refers to one wall. Yet, this perception could
shift in the next moment, when the line is perceived as a set of minimum shapes,
e.g. overlapping lines or several line segments in U 11 . They become more apparent
later when some segments are developed into two-line drawings (e.g. to define a
wall-thickness), while others are erased (e.g. to indicate openings in a wall).
The same goes for other algebras, for instance, during a brief moment before
more interior walls are drawn, the empty rectangle might be embedded with smaller
rectangles in algebra U 22 (e.g. for subdividing the open plan space to create more
rooms; see illustrations in Laseau 2001). While this analogy seems trivial, the mag-
nitude of perceiving maximum and minimum shapes and how this may influence
one’s preference in the design process should not be underestimated, particularly
when considering the required materials for the construction process.
3.1 Seeing a House and Its Parts
A simple house in Fig. 7 provides an example of such magnitude where the house is
perceived as an assemblage of points, lines, faces, and solids. One of many possible
ways to see the house is by considering one maximum solid, i.e. occupying the whole
volume of the house. The number of the maximum faces for this view is seven and
the maximum number of lines is 15. Another way of perceiving such a house is to see
the roof’s base plane split the house into two solids. Note that this means we perceive
two overlapping lines where one set of lines belongs to the roof and the other lines
belong to the wall. Already, we see a contradiction in perceiving the house. One has
15 lines for one solid and another has 34 lines derived from the two solids (see the
below images in Fig. 7). As the line would be translated into posts or beams and the
face into the wall panel, each view constrains the later design development in timber
construction.
For example, would each post and beam be made by one solid wooden beam or
a pair of two spaced beams? As spaced beams consist of two lumber beams, the
question could be paraphrased into: would one rather develop the house with 15
or 34 lumbers? Another question could be whether the wall should be constructed
with timber logs or framed with several timber studs and plywood panels? Such
questions require further translation across algebra because the descriptions (i.e. the
words “post”, “beam”, and “panel”) are embedded with a particular way of seeing the
shapes. In timber log construction, a face is perceived as an assemblage of horizontal
lines stacked on top of one another (Fig. 8). In timber frame construction, a face
Fig. 7 Embedding maximum and minimum shapes in a design. Different perceptions offer different
consequences for construction, such as a different amount of timber and connections
Fig. 8 Consequences of perceiving the shape in algebra U 13 differently that leads to a different
number of components and construction methods
is seen as multiple layers of subdivided faces (U 23 ) and some of the layers can be
perceived as an array of vertical studs or horizontal cladding panels (Steiger 2007).
The perceptual shift in Fig. 8 leads to various design outputs that are compatible
with the typical engineered timber shapes. They are not, however, driven by the tree’s
geometry in its natural shapes as such geometry is ignored the moment a house is
planned to be built with the cuboid timbers. As can be seen in Fig. 9, the perceptual
process of seeing different shapes in a tree becomes more converged and singular
as it is translated into the manufacturing process. Similar reduction occurs in the
house’s design as the numerous ways to see a house are reduced by embedding the
shapes with the mainstream timber products. Both views conclude in a one-on-one
relationship where the tree’s straight lines correspond with the house’s straight lines.
This relationship aligns with Gestalt’s principle regarding the way in which a “good”
shape is perceived as a whole in the mainstream timber architecture.
864 R. Muslimin
Fig. 9 A line for a line: a converged perception between tree geometry and architectural timber is
accumulated into a one-on-one relationship between the maximum lines in a tree and the maximum
line in a design
4 Seeing a Tree in a House and a House in a Tree
The relationship diagram in Fig. 9 shows an important map of the shape relationship
between architectural design and timber material in regard to visual perception.
Generally, the way one perceives a house during the schematic phase might cascade
from larger to smaller algebra, e.g. from U 3 to U 2 to U 1 . The house’s shape might first
be perceived as a whole, then the perception shifts to the parts (wall column opening,
etc.). Later, as the design is put forward for the construction phase, the perception
might be reversed from smaller to larger algebra for each part, e.g. U 1 to U 2 to U 3 ,
where all lines and faces are imbued with thickness and material properties (e.g. for
designing the connection). A similar process occurs in the timber industry, where a
tree is perceived as a whole in nature (U 3 ) before being reduced into smaller algebras
U 2 and U 1 , i.e. one line for a straight cylinder. Then, in the manufacturing process,
the perception is shifted back to U 3 (e.g. for producing plywood, lumber, and timber).
In perceiving the maximum and minimum shapes, a typical timber manufacturing
process tends to focus on the maximum shapes in each algebra, for instance, more
effort is applied to processing the straightest trunk, whereas the house design process
switches between maximum and minimum shapes. Visual perception in the early
phase may focus more on the maximum shapes as these can be observed through one-
line sketches or thick-less CAD models in the schematic design drawing. Minimum
shapes appear more in the later phase by splitting the maximum shape to suit more
functions, the size of the materials, or the connection types.
As such, the flow along algebra in these processes seems to occur linearly. This
suggests a procedural fixation, which is constrained by the order of sequences
imposed during the design process, e.g. from the schematic to the development
phase, or from one-line sketches to construction drawing, or from the skin to skele-
ton structure (this fixation is denoted with the arrows in Fig. 9.). There appear to be
perceptual boundaries between these phases that might constrain one’s perception
and modification of a shape. For example, our perception of a shape that has been
defined as a maximum shape in algebra U 33 (e.g. using single solid timbers for the
roof’s beam) may become solidified in such an algebra and is less likely to be per-
ceived as the other algebra or minimum shape (e.g. to consider other alternatives for
the structure, such as using a truss or spaced beam).
To better incorporate the tree’s geometry in the design, such fixations need to be
minimized. We outline three preliminary approaches. The first links the minimum
shapes from both the tree and the design geometry. As the number of minimum
shapes is larger than the number of maximum shapes, these links will decentralize
the previous one-on-one relationship and change it to a many-to-many relationship
(Fig. 10). In this relationship, the design shape is no longer associated with only
straight lines, but with any shape that can be extracted from a tree. Second, ensuring
the perceptual translation between the minimum and maximum shapes can go back
and forth across scales and design phases. For example, the shapes involved during
the construction detailing phase (e.g., joinery), which normally occurs in the later
phase, do not have to be limited to cuboid timber as one can find alternatives in the
natural geometry.
Third, we minimize the fixation by allowing shapes to be freely perceived from
higher to lower algebra and from lower to higher algebra. The former would allow
a solid shape, for instance, to be perceived as an assemblage of faces, lines, and
points. Yet, the latter process might lead to a wider variety of designs as the higher
dimensional translation of the shape does not have to follow the original one. The
higher dimension shape of a set of points, for example, could be any lines and/or
faces that share the points as its boundaries (see the two images at the bottom right in
Fig. 10). The shape conversion from high to low algebra allows one to see different
ways to discretize a tree or a design, while the conversion from low to high algebra lets
one see how a shape could be coupled with various higher-dimensional shapes (e.g.
for finding a different tree’s branch that can be supported by several points in a
design). This perceptual switch would increase the opportunity to match the house
and the tree’s geometry.
4.1 Hybrid-Timber
The output from Fig. 10 can be identified by different characteristics, such as techni-
cal, organic, and hybrid configurations (Fig. 11). The technical configuration refers to
866 R. Muslimin
Fig. 10 The matrix of seeing maximal-minimal shapes across the algebra of shapes in tree geometry
and architectural design works in a non-linear manner. The shape can be perceived from either higher
to lower algebra or vice versa. By expanding the different ways of seeing the design and the tree,
the possible ways the tree geometry can fit the design increase
the basic configuration normally found in the mainstream architectural timber where
a tree is seen in its maximum shape and assembled throughout the linear process. In
this configuration, the majority of the timber elements for a house are derived from
the typical ways of seeing a tree (i.e. cuboid).
The organic configuration refers to the natural composition that utilizes the min-
imum shapes from a tree’s geometry, including any shape within the algebra of
shapes that can be perceived from a tree. Through the matching procedures, the
tree’s shapes would find the part of a house that it sees fit to occupy. The number
of shapes from this many-to-many relationship is numerous, for example, there are
many tree branches that can be aligned and oriented to fit an intersection of a post
and the beams of a house. Similarly, there are many intersections in a house that can
be occupied by a tree branch.
The hybrid configuration refers to the combination of the two, where both min-
imum and maximum shapes are utilized to compose a house design. This could
become handy as the organic configuration might compromise the house’s function-
ality due to its arbitrary forms. The straight lines or the planar surfaces of the cuboid
timbers could neutralize this issue, e.g. by using the straight line for the beam and the
organic line for the joints, or by using the organic line for the beam and the straight
line for the joints.
Fig. 11 Hybrid approach to utilizing the tree’s geometry and manufactured timber in design. The
design in the left image characterizes a technical configuration where the structure mainly consists of
cuboid timbers and organic forks from the tree serve as the timber joints. The design on the far right
characterizes the organic configuration, where the design emerges from the tree’s geometry. The
designs in-between the two characterize the hybrid between the technical and organic configurations.
The below tree images show the parts of the tree that are being used for each design. The magenta
color indicates the fork of the branches for the joints; the thin, light grey line indicates tree branches
that were cut off and that have become part of the house; and the dark grey line indicates the parts
of the tree that were not touched. By using only parts of the tree’s geometry, this approach could
let the tree regrow and provide alternative designs for architectural timber
5 Conclusion
This chapter has discussed the way in which the visual perception of a tree’s geome-
try affects the timber manufacturing process and the mainstream paradigm of using
timber in architectural design. By considering the theoretical frameworks of shape
representation and visual perception to interpret how we see things in a certain
way, we can anticipate their influence on the digital design and manufacturing pro-
cesses. Several examples from both material and design viewpoints have illustrated
how perceptual fixations can lead to particular timber architecture design. Percep-
tual fixation from the material view is quite resilient and can therefore narrow one’s
approach to using technology related to tree geometry, as discussed in relation to com-
puter vision, algorithmic design, and manufacturing machinery. Fixation in design
shows a similar inclination in terms of how perceiving a group of parts as a whole can
limit the variation in timber shapes and might prevent architects from considering
the tree’s geometry in nature. To minimize this fixation, this chapter have showcased
alternative ways of perceiving trees and design geometry that could help the designer
to reevaluate the other parts of a tree that might be applicable for building and to
rethink the parts of a building that could harness the underutilized parts of a tree’s
geometry.
868 R. Muslimin
Acknowledgements The tree images in Figs. 3–11 are based on a model provided by Xfrog (Xfrog
2010), a courtesy of TurboSquid.com.
References
Abrahamsen RB, Malo KA (2014). Structural design and assembly of “Treet”—A 14-storey timber
residential building in Norway. In: World conference on timber engineering 2014, p 8
Adams J (ed) (2013) Carvel building in retrospect. In: A maritime archaeology of ships, innovation
and social change in late medieval and early modern Europe. Oxbow Books, pp 175–192
Born L, Jonas FA., Bunk K, Masselter T, Speck T, Knippers J Gresser GT (2016). Branched structures
in plants and architecture. In: Biomimetic research for architecture and building construction.
Springer, Cham, pp 195–215
Brell-Cokcan S, Reis M, Schmiedhofer H, Braumann J (2009) Digital design to digital produc-
tion: flank milling with a 7-axis robot and parametric design. In: 27th ECAADe Conference
proceedings on computation: the new realm of architectural design, pp 323–330
Ernest E (1989) Process for manufacturing corrugated plywood composites. US4816103A
Klein J (2017) In Poland’s crooked forest, a mystery with no straight answer—The New York Times
Koffka K (1922) Perception: an introduction to the Gestalt-theorie. Psychol Bull 19:531
Laseau P (2001) Graphic thinking for architects and designers. Wiley
Ludwig F, Schwertfreger H, Storz O (2012) Living systems: designing growth in Baubotanik. Archit
Des 82:82–87
Maloney TM (1996) The family of wood composite materials. For Prod J Madison 46:18
Marr D, Ullman S, Poggio TA (2010) Vision: a computational investigation into the human repre-
sentation and processing of visual information. MIT Press
Mattheck C (1998) Design in nature: learning from trees. Springer-Verlag, Berlin
Muslimin R (2014) EthnoComputation : on weaving grammars for architectural design. Thesis,
Massachusetts Institute of Technology
Muslimin R (2012) Recursive embedding of gestalt laws and shape grammar in the weaving design
process. Presented at the the 30th eCAADe, Prague, Czech Republic, pp 443–449
Poirier E, Moudgil M, Fallahi A, Staub-French S, Tannert T (2016) Design and construction of a
53-meter-tall timber building at the University of British Columbia. In: Proceedings of the world
conference on timber engineering, Vienna, Austria
Prusinkiewicz P, Lindenmayer A (1996) The algorithmic beauty of plants. Springer-Verlag, New
York
Ramage MH, Burridge H, Busse-Wicher M, Fereday G, Reynolds T, Shah DU, Wu G, Yu L, Fleming
P, Densley-Tingley D (2017) The wood from the trees: the use of timber in construction. Renew
Sustain Energy Rev 68:333–359
Raumonen P, Kaasalainen M, Akerblom M, Kaasalainen S, Kaartinen H, Vastaranta M, Holopainen
M, Disney M, Lewis P (2013) Fast automatic precision tree models from terrestrial laser scanner
data. Remote Sens 5:491–520
Reeve MM (2012) Gothic. Stud Iconogr 33:233–246
Rosette J, Field C, Nelson R, DeCola P, Cook B (2011) A new photon-counting lidar system for
vegetation analysis. Proc Silvilaser, 1–8
Runions A, Fuhrer M, Lane B, Federl P, Rolland-Lagan A-G, Prusinkiewicz P (2005) Modeling
and visualization of leaf venation patterns. ACM Trans Graph (TOG) 24:702–711
Scheurer F (2010) Materialising complexity. Archit Des 80:86–93
Schindler C, Tamke M, Tabatabai A, Bereuter M (2013) Serial branches
Schindler C, Tamke M, Tabatabai A, Bereuter M, Yoshida H (2014) Processing branches: reacti-
vating the performativity of natural wooden form with contemporary information technology. Int
J Archit Comput 12:101–115
Self M (2016) Hooke Park: applications for timber in its natural form. In: Menges A, Schwinn T,
Krieg OD (eds) Advancing wood Architecture: a computational approach
Self M, Vercruysse M (2017) Infinite variations, radical strategies. Fabricate, 30–35
Steiger L (2007) Basics timber construction, 1st edn. Birkhäuser Architecture, Basel, Boston
Stiny G (2006) Shape: talking about seeing and doing. The MIT Press
Stiny G (1993) Boolean algebras for shapes and individuals. Environ Plan B Plan Des 20:359–362.
https://doi.org/10.1068/b200359
Strahler AH, Jupp DL, Woodcock CE, Schaaf CB, Yao T, Zhao F, Yang X, Lovell J, Culvenor D,
Newnham G (2008) Retrieval of forest structural parameters using a ground-based lidar instrument
(Echidna®). Can J Remote Sens 34:S426–S440
Tittmann P, Shafii S, Hartsough B, Hamann B (2011) Tree detection and delineation from LiDAR
point clouds using RANSAC. Proc SilviLaser
Wagemans J, Elder JH, Kubovy M, Palmer SE, Peterson MA, Singh M, von der Heydt R (2012a)
A century of Gestalt psychology in visual perception: I. perceptual grouping and figure-ground
organization. Psychol Bull 138:1172
Wagemans J, Feldman J, Gepshtein S, Kimchi R, Pomerantz JR, van der Helm PA, van Leeuwen
C (2012b) A century of Gestalt psychology in visual perception: II conceptual and theoretical
foundations. Psychol Bull 138:1218
Wallace IH (1968) Piet Mondrian : the evolution of his neo-plastic aesthetic 1908–1920. University
of British Columbia. https://doi.org/10.14288/1.0104386
Wertheimer M (1938) Laws of organization in perceptual forms
Wertheimer M (1923) A brief introduction to gestalt, identifying key theories and principles. Psychol
Forsch 4:301–350
Xfrog (2010) XfrogPlants blue-gum eucalyptus [WWW Document]. TurboSquid. URL https://
www.turbosquid.com
Constructing Correctly in Wood: New
Insights into Timber Technology
Approaches Through Purist
and Liberalist Schools of Thought
Gabriel Tang and John Chilton
Abstract Conventionally, technology-based articles focus on methods by which

architects and engineers designed and built to present new methods, materials to
evidence novelty in technical terms. This paper does not do that. Instead, through a
current overview of past and present timber practices, it will present a new cultural
perspective by looking at timber technology from purist and liberalist approaches.
Indicating a moralistic sensibility of what “constructing correctly” in wood means to
them, with these two attitudes implying inherent values, this paper seeks to project
a new cultural dimension on technology. More importantly, the approaches convinc-
ingly reflects our relationship with digital technology, as timber culture and tradition
come to terms with the inevitability of the digital age.
Keywords Complex timber structures · CNC and woodworking technology

Wood design · History of representation in wood design · Timber gridshells
Wooden architectures and timber structures survey
1 Introduction
Recent research by the authors into the design, fabrication and use of timber gridshells
revealed contrasting polemical views of architectural timber use. Seen in the devel-
opment of doubly-curved, thin shell structures over the last 50 years, these attitudes
resonate with Pier Luigi Nervi’s views on the “correct” application of technology,
and are here categorized as purist and liberalist approaches to timber technologies.
G. Tang (B)
Department of the Natural and Built Environment, Sheffield Hallam University, Sheffield, UK
e-mail: g.tang@shu.ac.uk
J. Chilton
Department of Architecture and Built Environment, The University of Nottingham, Nottingham,
UK
e-mail: john.chilton@nottingham.ac.uk

872 G. Tang and J. Chilton
This technology-related polemic of being purist in terms of timber gridshells

was first identified and mentioned in the preface of a recent publication on timber
gridshells, namely:
… There are ‘purists’ who accept as gridshells only… those formed from initially flat deploy-
able grids of thin timber laths. However, nowadays the term is also freely and widely applied
by many to be free-form timber grids assembled from sometimes relatively massive, single-
or double-curved, engineered timber components, connected by sophisticated metal jointing
systems. (Chilton and Tang 2017, xviii)
Clearly, technology-based definitions change and evolve alongside technological

advances. The variation of cultural attitudes to technology held by different societies
marks an intriguing phenomenon. Within the immediate realm of timber technology,
where the acceptance, or indeed a simultaneous partial rejection, of technology in
timber practices can unite or divide opinions, how do contemporary timber practices
negotiate the influence of digital “advancements”? Similarly, how do timber arti-
sans working with traditional methods respond to the introduction and infiltration of
digital ideas? In the same grain, and vice versa, how do high-tech computational man-
ufacturers of precision components learn the rudiments and assimilate the wisdom
and intuitive material appreciation of the craftsperson?
What constitutes constructional correctness (as proposed by Pier Luigi Nervi,
see below) must be prioritised to inculcate an understanding of the influence of
technology on craft-based practices which combines both art and science. These
are important questions deserving discussion and a considered elaboration because
a confident appreciation of the different viewpoints will impact on future methods
and technique development influencing future timber works and affect the way we
restore/renovate past historic timber structures as well.
1.1 Constructing Correctly
In 1955, the reinforced concrete shell designer/constructor Pier Luigi Nervi raised
the notion of constructing correctly through his acclaimed publication Costruire
Correttamente (Nervi 1955), a theme he continued to promote in his influential
writings in the 1960s. In his book Aesthetics and Technology in Building he suggested
that objective technical issues “…stability, durability, function and … economic
efficiency…” are necessary conditions but that there is also a subjective aspect, which
includes the aesthetic appearance of the structure (Nervi 1965, pp. 2–4). Although his
remarks concentrated on his work in reinforced concrete, the idea of constructional
correctness is appropriate and applicable to all materials and are timely in the field
of timber technology. The term questions the relationship between historical and
contemporary materials/products, construction methods and the context in which
they are applied, including timber. The title of this present chapter borrows from
Nervi’s idea, to launch some intriguing questions about our relationship to technology
when working with wood.
Constructing Correctly in Wood: New Insights into Timber … 873
Evidently, timber technology has advanced rapidly in recent years to the bene-
fit of design, manufacturing and construction capabilities. This sees the previously
unimaginable becoming constructible, especially with regards to increased accura-
cies of curved geometries. However, like the many other technologies that have made
modern living easier (such as mobile telephones, apps, electricity, wi-fi and the inter-
net), these “improvements” in timber technology are also guilty of disrupting our
fundamental relationship with timber as a construction material. Some might argue
that this may also change our material understanding of timber itself.
“Constructing Correctly” is timely and relevant, as we move towards new modes
of production and as timber as a construction material evolves. This perspective
is highlighted in the development of timber gridshells. Using design, fabrication
and construction of timber gridshells over the years as examples, philosophies and
approaches taken by designers with regards to this traditional construction material,
are critically questioned to give a techno-cultural perspective of how we view the
more traditional or new computer-aided methods of working with wood. To instigate
a healthy debate, this chapter identifies alternative approaches to negotiate between
the polemics and better understand the concerns and deliberations from both sides, at
the same time reflects on the way timber technology is viewed, treated and applied.
1.2 Constructing Correctly in Timber
As represented in Fig. 1, “Constructing correctly” in timber suggests different ele-

ments of discussion. First, the material in question (in this case, timber) includes an
understanding and appreciation of physical properties. Secondly, the method, which
is concerned with how timber is used. These elements are influenced by personal
experiences. Thirdly, this correctness is governed by the context in which material
and fabrication methods are applied. In other words, the same technical solutions may
be viewed differently, when assessed in different countries and/or socio-economic
environments. A fourth element is the lens through which the judgment is viewed,
including the opinions and values of the person putting forward his or her views.
Although an objective judgment can be made, each person has the freedom to exer-
cise his/her opinion on whether construction has been carried out “correctly” in
terms of design, use, material treatment and/or construction method, aesthetics and
appropriateness to context.
2 Context
Aspects of timber technologies described in Fig. 2 below present a perspective on

how technology is solving previously unsolvable problems. It is widely accepted
that digital innovation negates the need for a fundamental understanding of timber,
Fig. 1 The idea of ‘Constructing Correctly’ is shaped by the values the individual places on material
understanding and constructional methods within specific contexts such as location, aesthetics etc.
Figure © Gabriel Tang
Fig. 2 The Purist and Liberalist camps exhibit different attitudes to technology. Liberalist attitudes
pertain to a view accepting computerization, digitization for increased productivity. On the other
hand, Purist attitudes value traditional methods, timber characteristics and structural behaviour.
This categorization is not binary, but allows each individual to position themselves on a varying
scale in relationship to each school of thought. Figure © Gabriel Tang
leading to the need to touch, see, smell or be in physical contact with the material
becoming redundant. Ultimately, this breeds a detached relationship from material
understanding.
A level of subjectivity about the concept of constructional correctness is embedded

within the discussion in this paper. What is judged as being correct or otherwise is
unavoidably coloured by our personal stance of “constructional correctness” shaped
by our personal experience.
Therefore, the following questions are posed: what does it mean to construct with
timber in digital age of the 21st Century? Do we regard material completeness as
being correct construction? For instance, should a “correct” timber building be on
more than half of the primary structure’s material being timber? Or should this be
measured against what is the correct method of constructing structure?
2.1 Purist View
In this research, through conversations with architects, engineers and builders, it was
observed that their descriptions of timber gridshell structures described their stance
and view of timber technology.
On the one hand, there are those, the purists, who believe in their perceived “cor-
rect” manner of construction, holding an innate understanding of timber’s physical
and mechanical characteristics, such as respect for grain direction and accommo-
dation to bending characteristics. In this school of thought, value is placed on the
skills of the craftsperson and fundamental material understanding of timber. Their
view of constructing correctly may also be steeped in a preference for material purity
(avoiding the excessive use of non-timber fastenings or adhesives, and/or complex
fabrication and construction techniques).
The purist view respects traditional skills, craft and tooling in creating timber
structures, be it simple or complex, imbued with an intuitive understanding of material
properties and handling gained through years of experience working with wood.
Purists place value on human craftsmanship, and may be opposed to the machine
and mechanization.
In the context of this paper, a purist also believes in the fundamental understanding
of material and holds a traditionalist approach steeped in historical precedent and
rationale. The purist is averse to the influx of digital influences in timber technologists.
2.2 Liberalist View
On the opposite end of the spectrum, a contrasting approach is observed to incor-

porate a standpoint that celebrates innovative construction and fabrication methods
largely brought about through digital technologies. It embraces the use of engineered
timber products with increased strength and a typical ability to withstand thermal
and moisture-induced deformations. Products such as cross-laminated timber (CLT),
laminated veneer lumber (LVL) and glue-laminated (glulam) timbers, fast replacing
steel or concrete to renewed claims of sustainability, are examples of how technology
is changing our understanding of timber as a material. Held in high reverence in the

liberalist school of thought, these products may not be viewed in the same light by
the purist camp.
Alongside this phenomenon of new engineered timber materials, as machining
technologies progress and new engineered timber products emerge, doubly-curved
structures (such as timber gridshells) are no longer limited by the discernible charac-
teristics of sawn timber and the traditional methods of using them. Instead, computer
numerically controlled (CNC) milling machines are capable of rapidly producing
single- and double-curved sections with repeated high geometric precision to a con-
sistent standard not seen previously using traditional hand tools. Liberalists are pos-
itively influenced by the importation of technologies: machines, digital tools such as
parametric design and digital form-finding, new engineered timber products, adhe-
sives and fastenings.
This approach differs from the purist way of dealing with timber through the
use of new machinery, associated method innovation and material improvements.
With this new technology, timber structures can be accurately designed, precisely
manufactured and efficiently assembled with improved adhesives and fastenings,
signifying a shift in the function of the craftsperson in design and construction terms.
To some, this also elevates the role of digital technology, often at the cost of
de-skilling (craftspeople) and de-materialization (of timber).
2.3 Spectrum
In between these extremes there lies a spectrum upon which the individual is able to
take a stance. An example of the benefit of acceptance of digital fabrication, whilst
using timber in its most basic form, is illustrated by the 10 m diameter reciprocal
frame roof installed at Hill Holt Wood, near Newark, UK, Fig. 3. Here, locally-
sourced debarked tree trunks were used to create interlaced spirals of mutually-
supporting beams. Although the tree trunks were irregular in cross-section, once a
reference centreline had been established precise CNC cutting and drilling of the
simple lap joints was possible (Chilton et al. 2009). This would have been extremely
difficult to achieve using a strictly traditional ‘purist’ method of setting out and
cutting of joint profiles. This acceptance of digital technology, combined with an
understanding of traditional woodworking methods is reflective of the inclusivity of
a balanced approach.
3 Tectonics of Timber Assemblies
The word tectonic [deriving from the root-word of the Greek word tektōn (τšκτων),
denoting a building artisan, in particular a carpenter (Wikipedia 2018)] is relevant
here. “Tectonic” has evolved to mean “the art of joining things” (Borbein 1982) and
Fig. 3 Reciprocal frame timber roof at Hill Holt Wood, near Newark, UK, assembled from CNC
machined tree trunks. Photograph © John Chilton
is highly relevant to timber technology. The need to extend and join timbers to each
other motivated timber technology.
Assembly Technology: A Crafted Dimension
Timber technology developed as a result of building from roughly cylindrical forms
(tree trunks). The structural bundling of organic cells, a composite of cellulose fibres
and lignin, provides strength to the material. Traditionally, trees are harvested and
machined into planks and posts (Fig. 4). Long building elements usually run along
the grain of the timber to avoid cutting through fibres and maintain their superior
structural capabilities.
Throughout history, civilizations endeavoured to join discrete timber sections, at
first without the use of metal fastenings, nails or adhesives. These inventive methods
and innovations were driven by the lack of metal fastenings and jointing technology.
These solutions and ancient wisdom were also limited by tools available for cutting
timber members into shorter, more manageable sections.
In Norway and Northern Russia, logs are primitively shaped and with ends cut into
throats and notches slot together without the need for metal nails or additional timber
dowelling. Traditional Norwegian stave construction was developed to connect round
timber sections to form locked and stable joints (Herzog et al. 2012). Similarly, in
Japan and China, ancient timber structures are also constructed through precise hand-
crafting, cutting and shaping to form connections that secure the discrete structural
elements (Herzog et al. 2012; Sumiyoshi and Matsui 1991), without mechanical
Fig. 4 Harvested raw timbers being debarked, processed and sawn into planks and board materials.
Guizhou, China. Photograph © Gabriel Tang
Fig. 5 Traditional tenon and mortise joints, Maogong, China. Photograph © Gabriel Tang
fixings, Fig. 5. The application of technology here is low and the quality of the
construction depended on the skill and experience of the artisan, the craftsperson.
Although apparently primitive in method, the construction was based on fundamental
understanding of timber. To the purist, this can be deemed a “correct” way of building
in a specific context without excessive metal fastenings.
To analyze and understand the intent and procedures associated with timber pro-
cessing, it is imperative to form a holistic view of constructing correctly. The evo-
lution of timber technology stemmed from the need to join and extend. Often, in
pre-digital societies, this knowledge evolved as it was passed down from master to
student over many generations. Therefore, tectonics, recognized here by Borbein’s
(1982) definition, as a way of joining things together, is clearly integral to the art,
and science, of timber technology.
3.1 The Changing Characteristics of Wood
With the influence of scientific advances on timber technology, the philosophical

question “What do designers want timber to be?” is an intriguing question. In recent
years, timber has established itself as a key material for use in sustainable construction
due its low embodied energy, the carbon that it sequesters, its ability to be reused
and the fact that it is a renewable resource. This is demonstrated in projects such
as the GlaxoSmithKline (GSK) Sustainable Chemistry Building at the University of
Nottingham, Fig. 6, which is expected to be ‘carbon-neutral’ after 25 years due to
the renewable energy technology incorporated and the extensive use of engineered
timber in its construction (Chilton 2015).
Fig. 6 The ‘carbon neutral’ GlaxoSmithKline (GSK) Sustainable Chemistry Building, University
of Nottingham, UK, under construction, 24th July 2014, which has predominantly timber structure
and cladding above ground. Photograph © John Chilton
Of all construction materials, especially when compared to steel and concrete,

wood from a living tree is most closely linked to human beings physiologically. A
natural material, figures resulting from growth rings and the way in which the timber is
cut, irregularities and imperfections such as knots and shakes, add to its visual charm,
lending it a psychological warmth. Timber is also breathable and reacts to relative
humidity, expanding and contracting in volume. This variability and environmental
responsiveness is often accepted as an attractive characteristic of green timber but
requires careful detailing to accommodate possible movement.
However, timber as a natural material is shedding its shortcomings of being a
material marred by (sometimes beautiful) defects such as knots, and lacking in uni-
formity due to the opportunities offered by engineered timber-based materials.
New technology has resulted in a shift in structural ability and pre-conceived
material understanding. Whilst retaining their timber appearance, new timber prod-
ucts such as CLT are behaving more like steel, technology generates new materials
by drawing on qualities from the best of both. This mis-match, often frowned upon
by the purist school, sees technology solving problems, which pose the question
of material honesty and their relationship to the materials’s basic provision, both
structural and aesthetical. These unfamiliar qualities quite easily split purists from
liberalist schools with the latter being receptive to new improvements through new
technology.
3.2 New Timber Products
Engineered timber products such as cross-laminated timber (CLT) are becoming

more widely available in the market. With key benefits of enhanced strength and
reduced dimensional movement, when compared to raw timber, the engineered
timber products appear to be very attractive indeed. Manufactured in large (up to
16.6 m × 2.95 m) boards, consisting of a minimum of three plys of crosswise-
orientated timbers glued together (KLH Massivholz GmbH 2012), extensive struc-
tural elements such as walls and floors, can be assembled easily and rapidly. Building
openings such as windows and floors can be robotically milled to high precision, as
used in the GlaxoSmithKline (GSK) Sustainable Chemistry Building, Fig. 6. These
new products are also more thermally and moisture-stable materials minimising the
uncontrolled movement of the material.
The issue is how this technology changes our acceptance and recognition of what
timber is. This is an interesting phenomenon, as technology dematerializes timber in
its application in architecture, as engineered timber products enhance the properties
of plain timber.
3.3 Fastenings
The capabilities of fastenings and attachments of one timber section to another par-
allel the development of cutting technologies. These days, contemporary architects
are choosing to celebrate the art and the honesty of attaching different sections of
timbers together.
The use of iron straps, coach bolts and nails which secured cathedral roof trusses
in the middle-ages in Europe saw new material, construction tools and methods in
improving construction. What was pure and completely made of timber was now
married to technology.
Of course, the use of metal plates, bolts, screws and nails is now common, but
they are often held in hostile regard by purist timber craftspeople. A recent exam-
ple of such thinking is seen in the Tamedia Headquarters Building, in Switzerland
(Fig. 7), completed in 2013, where, despite embracing the use of engineered timber,
the architect Shigeru Ban tried to minimize the number of metal connections in the
7-storey office building. Key joints between floor beams and columns in the multi-
storey frame were formed with large transverse oval glulam timber dowels spanning
between the columns (Antemann 2014).
Fig. 7 Interior of the Tamedia Headquarters Building, Switzerland. Source https://commons.

wikimedia.org/wiki/File:Tamedia-Verlagshaus.jpg
Fig. 8 The Timber Wave

Sculpture designed by
Amanda Levete Architects
and constructed by Cowley
Timberworks, Victoria and
Albert Museum, London,
UK, 2011 demonstrates the
expression of metal
fastenings and flitch plate
connections. Photograph ©
Gabriel Tang
Express or Repress
The philosophy of material and constructional honesty has been discussed since
Adolf Loos’ essay Ornament and Crime that laid down the foundation for new
architectural movements with regards to ornamentation (Loos 1913). Timber fas-
tening design raises questions of how these “foreign” metal connections should be
treated—either hidden from view or be acknowledged as tectonic and visual features.
The honesty of expressing timber connection responds to our aesthetic values and
where we place this. Flitch plates securing two or more pieces of wood together
are often hidden from view, while pinned connections are expressed. Both exposed
and concealed metal connectors are clearly illustrated in the Timber Wave Sculpture
designed by Amanda Levete Architects and constructed by Cowley Timberworks,
exhibited outside the Victoria and Albert Museum, London, UK, in 2011, Fig. 8. The
proliferation or excessive use of metal fastenings in timber connections, to the tech-
nology purist, evades the question of timber joining relying on technology rather than
through skills and patience. To express or conceal fastenings is a bone of contention
as their visual proliferation may suggest the dilution of craftsmanship.
4 Timber Gridshells
The evolution of timber gridshell design and construction not only reflects the devel-
opment of timber technologies, it also allows us to learn about ourselves in terms of
the values that we place on our relationship with technology.
Ideas associated with innovation in the design and manufacture of timber assem-
blies discussed earlier are clearly applicable to the development and evolution of
timber gridshells as a genre. The notion of the liberalist versus purist traditions
reflects our position in relation to material and construction methods used.
Gridshell structures are beautiful and expressive. Through their tectonic expres-
sion of doubly-curved volumes, they define interesting curved spaces. Their devel-
opment is a close reflection of the advance in architectural technology over the last
50 years, from the first engineered timber gridshell, constructed in Essen, in 1962
(Hennicke and Schaur 1974) to present day structures, such as the French Pavilion
for the Milan Expo, in 2015 (Chilton and Tang 2017) made possible through the
extensive help of digital technology.
4.1 Definition
Supporting either a purist or liberalist view of contemporary timber and digital tech-
nology, opinions on the definition of what constitutes a timber gridshell are divided.
The classic and highly-specific example can be illustrated by the definition of a
gridshell given in 1974 by Hennicke and Schaur, namely:
The gridshell is a spatially curved framework of rods and rigid joints. The rod elements form
a planar grid with rectangular meshes and constant spacing between the knots [nodes]. The
form of the gridshell is determined by inverting the form of a flexible hanging net. To invert
the catenary so that it becomes the thrust line of an arch free of moments is an idealization.
Analogously, inverting the form of a hanging net yields the support surface of a grid shell
free of moments. (Hennicke and Schaur 1974, p. 26)
This definition, detailed and prescriptive, supports a purist view. It records

the method/technology/technique applied in the 1970s to create efficient gridshell
shapes. In terms of components, it defines a spatially curved framework of rods and
rigid joints based on a grid with regular spacing between the nodes setting up a mesh
with distinct square pattern. Additionally, it also describes the method by which this
flat mesh arrives at its doubly-curved shape; its deformation by inverting the form
of a flexible hanging mesh to generate a line theoretically free of moments. [In fact,
due to the difference between a flexible net and a grid of continuous timber laths,
moments are induced in the gridshell—it is a bending active structure.] This defini-
tion was made at the infancy of digitization and described the derivation of efficient
shapes through physical model-making.
Of course, over 50 years on, in 2018, digital form-finding via software such as
dynamic relaxation or particle spring methods, exerts their influence on the design
Fig. 9 Interior view of the Kaeng Krachan Elephant Park shell, completed in 2014
at Zürich Zoo. Source https://commons.wikimedia.org/wiki/File:Zoo_Z%C3%BCrich_Kaeng_
Krachan_Elefantenpark_2.JPG. By Albinfo [CC BY-SA 3.0 (https://creativecommons.org/licenses/
by-sa/3.0)], from Wikimedia Commons
methods, superseding the method by which ‘moment-free’ shell forms are generated,
providing time and material savings when compared to designing through physical
model-making.
In 2000, Steve Johnson, of Edward Cullinan Architects (now Cullinan Studio)
presented his interpretation of what constitutes a gridshell:
A shell is a natural, extremely strong structure. A gridshell is essentially a shell with holes,
but with its structure concentrated into strips. (Johnson 2000)
Johnson’s definition appeals to liberalist thinkers by an implicit approval of con-

temporary technology. The definition described what gridshells are, and was not
overly concerned with how they were designed or built. In this case, design methods,
structural analysis and physical construction were left open-ended. The definition
left room for improvement that enabled innovations and new techniques to be devel-
oped in the future. This description is hence sufficiently broad to include projects
such as the Kaeng Krachan Elephant Park perforated timber shell, completed in
2014 at Zürich Zoo (Fig. 9), which, although appearing to be a gridshell made from
wide strips with irregular grid geometry, is actually formed from approximately 600
CNC-cut flat panels of CLT in three layers (Lennartz and Jacob-Freitag 2016, p. 31).
From these definitions, two main families of gridshells are commonly derived:
those made from gridmats of long, thin timber laths—a purist application of tim-
ber—that are actively-bent during deployment; and those that are assembled from
thicker, shorter, more rigid components assembled directly in the final shell firmly
supporting a liberalist attitude. Rigid timber gridshells free the designer from con-
straints experienced by their actively-bent predecessors, in particular the regular
grid and the requirement for deployability and shaping of the gridmat—a result of
technological improvements—supporting a liberalist standpoint.
4.2 Technology
Advancement in timber technology has directly affected the appearance of timber

gridshells and how they are designed and constructed.
Mechanized Lath Production
The development of improved timber adhesives allowed long laths of timber to be
produced rapidly and efficiently. During the construction of the Weald and Down-
land Jerwood gridshell (Fig. 10), digitized machinery, specifically a Grecon/Dimter
SUPRA E continuous feed finger-jointing machine (Harris et al. 2008), was used
to detect and remove knots and other imperfections in the oak laths. The resulting
laths had a reduced propensity to break when bent during the gridshell forming,
although there were still 145 breakages in the 10,000 joints (Harris et al. 2003,
p. 438). Automation and mechanization has influenced the efficiency and methods
by which these structures are designed, parts pre-fabricated and joined together.
Design Control and Production of Curved Elements
As gridshells derive strength from curvature, the production of timber components
curved in more than one direction remains a perennial challenge. This is achieved
with increasing ease due to the assistance of new timber technology and digital
manufacturing. Precision routers, milling machines and cutting of timber sections
have also made timber elements with complex surface definitions possible.
For example, the production of curved elements using specialized routers allows
the designer to produce complex timber sections accurately. It was reported that for
a 20 m timber member fabricated for the Pompidou Metz, a 3–4 mm tolerance was
measured (Döbele 2017). One may see this integration of computer aided design
and fabrication process as leaving the craftsperson/builder in a precarious position;
and for purists to question the true value of the timber craftsperson, in striving for
absolute precision and efficiency in a material that inherently and traditionally bears
a higher degree of tolerance and movement.
Safety and Method of Gridshell Construction
In early examples, timber gridshells were constructed by deploying gridmats of
timbers with square grid patterns. For the Mannheim Multihalle 1975, (Happold and
Liddell 1975), Fig. 11, western hemlock laths were end-to-end finger jointed into
lengths up to 40 m before assembly to form the gridmat. The regular gridmat was
then deformed through shearing the grid and actively bending the laths to form strong
Fig. 10 The Weald and Downland Jerwood Gridshell, Singleton, Chichester, UK designed by
Studio Cullinan and constructed by Green Oak Carpentry Company 2003. Photograph © Gabriel
Tang
shell forms with double curvatures. Once the predicted shell geometry was achieved,
the grid was fixed at the edges and restrained with diagonally placed pairs of 6 mm
diameter stainless steel cables (Fig. 12) to prevent it from deforming back into its
original flat mat. This method of installation, however, induced high bending stresses
in the timber laths, which resulted in many breakages.
This revealed several shortcomings when judged by current construction standards
of safety. Similar sized gridshells erected for the Silk Road Expo’ in Nara, Japan, in
1988 (Sakamoto 1992), used an alternative technique, where prefabricated sections
of pre-bent laths (Melaragno 1991) were craned into position and connected in the
air. For the Weald and Downland Jerwood gridshell, shown in Fig. 10, a new method
was employed, by which the flat gridmat was assembled at high level and deformed
into the final form through gravitational force and jacking (Harris et al. 2003).
In the case for the Savill Garden gridshell (Windsor Great Park, UK, 2005)
(Fig. 13), the larch gridmat was assembled at roof level. In order to reduce the
possibility of breakage, laths were bent to a much reduced curvature compared to the
Mannheim Multihalle, and restrained with a double layer of 12 mm plywood boards
(Harris et al. 2008; Chilton and Tang 2017).
More recently, shorter and thicker components, prefabricated and pre-formed to
curved gridshell geometry, have been assembled directly on temporary supports or
Fig. 11 Mannheim Multihalle was created by pushing up gridmats from the ground (Photograph
© Gabriel Tang)
assembled into larger sections on jigs before craning and connection to the previously
installed sections. Examples include: Pompidou Metz, France, 2010 (Lewis 2011),
Haesley Nine Bridges Golf Clubhouse, South Korea, 2010 (Scheurer 2010), France
Pavilion Milan Expo, 2015 (Scheurer et al. 2015), and La Seine Musicale, near
Paris, 2017 (Chilton and Tang 2017). These developing methods, are enabled only
by advances in digital fabrication and engineered timber production, have greatly
improved the safety of the construction process, in light of lessons learnt from the
Mannheim gridshell.
Off-Site Construction
Pre-fabrication and off-site construction using engineered timber has played a major
part in the movement from purist to liberalist attitudes to the realization of timber
gridshells. It was implemented for the gridshell canopies produced for the Hannover
World Expo 2000 designed by Thomas Herzog and engineered by Julius Natterer.
Each of the approximately 19 m by 19 m quadrants of the canopy was manufactured
off-site and then delivered to site for assembly (Natterer et al. 2002).
Computer technology enabled sections of timbers with bespoke straight sides
and specially curved sides to be constructed and holes milled for connection. Being
able to divide digitally a curved surface into smaller and discrete components also
Fig. 12 Mannheim Multihalle was restrained by a diagonal network of twin 6 mm diameter stainless
steel wire ties installed at 4.5 m centres. (Happold and Liddell Happold and Liddell 1975) after the
gridmats were pushed up from the ground. Photograph © Gabriel Tang
meant that traditional method of actively bending with fixed intersections is no longer
necessary (Chilton and Tang 2017).
Digital technology used in design and fabrication is effective in shortening con-
struction time, removing the need for manoeuvring space (in deployable gridshells)
bearing a positive impact on project cost and complexity in construction. It has
allowed designers to incorporate complex ‘woven’ jointing into gridshells, such as
at the Haesley Nine Bridges Golf Clubhouse and La Seine Musicale, whilst min-
imising the use of mechanical fixings.
5 Discussion: Constructional Correctness
5.1 Structural Values
Gridshells designed in accordance with the ‘purist’ definition allow the structure
to transfer dead and live loads efficiently, with no bending moments (theoretically)
under the primary load condition (usually shell self-weight). Through their doubly-
Fig. 13 The Savill Garden gridshell completed in 2005 was restrained by a double layer of plywood
working in shear. Photograph © Gabriel Tang
curved geometry, if correctly determined, they can do this whilst avoiding buckling in
compression. Traditionally, this allows shells to achieve spectacular thinness relative
to their span.
Such structural logic and purity allows bending-active gridshell structures to span
considerable distances. However, gridshells conforming to the looser definition of ‘a
shell with holes’, especially more recent examples formed by connecting rigid timber
components, are heavily dependent on digital fabrication techniques. Because curved
components can be fabricated directly, they no longer need to be flexible—in fact they
are thicker and can resist bending—the prime structural concern to create a wide-
spanning column free environment is no longer the motivation of the shell designer,
but the typology becomes valued primarily for its aesthetics. The application of
timber gridshells with internal columns can be applied to projects not restricted to
clear spanning purpose as very aptly demonstrated by Shigeru Ban’s Haesley Nine
Bridges project in South Korea (2010), Fig. 14.
The question of whether this construction is correct is a subjective one, highly
dependent on the background and understanding of such structures by the individual.
The purist may prefer the gridshell to be thinner to enclose a barrier free environment
internally. A liberalist view sees the aesthetic and design benefits of how it can create
an interesting architectural experience.
Fig. 14 The Haesley Nine Bridges building, South Korea by Shigeru Ban, 2010. Copyright Jong
Oh Kim https://www.arch2o.com/haesley-nine-bridges-club-house-kyeong-sik-yoon/
5.2 Material Values
Structural purity (pure compression action) for gridshells is often not the main driver.
Aesthetically and poetically, the Savill Garden gridshell (2005), Fig. 15, appears to
hover at an approximate height of 10 m above the ground. To allow this, the shallow
gridshell was restrained by a 400 mm diameter steel ring beam, in turn supported by
V-shaped 8 m long circular hollow section steel legs.
To some purists, the use of the highly conspicuous ring beam would suggest a
demeaning use of a (relatively flat) gridshell as their strength generally lies in their
shape. However, the Savill Garden gridshell was designed to realize an aesthetic
vision and architectural concept, thereby demonstrating the adaptability of gridshells
in timber, accepted by liberalists in recognizing the dexterity in accommodating
different concepts. On the other hand, some purists would frown upon the excessive
use of non-timber materials in the timber gridshell structure, questioning the material
identity of the structure.
In more rigid gridshells, digital technologies enable the milling of glulam timbers
into complex curves in more than one direction. Although the milling process can
be programmed to minimise cutting of continuous wood fibres, timber components
are not necessarily shaped with the same respect for the material as they would have
been in a pre-computer age. These timber sections often need to be of greater size
Fig. 15 The Savill Garden gridshell, completed in 2005, is raised above ground by 8 m long steel
columns and the shallow gridshell structure is restrained by a 400 mm diameter tubular ring beam.
Photograph © Gabriel Tang
in order to accommodate bending moments that result from geometries that did not
necessarily conform to natural load paths.
This also implies that the idea of the shell harnessing the shape (curvature) to
produce a structurally efficient form for load transfer may be compromised. Purists
would reject this. Conversely, this may be acceptable to the liberalists who appreciate
the potential architectural merits of the less than structurally perfect gridshell form.
5.3 Aesthetics
One major aesthetic concern of timber gridshells rests with their external cladding,
a problem perennially experienced by all doubly-curved surfaces. They lend them-
selves to the application of flexible membranes that can be patterned to conform with
curvature. Some may be quick to point out that the roof of the Weald and Downland
gridshell, with flat boarding being too complicated or even hiding the elegant doubly
curved structure beneath.
Technological advancement in materials has allowed gridshells to become lighter
visually by widening of grids. Compared to earlier gridshells, based on a 0.5 m ×
0.5 m grid, the Waitomo Glowworm Caves Visitor Centre gridshell, 2010, (NZ Wood,
undated) in New Zealand, appears much lighter and more transparent. Some may
question the identity of the building, whether it is in fact an ETFE roof (purist) rather
than a timber one (liberalist) by the sheer visual appearance (timber grid size of
4.25 m) and the large expanse of the membrane material.
Timber tectonics and connections imbue ideas of construction methods and mate-
rial understanding. The connections are traditionally expressed in deployable grid-
shells by virtue of their gridmat and deformation process. Timber gridshells, due to
their constructional nature are tectonically expressive. The methods by which they
are constructed or put together can become a design feature easily.
5.4 Economy
Material economy is an important issue in our resource conscious environment. Tra-

ditional actively-bent timber gridshells are constructed from rectangular-sawn laths
with little wastage of material—this fulfils a purist view of their creation. How-
ever, digitally controlled processes now permit knots and other imperfections to be
detected and removed from laths. Subsequently, finger-jointing with modern adhe-
sives create long laths with enhanced properties compared to the base material—thus
requiring a more flexible interpretation of the purist view.
In the manufacture of the rigid gridshells, laminations of small dimension are sawn
and planed smooth before assembly and gluing to create larger glulam sections.
Subsequently, in order to form double-curved components, such as that found in
the Pompidou Metz and Haesley Nine Bridges Golf Clubhouse roofs, requires the
removal of a substantial volume of material, potentially milling across and against the
grain. However, with the aid of digital control of the milling processes it is possible to
optimize the orientation of the final piece within the original glulam timber block to
minimise the impact of this. These digitally-controlled processes are more acceptable
to the liberalist viewpoint.
6 Conclusion
Pier Luigi Nervi’s notion of constructing correctly is relevant and significant in con-
temporary technology discourses. It has been argued here that his idea is subject
to the values we place on material, method and the immediate context. Key timber
gridshell projects have illustrated how different variations, developments and overall
evolution of these structures are met with different attitudes depending on an indi-
vidual’s value on technology. One’s view is not binary in nature, but sits on a sliding
scale between a purist or liberalist standpoint concerning the application of digital
timber technology.
This discussion proves that the degree of acceptance of technological change in

timber architecture is subject to our personal position in terms of values we place on
material and technology. As we negotiate between increased use of digital design,
fabrication and our craft tradition, this represents a cultural understanding of the
craftsperson’s relationship to technology, also raise the question of what, in our
opinion, represents correct or incorrect construction.
Technology can be considered a double-edged sword. The influence of digital
technology on timber practice is clear. Timber gridshells, intrinsically linked to tech-
nical capability, have evolved in the last half century—changing in shape, size and
complexity—to document this change of timber technology and reflect the way
such structures are designed, components fabricated and eventually assembled. Their
design and construction method divides and unites opinions. To some (the purists),
the traditional methods of designing and construction qualifies a gridshell as a pure
structure. To others (the liberalists), the use of design tools, new materials and design
methods are acceptable to qualify similar structures as gridshell structures. Signifi-
cantly, these new materials and methods are the direct or indirect results of digital
technologies.
Therefore, digital tools of design and manufacture will continue to exert a strong
influence on timber technology. As the digital wrestles with traditional ways of
designing and constructing, the understanding of timber as a material must not be
neglected by the next generation of designers and craftspeople which oversees a
continued celebration and healthy adaptation of the characteristics and qualities of
timber in both natural and engineered forms, in the spirit of Constructing Correctly.
References
Antemann M (2014) Seven storey wood office building in Zurich. Detail: review of architecture
and construction details: timber construction 2:174–180
Borbein AH (1982) Polyklet. Na
Chilton JC, Cowley G, Westmuckett J, Gilroy-Scott B (2009) Reciprocal frame timber roof at Hill
Holt Wood—a case study. In: Chilton JC, Mungan I (eds) Timber structures: from antiquity to
the present. T.C. Halíç Üníversítesí Ístanbul, pp 383–394
Chilton J (2015) Solid timber in carbon neutral construction in the UK. Future visions—international
association for shell and spatial structures symposium 25:8
Chilton J, Tang G (2017) Timber gridshells: architecture. Routledge, Structure and Craft
Döbele T (2017) Email communication with Gabriel Tang, 16 May 2017
Happold E, Liddell I (1975) Timber lattice roof for the Mannheim Bundesgartenschau. The structural
engineer 53(3):99–135
Harris R, Romer J, Kelly O, Johnson S (2003) Design and construction of the Downland Gridshell.
Build Res Inf 31(6):427–454
Harris R, Haskins S, Roynon J (2008) The Savill Garden Gridshell Design and construction. Struct
Eng 86(17):27–34
Hennicke J, Schaur E (eds) (1974) IL10: Gitterschalen—Grid Shells, Institut für leichte Flächen-
tragwerke (IL)/Karl Krämer Verlag, Stuttgart
Herzog T, Natterer J, Schweitzer R, Volz M, Winter W (2012) Timber construction manual. Walter
de Gruyter
Johnson S (2000) Gridshells and the construction process. http://www.wealddown.co.uk/explore/

buildings/further-reading/gridshells-construction-process/. Accessed 3 July 2018 at 17.02
KLH Massivholz GmbH (2012). Made for building—built for living: cross-laminated tim-
ber. Available at http://www.klhuk.com/media/33471/klh_component%20catalogue%20for%
20cross%20laminated%20timber_version%2001_2011.pdf. Accessed 3 July 2018 12.24 pm
Lennartz MW, Jacob-Freitag S (2016). New Architecture in wood: forms and structures. Birkhäuser,
pp 26–33
Lewis B (2011) Centre Pompidou Metz: engineering the roof. Struct Eng 89(18):p20–p26
Loos A (1913) Ornement et Crime, Les Cahiers d’aujourd’hui (French)
Melaragno M (1991) An introduction to shell structures: the art and science of vaulting. Van Nostrand
Reinhold, New York, N.Y., p p110
Natterer J, Burger N, Müller A, Natterer J (2002) The Roof Structure Expodach at the World
Exhibition Hanover. In Proceedings of the fifth international conference on space structures, No.
IBOIS-CONF-2002-001, pp. 185–193
Nervi PL (1955) Costruire correttamente. Ulrico Hoepli, Milano
Nervi PL (1965) Aesthetics and technology in building. Harvard University Press
NZ Wood (Undated) Case studies: Waitomo Caves Visitor’s Centre. http://www.nzwood.co.nz/case-
studies/waitomo-caves-visitors-centre/. Accessed 30 July 2018 at 13.01
Sakamoto I (1992) Wooden spatial structures in Japan. Bull IASS 33(2):109–119
Scheurer F (2010) Materialising complexity. Arch Des 80(4):p86–p93
Scheurer F, Simonin L, Stehling H (2015). ‘Energy for life’ the timber structure of the French
Pavilion at EXPO 2015. In: Proceedings of the international association for shell and spatial
structures (IASS) symposium 2015, Amsterdam
Sumiyoshi T, Matsui G (1991) Wood joints in classical Japanese architecture, translated by Ferenc
Kovacs. Kajima Institute Publishing Co Ltd, Japan
Wikipedia (2018) Tektōn. https://en.wikipedia.org/wiki/Tekt%C5%8Dn. Accessed 24 July 2018,
11.06 a.m
The Emergy of Digital Wood
Chris Beorkrem
Abstract The construction industry and building sector are now widely known to be
one of the biggest energy consumers and carbon emitters. Some architectural agen-
das for sustainability focus on ‘energy efficiency’ or buildings that minimize their
energy intake during their lifetime—through the use of new ‘greener’ elements, such
as more efficient mechanical systems or more insulative wall systems. The focus on
the efficiency of one aspect or system of the building, versus the effectiveness of the
whole, leads to ad hoc ecology and results in the familiar “law of unintended conse-
quences.” This chapter explores the ways in which researchers have been exploring
the use of technology to expand the use of wood while minimizing the ‘emergy’
or energy memory of the material being consumed. These explorations are two-
fold; using wood as a responsive material which can adapt to its environment with
no external energy, improving the performance of a building, and the creation of
construction methods and joinery which responds to the inherit “liveness” of the
wood. This chapter will explore the notion of emergy through a series of case studies
which exemplify research that can capitalize on dramatic advancements in the use
of scanning methods and robotic technology to greatly expand the performance of
wood.
1 Introduction
21st Century explorations in timber construction typically explore two general issues.
The first is climate change, a subject the design profession, has too often approached
with solutions applied as active solutions applied to a project. Typically remediated
with expensive technology as much of the world sits comfortably “atop 300 years of
flaming carbon” (Mann 2014). The second issue is the global trend towards urban-
ization. According to the United Nations, 54% of the world’s population lives in
urban areas. This proportion is projected to be 66% by 2050 (United Nations 2014).
C. Beorkrem (B)
UNC Charlotte, Charlotte, USA
e-mail: cbeorkrem@uncc.edu

896 C. Beorkrem
The need for architects, planners and urban designers to work to accommodate this
shift will be overwhelming in the coming decades.
As we build, instead of focusing on energy ‘efficiency’ in buildings, we should
also look to create systems or hierarchies of energy within our design methods which
use energy more responsibly. Kiel Moe’s take on this Convergence, (from his book
of the same name) is based on “the thermodynamic premise that architecture should
maximize its ecological and architectural power.” This methodology relies on the
understanding that all matter is captured energy. Moe goes on to talk about emergy,
entropy, and exergy, from a material perspective. The embodied energy within a
material, combined with its captured bio-geophysical factors, is its emergy, or energy
memory. For Moe, a building with high emergy content is not bad thing, as the
building has a greater capacity for energy feedback. Quality of the feedback is a
key design concern, in order to maximize the capacity for energy to do work, or
its exergy. In any system, some energy will become bound dissipated energy, or
entropy. A successful ecological architecture would have high emergy, high-quality
exergy, and low entropy. This maximum power design decreases overall energy use
through a series of nested hierarchies that “maximize the intake, use and feedback
of useful energy.” This methodology anticipates, rather than reacts, to the energy
issue in architecture. By reducing the amount of material used, reducing material
waste, using local and recycled materials, and increasing the lifetime(s) or use(s)
of the material, we can design systems and buildings that minimize emergy while
“maximize[ing] the intake, use and feedback of useful energy” (Moe 2013).
Moe says architecture relies on the understanding that all matter is captured energy.
His ideas on convergence are based on “the thermodynamic premise that architecture
should maximize its ecological and architectural power.” Rather than focusing on
making buildings “less bad” by singularly focusing on the performance or efficiency
of specific materials, he calls on designers to think of the building as an ecosystem
within an energy hierarchy. Thermodynamic principles state that within a system,
energy is neither created nor destroyed, but it changes from one form into another.
A designer’s goal should be to better understand the energy hierarchy, how systems
within architecture fall into this hierarchy, and how different forms of energy can be
used as inputs and outputs of these assemblies (Moe 2013).
Unlike other building materials, wood is alive. Its randomized grain patterning
and warmth can affect the occupant of a space in ways that few materials can (Bern-
heimer 2014). In the commercial building industry, wood has traditionally been used
as an option only for interior finish. However, timber is a competitive solution for
developing urban environments in the 21st century because of its ability to simultane-
ously act as structure, fireproofing, and cladding, all while improving environmental
benefits, ease of construction (related to fabrication, assembly, time and cost) and
can be used for its aesthetic material characteristics.
Moe relies on the traditional notion of architecture as the application of a hylomor-
phic method, first coined by Gilbert Simondon and later reframed by Gilles Deleuze
and Felix Guattari, which defined the notion that forms are imposed upon inert mat-
ter, independent of their material characteristics. Manual Delanda, in his article ‘The
Philosophies of Design: The Case of Modelling Software’ contrasts the notion of
The Emergy of Digital Wood 897
design as a purely cerebral idea which ‘imposes its will’ on predictable materials,
with an idea that materials might be allowed to have ‘idiosyncrasies’ and be allowed
to participate in a design process. And Branko Kolarevic calls the process of using
a materials character not in a responsive way not ‘subtractive but transformative”
(Kolarevic 2003). Achim Menges articulates the idea of ‘machinic versus material
morphospace’ which is defines formal typologies driven by a designer’s emphasis
on the machine or the material (Menges 2012). The continuous loop of first highly
processing a material and then distributing across regions or even hemispheres only
to be reprocessed is defined by Mario Carpo as “A wasteful redundancy becomes
apparent, in which material is processed two or more times to achieve characteristics
that may already be present in the original material.” (Carpo 2011, 105). And in
the process, wood’s internal fibers are cut over and over again, greatly reducing its
strength. Though thoroughly considered by this centuries most prominent philoso-
phers and researchers, this high level homogenization of materials is often driven
by codes and regulations. From the limitation to use two-by construction, plywood
or glue laminated beams there is a long history of using code and regulation-based
restrictions to limit opportunities for designers to use or test wood beyond its legal
intended use.
The structure of wood is varied in ways that most other materials is not. Driven
by the variability in its natural characteristics wood products are typically supported
in two directions (which creates a diaphragm). In most manufactured conditions this
is manifest as alternating layers of members on their perpendicular (but this can vary
in some conditions represented in this chapter). CNC products in wood typically use
plywood, or laterally braced elements tied together with glues or nails and screws.
These connection types greatly increase the embodied energy of the material by
rendering them useless after their initial installation or by adding a massive amount
of embodied energy through their assembly and the addition of mechanical fasteners.
This chapter will explore reductions in a design’s emergy through the analysis of
a series of contemporary applications of digital technology on wood. These include:
• Hygroscopically Enabled Responsiveness ‘Particles’-Karola Dierichs, Dylan
Wood, David Correa, Achim Menges-Institute for Computational Design-
University of Stuttgart
• Craftsman based robotic machine learning-Giulio Brugnaro, Sean Hanna, The
Bartlett School of Architecture
• MAS House-A Robotic Pavilion—ETH Zurich
• Hermes Rive Gauche-RDAI Architects-Klaas De Rycke, Louis Bergi, Ewa
Jankowska-Kus
• Bandsawn Bands-Ryan Luke Johns and Nicholas Foley
• Wood Chip Barn-Architectural Association Design + Make Program-Hooke park
• Sawmill Shelter-Architectural Association Design + Make Program-Hooke Park
These projects focus on one of three methodologies for the reduction of a mate-
rial’s emergy; the emergy of a material’s assembly, the reduction of emergy by min-
imizing raw material production, and the reduction of Emergy through a response to
a material’s nuanced character.
898 C. Beorkrem
2 Environmentally Responsive Wood
The research of Achim Menges and others at the Institute for Computational Design
at the University of Stuttgart (ICD) into the use of hygroscopically enabled responsive
wood elements exemplifies the idea that wood can respond to its environment in ways
that no other material can. Their research, led initially by Steffen Reichert, represents
a single wing of a diverging set of ideas framed around methods for controlled forming
of wood panels based on changing relative humidity in the elements environment.
These experiments have explored the use of these materials to create façade systems
which can respond to the external conditions, opening up the façade to respond to
increasing humidity levels, as well as an installation which represents more dynamic
changes in the environment (Pompidou), and more recent research which creates
multi-layered systems of large scale panels composed of multiple layers and pieces
of wood.
In each iteration of these systems a piece, or pieces of wood veneer were used
to create a mechanically responsive system which relied on no outside mechanical
control devices. The relevant variables for each system were the fiber orientation
which controlled the direction of movement, and the ratio of the thickness of each
piece, length, width and height. Lastly, the porosity of the material affected the
amount of movement based on how much change in moisture content the material
was capable of undergoing the material would respond with more or less dramatic
movements. The products of this research have a unique connection to the growth
of wood, capitalizing on our ability to unwrap veneers in ways that render wood
unique compared to any other material in our world. This combination of machine
and material has over the last century been used to create a material (plywood) which
is capable of meeting the highest standards for consistency and structural capacity
but also created a tremendous burden on our environment. By seeking out methods
for using veneers in their raw state this research minimizes the embodied energy of
the use of veneers while capitalizing on their most unique characteristic, their ability
to move.
Initial studies done under the direction of Achim Menges by Steffen Reichert,
indicated that fully reversible movement of such panels could be undertaken in less
than 20 s. Initial studies used these panels to curl away from a substructure to reveal
an opening. These elements were also created as single piece of varying thickness
whereby the substructure was a part of the veneer with a thicker ratio of length
and width, creating a base which moved less than the thinner proportion elements,
resulting in a system made of singular integrated static and dynamic components.
The example from this research which best exemplifies the notion of minimized
emergy is the use of hygroscopically enabled components as a set of actuated granular
materials. This research capitalizes on the controlled movement which was developed
as part of the early research to create small yet appropriately scaled elements which
work together to create properties which individually they could not attain. The
designers of the system argue that the product of the research is relevant for two
reasons, the first that it is capable of being “loose and pourable” as well as easily
Fig. 1 Low wood moisture content (WMC) and high wood moisture content modules (left) and a
close up cluster of low WMC modules expanded together to create a nest. Image K. Dierichs, D.
Wood, A. Menges, ICD, Domaine DeBoisbuchet, University of Stuttgart
packed, in one state and then highly entangled in another. Secondly, and probably
more importantly the system is disassemblable, as in it can be taken apart and easily
packaged and reassembled into the same or completely different configuration with
no external mechanical energy.
The system explored here uses wood bilayers which when exposed to higher rel-
ative humidity curve in two directions to create a non-planer hook shapes-idealized
for its ability to entangle itself with others. The research worked both through sim-
ulation and experimentation, using real assemblies of bilayer wood systems as well
as the Kangaroo 2 library and a commercial Finite Element Analysis (FEM) library.
This combination of research tools allowed for the research to be both analytical
and intuitive. The goal of each experiment was to find a set of systems which would
possess formal states that are flat-packed and linear as well as loose-packed entan-
gled three-dimensional forms which could consume a large amount of space when
combined.
The first case study modules were manufactured from a combination of beech
veneer and maple veneer in different thicknesses. The layers were glued under vac-
uum pressure in sheets. These sheets were then cut at a 30° angle to the grain of the
bilayer system. This angle creates the hook effect of the system. The particles can
be combined with linear scraps of wood to expand the size of the system without the
need for the system to include 100% actuated particles. Initial tests were capable of
creating a beam of actuated particles capable of spanning under self-weight over 1 m
(Fig. 1). Subsequent systems recombined the layers to explore combinations of per-
pendicular and diagonal grain maple with the beech veneer substrate (Figs. 2 and 3).
Various explorations of both hand organized, poured and robotically organized
systems were explored to test the effect of random and organized systems of deploy-
ment. The relative emergy of the system is lowered when less energy is used in
the layup of these systems. Due to their ability to be reconfigured and recycled as
well as their ability to self-organize into a structural logic, using only the ambient
900 C. Beorkrem
Fig. 2 The expanded modules with low WMC were capable of interweaving and spanning over a
meter in length. Image A. Wolkow, ITECH, University of Stuttgart
Fig. 3 A cable mounted robotic system deployed the modules across a floor with high WMC to
allow for quick deployment, as the elements dried out they expanded to create a low wall with much
higher volume. Image K. Dierichs, D. Wood, A. Menges, ICD, Domaine DeBoisbuchet, University
of Stuttgart
air conditions, these systems represent a broad range of emergy conscious design
parameters. When combined with simulation techniques and scaled using some of
the parallel research into hygroscopically enabled surfaces this system could have a
transformative effect on design and construction.
Additional contemporary research has delved into the possibilities for accom-
modating complex forms by laminating layers of shifting grain wood components
to pull the geometry into forms which match the desired outcomes. While it uses
sheet material to derive the components the Hermès Rive Gauche pavilions by RDAI
Architects is one example where the geometry of each component curves when flat-
tened and then is laminated to slightly contrasting components, which when secured
together form a doubly curved and twisting geometry. When pinned to each other
these components combine to create the final form.
3 Machine Learning Robotics
This second example represents a working methodology for approaching the use of
machine learning for examining the typically nuanced and complex possible patterns
found in cuts of wood. While seemingly infinite the typologies of grain in wood panels
do follow patterns and rather than measure them through scanning or simulation, the
authors created a system which instead is based upon the reaction and results of
typical woodworking tools. This test provides a method for analysis focused more
on the product of the test than the material being explored. The research eventually
employed the use of a machine learning tool to analyze a database of robotically
produced cuts in the wood to create a predictive model for what results are likely to
be produced by each cut. By using materially responsive tooling rather than typical
CNC tooling the production of these parts could avoid the use of highly refined
homogenous materials.
This research defines a new approach to the use of robotics and machine learning
which allows for the robotic tool to begin to understand and react to a materials
nuance in real-time. While in its infant state it does begin to allude to the idea that
automated tools might be able to respond to material characteristics in real-time,
avoiding waste and increasing the efficiency of their use, and therefore decreasing
the embodied energy needed to robotically manufacture building components in
wood. This avoids the typical pitfall of treating materials as an ‘inert receptacles for
imposed form’ (De Landa 2002) (Fig. 4).
The system these researchers used was initially trained by asking an experienced
wood worker to use a chisel and gouge which were outfit with a series of reflective
markers which could be scanned by a set of motion capture cameras. Additionally,
the tool was mounted to a reciprocating electric tool providing the ability to monitor
the force that is being applied by the wood worker in newtons. The combination of the
recorded movement and force were used to create a dataset which was subsequently
analyzed using a supervised machine learning algorithm or neural network (Brugnaro
and Hanna 2017).
To expand the data set, and therefore improve the latency of the neural network,
the ‘expert map’ of the initial woodworker tests were used to reduce the range of tests
that a robotic arm could emulate to test further. These tests were analyzed and added
to the data set to create a much broader understanding of how the wood reacted to the
various typologies of moves. Each test set was upwards of 1500 cuts and explored
varying tool angles and cut depths, within the range of the original work worker data
set (Fig. 5).
The results of this work pave the path for how human craftsmanship can be
emulated using computational tools. While the results leave more questions than
902 C. Beorkrem
Fig. 4 Motion capture enabled chisel tool mount on reciprocating electric tool. This was used to
measure the movement and force applied by craftsman which was then mounted onto a robotic arm
to measure the same range of movements many times over. Image Giulio Brugnaro
answers for how we might proceed to create a better method for carving tools, the
premise that we might find much more effective and efficient methods for using and
creating custom wood products without requiring highly refined and manufactured
wood products is compelling. If we can begin to learn from and respond to the innate
characteristics of a material while also employing advanced technology, we can avoid
the pitfalls which have plagued much of automated construction. And by embracing
and celebrating the raw characteristics of wood and finding methods for accurate yet
responsive methods for working with it, we can dramatically reduce the embodied
energy required to make custom wood products, while increasing the potential for
expressing its beauty.
4 Adaptability Through Digital Scanning
Another exploration which attempts to reduce the Emergy of a wood product through
technology is ‘Bandsawn Bands’ by Ryan Luke Johns and Nicholas Foley. This
project attempts to use a bandsaw mounted to a robotic arm to responsively cut
irregular wood fitches. The responsive technique minimizes the amount of waste and
once again works to express the grain and irregularities of the material rather than
work through it, to find rectilinear geometries (Fig. 6).
Fig. 5 Chisel mounted to robotic arm and resulting tests which were used to train a neural network.
Image Giulio Brugnaro
The process for optimizing the capabilities and therefore the forms they are capa-
ble of rendering, begins with a three-dimensional scan of the base flitch, using a
Microsoft Kinect. The flitch is then ascribed a factor of thickness which is defined
by the designer, typically driven by the structural needs of the application. The thick-
ness dimension is then used to divide the flitch shape into a maximum number of
sub-surfaces (Fig. 7).
The tangents of each cut through the form are paired with the edge of another cut
profile, to ensure that there is a continuous surface creation. The limitations of the
bandsaw mean that the profile cut must result in a ruled surface.
The researchers point out that this method emulates the methods of a butcher
much more than they do approximate those of any kind of typical sawmill operation
(following the edge of a bone rather than cutting through it). The found geometry
results in a particular use and permutation that could be scaled by adding in more lay-
ers as in the production of a three-dimensionally curving CLT panel, or by expanding
the use of single tree to create larger surfaces of the same material. Additionally, the
surfaces of the material which are exposed through these ‘aware’ cut lines, expose
an aesthetic not typically found through typical sawmill operations.
Another project which applies a similar approach at a different scale is the Sawmill
Shelter that was designed and built by students and faculty at the Hooke Park facility
run by the Architectural Association. This project used a sawmill (the same one they
were designing a shelter for) to slice parallel cuts into raw cut and dried trees from
the site. These parallel cuts enable the students to then steam the cut end of the tree
904 C. Beorkrem
Fig. 6 Robotic arm with band saw end-effector slicing raw cut wood flitch, following the natural
grains of the material. Image GREYSHED-Ryan Luke Johns & Nicholas Foley
Fig. 7 Slices of flitch cut using mapped logic with robotic arm mounted band saw. The results are
doweled together to create a smooth doubly curved surface. Image GREYSHED-Ryan Luke Johns
& Nicholas Foley
and bend it into a hook shape which matched profiles necessary for the design. Once
formed and dried the tree was connected back together to permanently ‘find’ the
hook shape. These members were used to anchor tension members over the roof.
Fig. 8 The exterior skin of the MAS house was scanned and then robotically assembled to optimize
the location and use of each individual panel. Image Prof. Philipp Eversmann
4.1 MAS House
The MAS House by students and faculty at the Master of Advanced Studies program
in Digital Fabrication at ETH Zurich exemplifies another approach for minimizing the
amount of embodied energy necessary for using wood components. This approach
automates a process of scanning and appropriately installing non-uniform wood
shingles onto a complex geometry. The house is a two-story construction which was
intended to implement an automated construction method using multiple robotic
arms and tools to manufacture complex trusses. While this system in novel and
compelling in its own right, the methods developed for skinning the system are where
the designers/researchers used an approach which again introduces a method that
represents a new technological approach capable of lowering the embodied energy
of the construction by adapting non-uniform components to the system (Fig. 8).
The project’s skin was conceived with the goal of minimizing how material was
consumed in the process of manufacturing which began by manually splitting the
wood to create a natural texture on the face of each shingle. Manually splitting
on the natural grains of the shingles also increases the durability and life of each
shingle. The robotically manufactured truss system was dense enough as to not
require a substructure for the exterior shingles. These shingles were assembled onto
the prefabricated truss components with a setup that included a custom end effector
and process which included; a sensor for scanning the shingles, a vacuum gripper
and two nail guns driven by a pneumatic trigger.
906 C. Beorkrem
Fig. 9 The custom-made end effector for the robotic arm, scanned each piece, tested it against
available locations and then gripped it using a vacuum gripper in the appropriate location so that
the two pneumatic air guns were in the correct relative location to mount on to the structure of the
house. Image Prof. Philipp Eversmann
The system ran a linear light-reflex scanner along the face of a shingle loaded into
the ‘shingle-feeding station’, to derive the thickness. This thickness is uploaded to a
system which has the prefabricated structural information and Python script is used
to determine the appropriate placement of the shingle as well as the gripper location
appropriate for ensuring the nail guns will fire into the structure to secure the shingle
to the structure (Fig. 9).
This automated tool and its relative scripting process works to minimize the
amount of material processing which is necessary to use CNC specific tools, and
increases the durability of the system, while positioning each component in a glob-
ally computed location. The system represents a strong set of methods for adapting
tools to materials in an effort to reduce the embodied energy necessary for the system.
In a prefabrication facility this setup could further increase the usability of individual
panels by sorting them to different assemblies across multiple designs.
4.2 Wood Chip Barn
The Wood Chip Barn was designed and built by students and faculty at the Hooke
Park facility run by the Architectural Association. This project’s primary feature is a
three-dimensional truss which is manufactured from tree branches which otherwise
would have been milled into far smaller components but instead constituted the
appropriate angles for accommodating the necessary loads for this geometry. The
Fig. 10 The completed wood chip shed structure, with an approximately 25 m span made of ‘Y’
shaped raw cut branches. The final truss shape was 10 m wide, 25 m long and its maximum height
reaches approximately 8.5 m. Image Valerie Bennett
researchers undertook the project with the premise that the ‘Y’ shapes typical of
most tree growth, which are typically avoided by most sawmill operations, and in
this instance actually constitute a structural opportunity (Fig. 10).
Similar to other examples in this chapter this system requires the three-
dimensional scanning of each component, but in contrast with some of the other
examples, this system used a method of photogrammetry which over the last few
years has become increasingly accurate. The system takes a series of photographs of
the object and stitches them together to form an approximation of the object in the
photographs. The designers noted that the results for this method had an average error
of 3–15 mm. The resulting models had transverse contours cut at regular intervals,
the centroids of which were used to identify the spline at the center of each member.
Using an algorithm, an ideal location for each component was found. The com-
ponents were then milled using a robotic arm, which followed the grain of the wood
to create a relatively smooth finish to each component and located holes for securing
the components together. The components after being finished still expressed the
surfaces of their original geometry which gave the overall look of the system an
organized yet natural aesthetic.
908 C. Beorkrem
5 Conclusion
Computational methods and digital fabrication techniques are changing the way
that buildings are conceived, materials are used, and assemblies are constructed…
Emerging technologies allow buildings to anticipate a future adaptation or disas-
sembly. “Approximately 61% of all construction projects are retrofit projects” (U.S.
Green Building Council 2015). “One billion square feet of buildings are demol-
ished and replaced with new construction each year” (Preservation Green Lab of
the National Trust for Historic Preservation 2011). Once two materials are glued or
adhered together, the material energy becomes bound, dissipated, and the likelihood
of energy feedback is null. Designing for disassembly would decrease the amount
of material waste sent to landfills and would increase the number of recyclable and
reusable material assemblies, dramatically reducing the Emergy of any given design.
By considering how technology can afford new adaptations of old materials we can
greatly expand the impact of using these materials on our environment.
The variety of projects described here encompasses at least a partial turn away
from the hylomorphic habits of typical uses for wood construction. The use of energy
throughout the design process affects the performance of design and its perception
by society. The use of computational tools, even with all their naïve flaws allow
for designers to process the intimate and individualized kinds of information that
only nature can provide in ways that start to create meaning. By considering the
consumption of energy, by both the planet and our imposition upon it we can begin
to understand the nuanced impact that each design decision has upon our inhabited
environment and our world for years to come.
References
Bernheimer A (2014) Timber in the city design and construction in mass timber. First ed. ORO
Editions
Brugnaro G, Hanna S (2017) Adaptive robotic training methods for subtractive manufacturing.
In: ACADIA 2017: disciplines & disruption proceedings of the 37th annual conference of the
association for computer aided design in architecture (ACADIA), pp 164–169
Carpo M (2011) The alphabet and the algorithm. MIT Press, Cambridge
De Landa M (2002) Philosophies of design: the case of modelling software. In: Verb: Architecture
Bookazine. Actar, 1 Jan 2002
Design + Make: Sawmill Shelter-Hooke Park, Architectural Association, London. http://pr2017.
aaschool.ac.uk/Sawmill-Shelter
Devadass P, Dailami F, Mollica Z, Self M (2016) Robotic fabrication of non-standard material. In:
ACADIA 2016: Posthuman frontiers: data, designers, and cognitive machines. In: Proceedings of
the 36th annual conference of the association for computer aided design in architecture (ACADIA)
Dierichs K, Wood D, Correa D, Menges A (2017) Smart granular materials: prototypes for hygro-
scopically actuated shape-changing particles. In: ACADIA 2017: disciplines & disruption [Pro-
ceedings of the 37th annual conference of the association for computer aided design in architecture
(ACADIA). ISBN 978-0-692-96506-1] Cambridge, MA, 2–4 Nov 2017, pp 222–231
Eversmann P, Gramazio F, Kohler M (2017) Robotic prefabrication of timber structures: towards
automated large-scale spatial assembly. Constr Robot 49–60
“Green Building Facts.” Articles|U.S. Green Building Council. US Green Building Council, 23 Feb
2015. Web. 22 May 2015. http://www.usgbc.org/articles/green-building-facts
Johns R, Foley N (2014) Bandsawn bands feature-based design and fabrication of nested freeform
surfaces in wood in robotic fabrication in architecture. In: McGee W, Ponce de Leon M (eds) 17
Art and design
Kolarevic B (ed) (2003) Architecture in the digital age: design and manufacturing. Spon Press, New
York
Mann C (2014) How to talk about climate change so people will listen. The Atlantic
Menges A (2007) Responside surface structures. In: Beesley P, Neumann O (eds) Future wood:
innovation in building design and construction. Riverside Architectural Press, Vancouver, pp
74–79
Menges A (2012) Morphospaces of robotic fabrication: from theoretical morphology to design
computation and digital fabrication in architecture. In: Rob|Arch 2012: robotic fabrication in
architecture, art and industrial design (2013). Springer, Vienna, pp 28–47
Moe K (2013) Convergence: an architectural agenda for energy, 1st edn. Routledge, N.P. (Print)
Reichert S, Menges A (2010) Responsive surface structures, Bionik: Patente aus der Natur. In: Pro-
ceedings of fifth bionics conference, Bionik-Innovations-Centrum (B-I-C), Bremen (Germany)
22–23 Oct 2010, pp 28–35
The Greenest Building: Quantifying the Environmental Value of Building Reuse. Rep. Preservation
Green Lab of the National Trust for Historic Preservation, 2011. Web, 22 May 2015
United Nations (2014) World urbanization prospects, the 2014 revision highlights. The UN Depart-
ment of Economic and Social Affairs, 7
Housing Prototypes, Timber Tectonic
Culture and the Digital Age
Colabella Sofia and Gardiner Blair
Abstract Arguably the balloon frame exemplifies the commencement of the embed-
ment of structural performance within timber construction standardisation and a sys-
tem innovation responding to socio-technical issues in domestic construction. Three
recent residential architecture prototypes which embrace digital design to fabrication
are discussed as continuing this tradition. Held as exemplars of the capacity poten-
tial of digital design to file-to-factory these projects offer an opportunity to reflect
on questions related to material culture, the social networks of construction and the
boundaries between architecture, structure, materials, and construction. This chapter
raises a series of discussion points centred around the role of timber-based products,
in a digitally enabled domestic construction industry.
Keywords Housing prototypes · Digital design · Timber innovation

Digital fabrication · Tectonic culture · CAD/CAM
1 Introduction: Fast Wood in the Modern World
Timber, arguably the second-oldest material worked by the human hand, after stone,
has been used as one of the necessary building materials for millennia. It, therefore,
has an embedded history with architecture especially in the development of domes-
tic architecture. Its use derives from locally sourced species, autochthonous cultures,
traditions, experiences and skill sets. These broader characteristics have influenced
specific regional forms of domestic timber building (Dangel 2016). Its historical,
social roots lie with carpenters responsible for developing a comprehensive design
that united the roles of the architect, structural engineer, and fabricator in one person.
After essential design features had been agreed upon with the client, the carpenter
C. Sofia (B) · G. Blair

Faculty of Architecture, Building and Planning, Melbourne School of Design, Parkville, Australia
e-mail: s.colabella@unimelb.edu.au
G. Blair
e-mail: b.gardiner@unimelb.edu.au

912 C. Sofia and G. Blair
followed customary rules of practice for selecting the type of construction, determin-
ing the building’s geometry, and developing the details. The work of the carpenter
pointed to the overall composition and proportions rather than precise measurements,
which resulted in no two structures being precisely alike (Dangel 2016). Industriali-
sation saw a shift in this mode of production which fragmented the process to increase
production efficiencies and establish a division of labour (Pfammatter 2008). Indus-
trialisation also contributed to the development of new timber-based products, which
have given rise to structural systems and use that go beyond the limits of traditional
wood construction.
Setting aside the licit doubt on the real sustainability of harvesting trees, the use
of managed forests as the source of certified timber products has gained norma-
tive acceptance. The use of global certification schemes, exemplified by the 1993
Forest Stewardship Council—FSC, and Programme for Endorsement of Forest Cer-
tification, has widened from protection of forests to the incorporation of the social,
economic, and cultural implications that arise from the commercialisation impacts
of timber (Falk 2005). The preserve of this mindset is supported by the use of renew-
ably harvested softwoods and its application in domestic construction arising from
supply capacity maintenance in those industry sectors, historically embedded to the
timber frame, and its continued relevance in contemporary labour markets interest
in low skilled assembly techniques.
The advent of digital tools, it is purported, offers the capacity to continue the trans-
formation of the timber tradition and its application in the domestic construction sec-
tor. These influences range from design conceptualisation to advanced applications
which combine and integrate architectural and structural engineering perspectives.
Digital working methods, such as CNC machines and associative parametric mod-
elling, have entered the shop floor, but some have highlighted that they are not a
wholesale replacement for what came before them (Smith 2008).
In this chapter, the interest in the continued relevance—in the digital age—of
timber as a housing material of choice is interrogated as are the implications of
digital design to fabrication in the residential construction sector. Tracing the bal-
loon frame as exemplifying, in the modern age, the commencement of the interest
in the embedment of structural performance within timber construction standardi-
sation, three recent examples that develop this interest through the digital design-
digital fabrication nexus are discussed. The first two have been chosen in the context
of the MoMA exhibition 2008 entitled ‘Home Delivery: Fabricating the Modern
Dwelling’, curated by Barry Bergdoll and Peter Christensen: (1) ‘Burst*008’ beach
house, designed by Jeremy Edmiston and Douglas Gauthier; (2) the ‘Digitally Fabri-
cated Housing for New Orleans’, from the MIT School of Architecture and Planning
guided by Professor Lawrence Sass and (3) ‘WikiHouse’, an open-source project
developed by Alastair Parvin. The three have all been held as exemplars of the extent
digital design can offer a chance to cross the boundaries between architecture, con-
struction and structure by offering tools for the customisation of the components
and the control of the constituent components. They have also been put forward as
being at the forefront of successful models of digital design and digital fabrication
Housing Prototypes, Timber Tectonic Culture and the Digital Age 913
in the use of timber-based materials, geared to low-cost, low skill labour use as a
transferable housing solution.
A series of questions arise, though not necessarily answered.1 Do digital tools
represent true innovation in timber building or just another iteration? Do digital
tools integrate design and making, craftsmanship and structural compatibility for a
material historically connected to these modes of production? Do such tools offer
the opportunity to reconfigure links in the construction supply chain and offer the
designer to actively engage in the making of buildings and the makers of buildings
engaging with the conceptual design of them? Do digital tools offer unprecedented
access to intellectual capital and information exchange which facilitates permeation
of capacity building access via a construction material which has traditionally cir-
cumvented formally driven social structures?
2 Tectonic Culture in Timber Architecture: From

the Industrial Revolution to the Engineered Timber
Innovation may be described as a non-trivial change via an improvement in a process,

product or system to the entity introducing the change (Freeman 1982). A selected
example of these forms of innovation in the 19th and 20th century suggests that
changes to a ubiquitous construction material as timber has the potential for signif-
icant and widespread impact if picked up and applied due to its embedded nature
within the industry as a whole. However, to do so effectively, certain conditions need
to be met which are related to the structural conditions and frameworks which are
unique to the construction industry.
The industrialisation of timber production offered an alternative to the archi-
tecture of heavy timber components, which prescribed a careful fitting of compo-
nents through complex joinery (Dangel 2016). Amongst the most significant innova-
tions in timber construction, whether attributed to George Snow or Augustine Taylor
(Sprague 1981) in the early 19th century, were the development of the balloon frame
and its refinements in the platform frame and panel-braced frame. Facilitated by
the advent of the steam-powered circular saw, which made boards and small wood
framing members inexpensive and readily available, and by the industrial mass pro-
duction of cheap nails (Dangel 2016). This innovation responded to the demands of
territorial expansion and development, dictates of timber scarcity, and permitted the
use of unskilled labour and minimal use of specialist tools.
Due to its flexibility, the speed of erection, and low cost, platform framing pre-
vails as the most commonly used construction type for residential and small-scale
commercial buildings in many parts of the world. The further refinement of this form
1 As Mario Carpo argues in his book “The Alphabet and the Algorithm” (2011), indeed, ‘Technolo-
gies change rapidly “new” technologies in particular. To predict, and even interpret, new develop-
ments in cultural technologies on the basis of their recent history is risky, as one needs to extrapolate
from a curve that is too short and build on evidence that has not been sifted by time’.
of timber construction has not arisen due to the system but due to other pressures
which impact developed economies. Amongst these is the cost of construction project
financing which has led to solutions that are geared towards the reduction of the con-
struction program. Equally important is the cost of on-site labour and the reliance
on competitive subcontracting which are quick to adopt solutions that reduce on-site
labour time. Hence, we have seen the fragmentation of the supply chain through the
rise of prefabricated timber framing systems and the demarcation of factory-based
and on-site labour. This potential has applied to the constituent parts of the build-
ing but has failed to develop the early 19th century’s interest, through timber and
steel prefabricated housing as a response to colonisation and 20th-century examples
of mass consumption such as in the ‘Sears Catalog Homes’ in expanding it to the
building as a whole.
The early application of the balloon frame such as George Snow’s warehouse
building (1832) and St Mary’s Church both in Chicago (1833) has been argued
as arising from a response to the demands of rapid urbanisation, accessible timber
resources and a shortage of skilled carpenters. The design response was for the
systematisation of the use of scantling timber and nails as the primary connecting
method (Sprague 1981). Since Giedion’s interest in this form of construction (Giedion
1941), the balloon frame and its development have been seen as being the most
influential innovation in timber construction Over time, this system became fully
industrialised; whole houses came to be packed and shipped as a kit of numbered
pieces for fast and ease of assembly. The invention of the balloon frame, and its
development into the ‘Platform frame’, using one story high studs, has also been
claimed as the antecedent of the skyscrapers. Pfammatter suggests that they hold a
common structural analogy and arise due to common spatial and economic demands
(Pfammatter 2008). However, fire regulations in response to the 1871 Chicago Fire,
new processing techniques for iron, steel, and concrete, as well as the appearance of a
new architectural typology, confined timber to a marginal niche of low-cost material
for small residential buildings.
The invention of glued laminated timber channelled a significant upturn in timber
construction beyond the confines of domestic construction. The acquisition in 1906
by Otto Hetzer, a Weimer master carpenter, of a patent for the production of beams
composed of a number of lamellas glued under pressure, allowed Hetzer to develop
a new method of timber construction. By using smaller sections and less structurally
refined wood, this new material unbounded the selection of timber from the dictates
of species and trunk size and heralded a widespread use of engineered timber con-
struction. Hetzer’s patent was further developed into the ‘Glulam’ product which
has become a staple response in contemporary long-span timber applications. Het-
zer’s invention revolutionised the use of timber and offered an apposite, predictable,
and reliable transposition; it eventually led to the further evolution of a large set of
engineered timber-based products. Glulam, allowed the separation of the structural
components from the spatial enclosure, expanding the possibility of plan design and,
by doing so, competing in a niche traditionally dominated by concrete systems (Falk
2013). However, the output still replicated traditional formal architectural expression
and took time to percolate through the strata of building typology, eventually serving
Fig. 1 Patent ‘Building Structure’ used for the General Panel System, No. 2,355,192, granted to
Konrad L Wachsmann and Walter Gropius on August 8, 1944. From left to right: sheets no. 1, 2
and 4. Source Google Patents [online]. Available at: https://patentimages.storage.googleapis.com/
74/c5/31/093da0152d79c0/US2421305.pdf
as major load bearing elements in domestic buildings rather than as a complete sys-
tem of construction. Further iterations of timber-based products, such as plywood,
allowed the evolution of light frame constructions into the panel construction system.
In 1941 the German emigre architect Konrad Wachsmann revisited his earlier
involvement in the design of prefabricated wood houses by collaborating with another
United States emigre, Walter Gropius to develop the ‘General Panel System’. Wachs-
mann and Gropius’s ‘Packaged House’ system sought to advance wood-frame, pan-
elised houses into a mass-produced industrial product as a response to wartime
demand (Herbert 1984). The idea lay in developing a modular timber-based building
system which was three-dimensionally symmetrical where it was possible to form
any combination of connections in any direction (Fig. 1). Despite a factory geared
for production of 10,000 houses per day, significant investment by the US National
Housing Agency, a combination of malfunctioning equipment, Wachsmann’s punc-
tilious nature and a lack of opportunistic sales and marketing skills contributed to
sales of just fifteen houses by 1948 (Davies 2005).
The ‘Packaged House’ is a telling reminder of the risk and degree of intellectual
and capital investment required in developing a system of timber construction that
lies outside the normative operative system of the construction industry. A timely
solution to the unique dictates of wartime demand took too long to respond, by the
time it could, the normative operative system of the construction industry fell back
to overwhelm the potential of the widespread application of timber prefabrication as
a housing solution.
The industry adopted Hetzer’s ‘Glulam’ invention as it could be integrated, by
substitution, into an existing already established system of residential construction.
The ‘Packaged House’ did not as it asked the industry to change the system. The bal-
loon frame represents an interesting study of construction innovation as it established
a different architectural paradigm (a rhythm dictated by ready-made timber framing
members, carrying an equal distribution of the building’s vertical compressive load)
concealed into a recognisable shell, indifferent to the rhythm of the structure. Most
of all, the invention of balloon framing anticipated the idea of using a systemised
structural solution that could be used by non-specialist users for architectural manip-
ulation. Advocates of digital-based production argue that through parametric design
software geared for the non-specialist, a similar revolution in capacity building in
the residential construction sector is available. Timber-based materials are viewed as
the most appropriate means by which such broad-based capacity may be delivered.
3 A Question of Tolerance
Early attempts to integrate human and mechanical outputs via computer-assisted

processes raised questions on the nature of this engagement and its potential ramifi-
cations. The ‘M.I.T. Computer-Aided Design Project’ at the MIT Lab, active between
1959 and 1967, sought to evolve machine systems to permit the interaction of human
designers and computers in design applications (Cardoso Llach 2015; Ross 1960).
On the manufacturing side, they sought to transfer human engagement to an auto-
mated and controllable digital process. While the CAD process turned into a compli-
cated definition of mathematical and geometrical codes, the translation of the manual
craft operating a milling machine resulted in a complex socio-technical system that
demanded new skills and new kinds of craft relocating work shifting from the shop
to the programmer’s desk (Cardoso Llach 2015). Today, the interaction of human
and machine has not made craft obsolete but has made it accessible to the designer.
It has gained new relevance in the production process not only in the translation of
design conceptualisation but the capacity to test, verify and adjust the design idea
and then deliver it through a CAD/CAM interface. Computer-aided tools enable fully
automated controlled machining, to provide tolerances and preciseness previously
challenging to produce. This accuracy has allowed the construction of complex tim-
ber forms, as witnessed by the ICD/ITKE Research Pavilion 2011, 2015–16 and the
robotically fabricated lightweight timber shell of the Landesgartenschau Exhibition
Hall 2014 (Menges et al. 2017).
The architectural historian and critic Mario Carpo has questioned how and if the
‘digital revolution’, through its new digital technologies, is significantly changing
the way architecture is and would have been designed and made (Carpo 2011). The
interplay of technology and tool undoubtedly have shaped new design methods and
perspectives, both in the early stages of design and in the fabrication and construction
phases. The computer has expanded the capacity for formal manipulation and visual-
isation. Verification of these formal transpositions may occur in real time, integrating
structural and performance considerations. The translations of these are now possible
to be extended directly into a manufacturing and fabrication process, with construc-
tion strategies able to be tested before committing to delivering architectural output.

By way of just one example in its application to timber construction, a critical aes-
thetic look and structural feature of timber construction is its jointing method. Timber
joints play a decisive role in the structural stability of the entire system, as well as
in the appearance of the timber system where exposed. Their detailing and delivery
have been based on the tacit and explicit knowledge of trade-based craftsmanship
and expertise. Digital design and fabrication have provided a capacity for greater pre-
cision, accuracy, expressive potential, and resolution to timber joints. This identity is
enabled by the millimetric precision allowed by the digital aids in conceptualisation
and the manufacturing processes in the transference of design idea to delivery (Garcia
2014). The inverse of this argument is the nature of construction lies in its built output
and consideration equally applies to the operative parameters that come to bear once
the digital interface has been traversed. What may be conceptualised or suggesting
ersatz legitimacy in the digital domain ultimately needs to have an output. This out-
put is subjected to the vagaries of a natural material such as timber, no matter how
engineered, the environment, the normative processes of the construction industry,
the realities of the site and its tolerances, and the human labour that participates in its
delivery. Equally the premise and objectives that underlie the uptake of the digital-
design-fabrication trajectory whether explicit or implicit requires consideration in
whether it is taken-up, appropriated and modified or marginalised.
4 Non-serial Timber Architecture: Three Examples

of Digital Timber for Prefabricated Houses
The investigation of such themes may be viewed via the prism offered by a 2008
MOMA exhibition, and the recent interest in ‘open-source’ architecture. The first two
arise from The Museum of Modern Art in New York exhibition entitled ‘Home Deliv-
ery: Fabricating the Modern Dwelling’. The event, curated by Barry Bergdoll and
Peter Christensen, was an attempt to outline the history of the prefabricated house, and
in the process display, five prototypes of this housing form to address different char-
acteristics of factory-produced architecture innovation (Bergdoll and Christensen
2008); two used plywood as the primary structural material. The ‘Burst*008’ beach
house, designed by Jeremy Edmiston and Douglas Gauthier and the ‘Digitally Fabri-
cated Housing for New Orleans’, from the MIT School of Architecture and Planning
guided by Professor Lawrence Sass were both designed and fabricated using digital
tools, but conceived following opposite approaches to customisation, serial produc-
tion and preciseness. The third open-source example arose a few years later, in the
form of the ‘WikiHouse’ project which used the similar digital technologies and ply-
wood as its construction material. All three demonstrate the capacity for digital tools
to stretch the domain of timber architectures through a mixed community of design-
ers, users, students, fabricators, makers and tinkerers. All three sought to embed the
needs of the end-user into an automated design process, using a systemised form
of timber construction. ‘Burst’ sought to do so via an automated bespoke capacity

for individual clients, which factored formative manipulation and integrated perfor-
mance correlations. ‘The Housing for New Orleans’ project sought to provide a rapid
response capacity to natural disasters which took into account architectural design
preferences into its generative script using simple construction techniques. ‘Wiki-
House’ sought to devise an open-source digital template that could be manipulated
by the end-client directly to produce customised solutions in a defined domain.
4.1 Burst*008 House: Customisation and the Bespoke
‘Burst*008’ designed by New York-based Jeremy Edmiston and Douglas Gauthier

(as Systems Architects), had its antecedent in an earlier project, the ‘Parish House’
(or ‘Burst*003’) designed by the same team and built in 2005 on the east coast of
Australia as a ‘holiday retreat’.
This earlier building was an opportunity to investigate the potential presented
by integrating digital techniques in design, environmental performance assessment,
fabrication and assembly utilising plywood. The challenge was not simply to offer
a pallet of predisposed options but to investigate the nature of serial production in
the utilisation of computer-aided manufacturing to generate further permutations of
response capacity in human-machine interaction. Its premise lay in appreciating and
responding to the needs of client and site but purporting the use of an algorithmic
approach in generating a design response.
The building is rectilinear, two spaces deep with three-bedrooms and amenities on
one side of the long central axis which faces the street and living, kitchen and dining
areas on the other with an adjoining external deck facing the rear yard (Fig. 2). The
street façade is windowless, with the rear façade floor to ceiling glass. Customisation
lay in its programmatic planning and in being elevated 1.5 m above the ground due to
local flood levels. Its generative form was in manually defining the external outline of
the two short-end elevations and allowing the structural load-bearing CNC-cut 50 mm
plywood rib divisions to be parametrically generated responding to the provision of
clerestory strip windows, natural light access, and passive ventilation. The ribs form
a two-way oblique grid, the intersecting points connected with custom-made, flexible
steel X-brackets (Figs. 3 and 4).
As Iwamoto has commented, ‘the system is founded in the belief that prefabri-
cation is at the scale of the construction not at the scale of the building’ (Iwamoto
2009) and it is at this level that the digital design and digital fabrication nexus needs
to be understood. Assembly and installation of digital outputs cannot be digitised, no
matter the robustness of the model or the accuracy of the CNC machine, prefabrica-
tion is only part of the construction story. The formal articulation of the architecture
and the optimisation of the plywood sheets resulted in 1100 non-identical timber
sections (as shown in Fig. 4). Although it is argued that prefabrication assembly
requires low-skill labour and in this case, architectural students provided some sup-
port, the skill set lay not in assembling a perfectly fitted and coordinated set of parts
Fig. 2 ‘Burst*003’, 2005, on the east coast of Australia: on the left, the view from the rear yard.
On the right, view of the undercroft, where the structural system is fully visible. Source Gauthier
Architects archive [online]. Available at: http://gauthierarchitects.com/work
Fig. 3 ‘Burst*008’ at the MoMA Exhibition 2008. On the left, end elevation view. On the right,
oblique view of the frame. Source Gauthier Architects archive [online]. Available at: http://
gauthierarchitects.com/work
but in the logistics of having to deal with these parts in on-site realities. The aspect
of architectural students not being considered as ‘skilled’ when it comes to digital
design and the interpretation of information so produced may also be questioned.
One could argue that they are highly skilled, potentially far more in such
techniques than the traditional tradesperson. Regardless, the critique often laid at
the architect’s feet in appreciating the nature of construction becomes ever more
paramount where digital modelling for fabrication outputs shifts upstream into the
Fig. 4 ‘Burst’ system. On the left, the components of the kit. On the right, structural components
made of CNC manufactured pieces and the X-Clip connector. Source Gauthier Architects archive
[online]. Available at: http://gauthierarchitects.com/work
design phase. Resolution of the entire system and process is required at the design
stage, with tacit knowledge of these types of folded structures still in their infancy
in the general construction industry and relatively unknown at the scale of a single
residential house. The accuracy requirement of folding structures and the tolerances
of the building site are different from those available in the digital environment. By
way of example, an issue that manifested itself in both the ‘Parish House’ and the
MOMA exhibit lay with the innate lateral instability of planar plywood elements
which relies on the connection system to provide rigidity to the structure. Although
considered at the design stage, the realities of such issues as constructability—when
one attempts to connect architectural design and building directly such shift requires
the designer to accept a broader pallet of responsibilities (Buri and Weinand 2011).
The ‘Parish House’ and the subsequent ‘Burst*008 House’ for the MOMA exhi-
bition where the system was further refined, suggest that customisation and opti-
misation are available via digital means. In formally complex buildings such as the
‘Burst’ projects, the challenge of developing a building system for potential mass
consumption which is not embedded into the normative industrial practice of con-
struction has ramifications which go beyond solving the digital design side of the
equation.
4.2 Housing for New Orleans: We Want a Shotgun House
The ‘Digitally Fabricated Housing for New Orleans’ by Prof. Lawrence Sass (Prin-
ciple Investigator, MIT School of Architecture and Planning) in conjunction with
MIT researchers, was designed to not only respond quickly to such natural disas-
ters as Hurricane Katrina2 but also to investigate the potential for design variation
2 On August 29, 2005, Hurricane Katrina made landfall in New Orleans. It was an extremely destruc-
tive Category 5 hurricane that caused catastrophic damage along the Gulf coast from central Florida
to Texas, much of it due to storm surge and levee failure.
capacity via interlocking structures, for unskilled site assembly (Sánchez-del-Valle

2011). The predecessor of this experiment is the Instant House developed by Sass
and Botha (2006) from M.I.T.’s Design Lab, aimed at studying how digital design
and fabrication can be employed in an emergency housing scenario (Sass and Botha
2006).
The formal expression or ‘styling’ of the New Orleans project was the vernacular
typology of the ‘shotgun house’. Sass chose this traditional form due to it being an
identifiable part of a place-bound collective imagination utilising a familiar cultural
trope in affording social acceptance of the use of digital tools in housing construc-
tion applications (Sánchez-del-Valle 2011). The ‘Digitally Fabricated House for New
Orleans’ represented an opportunity to demonstrate a craft-like experience, open to
customisable variation using digital design and fabrication. The timber frame con-
struction of this typology, being one room wide and four rooms long suited the use
of ‘planar construction’ outputs of digital manufacturing from CAD modelling. A
characteristic feature of this type is its ornate front porch, a distinctive symbol, and
an opportunity for the demonstration of the carpenter’s craft; the shotgun house was
also the place where carpenters and masons could pass their knowledge and skills,
allowing apprentices to learn their trade (Upton and Vlach 1986). The ‘Digitally
Fabricated House for New Orleans’ used digital design and fabrication techniques
to display that variation, distinctiveness and personalisation is available and is not
limited to prefabrication of repetition (Fig. 5). Its delivery methodology lay in the
production of a 3D design model, from which plywood cutting optimisation to suit
standard plywood sheet sizes were produced. Flat packed and bundled for assembly
by modular componentry, connections of the panels relied on friction-based inter-
locking sheet members (Sass 2007). Cognisant of the labour implications of on-site
assembly, it was built with the help of MIT students, using only rubber mallets.
Constructed out of 7000 VNC-cut plywood components, locked together by friction,
utilising no nails or screws and requiring little measuring or cutting on the worksite
(Fig. 6). With a labour force of four persons, the complete onsite project process took
23 days.3
However, as in any prefabrication system, the efficiency of on-site assembly relied
on significant upstream design resolution, detailed documentation, checking and veri-
fication and pre-planning. The team acknowledged that, for ‘smooth’ on-site assem-
bly process, it required a considerable time commitment for the CAD modellers.
Ostensively as the tasks were going beyond traditional construction documentation
as a communication method of design and performance intent, the modellers had
to accept the responsibility for tasks normally given over to and the remit of those
that follow design in the construction supply chain. They had to engage in the more
detailed tasks of consideration for structural integrity, constructability and buildabil-
ity, resource optimisation, resource planning and logistics. As a prototype, it revealed
that further development, integration and refinement was required for many of these
CAD-related functions which lay in the upstream domain of design.
3 Thebase structure was assembled in 18 days and the white ornamentation components were
assembled in 5 days. Data available on http://ddf.mit.edu/news/2014/project-summary.
Fig. 5 The digitally fabricated house for New Orleans at the MoMA Exhibition 2008: front and
lateral view. Source Professor Lawrence Sass, photo by Suzanne Camarata Ball
Fig. 6 The ‘Digitally Fabricated House’ for New Orleans at the MoMA Exhibition 2008: details
of the CNC components of the porch and 3D view of the components. Source Professor Lawrence
Sass, photo by Suzanne Camarata Ball
The ‘Digitally Fabricated House for New Orleans’ demonstrated that the capacity
of Computer-Aided Manufacturing had challenged the climate of serial production by
making possible a process of home delivery with digital fabrication machinery as an
alternative to craft and factory-based prefabrication (Fig. 7). In this capacity, it may be
Fig. 7 The ‘Digitally Fabricated House’ for New Orleans at the MoMA Exhibition 2008: front
view and details of the CNC components of the balustrade. Source Professor Lawrence Sass, photo
by Suzanne Camarata Ball
considered as a contemporary manifesto of the ‘simple technology’ envisioned by the

architect and urban designer Yona Friedman to give people in difficult circumstances
tools and ideas in access to particularising space and facilities (Friedman 2006).
4.3 WikiHouse: Open-Source Customisation
Open-source business models have been put forward as an effective strategy for
solving local issues by harnessing collective knowledge and applying this to specific
situations (Garcia 2016). The idea of a democratisation of architecture through open-
source means and citizen-centred design is not new and may be found in the ‘Mobile
Architecture’ concept of Yona Friedman in the late 50s (Friedman 2006), as well
as in the utilisation of automation of the design process in Nicholas Negroponte’s
research in the mid-70s (Negroponte 1975). One of the most recent manifestations
that coalesce these early trends is the ‘WikiHouse’ project. During a TED talk in
2013, Alastair Parvin, one of the co-founders of the project,4 argued that built envi-
ronment design carry a social responsibility that is universal, while architects, by the
nature of their protectionist approach, limit their skill set availability to a very small
and privileged market sector. The ‘WikiHouse’ approach advocates for the use of
open-source digital-based platforms enabling access to creative commons licensed
4 TheWikiHouse Project was initiated in 2011 by Alastair Parvin and Nick Ierodiaconou of 00
Architects, a London-based design practice, in collaboration with Tav of Espians, James Arthur
now with 00 and Steve Fisher of Momentum Engineering.
Fig. 8 ‘WikiHouse’—instructions for the assembly of ‘MicroHouse’, one-bed house type, designed
using the WikiHouse building system, initially developed in the UK for European contexts, by 00
Architects. The numbers 2, 5 and 7 refer to different stages of the assembly. Drawings licensed
under CC-BY-ND. Source WikiHouse [online]. Available at: https://wikihouse.cc/
Fig. 9 ‘WikiHouse’—details of the component ‘Wren’, CNC manufactured using structural-grade

timber panel materials developed by 00 Architects. Drawings licensed under CC-BY-ND. Source
WikiHouse [online]. Available at: https://wikihouse.cc/
information for the construction of houses (Fig. 8). A freely available open-source
library of solutions of plans and details of connections (such as the ‘Wren’ connec-
tor, Fig. 9) and structural components, applied to timber construction cut using CNC
machining, and applicable and modifiable through ‘Google SketchUp’ to anywhere
in the world which has access to such resources allows broader access and a shared
feedback loop to improve outcomes (Isaacson 2013).
A display of one of the last iterations of the system is ‘WikiHouse 4.0’, designed
by 00 Architects in collaboration with Arup and The Building Centre for the Lon-
don Design Festival 2014. The two-storey prototype, incorporating energy efficiency
measures, was built in twelve days by a team of volunteers (Grahame 2014). The
engineering consultancy firm Arup also developed for this occasion an open-source
heat recovery unit called ‘Open MVHR’, which can be built with 3D printed com-
ponents and aluminium sheet from beer cans (Mok 2014).
This last iteration (Fig. 10) demonstrates the evolving potential of customisation
in incorporating other inputs such as solving local energy-efficiency issues and using
Fig. 10 ‘WikiHouse’—assembly process of the structural system for 1-3 storey buildings, initially
developed for European contexts, by 00 Architects. Drawings licensed under CC-BY-ND. Source
WikiHouse [online]. Available at: https://wikihouse.cc/
locally appropriate recycled materials. Such inputs have a universal cache in accu-
mulating various permutations of particular operative conditions to an adaptable base
model to be responsive to that which is local, and which is project specific. However,
although open-source web-based generative information permits the expansion of
inputs beyond geographic boundaries; its limitations lie within access to the tech-
nologies of its containment and resources at hand and at that time (Parvin 2015).
Digital platforms and tools may narrow the gap between design creativity, demand
for architecture and access to its production, and offer a method to (re)democratise
housing design and production (Anderson 2012) but only in terms of those who have
access to these tools of production.
5 Themes
The three projects offer insight into different modes of interpretation of digital-based
transference of design to fabrication. These include such issues as craftsmanship,
domains of practice, production, participation and knowledge.
Despite their limitations, the projects presented demonstrate that a process util-
ising digital-design-to-digital-fabrication techniques opens alternative avenues for
kit home production. Leveraging digital platforms, architects can embed parametri-
cally responsive design, engineering simulation, fabrication outputs and implemen-
tation into the creative process. Team-based bespoke amalgamations which dissolve
through the project development and implementation cycle, a feature of the con-
struction industry, does not afford broader shared knowledge. The view that reten-
tion of design thinking and delivery expertise offers competitive advantage lays the
construction sector open to criticism in not being innovative and offering low pro-
ductivity. An alternate view lies in the enabling presented by digital tools and the
potential for design lying in a sharing of responsive design solutions. Much debate
has passed in professional associations in the construction sector with the advent of
building information modelling on the legal and risk ramifications of a single shared
digital platform. The collaborative platform afforded by digital communication meth-
ods has already established a quasi-open-source platform, albeit contained within a
project-based system. However, cloud-based applications such as those that apply in
document and workflow management offered by external commercial entities permit
utilisation of meta-data for further commercial offerings.
Major barriers exist to such housing propositions as ‘Burst*008’, the ‘Digitally
Fabricated House’ and ‘WikiHouse’ as a simple solution for housing. These span
such things as access to capital for the investment required to engage in a digitally
driven economy, to purchase equipment, education access for computer literacy, the
technical knowledge, to operate tools like a CNC router or to customise a 3D model
and the ability to navigate the regulatory environments of planning and building
compliance. Despite the uncertainties and difficulties in grounding digitally engaged
projects, discussion behind their concept is worthy of debate and trial. The building
material that satisfies many of the criteria of digital transmutations in housing and lies
as the basis of experimentation which is historically accepted, universally understood,
accessible and workable is timber and its processed derivatives.
5.1 Craftsmanship
The Oxford English dictionary defines craftsmanship as a ‘skill in a particular craft’

or ‘the quality of design and work shown in something made by hand’, its ‘artistry’.
It also defines craft as an ‘activity involving the skill in making things by hand’. The
intent of timber engineering in plywood, for example, is the removal and regulari-
sation of the caprice of natural timber fibres in its grain by flattening and rotating
the material. One could argue to what extent the engineering of timber removes the
artifice of the craft of understanding the raw material in the manner in which behaves
and the extent to which a particular piece can be utilised. Craft, in this scenario,
no longer connects knowledge of the material or the responsiveness of the hand
(removed from the making) to the nature of the material’s properties. Crafted takes
on a figurative form of the verb, not in the nature of craft as a noun—but in giving
primacy to the process. Does digital design to fabrication then remove the acts of craft
and craftsmanship? David Pye and Malcolm McCullough (Pye 1978; McCullough
1998) have argued for a reconceptualisation of digital tools as an aid to the artisanal
tradition of designers. Digital means of design and fabrication both reconnect and
extend this tradition. The projects each display facets of this approach, equally they
offer an insight into the questions around the political economy of design and the
sociological and economic systems that come to bear within architecture and the
construction industry.
‘Burst*008’ presented a prototype, an outline for a potential system for mass cus-
tomisation and variation responsive to site and program, through formal architectural
manipulation. Its offer was to provide a quasi-parametric design tool for flat-pack
production not necessarily a fully resolved structural building solution. In this, the
hand of the constructor was still required, not only to assemble a predetermined kit
of parts but to have knowledge of how to bring the system together into a cohe-
sive whole. As with most timber-based folded structures, the formal complexity of
the timber planer elements relied on the node plate connectors to bring structural
stability to the system. The knowledge required to achieve this, lay as with most
building construction, in the constructability skill of the assembly team—in the craft
of knowing how to coordinate the process and how to put things together.
‘Digitally Fabricated Housing for New Orleans’ attempts to replicate, by mechan-
ical means, a signature or hand of the traditional craftsperson. It recognises the value
of the hand and its local variation and interpretation of timber construction. Digital
design and fabrication permit the recording of a local vernacular and the capacity to
replicate or reinterpret this into an affordable recognisable and accessible typologi-
cal output. By doing so, it places a value on the local building tradition and argues
that digital design and fabrication offers the means to retain a local and culturally
significant building design response that may be applied even in extreme conditions
such as those arising from natural disasters such as Hurricane Katrina.
‘WikiHouse’ harks back to the tradition of architectural pattern books where
template designs were shared and made available to replicate these archetypes for
implementation to a broad audience and geographic location. The crafting lies in
the process of interpretation of the template and the development of this template in
open source design platforms as customisable and adaptable uptakes framed within
the bounds of the template. As with architectural pattern books, the use of the tem-
plate does not preclude local variation and adaption but serves in part as an enabling
mechanism. The ‘WikiHouse’ approach replicates the traditional pattern book by
presenting itself as a reference where the initial underpinning design has been estab-
lished remote from the source of application. Unlike the intent of the original pattern
book, it accepts the implementation of the pattern in itself, which becomes a potential
source of development and refinement through open-source means.
5.2 Domains of Practice
The projects provide insight into differing perspectives on how digital design and
fabrication can operate at various domains of practice, in the nature of the architec-
tural client, design development and testing, within varying commercial realms and
at vastly different scales of output.
A commonality of theme in the three projects is the argument for digital-based

production mechanisms is in its universal application in the provision of accessible
design thinking, through a globally comprehensible communication means delivered
via a portable means of controlled production. The residential client may be bespoke
(Burst), tailored to a specific community (New Orleans House) or a customisable
prêt-à-porter (WikiHouse). The internet offers the architect not to be geographically
bounded nor at the mercy of varying capacities of building knowledge, labour, prac-
tice, and technologies. Production may equally operate as a singular typological
output, customisable on a mass scale or generically applied through mass production
or adapted through the world-wide-web of tinkerers. The projects, therefore, are rep-
resentative of the optimism of digital design to fabrication and its usefulness in its
flexibility to deal with scale in residential applications. This local-to-global scale of
engagement is facilitated in no small part not only due to the digital aspect of process
application but also via the utilisation of embedding a material that is universally
familiar and accessible. Timber and its derived products fulfil a critical role in these
investigations into contemporary technology-based solutions to a diversity of client
for responsive residential building production.
This digital-based form of design and production offers access to another kind of
client, not defined as the end user but as an active agent, operating further along the
supply chain and directly involved building. Once design and fabrication are directly
linked, it offers the capacity for those involved in fabrication not to be de-skilled
or re-skilled facilitators of the execution process but to be active agents in the man-
agement of design. This shift is evident in commercial building procurement where
the expertise of the builder is valued by the client and early contractor involvement
up the supply chain is becoming the norm. Just as it is argued that digital design
to fabrication reconnects the designer to the construction process, it equally offers
passage for the contractor to be an active agent in design.
The projects, therefore, are representative of design-driven innovation where the
seed of the innovation lies within the domain of a design intent output. The projects
are indicative of designers adapting to the availability of design tools and fabrication
systems to promote design solutions to conventional design problems. They also rep-
resent the dilemma and limitations that are structural from an industrial perspective,
for regardless of the rhetoric of the possibilities of digital processes to take advan-
tage of the scale of application they require capital and investment from broader
industry sectors beyond the remit of the designer. Timber offers an accessible and
cost-effective entrée, for the three project prototypes—they are, however, still await-
ing uptake within the broader construction industry. Materials and technology do play
a part in the story, robotic 3D printing of concrete or robotic automation in masonry
offer benefits to the manufacturer and the constructor which are commercially trans-
ferable to the commissioning agent. The efficiency of the timber platform frame and
its continued use in prefabrication, catering to an industrial sector characterised by
a fragmented supply chain, is an embedded system of residential building practice
where substantial innovation is difficult to penetrate.
This does not devalue the opportunities presented by the digital environment for
residential applications. The great advantage of these systems is in the developmental
sophistication of the tools to be increasingly responsive to the intuitive nature of

architectural design conceptualisation. Such tools also permit the rapid investigation
of multiple and simultaneous design investigation to nth-degrees of complexity that
may be immediately checked via parametric inputs. The connection to fabrication
systems facilitates rapid prototyping and testing, tapping into an iterative process for
performance improvement and client ratification.
5.3 Kit Home Production
The three examples can be interpreted as a demonstration of what prefabricated

housing can achieve by mining the possibilities of the computer. The architects,
to varying degrees engaged in architectural form making, share a larger interest to
develop a system of digital-design-fabrication production that may be applied in the
single-family housing sector utilising timber as its means of delivery.
Tailor-made production systems can be made broadly available through a network
of shared knowledge. Recently, new user-friendly software tools, such as ‘Grasshop-
per 3D’ and its plug-ins, which define the geometry of structures parametrically,
form a large and shared database of possible solutions and details. Negroponte’s
argument for the nexus between the human designer and the computer, as designer,
is reignited (Negroponte 1975). Democratising the production of architecture is mir-
rored in broad-based access to digital tools and integrating these into the supply
chain or in shifting the links in the supply chain. Curiously, the use of digitally-based
technologies has looked to a very traditional building material to advance its cause
in the construction industry—timber, timber engineered and timber-based products.
Kit home production has shifted from a factory packaged export with limited option-
ing to a locally enabled production method that is customised to a particular client
and site requirements. In part, this has arisen due to the means of production being
broadly disseminated, via CNC machines, the advent of sheet-based timber products
of known behavioural characteristics and properties and the development of software
tools that embed professional knowledge for the layperson.
Digital tools present a further permutation to the do-it-yourself (DIY) building
culture characteristic of the Anglosphere (see Peters 1989 for America and Dingle
2000, for Australia) but equally evident in those manifestations of self-resilience,
need or desire of many locales. This DIY approach through digital distribution offers
access to alternative methods of design realisation and construction, where one may
engage commensurate with one’s knowledge base or interest. This does not mean
that it removes the aspect of privileged knowledge, but both adds and provides access
to design conceptualisation and delivery as a kit of parts. The architect may still be
required to provide the skills commensurate with the profession, however, in the post-
representational space of the digital environment one either acquires or engages the
expertise required by the translation tools of the digital environment—3-D modelling,
algorithm editors, coding.
The kit, as used in the project examples where material behaviour is controlled
using plywood and in its parametric conditioning of performance opens the oppor-
tunity for a wider audience of enablers, participants, and assemblers. Digital design
to fabrication in its predetermined and coordinated components and its application
of simplified and consistent connection systems reduces the reliance on the crafts-
manship of trade-based experts, such as carpenters or metalworkers. Such tools are
extending to the building services sector, where digital means permits off-site pre-
fabrication to lower skilled on-site tradespersons (Korman and Lu 2011). For this to
be effective, however, digital design to fabrication requires preciseness and unpre-
dictability and deviation from the predetermined path to be removed. To utilise the
digital aspect of design, it is best to minimise the vagaries of the material and limit
the capacity of the DIY participant to intervene. For the system to operate, its con-
stituent parts and the process by which these parts come together are best when they
are predictable and maintained for as long as possible in the environment of the
digital-design to fabrication domain. Where transference ultimately occurs, as with
any prefabricated building system, axiomatic structures of communication, and com-
prehension come to play by necessity. Industrial engineered timber and digital-based
building production therefore retains the specific material culture of timber and adds
another chapter to evolving building practice.
5.4 Participation as Apprentice
The advent of the platform frame and its dissemination was a response to capacity
building using elementary components, simply and cheaply connected without the
requirement for skilled labour. The system made building construction accessible to
a far broader labour force not bounded by trade-based knowledge. It also responded
to underdeveloped supply capacity both in the availability of materials, labour and
in the maturity of a construction sector, in meeting demands of land development.
The iteration of timber frame prefabrication required another link to be added to the
supply chain. Efficiencies in this system of production lay in minimising complexity
and encouraging repetition and systematisation. Cost efficiencies in material and
labour and reduction in construction time in the highly competitive single-family
home residential market sector encourages standardisation and limits formal design
agency to manipulation of the ‘face’ of the object. One may, therefore, speculate
on the rise in interest in the digital design and digital fabrication nexus. It offers a
potential contraction of the supply chain by linking design directly to a fabrication
output, by embedding the translation interface into the ‘software’, without the need
for an additional construction compositor. It, therefore, presents an opportunity for
some formal but limited design manipulation but endless repetition. It, on the other
hand also presents an opportunity for mass customisation, high design complexity
and performance integration.
The three projects all stake a claim for the applicability of their respective systems
in the use of non-skilled workers. All systems it is argued can be assembled by a
small group of people with limited construction experience; the lightness of timber
allows that. An argument that may be equally made for the traditional timber stud
framed building. This feature of labour efficiency has been one of the mainstays
of this form of construction which has permitted the longevity of this ‘stick-based’
system in domestic construction in those economies that have embraced the timber
frame. Digital production does not alter this, assembly system methodology however
does. As suited as it is to the potential removal of the reliance of tacit construction
knowledge, and thereby locates participation as apprentice, it also facilitates the
removal of the trade apprentice as participant. For all its advocacy for an opening
of access to non-skilled labour in the domestic construction process, it also presents
another method of labour deskilling in the domestic construction sector.
The ‘Parish House’ and the M.I.T. MoMA project were built in part with the
assistance of architecture students, therefore, exemplifying the robustness of the
building systems for unskilled labour assembly. The projects exist as points in time
examples of the digital pathway in construction and equally as research investigations
and prototypes. In these terms reflecting a research method of the utilisation of the
academic environment as test beds for building research practice before potential
adoption by industry.
The case for being unskilled required some qualification—this may have been the
case as the system removes the necessity of expertise in timber carpentry. However,
such students potentially may have been overqualified as workers for the manipula-
tion of digitally enabled communication platforms, as well as the interpretation of
3D drawings and optimisation of complicated plywood shapes. By doing so, they
were anything but unskilled. Two German words are used to describe the nature of
experience; Erlebnis in the participation of a memorable event—to have an experi-
ence and Erfahrung, to obtain a skill—to gain experience. Both describe the nature of
experience that is used in student participation of digital supported learning such as
promoted by the ‘Parish House’ and the MIT MOMA project. However, they overlook
Fähigkeit—the capabilities and abilities that students possess in being digitally liter-
ate. In its open-source approach, the ‘WikiHouse’ project, presents alternate oppor-
tunities for participation in all three approaches to the engagement of experience.
The aspect of the use of unskilled labour is reflective of an industry shift in cost
management arising in those economies where the balance of project cost lies on
the labour side of the ledger rather than in the costs of materials or equipment. Such
profiles further seek to leverage returns in reducing construction time as much as
possible. The digital-design to fabrication process, therefore, is seductive for those
construction economies that seek to minimise labour cost or reduce construction
time. Capital investment in tooling is set against the costs associated with production
processes reliant on skilled labour or where substantial reduction of on-site time may
arise from such investment. In those economies where construction labour costs are
lower than material and equipment, there is little incentive on an industrial scale
for the capital investment and training required regardless of the potential low-entry
investment offered by the digital environment. In such economies, the percolation of

this type of construction innovation may take longer to have an impact locally, even
though these economies may offer such services externally.
5.5 Cross-Disciplinary Knowledge and ‘Simple Technology’
Human-machine interaction offers an avenue to embed building performance criteria

into customisable components. Digital tools facilitate the envisioning of complex-
ity, both formally and in performance terms, but also provides the means for it to
be simplified, analysed, managed, and delivered. The outputs may be packaged to
suit an industrially mature construction environment or market sectors that would
benefit from simplicity or opt for more traditional means of transposition. Mass-
customisation in timber-based architectures leverages these advantages to enable
Slessor’s idea of ‘uniqueness’ as an economic possibility beyond repetition and the
‘creative capabilities of electronics’ (Pine 1993) to much broader-based applica-
tions. The question lies in the capacity of digital design to digital fabrication to offer
a high degree of customisation, formally and architecturally, in site specificity, and
in performance criteria, while integrating resource efficiency and rapid deployment
capacity.
However, one of the key features of the construction industry is that it has a
fragmented supply chain characterised by small businesses and independent con-
tractors. This structural industrial form presents both challenges and opportunity in
both training and investments across all market sectors to embrace the offerings of
digital environments fully.
Digitally-based design and analytical tools permit the harnessing of the innate
properties of timber and timber-based products for the form making capacity offered
by the digital design environment. The elastic properties of wood may be utilised
in rib shell structures or weave configurations taking advantage of the redundancy
of the weave as a system. The strength characteristics offered by glue-laminated
plywood and shape profiling lend itself to applications from traditional load-bearing
arrangements to folded structures (Buri and Weinand 2011). The environmental and
economic advantages of timber-based products coupled to its low-skill labour utilisa-
tion further weigh-in for the continued relevance of timber as a material in emerging
parametric design and digital fabrication methods.
It could be argued that the three houses exemplify the use of digital tools to offer
access to intellectual capital and information exchange via a construction material
that, in its ancient configuration, traditionally circumvented formally driven social
structures.
6 What’s Meant to Be Will Always Find Its Way
The ubiquity of technology has altered material culture, with capacity building under-
going exponential expansion to the possibilities and permutations of its application. It
could be argued that the three projects exemplify the potential of digital tools to offer
access to intellectual capital and information exchange via a construction material
that, in its ancient configuration, traditionally circumvented formally driven social
structures. Is there potential for a wider spectrum of the social structure of building
to present itself? The linear supply chain of building is potentially disrupted—where
the direct links of design conceptualisation are integrated with fabrication and assem-
bly method. A challenge is presented to the designer on several fronts as it not only
requires knowledge of construction but also requires the designer to participate in the
delivery of building. Equally, it presents the constructor with a further opportunity
to move up the supply chain and into the conceptualisation and design production
phase.
In utilising digital tools in the service of architecture using a natural material
like timber; one still confronts the unpredictability of such materials. Analysis and
simulation take one just so far until one faces a requirement for tacit material and
construction knowledge. The three projects discussed here have used plywood as the
primary structural material, in this use of engineered timber one is further removed
from an appreciation of natural material properties. However, in terms of building
knowledge, the use of digital design and fabrication in timber still presents a Janus-
like opportunity, it permits one to look to the future while simultaneously arguing
for an appreciation of the lessons of the past that are embedded in an understanding
of tacit knowledge of building.
Furthermore, timber whether engineered or otherwise lends itself in the digital-
design to fabrication nexus of production scale. It can equally be utilised for mass
production, mass customisation or the bespoke. The three examples engage with the
nature of scale and production method. They offer a capacity for a mixed community
of designers, of student leverage and engagement, of manufacturers, fabricators,
assemblers and tinkerers operating in residential building supply, be it as a generic
globally applied template, as an enabler to a specific community of practice or as a
customisable bespoke singular product.
Digital-design-to-digital-fabrication does not negate appreciation of materiality. It
highlights the difference between point-in-space robotic construction of masonry or
sprayed concrete and timber, where the behaviour of the material and its connections
still retain the input of the hand. Intuitive and tacit knowledge remains, tied to the stick
construction assembly method of timber, whether it be in the adjustment of the slip
joint connections of a timber gridshell, the rubber mallet and crowbar of the ‘Digitally
Fabricated House for New Orleans’ or ‘WikiHouse’ or the node connections of
‘Burst*008’.
The saying What’s meant to be will always find its way is often quoted as an
aphorism for a statement of will and applied intent. The inevitability of the digital
age and its will of propulsion, the ubiquity of the internet of things will eventually
find an avenue for the implementation of digital-design-fabrication to be endemic in

the timber construction sector. However, the saying equally may be applied to the
concept of time. Perhaps the applicability of a traditional construction material like
timber remains timeless and a contemporary construction solution equally suited to
the digital age. What’s meant to be will always find its way.
Acknowledgements The catalyst for much of the thoughts outlined in this article arose through
numerous discussions with our colleagues at the University of Melbourne, Prof. Paolo Tombesi
has been a feature of many of these conversations, for which we are appreciative. Some of the
information on ‘Burst’ House*003 and *008 are gathered from a research project on the indus-
trial potential of digital fabrication technologies commenced by the late University of Melbourne
academic, Professor Bharat Dave, and concluded by Paolo Tombesi—chief investigator, Blair Gar-
diner—investigator, Sofia Colabella—research assistant; (Tombesi et al. 2016).
References
Anderson C (2012) Makers: the new industrial revolution. Cornerstone Digital, Rushden
Bergdoll B, Christensen P (eds) (2008) Home delivery: fabricating the modern dwelling. The
Museum of Modern Art, New York
Buri H, Weinand Y (2011) The tectonics of timber architecture in the digital age. In: Kaufmann H,
Nerdinger W (eds) Building with timber paths into the future. Prestel Verlag, Munich, pp 56–63
Cardoso Llach D (2015) Builders of the vision: software and the imagination of design. Routledge,
London
Carpo M (2011) The alphabet and the algorithm. The MIT Press, Cambridge MA
Dangel U (2016) Turning point in timber construction: a new economy. Walter de Gruyter GmbH
Davies C (2005) The prefabricated home. Reaktion Books, London
Dingle T (2000) Necessity the mother of invention, or do-it-yourself. In: Troy P (ed) A history of
European housing in Australia. Cambridge University Press, Cambridge, pp 57–76
Falk A (2005) Architectural aspects of massive timber—structural form and systems. Doctoral thesis
Luleå University of Technology, Sweden. https://www.diva-portal.org/smash/get/diva2:999695/
FULLTEXT01.pdf
Falk A (2013) Cross-laminated timber: driving forces and innovation. In: Cruz (ed) Structures and
architecture: concepts, applications and challenges. Taylor & Francis Group, London, pp 511–518
Freeman C (1982) The economics of industrial innovation. Francis Pinter, London
Friedman Y (2006) Pro domo. Actar, Barcelona
Garcia M (ed) (2014) Future details of architecture. Architectural design (AD) series. Wiley, New
Jersey, pp 14–25
Garcia M (2016) Architectural patents and open-source architectures: the globalisation of spatial
design innovations (or learning from ‘E99’). In: Architectural Design (AD), vol 86, no 5. Wiley,
New Jersey, pp 92–99
Giedion S (1941) Space, time and architecture: the growth of a new tradition, 1st edn. Harvard
University Press, Cambridge
Grahame A (2014) How to build your own sustainable house. In: The Guardian, 23 Septem-
ber https://www.theguardian.com/sustainable-business/2014/sep/23/build-own-sustainable-eco-
house-wikihouse
Herbert G (1984) The dream of the factory-made house. Walter Gropius and Konrad Wachsmann.
The MIT Press, Cambridge MA
Isaacson B (2013) WikiHouse democratizes design for inexpensive, easily built homes. In: The Huff-
ington Post, 1 March https://www.huffingtonpost.com.au/entry/wikihouse-design-inexpensive-
homes_n_2790860
Iwamoto L (2009) Digital fabrications: architectural and material techniques. Princeton Architec-
tural Press, New York
Korman T, Lu N (2011) Innovation and improvements of mechanical, electrical and plumbing
systems for modular construction using building information modelling. In: Proceeding of the
architectural engineering conference, Oakland, California, pp 448–455
McCullough M (1998) Abstracting craft: the practiced digital hand. The MIT Press, Cambridge,
MA
Menges A, Schwinn T, Krieg OD (eds) (2017) Advancing wood architecture: a computational
approach. Routledge, London
Mok K (2014) WikiHouse: open source sustainable house designs that anyone can build. In: The
New Stack, 1 Oct
Negroponte N (1975) Soft architecture machines. The MIT Press, Cambridge MA
Parvin A (2015) An open manifesto for housing, https://medium.com/@AlastairParvin/an-open-
manifesto-for-housing-a56cf4eed2a1
Peters T (1989) An American culture of construction. Perspecta 25:142–161
Pfammatter U (2008) Building the future: building technology and cultural history from the indus-
trial revolution until today. Prestel, London
Pine BJ (1993) Mass customization: the new frontier in business competition. Harvard Business
School Press, Boston
Pye D (1978) The nature and art of workmanship. Cambridge University Press, London
Ross DT (1960) Computer-aided design: a statement of objectives. Technical memorandum,
MIT Electronics Systems Laboratory. http://images.designworldonline.com.s3.amazonaws.com/
CADhistory/8436-TM-4.pdf
Sánchez-del-Valle C (2011) Digital craft, design and the work of Lawrence Sass. Int Rev Afr Am
Art 23(3):37–40
Sass L (2007) Synthesis of design production with integrated digital fabrication. Autom Constr
16(03):298–310
Sass L, Botha M (2006) The Instant house: a model of design production with digital fabrication.
Int J Archit Comput 4(4):109–123
Smith M (2008) Associative design in fabrication. In: Kolarevic B, Klinger K (eds) Manufacturing
material effects: rethinking design and making in architecture. Routledge, New York
Sprague PE (1981) The origin of balloon framing. J Soc Archit Hist 40(4):311–319
Tombesi P, Gardiner B, Colabella S (2016) Is conventional knowledge enough? In: Chien S et al
(eds) Living systems and micro-utopias: towards continuous designing. Proceedings of the 21st
international conference of the association for computer-aided architectural design research in
Asia CAADRIA 2016, pp 871–880
Upton D, Vlach JM (1986) Common places: readings in American vernacular architecture. Univer-
sity of Georgia Press, Athens
Performative Architecture and Wooden
Structures: Overview on the Main
Research Paths in Europe
Angelo Figliola and Alessandra Battisti
Abstract One of the main aspects investigated in the European research context
on Performative Architecture is related to the use of digital innovations in wood
structures construction of units and technological systems as well as architectural
organisms at 1:1 scale. To analyze the different approaches the contribution proposes
a series of case study and the results of two applied research, the 1 to 1 scale pavilions
Fusta Ròbotica and Digital Urban Orchard. The case studies are selected verifying
the correspondence to the following parameters: the presence of a performance-based
process through which explore informed architectures; the use of low-engineered and
natural wood and the engineered one; the materialization of the digital model through
innovative manufacturing processes, specifically robotic fabrication. The contribu-
tion allows gathering pros and cons in the three different investigative macro areas:
performance-based design, material culture, and fabrication process. This analytical
investigation helps to create a clear research scenario around the topic of digital wood
design as well as the definition of an innovative pathway for future researches, look-
ing forward the assimilation of these innovative concepts in the building construction
sector.

Parametric timber engineering · Complex timber structures
Optimization of wood architectures
A. Figliola (B)
School of Architecture and Design, University of Camerino, Ascoli Piceno, Italy
e-mail: angelo.figliola@uniroma1.it
A. Figliola · A. Battisti
e-mail: alessandra.battisti@uniroma1.it

938 A. Figliola and A. Battisti
1 Introduction: Performance-Based Process and Wood

Architectures
The recent publication Parametricism 2.0 (Schumacher 2015) has highlighted the
beginning of a new testing phase that targets the use of computational algorithmic
tools for the resolution of environmental and social matters, returning to deal with
issues that warranted the birth of digital computing itself in the 70s (Frazer 2016).
The ability to process information, and then to use the data as guiding elements of
the design process (Deutsch 2015), opens many and largely unexplored possibilities
for environmental and technological design. The theoretical assumption defines the
architectural shape as a result of a diagram of forces within a morphogenetic process
(Hensel et al. 2010; Hensel and Menges 2006) and offers new investigation fields in
relation to the possibility of creating performance-based (Kolarevic 2003; Kolarevic
and Malkawi 2005; Grobman and Neuman 2012; Oxman 2009) or performance-
oriented architectures (Hensel 2013). The performance requirements made clear by
codes and regulations raise the issue of performance as a design focus at the cen-
tre of the debate. So far, the first digital era has interpreted the performance as a
necessary antidote against formal arbitrariness generated by digital processes or as
a representation of a new hyper-functional complexity. Today, performance-based
architecture represents an absolute necessity, ethical obligation of the profession due
to environmental problems1 that afflict the earth. To operate within this scenario,
the technological innovations are essential to ensure that the performances do not
remain only numerical parameters but represent a source of formal exploration and
process information. Hence, the objective of the generative process is the exploration
of design solutions that can be optimized in relation to a space of design possibilities
defined by the designers through the formulation of a meta-project (Kolarevic 2015).
This design tool is based on the definition of geometric variables and constraints,
genotype, and design goals to achieve through an optimization process. The genera-
tive process offers the opportunity to explore complex and informed geometries in a
flexible and relatively fast way, transforming the material from a passive receiver to a
design agent as driving element of creative process (De Landa 2015). The transition
from passive to active receiver is based on the exploration of its mechanical, structural
and behavioural properties to inform the design process (Menges 2012). The appli-
cation of the methodology involves the overcoming of the typological paradigm in
favour of a continuous formal variation that changes depending on system boundary
conditions. The customization of the form can be linked to a responsive interpre-
tation of local and regional variations in characteristics (Yuan 2015) thanks to the
data-driven strategy and the ability to interconnect design and fabrication in a sin-
gle workflow. To transfer an informed architecture from the digital to the physical
world, it is necessary to integrate various skills and technologies used in other areas
of architecture to implement standard processes. The birth of the first architectural
1 Theenvironmental problems like global warming, acid rain, air pollution, urban sprawl, waste
disposal, ozone layer depletion, water pollution, climate change and many more affect every human,
animal and nation on this planet.
Performative Architecture and Wooden Structures … 939
robotic laboratory at the ETH in 2005, with the direction of Gramazio and Kolher,
marked a new course for the digital fabrication in architecture by introducing inno-
vative design paradigms that soon became cornerstones of the main researches issues
(Gramazio et al. 2014). The generic industrial machine becomes a design tool able to
convert virtual models in material systems through a single computational workflow
that introduce the innovative concept of digital materiality (Gramazio and Kohler
2014). The potentials of this design methodology, extension of the file-to-factory
(Sheil 2012) concepts, are investigated with respect to several lines of research that
have in common the development of processes related to the definition of informed
architectures through a digital path that includes innovative computation methods,
material computation and digital fabrication. The digital-material relationship allows
to combine the research of new formal codes with performative aspects that ensure
the manufacturability of what is generated in the digital space and apply it to the
scale of architecture (Gramazio and Kohler 2008). The integration of the concepts
listed above opens the post-industrial era based on the customization of architectures
in relation to performance exploration and optimization. Moreover, the hybrid space
of interaction between designer and machine is contaminated with other disciplines
in order to investigate different digital fabrication strategies but also to stimulate
creativity through a fruitful collaborative process. The interdisciplinary defines a
post-industrial phase in which issues related to analogue design and manufacturing
methods are interrelated with disruptive digital technologies.
2 Research Lines in the Field of Performative Wood
One of the main aspects investigated in the European research context on Perfor-
mative Architecture is connected to the use of innovative technologies in the field
of wood structures for the production of technological systems as well as architec-
tural organisms at 1:1 scale. Wood is one of the architectural longest-lived materials
because of its physical and mechanical properties, flexibility in structural applica-
tion which allow to use it indifferently to traction, compression and flexion stresses.
Compared to traditional wooden applications those related to digital computation,
generative design and robotic fabrication, open new horizon of research with dif-
ferent purposes, sharing the same investigative tools (Fig. 1). An overview on the
theme is outlined by the texts Advancing Wood Architecture (Menges et al. 2016)
and Advanced Timber Structures (Weynand 2016) that identify the main topics and
the research centers involved to define the possible future research directions. The
investigations conducted can be summarized in two distinct approaches: the first
concerns the use of engineered wood’s elements to attempt to exploit its physical
and mechanical properties; the second approach provides the use of wood as natural
and low-engineered material for structural applications that benefit from the complex
relationship between computational design and digital fabrication.
Fig. 1 Informed architectures and wood structures. Overview of the main research topics
2.1 Wood Computation: Exploit Material Behavior as Design

Agent
Regarding the first approach, one of the focus of research consists of the exploration
of mechanical properties of the material to transform them into design opportunities
in the formal generation process. The computational method becomes the mean by
which to discover advantages offered by wood in relation to its complex and non-
linear behavior. The physical and mechanical properties of this material, especially
if used in the form of thin panels handled by a rolling process, such as heterogeneity,
anisotropy, hygroscopic and irregularity, can expand the range of design possibilities
and stimulate the creative process. The studies on wood computation has been for
some years part of the research conducted by Achim Menges at the University of
Stuttgart, Institute for Computational Design, ICD, within the Performative Wood
research line; by Michel Hensel at the Research Center for Architecture and Tectonics
at the Oslo School of Architecture and Design. The investigations on the topic concern
two performance aspects of the design that can benefit from the material qualities:
the structural behavior and geometric-formal responsiveness (Persiani et al. 2015).
Within this research field can be identified different lines of research, which can be
summarized as follows:
– Activation of the anisotropic properties of wood, to set different structural behav-
iors in relation to differentiation of stiffness in the direction of the fibers;
– Use of elastic properties of wood as a factor design, through bending processes

(La Magna and Schleicher 2016) to design not standard structures;
– Exploitation of irregularity, through selective processes of digital fabrication to
eliminate portions of inactive and inefficient material;
– Use of hygroscopic properties of wood to actuate morphological transformation
processes.
2.2 Wood Computation: Smart Assembly, Natural

and Low-Engineered Materials
A further aspect of research on the application of innovative technologies in wood

constructions concerns the design and the prototyping of new spatial and structural
configurations. This approach benefits from the digital design and robotic manufac-
turing with special attention to the smart assembly. The Institute of Technology in
Architecture, ITA, of the Federal Polytechnic of Zurich, ETH, is one of the most active
research centers on the topic of complex and optimized wood structures explored
through the construction of a series of prototypes made in collaboration with con-
struction industries. The recent research developments on the topic, focus on the
possibility to extend the sensory capacity of the robot through a series of sensors
and a feedback loop system between the machine and the material system. Talking
about feedback loops means building relationships and designing a dynamic and non-
linear systems capable of guaranteeing adaptability to the local scale. The machine
becomes extension of the craftsman who perceived and analyzed physical properties
of a material through his own hands. The sensing ability of robots can be considered
the main innovation of the discipline. This methodological approach constitutes a
design response to theory of soft system or adaptable systems and in continuous
evolution whose dynamism is constantly fed by a flow of information coming from
outside. The ability to act and react dynamically makes the system responsive to
variations due to the interaction between the parties involved. These aspects are
investigated by the Institute for Computational Design in Stuttgart in a research path
defined as Fabrication Agency. Another field of investigation conducted mainly at the
Architectural Association, AA School of Architecture in London and IAAC, Institute
for Advanced Architecture of Catalunya, concerns the use of low-engineered and nat-
ural material present in great abundance in nature to create complex and optimized
wood structures (Menges et al. 2016). The project Woodchip Barn from AA is a man-
ifesto of this specific approach. The role of digital computing and the importance of
data-driven strategy is evident: the scanning of the shrub and the consequent creation
of a digital catalogues based on dimensional and morphological parameters represent
the true innovation of the design process. Through the three-dimensional scanning
and reconstruction of the elements it is possible to realize a complex surface starting
from natural curvature and dimensional parameters of the trunks, avoiding expen-
sive industrial bending processes. The computational process allows to integrate in
the project the data related to the natural material used, the structural performances
and the parameters of the construction site. The relational models obtained by the
generative process define the concept of informed morphology.
3 Performative Wood: Case Studies
To analyze the different approaches the contribution proposes a series of case study
and the results of two applied research, the 1 to 1 scale pavilions Fusta Ròbotica and
Digital Urban Orchard. The case studies are selected verifying the correspondence
to the following parameters:
– the presence of a performance-based process through which explore informed
architectures;
– the use of low-engineered and natural wood and the engineered one;
– the materialization of the digital model through innovative manufacturing pro-
cesses, specifically robotic fabrication.
All the projects are the results of a performance-based process through which
the generative process is informed by performative parameters (e.g. structural, envi-
ronmental) (Fig. 2). This methodology enables the exploration of innovative for-
mal codes using different typology of wood as construction material. The presence
of innovative digital fabrication tools such as robots implies that the Universities
involved in the survey are the one equipped with robotic fabrication laboratories and
who work to realize 1 to 1 scale projects (Fig. 3).
Fig. 2 Informed architecture: computational workflow between shape generation and manufactur-
ing processes
Fig. 3 Informed architecture: universities active in the research paths on informed architectures
3.1 Wood Chip Barn: Exploring the Potential of Natural

Material
The case study Wood Chip Barn2 (Fig. 4) represents the real transposition of a
complex functional program through the use of a natural material, large wooden tree
trunks, and an informed computational process. The project comes from the selection
of 250 tree trunks grown in the Hooke Park forest that surrounds the homonymous
Campus.3 The tectonic and structural experimentation is informed by the 3D scan
of the bifurcated wood trunks to extract the geometric characteristics and organize
the elements in the three-dimensional space. From the scanning of the elements the
following data are extracted: volumetry, central axes of the trunks with respect to the
bifurcations present as well as points of intersection between the axes themselves.
The extracted information is used to create a digital catalogue of solutions for the
spatial organization of the trunks in the formation of the main beams. The spatial
organization of the trunks is optimized to achieve a structural system able to increase
the structural stiffness and minimize shear stresses. The optimization process relies
on the use of genetic algorithms through the definition of the fitness value, target,
and variable parameters, genome. The arch structure, composed of different beams
connected in the central points of the bifurcations, is anchored to the ground by foun-
dations in concrete while the beams are pre-assembled by dividing the arc into two
2 AA Wood Chip Barn. Students: Zachary Mollica; Swetha Vegesana; Sahil Shah; Vivian Yang;
Mohaimeen Islam. Tutors: Martin Self, Emmanuel Vercruysse, Jack Draper, Charley Brentnall,
Toby Burgess. With: Pradeep Devadass (Robotic Developer), Arup (Engineering).
3 Hooke Park is a research campus in Dorset for experimental wood architectures of the Architectural
Association of London, AA.

Fig. 4 AA, Woodchip Barn, Architectural Association: pavilion robotically fabricated and made
by natural and low-engineered wood. © AA, Architectural Association, Valerie Bennett
halves to be assembled in situ (Mollica and Self 2016). The robotic manufacturing
process that involved the use of a Kuka KR 150 robot with a milling spindle made
it possible to create customized male-female joints between the different elements
that make up the main beams with six hours of processing in the robot cell. Thanks
to the use of the subtractive milling process, it was possible to realize male-female
interlocking joints as well as joints for the installation of the secondary structure on
which the wooden covering was installed. The proposed case study demonstrates how
with new production technologies, together with a computational process informed
by real-time data, is possible to realize optimized architectonic bodies that use non-
engineered materials. The prototype, is the result of a data-driven computational
process able to use the material and structural performances as guiding parameters
of the design, the acquisition of real-time data as feedback loop between the digital
environment and the real one, and the genetic optimization to determine the best
spatial configuration of structural elements. Through the 3D scanning process, the
natural material, bifurcated wood logs, is digitized to acquire the dimensional and
geometric parameters necessary to inform the digital model (Fig. 6). The operative
methodology has allowed to realize an optimized structure able to respond to perfor-
mance parameters derived from the material, from the project site as well as from the
structural conformation. In conclusion, the methodology applied to the Wood Chip
Barn project has allowed to obtain an informed geometry whose materialization has
been possible thanks to a robotic manufacturing process (Fig. 5) through which to
realize the customized connections (Fig. 7) in relation to the optimized spatial con-
figuration. Nonengineered material and a non-industrial production setting open up
new tectonic speculations and formal codes for sustainable architectures.
Fig. 5 AA, Woodchip Barn: complex joints made by robotic milling process. © Genny Spyridonos
Fig. 6 AA, Woodchip Barn: complex joints made by milling process informed by data-driven
design. © Swetha Vegesana
3.2 ICD/ITKE Research Pavilion 2011: Discover

Opportunities Offered by Wood Complex Behavior
The research pavilion 20114 is produced by the Institute for Computational Design
and Construction, ICD, Stuttgart (Fig. 8) in collaboration with the Institute of Build-
ing Structures and Structural Design, ITKE, and other industry partners. The aim of
4 Institute
for Computational Design—Prof. Achim Menges, Institute of Building Structures and
Structural Design—Prof. Jan Knippers. Concept and System Development: Oliver David Krieg,
Boyan Mihaylov.
Detail Design and Fabrication and Construction: Peter Brachat, Benjamin Busch, Solmaz
Fahimian, Christin Gegenheimer, Nicola Haberbosch, Elias Kästle, Oliver David Krieg, Yong Sung
Kwon, Boyan Mihaylov, Hongmei Zhai. Scientific Development: Markus Gabler (project manage-
ment), Riccardo La Magna (structural design), Steffen Reichert (detail design), Tobias Schwinn
(project management), Frédéric Waimer (structural design).
Fig. 7 AA, Woodchip Barn: performances, materials, manufacturing processes. A design method-
ology
the investigation is to propose a wooden structural system at 1 to 1 scale as result

of a computational and manufacturing processes informed by a series of performa-
tive parameters extracted from the analysis of a specific biological organism (Jodido
2013). Specifically, the research involved the study of the structure of the skeletons
of the Echinoid, sea urchin, in relation to their composition and the rules of aggre-
gation. The biological system is analysed to extract the physical and geometrical
principles to be transferred to the structural-material system through the compu-
Fig. 8 ICD, research pavilion 2011. © ICD/ITKE
tational process. The analysis carried out on the organic organism has allowed to
extract the fundamental geometric principles including the modularity of the system
that allows a high degree of adaptability and a performative structural system (Knip-
pers et al. 2012). This organizational principle is transferred to an optimized surface,
obtained through a form-finding process integrated with structural analysis, through
the tessellation of the same with hexagonal geometric components. The hexagonal
components are assembled with two different strategies: the plates that make up the
single three-dimensional technological unit are connected by structural glue while
the assembly of the different units takes place thanks to the realization of 100,000
geometrically different joints and subsequent nailing process. The structure turns
out to be so hierarchical and heterogeneous, since the size of the hexagonal cells is
informed with respect to the curvature of the surface, anisotropic, as each hexago-
nal unit is oriented with respect to the mechanical-structural stress determined by
the distribution of loads on the surface. The construction optimized in relation to
structural performance parameters, was achieved through a robotic manufacturing
process and a mechanical processing of the subtractive, milling type, through which
it was possible to guarantee the customization of the various connections, defined as
finger joint. The proposed methodology has allowed the creation of an architectural
body of considerable size using only plywood wood in panels with a thickness of
6.5 mm, creating 850 geometrically different components and 100,000 finger joint
joints between the different elements. The research pavilion is the result of the trans-
Fig. 9 ICD, research pavilion 2011: tailored finger joints and assembled hexagonal cells. ©
ICD/ITKE
fer of biological principles in the project through digital computation and a robotic
manufacturing process (Fig. 9). The proposed methodology is based on the informa-
tion of the process with respect to data obtained from the analysis of the biological
organism under investigation: scientific data, such as modularity and adaptability
of the natural system, are transformed into geometric rules used to discretize an
optimized surface through a form-finding process (Menges 2013). The result of this
process is a heterogeneous responsive morphology to realize an architectural struc-
Fig. 10 ICD, research pavilion 2011: performances, materials, manufacturing processes. A design
methodology
ture in which the performance characteristics respond to variable external inputs. The
possibility of realizing this structure through the assembly of different units made of
extremely light wood panels, confirms the excellent structural performance that the
architectural body can guarantee even if it requires a ground anchor to withstand the
horizontal wind load. The major problems derive from the complex system of joints,
the result of the heterogeneity and anisotropy of the material system, which makes
the management of the production process complex compared to the generation of
the manufacturing code and the speed of the production process (Fig. 10). One of
the possible implementations may derive from the simplification of the tessellation
geometry and the connection system as well as the use of the entire volume of the
hexagonal units.
3.3 ETH, the Sequential Roof: Robotic Manufacturing

and Smart Assembly
The proposed case study5 presents a responsive technological system (Fig. 11) for
structural applications in 1:1 scale to be applied to free-form surfaces realized by the
Gramazio and Kohler Research Group as part of the research on Additive Robotic
Fabrication of Complex Building Structures (Gramazio and Kohler 2014). The 168
reticular beams, composed of 23 layers assembled for a total thickness of 1,145 m
with three types of sections, 50 mm × 115.140, 180 mm, and two support points,
were used to create a coverage area of 2,308 m2 whose morphology was determined
thanks to a process of structural, energetic and constructional optimization. The
structural analysis allowed to optimize the morphology compared to the multiple load
conditions while the density of the structural elements was determined after a series
of tests on the permeability of natural light. The studies and analyses were conducted
through digital models informed through evaluations on scaled physical models used
to perform load analysis and, through an artificial sky dome, natural lighting. This
methodology has allowed to inform the morphology with respect to the data obtained
both in digital and physical environments and the resulting customization requires
an innovative and efficient constructive approach (Krammer 2016). The construction
process definition required a series of tests with a multi-functional end-effector able
to position the structural elements in wood and automatically apply the polyurethane
glue to then evaluate the tolerance of joints after operations of fixing. Thanks to a 1:2
scale model it was possible to determine the criticality and potentiality of the process
in relation to structural stability before scaling the whole process on an industrial
level. The proposed technological system explores the application of an informed
design strategy with respect to structural and environmental data for the realization of
a non-standard reticular structure. The checks carried out following the construction
of the prototype have allowed us to test the actual veracity of the simulation and of
the structural and environmental optimization conducted in the digital and physical
environment, defining the process ready for possible applications to the architectural
5 Collaborators: Aleksandra Anna Apolinarska (project lead construction), Michael Knauss (project
lead building project), Jaime de Miguel (project lead preliminary project), Selen Ercan, Olga Linar-
dou.
Selected experts: Dr. Lüchinger + Meyer Bauingenieure AG, SJB. Kempter Fitze AG, Prof. Dr.
Josef Schwartz (Chair of Structural Design, ETH Zurich), Prof. Dr. Andrea Frangi (Institute of
Structural Engineering, ETH Zurich), Estia SA (EPF Lausanne), ROB Technologies AG. Selected
contractors: Arch-Tec-Lab AG, ERNE AG Holzbau.
Fig. 11 ETH Zurich, Sequential roof, Arch_Tec_Lab of the Institute of Technology in Architecture
(ITA). © Angelo Figliola
scale in the structural field, as well as for multi-layer architectural wrappers in wood.
The proposed methodology (Fig. 12) and the resulting responsive structural model
will be applied for the realization of the coverage of the Arch_Tec_Lab of the Institute
of Technology in Architecture (ITA), ETH Zurich; the management of the robotic
manufacturing process for the realization of the roof was managed by ERNE AG
Holzbau through Gantry Robot and an industrial endowment of this type:
– 7 axes (2 parallel);
– dimensions: 48 m length, 5.60 m width and 1.40 m height;
– 3 zones, each zone is 14 m long;
– maximum speed of 120 m/min (7 km/h);
– maximum load capacity of the gripper 250 kg;
– tools: circular saw, milling, closing, lifting;
– machine for wood, metal and other materials;
– ability to determine the assembly sequence independently.
Fig. 12 ETH, Sequential roof: performances, materials, manufacturing processes. A design

methodology
3.4 Robotic Softness: Behavioural Fabrication Process

of a Woven Space
Robotic Softness (Fig. 13) develops a system of relationships, feedback loop, which
regulates the formal generation with respect to performance parameters identified
thanks to a customized and highly specialized production process (Brugnaro et al.
2016). The process starts from the study of the natural phenomenon of nesting,
typical intertwined structures used as nests from birds, to analyze and understand the
logic that regulate the formation of such complex systems. Compared to the other
proposed cases, Robotic Softness, introduces a new concept that refers to soft systems
Fig. 13 ICD, robotic softness. © Giulio Brugnaro
characterized by a non-linear relationship based on continuous feedback both at the

global and local scale. Just as in the volatile world the formation of the interweaving
that constitutes the nest is guided by environmental and structural performances, the
digital and production process that led to the prototype is informed on the global
scale by evaluation parameters such as light permeability, density and thickness of
the filaments, while at the local scale are the geometric parameters to inform the
process with kinect sensor able to return in real time the data coming from the
physical environment. Thanks to the interaction between digital and physical world,
the industrial robot can be considered as an agent able to mediate with respect
to the global scale and geometric parameters at the local scale in a process that
defines only the boundary conditions and the rules general without defining the final
Fig. 14 ICD, Robotic softness: performances, materials, manufacturing processes. A design

methodology
morphology. To develop the prototype the generic industrial machine is equipped

with a highly specialized end-effector able to mediate the action with respect to
the conditions of the local system in formation and the material used. The robotic
system Agent Based System (ABS) allows to adapt, set global parameters of boundary
conditions and general rules of the system, the virtual model to the materialization
of the same in relation to data obtained in real time. The case study introduces
an operational methodology in which the subject is in continuous transformation
and its responsive behavior is regulated by real-time feedback with the physical
environment. The environmental and structural performances constitute the essential
parameters of the data-driven strategy in such a way as to dispose the material,
and consequently increase the density of the nesting, where strictly necessary. This
methodology allows to reinterpret an ancient method of processing materials, such
as weaving, and an ancient material such as wood, through the fusion between digital
model and physical environment just as happens in the formation of natural systems
(Fig. 14). The proposed methodology and the geometry obtained presents a high
degree of responsiveness with respect to external inputs, determined by the designer
during the definition of the meta project, with the optimization of the model in
relation to the desired performances. The realization of a customized end-effector
and the presence of a kinect sensor introduces the concept of cyber-physical making
in which the machine is not merely a performer of repetitive actions but interacts with
the physical environment in a state of feedback loop between digital environment and
physical. A responsive system of this type may be suitable for the realization of multi-
layer casing elements capable of modulating natural light and solar radiation while
guaranteeing energy savings and optimized use of the material. Among the critical
issues we can mention the scalability of the system that needs further testing and
experimentation as well as the difficulty of managing the generative computational
process and the robotic manufacturing process.
4 Applied Research: Fusta Robotica and Digital Urban

Orchard
The prototypes, made in collaboration with industrial partners, represent the results
of transdisciplinary experiments in which environmental, structural and material per-
formances inform the computational process and the robotics manufacturing.6 Fusta
Robotica, is the outcome of a tectonic exploration deriving from non-engineered
material, Digital Urban Orchard is the formal expression of a complex functional
program arising from the relationship amongst form (shape), function and context.
The practice of a new design paradigm based on the information of the process
that sees environmental, structural and material performances as a factor driving the
entire design process. The performance criteria inform the computational process,
subsequently materialized using an anthropomorphic robot that is able to transpose
informed digital models into physical reality, through nonindustrial settings and using
irregular and low-engineered elements (Figliola 2016).
6 The protorypes, Fusta Ròbotica and Digital Urban Orchard, was developed during the Open Thesis
Fabrication 2015/2016 program at Institute for Advanced Architecture of Catalunya, IAAC. OTF
team 2015–2016: Areti Markopoulou, Alexandre Dubor, Silvia Brandi, Djordje Stanojevic, Maria
Kuptsova//Andrea Quartara, Angelo Figliola, Monish Siripurapu, Ji Won Jun, Josep Alcover Llubia,
Yanna Haddad, Mohamad Mahdi Najafi, Fathimah Sujna Shakir, Nada Shalaby.
4.1 Fusta Robòtica: Material—Informed Design
The Fusta Robòtica7 pavilion (Fig. 15) is the first low environmental impact wooden
structure built using robotic manufacturing in Spain. It was born from a collaboration
between the IAAC, Institute of Advanced Architecture of Catalunya and Serradora
Boix Srl as a tectonic testing to be exposed at the “Semana de la fusta 2015”,8 with
the intent to show the potential derived from the application of digital computation
and robotics manufacturing in the construction of wooden structures. The objectives
of the research were represented by the promotion and enhancement of the Catalan
wood, Mediterranean pine timber, by an exchange of knowledge between industries
and research centres for the innovation of the production of the model in order to
test new formal codes for a sustainable design. The material used, highly deformable
and non-engineered to be used in construction, consists of small simple and irregular
wooden profiles and is employed for the production of industrial or biomass pallets.
The square section profiles the industry has made available for the project have the
following dimensions: 38 mm × 38 mm × 2,000 mm. Another aspect to be integrated
in the design process is the method of production and the production area with their
tools. The pavilion was built with anon-industrial setting, represented by a Kuka
KR-150 industrial robot equipped with pneumatic gripper, a device for the storage
Fig. 15 IAAC, Fusta Ròbotica, Setmana de la fusta 2015, Barcelona. © Ji Won Jun
7 FustaRobòtica Pavilion is a project of IAAC, developed with the generous sponsor of Serradora
Boix; in collaboration with Gremi de Fusters, Tallfusta, Incafust, Mecakim, Decustik.
8 Setmana de la Fusta is an annual event aimed to promote the use of wood Catalan that takes place
in Barcelona.
Fig. 16 IAAC, Fusta Ròbotica: set-up of the robot cell and organization of the working space. ©
Angelo Figliola
of the wooden profiles and a circular saw arranged on a rotary table (Fig. 16), within
the university digital fabrication laboratory. The entire design process was informed
by the mechanical properties of the material that were extracted through a series
of analog tests necessary to the understanding of the material and structural sys-
tem behaviour. Among these, we can mention the excessive bending of the wooden
profiles, due to the variation of the curvature following the drying process, and the
lack of structural rigidity of the profiles due to the mechanical characteristics of the
material. The manufacturing method adopted also contributes to inform the design
process, with the minimum and maximum workable dimensions of profiles and com-
ponents that can be aggregated depending on the work area, the characteristics of
the robot and their handling. In relation to this, the design process was informed
using the material performances as a design input. It is possible to avoid structural
problems due to excessive bending of the wooden profiles using a redundant, hyper
static structure. Nailed joints allowed to maximize the resistant section of the com-
ponents, in correspondence of the structural nodes. The discretization of the shape in
eight sections with a constant thickness optimized the working space and weight of
the robot cell, avoiding collision problems and facilitating assembly. The pavilion,
formed by about 1,000 square section wooden profiles, is the result of the elaboration
of a complex geometry, hyperboloid, in which the rotation of geometric continuous
elements has allowed to obtain a dynamic spatial configuration, a manifesto of the
potentials resulting from the use of material in the production of complex structures
(Fig. 17). At the same time with the analog test development on the material system,
the algorithm was developed to transpose the 3D solids of the digital model in sim-
ple geometric elements, such as lines and planes, necessary for the definition of the
various processing stages. Thanks to the direct connection between the parametric
model and tool manufacturing, the various stages of the production process have
been defined. They can be summarized as follows:
Fig. 17 IAAC, Fusta Ròbotica: computational design process. © Open Thesis Fabrication
– Taking of the wooden profiles from the storage device;

– Cutting of profiles to the corresponding size of the digital model;
– Profile deposit on the assembly platform.
Each stage of production included the manufacture of a half-arch to facilitate the
operations of manual assembly and the transport to the installation site. At the end
of the production process 940 pieces of wood in 8 arches, divided into 16 parts were
processed and assembled in 35 h of production. The eight sections that make up the
roof were assembled at the university laboratory and aggregated on the site of the
installation.
4.2 Digital Urban Orchard: Form Follows Data Flow
The Digital Urban Orchard research project9 (Fig. 18) involves the construction
of a functional prototype to be implemented in urban public spaces within the self-
sufficiency programme of the city of Barcelona, which stems from the relation among
form, function and application context for a new concept of space of socialization
and food production. As second part of the project, the pavilion hosts a hydroponic
cultivation system and an adaptive silicone skin able to ensure the indoor comfort
conditions that are essential for the plants growth. The need to design a stable yet
lightweight structure and to ensure maximum solar gain for a proper growth of crops,
at the same time, required multiple responsiveness able to get the proper compliance
with the performance required by each of the single parameters listed above. To
integrate the functional, structural and environmental-energy performance criteria,
and inform the design process, the data-driven strategy was necessary to correctly set
the genetic optimization by defining the genotype, the geometrical characteristics of
the shape and the phenotype or quantitative parameters by which the genotype can be
modified. The flexibility of the parametric model allowed to structure the meta project
through the clarification of invariable parameters and genotype variable geometric
Fig. 18 IAAC, Digital Urban Orchard 2016. © Andrea Quartara
9 Digital
Urban Orchard is a project of IAAC, developed with the generous sponsor of of Merefsa,
supplying the silicone and with Windmill, in particular thanks to Josep Ramon Sole and Álvaro
Romera for the structural consultancy.
Fig. 19 IAAC, Digital Urban Orchard: digital catalogue of informed architectures. © Angelo Figli-
ola
data, which may vary within a range aptly defined by the designer in relation to the
values of the phenotype or rather quantitative parameters of performance analysis.
The final shape has been selected from a catalogue of design solutions (Fig. 19), the
result of genetic optimization and creative process which included the integration of
different parameters:
– solar radiation on the surface of the orchard;
– solar radiation on the inclined surfaces where the plants are placed;
– wind pressure on the outer surface;
– minimum and maximum size of wood profiles that can be made with respect to
the setting used;
– mechanical and physical properties of materials.
The process of genetic optimization was handled varying the geometric curves,
two base ones on the x, y plane and a higher one, from which a surface is generated by
the creation of a Loft and the inclination of planes that host the hydroponic system.
The analysis of solar radiation on an annual basis, and the subsequent optimization,
have made it possible to determine the overall shape and inclination of wooden shafts
that host the hydroponic system. In parallel to the process described and thanks to
a form searching process, the CFD analysis allowed minimizing the wind pressure
on the outer surface of the pavilion in order to ensure the structural balance. The
adopted structural principles are the same as those used in the Fusta Robòtica pavil-
ion: the hyper static structural pattern, generated by the alternation of diagonals and
Fig. 20 IAAC, Digital Urban Orchard: set-up of the robot cell and organization of the working
space. © Open Thesis Fabrication
elements able to ensure structural rigidity, is a complex system that performs the
structural function. It is designed as a support plan for the hydroponic system, as
support for the silicone skin and as space-functional furniture. The density of the
structural pattern responds to optimization logics for solar access into interior spaces
and considers almost total transparency at the top of the pavilion. The final shape has
been discretized through 6 types of sections, for 12 components total. Three manu-
facturing strategies have been defined depending on the size of the sections and the
work platform. They involve the construction of the entire section or the assembly of
two/three parts of the final section with a total of 30 assembled parts. To maximize
the resistant section, we used 2,524 nails in nailed joints with a collaborative pro-
cess between manufacturing robotic and manual finishing. The structural analysis,
conducted in cooperation with the engineering firm Windmill-project partner on a
typical section under various load conditions, has allowed validating the structural
choice made despite showing a high displacement due to the horizontal pressure of
the wind in extreme conditions as set forth by the legislation. In the Fusta Robòtica
pavilion the production process was implemented at all stages in order to reduce
material consumption and expand the range of achievable geometry. Implementa-
tions concerned the customization of end effector, pneumatic gripper and tools used
for the production such as the circular saw and the device for the storage of wooden
profiles (Fig. 20). The customization of the circular saw has allowed to create spatial
cuttings (Fig. 21) in three dimensions. Thanks to a new wood provider, the length
of the profiles was diversified in order to reduce waste material. Realized with 1,681
profiles, the pavilion is the result of 52 h of robotic and 24 h of manual assembly
coming from the information of the process and the optimization of the performance
completed in a production process that can control the complexity and transform it
Fig. 21 IAAC, Digital Urban Orchard: complex assembly thanks to the spatial cutting process. ©
Andrea Quartara
into design opportunities while ensuring rapid execution, automation and only 2% of
material waste. Finally, during construction, the silicone wrap production benefited
from the collaboration with the silicone industry Merefsa S.R.L. that provided the
necessary material and laser cutting tools for the proper production of components.
5 Applicability of Systems to Architectural Practice
The analysis of the case studies and applied researches, leads to a critical reflection
on the possibility of introducing the technological systems proposed in the architec-
tural practice, explaining the level of applicability and identifying the main critical
aspects of the processes. The analysis represents a fundamental step for delineating
innovative pathway for future researches on the topic. In fact, if it is possible to state
that technology is the answer to the problems of design processes, citing the words
of Price in 1966,10 it is essential to define clearly the question, to ensure that the
contribution is not nullified within a generalist thought that sees the use of emerging
technologies as a solution to all critical design issues. One of the questions underlying
the research concerns the concrete possibility of making a technological transfer from
the academy to the actors involved in the design and construction process. Among the
technological systems studied, which of them can be introduced in the architectural-
executive practice? What are the critical issues that prevent a concrete application
outside the experimental field of academic research? The first consideration to make,
10 Technology is the answer but what was the question? Cedric Price, 1966.
for a correct reading of the application context, concerns the continuous interference
and contamination between the technical elements that define the building systems.
In this regard, some of the systems analysed present technological solutions that go
beyond the definition of the common classes of technological units and technical ele-
ments that characterize the architectural practice. It is necessary to carry out a work
of synthesis to clearly outline their field of application and, consequently, to be able
to analyse their level of applicability with reference to the common technological
classes. The reference categories have been defined as follows:
– Technological unit, as an element that identifies itself with a grouping of functions
necessary for obtaining environmental performances;
– Technological system, as a structured set of technological units;
– Architectural organism, as a unitary and structured set of technological systems.
The following application fields correspond to the reference macro-categories:
– Bearing structures/structural systems;
– Vertical and horizontal enclosures;
– Internal partitions;
– External partitions.
For what concerns the use of wood as primary material, potential applications
concern structural technological systems able to exploit the potential offered by the
use of complex computational processes and digital manufacturing techniques.
5.1 Wood Structural Systems
The survey conducted on the implementation of informed structural systems has

demonstrated the possibility of concrete applications in architectural practice. The
experiments that present a greater level of applicability are those that involve wood
as a construction material in relation to:
– a computational processes of structural optimization;
– experimentation on complex spatial configurations derived from the implementa-
tion of innovative digital manufacturing methods.
The research paths on informed wooden structures benefit from a consolidated
know-how for what concerns the properties of the material and the related construc-
tion technologies by all the actors involved in the design process. The collaboration
between academia and construction industries lead to introduce innovative technolo-
gies like digital computation and robotic manufacturing to fully explore new formal
codes related to environmental sustainability able to positively act on performative
criteria such as the carbon footprint and the embedded energy of the processes. The
experiments examined employed engineered material such as laminated and profiled
panels and low-engineered materials while digital exploration concerned computa-
tional processes of structural and multi-objective optimization as well as the study of
principles of biomimesis to be abstracted and transferred in the design of complex and

organic structures. The digital manufacturing processes involved in the construction
of wooden structural systems are different and can be summarized as follows:
– Subtractive processes, for the realization of differentiated joints derived from com-
plex, optimized spatial configurations;
– Additive processes,11 for the assembly of components by means of complex and
informed spatial sequences;
– Combined, subtractive and additive processes, in which the construction sequence
involves machining the components before their assembly.
6 Performative Wood and Structural Systems: Towards

an Application in Architecture
Some of the wood structural systems analysed experimented the integration between
different technological systems such as main structure, secondary structure, insula-
tion, finishing and technological systems. The capability to integrate complex tech-
nological systems through a digital workflow demonstrates the applicability of the
methodology and validates the processes in relation to the management of tolerances
derived from the use of innovative tools in the production. Projects such as Sequential
Roof , ETH Zurich, roof structure of the Arch_Tec_Lab of the Institute of Technology
in Architecture (ITA), as well as the entire production of the research path on additive
processes for the construction of complex wooden structures of the Swiss univer-
sity, Woodchip Burn of the AA, Architectural Association of London, and the 2011
Research Pavilion of the ICD, basis for the construction of the Landesgartenschau
Exhibition Hall of 2014, have demonstrated the efficiency of the computational pro-
cess in the exploration of different design solutions and the concrete applicability of
technological systems in relation to complex systems that characterize an architec-
tural organism. The described experiences can be considered as proof of concept for
what concerns:
– The scalability of processes and the integration of main and secondary technolog-
ical systems;
– The potentiality of the morphogenetic process and digital computation in the trans-
fer of biological principles in the virtual space, and subsequently physical;
– The advantages derived from genetic and heuristic optimization processes in the
design of performance structural systems;
– The potentiality of the customization of structural joints for the construction of
optimized structures;
– The management of tolerances resulting from the digital manufacturing processes
used.
11 The assembly of the components can be assimilated to other additive processes.

A further check was carried out through the construction of two research pavil-
ions, Digital Urban Orchard and Fusta Robòtica, made at the IaaC in Barcelona. The
design of the two pavilions has highlighted a further aspect regarding the applicability
of structural wooden systems in architectural practice: the control of formal genera-
tion through the parametric process and the possibility of customization offered by
the robotic fabrication, allows to overcome, or better expand the concept of structure
to that of an integrated system that manages to aggregate primary and secondary
structure, systems and furnishings through specific morphological configurations.
The research on the topic presents a high level of applicability also for the direct
involvement of the industries in the sector stimulated by the sharing of know-how
and the possibilities offered by the combination of a traditional material such as
wood and digital processes of performance simulation and digital manufacturing.
Compared to the projects mentioned above, other design experiences have a more
experimental vocation, in some cases mere speculations, which corresponds to a low
level of applicability in architectural practice, mainly in relation to the scalability
of processes. Nevertheless, starting from the analysis of the projects, it is possible
to extract theoretical concepts and future applications that can be transferred to the
design processes: one of these is the investigation of the soft systems, which con-
trasts with the hard systems discussed above, based on the adaptive relationship of
feedback loop between performance parameters and formal generation. The studies
that combined responsive morphologies with the robots’ introduction directly on
the construction site showed clearly the potentiality of their possible application to
realization of structural technological systems, or vertical, performance closures at
the architecture scale.
6.1 Performative Wood and Structural Systems: Bring

Complexity in the Design Process
As highlighted in the analysis of the case studies, the critical points of the method-
ology proposed is represented by the complexity of the computational process, the
intrigued material system and the managing of the manufacturing process. In rela-
tion to the computational process, the first problem concerns the designers and their
background on the topic. To manage such a complex process, the designer need
to possess integrated and advanced skills (Figliola 2018) that allow managing the
entire design process from design to construction. In fact, the data-driven strategy
envisages the integration of the formal generation and digital manufacturing process
within the same workflow and this implies a greater management complexity com-
pared to the common design processes. A further critical aspect is the complexity
deriving from multi-objective optimization paths and the consequent discretization
of the parametric model necessary to limit the calculation times and to guarantee the
readability and usability of the obtained data that allow to choose an optimal solution
in a digital catalogue. The optimization of morphology based on the information of
the computational workflow brings to complex manufacturing processes that need

a customization of tools and tasks and involves the design of complex communi-
cation protocols between digital model and machine as well as the engineering of
custom-made end-effector. The management of the process defined as cyber physical
making requires the possession of specialized skills on the border between mecha-
tronics, electronic engineering and computer science not easily available among the
sector’s operators. The paradigm shift in progress leads to a further reflection on the
materials used in the production processes in relation to the physical characteristics
and suitable mechanical properties. The analysis of the state of the art related to the
issue shows that one of the major problems concerns the technological-constructive
process related to the implementation of the components: the management of toler-
ances, the result of non-engineered manufacturing processes that involves customized
tools necessities the design of communication protocols based on the feedback loop
relationship through which to adapt the production process to the digital model and
vice versa. Tolerance becomes an important design theme on a par with the scalability
of the processes and the study of the structural connections between the technolog-
ical units and the integration with the other constructive elements that constitute an
architectural organism. The definitive introduction of the design methodology and
the proposed technological systems in the architectural practice passes for the res-
olution of technical problems and for the rethinking of the design and construction
process in its complexity. The first step concerns the centrality of the production
process that becomes an integral part of the design process helping to define the lim-
its of the formal generation process and ensuring the feasibility of the components.
To apply this methodology, it is necessary to have a structured knowledge regards
the technical specifications of the instruments, the organization of the working area
(defined as robot cell) and the characteristics of the tools used to perform a specific
and unique task. The second aspect concerns the development of a new material
sensitivity aimed at overcoming the consolidated relationship between material and
production process that characterizes industrial processes. The material acquires a
high specific weight in the preliminary phase of the design and its correct management
depends on the success of the operation. The digital process is informed about the
physical and mechanical properties of the materials and the constraints derived from
the manufacturing method used, which consequently expands the range of design
possibilities and introduces a new performative layer. The conjugation of the two
terms apparently opposed as digital, representation of the virtual environment, and
material, as something tangible and concrete, finds complete expression thanks to the
computational-algorithmic approach. The customization of the production process
and the renewed materiality imply a redefinition of the constructive logics as a com-
bination of the performance of the material, expression of its technical characteristics
and of its mechanical properties, and of innovative manufacturing methods.
7 Conclusion and Outlook
Through case studies and experimental outputs was verified the application of com-
putational workflow for performative architecture (Figliola 2016). The integration
of single and multi-objective optimization processes and innovative production tech-
nologies allow to use the material according to the performative parameters used to
guide the design process. In this regard, the additive processes represent the most pro-
ductive method that is suitable to optimize the material resources: the possibility to
compose generic elements, considering all degrees of freedom of movement that the
instrument provides, promotes the exploration of complex aggregative systems for
high-performance building components, managed by algorithms. A further imple-
mentation is represented by the reintroduction of vernacular materials in designing,
such as wood, and by low-engineered approaches that facilitate the development of
new formal codes for the sustainable design that can extend the concept of perfor-
mance to the materials used. Going from materials to the organization of the produc-
tion process, the results showed the potential of digital fabrication in the development
of the design complexity, using non-industrial setting in university laboratories. The
efficiency of production processes is guaranteed by the power of the instrument used
that allows to reduce time and production costs as well as optimize the use of human
resources in the entire production chain. The use of robotic manufacturing supports
the development of a technological thought based on the prefabrication of building
systems to be realized dry and completely reversible through the use of eco-friendly
and especially recyclable materials. The greatest potential, for what concerns the pro-
duction process, is represented by the realization of units and technological systems
whose minimum and maximum size is linked to the characteristics of the instrument
used, to be aggregated on the project site through dry processes, minimizing the noto-
riously high expenditure of resources for this stage of the architectural design. This
aspect could be further implemented with the introduction of operating instruments
directly on the project site which, in fact, would even terminate the consumption of
resources associated with the transport of components. That would encourage the
testing of innovative aggregation systems that enhance the performative architecture
parameters and, at the same time, stimulate design creativity to overcome the open
industrial prefabrication featuring contemporary design processes. The flexibility of
parametric models allows to examine and manage the complexity resulting from the
aggregation of components. Aspects related to material and digital manufacturing
technologies, together with the notion of informed architecture, are the main inno-
vations of the proposed methodology as they reintroduce aspects of the project that
the first digital era had identified as consequential and non-integrated processes. The
informed architecture leads to the integration of the parameters concerning geometry,
material and manufacturing from the early stage phase of the project giving two direct
consequences: the reduction of the space of design possibilities relating to the power
of digital computation and the introduction of the manufacturability limit that binds
the digital model to the physical space. Further developments concern the quanti-
tative analysis of the benefits arising from the adoption of the project methodology
in the context of environmental certification protocols in the entire design process,

with a focus on the performance of materials and the construction phase, as well as
the quantification of the impact on grey or hidden energy of processes.
Acknowledgements Fusta Robòtica Pavilion is a project of IAAC, developed with the generous
sponsor of Serradora Boix; in collaboration with Gremi de Fusters, Tallfusta, Incafust, Mecakim,
Decustik.
Digital Urban Orchard is a project of IAAC, developed with the generous sponsor of Merefsa,
supplying the silicone and with Windmill, in particular thanks to Josep Ramon Sole and Álvaro
Romera for the structural consultancy.
OTF team 2015–2016: Areti Markopoulou, Alexandre Dubor, Silvia Brandi, Djordje Stanojevic,
Maria Kuptsova//Andrea Quartara, Angelo Figliola, Monish Siripurapu, Ji Won Jun, Josep Alcover
Llubia, Yanna Haddad, Mohamad Mahdi Najafi, Fathimah Sujna Shakir, Nada Shalaby.
References
Brugnaro G, Baharlou E, Vasey L, Menges A (2016) Robotic softness. In: Velikov K, Sean Ahlquist
S, Matias Del Campo M (eds) Proceedings of the 36th annual conference of the association for
computer aided design in architecture, Michigan, Ann Arbor
Deutsch R (2015) Data-driven design and construction: 25 strategies for capturing, analyzing and
applying building data. Wiley, Hoboken
De Landa M (2015) The new materiality. In: Menges A (ed) Material synthesis: fusing the physical
and the digital. Wiley, London, pp 16–21
Figliola A (2016) Post-industrial robotics: esplorazione di architetture informate nell’era post-
digitale. Techne. https://doi.org/10.13128/techne-xxxxx
Figliola A (2018) The role of didactics in the post-digital age. Agathon. https://doi.org/10.19229/
2464-9309/342018
Frazer J (2016) Parametric computation. In: Block P, Shajay B (eds) Parametricism 2.0. Wiley,
London, pp 18–23
Gramazio F, Kohler M (2008) Digital materiality in architecture. Lars Muller Publisher
Gramazio F, Kohler M (2014) Made by robots: challenging architecture at a larger scale. Wiley,
London
Gramazio F, Kohler M, Willmann J (2014) The robotic touch: how robots change architecture. Park
Books, London
Grobman Y, Neuman E (2012) Performalism: form and performance in digital architecture. Rout-
ledge, London, New York
Hensel M, Menges A, Weinstock M (2010) Emergent technologies and design. Routledge, New
York
Hensel M (2013) Performance-oriented architecture: rethinking architectural design and the built
environment. Wiley, Chichester
Jodido P (2013) ICD/ITKE research pavilion 2011. In: Jodido P (ed) Wood architecture now!
Taschen GmbH, Köln, pp 194–199
Kolarevic B (2003) Architecture in the digital age—design and manufacturing. Spon Press, New
York
Kolarevic B, Malkawi A (2005) Performative architecture: beyond instrumentality. Spon Press,
New York
Kolarevic B (2015) From mass customisation to design ‘democratisation. In: AD Architectural
design J(238), pp 48–54
Knippers J, Menges A, Gabler M, La Magna R, Waimer F, Reichert S, Schwinn T (2012) From nature
to fabrication: biomimetic design principles for the production of complex spatial structures. In:
Hesselgren L, Sharma S, Wallner J, Baldassini N, Bompas P, Raynaud J (eds) Advances in
architectural geometry 2012. Springer Wien, New York, pp 107–122
Krammer M (2016) Individual serialism through the use of robotics in the production of large-
scale building components. In: Reinhardt D, Saunders P, Burry J (eds) Robotic fabrication in
architecture, art and design 2016. Springer International Publishing, pp 461–467
La Magna R, Schleicher S (2016) Bending-active plates: form and structures. In: Adriaenssens S,
Gramazio F, Kohler M, Menges M, Pauly M (eds) Advances in architectural geometry 2016. Vdf
Hochschulverlag AG an der ETH Zürich, pp 170–188
Menges A (2012) Material resourcefulness: Activating material information in architectural design.
In: AD Architectural design J(216), pp 34–44
Menges A (2013) Morphospaces of robotic fabrication. In: Brell-Çokcan S, Braumann J (eds)
Rob/Arch 2012: robotic fabrication in architecture, art and design. Springer, New York, pp 29–47
Menges A, Schwinn T, Krieg OD (2016) Advancing wood architecture: a computational approach.
Routledge, New York
Mollica Z, Self M (2016) Tree fork truss: geometric strategies for exploiting inherent material form.
In: Adriaenssens S, Gramazio F, Kohler M, Menges M, Pauly M (eds) Advances in architectural
geometry 2016. Vdf Hochschulverlag AG an der ETH Zürich, pp 138–154
Oxman R (2009) Performative design: a performance-based model of digital architectural design.
Environ Plann B Plann Des J36(6):1026–1037
Persiani S, Battisti A, Wolf T (2015) Recurring moving patterns in nature for a biomimetical
optimization of autoreactive systems. In: Proceedings of the international association for shell
and spatial structures (IASS) symposium 2015, Amsterdam
Schumacher P (2015) Parametricism 2.0. AD architectural design, J(240)
Sheil B (2012) Manufacturing the Bespoke. AD Reader. Wiley, United Kingdom
Weynand Y (2016) Advanced timber structures. Birkhauser, Basilea
Yuan P (2015) Parametric regionalism. In: AD architectural design J(240):92–100
Adaptive Timber Towers.
An Evolutionary Prototype for the 21st
Century Skyscraper
Alessandro Buffi and Gian Maria Angelini
Abstract The biological metaphor introduced by computational design, as well

as the increasing theoretical framework of parametric architecture, opens new per-
spectives for both wood and skyscraper design. At the intersection of architecture,
biology, and computer science, avant-garde designers are reshaping the historical
relationship between nature and architecture fostering a natural approach to design
which results in: structural lightness, rational use of energy and elegance. Wood
seems to be a perfect material to engage with this new period of design research,
and while timber towers are getting higher, there is a strong interest in understanding
to which extent wood can represent a valuable alternative to those materials that
have characterized the recent architectural debate. Analyzing the emerging type of
the wooden skyscraper within the context of globalization, technological advances,
and ecology; the authors present their vision for the 21th-century skyscraper. Based
on parametric design and evolutionary principles, the proposed model can adapt to
different contexts and conditions, providing different solutions as the result of the
interaction with the surrounding environment.
Keywords Parametric design · Form-finding · Environmental design

High-rise buildings · Complex timber structures · Evolutionary algorithms
Wooden skyscraper
1 Introduction
The biological metaphor introduced by computational design, as well as the increas-

ing theoretical framework of Parametric Architecture (Schumacher 2011), open new
perspectives for both wood and skyscraper design. At the intersection of Architecture,
A. Buffi (B) · G. M. Angelini

Department of Civil and Environmental Engineering, University of Perugia, Perugia, Italy
e-mail: info@alessandrobuffi.com
G. M. Angelini
e-mail: gianmaria.angelini92@gmail.com

972 A. Buffi and G. M. Angelini
Fig. 1 James Watson and Francis Crick posing next to their DNA double helix model (1953)
Biology, and Computer Science, avant-garde designers are reshaping the historical
relationship between Nature and Architecture fostering a natural approach to design
which results in: structural lightness, rational use of energy and elegance. The back-
ground of such a natural approach to design comes from the nineteenth century’s
form-finders such as Robert Maillart, Pier Luigi Nervi, Sergio Musmeci, Eduardo
Torroja, Riccardo Morandi, Felix Candela, Heinz Isle, Eladio Dieste, Buckminster
Fuller, and Frei Otto (Pone 2011). We owe them the birth of the so-called hands-off
approach to design, an approach in which decisions are not taken directly by the
architect, yet following a cognitive process supported by both analytic and digital
models (Bianconi et al. 2017).
Generative design systems give new impetus to this field of research allowing the
designer to emulate natural processes or even invent new ones to create architectural
organism characterized by the same efficiency and beauty of natural systems. In
particular, the efficiency of living systems is the result of a slow evolutionary process,
as explained by Darwin’s theory of evolution and the possibility to apply this process
to Architecture through Genetic Algorithms puts in the hands of the designer a
powerful tool (Bianconi et al. 2017). Starting from the 1990s shift has been noticed
in the way avant-garde architects have used new technologies of evolutionary biology,
to address or depict the increased complexity that is noticed in today’s architecture
(Fasoulaki 2007). This shift has been fostered on the one hand by the advances in the
field of genetic (Fig. 1) and computer science, with the development of bio-inspired
algorithms, on the other by the experimentation of CAD technology (Fig. 2) in
different design fields, with the development of innovative representation techniques.
With the windward of digital representation, the algorithm becomes a tool to
configure, rather than a statical morphology, a generative model characterized by
diversity and thus by adaptive behavior. Through the use of digital design tools, the
designer can effectively handle heterogeneous information and complex associative
rules; this opens to a series of optimization and form-finding strategies used to explore
Adaptive Timber Towers. An Evolutionary Prototype … 973
Fig. 2 Ivan Sutherland

working at the computer
TX-2, MIT Lincoln Labs
(1963). The Sketchpad
Interface
the generative potential of the model and find optimal solutions. Furthermore, while
generative systems allow to visualize and evaluate thousands of design options, a
new assembly line allows creating thousand of variations of the same product, thus
developing a direct link between what can be designed, and what can be built.
Within the emerging context of Industry 4.0 (Paoletti 2018), while robotics and
smart manufacturing are reshaping the relationship between architecture and indus-
trial production, the design, and construction process of the skyscraper has to be
rethought. This building type historically based on mass production and economy
of scale, is gradually moving to a new paradigm of industrial production based on
custom and nonstandard elements characterized by a high-level of complexity. Even
though the building industry has been slow in adopting this new paradigm (Ben-
ros and Duarte 2009), by applying techniques borrowed by automotive, naval and
aerospace industry (Portoghesi 2007) the architect can go beyond the limitations
imposed by mass-production moving to a design-driven manufacture that provides
the designer with a high degree of freedom.
Timber structures become a central theme within this innovative trend and, in the
light of the latest developments, wood has the potential to become the most advanced
building material of 21st century. Thanks to its workability and its strength-to-weight
ratio, it becomes an excellent material for structural expression and formal exper-
imentation. Furthermore, its carbon-storing properties associated with the benefits
on human health and comfort, make it more than an alternative to traditional con-
struction materials. For this reason, different projects around the world have been
built, triggered by the advances of engineered wood, and while timber structures are
getting higher, there is a strong interest in understanding to which extend wood can
represent a valuable alternative to those materials that have characterized the recent
architectural debate.
2 The Skyscraper as an Adaptive Organism
Adaptive capacity or adaptation has become a key ambition of the contemporary avant-garde
trend that might suggest a comparison with natural organic systems. (Schumacher 2007)
In the design of a skyscraper, “productive machine” of enormous complexity

(Willis 2004), the subjectivity of the individual contribution is subordinated to the
objectivity of the result. This involves every aspect of the construction, affecting its
aesthetic appearance “no longer product of an a priori creative idea, but result of a
long process of elaboration and selection aimed to identify the best solution” (Biraghi
2008). In this sense, within this typology, form-finding strategies reach their maxi-
mum expression, supporting the designer in a complex exercise of problem-solving.
In this process, “potential performance improvement can be explored using para-
metric multi-objective optimization aided by sophisticated evaluation tools, such as
computational fluid dynamics and energy analysis software, to visualize and explore
skyscraper generative potential (Imam and Kolarevic 2016).”
Within the context of globalization, the skyscraper is populating our planet as
a ubiquitous universal building type, engaging with widely different contexts and
climate conditions (Fig. 3). The only ways to make this process sustainable is through
an effective adaptation to the site-specific climate conditions; namely by the use of
traditional and vernacular architecture elements or by a conscious bio-climatic design
approach based on both active and passive means (Yeang 1999). Skyscraper energy
efficiency can be promoted by considering a climatically responsive design, where
the orientation, the thermal properties of the building envelope and the effect of
altitude become the main design tools (Saroglou et al. 2017). Strategies including
building configuration and orientation, the location of the service core, the design of
the building envelope to incorporate sun shading, integrated plantings, and the use
of natural ventilation (Howeler 2005).
Ecological skyscrapers belong to an emerging area of design research char-
acterized by innovative interdisciplinary explorations achieved through the total
Fig. 3 The Global context: a variegated realm characterized by site-specific environmental forces.
World map of Koppen-Geiger climate map (Murray C. Peel)
Fig. 4 Hōryū-ji in Japan, is

one of the oldest wooden
buildings in the world
integration of architecture, and engineering (Jahn and Sobek 1999) by blurring the
boundaries of disciplines like biology and computer science. In particular, ecology,
as a branch of biology, suggests the study of the relationship between organisms
and their environment, a concept that suits surprisingly well also the discipline of
architecture (Hensel and Menges 2006). Like a natural organism, the skyscraper is
composed of a series of deeply integrated systems, each of which, interacting with
its environment, contributes to defining the performances and the outcome of the
building. Instead of assembling rigid and hermetic geometric figures, “each of this
element should be parametrically malleable and participate in a dynamical play of
mutual responsiveness as well as contextual adaptation (Schumacher 2010).”
By looking at the skyscraper as an adaptive organism, bio-inspired form-finding
strategies can guide toward high performances, supporting the designer in a complex
problem-solving process that takes into account heterogeneous data such as geom-
etry, dynamic system of forces, and environmental conditions. Just like in Nature,
where every living organism, through a slow evolutionary process, is subject to the
logic of adaptation; in parametric design, tools like Genetic Algorithms (GA) can
identify, through a fitness-based process, the optimal solution from a bioclimatic,

structural and functional point of view (Causarano 2014). Since, in the late 60s,
Ingo Rechenberg and his colleagues applied for the first time Genetic Algorithms to
aerodynamic wing design (Rechenberg 1973) several designers mainly in the field
of engineering started to use these fitness-based algorithms to solve complex design
problems. Used as virtual prototyping tools in the field of mechanical and aerospace
engineering genetic algorithms belong to the metaheuristic class of Evolutionary
Algorithms and “are commonly used to generate high-quality solutions to optimiza-
tion and search problems by relying on bio-inspired operators such as mutation,
crossover, and selection (Mitchell 1996).”
3 The Era of Timber Towers
For more than a century, the skyscraper constructed with concrete and steel has had
a leading role within the metropolis skyline. Those materials, given their exceptional
structural properties, have represented the constructive solution for the designer of
the modern era, and thank to their seismic resilience and high load-bearing capacity
skyscrapers have achieved extraordinary heights. Unfortunately, steel and concrete
industries appear to be among the highest carbon footprint industries in the world. It
is assessed that altogether they are responsible for the 8% of global CO2 emission.
Therefore, it is worth asking whether there are alternative materials to steel and con-
crete which can guarantee the same static behavior while reducing its environmental
impact.
The background for the global diffusion of timber towers can be found in ancient
architecture; indeed wood has been one of the oldest construction material in the his-
tory of architecture, and Tall Wood buildings have existed for centuries. For instance,
tall pagodas in Japan were built up to 19 stories in wood 1400 years ago and a few
still stand today in high seismic and wet climate environments (Green 2012).
The Barsana Monastery (Fig. 5), standing at 56 m tall, is considered the tallest
wooden structure in Europe. Whereas, Asia, within his historical tradition, offers
a wide series of examples. The typical pagoda structure besides being an excellent
example of the durability of wooden constructions, is also a type of structure that
develops upwards with aesthetical purpose. The five-story pagoda of Hōryū-ji (Fig. 4)
in Japan, standing at 32.45 m in height, is one of the oldest wooden building of the
word. Originally built in the 7th century, the center pillar of the pagoda is estimated
to have been felled in 594 (“Web Japan” 2007). Most recently, after the burns of the
19-story height Jiulong Temple in Mianzhu in December 2017, the Tianning Pagoda
in Changzhou (Fig. 6) is the tallest pagoda in the world, with 13 stories and a height
of 153.79 m (China Daily 2007 ).
In recent years, once studies and research proved the efficiency of Cross Lami-
nated Timber (CLT) (Ceccotti et al. 2006), its usage for multi-story buildings sprout
across the globe with extraordinary results, from Norway to Japan. In 2009, the first
residential building constructed entirely in timber was completed in London. The
Fig. 5 The Barsana

Monastery is considered the
tallest wooden structure in
Europe
Stadthaus, 24 Murray Grove is a nine-story high-rise designed by Waugh Thistleton

Architects which could be considered as a pioneering project for its unique cross-
laminated structural system. Made with prefabricated panels including cut-outs for
windows and doors, this building was assembled within nine weeks, revealing the
benefits regarding time-saving related to wooden structures (Thompson et al. 2009).
In the meantime, following the first results, several CLT structural system have been
developed around the world. The Brock Commons Tallwood House, completed in
2017, at the University of British Columbia in Vancouver stands 53 m high and is the
tallest timber building constructed to date. Designed by the Canadian practice Acton
Ostry Architects Inc., the 18-story skyscraper is composed by a concrete podium and
lift cores, while glulam timber columns hold up cross-laminated timber floors and
walls (Naturally Wood 2017).
In Europe, Norway is assuming a leading role, indeed it hosts Treet the tallest
European wooden residential building, while the Mjøsa Tower that will be in 2019 the
tallest wooden building in the world is under construction. Treet has been completed
in 2015, this 14-story residential building located in the city of Bergen, is composed
by load-carrying glulam trusses and two intermediate strengthened levels. A system
of prefabricated building modules are stacked on top of the concrete basement and on
Fig. 6 Tianning Pagoda in

Changzhou, China is
currently the tallest pagoda
in the world
top of the strengthened levels, while glass and metal sheeting on the facades protect
the structural timber from weathering (Abrahamsen and Malo 2014).
Looking to the next future of timber technology, several exemplary projects can
inspire the design of next-generation timber towers. Arup developed a project called
Haut to create a 21-story tower in Amsterdam in 2020. Meanwhile, C.F. Moller
and Anders Berensson Architects plans for respectively a 34-story and a 40-story
skyscraper, both in Stockholm; and PLP Architecture and Cambridge University’s
proposed for an 80-story, 300-m high wooden building integrated within London’s
Barbican Centre. Furthermore, giving the results already obtained in different projects
and the advancement of timber technology, this positive trend doesn’t look to stop.
To date, the most ambitious timber tower project is represented by the W350
Sumitomo, in Tokyo. Called to be the tallest timber structure in the world, with its
350 m and 70 story it is estimated to cost 600 billion of yen and to be completed
in 2041. The planned structure is a hybrid wood and steel structure made from 90%
of wooden materials. In this visionary project, the interior structure of a pure wood
will produce a calm space that exudes the warmth and gentleness of wood, creating
like a forest, a habitat for living things (Sumitomo Forestry 2018). With the aim of
“Change Cities in Forests,” these ambitious project reflects the clear direction that the
Japanese building industry is going to undertake in the near future. Indeed, Japan in
order to promote forest regeneration, since 2010 is encouraging the shift to wooden
structures for public buildings through the “Act for Promotion of Use of Wood in
Public Buildings”.
Meant to be reversible structures that can be almost fully recycled at the end of
their life cycle, wooden skyscrapers become a vehicle to enable the sustainability of
our planet. With the global population growth and increasing rates of urbanization,
the demand for new high-rise buildings is constantly rising. If these new buildings
are made of wood, they will not only act as a long-term carbon reservoir, but they
will generate fewer emissions during their entire life cycle, representing meanwhile
a healthy place to live or a stimulating workspace environment. The benefits related
to wooden environments are multiple and strictly connected to the psychological and
physiological well being. For instance, the feeling of natural warmth and comfort
related to wood have the effect of lowering blood pressure and, heart rates reducing
stress in favor of productivity and positive social interaction. Furthermore, wood
performs as a moisture regulator, and sensible improves air quality within an interior
space (Bergs 2002). In this sense wood is a perfect material to design an architectural
organism whose adapt to both the environment and its inhabitants.
4 Evoluzione: Vision and Design Approach
Evoluzione (Fig. 7) is a design proposal presented by the authors at the SKYHIVE

Skyscraper Challenge 2018 and awarded the BB Green Award. The concept of the
project is to “propose, rather than a form, an architectural model characterized by
variety and thus by adaptive behavior; an iconic prototype for a sustainable human
evolution. The context of the project is planet earth, a variegated realm character-
ized by a range of site-specific environmental forces capable of informing the design
process. Following the rules of nature, a form-finding strategy based on parametric
design and evolutionary principles minimizes the use of matter and energy consump-
tion through the specific climate data interpretation. In this context, natural lighting,
strictly related to the office workers well-being, becomes one of the major driving
forces. The result is a skyscraper capable of meeting the requirements of flexibility,
comfort, and sustainability related to a working environment in continuous evolu-
tion.”
Fig. 7 Evoluzione: adaptive wooden skyscraper in Santiago de Chile, Miami, Dubai and Hong
Kong
4.1 Materials and Methods
As a method that can be used by designers in the early stage of design, the form-
finding process developed in this research emerges from a bottom-up process that
combines analysis, and experimentation in generative systems. Trying to put the
geometry in relation with his environment the researchers create, through the defini-
tion of an algorithm, an integrated design process that encompasses three steps:
• Generative Model: definition of a generative model able to create a variety of
feasible architectural solutions
• Fitness Functions: definition of the performance criteria and use of different anal-
ysis tools (fitness functions) to evaluate in an iterative process the performance of
each solution
• Optimization: use of an evolutionary multi-objective optimization based on
Genetic Algorithms to find within the search space an optimal solution
The optimal form that emerges from such a design approach results from an inter-
action with a heterogeneous set of climate based data that affect both its energy and
structural behavior. In this project, this integrated system was created entirely with
the Rhinoceros’s plug-in Grasshopper introducing in each phase of the project differ-
ent add-ons for analysis and optimization. The Grasshopper’s add-ons Honeybee &
Ladybug (Roudsari and Pak 2013) were used to inform the process with climate data
and advanced energy analysis, Karamba (Preisinger 2006) was used to evaluate the
weight of the wooden structure while Octopus (Bader and Zitzler 2008) to perform
evolutionary multi-objective optimization.
Fig. 8 The adaptive model is composed by three deeply integrated systems: the core, the cluster,
and the envelope
4.2 Generative Model: Adaptive Wooden Skyscraper
In the context of computational design, a generative design system is a group of

geometrical rules that enables to go from a set of parameters (genome) to geometry
(phenotype). This is the phase of the design process in which the designer has more
control over the outcome, and it’s only “through a well-designed system of rules,
that generative design systems have the capability of maintaining stylistic coherence
and design identity while generating different designs. (Granadeiro et al. 2013)” In
this sense, the proposed free-form organism is composed of three deeply integrated
systems: the core, the cluster, and the envelope (Fig. 8). They adapt their dimensions
and functions to the context to maximize energy efficiency and comfort.
• Timber Structure: The free formed structure is made of Cross Laminated Tim-
ber and Glulam. The use of this renewable material sensibly reduces the carbon
footprint of the skyscraper.
• Responsive Core: An artificial intelligence is integrated into the concrete core
structure. In addition to its structural and distributive function, it acts as a ther-
moregulator by controlling temperature, humidity, and air quality.
• Cluster System: The flexible cluster system is capable of adapting to the require-
ments of smart working by offering a wide variety of spatial solutions.
4.3 Fitness-Based Optimization
The right balance between the three described systems is obtained through an opti-
mization process aiming to maximize energy efficiency and comfort while minimiz-
ing structural weight. Indeed, while some characteristics of the architectural organism
are defined directly by the designer through the generative model, other qualities are
supposed to emerge through a well-defined optimization. Inspired by the evolution-
ary process, the algorithm used in this research is based on the Hype algorithm of
Octopus, a Grasshopper’s add-on for multi-objective optimization.
In this project, as in many other architectural projects, it was useful to consider
more criteria simultaneously: the weight of the structure, the energy performances,
and the visual comfort. However, these are only some of the performances that may
interest the designer, and in theory, each parameter that affects the Architect’s choices
during the design process can be added to the concept of fitness. It remains an open
list, and along with the advancement of performance simulation, new criteria and
analysis tools can be used to inform this process.
4.3.1 Lightness Through Structural Optimization
The structural system proposed in this project is inspired by the Australia Square
Tower designed by Harry Seidler in collaboration with the Italian Pier Luigi Nervi.
By moving from a cylindrical-shaped to a free-form structure, this research proposes
a reinterpretation of this structural model characterized by a central core, exterior
perimeter columns, and sculptural ribbed floors. The decision to liberate the form
from his original circular and highly efficient design (Desideri et al. 1979) is related
to the desire of obtaining an architectural organism that can be shaped by light.
The optimization process aims to find within the possible geometrical configura-
tion of this structural model the one that implies less use of material. The optimization
method encompasses the following steps: (i) for each formal solution, a structural
system is defined; (ii) an optimization process finds the optimal cross section for the
elements of the structure; (iii) the cost of the structure is calculated starting from its
weight; (iv) the cost of the structure is used as fitness in a genetic optimization.
Firstly, an associative logic defines the position of each structural element starting
from the free-form shape of the skyscraper (Fig. 9). With the parametric structural
engineering tool Karamba, these elements are used, to define a Finite Element model
able to evaluate the structural efficiency of each design solution. Then, after the
definition of loads, supports, joints, and material properties; the component Opti-
mize Cross Section is used to calculate the optimal cross-sections of the structural
elements.
Fig. 9 Structural fitness: CLT slab dimensioning through FEM analysis Karamba parametric
engineering
The construction cost is approximated by evaluating the mass of the obtained

structure and is used as a performance criteria. By iterate this process for differ-
ent design solutions the generative model shows great variability in performances,
highlighting different structural behaviors. These results are used along with energy
performances to inform the design process and are used as fitness in a multi-objective
optimization.
4.3.2 Environmental Optimization: Light, Energy and Comfort
With the aim to adapt the skyscraper to its environment the Grasshopper’s add-ons
Honeybee & Ladybug were used to inform the design process with climate data and
advanced energy and daylight analysis. The main goal in this phase was to optimize
natural lighting and visual comfort. Whereas, another environmental aspect was the
optimization of passive solar energy regulating the contribution of solar gain and
shading, depending on the specific climate conditions.
Light strictly related to the office workers well-being becomes one of the major
driving forces of this fitness-based optimization. The goal is to reinforce circadian
rhythms and reduce the use of electric lighting by introducing natural daylight into
space. Appling the LEED (USGBC 2013) guidelines for lighting, this design phase
Fig. 10 Fitness variability: daylight analysis for spatial daylight autonomy (sDA) evaluation
is driven by two Annual Daylight Performance Metrics: spatial daylight autonomy

(sDA) and Annual Sun Exposure (ASE) (IESNA 2012). The first one defines the
lighting autonomy of the space and is evaluated through an annual Daylight simu-
lation conducted with Honeybee and Radiance (Fig. 10); while the second one, is
related to glare potential and is calculate through a sunlight hours analysis based on
Ladybug. This two value are simultaneously optimized, and different design solu-
tions are tested to maximize sDA and minimize ASE.
The goal of an environmental optimization is to ensure a satisfactory comfort with
the minimum use of energy by adapting the architectural organism to his context
and his inhabitants. To this end in this research, an energy model is associated to
each architectural organism, individual of a genetic population. The fitness of each
solution is evaluated with Honeybee by calculating the overall energy consumption
as the sum of Heating, Cooling, Lighting. In this phase, through visual scripting,
the researchers can define in detail Constructions, Thermal Zones, HVAC systems,
schedules and climate data achieving a feasible energy model that can be used in an
automated process for environmental and multiobjective optimization.
The result obtained by both lighting and energy analysis are used along with
structural performance to inform the optimization and design process. Once again,
the variability in fitness highlights the adaptive capacity of the model and the impor-
tance of exploring skyscraper generative potential to obtain architectural organism
characterized by rational use of matter and energy. In this sense, genetic algorithms
support the designer in finding an optimal solution and compare different perfor-
mance (Fig. 11).
Fig. 11 View from the

interior space: double height
atrium
5 Conclusion
The innovative techniques of representation introduced by parametric design open

countless scenario for further research and experimentation. Wood seems to be a
perfect material to engage with this new period of design research, and while con-
cerns with the environmental impact of the building industry push the exploration of
alternative structural solutions, a new era for timber technology begins. The built and
unbuilt projects proposed in this paper highlight the global explosion of timber tow-
ers: a type that has the potential to become central to contemporary architecture. In
the 21st century, the importance of skyscrapers will continuously increase, and even
if it has been traditionally seen as a building type outside the mainstream of archi-
tecture culture, it is destined to ignite the theoretical and artistic debate, becoming
the most highly discussed building type of its era.
By looking at the skyscraper as an adaptive organism, designers have the poten-
tial to guide the sustainable growth of our cities, developing site-specific solutions
to a design problem of international nature. In this sense, generative design and
bio-inspired optimization strategies can be used as a mean for bio-climatic design,
highlighting the generative potential and adaptive capacity of an architectural model.
This design approach, apply surprisingly well to skyscraper design, where it can
lead to considerable savings regarding energy and material. Moreover, along with
the advancement of performance simulation and building information modeling, this
method can be improved to obtain a fully integrated design process that help designer
to achieve both performance and aesthetic expression. To conclude, the interdisci-
plinary approach based on parametric design and evolutionary principles, proposed
in this research, demonstrate how designing following the rules of nature can result in
the definition of architectural organisms characterized, like natural ones, by lightness,
efficiency, and multifunctionality.
References
Abrahamsen RB, Malo KA (2014) Structural design and assembly of “Treet”—a 14-story timber
residential building in Norway
Bader J, Zitzler E (2008) Hype: an algorithm for fast hypervolume-based many-objective
optimization
Benros D, Duarte JP (2009) An integrated system for providing mass customized housing. Autom
Constr 18:310–320. https://doi.org/10.1016/j.autcon.2008.09.006
Bergs J (2002) Effect of healthy workplaces on well-being and productivity of office workers, B
en R Adviseurs voor Duurzaamheid, Amersfoort
Bianconi F, Filippucci M, Buffi A (2017) Drawing, model and nature: form-finding one hundred
years after on growth and form. In: Empler T (2017) 3D modeling and BIM, Sapienza Università
di Roma, Rome. ISBN: 978-88-49645019
Bianconi F, Filippucci M, Buffi A (2017) Optimization and evolution in architectural morphogen-
esis. In: D’uva D (ed) Analyzing form and morphogenesis in modern architectural contexts. IGI
Global, New York. ISBN: 978-1-5225-3993-3
Biraghi M (2008) Storia dell’architettura contemporanea 2, Piccola Biblioteca Einaudi, Turin
Causarano RM (2014) Sviluppi del pensiero sistemico nell’architettura contemporanea. Il principio
di organizzazione/autoorganizzazione nel progetto architettonico, Dottorato di ricerca XXV
ciclo, Università La Sapienza di Roma
Ceccotti A, Lauriola MP, Pinna M, Sandhaas C (2006) SOFIE project—cyclic tests on cross-
laminated wooden panels. In: Proceedings WCTE, Portland, USA
China Daily (2007) World’s tallest pagoda inaugurated, 30 Apr 2007. Chinadaily.com.cn. Retrieved
on 22 May 2008
Desideri P, Nervi PL, Positano G (1979) Pier Luigi Nervi. Zanichelli, Bologna
Fasoulaki E (2007) Genetic algorithms in architecture: a necessity or a trend? Paper presented at
10TH generative art international conference Milan, Italy
Granadeiro V, Duarte JP, Correia JR, Leal VMS (2013) Building envelope shape design in early
stages of the design process: integrating architectural design systems and energy simulation.
Autom Constr 32:196–209
Green MC (2012) The case for tall wood building, how mass timber offers a safe, economical
and environmentally friendly alternative for tall building structures, mgb ARCHITECTURE +
DESIGN
Howeler E (2005) Skyscraper design of the recent past and for the near future. Thames & Hudson
IESNA Lighting Measurements Committee, IES LM-83-12 (2012) Approved method: IES spatial
daylight autonomy (sDA) and annual sunlight exposure (ASE). Illuminating Engineering Society
of North America, New York
Imam M, Kolarevic B (2016) Towards resource generative skyscrapers. Council on Tall Buildings
and Urban Habitat research paper
Jahn H, Sobek W (1999) Archi-neering. Hatje Cantz Verlag, Leverkusen, pp 12, 92
Mitchell M (1996) An introduction to genetic algorithms. MIT Press, Boston, Cambridge
Naturally Wood (2017) Construction of a tall wood building Brock Commons Tallwood House:
construction overview. University of British Columbia
Paoletti I (2018) Mass customization in the era of industry 4.0: towards immaterial building
technology. In: Hemmerling M, Cocchiarella L (eds) Informed architecture. Springer, Cham.
https://doi.org/10.1007/978-3-319-53135-9_8
Pone S (2011) Le tecniche della leggerezza. In: Materia, 70
Portoghesi P (2007) The envelope and the visage. Materia 55
Preisinger C (2006) Karamba parametric structural modeling—user manual
Rechenberg I (1973) Evolutionsstrategie. Stuttgart, Holzmann-Froboog, Stuttgart
Roudsari MS, Pak M (2013) Ladybug: a parametric environmental plugin for grasshopper to help
designers create an environmentally-conscious design. In: Proceedings of the 13th international
IBPSA conference Held in Lyon, France
Saroglou T, Meir IA, Theodosiou T, Givoni B (2017) Towards energy efficient skyscrapers. Energy
Build 149:15
Schumacher P (2007) Arguing for elegance. In: Architectural design
Schumacher P (2010) The parametric city, London 2010. In: Zaha Hadid—recent projects, A.D.A.
Edita, Tokyo
Schumacher P (2011) The autopoiesis of architecture, a new framework for architecture. Wiley,
West Sessex
Sumitomo Forestry Report (2018) New development concept W350 plan for wooden high-rise
building, Published Feb 2018
Thompson H, Waugh A, Weiss K-H, Wells M (2009) A process revealed. Murray & Sorrell FUEL
Design & Publishing, London
USGBC, LEED v4 (2013) Reference guide for building design and construction (v4). US Green
Building Council, Washington DC
Web Japan, sponsored by the Ministry of Foreign Affairs, Japan (2007) One hundred years older
than supposed? World Heritage Pagoda
Willis C (2004) Empire state building, 21 mesi per costruire il crattacielo più alto del mondo,
Electra, Milan
Yeang K (1999) The green skyscraper, the basis for design sustainable intensive buildings. Prestel,
Munich
Knowledge-Based Design
in Industrialised House Building:
A Case-Study for Prefabricated Timber
Walls
Graham Day, Eugenia Gasparri and Mathew Aitchison
Abstract This chapter illustrates how the adoption of a knowledge-based engi-

neering approach may provide a powerful tool for the industrialised house building
sector to manage the complex and multidisciplinary nature of design, fabrication and
installation. The research focuses on timber technologies and prefabricated timber
components, which are frequently selected in preference to other industrialised build-
ing systems because of the advantages they offer in terms of weight, workability and
sustainability strategies. A knowledge-based engineering methodology is explored
for the design of prefabricated timber-framed external walls, encoding both “explicit”
and “tacit” knowledge into a digital three-dimensional model. Results demonstrate
how such an approach could significantly change common design practices by shift-
ing the major phase of design effort to earlier stages in the project cycle, thereby
minimising re-work, reducing data fragmentation and potentially removing the need
for drawings. A key finding of this paper is that model interoperability, maintenance
and reuse becomes unlikely if an agreed methodology, including a description logic,
is not adopted. Despite the need for a rigorous approach, the ability to capture, manage
and reuse design knowledge could be of significant benefit to emerging industrialised
house building ventures.
Keywords Prefabrication · Timber-framed walls · Digital-twin

Generative design · BIM · KBE
1 Introduction
The architecture, engineering and construction (AEC) industry is facing many sub-
stantial challenges, not least the call to become a more sustainable and more produc-
tive sector. Moreover, the AEC sector has historically evolved at a slow pace as it is
profoundly risk-averse (Aitchison et al. 2018).
G. Day (B) · E. Gasparri · M. Aitchison

The University of Sydney, Sydney, Australia
e-mail: graham.day@sydney.edu.au

990 G. Day et al.
The construction sector is responsible for about 40% of the global CO2 emissions,
as well as being characterised by the inefficient use of materials resulting in high
wastage. In recent decades, many industrialised nations have endorsed initiatives
and protocols to reduce global warming, resource depletion and promote sustainable
growth. The Kyoto Protocol commits its parties to binding emission reduction targets.
This, for example, led to the European Commission setting specific targets for the
construction industry, including reducing energy consumption and promoting the
sustainable use of material resources. Life-cycle assessment analytical tools are now
commonly being used to analyse the environmental impacts of projects.
Several recent studies have highlighted the remarkably low increases in productiv-
ity within the construction industry (Egan 2014; Woetzel et al. 2017). These reports
provide an insight into the current status-quo, including an understanding of the
potential causes of this stagnation. Numerous actions have been suggested to gain
the required improvements. A need for value-oriented thinking and a lean approach
across the entire value adding process have emerged as fundamental factors. The
authors suggested that integration and transparency should be favoured over frag-
mentation and opacity. A major conclusion that can be drawn from this discourse
is that the construction industry needs an extensive overhaul. Today, tools and tech-
nologies that enable transparent collaboration and process integration are now readily
available and accessible to the construction industry.
New sustainable materials and high performing construction technologies, such
as engineered timber products, are gaining currency across the globe. Many countries
have put in place new regulations to remove the barriers that prevent their use. An
example of this paradigm change can be found within the radical revision of the fire
code in Singapore, to allow the use of Cross-laminated timber in high-rise buildings
(Chuan 2013), which echoes the earlier changes to the building code in Sweden in
1995 (Nord 2008).
Advances in the field of information and communication technologies (ICT), the
“Internet of Things” (IoT) and Big Data management have already radically changed
the way we live and work today. They will also enable a deep transformation of the
industrial practices across all major sectors. There has not been widespread adoption
of these technologies within the construction industry and in many cases, it is not
clear how they can be safely and effectively exploited to realise their potential.
Developments within the timber industry suggest a resolve to transform design,
fabrication and construction processes. Embracing the use of new-generation compu-
tational design tools will enable the AEC industry to manage complex projects across
all processes by sharing and reusing knowledge. Moving away from a bespoke design
approach in favour of a more systematic and standardised method would significantly
increase efficiency and boost productivity. This would mark an industry-wide change
towards industrialisation and automation.
Knowledge-Based Design in Industrialised House Building … 991
2 Industrialised House Building
The history of industrialised construction is characterised by numerous failed

attempts. A common perception of prefabrication is that it is often associated with a
lack of flexibility, extreme standardisation and even ugliness (Smith and Timberlake
2010). This is mainly attributed to the use of prefabrication as a necessity or on a
contingency basis. Key examples of utilitarian industrialised housing through history
are colonial settlements, mining or oil rigs industrial applications, military camps,
emergency and disaster relief housing (Aitchison et al. 2018). These examples are
often located in remote areas, and by their very nature are related to the concept of
impermanence and mobility. As such, aesthetic or energy performance concerns are
often irrelevant or are only marginally addressed during the design phase.
A major obstacle to industrialisation and the social acceptance of prefabricated
buildings in Europe, can be found in post-World War II construction where, exten-
sive and uncontrolled residential developments were driven by a necessity to quickly
provide inexpensive shelter. This approach gave rise to an often defective, dull build-
ing stock characterised by poor energy inefficiency, early decay and lack of general
appeal.
More recently, there is growing evidence that prefabrication is a likely candidate to
address contemporary concerns around population growth, mobility, rapid urbanisa-
tion, housing shortages and the affordability crises. Contemporary designs focus on
customisation and product flexibility. For example, the Swedish government encour-
aged and supported collaborative projects between industries and research institutes
which has led to a systematic and strategic development of the national marketed-
oriented IHB sector (Stehn 2009).
An industrialised manufacturing approach to house building promises to address
important factors such as environmental related issues including resources use, waste
minimisation and carbon footprint reduction. Furthermore, the shift to a production
system approach would have beneficial impacts on productivity in construction,
favouring integration and fostering quality across the whole process (Woetzel et al.
2017). The Japanese industrialised house building industry provides a valuable exam-
ple where the influence of a well-established manufacturing industry contributed to
the rapid growth of one of the world’s most successful prefabricated housing indus-
tries, providing a quality-oriented product approach and highly automated factory
workflows (Gann 2010).
History has shown how industrialisation of the construction sector is a very com-
plex issue. Many factors must be considered and coordinated. Today, more than ever,
the building industry appears increasingly ready for this challenge. Recent advances
in computational design tools, digital and automated fabrication technologies, new
management and control approaches provide significant opportunities for the success
of the prefabrication market. Integration across the entire process from project incep-
992 G. Day et al.
tion to building occupation is crucial for the industrialised house building industry.
The primary issues preventing an industry transformation may be attributed to the
very nature of the architecture, engineering and construction (AEC) industry, for
example:
• Design definition during the construction process. Efficient industrialised man-
ufacturing relies upon a complete understanding of component parts, assembly
methods and manufacturing processes being undertaken along with the associated
logistics. Any assumptions or ‘unknowns’ will result in unwanted and unpre-
dictable outcomes, frequently causing considerable out-of-budget expense and
presenting execution risks to the overall process. This also applies to prefabri-
cated buildings, where constraints and procedures dictated by the manufacturing
and assembly of single components must be carefully considered during the early
design stages. Thus, design definition during the construction process is not a
viable option.
• Poor communication among stakeholders. The AEC industry is intrinsically char-
acterised by fragmentation. The number and variety of participants involved is
likely greater than in most of other industries. Clients, contractors, design teams,
consultants, project managers, suppliers, municipalities and occupants all con-
tribute to the design. Many diverse processes involving many stakeholders must be
managed and integrated. Every process carries some level of project risk, depend-
ing upon the level of developed detail. Within the industrialised house building
sector, processes such as architectural design, fabrication, transportation and con-
struction must be perfectly integrated to prevent re-work, increase productivity
and mitigate risk. A segmented flow of information or a lack of communication
between involved stakeholders may jeopardise the successful completion of a
project.
• Project-to-project industry approach. The reconfiguration of teams on a project
basis leads to very limited long-term organisational memory with little or no shared
experience. The problem is compounded further by a lack of tools to transfer
knowledge between the fragmented teams or learn lessons from the collective
experience (Holzer 2009). This situation results in knowledge being fragmented
and individualised within project participants who may never work together again.
It is evident that digital technologies for collaborative design and design for fab-
rication are necessary to fully benefit from a mass-production approach. These tech-
nologies exist and have the capability to ensure that prefabricated buildings will no
longer be the exception but might become the ‘rule’.
3 Timber and Industrialised House Building
Timber has been used in construction since ancient times. Throughout history, all
civilisations have developed their own way of transforming timber products to build
shelters, villages and later, cities. This is not only due to material availability across
the globe, but also to the ease of production and workability that timber provides.
Today, timber products are increasingly chosen as more environmentally-friendly and
efficient options for large-size projects and tall towers, in preference to other con-
struction materials such as concrete or steel. The reasons behind timber construction
“renaissance” and “growth in scope” are many.
Recent environmental concerns have pushed many countries to promote and invest
in the sustainable development of cities. The first order benefit of timber is that it is a
carbon sink. It is the only construction material that sequesters CO2 from the atmo-
sphere, rather than increasing its emissions (Robertson et al. 2012). Several studies
have also shown how timber buildings contribute to improve indoor comfort, a reduc-
tion in occupants’ stress, including lower heart rates and where people generally feel
better (Sakuragawa et al. 2005). Furthermore, engineered timber products provide
excellent structural performance and enable the transformation of the timber indus-
try towards an efficient industrialised approach. It has already been demonstrated to
provide consistent increases in on-site safety and quality, as well as reducing con-
struction time (Patterson 2014). Timber-based building elements and systems are
highly suited to offsite production. They can be assembled to form prefabricated
large 2D Panelised components or, fully finished 3D modular elements which can
be installed on-site with no additional work. The lightweight nature of the material
reduces building weights and offers advantages for transportation and handling of
large elements within factories and on construction sites.
Recent advances in timber production techniques, processing technologies and
design tools have invigorated the timber construction sector. Timber manufacturing
and housing companies around the world have been stepping away from the tra-
ditional craft approach that was typical of the timber construction sector and have
undertaken a systematic modernisation of their factories by establishing digitised and
automated manufacturing capabilities. Lendlease Group, a large global construction
company, recently invested in DesignMake, a manufacturing business for prefab-
ricated timber buildings, to consolidate their leadership in the field of engineered
timber buildings across Australia.1 Other recently established companies operating
in the field of industrialised construction, such as Katerra, have selected mass timber
as their flagship structural material. Katerra intends to transform the architecture,
engineering and construction industry within North America in coming years, advo-
cating process integration and innovation in terms of technologies, smart design and
supply-chain management, among many other improvements (Woetzel et al. 2017).
To ensure its success, the industrialised building market has to meet contem-
porary challenges in terms of productivity increases across the whole process but
also sustainable development, reducing waste and resources depletion. Timber, by
its very nature, can be easily transformed and manufactured, and structures quickly
assembled to form large and lightweight prefabricated components. Timber systems
and components also enable flexible and customisable design both in space and
1 Asper the acknowledgement below, the authors are engaged in a Commonwealth Government
Cooperative Research Centre—Project with industry partners Lendlease and DesignMake, titled,
“Innovation in Advanced Multi-Storey Housing Manufacture”.
994 G. Day et al.
time. Finally, timber is an environmentally sustainable renewable resource and it can

be recycled or even reused. All the previous considerations underline how timber
technologies can play a pivotal role in the future of industrialised house building.
4 Knowledge Management, Timber and Industrialised

House Building
The construction industry is well known for relying upon experienced individuals or
teams to design and realise outcomes on a project by project basis. The knowledge
and experience gained is often undocumented but represents a valuable and strategic
business asset. Knowledge transfer takes considerable time and frequently relies upon
experts not being ‘turf-protecting’ or ‘knowledge-hoarding’. The ability to access
and reuse experience has been demonstrated to provide enormous productivity gains
within industrial processing and manufacturing (O’Dell and Grayson 1998).
The definition of knowledge and expertise depends heavily upon its context. A
relevant interpretation is provided in the framework of industrial capability transfer
between two factories on different continents (Bender and Fish 2000). A knowledge
hierarchy (Fig. 1) illustrates how data, information and knowledge is transformed
into individualised expertise. A consequence of this transformation is that knowledge
transfer becomes more difficult because people themselves ‘need to be transferred’
to realise or teach the benefits of their experience.
The term knowledge-based engineering (KBE) is attributed to the launch of a now
outdated software called Intelligent Computer Aided Design around 1984 (Sandberg
et al. 2008). Initial KBE applications focused on ‘rule capture and knowledge with
geometry being one of many kinds of outputs’ (Knutson n.d.). During the 1990s it
Fig. 1 Knowledge hierarchy (Bender and Fish 2000)

was recognised that new methods were needed to formalise knowledge acquisition
through delivery using such systems. Two primary methodologies emerged. Com-
monKADS (Kingston 1998) provided a model for knowledge-based planning and
was primarily developed for the artificial intelligence industry while the Methodol-
ogy for Knowledge-Based Engineering Applications, also known as MOKA (Stokes
2001) was developed by the European Aerospace and Automotive sectors.
The purpose of a knowledge-based engineering system is to encode domain expe-
rience, knowledge, information and data to enable retention and reuse by attempting
to eliminate the individualised nature of knowledge (Fig. 1). KBE systems differ
from expert systems in that they are independent of specific domains (LaRocca
2012). Numerous academics and individuals have tried to provide an unambiguous
definition of knowledge-based engineering, suffice to say that there are differences in
opinion. A review of knowledge-based engineering research challenges (Verhagen
et al. 2011), discusses many positions, including the use of ‘dedicated software lan-
guage tools’ and ‘the merging of object-oriented programming and computer aided
design’. It is suggested that more contemporary definitions ‘Focus on the automa-
tion of repetitive engineering tasks whilst capturing, retaining and reusing associated
knowledge’. The emphasis on ‘knowledge capture, retention and reuse’ differentiates
the technology from individual ‘standalone’ design solutions.
O’Dell and Grayson demonstrated that leveraging existing knowledge has signif-
icant benefits however, an environment or framework that enables and encourages
knowledge transfer is also vital for success. MOKA provides such a framework
including compelling arguments for the adoption of knowledge-based engineering,
telling us that routine non-creative design can be significantly compressed (Fig. 2)
and that design processes that took weeks were reduced to minutes and those that
took years of effort became less than a day. Clearly, the application of these methods
and techniques within the industrialised house building industry is worthy of inves-
tigation. For the present research, the term knowledge-based design (KBD) is used
as an alternative for knowledge-based engineering. KBD is considered to be a more
appropriate description, as it captures the essence of the architecture, engineering
and construction (AEC) industry and the diversity of factors involved in the design
process, which are not exclusive to the field of engineering.
Even though KBD methods have been available for many years, there has not been
widespread adoption within the industrialised house building or architecture engi-
neering and construction industries. Within the timber industry, this is likely due to the
established design tools being domain specific, expert systems that address specific
timber design and processing needs. Additional barriers may also include the cost of
KBD capable systems, skills shortages and the perception that the relevant software’s
were or are ‘too technical’. There is, however, evidence that knowledge-based design
might be making inroads into the industrialised house building industry, particularly
in Sweden (Manutelligence 2017; Sandberg et al. 2008) and the mass timber building
industry in Vancouver (CadMakers 2018). The benefits of such an approach are likely
to be design integration; detailed design automation; standardisation of manufactur-
ing/assembly process; multi-product production lines and knowledge management.
996 G. Day et al.
Fig. 2 Achievable creative design time allocation (Verhagen et al. 2011)
The combination of these improvements is likely to result in more efficient design,

less waste and less rework.
4.1 Current Industry Practices
Parametric, three-dimensional design tools including Building Information Mod-

elling (BIM) are now ubiquitous in the architecture, engineering and construction
(AEC) industry. There is no doubt that these tools have provided significant benefits.
The integration of many BIM tools with the detail design/manufacturing/fabrication
environments is in contrast, very fragmented (Smith and Timberlake 2010) and fre-
quently requires separate software solutions or ‘plugins’ with tailored functionality.
This approach is very common within the timber industry as it allows stakeholders
to adapt their tool sets to suit their needs, yet data exchange is often achieved by con-
verting to an appropriate file format for subsequent processes. This is accomplished
either by direct conversion or open third-party formats such as Industry Founda-
tion Classes (IFC). Consequently, this method of operation requires the storage and
configuration management (coordination) of the many differing file formats which
are derived during building design, manufacture and construction phases. Data con-
version between file formats invariably results in data fragmentation; loss of model
Fig. 3 Example design and part manufacturing digital workflow
functionality and loss of embedded knowledge. Contemporary ‘cloud’ technologies

now provide central, accessible storage repositories, but they do not solve the fun-
damental issues surrounding the need to convert data between native formats.
The work flow illustrated (Fig. 3) is typical of BIM to fabrication in the timber
industry. It is by no means exhaustive, but it serves to illustrate that the processes
are disconnected or siloed, and often rely upon manual intervention. Any upstream
changes will result in the inefficient duplication of previous downstream efforts.
This lack of design, simulation, fabrication, data and knowledge integration may
deliver an adequate working methodology for ‘one off’ buildings, where productivity
expectations are relatively low, but it provides for a very complex formula for the
manufacture and construction of mass customised, industrially fabricated timber
buildings.
In 2005, Patrick MacLeamy suggested that systemic problems could be over-
come by commencing design activities earlier. The MacLeamy curve (Fig. 4) shows
that earlier design should have a direct correlation with reducing project delays,
wastage and cost (Light 2011). On the surface, and with hindsight, it would appear
to be a common-sense albeit a ‘band aid’ solution. Later Holzer provided evidence
suggesting that there were deficiencies in such an approach ‘because of changes
through BIM on operations, the distribution of effort, impacts on design quality’
and questions surrounding the validity of the data used to generate the curve in the
first instance (Holzer 2011). This position is also supported by the observation that
current ‘Parametric design processes will only slightly affect the economy of the
whole AEC industry because, it is a labour intensive and slow process’ (Andia and
Spiegelhalter 2017). The McKinsey Global Institute also concluded that ‘despite the
adoption of BIM, improvements in digital design technologies are still needed to
increase productivity’ (Woetzel et al. 2017).
998 G. Day et al.
Fig. 4 The MacLeamy curve (Light 2011)
4.2 Towards Level 3 Building Information Modelling

and Knowledge-Based Design
Notwithstanding the observations made in the previous paragraph, the Boston Con-
sulting Group (Gerbert et al. 2016) discuss the integration of digital technologies
throughout the full design, construction and building operation lifecycle suggesting
that significant cost savings are possible. For some, the concept of this type of tech-
nology is often referred to as BIM Level 3. A full specification has yet to be defined
(NBS 2018), however, it can be anticipated that it will have similar characteristics to
the model proposed by the Boston Consulting Group who, suggest that ‘Parallel and
robust design and engineering; real-time data sharing, integration including coordi-
nation across stakeholders and BIM-enhanced operations/maintenance’ will be the
foundations for future systems. It can therefore be proposed that the fully integrated
collaborative workflow illustrated below (Fig. 5) will be a likely outcome and is
a model that is more commonly referred to as product lifecycle management (PLM)
in industrial sectors. The timber construction industry is well placed for being able
to capitalise on such technologies.
Fig. 5 Integrated (PLM) workflow
Fig. 6 Cost, knowledge and design freedom (Verhagen et al. 2011)
Much of the justification for the development of knowledge-based engineering

(KBE) within the aerospace and automotive industries was broadly related to enabling
much greater time to quickly explore design alternatives. This is illustrated by a graph
(Fig. 6). The graph suggests similar possibilities to MacLeamy’s argument, in that
earlier knowledge acquisition (design) increases design freedom and reduces the
committed costs (Verhagen et al. 2011).
The convergence of factory processed timber building elements and assemblies,
integrated design and collaboration tools, and the use of knowledge-based design
(KBD) methodologies offers a favorable environment which could address the design
innovation and productivity dilemma in the construction industry. The combination
of these three factors provides the foundations for the exploration of KBD within the
context of industrialised timber building manufacture at the Innovation in Applied
Design Laboratory (IAD) at the University of Sydney. This work is part of a collab-
1000 G. Day et al.
orative project in conjunction with Lendlease and DesignMake, looking at strategies

to boost productivity and foster architectural design and process integration in the
context of multi-storey timber housing.
The employment of knowledge-based design (KBD) methods should accelerate
routine design activities. The use of KBD or automated routine design within a PLM
environment might also justify the development of single, complete design models
of buildings or ‘digital twins’ that encompasses every detail. This is significant for
numerous reasons because digital twins will enable the design of the overall man-
ufacturing, construction and logistics processes. In the context of KBD however,
the design processes will be highly predictable and given the appropriate impetus
could assist in improving operator learning, processing, tooling, machinery, design
for manufacture and assembly, ‘debugging of design data’ (knowledge), material
costs and rates of production (Delionback 1975). All of which contribute to a much
more productive manufacturing environment.
The contribution of timber (the product) within this equation cannot be under-
stated. Timber can be quickly processed in a repeatable manner with close manufac-
turing tolerances being reliably achieved. The ability to deliver parts and assemblies
which conform to the desired specification in terms of form, fit and function is central
to the success of an industrial KBD system that provides detailed design rather than
just ‘design intent’. The investment in resources, process and time will amount to
naught if the product does not fit at the construction site.
5 Case Study: Timber-Framed Wall
The purpose of the case study was to test knowledge-based design methods within
the context of industrialised house building using a product lifecycle management
(PLM) design environment. The intention being to identify potential shortcomings
and improvements in our ideas and methodologies for future work, including assem-
bly, component and geometry properties and the relationships between them. Fur-
thermore, the study also provided the ability to test the claims relating to routine
design productivity including a means of investigating wall edge connections and
the automation methods for wall instantiation within building models.
It is understood that there are many commercial, off the shelf, domain specific
timber/wall framing tools which include embedded knowledge (Para 4.0). This case
study is however a part of a broader investigation into technologies that will assist
industrialisation techniques/methods, knowledge acquisition, implementation and
reuse within the timber construction industry.
Fig. 7 Timber-framed panel
5.1 Material and Tool Selection
A timber-framed wall was selected as a case study for several reasons. It is a

ubiquitous, highly configurable, easily understood, basic building element. From
a knowledge-based design (KBD) point of view, it is an assembly which can be
described by requirements and rules (knowledge).
The design of a prefabricated timber-framed wall (Fig. 7) may appear simple
however, the acquisition and encoding of the associated knowledge is a deceptively
complex activity. The non-recurring time associated with the activity can easily be
justified due to the repetitive nature of the assembly. The amount of information
flowing from the early stage of the design up to the manufacturing and construction
process is significant and often fragmented even within a single design team. Timber-
framed walls must satisfy various performance requirements, each being defined by
a different specialist based on their own experience, clients’ directions, national
codes or specifications. Moreover, when dealing with prefabricated construction, the
design phase must also be informed by logistic and manufacturing constraints. For
example, maximum panel length might be determined by limitations imposed by
transportation, manufacturing equipment, construction site handling and assembly
strategies. Encoding this knowledge within a template model provides the ability to
standardise and automate the routine design process and therefore control variability
within fixed well understood boundaries.
1002 G. Day et al.
The product lifecycle management (PLM) platform selected for the case study
was Dassault 3DExperience® as it provides an interconnected design, analysis, man-
ufacturing and data management environment, including KBD capabilities. It may
even be considered an early example of the direction that BIM level 3 may evolve.
The tool is not widely used in the industrialised house building industry however,
its lineage would suggest that it has a good chance of facilitating the productivity
gains previously discussed. It was also used to great effect during the design and con-
struction of Brock Commons at the University of British Columbia by CadMakers, a
company who specialise in virtual construction (CadMakers 2018). Other candidate
software solutions were Siemens NX® which delivers similar modelling, simulation,
KBD and manufacturing capabilities, but it is not integrated into the wider con-
struction/BIM environment via Industry Foundation Classes (IFC) and Parametric
Technology Corporation Creo® which also lacks BIM/IFC integration.
5.2 Method
There are numerous approaches to designing timber-framed panels using the

3DExperience® platform. The overarching needs and requirements of the model
will likely determine the method used. The model complexity could range from a
‘lightweight’ conceptual wireframe representation which calculates overall compo-
nent sizes and quantities in the form of a user defined feature (UDF) (Hoffmann and
Joan-Arinyo 1998) to a fully detailed assembly representation or ‘digital twin’ capa-
ble of representing every component, including, connections, size, configuration,
physical properties and quantities including Engineering Bills of Materials (what
was designed) and Manufacturing Bills of Materials (what was designed plus all
of the necessary processing materials and equipment). The approach also enables
mechanical and thermal simulation along with the determination of the manufactur-
ing processes and assembly methods. For the purposes of exploring a knowledge-
based design approach, a fully detailed assembly representation provided the most
appropriate solution.
5.2.1 Knowledge Capture
Methodology for Knowledge-Based Engineering Applications (MOKA) is a fre-

quently cited source for knowledge-based engineering (KBE) and was used to define
our initial knowledge-based design (KBD) strategy. The knowledge associated with
the construction of a timber-frame panel was acquired from many sources, including
domain experts familiar with structural design and manufacturing. Product speci-
fication information and our own understanding of the construction methods also
contributed to the acquired knowledge. This domain knowledge was initially docu-
mented in the form of sketches, drawings and text acquired during unstructured inter-
Fig. 8 Timber-framed wall

classification and structure
views and was collated and uploaded to the product lifecycle management (PLM)
server storage environment.
Attempts to use the MOKA beyond the initial capture stage proved unproduc-
tive. MOKA advocates modelling knowledge using tools such as Unified Modelling
Language however, this approach results in a significant duplication of effort and is
likely a result of MOKA being intended for use with considerably different toolset
than those available today (Knutson n.d.; Stokes 2001). While the organisation of the
acquired knowledge is very similar, the 3D interactive nature of the selected design
tool and the ability to define product structures, embed knowledge and test within
the design environment renders many of the MOKA layers redundant. Our strat-
egy became less formal, but recognised the need for a framework for the numerous
activities including the classification of the assembly, parts and materials, require-
ments analysis, functional design, interface, testing strategy, integration testing and
maintenance of the developed tools (Lovett et al. 2000).
Based upon our acquired understanding of the wall from the knowledge acquisi-
tion phase, the assembly structure classification was defined and constructed (Fig. 8)
within the modelling environment. The Industry Foundation Classes (IFC) standard
provided a means of developing the ontology. In the case of a timber-framed wall,
the assembly and part classification is reasonably simple and can be developed using
the IFC Data Dictionary as a guide (BuildingSMART 2018).
1004 G. Day et al.
Fig. 9 Input line nomenclature
5.2.2 Model Development
The development of the model was an iterative process which highlighted an early
requirement to understand the context in which the model is to be instantiated. In this
case, the wall inputs consisted of a named line to describe the panel configuration,
panel length and the bottom of the panel and an input plane to provide the top limit
of the panel. This approach evolved from the use of two-dimensional (2D) floor plan
sketches which are typically used to define wall layouts. Therefore, it was necessary
to build the panel iteratively to fully understand what needed to be defined by the
line nomenclature (Fig. 9), and what could be automated.
Each 2D line associated with a wall location was identified using the previous
method. The combination of a line, top limit (plane) and encoded name provided the
ability describe the location and configuration of multiple panel types with significant
variations in configuration. A script written using the inbuilt enterprise knowledge
language decomposes the line name and creates the necessary input parameters and
geometry for each wall panel in the form of a user defined feature (UDF) (Hoffmann
and Joan-Arinyo 1998). The development of this process demonstrated the need for
good documentation and configuration management. For example, changes to the
line nomenclature will also need to be reflected within the input script and the UDF.
If the three elements become uncoordinated for any reason, the designed process will
not function correctly.
The line nomenclature also provides a means of reducing the complexity or dimen-
sionality of the panel templates. For example, a template that could adapt to any type
or circumstance would have an incomprehensible number of configurations. The
Pre-definition of panel types and features, aided a significant reduction in complex-
ity but requires the development of multiple templates. This approach does however,
rely upon flawless coordination with the input UDF along with a need for additional
conditional rules to control flow.
As discussed, the panel structural hierarchy has three primary levels, the purpose
being to classify the assembly in a logical manner that represents how it is assembled.
Each level provides the appropriate containers for knowledge storage, geometry,
documentation and material definition associated with each assembly or individual
component. The template model descriptions must also follow strict logical naming
convention to enable automation and identification of individual assemblies and parts.
All the sub-components of the panel model were developed as reusable
knowledge-based content. This is an important aspect of knowledge-based design.
Fig. 10 Stud template construction
An example being the stud and plate templates (Fig. 10) where embedded inputs and
rules controlled every aspect of the part location, orientation and geometry.
Defining all the panel components as adaptable templated parts has two major
benefits in that it allows us to focus on the component input definitions rather than
the repetitive aspects of modelling a significant number of individual components.
Furthermore, the component templates can be easily reused or evolved in future
models if the interface requirements are clearly described and understood.
The overall development of the assembly was dominated by the structural layout of
panel. On the surface this appeared to be a reasonably straightforward task however,
exceptions to the acquired knowledge soon revealed themselves and illustrate the
potential complexities in acquiring and decoding domain knowledge even within
simple assemblies. The structure requirements, relations and constraints (Table 1)
were initially generated using the explicit knowledge acquired from relevant domain
experts. As the model developed, the need for additional rules became obvious.
The tacit knowledge was not acquired during the knowledge acquisition phases,
but became apparent requirements as the model progressed. This resulted in some
rules needing exceptions depending upon specific circumstances with solutions being
implemented to allow for these changes. The most likely reasons for these occur-
rences is a combination of assuming that we had acquired all the necessary domain
knowledge and not understanding the ‘know-how’ that would normally be part of
the tradespersons domain. The fact that we could easily identify the shortcomings is
significant because the overall benefits become dubious if the model does not behave
correctly and represent reality.
Conditional rules, inputs and checks are used throughout the model and provide a
means of interpreting and implementing the many requirements (Fig. 11). They define
1006 G. Day et al.
Table 1 Stud/plate—requirements, relations and constraints

S. No. Requirement Description/comment Knowledge type
1 Stud material specification: Defines physical Explicit
MPG 10 or 12 parameters
2 The stud wall consists of A basic undefined Tacit
an outer frame with requirement
horizontal plates and two
vertical studs. Intermediate
vertical studs are located
between the outer studs
3 Stud widths could be Stud width rule Explicit
70 mm, 90 mm, 120 mm or
140 mm
4 If the stud width is 70 mm Basic stud thickness rule Explicit
then the thickness is
35 mm, else it is 45 mm
5 The distance between the Stud spacing. The rule also Explicit
studs may be 400 mm or assumes a surface sheeting
600 mm depending on load width of 1200 mm
case
6 The distance between the Stud spacing rule Explicit
second last and last stud is
variable due to wall length
divided by stud pitch not
being an integer
7 The second last stud cannot Clash prevention Tacit
interfere with the last stud
or a connection reinforcing
stud if it is present
8 Sheeting Joints within the Stud spacing rule Explicit
panel should always be on
stud centrelines
9 The maximum top/bottom Material size restriction Explicit
plate length is 6 m and calculation of
bottom/top plate joint
positions
10 The maximum wall length Transportation limitation Explicit
is 12 m
11 Bottom and top plate nail Nail plate requirements Explicit
plate joints must not
coincide with stud
positions
12 The nail plate size depends Nail plate requirements Tacit
upon the stud width
(continued)
Table 1 (continued)
S. No. Requirement Description/comment Knowledge type
13 Jack stud pitches should Jack stud rule Explicit
observe the overarching
stud pitch rule regardless
of the opening position
14 Perpendicular end End configuration rule Explicit
connections must be
reinforced with an
additional internal stud
15 Intermediate wall Intermediate wall Explicit
connections along the wall connections rule
length shall be reinforced
with additional internal
stud(s)
16 If an intermediate Intermediate wall Tacit
connector coincides with connections rule
an existing stud the full
reinforcing may not be
required
17 To accommodate end End connection Tacit.
connections, the sheeting requirement. This
may need to overlap the requirement impacts the
entire first stud or last stud pitch of the second stud. If
or extend one stud the sheeting does not
thickness beyond coincide with the centreline
of the first stud the pitch of
the second stud must be
varied to account for the
change in sheeting location
18 Wall end offsets need to End Offset rule may alter Tacit
vary depending end stud position
configurations
19 A wall bottom offset Bottom offset rule. Tacit
parameter is required to Changes the stud length
account for conditions such
as sole plate thickness or
recessed slab edges
20 A wall top offset may be Bottom offset rule. Tacit
required to account for Changes the stud length
conditions such as
assembly spacers
21 The panel configuration Top and bottom plate Explicit
changes if it exceeds 6 m in extensions and appropriate
length width nail plates are
required
1008 G. Day et al.
Fig. 11 Stud end conditional rules
the behaviour, by driving the skeleton geometry and determining the conditions in
which specific components are needed.
Many of the rules follow simple ‘If then, else’ statements, changing dimensions
when appropriate. Other rules use enterprise knowledge language scripts combined
with resource tables to select specific configurations and load or unload components
to or from the model. This approach differs from simply hiding components in that
the model physically changes to reflect the selected configuration and not just appear
correct.
More sophisticated rules use enterprise knowledge language to determine the cur-
rent panel configuration and select the appropriate component configuration depend-
ing upon input location. This is illustrated by the intermediate connectors (Fig. 12)
which also provides a good example of tradespersons ‘know-how’. The model must
account for variable stud positions due to pitch, end conditions and the position
Fig. 12 Intermediate and reinforcing stud locations
of the intermediate connection (if needed), to determine and implement the correct
solution.
In comparison, the rules defining the external sheeting, cladding and insulation
were relatively straightforward. Edge condition rules within the structural frame
determined the sheeting and cladding overlaps and insulation sizes were determined
by the calculated stud spacing.
The development of the detailed model also provided the ability to consider the
physical properties of each component and their contribution to the overall mass
(Fig. 13). Although not tested this approach also enables the simulation of the walls
structural and thermal performance within the product lifecycle management (PLM)
environment.
5.2.3 Model Testing and Instantiation
Testing of the model functionality to demonstrate that it performs in accordance with

the rules and to ensure that there is no unexpected behaviour is vitally important. Test-
ing was performed during model development and later using a test matrix to ensure
that all potential configurations worked correctly. Limiting the model behaviour using
different panel types provided significant advantages at this stage by reducing the
necessary test combinations.
To achieve the knowledge-based design (KBD) criteria of retention and reuse, the
‘Template’ model was instantiated within a test environment consisting of the input
criteria discussed earlier (Fig. 14). This was achieved using an enterprise knowledge
language action which interrogated the user defined features, selected the appropriate
1010 G. Day et al.
Fig. 13 Material properties
template model and passed the parameters to the panel template inputs. This process
determines the location, configuration and size of each panels and if necessary assigns
individual references to each component. This process also highlighted a need to
coordinate and document the input criteria and logical naming conventions used
to ensure that the templates work as designed. The overall instantiation time for
each detailed wall panel was approximately two and a half minutes including the
associated documentation.
5.3 Results and Discussion
The case study has shown that knowledge-based design (KBD) within a product
lifecycle management (PLM) environment offers a potential solution for developing
routine and highly customised automated design solutions for the emerging mass-
customised timber construction industry. The technology provides high levels of
design automation for detailed assemblies, components, geometry and associated
two-dimensional documentation. The combinations of technologies and methods
may offer a way of achieving the goal proposed by MacLeamy in 2005. The process
does, however, require a non-recurring early investment in knowledge acquisition
and the development of template models and applications.
Published KBD methodologies and techniques are not necessarily valid for the
KBD tools used in this study. Methods and tools that work in parallel with current
integrated PLM tools are needed and are being explored by the IAD Lab. It is clear
Fig. 14 Instantiated panels
that design knowledge needs to be accessible to all stakeholders regardless of their

role and it is recognised that using a proprietary PLM database may provide a barrier
to some.
Despite the vast advantages offered by KBD approaches, they are still unlikely to
offer fast ‘off the shelf’ benefits for ‘one off’ assemblies. The use of KBD must be
justified and this usually requires a reasonable level of repetition of the part or assem-
bly type (walls, facades, floors, systems for example) to ensure that the investment in
nonrecurring design effort is worthwhile. The ability to process timber easily using
CNC technologies also helps justify the use of the technology as large close tolerance
parts can be manufactured without the need for significant manual intervention. Once
the knowledge has been acquired and implemented, KBD provides the possibility
of significantly reducing the routine design time. It is not a panacea, but it provides
a tangible means of exploring design alternatives and achieving detailed outcomes
during earlier design stages. The study demonstrated however that acquiring all the
information vital to the performance of the model during the knowledge acquisition
phase was not an easy task. In this case study, the problems and solutions were simple
to identify, but this may not always be the case.
It can be foreseen that individuals with industrialised house building (IHB) design
skills or timber processing knowledge are unlikely to have skills in KBD and vice
versa. Therefore, the IHB industry may need to evolve its skill base by recruiting
individuals with a combination of software programming as well as design skills to
take advantage of such tools. This is reinforced by the concept of the “Knowledge
Engineer” which is often discussed within knowledge-based engineering (KBE) lit-
1012 G. Day et al.
Fig. 15 Stud parent/child relationship graph
erature. Knowledge Engineers are trained in the methods and techniques necessary to
elicit, interpret and formalise acquired knowledge in a KBE environment. Although
uncommon, similar parallels can be found within the architecture, engineering and
construction industry with specialist design groups and companies that are now taking
advantage of knowledge-based design tools to increase the efficiency and accuracy
of building design.
The case study revealed that describing model geometry in a logical manner that
enables the interpretation and understanding of its function for maintenance and
reuse is an essential requirement. The BuildingSMART® dictionary provides defi-
nitions of geometry for file translation purposes but is not intended for developing
model geometry. Despite attempts, our findings showed that the development of
Description Logics (DL) for the geometric elements (points, lines etc.) may have to
be limited to component or assembly interfaces (inputs or outputs) rather than all the
construction geometry within individual models. This finding is also supported by
research undertaken by the National Institute of Standards and Technology (NIST),
telling us that ‘Restriction to the domain of Description Logics may be impossible if
a very low level of abstraction is to be achieved’ (Patil et al. 2005). A consequence of
not assigning descriptions to geometry is that models are constructed using default
names which can be self-defeating in the case of a KBD application. At present,
the process relies upon the model developer to describe the function of the geom-
etry within its properties. The screenshot (Fig. 15) is an example which illustrates
the parent/child relationship for the construction of a stud template model. Descrip-
tive names have been assigned to the construction geometry, parameters and rules
to assist with understanding the function of the model. The formulas have retained
their default names.
The method for reusing a template model may not be immediately apparent to a
subsequent user, without in-depth interrogation to discover the necessary resources
or input requirements. This further reinforces the need for clear, unambiguous
and accessible method of documenting all the associated files including logical
input/output requirements to ensure that the context for reusing templated mod-
els (assets) are formatted correctly and model resources correctly addressed. The
case study showed that documenting assembly templates, user defined features and
scripts, once a practical level of maturity had been achieved, provided a reason-
able solution. The longer-term maintenance and reuse of the model depends heavily
upon many circumstances. Future users need to know that the model(s) exist and
understand the models’ structure and construction. Furthermore, they will also need
a detailed knowledge of the model inputs including its context, related scripts, user
defined features and instantiation methods. Without this information, users might
be inclined to reconstruct the models rather than reuse them (Davis 2013). Explo-
ration of the most appropriate methods and medium for this activity is ongoing and
includes more philosophical discussions relating to whether a knowledge repository
should be within a single proprietary PLM environment.
Regardless of our observations, well executed KBD within a PLM environment
offers the potential for design automation and mass customisation. In the case of
industrialised house building it can be expected that given sufficient time to develop
the reusable content, it will provide highly automated building design and manufac-
turing process capabilities especially within the mass timber industry. We can foresee
that the combination of KBD and PLM technologies will deliver a single collabora-
tive building design and development environment that can provide the geometry for
semi-automated machining processes and engineering simulations without the need
to convert or transfer data files. A continuous emphasis on model accuracy, detail
and quality will encourage further automation to avoid non-value adding repetitive
design or process tasks and provide the basis for many additional benefits including
construction, logistic and factory simulations.
6 Conclusions
Knowledge-based design (KBD) enables high levels of routine and customised

design automation of detailed timber assemblies, components and associated two-
dimensional documentation. The use of KBD tools and processes need to be justified
as they require a non-recurring early investment in knowledge acquisition and the
development of template models and associated applications.
To achieve maximum return on investment, there is a prerequisite for a clear,
unambiguous and accessible method of documenting all the acquired knowledge; the
associated model resources; logical input and output requirements and management
of the configuration and control of content. The product lifecycle management (PLM)
database provides support for some of these activities, however established software
development methods may also provide additional direction. This approach requires
foresight, appropriate tools and a significant investment in time and effort. Overall
design costs will likely increase in the short term before the benefits are realised
through reuse of resources and the development of highly detailed models which
have the potential to significantly reduce project risk. The longer-term ability to
understand, maintain and reuse of the acquired knowledge is fundamental to the
1014 G. Day et al.
success of a KBD methodology. Without easily accessible information, users might

be inclined to reconstruct models rather than reuse them, thus the ability to capitalise
on the non-recurring investment is diminished.
Despite the potential benefits, the main challenges in adopting a knowledge-
based design approach in the future IHB timber industry are likely to be a need for a
meticulous approach to detailed design, a shortage of skills, and a need for a common
vocabulary to assist with the standardisation of knowledge-based models.
Our future research direction is guided by the need for a methodology which
combines high quality knowledge acquisition and cataloguing techniques with an
accessible ‘knowledge platform’. This may result in some duplication of effort how-
ever, it is our view that ‘raw knowledge’ needs to be easy to document, agree on and
should be open to all project stakeholders.
The convergence of large scale factory processed timber, highly integrated design
tools and a suitable KBD methodology may provide the appropriate mix of process,
product and resources to provide the foundations for increased design, manufac-
turing and construction productivity. It is, however, anticipated that it will initially
be in the form of highly resolved designs which will enable predictable fabrication
and construction workflows. As methods evolve, it can be foreseen that a reduction
in routine design time will be achieved without sacrificing design quality. Combin-
ing this approach within the timber industry may provide a technological leap that
addresses some of the big and seemingly difficult challenges facing the architecture,
engineering and construction (AEC) industry. In our work, we have observed an
emerging willingness within the AEC industry towards the type of transformation
we outline above. In closing, we hope to have demonstrated, in a small way, how the
industry can take steps towards these valuable goals and have pointed towards some
of the opportunities and difficulties on the way.
Acknowledgements This research was conducted as a part of CRC-P50578 in conjunction with

Lendlease and DesignMake. The CRC Program supports industry-led collaboration between indus-
try, researchers and the community. We would like to thank Richard Neuhercz, Dinah Zhang and
Gregor Rossegger for their insight and support.
References
Aitchison M et al (eds) (2018) Prefab housing and the future of building product to process. Lund
Humphries, London
Andia A, Spiegelhalter T (eds) (2017) Post-parametric automation in design and construction.
Artech House, Boston
Bender S, Fish A (2000) The transfer of knowledge and the retention of expertise: the con-
tinuing need for global assignments. J Knowl Manag 4:125–137. https://doi.org/10.1108/
13673270010372251
BuildingSMART (2018) BuildingSMART data dictionary. BuildingSMART. http://bsdd.
buildingsmart.org/. Accessed 19 Jan 2018
CadMakers (2018) Virtual construction. https://www.cadmakers.net/. Accessed 8th Feb 2018
Chuan CK (2013) In-principle acceptance. Singapore Government, Singapore
Davis D (2013) Modelled on software engineering: flexible parametric models in the practice of
architecture. RMIT University
Delionback L (1975) Guidelines for application of learning/cost improvement curves. George C.
Marshall Space Flight Center, Alabama
Egan J (2014) Accelerating change: rethinking construction. http://constructingexcellence.org.uk/
wp-content/uploads/2014/10/accelerating_change.pdf. Accessed 1st May 2018
Gann DM (2010) Construction as a manufacturing process? Similarities and differences between
industrialized housing and car production in Japan. Constr Manag Econ 14:437–450. https://doi.
org/10.1080/014461996373304
Gerbert P, Castagnino S, Rothballer C, Renz A, Filitz R (2016) The transformative power of
building information modelling. The Boston Consulting Group. https://www.bcg.com/en-au/
publications/2016/engineered-products-infrastructure-digital-transformative-power-building-
information-modeling.aspx. Accessed 17th Nov 2017
Hoffmann CM, Joan-Arinyo R (1998) On user-defined features. Comput Aided Des 30:321–332.
https://doi.org/10.1016/S0010-4485(97)00048-1
Holzer D (2009) Sense-making across collaborating disciplines in the early stages of architectural
design. RMIT University
Holzer D (2011) BIM’s seven deadly sins. Int J Architectural Comput 9:463–480
Kingston JKC (1998) Designing knowledge based systems: the CommonKADS design model.
Knowl Based Syst 11:311–319. https://doi.org/10.1016/S0950-7051(98)00071-9
Knutson S (n.d.) Knowledge based engineering. Stanley Knutson. http://www.stanleyknutson.com/
kbe-history.html. Accessed 28th Mar 2018
LaRocca G (2012) Knowledge based engineering: between AI and CAD. Review of a language
based technology to support engineering design. Adv Eng Inf 26:159–179. https://doi.org/10.
1016/j.aei.2012.02.002
Light D (2011) BIM implementation—HOK BuildingSMART. National Building Specifi-
cation. https://www.thenbs.com/knowledge/bim-implementation-hok-buildingsmart. Accessed
12th Apr 2018
Lovett PJ, Ingram A, Bancroft CN (2000) Knowledge-based engineering for SMEs—a methodology.
J Mater Process Technol 107:384–389. https://doi.org/10.1016/S0924-0136(00)00728-7
Manutelligence (2017) Product service design and manufacturing intelligence engineering platform.
Manutelligence. http://www.manutelligence.eu/. Accessed 24th Sept 2017
NBS (2018) BIM levels explained. National Building Specification/RIBA Enterprises Ltd. https://
www.thenbs.com/knowledge/bim-levels-explained. Accessed 28th Feb 2018
Nord T (2008) Prefabrication strategies in the timber housing industry: a comparison of Swedish
and Austrian markets. Doctoral thesis, comprehensive summary, Luleå University of Technology
O’Dell C, Grayson CJ (1998) If only we knew what we know: identification and transfer of internal
best practices. Calif Manag Rev 40:154–174
Patil L, Dutta D, Sriram D (2005) Ontology formalization of product semantics for product lifecycle
management. National Institute of Standards and Technology, Gaithersburg. https://dx.doi.org/
10.6028/NIST.IR.7274
Patterson D (2014) Completed 10-storey apartment in Australia: Forté from an owner/developer per-
spective. Lendlease. http://www.woodworks.org/wp-content/uploads/TTWB-2014-Patterson-
10-story-apt-Australia.pdf. Accessed 11th Apr 2018
Robertson AB, Lam FCF, Cole RJ (2012) A comparative cradle-to-gate life cycle assessment of mid-
rise office building construction alternatives: laminated timber or reinforced concrete. Buildings
2:245
Sakuragawa S, Miyazaki Y, Kaneko T, Makita T (2005) Influence of wood wall panels on physi-
ological and psychological responses. J Wood Sci 51:136–140. https://doi.org/10.1007/s10086-
004-0643-1
Sandberg M, Johnsson H, Tobias L (2008) Knowledge-based engineering in construction: the pre-
fabricated timber housing case. J Inf Technol Constr 13:408–420
1016 G. Day et al.
Smith RE, Timberlake J (2010) Prefab architecture: a guide to modular design and construction.
Wiley, Incorporated, Hoboken, United States
Stehn L (2009) Industrialized timber housing in Sweden, Internationales Holzbau-Forum 09. http://
www.forum-holzbau.com/pdf/ihf09_Stehn.pdf. Accessed 12th Apr 2018
Stokes M (ed) (2001) Managing engineering knowledge: MOKA: methodology for knowledge
based engineering applications. Professional Engineering, London
Verhagen WJC, Bermell-Garcia P, van Dijk REC, Curran R (2011) A critical review of knowledge-
based engineering: An identification of research challenges. Adv Eng Inf 26:5–15. https://doi.
org/10.1016/j.aei.2011.06.004
Woetzel J et al (2017) Reinventing construction: a route to higher productivity. McKinsey Global
Institute, Shanghai
(Re)construct with Wood, The Case
Study of Amatrice’s Prefabricated Bus
Station Designed In BIM Environment
Maria Grazia Cianci, Daniele Calisi, Francesca di Benedetto

and Matteo Molinari
Abstract The project stems from an agreement between the regional transport com-
pany COTRAL spa and the Department of Architecture of the University of Roma
Tre. The need for a new bus exchange node arises following the earthquake of August
24th 2016 that destroyed the city of Amatrice and damaged the neighboring areas.
The project is designed for areas subjected to earthquakes, it is therefore an oppor-
tunity to tackle the difficulties at national level starting from the solution of the local
problem. It is a wooden building, environmentally aware, combining active and pas-
sive strategies together with the technological innovation of the modular systems of
which it is composed, with the objective of reducing costs and consumption. The
drafting of the project documents, the choice of technology to be used in the nodes
and the prefabrication system was all realized in the BIM environment. The aim of the
University convention with COTRAL spa is not to deliver only the graphic works, but
a complete package that simultaneously considers the project in three dimensions,
the detailed nodes, the quantification of the materials used, and the solar contribution
given by the its geo-location. The opportunity arising from the agreement, therefore,
lends itself to become a case study, a virtuous example of research on how architec-
ture can intervene quickly and concretely on the territory. An intervention with an
open look at the future of design and representation.

Wood joint systems · History of representation in wood design
M. G. Cianci (B) · D. Calisi · F. di Benedetto · M. Molinari

Roma Tre University, Rome, Italy
e-mail: mariagrazia.cianci@uniroma3.it
D. Calisi
e-mail: daniele.calisi@uniroma3.it
F. di Benedetto
e-mail: francesdibenedetto@gmail.com
M. Molinari
e-mail: mat.molinari@outlook.it

1018 M. G. Cianci et al.
1 Introduction
The project, coordinated by prof.ssa arch. Maria Grazia Cianci was born from an
agreement between the regional transport company COTRAL spa1 and the Depart-
ment of Architecture of the University of Roma Tre. The request provides for the
drafting of a final design for a bus terminal in the town of Torrita, Amatrice (Fig. 1).
It is a wooden building, observant to the environment, combining active and pas-
sive strategies together with the technological innovation of the modular systems
of which it is composed. The aim is the reduction of costs and consumption both
in terms of construction and management. It is a concrete commitment to the envi-
ronment. The agreement is part of a study by the Department of Architecture in
collaboration with COTRAL spa, for the recovery of the parking areas. The theme
dealt with in the following chapter is the second project carried out in collaboration
between the transport company and the University as the OPEN II level Master2 and
Department of Architecture, after the successful intervention for the redesign of the
external spaces of the Ponte Mammolo station in Rome and of the related waiting
room (Fig. 2). Both agreements are part of a new intervention plan of COTRAL spa
in the restyling of the company at the level of staff, means of transport and stations
(Cianci et al. 2018).
The two agreements stipulated so far allow the architecture department to carry
out the research in progress in the areas of design, representation and technology.
The work, begun with the Ponte Mammolo station, links the traditional design with
the modus operandi of the Building Information Modeling (BIM). The research
aims to understand the differences and limitations of each of the systems. With the
second convention, research has reached very high levels of detail and depth. Italian
legislation is evolving to adapt to European and international laws, where the use of
BIM in design is becoming a practice. Only now in Italy it is starting to legislate in
the construction sector.
This project has the objective, to highlight how the use of Building Information
Modeling, applied on a relatively small building, does simplify the realization and
management of the different personalities involved, and even it opens up other ways
of design that would require a complete rethinking of the project through CAD
software.
1 COTRAL spa, (Compagnia Trasporti Laziali), is a company founded in 2001 aimed at suburban
and interurban public transport in the Lazio Region; currently it is the largest and most important
road transport company. The annual flow of travelers on the line is 100 million.
2 At the Department of Architecture is active the curricula of a Master’s Degree in Landscape
Architecture, OPEN, directed by prof.ssa Maria Grazia Cianci whose purpose is to train subjects
who have specialist skills of an interdisciplinary and multidimensional nature regarding the themes
of representation and design of the open space, particularly in urban areas, increasingly central to
the culture of the project and fundamental for defining the quality of the habitat.
Ponte Mammolo convetion.
Head of the Convention: Prof.ssa Maria Grazia Cianci
Working group: Prof. Maria Grazia Cianci, Prof. Arch. Emanuele Von Normann, Prof. Arch.
Daniele Calisi, Arch. Giulia Bassi, Arch. Matteo Molinari, Arch. Marta Rabazo Martin.
(Re)construct with Wood, The Case Study … 1019
Fig. 1 Design floorplan, location of the project intervention within the lot and external arrangement
of the area
2 The Design Area: Amatrice
On the night of August 24th 2016, at 3:36 am there was a shock of magnitude 6.0 with
an epicenter located in the Valle del Tronto, in an area between the municipalities
of Accumoli and Arquata del Tronto. In the months to follow, from October 2016
to January 2017, there were a series of strong shocks with magnitudes between
6.5 and 5.5°. The area affected by these seismic events is a place strongly active
seismologically; and this zone extends to the city of L’Aquila, where the earthquake
of 2009 caused the destruction of the city and more than three hundred dead (Fig. 3).
The seismic sequence has completely razed the country of Amatrice, where only
some buildings have withstood the earthquake, but bringing back such damage that
they can no longer be restored. The seismic event has attracted an interest not only
on a human level and on solidarity, but also on an academic level. The ways in which
to intervene on Amatrice and other cities damaged, the reconstruction of the city
and the possibility of securing the Italian historical heritage have become a topic of
debate and research within the Italian faculties of architecture and engineering but
also among professionals.
Intervening and studying the situation that has hit central Italy in the last decade,
is essential to prevent future accidents not only in the Amatrice area but also in the
historic towns of the Italian territory. The choice of interfacing with the situation of
Amatrice is part of a wider project of reconstruction of the city’s fabric. Through
Fig. 2 Photographs of the project at the Ponte Mammolo station in Rome. From top to bottom ante
operam image of the waiting room, post operam image
Fig. 3 Main areas of seismic interest on the Italian territory. Detailed map on the location of major
seismic events in design area. Seismic activity in the region starting from 24 August 2016. Below
post-earthquake images of the historical urban centers after the earthquake
the agreement with the COTRAL spa, transport company of the Lazio Region, it
was possible to start a virtuous intervention to participate in the reconstruction of
Amatrice.
The possibility to intervene in an earthquake area, to construct a new architectural
artifact, is a way to apply the university departmental research in progress and to
make an empirical element to topics up to now theoretical. The design of a new
exchange node, a link between Amatrice, the neighboring countries and the city of
Rome makes it possible to restore 100% a connection that, to date, due to the natural
disasters that hit the area, is not in operation in necessary quantity.
The new station is not designed to be located in Amatrice, or at least for an initial
period of time, but in the town of Torrita, a neighboring county. The need to place
the new intervention in a nearby town stems from the recovery and planning of
the new Amatrice. For this reason, from the beginning the design idea involves the
design of a removable station that can be transferred in the near future within the
new urban fabric of Amatrice. This request for a non-fixed architectural product, but
transportable in a different area, strongly influenced the construction technique, the
material to be used and the compositional design (Fig. 4).
3 The Design Technology: The Wood
There were many possibilities to design a prefabricated building that could be dis-
mantled and remounted in a seismic area, all the choices were examined but only one
was the most logical and comprehensive of the place where it was to be installed. The
Italian buildings, in particular the historic towns, had always been strongly character-
ized by a construction technique in masonry with horizontal structural elements and
in some vertical wooden cases. The decision to carry out a project with a predom-
inantly wooden construction technology was almost natural, not only for historical
and cultural reasons, but also for the good resistance to seismic events and the possi-
bility of working with prefabricated elements that are easy to dismantle and assemble
(Blass 2004).
Taking a brief historical excursus on the use of wood and its use over time, it
is possible to identify precise phases, ignoring the prehistoric timeline of the first
dwellings.
• I sec. B.C.: Vitruvio describes the types of timber
• 4th century A.D.: introduction of the trusses
• 14th–18th centuries: case blockbau and fachwerk in central Europe
• 16th–17th century: Leonardo studied the inflexed beam, Galileo solved the prob-
lem almost exactly
• 1825: laminated wood system Emy
• 1832: Balloon framing in Chicago
• 1905: multilayer wooden panels
• 1915: laminated wood
• 1933: synthetic gluing resins
The study of historical building, in the seismic field has also led to note that
architectural artifact made of mixed technique, masonry and wood, suffered minor
damages compared to other types of construction techniques that made up the other
buildings. The closest example in recent history is the earthquake of L’Aquila, where
the whole part of the historic center, with buildings made in mixed technique, had
largely been damaged or destroyed, but the damages were of lower gravity compared
Fig. 4 Framing of the design area, Torrita, in the municipality of Amatrice in the province of Rieti.
Detail of the regional CTR of the province of Rieti. Below an overview of the lot
to the peripheral part built in the twentieth century in reinforced concrete. The his-
toric center of L’Aquila has been destroyed several times throughout history, first
in 1259 by Manfredi di Sicilia and in 1703 by an earthquake measuring 6.7 on the
Mercalli scale. The earthquake damaged much of the city and caused the destruction
of many historic and common buildings, with a count of 6000 dead, about 10% of
the population of the time. There were two approaches in the reconstruction of the
damaged and destroyed buildings, the first for those of the poorest classes to rebuild
them as they were before the earthquake, the second for the richest classes was to
readjust the city of L’Aquila to the Baroque style used at the time in Italian peninsula.
The reconstruction of some historic and noble buildings has led to the use of wooden
elements, structural and not, in particular in mixed façade with masonry. This has
led to a strong structural resistance to seismic shocks both to horizontal and vertical
thrust waves. This preliminary study led as a first mentioned to choose wood as a
building material (Fig. 5). Over the centuries, this material in the use of buildings
has undergone considerable technological innovations, making it possible to cre-
ate multi-storey buildings using entirely wooden structures. For this reason, for the
requests of the customer for prefabricability, attention to the environment, reduction
of costs and the timing of the construction site, it was decided to use two different
technologies: the XLam and a punctual structure in laminated wood (Galletti 2004).
From a static point of view, XLam works like the load-bearing masonry, so it
provides better performance when the structure has a box-like layout and the two
sides have ratios that recall the square. The use of this construction technique also
for horizontal partitions, allows to eliminate the presence of horizontal connecting
elements (such as beams) and to have more lights than traditional buildings in wood or
reinforced concrete. Because of these static limitations, of a box-like configuration,
the stalls have been designed with a laminated beam-pillar frame structure. Laminated
wood, like XLam is made of glued wooden layers and allows a greater resistance to
flection stress than normal construction techniques (Pryce 2005). Given the width of
the span, it was decided to design the stalls as a portal with three hinges, connecting the
various frames to each other with steel tie-rods in order to avoid spilling phenomena.
4 The Design: A New Station for the Town of Amatrice
The design guidelines provided by the company were very specific regarding the
size of the rooms and intended use, leaving in return a complete design freedom
with respect to form, materials and construction technologies. The building includes
the housing of a waiting room with toilets, a collective space for COTRAL employ-
ees, and rooms for cleaning staff. In order to avoid the overlapping of the different
functions, the design focused, first of all, on the division between spaces for users
and spaces for the managers of the station, assigning to the former a larger build-
ing of about 90 m2 , while to the latter the smaller one of about 43 m2 . In the first
block, in addition to providing a large waiting room, are contained the bathrooms
and according to regulations and thanks to the slope of the roof, also a second level
walkable, above the toilet area. The second block maintains the functional division
of accesses. The first part of the building, intended for COTRAL employees, has
a collective space, a dressing area, toilets and a full access to the parking area so
that the control tasks on vehicles approaching the stalls can be carried out. The sec-
ond part, dedicated to the cleaners, is placed more externally, and also provides a
dressing room and service bathrooms. The two blocks are divided by the access
and distribution ladder, perfectly aligned with the connection path to the bus parking
stalls (departures and arrivals). The slope of 30%, which the Lazio Region legislation
Fig. 5 From top to bottom: cartography of the city of L’Aquila with the current status of the post-
earthquake shipyards of 2009 published on the institutional website of the municipality, photographs
of some buildings in the historic center damaged and destroyed, historical documents on the anti-
seismic level in the Bourbon-based era on the “baraccata house” with a wooden core and a curtain
wall
provides for areas with high snowfall, has strongly characterized the project: the two
neighboring buildings have opposite slopes, so as to ensure the fall of any accumu-
lation of snow in green areas contiguous and not directly accessible by the public.
The same concept has also been applied to the covers of the stalls. In particular, the
slope of the building of the waiting room is such as to allow the creation of a sec-
ond level, inspectable (average height 1.95 m), above the service bathrooms (useful
height 2.40 m) to be exploited for placement of all the facilities. The choice of the
double pitch was decisive to avoid excessive heights with consequent problems of
deforming optics in limited environments (for example the bathrooms) and dispro-
portionately high, but also to avoid waste of volumetries that would be unnecessarily
lost with false ceilings at 2.40 m (Fig. 6).
Some design choices also concern internal—external permeability: the waiting
room is completely glazed, to allow users to visually check the arrivals and departures
of the buses. However, the lighting factor has also been studied: the south-west wall,
glazed in the waiting room, is also the one most exposed to the rays of the sun in the
afternoon, so a system of wooden slats is provided, beside the aesthetic aspect, they
have the function of protecting from glare. Also, the common room of COTRAL
employees has two glass walls, to allow, in this case, the control on the vehicles
arriving at the terminal (Fig. 7).
The permeability of the two blocks is also accompanied by an attention in the
choices of covered paths and areas of waiting for a short rest stop: the access ramp
for the disabled that fills the difference of 1.20, all external to the building, has been
designed repaired and covered by the latch of the floor and vertical wooden slats.
The same ramp is divided, according to regulations, into two portions, with landing
for parking and maneuvering. Even the public rest area, outside the waiting room,
has been imagined covered and sheltered from the natural elements. Equal attention
was paid to the second building dedicated to the COTRAL staff.
Similarly, also the bus stalls, for the ascent and descent of the passengers, have been
covered by wooden double-pitch structures, respecting the slope of regulations. The
height of these covers is naturally dictated by that of the buses that must necessarily
pass below them. However, a walkway over the pedestrian distribution route to the
stalls at a lower level has also been designed, which also has the COTRAL color
characteristics as a distinguishable sign of the company on site. The roofs of the
stalls, which follow the material characteristics of the main building blocks, are full
above the pedestrians stationing, but have been designed with a fixed brise soleil
system to be covered externally, to ensure the passage of light (with light and dark
effects), but in any case, permeable to atmospheric agents. The paving of the stalls
is imagined as a continuation (without interruption) of the roof, with some custom-
designed accessories, such as benches, recycling buckets, planters and green areas
(Fig. 8).
The project is replicable in every context with problems similar to those of Ama-
trice, is a small building, equipped with a system of dry construction, light and fast
to assemble, energy efficient. It consists of a constructively fast structure both during
production at the factory and during assembly. It is completely prefabricated exclud-
ing the foundation slab and therefore the walking surface. The structural elements
Fig. 6 On top the roof design floorplan, on the bottom project floorplan of the service block for
the staff and waiting room
are prefabricated in wood in order to make assembly on site easier and faster. The
external façades follow a regular and modular design, they are covered with larch
wood lamellas, glazed panels and colored plaster cuts. The simplified type is easily
duplicated. The study of the façade holes has led to the design of two types of frames
depending on the environment they serve. For public areas, the window consists of
completely transparent glass elements, for those for the exclusive use of personnel,
it is designed in such a way as to shield the transparent part with the slat system used
on the façade, interspersing with a yes/no system, in order to allow the light to filter
through the slits with an interval of 10 cm. The alternation of the slats allows a sense
Fig. 7 South elevation of the building and longitudinal section on waiting room and staff block
of continuity in the façade. The building, as previously mentioned, was designed to

be eco-friendly and have the least possible impact on the environment. This partic-
ular attention in the project, is not dictated only by a choice of designers but also by
regional laws that impose energy and environmental standards for all new buildings.
The desire, from a environmental point of view, in the design phase, was to make
the building independent. This energy independence was achieved through the use
of different strategies: a photovoltaic system, a rainwater collection and recycling
system, a type of external coat insulation (specific for each type of exposure) and
a thermal inertia of the opaque casing both horizontal and vertical to allow thermal
displacements suitable for the use of the place. The pitched roof of the waiting room,
exposed to the south-east, allows following the inclination of the roof to position
Fig. 8 Project profiles of the intervention area, with in evidence (blue line) the design ground
movements
the photovoltaic panels. Taking into account these factors, the photovoltaic system
designed, allows such a production of electricity that does not require an attack on
the regional network. The wall pack of external infill is made with a ventilated wall,
with an external coat insulation positioned in part of the air gap. The positioning
of the insulation, the use of the ventilated wall derives from the regulations relating
to the climatic zone in which it is located, that is F3 n this case. This technological
solution has allowed to obtain not only the transmittance values, in respect of the
limits requested by law4 but to have such a high-performance building to obtain stats
3 Introduced by the Decredo of the President of the Republic n. 412 of 26 August 1993. Subdivision
of the climatic communes in 6 zones. These areas are defined by specific values such as temperature
difference of the internal and external environment (GG Graco Day measurement unit) and exposure
to the sun.
4 Regarding the minimum superficial mass: Legislative Decree 192/2005 (Annex I art9 lett. B)
imposes “in order to limit the energy requirements for summer air conditioning and to limit the
internal temperature of the rooms”, that the mass surface of the walls is more than 230 kg/m2 .
Regarding the maximum transmittance for vertical opaque in climate zone F following the
Directive 2002/91/ EC in 2010 the maximum value has been decreased to 0.1. For horizontal and/
or inclined closures is less than 0.18 W/m2 k.
for a class A+ building. This is achieved by merging passive technologies such as the
building envelope and orientation, with energy production tools such as photovoltaic
panels (Rozman and Fajfar 2009).
The structural technologies used within the project, as explained in the paragraph,
are mainly three, reinforced concrete structure for foundations, xlam panels for wait-
ing room buildings and staff rooms and laminated wood structure for the waiting stalls
for the buses. The different technologies used mainly derive from the geometry of the
environments and the design development of each individual element. The decision
to use xlam as a support structure for the waiting room derives from the planimetric
layout of the building. The xlam as the load-bearing masonry has a structural behav-
ior that is performing when used in a box-like static scheme (Ceccotti 2008). The use
of this technology requires the use of a greater quantity of wood than the platform
frame technique but is more efficient in the insulation and management of thermal
bridges. In fact, in this way it is possible to have a continuous wooden wall, without
interruptions on the ground, where it is possible to apply an outer coat as in the case
in question. The only problem of insulation is found in the connection between the
load-bearing structure in cement and the wooden panel, where there is a discontinuity
in the installation and in the types of material used. To overcome this problem, it was
decided to use an internal insulation in connection with a prefabricated wooden floor
with a linear development of 150 cm to dissipate the effect of the thermal bridge.
Among the possibilities offered by this technology, one of the fundamental ones in
this project is not having to use the beams, but a structural panel in xlam. This given
the slope of the 30% roof was a winning solution. The possibility of not having a
false ceiling at a distance from the partition in such a way as to have to cover the
beams or to have structural connecting elements such as not to allow the accessibility
of the room allowed a greater use of the spaces.
In the design of the stalls, the problems encountered in the waiting room were not
present. As external covering elements, there has not been a design need in terms
of thermal insulation. From the design point of view, each stall is thought of as
a structural element, composed of a laminated wood structure, with a main beam
system with an important section, and a system of secondary beams, which served
as a link between a stall and the other. The secondary beams, interspersed with false
beams, had a triple function: structural element, sun-shading element and support
for the substructure of a glazed roof, in the points of connection between one stall
and another. Structural laminated wood was used given the important light free to
cover, 14 linear meters.
5 The Way BIM Is Used in the Design of the Station
One of the fundamental points in the design of the new Amatrice station for COTRAL
spa is not only the compositional aspect but also the way in which it was designed and
finally represented. The representative aspect is a fundamental point in the design.
Fig. 9 The two project proposals of the preliminary design phase. Proposal 1 solution adopted with
larch lath strips and xlam structure. Proposal 2, covering in strips of wood and polycarbonate with
frame structure in laminated wood. Proposals made using CAD drawing software and dedicated
modeling and rendering programs (Livingston 2008)
The specific software used, their ability to manage a complex project and to arrive
at an appropriate scale of detail have greatly influenced the design process.
The project has developed into three designing phases, each one strictly connected
to a specific mode of representation. The first phase includes a specific study of
the intervention area, the context and an analogical elaboration using planimetric
and three-dimensional sketches. This first part of the work, not addressed to the
client, but aimed at project reasoning, remained internal to the design group and
was fundamental to develop the ideas in a more complex way. The second phase
corresponds to the preliminary project, aimed at the client, to transmit the design
idea and receive the approval, following changes to continue to draft the project in
the final phase of execution (Fig. 9).
In this part of the work, implemented by means of CAD and three-dimensional
modeling programs, the project was started by exemplifying the construction tech-
nique, the volumetric and architectural characteristics of the building and the stalls.
Given the requests of the client, two design ideas were formulated based on different
technologies, one designed with a prefabricated frame structure in laminated wood
and one with a supporting structure in XLam. The possibility to work with specific
representation software ranging from two-dimensional programs to rendering soft-
ware has made possible a direct transmission of the architectural object to people not
related to the subject (Mclaren 2008). The graphic works produced with this working
methodology have been used to request the necessary permits to proceed with the
practices aimed at obtaining the area concession and the documents to start the work
(Fig. 10).
Fig. 10 On the right render of the first proposal, on the left of the second
The third phase is the most complex and contains all the design ideas developed
in the project of the new COTRAL bus station. The need to design a prefabricated
wooden building, which involved the interaction of several personalities on the same
design projects, led to the decision to work in the BIM (Building Information Mod-
eling) environment. Starting to work in BIM also falls within the law, signed by
the Ministry of Infrastructure and Transport, called the BIM Decree. The legislation
requires the use of the BIM from 2019 in public contracts with an amount exceed-
ing 100 million euro, to reach 2025 including also public contracts with an amount
exceeding 100 thousand euro. The decree is aimed at aligning Italy with European
countries such as the United Kingdom, Germany, Finland, Norway and Denmark,
where this working method has been consolidated for years and has started to bring
notable results both in terms of construction times and savings in construction costs.
At the same time, the legislators are updating the UNI standard, specifically the UNI
11337, the first Italian technical standard on BIM. To date only three parts have been
written 1-3-5.
The UNI1137 standard derives from the implementation of the Italian legislative
body by the European Directive 2014/24/EU on Public Procurement. In Article 22
paragraph 4 of the Directive there is a direct reference to the Building Information
Modeling not entering into the specific of the methodology to be used with the words
“For public works contracts and design contests, Member States may require the use
of specific electronic tools, such as building information electronic modeling tools or
similar […]” The Italian legislation, in reference to the international ones, specifically
the US, defines the figure of the” Level of development of Digital Objects “(LOD), of
the” Item development level “(LOG) and” Informational attributes “(LOI). Unlike
the US legislation in which the LODs are defined in the numerical order of the
hundreds (e.g. LOD 100, LOD200…) the UNI1137 defines the LODs through the
use of alphabetic characters (e.g. LOD A, LOD B…). The LOD defines the degree of
definition of the BIM project according to the deepening of the LOG and LOI. The
definition of the project in the BIM area is therefore dictated by two main factors: the
geometry (LOG) and the metadata (LOI) of each architectural element. The Italian
legislation defines seven project LODs, from “A” to “G”, where “A” refers to the
symbolic object and “G” to the updated object. The LOD “A” refers essentially to
a two-dimensional geometry with essential metadata. The legislation in addition to
allowing the contracting authorities freedom in choosing the allocation of the LOD
in the tender for public works allows to design not in the specific LOD, specifying at
the beginning the degree of detail of the project. For example, you can design with
a degree of detail greater than D but less than E, in which case you enter a numeric
factor after the identifying letter (Pavan et al. 2017).
In designing the station, specific LODs were considered within the BIM work
area, even though it was not required by the contracting authority, in order to have a
defined and coordinated level of planning among all the professionals involved in the
project. The reference LOD chosen for the design is LOD D.2 because of the degree
of details and metadata included in the project and specifically in some individual
design elements.
Designing in BIM methodology, with a common LOD between all the working
groups, allows to have a final result congruous in all its aspects, without having model
elements, designed with a different detail than the others. This allows not only better
coordination but also a homogeneous design yield.
The possibilities offered by this working methodology and dedicated software
allows not only an instant collaboration within the same design group but also an
interaction with personalities such as engineers and installers on the same model. For
this reason, both for diffusion and for interoperability with the most common CAD
programs, it was decided to work with the Autodesk Revit program. The program
allows you to set up shared coordination files, which help the mentioned personalities
to work in sync and eliminate exponentially the possibility of the error. Working with
this methodology, two main files were set up, a coordination file and a working file
within the design group. The former has the function of a container, loaded on a server,
it is a hub in which the structural, plant and architectural work files are loaded. The
second is the architectural processing file, always a shared file, but only within the
design group, where it is possible to work simultaneously. Working in BIM was not
a choice dictated by the interoperability of the file, but also by the ability to control,
simultaneously in the three dimensions, the spatial development of architectural
design. Another element, already dealt with in the design of Ponte Mammolo, is the
ability to manage not only the three dimensions that underlie the design, but also
the so-called seven dimensions of BIM (Kymmel 2008). In fact, in addition to the
canonical dimensions when this modus operandi is used, the following are taken into
consideration: time (duration analysis), costs, management (management phase of
what has been achieved) and sustainability. To allow this type of work two specific
tools have been used, the first is that of the workset, the second the one of the phases.
Workset allows to create within the same work file the “layers” in which it is possible
to work. A workset is assigned to each design group, so that it does not have to go
to affect or modify the reciprocal work. The workset used were the following: state
of the art, architectural, structural and plant engineering. This subdivision allows,
in addition to a clean job, without overlapping, a simultaneous connection between
the architectural, structural and plant design within the coordination file. In support
Fig. 11 Screenshot of the BIM modeling program highlighting the shared model and the workset
used
of this division of labor, and for a preparation of the workings of the construction
site each file has been divided into phases, shared. The instrument of the phases,
allows to insert each process in the specific time frame. The phases used are the
following: state of the art, demolitions/removals, structure, rustic, finishing, external
arrangement and project status (Fig. 11).
The Regione Lazio since 2008 has passed a law on green building, where it
imposes specific characteristics within each new construction project. Respecting
this law requires respect and management of the additional dimensions used in BIM.
The will and the duty to design a building in class A+ , not only for legislative reasons
but also for a question of image of attention to the environment of society, has led
to the need to develop solar analysis, irradiation and transmittance of the envelope.
The possibility, through third-party plug-ins (in this case study Formit was used)
made it possible to calculate the specific solar load of the building using the geo-
referencing tool inside the software and consequently design a dynamic envelope
that paid attention to the environment.
One of the motivations that led to work in this area in addition to sharing and
multidisciplinarity of the file is the ability to have a systematic control on planimetric
and elevations. In fact, the ability to “trace” a wall septum but at the same time
through parametric algorithms to develop it in elevation allows a simultaneity of the
processes that is not present in the traditional design. The possibility of sectioning
with a theoretical number of infinite planes the building with a perfect correspondence
between plan and section allows to have a millimeter control on every part of the
project.
Another fundamental aspect in the drafting of not only the elaborate but of con-
structive details is the design of specific parametric elements for the project; a custom
design in some cases in others directly provided by the manufacturer. Working in a
parametric way, on the envelope, the structure, the finishes and other elements of the
building allowed for instant modifications, and visible changes in all the dimensions
of the project (Peterson et al. 2001). The wood prefabrication technology, in this
case XLam, lends itself even more than the other construction techniques to this area
of work. When we talk about prefabrication we think of several factors at the same
time, the size of these parts, the transport, the assembly and the coordination of the
construction site. The realized model, integrated then in the execution phase with
the information provided by the manufacturer, allows to understand all this and to
facilitate and speed up the management of the building site (Fig. 12).
BIM is an innovative work tool for many reasons, not only in the field of construc-
tion methodology, but as mentioned previously for the unique modes of representation
that can be used starting from the architectural whole to arrive at the single element
of detail. The digitalization of the design process in the last decades has opened up
many paths in the field of representation, and the field of wood design, specifically
with prefabricated elements, has undergone constant innovation. Using a BIM scope
of work allowed not only a geometric representation but also metadata, allowing the
architectural element to be graphicized not only from the point of view of the three
dimensions. Working three-dimensionally, it is not a specific BIM design tool, in
the architectural panorama, there are many dedicated software that allow this modus
operandi, the substantial difference consists in the data that each single element con-
tains. The design of each individual element is no longer a representation based on
Cartesian coordinates within the space but includes a series of metadata that lead to
an overview of the whole construction process.
6 The Design and BIM: The Model Applied to Other

Realities
Given the strong component of prefabrication, starting from the design of the external
infill panels, the structure, the vertical, horizontal partitions, up to the external seats
of the stalls. The most used tool within the project is that of families. Families
are single parametric elements, included in the project, interchangeable and subject
to different modifications depending on the type. Given the prefabrication of the
building, reasoning in this way was a natural result. An example to be considered
is the external shell of both the stalls and the waiting room and office area. In the
case of the latter, the cladding was made with the curtain wall system. This allows
you to draw a wall without thickness that can be divided into a parametric grid. This
grid identifies panels, each panel can be assigned a specific family, in the case of the
building that of the larch wood lamellas. By assigning to each grid space this type of
Fig. 12 Three-dimensional images taken from the BIM work model of some of the phases used,
from top to bottom: foundations, structure, finishing
Fig. 13 Construction details of the waiting room envelope. On the left, the sky section of the waiting
room block. Right, from top to bottom: detail section of the frame, detail plant of the window, corner
solution outer casing
panel it is possible to obtain a parametric façade, which can be modified, but which
reflects the reality by size of the slats and spacing (Fig. 13).
The use of the grid and the curtain wall, allows an instantaneous positioning of the
substructure in uprights and crosspieces. The division of the grid, through elements
belonging to x and y axes, defined as horizontal grid and vertical grid within the
program, allows an almost instantaneous positioning of the structure of the uprights
and crosspieces.
Fig. 14 Image of the parametric family used to make the envelope of the stalls. In the table you
can find all the information related to the material, dimensions and number of lamellas used
What are the simplifications that this type of design brings, not only to the drafting
of the project, but also to the computation of the elements used? The reasoning for
individual elements, each one characterized by a code, characteristics of the material,
design notes, technical specifications and costs, leads to an all-encompassing vision
of the project entity. There are many ways to get this study. Traditionally, in CAD
programs, it is done manually, through the use of specific blocks and filters. Through
these instruments it is possible, through long counting procedures, to have an estimate
of the materials used and finally to obtain construction costs. Working in BIM the
procedure is the same but automated through the schedules. This tool, used to create
real tables which once revised lead to obtain a precise estimation calculation simply
extrapolating a file from the program. The flexibility of these schedules and the
interaction between three-dimensional model and table is such as to allow a control of
the design elements. Any action carried out on the table will affect the project. Using
the phases, in conjunction with the tables, it allows to obtain, already without setting
the reading fields, the different constructive elements cataloged by work phase. For
example, in the table structure will be shown all the elements “Drawn” within that
phase. The way in which they are then cataloged must be set, for example the choice
whether to show each single piece used or how in the metric computations filter and
group all the identical pieces under one item. By entering the cost of each element,
the building construction costs are instantaneous. Each entry in the schedules shows
a partial cost of the subcategory and finally a total cost, to which the other items that
influence the cost of the project can be applied, such as the costs of the designer,
VAT, safety percentage. (Fig. 14).
Working with prefabricated elements in the BIM field has made possible to auto-
mate the work of identifying the components of the architecture that with the normal
project software would have required multiple processes of architectural design. The
use of the schedules for the identification of the components of the project is an index-
ing of elements already present at the three-dimensional level and not just metadata
within the project. Designing with wood is a technique as mentioned in the previous
paragraphs used for hundreds of years, standardized from a certain point of view.
The working methodology is not influenced by the planning in a BIM area but by
the possibility of representation and communication clearly superior to today’s stan-
dards. As mentioned above, working in BIM means facing different dimensions of
the design at the same time, not only in terms of geometry but in terms of timing, costs
and environmental impact, this means that every single component, depending on the
degree of detail, has precise properties. By focusing on the representative aspect of
the individual parts of the building, the possibility of realizing the precise schedules
of the components through the tool, greatly simplifies the workflow of the single
parts of the architecture. Examining two case studies used in the BIM model of the
station, one can see the representative gap with traditional representation techniques.
The first case study concerns the Xlam structure of the three-dimensional model,
which is the responsibility of two design groups, the structural and the architectural
ones. The design in xlam envisages working with panels of different dimensions but
standardized by the manufacturer of the aforementioned materials (Fig. 15). The cat-
aloging of the panels as an assembly within the dedicated software used for modeling
allows to have tables of specific elements that include all the features necessary for
processing in the factory. Once all the data have been inserted, the program allows
to create these tables that contain not only the panel with its own geometry drawn
in plan and section, axonometric perspective, but also to attach schedules with the
specific information of the material used. This tool allows to represent the structural
panel according to the orthogonal projections, highlighting not only the geometry
of the panel, but also the drilling points of where the joints will be positioned to be
used on site.
The second case study examined concerns the envelope of the stalls, dealing as
an assembly with this specific element of the project has not only allowed to create
detailed tables aimed at the construction of the project but also to make elaborate
aimed at communicating the construction methodology used. Isolating the element,
not as a component of a generic family, but as an assembly of the project, has made
it possible to create a constructive axonometric that exemplifies not only the wooden
profiles used but also the coupling modes between slats and wooden substructure,
but also between the latter and the load-bearing structure of lamellar wood stalls.
This method of representation, designed as an instruction manual for mounting the
enclosure, has proven to be a winning methodology of representation on several parts
in dealing with: the client, the manufacturer and the company (Sebastian 2011).
Finally, the model can be used for building management and maintenance fol-
lowing the conclusion of the project. In fact, once the structural architectural model
is integrated with the plant engineering part and the transmittance values of each
Fig. 15 Screenshot of the BIM modeling program highlighting the use of assembly. In the Image
the modeling tool is used of one of the xlam panels
material are inserted, the lighting and safety systems used can have in real time the
energy situation of the building and the maintenance status of the building.
The new station was born as a project for the reconstruction of the city of Amatrice
as a primary goal. The desire to revive a new image for the company in the area does
not stop at the single architectural building, but wants to propose a standard, as a
reference image for the company.
The need to reuse the design of the building not in one place but in several parts of
the region has imposed strong design choices and composition of the project’s matrix
file. How can we propose an architectural artifact, designed to resist seismic events
in a specific area with specific terrain, in other places? There were two solutions, one
from the design point of view and one concerning the way in which the project was
drafted.
The illustrated needs have led to dividing the project into three different elements,
not only on a structural level but also on a compositional level. The first consists of
the basement, an element in reinforced concrete, consisting of an overhead beam
system for the building and a concrete parterre for the stalls. This is a constituent
component, designed in such a way as to raise the building from the ground and level
it, incorporating the altitude jumps. The second part is the structural and finishing part
of the building, which is hooked by means of anchoring systems that can be removed
from the base. Anchoring systems are metal hinges designed to make the building
respond to seismic waves. The third element that makes up the intervention are the
stalls, a lamellar wood structure anchored to the base. The stalls are designed as a
modular element, repeatable but also reducible according to the needs of the station
and the lot. These three elements, which in correlation between them make up the
project for Amatrice, are therefore designed to be modified for different types of land
on which the station must be built in adjacent countries. The element of prefabrication
is then found in the elevation component in wood and not in the base part that for
safety and regulatory reasons it was decided to design as a fixed baseboard.
The second solution, implying the way in which the project was drafted, brings
BIM into the field, the cornerstone of all the definitive/executive design development.
As explained in the previous paragraph, the working model has been divided into
specific worksets connected to each other, differentiating not only the phases but the
structural project from the architectural one. In this way, through the use of specific
constraints and parameters it was possible to connect the development of the casing
infillings and internal partitions to the structure. Exemplifying the concept with an
example extraneous to the work area, a reinforced concrete pillar can be used as a
reference. In a BIM design, this pillar is produced by structural engineers in a specific
BIM file, the finishing part instead of the architectural design group. So, the finish
must be imagined as an empty container (architectural column), parameterized and
bound to the space of the pillar, when the structural element varies, the dimensions
of the finish vary. The same principle has been applied in setting the BIM file. The
stalls are also designed as modular elements, equal to each other and repeatable, in
the same Revit project, the setting is such that as one varies, the others also change.
This is because they are constrained and parameterized in such a way as to cause the
planimetric and raised elevation to result harmoniously.
This design mode, in relation to the level of prefabrication of each element not
only allows to lower costs, but reduces the design expenses for future stations, as the
only element that needs to be recalculated is the structural element. In elaborating the
prefabricated elements, it was decided to use, where possible, standard components
already existing on the market and not to provide an excessive addition of costs in
proposing a unique design (Fig. 16).
7 Conclusion
The project examined in this chapter is currently in tender in order to allow it to be

built quickly and reactivate efficiently the bus line station in the area of Amatrice. This
latter, like Ponte Mammolo, due to the strong aesthetic impact of the waiting room,
is a pilot project to develop a network of environmentally sustainable buildings in the
Lazio region, with an observant eye to not design an energy eating but a conscious and
well-defined and recognizable architectural style. A way to demonstrate the serious
change of a company, which is being renewed to promote an excellent service on the
territory.
The possibility offered by COTRAL at the university is a unique opportunity
for the development of internal research within the department, which unfortunately
remain in many cases only theoretical. Being able to experiment with new techniques
of design and representation, allows the formation of new professional skills, both
internal and external to the department. The focal point of the research being dealt
Fig. 16 Three-dimensional images extrapolated from the BIM work model, Perspective South view
and longitudinal perspective section
with is not only in the design of a single product, but in being able to use it, due to
the way in which it was designed in several situations. The final aim of the research
is to highlight the characterizing and new way in working the BIM.
The designers’ obligation is to no longer reason only in producing two-
dimensional executive drawings aimed at describing a three-dimensional idea of
the building; the use of BIM requires an executive three-dimensional model of the
architectural artefact. This model is not obtainable only through the use of specific
software, so called BIM Oriented, but through a broader reasoning in the organiza-
tion of work. In Italy this reasoning is entering the design teams today because the
state is imposing specific laws. But even today there is the incongruous use of BIM
software, using them as CAD drawing tools and not design and collaboration. The
“traumatic” passage that we are experiencing in these years from CAD to BIM, is
a change imposed both for economic and practical reasons, but is similar to what
happened in the twentieth century with the imposition of abandoning the practice
of drawing to free hand to go to the digital one. A transaction that has forced the
class of architects to change the working methods, but also to integrate them and not
necessarily to abandon one for the other.
Acknowledgments Paragraphs 1 and 5 are written by Maria Grazia Cianci, 2 and 6 by Daniele
Calisi. 3 by Francesca di Benedetto, 4 by Matteo Molinari.
References
Blass HJ, Fellmoser P (2004) Design of solid wood panels with cross layers. In: 8th world conference
on Timber Engineering, WTC 2004, Finland
Ceccotti A (2008) New technologies for construction of medium rise buildings in seismic regions:
the xlam case, IABSE structural engineering international, special edition on tall Timber buildings
Cianci MG, Calisi D, Molinari M (2018) Il BIM come strumento di controllo: Recupero delle aree
di stazionamento bus e dei locali passeggeri nei nodi di scambio COTRAL di Ponte Mammolo a
Roma, Atti del convegno 3D Modeling & BIM 2018, Roma, Italia, 2018. ISBN 978 88496194 1
6
Galletti P (2004) Civiltà del legno. Per una storia del legno come materia per costruire dall’antichità
ad oggi. CLUEB, Bologna, Italia
Kymmel W (2008) Building information modeling. McGraw Hill Construction Series
Livingston C (2008) From CAD to BIM: constructing opportunities in architectural education. In:
Ettouney M (ed) AEI 2008 Denver: ASCE
McLaren S (2008) Exploring perceptions and attitudes towards teaching and learning manual tech-
nical drawing in a digital age. Int J Technol Des Educ
Pavan A, Mirachi C, Giani M (2017) BIM: Metodi e strumenti, progettare, costruire e gestire nell’era
digitale. Tecniche Nuove, Milano
Peterson F, Hartmann T, Fruchter R, Fischer M (2011) Teaching construction project management
with BIM support: experience and lessons learned. Autom Constr
Pryce W (2005) L’ architettura del legno. Una storia mondiale. Bolis editore, Bergamo
Rozman M, Fajfar P (2009) Seismic response of a RC frame building designed according to old
and modern practices. Bull earthquake Eng
Sebastian R (2011) Changing roles of the clients, architects and contractors through BIM. Engi-
neering, construction and architectural management
Customizable Social Wooden Pavilions:
A Workflow for the Energy, Emergy
and Perception Optimization in Perugia’s
Parks
Marco Seccaroni and Giulia Pelliccia
Abstract The research aims to generate a workflow, which subdivides the complex
problem of optimizing the buildings energy consumption in smaller problems that can
easier be solved. The workflow starts from the definition of the insertion context of the
building, which influences it principally regarding the climate, the sun exposure and
the shadings. The successive step is choosing one or more optimal wall stratigraphies
which show the best combination of different parameters, like cost, transmittance,
thickness and emergy. The last step concerns the optimization of the shape as a
function of the previously defined stratigraphies and of the energy consumptions for
lighting, heating, cooling and electrical equipment.
Keywords Wall stratigraphy · Energy consumption optimization

Emergy optimization · Context perception
1 Introduction
The proposal comes from a collaboration between the Department of Civil and Envi-
ronmental Engineering of the University of Perugia and the Municipality of Perugia,
in relation to a convention aimed to create wooden structures for citizenship to be
placed in the parks of the town, named “Generative design and modelling for the
optimisation of architectural forms of wooden structures”.
The input derives from the necessity of the Municipality of Perugia to realize
multifunctional spaces in the 92 parks located in its territory. A mass customization
approach ensures a simple replicability of the common characteristics applied in
M. Seccaroni (B) · G. Pelliccia

Department of Civil and Environmental Engineering, University of Perugia, Perugia, Italy
e-mail: marco.seccaroni@gmail.com
G. Pelliccia
e-mail: giulia.pelliccia@outlook.it

1046 M. Seccaroni and G. Pelliccia
different contexts and can bring benefits to an ecological, territorial and economical
point of view, which can be supported and used by the public administration.
The construction sector is generally responsible of the 30% of the carbon footprint,
it consumes over the 40% of the whole energy produced in its lifecycle (OECD 2003)
and, in the same time, it uses more raw materials than any other economical activity.
These considerations prove how the construction sector is unsustainable and needs a
trend reversal. Furthermore, an increase of the world population and the consequent
increasing in the request of buildings and infrastructures could further increment the
consumption of non-renewable materials, as well as the waste production. Thus, the
use of more sustainable construction materials represents an important contribution
to the eco-efficiency of this sector and, consequently, to a more sustainable develop-
ment (Sposito and Scalisi 2017). A similar approach can significantly influence the
reduction of the energy consumption (Thotmark 2006) and of the carbon footprint
(Gonzalez and Navarro 2006).
A physical parameter for the evaluation of the energy necessary for the production
of a specific good or service is the emergy. The emergy synthesis is an environmental
evaluation methodology developed by Howard T. Odum at the University of Florida
in the Eighties. It is based on the thermodynamics and on the theoretical ecology,
with particular reference to the general systems theory of von Bertalanffy (1968) and
to the theory of dissipative structures of Prigogine (1947).
This methodology can highlight the dependence relationships between the natural
ecosystem and the human economy and allows to calculate a set of indicators for the
evaluation of the performances and of the environmental sustainability of a territorial
area or a production process (Franzese et al. 2003b).
The “ecodynamic” methodological approach (Franzese et al. 2003a, c, 2005,
2009) focuses on the systemic properties and determined a paradigm shift (Kuhn
1962) in the environmental sciences, definitely innovative compared to the standards
of neoclassic economy and reduction biology. The concepts of solar emergy and solar
transformity (Odum 1988, 1996), are the basis of that methodology, which aims to
determine the performance, the impact and the environmental sustainability of the
investigated system.
The importance of similar objectives is confirmed by the Directive 2010/31/EU
of the European Parliament and of the Council of 19 May 2010 on the energy per-
formance of buildings (EPBD), which established that, by 31 December 2020, new
buildings must be nearly zero-energy buildings, whereas for new public buildings
the term is set on the 31st of December 2018.
Different kinds of approaches are used to reach the maximum energy efficiency
of the building. Passive design strategies are widely exploited and are becoming
an essential part of the design process: the optimal sun exposure and the natural
ventilation, especially, are two basic principles which influence the project. The
improvement of the energy efficiency is an economic and environmental matter, but
concerns as well the comfort of the occupants. A technological approach can also be
used, based on an I.P.O. model (Input Process Output), which is the starting point of
every artificial intelligence and is composed of sensors, processors and actuators that
automatically and mechanically act on some elements of the building to adapt the
Customizable Social Wooden Pavilions: A Workflow … 1047
indoor conditions to the external ones (Negroponte 1970). However, independently

from all the strategies that can be used, the selection of the materials which compose
the stratigraphy of the walls is one of the preliminary steps that every designer must
deal with. Generally, each kind of structure has typical stratigraphic characterisations
which usually ensure a good energy efficiency of the building.
2 The Optimization Applied to a Real Case
The project is developing by the research group coordinated by Fabio Bianconi

and Marco Filippucci and composed by Elisa Bettollini, Marco Bifulco, Alessandro
Buffi, Lorenzo Ciculi, Michela Meschini and the authors Marco Seccaroni and Giu-
lia Pelliccia. The further collaboration between the research group and the timber
construction company Abitare+, with a Post-graduate Research Scholarship titled
“Parametric representation studies and BIM modelling finalized to the analysis and
the energetic optimization of timber architectures”.
The functional objective that binds to the definition of a meeting place and garri-
son is also associated with the desire to connote the place with a space that gives it
identity by receiving it from those who use it. The architectural form wants to be an
opportunity to rethink the relationship between architecture and community, in the
mutual loss of forms and meanings inherent in the excess and approval of the pro-
cesses of standardization and stereotyping of contemporary culture (Aravena 2007;
Cruz and Boddington 1999). These are architectures also placed in an environmental
context of value for the city, in those spaces with a strong social identity and therefore
also cultural function, places that are designed to promote a new way of building in
harmony with the environment and nature.
From what has been shown, there are three characteristics that want to determine
the architecture of these pavilions: the recognizability of the architectural form as a
denoting element of the action of the City Municipality, the ability to characterize
these spaces as a function of the community to make them their own an finally
sustainability and the low environmental impact of the chosen solutions.
The area of intervention is in the Chico Mendez park of Perugia. The park has
a regular grid of tall trees and carriageable/pedestrian streets, only used as a pedes-
trian transit area to join the Minimetrò stop Cortonese to the receptive poles in via
Cortonese. Inside the park there are a play area for children and some others spaces
used for sport activities, while there are no restoration or accommodation structures.
The project concerns the realization of a wooden pavilion (Fig. 1), destined to
host a local association, and of a green square in front of it. The intervention has
several goals to achieve: the green square aims to realize a rest and socialization area,
as well as a reception structure for the close zones, in a space that was previously
perceived as a passing area. The park will so be strongly linked to the territory and
returns a collective space to the close residents.
The pavilion is composed of a regular plan building and of a porch facing south.
The external wooden coating develops from a stone basement and those natural mate-
Fig. 1 Mock-up of a multifunctional space in a park of the Municipality of Perugia
rials make this kind of structure not impactful in such a natural context. The elevation
aspect is contemporary, in line with the close Minimetrò station of Jean Nouvel. The
internal space is divided in four: a toilet, an open space for the association meetings
and parties, an office and a room, which can be used as an archive. The roof has the
peculiarity to be oriented and inclined in order to be visible from the Minimetrò: it is
possible to install on the roof some advertising sheets to sponsor events or products,
which could also constitute a self-financing system for the pavilion. Photovoltaic
panels would have a very low return, due to the shadows of the surrounding trees.
The current altimetry of the terrain will be preserved with the new external arrange-
ment: instead of the walkway in the park, a pedestrian path will be created, where
benches in oxidized concrete will be placed. The successive design levels consider
and elaborate the informations obtained from the described metaproject approach
(Fig. 2).
The research aims to create new tools to achieve a mass customization of architec-
tural forms. The goal is to create a family of pavilions by drawing their relationships
with the schematized surrounding environment, using the tools of parametric and
generative representation and in particular Rhinoceros, with its visual aid for script-
ing Grasshopper and its evolutionary solver plug-in Octopus, and integrating the
instruments of the different analysis necessary to evaluate the sustainability.
Fig. 2 3D model of the load-bearing structure of the wooden pavilion realized with the software
SEMA
3 From Real to Virtual and Vice Versa
The first step is choosing which elements can be discretized to define the context
of the parks in Perugia, identifying a volume of influence that varies according to
the measures of the pavilion to be designed. This volume is the minimum portion
of space necessary for the construction of the building, considering the context, and
every face of the volume must be characterized and schematized (Fig. 2). The square
base overlaps with the terrain: it will be flat if the ground is flat or inclined by a defined
angle if it is sloping. The side faces define if there are obstacles (trees, walls etc.)
and establishes their entity. A solid with openings generated by the place will then
be obtained. Being in a metaproject state, it is possible to approximate in this way
the reality. For example, the relief was measured by steps instead of using express
tools.
Another parameter to consider is the orientation, as it is possible to rotate the
volume according to the cardinal points. These elements are real measurements that
reflect the place, another important parameter is the perception, or better the space
identification of those valuable elements on which we should maximize the view
(attractors) or the opposite, the detractors, i.e. where the vision should be minimized.
Only with these four parameters, almost an infinite number of solutions could be
created.
Fig. 3 Fractal geometries, as parameters increase there is an increase in branches
Fig. 4 Some of the possible configurations
These solutions can be represented using a fractal logic, instead of using the
classical set theory; n-possible architectural solutions can be draw through the fractal
ramifications. In fact, for example, assuming that the four variables are defined in the
natural numbers, between 1 and 10 values, combining them linearly it will be possible
to get a finite number of solutions (104 10,000). These values may, however, also
increase, with subsequent exponential growth of solutions. In addition, the fractal
provides a dynamic representation that also ensures a progressive accuracy of the
related variables, where the increase of the same follows the same fractal logic in the
connected microcosms. The range increasing of parameters gives a greater amount
of solutions, which is represented in (Fig. 3).
These parameters and, therefore, these geometries (flat surfaces and points) can be
easily transposed in a parametric approach. It goes without saying that it is possible
to recreate a schematized virtual model, comparable to the real one (Fig. 4).
4 Stratigraphy Optimization
The form finding process of the pavilion derived from the optimization of the orien-
tation, which depends on the sun path and on the context, and from the optimization
of the external environment perception, using generative algorithms in order to find
the solution which maximizes the view on the attractors and minimizes the view on
the detractors. In this second phase, the goal is the selection of different insulating
materials, between the most used and diffused and according to the wooden structure
of the pavilion, for the optimization of the energy efficiency, the emergy efficiency
and the costs. The structure was realized with CLT Cross Laminated Timber panels,
which guarantee a quick assembly on site and a safe and antiseismic construction;
furthermore, CLT is a sustainable material and offers a good degree of thermal and
acoustic insulation. The timber construction company Abitare+ generally use a stan-
dard stratigraphy for its buildings, composed of 1 cm plaster, 8 cm XPS, 10 cm
CLT (5 layers), 7 cm glasswool and 2.5 cm drywall, which has a thermal transmit-
tance equal to 0.192 W/m2 K, a cost of 185 e/m2 and 2,05 E + 14 eJ emergy. The
costs were calculated considering loading, transport, unloading and assembly (Prez-
zario Regionale Umbria 2017), while the emergy includes materials, transport and
assembly (Pulselli et al. 2007, 2008, 2009; Keena et al. 2018).
In order to find the best combinations of materials to improve the energy efficiency,
a typical stratigraphy for the external walls was selected, after a comparison between
the most used by the majority of the Italian wooden building constructors: 1 cm plas-
ter, variable external wall insulation systems, 10 cm CLT panel, variable insulations,
variable interior wall finishes. For each material of the stratigraphy, thickness [m],
conductivity [W/m K], density [kg/m3 ] and specific heat [J/kg K] were provided in
order to launch a simulation of the thermal transmittance [W/m2 K] of the final exter-
nal wall with Ladybug and Honeybee (Pak et al. 2013), an open source environmental
plugin which connects Grasshopper with the engines EnergyPlus and OpenStudio,
for the thermal simulations.
For the variable external wall insulation system, a list of five different materials
were considered: XPS, cork board, rockwool and two kinds of wood fibre panel, with
different conductivity and density. For the insulation, four materials were selected:
glasswool, cellulose fibre, rockwool, wood fibre panel. For the interior wall finishes,
at last, drywall or cellulose fibre gypsum board were tested. An optional additional
layer was also considered between the plaster and the external wall insulation sys-
tem, consisting in a BetonWood panel, i.e. a performing material made with wood
processing wastes, iron oxide, water and Portland cement. Keeping the conductivity,
density and specific heat values as a constant for each material, the only variable is
the thickness, which was selected between the commercial ones for a list of producers
(Table 1).
After this first part concerning the documentation and selection of the materi-
als and their thermal properties, the optimization process of the stratigraphy was
conducted with the plugin Octopus for Grasshopper, which allows the application
of evolutionary principles to the parametric modelling, in order to optimize some
Table 1 The materials used for the different stratigraphies, with their properties
Conductivity (W/mK) Density (kg/m3 ) Specific Thickness (m)
heat (J/kg K)
Exterior Wall Plaster 0.075 1600 1000 0.010
Finishes
External Wall Betonwood 0.26 1350 1880 0.010
Insulation 0.016
System 0.022
XPS 0.035 35 1450 0.060
0.100
0.140
0.180
Cork board 0.039 120 1900 0.040
0.006
0.100
0.140
Rockwool 0.039 150 1030 0.020
0.030
0.040
0.050
Wood fibre 0.037 110 2100 0.080
panel 0.140
0.200
0.260
Wood fibre 0.038 160 2100 0.060
panel 0.100
0.140
0.180
Structure 5 Layers CLT 0.11 512 1600 0.100
Insulation Wood fibre 0.038 160 2100 0.060
panel 0.100
0.140
0.180
Glasswool 0.039 13 1030 0.045
0.070
0.095
0.120
Cellulose fibre 0.038 180 2150 0.160
0.180
0.200
0.220
Rockwool 0.035 60 1030 0.040
0.060
0.100
0.140
Interior Wall Drywall 0.2 900 1000 0.0095
Finishes 0.0125
0.015
0.018
Cellulose fibre 0.32 1150 1100 0.0010
gypsum board 0.0125
0.015
0.018
Fig. 5 The stratigraphies analyzed up to the 30th generation
features of the project depending on selected parameters. The thermal transmittance,

the addition of the singular costs of the variable materials (thus, excluding plaster
and CLT) in e/m2 and their emergy were the three goals to minimize, while the
variable parameters to be combined were the different materials that can compose
the stratigraphy (the external wall insulation system, the insulation and the interior
wall finishes) and some of their possible commercial thicknesses. The optimization
process analysed 30 generations in a 3D chart where the x-axis is the thermal trans-
mittance, the y-axis is the cost and the z-axis is the emergy: the closer the solution
is to the origin, the more it is optimized.
Between the 12800 possible combinations of materials and thicknesses, Octopus
analysed 30 generations (Fig. 5) and found the 107 solutions belonging to the Pareto
Front (Schaffer 1985); (Fig. 6), which is a subset of optimized solutions belonging
to the set of all the stratigraphies. These 107 data represent less than the 1% of
the whole possible solutions and excluded the BetonWood panel, the cork board
and the higher density fibre panel. The graphs in Figs. 4 and 5 compare all the
optimized stratigraphies to each other and to the standard stratigraphy used by the
timber construction company, which is highlighted in yellow.
Fig. 6 The 107 best stratigraphies belonging to the Pareto Front
The graph (Figs. 7 and 8) shows that some different configurations are cheaper,
have better U-values and minor emergy than the standard one, with similar or smaller
thickness.
5 Form Finding
The final optimization process aims to make a further selection between the Pareto
Front stratigraphies, in order to have the minor emergy and energy consumption
of the pavilion regarding heating, cooling, lighting and, at the same time, the minor
initial cost. Once defined the context and the optimised stratigraphies, the algorithms
can generate architectural shapes through geometric rules. In this specific case, the
imposed parameters allow to produce a certain number of solutions, starting from a
fixed rectangular plan. These solutions differ from each other for the position of the
windows, for the typology and inclination of the roof and for the orientations of the
whole pavilion. The previously defined stratigraphies can be applied to every shape
generated by the algorithm and for each of them the construction cost, the emergy,
the energy consumptions due to illumination, heating, cooling and the electrical
equipment can be calculated. The perception coefficient can also be evaluated, with
Fig. 7 The different materials and thicknesses of the optimized stratigraphies with their cost, U-
value and emergy
a specific component of HoneyBee, viewAnalysis that allows to exploit attractors

and detractors as inputs. The algorithm dissects the analyzed form with a virtual
plane placed at eye level (1.7 m) and the component divides the surface into a grid
where it evaluates in percentage from how many attractor points each portion of the
discretized surface is visible. The algorithm then calculates the average value of all
the elements of the grid for the different attractors, thus arriving at a synthesis value
determined by the difference between the mean values of attractors and detractors.
Therefore, the final optimized solutions represent the best configurations for a
specific place.
Fig. 8 A 3D chart showing the relationship between costs, thicknesses, U-values and emergy of
the 107 optimized stratigraphies and the standard one used by the timber construction company
In Fig. 9 the whole combinations are represented up to the 30th generation, while
in Fig. 10 the Pareto Front identifies the 120 best pavilions. Among them, 6 solutions
were selected: one is characterised by a high initial cost, medium emergy and low
energy consumptions, another one has low initial cost, high energy consumptions
and low emergy, while the 4 others are closer to the origin. In this way, different kind
of optimized solutions are studied.
The more economically advantageous configuration between these 6 selected
pavilions can be identified in relation to its expected lifecycle. An autonomous natural
gas boiler with a cost of 0.1e/thermal kWh and an electrical equipment with a cost
of 0.2 e/kWh are considered for the electricity expense. The costs of the 6 pavilions
in relation to their expected lifecycle. For example, pavilion 1 can be discarded, as it
has the highest initial cost of construction and at about 25 years it intersects pavilion
6, which has a lower initial cost. If the lifecycle is expected to be longer than 25 years,
pavilions 2, 3, 4 or 5 should be better solutions (Table 2 and Fig. 11).
The algorithm identified the unique following stratigraphy, which is optimized
for the specific context: 1 cm plaster, 18 cm XPS, 10 cm CLT, 22 cm cellulose
fibre and 1.25 cm drywall. This is the solution characterised by the lowest energy
consumptions, in comparison to others which are thinner and cheaper, like the timber
construction company’s one that, on the other hand, has a U-value 2.3 times higher.
Fig. 9 The possible combinations of emergy, energy consumption and initial cost of the pavilion,
depending on the different previously optimized stratigraphies
Fig. 10 The 120 best pavilions of the Pareto Front with the 6 ones selected for the more economi-
cally advantageous configuration analysis
Table 2 The costs in e of the 6 pavilions

Pavilion Cost in 10 years Cost in 20 years Cost in 40 years
1 146.9635 200.2577 306.8461
2 116.2116 170.4454 278.9129
3 122.6598 181.3333 298.6803
4 117.3614 174.7938 634.2535
5 118.0091 174.9383 630.3722
6 125.3388 192.5769 730.4816
Fig. 11 The trend in 40 years of the costs of the 6 selected pavilions, considering the emergy and
the costs of the materials only for the construction phase, while the illumination, heating, cooling
and electrical equipment are calculated up to 40 years
A good compromise could be a cheaper stratigraphy, which has thickness, U-value

and emergy between the two others (Fig. 12).
Furthermore, the optimization process allowed to have a specific distribution and
dimensioning of the openings (Fig. 13), due to the orientation of the construction and
to the presence of attractors and detractors points in the surrounding environment
(Fig. 14).
This research presented a metaproject approach, that aims to constitute an exam-
ple of workflow to follow when dealing with similar simulations and that can be
adapted to the designer necessities, depending on the required parameters. After fol-
lowing the provided guidelines, the project needs to be further developed to reach a
more appreciable architectonical quality. The openings, for example, can be slightly
changed in their dimensions and shape, in order to obtain more regular windows.
Fig. 12 Comparison between three different stratigraphies. Top: one of the best solutions obtained
from the first optimisation process, characterised by the best combination of the four parameters.
Centre: the timber construction company’s one. Bottom: the one obtained from the last optimisation
process, which considered the initial cost and energy and emergy consumptions
The role of the designer remains central in the whole process, especially in the
first part of the project, in the definition of the geometric rules, which determine
the shape and control all the successive steps. The selection of the parameters to be
optimized and the choice of the solutions to be used from the Pareto Front still belong
to the designer, depending on the specific project and surrounding conditions.
This methodological process is based on generative algorithms and can constitute
a DSS (Decision Support System), which therefore helps the designer in the decision-
making process (Elbeltagi et al. 2017; Benis et al. 2017)
Eventually, the project can be detailed through the BIM modelling, in order to
create exportable stratigraphies that can be used in other design situations.
6 Conclusions
The aim of the research was the creation of a specific workflow to divide a com-
plex problem, such as the optimization of the energy consumptions of a building
Fig. 13 The 6 optimized pavilions
Fig. 14 The context with green attractor points and red detractor points
as a function of various parameters, in more confined parts to deal with separately.

All the parameters used can be changed from time to time, depending on what is
requested. The performance of the stratigraphy can be evaluated not only through the
thermal transmittance, but also through others parameters, like the thermal inertia.
Furthermore, the described process should be extended to the definition of the roof
and the floor stratigraphy.
The algorithm could be improved considering different kinds of thermal and elec-
tric systems, in order to evaluate which of them give the smallest energy consumption.
Using renewable energy systems could further give different results.
All the 6 optimized pavilions have a flat roof, because it is characterized by a minor
surface and, therefore, by a minor cost. Considering a photovoltaic system, the roofs
Fig. 15 Render of the project for the park in the Municipality of Perugia
inclinations could be different, as the surface saving of a flat roof, in economic terms,
could be inferior than the saving obtained from photovoltaic panels.
The emergy was calculated only for the construction phase, while it could be
considered also for the energy consumptions during the whole lifecycle of the pavil-
ion. Eventually, a more developed algorithm for the evaluation of the perception
coefficient could be used.
In conclusion, all the previously exposed design indications need to be further
developed both from the technical and the compositional point of view, in order to
reach an architectonically effective and functional shape (Fig. 15).
References
Aravena A (2007) Alejandro Aravena: progettare e costruire. Milano, Electa

Benis K, Reinhart C, Ferrão P (2017) Building-Integrated Agriculture (BIA) in urban contexts:
testing a simulation-based decision support workflow
Cruz T, Boddington A (1999) Architecture of the borderlands. Wiley, London
Directive 2010/31/EU of the European Parliament and of the Council of 19 May 2010 on the energy
performance of buildings
Elbeltagi E, Wefki H, Abdrabou S, Dawood M, Ramzy A (2017) Visualized strategy for predicting
buildings energy consumption during early design stage using parametric analysis. J Build Eng
13:127–136
Franzese PP (2009) 6—anni di Ecodinamica. In Raccolta di articoli in CD-ROM. Supplemento a
Biologi Italiani, A cura di Pier Paolo Franzese. Available online http://www.onb.it/ecodinamica.
jsp
Franzese PP, Scopa A, Riccio A, Barone G (2003a) Studio di sistemi complessi: la prospettiva
ecodinamica in chimica-fisica ambientale. Biologi Italiani 11:39–45
Franzese PP, Dumontet S, Scopa A (2003b) L’analisi emergetica: una metodica termodinamica per
la valutazione della sostenibilità ambientale. In: Biologia Clinica, Ambiente, Sicurezza e Qualità:
obiettivi di una professione che evolve. Proceedings of the XVI international conference of the
association of Italian biologists, Abano Terme, Italy, Dumontet S, Landi E, Pastoni F
Franzese PP, Riccio A, Scopa A (2003c) Valutazione della sostenibilità ambientale: verso un approc-
cio ecodinamico. Biologi Italiani 7:58–60
Franzese PP, Comar V, Russo GF, Ulgiati S (2005) Analisi dei sistemi ecologici: approccio
modellistico-ecodinamico. Biologi Italiani 6:41–48
Gonzalez M, Navarro J (2006) Assessment of the decrease of CO2 emissions in the construction
field through the selection of materials. Build Environ 41:902–909
Keena N, Raugei M, Aly Etman M, Ruan D, Dyson A (2018) Clark’s Crow: a design plugin
to support emergy analysis decision making towards sustainable urban ecologies, Ecological
Modelling. vol. 367(C), Elsevier, 42–57
Kuhn T (1962) The structure of scientific revolutions. University of Chicago Press, Chicago
Negroponte N (1970) Soft architecture machines. The MIT Press, Cambridge
OECD (2003) Environmental sustainable building—challenges and policies Paris, France, p 194
Odum HT (1988) Self organization, transformity and information. Science 242:1132–1139
Odum HT (1996) Environmental accounting: emergy and environmental decision making. Wiley,
New York, p 370
Pak MH, Smith A, Gill GN (2013) Ladybug: a parametric environmental plugin for grasshop-
per to help designers create an environmentally-conscious design. Procedia—Soc Behav Sci
216:948–959
Prezzario Regione Umbria (2017)
Prigogine I (1947) Study of thermodynamics of Irreversible Processes. Wiley, New York
Pulselli RM, Simoncini E, Pulselli FM, Bastianoni S (2007) Emergy analysis ofbuilding manu-
facturing, maintenance and use: em-building indices to evaluate housing sustainability. Energy
Build 39(5):620–628
Pulselli RM, Simoncini E, Ridolfi R, Bastianoni S (2008) Specific emergy of cement and concrete:
an energy-based appraisal of building materials and their transport. Ecol Indic 8(5):647–656
Pulselli RM, Simoncini E, Marchettini N (2009) Energy and emergy basedcost–benefit evaluation
of building envelopes relative to geographical location and climate. Build Environ 44(5):920–928
Schaffer JD (1985) Multiple objective optimization with vector evaluated genetic algorithms.
Genetic algorithms and their applications: proceedings of the first international conference on
genetic algorithms. pp 93–100
Sposito C, Scalisi F (2017) Sustainable architecture: the eco-efficiency earth construction. Eur J
Sustain Dev 6:246–254
Thormark C (2006) The effect of material choice on the total energy need and recycling potential
of a building. Build Environ 41:1019–1026
von Bertalanffy L (1968) General system theory. George Braziller, New York, p 295
Empathic Architecture: Digital
Fabrication and Community
Participation
Hiroto Kobayashi and Don O’Keefe
Abstract This paper examines a new construction method in engineered wood

material, including plywood and LVL (Laminated Veneer Lumber), using computer
numerically controlled routers to build simple buildings in a quick and inexpensive
way. With the method elaborate on here, there is no need to use skilled labor or
sophisticated construction equipment. It provides an effective way of rebuilding in
the wake of natural disasters. The primary innovations of this method are in ease
of construction and transportation by using flat, portable, and durable engineered
wood products, application of the traditional wisdom in wooden carpentry, and the
efficiency of digital fabrication technology. In the case of disaster relief, using this
construction process as a method of organizing community is essential for successful
implementation. The experience of the Great East Japan Earthquake and Tsunami in
March, 2011 in north-east Japan highlights the importance of bringing both technical
and social skills to disaster reconstruction.
Keywords Empathy · Ethics · Engineered wood · CNC machine

Public participation · Mutually-built · Self-built · Disaster reconstruction
Inclusivity
1 Introduction
In the last two decades we have been facing significant difficulties in maintaining
‘ordinary’ and peaceful order of our lives, socially and spatially. Natural disaster,
terrorism, large scale migration, and other unpredictable issues have made it increas-
ingly difficult to foresee the future form of architecture or the city. Instead of planning
decades into the future, we have to be prepared to adapt to what is happening in front
H. Kobayashi (B)
Graduate School of Media and Governance, Keio University, Fujisawa, Japan
e-mail: hiroto@sfc.keio.ac.jp
D. O’Keefe
Graduate School of Design, Harvard University, Cambridge, USA

1064 H. Kobayashi and D. O’Keefe
of us at any moment, and react to new realities spontaneously. We cannot give up

trying to improve our future, and yet we have to accept the inevitability of drastic
change.
Being resolved to adapt to these unpredictable changes, and revise our own idea of
what is ‘normal,’ we also have to ask ourselves: what constitutes contentment in life?
What kind of life goals would we like to pursue collectively; and, how can we make
them happen? Now, we are facing a time in history in which individuals, more than
groups, are starting to express their hopes explicitly. In daily life, and increasingly
through the internet and social media, we express our ideas for our better lives, and
our individual desires. But in order to create a coherent response to the challenges
ahead, we must also recover an ability to hope and act collectively.
Unfortunately, it seems that many of the conventional channels for collective
action are narrowing, in recent times. Political participation has been decreasing
in many developed countries, and though leaders are directly elected, elections are
affected by rising global tendencies linked to populism and immigration that seek
to create walls between abstract groups like “us” and “them.” Many of the political
structures and geographical borders around which we have structured our notions of
self seem to be deteriorating, so we must rely on grassroots action to create a new
sense of community and purpose.
In this indeterminate period, we must have an honest dialogue to establish a clear
direction to a future we are working toward as a society. Narrow and technocratic
responses to challenges of this scale are insufficient. While maintaining empathy for
each individual life and their views, we can still construct a shared vision for the
future of our community as a whole.
2 The Role of Architecture
In light of the uncertainty described above, how should we define the role of archi-
tecture in defining and improving community? One idea is to shift the major focus of
our field from the aesthetic qualities of buildings to their economic, environmental,
and social performance. This would also have implications for how architects, crit-
ics, academics, and even prize committees evaluate architecture. In order to change
architecture, we must change the methods by which architecture is produced, pub-
licized, and evaluated. All of this implies that architectural education will also have
to change.
Empathic Architecture: Digital Fabrication and Community … 1065
2.1 Architecture in Service of Economy, Environment,

and Society
Economically, the role of architecture in the modern city is clear, but the role of
architect is somewhat harder to define. What does the architect add to the economic
equation that is not already offered by contractors, developers, and realtors? Archi-
tects often focus on eliminating or mitigating unwanted externalities of development;
we shape and clad buildings so as to make them sympathetic to the surrounding
environment; we try to make them efficient in their use of materials and energy. But
these respond to perceived social and environmental problems, as we will address
presently.
The question is: should the architect attempt to play a part in the ongoing economic
development of the city, in addition to the spatial manifestation of that development?
We believe so. The growth of the city cannot be considered in isolation from the
movement of the global economy. Industrial development, economic revitalization
and their economic effects are hugely important for how cities and rural areas change.
These structural changes should be considered in tandem with the future demography
of the world population. As internal and external migration increase, larger cities are
becoming the only place where many people can earn money and have a fulfilling
life. It seems continued population concentrations are inevitable, as people seek to
enter the emerging global middle class, but have we studied this problem enough to
know that? Given the scale of modernization and all of the cultural and spatial lenses
through which it is filtered, it is difficult to say.
What is certain is that the shape and direction of the global economy will have
a significant impact on the built environment, and thus on the daily lives of people
around the world. Architects should work with allied professionals, in planning, urban
design, government, media, and business, to help ensure that the built environment
is contributing to the just distribution of resources throughout the built environment.
This means taking a keen interest in economic development, transit, employment,
and public health.
Energy, ecology and technology can be considered as the primary constituents of
environmental practices in architecture, each of which affects and changes our lives,
spatially and physically. Our course, the energy efficiency of buildings with respect
to lighting, heating, cooling, and other mechanical systems is a primary concern.
Recently, more focus has been given the important issue of embodied energy in
architecture. We must continue to measure and work to reduce the energy used in the
production, transportation, and assembly of building materials. This is where tech-
nology comes in; by harness digital fabrication and intelligent logistics services, we
can reduce the impact of architecture on the environment while providing increased
design and construction flexibility.
The social role of architecture has always long been a topic of public debate
in the profession, but we must work to extend the benefits of architecture to all
corners of society, not just those who can afford to commission architects themselves.
Recovery from disastrous situations, be they social or natural, is a pressing need that
architecture can help to address today. Natural disasters have a larger proportional
effect on people in developing countries, on those who do not have resources to
provision for disaster prevention and response (Kreimer 2001; Yamamura 2015).
Developed countries have their own difficulties, as the political effects of migrants
in European cities have demonstrated.
When one considers that climate change will increase the number of natural dis-
asters, and also the number of migrants, then the inseparability of these categories
become obvious (McCarthy et al. 2001). It is then clear that, if we seek to meaning-
fully contribute to the resolution on these problems, architects must simultaneously
address economic, social, and environmental challenges. Ignoring these responsibil-
ities, even in part, will degrade the status of our profession and possibly bring about
catastrophic failure in the future built environment. And only by addressing all of the
above challenges, not only those that are explicitly environmental, can we achieve
true sustainability.
2.2 Reacting to Unpredictable Situations with Architecture
As we have discussed, in order to construct a sustainable future for society, we

must recover a sense of collective imagination and action. However, it has become
imperative to allow space in society for those with unconventional identities and
circumstances to make their individual voices heard. If we are to pursue our hopes
and desires both individually and collectively, then we must overcome the perceived
binary conflict between the individual and the collective. Therefore, action should
be organized in a relatively non-hierarchical way, from the ground up, allowing each
individual to make a contribution suited to their capabilities. A sincere effort based
on honest feeling is the foundation of any critical or speculative attempt to revise our
relationship with architecture and the city.
At least in developed countries, our daily needs are increasingly provided for by
impersonal institutions like the state or large corporations, but we are still obliged to
work together spontaneously in extreme situations. In the case of natural disasters
such as earthquakes, tsunamis, hurricanes, floods, volcanic eruptions, wildfires, etc.,
people commonly come the aid of strangers, and immediately realize that mutual
aid is the best path to preserve the stability and civility of society, even when one’s
individual needs are taken care of.
Architecture has a role to play in the aftermath of natural disasters as well. When
homes, schools, and places of work destroyed in disasters, they also interrupt the
functions that took place their and delay the recovery of the community. It is essen-
tial to prepare alternative places to live, work, and gather quickly after a disaster, but
resources are limited. Because of the lack of availability of skilled labor in these sit-
uations, a self-built construction methodology should be investigated and promoted.
We believe that, like disaster relief coordination and logistics efforts, disaster
relief architecture can be revolutionized with the application of digital technologies.
Recently, structural engineers and architects have begun to explore the power of
iterative computational and parametric design to test large numbers of spatial and
tectonic alternatives. Computers, web-based databases, and social networks have also
enabled designers to create imaginative virtual spaces and speculative proposals that
regularly feature in academic and professional publications. While these techniques
undoubtedly provide many benefits, this change in the process of realizing spatial
ideas has had extensive effects on the prevailing notion of professional responsibility
and agency.
Architecture has, to an extent, become distant from the people it is intended to
serve. Though architects are aware of the most pressing issues of society, they find
it increasingly difficult to break out of the narrow professional channels of special-
ization that they fall into. It has become easier to restrict oneself within a border,
however vague, then to extend oneself into adjacent fields and difficult professional
situations. The barriers to this action are not just disciplinary subgroups like “health
care architecture” or “residential architect,” but also notions of speculative and aca-
demic practice that tend to limit the interaction of the architect with those outside
their immediate professional sphere. Rather than define the role of the architect by
a set of pre-determined capabilities, we should define ourselves based on a set of
issues or challenges in the world at large that we intend to address.
Our notions of the status of the client are as badly in need of revision as those
of the architect. In particular, the client should not be regarded as an individual that
approaches the architect. Architects need to be active in the communities they serve
before a specific commission materializes. This is particularly clear with regard to
the above example of disaster relief architecture.
The predominant form of architectural commission today, like professional spe-
cialization itself, seems to bolster self-defensive notions and a protective sense of
ownership of space. If we seek to create more equitable access to space, and foster a
sense of community and mutual responsibility for its maintenance, we need to create
new ways for clients and community stakeholders to participate in the architectural
process. In particular, we should encourage people use their body to examine how
they can physically contribute to the construction process and, together with others,
to forge a collective sense of ownership of their architecture.
With the above in mind, this paper attempts to present modest but realistic and
impactful ideas to address the above challenges through architecture.
3 Veneer House Project
The Veneer House Project began in the wake of the 2011 Great East Japan Earthquake
and Tsunami. The disaster displaced tens of thousands of people and destroyed
buildings along the coast of northeast Japan (Fig. 1). After witnessing this disaster,
we resolved to find a way to ameliorate some of this damage through architecture.
The rebuilding efforts in the affected area were a priority, and a number of large
scale building and town reconstruction projects were initiated by local and national
level government agencies. Thus, rather than focus our contribution on the design of
large scale buildings, it seemed that a greater impact could be made by helping to
simplify and expedite the construction of small scale buildings.
In the wake of the 2011 disaster, our laboratory at Keio University in Japan began
developing a strong and flexible structural system based on Computer Numerically
Controlled (CNC) routed plywood components. This system allows a structural frame
to be assembled quickly without advanced tools or a prior knowledge of architecture.
Given the abnormally high demand for contractors and construction workers during
the rebuilding efforts after the disaster, this construction system proved helpful in
reducing the time and labor costs. The system also requires less specialized equipment
than conventional building, and large portions of the construction can proceed without
any power tools at all.
Fig. 1 The great East Japan earthquake and tsunami in march, 2011
Moreover, the involvement of final users in the construction process engenders a

close relationship between the user and the architecture, thereby increasing mutual
attachment. There is a promising and inherent potential in the active use of self-built
structures after completion. Creating a sense of ownership over a building is akin to
creating a sense of civic duty for the creation and maintenance of spaces, public and
private.
3.1 Embodiment Design and Self-Built Architecture
The first Veneer House, intended to the test these suppositions, was realized in the
village of Minamisanriku, Miyagi, Japan in 2012 (Fig. 2). Minamisanriku sustained
significant damage in the 2011 disaster, which left the community in need of a place
to gather and recover their shared identity and resolve to overcome the damage. The
program included a meeting space as well as a public bathing facility. Because the
building was an inherently public undertaking, and one that everyone in the town
was aware of and committed to, it was an ideal staging ground for an experiment in
collaborative construction. Because the small, prefabricated plywood components
used in the design were lightweight and wieldy enough to be held by one or two
people, many locals felt comfortable participating in the construction. The simple
assembly process reduced the amount of time needed on site, compared with con-
ventional construction, which also reduced the time burden on citizen participants
to an acceptable level. Though imperfect, this was an auspicious beginning to our
investigations into self-build construction.
In the field of architecture, so-called embodiment design tends to be associated
with self-built methodologies. Encouraging more people to construct buildings by
themselves for recovery projects after natural disasters is one such example of embod-
iment design, as is monodzukuri ‘fabrication.’ In part, this is also a manifestation
of a critical attitude responding to our overreliance on advanced technologies at the
expense of human contact with the construction process, even for those in the pro-
fession. This itself is linked to an uneasiness with the widening reach of technology
in our everyday lives and the acceleration of related social changes.
We should in no way seek to turn back the clock or align ourselves with Luddism.
Even so, we should take such concerns seriously, and examine the effects of the
alienation of the public from the process of construction and design on the experience
of the built environment. The Veneer House Project seeks to implement advanced
fabrication technologies, not to mystify or exclude the public, but to simplify the
construction process and thus open it to everyone.
Fig. 2 First Veneer House in Minamisanriku as a community center for the tsunami victims (2012)
3.2 Logistical and Material Sustainability of Engineered

Wood
In 2013, our Lab realized another structure, the Maeamihama Veneer House, in the
disaster stricken Ishinomaki area of north-east Japan (Fig. 3). The project included
a meeting space and storage area for local fishermen, who themselves constructed
the building in the afternoons after fishing in the morning. For this project, the con-
struction process was illustrated in advance with a manual complete with diagrams.
This functioned as a kind of informal construction documents package, allowing the
fisherman to reference and fully understand the process. The project not only helped
restore the sense of community that was damaged by the disaster, but helped the
fisherman in their daily livelihood.
As described above, these projects functioned in social and economic dimensions,
but they also involved an environmental strategy. The use of digital fabrication tech-
nology and a simplified construction system can reduce waste and mechanical equip-
ment use on site. For an oceanside site like that of the Maeamihama Veneer House,
and for other sensitive sites, this can be an important factor. More importantly, both
projects used engineered wood products made from local forest thinnings. These
projects help promote the resumption of forestry activities in the area, which is a
renewable and sustainable method of material production, if properly managed.
Indeed, we believe material selection to be of primary importance to the Veneer
House environmental strategy. In the future, the relationship between design, con-
Fig. 3 Maeamihama Veneer House, first Veneer House using CNC router to prepare pieces for
assembly and built by local amateur community members (2013)
struction, and local natural resource reserves will become critical in evaluating the
environmental importance of buildings. We must think across multiple scales, includ-
ing about how precious resources like rare earth minerals as well renewable materials
like timber can be most effectively used. Logistical efficiency, a factor of the weight
and proximity of materials, must be considered alongside embodied energy, extrac-
tion costs, and material durability.
Considering the group of concerns listed above, wood is an ideal material for build-
ing construction. Not only is material affordable and renewable, it also sequesters
carbon from the atmosphere and thus combats the effects of climate change. It can
also be produced in many locations, meaning that transport efficiency is likely to
be high. Our team at Keio University uses engineered wood products such as ply-
wood and LVL (Laminated Veneer Lumber) to enhance the usage of timber and
promote forestry as a source of sustainable employment. Furthermore, engineered
wood products can be made using forest thinnings, rather than clear cutting or cutting
of old growth forest as is sometimes required when using natural timbers of large
dimension.
Fig. 4 Wedge-locked joinery in plywood
3.3 Traditional Techniques Simplified with Digital

Technology
For us, the use of wood has also opened a door to another important resource: the
accumulated wisdom of carpenters and craftspeople. The study of wood joinery
systems, in particular, have helped us to increase the efficiency and applicability of
the Veneer House system.
In the earlier Minamisanriku and Maeamihama projects, we used a notch cut
system that, while easy enough to assemble, still required the use of a crane to place
preassembled plywood beams. Furthermore, the shapes for the initial Veneer House
at Minamisanriku, which had to be cut on a table saw rather than a CNC router, did
not take full advantage the flexibility and accuracy of digital fabrication. Though a
CNC router was employed in the Maeamihama Veneer House, it was not used to full
effect.
The next innovation in the development of the Veneer House project was the
introduction of wedge-locked joinery into the system. Using traditional Japanese
joinery techniques as a precedent, our lab at Keio University developed a way to create
rigid assemblies of plywood ribs and structural panels, held together by plywood
wedges (Fig. 4). The updated systems were lighter and easier to assemble than
previous iterations, and a crane was no longer required, even to construct the roof. In
conventional carpentry, joinery is a laborious process involving the hand finishing of
each joint with chisels. Because the Veneer House assemblies are composed only of
flat components, the CNC router can cut the all components needed for construction,
including wedges and complex joints. Additionally, engineered wood products have
fine tolerances and, due to cross lamination of layers of wood grain, they resist
bending due to weather exposure.
The resulting system was robust enough to function as a complete structural frame,
and yet simple enough to allow anyone to participate in construction (Fig. 5). In fact,
the assembly of the revised structural frame does not require the use of glue, nails,
or power tools of any kind. The modular design of panels, constrained as it is by the
standard dimensions of plywood sheets, also ensures that components are portable
and wieldy (Fig. 6).
Fig. 5 Collaborative construction work
Fig. 6 Panel layout for Cogon Day School

Fig. 7 The Cogon Day School, the first Veneer House project to fully employ digital fabrication
system (2014)
The Cogon Day School, built in 2014 on Bohol Island in the Philippines, was the
first Veneer House project to fully employ this digital fabrication system (Figs. 7 and
8). The school building is a first for the small village of 700 people. The village was
affected by a severe 2013 Bohol earthquake and Typhoon Haiyan, and the project
helped to catalyze their recovery.
The structural frame was assembled in the course of only two days, including
workshops for local student who would later use the school. Some participants we
as young as kindergarten age, but all could participate in the workshop with the
aid of scale models and mockups (Fig. 9). Older children and other adults from
around the village helped to complete the structural frame and to clad the exterior in
local materials. By involving students and their parents in the construction and even
the design of the structure they would later inhabit, the project forged an intimate
relationship between building and user.
Based on our experience in Japan, we believed that broad community participation
would create a sense of ownership and agency that would strengthen the will of the
community to maintain the building and perhaps go on to collectively address other
needs. In particular, we felt that the experience would be a lasting one for the children,
and perhaps have some long term impact on them.
Fig. 8 CNC router at Bohol Island State University used by Keio University students
For the Cogon Veneer House, fabrication as well as construction was carried out
locally. Components were CNC routed at a fabrication lab at Bohol Island State Uni-
versity using locally produced plywood sheets. Only data was prepared in Japan and
sent to the Philippines, and no materials had to be shipped internationally, illustrating
the logistic efficiency the system can achieve.
Because of the ease and speed with which the Veneer House system can be imple-
mented, and because of its ambition to stitch torn communities together, disaster
relief applications became a main focus of our work. The Manawhari Learning Cen-
ter (2013) addressed flooding conditions in the rural Ayeyarwady region of Myanmar
(Fig. 10). The building comprises a flexible classroom space and a veranda for the
use of local children, many of who also participated in construction. The Charikot
Veneer House (2015) was built in Nepal in response to severe earthquakes occurring
earlier that year (Fig. 11). The flexible and lightweight wood construction system
performs better in seismic events than the static and heavy stone masonry buildings
common in the region.
Fig. 9 Kids’ workshop in Cogon village
3.4 Assemble/Disassemble; Flexibility and Adaptability
In addition to ease of assembly, the Veneer House systems also allows for easy
disassembly. In the Charikot Veneer House, the entire structural frame was assembled
in a factory as a test, before being dismantled and moved to the site to be assembled
once again (Fig. 12).
More recently, this ease of disassembly has been exploited in a number of tem-
porary pavilions most notably the Veneer Beach Houses. Every summer in Japan,
umi-no-i.e., or beach houses, are erected along many popular beaches in Japan. These
are used as restaurants, music venues, or other temporary venues for the summer fes-
tivities. in 2015, we completed the Veneer Beach House in Shichigahama, Miyagi
Prefecture, Japan (Fig. 13). Shichigahama beach, formerly a major tourist attraction,
is located in the region most affected by the 2011 earthquake and tsunami. This
Beach House serves as a venue for concerts and other festivities, helping to bring
vitality back to the local community.
The second Veneer Beach House, built in 2017 in Higashihama near Enoshima
Island, Japan, houses a clinic and office for lifeguards, as well as a temporary radio
station (Fig. 14). The Veneer House system makes it easy to assemble and disassemble
the structures each summer, and store the flat components in a small space during
the rest of the year. Additionally, because there are no nails are screws used in
the assembly, there is no chance of losing nails in the sand and endangering beach
visitors.
Fig. 10 Manawhari Learning Center in Myanmar, where local building technique is applied to
exterior finish to maintain their local architectural culture (2013)
With these beach houses, the number of kinds of pieces required for assembly
is less than ten, which greatly simplifies and accelerates the construction process
(Fig. 15). Compared with the first Minamisanriku Veneer House which required more
than 100 kinds of pieces, the beach house assembly is less complex and is proceeds
with fewer mistakes. Veneer House types with smaller number of components and
fewer types of components are most suitable for applications in which the structure
will be dismantled and reassembled several times. These components can be easily
replaced or even reused in other ways.
3.5 Developing the Joint System to Accommodate Local

Cladding
The Veneer Beach Houses also include a flush joinery system, the latest development
in the progression of the Veneer House technology, which leaves the exterior finished
surface of the structural frame smooth (Fig. 16). This allows exterior cladding to be
fixed to the frame without the use of an additional substructure. In the case of the
Fig. 11 The Charikot Veneer House in Nepal for disaster relief of earthquake (2015)
Veneer Beach House, a large custom tarp is stretched over the extent of the frame,
providing adequate weather protection in the summer months.
With the flush joint, other forms of membranes can easily be fixed on the exterior
of the structural frame, including regionally sourced materials like amakan woven
bamboo used earlier Veneer Houses in Myanmar and the Philippines. Although the
underlying structural technology and joint details can be applied to structural frames
of various shapes and dimensions, we are conscious that the Veneer House system
still represents a new and thus unusual or foreign construction method for people in
most contexts.
Fig. 12 Pre-assembly in plywood factory
Making provisions for the application of local cladding materials helps to integrate
these structures into varied contexts. We benefit by learning from the accumulated
knowledge of the community how best to protect the structure from local conditions,
and we also believe this helps the community develop a relationship with the building.
Using local materials as an exterior finish also allows local people to maintain the
building continuously by themselves, protecting the comparatively durable veneer
structure on the interior. This hybridization of local techniques and global technology
is one promising direction in the future architecture.
3.6 Agile Architecture
The agility of the Veneer House system, both in production and construction, is one of
its principle strengths. Given the pace of change in contemporary living and working
conditions, we feel that Veneer House can make a contribution to a wider range of
situations than we have, or even could, anticipate. Even the notion of permanent
architectural solutions may come to seem outdated, as temporary and flexible forms
of housing are desired or demanded by circumstance.
Fig. 13 Framework of Shichigahama Veneer House (2016)
Economic, social, and environmental problems will continue to be those we are

most interested in addressing. With climate change, natural disaster, and the refugee
crisis continuing into the foreseeable future, there may be more need than ever for
temporary and flexible architectural solutions.
At the same time, we have attempted to experiment with the Veneer House system
in entirely new contexts. As part of an ongoing relationship between Keio Univer-
sity and the Polytechnic University of Milan, we have constructed pavilions in Vis,
Croatia (Fig. 17) and Slovenj Gradec, Slovenia (Fig. 18). Each has been a chance for
students from both countries to interact with local craftspeople and citizens, making
the construction and use of each pavilion a change for cultural interchange. It has also
been a chance for us to experiment with structural system itself; the open, column
grid system in the Slovenj Gradec Veneer House does not require shear walls.
At an even smaller scale, our lab at Keio University has developed a series of
kiosks and booths used for both temporary and permanent installation. We erected a
temporary kiosk in the earthquake stricken Japanese city of Kumamoto, which was
used for a local event after the disaster (Fig. 19). We have also deployed temporary
structures at beaches for summer festivities, and in various exhibitions and industry
events dealing with wooden architecture and digital fabrication. Finally, we have
created a series of interior booths that can be assembled inside offices to create private
spaces for conversation or individual work (Fig. 20). We believe the flexibility of the
Fig. 14 Interior view of No-nail Veneer Beach house in Higashihama, Fujisawa (2017)
system makes it ideal for augmenting the ever changing interior landscape of the
contemporary, open plan office. In each of these applications, the participation of the
users in the construction process is an integral part of the system. Also, by distributing
these booths kits in the marketplace and keeping a large stock of kits in storage, we
hope to be able to offer as many temporary rest spaces as possible in the case of an
unpredicted disasters.
As Veneer House technology continues to develop, the quantity and variety of its
potential applications grows apace. The agility of the technology helps it to adapt
and grow along with the changing demand and conditions we have come to expect.
What is most crucial, we believe, is maintaining an empathetic stance toward the
built environment and the communities that inhabit it. Only by first focusing on
the underlying social, economic, and environmental challenges before use can we
develop and apply this technology to its highest form.
Fig. 15 6 key components of no-nail Veneer Beach house
Fig. 16 Original flush joinery system

Fig. 17 Vis Veneer Tea house in Vis, Croatia (2016)
Fig. 18 Slovenj Gradec Veneer pavilion in Slovenia (2017)

Fig. 19 Kumamoto Veneer House is used as a secondhand kids’ bookshop (2017)
4 Conclusion
In the history of Japanese architecture and construction, people worked together to

build each community member’s house in a village in turn. This was a collaborative
working system based on mutual aid in a small and autonomous community called
‘Yui.’ By this method, people could complete a large amount of construction work
that could not be done alone. This type of working system promotes mutual under-
standing and respect among the community members and sense of ownership of the
community itself through collaboration.
The Veneer House building system tries to realize a contemporary version of
Yui by providing a simple building method by which any and everyone can work
together to provide a place for themselves. Is has been our observation that self-built
structures create a strong sense of unity among the participants, a sentiment that
extends to the building itself. Committing to build something collaboratively fosters
a sense of ownership for each person and collaborative work can forge a new sense
of unity in a community recovering from a calamity. It is the ambition of the Veneer
House project to simultaneously promote both a notion of a community around a
building, and an individual sense of commitment to it.
Fig. 20 Meeting booth (2.3 m × 2.3 m × 2.3m)
Mutual understanding occurs not only between community members, but also
between the local people of the construction site and us. Using local materials and
techniques helps us understand the different cultural backgrounds we encounter, and
create a sense of empathy that extends beyond any one culture, time, or place.
Bibliography
de Waal F (2017) The age of empathy: nature’s lessons for a kinder society, no 6. Kinokuniyashoten,
p 128
Kobayashi H et al (2017) Rethinking resilience, adaptation and transformation in a time of change,
1st edn, no 1. Springer, pp 365–385
Kreimer A (2001) Social and economic impacts of natural disasters. Int Geol Rev 43(5), pp 401–405.
https://doi.org/10.1080/00206810109465021
McCarthy, JJ et al (2001) Impacts, adaptation and vulnerability. Contribution of working group II
to the third assessment report of the IPCC, Intergovernmental Panel on Climate Change. www.
ipcc.ch/ipccreports/tar/wg2/index.php?idp=450
Yamamura E (2015) The impact of natural disasters on income inequality: analysis using panel data
during the period 1970 to 2004. Int Econ J, 29(3):359–374. https://doi.org/10.1080/10168737.
2015.1020323
Non-orthogonal Light Timber Frame
Design: Using Digital Manufacturing
Technologies to Facilitate Circular
Economy Architecture
Gerard Finch and Guy Marriage
Abstract Orthogonal structural timber framing is the predominant method for

building low density residential buildings in a large proportion of developed coun-
tries. Today this framing system is highly refined to be economically advantageous—-
making use of low-value and widely available materials. However, this construction
product largely ignores the emerging ‘Circular Economy’ (CE) sustainability agenda.
At the end of a buildings life, and when deconstruction is attempted, most mate-
rials integrated into an orthogonal frame are irreversibly damaged. Furthermore,
deconstruction is time consuming and yields very few valuable materials. Thus, this
research questions the suitability of conventional framing methods to achieve true
life-cycle sustainability and suggests a series of radical non-orthogonal solutions in
response. These solutions are centered around maximizing the recovery of all mate-
rials attached to (and located in) the structural frame at the end of a buildings life.
Non-orthogonal frames are the superior solution as they are generally inherently
resistant to lateral loads and can be dynamically modulated to fit within many dif-
ferent building conditions. The research uses computer-aided fabrication technology
to integrate jointing and assembly conditions in the non-orthogonal timber frame
geometry that substantially speeds up end-of-life deconstruction.
Keywords Sustainable design · Non-orthogonal structure

Circular economy construction · Timber frame
1 Introduction
This chapter investigates the advantages and potential limitations of non-orthogonal

timber frame solutions to combat life-cycle construction waste. The role of computer-
aided design and fabrication tools for timber is a critical component of this argument
as such tools facilitate the making of these sophisticated non-orthogonal structural
G. Finch (B) · G. Marriage

Victoria University of Wellington, Wellington, New Zealand
e-mail: ged.finch@vuw.ac.nz
1088 G. Finch and G. Marriage
geometries. This chapter gives a background to traditional, orthogonal, light timber

frame construction systems and the waste problem they represent. Conditions of a
superior product are suggested in the framework of a Circular Economy (CE) and
existing non-orthogonal solutions are illustrated. Learning from these solutions an
innovative diagrid structural frame is proposed. Salient design decisions are discussed
in terms of fabrication, integration, material and Circular Economy (waste-free)
performance. Finally, the discussion outlines key challenges that must be overcome
if non-orthogonal timber framing solutions are to be adopted by the mainstream
building industry.
2 Background
2.1 Platform Framing and Waste
Platform light timber framed (LTF) construction is the most popular building method
for residential dwellings in the United States of America, Canada, Australia and
New Zealand (Alderman 2013; Curtis 2015). In these countries, traditional platform
framing methods that became favored in the 1950s continue to be used extensively
today (McKeever and Phelps 1994). Platform framing encompasses a timber bottom
plate, vertical load bearing members, horizontal blocking and a top plate (Fig. 1).
These components are nailed together to form a gravity load resisting frame. The
structural members of this frame are strictly orthogonal in the sense that they meet
at 90-degree angles. Platform LTF also makes use of a fast growing and quickly
replenishable natural resource to keep costs down. The translation of the raw material
into a useful product is straightforward and inexpensive. Timber logs are sawn into the
desired cross-sectional dimension and dressed for use on construction sites. For added
resilience this cut material is pressure treated and dried before leaving the sawmill.
Once on site, lengths of timber can be carried, cut and fixed by hand using simple
tools. Collectively these factors ensure the continued widespread use of platform
light timber framed construction.
Platform framing replaced balloon framing as the availability of longer timber
members reduced. This transition from balloon to platform framing is just one of
many modifications that have been made to ensure LTF remains an economically
and technically attractive construction solution. The chemical treatment of all, or a
selection of timber members (depending on country), is now mandatory (Ramage
et al. 2017; DoBH 2012). This modification reflects the increased use of fast-growing
softwood trees. Faster growing timber has a more open cell structure that is less
resilient to moisture and is therefore more prone to rot and fungal attack (Ramage et al.
2017). The layers we fix to the structural frame have also changed substantially. Tar-
based building paper has been replaced by sophisticated plastic composite building
wraps. In some instances, this fabric wrap has been replaced with rigid-air barriers
(RAB) made of thin plywood, oriented strand board (OSB) or fibre-cement sheeting.
Non-orthogonal Light Timber Frame Design … 1089
Fig. 1 Conventional platform light timber framing assembled (left) and disassembled (right)
Adhesives, silicone sealant and waterproof tapes are also widely used to achieve a
waterproof and airtight building envelope. Increasing energy efficiency requirements
has also led to the mandatory inclusion of insulation. To achieve higher energy saving
performance levels (such as those stipulated by Passive House) the dimensions of
structural timber members are now often increased to allow for thicker insulation.
Fibrous batt, foamed sheet, and/or ‘blown-in’ insulation products are all widely
available to fill this void space. Alternatively, the width of the orthogonal structure
can remain unchanged and rigid polystyrene sheets can be screwed and glued to
the external surface of the frame to provide this added insulation. Also contributing
to the overall increased envelope (wall) thickness of a modern structure is the (now
common) addition of an external drained and ventilated cavity. This prevents moisture
from tracking directly through the cladding and into the structural frame. As such,
although we are utilizing the same basic structural arrangement that we did almost
70 years ago, the collective assembly is vastly modified.
With these modifications comes major waste and environmental concerns. Where
historically materials were predominantly high-value untreated and unprocessed
products, today most of the materials that we specify are low-value composite prod-
ucts fixed in ways that makes recovery or recycling impossible. An example of these
increased waste problems is the bracing of the structural frame against lateral loads
such as wind and earthquakes. Platform framing was once braced, kept true and
square, through diagonally located timber members. These were lengths of timber
rebated into the stud and blocking members and fixed with nails at every structural
intersection. This technique has been made redundant by the construction industries
adoption of sheet material bracing. Although sheet bracing is an intelligent modifi-
cation—using an already specified sheet product (wall-lining/sarking) as a bracing
element—such an approach almost guarantees that both the structure and the lin-
ing materials are irreversibly damaged when material separation at end-of-life is
attempted. Plasterboard is the predominant sheet material used to form this bracing
element. In New Zealand this construction approach can be found in more than 85%
of all new residential buildings (Curtis 2015). The material, regardless of its use
as a bracing component or not, is screwed/nailed and glued to the structural frame.
The joints between plasterboard sheets are filled with plaster and then painted so
to form a continuous monolithic internal finish. This material is all but destroyed
when separation is attempted. Mixed with adhesives, plaster, jointing tape, paint and
screws/nails this material is both awkward and expensive to recycle.
Waste plasterboard, together with other construction materials, makes up 40% of
all the world’s waste (EPA 2014). 39% of this construction and demolition (C&D)
waste can be attributed to solid timber and wood fibre products (Forbes 2018; Inglis
2009). These are unprecedented volumes of waste production with far reaching con-
sequences. Treated timber (such as chromated copper arsenate (CCA)), if improperly
disposed of, can leach chemicals into the environment “in concentrations 500 times
higher than safe background levels” (Finch 2018; Parisio 2006). Although we have
alluded to modern construction methods as the cause of these waste problems further
clarification is required. Is the waste a product of a specific feature of a specific part
of modern construction? Or is it a generalised issue, one that is a result of many poor
decisions?
This study identifies the structural core and the associated fixings of a given
construction system as the key to preventing waste by unlocking the potential for
direct material reuse. Due to the structural systems role in supporting the building
envelope a significant portion of all fixings in a building are to these structural
members. The way in which the structure accepts fixings will dictate if materials,
including the structural material itself, can be effectively recovered at the end of
its useful life. If the structure inherently integrates a reversible fixing system, it is
likely that more materials will be recoverable. Furthermore, the structural system
often dictates the modulation of materials within a building. Waste at the time of
construction is almost entirely a result of the supplied material clashing with the
spatial specifications. At the end of a buildings life modulation can also result in waste
even if the structure facilitates entirely uncompromised material recovery. Unusual
modulations, segments designed for a specific use case or inconsistent divisions of
space, all result in material units that are unlikely to have an attractive reuse value.
This argument is supported by historic material reuse situations. Clay bricks have
endured, in a large part, because of their universal adaptability (Webster and Costello
2005). Every module is the same and therefore old and new can be merged to form
an effective system. Again, however, the modern adoption of rigid cement-based
mortars over more flexible and removable lime-based mortars, now restrict the reuse
of this historic material. In this sense the modular, fixing and structural conditions
of any structural system are fundamental to advance lifecycle material management
in the building industry.
2.2 Fixing the Problem
The pre-eminent waste elimination strategy today is referred to as ‘a Circular Econ-

omy’ (CE). In this framework all elements of a product, e.g. a wall assembly, are
designed with the capacity/intent to be without waste in all stages of their lifecycle.
This means that the elements that make up a given product must be able to be recy-
cled without a loss of value, reused directly or reintegrated into a natural biological
cycle without harm to organic or human systems. Circular Economy ideas translate
into a set of criteria for building and construction and these can be used to measure
the waste-free performance of a given construction system (Table 1). These criteria
also reinforce necessary design features that ensure material recovery. The failure of
conventional platform framing in a CE is underscored by a comparison between it
and these CE criteria.
Table 1 Construction circular economy criteria and performance of selected systems

Circular economy criterium Platform framing
Mass standardisation of components to improve direct reuse economic ✓
viability
Geometry enables the independent layering of materials to facilitate easy
separation
Fixings are exposed or have a sense of inherent logic to facilitate separation
Fixings require no tools or are standardised to require only non-specialist
tools
Fixings are detailed to be easily removed/reversed without contaminating or
inflicting irreversible damage to the primary materials
Materials are detailed as to avoid the need for compromising bonded
coatings or treatments
Materials remain aesthetically desirable/can be easily restored to ‘as-new’
conditions
The use of composite/compromised or inseparable hybrid materials must be
eliminated
No chemical, composite or adhesive-based fixings can be used
Materials retain performance through multiple reuse cycles or easily
facilitate end-of-life management with minimal loss of value
Modules and material components are easily reconstituted or formed in a
way that is economically attractive to reuse
Deployed materials are scalable based on a minimum divisible unit
dimension
Platform LTF achieves only one of the twelve criteria for a construction ‘product’
suitable for a Circular Economy. And notably, this ‘modularity’ achievement is not
a result of a will to ensure material reuse but instead financial drivers that call for
the ‘mass standardisation of components’ to make production economically attrac-
tive. By coincidence this standardisation also helps to contribute to the ‘improved
direct reuse economic viability’ of the products materials. Furthermore, achieving
in one specific CE category does not represent CE success. To ultimately achieve
a successful CE product significantly more compliance is necessary. In the case of
platform LTF the advantage of efficient modulation is compromised by single use
fixings between the structural timber members that cause critical damage on removal.
The lack of capacity for LTF to support the reversible fixing of other layers in the
building envelope is also compromising. Similarly, the lack of inherent lateral load
resisting capacity in platform framing complicates the potential reversible fixing of
these supplementary layers. In platform LTF massively standardized components
are also often compromised in situ due to a need to add an opening, wall junction
or architectural feature. Each modification results in a non-standard modification to
all materials in the envelope. This is because the structural configuration demands
continuous ‘full height’ lengths of timber. In this sense although platform framing
does use massively standardized components the ‘deployed materials’ are not always
‘scaled based on a minimum divisible unit dimension’.
2.3 Examples of Best Practice
There are a range of existing sophisticated timber frame solutions that address many
of the issues raised by the Circular Economy criteria for platform framing. Click-
Raft, an experimental, lightweight, computer numerically controlled (CNC) router
cut plywood construction system represents a significant performance improvement
over platform LTF (Fig. 2). Click-Raft is a wall and flooring structural product
developed by New Zealand Architect Moller (2016). The system has been used
in a small selection of residential buildings as structural flooring and wall framing.
Prefabrication of pods has also meant the use of Click-Raft in commercial instances
for enclosed workspaces. What makes Click-Raft unique is the way in which it
forms a structural diaphragm. The Click-Raft system puts “plywood under stress to
create tension and compression in both planes.” (Marriage 2016). This means that the
solution does not need a sheet material rigidly fixed to its face to achieve lateral load
resistance. This simplifies the wall envelope by ensuring each layer can perform its
necessary role independently of other material. Moller’s system is assembled using
slotted plywood-only joints at regular intervals down the length of each ‘click-leaf’.
This assembly method, unlike platform framing, ensures that the structure can be
separated without risk of damage to the principal materials. Click-Raft’s geometry
helps to lock each joint in place and gains strength and inherent self-bracing rigidity
by stressing the members into sinusoidal curves that have both a structural logic
and a graceful appearance. The jointing condition, the fact that the structure can
Fig. 2 Click-Raft—a plywood light timber framing solution developed by Chris Moller
be used vertically or horizontally, and the inherently lateral load resisting structural
arrangement results in a structural solution appropriate for the circular economy.
The key characteristic that differentiates Click-Raft from Platform LTF is its
non-orthogonal geometry. Click-Raft deliberately exploits the inherently structurally
stable advantages of this alternative formal arrangement. Interestingly, another frame
solution designed specifically for the circular economy, dubbed WonderFrame® ,
also exploits a non-orthogonal geometry for structural integrity (McDonough 2016)
(Fig. 3). This highly modular lightweight frame was developed by McDonough and
Associates for the ICEHouse® (Innovation in the Circular Economy). WF® can act as
structural walls or span horizontally for roof/flooring elements just like Click-Raft.
The frame is modulated across a linear grid with an intersecting triangulated space
frame element. The angular components are fundamental to a standardised unit of
WF® hence our definition of the system as a non-orthogonal structure. Although the
only built example of WF® is metal, the intention of the designers is to use “locally
available materials to make cost effective components that can be assembled into
safe, secure, expandable and attractive homes” (McDonough 2016). The triangulated
geometry is bolted together with a single fixing at every intersection and due to the
intrinsic stability of the structure can be disassembled from any point in situ. It is not
clear however how this point fixing translates into a ‘local material’ such as timber.
Fig. 3 Authors scale model of WonderFrame® —a three dimensional triangulated grid frame orig-
inally design in steel by McDonough and Associates
Click-Raft and WonderFrame® are examples of the potential improved circular

economy performance found in non-orthogonal frames. In both instances layers of a
building envelope are freed from traditional dependencies. The resulting distinctive,
separate, building layers promote rapid material recovery at the end of a building’s
life. Both solutions rely on prefabricated members with intricate in-built jointing
parameters. These parameters are only made possible thanks to the onset of widely
available, low cost, computer aided design and fabrication processes (CAD/CAM).
The rapid deconstruction of Click-Raft is principally facilitated by the precisely
cut timber slotted connections and the reliability of this accurately formed joint.
In this capacity digital wood design and fabrication has unlocked affordable and
advanced structural solutions that have the potential to address key waste concerns
in the construction industry. The ability for CAD/CAM to deliver timber elements en-
mass with inherent assembly features can thus work to eliminate the toxic composite
materials widely used in today’s building industry.
The Circular Economy successes of Click-Raft and WonderFrame® can be mea-
sured by the criteria set-out at the beginning of this section (Tables 1 and 2). This
evaluation demonstrates the outstanding performance of these two non-orthogonal
systems in respect to ensuring material reuse at end-of-life and eliminating any poten-
tial lifecycle waste. There are two noteworthy omissions however that could impact
the material recovery potential of either system. The first concern is the lack of inte-
grated capacity for the reversible fixing of building linings. Click-Raft’s non-linear
geometry and slender members make the fixing of linings and claddings difficult
with an additional element required to enable this connection. WonderFrame® is not
limited in this instance as each structural connection point also has the potential to
adhere linings. However, there is no clear methodology to date for connecting tra-
ditional timber weatherboard/batten linings to this large triangulated geometry and
it is likely that additional members would also be required. The second concern for
Click-Raft is its reliance on ‘full-length’ components. Any openings, windows or
doorways require the full-height components to be cut to size while also needing to
have a beam or lintel installed above the opening to transfer any loads from roofing
or flooring elements above. The discontinuous geometry found in WonderFrame®
however allows windows, doorways and other openings to be articulated throughout
the structure without the need for additional components or larger spanning mem-
bers. Total triangulation means that Greater levels of potential flexibility found in the
WonderFrame® system (due to its discontinuous non-orthogonal geometry) arguably
makes it the ideal CE solution.
3 Non-orthogonal Timber Frame Design Experimentation
Circular Economy analysis of WonderFrame® and Click-Raft suggest that existing

non-orthogonal systems vastly out-perform conventional platform framing. How-
ever, this analysis also reveals that current non-orthogonal timber solutions have
limitations that prevent total integration into a Circular Economy framework. The
ability to attach linings in a reversible way without damage or introducing auxiliary
components is necessary requirement that currently eludes Click-Raft. Likewise,
achieving discontinuous geometric modulation that facilitates openings is a complex
but necessary next step to achieve CE success in a timber system. In response to
these shortfalls a series of iterative design-led research experiments were undertaken
(Fig. 4). These experiments explored a range of discontinuous non-orthogonal struc-
tural frames utilising timber-only connections. In this study non-orthogonal frames
are defined as a structural system with an integral element of the assembly at an angle
other than 90° to an associated component.
Table 2 Construction circular economy criteria and performance of selected systems

Platform Click-Raft WonderFrame®
framing
Mass standardisation of components to improve ✓ ✓ ✓
direct reuse economic viability
Geometry enables the independent layering of ✓
materials to facilitate easy separation
Fixings must be exposed or have a sense of ✓ ✓
inherent logic to facilitate separation
Fixings require no tools or are standardised to ✓ ✓
require only non-specialist tools
Fixings are detailed to be easily removed/reversed ✓ ✓
without contaminating or inflicting irreversible
damage to the primary materials
Materials are detailed as to avoid the need for ✓
compromising bonded coatings or treatments
Materials remain aesthetically desirable/can be ✓ ✓
easily restored to ‘as-new’ conditions
The use of composite/compromised or ✓
inseparable hybrid materials must be eliminated
No chemical, composite or adhesive-based fixings ✓
can be used
Materials retain performance through multiple ✓ ✓
reuse cycles or easily facilitate end-of-life
management with minimal loss of value
Modules and material components are easily ✓ ✓
reconstituted or formed in a way that is
economically attractive to reuse
Deployed materials are scalable based on a ✓
minimum divisible unit dimension in all directions
3.1 Conceptual Design
The range of structural frame design variations possible when orthogonal constraints
are rejected is extensive. Figure 4 documents just a small selection of these non-
orthogonal frame solutions that assemble using only integrated timber joints. All
systems have inherent lateral load resistance due to their triangulation and are mod-
ulated across a divisible unit that allows openings to be introduced without compro-
mising components. The systems tested a range of slotted, mortise and tenon, pin
and compression connections—as well as geometric patterns that made the physi-
cal assembly process smooth. Experimentation highlighted the increased levels of
complexity present in non-orthogonal structures that are not found in liner, orthog-
onal systems that could in-fact act as a barrier to deconstruction. For this reason,
experiments aimed to identify geometric conditions with inherent assembly logic, as
Fig. 4 Conceptual non-orthogonal design experiments. Clockwise from top left: ‘X-Frame 1’
900 mm square module. ‘Pattern Frame’ 700 mm triangular module. ‘Reciprocal Frame’ 1200 mm
hexagonal module. ‘Click-Raft Cladding’ 400 mm non-linear module. ‘Reciprocal Span’ 1200 mm
hexagonal module. ‘Braced Frame’ 800 mm square module
few individual components as possible and conditions that prevented the accidental
incorrect assembly of elements.
Top performing conceptual design iterations included the X-Frame solution. This
was largely successful for the very few number of differentiated components that were
required to assemble a section of structure. Like WonderFrame® this triangulated grid
allowed components to be assembled and disassembled in situ and created a repeating
grid for fastening linings. X-Frame was less successful when orientated horizontally
as the joint between the discontinues members was prone to large amounts of deflec-
tion. Both vertical and horizontal spanning is desirable for a CE building solution as
it minimises the quantity of differently formed components. Reduced quantities of
varying components then translate into simplified construction systems that are easy
to reuse and modulate throughout a building. The amount of defection expected in
a structural system was estimated using structural simulation software and though
the building of scale models. Reciprocal Frame and Span iterations where intended
to eliminate issues of deflection when orientated horizontally however the proposed
connections between elements were incorrectly detailed to achieve spans with low
amounts of deflection. Horizontal deflection caused catastrophic failure in Braced
and Pattern frame design iterations.
The conceptual non-orthogonal experiments (Fig. 4) focused largely on deliver-

ing a discontinuous structural geometry that was inherently resistant to lateral forces
and could both stand vertically and span horizontally. These tests were rapid and
touched on a range of different geometric approaches. Such tests aimed to examine a
spectrum of possible solutions early that could then be refined at a later point. Many
of these experiments were inspired by existing structural geometries and simply rein-
terpreted in a circular economy framework. For example, “Reciprocal Frame” and
“Reciprocal Span” (Fig. 4) were both inspired by work completed at the Hachinohe
Institute of Technology by Pei-Shan Chen (2014). Chen’s development of ‘1.5-Layer
Space Frames’ was seen to allow for rigid planer frame structures made of many
repeating modules. This study took the underlying structural rules and reinterpreted
them on a domestic scale with timber materials. While the authors adaptions were
unsuccessful (jointing members required further detailing to achieve adequate stiff-
ness) the experimentations provoked further studies and informed the development
of the final outcome. This process of adapting existing solutions continued to the
“Braced Frame” test (Fig. 4) where the system was designed to work with very
low cost materials such as straw and lime plaster. “Braced Frame” was therefore
a modulated adaptation of ‘ModCell straw technology’—a prefabricated straw and
laminated timber cell building system developed in the United Kingdom over the
last 5 years (ModCell 2017). Author modifications included horizontal modulation
to facilitate the incorporate of varying aperture openings and/or heights. While this
solution showed some potential, the inability of the system to span horizontally meant
that a second set of components would always be required. Again, this is seen as
undesirable as more varied components increase system complexity and constrain
reuse flexibility.
In order to develop these conceptual variations and tests a rapid prototyping
methodology was established. Ideas where modeled digitally in many small sep-
arate iterations, then translated into their respective components and fabricated at
1:10 scale using medium density fiber board (MDF) and computer numerically con-
trolled laser cutting equipment. Once discrete components of a model had been
tagged the process of nesting and calculating the machine paths for fabrication was
entirely automated. This process also enabled many different designs to be testing
simultaneously. Often one concept had four or five variations within itself that would
be quickly tested before an entire wall was formed. It is worth noting that at this con-
ceptual testing stage the design did not take advantage of parametric modeling tools
to generate the forms. The reason for this was that speed and simplicity were key
driving factors. Each concept only had a small number of unique pieces (maximum
of five) and it was the expectation that the materials were unlikely to change during
initial tests. Regardless of the integration of parametric modeling, the ability to cre-
ate highly detailed three dimensional digital models that could be quickly changed,
fabricated and developed was key to the conceptual design process of this research.
3.2 Developed Design
Testing of the conceptual ideas in complex systems worked as a process to identify

potential system limitations (Fig. 5). These collated tests highlight key areas of
complexity, such as wall to wall junctions and wall to roof/floor junctions, all of which
need to be resolved and designed into the geometry. X-Frame, as the most successful
conceptual design iteration, was developed further in two separate iterations. The
first iteration (Fig. 5, left) was an attempt to use the X-Frame system for all main
elements of a building and understand how it would work at each intersection. The
developed design test also examined how internal linings, claddings and water proof
layers would adhere to the non-orthogonal structure. Construction of a scale model
revealed the need to ensure that offset at any intersection (where two or more walls
meet) was managed through the design of the frame itself. Modeling also identified
the need to ensure that the perimeter of the non-orthogonal system can connect easily
with perpendicular structure(s). Iteration two (Fig. 5, right) was a further developed
design that attempted to simplify the non-orthogonal structural frame solution. Here
slender posts replaced the X-Frame matrix in vertical orientations. This allowed the
X-Frame structure to span horizontally and be easily expanded in any direction as
necessary. Although this approach had a valid intention it introduced another set
Fig. 5 Developed non-orthogonal design experiments in complex system assemblies. ‘Full X-

Frame’ (left) was an experiment in using the X-Frame concept for the entire structure. ‘X-Frame
Post’ (right) was an experiment in using the X-Frame concept for only the lateral structural elements
of components which were largely restrictive. These posts re-introduced orthogonal

geometric constraints such as fixed lengths and a lack of lateral load resistance.
Testing of these proposals in a realistic spatial situation is necessary to highlight key
drawbacks.
3.3 X-Frame 3-7
X-Frame represents the culmination of developed design experimentation. The solu-

tion performed best out of all conceptual iterations against CE criteria (Table 3). A
process of ‘refinement design’ for the X-Frame system was then undertaken to more
closely associate the system with the aims of a CE while also retaining high levels of
architectural freedom (Fig. 6). X-Frame 2 (Fig. 6, top left) was the proposal used in
the ‘X-Frame post’ developed design test and had a single four-point connector and
one spanning component. Although this design lacked the necessary structural detail-
ing to work dynamically in a building it did work to lay the foundation for the overall
X-Frame system. Further iterations (X-Frame 3 to X-Frame 7) then developed the
system to fulfill as many CE requirements as possible. Salient developments centered
around the span connection joint. It was essential that this joint could act as a rigid
point to transfer loads when used in both horizontal and vertical orientations. The
single connection component was replaced by three perpendicular plywood plates.
Two of these plates were designed to fit over the edges of the spanning components
to lock the four members together. One plate was slotted into the center of each span
to ensure correct alignment and minimal deflection for horizontal loadings. Consul-
tation with engineers suggested that parallel members should be introduced to span
this central joint to provide additional resistance to deflection (first introduced in
Fig. 5, centre left). This change, as well as a series of refinements to the perimeter
of the system and the addition of a central peg when two span components intersect
all aimed to improve the end-of-life material recovery potential of X-Frame.
The final X-Frame design iteration (X-Frame 7) was fabricated extensively at full
scale and assembled (Fig. 7). The assembled prototype tested the non-orthogonal
structural systems ability to form internal and external corners, to span horizontally, to
cantilever and to form structural roof/wall and roof/floor connections. Conventional
mechanical fasteners, in the form of M6 70 mm bolts with washers, were used in
this structure to secure the additional components that spanned across the four-point
connections. Bolts were also used at the perimeter of the frame to secure the two
edge components to one another. Four unskilled laborers assembled the structure
over 10 h using only step ladders, spanners and mallets. The prototype was then
deconstructed by four different laborers using the same tools over a period of five
hours. This construction and deconstruction process is considered one effective ‘use
cycle’ for the system. Over this cycle 97.75% of all components that were fabricated
were directly reusable.
Table 3 Construction circular economy performance criteria analysis of conceptual non-

orthogonal solutions
Platform X-Frame Pattern Reciprocal Braced Reciprocal
framing 7 frame frame frame span
Mass standardisation of components to ✓ ✓ ✓ ✓ ✓ ✓
improve direct reuse economic viability
Geometry enables the independent layering ✓
of materials to facilitate easy separation
Fixings must be exposed or have a sense of
inherent logic to facilitate separation
Fixings require no tools or are standardised ✓ ✓ ✓ ✓ ✓
to require only non-specialist tools
Fixings are detailed to be easily ✓ ✓ ✓
removed/reversed without contaminating or
inflicting irreversible damage to the
primary materials
Materials are detailed as to avoid the need ✓ ✓ ✓ ✓ ✓
for compromising bonded coatings or
treatments
Materials remain aesthetically desirable/can ✓ ✓ ✓ ✓ ✓
be easily restored to ‘as-new’ conditions
The use of composite/compromised or ✓ ✓ ✓ ✓ ✓
inseparable hybrid materials must be
eliminated
No chemical, composite or adhesive-based ✓ ✓ ✓ ✓ ✓
fixings can be used
Materials retain performance through ✓ ✓ ✓
multiple reuse cycles or easily facilitate
end-of-life management with minimal loss
of value
Modules and material components are ✓ ✓ ✓ ✓ ✓
easily reconstituted or formed in a way that
is economically attractive to reuse
Deployed materials are scalable based on a ✓ ✓ ✓ ✓ ✓
minimum divisible unit dimension in all
directions
4 Non-orthogonal Analysis and Critique
The remainder of this chapter will explore the functionality of non-orthogonal timber
frame solutions (predominantly X-Frame) that are made possible by CAD/CAM.
This discussion examines the advantages and limitations of non-orthogonal frame
design in respect to CE performance and real-world construction demands. As part
of this analysis the role of timber in these systems and role of digital design and
fabrication is critiqued.
Fig. 6 Developed non-orthogonal design experiments. Iterations of X-Frame 2 through 7 (left to

right, top to bottom)
4.1 Material
This study controlled its scope by limiting material and fabrication methods. The
choice of Plywood as the principal material was driven by local conditions (New
Zealand) and overall performance. Engineered structural sheet timber (Plywood),
is a low carbon, low cost, quickly growing and easily renewable building material.
Coupled with strategic design, engineered plywood has the potential to be precisely
machined using widely available CNC routing technology. Plywood can sustain
multiple reuse cycles and has a stable, typically local, supply chain. Additionally, as
an engineered timber product, plywood has a more consistent strength performance
measure versus conventional strength graded stud framing. Plywood also has the
potential, depending on how the timber product has been chemically altered, to be
sustainably managed in a cradle-to-cradle process if the product becomes damaged
or is discarded. These factors make the material ideal to use for the purposes of
non-orthogonal light structural frames in a CE.
Alternative timber products, including solid dressed timber, wood-plastic com-
posites and acetylated treated lumber, have not been explored in this study but all
have the potential to be assembled into non-orthogonal frame solutions. The use of
these materials would not be difficult as there is no need for a significant change in
the proposed non-orthogonal geometries. The use of solid dressed timber may also
Fig. 7 Completed X-Frame 7 prototype
be a way to reduce the embodied energy cost of the system by eliminating the petro-
chemically derived adhesive products required in plywood. However, a redesign of
the joint conditions would likely be required. Solid timber products can be shaped by
CNC machinery in a similar way to plywood, but they are prone to higher levels of
dimensional variation. This, as well as the deeper cross-sectional area, would need to
be considered in the redesign of any connections using an alternative timber product.
4.2 CAD/CAM, Non-orthogonal Geometry and Jointing
Computer-aided manufacturing (CAM) processes allow affordable detailing of struc-

tural members that integrate self-bracing capacity. This means that we can reimagine
how sheet materials are fixed to the building, and by doing so, make strides to elimi-
nate a significant portion of C&D waste (an estimated 18%). Self-braced wall struc-
ture should result in total (100%) recycling of post-consumer C&D waste through
adhesion techniques that are easily reversible and the removed need for additives in
gypsum board products (such as fiberglass). Self-bracing relies on the frame being
inherently stiff in every direction for every sized module. The geometric implications
of this physical requirement are either totally rigid 90-degree junctions or members
that act across the plane of a given structure (e.g. WonderFrame® ). Ensuring rigid
perpendicular connections in a continuous timber frame structure is difficult and can
result in the concentration of forces that cause catastrophic failure to a connection.
Click-Raft avoids this by replacing larger members with many slender elements. This
distributes the load in a way that takes full advantage of the strength of plywood.
Generally non-orthogonal assemblies of angular, curved and/or intersecting mem-
bers distribute forces in a way that ensures total resistance to lateral loads without
the need for rigid 90-degree joints.
The limitation of timbers capacity to form rigid 90-degree joint conditions is
exemplified by orthogonal plywood structural solutions. These configurations rely
on continuous sheet materials perpendicular to the gravity load resisting members to
achieve their lateral load resistance. Perpendicular sheet materials are also necessary
to resist buckling of the slender vertical members. Two existing orthogonal ply-
wood structural systems using sheet materials for bracing and available today are the
WikiHouse ‘Wren’ System and Facit Homes ‘Facit Chassis’ system. The WikiHouse
product is an opensource CNC router cut structural grid that uses portal-like frames in
the transverse direction that are made from three-dimensional plywood box-beams.
Between these portal frames continuous sheet material is fitted to provide lateral
stiffness. The Facit Chassis system uses the same CNC fabrication technology and
plywood material but instead modulates the structure into a series of panels. The
plywood panel structure is used for vertical components and the floor with com-
posite Orientated Strand Board (OSB) and solid dressed timber ‘I’ beams spanning
horizontally to support the roof. In both products sheet materials are used stiffen the
frame against horizontal loads through digitally fabricated teeth that interlock into
any available adjacent edge. It is rare for this sheet to be left exposed due to the
visible perimeter teeth. Plasterboard is either direct fixed (glue and screwed) to the
structure or fixed onto slim timber battens over top of the plywood bracing element.
Non-orthogonal frame systems use material more efficiently by eliminating the need
to have two layers of sheeting (one structural and one visual). Non-orthogonal frames
also ensure that what sheet product is specified is recoverable and reusable in many
different instances as it is not required to perform structurally.
Orthogonal frames can also use plates attached directly to the 90-degree junction
to create a rigid ‘moment’ joint capable of transferring lateral loads to the foundation
(like Fig. 8). This bracing solution does not require the transition to non-orthogonal
structural solutions or the rigid fixing of sheet linings. This proposal was trailed in
‘Click-Lock’—a prefabricated plywood framing proposal developed by Guy Mar-
riage and Jae Warrender in 2016 (Fig. 8) (Marriage 2016). In Click-Lock the perpen-
dicular plates were continuous and used to lock a central flange in place (Fig. 8). The
reality of this solution however is that timber, and even engineered timber products
such as plywood, have insufficient local compressive strength to resist the transverse
forces in a concentrated single connection. If this Click-Lock system was to be used
as a bracing element in a building engineers mandated that full surface linings would
be required to be rigidly fixed to the frame. Non-orthogonal structural frames do
not have such issues as the distributed load pathways largely eliminate the poten-
tial for this local deformation of timber joints under stain. This knowledge suggests
Fig. 8 Click-Lock system with perpendicular plates
that non-orthogonal structural solutions are more suitable as an intrinsically braced

structure for the CE.
Non-orthogonal structural systems introduce an additional level of geometric
complexity. Integrated jointing capacity aims to counteract this complexity through
simplifying the assembly of these systems while simultaneously meeting a range of
circular economy criteria. Components that are designed to join together without the
need for additional materials also help to achieve the aforementioned assimilated
bracing capacity. This capacity is made possible through connections that respect
the inherent structural properties of the material (plywood) they are formed from.
Integrated jointing like this is again exemplified in the Click-Raft system where slot-
ted connection points give logic to how elements need to be arranged to facilitate
assembly. In X-Frame, the non-orthogonal system developed in this study specifically
for the CE, integrated joints are used to collect the principal triangulated spanning
Fig. 9 X-Frame
members at nodes. A thin ‘plate’ of plywood with appropriately located penetrations

fixes over both sides of the junction, perpendicular to the triangulated spans (Fig. 9).
This is advantageous as only these two identical plates are required to align and lock
together up to eight converging structural members. Each mortise is elongated to pre-
vent joint failure and the separation of planes ensures that all converging members
are pulled into planar parallel arrangement. The plate also facilitates the reversible
fixing of linings and claddings to either side of X-Frame (see Sect. 4.3). Thus, the
added geometric complexities of non-orthogonal structural frames can be ‘mitigated’
through integrated jointing that also accentuates the overall assembly and circular
economy performance of the system.
Joints integrated into the structural components rely heavily on friction. An exam-
ple of these joints can be seen in Fig. 9 when a mortise and tenon connection helps
to lock together the span component with other converging spans. These joints are
only possible en-mass because of precision CNC fabrication where the tolerance for
a successful adhesion can be measured in the tenth of a millimeter. There are obvi-
ous issues with this approach. The timber product used to form these systems often
varies in thickness by more than the tolerance of CNC machinery (and the built-in
design tolerances) resulting in failed connections. For example, a tenon joint with no
applied loadings is held in place only by the friction between the two parallel edges
of the abutting timber. If the width of the plywood is thinner than expected the tenon
is only held in place by friction at either end of the joint, resulting in a significantly
weaker connection. To counteract this problem, we can design the joint so that the
parameters controlled in 2D fabrication are exact and the parameters that are not
adjustable (i.e. plywood thickness) can be absorbed. Subtle chamfering details on
the edge of a tenon will guide the component into alignment and guarantee adhesion
while not requiring physical changes to the mortise or tenon size. An issue may arise
however after separation where either part of the joint has deformed to an extent that
it will not function correctly in future uses (adequate friction). Environmental cir-
cumstances can also affect the thickness and size of timber cells resulting in further
compromised friction joints. For homogenous connections with timber, that is joints
that use exclusively wooden elements, pin and peg based connections are the most
reusable and proficient in aggregating inconsistencies. Such joints however are more
expensive to fabricate, use more timber and are typically bulkier in their detailing.
It is also important to note that issues with homogenous friction-only joints are not
unique to non-orthogonal frames. Both WikiHouse and Facit systems use the same
friction based jointing methods and, in the authors reuse tests of the WikiHouse
system, similar ‘inadequate friction’ situations arose.
4.3 Material Efficiency
Non-orthogonal timber frames offer the potential for increased lightness of the struc-
tural system. This is beneficial in terms of physically less materials required and
therefore cost savings, and generally easier workability during assembly. Decreased
weight is a result of the use of slender components in an optimised matrix that
essentially eliminates ‘collateral’ material deployment. For example, in a load bear-
ing floor spanning 2.4 m Click-Raft is 25% lighter than conventional timber beams
(Marriage 2016, p. 689). This is impressive considering plywood is 50% heavier per
cubic meter than un-engineered wood products (Dressed Pinus Radiata 400 kg/m3 ,
Pinus Radiata Plywood 600 kg/m3 ). X-Frame however does not offer such weight
savings. The thicker 17 mm plywood and redundant design inclusions (i.e. excess
material that allows the system to turn corners, span horizontally and stand vertically)
add to the weight of the solution. Calculations based on physical tests indicated that
one square meter of wall using X-Frame weights 15 kg versus 7.1 kg/m2 for 90 mm
by 45 mm platform framing. It is worth noting however that if platform framing was
specified at the same thickness as X-Frame (200 mm) the weight difference would
be negligible. Furthermore, future structural tests may validate the use of 12 mm
structural plywood in X-Frame, effectively reducing the weight of the system by one
third. As an aside, because X-Frame is conceived in a digital design environment,
the entire system can be parametrically changed parametrically to adopt a different
material thickness.
Digital processes can also be used to radically change the geometry of an individ-
ual structural component to deliver superior material efficiency. The huge reduction in
man-hours necessary to fabricate complex shapes (thanks to CAM processes) means
Fig. 10 X-Frame components failing due to assembly parameters
that we can now set the geometry of an object based on the efficient deployment
of materials, rather than efficient use of working hours. In X-Frame, the junction
elements are formed as to only have the minimum necessary material around a given
functional moment. A ‘functional moment’ might include a bolt hole, tenon, mortise
or a slotted connection. The minimum material quantity can be determined by the
capacity of the material to resist the applied loads for each functional moment. In the
case of X-Frame these specific measures are determined through collaboration with
structural engineers, software simulations and material manufacturer guides. Once
the technical specifications have been set they can become part of the intelligence
of the digital design and applied to the geometry through automated scripting. The
result is components that are significantly lighter and, depending on the resulting
geometry, more cost effective to fabricate.
The slotted joint (discussed in Sect. 3.2) that Click-Raft uses is a point of dis-
cussion. This ‘slotted joint’ occurred in all X-Frame iterations from one to seven
when four span elements converged. The slotted joint ensured each converging com-
ponent could be quickly aligned in the correct position. The outer plywood plates
could then be fixed easily once the frame is in its final position. The simplicity of
the slotted joints fabrication and the rapid construction times that it facilitated was
very advantageous. This joint however was often awkward when larger segments
were being assembled due to the tension exerted by needing to bend the plywood
span outwards. This meant putting increased strain on the weakest part of the span
element in the timbers weakest direction, often resulting in total failure (snapping)
(Fig. 10). Later iterations (X-Frame 9) removed this central element and thus when a
section of X-Frame is lying flat all joints separate without difficulty or risk of failure.
Is there a comparable ‘failure of member due to fixing’ issue for orthogonal

structural systems? While conventional light timber platform framing does not rely
on any elements under constant strain, separation of components still results in major
damage to the principal materials. The wire nails fixing structural members in place
splinter the wooden components on removal. Coupled with the residue left on the
studs from plasterboard adhesive the time required to prepare the material for reuse
often exceeds the cost of buying a new piece of timber (Forbes 2018).
4.4 Assemblage Challenges
Structural systems not complying with rectangular geometry face obvious challenges
where they intersect conventional orientations. The current and perhaps very valid
‘expectation’ of flat floors and vertical walls means that the system must negoti-
ate a 90-degree corner regardless of its form. The frame also needs to be able to
negotiate the intersection of other vertical members and their respective edge condi-
tion. Furthermore, conventional window configurations (rectangles) and the resulting
orthogonally framed openings that they require presents a significant challenge to
non-orthogonal solutions. This could be a significant limitation of Click-Raft as
the curved geometry never leads to a ‘orthogonal conclusion’ on two of its edges.
Installing a window in a conventional way would mean the use of a large amount
of awkwardly shaped blocking supported by the vertical Click-Raft members. How-
ever, this issue has typically been avoided in existing Click-Raft use cases. Instead,
a polycarbonate cladding has been bolted to the outside of the frame creating a con-
tinuous translucent envelope. Yet where openings and transparent insulated glazing
units have been required the Click-Raft system is replaced entirely by conventional
LTF at the perimeter. These solutions are not desirable for a Circular Economy as
they represent the need for bespoke elements and a reduction in flexibility. This
leads to a system that is not economically attractive to reuse due to the increased
presence of massively varied components. X-Frame performs better in this respect
due to its triangulated geometry. Working with only two more individual members
than Click-Raft, X-Frame can perform rectilinear transitions both at its edge and
internally.
Industry’s current preoccupation with orthogonal structural systems means that,
like windows and doors, most materials are formatted to work within a platform
framing ‘framework’. Any non-orthogonal approach that does not work within these
constraints is therefore compromised. This is true for the interaction between off-the-
shelf insulation products and Click-Raft. Although insulation sheets can be fitted to
the exterior of the frame conventional practice dictates infilling the cells of the frame
with an insulation product. Fitting conventional products between the curved cellular
structures is a labour intensive and time-consuming task. In the case of Click-Raft,
external insulation fixing may be preferable as it would retain the sinusoidal curve
pattern internally, while performing the necessary functional thermal control. This
configuration, however, would result in a wall almost 300 mm thick which could be
Fig. 11 Click-Raft with chamfered baton to attach flooring finish
undesirable in many situations. Fixing insulation externally would also require an

additional batten or bracket mounted in or on the Click-Raft system. X-Frame is again
more easily supplemented with conventional building products due to its 600 mm
spaced diagonal grid. A single cut to a polyester batt allows it to fit snugly within the
frame. Both Click-Raft and X-Frame could also employ blown-in insulation products
such as cellulose, glass wool or EPS beads. This solution would fill the unusual cavity
shape of Click-raft easily. This discussion highlights that conflicts brought about by
non-orthogonal structural systems can be easily mitigated using a range of existing
construction products.
While Click-Raft’s two-piece design is favorable in terms of simplicity and joint-
ing to itself, the ability for the system to accept linings and claddings is also lim-
ited. This is a concern given that successful Circular Economy design is as much
a consequence of successful integration as it is a consequence of a single product.
That said, computer-aided manufacturing makes it possible to eliminate such issues.
Mortise details can be added to any component easily through parametric or stan-
dard digital modeling processes. These reception points enable further homogenous
‘wood-only’ connections between the structure and supplementary building enve-
lope layers. Click-Raft has partially utilised this approach in the past by using a
chamfered CNC router cutting head. A chamfered cut was made to the edge of the
click-leaf components to form a set of parallel notches aligning perpendicular to the
click-leaves (Fig. 11). This allowed an appropriately chamfered batten to be threaded
through the leaves and act as a fixing point for screws from the plywood flooring
above.
While not a pure Circular Economy solution this detail demonstrates the potential
of CNC fabrication, timber and intelligent design to facilitate integrated material
connections. This idea was explored in far greater depth throughout the development
of X-Frame. Oversized plywood pegs oriented perpendicular to the direction of the
wall were designed to thread through the X-Frame locking plates. This allowed a
plywood waterproof sheet lining to be compressed to the structure itself with a small
plywood pin. Compression joints at regular intervals ensured complete adhesion
while also enabling rapid and damage free deconstruction. Further detailing of this
oversized peg allowed a visual rain-screen cladding to be fixed over the waterproof
layer using only another small plywood pin at regular intervals. Digital design and
manufacturing is again assisting in total material reuse through easily reversible
homogenous junctions.
4.5 Aesthetics
Non-orthogonal geometries can be incorporated in an architectural way that exhibits

their structural geometry as a formal feature. In all Click-Rafts installations the
structure has remained exposed from at least one side. The slender plywood ele-
ments and the gentle sinusoidal curves that intersect without secondary materials
form an eye-catching structural matrix. Exposed structural elements like this have
been popular in traditional architectural styles but are typically hidden in modern
light timber framed buildings. Furthermore, and importantly, the visual spectacle
that these non-orthogonal structures can be associated with a ‘formal sustainability’.
This expressive formal embodiment of sustainability is rare, yet significant, as it
aligns with architectural theorists’ views that “sustainability should be revolutionary
to architectural form” (Heymann 2017). Critically, the exposed geometries are also
serving a useful CE purpose. They inadvertently eliminate the need for an internal
plasterboard lining and aid in communicating the logic as to how the structure and its
supplementary layers can be separated. These features make end-of-life deconstruc-
tion and material recovery safer, faster and less wasteful. X-Frame’s woven diagrid
asks to be left exposed. X-Frame 7 facilitated this exposure by incorporating softer
curved components and two distinct 90 mm cells (Fig. 12). Multiple cells allowed
half of the structure to be filled with insulation and an internal lining while keeping
the other half exposed to the interior. The non-orthogonal geometry is expressed,
thermal and/or acoustic performance is achieved, and the system is significantly
faster to deconstruct.
4.6 Structural Limitations
Non-orthogonal structural systems have the potential to span horizontally and verti-
cally using the same structural elements. This is a Circular Economy advantage as it
Fig. 12 X-Frame 7s multiple cell structure
means less differentiating components are required in the overall system. Less, and
more flexible components, mean better recovered material resale rates as the same
piece can be used in many different situations. To enable this flexibility the depth
of the individual spanning elements must be increased to match the corresponding
maximum span length of the assembled system. To ensure maximum interoperabil-
ity the depth of a spanning component should not change between horizontal and
vertical configurations. Horizontal spanning also requires stiffening of the structure
against live face loads such as people walking across a floor. These loads far exceed
the normal forces exerted on the face of a building by wind and therefore joints
must be resilient and stiff. Click-Raft’s maximum span to date is 2.4 m before being
supported by a structural beam. X-Frame prototypes have spanned 5.4 m however
this span length has not been tested under live loads. It is likely that the deflection
caused by the network of bolted and timber only joints will exceed safe maximums.
This is where WonderFrame® excels as rather than adjusting the depth of the geom-
etry another layer can be bolted into the first without any additional components to
provide the adequate strength. Connections between the smaller components do not
deflect due to their metal to metal bolted detailing. An alternative geometry, made in
timber and digitally fabricated, that is expandable in its depth to achieve higher load
capacity is in the scope of forthcoming research. This research will also quantify
the deflection and maximum span potential of the X-Frame system in its current
format. Based on the author’s experience with the system it is likely that deflection
in the joint due to the inadequate connection detailing and the slender, soft, plywood
material will be the ultimate limiting factor.
Underpinning all design explorations and discussion in this research was a clini-
cal understanding of the theoretical structural capability of non-orthogonal framing
solutions. The use of diagrid ‘nodal’ geometries was based on extensive empirical
evidence of the superiority of triangulated member arrangements and their ability
to distribute forces (Moon et al. 2007). In the case of X-Frame, almost all vertical
structural elements have been eliminated with “the diagonal members in diagrid […]
carrying gravity loads as well as lateral forces” (Moon et al. 2007). The result is a
structural matrix that is constantly working to distribute forces evenly. Dense spacing
between each node in the X-Frame is also purposeful and fundamentally driven by
the buckling capability of plywood as well as efficient modulation in architecture.
The maximum distance a member spans before being intersected by a supporting
element is 600 mm. This all but eliminates the possibility of buckling of the slender
timber members.
5 Conclusion
Computer-aided manufacturing makes possible this new category of non-orthogonal

light timber frame design. Now intricate and detailed individual components can
be mass fabricated, at competitive economic rates and with low-skill labour. From
this new fabrication capacity comes the ability to move away from strictly orthog-
onal structural frames and exploit the advantages of non-linear and self-stabilising
geometries. The optimised frame geometries are based on a fundamental and exact
relationship between material, form, function and assembly. This complex asso-
ciation includes the detailed consideration of weatherproofing, seismic structural
resistance, service integration and jointing operations.
If designed appropriately, these patterns can be geometrically rationalised to be
flexible and suitable for vast range of different building situations and therefore,
through appropriate jointing details, be reused again and again. X-Frame is a clear
demonstration of the versatility of a modulated diagrid. The same discontinuous
structure can be used across the floor, walls and ceiling of one structure and then dis-
assembled and rebuilt in a different configuration using the same components. Direct
reuse of the frame eliminates end-of-life material waste and adds value to the framing
product. Awareness in the design phase of the performance capabilities of plywood,
computer-aided manufacturing potential and real-world assembly/disassembly con-
straints informs the reusable jointing details and thus results in a more intelligent
frame product.
Acknowledgements The authors acknowledge the generous support of the New Zealand Institute
of Building’s Charitable Trust, Carter Holt Harvey Limited, Makers of Architecture and Victoria
University of Wellington in facilitating the fabrication and testing of non-orthogonal structural
solutions. This work was funded in part by the Building Research Levy.
References
Alderman D (2013) Housing and construction markets. UNECE/FAO Forest Products Annual Mar-
ket Review, 2012–2013, 118
Chen P (2014) Innovation of 1.5 layer space frames. J Struct Eng 60(B):153–158
Curtis M (2015) Physical characteristics of new houses in 2014. Building Research New Zealand
(BRANZ) Study Report 330, 5
DoBH (2012) Dealing with timber in leaky buildings. A Guide for Builders and Building Profes-
sionals, Department of Building and Housing, New Zealand Government, 5
EPA (2014) Advancing sustainable materials management: 2014 fact sheet assessing trends in
material generation, recycling, composting, combustion with energy recovery and landfilling in
the United States. Environmental Protection Agency, United States of America, 16
Finch G (2018) Defab: prefabricated architecture for a circular economy. Unpublished thesis, Vic-
toria University of Wellington, 1
Forbes N (2018) Reusing Pinus radiata structurally in New Zealand. Unpublished thesis, Victoria
University of Wellington, 12
Heymann D (2017) The ugly pet. Places J, September 2017. Accessed 30 May 2018. https://doi.
org/10.22269/170917
Inglis M (2009) Construction and demolition waste—best practice and cost saving. Ministry for
the Environment, 1–12
Marriage G (2016) Experimental construction in a timber house. In: Aurel M (ed) Proceedings of
the 50th international conference of the architectural science association 2016, pp 685–694
McDonough W (2016) ICEhouse™ Davos, Switzerland, William McDonough + Partners, Archi-
tecture and Community Design. Available from: mcdonoughpartners.com/projects/icehouse/.
Accessed 1st May 2018
McKeever D, Phelps B (1994) Wood products used in new single-family house construction: 1950
to 1992. For Prod J 44(11/12):66–74
ModCell (2017) ModCell Systems; About Us. ModCell Straw Technology. Available from:
modcell.com/about-us/. Accessed 5 June 2017
Moller C (2016) CMA + U. Available from: cma+u.com. Accessed 21 Jan 2018
Moon K, Connor J, Fernandez J (2007) Diagrid structural systems for tall buildings: characteristics
and methodology for preliminary design. Struct Des Tall Spec Build 16(2):205–230. https://doi.
org/10.1002/tal.311
Parisio S (2006) Arsenic & old landfills: what we have learned from post-closure groundwater
monitoring at inactive landfills in NY State. New York State Department of Environmental Con-
servation, 2006 SBRP Workshop on Arsenic and Landfills: Protecting Water Quality, 18
Ramage M, Burridge H, Busse-Wicher M, Fereday G, Reynolds T, Shah D, Wu G, Li Yu, Fleming

P, Densley-Tingley D, Allwood J, Dupree P, Linden, P, Scherman O (2017) The wood from the
trees: the use of timber in construction. Renew Sustain Energy Rev 68, Part 1,2:333–359
Webster M, Costello D (2005) Designing structural systems for deconstruction: how to extend a
new building’s useful life and prevent it from going to waste when the end finally comes. In:
Greenbuild conference, 9
Timber Plate Shell Structures: A Digital
Resurgence of Traditional Joining
Methods
Christopher Robeller
Abstract Thin shell structures such as vaults and domes, allow for a column-free
construction over a large span. Such surface-active structure systems have been of
interest for architects and engineers over centuries, both due to their structural effi-
ciency and their elegant appearance. Only thin cross-sections and therefore small
amounts of materials are required, due to the curved or even double-curved shape,
which minimizes bending moments and provides a membrane action. However, the
design and production of these thin shell structures has always been challenging. In
particular, the construction of shells usually requires very time-consuming and costly
formwork. Over time, advances in technology and new materials have changed the
way vaults and shells have been designed and built, leading to new structural typolo-
gies. The construction sequence and assembly methods have always had a decisive
role and a great influence. This article will describe the background and the ongoing
developments in the design of plate shell structure, with a detailed focus on timber
folded plates and a segmental plate shell system. The digital age has brought entirely
new possibilities and materials, which may remove the cost and time constraints
in the design, fabrication and production those efficient and elegant, surface-active
structures. While the inspiration for new assembly methods and building material
wood date back to the beginning of building history, the efficient realization can now
be achieved through algorithmic geometry processing.
Keywords Computational design · Digital fabrication · Design for assembly

Lightweight structures · Shell structures
C. Robeller (B)
Digital Timber Construction DTC, TU Kaiserslautern, Kaiserslautern, Germany
e-mail: christopher.robeller@architektur.uni-kl.de

1118 C. Robeller
1 Plate Shell Structures in the Digital Age
This paper describes the background and the ongoing developments in the design
of timber plate shell structures. Recent advances in this field of research are driven
by computational design methods and digital fabrication technology, which allows
for more geometrically complex structures to be realized in a cost- and time effi-
cient manner. Structurally efficient shell and spatial structures present particularly
interesting case studies for these new technologies. Much like in natural structures,
these lightweight building systems use form to reduce the amount of material needed.
While the later sections of this article focus on timber plate shell structures and their
relevance beyond the field of space structures, the following sections will provide a
background and context to these developments. Over decades, shell and spatial struc-
tures have been at the forefront of technological advances. Computational design,
production and engineered wood products mark the latest step in their evolution.
2 Origins of Braced Vaults
Barrel vaults have been used throughout the history of buildings, dating back to the
antique, the Roman and the Persian architecture. These stone and brick structures had
a very high self-weight, which required massive walls and foundations to support the
reactions. This is particularly problematic due to the cylindrical shape, which creates
considerable horizontal forces. The other disadvantage of the stone and masonry
vaults in general was the very labor-intensive construction of formwork. These two
issues of dead load and formwork were addressed with new material and technology
in the 19th and 20th century. Barrel vaults with a greatly reduced self-weight became
possible with iron bars and trusses during the industrialization period, where prefab-
rication also became an increasingly important topic. One of the earliest and most
important “braced vaults” and prefabricated structures was Joseph Paxton’s famous
Crystal Palace for the 1851 Exhibition in Hyde park, London (Fig. 1a).
In his Book “Analysis, design and construction of braced barrel vaults”, Stanislav
Makowski describes one of the main reasons for the popularity of such barrel type
vaults at this time as the ability to construct them with “identically shaped members”
(Makowski 1963). This trend towards the efficient mass-production with a large
amount of equally shaped elements could be observed throughout the industry in
this period. The same reasons led to the development of the first lightweight braced
barrel vault using prefabricated wood elements by Friedrich Zollinger, which he
patented in 1906 (Fig. 1b). Mainly used in a time of great scarcity between the first
and second world war, this system allowed for roof constructions with short wood
elements. The overall amount of wood was reduced by 40% compared to a regular
roof structure.
Timber Plate Shell Structures: A Digital Resurgence … 1119
Fig. 1 a One of the earliest braced barrel vaults using iron bars: Crystal Palace, London, 1851.
b Braced Barrel Vault using wood: Hangar Roof near Munich, 1908. c Braced Barrel Vault using
reinforced concrete: Orvieto Hangar, P. L. Nervi 1935
Since the Zollinger “lamella roof” was only diagonally connected, the cross-
bracing was achieved through the cladding with wooden boards, which were nailed
as a layer on top of the lamellas. Similar ribbed shell vault structures using steel-
reinforced concrete were constructed in the early 20th century, including the remark-
able Architecture by Pier Luigi Nervi, such as the famous Demanio Aeronautico
Hangar in Orvieto (Fig. 1c). In this building and other structure of Nervi, precast
concrete plates with protruding rebar elements were connected on site through weld-
ing. Making it one of the earliest prefabricated vault structures using plate elements.
Due to the lack of calculation methods for such complex assemblies, elaborate phys-
ical models were used for the development of the structure and for their analysis.
However, the construction of such concrete barrel vaults still required elaborate form-
work and was rather slow and expensive. Another issue with the concrete structures
was that it was rather difficult to include openings for roof lights. This was a clear
advantage of the steel vault structures.
3 First Corrugated Vaults
Another geometrical variation for lightweight vault structures takes advantage of

folded plate elements. The structural principle of folded plates is a combination of
three structural actions in one structural elements: (1) A horizontal plate action, (2)
A vertical slab action, and (3) A frame action, when it is clamped on the sides. This
structural principle had first been discovered in the 1920s, in the early construc-
tions with reinforced concrete. In the design of coal bunkers, concrete plates were
connected in the for of an inverted pyramid. These oblique plate assemblies were
observed to be stable without any cross-bracing. The principle was used in many
building structures in the following decades. It still required (with few exceptions of
experimental prefabricated structures) an in situ casting of the concrete and form-
1120 C. Robeller
work, however the planar formwork was considerably simpler that the formwork for
barrel vaults or shell structures.
An early system using plates but not requiring any formwork were the called
“Corrugated barrel vaults”, which were first proposed with the advent of glass-fiber
reinforced plastics plates in the 1960s. They combined the principle of folded plates
with those of the diagonally braced barrel vaults. In his paper “the analysis of folded
plate structure in plastics”, B. S. Benjamin emphasizes the decisiveness of the ease-of
prefabrication and assembly: “Plastics as materials are ideally suited to prefabrica-
tion, and this concept is one of the greatest importance in the choice of a suitable
system. For ease of the prefabrication, the basic folded plate unit should consist of
only two or three component plates.” Benjamin first presents a corrugated plate vault:
“The folded-plate barrel vault is one of the most interesting structural shapes avail-
able for the roofing of large areas. It has the advantage that it can be built up with only
one type of unit, which makes it ideal for prefabrication and hence for a material like
plastics.” (Makowski 1963). In a retrospective, only very few such structures were
realized with plastics, which provides a low weight but also a relatively low strength
and challenging details for the connections between the prefabricated elements. One
of the well-known examples is a movable lightweight structure for a sulfur extrac-
tion facility in Italy, designed by the architect Renzo Piano in 1966 (Pizzi and Piano
2003).
Even though the shape of these corrugated vaults was considered efficient in
structural terms, a construction with plastics did not allow for large spans and its
construction with concrete would have been too difficult and costly. Instead, the vast
majority of concrete folded structures was straight extrusions of a folding profile, so
called prismatic folded plates. The formwork for such designs could be realized in a
cost-efficient way (Fig. 2).
An entirely new plate material for building applications became available with
engineered wood plates such as glue-laminated plates and cross-laminated timber
(CLT) in the 1990s. A first folded structure using 100 mm thick glulam plates was
designed and realized by Regina Schineis in 2004 for a music rehearsal hall in
Fig. 2 Corrugated Barrel Vaults made from thin fiber-reinforced plastics plates (Makowski 1986)
Thannhausen, Germany (Schineis 2004). This building with a span of 10 meters was
first taking advantage of the possibility to realize more complex shapes with the
wood plates. Both the walls and the roof were designed as prismatic folded plates,
connected along the longitudinal edges of the structure.
While the prismatic folded plate shape of the Thannhausen structure was well
suited for glue laminated wood elements, cross-laminated plates presented a new
opportunity to revisit the structural designs and topologies that were experimented
with in the 1960s with plastics, such as Benjamin’s and Makowskis corrugated bar-
rel vaults. Thanks to the cross-lamination of plywood and CLT, which provide an
orthotropic structural behavior similar to the fiber reinforced plastics, the thickness
of the plates could be greatly reduced.
A first experimental corrugated barrel vault prototype made from 20 mm thin
plywood plates was built and studied in 2006. Its shape was inspired by the Japanese
Origami Art of paper folding, namely the so-called Miura Ori pattern. Just like the
experimental structures of Makowski and Benjamin, the design was taking advantage
of the repetition of only 4 basic units, which were assembled into a barrel vault
consisting of 120 elements in total. Both the cutting and assembly were carried out
using custom built templates, the connections between the plates were realized with
self-tapping wood screws in the miter jointed side faces of the plates. The studies of
the structural behavior revealed that similar to the structures with plastics, the large
amount of plate-to-plate connections were challenging to realize with a sufficient
stiffness, ease-of-assembly and cost efficiency.
Following the sustainable, lightweight plates made from wood, another key tech-
nology for the realization of corrugated vaults was introduced with the proliferation
of digital technology in 2000s. While CNC fabrication technology had been available
in factories for a few decades, its full potential for building applications was made
accessible to architects at this time, through the availability of architectural CAD
with application programing interfaces and parametric design applications.
4 The Digital Age: Individualized Production and New

Plate Materials
One of the most important factors in the design of shell structures is the formwork. It
is decisive for the total cost, complexity and construction time of the structure, any
may often foil the potential efficiency of thin shell vaults. The idea of Benjamin and
Makowski’s corrugated vaults was to entirely get rid of the formwork, working with
a plate material instead of casting concrete. Due to the fabrication with templates,
these structures however were limited to the repetition of identically shaped elements,
which greatly constrains the variability of the overall shape of the vault. It results
in the circular arc shaped cross section and single curvature of the barrel vaults.
Investigations showed that large deflections are to be expected at the edges of such
1122 C. Robeller
Fig. 3 a Barrel Vault, 4 different shaped plates. b Double-curved shell, 140 different shaped plates
design, but more efficient shapes could not be realized with these methods. This was
also the case for the first experimental corrugated vault built from cross-laminated
wood plates.
New possibilities opened up with the proliferation of digital technology. CNC
machines and robots do not require templates and guides. Figure 3 shows a barrel
vault made from 140 plate elements. The left side image shows how such a mass-
produced, single curved vault which could be realized with the only 2 basic units, the
main unit is produced 124 times, the other one 16 times. On the right side however,
we see the same design with a parabolic cross-section instead of a circular one (blue
curves), as well as a second curvature in the longitudinal cross-section (red curves).
This double-curved shape is not only structurally more efficient, it also shows the
variability of such an individualized production, allowing for a variety of architectural
applications.
As shown in the top view of the plates at the bottom of Fig. 3, the double curved
overall shape of the structure is achieved through small variations in the shape of
each plate. The fabrication of such an individualized production required two tech-
nological innovations. First, CNC machines do not require templates, for them the
production of the 85 individually shaped plates shown in Fig. 3b is not more time
consuming or more expensive than for the 2 mass-produced plates in Fig. 3a. The
only step in the fabrication that could take more time in the production of such an
individualized series of plates, is the generation of the machine G-Code, however
this depends on the programing technique. The same applies to the work of the archi-
tect and engineer. Manual drawing using CAD, CAE or CAM software would be
very time consuming because these programs were optimized for products and build-
ings, which were influenced by mass-production fabrication processes ever since the
industrialization. In this context it is particularly remarkable that the very first CAD
system “Sketchpad”, which was developed in the 1950s at MIT, was an exception.
This first CAD system was a fully parametric CAD. Robert Aish, who was one of
the pioneers to bring parametric functions back into architectural CAD forty years
later, wrote that it was “a career changing experience”, when he first got in touch
with the predecessor of this system in the 1960s (Peters and Peters 2013).
While the widely known architectural CAD mostly imitates the traditional pro-
cesses of manual technical drawings, with the added benefits of much simpler revi-
sions and increased precision, the parametric CAD allowed for the definition of form
through its underlying logic and parameters instead. Ultimately, the two approaches
in generating and processing geometry are very different, which will inevitably reflect
in the architecture that is created with these tools.
5 Design for Assembly: Prismatic Joints as Assembly

Guides
While the fabrication and design tools have evolved allowing for the production of
advanced, individualized structures, a major question remained how they could be
assembled. It was already shown in the work of Benjamin, Makowski and Huybers,
that the assembly of a large amount of thin plates is challenging. Generally, the
connections in such structures have two roles. The first one is to transfer stresses,
which is critical to actually benefit from the advantageous shape of a vault and its
very thin, lightweight plates. In the case of the corrugated structures, the rotational
stiffness at the ridges is very important. The first prototypes with cross-laminated
wood plates showed that it was difficult to achieve the required stiffness with state-
of-the-art steel fasteners such as self-tapping screws.
The second critical role of the joints, apart from the transfer of stresses, is to
allow for a fast and precise assembly of the structure. Considering the large amount
of plate elements in the corrugated vaults, these so called “locator features”, as they
are called in mechanical design, are particularly important for such structures. Joints
without locator features, such as the self-tapping wood screws in the first timber
folded plate prototypes do not provide any help with the correct alignment of the
plates. Additional templates were used for this purpose, but this will only work if
all plates meet at the same fold angle (Fig. 3a). So, this method cannot be applied
individualized, freeform designs (Fig. 3b).
Considering that the individualized cutting of large amounts of differently shaped
plates is carried out on CNC machines anyway, the possibilities of these machines
should be used for the individualized production of the connectors as well. In the
domain of timber frame structures, which are built from linear wood elements like
beams and posts, this integrated production of the joints, in one process along with the
cutting of the beams, has already become a state-of-the-art method. The connections
are mostly through-tenon joints, which are similar to those of traditional handcrafted
wood structures, but they have been modified and optimized for the cutting with
rotating, cylinder-shaped milling tools. As a result of these adaptations, the CNC
fabricated through tenons have a rounded cross-section, because their counterpart
holes can be milled like this more easily. The traditional handcrafted joints however
1124 C. Robeller
had a rectangular cross-section, which was the result of the counterpart holes being
cut with chisels.
While these CNC-tenons for linear wood elements have been developed for wood
joinery machines like the Hundegger K1 beginning in the 1980s and widely used for
multiple decades, the development of similar CNC cut joints for plate-shaped wood
elements had only begun in the 2000s. two of the earliest architectural projects that
experimented with such methods, were the Swissbau Pavilion 2005 developed by
ETH Zurich and the 2007 Serpentine Gallery Pavilion of Alvaro Siza. In the Swiss-
bau Pavilion, OSB plates were connected with miter joints and dovetail-shaped Hoff-
mann Schwalbe connectors. The Serpentine Pavilion was demonstrating a reciprocal,
Zollinger-like construction of laminated veneer lumber plates, which were connected
with through-tenon joints. However, both structures were grid shells rather than the
plate shells. For the connection of plate shells, a connection with multiple tabs and
slots would be required, such as dovetail joint from traditional handcrafted cabinet-
making. Such joints did not exist in traditional handcrafted carpentry, because there
were no plate shaped elements.
The cutting of multiple-tab-and-slot dovetail joints with a CNC machine on cross-
laminated plywood was first demonstrated by Simek and Sebera in 2010, aiming at
applications for furniture. A first prototype using these joints on CLT plates for
an architectural application was presented in a curved-folded prototype structure
in Mendrisio, Switzerland in 2013. In this first structure, the joints were used in a
combination with a 1K PUR adhesive. They allowed for a precise alignment of the
curved plates and also provided the necessary rotational stiffness. The use of such
joints without glue, which would allow for much more versatile applications, was
first investigated in 2014. This study compared the stiffness of the dovetails with
other state-of-the-art connectors, concluding that a strength equivalent to a screwed
connection in the side faces of the plates can be achieved. The possibility to use
such joints on cross-laminated plates allowed for entirely new applications. The
traditional handcrafted dovetails on solid wood boards had a major constraint, as
they could only be applied to one side of the wooden boards, across the fiber of the
wood. This constraint is due to the highly anisotropic nature of the material wood. The
strength across its fiber direction is approximately 20 times higher than in the other
direction. On cross-laminated wood plates however, the alternating, perpendicular
configuration of the layers of wood veneer (LVL) or wood boards (CLT) results
provides an orthotropic behavior.
In 2012, a first experimental investigation of the mechanical strength, aiming at the
use of such joints in building structures, showed that indeed, a stiffness compared to a
connection with wood screws can be achieved with CNC cut dovetail joints on cross-
laminated wood plates. An even higher stiffness, as shown by the experiments, could
be achieved with glued connection. However, for flat-pack “IKEA-style” lightweight
and transportable construction systems, which are becoming more and more impor-
tant in an increasingly pre-fabricated building construction environment, on-site con-
nections are particularly important. Here, glued connection cannot be used, because a
controlled environment is required for the curing of the adhesive, such as temperature
and humidity, to ensure a safe and reliable connection.
The first CNC dovetails were cut with 3-axis CNC technology, which requires the
pin part of the joints to be cut on a vertically positioned plate. For our “individualized
production” application, as it was proposed in Fig. 3, this would have been very
complicated to implement, since all plates have both pin and tail elements on one of
their edges, meaning that the plates would have to be re-positioned during the cutting.
Such re-positioning and clamping of parts is not only very time-consuming, but even
more importantly, the required precision of the parts of less than a millimeter cannot
be achieved when the plates are moved manually during the fabrication process. The
solution to this problem was to cut the pin part of the dovetails with 5-axis CNC
technology, which was first implemented in the above mentioned Curved Folded
wood Pavilion in 2013. In contrast to the traditional handcrafted dovetail joints and
also the first 3-axis cut CNC dovetails, the plates in the Mendrisio prototype were
not connected orthogonally. The geometry of the dovetails allows to connect plates
at a range of dihedral angles, which results from the maximum possible inclination
of the cutting tool. Such geometry is possible with dovetail joints, it was known in
traditional crafts, but very difficult to produce with hand tools. Hence, the digital
implementation took particular advantage of this technology.
The Mendrisio Pavilion can be described as a one-directional composition of 5
plates. In each of its connections, two neighboring plates were joined along one
common edge. This is very similar to the “drawer” configuration in traditional cab-
inetmaking. However, for most segmental shell structures, such as the corrugated
shell shown in Fig. 3, the surface is segmented in two directions. There are many
different possibilities for this “tiling” or “tessellation” of the structure, but all of
them will require that two or even three edges must be connected simultaneously,
in one step, during the insertion of the plates. Using prismatic connections like a
dovetail joint, which allow only for one insertion direction, this was not possible
with standard dovetail joints and presented a major challenge, aiming at the use of
such joints in two-directional plate compositions. The problem results from the fact
that the insertion direction of traditional dovetail joints was always perpendicular to
the connected edge line. However, it is possible to join plates diagonally, as shown by
traditional Japanese joints such as the decorative “Nejiri Arigata” connections, which
were used for gift boxes in Shinto Shrines in Japan for centuries. A new digital, algo-
rithmic method was now developed based on this principle, which was demonstrated
in the Interlocking Folded Plate prototype in 2014. This structure, very similar to
the example in Fig. 3, consisted of 156 individually shaped plates, which were all
connected with prismatic, dovetail type joints. Figure 4b shows, that the insertion
vectors of the individual joints had to be rotated, to be parallel with the other joints
on their plate, which had to be connected at the same time. The research project
presented a method, which allowed to calculate each of these joint rotations, and to
confirm whether it lies within the constraints of fabrication and assembly. It must
be noted that the rotation of an insertion vector is only possible within a so called
“rotation window”, which represents the 3d vector subset of all possible insertion
directions for this joint. With this being calculated for each of the simultaneously
connected edges, the intersection of the individual vector subsets can be obtained,
which will indicate whether these edges can be connected simultaneously using this
1126 C. Robeller
Fig. 4 a Interlocking folded plate prototype (Robeller and Weinand 2015). b Interlocking folded
plate schematic (Robeller and Weinand 2015)
joining method. Figure 4 shows that the angles between the simultaneously joined
edges in this prototype were rather obtuse, which is beneficial for the methods, since
the assembly directions do not have to be rotated slightly. The image also shows that
there are multiple edges in this structure, where no other edge must be joined at the
same time (basically all edges oriented along e2). On these edges, standard dovetail
joints were used.
6 Double-Layered Corrugated Vaults
The “Interlocking Folded Plate” project first took advantage of the algorithmic joint
and CNC code generation, which was programed with a CAD application programing
interface. It allowed to process 156 individually shaped plates. Technical details on
this prototype, including a load test and the comparison of deflections with a singly-
curved vault are provided in a separate technical paper (Robeller and Weinand 2015),
as well as further information on the choice and influence of the fold pattern (Stitic
et al. 2015) and on the general semi-rigid behavior of the joints that were used (Roche
et al. 2015).
The potential of this new method could now be further explored, through the
application of the existing geometry generation program to other prototypes. A par-
ticularly promising application was the development of double-layered structures,
due to the fact that cross-laminated wood plates such as LVL provide a high strength
at very thin cross-sections of only a few centimeters. However, connecting such

plates directly along their edges is very difficult with standard connectors such as
screws. Side distances must be respected, which are typically five times the shaft
diameter of the screw to either side. This may quickly result in a situation where
the plate thickness must be increased, only because of the connectors. Integral joints
such as dovetails or through tenons are ideal for such constructions, allowing the
use of two layers of very thin plates instead of one layer of thick plates, keeping
the structure lightweight. A second major advantage of double-layered structures is
the possibility to put an insulation in between the two layers. This may considerably
reduce the cost of a design, because especially on folded or curved roofs, custom
3d-cut elements will be required, which can sometimes be more expensive than the
timber construction. Additionally, such 3d-cut insulation elements will typically be
made from fossil resources, while the insulation between two thin layers of wood
plates can be made from recycled paper, which is simply blown into holes with a
radius of approximately 100 mm, which are closed airtight with tape afterwards.
The first prototype for such a double layered timber folded plate was realized for
a formwork construction for a corrugated vault made from Ultra-high-performance
fiber reinforced concrete (UHPFRC), in a collaboration with the MCS group. It
was very similar to the interlocking folded plate, simply adding an offset second
layer at a constant distance of 20 mm, where the concrete was poured in afterwards.
While this was not a structural application of timber plates, it showed the geometri-
cal challenges in generating the offset, which would be important for the following
developments. Figure 5a, b show the first experiments in the development of a fully
integrally connected, double layered corrugated vault. A direct connection is estab-
lished between the plates of the inner layer and the plates on the outer layer at every
edge—through a penetration of the plates through one another using newly devel-
oped “Double-Through Tenon Joints” (DTTJ). This follows up on the principle of a
puzzle of mutually interlocking pieces, where the last element forms a “keystone”.
The entire assembly can only be put together in a specific sequence, and only be
disassembled in that exact reverse order of assembly.
7 Large-Scale Application
The application of a double-layered, fully integrally connected corrugated vault to

a building scale structure presented several challenges, which had to be solved and
optimized. In the geometry of the joints, the most important change is shown in
Fig. 6. Here, chamfers were added to allow for an easier insertion of the plates.
This is the case both for the upper and the lower part of the tenon, since it connects
with two layers of plates. The left side of the figure shows, how the contour of the
joint is defined parametrically in the algorithm through a center point p0, and ten
points which are calculated based on the plate thickness and distance between the
two plate layers. The right side shows the final joints cut on cross-laminated wood
plates. In contrast to the previous projects, the building scale project was realized not
1128 C. Robeller
Fig. 5 a First laser cut model of a double layered, integrally attached Folded Plate Barrel Vault. b
The basic units of a first laser cut model
Fig. 6 a Double-though-tenon joint DTTJ (Robeller et al. 2017a). b Full-scale CLT elements with
DTTJ (Robeller et al. 2017a)
with laminated veneer lumber (LVL), but with cross laminated timber (CLT) plates.
While the LVL plates are made from 3 mm thick, peeled veneer layers, CLT plates
are typically made from layers with a thickness of at least 20 mm thick wooden
boards. These boards are typically not glued along their sides, leading to a rather
lattice like structure in mechanical terms. For our integral, form fitting joints, these
gaps between the solid wood boards are problematic because their position cannot
be controlled and may end up in critical areas of the connections. This makes LVL
a generally more suitable material for integral connectors. However, the majority of
LVL plates is not available with high quality visual surfaces since they are typically
cladded or covered afterwards. The elements in Fig. 6b are made from CLT, which is
produced in a special process that allows the sides to be glued with one another. The
plates are therefore as homogenous as LVL plates, with an even superior mechanical
strength and a visual quality surface finish.
The second major problem in the building scale implementation of the system
was production times. While in the fabrication, time-consuming production was
not a problem, the time efficiency was crucial for the production of five hundred
large components for a full-scale structure. An efficient production, which would not
create additional cost compared to a state-of-the-art constructzion system, required
the cutting time to be drastically reduced by two thirds of the first cutting algorithms.
The solution to this problem was found in a hybrid strategy. First, it is important to
know that the most efficient tool for wood processing are saw blades. This is due
to the large diameter and the large number of blades, which combine a high cutting
velocity with a long cool down time in between cuts of the individual blades. Milling
tools however are always a compromise, but they are often necessary for certain
cuts. For example, the contours of plates with dovetails or through tenons will show
various concave corners, where only a milling tool can be used. Therefore, the first
CNC cutting algorithms for these joints were cutting the entire path with milling
tools. However, Fig. 6a shows that parts of the contour could still be cut with a saw
blade. For example, the segment at the tip of the joint, where two convex corners
follow each other. This basic principle was built into the algorithm, which would
now use a hybrid cutting strategy, first using the saw blade for all possible cuts, then
cutting the remaining contours with a milling tool.
Another challenge in the building scale implementation was presented with the
digital workflow. The complex shaped plate contours with all the angular tenons and
cutouts could not be manually programed with CAM software, which the timber con-
struction companies typically use for the production of simply shaped rectangular
elements. In the research laboratory, this problem was solved through the develop-
ment of an algorithmic tool for the automatic generation. Of the G-Code, directly
through a CAD application programing interface. This “offline generation” of the
G-Code puts the entire responsibility of the machine programing and operation in
the hands of the programmer, in this case an architect. The machine operator will
typically not be able to check the G-Code and can only start the machine and perform
an emergency stop, if necessary. In an industrial context, this is not an efficient strat-
egy, since it increases the risk of expensive damages to the machine, which will also
cause production delays, and it does not take advantage and include the experiences
technicians which usually operate their highly individual and relatively complex
machines. Additionally, a cutting file must typically contain multiple elements, to
reduce waste material. The plate elements must be nested on the raw material sheet
formats (which are usually given by the length of the CLT press and the width of
the CNC machine). Errors are not uncommon, so it is often necessary to re-cut one
of the plates, which will change the nesting and the remaining production planning.
Such cases clearly show the disadvantage of delivering static, pre-generated G-Code
files. Instead, an app was developed, which allowed the technician at the workshop
to generate the G-Code files himself, including a simulation of all machine motion,
which is shown in Fig. 7.
1130 C. Robeller
Fig. 7 A 5-axis CNC simulator, project-specific development for the factory
8 Application to Other Types of Shells
The previous prototypes were focused on corrugated and folded structures, especially
because plate-to-plate edgewise connections are easier to realize on folds and ridges,
where the dihedral angle between the plates is somewhere in the range of 45° to
135°. However, there are other structural typologies using plate elements, such as
segmental plate shells. Such structures and adequate topologies have previously been
studied with other materials such as glass. Tiling strategies have been extensively
studies in the field of computer science, offering a variety of useful algorithms and
methods. Generally, segmental plate shells allow for highly efficient lightweight shell
structure. The ideal static surface shapes, working with in-plane membrane forces
can be followed more closely than with folded shapes. However, the most critical
and weakest point in timber plate structures always lies in the connections (Fig. 8).
A fully integral joining system was presented with the double-layered timber plate
shell shown in the Figs. 9 and 10 (Robeller et al. 2017a). The system is an evolution
of the integrally attached folded structures, with the aim to construct a two layered
curved shell from box-shaped timber elements. In particular, the aim was to avoid
double-plates in the vertical layer, as it would happen when typical 6-sided boxes
out of plates would be assembled. Instead, this new system is consisting of boxes
with only four sides. It is possible to connect these plates completely integrally,
through a new detail which connects three plates in one joint. Here, the intersection
Fig. 8 A first building scale implementation of a double layered, integrally attached corrugated
vault from timber plates (Robeller et al. 2017a)
Fig. 9 A fully integrally attached system for a double-layered segmental plate shell structure
(Robeller et al. 2017b)
space of the three plates is divided equally between the three neighboring plates.
An algorithm helps to optimize this division process automatically. After a series of
small scale prototypes, a building scale implementation of this novel construction
system is currently on the way.
1132 C. Robeller
Fig. 10 First prototype using 15 mm birch plywood plates (Robeller et al. 2017b)
9 A Toolkit for the Design of Timber Plate Structures
In the development of these integrally attached structures, many geometric functions

are required, which are currently not found in any CAD package. Instead, the func-
tions were custom built for these structures, for their design, analysis and fabrication
with various computer-controlled machines. Selected functions from this computa-
tional library for timber plate structures has been made available for download via
the food4rhino app store. For example, the package allows to create corrugated vaults
based on NURBS surfaces, including the generation of the 3D plate geometries with
all angular cuts and different types of joints. The package also includes functions
for the generation of the G-Code, which can be adapted to work with a variety of
machines. It is also compatible with the KUKA prc app, which allows to use the
fabrication methods with KUKA industrial robots.
10 Technology for More Efficient and Sustainable

Structures
While this research is primarily focused on the exploration of new digital-technology-

enabled possibilities in the design of shell structures, it is also highlighting the largely
underestimated potential of wood as a high performance, sustainable building mate-
rial. Wood is a particularly important construction material, not only for its carbon
dioxide storage, but also for its outstanding weight-to-strength ratio and low energy
processing. With its low weight, it is easily transportable and therefore ideal for pre-
fabricated structures. We can take advantage of the already existing CNC technology
in the woodworking companies, while we are looking forward to the upcoming next
generation of on-site robotic fabrication.
Currently, building with timber is still slightly more expensive than for other
materials and requires particular expertise and skilled workmanship for the design,
dimensioning and especially the on-site assembly. The state-of-the-art alignment
and connection of timber parts with screws and steel plates is time consuming and
manual labor intensive. Still, the on-site assembly of wood constructions is much
quicker and cheaper than for other materials, but much of time that is won on site
is currently lost in the planning phase and cost, where wood is more complex, time
consuming and expensive than its competitor materials. Furthermore, form-fitting
wood-to-wood connectors are ideal for a simplified disassembly and for re-use of
structures, as well as for their recycling, where structures made solely from wood can
be easily brought into energy production (Lambert and Gupta 2004). In this context,
digitally planned assembly and disassembly sequences are very promising.
Finally, computational design is providing us the tools to make this change that
can make a big impact towards more sustainable building structures. It enables us to
manage the complexity of the material wood, as well as the challenging design and
fabrication of timber structures. We can exploit the potential of modern fabrication
technology and re-visit the wealth of knowledge in traditional woodcraft.
References
Lambert AJD, Gupta SM (2004) Disassembly modelling for assembly, maintenance, reuse, and
recycling
Makowski ZS (1963) Structural use of plastics in stressed skin construction. Appl Plas Fab, 47–52
Makowski ZS (ed) (1986) Analysis, design and construction of braced barrel vaults. CRC Press
Peters T, Peters B (2013) Inside smartgeometry: expanding the architectural possibilities of com-
putational design
Pizzi E, Piano R (2003). Birkhäuser
Robeller C, Weinand Y (2015) Interlocking folded plate—integral mechanical attachment for struc-
tural wood panels. Int J Space Struct 30(2):111–122
Robeller C, Gamerro J, Weinand Y (2017a) Theatre Vidy Lausanne—a double-layered timber folded
plate structure. J Int Assoc Shell Spat Struct 58:295–314
Robeller C, Konakovic M, Dedjier M, Pauly M, Weinand Y (2017b) Double-layered timber plate
shell. Int J Space Struct 32(3–4):160–175
Roche S, Robeller C, Humbert L, Weinand Y (2015) On the semi-rigidity of dovetail joint for the
joinery of LVL panels. Eur J Wood Wood Prod 73(5):667–675
Schineis R (2004) Gefalteter Klangkoerper Musikprobensaal Thannhausen/Thannhausen Rehearsal
Room. In: 10 Internationales Holzbau Forum (IHF), Garmisch-Partenkirchen
Stitic A, Robeller C, Weinand Y (2015) Form exploration of folded plate timber structures based
on performance criteria. In: Elegance of structures: IABSE-IASS symposium
Computationally Derived
Cross-Laminated Timber Reinforcement
and Construction
Jefferson Ellinger, Chris Beorkrem and Calum Dodson
Abstract Computationally derived Cross-Laminated Timber reinforcement lever-

ages advanced computation to design and manufacture hybrid panels of Glass-Fiber
Reinforced Plastic (GFRP) and Cross Laminated Timber (CLT). CLT is an emerging
technology for all scales of design because of its environmental and performative
benefits. The major factor affecting their adoption more holistically is spanning lim-
itations, which often result in limited building possibilities. In this study innovative
computational techniques are used to modify traditional CLT panels by removing
material where analysis suggests it is unnecessary and to tactically introduce GFRP
material where performance is most affected. The computational rethinking of the
composite make-up begins by looping design and analysis programs together during
design to strategically modify the panel for specific load conditions. This is followed
by the coupling of robotic manufacturing techniques at both the panel assembly and
modification levels in the pursuit of accurately and efficiently constructing the panel.
The designer ultimately modifies the unit assembly allowing for more formal free-
dom with increased performance characteristics. Additionally, the strength to weight
ratio of wood and GFRP versus more common materials suggests a desirable rela-
tionship of increased strength similar to reinforced concrete with only a fraction of
the weight increase.
Keywords Cross-laminated timber (CLT) · Glassfiber reinforced polymer (GFRP)

Finite element analysis · Structural simulation · Mass timber
Parametric structural analysis
J. Ellinger (B) · C. Beorkrem

UNC Charlotte, Charlotte, USA
e-mail: jellinge@uncc.edu
C. Beorkrem
e-mail: cbeorkrem@uncc.edu
C. Dodson
International Beams LLC, Sarasota, USA
e-mail: calum.dodson@ibewp.com

1136 J. Ellinger et al.
1 Introduction
Cross-Laminated Timber (CLT) is an emerging material technology in the mass tim-

ber category for the building industry gaining interest and popularity because of
its environmental and performance benefits over similar products. CLTs are con-
structed of small size dimensional lumber elements adhered to each other in alternat-
ing orthogonal layers. The young growth wood members are sliced and aggregated
into large panels that are capable of being used in many building applications and
approach the strength of traditional construction material systems such as steel and
reinforced concrete but with a substantial reduction in weight. Using CLT technol-
ogy to construct buildings also transforms what is a CO2 producing endeavor using
traditional materials such as steel and concrete to one that sequesters CO2. As an
example, CLTs are ¼ the weight of reinforced concrete when used in similar appli-
cations, while having a net reduction in CO2 rather than having concrete’s high net
production.
Projects constructed of CLT material logics have successfully taken advantage
of the material’s ability to integrate structure, cladding and interior finishing within
a single product, greatly reducing time and expenses when compared to alternative
material systems. However, existing CLT panels have limited spanning capacity due
to challenges in meeting the allowable deflection criteria for most building appli-
cations. This issue effectively limits their application to fewer building types than
do traditional building materials, typically in modular types of construction such
as hotels and small apartment structures. The research presented in this chapter
strategically introduces Glass Fiber Reinforced Polymers (GFRP) to the CLT panel
construction as a method for increasing the spanning capacity of CLT panels to open
more opportunities to take advantage of the benefits over traditional building mate-
rials that CLT offers. By computationally deriving the custom panels specifications
to their loading conditions the amount of GFRP used and the embodied energy of
production can be reduced.
1.1 Computational Approach for Anistropic Materials
Manuel de Landa has proposed a design strategy which begins with the conceptual or
cerebral genesis of form, imposed on materials (De Landa 2001). This design philos-
ophy treats materials as active participants in the creation of form. This methodology
is more in tune with a holistic energy-conscious design strategy. De Landa begins to
organize and classify materials based on their behavior and the characterization of
their phase transitions. If we can create a logic or family tree of material behavior, we
can better understand the embodied energy, the intricate properties of that material,
and how those materials might work within a larger infrastructure.
De Landa suggests that the development of isotropic materials through intellectual
transfer of knowledge through written text allowed for the holistic calculation of the
Computationally Derived Cross-Laminated Timber Reinforcement … 1137
material performance. In other words, the mass production of homogeneous materials

developed over the ages gave rise a top down approach to design because the material
performance within the manufactured form and assemblies could be easily predicted
through rudimentary calculation. This kind of predictability led to a material design
culture where standardized material elements became the units of design, a kind
of metascale kit of parts. The alternative method for design that De Landa’s idea
is that by using a ‘craftsman-like’ approach associated with design methods where
impurity, imperfection or heterogeneity in material construct (anisotropic) is used
to inform the ultimate form the developed material would take on. This argument
primarily resides as a suggestion for how to approach design problems in general.
By understanding the inherent differences within the space the designer is working
within new solutions that have incorporated those differences will achieve unique-
ness unavailable to alternative design approaches. De Landa posits that this method
of understanding material behavior in the design process lost favor because this kind
of knowledge was transferred through an apprenticeship model, one which necessary
lost favor as advanced technologies such as the printing press allowed for the transfer
of knowledge through intellectual means. However, through advanced computation
and modern technologies available to us now we can once again embrace the nuanced
complexities only available through apprenticeship previously. With computational
feedback loops the anisotropic material sensibility can now be used to take on the
design of the material assembly itself. This method would presuppose a bottom/up
approach where an understanding of a non-uniform material would influence the
design of the material itself to meet the performance requirements for a specific
condition, in the case of this study, span and load conditions. This idea would lead
to a material assemblage that has both a material performance advantage over stan-
dardized materials while also affording an aesthetic alternative to the kit of parts
approach to building design.
2 Benefits of Mass Timber
Interest in mass timber construction stems from two general issues. The first is climate
change, where the building and construction industry has contributed substantially
to its degradation. The first generation solutions to this problem have been to apply
bandaids as supplemental technologies to the building, adding to the complexity and,
ironically, the initial carbon footprint with these expensive technologies as much
of the world sits comfortably “atop 300 years of flaming carbon” (Mann 2014).
The second issue is the global trend towards urbanization. According to the United
Nations, 54% of the world’s population lives in urban areas. This proportion is
projected to be 66% by 2050 (United Nations 2014). To meet the building demand
to house and provide services for this population as it migrates to the urban centers
around the world incredible amounts of carbon will be emitted. Instead of ‘business
as usual’ construction techniques where mining, manufacturing, transportation of
raw materials to building sites continues to stress the environment and new regional
supply chains that are instead sequestering CO2 could be deployed (Organschi).
The technology originated in Europe where it is more commonly used than in the
US. This is quickly changing; however, as the Timber Innovation Act was reintro-
duced in the U.S. House and Senate on March 7, 2017. This proposed legislation will
support research for timber buildings over 85 ft. in height. The major factor affect-
ing their adoption more holistically are its spanning limitations, which often result
in compromised formal possibilities. In “Twenty-First-Century Timber”, Andrew
Waugh provides further reasoning for the use of mass timber in urban environments.
“Pound for pound, it is four times stronger than steel. Buildings in timber rise quickly
and accurately.” Waugh concludes the essay by discussing how mass timber can affect
future architectural discourse.
2.1 Emergence of Mass Timber as a Building Material
“The tectonic expression of mass timber is still developing. New technologies

demand new forms. For centuries, material innovation has been a means with which
to revitalize the traditions of architecture.” The phenomena of mass timber will
continue to revolutionize the architecture and construction industry with presently
unclear limits (Waugh 2014). Mass timber is a competitive solution for developing
urban environments in the 21st century because of its ability to simultaneously act
as structure, fireproofing, and cladding, all while improving environmental bene-
fits, ease of construction (related to fabrication, assembly, time and cost) and for its
aesthetic material characteristics.
The Murray Grove Apartment building also by Waugh Thistleton located in East
London was the first timber high rise in the world, completed in 2009 (Fig. 1). The
architect, estimates that the wood in Murray Grove stores 186 tons of carbon while
the steel and concrete for a similar, conventionally built tower would have generated
137 tons of carbon dioxide during production (Waugh 2014). Using wood in this
instance actually nets a savings of 323 tons of carbon from the atmosphere.
Alan Organschi has researched the consequences in a proposed project entitled
‘Timber City’ which predominantly uses CLT construction, where possible, in the
construction of all the buildings to meet the projected habituation demands predicted
in the near term. According to their research, if typical steel and concrete build-
ing construction systems are replaced in the new city with standard CLT products
4720 tonnes of CO2 would be sequestered rather than 3210 tonnes CO2 produced
(Organschi). This net savings of 7930 tonnes CO2 could have substantial impacts on
the environment.
These projects suggest that if there is substantial investment in using renewable
mass timber in the construction of buildings a large amount of CO2 can be sequestered
in the built environment itself. Both of these projects use standard CLT construction
panels, which is limited to mostly building types that have a cellular spatial model;
apartments, hotels, etc. With the clear sustainable benefits afforded by using con-
Fig. 1 Murray Grove Apartments, and the diagrams of the assembly methods and construction
details using mass timber. From Author
ventional CLT panels the question is, can the panel itself be modified to achieve
performance capabilities to be used for more building types, therefore expanding
this impact into a whole host of new building programs and scales?
2.2 Material Specific Design Logic
The translation of wood products into isotropic materials that can be calculated with
degrees of certainty required for building construction are important to issues of life
safety, etc. but have been implemented for a century without awareness for the inef-
ficiencies this creates in the production of the built environment. The combination of
CLT and GFRP, due to the aforementioned advantages over traditional construction
systems, seeks to leverage the computational and technological advances now avail-
able to develop a technique for reinforcing the CLT panel with anisotropic properties,
not unlike those found in wood naturally. Further reductions to the embodied energy
of the system can be achieved by customizing panels to their relative loading and
assembly conditions.
Currently, CLT panels are constructed of alternating layers of orthogonally laid up
standard dimension boards of timber. Like standard plywood layups, CLT’s consist
of two strength directions, the 0° layer (major strength direction) layer and the 90°
(minor strength direction). By alternating the assembly in both directions an isotropic
character is generated in the material assembly, in that every panel assembly can be
understood to behave within an acceptable tolerance, and to build up a dimension
larger than would be possible to grow naturally, especially in a sustainable method
using new-growth timber. However, even with this manufacturing method a current
limitation to traditional CLT panels is the span capacity; maximized spans are roughly
Fig. 2 Patent diagram of a CLT/GFRP panel. From Author
25 ft, which is far less than typical commercial demands for the majority of building
types.
This project defines a computational method for tuning the assembly of the CLT
using two custom methodologies. First, by removing material where it is unnecessary
for the load condition not unlike the shaping of steel into an ‘I-section’ moving mass
to where it is more effective. A second means is to add Fiber Reinforced Polymeric
materials (FRP) to the assembly in specific locations to increase the capacity of
the combined material composition (Fig. 2). By combining these two approaches
an increase in their use through more adaptable material design can be executed.
Here this is achieved through the development of a digital technique of structural
evaluation and integration with BIM software to facilitate ease and efficiency in
design and manufacturing.
2.3 Composite Material Assemblage
Wood is a natural polymer comprised of a cellulose structure intended to absorb mois-

ture. However, this natural polymer does not have an advantageous strength to weight
ratio as compared to other engineered materials (Fig. 3). With their similar molecular
structure synthetic polymers can be used to ‘strengthen’ wood material, making the
new assembly more desirable as structural material. The polymers in GFRP absorb
and adhere to wood to form a strong bond because of the very nature of both of these
materials. Research conducted by the University of Maine in 1998 demonstrated that
laminated wood products reinforced with GFRP could significantly increase strength.
Fig. 3 Diagram of strength ratios of various materials. From Author
For example, a 2% GFRP reinforcement can increase the strength of laminated wood
beams by 70% (Dagher et al. 1998). A 1992 research paper published in the Journal
of Materials in Civil Engineering states that, “Advanced forms of wood construction
can enable contemporary wood structures such as bridges to be competitive with
those constructed from any other conventional material” (Nikolaos and Triantafillou
1992). Existing research claims that, “Fiber-Reinforced Polymers (FRPs) associated
with glulam beams provide significant gains in terms of strength and stiffness, and
also alter the mode of rupture of these structural elements” (Fiorelli and Dias 2006).
Despite these advances Dagher et. al. note that; While timber has been success-
fully reinforced over the past few decades using various materials and reinforcing
techniques, very few of these methods of reinforcing timber have reached the com-
mercial market (Bulleit 1984). As argued by Bulleit et al. (1989, p. 433), there are
several possible reasons for this lack of commercialization of reinforced timber: (l)
The material used to reinforce the wood was not commonly used in building applica-
tions; (2) the reinforcing material was too expensive; and (3) the fabrication required
an additional and, thus, cost-increasing step in the laminating process (Dagher et al.
1996). As the introduction of composite materials have entered the building space
over the last decade more readily, a revisit of the possibilities for their introduction
should be studied.
In a comprehensive study comparing various materials to reinforce wood products
with rods (Steiger et al.) concluded that FRP performs better than steel materials due
to higher tensile strength and its advantageous strength to weight ratio coupled with
corrosion resistance. Additionally, they state that the lack of yield strength is not an
issue as each material has a similar modulus of elasticity. It should also be noted that
the FRP rods in the study were cast and then bonded to the wood with a secondary
adhesive.
Little research exists which explores the use of GFRP reinforcement in planar
CLT panels, which could provide significant gains in terms of strength and stiffness
much like their application in glulam products. Looking at the strength to weight ratio
of wood and GFRP versus more common materials suggests a desirable relationship
of increased strength similar to reinforced concrete but with only a fraction of the
weight increase.
Despite these advances Dagher et al. note; the cost benefit of adding in FRP
reinforcement needs to out weight the material savings that can be accommodated
by removing material from the overall profile. This is where the commercialization
of these products will become more viable. Additionally, they note that the locations
where the FRP and wood interface would need to be monitored for fatigue due to
weathering overtime.
3 Control Samples and Testing
To determine the efficacy of using GFRP as a reinforcing material in a CLT panel

two panels were constructed as control samples for load testing under ASTM D790
standard test methods for flexural properties. In both cases a three layer CLT panel
was constructed using 0-90-0, 6 × 1 lay-up using the same dimensional lumber (for
these prototypes the planks were 1.125 by 3.00 ). Load Case 1.1 is the all wood
assembly and GFRP Load Case 1.2 is the GFRP reinforced assembly. The GFRP
Load Case 1.2 has a uniform layer of unidirectional GFRP laminated to the bottom
of the CLT panel assembly. The panels were loaded under requirements stipulated
in the ASTM D790 and recorded as control samples against which the optimized
panels produced using the computational structural design method could be tested
and compared (Fig. 4).
The two samples were produced and subjected to a uniform 250 lb/ft2 load (Fig. 4).
The total deflection of the member was recorded as a measure of the structural
performance of the system. The result demonstrates that Load Case 1.1 deflects a total
of 0.189 and GFRP Load Case 1.2, reinforced with GFRP, deflects 0.102 (Fig. 8).
Fig. 4 Manufacturing of the samples and load testing under ASTM D790 standards
GFRP Load Case 1.2 exhibits an 85% increase in structural performance under
load. The test results were enough to show that a single layer of GFRP reinforcing
substantially increased the capacity of the assembly with a 2% increase in weight
(66 lbs. for Load Case 1.1 compared to 67.2 lbs. for GFRP Load Case 1.2). With
these results the computational methods were developed in an attempt to achieve
greater efficiencies.
3.1 Computational Design and Manufacturing
The examples below constitute a preliminary set of prototype CLT panels reinforced
with GFRP. Each of the prototypes was completed to explore the possibilities for con-
figuring GFRP reinforcement to increase the loads and spans possible using CLTs.
The locations of the GFRP were defined using a workflow integrated with Autodesk
Revit (Fig. 5), where the user specifies a span condition and models a specific load-
ing condition. The overall geometry of the panel to be designed is imported into
Dynamo (a scripting and automation tool for Autodesk Revit) where a layup of indi-
vidual wood plank elements are automatically generated, based on a user defined
plank sizes (1.125 by 3.00 ). This individual member geometry along with a prede-
fined adhesive layer are used later for structural simulation. The workflow is designed
to create a recursive loop which allows for each panel configuration to be analyzed
for tensile strains. After the first load simulation and analysis is performed the algo-
rithm recommends locations and sizes for GFRP strands to be placed. The panel is
redefined with GFRP in those recommended locations and is designed to reanalyzed
the reconfigured assembly.
Fig. 5 Computational workflow diagram. From Author
Technically, the first iteration within the recursive loop develops a three-
dimensional geometric representation of the stresses within the traditional wood
panel. This generic layup is processed by Autodesk Simulation Mechanical’s finite
element analysis (FEA) solver by simulating the anisotropic material properties of
wood and adhesive. The analysis generates three dimensional voxel-based data rep-
resenting the stress profile of the part. This voxel data is then transferred back to a
Dynamo script through its application program interface (API) (Fig. 6) to be used to
inform the modification of the panel itself.
Using the imported data from Simulation Mechanical the Dynamo definition then
iteratively adds GFRP reinforcing where the tensile strains are the highest. The new
model of the panel, now consisting of a composite structure of GFRP and wood
is then repackaged and imported into Simulation Mechanical to be reanalyzed and
simulated with the new material configuration. The algorithm is designed to adapt
to any panel geometry with any loading condition to the unique stresses simulated
within the structural analysis software (Fig. 7). As a final step, the geometry is output
Fig. 6 Interpreting the imported Voxel data from simulation mechanical in dynamo to inform the
location of the GFRP fibers and removal of wood material. From Author
from Dynamo, packaged and imported into RhinoCAM and a tool path is generated
to be used to modify the physical CLT panel using a robotic or five-axis CNC mill.
The manufacturing of each panel is informed by the final output of the algorithm
depending on how much of the panel is to be altered. In the test case submitted here the
recommended output defined manufacturing the CLT using conventional processes
while strategically removing unnecessary cross members that were incurring no
force. The output G-code (CNC) is then used to cut troughs into the traditional
CLT panel using CNC milling robots and the resulting surface and ridges are filled
with unidirectional glass fiber reinforcement and polyester resin polymer (the GFRP
specifications).
3.2 Digitally Modified Panel Load Tests
This assembly was tested before (Load Case 2.1) and after (GRP Load Case 2.2) with
physical load testing (Fig. 8). The computationally informed and edited panels were
loaded again using ASTM D790 standard test methods for flexural properties. The
deflection of the panels were measured after the load was applied for the requisite
amount of time and compared to the previously tested control samples. As expected
the unreinforced all wood modified sample, Load Case 2.1, had the largest deflection
at 0.284 a 54% increase in deflection over the traditional panel, Load Case 1.1.
Load Case 2.1 was 21.2 lbs. lighter than the control sample Load Case 1.1. The
GFRP reinforced sample, Load Case 2.2, was tested and as expected it performed
better than the traditional CLT control sample panel, Load Case 1.1. The final GFRP
reinforced and modified test sample, Load Case 2.2, deflected a total of 0.16 and
weighs 53.2 lbs. This result indicates a 14% reduction in deflection or a 18% increase
in the load bearing capacity while being 25% lighter than the all wood control panel,
Load Case 1.1. When compared to the GFRP reinforced control panel, Load Case
1.2, the modified panel, Load Case 2.2, is 26% lighter but deflects 60% more.
Fig. 7 Several examples of resultant surfaces to be milled generated from different loading condi-
tions; from simply supported on left, one-way cantilever middle and two-way cantilever right
Fig. 8 Physical load testing and results. From Author
3.3 Observations
The results of the testing indicate that the benefits gained in terms of both a decrease in
deflection when tested and a reduction in weight as manufactured show the potential
for the modified CLT panels to be used in more building structure applications.
The suggestion is that the 18% increase in load bearing capacity could result in an
increase in the clear span lengths when compared to traditional CLT products. Also
when comparing the all wood samples to one another (Load Cases 1.1 and 2.1) the
removal of the material at the interior level of the assembly may have more impact
on the structural capacity than the simulation is suggesting and more study of that
aspect of the assembly might reveal a more nuanced solution to the removal of this
material. The study carried out here was primarily a study of the location and tuning
of the GFRP material in the assembly. This method approaches the creation of CNC
panels more as a robotic butcher might, rather than trying to use any typical kind of
saw mill operation.
The modifications tested here have shown that multiple performance benefits can
be achieved through this computational approach to the material design, specifically
when compared to the typical CLT panel but also in relation to standard building
materials. Possible other configurations to investigate include expanding the test
conditions into larger panels with deeper profiles and more layers of assembly. This
could afford a variety of new explorations including the combination of both sheet
and trough based reinforcement, as well as more complex methods of positioning
GFRP within the panel, along a multitude of axis and dimensions. Within a larger
panel the removal of material can become scaled to the size of individual members
resulting in much more efficiency and in panels which may have a more elegant expo-
sure to the interior of the building, defining edges and shadow lines which follow
the structural profile of the panel. Additional types of resin and fiber reinforcing can
also be tested for a variety of performances. In addition to testing for the immediate
performance results, fatigue testing will need to be undertaken to ensure that like
the resin used in laminating traditional CLT panel the resins in the GFRP matrix
and bond to wood are not compromised over time. Due to the nature of the assem-
bled construction embedding sensors within the panel matrix could produce a more
complete understanding of the panel reactions, from shear bond strength between
GFRP and wood to internal stresses accumulated under loading in different areas of
assembly to be compared to the digital simulations.
4 Conclusion
By integrating structural calculations into the schematic design tools, a more inte-
grated workflow can be created for analyzing and creating custom structurally respon-
sive CLT panels. The use of composite structures and systems will continue to affect
the design processes of architectural systems, improving their durability, capability
and visibility. CLTs limitations, in particular their limited spanning capacity, inher-
ently impact the commercial applicability of this material. The method for connecting
these panels, whether floor-to-floor or floor-to-wall is to use large screws and angle
brackets, primarily for cost efficiency and speed. Considering the use of strategi-
cally designed GFRPs with integrated connections, utilizing CNC milling for timber
modifications, the formal variability of this method may be able to overcome this
limitation (Figs. 10 and 11). Implementing GFRP for reinforcing and connections
has significant architectural implications because the GFRP can become volumet-
ric and break away from the planarity of the CLT. In this application, the GFRP
establishes its own shape, inherently curvilinear and organic while the CLT remains
planar, offering a formal dialogue between two materials that work together to create
a monolithic system.
As the research progresses, the focus widens, and simulation tools become more
advanced, allowing for a total system analysis, the possibility to create holistic assem-
blies could create even greater efficiencies. Meaning, the tool not only includes CLTs
as floor panels but also walls and vertical structures, and will possibly no longer be
made of planar assembled elements. Once complete the variability and versatility of
this technology can potentially be much more significant because the approach will
no longer come from the perspective of traditional trabeated systems but rather, as a
monolithic structural system. GFRP is a material that structurally wants to be volu-
metric rather than planar or linear; it also works best as a monolithic system, mostly
because of the material properties of the resin. The resin is a thermoset polymer as
it forms an irreversible molecular bond when catalyzed, this bond is no longer a
connection but can be integrated with other polymer materials including GFRP and
wood member(s).
The integration of GFRP into CLT’s can also enhance other conditions inherent in
CLT technologies, namely their uniform coloring and texture and difficulty integrat-
ing mechanical systems within their profile. By expanding the geometric possibilities
and examining the locations within each profile that are over-structured, material can
be removed to create sectional volumes which can be used to move mechanical pip-
ing, electrical and plumbing. Additionally, owing to the transparent nature of GFRP’s,
Fig. 9 Example interior view of CLT/GFRP Panel with Integrated Lighting. From Author
their application to the underside of a floor can be doubly used to contain integrated
lighting components (Figs. 9, 10 and 11). If the GFRP reinforcement takes shape as
curvilinear hollow core surface, the GFRP and CLT can work together to facilitate
electrical and mechanical requirements as well as post occupancy analysis sensors.
For example, the integration of conductive tape into the panels could afford for real
time monitoring of their performance during typical building maintenance schedules
and through seismic or wind load events.
The study described here suggests that using an anisotropic approach to design
for CLTs is possible due to the recent advances in computational analysis and sim-
Fig. 10 Parametric concept of simply supported panel. From Author

Fig. 11 Parametric concept of column bay system. From Author
ulation. Having the performance simulation available as immediate feedback to the

designer in a ‘schematic’ design phase makes it possible to integrate these kinds
of approaches to design. With this technique a savvy designer would use the mate-
rial design outcomes to immediately impact the overall form of the architecture.
Lastly, with an extension or modification of the technique suggested here the logic
could be applied to a number of material design applications where multiple perfor-
mance criteria can be solved as a single solution. Through that holistic approach as
DeLanda was suggesting, the use of advanced computation would facilitate a return
to the craftsman-like approach to design towards a more ecologically responsible
architecture.
References
Bernheimer A (2014) Timber in the city design and construction in mass timber, First edn, ORO
Editions
Dagher HJ, Shaler S, Abdel-Magid B (1998) Ultimate strength of FRP-reinforced glulam beams
made with douglas-fir and eastern hemlock, 19 Jan 1998, 1
Dagher HJ, Kimball TE, Shaler SM, Abdel-Magid B (1996) Effect of FRP reinforcement on low
grade eastern hemlock glulams. In: National Conference on Wood Transportation Structures,
Madison, WI, General Technical Report No. FPL-GTR-96
De Landa M (2001) Philosophies of design: the case of modelling software. ACTAR, VERB,
Processing, ACTAR, pp 130–14
Fiorelli J, Dias AA (2006) Fiberglass-reinforced glulam beams: mechanical properties and theoret-
ical model. Mater Res 9(3):263
Mann C (2014). How to talk about climate change so people will listen. The Atlantic, 1
Organschi A (2017) Timber CITY. http://timbercity.org
Plevris Nikolaos, Triantafillou Thanasis (1992) FRP-reinforced wood as structural material. J Mater
Civil Eng 4(3):300
Stepinac M, Rajcic V, Steiger R, Serrano E, O’Neill C, McPolin DO, Widmann R (2015) Strength-
ening of timber structures with glued-in rods, vol 97
United Nations (2014) World urbanization prospects, the 2014 revision highlights. The UN Depart-
ment of Economic and Social Affairs, 7
Waugh A (2014) Timber in the city, essay: in twenty-first-century timber. First edn, ORO Editions,
41
Beech Wood for Architectural Design:
Three Studies Case from an International
Design Contest Terres de Hêtre®
Giuseppe Fallacara, Antonio Pantaleo and Giuseppe Scaltrito
Abstract The present contribution proposes three projects and relative speculative
reflections, elaborated by the writers, concerning the important theme of the wooden
roof construction, showing the results of the three competitions organized by the
Terres de Hêtre® consortium, entitled: “L’architecture au service de construction en
bois de hêtre des vosges”. The three projects, object of the paper (awarded with the
third placement during the first edition and with the second placement during the third
edition), describe an effort to demonstrate new applications, characteristics, shapes
and unexpected unexpressed potentials of this traditional material for a new contem-
porary design; they are architectures that have the common characteristic of being all
a great challenge to the intrinsic limits of the wooden material: in all the outcomes
results the research of geometric complexity and the limit of the static-mechanical
resistance of the elements is evident. The architectural forms are developed starting
from a series of geometric and topological variations on the theme of the wood-frame
roof: the first with powerful overhangs on the model of the gull-wing structure, the
second on the tree-shape model and the third on the shell-ruled–surface model.
Keywords Wood design · Cross laminated timber (CLT) · Flexible wood
1 Introduction
Between 26th and 28th December 1999 two storms of exceptional violence quickly
crossed half of Europe and in particular France, from west to east; the most serious,
G. Fallacara (B) · G. Scaltrito

Polytechnic University of Bari, Bari, Italy
e-mail: giuseppe.fallacara@poliba.it
G. Scaltrito
e-mail: giu.scaltrito@gmail.com
A. Pantaleo
University of Bari, Bari, Italy
e-mail: antonio.pantaleo@uniba.it

1152 G. Fallacara et al.
incalculable damages were suffered by forests and the signs of the passage of the
two extra-tropical cyclones are still evident in the country: the fallen trees covered
a volume of 146 million cubic meters, while one million of hectares of forests suf-
fered immense deterioration. Since this moment and after the consequent collapse
of the furniture sector in the Vosges, beech has become a less appreciated species;
currently this resource is insufficiently valued in light of the needs of local industry:
despite undeniable qualities such as consistency and color, professionals perceive the
intrinsic constraints of this essence as its low natural durability or its large dimen-
sional variations. The forest represents 42% of the territory of the Pôle d’Équilibre
Territorial et Rural du Pays d’Epinal Coeur des Vosges (PETR) and consists mainly
of hard woods, but also has a significant portion of conifers (Fig. 1); it is a vast
repository of timber, an essential raw material for local industry, which offers most
of the many rural jobs that are essential for maintaining economic activity in PETR.
The map highlights the fairly homogeneous distribution of the forest on the territory
of the Vosges: the whole PETR is affected by these problems and therefore several
measures have been taken into consideration to practice a sustainable and respectful
management of all forest functions. Recent attention to environmental issues has led
to a reconsideration of wood as a suitable material for building practice (PETR du
Pays d’Epinal Coeur des Vosges 2016). Wood as a natural resource, present but not
widely used, and as a material for innovative building components and systems, is
promoted and used as an alternative to the use of building materials and systems that
are now consolidated in the construction sector. Today we ask ourselves, on interna-
tional level, on what the future of wood architecture could be and what could be the
best way to go towards exploiting a resource that has always been present in our civi-
lization, but not long considered a real alternative to the consolidated way of building
(Spinelli 2009). For this reason, PETR created and launched the Hêtre des Vosges®
brand in November 2011, which marked the birth of a rural pole of excellence in the
wooden compartment: the creation of the brand brought all the stakeholders in the
sector together in autumn 2012, marking the collective desire to enhance beech wood
beyond the single securities sector. In the spring 2013, PETR transformed the sim-
ple collective label Hêtre des Vosges® into Terres de Hêtre® , a real economic entity
created and developed for professionals (private and public producers, transformers,
planners) who engage in the process of valorization of beech wood from the French
forest heritage and sustainable management. The new economic approach of Ter-
res de Hêtre® has the strategic objective of developing a network of innovation and
marketing of products derived from beech and therefore optimizing the evaluation
of this hardwood in the construction and furniture market. More generally, the actors
of this innovative entity are committed to: (i) propose and practice dynamic forestry
to improve the quality of beech wood; (ii) assure raw materials for the processing
companies by developing sales through supply contracts; (iii) promote the marketing
of products; (iv) guarantee the development of local companies of primary and sec-
ondary transformation; (v) strengthen research to develop new markets; (vi) increase
design training in the construction and furniture sectors; (vii) increase awareness
among architects, design offices and control offices (PETR du Pays d’Epinal Coeur
des Vosges 2016).
Beech Wood for Architectural Design: Three Studies Case … 1153
Fig. 1 Zones forestières sur le Pays d’ Épinal, Coeur des Vosges
2 Main Characteristics of Beech Wood
Beech is an important and widely-used hardwood in Europe. Its hardness, wear-

resistance, strength, and excellent bending capabilities, together with the relatively
low cost, make this hardwood largely used in European market. Depending on soil
conditions, European Beech can grow to very large sizes, so that wide and long
lumber is commonly available for use. Beech is a semi-hard wood, not durable
(durability class 5), very sensitive to caries fungi, chromogenic fungi, molds and
insects such as termites, which can therefore only be used in class 2 (for beams
or internal coverings); it is also widely used in the construction of furniture and
furnishings (e.g. chairs, especially in folded wood, veneers and plywood, also folded
plywood, stairs and parquet with strips and mosaic) it is subject to high deformation
during drying and large dimensional variations during its utilization due to variations
in humidity. Beech wood is heavy (dry density 680 kg/m3 ) and hard (Brinell hardness
34 N/mm2 ) and unfortunately presents a very marked shrinkage. The poor stability
of the form when the humidity changes must be taken into account, especially in
the case of large sections. You work it easily and cleanly and thanks to the uniform
structure it is very good for milling, turning and notching. During drying, the beech
tends to sag and crack. For this reason, the stacking of the sheets and the drying
process should be carried out carefully. The technique of vaporization, concerning
the heating of wood, allows to overcome these problems: it is a process that consists
in subjecting the action of the saturated green wood vapor stacked inside a cell,
up to increase its temperature around 100 °C, to change the natural color of some
wood species (e.g. beech, cherry, walnut and various tropical species). The cells of the
dryers, being naturally adiabatic chambers absolutely watertight, can become an ideal
environment in which you can obtain saturated steam at all temperatures between 60
and 100 °C through an appropriate management of vacuum and temperature, without
the need to inject into the cell steam produced from the outside, but using the same
steam produced by the water extracted from wood, thus generating a process of self-
vaporization. The rate of temperature rise and heating time are important parameters
to be controlled. At the molecular level, the process allows a controlled degradation
of the cellulose and consequently a modification of the crystalline structure of wood.
In addition, lignin can also change through a cross-linking process: this type of
treatment determines a decrease in the hydrophilic character of the wood and a higher
resistance to fungi through the chemical modification of the molecules that compose
it. The main problem of this process is the decrease of the mechanical resistance
of the treated products that can be easily understood following the modification of
the molecular structure. After a suitable steam treatment, beech wood is cut and
leafed very well. In one meter long steamed pieces, the beech wood bends easily and
maintains the new shape. This has made possible for example the great worldwide
success of Michael Thonet, creator of millions of durable and cheap chairs (Fig. 2);
also not folded beech is used for cheap furniture. The light, almost white wood takes
on the well-known pink color through vaporization and drying. With the action of
light, the color turns to pale yellow. More and more often beeches form an optional
reddish brown heartwood with an irregular “starry” outline. Once described as typical
wood with a diffuse porosity, today it is known by some experts to be semi-diffused
porosity wood, since the pores in the late wood are less numerous and smaller. The
medullary rays are clearly visible in all the cutting sections and they characterize the
design of the wood, above all in the tangential section, where they appear in the form
of a few millimeters high thin spindles.
2.1 State of Art in Beech Laminated Veneer Lumber (LVL)
Typical applications of European Beech include: (i) veneers (usually as rotary cut
veneer for plywood boards, composite boards, press-glued multi-layer boards etc.);
(ii) wood for furniture (office- and school-furniture, period furniture, childrens bed-
rooms et al., chairs, tables, upholstery frames); (iii) wall panelling and parquet floor-
ing, construction wood for average wear used in motor and machine building indus-
try, in civil underground and surface engineering as well as hydraulic construction;
(iv) special wood for particle and fibre boards, pulp and paper industry, compressed
wood, sports equipment, barrels, crates, workbenches, handles, domestic appliances,
Fig. 2 Rocking chair made by Thonet-Mundus, number B 804 with feets support (in cata-
logue under chair no. B 829). Matus D, https://upload.wikimedia.org/wikipedia/commons/3/38/
Moravsk%C3%A1_galerie_02_-_Thonet-Mundus_k%C5%99eslo.jpg
components for shoes, wood shavings, toys, doorsteps, turned parts for stairs, wood-
block paving, parts for musical instruments. Other interesting applications include
the production of glulam and cross laminated timber. A novel and promising prod-
uct obtained from beech is the laminated veneer lumber (LVL). In this case, peeled
veneer layers of three mm in thickness are parallel- or cross-laminated and turned
into beams, boards, panels and flooring elements. The exceptionally high strength of
such material allows structures with significantly slimmer dimensions, compared to
softwood materials. The high surface quality that can be obtained makes this LVL ide-
ally suited for visible construction elements, while the cost-efficient manufacturing
technology places structures using beech LVL at similar price level as conventional
softwood structures. The main advantages of beech LVL are: (i) potential usage for
a variety of applications in timber construction, interior design, furniture and floor-
ing; (ii) high load-bearing capacity which reduces dimensions and cross-sections
in timber construction—that means more space, lower material consumption and
suitability for large spans; (iii) top-quality aesthetics and exceptional surface qual-
ity, with easy use as solid wood, and for visible construction elements; (iv) savings
connectors, and thus costs reduction, due to the high density and homogeneity; (v)
price-competitiveness through efficient manufacturing and economical use of mate-
rials. A comparison of average mechanical properties of beech, LVL beech and other
materials is reported in Table 1.
In addition, it should be also noted that, due to the current fundamentally ecologi-
cal conversion of forestry in Europe, the ratio of available quantities of hardwoods vs.
softwoods is expected to change considerably, and the available quantity of beech-
wood will increase in the next years (Dill-Langer and Aicher 2014), whereby the
problem of utilization, especially of the lower grades, should be properly addressed.
The widespread direct thermal utilization is not sustainable from an economical or an
ecological point of view. For this reason, structural beech wood and glued-laminated
beech is more and more introduced in the building sector (Aicher et al. 2001; Glos
et al. 2004; Freese and Blaß 2005). In 2011, a German technical building approval
(DiBt: Allgemeine bauaufsichtliche Zulassung (General building approval) Z-9.1-
679 on “BS-Holz aus Buche und BS-Holz Buche-Hybridträger (glulam made of
beech and beech hybrid glulam)” 1st issued (October 2009) was issued for glulam
made either entirely of beech wood laminations or as a hybrid of beech and softwood
laminations. However, structural products based on solid timber have some serious
drawbacks, which has limited their utilization so far. In particular, the pronounced
radial-tangential anisotropy of shrinkage-swelling properties and the very low dura-
bility cause problems in all conditions except if used in service class 1. Moreover,
the potential high strength capacity is only valid for high visual or machine strength
grades, which are not currently commercially available to a major extent. Moreover,
the significantly different machining and gluing properties of beech hardwood lead-
ing to problems in its integration into production processes that glulam companies
have laid out for handling softwoods is another major drawback for the widespread
utilization of beech wood in glulam. By processing beech round wood into LVL,
some of the mentioned drawbacks can be overcome or minimized. In fact, LVL
allows an homogenization of the year ring structure and thus there are less problems
with climate induced stresses and cracking. The lamination effect of veneer layers
enables the utilization of the high strength capacity of beech wood by additionally
using some fraction of lower grade material. A comprehensive compilation of the
major strength and stiffness properties of LVL made of European beech considered
Table 1 Comparison of bending strength, density, and modulus of elasticity of beech LVL in
comparison with other materials (Korkut and Guller 2008)
Material Bending strength Density (Kg/m3 ) Modulus of elasticity
(N/mm2 ) (N/mm2 )
Beech 30 680 13,000
Beech LVL 70 730 16,800
Spruce LVL 48 480 13,800
Glulam GL24 h from 24 385 11,600
spruce
Aluminium 120 2,700 70,000
Steel (S235) 235 7,850 210,000
here is given in (Knorz and Van Kuilen 2012). Product inherent, the panel product
LVL is mainly suitable for plate like structures. When used for beams or columns,
the cross-sectional dimensions are limited and can only be enhanced by further pro-
cessing of the panel product to glued laminated beam build-ups. In Tsoumis (1991),
it is shown that the assembly of LVL laminations to “glulam” cross-sections not
only overcomes the dimensional limits, but also enhances the load bearing capacity,
demonstrated by the compression strength parallel to fiber direction. Heat treatment
is known to improve the dimensional stability of wood by reducing hygroscopicity
and to enhance biological durability by reducing wood moisture absorption (Korkut
and Guller 2008). One of the main limitations of heat-treated wood is the decrease
in mechanical strength, making this wood unsuitable for most structural applica-
tions. On the other side, reinforcing technology is a feasible and effective method for
improving the mechanical properties of LVL, in particular when low quality wood
is used. Consequently, reinforced LVL can be an alternative to high quality solid
wood (Wang et al. 2015). In recent years, to reinforce laminated veneer lumber some
materials were used, such as glass and carbon fiber, aramid, natural fibers, steel,
aluminum, plastic, etc. Many researchers have studied the effects of carbon fiber on
the physical and mechanical properties of laminated wood or composite materials
(Wei et al. 2013). The bonding between the fiber material and the matrix is also
very important. The effectiveness of reinforcement depends on the bonding between
matrix and fiber material, making this a key factor in determining the final properties
of the composite material, especially its mechanical strength. In this case, the fiber
materials are responsible for the strength of the composites, while the adhesive is
responsible for the stress transmission. The fiber reinforcing materials technology
has been applied to a wide range of timber and glulam and also to heat-treated LVL.
In (Percin and Altunok 2017) the focus was on the evaluation of the effects of car-
bon fiber on some mechanical and physical properties of laminated veneer lumber
manufactured from heat-treated beech wood. Carbon fiber revealed a clear effect
on the improvement of mechanical (modulus of elasticity, compression and bending
strength) and physical properties (density, volumetric swelling) of heat-treated LVL.
3 International Design Contest Terres de Hêtre®
Pays d’Epinal Coeur des Vosges, in collaboration with the Communauté de Com-
munes de la Vôge, Vers les Rives de la Moselle and Terres de Hêtre® , launched, in
autumn 2012, a first architectural competition: the construction of eight residential
buildings made of beech wood (50–60 m2 ) on Lake Aître in Xertigny. The success of
this first edition (57 participants from all over Europe) led to the publication of a new
competition in 2014 that focused on the design of a cyclestation in the city of Epinal.
In 2017 a third competition was organized: the theme concerned the construction
of an intergenerational complex in the center of the village of Bouxières-aux-Bois.
The aim is to encourage the emergence of creative and innovative ideas and to show
everything that can be achieved with beech wood by exploiting its multiple qualities.
The three projects (awarded with the third placement during the first edition and with
the second placement during the third edition), describe an effort to demonstrate new
applications, characteristics, unexpected shapes and unexpressed potentials of this
traditional material for a new contemporary design; they are architectures that have
the common characteristic of being all a great challenge to the intrinsic limits of the
wooden material: in all the outcomes, in fact, the research of geometric complex-
ity and the limit of the static-mechanical resistance of the elements is evident. The
architectural forms are developed starting from a series of geometric and topological
variations on the theme of the wood-frame roof: the first with powerful overhangs
on the model of the gull-wings structure, the second on the “tree-shape” model and
the third on the striped shell model. The regulation provided for the maximum use of
beech wood for each type of element. All projects use beech wood to contribute to
the improvement of natural resources present in the area: forest products they must
participate in the dynamism of the local economic fabric, the maintenance of jobs
and the reduction of greenhouse gases.
3.1 Touristic Village Les ailes du Hêtre pour l’être
The first competition involved the construction of a group of wooden holiday accom-
modation (Figs. 3 and 4) with a high symbolic meaning for the client. In fact, in
addition to being a showcase for beech wood buildings, Xertigny is a city in which
the forest sector seeks to establish itself in a lasting way. The complex geometry of
Fig. 3 Village touristique “Les ailes du Hêtre, pour l’être”—plan, Fallacara G, Stigliano M, 3e
Prix du concours d’idées 2013 Terres de Hêtre
Fig. 4 Village touristique “Les ailes du Hêtre, pour l’être”—render, Fallacara G, Stigliano M, 3e
Prix du concours d’idées 2013 Terres de Hêtre
the structure, inspired by the wings of birds on the point of flying (as a wish for the
future glory of the village of beech), has created a markedly sculptural project,1 which
serves as a shelter under which to organize the functional program of the house; it has
come to life from a reflection around the static complexity of the gull-wings model,
consisting of punctual supports, from which parallel wooden arches depart and act
as a framework for the covering structure that is expressed in the double curved
shape of the two large wings (Fig. 5). The use of wooden rectilinear elements for the
structure of the wings represents an attempt to optimize and standardize the compo-
nents in the simplest way possible, in order to obtain complex morphologies in the
assembly phase (Fig. 6). The intervention has become an opportunity to reinterpret
the compositional vocabulary of the local rural heritage, soberly conjugated with the
current architectural language, proposing a simple and immediate building: a small
living container under a gull-wings cover. The image of the whole house is innova-
tive but in accordance with the principles of tradition (gable roof, natural materials
and respect for the local environment) that are updated here to give a new value to
beech wood (Fig. 4). The two fundamental geometries of the house are represented
by the ruled surface of the roof and the truncated cylinder that contains the func-
tions of living. The roof surface has been designed as a structure of wooden beams,
rectilinear and curvilinear, which fit together to create a spatial network. A wooden
arch placed near the symmetry axis unloads the forces on two foundation pillars.
The entire structure of the roof, therefore, would unload on the ground through these
two points, minimizing the dimensions of the foundations; the cylindrical geometry,
instead, rests on a circular wooden beam anchored to the structural foundations of the
house. The residence, longitudinally, is oriented along the north-south axis to maxi-
1 Villagetouristique “Les ailes du Hêtre, pour l’être”—Fallacara G et Stigliano M, Bari—Italy, 3e

Prix du concours d’idées 2013 Terres de Hêtre.
Fig. 5 Village touristique “Les ailes du Hêtre, pour l’être”—basic geometries and possible vari-
ations of architectural composition, Fallacara G, Stigliano M, 3e Prix du concours d’idées 2013
Terres de Hêtre
mize the contribution of solar energy, to the advantage of the large opening front, and
create a natural greenhouse effect for heating the house during the winter while it is
protected during the summer from the cover when the sun is high. The gull-wings
roofing project guarantees, in fact, to the interior rooms the greatest possible contri-
bution of daylight, allowing to insert along the perimeter some larger windows. This
also makes it possible to give more value to the interior spaces, which, even in the
modest dimensions in the plan, face outwards with greater airiness. The living spaces,
well lit by large windows to the south, are minimal and functional, and their interior
design is original and designed to be entirely made of beech wood. The interior space
is functionally divided into three parts: kitchen, dining room and living room to the
south, bedrooms to the north; between these two spaces a technical and storage axis
is included. The outdoor areas, covered and protected from the elements to the south
and to the north, are used to enjoy life in contact with nature. The heating system has
been designed with the use of a fireplace that can heat the whole house and create
a warm atmosphere in winter behind the large window facing the lake. The idea is
to draw on the local heritage that has some recurrent principles: in particular the use
of wood and pitched roof. The elevations are characterized by curved surfaces that
alternate between wooden walls and glazed surfaces. The dimensional compactness
and the shape with an excellent surface/volume ratio, represent the desire to achieve
the construction of a building that is highly performing from the energy point of
view, so as to minimize the amount of energy needed for its heating. The proposed
insulation materials, the heating system, the reasoning on thermal inertia, should
contribute to achieving a high level of environmental comfort without resorting to
fossil fuels.
Fig. 6 Village touristique “Les ailes du Hêtre, pour l’être”—wooden rectilinear elements for the
structure, Fallacara G, Stigliano M, 3e Prix du concours d’idées 2013 Terres de Hêtre
3.2 Cyclestation I Love vélo!
In the second case the structure of the building consists mainly of “trees” in lamellar
beech and micro-lamellar beech, which support a circular tensegrity system formed
by circular beams in lamellar beech, uprights in beech and post tension through
steel cables (Fig. 7). All the connections between the elements are designed to be
made without the use of glues or metal supports, but simply with wood-wood joints
(Figs. 8 and 9), recovering the traditional technique and using modern technologies
to make them (three-dimensional modeling of the joints and realization through
numerical control milling cutters). Cyclestation is, by definition, a particular type of
modern building that does not have a formal historical tradition. For this reason, it
can be considered an exception and a singularity in the architectural design process.
The proposal for this project2 stems from the desire to create, in the tourist village,
a unique architectural excellence that interprets the theme of nature and cycling.
The project involved a building surrounded by trees, integrated and protected by
nature. There are four main themes: (i) transparency and immersion in nature; (ii)
lightness and warmth of the building; (iii) tree-shape modular structure; (iv) use
of the principle of circular tensegrity for the construction of the roof; the main
structure of the building consists of two trees that support a circular tensile structure
formed by a circular beam in lamellar beech, four uprights in beech and post tension
2 Maisondu vélo “I Love vélo!”—Fallacara G, Stigliano M, Boccadoro N, Barberio M et Colella

M, Bari—Italy, Concours d’idées 2014 Terres de Hêtre.
through steel cables (Figs. 10 and 11). The tree is a “Y” shaped structure that is
repeated four times during assembly. The “trunk” is constructed from curved lamellar
beech beams (lamella thickness of two mm), while the “ramifications” of the tree
are made of curved laminated beech beams (lamella thickness of 5 mm). Therefore,
the branches are more flexible and therefore able to withstand the optimal dynamic
stresses of the structure. The connections and joints between the various parts of
the tree structure are made of specific elements of galvanized iron and mechanical
bolting. The building, totally immersed in its natural environment, consists of an
outer shell of transparent and translucent walls in Plexiglas with a honeycomb. This
condition allows you to use the greenhouse effect during the winter season and isolate
the building in the summer. The transparency of the building would allow to see from
the outside, the interior spaces and the main tree structure made entirely of beech
wood. For those who are inside the Maison du Vélo, the difference between interior
and exterior space is reduced to a minimum. Structural trees are totally integrated
into the host nature. The outdoor area would have been planted with beech trees
in order to protect the house. The transparency and lightness of the construction,
with reference to its external skin in contact with the atmosphere, was imagined
through the use of specific and innovative technical materials. The cover would be
made of technical fabric Ethylene-Etrafluoroethylene (ETFE) reinforced with beech
slats. This material is not expensive and offers a very high thermal performance.
The external façade, transparent and curvilinear, would have been realized with
cellular Plexiglas sheets, resistant and with high thermal performances. The upper
part of the curved facade is covered by a transparent photovoltaic adhesive in order
to contribute to the building’s energy needs. The natural ventilation of the building
would have occurred thanks to the mechanized opening of the central roof eye.
The roof of the cyclestation is a tribute to the invention of the bicycle wheel that
passes under the name of circular tensegrity. A definition that well explains this
architectural concept, is the one that Buckminster Fuller used to refer to this type
of system, that is “islands of compression within an ocean of tension” (Krausse and
Lichtenstein 2017). It is a very light and at the same time resistant structure consisting
of elements that work exclusively in compression (beech spacers) and elements that
only work in tension (post-tension steel cables). This structure keeps the ETFE fabric
cover. The project was developed using an innovative approach to bioclimatic design,
which led to a convergence of environmental sustainability objectives and economic
optimization, maximizing energy savings throughout the year through the use of
solar energy and high-performance skin building. Everything is defined by a single
term: eco-responsibility of the building. The key principles are: (i) compact shape
(the cylindrical surface has a good volume/area shape factor); (ii) free contribution of
the sun (exploiting solar radiation and the greenhouse effect during the winter); (iii)
sun shading (during the summer through mobile wooden elements); (iv) controlled
natural ventilation; (v) use of high thermal performance materials for the roof (ETFE
fabric membrane with double air chamber and aluminum blades that reflect sunlight
without heat accumulation).
Fig. 7 Maison du vélo “I Love vélo!”—structure, Fallacara G, Stigliano M, Boccadoro N, Barberio

M, Colella M, Concours d’idées 2014 Terres de Hêtre
Fig. 8 Maison du vélo “I Love vélo!”—plastic model, Fallacara G, Stigliano M, Boccadoro N,

Barberio M, Colella M, Concours d’idées 2014 Terres de Hêtre
Fig. 9 Maison du vélo “I Love vélo!”—plastic model, Fallacara G, Stigliano M, Boccadoro N,

Barberio M, Colella M, Concours d’idées 2014 Terres de Hêtre
Fig. 10 Maison du vélo “I Love vélo!”—section, Fallacara G, Stigliano M, Boccadoro N, Barberio

M, Colella M, Concours d’idées 2014 Terres de Hêtre
Fig. 11 Maison du vélo “I Love vélo!”—tensegrity system, Fallacara G, Stigliano M, Boccadoro

N, Barberio M, Colella M, Concours d’idées 2014 Terres de Hêtre
3.3 Intergenerational Center La dolce vita
The third case concerns the construction of an intergenerational center in the village
of Bouxières-aux-Bois. The project3 stems from the desire to combine aesthetic
complexity and structural simplicity through the use of beech wood. The architectural
structure consists of a circular plant (29 m Ø), with the presence of an Italian-style
garden that extends all around, from which one enters the center where the housing
units for the elderly are provided with an independent entrance for the disabled and a
classroom for the nursery school (Fig. 12). The choice of materials and construction
details are based on the use of beech wood for all components, from beams that
make up the roof structure, to XLAM panels, to wall and floor coverings, to door
and window frames. The fundamental elements are the large roof, with wooden
beams and metal cladding, the elliptical geometry in which the living spaces are
contained and the classroom that represents the heart of the center (Figs. 13 and
14). The project, as a whole, aims to capture the complexity of interactions between
children and the elderly: the goal is also to build a living community, in which
to combat the sense of loneliness that often guests of retirement homes can try.
At the same time, however, thanks to this intergenerational project, children have
3 Complexe intergénérationnel “La dolce vita”—Fallacara G, Pantaleo MA, Scaltrito G,

Acquafredda M, Reverte V, Bari—Italy, 2e Concours d’idées 2017 Terres de Hêtre.
the opportunity to develop a greater sensitivity towards people with disabilities or

specific pathologies. The roof is geometrically obtained starting from a complex ruled
surface using rectilinear beams of lamellar wood that make up the primary frame.
This transfers the loads to the load-bearing structure consisting of prefabricated,
self-supporting panels, in curved X-LAM (engineered wood product, made with
the innovative technology of the Radiusplatten (RAP) and composed of mutually
crossed and glued wooden boards, resting in appropriate housing, pre-shaped on
the panels in question; this allows a quick and very precise assembly, decreasing
the construction time thanks to the dry laying by means of mechanical connections
(Figs. 15, 16 and 17). In this case the technique of parametric kerfing was used for the
creation of flexible wood panels for architectural use: different patterns were carved
on the wood so as to “break” the structural lines (or fibers) allowing to bend the
same material (Figs. 18 and 19). In particular, etching grids on both sides make the
panels adaptable to three-dimensional surfaces similar to the intrados of the roof. The
intent of the project, in fact, was also that of exploring the bending capacity of wood
panels through digital methods of manufacturing and computational simulation to
evaluate the potential and possibilities offered. The elliptical geometry of the houses
has been designed to facilitate the fluidity of movements for people with reduced
mobility. Inside, the house has a large living area, which serves as a link between the
entrance and what can be defined as the service block of the house, which houses the
bedroom, bathroom, kitchen and technical room; also in this case it is an extroverted
house, built towards the outside, to attract heat and solar radiation, guaranteeing
ventilation and natural lighting of the main environment. The aim is to maintain a
constant internal temperature, thanks to a load-bearing structure that is so thermally
performing, innovative and easy to assemble and disassemble. The wall pack is then
completed with a layer of reflective thermal and acoustic insulation panels with a
very low emissivity value. This allows, during the cold seasons, the containment of
heat produced by heating, preventing the dispersion to the outside, and, during the
summer seasons, prevents heat from entering the house.
Fig. 12 Complexe intergénérationnel “La dolce vita”—plan, Fallacara G, Pantaleo MA, Scaltrito
G, Acquafredda M, Reverte V, 2e Concours d’idées 2017 Terres de Hêtre
Fig. 13 Complexe
intergénérationnel “La dolce
vita”—render, Fallacara G,
Pantaleo MA, Scaltrito G,
Acquafredda M, Reverte V,
2e Concours d’idées 2017
Terres de Hêtre
Fig. 14 Complexe
Terres de Hêtre
Fig. 15 Complexe
vita”—section, Fallacara G,
Terres de Hêtre
Fig. 16 Complexe
vita”—detailed section,
Fallacara G, Pantaleo MA,
Scaltrito G, Acquafredda M,
Reverte V, 2e Concours
d’idées 2017 Terres de Hêtre
Fig. 17 Complexe
Terres de Hêtre
Fig. 18 Complexe intergénérationnel “La dolce vita”—kerfing test, Fallacara G, Pantaleo MA,
Scaltrito G, Acquafredda M, Reverte V, 2e Concours d’idées 2017 Terres de Hêtre
Fig. 19 Complexe intergénérationnel “La dolce vita”—kerfing test, Fallacara G, Pantaleo MA,
Scaltrito G, Acquafredda M, Reverte V, 2e Concours d’idées 2017 Terres de Hêtre
4 Conclusion
The concept shared by all three projects comes from the desire to imagine great struc-
tural and aesthetic performance thanks to the use of wood that intrinsically possesses
qualities of strength and beauty. Although the three projects are specific responses
to defined contexts, they aim to establish new paradigms that can be generalized
to innovate and update the art of wooden construction. Given the characteristics of
strength and lightness, using laminated wood as a construction system, in this type
of projects, is certainly a favorable choice. The perspective of buildings made of
beech wood is ever closer: the competitions have been translated (Les Woodies of
Xertigny and Maison du Vélo of Epinal) or will translate (complexe intergénéra-
tionnel of Bouxières aux Bois) into the construction of operational buildings. The
idea behind the first competition was the creation of synergy between technical per-
formance and projects reproducibility. The construction, for 1.2 million euros, of the
seven houses took place entirely in local wood. The reproducibility of the structures
and the pre-assembly have been designed by the selected company, so as to lay the
foundations for an efficient standardization of the industrial process (Fig. 20). On
this site, four building systems were used: (i) experimental beech laminate for the
reception building (Fig. 21) with thermo-heated beech cladding and external skylight
(Fig. 22); (ii) pitched roof with large overhangs to protect the beech on the façade and
in the structure of the buildings called Chalot; (iii) visible internal reinforcements in
beech, external pine cladding and heat-treated beech roof panels for Kapla models
(Fig. 23); (iv) Tetris model walls made of beech wood that play a structural role but
also an internal layout role (Fig. 24). The tourist complex has become a technological
showcase for beech buildings but also for local wood construction. The project for the
cyclestation, on the other hand, seeks to respond to the need for a physical location
for the rental of bicycles and the need for innovation in using beech (Figs. 25 and
26). The horizontal and vertical elements are made of laminated beech wood. The
connections between the uprights and the beams are made by bolting (Figs. 27 and
28). A whole log of beech is the central pole of the structure (Fig. 29). The interior
is in beech with custom furniture and partitions, which give the building a unique
character. The beams and the floor of the mezzanine are also of beech to confirm the
warm atmosphere inside the Maison du Vélo. Outside, the beech is used as a coating
thanks to a heat treatment, which gives the wood a brown color that is well suited
to the landscape (Fig. 30). The studies on the wooden structure were carried out by
the Centre Régional d’Innovation et de Transferts Technologiques des industries du
bois (CRITT) and the structure was created by the company Passiv’Home (Remire-
mont, Vosges). All the interior furnishings were made by the Houillon woodworking
(Dompaire, Vosges). Terres de Hêtre® also acts as an incubator and promoter for
several start-ups that make it possible to produce new products deriving from the use
of local wood. Terres de Hêtre® has, in fact, also financed some studies conducted
by Laboratoire d’Étude et de Recherche sur le Matériau Bois (LERMAB) whose
objective was the feasibility analysis of a beech wall pack assembled through a spe-
cific preparation of the contact surfaces; this analysis should lead to the realization
of some patents. The overall appearance of the building package takes into account
that roofing elements or flooring can decline on the same principle. It should be
noted that this development has endeavored to remain strictly within the framework
of the construction regulations in order to make immediate access to the market. The
composition of the proposed building package (Fig. 31) aims to: (i) use the beech of
the Vosges but also other local species as needed, depending on availability; (ii) max-
imize the value of timber; (iii) reach the Passivhaus standard; (iv) use the principle
of an instant mechanical assembly with resistance equivalent to a gluing assembly;
(v) homogeneously distribute the forces in walls, floors and roofs with wood/wood
joints; (vi) feed the ongoing reflection on the complete robotization of the produc-
tion of this type of wall in order to reduce costs thanks to the mechanized assembly
developed for maximize its production flexibility. The environmental goals set for
the coming years are largely based on the decisive contribution of forests and tim-
ber (PETR du Pays d’Epinal Coeur des Vosges 2016). This sector is an example of
sustainable development and circular economy. From the forest to the industrial land-
fill, this sector, with almost 450,000 jobs in France, creates ecological, economic and
social value. Wood is a material that has become strategic because of its renewable,
recyclable nature and its contribution to the fight against climate change. Slightly
energetic in its highly insulating implementation, it is established as one of the basic
materials in the design and construction of buildings with high energy performance
and zero environmental impact. The architecture and contemporary design, discover
every day new virtues for this material with surprising modern features and in all this
Terres de Hêtre® represents a virtuous experience and a model to be imitated. The
entire territory, in fact, greatly benefits from the dynamics around the exploitation
of wood and establishing itself in a rural area has helped to promote the synergy
between urban and forest areas.
Fig. 20 Les Woodies—Chalot model, Xertigny, Scaltrito G
Fig. 21 Les Woodies—Frame mixed beech (structure)—spruce (siding) of the structure of recep-
tion of holiday houses in Xertigny, Remy Q
Fig. 22 Les Woodies,

Thermo-treated beech
cladding, Remy Q
Fig. 23 Les Woodies,

Terrace in heat-treated beech
and pine cladding—Kapla
model, Remy Q
Fig. 24 Les Woodies, Internal layout—Tetris model

Fig. 25 Maison du Vélo, Epinal, France, Remy Q, Hamard JF
Fig. 26 Maison du Vélo, Epinal, France, Remy Q, Hamard JF

Fig. 27 Maison du Vélo—bolting, Epinal, France, © 2018 Terres de Hêtre

Fig. 28 Maison du Vélo—bolting, Epinal, France, © 2018 Terres de Hêtre

Fig. 29 Maison du Vélo, Epinal—log of beech, central pole of the structure, France, © 2018 Terres
de Hêtre
Fig. 30 Maison du Vélo—heat treated beech coating, Epinal, France, © 2018 Terres de Hêtre
Fig. 31 Proposal of a standard wall composition in cross beech strips, © 2018 Terres de Hêtre
References
Aicher S, Höfflin L, Behrens W (2001) A study on tension strength of finger joints in beech wood
laminations. Otto-Graf J 12:169–186
DiBt: Allgemeine bauaufsichtliche Zulassung (General building approval) Z-9.1-679 on “BS-Holz
aus Buche und BS-Holz Buche-Hybridträger (glulam made of beech and beech hybrid glulam)”
1st issued (October 2009): Approval holder: Studiengemeinschaft Holzleimbau e.V., Wuppertal,
Germany (2011)
Dill-Langer G, Aicher S (2014) Glulam composed of glued laminated veneer lumber made of beech
wood: superior performance in compression loading. In: Aicher S, Reinhardt HW, Garrecht H
(eds) Materials and joints in timber structures. RILEM Bookseries, vol 9. Springer, Dordrecht
Freese M, Blaß HJ (2005) Beech glulam strength classes. In: Proceedings meeting 38, CIB working
commission W18-timber structures, paper CIB-W18/38-6-2, Karlsruhe
Glos P, Denzler JK, Linsenmann P (2004) Strength and stiffness behaviour of beech laminations
for high strength glulam. In: Proceeding meeting 37 CIB working commission W18-timber
structures, paper CIB-W18/37-6-3, Edingburgh
Knorz M, Van Kuilen JW (2012) Development of a high-capacity engineered wood product—LVL
made of European beech (fagus sylvatica L.) In: Quennville P (ed) Proceedings world conference
on timber engineering, pp. 487–495, Auckland
Korkut DS, Guller B (2008) The effects of heat treatment on physical properties and surface rough-
ness of red-bud maple (Acer trautvetteri Medw.) wood. Bioresour Technol 99(8):2846–2851
Krausse J, Lichtenstein C (2017) Your private sky R. In: Buckminster Fuller: The Art of Design
Science. Lars Müller Publishers, Zurich
Percin O, Altunok M Eur (2017) Some physical and mechanical properties of laminated veneer
lumber reinforced with carbon fiber using heat-treated beech veneer. J Wood Prod 75:193
PETR du Pays d’Epinal Coeur des Vosges (2016) Charte forestiére de territoire. Socosprint
Imprimeurs, Epinal, November, p 2016
Spinelli A (2009) Costruire in legno nell’architettura contemporanea—Il sistema costruttivo a pan-
nelli di compensato strutturali di legno regionale. Un contributo per la valorizzazione della filiera
piemontese. Dissertation, Politecnico di Torino
Tsoumis G (1991) Science and technology of wood: structure, properties and utilization. Chapman
& Hall, 115 Fifth Avenue, NY, US
Wang J, Guo X, Zhong W, Wang H, Cao O (2015) Evaluation of mechanical properties of reinforced
poplar laminated veneer lumber. BioResources 10(4):7455–7465
Wei P, Wang BJ, Zhou D, Dai C, Wang Q, Huang S (2013) Mechanical properties of poplar laminated
veneer lumber modified by carbon fiber reinforced polymer. BioResources 8(4):4883–4898
Brise-Soleil House: Developing Software
Tools to Enable the Integration of Design
and Fabrication in Timber
Jonathan Nelson, Chris Knapp, Rory Spence and Joel Hutchines
Abstract This chapter presents a system for seamlessly integrating design, fabri-
cation, and assembly of a geometrically complex timber veil for a house in Port
Moresby, Papua New Guinea. It explains how, through the development of bespoke
software tools, a designer can iteratively work in real-time and have the ability to
pass model data through a comprehensive automated system, generating the output
of fabrication-ready data for a highly customized project.
Keywords Wood design · Parametric design · File-To-Factory

Complex timber structures · Wood joint systems
1 Introduction
The use of parametric and generative tools is no doubt increasingly common in

the architectural space, and the barriers to generating and controlling increasingly
complex forms is lowering with the availability of new digital tools and developing
skillsets of designers. Concurrently, access to powerful digital numeric control (NC)
machinery able to mass customize large numbers of self-similar but unique parts
is also increasing. However, the workflows necessary to seamlessly bridge between
design and fabrication are not yet clearly developed for a variety of reasons, including
J. Nelson (B)
The University of Sydney, Sydney, Australia
e-mail: jonathan.nelson@sydney.edu.au
C. Knapp
Western Sydney University, Sydney, Australia
e-mail: c.knapp@westernsydney.edu.au
R. Spence · J. Hutchines
Studio Workshop, Mermaid Beach, Australia
e-mail: rory@studioworkshop.com.au
J. Hutchines
e-mail: joel@studioworkshop.com.au

1184 J. Nelson et al.
the need to develop a solution for the uniqueness of required method and execution
of any given project, the perceived lack of cost and quality control on geometrically
complex projects, and difficulty of negotiating tolerance between human and machine
work in respect to the former.
In an effort to address these issues in a real-world application scenario, the case-
study project below demonstrates the testing of the hypothesis that at the current
level of technology it is practical to develop custom computational tools that bridge
between design and fabrication on complex projects and that the tools can also extend
into the assembly and installation phases while controlling costs and quality. It also
demonstrates that by incorporating elements of design, new issues are encountered,
such as accommodating change and flexibility (Woodbury 2010; Davis 2013) which
lead to innovative adaptations within the system.
Of particular note, this project addresses the need for mediation between users
and developers of a bespoke tool and their potentially dissimilar skill sets and needs
without compromising the tool’s efficacy or functionality. Through the implementa-
tion of a simple, familiar interface for the designer to engage with custom tools in an
intuitive and flexible manner—in this case a ‘sculpting’ solution is employed—mak-
ing it possible to control the solution space of a generative system without disrupt-
ing the automated workflow [which is a constant issue in the industry as noted by
Sass (2012)]. This contrasts with many previously developed file-to-factory systems,
effectively demonstrating that a designer can iteratively work in real-time and have
the ability to pass model data through a comprehensive automated system, generating
the output of fabrication-ready data for a highly customized project.
The project described herein is an iteration by the design and fabrication office,
Studio Workshop, which is a practice based in Australia. The office is comprised
of architects who all approach the discipline from a background in construction,
technology, and design. The process described in this paper of developing a custom
digital design interface and CNC workflow to construction is entirely an in-house
process that has been designed and developed by the authors. This approach to
making architecture has been developed through a series of commissioned projects
and experimental research prototypes which have grown in scale and complexity
over the past decade. Important to the outcome of each project, including the Brise-
Soleil House, is the seamless integration between conception and fabrication that is
shared and understood by the entire practice which allows ‘design’ to be informed
by ‘fabrication’ and vice versa. Many of the processes described herein are not ‘new’
except that in this instance they are housed within a singular practice entity and are
tested through ‘real world’ projects. Studio Workshop is not a ‘design/build’ practice
nor a ‘prefabrication’ builder; rather, it is an architecture practice that conceives of
design through the constraints of materials and processes, much like in the dictum
“Tools shape materials which shape forms, and not the other way around” (Barkow
and Liebinger 2009).
Brise-Soleil House: Developing Software Tools to Enable … 1185
2 Context
Papua New Guinea has a rich tradition of building with timber in surprisingly sophis-
ticated ways. Vernacular dwellings such as the stilt houses of the coastal lowlands
feature thatched rooves and woven wall panels amid a tracery of roughly shaped
timber. Their vernacular sea-faring canoes (lakatoi) are also quite robust, featuring
intricately carved prows and wave splitters, and have been refined by many genera-
tions of development.
However, the effects of rapid modernization and industrialization in the country
have worked to favor extreme speed and economy and as a result most traces of these
traditions and skills, which are ostensibly labor and time intensive, are conspicuously
absent. Rough-hewn stilt houses with woven skins are being displaced by steel,
blockwork, and fiber cement dwellings and traditional canoes have been largely
supplanted by modern fiberglass watercraft. Along with modern materials, or course,
comes other modern dependencies; the houses necessarily require air conditioning
equipment due to a lack of adherence to environmental principles and the boats
require motors due to lack of keel or sail.
The Brise-Soleil House seeks to mediate between these contexts. The site-cast
concrete shell of the building is simple and prismatic, left raw and unfinished inside
and out, echoing the more modern structures on the island. A timber wrapper pays
homage to local traditions of manual craftsmanship, utilizing a system of complex
joints and a high degree of surface articulation which formally resembles the tidal
motion of the nearby sea. But it does so though a digital lens; updated to address
the capabilities of contemporary software and hardware tools and their ability to
manage complex systems and to mass-customize intricate geometries with embedded
assembly logics. Furthermore, the design of the wrapper engages a broader discussion
of screens and veils in equatorial architecture found in the canon of Modern and 20th
century works by Ossipoff, Ferrie (Fig. 2), Rudolf, and others (Fig. 1).
Fig. 1 Papuan Lakatoi

Boats
Fig. 2 James Ferrie
3 Project Details
The Brise-Soleil House is a compact 2-bedroom, 160 m2 dwelling situated at the top
of a steep, west-facing block near Port Moresby Harbour overlooking a former cargo
terminal. The house is a cast-in situ concrete building partially clad in an undulating
wave-like timber wrapper, or brise-soleil as referred to in the Corbusian tradition,
which provides shading, privacy, and ventilation to the exposed volume of the house.
The wrapper flattens as it continues around the building to become a full-height
operable screen for the upper level gallery, controlling the western sun and capturing
views to the Coral Sea beyond.
The timber wrapper and doors are constructed from 50 × 50 mm rough-sawn
acetylated pine timber (under the trade name “Accoya”) which is glued and doweled
at digitally-executed complex half-lap joints. The wrapper and doors, with a total area
of over 80 m2 , contain over 2200 pieces in total and over 700 unique pieces with CNC
milled half-lap joints, often on both ends. All milling, joining, finishing, and assembly
into panels took place at the designers’ workshop on the Gold Coast of Australia, and
was shipped via container to Papua New Guinea for installation. The entire assembly
has been organized as a system of seamless 1200 mm panels and doors which serves
not only to aid in transportation and erection, but also to minimize the need for skilled
labor on-site. Though each panel was unique the installation method was identical,
which took approximately 2 days for a crew of 3 laborers working exclusively by
hand on scaffolding. The 8 door panels were similarly installed in one day.
4 Bespoke Tool Development
As mentioned previously, this project is a continuation of the authors’ work which

focuses on the integration of a design, fabrication, and assembly feedback loop
within a given project. This is accomplished through incorporation of a strategy to
develop comprehensive bespoke tools which facilitate much greater control over
all parts of a project and to greatly increase flexibility in addressing the particular
parameters of a project, such as functional, environmental, or cost-related goals.
While these methods are much less common in the AEC industry than in other
industries such as automotive, maritime, aerospace, and industrial design, they are
still employed in various scales by forward-thinking designers. However, the cost
and time requirements of developing these solutions has usually relegated them to
larger projects or the academic arena where they can be effectively subsidized, in the
case of the former, or not have to conform to applicable building standards, in the
case of the latter. That said, recent advances in software technology and equipment
accessibility are helping to lower the barriers to developing bespoke and project
specific tools. Studio Workshop seeks to further bridge the gap between design and
production by both developing these software tools in-house and fabricating them as
well in our own on-premises workshop.
A major challenge to the development of these tools is that since it accommodates
issues of design, it must be inherently flexible, as design by definition implies change
(Woodbury 2010) and that changes can have great effect on downstream processes. To
address these issues, the tool is developed through a process more closely following
software engineering practices than architectural logic (Davis 2013) and features
real-time data validation, interference checks, and reporting measures to advise on
the status of the tool in process as discussed previously by Scheurer et al. (2005).
Specifically, the Brise-Soleil House tool is intended to be a comprehensive system
that can directly address project phases from design development through assembly,
and indirectly address other phases such as transport and installation in situ. The tool
operates within 4 distinct stages which are seamlessly linked through one compre-
hensive digital environment that incorporates the following sections:
• Form Development: The designer is able to directly and intuitively tweak and
sculpt the nuances of the form in real-time via a representation of the underlying
base surfaces. This includes automatic correction of sculpting moves which result
in impossible formal arrangements.
• Design Development: The designer is able to verify the fidelity of their sculpting
and manipulate more fine-grained aspects of the assembly through a 3d line-based
representation (skeleton) showing generally the method of assembly (there are
two distinct assembly type choices), member spacing and overlap, and get rough
estimates of quantities of timber required and costs. This stage takes longer to
update than the previous stage but is still relatively fast.
• Finalization: The designer is able to see a full 3d model of the final design to assess
the fabricated outcome and judge the number of pieces, thickness of members, and
light transmittance. Voids for windows are cut, substructure is added, and members
are mitered at the correct angles for assembly. This process is too slow for real-time
designing, but still only takes a few seconds to update.
• Fabrication: At this point the details are processed; features such as half-lap joints
are calculated, dowel holes are drawn, tolerance and milling bit diameters are taken
into account, and the members are numbered and nested into raw material sizes
and output as fabrication-ready files. This process is discussed in more detail in
Sect. 5 below.
The following sections further illustrate the details and functionality of each of
these constituent parts of the tool. In order to allow real-time updates during the
sculpting and development phases there is an internal switch within the tool that
allows phase selection, effectively blocking later parts of the tool from processing
until the designer is ready, though the designer is free to move both forwards and
backwards within the tool as needed. All checks and error-prevention devices are
included in the current phase so that a designer cannot easily bypass phases with an
impossible structure.
4.1 Form Development
One of the major difficulties of algorithmic development is its interface with design
methodologies, specifically in relation to traditional creative methods and “design-
erly” ways of working (Aish 2005). It is true that working purely algorithmically
necessitates a change in design paradigm and as a result may arguably be foreign
in method and result than what many are used to, but that can be seen as an oppor-
tunity rather than a hinderance. As a prominent example, Frank Gehry, who’s firm
is renowned for incorporating advanced digital methods and working across design
and fabrication phases famously still sculpts his works by hand with paper and is
admittedly a luddite (Lindsey and Gehry 2001). His team has evolved hybrid ways
of working across physical and digital processes during the design phases which
serve to incorporate both Gehry’s methods of designing and ensuring constructabil-
ity in subsequent phases, notably by constricting Gehry to working with paper strips
(Shelden 2002).
With the Brise-Soleil House tool, the authors are experimenting with translating
sketches and loose ideas to form a way of working which is then built into an intuitive
software tool that can support a designer and the way they are accustomed to working-
yet, this tool can be easily changed and reversed if desired. In this case, the tool
took on a sculpting-like approach, much like a digital form of clay with a push-pull
interface. While this may at first seem excessive, it is precisely because of the relative
simplicity and speed of the toolmaking process that these don’t have to be reused
from project to project to amortize their development cost.
In the initial stage, the tool uses software drawn curves that are directly referenced
into the tool environment, in this case McNeel’s Rhinoceros and the visual scripting
environment Grasshopper. The designer can easily manipulate the curves via their
control points and the tool will update a simple surface representation of the brise-
soleil (Fig. 3) which is generated in real-time and allows the designer to intuitively
edit the overall formal characteristics of the geometry.
At this stage the tool also attempts to correct the designer’s work which may
exceed certain physical criteria. For instance, the designer can sculpt and pull the
curves out of bounds with regards to the maximum projection from the base surface
or into a configuration that has too little continuity to enable the blocks to span two
adjacent sticks. To avoid this condition, the tool will automatically remap the curves
Fig. 3 Sculpting
environment of the tool. The
curves influence the
articulation of the
represented form
into the closest acceptable (buildable) configuration for the purposes of continuing
the algorithm, the parameters of which are set by the designer within the tool. If the
designer deletes or irreversibly contorts the curves the closest appropriate curves can
simply be reintegrated into the initial inputs and the designer can start over.
4.2 Design Development
Once the general form is acceptable, the designer can look at the particulars of
construction and some of the general details that influence the design development
side of the work such as the sizing of members, their functional characteristics and the
relative impact on things like quantities and budget. The designer can choose between
two different construction types, a staggered and woven, comb-like arrangement of
members or an inline series of jointed members with spacing blocks, the relative
merits of which are all recorded equally into the tool’s output.
This section relies heavily on mathematical rather than geometrical processing
methods, which are generally the norm in the Grasshopper environment. However, in
order to efficiently process quantitative representations without affecting the usability
of the sculpting features, it was necessary to develop simple functions that estimate
member quantities, lengths (smallest, longest, average, mean lengths), total estimated
lengths of timber required (including predefined overage percentage) for processing,
estimated costs, and estimated light transmission percentages. These provide valuable
cost and functionality planning resources, whereby a designer can easily edit the
opacity/transparency of the screen or adjust the sizes and level of curvature to fine
tune material costs and process complexity via the number of unique pieces in the
system. Figure 4 shows the environment during this phase.
Fig. 4 The design

development interface shows
approximate layout of the
sticks and their resolution in
relation to the overall form
4.3 Finalization
The next step is to develop detailed geometric and construction-related details for
fabrication. This process is relatively resource intensive and not useful for real-time
sculpting, but for verification it is further split into two sub-phases. The first shows
a full 3d representational model of the brise-soleil for final verification as shown in
Fig. 5. This gives accurate sizing and trimming of the members to give a very close
approximation of the final result.
The second sub-phase is again more time intensive and not particularly useful on-
screen to the designer. It develops the fabrication and assembly details that are suited
to being done in a 3d environment such as the half-lap joints for connecting pieces
in line. These details are calculated in situ geometrically in 3d and 2d representation
are extracted from them, adjusted for manufacturing parameters such as tolerance
and router bit size. While this does not strictly need to be calculated geometrically
it was efficient during development to visually verify that joining pieces are equally
Fig. 5 Full geometry model

for verification in the
finalization stage
extracted and adjusted so that they fit together. Simple additional checks such as
matching areas of cutouts to adjacent pieces are done and errors are thrown to the
user if there are non-matching adjacent pieces or if the proper counts of features are
not found within the algorithm.
Finally, the tool gathers groupings of 2d curves representing member perimeters,
the lap joints, dowel placements, screw divots, and numbering text and lays them
out on a 2d plane for nesting, which occurs in the next phase. At this point one final
check for matching quantities is performed.
Another significant benefit to this system was that the entire process was nearly
paperless; the only documentation required were verification sheets for milling oper-
ations in the workshop and two pages of details and drawings for onsite installation.
The tool and 3d model produced all other data for the project which was managed
digitally. This phase is also useful as an assembly guide as each numbered piece and
each larger assembly panel is numbered in the model which allows the assembly to
be verified digitally instead of relying on a set of assembly drawings.
5 Fabrication and Assembly
The integrated nature of the developed tool addresses a number of common pain
points within the industry which occur during the transitions between the design
phases and the fabrication and assembly phases. The loss of design intent through
poor transmission sources (paper, simple 3d models) can hinder the seamless trans-
mission of information and lead to unintended or unworkable results. Likewise, the
need to remodel or redraw portions of a design to suit fabrication can lead to expen-
sive delays through unnecessary rework. This is further exacerbated if any design
changes are made which necessitate editing of the reworked drawings and models.
The management of complexity and installation methods is also integrated into
the script. This provides a level of automated control of the number of unique parts,
assembly methods, tolerances, and sequencing that allows for the use of unskilled
labor to quickly and reliably assemble geometrically complex pieces. The following
sections detail some of the approaches and process employed in this project.
5.1 Documentation
One of the key potential benefits of an integrated process is the ability to avoid
the extensive quantity of paper documentation traditionally required in architectural
projects. As a means for transferring information, paper is a poor choice since it
only conveys that which is explicitly highlighted as important. A digital model is
better in that it inherently captures all the geometrical information used to create it,
a high level of accuracy, and, in the case of rich formats such as IFC, meta data.
As previously described, the script used for the Brise-Soleil House is completely
Fig. 6 Brise-soleil assembly diagram with rules
open and continuous, allowing access to any data at any point in the script and the
design intent is fully resolved and seamlessly translated into fabrication data. This
eliminates the need for much of the documentation as traditionally required such as
shop drawings and assembly diagrams. Documentation overall was very minimal;
in total there was only one assembly diagram outlining the simple rules for the
numbering and assembly system which was a 3d drawing of the organization of the
entire wrapper divided into section and including a few instructions (Fig. 6).
The only other paper involved in the process were check sheets for the CNC
process to validate completed batches and identify pieces to be recut, and installation
details for site-based work.
5.2 Fabrication
In order to quickly process dimensioned timber for this project, a custom jig of
aluminum angles, constrained by the dimensions of the CNC router, and screwed to
an MDF spoil-deck was built which could accommodate the raw materials with little
extra fastening and positioning, and would allow for rapid batching of processed
parts. Aluminum was chosen because the router bit could cut through it with little
or no damage to the bit and allowed for extra margin of error. This jig was balanced
with the bespoke tool to ensure accuracy by outputting a scribing routine based on
its internal nesting configurations to locate the positioning of the aluminum angles
on the jig.
The actual conversion of curve geometry into G-code routines is not totally auto-
mated for a few reasons. Firstly, the constraints of developing a postprocessor for
a single machine were beyond the scope and budget of the project and would have
Fig. 7 CAD output showing one batch for CNC processing. Each process is shown in a different
color
limited the use of the tool to a single machine. Secondly, the CAM software used was
relatively robust in itself and could prepare templates allowing for quick processing
of similar elements. And finally, the need to manually input the curves into the CAM
software provided an additional layer of human supervision to ensure that the script
was functioning properly.
However, all of the parts are pre-positioned for the jig and CAM work in a related
software package is simple, taking only a few minutes per batch. As shown in Fig. 7
pieces output for CAM work are pre-colored by router operation—profiling, pock-
eting, drilling, and numbering—to make fast selection as the software allows quick
selection of elements by similar color and to provide differentiation for human eyes
and yet another layer of error checking and redundancy in the process. Once con-
verted into G-Code for the CNC router the file runs as a continuous process including
all operations which took about 40 min per batch at a relatively conservative feed
rate.
Once the timber sticks are routed they are removed in one continuous piece and
a new batch of sticks is inserted into the jig for processing. The sticks are then cut
by power saw into the final piece shapes using the pocket ends as a guide.
5.3 The Jigsaw Effect
When dealing with highly complex and bespoke assemblies and, in particular, those
that would be considered generative designs, there is a marked propensity for the
exacerbation of the abilities of equipment to produce mass-customized designs.
While it is true that most CNC equipment can make an unlimited number of unique
Fig. 8 Sticks from the CNC router (left), and trimmed for assembly (right)
pieces with little or no extra effort on behalf of the machine, the management of a
large number of pieces with little geometrical distinction quickly becomes an overly
complex task for logistical reasons, and, objectively may add little to the overall
project relative to the required effort. Based on previous experience, the authors
have learned that it is far better to use as many identical parts as possible where
practical to avoid searching for a particular unique set of geometrical attributes on a
part or complex numbering systems which the authors liken to assembling a jigsaw
puzzle, hence the term “The Jigsaw Effect”.
The overall complexity of the brise-soleil screen makes this a delicate proposition
since each twig in a vertical stick is unique based on two conditions: the general results
of the curve-based sculpting tool and the angle of intersection with its adjacent twigs.
However, the joining blocks between twigs could easily be made identical, resulting
in the conversion of hundreds of potentially unique parts into just one simple block
of consistent dimensions, which could in turn be applied to the entire screen and the
balcony doors. This has the flow-on effect of removing these parts from the CNC
processing chain which, in addition to reducing milling time and effort, enabled a
less precise but faster fabrication method (manual cutting of blocks on a drop saw)
which could be completed concurrently (Fig. 8).
To further lessen the jigsaw effect, a batching strategy was employed for nest-
ing and processing pieces within their jigs. Each panel was nested and processed
separately which allowed the easy collection of sticks and twigs within a panel to
avoid misplacing pieces and reduce sorting to a minimum. This leads to a slightly
greater amount of physical waste generated but when compared with the potential
effort waste in organizing pieces the amount was within acceptable limits, especially
when compounded with the other waste minimization strategies employed here.
5.4 Assembly
Another significant advantage of the bespoke digital tool workflow is that assembly
methods and logics can be integrated into the earlier design and fabrication stages
to simplify the process of positioning fixing complex parts allowing for lower skill
level requirements.
As previously seen in Fig. 6, the part numbering system provides all necessary
information as to the location of a part within the larger assembly. The number
goes in descending order and forms a hierarchy: Side, Panel, Stick, and Twig—with
there always being 5 twigs. Twigs are fastened together at the routed half-lap joint
using timber screws into the Stick level of hierarchy. The timber is soft enough to
avoid having to pre-drill the hole completely and just a small divot for location was
required. While the tolerance was small enough (0.5 mm) in the lap joints to limit
the range of motion in the joints, a secondary system employing an elongated dowel
or biscuit was used which would assist in orientation and was snug while gluing.
This helped lock in the angles with less than 1 degree of error. As a final failsafe, the
script output a series of scribed guides that were CNC engraved onto MDF sheets
that would serve as a guide to stick fixing angles.
The process in total was very rapid; assembling a stick from 5 twigs took about
10 min for one person to complete. Assembling Sticks into Panels was similarly
efficient; the flat ends of the sticks allowed for propping up on assembly tables and
blocks and sticks could be assembled in order to complete a panel. This process took
two people about an hour to complete per panel (Fig. 9).
6 Transport and Installation
While the bespoke tool for the Brise-Soleil House does not have a specific phase or
section that addresses issues of transport and in situ installation, there are methods
and parameters included that account for potential issues and in effect exert an impact
beyond the digitally controlled. As stated, the lack of skilled onsite labour and the
harsh environmental conditions combined for a good deal of potential issues which,
because of their distance and lack of skill can represent considerable risk for a
complex project.
As the screen and door panels were manufactured in Australia and shipped to
Papua New Guinea, and installed at height, they necessitated considerable handling.
To this end the tool allows for the panelization of the brise-soleil and adjustment
of door sizes to allow for handling by two people comfortably, even at awkward
Fig. 9 Panels being assembled. Prototype panel can be seen in background
positions and on scaffolding; panels weighed an average of less than 35 kilos each.
Through a combination of full-scale prototyping with acetylated pine, having the tool
calculate the weight of the panels, and inputting as-built site conditions for sizing it
was determined that panels of approximately 1200 mm would divide evenly through
the screen (Fig. 10). With only minor need for packing out with slightly larger or
smaller infill blocks that would be cut on site any tolerance issues could be made
up cleanly and easily. The 2800 height allowed for easy placement into a shipping
container which also carried interior finishes and whitegoods.
In situ installation was very simple and took less than two days for a crew of two
people. The weight was such that panels could be lowered from the roof by hand
into place and secured from the top into a steel angle fixed to steel support columns
within the building wall (Fig. 11). Due to the potential of different coefficients of
thermals expansion in the materials used—steel concrete, and timber—the bottom
of the panel rests on another steel angle cushioned by a rubber gasket and a bead
of silicone. Finally, a powder-coated metal fin surround hides the fixings from view
and prevents water from seeping behind the screen.
Fig. 10 Panels arranged at workshop prior to shipping (left) and onsite awaiting installation
Fig. 11 Installation in situ on scaffolding

7 Discussion and Conclusions
It is argued that we are now at a stage in the industry where the development of the
software tools and design protocols necessary to achieve works such as the Brise-
Soleil House can today be rapidly executed and economically beneficial to small
scale professionally-delivered projects, whereas previously the majority of similar
examples were either developed in academic scenarios with few constraints or on
very large projects with grandiose teams and budgets (canonically, the Guggenheim
Bilbao and its antecedents). These small-scale tools for small scale projects offer
many benefits when trying to manage complex projects: they provide constantly
updated quantity and cost estimates, environmental feedback, waste analysis, and
integrated error checking and embedded intelligence to become a comprehensive and
data-rich design environment that is also capable of capturing the implicit quality of
a designer’s work (Figs. 12 and 13).
Further, and in a more general sense, the ability for designers to integrate and
control downstream operations intuitively and interactively by not only designing
what is to be built, but how it is built, through a meta-designing of the processes that
encode the information of design, fabrication, and assembly, represents a paradigm
shift in the practice of design itself in the AEC industry through a new model of
practice. As noted by Sharples (2011), the process itself is the model.
Fig. 12 Completed Brise soleil and doors

Fig. 13 Completed house
Acknowledgment All works represented herein including designs and software tools, and their rep-
resentations are protected intellectual and commercial property of StudioWorkshop Design Group
Pty. Ltd., Queensland, Australia under international and local copyright law.
References
Aish R (2005) From intuition to precision in digital design: the quest for new paradigms, 23nd
eCAADe Conference Proceedings, Lisbon, Portugal September 2005, pp 10–14
Barkow F, Liebinger R (2009) An atlas of fabrication. Architectural Association, London
Davis D (2013) Modelled on software engineering: flexible parametric models in the practice of
architecture. Ph.D. dissertation, RMIT University
Lindsey B, Gehry F (2001) Digital Gehry: material resistance digital construction. Springer, New
York
Sass L (2012) Direct building manufacturing of homes with digital fabrication. In: Gu N, Wang
X (eds) Computational design methods and technologies: applications in CAD, CAM and CAE
Education, Hershey P, IGI Global, pp 71–82
Scheurer F, Schindler C, Braach M (2005) From design to production: three complex structures
materialised in wood. In Soddu C (ed) 6th International conference generative art, Milan, 2005
Shelden D (2002) Digital surface representation and the constructability of Gehry’s architecture.
Ph.D. Thesis, Massachusetts Institute of Technology, Boston Massachusetts, USA
Sharples C (2011) Barclays center in Tarron J (ed) ACADIA 2011: integration through computa-
tion, proceedings of the 31st annual conference of the association for computer aided design in
architecture (ACADIA), vol 2, pp 172–173, Banff, Canada, 2011
Woodbury R (2010) Elements of parametric design, Routledge, New York, NY
Towards a 4.0 Mass Customized Wooden
Housing in the Mediterranean Area: The
Ecodomus Project
Micaela Colella and Giuseppe Fallacara
Abstract The paper describes the necessary transition from the consolidated prac-
tices of contemporary residential architecture to innovative sustainable design and
construction practices, which can be pursued by applying the principles of digital
fabrication to the construction industry. The contribution is divided into two parts:
the first investigates the theoretical-critical assumptions of research and the second
part describes a design experience that led to the creation of a housing prototype
with an irregular geometry made up of CLT panels. The paper will explain the need
to overcome the current phase of experimentation using digital fabrication technolo-
gies applied to small pavilions with complex shapes or high budget, to move to their
application on residential building. The application of these technologies to residen-
tial architecture is desirable both for the diffusion and for the relevance of the change
it would produce.
Keywords Wood design · Wood joint systems

Large-scale cross-laminated timber (CLT) projects · Industry 4.0
Mass customization
1 Introduction: The Paradigms of Industry 4.0
Today we are on the threshold of the fourth epochal turning point in the incessant
and inescapable race towards technological progress. The new nodal point in techno-
logical evolution in Europe has taken the name of “Industry 4.0” since 2011, when
Kagermann et al. (2013) presented their strategic proposal at the Hanover Fair to
strengthen the competitiveness of the German manufacturing industry. The name
“Industrie 4.0” was subsequently adopted by the German Federal Government.
M. Colella (B) · G. Fallacara

Department of Sciences of Engineering and Architecture,
Polytechnic University of Bari, Bari, Italy
e-mail: mc@newfundamentals.it; micaela.colella@hotmail.it
G. Fallacara
e-mail: giuseppe.fallacara@poliba.it

1202 M. Colella and G. Fallacara
The new scenarios outlined by the advent of the fourth industrial revolution are
changing the way of thinking about society and the world of work in the near future,
with the definition of new professionals. However, industry is the field most affected
by the changes, with a rethinking of its management, production methods and prod-
ucts themselves. There are several principles that raise the factory to 4.0, opening
up new and innovative scenarios. The pivotal point of the ongoing revolution is rep-
resented by the fusion between the virtual and the real world in the new concept
of Cyber-Physical System (CPS). In short, CFS are defined by objects that incor-
porate technological devices capable of creating an interface between the physical
and digital world. The network in which CFS interact and cooperate with each other
according to predefined patterns, is called the Internet of Things (IoT). In addition to
the IoT, the Internet of Services (IoS) determines a leap in scale, in which the CPS
are constituted no longer by individual objects, but by individual activities in the
company’s value chain. Therefore, the development of this technology allows a new
way of business management, characterized by a dynamic distribution of activities.
The aforementioned principles describe a highly flexible production model, capable
of modifying and reorganizing production in a short time, thanks to the high degree
of modularity that characterizes its production components. This aspect, together
with the ability to collect and process in real time an enormous amount of data, will
make it possible to implement a fully customized industrial production, thus able to
satisfy the needs and desires of the consumers. Therefore, the specific desires and
needs of unique consumer will return to occupy the central role they played in craft
production. This scenario is defined mass customization. The emergence of this new
conception of industrial production destined to constitute the post-Fordist paradigm
in the 21st century economy, dates back to the Nineties of the last century. The term
mass customization coined by economist Stanley Davis in (1987) can be defined
as «producing goods and services to meet individual customer’s needs with near
mass production efficiency» (Tseng and Jiao 2001). The products are then tailored
and no longer designed as standard production based on market forecasts. They are
characterized by the controlled quality of an industrial product (Paoletti 2010), but
ensuring at the same time lower production costs and selling prices.
The mass production, characterized by the storage of large quantities of products to
be sold in the future, may lose sense if we are able to know in “real time” the demands
and necessities of consumers and, at the same time, the factory is able to vary the
good produced for each production cycle with great flexibility. This new paradigm
paving the way for on demand and customized industrial production, returning to the
production volumes and the customization of goods closer to the craft world than to
industrial mass production. Consequently, even considering the only changes related
to the use of these technologies in industrial production, the benefits deriving from
it are numerous and of considerable importance: the possibility of reducing waste
related to production excess, the optimization of energy and materials used and the
reduction of space for storage of products. Unlike previous Industrial Revolutions,
which had as their main result the improvement of living conditions through increased
productivity and profits, Industry 4.0 has the potential to significantly improve the
Towards a 4.0 Mass Customized Wooden Housing … 1203
living conditions of humanity (and probably of planet Earth) not through an ever
increased production, but by optimizing the resources available in a more conscious,
targeted and personalized production.
2 Industry 4.0 Principles for Architectural Design

and Construction
Despite the fact that in recent decades there has been an area of phlegmatic and
opposed technological progress, with respect to, for example, the rapid evolution of
the automotive industry, we believe that AEC industry is destined to become one
of the main fields of application of Industry 4.0 principles. According to Achim
Menges,1 «Making in architecture has of course always differed from manufactur-
ing in other fields, and some building processes, especially in on-site construction,
have managed to escape the economy of automation and industrialisation altogether.
But there is no question that, in the same way that previous technological advance-
ments in other industries have had a major influence on the development of those
disciplines, the ideas, logics and technologies of Industry 4.0 will have a signifi-
cant impact on both the conception and production of architecture. It is important to
note, though, that the Fourth Industrial Revolution differs from the previous three,
because its primary objective is not increased productivity, but instead higher levels of
flexibility, adaptability and integration. This industrial transformation may therefore
have an even greater effect than its precursors in terms of its relation to architectural
design» (Menges 2015). There is no doubt that the most important innovations in
the design field in the last thirty years were the diffusion of CAD (Computer Aided
Design) and CAM (Computer Aided Manufacturing) systems. A revolution whose
first fundamental experiments can actually be traced back to the Sixties and Sev-
enties, at the Massachusetts Institute of Technology, thanks to the work of leading
researchers such as Sutherland (1964), who developed SketchPad, the progenitor
software of modern CAD systems, and Nicholas Negroponte, founder of the Archi-
tectural Machine Group, a research group dedicated to the study and development
of the first human-machine interfaces (Negroponte 1970). Despite this long history,
the still predominant use of CAD tools is simply represented by a “computerized”
version of traditional drawing techniques (Kostas 2006). However, since the early
2000s, a new digital revolution (perhaps fully digital) spreads: from computer aided
design towards computational design. As stated by T. Kostas, while computeriza-
tion in recent decades can be traced back to the digitization of consolidated, defined
and predetermined processes to improve efficiency, accuracy and workflow, on the
contrary, the computational design concerns the algorithmic exploration of indeter-
minate processes. Since the 2000s, the whole AEC industry has been involved in
this innovation process through the introduction of parametric-computational design
softwares. These software allow a rapid variation of the design outcome depending
1 Director of the Institute for Computational Design and Construction (ICD), University of Stuttgart.
on the variation of the design parameters, while requesting a greater commitment

during the design phase. The use of these tools take full advantage in digital fabrica-
tion,2 precisely because of their ability to overcome the static nature of computerized
design processes and to actively contribute to the optimization of construction. In
this regard, Menges and Ahlquist (2011) noted: «where computer-aided processes
begin with the specific and end with the object, computational processes start with
the elemental properties and generative rules to end with information which derives
form as a dynamic system».
Why is it important to emphasize this passage from our point of view? This pas-
sage is necessary because computational design is the only design process able to
communicate profitably with the virtualization processes typical of Industry 4.0. In
fact, «truly computational making entails the unfolding of an explorative process
of materialisation driven by cyber-physical feedback, which extends design rather
than merely realising it» (Menges 2015). These new “smart” machines are able, not
only to perform the tasks according to predefined schemes, but to develop alterna-
tive answers in order to improve the efficiency of the processes and consequently
reorganize the production. In this way they become part, with their contribution, of
the computational design process. The interconnected systems are able to recalibrate
the production plan, re-launch the production on other machines and optimize the
processes. In this way, humans are supported by the machines themselves in their
complex management. While in the past for each change in the production chain it
took weeks of tests managed by highly qualified personnel, thanks to virtualization
(i.e. the use of models of simulation), the preliminary tests of production take place in
the virtual environment and the time in which the machines are stationary is reduced
notably, with an enormous saving in economic terms (Blanchet et al. 2014).
Referring to the technologies characterizing the Industry 4.0 and imagining an
application of their principles in the field of architectural design, it is possible to
outline the feasible scenarios that could be configured in the architecture of the 4.0
era. Starting from the principle of virtualization, it is therefore possible to identify
at least three levels of virtualization:
– project virtualization, i.e. predictive processes able to cover several factors simulta-
neously, thanks to the increasingly sophisticated development of three-dimensional
and computational modelling programs. To cite just a few examples, it is possible:
to virtualize the climatic behaviour of a building, simulating the position of the
sun in a given time, of a given day, in a precise position on the terrestrial globe in
order to calculate the consequent irradiation (clear sky); to virtualize the phases of
the construction process and the entire life cycle of a project, thanks to the Build-
ing Information Modelling software (BIM); or analyze the structural behaviour of
buildings, and so on;
– virtualization of the CAD/CAM manufacturing process, i.e. prediction, control and
preventive verification of the Computer Aided Manufacturing phases, referring
both to the single processing and to several consequential processes;
2 For
digital fabrication, we mean a process in which the product is obtained starting from a digital
model and is manufactured through a machine controlled by a computer.
– virtualization of the production process, in which the entire production process of

the factory is to be foreseen and verified in the virtual environment, thanks to the
advent of Industry 4.0 technologies.
The affirmation of an “integrated virtualization” can therefore represent a possible
way, since the designers are already familiar to a working approach with an important
component of virtualization, an aspect that would be a key element for the interac-
tion between the design process and the industrial realisation. In other words, the
designer would be allowed to contribute to the management of the industrial process,
through a type of integrated design that is able to contemplate the manufacturing and
industrial production methods right from the early design stages. Furthermore, the
computational project of the designer is subjected to a “critical” analysis by the start
machines involved in the various phases of the design and manufacturing process,
which can elaborate and report possible modifications aimed at optimizing a process
(for example, the optimal use of a material according to its performance, or to reduce
waste, etc.).3 This is an aspect that brings with it a series of critical issues, mainly
related to the interchange of data between one process and another, hence the need
to be able to work on common software platforms or that are able to interface with
each other. The whole process would become “informed”—and not simply comput-
erized—enriched and guided by the “cloud of data” elaborated in real time, in a state
of interconnection between the design tools and the manufacturing and construction
tools, in a continuous exchange of data and control of physical processes.
In this context, the role of the designer should not be considered as a marginal
role. The “cloud of data” can be a tool of extraordinary potential for the architect
who can boast good skills as a director of the design and construction process.
By imagining the application of the technologies characterizing the 4.0 industry to
the field of architectural design, the architect’s work would naturally be supported
by the potential of the new tools, but would become more difficult and greedy of
responsibility, as the project would become increasingly integrated, fruit of a holistic
concept that could no longer be ignored. It is sufficient to think about the possibility
of designing with the support of a software tool always connected, able to inform
and consequently condition the project with data related to the project site, to the
processes that see it involved, for example to environmental data (climatic data,
seismic risk, presence of electromagnetic fields, etc.), and having to make sure that
every condition falls within the project with an adequate architectural response. At
the same time, the work of the designer is not limited to the production of drawings
useful to the construction site, but could (and should) allow the management of
manufacturing and construction processes through simulation tools of the processes
themselves, which, in the last instance, would also allow the optimization of available
resources, be they material, energy, economic, etc.
The characteristics outlined so far, although probably not fully described given
the topical nature of the theme, lead us to realize how essential the exclusive use of
prefabricated dry-mounted components is. This condition is absolutely essential for
the following reasons:
3 In reality, the possible cases are so many that it is difficult to identify them a priori.
– total adequacy in relation to digital and computational design processes;

– total adherence to dry digital or robotic manufacturing processes, especially com-
pared to wet or mixed systems, which are notoriously characterized by inaccurate
and not very controllable manufacturing and construction processes;
– greater precision in the manufacturing phase of building components, in a con-
trolled environment and under constantly optimal conditions;
– greater precision in the construction/assembly phase of building components;
– greater adherence between the structural and energy designed performance and
the actual ones, thanks to the precision of all the phases of execution and the use
of elements with certified performances;
– less production of waste material during manufacturing and construction phases;
– possibility, if opportunely foreseen in the design phase, to be able to replace parts
in the future that are no longer suitable for the function to which they must perform.
3 Digital Designed Housing: From the FabLabs to Industry

4.0
In the light of the considerations mentioned above, it is opportune to reflect on “mass

customization 4.0”, probably the most relevant aspects that Industry 4.0 could imple-
ment and which can lead to exceed the limits of the prefabrication processes as we
have known them until now. In the range of changes that the advent of new tech-
nologies will make possible in the architectural field, certainly the full realization of
mass customization will be one of the most important goals. Today, as mentioned
above, the enormous potential offered by the tools and principles of Industry 4.0, have
paved the way for a high flexibility of the factory that allows a customization at all
levels of industrial production. According to Ingrid Paoletti (2018): «Digitalization
of manufacturing in the era of Industry 4.0 is slowly but strongly challenging also
the construction sector. The possibility to introduce Mass Customization of products
is possible thanks to two very important factors: computational design process and
new advanced manufacturing machines. These two advancements can influence the
construction industry, towards 4.0, pushing to deliver not only customized products
for special projects but also to increase the quality of building construction in gen-
eral with techniques that can be designed ad hoc and personalized in material and
technology with a cost near to standard production». In the last years «researchers
and designers from all over the world are developing experimental architecture, in
prototypes up to the scale of small pavilions, in collaboration with industrial pro-
ducers, and with the use of advanced fabrication machines. This global research
has provided the opportunity to analyze, study and share unconventional solutions.
Nevertheless, the biggest challenge in the future will be the implementation of these
results at the building scale, addressing the complexity and high standards typical of
the built environment of our age» (Naboni and Paoletti 2015).
On this point, the pioneers of robotic manufacturing, Gramazio et al. (2014),

from ETH in Zurich, argue that the use of advanced design and construction tools
are able to stimulate the development of the concept of personalization and have
the potential to create a large amount of new building components that, having the
possibility of being produced only once in an exclusive way, can achieve a high
level of performance and customization, overcoming the stagnation resulting from a
repetitive production in buildings.
However, it would be important that digital fabrication overcome the experimen-
tal phase related to the construction of small pavilions, but that it would spread as
a method to be adopted for the construction of residential buildings and not only
of daring constructions related to high budget and complex shapes. In fact, the pro-
cesses of advanced customization of the “non-standard4 ” architecture refer mainly
to projects where the complexity is essentially geometric. But if the personalization
process itself is the important criterion, then it can also be applied to projects that are
simpler from a geometric point of view, as is commonly the case for homes. Looking,
for example, at the proliferation of digitally manufactured temporary pavilions, it is
possible to state that much research has been done in recent years on the subject of
advanced customization, while the theme of customization of building components
by simpler geometry definitely require more attention by researchers and designers.
Technology only changes the world when it becomes necessary and available to
everyone. There is no doubt that the next logical step in the advanced technology
4.0, is the dissemination and accessibility of this potential to the broader segments
of society. In order to achieve this diffusion one should reasonably go beyond the
application in the production of goods and services, which already would have an
extraordinary positive impact on the sustainability of the economic model, also in the
field of construction. And in the latter context, it is expected to exceed the explorative
phase applied to small pavilions with complex shapes or high-budget buildings, to
move towards the application on residential buildings. This would be more desirable
both for the spread and for impact of change that it would produce. Just think of the
energy savings due to the efficiency of the new houses themselves and their optimized
construction process, multiplied by the numbers of housing.
For the application of the principles of Industry 4.0 to residential architecture, an
unavoidable assumption is constituted by the overcoming of the current system of
the construction market, where the management of the construction companies plays
a decisive role, moving the reins of the whole process. In recent years, probably to
overcome this stagnant system, whose choices are dictated to a predominant degree
by economic interests, numerous prototypes of innovative homes have been created,
4 “Non-Standard Architecture” is the name of a very important exhibition, organized at the Center
Pompidou in Paris by Frédéric Migayroue and Zeynep Mennan between December 2003 and March
2004. In this exhibition a series of architects and designers of the first digital age, such as UNStudio,
Kas Oosterhuis (NOX), Greg Lynn and Objectile, exhibit a series of projects that empowered the new
digital architecture. In this exhibition the complex shapes possible thanks to the three-dimensional
modelling tools are enhanced and brought to the fore.
designed by research centres and produced in “self-build”, thanks to the use of the
typical Fab Lab tools.5
The Fab Lab House (2010) is the first significant example of an experimental
dwelling fabricated by Fab Lab tools, a prototype of a low environmental impact
house co-produced by the Instituto de Arquitectura Avanzada de Catalunya (IAAC)
and the Center for Bits and Atoms del MIT, winner of the Solar Decathlon Europe
2010. With this experimentation it was intended to demonstrate the possibility of
obtaining an energetically self-sufficient house, using the instruments easily avail-
able in any digital fabrication laboratory, thus reducing the costs of distribution and
transport (Gallanti 2011).
Another significant example is the WikiHouse (2017), an open-source project
born in 2011 thanks to Alastair Parvin and Nick Ierodiaconou, which aims to make
the construction of small sustainable wooden houses easier and more democratic.
The project is based on the idea of giving the possibility to anyone who wishes, to free
access a library of three-dimensional models of houses and modify them according
to their needs, obtaining automatically the files useful for production by CNC milling
of all the pieces necessary for the construction of the house.
Digital manufacturing tools are now at the core of the so-called Industry 4.0.
The application of the principles of Industry 4.0 to architecture will necessarily pass
through computational design and subsequent manufacturing and digital construc-
tion. In relation to the principles underlying these two frameworks, it is possible to
identify two constructive processes that, not surprisingly, are becoming the focus of
research and experimentation of research centres, and of public and private funding
and investments. We refer to additive manufacturing and a newly conceived prefab-
rication, which is completely customized. In our opinion, these two scenarios will
deeply influence the design of the homes of the near future. This article focuses on
the second axis of research, which sees wood as the main material for the design and
construction of mass customized prefabricated homes.
4 Towards a 4.0 Customized Prefabricated Housing
The first experiments of self-build homes made by using the Fab Labs’ tools have
taken advantage the potential of digital manufacturing tools for the construction
of prefabricated elements to build affordable customized solutions. However, for an
effective diffusion of housing solutions obtained by assembling custom prefabricated
constructive elements, it is necessary that the principles of mass customization (in
its new meaning deriving from the affirmation of Industry 4.0) become part of the
5 As is well known, FabLabs began to spread starting from the opening of the first digital manufac-
turing laboratory in 2001 at MIT, based on the intuition of Neil Gershenfeld about the importance
of direct experimentation by students through the prototyping of their projects with CNC machines.
Over the years, the instrumentation typical of the Fab Lab has extended to include numerous tools,
including: CNC milling cutters, laser cutter machines, 3D printers, welding stations and electronic
processing.
mainstream best practices of the companies that build prefabricated wooden homes.
The prefabricated house is a typology for which today a certain interest has returned
thanks to its ability to respect high quality standards (certifiable) with relatively low
building costs. The construction of these buildings is managed almost exclusively by
companies specializing in wood processing. In the field of residential prefabrication,
there is currently no technological culture shared between companies, because each
of them develop technical solutions and patents independently and in competition
with each other. This does not contribute to spread a solid theoretical and technical
culture on the subject and entails the need of clients and designers to turn to a company
to transform their design idea into a built architecture. This reduces the possibility of
control of the project by the designer who must limit himself to elaborating a general
architectural project that will then be engineered by the company itself.
Looking at the Italian market, according to a comparison carried out by the authors
taking into consideration 17 partner companies of the association Lignius,6 (belong-
ing to the Lignius Certified Member and Lignius Member categories), it has been
observed that companies in this field propose technical solutions which generally
comprise one or two structural systems—a frame system and a massive system with
CLT walls—and from two up to four wall systems. In this way, each company devel-
ops its own certified technological solutions and its patented building system, so
that the construction of homes becomes “serial”, optimizing construction timing and
costs. Furthermore, most companies offer catalogs of single-family or semi-detached
housing solutions, i.e. predetermined designs that can be chosen by the customers
or subjected to very limited modifications. Most of these companies add to the sale
of catalog houses with the design of customized houses using their construction
systems, or they realize projects developed by external designers, adapting later the
original project to the use of their technical solutions. The houses built according to
this last modality, even if designed by highly qualified professionals and perfectly
realized according to the highest technological standards, will always present a con-
ceptual weak point, because they were born from an adaptation process rather than
a totally customized design. In other words, they are not generated considering the
specific climatic and historical-cultural conditions of the place where the house will
be built. Only a small part of these companies does not propose any predetermined
solution, neither architecturally nor technologically, making available its know-how
for the realization of a completely customized project. In the latter case, the role of
the company is not represented by the sale of a product, but by a service subordinated
to the architect’s work.
If we consider the almost total absence of prefabricated houses in southern Italy,
it is easy to deduce that if in this area we wanted to build a prefabricated dwelling, it
would be necessary to turn to one of the many companies based in the northern Italy,
which probably develop technological solutions far from the needs of the Mediter-
ranean climate. If we consider then the possibility (anything but remote) of a direct
relationship between the company and the client that totally excludes the architect,
6 Ligniusis the Italian National Association of Prefabricated Wooden Houses (Associazione

Nazionale Case Prefabbricate in Legno), www.lignius.it.
the problem will extend to include, in addition to the possible technical inadequacy,
even the inconsistency of architectural solutions with respect to the historical and
cultural context.
However, it is of fundamental importance that the irreplaceable work of the archi-
tect should not be overlooked in the affirmation and diffusion of prefabrication, as
a means for the transition to a building process and a more sustainable life cycle
of buildings. This is necessary in order not to re-establish a model of design and
construction based on standardization which, although it should be in principle more
efficient and ecological, would once again be de-contextualized with respect to the
physical and historical-cultural environment. The fundamental role in the transition
towards an architecture produced by means of mass production but with the unique-
ness of a craft product, lies in the work of the architect. The designer, through com-
putational design, is at the centre of the new design-to-production process allowed
by the principles of the 4.0 industry and digital manufacturing.
The framework proposed here has already been experimented by some avant-garde
architects. The first results in this sense were obtained by a Californian company
called Cover, founded in 2014 by two architects, Alexis Rivas and Jemuel Joseph
(McKnight 2017). The two architects, driven by the conviction of the usefulness
of digital tools for the optimization of the design and construction process, have
created their own design software which, through algorithms, informs the project of
any possible conditioning such as: geospatial information, the path of the sun for the
determination of the position of the walls and glazed surfaces, building regulations,
the details of the construction site and the lifestyle of the clients.
The projects developed by the designer with the help of this software, after having
passed customer scrutiny, moves to a phase of computer optimization of the man-
ufacturing process of the components. Subsequently the parts of the building are
tailor made at the factory and simply dry assembled on site. By moving the build-
ing process at the factory, the company claims to have reduced the production of
waste material by 80%. Moreover, the use of digital technology not only facilitates
the construction of energy-efficient homes, but the precise execution in a controlled
environment ensures that what is fabricated is in line with what was planned during
the design phase.
Therefore, the construction of prefabricated dwellings lends itself perfectly to
becoming a fully digitalized process. The designer develops his own (computational)
project, which is transmitted to the digital manufacturing machines, optimizing the
process for the realization of the components, in order to reduce waste of any kind.
All components are made and finally assembled according to a precision that is
only possible in an industrialized process, thus reflecting the high quality standards
required by the project in the finished construction.
In this perspective fits the design experience of the Ecodomus housing prototype.
With the construction of this prototype some technical solutions related to prefab-
rication have been tested and linked to the potentialities of Industry 4.0. Even if
there is not yet a full realization of the production processes made possible by the
fourth Industrial Revolution, it has been possible to experiment with the design and
construction of non-standard prefabricated elements, to be introduced in the market
of prefabricated wooden houses.
5 Explorative Research: Introducing Non-standard

Prefabricated Elements with the Ecodomus Project
5.1 Design Goals and Principles
The design of Ecodomus represented an important moment of project synthesis of

the theoretical evaluations and reflections that will be reported in the first part of the
chapter, as an example of design influenced directly by the location, bioclimatic and
historical-cultural aspects and of industrial production through the use of wood to
create highly customized construction solutions.
Ecodomus is a house made of structural wooden panels, cut with CNC machines,
designed to respond adequately to the needs of a contemporary house in Mediter-
ranean climate, paying particular attention to the optimization of solar exposure,
the use of materials for the purposes of internal comfort and sustainability of the
building, as well as energy efficiency and structural optimization (Fig. 1).
This housing solution arises from the idea of obtaining an architectural typology
based on minimal modules with irregular geometry that can be aggregated accord-
ing to different configurations that planimetrically reproduce the complexity of the
historical agglomerations of the Mediterranean tradition (Fig. 2). This planimetric
Fig. 1 Rendering of Ecodomus

Fig. 2 Examples of triangular based aggregation patterns
arrangement does not represent a purely aesthetic choice, or an ideal reference to the
past, but finds an explanation in the advantages that the historical agglomerations
derived from such a conformation. The close and irregular arrangement allowed for a
mutual shading of the buildings and the streets, which avoiding direct irradiation were
more liveable even in summer. Furthermore, designing with an irregularly shaped
floor offered the opportunity to test the potential of digital manufacturing tools (CNC
machines) in order to simplify and make possible the manufacturing of non-standard
elements. This potential opens up new scenarios in the development of irregularly
shaped housing modules, useful in those contexts where it is not possible to install
modules with regular shapes due to the complexity of the basic lot geometry (such
as in the historical centres of the Mediterranean).
From a climatic point of view, Ecodomus is designed for the Mediterranean cli-
mate is characterized by mild and rainy winters and hot summers with long periods
of drought. The temperature variations during the day are minimal, as the tem-
perature variation during the year is rather limited. The adaptation of buildings to
these peculiar climatic conditions must therefore provide an answer to two opposite
requirements: on the one hand, to favour the solar thermal contributions and to con-
tain the thermal dispersions in the cold season, and on the other hand, to reduce the
sun exposure and promote natural ventilation in hot season. Hence the importance
for a dwelling in Mediterranean climate to be constituted by a building envelope
having a high thermal inertia, able to defend the internal environment not only from
Fig. 3 Full-scale model of the experimental wood-only connection
the winter cold but also and above all from the summer heat. For the building enve-
lope of Ecodomus, therefore, cross-laminated timber (CLT) panels were used, cut
with numerical control machines, in order to make the wall load-bearing again. The
structural panels were used to create a box system, braced in the middle by two
transverse walls. The choice of this construction system was dictated by numerous
factors, including: a considerable reduction in construction time, as well as excellent
performance of resistance to earthquakes, fire resistance, living comfort and thermal
insulation.
The most innovative aspect of the prototype is represented by the study of a new
type of structural joint that allow at the same time, a fast and precise assembly and
a better three-dimensional connection between the walls and floors, using the tra-
ditional dovetail joints taken up by traditional woodworking, in order to realize the
connection between these parts. These joints were then completed with wooden big
dowels which, inserted perpendicular to the direction of the dovetail joint, strengthen
the connection, however allowing a dissipation of energy in the event of an earth-
quake, through the elasticity and homogeneity of the material. This experimentation
is aimed at reducing the use of screws and metal plates to connect the CLT panels
in favour of a total material and structural homogeneity of the load bearing enve-
lope (Fig. 3). A wood-only connection, in fact, avoiding the use of metal elements
(nails, bolts, brackets) avoids the effects of wear caused by the interaction between
two materials with deeply different physical-chemical properties, obtaining a more
homogeneous and solid junction compared to a wood-steel one.
The basic module that constitutes the house has a gross surface area of 62 m2 . This
surface has been evaluated by the designers as the minimum surface within which,
Fig. 4 Plan of the single module and the double module
considering the peculiar planimetric geometry (parallelogram), all the fundamental

functions of living can find an adequate answer. A larger surface can be achieved
by aggregating multiple modules (Fig. 4). Functional distribution solutions have
therefore been studied both for the single module and for aggregations of several
modules. As the space needs of its inhabitants vary over time, the dwelling can
therefore respond by increasing its surface with the construction of an additional
module.
The basic module is characterized by a floor at the ends of which are placed two
triangular verandas which rectifying the interior spaces in addition to performing a
shading and filtering function with the exterior. Inside the rectified floor are placed
the two main rooms, the living room and the bedroom. At the centre the bathroom
is placed which, in the case of aggregation between multiple modules, acts as a
distribution node between them.
5.2 Bioclimatic Design
In order to obtain an energy-efficient house, it was decided not only to provide for
the installation of photovoltaic and solar-thermal systems, but to reduce the energy
needs of the building to a minimum. To pursue this goal, it was necessary to con-
sider all the design choices, starting from the shape of the architectural organism
which, determining the quantity of heat loss surface, decisively influences its energy
performance; in fact, we chose a compact volume with few openings.
Much attention has been paid to the choice of orientation. Considering that the
construction of the prototype would take place in the city of Termini Imerese, in the
province of Palermo (Italy), we proceeded to virtually simulate the solar exposure of
the building at three different times of the day (8:00 am, 12:00 am, 4:00 pm) as the
season varies (Fig. 5). Through this method of verification, an inclination of 9° on
the north-south axis was chosen as the optimal orientation, with respect to the site’s
latitude (13.46°), in such a way as to favour the passive behaviour of the building.
The position of the rooms, was chosen by evaluating the use of the spaces during
the day with respect to the orientation (Figs. 6 and 7). The living area, because is the
part of the house occupied for a greater number of hours during the day, is located
on the southern front, in order to exploit the free contributions of solar radiation in
the winter. The house, in fact, has a large glazed opening on the southern façade that
allows a greater capture of solar energy in winter, when the sun’s rays, being lower on
the horizon, penetrate directly into the house; while in the summer, as the sun is higher,
the sun’s rays break on the door’s triangular cover shading the interiors. The large
south window is set back by 2 m from the building’s edge, ensuring internal thermal
comfort even during the summer, avoiding solar radiation to penetrate directly and
produce overheating effects and glare.
The sides exposed to the east have good sunlight during the morning hours, pro-
longed in winter and more intense in summer. The sides exposed to the west have
medium sunlight during the afternoon hours, prolonged in winter and intense in sum-
mer; in order to avoid overheating on this front, shading systems have been designed.
In the single-dwelling version, the east and west facing sides are completely enclosed,
with the exception of a small west opening of the bathroom. The north-facing sides
are hit by the sun’s rays only at dawn and at sunset with a sloping inclination to the
ground, so in the two-module version to the north there are no holes, while in the
single-module version it has a modest opening size.
Fig. 5 Simulation of solar exposure
A further measure aimed at improving the passive behaviour of the building is the
choice of the white colour for the opaque building envelope. The white walls in fact,
are able to reflect more than 90% of visible solar radiation. The use of moucharabieh
is the last measure aimed at controlling solar radiation. Their use has been taken up,
revised and updated, from the Arab-Mediterranean tradition, in which this device of
solar shading and natural ventilation is traditionally used in the Arab culture to cool
the rooms and decorate them at the time same (Fig. 8).
5.3 Study Models of Assembly and Joints Details
In order to obtain an energy one of the undeniable advantages of prefabrication, in

addition to precision and speed of execution, is the ability to manage the entire build-
ing construction process from the very beginning design phase. And perhaps even
more important, is the attitude of who designs prefabricated structures to evaluate
all the appropriate modifications to the project in order to optimize its construction.
For the Ecodomus project, the construction was simulated in advance, both
through three-dimensional modelling (fundamental for the previously illustrated
Fig. 6 Simulation of the solar exposure of the interiors, in the two-module solution, during the
summer
solar exposure simulations), and through physical models of study (Fig. 9). A first
full-scale study model was necessary to verify the feasibility of the wood-only con-
nection, as hypothesized in the preliminary design phase. It has been decided to
simulate in real scale the wooden joint in order to proceed subsequently in safety to
the realization of the entire structure. The construction of the model has given satis-
factory results, both in terms of feasibility of construction, both in terms of strength
and solidity of the joint (Fig. 3). Further investigations were carried out, naturally, to
verify the structural behaviour of the node through the use of structural engineering
software, which is not the object of this discussion.
With a second study model, all the CLT structural panels and the metal foundation
portals have been reproduced in scale using 3D printing (Fig. 10). For a prefabricated
structure the exact consequentiality and the precision of assembly are aspects of fun-
damental importance and, given the size and weight of the elements to be assembled,
any uncertainty during the construction phases should be avoided. In this way, it
was possible to verify the correctness of the assembly sequence of the parts assumed
beforehand.
Fig. 7 Simulation of the solar exposure of the interiors, in the two-module solution, during the
winter
Fig. 8 Rendering of Ecodomus
5.4 The Construction of the Prototype
One module of the Ecodomus housing prototype was built on a land owned by the
company Gervasi Srl in Termini Imerese. For its construction CLT structural panels
were used. The CLT panels were made with five layers of solid timber, in which each
layer of boards is orientated perpendicular to adjacent layers and sticked together by
using formaldehyde-free glues (E1 category).
Fig. 9 3D simulation of assembly phases

Fig. 10 3D printed model of the structure for the simulation of the assembly phases
The CLT wall panels were installed with all the openings already cut with CNC
machines (Fig. 11). Their installation was greatly facilitated by the connection
through the dovetail joints allowed to connect precisely and quickly the different
panels together (Figs. 12, 13 and 14), later connected with big wooden dowels (as
required by the experimental only-wood joint developed for this project) (Fig. 15).
The wooden “box” is raised above the ground by means of steel foundation portals
(braced by metal cables) to allow the ventilation of the floor and to avoid rising damp,
which could ruin the wooden structure itself. The foundation portals are connected
to the floor slab panels by means of a bolted connection, with the interposition of an
EPDM rubber seal layer. The whole is supported by micropiles foundation fixed in
the ground.
Following the installation of the floor slab panels, the load-bearing walls and
the roof panels are laid, starting from the central modules and continuing with the
external ones, recreating the box-like structure (Fig. 16). The CLT panels were then
coated both internally and externally to make the finished wall. Internally, the walls
were completed with kenaf and hemp fiber insulation panels and a plasterboard finish
layer. Instead externally, a layer of hempcrete insulation has been realized, finished
with a layer of bio-lime and pozzolan (Fig. 17).
Fig. 11 Photo of the assembly of the floor slab on site
Fig. 12 Installation of CLT wall panels facilitated by the connection through the dovetail joints
Fig. 13 Installation of CLT wall panels facilitated by the connection through the dovetail joints
Fig. 14 Installation of CLT roof panels
Fig. 15 The CLT panels connected with big wooden dowels

Fig. 16 The construction site during the completion of the work
5.5 Final Considerations and Future Developments
From the experience of prototype construction it has been possible to see how the
realization times, compared to traditional buildings, are more than halved, speed
that translates into an appreciable economic saving. Among the various construction
phases, the realization of the insulation layer with hemp fiber and sprayed lime is
improvable. A possible alternative is the use of ready-made hempcrete insulating pan-
els, the use of which would also improve their reversibility. The low environmental
impact of the building is favoured by the almost exclusive use of wood for structural
components and natural, biocompatible or recyclable materials for the remaining
components. In addition, the high level of mechanical properties and energy perfor-
mance of these materials, whose production and installation does not require high
consumption of energy and resources, providing moreover reduced management and
maintenance costs, give the building high sustainability capabilities in the broadest
sense of the term. However, it must be underlined that to make the prototype it was
necessary to transport CLT panels produced in Austria to Sicily. To further increase
Fig. 17 Realization of hempcrete insulation and final layer of bio-lime and pozzolan
the level of sustainability of the process, it would be desirable to use materials that
do not require long road transport. Ecodomus (Figs. 18 and 19) seems to confirm the
widespread tendency to design prefabricated dwellings only of the isolated single-
family type, an aspect that probably partly determines the lack of interest on the
subject by large sections of users. However, with this project, in addition to having
provided different planimetric configurations, it was considered the possibility of
realize multi-storey housing units. Certainly, to carry out such an experimentation
would require more resources, both intellectual (more complex structural calcula-
tions) than economic ones, but, considering the growing concern for the consumption
of the soil, it would be an important evolution of this type of experimentation.
6 Conclusion
In conclusion, this prototypical experimentation, while certainly not taking on the

value of completed demonstration of the topics dealt with, has constituted an impor-
tant moment to increase the awareness of a complex issue such as the feasibility
of creating fully customized prefabricated houses, through the use of digital man-
ufacturing tools and the potential of the principles of Industry 4.0. The irregular
geometry of the module (oblique parallelepiped) and the study of complex spatial
Fig. 18 Photo of the finished prototype
aggregations and complex three-dimensional joints has demonstrated the potential

of mass customization applied to large construction elements. It has been found that
the use of digital manufacturing tools and the potential of the principles of Industry
4.0. Thanks to the great production flexibility that characterize it, can allow the use
of prefabrication (understood as industrial production of building components to be
assembled with simple stages of assembly on site) for the construction of buildings
that, in addition to benefiting from all the advantages that characterize this con-
struction technique, such as speed, economy and precision of execution, can enjoy
a customization at all levels of construction, from the adaptation to the particular
geometric configuration of a lot up to the choice of technological solutions adapted
to the climatic conditions of the project location.
Fig. 19 Photo of the finished prototype

References
Menges Achim (2015) The new cyber-physical making in architecture: computational construction.
Arch Des 85(5):28–33. https://doi.org/10.1002/ad.1950
Blanchet M, Rinn T, Von Thaden G, De Thieulloy G (2014) Industry 4.0: The new industrial
revolution-how Europe will succeed. Hg v Roland Berger Strategy Consultants GmbH München
Abgerufen am 1105 2014, unter http://www.rolandberger.com/media/pdf/Roland_Berger_TAB_
Industry_4_0_20140403.pdf
Fab Lab House (2010) Home. http://www.fablabhouse.com/en. Accessed 1 Oct 2017
Gallanti F (2011) Fab Lab House. I raggi del sole curvano: un prototipo di casa auto-sufficiente real-
izzato dagli studenti di IAAC e MIT. http://www.abitare.it/it/design/2011/01/31/fab-lab-house/.
Accessed 15 Oct 2017
Gramazio F, Matthias K, Willmann J (2014) The robotic touch. Park Books, Zurich
Kagermann H, Helbig J, Hellinger A, Wahlster W (2013) Recommendations for implementing the
strategic initiative INDUSTRIE 4.0: securing the future of German manufacturing industry; final
report of the Industrie 4.0 Working Group. Forschungsunion
Kostas T (2006) Algoritmic architecture. Elsevier, Oxford
McKinght J (2017) Cover designs backyard studio in LA using computer algorithms. https://www.
dezeen.com/2017/10/23/cover-computer-algorithm-design-prefabricated-backyard-studio-los-
angeles/. Accessed 28 May 2018
Menges A, Ahlquist S (2011) Computational design thinking: computation design thinking. Wiley,
London
Naboni R, Paoletti I (2015) Advanced customization in architectural design and construction.
Springer, Milan
Negroponte N (1970) The architecture machine. MIT press, Cambridge
Paoletti I (2010) Costruire le forme complesse: Innovazione, industrializzazione e trasferimento per
il progetto di architettura. Maggioli Editore, Santarcangelo di Romagna
Paoletti I (2018) Mass customization in the era of industry 4.0: towards immaterial building tech-
nology. In: Informed architecture. Springer, New York, pp 77–87
Sutherland IE (1964) Sketchpad a man-machine graphical communication system. Simulation
2(5):R-3–R-20. https://doi.org/10.1177/003754976400200514
Tseng MM, Jiao J (2001) Mass customization. In: Handbook of industrial engineering. https://doi.
org/10.1002/9780470172339.ch25
WikiHouse (2017) Library. https://wikihouse.cc/library. Accessed 15 Oct 2017
HOUSE 1 Protostructure: Enhancement
of Spatial Imagination
and Craftsmanship Between the Digital
and the Analogical
Dieter Dietz, Dario Negueruela del Castillo, Agathe Mignon

and Julien Lafontaine Carboni
Abstract Conceived around the concept of protostructure, HOUSE 1 deploys a

strategy to answer a daring but simple question: How could we design a house
between almost 300 people? The unique pedagogical framework of ALICE, first year
Architectural Design course, proposes the integration of a series of full scale physical
wooden constructs, enacting collaborative thinking and drawing on collective spatial
knowledge. The protostructure constitutes at once both a material and immaterial
open source support for the individual and collective interventions by the students.
Its material dimension as a physical construction is invested and complemented by
the immateriality of the guiding scheme. In this article, we review the steps in the
development of the theoretical model and physical implementation of HOUSE 1
and discuss its relevance with regards to the relation between analogical and digital
modes of engagement, pedagogical frameworks and spatial cognitive strategies. This
implementation of the protostructure shows its potential as a tool to approach wood
design, through a combination of digital and analogical processes, enhancing the
deployment of spatial cognitive strategies with the use of wood as a material through
and with which to think about space.
Keywords Protostructure · Collaborative design · Pedagogy · Craftsmanship

Open source · Database · Spatial imagination · Parametric design
Fabrication strategies
1 Introduction
Within contemporary approaches to wood design through digital processes, con-

siderations of how tools foster interaction and remain appropriable constitutes a
promising line of research, albeit somewhat secondary. In order to fully incorporate
D. Dietz · D. N. del Castillo (B) · A. Mignon · J. L. Carboni

ALICE Space Design Workshop, Federal Polytechnic School of Lausanne, Lausanne, Switzerland
e-mail: dario.negueruela@epfl.ch

1230 D. Dietz et al.
and consider how these aspects may contribute significantly to digital wood design
and architecture, we must look at educational and participatory design frameworks,
where such concepts are most developed. Responding to these potentials and chal-
lenges, at ALICE we have developed and implemented, together with students of
architecture, the theoretical, material and methodological framework that we intro-
duce in this chapter around the built case-study of HOUSE 1. These experiences
explore the full potential of combining a generative design scheme with a strong
participatory and engaging dimension.
In convergence with the recent paradigm shift favouring agency and empower-
ment, students of architecture tend to be increasingly considered as active agents in
their learning process. Deep spatial and social engagement are regarded as roots of a
collective framework, enacting collaborative thinking and drawing on spatial knowl-
edge. We contend that architecture through the complexity of the design and archi-
tectural design, sharpened by globalized problematics and digital mutation should
be accounted for as a collective process, place of knowledge-sharing, as well as a
support for a personal and collective individuation.1 Today the natural complexity of
design processes is enhanced in view of globalized problematics and digital muta-
tion. In that scope, first-year students in architecture at École Polytechnique Fédérale
de Lausanne (EPFL) are asked to engage themselves with space—to act on it and to
conceive it, using different tools to project and imagine it. As such space is presented
as a source of challenges, and not of theorems. Conceived around the concept of pro-
tostructure,2 the pedagogical framework of the Atelier de la Conception de l’Espace
(ALICE laboratory), has the objective to foster student skills through individual and
collective intervention. It deploys a strategy that seeks to answer a daring but simple
question: How could we design a house between almost 300 people?
As an answer to this question, the ALICE first year Architectural Design Course
proposes the integration of a series of full scale wooden constructs. HOUSE 1,
developed as the first of a series, is conceived around the concept of protostructure,
1 For Simondon, individuation is a synonym of ontogenesis. In his article Fifty Key Terms in the
work of Gilbert Simodon, Jean-Hugues Barthélémy defines the simodonian ontogenesis as such:
‘This term is first of all a synonym of individuation, because individuation, for Simondon, is
genesis. In biology, ontogenesis is also the genesis of the individual; in this case, it is distinguished
from ‘phylogenesis’, which is genesis of the species. However, Simondon also applies this term to
philosophical theory itself, because the ‘knowledge’ of individuation is ‘individuation of knowledge’
(Simondon 2005; 36). This is the properly Simondonian mode of overcoming the subject/object
opposition in view of a non-objectifying philosophical ‘knowledge’. […] In the second instance, it
is the term ontogenesis itself that is enlarged in order to refer to the ‘becoming of being’ (Simondon
2005; 25) in general, and thus to individuation as the genesis of the individual and its associated
milieu.’ (Barthélémy and Boever 2012). For a discussion of Simondon’s theory in relation to the
concept of protostructures and protofigurations, see Lafontaine Carboni, 2018, Cité de l’Architecture
et du patrimoine.
2 By protostructure we refer to that which through its use and interaction with, allows for the
extension of one’s cognitive and agential capacities. It facilitates the mobilization of additional
emergent resources, before unnoticed or unexplored. For a more detailed definition, see pages 3
and 4.
HOUSE 1 Protostructure: Enhancement of Spatial Imagination … 1231
developed at the core of the research activities of the lab.3 The protostructure con-
stitutes at once both a material and immaterial support of students’ interaction. Its
material dimension as a physical construction is invested and complemented by the
immateriality of the guiding scheme. Furthermore, the three-dimensional field of
the protostructure locates every individual proposition or intervention in relation to
the ones from the other participants, vertically and horizontally. It embodies each
collective situation in a physical medium. At the same time, a temporal framework
locates every action into a broader system. Thereby a set of rules is shared by all
actors, enabling them to interact within a gravitational system of structural logic and
to operate with similar vocabularies (types of response) in order to construct common
knowledge.
The concept of protostructure has been developed by ALICE over several years
in the context of diverse teaching and research formats. At its outset, there are two
major intentions. First, as already mentioned, the concept of protostructure is oper-
ating as a hinge between physical articulation and conceptual idea. Linked to a
temporal framework it literally constructs pathways from idea to built form. Sec-
ond, protostructures are able to federate collective action. The idea of structural
supports engages the spatio-temporal dimension of people appropriating space in all
its dimensions.4 As such, it opposes the idea of the territory as a tabula rasa, where
architectures would be placed as objects one next to the other, implying exclusive
individual or collective ownership of the planet’s surface. In conceiving a project in a
3-dimensional grid space, a project will automatically be, not only next to the other,
but also in mutual relationship with every aspect around it, often superimposed with
other projects below and above. This relationship will automatically ask for nego-
tiation between all actors involved in the conception process of the overall spatial
construct.
Before starting the HOUSE 1 program, ALICE proposed different kinds of struc-
tures at model-scale concerning smaller groups of students. Called Champs, or
Matrice,5 these structures are comparable from one workshop group to the other, but
not superimposable. Even though conception challenges did not include full-scale
embodiment, the common structural basis allowed for collective appropriation and
communication based on a same basis. The definition of the concept of protostruc-
ture forged by the lab has later concretized this research. Defining simultaneously the
physical form of a structure (full-scale wood assembly) and a temporal framework
(temporality, program and protocol), the neologism was introduced with the inten-
tion to allow for plural experiences and experimentation under a same denomination
(Fig. 1).
3 HOUSE 1 is the final project of the teaching program “Inside Paris”, conducted during the school
year 2015–16 by the laboratory.
4 Certain similarities with the radical propositions of Price, Friedman, Constant and others can be
seen here.
5 Called “matrice” or matrix during the years 2010–11, 2011–12 and 2012–13, for the teaching
programs Villes Archipels, Ville Entière and Airborne: Non-Site/Earthwork, “champ” or field for
the 2013–14’ program Where Are We Now?, and finally protostructure for the last three years
programs, Inside Paris, Inside Zurich and Inside Brussels.
Fig. 1 HOUSE 1 final critics day. © Joanne Nussbaum, ALICE EPFL
Mignon has defined the term as such6 : ‘The term protostructure is constituted by
the prefix proto and the substantive structure. The part of the prefix coming from the
Ancient Greek doesn’t need hyphens; protostructure is so a single word. We equally
notice that the prefix is Greek while the substantive is Latin. As such the word is a
hybrid (in line with sociology for example).
Structure is coming from the Latin structura, itself derived from the supine struere,
the action of assembling, piling up, but also building, arranging or weaving. The
current substantive also indicates the assembly and the resulting form, or even the
organisation of elements. It tends to communicate the idea of a set of systems that
could be either physical, mental or informational. In the Dictionnaire d’histoire et de
philosophie des Sciences (Dictionary of History and Philosophy of Science), Lecourt
(2006) uses an architectural metaphor in order to illustrate the hard to define term:
It is the hinge between plural parts, by means of rigid links often hidden. It is the skeleton or
the framework of a building, which provides its stability and cohesion, but not necessarily
its meanings. The structure is indifferent to superficial charms of appearance and dressing,
it is behind the scenes.
The prefix proto, coming from the Greek form πρîτoς, means first. It is composed
of the basis pro, forward, and the superlative tatos. It points at a primary character of
the substantive, holding potential of a dynamic row of events. This notion of process
is also apparent in prototype, the first model of a series. It is also distinguished from
6 Mignon, Agathe. Protostructure, archeology and hypothesis of a support structure. Doctoral thesis
(on going), ALICE EPFL.

the infrastructure that suggests a static point of view, without meaning of further
action or temporal proceeding.
The neologism protostructure not only captures a process of creation and con-
struction, but it also aims to establish a transversal view upon a series of realizations
and architectural theories. A protostructure designates a system or a building in a
primary state. The prefix introduces the temporality placing the structural element as
a support and an object of the process, in a state prior to any evolution. A protostruc-
ture possesses or demonstrates the capacity to adapt; its own nature is subordinated
to future uses; it will evolve, change appearance, function, become a trace again or
even disappear.
As it constitutes the physically and conceptually necessary base to every process of
creation, it institutes a relationship of control in imposing some of its characteristics.
We can use the metaphor of the white page to illustrate the concept of protostructure:
In a primary state, the white page is defined by the format, weight, grain, etc. Once
used as a support, its potential forms are infinite. The primary characteristics are
nevertheless not erased; they exist through the influence they had on the process: the
quality defines the types of technique, the format the limit, etc. Support and uses are
bound in a relationship of interdependence; the protostructure configured it ahead.
Since 2015 ALICE works with the concept and physical articulations of proto-
structures. HOUSE 1 was built in summer 2016 and was the first project realized on
the basis of this principle. The specific experience of the house as a configuration
exercise offered the possibility to explore principles of modelization inherent in dig-
ital technologies. These dynamics allow to intuitively establish connections between
different conceptions of a given space—analogical, digital and material. The proto-
structure of HOUSE 1, similar to the American balloon-frame system, consists of
a construction system based on orthogonal assemblage of discrete and standardized
units in three dimensions, revealing space in its Cartesian form. Each assembly is
thus a point, each wall a plane, within a grid that defines a primary rhythm. The inter-
action between the experience of physical space and the reading of a vector-based
space is mutual. It is based on the capacity of wood to offer precision both at the level
of its drawability—manual or digital—and of its constructability—by hand or with
automatized machinery. It is this precision that allows to acknowledge and perceive,
the interdependency between details and the whole. The assembly work becomes
the key operation in this system through the analogical engagement with each of the
smallest components (Fig. 2).
In this article, we review the steps of the development of the theoretical model
and physical implementation of HOUSE 1. The first question, ‘How could we design
a house between almost 300 people?’ has been enacted by the protostructure, and
experimented through HOUSE 1. In the context of this article we will now ask
whether and in what way the materialization of this process follows principles of a
digital mode of thinking. How does wood, the only material used during the concep-
tion and construction, permit to bound the digital/virtual and the physical dimensions
of design?
As part of HOUSE 1, the translation of the operating principles of the digital to
wood design will be investigated as a chance to favour agency and empowerment of
Fig. 2 Construction © Aloys Mutzenberg, ALICE EPFL
students in the learning process and architectural conception, giving birth to a shared
and collective spatial knowledge. In this respect, the concept of protostructure allows
us to unveil how this particular engagement with wood design and construction has
led to personal empowerment and agency of each of the participants.
To engage with the stated research questions in context of the HOUSE 1 exper-
iment, ALICE gathered the physical and conceptual work of 227 students, stu-
dio directors and student-assistants. For HOUSE 1 a particular protostructure was
designed by members of the ALICE research team: A cubic light structure of 11
by 11 by 11 m. The design of the structure is such that it is able to support its own
weight. However, projects need to rigidify the structure in order to be functional,
and superimposed designs will become mutually dependant as loads will have to
be transferred via neighboring projects. Importantly, the protostructure contains the
genetic code for the future interventions, allowing it to act at complementary levels.
On the one hand, it acts as a collective material support, to be shared, and enabling
the physical construction of one project by a large group. On the other, it acts as a
research support and as an experiment set-up in respect to the posed research ques-
tions. HOUSE 1—conceived, designed, and constructed by 227 people—is the final
realization of the Inside Paris program, developed in 2015–16 by ALICE and the
first-year design studio in architecture, EPFL.
Below we will first discuss the protostructure, as a physical support, a virtual space
of projection and a social space of interaction. In that sense, the wood construction
and assembly system permits to apply a series of digital concepts, as the location
system using reference points, or the decomposition of an organisation in the aim to
facilitate group cohesion within a shared space. In addition, we discuss the role of
the protostructure in enhancing spatial cognitive abilities. Secondly, we will focus
on the learning and creation process prolonging the extended metaphor between
the empirical approach to permit the agencies and the model of data management
massively used in the digital fields. Once more, wood will be investigated as a medium
enabling us to lead complex spatial research involving a considerable number of
actors. The choice of wood fostered an open and collaborative process, enabling a
shared language at the origin of a common database of spatial solutions.
2 HOUSE 1, (1) Between Conceptual, Material and Digital

Space
Nowadays, the industrial standardization of timber permits to achieve structures with

an astonishing precision; a thoroughness of execution in details is obtainable as much
as a hyper-regular three-dimensional structure. At the end of the nineteenth century,
the balloon-frame construction system-or Chicago construction-, merged from the
industrialization of sawmills and hardware factories (Pizzi 2003; Turan 2009). Mas-
sive industrial production of standardized wood pieces of low cross section and steel
nails started as a result of the diffusion of the mechanical saw, the invention of
multi-bladed saw and the development of a nail-production machine (Fig. 3).
Timber is perfectly parallelepipedic and makes it possible to trust pre-designed
assembling. In opposition to a wood-wood notch, nailing facilitates right-angled
corners through, for instance, the nailed boards method. Also, the frame method
consists in the pre-fabrication of walls, floor and roof using smaller rectilinear cross-
section, regularly spaced out. Amongst others, these methods are highly imageable
and use a simple Cartesian system of reference (orthonormal). As a trustable material
from its virtual to its actual uses, wood permits a quasi-immediate enaction in space
—conceptual, material or digital.
The protostructure of HOUSE 1 is entirely designed in pieces of wood of 10 ×
3 cm, 500 cm long, used in one full piece. This scheme originates in a reinterpre-
tation of the balloon-frame. Its departing hypothesis postulated that its design and
construction in wood allowed for an intuitive and quasi-immediate interaction with
it. Moreover, we hypothesized that these characteristics set up the grounds for the
development of a truly collective project built upon genuine collaboration, enabling
us also to lead a translation of digital modes of thought into actual space manufac-
turing (Fig. 4).
Height, width and length are equally of 11 m. Six vertical frames, two delimiting
opposite sides and four inside perpendicularly oriented, delineate one-metre-squares,
themselves composed of upright members also regularly placed every 70 cm. Two
horizontal frames offered intermediary levels at 4 and 6 m off the ground. The general
form emerges from the use and assembly of linear wood pieces, which are right-
angled and parallelepipedic. The repetition of these constructive elements outlines a
Fig. 3 HOUSE 1, plan and elevation of the protostructure with a numeral reference system (X; Y;
Z) CAD drawing. © ALICE EPFL
network of supports in several directions. Assembly nodes locate precise reference

points, also composing a self-positioning system in three-dimensional space.
The CNRTL (Centre National de Ressources Textuelles et Lexicales) defines
Cartesian space as an ‘axial system permitting the location of every point in a plane
using a pair of two numbers called Cartesian coordinates’. In this sense, the proto-
structure can easily be translated in the Cartesian system: every node is a point, every
wood pieces a line, every frame a plane. Each row can be located relatively to the
whole, and qualified by its direction. Thereafter, we can notice that most of digital
representations of architectural space are using exactly the same system; a point is
defined by three reference vectors.
The protostructure offers a physical support for intervention but also a system
of reference points, which favours comprehension and reading of space. In using
similar codes in a full-scale construction, HOUSE 1 provides some landmarks when
Fig. 4 Axonometric view of the two wood assembly, located using vectors in the tridimensional
grid of the protostructure. Pencil drawing on paper. © ALICE EPFL
building and utilizing it. A particular focus has been made on the notion of assembly;
the smallest unit of the protostructure’s genetic code enables to understand the whole
building, as much as all the merging complex creation. The small cross-section of
the American balloon-frame capacitates students to manipulate it easily, without any
need for particular knowledge or physical ability. In this particular case, only two
types of assembling were used: first, surface against surface, realized by pre-drilling
and mounting nuts; second, between a surface and the largest side of the timber,
realized in situ by using screws (Fig. 5)
Those two simple actions of screwing or tying up do not require particular knowl-
edge or skills. Thus, the manufacturing can be decomposed in terms of time (prefab-
rication) and workforce (shared between a large number of students).
Every kind of intervention on the protostructure, every kind of appropriation
through individual or collective projection can be translated to the assembly through
a value of 0 (Nuts) or 1 (Screws). The coupling of coordinates, length and assembly
types, are the kinds of information that allow the construction of the building. There-
after, it is reasonable to imagine that we can translate the HOUSE 1 project into a
series of codes that contain the students’ interactions.
The imageability of the protostructure, the three-dimensional orienting reference
system, the common language of assembling and mounting, its simple manufacture,
and the constitution of a database gathered, facilitates the translation from one space
to the other. The protostructure, as the conceptual, material and digital structure of
HOUSE 1, incarnates the interdependence of detail and whole, additionally allowing
individuals to identify the means and the function of their action at both individual
and collective level.
The first part was angled on the frame offered by the protostructure as a support
and a mediation device between spaces, to its constructive and qualitative charac-
teristics. As we shortly introduced it, the protostructure aims to favour agency and
empowerment, for students to be considered as active agents in their learning pro-
cess. In order to proceed with an analysis of this experience, we will focus on the
designing process in the studio throughout the academic year.
3 HOUSE 1 Protostructure (2), a Tool for Enhancing

Cognition, Imagination and Collaboration
A second relevant aspect of the protostructure consists in its role as a cognitive aid.
Spatial cognition in humans and other animals is a multimodal process taking advan-
tage of several aspects, among which, how the embodied space of the body mediates
in the cognition of its location in relation to an external physical environment (Tver-
sky 2009). Cognitive aids, in particular internal and external representations play
a non-trivial role in allowing for good and precise orientation and navigation in
space (Portugali 2005), through fixing and offloading cognitive information. The
method of working with the concept/tool of protostructure for HOUSE 1 was delib-
erately designed to take advantage and explore the role such a tool could play in
enhancing individual and collective spatial cognitive capacities. Cognitive aids are
Fig. 5 Four selected student proposals for interventions located in the protostructure. Pencil draw-
ing on paper. © ALICE EPFL
of accrued importance in dealing with particularly complex and detailed cognitive

and sensorimotor tasks, and especially so in collective settings where there is a need
to develop precise and concise sets of signs and codes to mediate such interaction.
Moreover, it has been noted that, in human cognition, spatial relations are typically
qualitative, approximate, categorical, or topological rather than metric or analogue

(Tversky 1993). Usually, this results in imprecisely held conceptions of space, which
are non-reconcilable with physical three-dimensional space.7 The protostructure, as
proposed in the HOUSE 1 project, attempts to facilitate an incremental and iterative
development of spatial cognitive capacities in students and participants through its
direct appeal to what is, at once, a conceptual and material three-dimensional grid
of reference. Moreover, the chosen scale and material—wood—have an impact on
the capacity to interact with this emergent and iterative protostructure as an external
spatial representation, by involving the body and its senses into the venture. “The
space around the body, that is, the space immediately surrounding us, the space
that functions for direct perception and potential action, is conceptualized in three
dimensions constructed out of the axes of the body or the world” (Tversky and Hard
2009). By developing the protostructure as an immersive interface, a particular fea-
ture is attempted: the embodied dimension of spatial cognition extends from the
body into its immediate surroundings, which is graspable and apprehendable thanks
to its simple yet generative orthogonal coordinate system, and through the use of
discretized structural wooden members. An additional relevant feature of HOUSE
1, is its materialization of both a three-dimensional orienting reference system and
of students’ projects, which behave as salient features, allowing for a very precise
spatial orientation that profits from both features.8
This role of the protostructure as a cognitive aid through its role of external repre-
sentation is well captured by the concept of scaffolding,9 which proposes that human
cognitive capacities both depend on and have been transformed by environmental
resources.
The interaction between the material and conceptual dimensions of the protostruc-
ture, moreover, speaks of our capacity to imagine and manipulate spatial organiza-
tional structure as a particular ‘cognitive interface’ that mediates between abstract
or non-perceptual knowledge and the ‘real world’ (Freksa 1991). In the case of
HOUSE 1, students showed the ability to “use” and practice space as grounds for
the translation of different sets of concepts, relations and assemblages. In this way,
students learnt to grasp certain more abstract and complex concepts and categories
by understanding them in terms of spatiality.10
7 Like the cases where individuals’ mental maps of locations wrongly capture distance and arrange-
ment of objects due to biases, preferences or relative importance.

8 This complementarity poises HOUSE 1 as a rich laboratory where to test postulates from both
cognitive maps theories (originating in the work of Tolman from 1948) and shape-parameter theories
of spatial navigation (ref). The discussion about these revolve around the question of whether an
“initial sense of place and direction arises from the overall geometry of the framework or from the
perception of distinctive landmarks within that framework” (Gallistel 2017).
9 This idea, originally present in the work of Vygotsky on the Zone of Proximal Development, has
been later further developed by Clark (1997) and Sterelny (2013, 2010). These theories clearly
imply a non-internalist and non-nativist view on human cognition and agency.
10 The thesis of Freksa (1991) states that “rather than generalizing by forming a common abstraction
for various domains we generalize by forming suitable analogies to a well-understood concrete

domain”. This presupposes this ability as the more common capacity of human cognition, against
4 Digital Analogies: (1) Database Approach of Education

in Architectural Design
Before any full-scale construction, the protostructure is introduced to students as a

problematic space to interact with, through the use of a set of tools. From the first
exercise to the HOUSE 1 manufacturing, the conscious limitation in terms of materi-
als, in correlation with the plethora of students’ experimentation, lead us to consider
the amount of individual solutions, inventions, and interventions as a database. We
consider the term database in its broader sense: a set of related data and the way it
is organized. Through this analogy, we are able to afford the complexity of a project
involved with 300 designers (classification method, queries, entry, hypertexts…).
In order to conceive their projects, students were invited to use two different cross-
sections; the first is the 10 × 3 cm section of the protostructure, the other is 2 × 5 cm
width and length. In a first phase, students are asked to intervene separately—without
any need to locate their project in comparison to others, they just had to locate
themselves in the protostructure. This exercise resulted in more than 100 different
proposals of similar scales, with the same material and in the same context (Figs. 6
and 7).
Students are also asked to produce a set of drawings and a model of the proto-
structure in order to include it in their projects. A plan, an elevation per student, and
one 1/10 model per studio constitute the standard basis fostering critical capacities
and spatial analysis skills. Therefore, students rapidly become experts in extract-
ing information from those drawings, as well as dexterous in re-constructing and
enacting conceptually, digitally or materially to proposition or intervention.
Architectural design exercises have neither right nor false answers, as proposals
that followed the protocol constitute in themselves a coherent entity, and provide
many possible answers to the same input. Since the protostructure is designed by
the repetition of a common detail, we can thereafter classify and analyse locations
in the reference system as reactions to the basic assembling and behaviours in the
protostructure. In this manner, a homogenous breadth of dispositifs stems from the
narrow range of the cross-sections. First, the standardization of the modality of
drawing the project; secondly, the defined set of tools facilitates the concretization of
an index of possibilities; and thirdly, the sharing and access to all this data. As a result,
not only are projects easily identifiable and readable as the nature of intervention is
the same—by nature, we mean method and goals—but also students become well
aware of the languages they used.
the commonly held belief among design professionals that this transformation or translation of
complex abstract thoughts into space is one of the key abilities in design and architecture, acquired
through practice and education.
Fig. 6 Series of 15 variations with the wood pieces assembled with screws. Pencil drawing on
paper. © ALICE EPFL
Fig. 7 12 Composition research drawings. Pencil drawing on paper. © ALICE EPFL
5 Digital Analogies: (2) an Open Source Architectural

Design
Subsequently, 12 projects are selected from the database in order to be developed,

designed and constructed by each studio in the full scale house. The graspability
of wood and the design codes enable the database, to be employed as a common
language favouring dialogue and exchange. This is, in fact, the basis of an Open
Source (OS) system (Open knowledge International n.d.). The implications of such
an approach are non-trivial, for it lays the grounds for the perfectibility of the project,
allowing for its projection into an active future life. Such accessibility and perfectibil-
ity do constitute a radically different status of the project with regards to what tradi-
tionally in the world of architecture have constituted fixed and immobile built works.
Thanks to the versatility and malleability of wood as a construction material, the
project acquires a different status between the mental model, graspable through a
quick induction of its structuring rules, and the built yet dynamic dimension. The
implications for participation and inclusivity are difficult to over-emphasize, as fol-
lowing the principles of OS systems, any user can then become a participant. In
fact, this approach to architectural and urban OS systems have seen a fertile scene
of proposals, debate and experimentation in the last years (see for instance Ratti
and Claudel 2015). From the privileged position of a hybrid setting combining the
Fig. 8 Plan of the location of the 12 interventions within the protostructure of HOUSE 1, identifying
«zones of tension». CAD drawing. © ALICE EPFL
academic framework and individual agency, it is interesting to note how this specific
pedagogical and projective experience combines exposure to the reality of the physi-
cal dimension of its construction and location, with the social aspect of participation,
empowerment and negotiation (Figs. 8 and 9).
The OS model of knowledge-sharing favours a horizontal hierarchy even in the
case of a considerable amount of people. Likewise, the multilayered human orga-
nization encourage a re-appropriation of the conceptual material provided through
different level of proximity. Indeed, the studio director is the referee of the official
academic program that introduces fixed output and tools, but requires individual inter-
pretation of the path. Student assistants (year four and five) lend levels of meaning
of the tasks to students, as well as airbornes (year two), encouraging student-to-
student learning. The same conceptual aids and program are described and handled
through several lenses, which encourage an active position facing the exercise; an
Fig. 9 Diagrammatic section of the location of the 12 interventions within the protostructure of
HOUSE 1, identifying «zones of tension». CAD drawing. © ALICE EPFL
intentional gradient of vagueness or ambiguity puts the objectives into perspective.

For instance, the concept of protostructure is a strong input by ALICE researchers.
After having been theorized by the lab, each agent transcribed the notion according
to her/his role; students addressed the input by producing their own definition, con-
tributing to a plurivocal theoretical production. Looking at their publication ALICE
Blog, we noticed a strong appropriation of the terms. Resources—conceptual and
material—become available thanks to the concourse of plural actors, both in terms
of individuals and meanings.
The database is composed of a set of individual and collective answers to the
exercise, as well as of conceptual and material inputs which are thus shared as an
OS system, but tries to encourage an individualized navigation system. The data
is either graspable or re-organizable, always in the process of transformation; this
in-between data and navigation system aims to set up a space for the inscribing on
the self into a collective experience, as a system of footnotes or reading notes in the
global experimentation. Through these uses, we claim that a database can be said to
be OS only if the navigation system enables individuals to appropriate the whole, by
inventing new meanings, and by mounting and dismounting new boards, leading to
a perception of the self.
6 Analogic Digitals: Towards Empowerment in Open

Source Systems
Knowledge is open if anyone is free to access, use, modify, and share it — subject, at most,
to measures that preserve provenance and openness. (Openknowledge)
In open movements, open is equivalent to free; open source software (OSS), free
software… With reference to the definition of the OS, which it is derived from, the
formulation ‘makes precise the meaning of ‘open’ with respect to knowledge, pro-
moting a robust commons in which anyone may participate, and interoperability is
maximized’ (Open knowledge International n.d.). Kelty sustains that the values of
OSS ‘reach directly into the heart of the legitimacy, certainty, reliability and espe-
cially the finality and temporality of the knowledge and infrastructures we collec-
tively create’ (Kelty 2008: 6–7). New overlooks on the relationships between data,
empowerment and agency, raised from social analysis of such movements (Couldry
and Powell 2014; Baack 2015). This organization model can be applied in other
domains (Demil and Lecocq 2006; Matei and Irimia 2014; Weber 2005) and tend
to favour empowerment. Nevertheless, apart from the potential criticism of reticular
forms of control (Stiegler 2008) derived from Foucauldian methods of analysis that
can be associated with Big Data, OSS projects unveil empowerment problematics
through the lens of usability (Rajanen and Iivari 2015). Thanks to the physical exper-
imentation of OS organization in collective architectural design, we can propose a
new insight on the potential shortcoming articulated on the notion of participatory
and egalitarian settings through open code, enhanced by graspability of the tools. An
inverted analogy from HOUSE 1 to digital mode of thought allows us to investigate
where personalization ends and empowerment starts.
As we briefly introduced, we argue that opening the code may not be sufficient;
the conflict between the free software movement and the open source can shed light
on specific aspects. Stallman stated that ‘The two terms describe almost the same cat-
egory of software, but they stand for views based on fundamentally different values.
Open source is a development methodology; free software is a social movement. For
the free software movement, free software is an ethical imperative, essential respect
for the users’ freedom’ (Stallman 2007).
To summarize, the different settings to favour empowerment of the protostructure
are the following:
– High usability of the reference system through a Cartesian model with high image-
ability and interdependent framework through repetitive operations.
– Simple manufacturing of the wooden structure, demanding no particular knowl-
edge or physical capacities providing high agency of participants, and narrow
range of assembling, easing the reading of the whole.
– Common representation tools—plan, elevation, model—at the same scale,
enabling a comparable data input and graspability of conceptual input through
multiple proximities of levels of interlocutors, and clear outputs; Individualized
reading/navigation system into the database of previous experimentations.
– Tool to enhance cognition, imagination and collaboration, through the appeal to

both embodied and extended dimensions. This facilitates a direct and fluid feed-
back between internalized and external spatial representations, where collabora-
tion is key to accomplish shared knowledge.
Not all of these allow to draw an analogy with OSS or FreeS. Nevertheless, both
need social interaction—birth of an ethical group—and openness of the code, plus
high agencies are central in the HOUSE 1 project. We contend that the protostruc-
ture is a development methodology, but above all an operating system for learning
processes through student-to-student and social interactions.
Rajanen and Iivari (2015) pointed out the lack of research addressing the ‘impor-
tant issues as regards power and politics in OSS’. At the opposite of empower-
ment resides Alienation. In the book “On the mode of existence of the technical
object”, the French philosopher Gilbert Simondon is giving an interesting defini-
tion of ‘Alienation’. He proposes a system of analysis consisting of three concepts,
‘Element/Individual/Set’, to understand genesis, evolution and transformation of
technical objects. In the words of Jean-Hugues Barthélémy, ‘the levels of analysis
in MEOT are classic, as the element composes the individual, and the individuals
compose the set’ (Barthélémy and Boever 2012: 210). Elements exist in two types,
the tool and the instrument. Both participate in the problem resolution of the indi-
vidual/technical individual and compose it. By using this system, Simondon is able
to formulate the ground of an industrial ‘psycho-physiological alienation’. At the
level of technical individuals, beings are able to perceive through instruments and
to act through tools. According to Simondon an important shift happens when uses
of technical individuals become industrial. He suggests that the industrial use of
the machine is destroying the system of ‘coupling between the inventive and orga-
nizing capacity of plural subjects’ (Simondon 1958: 342). Indeed, the role of the
worker changed: tool holders became an auxiliary of the machine, descending from
the level of the individual to that of the element. It is the contention of Simondon
that replacing humans with machines at the rank of the individual, is the cause of a
psycho-physiological alienation as much as the Marxist socio-economical one. Yet,
‘today, technicity has a tendency to reside in sets’ (Simondon 1958: 16). OS and FS
movements are perfectly illustrating this shift into modes of invention in technical
culture; empowerment is now processing at the level of sets.
Students are thus provided with tools to act first on the element level through
wood construction, thereafter on an individual level with architectural and spatial
invention, and finally on the set using diagonal and individualized reading of the
conceived database. With this, we aimed to go further with an open code of the
protostructure in order to engage students in concretizing invention at each level of
the ‘technical object’. With the objective of setting up a collaborative project, H1 does
not simply provide material support in the form of a wooden structure, but frames the
capacities of the material within a specific and overall logic of an open system. Thus, a
set of rigorous rules constitutes, beyond simple constraints to the project, a generative
playing field for the emergence of a community of participants and actors around
the enactment of a common objective and the resulting shared knowledge. The very
provision of these sets of generative rules and their materialization into a physical
support provide for the spatial conditions of a network. This, in turn, allows for a
number of possible combinations already contained in potential. This “virtuality” is,
thanks to the simplicity of the setup, the versatility and ease of transformation of
the material, which permits students to re-act at each level of the technical object,
inventing and imagining new solutions.
Finally, protostructure as both a concept and an operational setting between the
virtual and the actualized, the digital and the physically built, articulates space as a
radically different category. Beyond hegemonic stances that tend to consider space as
an extant object, protostructure poses an agential turn, proposing to consider space
as a capacity (Jiménez 2003; Negueruela Del Castillo 2015) (Fig. 10).
7 Epilogue: HOUSE 1 in Versailles (500–1000 Words)
At the same time as the HOUSE 1 project was launched at EPFL, first meetings
took place between ALICE EPFL and the Ecole Nationale Supérieure d’architecture
of Versailles in late 2015. A collaboration was launched between the schools with
Edouard Cabey, studio director with ALICE EPFL, and Cédric Libert, professor at
Versailles School of Architecture. The idea was to expand the experience of collabo-
rative work across schools and to observe how what we described above as a genetic
code would lead to further projects and configurations in terms of project-outcome,
and on spatial and social levels. The project was scheduled to take place in parallel
in terms of phases of first encounters between students and exchange during cri-
tiques, and in sequence in terms of construction, in order to allow for HOUSE 1 to
be disassembled on the EPFL-campus site and the protostructure to be relocated and
reassembled in Versailles. Also, certain elements of some projects were incorporated
in the following process leading to the HOUSE 1–12 CITIES project that occupied
one of the courtyards of the Versailles School of Architecture between September
2016 and January 2019. The collaboration dispensed a course about the 12 imaginary
cities by Superstudio.11 During the summer, a collaboration between Versailles and
the HOUSE 1’s students took place, with the aim to propose 12 more interventions
in the same protostructure, following the course “Twelve Cautionary Tales…”. The
new protostructure design consisted of a reconfiguration of the HOUSE 1 proto-
structure elements from a cubic disposition to a more linear arrangement, with two
halves of the square plan folded up. Thus, the rhythm of the elements and vertical
proportions were similar to HOUSE 1 at EFPL, while the overall spatial articula-
tion changed. In a first step, a limited number of EPFL students came to mount the
protostructure and explained the coding system (reference, assembling, conceptual
approach, etc.); thereafter, the 12 interventions were conceived. Participants from
Versailles visited the former protostructure, but did not assist to the elaboration and
11 Twelve Cautionary Tales for Christmas, (12 Ideal Cities), (Superstudio, 1971).
Fig. 10 HOUSE 1–12 cities, second phase of HOUSE 1 experimentation, at the Ecole Nationale
Supérieure de Versailles. © Agathe Mignon, ALICE EPFL
construction; EPFL’s students were relied upon as carriers of knowledge from their
past experience, and were key in bounding both instances.
Beyond the creative support that the protostructure has provided twice, HOUSE
1–12 CITIES at Versailles enhanced students’ capacity to use a genetic code; as
ideally is the case in an Open Source System. They took over the hardware and rear-
ranged the cartesian reference system in order to foster new interactions. In fact, the
empowerment process we tried to put in place went further than we expected. Not
only did students co-design a database and a navigation system of high-usability,
they were also able to directly conceive and manufacture spaces of the protostruc-
ture at each of Simondon’s levels, from the element, throughout the individual space,
to the global set. Handling structural balloon-frame elements, they (h)ac(k)ted the
learning process, sharing a collective spatial knowledge through agencies of wood
design. In this sense, the choice of wood for designing and as a construction material,
allowed participants to intuitively understand the role of details in a complex con-
struction, as well as to emphasize the importance of a common language in realizing
a collaborative project.
In the context of this article, we would like to stress the aspect that the concept
of protostructure, including its conceptual and physical dimensions, and its accom-
panying spatio-temporal timeline making it operational, bear further implications in
terms of access to tools and integration of synthetic processes. As in any discipline,
architecture employs a long series of tools. Many of those pertain to the realm of
simulation (Cross 2001). We have seen that the protostructure federated a smooth
integration of nearly any tool involved in the process. From sketch to hand drawing,
from construction drawings to models at any scales to mock-ups in 1:1, from pencil
to Japanese saw to computer aided design. Importantly, diverse 3d-software became
an integrated part of the process, naturally adopted by students in learning from
one another without any frontal teaching involved. In the final HOUSE 1 configura-
tion, room-projects entirely developed by hand-crafts were realized next to projects
scripted via parametric design software. The integration was particularly smooth as
all actors would refer to the tools at hands and most appropriate for a given process,
almost always combining an array of them to get to the projected aims. The same
can be said of the blog, which was, as mentioned, an important element in providing
access to information in this Open Source system. The blog would contain any infor-
mation from text to scanned drawings, to vectorized data-files, to photos, all arranged
in chronological order and edited by the students themselves. In fact, through the
rigorous structure of the concept of protostructure both in time and in space, the
strengths of computational thinking and iterative processes can be observed as a nat-
ural behaviour and outcome in the context of complex design processes involving
many actors.
Finally, this framework for collaborative design and construction bears conse-
quences for the development of shared spatial knowledge, further reaching into
questions of identity formation. Through interaction with the haptic and conceptual
dimension of the protostructure, students develop their own nuanced and intentional
stance with regards to what it means to deploy and invest space. This transformative
mise-en-espace partakes of a genuine collaborative effort, forging this dimension into
the very definition of what new digital wooden design and architecture are capable
of.
References
Baack S (2015) Datafication and empowerment: how the open data movement re-articulates notions
of democracy, participation, and journalism. Big Data Soc 2:205395171559463. https://doi.org/
10.1177/2053951715594634
Barthélémy J-H, Boever AD (2012) Glossary: fifty key terms in the works of Gilbert Simondon.
In: De Boever A, Murray A, Roffe J (eds) Gilbert Simondon: being and technology. Edinburgh
University Press, pp 203–231
Clark A (1997) Being there: putting brain, body, and world together again. MIT Press, Cambridge,
Mass
Couldry N, Powell A (2014) Big data from the bottom up. Big Data Soc 1:205395171453927.
https://doi.org/10.1177/2053951714539277
Cross N (2001) Designerly ways of knowing: design discipline versus design science. Des Issues
17:49–55. https://doi.org/10.1162/074793601750357196
De Boever A (ed) (2012) Gilbert Simondon: being and technology. Edinburgh University Press,
Edinburgh
Demil B, Lecocq X (2006) Neither market nor hierarchy nor network: the emergence of bazaar
governance. Organ Stud 27:1447–1466. https://doi.org/10.1177/0170840606067250
Freksa C (1991) Qualitative spatial reasoning. Springer
Gallistel CR (2017) Navigation: whence our sense of direction? Curr Biol 27:R108–R110. https://
doi.org/10.1016/j.cub.2016.11.044
Jiménez AC (2003) On space as a capacity. jroyaanthinst. J Royal Anthropol Inst 9:137–153
Kelty CM (2008) Two bits: the cultural significance of free software, Experimental futures. Duke
University Press, Durham
Lecourt D (ed) (2006) Dictionnaire d’histoire et philosophie des sciences, 4e éd., rev. et augmentée.
ed, Quadrige. Dicos poche. PUF, Paris
Matei A, Irimia SI (2014) Open source governance—a more ambitious cousin of collaborative
governance. Int J Public Adm 37:812–823. https://doi.org/10.1080/01900692.2014.907315
Negueruela Del Castillo D (2015) The city as a prototype. A frame for pragmatic social action.
In: Bianchetti C, Cogato Lanza E, Enver Kercuku A, Sampieri A, Voghera A (eds) Territories in
crisis. Architecture and urbanism facing changes in Europe. Jovis, Berlin, pp 222–234
Open Definition 2.1—open definition—defining open in open data, open content and open knowl-
edge [WWW Document], n.d. URL https://opendefinition.org/od/2.1/en/. Accessed 25 May 18)
Pizzi M (2003) The invention of the balloon frame, how it affected architecture in the New World.
The case of Chile. In: Proceedings of the first international congress on construction history,
Madrid, 20th–24th Jan 2003. Presented at the first international congress on construction history,
I. Juan de Herrera, Madrid, p 10
Portugali J (2005) Cognitive maps are over 60. In: Cohn AG, Mark DM (eds) Spatial information
theory. Lecture Notes in Computer Science. Springer, Berlin, Heidelberg, pp 251–264
Rajanen M, Iivari N (2015) Power, empowerment and open source usability. ACM Press, pp
3413–3422. https://doi.org/10.1145/2702123.2702441
Ratti C, Claudel M (2015) Open source architecture. Thames & Hudson, New York, New York
Simondon G (2005) L’individuation à la lumière des notions de forme et d’information, Krisis.
Millon, Grenoble
Simondon G (1958) Du mode d’existence des objets techniques. Aubier et Montaigne, Paris
Stallman R (2007) Why “Open Source” misses the point of free software. GNU.org
Sterelny K (2013) Cooperation and its evolution. MIT Press
Sterelny K (2010) Minds: extended or scaffolded? Phenomenol Cognitive Sci 9:465–481. https://
doi.org/10.1007/s11097-010-9174-y
Stiegler B (2008) Prendre soin, La bibliothèque des savoirs. Flammarion, Paris
The Cathedral and the Bazaar [WWW Document], n.d. URL http://www.catb.org/esr/writings/
homesteading/cathedral-bazaar/index.html. Accessed 25 May 2018
The Open Definition—open definition—defining open in open data, open content and open knowl-
edge [WWW Document], n.d. URL https://opendefinition.org/. Accessed 25 May 2018
Turan M (2009) Reconstructing the balloon frame: a study in the history of architectonics. METU
J Fac Archit 26:175–209. https://doi.org/10.4305/METU.JFA.2009.2.10
Tversky B (2009) Spatial Cognition. Embodied and situated. In: Robbins P, Aydede M (eds) The
Cambridge handbook of situated cognition. Cambridge University Press, Cambridge, New York,
pp 201–216
Tversky B (1993) Cognitive maps, cognitive collages, and spatial mental models. In: European
conference on spatial information theory. Springer, pp 14–24
Tversky B, Hard BM (2009) Embodied and disembodied cognition: spatial perspective-taking.
Cognition 110:124–129. https://doi.org/10.1016/j.cognition.2008.10.008
Weber S (2005) The success of open source. Harvard University Press, Cambridge, Massachusetts
Part V
Parametric Transfigurations
and Morphological Optimizations
KODAMA: A Polyhedron Sculpture
in the Forest at Arte Sella
Kengo Kuma, Marco Imperadori, Marco Clozza, Toshiki Hirano,
Andrea Vanossi and Federica Brunone
Abstract KODAMA is the new “sculpture” conceived by Kengo Kuma for Arte
Sella—The Contemporary Mountain. The process shows the concept, the develop-
ment, and the construction of a complex structure in massive wood (Italian larch),
a new sculpture in this famous Land art park. After the first options, which have
been discussed between the Japanese and the Italian teams, a solution with lattice
blades has been chosen. The first phase was to produce different scale mockups,
both analogic and digital in order to find proportion and dimensions. All the struc-
ture which creates a kind of “porous sphere” (a polyhedron) is made only by one
wood section, connected without nails or screws. Old toys, Chidori-like, are the ref-
erence and several scaled mockups have been produced with different woods: cedar,
larch, fir, and oak. Once assessed the architectural/spatial shape the model has been
studied for its structural behavior thanks to complex software analysis, which have
been studied using the parametric model (Rhino-Grasshopper). Models have been
used for testing the structure after assessing the real loads (mainly snow and wind).
Full scale 1:1 mockup have been realized both analogic and with CNC machines by
master carpenters. The final shape is an ideal “Teahouse” space in the middle of the
forest, close to Villa Strobele, which will be the new core of Arte Sella, dedicated to
famous architects and their ways to conceive art and nature using wood as a creative
material.
Keywords Wooden land art · Chidori · Lattice blades · Parametric design

CNC-based cutting
K. Kuma · T. Hirano
Kengo Kuma Lab, University of Tokyo, Tokyo, Japan
M. Imperadori (B) · A. Vanossi · F. Brunone
Department of Architecture, Built Environment, Construction Engineering,
e-mail: marco.imperadori@polimi.it
M. Clozza
D3WOOD, Lecco, Italy

1256 K. Kuma et al.
Fig. 1 KODAMA, the name of the pavilion, embraces the spirit of the beautiful forest of Arte
Sella: “KO” means “tree”, and “DAMA” means “spirit” or “sphere” in Japanese language (Credit
Kengo Kuma)
“KODAMA” (Fig. 1) is a permanent pavilion, which has been installed in the forest
of Arte Sella Foundation in Northern Italy.
Architecture since the 20th century has been mainly built of concrete, which
is massive, thus it is limited to professional workers who can participate in the
construction process. On contrary, we are interested in developing a democratic
architecture which is made out of smaller and lighter elements so that it opens the
construction process to everyone no matter they are non-professional. Many of the
pavilions designed by us in the past (Attali et al. 2016; Frampton 2013; Kuma 2018;
Kuma and Daniell 2015; Kuma and Ferrari 2016; Liotta and Belfiore 2012; Ejiri
2018; Vanossi 2018), such as “CIDORI” (“Yure”, and “Neowa Dome”) (Fig. 2)
that are made of small wooden elements are based on this idea, and we explored how
this can be further developed in this project.
We came up with a nested spherical geometry as an overall form of the pavilion
after several design iterations. The idea of wooden nested sphere quietly sitting in
the forest was quite appealing to us (Attali et al. 2016; Frampton 2013; Kuma 2018;
Kuma and Daniell 2015; Kuma and Ferrari 2016; Liotta and Belfiore 2012; Ejiri
2018).
We had initially considered using bent plywood sheets to create the nested spheres,
however, the idea was discarded due to a durability issue of the plywood and instead
we attempted to update the structural system developed for “Neowa Dome” pavilion.
The dome of the pavilion, which is 6 m in diameter and 4 m in height, is built by
joining 18 mm thick plywood sheets with various lengths using wooden dowels. In the
project, we explored a transparent and lightweight structure by using thin plywood
KODAMA: A Polyhedron Sculpture in the Forest at Arte Sella 1257
Fig. 2 On the left, the CIDORI pavilion explored a new structural system based on CIDORI Lattice,
a Japanese traditional technique lacing thin rectangular wood into a lattice by making a special notch
in the wood (Ph. credit Kengo Kuma Lab); on the right, the Neowa Dome is a 4 m high, 6 m in
diameter transparent dome, which consists of pieces of birch plywood which width is 200 mm and
thickness is 18 mm (Ph. credit Kim Yongkwan)
pieces, whereas we pursued a contrasting expression in this project using thick solid
wooden panels.
Six identical wooden panels with notches jointed (Seike 1978) with each other
to become a basic unit component of the structural system developed in this project.
The system sometimes gives a solid, sometimes transparent impression depending
on the viewing angle. Dimensions of the panel and the overall sphere were carefully
studied in order to maximize this visual effect. In the final plan for Arte Sella, the
sphere is 5 m in diameter with a cavity of 2.5 m sphere, and the panel is 58 mm in
thickness, 300 mm in width and 1000 mm in length, connected with millimetric cut
tolerance.
The name of the pavilion, “KODAMA”, is associated with a spirit in the beau-
tiful forest of Arte Sella: “KO” of means “tree”, and “DAMA” means “spirit” or
“sphere” in Japanese.
Kengo Kuma
1 Between Eastern and Western Philosophy: The Design

Concept
1.1 The Void, the Emptiness
East and West philosophy meets in KODAMA very strongly in the beautiful sur-
roundings of Arte Sella—The Contemporary Mountain (http://www.artesella.it/en/),
and the forest of Villa Strobele (Montibeller et al. 2017). While the main elements of
1258 K. Kuma et al.
Fig. 3 Initial design study attempted to create a nested sphere by combining bent plywood sheet
units (Credit Kengo Kuma Lab)
Western philosophy are Earth, Air, Water, and Fire, the Eastern one has one more:
the Void, the emptiness (Tanizaki 1977; Tenshin 2015).
In KODAMA this void is amplified into three levels: the convexity outside the
volume, the concavity inside and the porosity, the emptiness in between (Kuma 2018;
Liotta and Belfiore 2012). In order to achieve these spatial effects as an effective
representation of the design concept, many sketches, and drawings, and mock-ups
have defined the first phase of the design process, towards a visual and practical
experimentation (Fig. 3).
Finally, the multiple interpretations of void—as convexity, concavity, and empti-
ness in the porosity of the sphere—have been created by massive larch blades that
now define the structure of the art piece.
1.2 One Element for a Complex System
58 mm thickness, 300 mm in width and 1000 mm in length are the three dimensions
of the single larch element (Fig. 4), whose repetition origins the wooden, porous,
and spherical volume of KODAMA, giving it the rhythm of a complex entity, but
being at the same time a simple geometry (Fig. 5) (Kuma 2018; Kuma and Daniell
2015; Kuma and Ferrari 2016; Liotta and Belfiore 2012).
Fig. 4 Dimensions of the panel and the overall sphere were carefully studied in order to maximize
the solid-transparent ambiguous visual effect of the structure (Credit Kengo Kuma Lab)
Fig. 5 From the backside of the pavilion, it looks like a solid sphere, giving a mysterious impression,
while the structural system becomes transparent and bright inside the cavity of the sphere (Credit
Kengo Kuma Lab)
1260 K. Kuma et al.
The concept of the wooden structure, indeed, had the aim to achieve a single and
iconic volume made only by one element, as a particle of the whole. It is repeated,
connected and exalted across the construction, following Kengo Kuma’s theory of
particles and diffused structures, which leaves traces in many of his works, all around
the world. In a way, it has also a reference to Hegel’s philosophy of the elements, the
particles and the whole together, as a synthesis (Frampton 2013; Kuma 2018; Kuma
and Daniell 2015; Kuma and Ferrari 2016; Liotta and Belfiore 2012; Ejiri 2018).
2 Between Digital Parametric Design and Analogic

Mockups: An Iterative Process
Behind the wooden structure of KODAMA and its capability to embrace and synthe-
size both the prime conceptual ideas of emptiness and complexity through simplicity,
there is an extremely intense and complex design process (Imperadori and Senatore
2011; Vanossi 2018), carried out after many experimental tests, mechanical calcu-
lations, and virtual simulations. It has been an iterative process, balanced between
theoretical validations and practical experimentations, towards the optimization of
the wooden elements, their dimensions, the specific material, connections and, thus,
mechanical behavior and final impact.
Digital and parametrical models were used in order to control the geometry, assess
the structural design, and help the management of wooden pieces’ production, classi-
fication, and assembly. Digitalize the process means to gain a better control over each
phase and procedures, transferring the earlier ideas and sketches into an engineered
knowledge (Menges et al. 2016; Tedeschi 2014; Vanossi 2018; Vanossi et al. 2012,
2013).
Once the general shape and dimensions of the outer volume have been defined,
a series of different options about the possible structural approaches have been
assessed (Aicher et al. 2014; Kuma 2018; Vanossi and Imperadori 2013; UNI EN
1995-1-1:2014, 1995-1-2:2005; UNI 11035-1:2010, 11035-2:2010, 11035-3:2010).
The chosen one is based only on one, simple, wooden element, produced through a
CNC-based cutting process and whose assembly and interconnections were thought
to be realized without any kind of steel joints and nails. According to this structural
configuration, every single wooden piece mechanically behaves like a blade, a plate
which is charged on its own layer.
In the end, the structure has been defined as spherical volume built up by the
interconnection of wooden plates (58 × 300 × 1000 mm), the repetition of a node
of 6 panels along the whole geometry. No steel nails are present in the structure,
except for the anchoring to foundations, realized using steel bars and bolts. The
elements are connected to each other by gravity and their joints are fixed thanks to
the deformations induced by humidity.
3 The Engineering Process: Structural Design

and Parametrical Modeling
3.1 The Reasons for the Structural Analysis of a Monument
KODAMA Pavilion is characterized by considerable dimensions, being approxi-

mately a sphere with a diameter of five meters; moreover, it is placed at about 1000 m
a.s.l. (above sea level) in the Alps, where the snow load is significant. People can
enter the pavilion in all weather conditions (with snow and wind) and, therefore, the
structure must be safe. The weight of the workers must be supported during assembly
and no parts should be accidentally broken. Lastly, the pavilion has required con-
siderable planning effort, the material and production have their own costs and the
structural calculation is the instrument to guarantee the maximum durability, as well
as safety. Hence, the structural analysis represents the tool for the evaluation of such
a complex geometry, in terms of both the stress distribution across the elements, the
presence of any findings, and the elaboration of problem-solving hypotheses.
The calculation approach is completely innovative if compared to ordinary
wooden structures (Aicher et al. 2014), linked to the traditional static schemes. The
peculiar features of the wooden geometries led to consider the orthotropic relation
among the elements, the entity of the contact surfaces and the rigorous compu-
tation of the structural strains under the combination of proper weight, wind and
snow loads, instead of the simplified static approaches from standards (Circolare
02/04/2011; CNR-DT 207/2008, 206/2007; D.M. 14/01/2008, 06/05/2008; UNI EN
1995-1-1:2014, 1995-1-2:2005; UNI 11035-1:2010; 11035-1:2010, 11035-3:2010).
The structural verification of the wooden panels is under the validation of the worst
parameter of wood resistance: pulling force, perpendicular to the wood fibers.
The aims of the structural analysis are therefore to define the thickness of wooden
panels, to understand the different behavior of different materials, to define constraint
reactions at foundations, to size them, and to calculate their metal connection plates.
3.2 Rules Definition for the Structural Calculations

and Loads Analyses
The structural simulations imposed the snow and wind loads, in order to understand
the distribution of stresses and check them over three different kinds of wooden
elements: chestnut, larch, and oak wood. The calculations have been performed
according to the European Code UNI EN 1995-1-1:2014 and the Italian standards of
DM 14/01/2008 and UNI 11035, being the construction installed in Italy (Circolare
02/04/2011; CNR-DT 207/2008, 206/2007; D.M. 14/01/2008, 06/05/2008; UNI EN
1995-1-1:2014, 1995-1-2:2005; UNI 11035-1:2010, 11035-2:2010, 11035-3:2010).
1262 K. Kuma et al.
An accurate analysis relating to the definition of the snow loads has been carried
out, beyond the imposed procedures and values (D.M. 14/01/2008, 06/05/2008; UNI
1995-1-1:2014, 1995-1-2:2005). The snow load considered by the Italian code (D.M.
14/01/2008, 06/05/2008), indeed, refers to the total amount of snow that could settle
down, throughout the entire winter season, directly on the ground, considered as a flat
surface. However, for small surfaces, suspended in the air, the situation is different.
Two examples are the fences and the branches of trees: the snow settles on them,
but in smaller quantities if compared to the ones on the ground; the surface of the
snow accumulation presents relatively small dimensions, and since there is no lateral
confinement, due to the instability, and the snow falls to the ground in the short
term. The action of weather elements improves this phenomenon: wind fosters the
cleaning of the aerial elements, which can, therefore, receive just the load related
to individual snowfalls; the sun also radiates on a larger surface and causes a faster
melting. Starting from the local nivological data that are collected a few kilometers
from the project site (Folgaria Passo Sommo), the design snowfall has been defined
and characterized by the greater height of the single snowfall of the year and the
higher density detected in the same year. In order to improve the safety conditions
of the final results, a reference period of 35 years has been considered, since the
wooden sculpture would certainly have a shorter life if considering the durability of
the material with which it is realized. According to a proper engineering practice,
the snow load for ULS (Ultimate Limit State) has been then reduced from 4.80 kPa,
imposed by the Italian code (D.M. 14/01/ 2008, 06/05/ 2008) to 1.85 kPa for the
specific project.
Besides all these considerations about the snow loads, the wind action was also
defined, and calculated on the planar projection of the entire model on the YZ plane
(about 15 m2 ); it has been divided on the total Z panels area (25 m2 ) and applied on
oriented Z panels (0.45 kPa). Following a precautionary approach, the load analysis
that sets the initial conditions of the structural calculations does not consider that the
pavilion has several cavities through which the wind can go through.
Finally, the design situations have been defined, according to the Italian stan-
dards (Circolare 02/04/2011; CNR-DT 207/2008, 206/2007; D.M. 14/01/2008,
06/05/2008; UNI EN 1995-1-1:2014, 1995-1-2:2005; UNI 11035-1:2010,
11035-2:2010, 11035-3:2010), and their combination factors, related to snow and
wind loads, have been properly set, along with the different directions (+x −x +y −y).
3.3 The Structural Analysis Process: FEM-Based

Simulations, Results, and Considerations
The first step was to investigate the structure as a single glued solid (Fig. 6), in
order to understand the worst situation. The analyzed geometry was characterized
by panels with a thickness of 50 mm, in larch wood, which has been considered as
an isotropic material. The worst condition was determined as 1.3x Dead load + 1.5x
Fig. 6 Mesh of first single solid glued (Credit D3WOOD)
Snow + 0.9x Wind (−x), thanks to the structural simulation by a FEM-based (Finite
Element Method) software.
This starting and simplified analysis, however, could not be considered as a very
accurate procedure. Therefore, the second step of the process was to define, at first, the
single “type-node”, as the connection of six panels of larch wood, with 50 × 300 mm
section and 500 mm’s length. Starting from this set model, the FEM-based analysis
tested two different possible scenarios of the “type-node”: a glued connection or with
contact surfaces, as in the real case (Fig. 7a). In order to set the boundary conditions
of the simulation environment, one panel has been fixed at the basement, while a
force of 500 N has been applied on the top of another panel. The results from the
first scenario showed that the glued node had a rigid behavior, and the stresses were
more contained. In the second case, since there were no tractive areas, the stresses
were greater and the deformation was much more marked (Table 1).
Another interesting consideration concerned the variation of the final results,
according to the parallel variation of the meshes’ dimensions in the 3D model,
enhancing this feature as an important parameter for obtaining the correct results. At
least 2 mesh per thickness was set as the minimum value (Table 2 and Fig. 7b).
Starting from these considerations and the obtained results, the next simulations
could be more representative of the reality: larch wood was considered as a 3D
Table 1 Comparison between the glued node and the node with contact surfaces
Glued Contact surfaces Difference
(%)
Displacement 0.138 mm 2.939 mm 2129.71
Stress 1.042 Mpa 2.795 Mpa 268.23
1264 K. Kuma et al.
Fig. 7 The FEM-based analysis, performed by D3WOOD, has been set at first on the “type-node”,
whose behavior has been set into two different scenarios. The upper part of this picture (a) shows
different displacements of the glued node on the left, and the node with contact surfaces on the
right, where is indicated the opening and rotation of the piece. The analysis enhances, moreover,
a dependency of the results (b) on the size of the meshes that describe the 3D model (Credit
D3WOOD)
Table 2 Variation of the results with the variation of the mesh size
Mesh dimension fx min fx max fy min fy max fz min fz max Displacement
(mm)
15 2.26 2.68 0.86 0.96 0.87 0.62 1.57
25 1.9 2.36 0.38 0.4 0.27 0.33 1.44
30 1.94 2.18 0.44 0.35 0.26 0.3 1.35
40 1.66 2 0.23 0.22 0.13 0.23 1.06
45 1.41 1.63 0.11 0.09 0.05 0.06 0.69
Medium value 1.834 2.17 0.404 0.404 0.316 0.308 1.222
Fig. 8 First FEM-based analysis of the entire model and global stress values (MPa) (Credit
D3WOOD)
orthotropic material and the stresses’ transmission occurred by surface contacts. The
results showed that the values of tension perpendicular to the grain were too elevated
(Fig. 8). Moreover, another interesting aspect was the relevance of the entrance
configuration, in relation to the flexional behavior of the entire structure, from which
derived high-stress values. In order to change this condition, the possible alternative
methods would be to increase the thickness of the panels or to change the model
geometry.
Besides, the definition of the wooden pieces’ dimensions considered also the
availability of market standardized log-slices and the processes to obtain the final
wooden element from the raw material. Starting with 65 mm commercial thickness
of slabs, the treatment with thickening planer, made possible to have 58 mm panels.
This choice involved a 15% increase on the mechanical resistance of the material,
comparing to the initial of 50 mm thickness.
The analysis of the first two different geometrical models showed that the position
of the ground restrictions was also decisive. Therefore, the structure would need more
supporting points at the ground level, especially in the most external parts. This
proposal would reduce the flexion effect of both the wind action the eccentricity of
gravity, and the snow load of the exterior panels, resulting in a 20% reduction of
stress values (Table 3a, b).
Furthermore, the calculations showed that also chestnut wood would be a possible
alternative to the selected larch wood, as a construction material because of its optimal
ratio between weight and resistant to traction perpendicular to the grain.
At last, definitive analyses have been carried out according to all the load com-
binations provided by the legislation in order to verify a certain homogeneity of
stress levels. These analyses allowed to determine also the reaction forces useful for
1266 K. Kuma et al.
Table 3 Comparison of the stress level after the addition of new constraint to the ground: model
with 58 mm thickness
(a) Model 58 mm
Larch wood; thickness 58 mm
min limit max limit
fx −9.55 −12.0 5.09 8.4
fy x panels −0.42 −2.16 0.28 0.24 1 panel over
stress limit
fz x panels −0.83 −2.16 0.56 0.24 5 panels over
stress limit
fy y panels −0.47 −2.16 0.40 0.24 3 panels over
stress limit
fz y panels −0.96 −2.16 0.17 0.24
fy z panels −0.77 −2.16 0.14 0.24
fz z panels −1.69 −2.16 0.47 0.24 1 panel over
stress limit
xy −0.6 −2.28 0.63 2.28
yz −0.52 −2.28 1.07 2.28
zx −1.48 −2.28 1.01 2.28
Disp. (mm) 6.02
Total number of panels over stress limit 10
(b) Model 58 mm
Larch wood; thickness 58 mm; additional restraints (after meeting in Tokyo)
min limit max limit
fx −4.92 −12.0 4.39 8.4
fy x panels −0.34 −2.16 0.24 0.24
fz x panels −0.70 −2.16 0.49 0.24 5 panels over
stress limit
fy y panels −0.37 −2.16 0.34 0.24 2 panels over
stress limit
fz y panels −0.72 −2.16 0.14 0.24
fy z panels −0.75 −2.16 0.13 0.24
fz z panels −1.56 −2.16 0.41 0.24 1 panel over
stress limit
xy −0.52 −2.28 0.5 2.28
yz −0.49 −2.28 0.9 2.28
zx −1.29 −2.28 0.75 2.28
Disp. (mm) 5.06
Total number of panels over stress limit 8
Fig. 9 Technical detail of the foundation (Credit D3WOOD)
dimensioning the anchor plates and the connections with the larch wood foundation
beams (Fig. 9). Hence, the anchoring plates have been constrained with three types
of threaded bars (diameters of 10, 12, 14 mm, with cl 8.8, depending on the level
of load) and strengthened to the wooden foundations by resin. These last elements
have been foreseen above all in order to guarantee a precise regulation of the starting
quote of different panels during assembly (Tables 4 and 5).
3.4 KODAMA in Numbers: Algorithms for the Production

and Assemblage
KODAMA Pavilion is a part of a long design process based on computational phi-

losophy (Liotta and Belfiore 2012; Menges et al. 2016; Tedeschi 2014). This method
offers the possibility to use algorithms for managing logical constraints between
geometry in buildings, defining a new design field where simple objects, like small
pavilions or temporary structures, could be items of a deep experimentation (Vanossi
et al. 2014a; b; Vanossi and Imperadori 2013).
As known, the use of algorithms in design became popular among the design
firms with the diffusion of the visual programming tools, that regroup in a blocks
part of simple algorithm definitions. The level of detail in the definition and the
language used for the scripting are the variables that make the difference between
parametric design, often based on managing parameters derived by a simple and
already-prepared algorithm, and computational design, based on a complete defi-
nition of an algorithm (Vanossi 2018). The advantage of this approach is that the
design object and its visual form, the three-dimensional model, are described in
1268
Table 4 Stress values and maximum displacements of different situations

Stress (MPa)
ULS
1.5 + 0.9 (−X) 1.5 + 0.9 (+X) 1.5 + 0.9 (+Y) 1.5 + 0.9 (−Y) 1.5 (−X) (wind) 1.5 (+Y) (wind)
(snow + wind) (snow + wind) (snow + wind) (snow + wind)
fx −4.92 4.39 −6.17 3.33 −4.74 4.24 −6.36 3.6 −4.14 2.74 −3.44 2.73
fy −0.75 0.33 −0.93 0.6 −0.78 0.47 −0.68 0.47 −0.55 0.24 −0.47 0.21
fz −1.56 0.49 −1.37 0.55 −1.69 0.42 −1.23 0.49 −0.78 0.34 −0.97 0.26
xy −0.51 0.5 −0.53 0.57 −0.56 0.48 −0.49 0.45 −0.3 0.34 −0.29 0.32
yz −0.49 0.94 −0.45 0.62 −0.56 0.78 −0.37 0.79 −0.25 0.6 −0.34 0.34
zx −1.29 0.76 −0.91 0.9 −1.12 0.79 −1.08 0.65 −0.8 0.51 −0.48 0.54
Displ. 5.06 mm 4.8 mm 5.65 mm 4.72 mm 5.17 mm 4.83 mm
K. Kuma et al.
Table 5 Envelope of reaction Envelope reaction forces (kN)

forces
TX TY TZ
1-1YZ 0.05 −0.08 0.06
1-2YZ 0.78 −0.6 2.61
1-4ZX 0.91 −0.32 7.1
1-5ZX 1.63 −0.22 4.49
1-6YZ 0.17 1.27 3.47
1-8YZ 1.04 −0.35 1.63
1-9ZX 0.14 0.34 4.77
1-10YZ 0.87 1.41 5.02
1-11YZ 0.78 −0.42 4.45
2-1YZ 0.7 −1.66 8.22
2-9YZ 0.44 −0.85 0.46
2-15YZ 0.68 −0.96 1.8
2-20YZ 0 −0.22 0.04
3-1YZ 0.42 −2.11 11.5
3-10YZ 0.79 0.5 1.52
3-16YZ 0.95 0.75 7.85
3-18YZ 0.26 0.9 1.66
3-24YZ 0.33 −0.35 −1.52
3-38YZ 1.4 −1.1 5.21
3-40YZ 0.53 −0.97 4.13
3-45YZ 1.23 −0.71 2.35
4-23YZ 0.42 0.26 0.79
4-29YZ −0.66 −0.61 7.93
4-41YZ 1 −1.7 6.51
4-43YZ 0.31 −0.97 12.35
a clear dynamic process, which overcomes the static visualization of a simple 3D

virtual element and describes the geometries of components and their connections
through specific rules. This way to define the design process allows a high level
of interoperability since the model can be used for several scopes, from responsive
design to production.
The project of KODAMA Pavilion has been developed, in its parametrical stage,
with the well-known visual programming tool Grasshopper for Rhinoceros. This tool
has been used as an integrated element of the design process, from the first stages, in
order to manage the high complexity of the structure, in its geometry and components.
Indeed, even if the wood joints are usually simple in their one-to-one connection,
they could become very complex when assembled in a three-dimensional structure.
Therefore, the relationship between components, from the simple connection to the
entire assemblage in a bigger structure, has been managed thanks to the creation of an
algorithm that defined the logic rules for the propagation of the joint into a spherical
1270 K. Kuma et al.
Fig. 10 KODAMA in number by different requirements (Credit Politecnico di Milano)
surface. The pavilion is composed, indeed, by two main dimensions in between the
algorithm: the simple and detailed connection, the so-called “type-node” (Fig. 4),
and the general shape of the pavilion spherical volume (Figs. 3 and 5).
The final geometry of the pavilion collects 335 components, each of whom has
been identified for different phases (Fig. 10). At first, the visual programming tool
has combined the geometries of all the wooden elements into the general spherical
volume, and every single wooden piece has been identified with a specific codifi-
cation. Thereof, the computational team of the course of Design Optioneering at
Politecnico di Milano wrote a newer optimization algorithm, in order to:
• analyze the properties of every component: the volume, the dimensions in X-Y-Z
directions, the number of knots, and the position of the connections;
• aggregate the components, by one—or more—of those properties.
In the second phase, the different types of elements have been discretized by
dimension, resulting in the aggregation of the 335 components into 154 components,
performed by the new algorithm. This number has been set according to the level
of discretization defined for the CNC-based production process and the final con-
struction phase. In particular, considering the fabrication issues, the optimization
algorithm has been defined with the aim of saving time, during the production phase.
Here, the process used a CNC-based machine, whose low cutting tolerances usually
assure the reduction of the number of variables in the cutting process itself and vari-
ances between the same dimensions of different processed components. Therefor, the
number of similar components changed, according to the tolerance: 154 represented
the balance (Fig. 11a). In the third phase, the components have been aggregated
by their position, according to the breakdown structure of the assemblage, whose
stages have followed a precise spiral sequence: from the bottom to the top, in ver-
tical assemblage, and following a circular aggregation in the horizontal way. Every
component has been identified with a proper unique name (ID), in order to have a
clear indication during the construction (Fig. 11b). In the fourth phase, the algorithm
considers a subdivision of the components aiming to facilitate their transportation in
homogenous pallets, subdivided by construction phases (Fig. 11c).
In conclusion, as described in the process above, the advantages of the algo-
rithmic design are multiples. Starting from the design phase, the information that
creates the geometries can be used and manipulated in several ways. Primarily, the
project can be manipulated easily in order to evaluate different options in the final
result, avoiding to waste time in recreating the visualization model. Moreover, the
computational definition can be used to manage the parameters of the object for the
fabrication needs, e.g. changing the thickness of all the components while preserving
their connections and their cutting profiles. By the other hand, as in all BIM process,
the curve of effort in computational design is inverse, compared to the traditional
design process (Vanossi et al. 2012, 2013; Vanossi and Imperadori 2013). The com-
putational approach requires, indeed, an initial effort for the definition of the logic
process that drives the geometries and their links. But afterward, all the information
input in the design can be used for managing all the following phases. In KODAMA
this advantage has been highlighted, using the database of the 3D model as the base
to create new algorithms for production and organization of the assemblage phase.
4 The Power of Experimentation: Physical Models

and Analogic Mockups Towards the Final Construction
Phase
Besides the mathematical calculations, the FEM-based analysis, and the parametrical
design, the real prototyping acquires a fundamental role in the construction process,
especially for such a complex and unique structure. Real mock-ups followed the
engineering analysis, in order to check the visual proportions, the reliability of con-
nections and the feasibility of the construction operations. Moreover, this phase gave
samples on which physical tests could confirm the model’s assumptions and prove
the FEM results.
Thus, an intense production of wooden mock-ups started, crossing the design
project together with the other analysis. The first one assessed the three-dimensional
geometry of the “type-node”, being a 1:3 scaled sample, made in larch wood (Fig. 12),
by D3Wood. Other small scaled mock-ups have been realized, testing the details
and connections. The process has been iterative and optimized, to refine and then
check the digital model, through detailed samples of both the single joint and the
whole spherical geometry. The optimization criteria involved, e.g., the thickness of
the single elements, starting from the commercial profile dimensions (65 mm), and
integrated them with the needs from the structural analysis and the construction fea-
sibility perspective, according to different wood essences. Between Italy and Japan,
1272 K. Kuma et al.
Fig. 11 The optimization algorithm categorizes the components of KODAMA Pavilion by different
associated properties: a by single name and b by assemblage requirements, c according to the
construction sequences (Credit Politecnico di Milano)
Fig. 12 The first mockup of the “type-node”: 1:3 scaled, made in larch by D3WOOD, July 2017
(Ph. credit D3WOOD and Politecnico di Milano)
those physical models were made in methacrylate, poliplat, wood-veneer or massive

oak wood, a multitude of samples and materials that physically testify the intensity
of the entire design process (Fig. 13).
4.1 From Mockups to 1:1 Prototypes: The Assessment

of the Structural Assumptions
The first sample in real scale has been realized by D3WOOD: impressive cedar
log Sections (3500 × 500 × 80 mm) have been cut, fixed and smoothed in order to
prepare the wooden boards; after that, the trimming and carving steps completed the
production process, till the assembling phase. The result has been a majestic piece
representing the single connection among the six wooden blades, impressive in its
real dimensions and physical power (Fig. 14).
After the definition of the set dimension by the parallel studies with the FEM-
based analyses, Ri-Legno, an advanced Italian company that works in the wooden
sector, started the first tests on the CNC-based processes of cutting. Once again,
the parametrical design integrated the process: the virtual model has been adapted
and used in the CNC machine to produce all the pieces for a 1:1 scaled sample of
the “type-node”, made in fir wood. This phase of the process has been essential in
order to understand the tolerances needed to guarantee the buildability of the wooden
structure, without any further steel connections or reinforcement (Fig. 15).
1274 K. Kuma et al.
Fig. 13 Numerous physical models, at different scales, have been realized during the design phase,
in order to study the geometry of the global volume and the details of connections; In particular, a 1:5
scaled yew wood mock-up has been realized as a high-quality sample of a wooden toy, remembering
the traditional origins of the wooden construction techniques and impressive abilities of Japanese
craftsmen, since ancient times, until nowadays (Ph. credit Kengo Kuma Lab and Politecnico di
Milano)
Fig. 14 Production process of the 1:1 scaled mock-up of the cedar wood “type-node”, performed
with SCM machines: Surfacer-thickneeser with spindle 520 mm, Saw-spindle moulder minimax st
5es, and CNC machining center Morbidelli Author 430S, by SCM Group. In the upper images, the
long side trimming and carving steps; below, the assembled 1:1 scaled prototype of the cedar wood
“type-node” (Ph. credit D3WOOD and Politecnico di Milano)
1276 K. Kuma et al.
Fig. 15 The CNC-based cutting process has been tested, in order to assess the tolerances needed
to guarantee the buildability of the wooden structure, without any further steel connections or
reinforcement. In this phase, the parametrical model resulted very useful for the optimization of data
and information fluxes, between different places and towards the CNC machines (HUNDEGGER
K2 5 Axis). In the right corner, the fir wooden 1:1 mock-up of the 4 type-nodes, assembled in
RiLegno industries (Ph. credit Politecnico di Milano and RiLegno)
As the design process moved forward, towards the definition and assessment of
every feature of the wooden structure, the experimentation field and the engineering
analysis integrated always more, on the same path. D3WOOD performed a structural
load test on a 1:1 larch wood prototype of the “type-node”, whose plates have been
charged according to the structural calculations and the FEM-based analyses. The
purposes of such tests were multiple:
• to check the correspondence between the values of strength and deformation

returned by the software and the real behavior of the structure, in order to val-
idate the model and FEM-based analyses;
• to understand the safety margin between the real project stresses and the possible
breaking point of the single node;
• in case of a crash point close to the design stress values, to check the local rein-
forcement with the use of full thread screws, in order to increase tensile strength
perpendicular to the grain;
• lastly, given the high degree of hyperstaticity of the node, to understand after the
breaking point if the node was able to guarantee a redistribution of strength.
The load tests have been set up following two different static schemes (Fig. 16).
The first one used a restraint opposite to the point of the pulling force, while the
second one used only a fixed support to the ground. The first test represented the
best simulation of the real conditions, since the second configuration (with a second
restraint to the ground) defined a too eccentric status of forces.
In conclusion, the results of these tests attested several positive aspects:
• the good correlation of strength and deformation between the FEM-based analyses
and the reality: initially, the displacement of the plates in the real model was higher
than the calculated one, due to the recovery of the tolerance of 3/10 mm of the
joints; closer to the breaking point, however, the real displacement was lower than
the calculated one;
• before arriving at the break-point, with 7 kN of pulling, the knot had remarkable
reserves of plasticization. The software at 6 kN of pulling calculates values of
stress ten times higher than those we have in the global model with real forces
at ULS. In case of centered force at the first break point with more than 7 kN of
pulling force, the hyperstaticity reserves influence the result and the dynamometer
had revealed still 4 kN of pulling (Fig. 17).
After having analyzed the maximum stress values, verified the margin to arrive
at a possible crash point and attested the considerable reservations of hyperstaticity,
the experimental results attested and overtook the general assumptions about the
structural behavior of the system, and allowed to avoid any local reinforcement
element. The pavilion would be completely faithful to its initial concept and devoid
of any metallic element except for the ground connection.
4.2 The Last Challenge: Assembly
The Pavilion has been designed as a completely reversible and environmentally

friendly system.
The foundations consist of five beams coupled in larch with a square section
of 200 × 200 mm, wrapped in a breathable fabric and parallel disposed of. Double
threaded screws, transversely inclined by 45° and arranged in pairs, were used to
make them integral. A 1:1 scale drilling template was produced by CNC machine for
1278 K. Kuma et al.
Fig. 16 Configuration of load testing on real 1:1 prototype of the single “type-node”, performed
by D3WOOD (Credit D3WOOD)
drilling on site the beams and inserting the threaded bars for fixing the connection
plates. By using a laser level, the quote of the metallic plates has been adjusted
to have the maximum precision (millimeter scale) for the starting position of the
wooden elements.
The realization of the larch structure has been based on a strict assembling scheme
that stems from the study of the parametric model. KODAMA is composed, indeed,
by 335 panels characterized by bidirectional joints; this complex geometry requires
a construction organized by phases. Through “a rings scheme” the realization of the
entire pavilion was allowed. This procedure provided a separate assemblage of each
ring that is jointed to the previous one in the vertical direction. The 335 larch pieces
cut by the CNC machine have been stocked into five different pallets, according
to the parametric model codes (Fig. 11c). In order to assess the feasibility of those
procedures, moreover, a last mockup in oak wood, 1:5 scaled, was produced and
assembled by D3WOOD.
Fig. 17 Results of load testing on real 1:1 prototype of the single “type-node”, performed by
D3WOOD (Credit D3WOOD)
The construction processes of KODAMA lasted five days, from April the 16th,
to April, the 20th, 2018. The first one has been used to place the foundation’s beams
made in larch wood, and to fix the metal plates. The wooden structure, completely
jointed without any metal connections, started on the 17th of April 2018 and contin-
ued for four days. D3WOOD craftsmen mounted the different rings on the ground, in
a lying position and once raised to the ground using three hoists have been oriented
so that the joints can be combined between the different levels (Fig. 18). The most
complicated situations involved 23 junctions on rings of about 5 m in diameter with
a mounting tolerance of 1 mm.
1280 K. Kuma et al.
Fig. 18 Realization process (Ph. credit Politecnico di Milano and D3WOOD)

After the assembly of each ring, the dimensions were controlled, provisional
supports were positioned and the perfect interpenetration of all the joints was verified.
By adopting all these measures to respect the geometry of the design, the assembly
was engaging and no changes were made to any element during construction (Fig. 19).
As the end of the path, a mockup of the sculpture has been exposed also as a project
in Tokyo, from March 2018, until May 2018. The exposition took place at Tokyo
Central Station, as a personal exhibition dedicated to Kengo Kuma and all his best
works: Kengo Kuma: a LAB for materials (Fig. 20) (http://kuma-lab.arch.t.u-tokyo.
ac.jp/). It has shown the potential of wood and how geometry and engineering can
optimize a single, simple element and make with it a complex concept and sculpture
(Kuma 2018).
4.3 The Role of Humidity: The Power of Nature
Wood is a living material. According to this, humidity represents an important factor

to focus on, especially considering the construction processes of wooden structures.
In the specific case of KODAMA Pavilion, it assumed a significant role, because of the
different conditions between the initial stage, when the nodes still have construction
tolerances, and the final one, when the nodes could be considered completely fixed.
The compensation of the tolerances was, indeed, a fundamental aspect in order to
guarantee the correspondence between the calculation model and the real mechanical
behavior, which is characterized by the transmission of the strength between the
elements through the contact surfaces. Larger tolerances would have determined not
only the kinematic phenomena, not provided for by the structural analyzes, but also
possible stagnations of water, with the acceleration of rot and decay of the material,
at the joints’ area, to the detriment of the durability of the pavilion.
In order to predict the final enlargement due to the increase of the humidity, the
starting content of moisture needed to be established. Therefore, a drying treatment
let the wood elements to achieve the 15% of humidity (EMC), which represents
the standard condition of a naturally dried wood, protected from rain. It is known,
indeed, those wooden elements, which are directly exposed to the weather, reach an
average moisture content of 24% EMC, meaning, therefore, a predicted increase of
9% as compared to the defined initial condition. Hence, taking into consideration
the radial deformation that is typical of the geometry of the wooden elements, this
increase of 9% of humidity means an expansion of the single panel, with the 58 mm’
initial thickness, by about 1.2 mm. These considerations led to define the width of
the milling process wider than 1 mm, as compared to the thickness of the panels, a
value that allowed to make the assembly of the elements easier.
1282 K. Kuma et al.
Fig. 19 In the bigger image, KODAMA Pavilion, in its wonderful location of Arte Sella—The
Contemporary Mountain (Ph. Giacomo Bianchi for Arte Sella); besides, some detailed views of the
wooden pavilion in the forest and its amazing scale, realted to the human body one (Ph. Toshiki
Hirano for Kengo Kuma Lab); below, on the left, the open ceremony, when Kengo Kuma, Emanuele
Montibeller, Marco Imperadori and Marco Clozza described the sculpture; on the right, the real
structure compared with the numerous mockups (Ph. credit Federica Iachelini and Clara Rinaldi,
Politecnico di Milano)
Fig. 20 Kengo Kuma presented the work at his personal exhibition: Kengo Kuma: a LAB for
materials, in Tokyo Central Station, Tokyo, Japan (Ph. credit Kengo Kuma Lab)
After two weeks from the end of the assembly, during the opening day of
KODAMA Pavilion, the checking procedures attested how just three days of rain had
been enough to block several pieces. Some elements seemed already glued to the
others, with perfectly-fitting contact surfaces and the assembly tolerances already
compensated for the volumetric expansion of the material, due to the increase of
moisture content of wood.
5 Conclusions
At the end of this intense path of design, engineering, and prototyping, KODAMA
proves to be a very poetic space (Fig. 19). Kengo Kuma, author of the new Olympic
Stadium for Tokyo 2020 (Frampton 2013; Kuma 2018), has conceived a landmark
into a landscape, where Nature accepts massive wood to “come back”, and to decay
and age with time and seasonal changes. This idyllic image is typical of the Japanese
philosophy, indeed: beauty achieved with ephemerality, decay, and imperfections
given by the time as Wabi Sabi beauty.
Space itself remembers a Japanese tea-house in the open air, approaching it
through big stepping stones from local mountains. The ceremony of tea and its
aesthetics described by of Okakura Tenshin (Tenshin 2015) find a full expression
within the project of KODAMA, as well as the importance of light and shadows, cre-
1284 K. Kuma et al.
ated by the structure which brings us also to the core of the philosophy by Junichiro
Tanizaki (Tanizaki 1977).
The real only way to feel all of this experience, as Bruno Zevi may have said, can
only be achieved by going around and inside KODAMA in Arte Sella, becoming a
part of it and the forest. Nature and human artifact join, in a very harmonious and
poetical way.
Acknowledgements This work is the result of a deep and intense collaboration between different
realities, which could be distant in a geographically perspective, but very close in their intellectual
nature and intentions. They involve both the academic field, the industry sector and the bigger design
realities, through the conception, management, production and realization of an art piece that has
already encountered a great success among people.
The authors, therefore, would like to acknowledge at first Arte Sella—The Contemporary Mountain,
for the opportunity to be hosted in such a beautiful place, where Art, artifacts, and Nature are
embraced in the same substance. We are very proud to be a part of this story.
Then, a special thank goes to the academic realities that started this international collaboration
between Politecnico di Milano and the University of Tokyo. At first, we would like to thank all the
members of Kengo Kuma Lab, for the primary conceptual ideas and the continuous rehashing of
materials. Prof. Jun Sato of the University of Tokyo worked on structural calculations in Japan, with
D3WOOD, which also cared about the simulations of the mechanical behavior from Italy. The com-
putational team of the course of Design Optioneering of Politecnico di Milano, Matteo Pedrana,
Leandro Robutti and Fabrizio Miele, helped in defining the algorithm code for the parametrical
design stage. Afterward, the prototyping phase would not be possible without the collaboration
between D3WOOD (Marco and Claudio Clozza) and Ri-Legno Srl, where Giulio Franceschini,
Lavinia Sartori and Giorgio Franceschini have been in chief of the CNC works. Then, the construc-
tion phase has been carried out by D3WOOD, with the support of Rotho Blaas Srl for the material
supply.
Finally, we thank all the students involved in the process, for their hard work: Federica Iachelini
and Clara Rinaldi of Politecnico di Milano, and Takahiro Hirayama, Masumi Ogawa and Ifan Yim,
from the University of Tokyo.
References
Aicher S, Reinhardt H-W, Garrecht H (eds) (2014) Materials and joints in timber structures. recent
developments of technology. In: RILEM book series, vol 9. https://doi.org/10.1007/978-94-007-
7811-5
Attali E et al (2016) Glass/wood: Erieta Attali on Kengo Kuma. Hatje Cantz Verlag Gmbh & Co
Kg, Berlin
Circolare n. 617 del 02/04/2011 del Ministero degli Interni - Istruzioni per l’applicazione delle
“Nuove norme tecniche per le costruzioni” di cui al D.M. 14 gennaio 2008
CNR-DT 207/2008 “Istruzioni per la valutazione delle azioni e degli effetti del vento sulle
costruzioni”
CNR-DT 206/2007 “Istruzioni per la Progettazione, l’Esecuzione ed il Controllo delle Strutture di
Legno”
D.M. 14/01/2008 Ministero delle Infrastrutture - Approvazione delle nuove norme tecniche per le
costruzioni (Gazzetta ufficiale 04/02/2008 n. 29)
D.M. 06/05/2008 Ministero delle Infrastrutture - Integrazione al decreto 14 gennaio 2008 di
approvazione delle nuove “Norme tecniche per le costruzioni” (Gazzetta ufficiale 02/07/2008
n. 153)
Ejiri N (2018) Norihiro Ejiri × Jun Sato. Traditional craftsmanship supporters and traditional wood
construction supporters | The advantages of engineering wood | New joints and traditional tech-
nique | Communication with craftsmen | Orthogonal interdigitating | Examination method of
complicated joints | Fire proof measures and the highness of specific strength. GAJ 152, Session
4: Wood Structure, Tokyo
Frampton K (2013) Kengo Kuma: complete works. Thames & Hudson, Reprint edition, London
Imperadori M, Senatore A (2011) Schematic design: tecniche ed esempi di comunicazione del
progetto, Il Sole24Ore, Milano
Kuma K (2018) Ja 109 Kengo Kuma: a lab for materials. Shinkenchiku-sha Co Ltd., Tokyo
Kuma K, Daniell T (2015) Small architecture/ natural architecture. Architectural Association Pub-
lications, London
Kuma K, Ferrari M (2016) Kengo Kuma. L’anti oggetto. Dissolvere e disintegrare
l’architettura. Parole di architettura, vol 1, Ilios. Retrived from https://books.google.it/books?
id=5Ux9jwEACAAJ
Liotta SJ, Belfiore M (eds) (2012) Patterns and layering. Japanese spatial culture, nature and archi-
tecture. Architecture & Interior, Gestalten, Berlin
Routledge, United Kingdom
Montibeller E, Tomaselli L, Bianchi G (2017) Arte Sella. The contemporary mountain. The new
beginning. Silvana Ed., Milano
Seike K (1978) The art of Japanese joinery. Weatherhill ed., Boston (Printed in China)
Tanizaki J (1977) In praise of shadows. Leete’s Island Books, Stony Creek
Tedeschi A (2014) AAD algorithms-aided design. Parametric strategies using grasshopper. Edizioni
Le Penseur, Brienza
Tenshin O (2015) The book of tea. CreateSpace Independent Publishing Platform
UNI EN 1995-1-1:2014 Eurocode 5—Design of timber structures. Part 1-1: General—Common
rules and rules for buildings
UNI EN 1995-1-2:2005 Eurocode 5—Design of timber structures. Part 1-2: General—Structural
fire design
UNI 11035-1:2010 Structural timber—Visual strength grading for structural timbers. Part 1: Ter-
minology and measurements of features
UNI 11035-2:2010 Structural timber—Visual strength grading for structural timbers. Part 2: Visual
strength grading rules and characteristics values for structural timber population
UNI 11035-3:2010 Structural timber—Visual strength grading. Part 3: Timber Uso Fiume and Uso
Trieste
Vanossi A (2018) Interaction and intersection between digital modelling and design in architecture:
different approaches in parametric design. In: Handbook of research on form and morphogenesis
in modern architectural contexts. Information Science Reference, Milano, pp 152–174, Feb 2018
Vanossi A, Parello G, Imperadori M, Bernardo C, Liotta SJA, Ito Y (2014a) BIM for archaeology,
use of BIM process and parametric model in a temporary shelter adaptable to the diverse needs of
archaeological sites. In: PPC 2014 conference proceedings. Monza, Mantova (Italy), pp 191–200,
May 2014
Vanossi A, Parello G, Imperadori M, Bernardo C, Liotta SJA, Ito Y, Occhipinti F (2014b) Architec-
ture for archeology: identifying new modular and flexible types of shelter adaptable to the diverse
needs of archaeological sites. In: XII Forum Internazionale Le Vie dei Mercanti –proceedings.
La scuola di Pitagora, Anversa, Capri, pp 457–466, June 2015
Vanossi A, Ciribini A, Tegon A (2013) From BIM requirements to facility management. In: 5D-
conference 2013—proceedings, pp 247–256
Vanossi A, Imperadori M (2013) BIM and optioneering in dry technology small-scale building. In:
ICT, automation and the industry of the built environment. Maggioli, Milano, pp 53–65, March
2013
1286 K. Kuma et al.
Vanossi A, Veliz A, Ciribini A, Balbo R (2012) Paths between real and virtual within a BIM-based
scheme. In: First UK academic conference on BIM—proceedings. Newcastle, pp 254–265, Sept
2012
http://www.artesella.it/en/
http://kuma-lab.arch.t.u-tokyo.ac.jp/
http://www.D3Wood.it/
http://www.ri-legno.it/
Wooden Byobu. From Architectural
Façade to Sculpture
Atsushi Kitagawara, Marco Imperadori, Ryosuke Kuwabara,

Federica Brunone and Mayuko Matsukawa
Abstract The Japanese Pavilion at EXPO Milano 2015, designed by Atsushi Kita-
gawara, has been visited by millions of people with a great success, especially for
its stunning wooden façade. Its naked structure is defined by only one element, a
12 × 12 cm section, whose larch came from Fukushima region, to show the health
of the forest after the terrible Tsunami. The façade was conceived with studs, con-
nected without any screws or nails, thanks to the “compressive-tension” effect, whose
concept interprets old Japanese wooden-based techniques of constructions and even
handcrafts of complex wooden toys. Analogic mock-ups and parametric models have
defined the entire design phase, leading to a simple form as the end of a complex
research path. The wooden structure was, then, elaborated and produced through
CNC processes, built for EXPO Milano 2015, and dismantled after the event. A new
life of this concept is becoming an art object: the Byobu. This is a traditional separat-
ing screen in Japanese houses, something that creates space and multiple possibilities.
Thus, the structural concept of EXPO Milano 2015 turns into several variations for
different places: University Byobu at Politecnico di Milano, Urban Byobu at Farm
Cultural Park in Sicily, Forest Byobu at Arte Sella, near Trento, and finally Museum
Byobu—Kigumi Infinity at Mori Art Museum in Tokyo. All these examples leave a
memory of a stunning structure, conceived to be temporary for EXPO Milano 2015,
now visible in several places between Italy and Japan.
Keywords Wooden lattice · Parametric design · Compressive-tension · CNC

Byobu
A. Kitagawara
Fine Arts Department of Architecture, Tokyo Geidai University, Tokyo, Japan
M. Imperadori (B) · F. Brunone
R. Kuwabara · M. Matsukawa
Atsushi Kitagawara Architects, Tokyo, Japan

1288 A. Kitagawara et al.
Fig. 1 Japan Pavilion at Expo Milano 2015, Milan, Italy, a project by Atsushi Kitagawara and
Atsushi Kitagawara Architects, Ishimoto Architectural & Engineering Firm Inc., Takenaka Europe
GmbH AND Galloppini Legnami, Milan (Italy), 2015 (Ph. credit Marco Imperadori)
1 Introduction
The Japanese Pavilion (Fig. 1), designed by Atsushi Kitagawara + Atsushi Kita-
gawara Architects (www.kitagawara.co.jp), Ishimoto Architectural & Engineering
Firm, Inc. and Ove Arup & Partners Japan Ltd./Arup Italia S.r.l., for EXPO Milano
2015, has achieved a resounding success, both among experts and visitors (Imper-
adori 2015a). The BIE—Bureau International des Expositions awarded the project
for its interior exhibitions, while the long queues of people outside the entrance,
waiting until 10–12 h, recognized its widespread appeal (Fig. 2).
The main architectural attractive was undoubtedly the magic wooden structure of
the outer envelope, a studs’ pattern created by the multiple connections of just one
type of element: a 115 × 115 mm squared glulam section, extruded for a length of
2.10 m by CNC-based (Computer Numerical Control) processes (Doust 2015).
The lattice structure resulted in a porosity that continuously changes the geomet-
rical perception of the surface, in a dynamic/kinetic dialogue with the users, inside
and outside, depending on view angles (Figs. 3 and 4). It determinates the quality of
spaces and light, giving a sense of vibration that we may call magical because of the
contrast between the mass and weight of wood and the emptiness left by the lattice
Wooden Byobu. From Architectural Façade to Sculpture 1289
Fig. 2 The Japan Pavilion was one of the main attraction for people, visiting Expo Milano 2015,
thanks to its outstanding architecture and interior exhibition. Below, Atsushi Kitagawara, Japan
Pavilion designer, in front of the impressive wooden structure of the façades, Milan (Italy), 2015
(Ph. credit Shigeru Ohno)
structure. That emptiness is, indeed, a typical element of the Japanese philosophy.
The void and the mass are alternated, in a mutual conversation between aesthetics
and structural needs, architectural expression and innovative, engineered technology
(Fig. 5).
How to achieve this? The process is complex (Cipollini et al. 2016; Doust
2015), a continuous endeavor of form-shape and structure-tectonics that starts
from sketches, then is verified by small mock-ups, then is controlled and modified
Fig. 3 The stunning wooden façade of the Japan Pavilion, Milan (Italy), 2015 (Ph. credit Shigeru
Ohno)
Fig. 4 The stunning wooden façade of the Japan Pavilion: different light, different perceptions,
Milan (Italy), 2015 (Ph. credit Shigeru Ohno)
with parametrical design, in order to check geometries, proportions, and issues;

once the advanced model and some options are assessed, it comes the time of the
real, full-scale modelling, to fix the real impact of the final construction. Finally,
CNC-based cutting processes applied to the optimized and simplified geometries of
Fig. 5 The stunning wooden façade of the Japan Pavilion: the perspective under the structure,
Milan (Italy), 2015 (Ph. credit Shigeru Ohno)
the wooden elements allowed the controlled assembly, directly on site, by Galloppini
Legnami company (Fig. 6).
The experience of EXPO Milano 2015 shows, therefore, an innovative step for
the implementation of diffusive parametrical structures’ design process and a success
for this kind of construction system itself.
After the event, the structure has been dismantled (Bottelli et al. 2015), but its
concept has had an artistic chance to remain. Study-mock ups and new samples of
the façade became art objects, spreading the structure’s beauty along iconic places
between Italy and Japan. Thanks to the parametric phase of the first design pro-
cess, that cleverly translates ideas, thoughts, and sketches in virtual models, several
examples could be produced.
These revalued objects are now iconic art pieces, gathered under the master con-
cept of the Japanese Byobu, a traditional separating screen in old Japanese houses,
here conceived as something that is able to create multiple spaces and possibilities.
It is a structural prototype at Politecnico di Milano—Lecco Campus, an urban scenic
set in Favara, Sicily (https://www.farmculturalpark.com/), and a wooden sculpture
into the Alpine forests of Arte Sella Land Park (http://www.artesella.it/en/), near
Trento, in Northern Italy; then, it is a division wall enhancing the entrance for an
exhibition at Mori Art Museum (https://www.mori.art.museum/en/), in Tokyo.
Fig. 6 The construction phase of the wooden façade, by Galloppini Legnami, Milan (Italy), 2015
(Ph. credit Marco Imperadori)
2 Methodology
This contribution reports the account of a methodological approach to designing

with wood as construction material (Aicher et al. 2014), directly described through
its application on a real design experience—the project and realization of the Japan
Pavilion wooden façade at EXPO Milano 2015—whose complexity is the result of
a deep endeavor of integration between different techniques and technologies. From
sketches (Imperadori and Senatore 2011), to scaled mock-up and technical drawings,
the design experience passes through the architectonical and aesthetic definition
and technological and structural validation, till the real experimentation with 1:1
prototypes. It follows an iterative path that connects and overlaps conceptual ideas,
to construction hypothesis and technological solutions, to be tested and verified, both
in the virtual reality and for real.
The process has been aided, indeed, by the most innovative design methods, such
as the parametric design or the finite elements structural calculations (Menges et al.
2016), and practices, e.g. the CNC-based production processes, enhancing the infinite
potential of wood’s applications.
This specific design experience, finally, has proved both the feasibility of the
methodological approach in the realization of the Japan Pavilion at EXPO Milano
2015, and its reliability, thanks to the following numerous applications of the same
design process to smaller cases, spread up all over the world.
Wood is definitely a material with an infinitive potential, thanks to its versatility
and flexibility in being the object of the most innovative methods and, therefore,
declined into the most different design applications (Aicher et al. 2014).
3 The Japanese Pavilion for EXPO Milano 2015: The

Wooden Diffusive Façade
Harmonious Diversity was the theme of the Japan Pavilion, by Atsushi Kitagawara
and AKA—Atsushi Kitagawara Architects, for EXPO Milano 2015—Feeding the
Planet, Energy for Life. This concept seeks to represent the variety of wisdom and
skills that portrays Japanese culture, embraced by the architecture of the Japan Pavil-
ion as a “bowl of diversity”. The Pavilion adopted, therefore, a three-dimensional
wooden grid as the naked structure of the outer envelope, whose design has been
developed by Ishimoto Architectural & Engineering Firm, Inc., with the collabora-
tion of Ove Arup & Partners Japan Ltd./Arup Italia S.r.l. for the structural design, and
Atelier2 and Prof. Marco Imperadori from Politecnico di Milano, as local advisors
(Figs. 1, 2, 3, 4 and 5) (Bottelli et al. 2015; Doust 2015; Imperadori 2015a).
Wood, as a renewable resource, connects the architecture to the issues of sus-
tainable maintenance and protection of forests. Besides, wood, as a construction
material, embodies itself the notion of sustainability. The recent policies of Japanese
government about environment issues and the attention to the topic from both public
and private associations has enhanced that a sustainable forestry management has
to pass through a responsible exploitation of wood, and the AEC sector is one of
the eligible field of application. Among the Japanese culture, indeed, the traditional
concept of Satoyama (Japan Satoyama Satoumi Assessment 2010)—natural ancient
woodlands that coexist with nearby populated areas, in Japan—fosters the idea of
recycling-oriented contemporary societies. Moreover, the raw Japanese larch wood
came from Fukushima region, becoming then glue-laminated timber in order to show
the health of the forest after the terrible Tsunami, in 2011.
Wood, as a construction material, is deeply linked to Japanese architecture. The
wooden diffusive structure of the façades derived from the fusion of the ancient
and traditional Japanese construction crafts (Seike 1978) with the most advanced
technology, and the modern analysis and application techniques, which guided the
lattice development (Betsky et al. 2013; Fukuyama et al. 2008; Imperadori 2013,
2015b; Kamachi et al. 2008). The result was an innovative construction that revalued
tradition through advanced technology.
3.1 From the Architectural Concept to Parametrical

and Structural Design
The complex process that follows the realization of the impressive wooden structure
started from a simple gesture: a sketch by Atsushi Kitagawara (Fig. 7).
Fig. 7 Japan Pavilion at Expo Milano 2015: the perception of the wooden façade structure,
impressed on washi paper by Atsushi Kitagawara, Tokyo (Japan), 2014 (Credit Atsushi Kitagawara)
In its simplicity, however, it represented an innovative vision on the deeply rooted

knowledge about the wooden framework systems of Japanese architecture (Imper-
adori and Senatore 2011). From Japanese joinery to lattice structures, indeed, sev-
eral publications and studies had already investigated those ancient and traditional
constructions and crafts’ techniques (Seike 1978), in order to understand the inner
mechanisms of assembly and structural resistance (Betsky et al. 2013; Fukuyama
et al. 2008; Imperadori 2013, 2015b; Kamachi et al. 2008; Kobayashi et al. 2008;
Song et al. 2008). Modern and contemporary architectures, then, testify that their
application is a consolidated and well-spread practice, mostly in the Far East experi-
ences (Betsky et al. 2013; Imperadori 2015b). Here, the new architectonical concept
impressed with black ink on washi paper (Fig. 7) reveals more than the idea of the
“bowl of diversity”. It portrayed a spatial improvement of the lattice structure—al-
ready undergone by Atsushi Kitagawara Architects with the Mengoshi construction
system and its multiple applications (Betsky et al. 2013; Imperadori 2013, 2015b).
This little shift from two to three dimensions, actually, meant to intensify the com-
plexity of all the construction aspects: its design and conception, the definition of
geometries and volumes towards the final aesthetic effect, the structural implications,
the management of the construction elements, and the feasibility of its realization.
The development of the project followed, then, a continuous production of small
mock-ups and drawings, in order to deeply study the geometry, the proportions
between masses and voids, the porosity of the lattice structure, and the aesthetics of
emptiness, that domains the entire concept (Fig. 8). This first design phase allowed
having an initial approach to the construction, towards the definition of the basic ele-
ments, in their geometrical features, shape, and dimensions (Cipollini et al. 2016).
At the end of these physical experimentations, the wooden structure has been com-
posed as a lattice of members, a grid of glue-laminated timber sticks, having all the
same section and length, and whose joints were regularly repeated across the three
dimensions, through all over the construction (Fig. 9). Each timber member has been
shaped, indeed, to present a number of halving joints (aigaki), repeated along its
length with a modular rhythm. The assembly of the diffusive three-dimensional lat-
tice structure, thus, results in the definition of four different timber profile (Fig. 10a):
• WG1, WG4, WC1, shaped with simple halving joints (aigaki), and whose different
codifications stood for a different allocation inside the grid structure;
• WC2, with a mirrored-WC1 profile, a key piece to progressively lock the wooden
structure during the construction phase;
• WG3 and WG4, whose halving joints’ profiles are modified with a diagonal
milling, in order to allow the insertion of the WC2 profile during the construction
phase.
Those timber elements are perpendicularly interconnected (WG4 + WG3 and
WG1 + WG2), creating horizontal grid plans (Fig. 10a); the vertical profiles (WC1
and WC2), then, connect the horizontally overlapped plans (Fig. 10b, c), finally
resulting into a three-dimensional grid of diffuse nodes, with specific and harmonic
proportions between masses and voids.
Fig. 8 The design process focuses on rethinking the tradition of Japanese wooden constructions,
in order to explore geometries, spatiality, textures of lights and shadows, from the first physical
small-scale models (Ph. credit Atsushi Kitagawara and Atsushi Kitagawara Architects)
Beside this huge and intense work on physical models, those proportions have
been experimented also into a parametrical environment, thanks to virtual 3D models,
directly realized within Grasshopper and Rhino (Fig. 11). These platforms allowed
evaluating several different geometries of the structure’s basic element, changing
its length, the numbers of joints and the final aesthetics appearance, according to
the management of a single parameter: the dimension of the stick’s squared section.
Thus, the digitalization and parametrization of the process helped to individuate the
final dimensions of the elements, the consequent porosity of the lattice structure,
in its relation with the environment of the project, the EXPO Milano 2015 site
Fig. 9 The final structure was experimented through the continuous realization of models, in order
to deeg the feasibility of construction beside the exploration of geometries, spatiality, textures of
lights and shadows (Ph. credit Atsushi Kitagawara and Atsushi Kitagawara Architects)
(Fig. 12). Moreover, the parametric model has been useful also for the validation of
the mechanical behavior, optimizing the architectural requirements with the structural
needs, and for the management of the CNC-based production and construction stages
(Menges et al. 2016).
This intense endeavor led towards the definition of the wooden structure, as either
a three-dimensional lattice of long elements, or a diffusive grid of structural joints,
or a potentially infinite repetition of “cubic units”, as the basic geometry that arose
from the intersections of the long members into nodes.
The “cubic unit”, indeed, has been the basement of the first structural investiga-
tions about the mechanical behavior of the grid, when shifted into the third dimen-
sion. This analysis has been conducted by Ove Arup & Partners Japan Ltd./Arup
Italia S.r.l.—Professor M. Kanada, as part of the Pavilion design team. Then, several
dynamic simulations, based on FEM (Finite Element Method) (Fig. 13), assessed the
mechanical performances of the entire grid structure, in its diverse configurations,
defined within the parametric environment and through the first stage of physical,
scaled mock-ups. In particular, the analyses aimed to verify the “compressive stress”
status of the material within the joint area, a specific condition which is inherited from
the traditional wooden construction techniques of Japan—epitomized in the Horyuji
Temple and theorized by Professor M. Inayama in his so-called Merikomi method
Fig. 10 a The wooden lattice is conceived as the repetition of a single stud—section 115 × 115 mm,
length 2.10 m. Different position inside the structure generate just four cutting profile, CNC pro-
cessed, simplifying the production and construction phases, beside the improvement of sustainability
of the entire construction. b The studs are joint, setting up different horizontal planes and verti-
cal axes. c The studs are fixed without screws or nails, where vertical element locks the structure
according the “compressive tension” theory, conceived by M. Inayama. (Credit Atsushi Kitagawara
and Atsushi Kitagawara Architects
and “compressive tension” theory (Imperadori 2013, 2015b; Song et al. 2008). It
is based on the anisotropic behavior of wood, especially on the low levels of its
compressive strength, when the material is loaded along the perpendicular direction
to its fiber line, and the high deformations, finally causing a dissipation and redis-
tribution of stresses across the statistically indeterminate grid structure. Therefore,
the FEM-based analyses proved the robustness and ductility of the structure, under
its own weight of 0.3 kN/m2 and a horizontal wind force of 0.5 kN/m2 , enhancing
how the 35° rotation of the horizontal plans improved this local effect. Moreover,
the diffusion of the wooden joints through the lattice structure causes a redundancy
of the structural constraints, resulting in a homogeneous stress status, without any
local tension. The so-defined structural scheme, a grid of punctual joints, indeed,
allows to maintain constant the overall thickness of the timber elements, combin-
ing both stiffness and elasticity; furthermore, it assures the resilience of the entire
construction system itself, thanks to its global structural behavior that follows the
mechanical principle of the so-called “resistance hierarchy”. At last, in order to prove
Fig. 11 Next to the mock-ups, the virtual and parametric modelling defined the complex design
process of the wooden lattice structure of the Japan Pavilion façade (Credit Atsushi Kitagawara and
Atsushi Kitagawara Architects)
Fig. 12 The virtual and parametric modelling helped to individuate the final dimensions of the
elements, the consequent porosity of the lattice structure, in its relation with the environment of the
project, the ExpoMilano 2015 site (Credit Atsushi Kitagawara and Atsushi Kitagawara Architects)
that, several additional structural tests on nodes have been done also at Politecnico
di Milano, by Prof. M. Colombo and Prof. M. Imperadori.
Finally, after all the physical and digital evaluations, the basic parameter of this
entire design phase has been set: 115 × 115 mm. This is the set value for the basic
glue-laminated timber stick, which presented a length of 2.10 m and eight aigaki
each, and whose assembly portrayed the magnificent wooden façade of the Japan
Pavilion (Figs. 2, 3, 4 and 5).
3.2 Technical Details and Construction Mock-Ups
Once the advanced model and some options have been defined, the design process
moved to the real and full-scale modeling, in order to sharpen the technical details,
and to assess the construction feasibility.
One of the unique features of this structure has been the self-binding assembly.
The three-dimensional lattice, indeed, did not need external supporting elements
during the construction phase. The process presented a breakdown structure that has
Fig. 13 The FEM output about the structural behavior of the wooden grid; it has been assessed,
indeed, through virtual simulation, according the specific geometries and material features, both
during the first design phase by Ove Arup & Partners Japan Ltd./Arup Italia S.r.l.—Eng. Mitsuhiro
Kanada and before the construction phase, by Galloppini Legnami, the Italian company that realized
the façade (Credit Ove Arup & Partners Japan Ltd./Arup Italia S.r.l.)
been directly experimented, checked and validated during the prototyping phase,
and whose progressive steps consisted of the quick connection of timber elements,
without any screw or nails, into the overlapped planar grids and vertical profiles
(Fig. 10). Therefore, the assembly of each physical model has followed the same
procedures: after the realization of the bottom supporting elements that gives the 35°
slope of the “springing plan”, the first flat grid (WG4 + WG3) could be fixed to it,
and host the vertical linear elements (WC1) into the aigaki shapes; then the second
planar grid (WG1 + WG2) is assembled and connected to the same vertical elements
(WC1). Each timber element is, therefore, bond by other eight elements, along with
the perpendicular direction to its own axes line. Finally, after two more flat grids
have been installed, the WC2 elements could be inserted into the so-defined 3D grid;
it locks the structure, thanks to rotation along their longitudinal axes. This operation
generates the inner status of tension.
The first set of 1:1 samples was made in pine wood, with a 100 × 100 mm section,
in order to evaluate the real impact (Fig. 14). This part of the design process has
been aided by the employment of CNC cutting machines (Fig. 15), within which the
digital 3D model of the parametrical design stage has been processed. The standard-
ization of timber pieces and the redundancy of the grid structure, indeed, allowed
the prefabrication of all the timber elements. Hence, the CNC-based production pro-
cess assured an optimization of the manufacturing, in terms of both precision (the
CNC cutting machine guarantees few millimeters of tolerance), and quickness of
execution (between 2.5 and 3 min per each element). Afterward, getting increasingly
closer to the final aesthetical and structural results, also the original material had
to be tested: the first glue-laminated larch wood mock-up—6.00 m high, 4.00 m
wide and 1.40 m deep—was realized by Gallppini Legnami—the Italian company in
charge of the engineering, industrial production, and construction tasks. As first, the
Japanese larch wood was definitely approved and considered suitable, by structural
checks and for CNC cuttings, according to the final production processes. However,
the assembly of the giant prototype (Fig. 15) attested a vertical misalignment of the
joints (3 mm per 1 m of height), being extremely useful in enhancing an important
issue that otherwise would have affected the construction phase on site. Therefore,
the constructor decided for the reinforcement of some nodes, nearby the footing and
the peak areas, without altering the structural concept of the construction system.
Besides, a last FEM-based analysis on the structural behavior was performed by Gal-
loppini Legnami, in order to confirm the assumption of the design phase, according
to the Italian and European standards (617 del 02/04/ 2011; CNR-DT 2008; CNR-DT
2007; 14/01/ 2008; 06/05/ 2008; UNI EN 1995-1-1:2014; UNI EN 1995-1-2:2005;
UNI 11035-1:2010; UNI 11035-2:2010; UNI 11035-3:2010). Finally, the last sample
was transported at the building site, before the final construction phase (Fig. 16).
Once assessed the wooden structure from all those different perspectives, the pro-
duction and construction of the wooden envelope could start. 20,000 Japanese glue-
laminated larch wood elements were collected in 187 packs (260 × 210 × 103 cm)
and shipped from Japan to Italy in order to be processed under the production chain.
Thanks to the restrained dimensions of the timber elements, indeed, their transporta-
tion and handling became more simply and easier: before the milling and cutting
process, each piece had a 2100 mm’s length and a 120 × 120 mm squared section,
while the final shaped element became 115 × 115 × 2100 mm, with a total volume
of 27,500 cm3 . The shipping condition of the material arose another important con-
sideration about the hygrometric behavior of wood and the processing techniques
and details: the initial average moisture content (10.4%) led to adopt a machining
tolerance of 1.5 mm, for the halving joints realization. Hence, all the elements passed
through a first planning phase, performed via a motorized planer, in order to reduce
the squared section and obtain the final chosen dimensions of 115 × 115 mm. These
chosen values guaranteed also the feasibility of the tolerances and the absence of
any defects that typically affect wooden elements. Then, the CNC-based process
elaborated the reduced timber beams in three different CNC machine, carrying out
14 different cutting processes, because the halving cuts were located on two surfaces
per each element. Here, the parametrical design stage helped the management and
the fluidity of the working steps, because of the possibility of codification of pieces
and counting the total amount of produced elements.
Finally, all the processed timber elements were carried to the site, where the lattice
structure, with 115 × 115 × 2100 mm CNC, cut studs, was assembled by Galloppini
Legnami, for EXPO Milano 2015. Here, the constructor realized a concrete founda-
tion, shaped with three steps, in order to assure the 35° slope of the wooden façade.
Then, three glue-laminated timber ground beams (GL24 h) were connected to it by
means of steel bars, and used to fix the first flat grid. Following the designed pro-
gressive phases, three differently high façades—12, 8, 6 and 4 m height, for 1.34 m
Fig. 14 The first Japanese pine mock-up, 1:1 scaled, realized through a CNC-based cutting process
(Ph. credit Politecnico di Milano)
depth and a maximum development of 23.68 m length—were assembled, according

to the architectonical drawings, finally defining the envelope of the Japan Pavilion
for EXPO Milano 2015.
Fig. 15 The first Japanese larch mock-up, 1:1 scaled, realized through a CNC-based cutting process
by Galloppini Legnami, the final constructor, in order to assess the feasibility of the construction
phase (Ph. credit Galloppini Legnami)
4 Re-thinking Process: The Wooden Byobu
Despite the intense, dense and complex process developed into order to accept and
win the challenge of EXPO Milano 2015, the structure was dismantled, at the end of
the event, since the entire exhibition was conceived as a temporary experience.
However, the original idea has had the chance to remain, especially thanks to
the innovative design process that had characterized the conception and realization
of the three-dimensional grid system. At first, the impressive and majestic mock-
ups—realized in order to assess the construction and the process itself—unveils
an intrinsic artistic potential. Then, the engineered practices of production by CNC
machines and fast settlement led to the realization of several samples, which could be
easily assembled, following the same procedures. Those prototypes had, therefore,
the chance to become impressive art pieces, under the master concept of the Japanese
Byobu.
Fig. 16 The last mock-up, 1:1 scaled, delivered at the construction site, to evaluate the real impact
and the possible construction issues (Ph. credit Atsushi Kitagawara Architects)
The Byobu is a traditional element of Japanese culture, among the handcrafted

furniture of old wooden houses. As every object belonging to Asian philosophy had
a hidden meaning and value, the Byobu is more than just a separating screen; it was
conceived as an element able to create multiple spaces and possibilities. Multiple
possibilities as multiple opportunities for the primitive wooden structure to be dis-

seminated and permeate places and people. This is the aim of the re-thinking process,
turning the wooden façade into a wooden Byobu, an art object eligible for outstanding
installation, between Italy and Japan.
4.1 The University Byobu
The first open-air sculpture was placed at the entrance of Politecnico di

Milano—Lecco Campus (Italy), in order to enhance and acknowledge the deep col-
laboration among the designer, builders and the research team of the university, as
the local supporter for the entire design, testing, and construction process.
The University Byobu is exactly one of the first mock-ups realized for studying
purposes by Galloppini Legnami (Figs. 14, 15 and 16) and Politecnico di Milano.
After it assembly, it has been transported from the factory to the university campus.
The reduced dimension of the global volume—3.50 m width, 3.20 m height and
1.34 m depth—and the porosity of its lattice structure facilitated all the handling
operations, avoiding the dis-assembly and reconstruction of the wooden sculpture.
Therefore, it has been installed as a prototype, as the first milestone of the Byobu’s
path (Fig. 17).
4.2 The Urban Byobu
The second experience of the Byobu dissemination takes place in Favara, a small city
near the Valley of the Temples of Agrigento (Sicily). The so-called Urban Byobu has
been located at FARM Cultural Park (https://www.farmculturalpark.com/), becoming
an iconic sculpture of I 7 Cortili (the Seven Courtyards), in the middle of the old city
center. This location is nowadays one of the most innovative places in Italy dedicated
to art and design, and the Urban Byobu has become the stage of artistic performances
and exhibitions, conferences and events, involving every year thousands of people
from all over the world.
The sculpture is, one more time, the result of wooden studs’ assembly, with a 115 ×
115 mm section, the same one defined for EXPO Milano 2015. Here, the majestic
volumes of the Japan Pavilion envelope have been re-calibrated and turned into a
smaller sample—3.50 m width, 3.20 m height and 1.34 m depth—that better fit the
closed courtyard. Its construction involved a group of selected students of Politec-
nico di Milano, among the Premio Compasso Volante—Ettore Zambelli workshop,
which took place at FARM in 2016. The wooden pieces were produced through a
CNC-based cutting process at Borgo Sesia (Northern Italy), according to the same
procedures of the Japan Pavilion façade, and then shipped to Sicily, transported to
Favara, and assembled over there, by Galloppini Legnami.
Fig. 17 One of the larch mock-ups, 1:1 scaled, has been delivered at the university campus of
Politecnico di Milano, in Lecco, and located at the entrance of the main building (Ph. credit Politec-
nico di Milano)
Among all the temporary architectures and installations that are coming in intense
succession at FARM Cultural Park, the Urban Byobu is now a constant iconic pres-
ence, whose wooden texture and lattice porosity perfectly match with the close space
of the courtyards and the consumed finishing plasters of their old buildings (Fig. 18).
Fig. 18 The larch Urban Byobu, at FARM Cultural Park in Favara (Sicily), 2016 (Ph. credit Politec-
nico di Milano and FARM Cultural Park)
4.3 The Forest Byobu
Ideally connected to the Sicilian experience, the third art piece has been called the
Forest Byobu, since it is located at Arte Sella—The contemporary mountain, a famous
land park close to Trento, in Northern Italy (http://www.artesella.it/en/). In this spec-
tacular location, numerous artists, designers, and architects gather unique and cre-
ative ideas to develop an ongoing dialogue with the natural world. Since 2017 Prof.
Imperadori is the scientific consultant for architecture and Atsushi Kitagawara was
selected to be the author of the first artwork.
The larch wood structure presented the same dimensions and proportions of the
Sicilian experience in order to enhance the connection between the two artistic
objects.
By contrast with the Urban Byobu, the Forest Byobu is conceived as a geometrical,
rationally define a wooden piece that goes back to the forest, into the woods. Here
the construction material is becoming a part of the environment and changing its
aspects and surface texture according to the natural outside condition. The CNC
processed larch wood—after the parametrical design—has been placed back in the
middle of real trees, creating a dialectic between Nature and Artifact, which is magic
and unique. Let us image birds or insects making their nests in the artificial, squared
branches of the structures, or mushrooms and spiders populating such ephemeral
geometries (Fig. 19).
4.4 The Museum Byobu—Kigumi Infinity
The last experience of the artistic application of the Byobu’s concept and structure
is the Kigumi Infinity at Mori Art Museum of Tokyo, for the exhibition “Japan in
Architecture: Genealogies of Its Transformation” (Fig. 20) (https://www.mori.art.
museum/en/).
In this case, the lattice structure has been installed indoor as a division wall, a
porous partition which enhances the entrance of the exhibition and architectonically
articulates the closed space: a real Byobu in the famous museum in central Tokyo.
Here, the parametric design helped in the newer definition of two majestic pieces,
more than five meter height, whose dimensions and respective proportions have been
set according to the golden ratio, in order to obtain an aesthetical perspective of the
indoor space. The two volumes, indeed, have fulfilled the height of the space, reaching
the false ceiling structure and being hung to it, in order to avoid possible instability
issues. The timber elements have been realized as massive 105 × 105 mm section
profiles of hinoki—Japanese cypress wood, processed by automated machines, then
sharpened by skilled craftsmen, in the Yamacho CO, Ltd factory, and delivered on
site.
For this experience, the parametric and engineered design that was already devel-
oped for the original project of the Japan Pavilion façade and then upgraded for the
Fig. 19 The larch Forest Byobu located at Arte Sella—The contemporary mountain, near Trento:
to the left, the Forest Byobu whitened by snow (Ph. credit Giacomo Bianchi for Arte Sella); in
the center, Marco Imperadori, Emanuele Montibeller, Atsushi Kitagawara, Valentina Gallotti and
Mayuko Matsukawa at the opening ceremony (Ph. credit Federica Brunone); to the right, a cat
enjoys the wooden geometries (Ph. credit Mayuko Maztukawa); below, the Forest Byobu in its new
location nearby Villa Strobele, at the foot of the forest (Ph. credit Marco Imperadori)
Italian experiences on Byobu’s realizations, led the management of the production

and construction phases. The codification of all the wooden members and the possi-
bility to extract data and information directly from the 3D virtual models have been
Fig. 20 Kigumi Infinity—the Museum Byobu is an installation for the exhibition “Japan in Archi-
tecture: Genealogies of Its Transformation”, 2018, Mori Art Museum, Tokyo: the pictures show the
different stages of the design process, from the scaled mockup realization in the AKA studio, with
Federica Brunone and Prof. Marco Imperadori, to the construction phases (Ph. Mayuko Matsukawa
for Atsushi Kitagawara Architects) and the complete installation (Ph. Shigeo Ogawa for Atsushi
Kitagawara Architects)
important elements to allow the communication between the Italian constructors and
local advisors, and the new Japanese clients, Mori Art Museum, and companies in
charge of the production and construction (Yamacho CO, Ltd and Maeda Corpora-
tion).
The Museum Byobu—Kigumi Infinity art piece closes the fil rouge that connects all
the samples, witnessing the fusion of Japanese tradition about wooden construction
and handcrafts, and innovative technologies.
5 Conclusions
This intense and complex design experience portrays and attests the availability to
integrate traditional techniques and crafts with the most innovative computer-aided
technologies, especially within a field of the architectural and construction engi-
neering sector—the wooden construction systems—that has been ignored for a long
time. These advanced technologies are representing a sort of Renaissance of wood,
highlighting the great potentials of this material, in design, architecture and engi-
neering. However, it would not be possible without a profound and deep knowledge
of the original and practical knowledge, handed down from generation to generation.
However, the FEM-based analyses or the parametric design represent powerful tools
that are able to prove the ongoing effectiveness of this ancient knowledge, about
wooden structures and their mechanical behavior, and their feasibility and use for
newer architectonical visions.
Within the development of the wooden façade of the Japan Pavilion for EXPO
Milano 2015, this fusion between tradition and innovation has represented not only
the validation of the specific construction techniques and its structural behavior,
but also the diffusion of its artistic and architectonical potential on a worldwide
panorama, beyond the Japanese geographical boundaries. This project has been an
occasion to touch the Japanese sense of aesthetics, and paved the way to other great
masters, Sensei, of the Japanese architecture, who are working on the same path.
In conclusion, all the re-using experiences are unified by the use of the same
structural concept, resulting in its own technical power and solidity as well as its
ephemeral and artistic potential. From a façade at EXPO Milano 2015, to a sculpture
in an old town or in a forest, or as inside division wall/sculpture, this concept shows
the evocative potential and flexibility of wood and its use, through dense, engineered,
design processes towards a synthetic result, which is extremely clear and variable
and where complexity is hidden by simplicity.
Acknowledgements The authors would like to deeply thank all the collaborators, researchers,
architects, engineers, designers, and experts that have been involved at different stages of the whole
design process, from the earlier conceptual idea to the latest realization. The entire work is the result
of an intensive production that last several years and involved design pros in a fruitful cooperation
between Italy and Japan.
A special thank goes to Atelier2—Arch. Valentina Gallotti, for the close collaboration with Prof.
Marco Imperadori, in supporting and supervising the design and construction processes; to Marco
Corbella from Ishimoto Architectural & Engineering Firm, Inc., as a part of the design team; to
Galloppini Legnami, for the collaboration in developing details and availability to make the structure
reliable for the re-use projects; to the team of Politecnico di Milano—Lecco Campus; to Andrea
Bartoli and Florinda Saieva for having embraced and welcome the structure at FARM Cultural Park,
and to Emanuele Montibeller and Giacomo Bianchi from Arte Sella, for hosting it in its beautiful
landscape; at last, to Mori Art Museum for having recognized the power of the construction and
chosen it as the icon of one of its representative exhibition.
In the end, we would like to acknowledge all the graduating students and trainees that have been
involved in the development of virtual parametric models and construction of the several mock-ups
and samples.
Bibliography
Aicher S, Reinhardt H-W, Garrecht H (eds) (2014) Materials and joints in timber structures. Recent
developments of technology. In: RILEM bookseries, vol 9, Recent developments of technology.
https://doi.org/10.1007/978-94-007-7811-5
Betsky A, Imperadori M, Kawamukai M, Sauerwein K (2013) Atsushi Kitagawara Architects. In:
Jaeger F (ed). Jovis Verlag, Berlino
Bottelli V, de Agostini F, De Giorgi M, Raggi F, Volpi D (2015) EXPO DOPO EXPO. Lo sguardo
di otto fotografi sulle eredità urbane e ambientali di sette Expo, SOLFERINO edizioni
Cipollini M, Bonannini E, Cinotti M, Sassu M (2016) Design, production, and installation of
wooden walls for the Japan Pavilion at Expo 2015. Buildings 6(4):43. https://doi.org/10.3390/
buildings6040043
CNR-DT 207/2008 “Istruzioni per la valutazione delle azioni e degli effetti del vento sulle
costruzioni”
CNR-DT 206/2007 “Istruzioni per la Progettazione, l’Esecuzione ed il Controllo delle Strutture di
Legno”
D.M. 14/01/2008 - Ministero delle Infrastrutture - Approvazione delle nuove norme tecniche per le
costruzioni (Gazzetta ufficiale 04/02/2008 n. 29)
D.M. 06/05/2008 - Ministero delle Infrastrutture - Integrazione al decreto 14 gennaio 2008 di
approvazione delle nuove “Norme tecniche per le costruzioni” (Gazzetta ufficiale 02/07/2008 n.
153)
Doust N (2015) Japan Pavilion—Atsushi Kitagawara. In: Arketipo, vol 94. Il Sole24Ore, pp 48–57
Fukuyama H, Kairi M, Hirisi H, Inayama M (2008) Shear characteristics of wood dowel shear joint
and practical application example. In: WCTE 2008: 10th world conference on timber engineering:
June 2–5, 2008, Miyazaki, Japan: Conference Proceedings, Miyazaki. Retrieved from: http://
support.sbcindustry.com/Archive/2008/june/Paper_167.pdf
https://www.farmculturalpark.com/
http://www.artesella.it/en/
https://www.mori.art.museum/en/
Imperadori M (2013). Rethinking Japanese tradition: wooden buildings. In: Jaeger F (ed) Atsushi
Kitagawara Architects. Jovis Verlag
Imperadori M (2015a) EXPO 2015: construction site of wonders. In: Area, vol 139. Milano
Imperadori M (2015b) Tensione compressiva. Pura Geometria. In: Modulo, vol 396. BE-MA
Editrice, Milano, pp 432–436
Imperadori M, Senatore A (2011) Schematic design: tecniche ed esempi di comunicazione del
progetto. Il Sole24Ore, Milano
Japan Satoyama Satoumi Assessment (2010) Satoyama-Satoumi ecosystems and human well-being:
socio-ecological production landscapes of Japan—summary for decision makers. United Nations
University, Tokyo, Japan. Retrieved from: http://archive.ias.unu.edu/resource_centre/SDM-EN_
24Feb2011.pdf
Kamachi K, Adachi K, Tanaka K, Inayama M, Inoue M (2008) Full-scale test and analysis on
wooden made shear plate fro timber structure. In: WCTE 2008: 10th world conference on timber
engineering: June 2–5, 2008, Miyazaki, Japan: Conference proceedings, Miyazaki. Retrieved
from: http://support.sbcindustry.com/Archive/2008/june/Paper_168.pdf
Kobayashi K, Inayama M, Ando N (2008) New estimation method on stiffness and strength of single-
shearing screw joints with structural panels. In: WCTE 2008: 10th world conference on timber
engineering: June 2–5, 2008, Miyazaki, Japan: Conference proceedings, Miyazaki. Retrieved
from: http://support.sbcindustry.com/Archive/2008/june/Paper_144.pdf
Routledge, UK
Seike K (1978) The art of Japanese joinery. Weatherhill ed
Song C, Inayama M, Ando N (2008) A Study on the deformation properties of wooden shear wall and
beam stiffened with inserted glass plates in response to applied load. In: WCTE 2008: 10th world
conference on timber engineering: June 2–5, 2008, Miyazaki, Japan: Conference proceedings,
Miyazaki. Retrieved from http://support.sbcindustry.com/Archive/2008/june/Paper_291.pdf
Circolare n. 617 del 02/04/2011 del Ministero degli Interni - Istruzioni per l’applicazione delle
UNI EN 1995-1-1:2014 Eurocode 5—design of timber structures. Part 1-1: General—common
UNI EN 1995-1-2:2005 Eurocode 5—design of timber structures. Part 1-2: General—structural fire
design
UNI 11035-1:2010 Structural timber—visual strength grading for structural timbers. Part 1: Ter-
minology and measurements of features
UNI 11035-2:2010 Structural timber—visual strength grading for structural timbers. Part 2: Visual
strength grading rules and characteristics values for structural timber population
UNI 11035-3:2010 Structural timber—visual strength grading. Part 3: Timber Uso Fiume and Uso
Trieste
www.kitagawara.co.jp
Experimenting the Use of Wood
in Contemporary Architecture:
Integrating Research into Practice
Salvator-John A. Liotta
Abstract This chapter is a take on contemporary works in wood seen from two
different points of view: academic and professional. The knowledge about digital
wood developed at different universities through Digital Fabrication Laboratories
and, when teaching architecture, it has proved to be effective but with certain limi-
tations when used for real constructions. In fact, translating the freedom of building
temporary architectures—which is usually one of the prerogatives of teaching archi-
tecture during design studio or workshops—into wood architecture that respect all
the constraints of real construction is a challenge. Parametric design is essential to
control process and final output. There are several aspects such as building codes,
costs, certifications, durability and expertise which have a strong impact on the fea-
sibility and final output of digital wood architecture. This chapter shows several
experiences where innovative ideas developed through research have been applied
both to temporary pavilions and real constructions in Japan, Italy and France. Two
tea houses, a pavilion for archeologists, a façade for a recreational youth center, a
modular furniture system customisable online and the interior design of the media
library of the Quai Branly Museum in Paris will be discussed. Integrating research
into practice is a way to try to make better buildings, that are appropriate to their
users, clients, context and time.
Keywords Parametric design and fabrication strategies · Pedagogy and practice

CNC and woodworking technology · Wood joint systems
Wood complex surface
S.-J. A. Liotta (B)

Faculty of architecture La Cambre Horta, Free University of Brussels, Brussels, Belgium
e-mail: salvator-john.liotta@ulb.ac.be

1318 S.-J. A. Liotta
1 Introduction
Computational software and digital fabrication technologies have transformed the

building and construction sector. A domain which, until the 20th century, was rather
static and closed. In recent years, there has been an enormous shift in architecture,
engineering and design.
Currently, architects, designers and engineers have to face increasingly complex
design requirements which leads to the question: how do practices acquire the ability
to do so? There are at least two ways: either by developing applied research on their
own or by including students trained in complex design (because they have partic-
ipated in workshops on digital design and construction or because they have taken
courses on parametric design). These are the two most viable methods to integrate
research into work and to prepare for the changing form of practice in architecture.
However, once this conceptual framework has been integrated into practice there are
some decisive issues left open such as budget limitations, building code regulations,
technical constraints, durability, expertise and machines used by the construction
companies involved in construction projects. These have a strong impact on the
feasibility and final output of the most advanced architecture.
This chapter considers contemporary works in wood seen from two different
points of view: academic and professional. It builds upon several years of research
experience at the Kengo Kuma Lab of the University of Tokyo, where the author
worked for a total of 4 years as a post-doc researcher and as an instructor and mem-
ber of the scientific committee of several workshops. These workshops aimed to
build small architectures in wood through the use of parametric software to test
and prove the coherency of the link between digital tools and traditional culture,
between technological innovation and a redefinition of traditions (Ko and Liotta
2011). After leaving the University of Tokyo, thanks to a Marie Curie IFF scholarship,
the author joined the National Research Center (CNRS) in Paris and then in 2016—as
an ass.professor—the faculty of architecture La Cambre Horta at the Université Libre
de Bruxelles. In 2012, the author became a partner at LAPS Architecture—an archi-
tecture office based in Paris which is characterized by a strong interest in developing
an applied research based approach to architecture. The office—which has won sev-
eral awards and has been invited two times to Venice Architecture Biennale—back
in 2012, was searching for a specific language through the integrated use of para-
metric design and the employment of different materials. In the past years, LAPS
Architecture has integrated digital wood research into private and public projects of
different scales. Through the study of different case studies, the goal of this chapter
is to show how research can be integrated into practice.
Experimenting the Use of Wood in Contemporary Architecture … 1319
2 Contemporary Works in Wood Between Pedagogy

and Profession. How to Integrate More Research
into Architectural Practices
2.1 Concrete Experience of Construction
Efforts in establishing research in the use of digital wood in architecture commenced

in 2009 with the first courses at the University of Tokyo introduced by Kengo Kuma
and related seminar courses that focused on a parametric approach to different materi-
als including wood. In Japan, architecture students have different courses and studios
where they are exposed to structural analysis and materials technology. Full-scale
constructions constitute an essential element of research by design in these studio
courses. Besides design studios, the study of architecture, materials and structure is
often reinforced thanks to targeted activities such as summer schools and hands-on
workshops with partner universities therefore promoting an exchange of experience
between students on design, space, structure, technology and materials.
Examples of successful courses include the participation of other universities such
as Columbia University and Politecnico of Milan. During these activities, students
were asked to conceive and build small pavilions to test all the aspects of translating
ideas into real projects. Research by design as a mode of inquiry had its own particular
modalities and one of the most effective strategies is to design and build a small
pavilion. Building a pavilion presents several positive aspects because it can be done
at reasonable costs, it is ideal to test new materials and it can be built by the students
themselves. Learning outcomes were very positive because, thanks to these activities,
students got a first concrete experience in the physicality of construction.
2.2 Research at Work
Among small-sized architectural offices, it is not common to have a research section

because it is time consuming and not immediately economically sustainable. How-
ever, LAPS Architecture has invested in research activities related to architecture and
the built environment to create a comprehensive approach to architecture. According
to Hensel and Nilsson (2016), practices interested in research develop a knowledge
that “involves specific ways of integrating architectural education and research, in
particular research by design, to support the development of related knowledge, con-
cepts, methods and frequently also the production of reliable empirical data, engaging
real-life design problems and seeking interaction with practice”. LAPS Architecture
has focused on the combined use of digital tools and materials, such as wood, U-
Glass, raw earth, integrated in the construction of private and public buildings.
For the author, the knowledge about digital wood developed at different uni-
versities through Digital Fabrication Laboratories and by teaching architecture has
proved to be effective but within certain limitations when used for real construc-
tions for public architecture. In fact, translating the freedom of building temporary
architectures—which usually is one of the prerogatives of architecture built at uni-
versity—into wood architecture that respects all the constraints of building codes is
a real challenge. The research section at LAPS Architecture relies on a team of usu-
ally two or three people according to the size of the mission. The work is organized
around two different types of projects: (1) basic research based on form finding and
material testing, and (2) oriented research aimed at solving real projects. Basically,
on one hand there is a research for aesthetic pleasure1 of searching for new geom-
etry, shapes, materials, assembly logic and structures. On the other hand there is a
research oriented at solving concrete constructive problems including building codes
and budget.
2.3 The Results of Research Based Practice
The chapter describes how knowledge developed through academic research is

increasingly integrated into a small sized architectural practice on different projects
sizes ranging from small to large. The article is organized diachronically which is
significant in that it addresses different projects and resourcing requirements that
are enabled by specific project sizes. Scale determines the type, scope and modes
of research of each project. With the aim of experimenting the use of wood in con-
temporary architecture, the projects covered include: two tea houses, a shelter for
archeologists, a partition and shelving system, two facades for public buildings and
an interior for a media library (see Table 1).
3 Research: Building Small Pavilions as a Strategy
A small pavillon does not require significant economic investment (often a private
sponsor or a research fund cover the expenses) but it is an invaluable tool to advance
applied research in the use of new softwares (i.e. for optimizing structures), in testing
new solutions, materials, forms, performances and assembly techniques. Moreover,
there is a rising demand in students wanting to participate in hands-on experience
where they are involved in the translation of concepts into reality. The main idea
involved in building a pavillon is to shorten the distance between conception and
1 Cyril Stanley Smith as early as the 1980s believed that aesthetic curiosity was key to both genetic
and cultural evolution. Without aesthetic curiosity, maybe we wouldn’t have survived at all, or we’d
still be back in the stone age. For the emeritus professor in Science of Materials at MIT, many of
man’s inventions have appeared in decorative applications: the wheel, for example, first appeared
in decorative jewellery and children’s toys. Smith proves that innovation and discovery are not born
under the pressure of need and that aesthetics takes the lead from function.
Table 1 Each project presents different characteristics concerning program, materials, costs, fab-
rication and regulations among other issues
Name, Geometric Materials, Production, Regulations, Image
Program Research Fabrication, Construc- Code,
Assembly tion, Cost, Professions
Durations of of the people
design and involved
construc-
tion, number
of people
involved
TEA Groove Plywood + Pre-cut at Prerequisites
HOUSE I, patterns and Tatami factory and and rules
Tokyo densities CNC cut + assembled defined by a
Temporary determining Screwing on site by 8 scientifique
pavilion the desired students. committee.
bending of 1500 US Students,
wood dollars. profession-
3 weeks als,
workshop architects
and
engineers
involved
TEA Curved Plywood Pre-cut at

Prerequisites
HOUSE II, surface CNC CUT + factory and
and rules
Tokyo tessellation Screwing + assembled
defined by a
Temporary composed of finger joints on site by 8
scientifique
pavilion 3D elements students
committee.
1500 USStudents,
dollars.
profession-
3 weeksals,
workshop
architects
and
engineers
involved.
Shelter for Investigation Pine Wood Partially Prerequisites
archeology, on reinter- CNC CUT + pre-cut at a and rules
Agrigento pretation of Screwing factory and defined by a
Temporary traditional partially on scientifique
pavilion bracket site. committee.
system Assembled Students,
through on site by 5 profession-
parametric students in a als,
design, week. 1600 architects
reconfig- e. 45 days and
urable grid of design + engineers
system 7 days of involved
construction
(continued)
Table 1 (continued)
Name, Geometric Materials, Production, Regulations, Image
Program Research Fabrication,Construc- Code,
Assembly tion, Cost, Professions
Durations of of the people
design and involved
construc-
tion, number
of people
involved
XME Investigation MDF-wood Pre-cut at Stability and
System on the composite factory and resistance of
Partition and possibility material + assembled structure,
shelving of applying Aluminum on site by Architects
system a grid joints profession- and
pattern for CNC CUT + als or by professional
generating wood and clients carpenters
three joint
dimensional connection
forms
through
parametric
design,
configurable
on line
Youth Investigation Outdoor Pre-cut at DTU +
Recre- on Pine Wood factory and spacing
ational parametric pre-coated assembled between
Center in surface CNC CUT + on site by 3 lamelles
Canteleu, waves Hand made professional restricted by
Normandy pattern finishing + carpenters recent rules
Educational formation Nailing about wood
facility and construction
variation in public
building.
Architects
and
engineers
involved
Media Investigation Amarante Pre-cut at Building
Library on surface wood, factory and code and
Quai pixelisation coated for assembled issues linked
Branly fire on site by 3 to shelves
Museum protection professional cleaning.
Interior CNC CUT + carpenters Architects
architecture Hand made and
finishing + engineers
Gluing involved
construction and offer students a knowledge which is based on a more organic inte-
gration between architecture, materials and structure.
This section of the chapter describes three pavilions, which were the output of two
workshops. One held at the University of Tokyo in 2011 and another at the Unesco
Heritage site of Agrigento in 2013. The workshops were organized with the following
structure: two universities invited, a topic linked to culture, a pre-defined budget, a
limited number of students, a two-phase period including design and construction,
wood chosen as main material and use of parametric software and digital fabrication.
The first two pavilions were part of a reflection on tea culture and the third pavilion
is a speculation on a temporary shelter for archeologists.
3.1 Approach and Practicability of Parametric Design

and Digital Fabrication
The two workshops provided a point of reflection within the academic setting to con-
sider the consequences of computational design when applied to the physical reality
of making, rather than stressing too heavily on the form-finding aspects of compu-
tational design. Engagement in a continuous process leads to a constant feedback
loop between the software, material characteristics, and contextual considerations.
The parametric tool allows for the input of an almost infinite amount of informa-
tion, producing mathematically logical, countless variations during the process, but
it obviously cannot determine any hierarchical design decisions. In an essay, Mat-
sukawa (2007) dissects the differences between conventional design process versus
the generative model, where diagrammatic process models between environment,
architect, building, user, and other basic criteria such as building codes, operate dif-
ferently and it is illustrated in examples including linear model or partial feedback
loop model. One of the most complicated process diagrams—where series of algo-
rithms are employed—leaves the architect with the “question” without a definite
answer.
The workshops became a framework for the assessment of design and available
technology, both of which were then experienced and evaluated by the participants as
well as the tutors, invited experts and guests. Each of the pavilion’s physical presence
and their feasibility on multiple levels were evaluated and reconsidered throughout.
The process also assessed the technology employed and the highly sophisticated
elements that can be produced, whilst relieving participants, who were inexperienced
in construction, of some issues including protection from weather, foundations, and
structural stability over an extended period of time, thanks to the temporariness
of pavilions. It also encouraged material experimentation, intuitive predictions for
structural elements, and a trial and error approach in detailing and assembly. The shift
in focus from satisfying the public at large to an increasingly personal scale gave
an ideal opportunity to test these digital tools. The small scale of the tea house and
its function, both of which are deeply rooted in culture and its sensibilities, may not
produce monumental visions single-handedly. The potential for individual insights

to be optimized by the use of parametric tools and processes pointed to a direction
where the framework of culture can serve as a starting point for selective optimization.
Design has evolved alongside technology in a mutually beneficial relationship. As
a result of one of the most significant innovations initiated by parametric modeling
software, participants were ultimately able to extract, edit, and abstract the traditional
aspects pertinent to their works and transform into their contemporary possibilities
of what a tea house and a shelter for archeologists might be.
The workshops organizers decided to use software such as Rhinoceros and
Grasshopper, because they offer architects a chance to manipulate tools that facili-
tate their imagination while engaging the challenges of design and construction. For
the duration of the entire process—from design, manipulations, detailing, to fabri-
cation—it became apparent that the use of a graphical algorithm editor tool allows
for instantaneous modifications and influences on architectural conditions and con-
text. On the other hand, it is easy to veer towards oversimplification in such models;
when a single line, for example, is further manipulated into endless variety of forms
and surfaces, it alone can seem to substitute complex architectural elements. How
would these simple manipulations translate to construction? What would this dia-
grammatic and methodical approach produce when combined with the culturally
meaningful function such as a tea house or an archeological site? The first attempts
on cultivating formal innovations resulting from digital design process applied to
construction of tea houses and a shelter for archeology, highlighted that parametric
design can be (i) a mechanism through which architects are able to produce new solu-
tions and renew traditional conceptual meanings, and (ii) a tool to retain architecture
converging with cultural values.
4 Digital Tea House Workshop
Held at the University of Tokyo, together with Columbia University GSAPP, Digital
Tea House was a joint workshop with the aim to design and build three pavilions
for hosting tea ceremonies. Issues addressed in the three-week workshop ranged
from applications of computational design, interpretations of tradition and culture,
structural stability, to practical solutions for quick physical materialization within
limited time and budget. The workshop was divided into two sections. The first part
introduced computational logic and concepts, which led to the second part where
explorations relating to the Japanese tea ceremony culture served as a pretext for
further exploring digital design and fabrication. Three teams, each comprised of 6–8
members, ultimately produced three full-scale tea houses to test out their concepts,
methodologies and materials. Several elements served to make comparisons and
analysis during the process and later in two distinct outcomes: baseline for common
software (Rhinoceros and Grasshopper), principal material (50 sheets of 9 mm and
12 mm thick 3 × 6 plywood), fabrication method (CNC routing), and budget (of
approximately 1500 USD, per team, excluding plywood and CNC fabrication costs).
Fig. 1 Traditional elements, above, were used as inspiration to create innovative interpretations
below
4.1 Tea House as a Cultural Backdrop
In the 16th century, the Japanese people gave life to a new culture which consisted
of elements such as the ‘Tea House’, ‘Sukiya Style Architecture’ and the ‘Wabi-Sabi’
aesthetic of transience (Nishi and Hozumi 1983). Kengo Kuma comments that the
changes thus initiated can be considered “a reorganization of all products where the
body is used as the reference. […] They thought that relying on anything else as
the reference would only add to the confusion” (Fischer and Schneider 2008). The
merciless discarding of traditional standards and cultural hierarchy prevalent in Japan
up to that point constituted a significant break with Japanese tradition. At the time,
the ‘Tea House’ thus represented an exceedingly avant-garde type of architecture
(Fig. 1).
Stemming from the use of the body as a point of reference, the tea house evolved
into a microcosmic situation where host and guest(s) meet and, at the same time,
an intellectual device through which one is made aware of the natural phenomena
occurring outside. Okoshi-ezu, a 3-dimensional chapter model which first appeared
in the 16th century, was often employed for the planning of tea houses as an effec-
tive visualization tool for potential effects resulting from minute alterations in the
placement of windows, doors and partitions within the extremely small space inside.
Effectively it was a preceding method of simulation (Moussavi and Kubo 2006).
Moussavi writes: “architecture needs mechanisms that allow it to become connected
to culture. It achieves this by continually capturing the forces that shape society as
material to work with it. Architecture’s materiality is therefore a composite one,
made up of visible as well invisible forces.” The traditional tea house is composed
of a variety of elements, including shoji2 screens or bamboo slats that serve as light
2 Shoji
is a vertical screen, typically composed of Japanese washi chapter stretched over delicate
wooden frame.
filters that underscore gradations of light, a small wooden sliding door at nijiri-
guchi3 entryway, tatami mats that signify where one may sit, and a recessed alcove
tokonoma4 for hanging scrolls and flowers. Kengo Kuma notes that in recent years,
we are starting to witness the generation of new types of creation due to both the
diversionary and personal. He describes that the digital is similar in the sense that it
is also a reversal of an idea, based on the notion that within the minimal lies a new
clue that can turn the world upside down (Kuma 2011).
4.2 From Context-Neutral to Context-Aware Design
How can we build a bridge between the digital environment to the physical envi-
ronment in which we live? First, we addressed the consequences of the designs -
how every part should be considered for its strength, weight, its assembly sequence,
and surface treatment. There are also questions regarding appropriateness and scale
of design, which often cannot be resolved in the isolated modelling phase. How
to reconcile disparities is one of the biggest challenges in the design process; this
workshop used the cultural function of the Japanese tea house as a starting point to
address the gap. The often scale-less nature of design through scripting is applied
via a traditional set of rules, such as the prototypical 4.5 tatami-mat scale of a tea
house, to be physically implemented. In her book about digital fabrication, Iwamoto
writes:
[…] the route from virtual to actual is one of constant calibration. […] Taking such exigencies
into account, one may observe a host of exciting digital-fabrication projects that have cropped
up, using material and constructive constraints to alter end results. Because of the nature of
designing in such a manner is improvisational […] relationships among the design, material,
fabrication, and assembly are intentionally kept flexible through the final building stage. The
design-build process fosters experimentation, where fortuitous “accidents” may lead to new
insights and unintended design consequences. (Iwamoto 2009)
Concurrently a kind of geometrical rationale embedded in parametric processes

can be contradictory to traditional cultural principles. The type of optimization that
can be clearly present in the relatively simplified model of the parametric process
may have a very contradictory set of values compared to the traditional sense of
esteem for subtle and intricate deformities and variances in nature. Perhaps with
practice such subtleties can be emulated, but are extremely difficult with the level of
computational skills that the participants possessed. However, layers of complexity
could have been incorporated and each of the pavilions represented such an attempt.
The attending groups did so in a visually pronounced manner, at macro rather than
3 Nijiri-guchi
refers to a small entryway specific to the tea house, approximately 60 cm in both
height and width. The tightness of the opening makes the guest crawl through with his or her head
down.
4 Tokonoma, a small recessed alcove within a room where guests are received, is where flower
arrangement and hanging scroll are commonly displayed.

micro scale for maximized effect. They tried to apply rationalistic principles to the
chosen rational set of traditional rules and to exact the perfect middle-ground, as
well as trying to tame the material to find the least confrontational approach. The
process of computational design and fabrication can amplify complexity as desired,
but perhaps more intriguing are considerations on what new values can be extracted
from the combinations of culture, tools, and materials, and what sets of cultural values
are left behind in the process of translation, or decoding. The development process
from context-neutral idea to context-aware architecture calls for multidimensional
views, and the deviations apparent in the outcomes each show different approaches
and emphasise the problem of the rationalization of traditional aesthetic sensibilities.
4.3 Pavilion “Nami-no-Ma (Space of Waves)”
Guided by strong aesthetic characteristics from the tea ceremony, the expression
of the beauty and imperfection of nature inspired by the tea bowl is translated to
plywood, which surrounds the basic 2-tatami traditional layout of the interior space
(Fig. 2). The initial concept was drawn from the slightly irregular traces left from the
process of throwing the tea bowl on the potter’s wheel. The bowl used in Japanese
tea ceremony favored controlled imperfection in the aesthetic of yuragi and yugami.
Yuragi is the slightest warping often from the uneven pressure of the kiln, which later
developed into a more as deliberate and artistically restrained distortion of yugami.
Pavilion Nami-no-Ma boldly translates the phenomenon of yugami in the same cal-
culated manner as the ceramicists of the past, with every layer of plywood by taking
advantage of the CNC router. Efforts to create a natural and flowing form from the
3-axis CNC routing, which is a flat surface fabrication, pushed for experimentations
with half-depth grooves in specifically calculated patterns on the 9 mm plywood.
The key challenge was in achieving the desired bend in a continuous curve following
a circular geometry of the plan. Enabled by close communications with CNC router
operators, tests initially began with grooves of different depths and stitch patterns of
varying lengths. The triangulated grooves eventually proved to be the ideal solution
for 3-directional curves to be fixated on site, whereas perpendicular grooves only
enabled 2-directional bending per panel. The resulting pavilion is a soft envelop for
the interior activity, accentuated by modest connections to the surrounding environ-
ment (Fig. 3).
Tea houses typically offer a limited level of openness to the outside. In this inter-
pretation, the varying thickness of the wall becomes the boundary between the tea
ceremony taking place inside and the surrounding nature, while the views are con-
trolled by the density and bending angle of each layer. The undulating waves (Fig. 4)
also facilitate the functions of tokonoma and nijiri-guchi, where the largest opening
in the pavilion is structurally reinforced beneath the lower curvature to support body
weight.
Fig. 2 Pavilion Nami-no-Ma
4.4 Pavilion “130008252010”
Designed for the precise moment of 13:00 on Aug. 25, 2010, when the tea ceremony
was to take place, the interior space is planned for specific lighting conditions where
fragmented shadows align with the cut-out straw mat floor pattern, serving as an
ever-changing ornament but also indicating where host and guests may sit. Angles
and depths of 192 triangular patterned sun-shading units over a semi-conical arch
structure, leaning forward at 35.5°, the latitude of Tokyo, have been designed with
considerations for the passage of time and changes in sunlight (Fig. 5). In the tea
house, window placement is calculated not only for ventilation and visual effect
on the walls, but also to create just the right play of light and shade when the tea
ceremony is performed.
The design of ‘130008252010’ was partially influenced by the traditional con-
cept of sabi, which refers to that art of imperfection celebrated in The Book of Tea
by Kazuko Okakura. Traditionally, according to Okakura, there is a compulsion to
Fig. 3 Exploded axonometric drawing shows composition of parts to be assembled
Fig. 4 Groove patterns and densities determined according to the desired bending (left), tests of
the groove patterns and bending characteristics with plywood
divide space into equal parts in Western art forms, to attain a sort of mathematical
and geometrical balance and perfection, and leaving nothing for the viewer to add
(Okakura 2007). Conversely, according to De Mente one of the traditional Japanese
ways is “to utilize asymmetry and austerity to achieve surprise and delight elicited
by the unbalanced, by the apparent randomness of things that allows the viewer to
complete the image” (De Mente 2006).
Pavilion ‘130008252010’ was almost entirely based on this concept of letting the
viewer complete the image. As the inclination of sunlight coincides with the pattern
Fig. 5 Pavilion 130008252010 (left), interior view (right)
designed on the floor, for only a few minutes per year, the tea house becomes a
continuous recreation of the image. This work is imbued with the spirit of a Zen
teaching called yugen, which refers to a type of beauty, in a harmony so delicate and
so right that only an attentive observer can see and perceive it.
The external surface of the pavilion is a smooth continuous curvature while the
internal surface is irregular, with varying depths of the triangular units. These units
posed a challenge in terms of connections due to their wide range of angles; fin-
ger joints resolved the issue by increasing the contact surface area, but these joints
would have been considered differently with a 5-axis router or a different tool to
enable smoother edge conditions where as many as seven angles congregate. Cor-
responding angled fins made of laser-cut white translucent chapter-plastic laminate
inserted into each unit further regulated desired lighting conditions. However half
of the fins were lacking in the end because of a miscalculation of time; while para-
metric manipulations are almost instantaneous, physical realization through labor
takes much longer. Pavilion ‘130008252010’ is essentially a shading device (Fig. 6),
materializing the natural phenomenon of daylight through their tea house. It is a
way of visualizing something as ephemeral as light via the incorporation of the 4th
dimension, time, by utilizing parametric software. Although traditional sequential
and spatial considerations were somewhat sacrificed by its singular focus, it was also
very poignant within the ever-changing interior conditions.
4.5 Output of the Tea House Workshop
Conceived and finished in a very condensed period of less than one month, com-
putational design and digital fabrication enabled the construction of each pavilion.
Every decision inevitably affected other aspects of the design and its construction,
especially since each pavilion was designed to be movable and easily reassembled
multiple times in different locations. Applying materials suitable for both design
and fabrication processes was crucial, since factors such as the joints and weight of
Fig. 6 Surface map of all triangular units, divided into alphabetical zones for organization (left),
exploded isometric (right)
preassembled parts inhibited students from directly realizing what was envisioned
initially. Each constraining factor pushed for creative and feasible solutions. The
output of the workshop clarified that, firstly, parametric processes are not contradic-
tory to traditional cultural principles; and second, how traditional elements of the tea
house can be decoded and formally reinterpreted through parametric design. From
a cultural point of view, aspects of the tea house were interpreted in a contemporary
digital manner and can be seen as reviving the tradition of avant-garde. This points
to the benefit of implementing present technologies of parametric design and digital
fabrication, but also with certain specificity. When the specificity is drawn from a
cultural reference, one seems to be more sensible of the complexities and rationales
that require interpretation. It was made clear that such traditional values and new dig-
ital technologies are not contradictory, as exemplified in the tea houses and enforced
by the final tea ceremony held inside in the forecourt at the University of Tokyo. On
the contrary, the projects of Digital Tea House proved that it is possible to innovate
and offer new interpretations of a long established tradition and aesthetic.
Fig. 7 The project in dialogue with the context through a respectful design based on an optimised
use of structure and material
5 A Pavillon for Archeologists in Agrigento, Italy, 2013
“Architecture X Archaeology” is a co-joint workshop held in 2013 in the Valle dei

Templi Archaeological Park in Agrigento, Sicily (Italy). Tokyo University, Politec-
nico di Milano and Università di Palermo explored the design and construction of
lightweight, temporary structures to shelter archaeological sites and excavation works
otherwise exposed to the weather. This peculiar goal emerged from discussions with
the Valle dei Templi Park agency, that highlighted the need of improved touristic
infrastructures in the area whilst, at the same time, respecting the landscape and the
archaeological remains and providing better working conditions for the archaeolo-
gists on site.
The introduction of additional elements in the landscape proved challenging,
because—even if the shelter itself is small—it inevitably transforms the site, altering
the relationship between ancient remains and surrounding landscape (Tusa 1996).
The project then attempted a dialogue with the context through the design of a
respectful architecture based on an optimised use of structure and material (Fig. 7).
The workshop was divided into two parts: a preliminary part, 45 days duration,
took place at each university. The second phase, 7 days long, took place on site.
The preliminary part introduced issues related to the complexity of designing within
archaeological sites, applications of computational design, structural stability, and
practical solutions for quick physical materialization of ideas within limited time and
budget. The second part explored the culture and nature of Akragas—ancient Greek
name for Agrigento—which served as a pretext for further exploring alternative
designs for archaeological sites.5
5 The office LAPS Architecture was in charge of the organization of the workshop.
5.1 The Site and Its Needs
Located in the Southern Italian island of Sicily, the Valle dei Templi Archaeological
Park, covers an area of nearly 1300 ha, and preserves an extraordinary monumental
and natural heritage that includes the ruins of ancient Greek city of Akragas and the
surrounding landscape. Listed in 1997 as a Unesco World Heritage Site, the Valley
of Temples is home to one of the largest archaeological complexes of Mediterranean
area, immersed in an outstanding rural landscape punctuated by centuries old olive
and almond trees. The Park Agency promoted the “Architecture X Archaeology”
workshop as an occasion of reflection about the construction of shelters, with the
related technical aspects—anchoring to uneven ground, run-off and collection of rain
water, transportability—and the cultural issues implied by the inevitable transforma-
tion of the original site.
5.2 Molecular Shelter: A Temporary Shelter

for Archeologists
While respecting the local context, the Molecular Shelter design reinterprets a con-
cept borrowed from Japanese traditional culture. In fact, the shelter takes inspiration
from the bracket system “斗痴: To-Kyou:” found in traditional wood Japanese tem-
ples (Fig. 8), where the roof plays a prominent role. The “To-Kyou” derived from
Chinese architecture and was reinterpreted in a Japanese style in the medieval era
(VIII–XII century A.C.). It is a stacking structure composed of wood materials and it
enabled architects to design flexible plans, structures and compositions of the temples
in that era.
The eaves—which extend far beyond the walls, covering verandas—have an aes-
thetic and practical function because they protect the building by carrying the rain as
Fig. 8 The traditional Japanese bracket system To-Kyou 斗痴 and a detail of the Molecular Shelter
Fig. 9 Left: Joint detail. Center: 1:1 mock-up. Right: Drawing of To-Kyou 斗痴 bracket system
and its contemporary reinterpretation
far away as possible from its perimeter. The roof’s weight is supported by a simple
bracket system. The shelter meets the requests of the Park Agency that asked for
a shelter with a roof as large as possible, designed to carry the rain as far away as
possible from the excavations, movable from site to site, and with the possibility to
be set up again in different locations and with minimum surface area at the column
base. The structure was validated by engineer Jun Sato through structural analysis
software to optimize it to the thinnest possible size (Fig. 9).
The traditional bracketing system was reinterpreted in the design, yet, it keeps
an extreme essentiality both structurally and aesthetically, and present an intrinsic
elasticity, which lessens the impact of lateral forces by acting as a shock absorber. In
the traditional “To-Kyou” bracket system the position of columns is fixed, but with
the system developed for the Molecular Shelter it is possible to change it thanks to a
grid ceiling-structure (Fig. 10). The shape of the inclination of the roof—to evacuate
rainwater—is borrowed from Tempio della Concordia, so to have a direct reference
to existing forms in the landscape. The shelter—being made of multiple units—is
light, easy to fabricate and assemble in a reasonably short time (it takes one day and
four people), movable, adaptable to the site, and modular. The structure is designed to
allow archaeologists to move the shelter without any help and to hang their working
tools.
The whole structure of the shelter is composed by a joint system of 4 small
struts, with constant section, bound along both X and Y direction beams with M6
screws. The screws add resistance against rotational movements due to horizontal
external forces. Being modular, the system allows future extensions. If requested
by particular site conditions, columns can be positioned at different points of the
grid. The structure weighs around 100 kg, and can be easily moved by four people.
The shelter was anchored to the ground by 4 concrete blocks of 20 × 20 cm. The
prototype shelter is made of local pine trees, cut and assembled with 1500, 6 mm
Fig. 10 Left: Exploded axonometric. Right: The grid structure allows for flexibility
size screws. It took five days to cut and prepare the pieces and one day to assemble
them. The final cost of the pavilion amounted to 1600 e.
Concerning computational aspects, programming was developed with Rhino,
Grasshopper and Python. This made it possible to organize the exchange of informa-
tion and optimize the calculation which enables parametric changes. In particular,
the use of parametric software proved to be essential for testing different size and
arrays of the grid structure, structure thickness and weight, materials length and num-
ber of elements used. As for the fabrication, a hand-made easy fabrication system
was preferred, which demands only simple holes by drilling and screwing instead
of sophisticated machinery hard to find on site. Moreover, due to limited vehicu-
lar access to the archaeological areas, it has proved strategic to use transportable
lightweight tools instead of heavy machinery.
Conceived and finished in a period of less than 45 days, both computational
design and digital fabrication enabled speed and a consideration of the challenges to
be undertaken. The Japanese and Italian students interacted profitably, both with the
local authorities and with the park functionaries and architects. The union of these
synergies—university, public functionaries and craftsmanship—represents a unique
moment of exchange and growth which opened new paths to be explored.
6 Practice
Increasingly complex design requirements demand practices to acquire the ability

to cope with them. LAPS Architecture has been developing applied research as a
tool for experimenting on its own by conducting applied research and by including
students trained in complex design in its team. As written above these are the two
most viable methods to integrate more research into real work and prepare for the
changing form of practice in architecture (Samuel 2017). Experimenting has some
costs, however, in the case of LAPS Architecture, the investment in research is a
calculated risk that is bearing good results in terms of integration of research into
practice and it defines the identity of the office. Issues such as budget limitations,
building code regulations, technical constraints, durability, expertise and machines
used by the construction companies involved in the projects are decisive for building
an architecture with an advanced level of complexity (Dye 2014).
This section of the chapter shows how the use of advanced digital modelling
solutions have been translated into solutions that respect all the constraints of building
code and different parameters such as structural soundness, fire risk prevention,
standard assembly logic, certified materials, budget and time constraints.
Architecture practices experience a paradox: to work they need to build projects, to
build projects they need to have already some built references, to have references they
need to build. But for a young architect there are not many possibilities to build since
the practice does not have references. One solution is to design and build temporary
installations which is a good strategy to test ideas, materials, etc. But once a practice
gets a project, usually, there is no time to experiment new solutions, because the
different phases of a project are very strictly regulated by public rules regarding the
due date for delivery i.e. for a project such as the media library discussed later in this
chapter (a 1000 m2 project), the schematic design was limited to 3 weeks, the design
development to 4 weeks, construction documents to 4 weeks. This means that there
is a very limited amount of time to experiment, test, analyse new ideas: basically,
quite often, an architect usually prefers to rely on concepts, materials, technologies
he/she already knows which are possibly certified and normed. All delays are subject
to economic penalties that an architecture office does not wish to pay just for the sake
of doing an experimental design. Basically, to avoid risk is to use standard protocols,
forms and materials and one of the possibilities for small size architecture practices
is to use ongoing, personal research and try to integrate the knowledge developed
when a real possibility arises. It does not mean that research previously developed is
adjusted randomly to a project but the sole fact of having a research mindset helps to
develop potential solutions faster. For LAPS Architecture office investing in applied
research (i.e. the production of X.me modular system) represented a way to diversify
its business and define its identity and style. Architectural constructions are often
quick in terms of drawing phases and slow in construction, impeding an immediate
feedback between ideas and reality. On the contrary, small projects or installations
offer architects the possibility to test ideas more immediately. The strategy used by the
office is clear: instead of investing only in medium or large size architectural projects,
the partners of the office decided to stimulate potential economic gain from interior
and product design. Invited by a product design company to think about an innovative
modular system for interior design, the office decided to focus on the development
of a system that used a simple vocabulary of repetitive customizable elements with
wood as the main material. Budget allowing, there are some legal protocols such as
the so called “ATEx”6 or “Appreciation of Technical Experimentation” that transfer

the research risks onto the contractor and make it easier for architects to promote
innovation and changes in regulations. In French Building Code regulation there are
several classes of public building that demand different types of wood according to
different safety and risk factors such as the number of people allowed in a building or
the number of floors. These factors define the class of the wood for structural loads
and fire prevention. For structures and facades there are different classes of wood
allowed. In the past years there has been an increasing interest in wood construction
and accordingly new rules and limitations have been updated to current demands
for safety and risk preventions. For those architecture offices integrating research
into practice is even more challenging because they have to manoeuvre within this
framework in constant evolution.
7 X.me System
The X.me project is an investigation on structure, assembly and fabrication pro-

cess in order to produce a customisable modular system for interior design and
furniture. Developed firstly as academic research, the X.me system evolved into a
product thanks to several prototypes built in real contexts. Resulting in a redefinition
of the classic chain designer-maker-distribution-client, the X.me system is modu-
lar and flexible and simple to configure (Liotta 2016). The system is self standing
and presents a solid structure, it is easy to assemble and disassemble resulting in
a reversible and recyclable system. A grid pattern has been extruded to create cus-
tomized elements that can be adapted to different contexts. The cells of the grid that
emerge are folded and then rounded for different interior design needs such as parti-
tions, bookshelves, sitting spaces, benches and tables. This process enables their final
form to suggest the rule of composition without explicitly revealing it thus resulting
organic in its aspect. The system could be composed of customised panels made
of different materials such as wood, sandwich MDF (Medium-density fibreboard)
or PVC (Fig. 11). Horizontal or diagonally cut, the panels are then combined and
assembled manually.
The X.me focuses on the possibility of applying a grid pattern for generating
three dimensional forms through parametric design. Generated by an orthogonal
grid which is rotated by 45 degrees and made of customized elements cut by a 4
axis CNC machine, the X.me system is assembled by using extruded aluminium
joints (Fig. 12). This simple system allows an infinity of different shapes to be
created therefore making the product highly adaptable to diverse contexts, even
6 ATEx is the acronym of “Appreciation Technique d’Experimentation”. It is a quick technical

evaluation procedure formulated by a group of experts on any product, process or equipment not
yet the subject of a technical certification and whose development requires experimental use on one
or more sites. There are three classes of ATEX according to the degree of innovation.
Fig. 11 A configuration of
the X.me system
though some improvements are still needed. An online configurator allows for real
time customisation, shopping and delivery.
7.1 Social and Technological Innovation
The X.me system is innovative from both technological and social aspects. The
technological innovation relies on the reduction of the system to its bare essentials
which allows everyone to intuitively design their own wood furniture and assemble
it without using nails, bolts or glue. Moreover, because of its freedom of form, the
system easily adapts to different contexts. Interestingly, the orthogonal rigidity of
the grid patterns permits a great degree of diversification on the Z axis.
The social and technological innovation relies on two aspects: the first concerns
the maker and the second the customers. The X.me is produced by MYOP, an Italian
company which has confederated several makers and artisans who—because of the
recent economic crisis—have had to greatly reduce their production and consequently
decrease part of their labour force. The company is a start-up that was created as an
on-line platform to facilitate the meeting between artisans and designers. The second
innovative aspect relies on the fact that the platform offers a different approach
Fig. 12 A detail of the

assembly system
to design and gives the possibility to the customers to design and order online a
customised solution of different products directly from the maker thus reducing the
cost of production, stockage, promotion and transport.
The development of the X.me system is here shown through the analysis of two
installations—at the Norman Castle of Favara, Italy, and at Italian Cultural Institute
in Paris. Besides showing different formal solutions, flexibility and adaptability, at
the same time the two installations underline some limits of the system. While it has
proved to be very intuitive to design and simple to assemble when composed of a
limited number of elements, it still needs improvements for large scale projects due to
the difficulty to connect the upper part of the system. The difficulty is due to the fact
that even slightly uneven floors affect the assembly by deforming the grid through
its weight. Because of the weight, the geometry of the grid varies infinitesimally
but enough to impede a smooth assembly of the system. To solve this issue, the
elements were cut thinner than the aluminum joint size, however some elements
needed to be hammered to slide into the joints. This proves to be a major problem
for commercialization, because only specialized workers can assembly the system.
Another solution is to send the furniture already assembled, but this will increase the
shipment costs because it will occupy more space than when it is sent disassembled.
Fig. 13 The X.me inserted in the sensitive historical context of the Norman Castle of Favara
7.2 Info Point, Bookshop and Ticket Booth at Norman Castle

in Favara, Italy, 2013
The project for the info point center of the Norman Castle of Favara was the first
translation of research on digital wood into an X.me build project. Several mockups
were built to test the resistance of the structure, different materials, dimensions of the
elements, different aluminum connectors and assembly. One of the first observations
concerned the use of wood. In fact, because of the unpredictability of the direction
of its fibres wood proved to be the least performative material.
After testing different materials such as wood, plywood and MDF, the last was
chosen. MDF—being a composite wood material—presents an homogeneous fibre
direction which is indispensable for the elements to easily slide in and out from the
aluminium connector. In this first project, the modular system integrated different
programs such as tables, benches, arches for partitions and shelves.
Due to the fact that the Norman Castle is listed as an historical building, the walls
could not be touched in any way. There the X.me system was particularly adaptable
because it is a self standing structural system that does not need to be fixed to the wall.
Because the shelves are nearly 5 meters high, the shelves needed a large base. For
functional reasons, the designers have preferred to enlarge the base of the modular
system by integrating in the design a bench and a desk (Fig. 13). Both were extruded
by 100 cm to help stabilize the entire system. The space presented an uneven floor
which did not enable a smooth assembly of the upper part of the system. Due to the
proximity from the factory, the system was partially assembled and then transported
on site.
Fig. 14 The modular system integrate different programs such as tables, benches, arches for par-
titions and shelves
7.3 Children Library in Paris, France, 2014
Situated inside XVIII century’s Hotel de Gallifet in Paris, the Italian Cultural Institute
hosts the Italo Calvino Library which has been extended to include a space dedicated
to childhood designed by LAPS Architecture office. Opened at the end of the ‘50s,
the Italo Calvino Library owns over 50,000 volumes, resulting in the largest library
among the network of the Italian cultural institutes. The Children Library offers over
1.650 books for kids and teenagers. The X.me system was selected as a bookshelf
system meant to characterize a space where children can read and discover. Besides its
modularity and flexibility, X.me was chosen because it can be assembled without glue
or nails and it can be entirely disassembled proving to be reversible and recyclable
(Fig. 14).
The system was designed to fit the space between the columns and arches of
the library (Fig. 15). Being an historical construction dating to the 17th century,
the walls and the pavements presented uneven surfaces. These irregularities affected
the rapidity of the assembly. The X.me envelops the columns and its back part
(the one facing the wall) presents different depths which follow the geometrical
variations of the colonnades. Besides testing for the first time the variations of the
side facing the walls, one more novelty of the design relates to the fact that some
horizontal planes were inserted into the structure. While in the X.me for the Norman
Castle of Favara, the horizontal structure was inserted into a single cell, here, the
design included a continuous horizontal plan covering four cells. To have such a plan
required customized cut coinciding to the intersection with the aluminum connector.
This time thanks to the limited height of the structure (less than 200 cm), there was
no major problem for the assembly of the system. However, the horizontal partitions
Fig. 15 The modular system can bear heavy loads
proved to not be efficient because the minimum variation of the uneven floor produced
slight distortions into the grid geometry enough to impede a smooth sliding in of the
horizontal plans. To solve this issue, the successively installed horizontal plans were
subdivided into single cell length units. They form a continuous plan interrupted only
by the joining lines. They resulted easier to assemble and aesthetically there was no
significant difference (Fig. 16).
In conclusion, these two installations served to test as many issues as possible
regarding functions, program, materials. Thanks to these prototypes, the knowledge
collected helped to finalise an online configurator and made evident that the X.me sys-
tem was ready to be commercialized. The X.me system can be designed by potential
customers on the website of MYOP company. The configurator uses WebGL soft-
ware, it is freely accessible online and allows the customer to chose the number
of cells, their depth and materials. It is written in Java Script and PHP. Customers
are free to configure their own design, to get an immediate price and delivery time.
Once the order has been finalized, the drawings are processed through Rhinos and
Grasshopper and then exported and arranged on panels and ready to cut by CNC
milling machines. What started as a research performed at a small size architecture
office turned into a system that today is on the market. These built references were
included in Laps Architecture portfolio and were fundamental in the selection for
the media library of the Quay Branly Museum discussed below.
Fig. 16 The modular system can incarnate different shapes
8 Recreational Centre for Children and Young People

in Canteleu, Normandy, France, 2016
The project for the recreational centre for children and young people in Canteleu is
located in a depressed area of Rouen suburbs. The rate of unemployment of this area
is around 25% (two times the national rate) with a presence of nearly 60% of social
housing built in concrete. The mayor of the city intended to give to the city a new
architecture that would renew part of the infrastructure and also represent a symbol of
hope for the city. The authors believe that once people meet beauty they do not want
to go back to ugliness anymore. The mayor required that the recreational centre for
children and young people (1) included a certain attention to aesthetics as a positive
message for the youth of the city, and (2) that the building had to be in wood for it is
a sustainable material, and thus bring a positive message in opposition to the existent
concrete landscape of the town. The project includes in its architecture these two
elements: a building which is recognisable thanks to a singular wood facade shaped
in form of a wave. The wave intentionally represents the energy and dynamism of
the youth of Canteleu. Today the local community refers to this new building as the
Wave (Fig. 17). The wooden wave is one of the possible results of a research on how
to generate an iconic architecture by mixing standard and non-standard elements
(respectively flat and curved lamellas) to reduce the costs of the project.
Fig. 17 The modular system can bear heavy loads and can incarnate different shapes
8.1 The Site and Its Needs
Located in the north of France, Canteleu is a small town of forestry and light industry
situated by the banks of the river Seine, just 3 km northwest and over the river from the
centre of Rouen. Canteleu has some historical monuments (such as the chapel) and
a cultural center (the Flaubert museum) and the rest is composed primarily of social
housing totalling a 57% of the total construction (which ranks Canteleu among the
first 20 cities in France according to % of social housing). In an attempt to revitalize
the services of the city and to change the reputation of Canteleu, the mayor of the
city since 2014 is re-investing the taxes of the city in the construction of cultural and
educational infrastructures. The territory of the municipality is 70% occupied by the
forest. With more than 4000 ha, the forest of Roumare is a real green lung for the
Rouen area. Installed on a poor soil consisting partly of silt and flint clay in the North
and alluvial sands in the South, this forest has developed a set of hardwoods such as
oak or beech, and a coniferous forest for forestry.7
8.2 Recreational Center
The design of the recreational centre connects with the local strong presence of wood.
Whilst the structure of the building is made in cross laminated timber, the facade
7 Siteweb de la Ville De Canteleu http://www.ville-canteleu.fr/.

Fig. 18 A detail of the facade
is made in pre-coated pine wood usable for the outdoor (Fig. 19). The facade was
partially pre-cut at factory and assembled on site by carpenters (Fig. 18).
The design includes 40 mm wide lamelles of wood which alternate two flats one
and a curved one. The distance between the lamellas is 10 mm, which is the maximum
distance allowed by the so called DTU-Dossier Technique Unifié (Unified Technical
Regulation). The original project presented a distance between lamellas of 40 mm
with 40 mm wide lamellas. The idea behind the design was to minimize the use of
wood by having a certain distance between lamellas. The fact that new regulations
for public building facades was changed between the competition phase and the
construction phase demanded the design to be changed to comply with the new
regulation.
Fig. 19 The technical

drawing with 150 mm thick
curved lamellas
The design of the wave on the facade was defined by some spline lines that
change direction because of the presence of some openings in the facade. For the
construction, initially only the flat lamellas were positioned and fixed on a horizontal
substructure composed of 4 battens distanced 1 m from one another. Subsequently,
the curved lamellas were nailed to the substructure. Originally the curved lamellas
were supposed to be more than 150 mm deep in the highest point, but there are no
standard nails long enough to fix such a deep element on the supporting substructure.
Therefore, the thickness of the lamellas had to be adjusted to 50 mm. This affected
the initial design making it more flat (Fig. 19). In the DTU there were no indications
concerning the depth of wood in facade, however the design changed for technical
constructive reasons. The lamellas were cut on site with a portable CNC machine
and sanded by hands before being put in place.
Concerning computational aspects, programming was developed with Cinema 4D
and Rhino. This did not help the exchange of information and optimization with the
construction company, since they only used Cad software. The use of parametric
software proved to be good for testing different size and arrays of wave design,
materials length, thickness and number of elements used. As for the fabrication,
a mix of CNC mill cut elements and hand-made fabrication was used. The overall
result is acceptable but it respects the original design only partially. In fact the curved
lamellas were resized and since they were hand cut on site do not exactly correspond
to the design.
9 Media Library of Quai Branly Museum, Paris, 2018
The project for the Media Library of the Quai Branly museum in Paris is an investi-
gation into wood, transparencies and the production of a sense of intimacy within a
public place. The project aims on one hand to update the aesthetics, ergonomics and
functionality of an emblematic place of research in Paris. On the other hand, the new
project resonates with the existing project of Jean Nouvel by using the same material
(amaranth wood) and a reinterpretation of the concept of selective dematerialization.
9.1 Context and Program
Strongly desired by former French president Jacques Chirac, a place of pilgrimage

for prehistoric art, anthropology and ethnological studies, the Quai Branly Museum,
Paris was built in 2007 and designed by Jean Nouvel.8 In 2018, the media library
was renovated by an intervention of LAPS Architecture. A center of excellence
in ethnology, the media library offers a wide selection of works on the arts and
civilizations of Africa, Oceania, Asia and the Americas and preserves patrimonial
documents of high historical interest: photographs, archives, funds and collections.
Located on the 6th floor, opening entirely on a panoramic terrace, the new project
responds to the need to reorganize the visual identity, ergonomics and functionality
of the space. The project of LAPS architecture includes the reorganization of an
area of 1000 m2 . and it aims at improving the workplaces of the employees and to
8 Jean Nouvel’s writings on the Quai Branly Museum helps to understand the design concept: “It
is a museum built around a collection. Where everything is done to bring to light the emergence of
the emotion carried by the “first object”; where everything is done, at the same time, to protect it
from the light and to capture the rare ray of sun indispensable to the vibration, to the installation
of spiritualities. It is a place marked by the symbols of the forest, the river, and the obsessions of
death and oblivion. It is a busy, inhabited place, one where the ancestral spirits of men meet who,
discovering the human condition, invent gods and beliefs. It’s a unique and strange place. Poetic
and disturbing. Building it can only be done by challenging the expression of our current Western
contingencies. Eliminate the structures, the pipes, the “joineries” of facade, the emergency stairs,
the railings, the false ceilings, the projectors, the bases, the showcases… If their function must
remain at least that they disappear from our sight and our consciousness, that they fade before the
sacred objects to authorize communion with the holy. (…) Only the result counts: the material at
times seems to disappear, one has the impression that the museum is a simple shelter without facade,
in a forest. When dematerialization meets the expression of signs, it becomes selective.”
better accommodate the media community—composed by researchers and scholars.

In the preliminary meetings with the client, gave the architects a great freedom to
manoeuvre within certain limits of budget and consistency with the existing project.
9.2 Main Concept of the New Project
LAPS Architecture office proposed a concept that is in dialogue with the existing
museum design both for material continuity and for the so called strategy of selective
dematerialization. The amaranth9 wood already used by Jean Nouvel in the museum
was also used for the media library. Even though it is more expensive than other types
of wood, the use of amaranth—with its rich dark shades ranging from bordeaux to
violet—is particularly good for three reasons: it is an hard wood suitable for public
spaces, it favors concentration and it induces a respect of silence. As for the design,
the use of pixels of wood characterizes the interior design by taking inspiration from
the existent design of the Museum southern facade characterized by an alternation of
rectangular panels of different colors. As a general principle, the new project aimed
to establish a continuity with the existent atmosphere through dematerialization. In
this sense, the new project is characterized by a pixelated texture which plays with
notions of transparency and visual effects. This expedient allows to feel the presence
of others but at the same time preserve the intimacy and concentration required by a
research center such as the media library.
The redesign of the interior space includes five zones: entrance lobby, magazines
consultation space, cabins for multimedia screening area, a special area for consulting
rare archives, and several scattered individual, semi-individual and group research
rooms. The idea is that the whole media library becomes not only a place for con-
sulting, learning and reading but also a place that offers space for research meetings
and seminars. Besides a general redesign of the spaces, the mission of the architects
included the furniture design of tables, chairs, partitions, rest seats, benches, video
and audio booths (Figs. 20, 21 and 22). Amaranth wood was used in all the above
mentioned objects and space.
The wood pixels were chosen for two reasons: acoustic and visual comfort. One of
the requests of the program was to improve the acoustic performances of the library. In
this sense the pixels positioned on vertical walls contribute to make non-continuous
surfaces that help to improve the general acoustic of the spaces for working and
reading. The acoustic tests were performed with dedicated softwares. As for the
visual comfort, in order to re-define the notion of “selective dematerialization”, a
fundamental aspect to solve about the partitions was to control a certain balance
9 Amaranth is an extremely dense and water-resistant wood. It is ranked as one of the hardest and
stiffest of the woods in the world. It is so durable that it can be used in applications that require
toughness, such as truck decking. The trees are prized for their beautiful heartwood which, when cut,
quickly turns from a light brown to a rich purple color. Exposure to ultraviolet (UV) light darkens
the wood to a brown color with a slight hue of the original purple. This effect can be minimized
with a finish containing a UV inhibitor.
Fig. 20 The entrance of the media library present diagonal surface with a certain degree of com-
plexity
Fig. 21 A detail of a vertical

surface
Fig. 22 The connection of

different converging planes
between intimacy and transparency (Fig. 23). The wood pixels were a perfect solution
to make the users feel alone yet together with other people (Fig. 25). The pixels
contributes to give a new identity to the media library and their disposition was
manipulated through parametric design software such as Grasshopper. After several
tests on mock-ups, a pattern was found and repeated on all the surfaces populated
by the pixels. Interestingly, even if repeated, the pattern does not reveal the rules of
composition resulting totally casual. The pixels have equal size 20 × 20 mm but two
different lengths: 100, 200 mm. The composition is quite simple, made by alternating
a line of pixels with a continuous wood element of the same size (Fig. 24). The
pixelated partitions are made by doubling the pixels on the same line for producing
a selective effect of depth. The pixels are inserted and positioned between the fixed
elements of wood. Translated into a real product by a furniture company that uses
CNC routing, the project is 100% faithful to the original design. This was possible
thanks to a smooth dialogue between architects and contractors.
Fig. 23 A dematerialized
partition
To conclude, the new media library, thanks to this intervention by LAPS Archi-
tecture, offers spaces that reflect the uniqueness of an emblematic place of research
within the Quai Branly Museum. If, on the one hand design freedom was given to the
architects, on the other hand the functional program presented very specific requests
related both to the functioning of the media library and to the ergonomic needs
required by the program. The use of computational design and digital fabrication of
amaranth wood proved to be the right choice to respond to the program.
10 Conclusions
The knowledge about wood developed at different universities through digital fab-
rication laboratories (Liotta and Belfiore 2012) and by teaching architecture proved
to be effective but within certain limitations when used for real constructions. In
fact, translating the freedom of building temporary architectures—which usually is
one of the prerogatives of architectural design at university—into wood architecture
that respects all the constraints of real construction proves to be difficult because of
several parameters that must be taken into consideration.
Fig. 24 Technical drawings of a pixelated partition
The chapter describes several experiences where innovative ideas have been
applied both to pavilions in an academic environment and in real constructions in
Japan, Italy and France. The discussed case studies offer a collection of issues that
researchers and architects face in architectural design either in an academic or a
professional environment.
The use of parametric design can, at times, promote a tendency where its users
can easily produce forms too complex to control with little regard to issues not
only of structure and feasibility but also economy, society or culture, in part due
to a fascination with new forms (Schumacher 2009). The chapter makes clear that
constraints including structure, material, budget, time, assembly, site, and function
help avoid the risk of designs that are impossible to be realized. When parametric
design is integrated into a process within a framework of real-life constraints, its
advantages are beyond mere stylistic choices or visual effects. Often, as we are
“limited to the constraints of borrowed code, most architects today are forced to
squeeze their imagination through the prefabricated structures of existing protocols.
To a certain extent this situation has begun to change with the introduction of custom
scripting where designers are able to tweak the constraints of their tools” (Silver
2009).
Fig. 25 Selective dematerialization effect in the context
This chapter proves that (1) the democratisation of combined use of digital and
fabrication tools have made it possible to conceive and realize complex design with
limited budgets whilst respecting the building code, (2) the design of the projects here
discussed has evolved alongside technology, within a mutually beneficial relationship
(3) integrating research into practice is a way to try to make better buildings, that are
appropriate to theirs users, clients, context and time and (4) integrating research into
practice opens up to several new possibilities before unimaginable.
Acknowledgements I am deeply grateful to Sarah Beal for making it all richer in many several
ways.
I am in debt with Kengo Kuma for revealing to me a world of ideas and for showing me the tools
for making them real.
I am grateful to Fabienne Louyot, my work associate for researching and practising architecture
together.
Parts of Paragraph 4 were published in the following paper: Ko, K., Liotta, S. (2011), Decoding
Culture Parametrically: Tea House Case Studies, International Journal of Architectural Computing
(IJAC), issue 04, volume 9, Dec. 2011. pp. 325–338 Parts of Paragraph 5 were published in the
following paper: Imperadori, M. Liotta, S. et al. (2015), New Architectural solutions for Archeologi-
cal sites. “Bio-based and Bio-inspired Environmentally Compatible Structures” by the International
Association for Shell and Spatial Structures (IASS), Tokyo, 2015 Parts of Paragraph 7 were pub-
lished on the following paper Liotta, S. 2016, Using a pattern as a 3D generator for producing
a modular and flexible system for architecture in “Chanllenges For Technology Innovation - An
Agenda for the Future”, Taylor & Francis.
Credits workshop Digital Tea House
Instructors: Philip Anzalone, Toru Hasegawa, Brigette Borders (GSAPP Columbia University);
Kengo Kuma, Yusuke Obuchi, Tomohiro Tachi, Salvator-John Liotta, Kaon Ko, Yoshii Nahoko
(University of Tokyo); Keisuke Toyoda, Yusuke Oono, Motoki Yamamoto, Alex Knezo (Noiz
Architects) Participants: Taichi Kuma, Anna Braverman, Kazami Furukawa, Akinori Hamada,
Yuta Ito, Nikola Nikolovski, Hiroyuki Tanaka “Nami-no-Ma”; Tan Toon Cheng, David Jenny,
Taichi Kuma, Reiko Nishiyama, Shuhei Tanaka, Haruka Tomoeda, Cui Xuan, “130008252010”.
Images credits: Taichi Kuma, Yuta Ito, Alessio Guarino.
Credits workshop Architecture x Archeology
Instructors: Marco Imperadori, Andrea Vanossi (Milan Politecnico), Salvator-John A. Liotta,
Yuta Ito (University of Tokyo) Fausta Occhipinti (Associazione Giardini in Campo) Students:
Eugene Kiang, Kaoru Yamaoka, Kosuke Nakakura, Salvatore Vinciguerra, Luisa Lo Faro, Antonio
Carmisano, Tommaso D’Angelo, Michele Piccolo, Federico Macedonio, Marina Tedesco. Scien-
tific Advisors: Kengo Kuma, Marco Imperadori, Salvator-John A. Liotta, Jun Sato. Organization:
LAPS Architecture, Paris.
Credits X.me project:
X.me System Design: LAPS Architecture, Salvator-John A. Liotta, Fabienne Louyot
Product Engineering: MYOP s.r.l. Riccardo Scibetta
Credits Recreational Youth Center:
Architectural Design: LAPS Architecture, Salvator-John A. Liotta, Fabienne Louyot
Credits Media Library Quai Branly:
Architectural Design: LAPS Architecture, Salvator-John A. Liotta, Fabienne Louyot
References
De Mente B (2006) Elements of Japanese design. Tuttle Publishing, Tokyo

Dye A (2014) Architects and research based knowledge. RIBA, London
Fischer V, Schneider U (eds) (2008) Kengo Kuma: breathing architecture. Birkhauser, Basel
Hensel M, Nilsson F (2016) The changing shape of practice: integrating research and design in
architecture. Routledge, London
Iwamoto L (2009) Digital fabrications: architectural and material techniques. Architectural Press,
Princeton
Ko K, Liotta SJ (2011) Decoding culture parametrically: tea house case studies. Int J Archit Comput
(IJAC) 9(04):325–338
Kuma K (2011) The significance of tea house in Japanese architectural education
Liotta S (2016) Using a pattern as a 3D generator for producing a modular and flexible system
for architecture. In: Challenges for technology innovation—an agenda for the future. Taylor &
Francis, pp 115–119
Liotta SJ, Belfiore M (eds) (2012) Patterns and layering: Japanese spatial culture, nature and archi-
tecture. Gestalten, Berlin
Matsukawa S (2007) What is algorithmic thinking? 10 + 1(48):155–160 (translated by author)
Moussavi F, Kubo M (2006) The function of ornament. Actar, Barcelona
Nishi K, Hozumi K (1983) What is Japanese architecture: a survey of traditional Japanese archi-
tecture. Kodansha International, Tokyo
Okakura K (2007) The book of tea. Stone Bridge Press, Berkley (first edition 1906)
Samuel F (2017) Supporting research in practice. J Archit 22(1):4–10
Schumacher P (2009) Parametric design. In: The patterns of architecture. Wiley, London
Silver M (2009) Pattern deposition from script to application. Archit Des 79(6):94–99
Tusa S (1996) Aree archeologiche e compatibilità produttiva, in proceedings of “Il paesaggio della
Valle dei Templi”. Agrigento, pp 15–26
Digital Design Thinking in Architectural
Education Testing Idea-Driven
and Science-Driven Design Processes
Towards Researching Polymer/Wood
Composite Structures
Andrei Gheorghe
Abstract The ongoing introduction of digital design tools in the architectural pro-
fession does not only influence the architectural design product, but also affects the
underlying design process. A “digital design process” taking full advantage of the
potential of digital design tools is different and can offer more potential than a paper
and pencil based design development process. Zaero-Polo (2001) notes, the design
process itself becomes more interesting than the design idea. This process aims to
explore the creative potential of machines and incorporate knowledge from multiple
disciplines, rather then apply pre-conceived design metaphors or design ideas. In
order to differentiate between traditional and new digital design development pro-
cesses, a distinction between idea-driven design processes and science-drive design
processes is introduced. This research argues that digital technologies help to foster
collaboration between architecture and related disciplines (e.g. engineering) in archi-
tectural education. Innovative digital architectural design does not only use digital
skills in the design process, but also calls for specific digital design and teaching
methods. This research combines digital architectural design processes with digital
design fabrication technology while teaching experimental workshops with students
at different international academic institutions. The idea is to explore the potentials
of this combined digital learning chain to initiate innovative architectural implemen-
tations.
Keywords Wood joint systems · Parametric design and fabrication strategies

Complex timber structures · Optimization of wood architectures · Wood design
Digital design education
A. Gheorghe (B)
Harvard University, Cambridge, USA
e-mail: a.gheorghe@uni-ak.ac.at
A. Gheorghe
University of Applied Arts Vienna, Vienna, Austria

1356 A. Gheorghe
1 Introduction
The ongoing introduction of digital design tools in the architectural profession does
not only influence the architectural design product, but also affects the underlying
design process. A digital design process taking full advantage of the potential of
digital design tools is different and can offer more potential than a paper and pencil
based design development process.
In order to differentiate between traditional and new digital design development
processes, the notion of “digital design thinking” and a distinction between idea-
driven design processes and science-drive design processes is introduced. It is argued
that digital technologies help to foster collaboration between architecture and related
disciplines (i.e. engineering) in architectural education. The widespread use of digital
technologies helps to bring related disciplines closer together, also because these
share the same digital tools and a common communication language. Architecture
curriculum will need to adapt to theses developments in future (Mark et al. 2001).
Historically, a wide range of tools has been used in the process of architectural
design. In the course of history, these evolved and lead to different professions with
separate modes of operation and evaluation. Mastering these tools is naturally essen-
tial in the critical process of design development. Much more importantly, these
tools are equally relevant in how the architectural design process is taught to the
next generation of architectural students. The pedagogical transfer is on one hand
defined by teaching the skills of use for the relevant tools, but is also dependent on
the mediation of a critical design process.
Dana Cuff lucidly describes the historic emergence of the relevant education pat-
terns (Cuff 1995) until the well-known current state: the setting of problems at the
start of the educational design process, the studio setting emulating the professional
environment, the graphic development of the design through sketches, the influential
pedagogical relationship with the studio master and the well-accepted evaluation
mode by jury review. Schön (1984, 1985) defines the architecture studio as an envi-
ronment to learn to make or perform, the respective pedagogy aims for the acquisition
of artistry through coaching and practice. He describes the design process in an archi-
tecture studio as “Reflection in Action”, an “artistry with uncertainty, uniqueness and
value-conflict in many domains of architectural practice.”
In the contemporary profession the use of IO tools (e.g. computer mouse, AR/VR
tools) and analytic algorithms (e.g. computer software) became increasingly relevant.
The application of design tools influences the education of future architects, because
in current education models the transfer of design knowhow happens by learning the
handling of design tools. However, the success of design tools cannot be measured
based on the aesthetic and formal properties of the design product itself, but on the
lucidity and consistency of the underlying design principles. Many master architects
follow this approach of prioritizing the design process over the design outcome.
As an example, Pelkonen (1999) describes Alvar Alto’s design process as a playful
discovery with autonomy in use of form and material. Alto emphasized the design
process, and not the formal outcome, with the primary challenge to keep design
Digital Design Thinking in Architectural Education 1357
control loose enough to allow for intuition and aesthetic talent to come into play
without having any pre-conceptions on the possible design solutions. Andreas Ruby
analyzed the work process of the Austrian architects Ernst J. Fuchs, who avoided
using a specific image as the starting point of the design process, as this would
pre-conceive and fix in advance the result of this process. Fuchs is furthermore
preoccupied with the initiation of a design development process, which would define
the formal qualities of the project in a continuous sequence (Ruby 1999).
The principles of digital design thinking are not new phenomena and are not
related to, nor do they require computer software (Carpo 2009). For the scope of this
work, digital design thinking is defined as a systematic way of designing by depen-
dencies and defining relations between architectural elements. Vitruvius described
in his De Architectura—Ten Books on Architecture (1999) the principles of archi-
tectural design of his time. Interestingly, the design elements and their respective
relationships were mostly not drawn or sketched, but described in written form. This
happened in a systematic order, explaining relationships and hierarchies. For exam-
ple, he defined heights of columns according to respective widths and related it to
other measurements. To quote another example of what we would call today design-
ing by parameters (Reas and Fry 2014), he explained the ideal grid positioning and
layout for a new city according to prevailing wind directions. These approaches are
very similar to the way a modern computer design algorithm would be conceived
and programmed.
It is furthermore interesting to see that along history, certain oeuvres of art devel-
oped mainly because of mathematical, geometrical or astronomical innovations in
science. Arabic architecture and art (non-figurative ornamentation) emerged and
developed because figurative drawings were not allowed. Art was derived from sci-
ence, main design tool of that period was the compass. To mention another example,
gothic architecture experimented with the flow of forces to create unprecedented
structures. Although all this happened intuitively through a tedious and dangerous
trial and error process, this systematic process of defining parameters and techni-
cal relationships displays a very concise design thinking process. The design idea
was often developed through defining and testing a material related, intuitive tech-
nological process, while incorporating the latest scientific findings in geometrical,
mathematical or astronomical disciplines. At the same time, a strong vision, or a spa-
tial idea was simultaneously developed, concepts of lightness, religious, immersive
sacral spaces was developed and tested in parallel.
This context sets the background to initiate a series to test with architecture stu-
dents digital thinking principles combined with the creation of full scale testing of
hybrid wood and composite wood/polymer concrete full scale applications.
2 Qualitative Research Approach
Qualitative results of the design experiments in this empirical research are obtained
through a mix of methods of assessments such as formal interviews with students
1358 A. Gheorghe
during and after the design development, and equally through discussions, evalua-
tions of student’s work and observation of progress (Oxman 1999). The aim consists
in observing “changes occurring in student’s thinking about design and the growing
skill to deal with the complexities of design thinking and the dynamic progress of
their performance”, and identifying phenomena such as depth of understanding of
the task in the development of development of modeling skills, the efficiency in
modeling representation.
Oxman describe the importance of the learning environment. She recognizes
(1999), that the classroom may provide an adequate spatial situation for directed
analytical education in problem solving, while the studio is traditionally viewed as
providing the creative ambiance of design making. Consequently, she used in her
extensive research on digital design education the computer laboratory as a design-
learning environment, coining the term “Cognitive Design Media”.
However, Oxman suggests that special design learning environments must be
developed to enhance and supplement digital design learning education. In recent
years, the multidisciplinary learning environment of an open-field experimental
research pavilion emerged as a digital design-learning environment for digital design
thinking.
2.1 Idea-Driven and Science-Driven Design Processes
When looking at the modus operandi of the architectural design development

throughout architecture history, two main components can be identified. A certain
design process (1) is put in place to arrive to a certain architectural outcome (2).
Mediator and facilitator between these two components are the architectural design
tools, which changed significantly in the past and will surely change in future. This
research argues that this mediation process of the design tools balancing the design
process and the architectural product can work in both directions. In many cases
throughout history, an architectural outcome or design goal was defined first, and
then a corresponding design process was developed to fit this formal vision. This
approach is defined here as an idea-driven design process. In other cases, primarily
a design process was developed based on systematic project parameters, and subse-
quently the formal outcome emerged as a fit after employing this work pipeline. This
research defines this as a science-driven design process.
science-driven design process: process > leads to > outcome
versus
idea-driven design process: outcome > leads to > process
Whereas in history, either one of these processes was the predominant design
approach, contemporary digital tools have the potential of integrating the advantages
of both approaches.
2.2 Idea-Driven Design Process
An ideal simplification of this case is the following approach: a certain architectural

formal vision, or architectural design product is firstly defined, and secondly the
architectural process is implemented to achieve this goal.
In many cases, although not always, this process tends to become a top-down
design implementation process, in any case a large amount of architectural design
experience and intuition are required for this method to be efficient or successful.
This method has the tendency to be a linear project development with fixed work
and design hierarchies, however, for many projects developed through this strategy,
a strong and unique architectural language is developed, which is identifiable either
by the time period or by the specific designer.
2.3 Science-Driven Design Process
An ideal simplification of this case is the following approach: the design thinking
process is developed first without any preconceived formal vision, and subsequently
the aesthetic outcome is derived from this process.
Many projects employing this method are bottom up design development process,
involving a range of scientific methods for evaluating and adapting design decisions.
Features of this process are interdisciplinary thinking, and a step-by-step systematic
development with feedback loops and evaluating steps.
2.4 Empirical Studies—Experimental Digital Design

Workshops
Menges (2011) describes form as a performative, dynamic system that emerges from
principles based on the behavior of material, methods of manipulation and assembly,
and interaction with environment. Researching a critical Digital Architectural Design
process therefore does not mean to research the from of the generated design, but the
emerging potential of its application. This means that the organizational capacities
of digitally generated content are able to improve the efficiency, environmental or
structural performances of the architectural design. We should not merely judge
geometrical structures based on their formal beauty, but have to take their possible
advantages in daily, pragmatic use into account. The main scope of this research is
oriented towards the design process.
As a first test ground of this research, digital methods were used to create an
experimental seminar/workshop series organized and operated in cooperation with
international institutions. These workshops were organized between 2011 and 2017
and aimed at offering and testing digital design concepts through the combined
1360 A. Gheorghe
application of idea and science driven design processes. The workshops focused
towards testing a specific pedagogy and digital education aspects. Various digital
design tools and fabrication strategies were tested. The workshops integrated differ-
ent digital design techniques in combination with structural expertise teaching, and
the output production contained full-scale installations.
The aim of the seminars was to test the methods on how to teach simultaneously
architectural design and digital technologies while fostering critical thinking (comb-
ing theory, design and technology). The participating students were encouraged to
use a wide range of analog and digital media in a creative way, such as parametric
3D models, renderings, animation, code-writing.
Following experimental workshops have been conducted to test the integration of
idea-driven and science-driven design process. A specific methodology was devel-
oped to test this approach (Fig. 1). This methodology including a selection process
of the built project through a democratic vote by the student group was refined
through different experimental workshops with master in architecture students and
external experts at the University of Applied Arts Vienna. The length of the work-
shops was 5–7 days, divided in an approx. 2.0–2.5 days for the project research
phase, 0.5–1.0 day for the project selection phase and 2.5–3.5 days for the project
development phase including fabrication.
2.5 Educational Setup
Digital Design Vision: The students were asked to research the interplay and influence
of material performance, shape behavior and atmospheric presence. In order to setup
and test the idea driven and science driven design process, different ambitions were
presented to the students. Morning sessions teaching software tutorials alternated
with afternoon sessions enforcing project design development in groups.
Setup of Science Driven Design process: Tutorials were organized in the morn-
ings to acquire the necessary digital software and technology skills. The tutorials
were selected to fit with the respective type of the structure. The tutorials cov-
ered different tasks in ascending complexity. The overall ambition was to create
form-active structures by analyzing the performative structural behavior. Structural
references, performative typologies, the relation of shape, geometry and structural
strength were discussed. Instructors lectured the software skills with the intention to
setup an understanding on how to make design decision based on performative eval-
uations and strategies to develop a design progress based on systematic decisions.
Students’ capabilities were strengthened towards building and handling complex
material matter systems while raising digital knowledge through 1:1 tutorial help
and independent software tasks.
Setup of Idea Driven Design process: Design sessions were organized in the
afternoons. The ambition was to create architectural artifacts, focusing not on pro-
grammatic requirements but aiming at unveiling architectural qualities. The students
were organized in groups of 3–4 persons and asked to develop three-dimensional
Fig. 1 Design methodology

1362 A. Gheorghe
concepts. Drawing, digital 3D software, model building techniques were allowed to

work with. Each group designed with intense help and under constant participatory
observation of the instructors structural modules of different sizes and configurations.
Different architectural references were discussed.
These spatial concepts were discussed while keeping their relevance to the digital
design vision as defined above in mind. Furthermore these artifacts were evaluated
against structural behavior and vertical growth strategies. The instructors initiated
and were involved in group discussions, offering deskcrit sessions if deemed nec-
essary. This happened with the intention to help students formulate a project vision
incorporating the digital tools at hand while simultaneously unveiling phenomeno-
logical architectural qualities.
After multiple feedback loops, involving material and fabrication testing, large-
scale digital artifacts were constructed as a group effort. In order to develop and
strengthen critical thinking, all students were asked to participate and contribute in
group discussions. Subsequently, a democratic voting by the students was initiated
to select one project to be built by the complete student body.
2.6 Workshop in Aarhus “Cones”
Type: Vertical Plated structure

Digital Design Tools: Rhino, Grasshopper, Karamba
Digital Fabrication Tools: industrial grade Hydroline IKC water jet cutter
Material: OSB (Oriented Strand Boards 25 mm)
Date and Place: 6/2013—Aarhus School of Architecture, Denmark
The workshop was organized in collaboration with the Aarhus School of Archi-
tecture in Denmark, focusing on experimentation with structural design simulation
and full scale fabrication, hereby exploring material behavior at architectural scale.
This workshop used of 140 m2 of OSB wood boards CNC milled by an industrial
grade water jet. The specific focus was to test architectural scale building with the
respective development of wooden joints and assembly details. The students were
trained to develop and test structural strategies (Karamba) suitable for the creation
of self-supporting structures.
Architectural students were divided in groups of 5 students and were asked to
develop initial design concepts (Fig. 2). Architectural references (Fig. 3) were dis-
cussed in groups (Gottfried Böhm, Livio Vacchini). The material behavior of the
OSB boards was taken into account.
The initial ideas were reworked towards creating vertical oriented structures
(Fig. 4) consisting of smaller modules to be fabricated by a water jet machine.
After developing overall surface options, multiple strategies for discretion and
subdivision the basic shape into structural components were discussed, programmed
in Grasshopper, and tested in Karamba for structural capacity (Fig. 5).
Fig. 2 Initial design ideas, created in the afternoon design sessions (idea-driven process)
Fig. 3 Design references
After the software tutorials and design deskcrits, a simple structure consisting of
two intersecting comes emerged of the voting process (Fig. 6). The basic structural
module was very simple, basically a rectangular OSB board with different cutting
angles and cutting length. A formal challenge constituted the design of the connecting
curves of the saddle between the cones in the digital fabrication preparation phase.
The necessary cutting files were extracted from the 3D model and produced
directly in Grasshopper. The flat elements were then exported in the CAM envi-
ronment of the industrial grade water jet used to cut the elements (Fig. 7).
The pre-cut and pre-labeled OSB boards were assembled during a timeframe
of three days (Fig. 8). The horizontal rings were laid out starting wit the lowest
one, and the vertical boards were attached using the precut slits. The final geometry
emerged through the correct positioning of the boards within the rings, no additional
measurements needed to be made. The correct assembly technique was tested using
a 1:5 model, which was fabricated out of cardboard with a laser cutter (Fig. 9).
The final structure of about 4.5 m height was exhibited in the exterior courtyard
at the Aarhus School of Architecture (Figs. 10 and 11).
1364 A. Gheorghe
Fig. 4 Initial structural concept
Fig. 5 Subdivision strategies, created in the morning tutorial sessions (science-driven process)
The structural system of the structure consists of horizontal rings with slits, which
stabilize the vertical OSB slats arranged radially (Figs. 12 and 13).
Fig. 6 Selected basic design model after the voting
Fig. 7 Flat structural modules (OSB boards)
The structural analysis of the final structure in the Grasshopper plugin Karamba
indicate the areas of maximum deflection areas (magenta) under gravity, vertical and
horizontal load (Fig. 14). The 3D model was translated into a structural FE model
and the particular parts were analyzed as shell elements.
Involved Design Team and Institutions: Andrei Gheorghe, Bence Pap, Alexander
Kalachev, Rasmus Gronbaeck, students and faculty of Aarhus School of Architecture,
Denmark.
1366 A. Gheorghe
Fig. 8 Buildup process
Fig. 9 Laser cut model in scale 1:5 as a reference
2.7 Architecture Challenge 14 “Polymer Folding Scapes”
Type: Beam structure

Fig. 10 Finished structure
Fig. 11 Finished structure
Digital Fabrication Tools: CNC wood saw

Material: laminated wood boards 19 mm
Date and Place: 9/2014, University of Applied Arts Vienna
1368 A. Gheorghe
Fig. 12 Structural system of the finished structure (horizontal rings and area in between cones)
Fig. 13 Structural system of the finished structure
The project represents (Figs. 15 and 16) research that explores the potential of
polymer concrete applications and wood/concrete composites in the production of
full-scale structures.
Fig. 14 Karamba analysis of the finished structure
Fig. 15 AC14 finished structure

1370 A. Gheorghe
Fig. 16 AC14 finished

structure
The handling of virtual simulation methods and the engineering of full-scale

structures using digital manufacturing were a primary focus of the project. The
polymer concrete material testing and application has been made possible through
the support of the RAMPF Group in Germany.
2.7.1 Structural Performance
Structural testing of joints in wood/polymer concrete composites (Fig. 17) revealed a

strong tension and bending ability of the cured material. Using this method, stiff con-
nection points between wood members could be achieved without the inlay of metal
sheets. Threaded rods were used in prototypical tests, but the adherence between
wood and polymer concrete proved to be high enough to omit the metal rods.
A hyperboloid surface (Fig. 18) was designed in Rhino Grasshopper and opti-
mized using Karamba and multicriteria optimization in Octopus taking into account
the architectural site context, structural behavior, assembly and construction limits
(Fig. 19).
Fig. 17 AC14 structural polymer/wood composite node
Fig. 18 Initial design sketches created during the afternoon design sessions (idea-driven process)
1372 A. Gheorghe
Fig. 19 Project buildup through radially arranged branches
2.7.2 Intrinsic Material Qualities
The precisely defined design strategy reacts with loosely defined material behavior.
The formwork was unfolded from the 3D model and laser cut on polystyrol sheets.
The wood members were fit at the joints in the formwork and the polymer concrete
was cast. After the curing process, a stiff wooden joint consisting of direct adhesion of
wood and concrete was achieved. The overall geometry emerged through the precise
folding process of the cast nodes. In order to keep the geometry in place, the ends
of the wooden boards were connected with strings of the right measurement length
(Fig. 20).
The precise, optimized overall geometry correlate with the very differentiated
aesthetic quality of the cast polymer concrete joints. This intuitive material behavior
enriches the digital precision resulting in artifacts of one-time unique aesthetic design
results. Material experiment and final production of the structure merge in one single
fabrication process (Fig. 21).
Fig. 20 One branch consisting of 5 wooden boards and 2 concrete nodes
2.7.3 Fabrication Process
A Grasshopper definition was developed to extract two types of nodes, the cast
3-points polymer concrete nodes and wood–wood connection joints between the
3-points modules. Grasshopper was used to compute the wood–wood joints at a
right angle to allow for easy assembly and fixation while erecting the final structure.
1374 A. Gheorghe
Fig. 21 Prototypical testing of the polymer concrete node cast in polystyrol sheet formwork
Fig. 22 Polystyrol sheet formwork connected with the wood boards before casting
A specific material mixture of the polymer concrete components combined with

precise time management was required to produce the structural stiff nodes (Figs. 22,
23 and 24).
Fig. 23 Demounting of the polystyrol sheet formwork after casting the polymer concrete
Fig. 24 Casting the polymer concrete in multiple branches
2.7.4 Results and Future Application
The project researches an alternative way of construction testing modular assembly.

The high strength polymer concrete was applied for the first time in a new way in
this project to achieve structurally stiff wood/concrete composites without any metal
inlays. Multiple applications of the wood/polymer composite were tested prototypi-
cally (Figs. 25 and 26), in future, this technology could be refined to be used in fully
functional small and large-scale construction.
1376 A. Gheorghe
Fig. 25 Polymer concrete/wood composite test to create local inflexibilities
Fig. 26 Polymer concrete/wood composite prototypical tests
2.7.5 Involved Design Team and Institutions
Andrei Gheorghe, Robert Thum, Wieland Becker, Peter Ortner, Robert Vierlinger,
Eirik Kjolsrud, University of Applied Arts Vienna, Hochschule Trier, Bollinger +
Grohmann Ingenieure.
2.7.6 Students
Ethan Kerber, Jiaxing Lu, Mehmet Onur Köklü, Julia Vishanenko, Clemens Neu-
ber, Amine Khouni, Alexander Walzer, Marc Winklhofer, Jakub Havlík, Lukas
Gschweitl, Juan Sepúlveda, Padure Andrei, Madalin Gheorghe, Anan Ashour, Tal
Friedman, Marin Katusic, Sorana Chiris, Eleonore Audi, Lidia Ratoi, Vlad Popa,
Klaus Marschallinger, Oliver Alunovic, Kevin Mazanek, Anna Varakli.
2.8 DigDesFab14 + 15 “NoHOME Pavilion”
Type: Modular shell structure

Digital Fabrication Tools: CNC Mill (3-axe), 250 × 125 cm bed size
Material: Laminated wood boards, 19 mm
Date and Place: 9/2014–11/2015, University of Applied Arts Vienna
This project achieved the construction of a prototypical type of structure (Figs. 27,
28 and 29) testing a new, innovative composite material consisting of wood and poly-
mer concrete (Gheorghe and Vierlinger 2017). This composite material application
was suitable for mid-scale, light-weight structures (pavilions) using pre-fabricated
elements with structural stiff nodes in wood and an unrestricted geometry (any shape
could be produced).
The research pavilion followed a specific structural strategy. It attempted to
develop and test a new hybrid construction technique using a combination of poly-
Fig. 27 noHOME Pavilion in Mariahilferstrasse pedestrian zone, Vienna

1378 A. Gheorghe
Fig. 28 noHOME Pavilion at Loft City, Ankerbrot Fabrik Vienna
Fig. 29 noHOME Pavilion at Expedit Halle, Ankerbrot Fabrik Vienna
mer concrete and wood as a composite material. Polymer concrete has not yet been
tested on large-scale architectural production. The Hochschule Trier completed some
full-scale mockups in Trier (2014, 2016) and in Vienna (2014) in cooperation with
the University of Applied Arts (Schober et al. 2016). These initial mockups can
be inscribed in a bounding box volume of about 3 × 3 × 3 m. In the DigDesFab15
research pavilion project, the polymer concrete material has been tested as a hybrid
material in combination and direct adherence with wood boards in a bounding box
volume of about 8 × 8 × 8 m.
2.8.1 Material Composition and Properties
As opposed to conventional concrete, the binding component of polymer concrete

is not cement, but polyester resin. A precise relation and mix of three components
(7% resin, 2% additive compound, 91% mineral aggregate) is required to ensure the
material stiffness. For this project, a total of 120 kg of Epument 140/5 A1 polymer
concrete was used.
Therefore, due to the exact mixture, no different material consistencies are possi-
ble, as this is the case with cement-based concrete. This proves to be a disadvantage
when applying and pouring the mixture. Another disadvantage is the delicate han-
dling due to resin-based components and the relatively high price as compared with
cement-based concrete.
However, the application of polymer concrete offers also large advantages. The
cured material is not brittle, has also high tension (not only compression) abilities,
and proves to have a high adherence to wood materials. Aesthetically, it can be
produced (depending on the additives) with different grain sizes, so that the surface
quality can be very smooth or coarse, depending on the formwork surface. Polymer
concrete does not need to be compacted by vibration.
2.8.2 Use of the Material in the DigDesFab15 Research Pavilion
In recent years polymer concrete testing and small-scale applications were conducted
at the Hochschule Trier (Prof. W. Becker, Prof. R. Thum) in Germany. The DigDes-
Fab15 project built upon the knowledge gained through this preliminary research in
Germany, however, this type of construction and structural typology has been tested
with this research pavilion for the first time at larger scale. The intention was to create
a fully self-supporting structure with structural stiff nodes (Fig. 30) on a rectangular
grid (without the need of using a triangulated grid). This is innovative, because under
normal conditions it is very difficult to achieve structural bending stiff nodes in wood
joints, therefore in many cases triangulated (truss) systems are used. This problem
naturally reduces the availability of cost effective design options for architects and
structural engineers.
2.8.3 Architectural Concept and Social Aspects
The “noHOME” project had a specific architectural concept. It examined the theme of
homelessness. Being homeless becomes unfortunate reality for an increasing number
of people in Vienna. Rising living expenses causes this problem to constantly grow,
although this situation is often not visible to our relative wealthy society.
The noHOME pavilion interprets sensory impressions caused by the absence of a
home. The aim of the design was to create a spatial shell, which does not satisfy the
basic needs of a home (protection, sleeping, eating, living), but reflects to the visitor
the subjective sensations that arise due to the lack of privacy. It was conceived as
1380 A. Gheorghe
Fig. 30 Structural stiff polymer concrete node
an architectural opportunity to draw public attention towards the social problem of

being without a home/homeland. The aim of the design was to raise consciousness
about implications of being homeless for the persons concerned and their perception
in society. In order to achieve this, the design explored and evoked subjective feelings
such as the lack of privacy and the missing of a protective interior, a certain ambiguity
between inside and outside with varying degrees of exposure, the perception of a
hostile environment leading to an aggressive contextual interface and a limitation of
space without the provision of shelter functions to weather conditions (Figs. 31, 32
and 33).
2.8.4 Educational Method
The design process of the research pavilion followed a systematic approach. Twenty-
four students were organized in four groups of six students each. These groups were
given a design task and were tutored towards developing initial design proposals.
The project design was developed from two points of interest. On one hand the
architectural brief asked for an examination with regard to the current situation and
living conditions of homeless people, and on the other hand a structurally feasible
Fig. 31 Detail of the structure in Mariahilferstrasse pedestrian zone, Vienna

1382 A. Gheorghe
construction technique and application of the polymer concrete in combination with

wood had to be developed.
For this project seminar, it was not required by the architectural brief to design a
water-proof shelter for rain protection. The main focus was to develop methods and
design concepts towards rising awareness for this social problem while the structure
is exhibited in public space.
The student group size was chosen to be large enough in order to have a criti-
cal mass within the group and allow for specialization of student tasks (modeling,
sketching, model building, rendering, etc.) but to be still small enough to allow for
efficient group dynamics.
The different student projects were discussed through multiple rounds of reviews
with invited external guests and members of the faculty at the University of Applied
Arts. The tutors provided parametric modeling, structural and architectural design
input (Fig. 34). After the four projects reached a certain design maturity and construc-
tion feasibility, they were discussed within the whole group, and after the discussion
a democratic vote was instantiated to choose one project to be further designed
in detail and constructed by all students. The architecture students developed the
required Rhino Grasshopper and Karamba analysis setup and definition files with
the help of experts from Bollinger + Grohmann Engineers.
Fig. 34 Strategies of concrete casting (parametric models)

1384 A. Gheorghe
2.8.5 Design Method
The structural system has been developed during the design tasks from the so-called
Zollinger system (Menges et al. 2016), where small modules (Fig. 35) are prefabri-
cated and joined to form a greater structure. A parametric model in the Rhinoceros
and Grasshopper software was created and tested virtually through various subdivi-
sion densities and node sizes. The overall structure was evaluated with the structural
analysis tool Karamba (plugin for Rhino Grasshopper), the different heights of the
lamellas in the pavilion arise from the structural analysis of stresses and deflections
under vertical and horizontal loading.
The Grasshopper definition loaded a NURBS surface created in RhinoVault. The
definition applies a rectangular subdivision and creates the corresponding four legs
of each node as sheets (or lamellas) of different structural heights (Fig. 36). The
Karamba plugin was used to apply gravity (vertical) and mesh (horizontal) loads to the
Fig. 35 Initial design concepts (idea-driven process)

Fig. 36 Initial subdivision strategies (science-driven process)
Fig. 37 CNC milling
structure. The mesh view component and corresponding material utilization results
in Karamba was linked to differentiate the height of the lamellas parametrically. The
rectangular mesh subdivision is optimized while keeping utilization below 60% of
the maximum material capacity (Fig. 40).
2.8.6 Construction Method
Laminated spruce timber boards (3 layered, 19 mm thick) were used for construction.
All the pieces from the 3D model were flattened and distributed on standard size
boards (1.25 m × 2.00 m) parametrically, using the Rhino plugin Grasshopper. The
pieces were cut by a 3axe CNC mill (Fig. 37) and the edges were sanded to avoid
injuries when exhibiting the final piece. The geometrically unique pieces (cut two
times for double layering) were labeled in the digital model and the students sorted
and labeled and grouped the adjacent pieces (node by node) immediately manually
after production. Each node was screwed together with metal screws as a temporary
fixture.
The three-dimensional geometry was predetermined by the shape of the assem-
bled pieces in each node. Neighboring pieces would only fit together on the angles
1386 A. Gheorghe
Fig. 38 Node fixation before polymer concrete casting
predetermined by the geometry of the milled pieces, therefore no measurements had

to be transferred from the 3D model to the physical model at this state. This means,
the four pieces for each node and the bottom sealing plate would fit together only
in one way. For helping the assembly process and for purposes of geometry control,
marker slits were milled in each piece, indicating where the neighboring piece would
need to be attached precisely (Fig. 38).
The screws only served as a temporary fixation, until the polymer concrete was
filled in the center of node for final fixation. Each node was sealed on the bottom
with a custom-cut plate out of a thinner (8 mm) plywood material, labeled with an
unique code. The project builds up on a rectangular grid, each node has four legs
around a central node cavity (reciprocal system) where the polymer concrete was
poured.
The polymer concrete (curing time about 8 h) adhered with the wood and therefore
created a structurally stiff node in wood, in a large range of geometrical variations.
Fig. 39 Finished nodes before assembly
Normally, it is very difficult to achieve stiff structural nodes in wood structures;

therefore this applied composite method might offer interesting potential in future.
All 122 nodes were fabricated using a production line (Fig. 41), were cast sepa-
rately (Fig. 39) and bolted on-site with screws. An assembly strategy was developed
to easily connect the nodes with each other through using milled geometrical ref-
1388 A. Gheorghe
Fig. 40 Karamba analysis of the finished structure (maximum deflection in red)
Fig. 41 Production line (cut elements, sorting, node fixation and polymer concrete pouring)
erences. In the final structure the wood sheets of the four node legs overlap, so the
final material thickness is 2 × 19 mm 38 mm. The overlapping pieces fits at certain
positions, this creates a reference when assembling the 1:1 structure. The assembly
pieces were added starting from the lowest row and the structure reached final struc-
tural stability after the arches were completed. During the construction, scaffolding
was needed for the assemblers to reach and mount the higher levels of the structure.
Equally, the scaffolding helped to stabilize the not completely closed arches of the
structure against horizontal loads (wind) during construction (Fig. 40).
2.8.7 Reflections
The pavilion (Figs. 42, 43 and 44) was decomposed and re-assembled at different
locations. The nodes could be disassembled node after node by unscrewing the con-
nections of the double-layered wood sheets. The quality of the hybrid wood board and
polymer concrete remained constantly good after being exposed to outside weather
and humidity conditions for several months. The wood was not treated in this exper-
iment, in order to not have a potentially structurally weak layer between the wood
surface and polymer concrete. Therefore the exposure of the raw wood to outdoor
conditions could not have lasted for more than 6 months. In future experiments, dif-
ferent wood treatments with the combination of polymer concrete application can be
tested. The color of the exposed polymer concrete changed significantly after being
exposed to UV light for a longer period. Multiple composite pieces and nodes were
tested for material failure, the breaking fissure occurred always within the laminated
wood board, and not in the polymer concrete/wood surface joint. Special attention
has to be devoted due to the fact that the polymer concrete has to be poured and (and
the material consistency of the polymer concrete cannot be changed), solutions have
to be found for keeping the fluid polymer concrete in contact with the wood until
curing (formwork, temporary fixation, etc.).
These construction techniques proved to be a structurally viable way to produce
stiff timber nodes in any geometrical layout for mid-scale outdoor construction within
this specific structural system. Normally, for structural stiff wood/wood connec-
tions steel inlays or special screws/bolted connections are required. For this research
project, the wood pieces were connected in a stiff structural node by pouring polymer
concrete. If the contact surface between polymer concrete and wood is large enough,
the hybrid polymer concrete/wood joint can be exposed to high tension and compres-
Fig. 42 Side elevation final structure

1390 A. Gheorghe
Fig. 43 3D view final structure
sion forces before material failure happens. Usually, the adherence is high, therefore
the material failure occurs in the wood component, without destroying the compos-
ite joint. One of the largest advantage of this composite construction technique is
that due to the fluid material consistency of the polymer concrete, any geometrical,
three-dimensional layout can be produced, as long as a temporary containment and
fixation can be implemented until the chemical curing process is completed (about
8 h). The composite material testing at middle scale (pavilion scale) has been suc-
cessful, however, for large scale, industrial architectural application at building scale,
further knowledge and research has to be done specifically in two areas:
A. assessing structural strengths (compressions/tension ability) of the composite
material
B. long-term durability (exposure to humidity and UV light) of the composite mate-
rial.
2.8.8 Involved Design Team and Institutions
Andrei Gheorghe, Robert Vierlinger, Jiaxing Lu, Christian Schwarzwimmer, Sophie

Gierlinger, Ceren Yönetim, Marko Margeta, Eva Blsakova, Carina Zabini, Anna
Sergeeva, Saba Nabavi Tafreshi, Nasim Nabavi Tafreshi, Julia Karnaukhova, Miro
Straka, Jiri Vitek, Mathias Bank Stigsen, Sophia Keivanlo, Johanna Jõekalda, Ioana
Binica, Leah Park, Nicolas Gold, Cecilia Sannella, Agnese Trinchera, Sven Win-
Fig. 44 Top view final structure
kler, Florian Smutny, Ziwar al Nouri, Banafsheh Fahimipour, Bernd Seidl, Johannes
Cziegler, Maria Korolova, Luca Beltrame, Niklas Knap, Fady Hadad, Katerina
Joannides, Andrea Carniti, Afshin Koupaei, Atanas Zhelev, Roxy Rieder, Min
Zhang, Dennis Schiaroli, Angelica Lorenzi, Johanna Jelinek, Mary Denman, Bar-
bara Schickermüller, Roman Hajtmanek, Adam Sebestyen, Kaan Karabagli, Institute
of Architecture, University of Applied Arts Vienna “die Angewandte”, Bollinger +
Grohmann Engineers, Loftcity Brotfabrik.
3 Conclusion
Due to the experimental nature of the projects, it was very helpful to develop and test
the underlying basic principles within an academic setting. The students benefitted
from the collaboration with structural experts, and became used to the simultaneous
1392 A. Gheorghe
development and combination of design expertise and technological or structural

principles. At the same time, the considerably amount of work involved in design-
ing, manufacturing and assembling the prototypical structures for testing purposes
only would not have been possible in a purely industrial setting without investing
substantial financial and human resources.
The empirical studies suggest that Digital Architectural Design knowledge is
conveyed in a very effective manner through short-term workshops with speci-
fied themes. This flexible “workshop” style teaching, often by interdisciplinary,
rapidly changing teams offers the advantage of benefitting from multiple, partly
not connected or distant sources (“cloud cluster teaching”) with the aim of providing
knowhow efficiently across disciplines.
In architecture, “digital architectural design” is often seen as a consequence of
the use of computer technology. In the broader sense, however, it requires a specific,
systematic way of thinking and the design process features some differences when
compared with the traditional design process based on paper and pencil. As most
architecture schools offer an introductory course to visual communication and con-
ceptualization based on analog, traditional hand-drawing, it seems equally beneficial
to offer a similar strategic education for digital design thinking, using a combination
of idea-driven and science-driven approach and aiming at using innovative digi-
tal methods to acquire creative and critical digital proficiency. This should educate
students to be critical designers beyond purely teaching digital skills and empowers
students with an innovative as well as creative use of the new conceptual possibilities
offered by digital technologies.
Acknowledgements Many thanks to the Institute of Architecture at the University of Applied Arts
Vienna, Hochschule Trier, Aarhus School of Architecture and Bollinger + Grohmann Engineers for
the support.
Copyright
The copyright of all images is retained by the respective teams (Workshop Aarhus “Cones”, AC14
“Polymer Folding Scapes”, DigDesFab14 + 15 “noHOME”).
References
Carpo M (2009) Aufstieg und Fall der identischen Reproduzierbarkeit - Zu Leon Battista Alber-
tis unzeitgemäßer Entdeckung digitaler Technologien in der Frührenaissance. In: Gethmann D,
Hauser S (eds) Kulturtechnik Entwerfen: Praktiken, Konzepte und Medien in Architektur und
Design Science. Transcript, Bielefeld, pp 49–63
Cuff D (1995) Architecture: the story of practice. The MIT Press, Cambridge
Gheorghe A, Vierlinger R (2017) DigDesFab15 research pavilion. Front Digit Humanit 4. https://
doi.org/10.3389/fdigh.2017.00018
Mark E, Martens B, Oxman R (2001) The ideal computer curriculum. In: 19th eCAADe confer-
ence proceedings of architectural information management. University of Technology, Helsinki,
Finland, Helsinki, pp 168–175
Menges A (2011) Integral formation and materialisation: computational form and material gestalt.
In: Ahlquist S (ed) Computational design thinking. Wiley, London

Taylor & Francis, London
Oxman Rivka (1999) Educating the designerly thinker. Des Stud 20:105–122
Pelkonen E-L (1999) Entfaltung der Phantasie - Materialstudien von Alvar Aalto. Daidalos
Architektur-Kunst-Kultur, Entwurfsstrategien 71:23–26
Reas C, Fry B (2014) Processing: a programming handbook for visual designers and artists, 2nd
edn. The MIT Press, Cambridge
Ruby A (1999) Espace Trouve - Über das Ent-Decken als Entwurfsstrategie. Daidalos Architektur-
Kunst-Kultur, Entwurfsstrategien 71:62–71
Schober KU, Becker W, Weber J (2016) Grouted joints in timber engineering. In: Proceedings of
the 16th world conference on timber engineering, Vienna
Schön D (1984) The architectural studio as an exemplar of education for reflection-in-action. J
Archit Educ 38(1):2–9
Schön D (1985) The design studio: an exploration of its traditions and potential. RIBA Publications
Limited, London
Vitruvius (1999) de Architectura - Ten Books on Architecture. In: Howe IR (edn) Cambridge
University Press, Cambridge
Zaero-Polo A (2001) Roller coaster construction. In: Verb 1: processing. Actar
Digital Construction of Timber
“Metabolism”
Yasushi Ikeda
Abstract Architecture tectonics with specific geometry is embraced as a major tech-

nique to achieve a physical relationship between engineering and social rationality in
terms of the use of space and resources. Nowadays, thanks to parametric geometry,
new architectural technologies are expected to evolve with digital innovation. Thus,
we are in the age of searching for comprehensive changes in architectural tectonics,
including materials and processes. We have started to think that some reflections of
modern architecture with industrial technology give an opportunity to explore the
tectonics in digital tern. Historically, the concept of “Metabolism” was presented
in 1960 to integrate the culture of Japanese traditional architecture with modernism
and this concept may be revived with digital construction technology. The concept
has an impact on the modularity, adaptability, compatibility, and transformability
of architectural elements. Presenting our experimental project with this idea, this
paper discusses the paradigm shift of architectural design and technology from mod-
ern construction based on mass production to digital construction based on smart
production.
Keywords Digital construction · Parametric geometry · Timber structure

Metabolism · Digital wood block
1 Geometry and Productivity of Tectonics
In many aspects, architectural technology has been a matter of geometry. It is not

only the mathematical definition of shapes, forms, and spaces but also a fundamental
tool to identify the performance of these properties.
For example, the geometry of the curved form of a Roman masonry arch was an
exceptional technology that spanned a space while supporting the weight above it in
a pure compression structure with discrete materials pushing against each other. Over
Y. Ikeda (B)
Keio University SFC, Fujisawa, Japan
e-mail: yasushi@sfc.keio.ac.jp

1396 Y. Ikeda
time, these arches have made a great impact on human civilizations with numerous
applications in our living environment. The observation of the development of arches
and vaults in many regions, as well as similar variants, shows us that geometries in the
form of arches can relate tectonic, iconography, and cognitive effects to influence the
feeling and behavior of people around the arch form. These critical characteristics of
architecture are integrated in the curvature of an arch. With a similar mechanism, the
Asian tradition of timber structures with posts and beams have been developed in the
form of a rectangular plan and triangular roof. The two traditions seem to use very
different tectonic solutions for architecture, but both are a result of the evolution of
productivity and usability of space over a long period. This phenomenon shows the
basic engagement between technology and design. We must consider solutions that
involve social usability, for certain programs or activities in a space, while considering
the technological limitations and potential of materials and construction when we
design something. Therefore, design could be defined as the result of the relationship
between activity and technology.
Traditionally, we think of architecture as a solution realized in one shape or form.
However, the use of parametric modeling has evoked a more flexible, adaptive, and
compatible type of solution, which has favored transformable objects as a whole
rather than one screenshot of a single concept model’s shape or form. In short,
digital technology has shifted the paradigm from the design of form to the design of
algorithms for transformation (Fig. 1).
There is not much to say about the novel freedom created by geometrical manip-
ulation with digital tools on the computer monitor (Woodbury et al. 2013). But
following that, the robotic manufacture of geometries represented by data and the
productivity of materialization has evolved radically in many cases. Therefore, the
technology of modern architecture can be re-examined in terms of this basic premise
of the digital age. What is the major difference between geometry and productivity
in the digital age compared to existing construction technology?
Fig. 1 Sift of target under

relation between technology
and design
Digital Construction of Timber “Metabolism” 1397
2 Searching for the Advantage of Tectonics in Parametric

Geometry
In 2010, our laboratory team challenged experimental project to investigate and

demonstrate the potential of digital geometry and digital fabrication. A pavilion with
500 non-standardized parts of laser cut plywood named “Porous Torus” was thus
created (Fig. 2). Our intention with this project was to achieve a shape with double
curvatures and simple joints using wood panels. A fluid double curvature surface is a
typical motif for computational geometry. However, when this type of surface needs
to be assembled with discrete parts, it is almost inevitable that they be subdivided
into irregular patterns. Thanks to digital fabrication methods, such as laser cutters,
the uniqueness of the parts does not affect the productivity as long as the data to
be processed by the machine is provided. However tectonic performance could vary
since there could be many different ways to subdivide the geometry; thus, there
are significant differences in the robustness, ease, and process of construction. In
other words, the algorithm to create series of unique parts is primarily a matter of
geometry and productivity. In this particular case, simplified joint details constrained
the minimum size of panels and the minimum distance between joints, which resulted
in an arbitrary layout of panels. This experimental project gave us a fundamental
understanding of non-standard fabrication as an opportunity in digital tectonics,
especially for double curvature surfaces and irregular subdivisions (Figs. 3 and 4).
Organic patterns with non-identical units have been observed in nature generally.
Human attempts to apply this irregular uniformity in traditional manufacturing and
architecture has been achieved by imitating it manually; however, most have been in
the form of regular geometrical patterns of identical parts in the modern age. There
Fig. 2 “Porous Torus” A experimental build of parametric form with non-standardized parts
1398 Y. Ikeda
Fig. 3 The detail of joint for interlocking units has certain geometrical constrains
Fig. 4 Simulation of layout

solution found
algorithmically for all panels
to meet a condition
are other inspirations for the potential of patterns with irregular uniformity coming
out arising from a practical project we have worked on at in a local town in China. The
project was built in 2011 as a Japanese style hot spring spa at a remote village in the
north east China, where neither contemporary technology nor skilled craftspeople
were available. We noticed that the local craftsmanship of traditional stone finishes
with irregular units had interesting features because it seemed to be easier to get a
superior hand-made finish by hiding irregular gaps between parts. We applied this
idea to the design of a ceiling wood louver, showing that it was smarter to keep
an allowance for error when sufficient accuracy in the architectural finish could not
be achieved (Fig. 5). After computer simulations on how builders selected irregular
units to add sequentially, we were convinced that a simple set of rules could generate
Fig. 5 Algorithmicaly scattered pattern for ceiling louver to mitigate blur an appearance of errors
Fig. 6 A result of partial simulation to predict craftman’s work of layout
sufficiently randomized patterns to meet aesthetic standards (Fig. 6). We asked local
builders to follow instructions on which units to add next to avoid alignment gaps
in the next line (Fig. 7). The result was a kind of algorithmic construction process
carried out with manual labor.
It was an unexpected surprise when we found that the finish was even better than we
planned because all the lighting fixtures were nicely bound between the deeper louver
units. This was done based on the local builder’s ideas without any suggestions from
our side. This was impressive because it may not have happened if we had specified
the complete pattern in drawings. In natural phenomena, heterogeneous patterns
generated by algorithmic interactions have advantages in terms of adaptability or
robustness compared to regular and periodical patterns. This experience gave the
insight that in construction, algorithmic processes using irregular patterns have the
potential to add adaptability.
1400 Y. Ikeda
Fig. 7 The instructions for craftman to avoid the consecutive gap while using 4 standardized length
of parts
3 Rethinking Geometry and Productivity of Modern

Architecture
An industrial technology has the economic potential to increase productivity regard-

ing performance predictability based on the regularity of material behavior. That is
why it prefer homogenized materiality and standardized use of identical unit. And
this economical point has been pursued while the industrial civilization spreading
globally with a methodology base on the concept of “standardization”.
In contrast, digital technologies help the realization of complex projects by
increasing the productivity in terms of the fabrication of unique units for assem-
bly and allow for irregular layouts with more tolerance. If we see this as a major
difference compared to Modern architecture, we can examine this advantage as an
evolution compared to the Modern style’s Geometry and Productivity.
Orthogonal geometry is a principle for construction in modern buildings; it became
a style when it showed the technological advantages of rectangular solids as follows.
1. The projected 2D graphic in the trihedral figure of objects makes the real length
and angle of objects measurable. This makes the scaled model and real fabrication
consistent.
2. Quantitative indexes of forms, such as the area, volume, and amount of parts,
can be identified with relatively simple multiplications. This helps us figure out
the physical performance with parameters like the weight and force.
3. The composition from the parts to the whole can be optimized with the repeated
arrangement of regular units. In other words, the rectangle is the most efficient
type of shape to construct buildings with identical parts.
The orthogonal geometry can maximizes the advantages of mass production
industry by minimizing the cost of manufacturing because rectangle is the most
efficient shape for construction using identical parts. The modern maestro architect,
Mies van der Rohe, advocated the use of rectangular parts in orthogonal coordination
geometry as a symbolic feature of “Universal Space” in the age of the popularized
and industrialized world. It is worth noting that a steel is recognized as the repre-
sentative material of modernism, although iron has been utilized much earlier. More
precisely, we should recognize that a wide use of mass-produced rolled economical
steel has had a critical impact on the modern architectural concept based on stan-
dardization with its homogeneity in material and identically in shape. That have been
disadvantages of wood as industrial products and therefor it will be a great chance
to use digital technologies to manage an uniqueness of all non-standardized parts of
natural materials without sacrificing productivity of collective assembly of it.
4 Japanese Traditional Timber Tectonics

and “Metabolism”
While this economical productivity is mostly taken as feasibility of architecture,

it must be discussed with the purpose or particular value of it from benchmark of
orthogonal geometry of modernism. We tried to illustrate the Rank of Digital Con-
struction in comparison of digital and industrial architecture as Fig. 8. Our consid-
eration is necessity of novel value of organic system of non-standerdized techtonics.
It will be a driving force of this paradigm sift. We are seeking it in a sustainability,
resiliency and adaptability of construction through our previous experience. With this
new paradigm, we may start exploring the use of materials with non-standardized
methodology that have other advantages in productivity rather than simply being mas-
sive and stable. Most potential alternatives are renewable, natural, and of botanical
material, such as timber. These materials can be processed easily without centralized
production. Also, the fluctuation in natural material can be managed with the digital
processing of individual properties of units. In addition, timber can be handled with
relatively smaller power in construction since it is lightweight.
Fig. 8 Rank of digital

construction in comparison
of digital and industrial
architecture
1402 Y. Ikeda
Japan, a country with an Asian tradition, has a different background in the tra-
dition of architectural culture and technologies; wooden materials have been used
since before the modern age. The rectangular and orthogonal coordination of geome-
try started to appear in post and beam timber structures from the beginning. However,
it has a very different philosophy in terms of the material use, tectonics, and form, as
well as the system of design. In ancient temples, timber structures have been main-
tained for more than 1300 years by replacing damaged parts periodically. Such parts
are connected with each other without nails or screws for easy assembly and disas-
sembly. This resulted in an algorithmic manner of assembly involving intersecting
joints by cutting unique incisions for each part or multiple interlocking brackets. All
parts are unique in shape and in material because of the heterogeneous nature of
wood. This algorithmic method determined how each unique shape was made and
led to the construction of considerable large buildings without blueprints; the selec-
tion of unique shapes was the most important skill for apprentices to successfully
become a carpenter. Focusing on this tradition, Japanese architects once presented the
manifest “Metabolism 1960 proposal for a new urbanism” (Kawazoe et al. 1960) at
the World Design Conference held in Japan in 1960. The group pointed out that
modernism could not include a self-organizing system for growth. On the contrary,
traditional Japanese architecture processes allowed whole systems to grow, repro-
duce, maintain their structures, and respond to their environments. They researched
Japanese traditional architecture in terms of the adaptability of function and scale.
However, computer simulations and the knowledge of complex sciences were not
advanced enough to implement it. It is now easier to explore the implementation
of “Metabolism” with digital technology by applying parametric geometries for a
tectonic system of architecture because we can use non-standard production methods
with digital fabrication and the non-linear simulation of heterogeneous structures.
We can also apply computing power to simulate and ensure the effect of metabolism
on an abundant number of units.
5 Revival of Japanese Metabolism with “Digital Wood

Blocks”
Regarding an idea of Revival of Japanese Metabolism with digital technologies, we

started to study on the conceptual model “Digital Wood Block” projects since 2010
to address the issue (Fig. 9). To demonstrate the concept in the beginning stages of
the project in 2010, we proposed architectural designs for a 7 story commercial use
building in Tokyo. In the model, every small wood block unit was identified with
computational coordination. We assume one of the advantages of timber as a con-
struction material is on-site workability. Wood is comparably lighter than similarly
sized materials and can be cut or curved with less effort than materials such as stone
and steel. These are reasons why it is the most popular material for non-professional
builders. These important usability characteristics should be retained in computa-
Fig. 9 Proposal of the

digital wood block building
based on the concept of
continuous construction with
reloadable modules assisted
by computational guidance
tional assistance of construction. On the other hand, even after start use building,
human activities are changing over time, especially for commercial use. Therefore,
it is an advantage in terms of usability if a material can be modified and adopted
by users themselves without professional services or major reconstruction requir-
ing a suspension in the use of a building. This is why we focus on the method of
computational modularity to organize a complete structure based on local conditions.
Regarding this issue, firstly we have developed an algorithm to generate supple-
mental modules needed for sufficient structural stability after a manual specification
input of the preferable layout of a structure (Fig. 10). We intended it to be a kind of
human computer interaction process for the design phase under the condition of it
does not have fixed global geometry and accept local modifying autonomously. Then
we found the same idea also create the possibility of allowing gradual modifications
by replacing small compatible modules in the maintenance phase. As a results our
proposal sift its target to a computational model with an interface called the “Mod-
ule Navigator” to enable human computer interaction in the process of construction
using wood blocks to allow endless communication between the human and com-
puter (Fig. 11). Then our proposal also can show potential of recycling modules and
adaptively changing architecture while maintaining structural assurance.
6 Development of Support Systems for “Digital Wood

Block”
The original proposal of Digital Wood Block Buildings extract many technological
topics is yet to be developed. The most typical criticism to that idea was unnecessary
human effort to manage unique parts and complex layout of it. It excess human ability
to recognize difference of parts and memorize all relative location of each parts. That
seems quite common issue of many projects on assembly with unique parts and also
1404 Y. Ikeda
Fig. 10 The process of obtaining layouts of wood blocks from human input for the baseline (yellow)
and the computational supplemental line (green)
Fig. 11 Viewer to retrieve information for each module to navigate the manual conversion work
of existing wood blocks
relevant issue in management of heterogeneous materiality. So our study moves to

more technical terms to realize “Module Navigator” to guide and assist human work
on non-standardized construction. Following this study, we have been developing
digital devices to move the project forward.
The most fundamental technology to achieve our ideal system is to identify all
modules in real space and get precise information on the current status, such as the
location and layout of modules. Adding to these two points, we also need to provide
information to support manual work during the construction process due to the diffi-
culty of handling enormous amounts of unique parts, based on previous projects. It
is a typical problem that human builders need more guidance and information when
they have to work on massive unique modules, simply because they cannot figure
out the uniqueness and the intricate combination of the modules. Regarding these
points, we made primitive and conceptual mock-ups of a handling system for digital
wood blocks to study and demonstrate our idea.
1. All unique wood block modules have Radio Frequency IDentifier (RFID) in
possible joinery positions.
2. By referring to the ID in the database builder, we can get the stored information
of parts, which will be indicated through Augmented Reality (AR) immediately.
3. By sending the ID to virtual models, builders can input the result of building up
blocks in an incremental process.
4. The computational simulation of the physical balance of all blocks can assess
the stability of the current build status and evaluate the most recent block added
to it.
5. Human builders can get feedback through AR and try another block if needed.
The dynamic data flow to synchronize the digital and physical model by tracking
part IDs is the most essential idea in this system (Figs. 12 and 13). This experiment
for a hybrid system between humans and computers contributed to the exploration of
the possibilities and limitations of digitally-enhanced construction methods. Origi-
nally, our system developed for providing information on unique parts for humans to
guide the collect location designed in a computational model in advance. It is com-
pared to a “jigsaw puzzle where the parts themselves tell the answer”. But during a
development we found a necessity of feedback from current status of physical model
to computer model to keep synchronicity of them. Then additionally, we found that
the capacity of the computational model to adopt human errors or intuitive decisions
in the real process relies on the results of human work being detected and fed back
to the computational model to amend the overall plan accordingly. It is compared to
“assembling work with a route navigation system” because it is keep on updating the
guidance information such as indications of the next block to build. When we under-
stood the same system for dynamic updating of computational models leads to the
development of further systems that suggest high-level recommendations based on
simulations for predicting problem in a complex status such as the physical balance
of all blocks. Finally, this stage of the system is compared to “a complex Jenga game
played with a computer” because human builders are able to act more intuitively at
every step. Both human and computer may not have a final goal but results of con-
struction are kept in certain condition through continuous interaction between them.
As a result it approaches a collaborative process with a computer It is remarkable if
the informative support of modules during construction can extract more interaction
from human builders because it shows the phenomenon of an integration of design
and construction.
7 Conclusion: Sustainable Digital Construction of Timber

“Metabolism”
Thanks to computational geometry, new architectural technologies are expected to

evolve with digital innovations. We can take further advantage of local and traditional
wisdom in the future because it sets us free from the universality of the mass produc-
1406 Y. Ikeda
Fig. 12 Demonstration for interactive non-standidized wood block pilling with RFID sensing and
AR guidance
Fig. 13 System diagaram of integration among uncertain human work on real block and virtual
model to predict result of the take
tion industry. Non-standardized but smart products give us more opportunities to use
local and natural materials in a sustainable way, allowing for new productivity with
intuitive human work. The sustainable tectonics of “Metabolism”, with this type of
digital construction, can make building an endless process. It is somewhat against
the conventional idea of the iteration buildings that have separated phases of design,
construction, and use. Current engineering, professional, and institutional regulations
are based on this fundamental structure of phased processes in realization through
the documentation of design. Mario Carpo wrote in his book “The Alphabet and the
Algorithm” that one of the “paradigm shifts” in the history of architecture was made
in the fifteenth century by Leon Battista Alberti’s invention of architectural design
as an identical copy of a graphic design drawing (Carpo 2011). The Renaissance age
produced many innovations including architectural designs based on scale drawings
through the re-discovery of ancient graphical sciences from Greece. Interestingly,
there was neither a definition of completion nor a distinction between design phases.
People changed and adjusted designs while they were constructing or even using
their building, just like the never-ending construction of medieval cathedrals. The
Renaissance age established a dogma of the human process of physical creation as
a consideration between the virtual model in the drawing and the real existence.
From the viewpoint of informatics, this was the beginning of the coupling between
simulation and materialization. We can see the rise of BIM (Building Information
Modeling) as a further descendant of this paradigm. The advantages of BIM, in
terms of changing our workflow through communication using data, allows both
the further prediction of the result of the transformation and a larger allowance for
unpredictable events. The result is an impact on the modularity, adaptability, com-
patibility, and transformability of architectural elements. If BIM is able to sense the
real element status and identify almost all the architectural components, the data col-
lected from this system, in conjunction with the interaction with users, can provide
suggestions for improvement. Our experiment in the digital wood block is primitive
example of this idea. In the future, these suggestions may include a smart manage-
ment of natural heterogeneous material, the reduction of energy consumption, the
efficient utilization and modification of space, through visual navigation predicted
by simulations. This process of Timber “Metabolism” requires the use of compu-
tational geometry for the continuous remodeling of the architecture to be built and
ubiquitous computing devices built in the architectural elements.
Acknowledgements The series of study in this paper have been done by Keio university Ikeda
Laboratory with many students and researchers:
Uich Abe, Kensuke Hotta, Yasushi Sakai, Yusuke Takami, Akito Hotta, Haruna Okawa, Shigeru
Miyako, Minori Aoyama, Jungseop Kim, Yoko Hagiwara, Ai Tanaka, Azusa Ino, Ahn Suh June,
Toshiko Shibuya, Terumoto Mizushima, Wataru Dio, Akihiko Tanigaito, Ririko Mizukami, Masaya
Simizu, Mai Yanagida, Togo Takagi, Takeshi Sato, Shinya Azuma, Yuma Mitsui, Kei Watan-
abe, Hikari Watanabe, Jung Daeho, Hikaru Ikeda, Tong Xing, Takaharu Suzuki, Shion Iida, Ran
Tateyama, Takuhide Hukuyama, Joaquim Silvestre, Gen Karouji, Jejong Lee, Erika Takeda, Shyam
S. Sirimalla, Mariko Muto, Erika Takeda, Yoshimi Takeda, Kurara Matsui, Hisayuki Ogiwara, Xiao
Yahan, Naoki Imanishi, Tan Xin, Aqui M. Cheddadi, Tabi Salma.
1408 Y. Ikeda
References
Carpo M (2011) The alphabet and the algorithm. The MIT Press
Kawazoe N et al (1960) Metabolism 1960: the proposals for a new urbanism. Bitjsutu Shuppan Sha
Woodbury R, Johnson B, Jabi W (2013) Parametric design for architecture. Laurence King
Publishing
Parametric Modeling of a Wooden
Folding Structure
Tuğrul Yazar
Abstract A-Chord is an experimental design project of a wooden structure, pro-

duced for the exhibition at the World Wood Day 2015 Festival in İstanbul. It is a
design family of folding structures, which are non-standard, light-weight, portable,
and re-usable. In this article, we describe a multi-model design process utilized in this
project. Then, we explain the application and feedback of the geometric, physical,
structural and parametric models. Finally, we introduce the production of a full-scale
prototype.
Keywords Parametric modeling · Light-weight structures · Folding structures

Contouring geometry · Wood
1 Introduction
A-Chord is an experimental design project of a wooden structure, produced for the

exhibition at the World Wood Day 2015 Festival in İstanbul (Fig. 1). The festi-
val is organized by the International Wood Culture Society (IWCS) with an aim to
gather designers and artists from around the world to celebrate their wood cultures.
A-Chord is designed and produced by an İstanbul-based design-research group POT-
plus, established by two architects and academics; Fulya Akipek and Tuğrul Yazar.
POTplus focuses on the contemporary methods of design and production, and studies
the relationship between nature, architecture and urban life through computational
design projects. In POTplus, we design and build experimental structures for exhi-
bitions, festivals, biennials, and workshops. We aim at transforming open public
spaces into fields of interaction and translate this information to architectural knowl-
edge through publications. In a previous paper, we introduce our design-research
approach underneath several projects including A-Chord (Akipek and Yazar 2015).
T. Yazar (B)
Faculty of Architecture, İstanbul Bilgi University, İstanbul, Turkey
e-mail: tugrul.yazar@bilgi.edu.tr

1410 T. Yazar
Fig. 1 Final installation of A-Chord
In this article, we take a closer look at the parametric modeling workflow of A-Chord
project.
The main objective of A-Chord project was to re-emphasize the innovative uses
of structural wood by a parametric modeling workflow. We aimed at creating a fam-
ily of non-standard and multi-functional structures which are light-weight, portable
and re-usable. These preliminary design decisions led us to analyse several exam-
ples with similar purposes. One of them can be seen at the doors of the Ernstrings
Warehouse, designed by Santiago Calatrava (1985). It is possible to study the basic
geometry underneath the flat-foldable system in his design. Another example is the
Rising Chair, designed by Robert Van Embricqs as a part of his furniture collec-
tions (Van Embricqs 2012). This project reveals further potentials of a light-weight
wooden structure, utilizing the same geometric principle introduced by Calatrava.
In both projects, it is possible to analyse the ruled surfaces, generated by the curved
paths of folding. These examples prove that a kinetic structure can be developed by
combining the geometry of contouring and the performance of folding. In A-Chord
system, wood was used as a light-weight material which is an essential quality of
folding performance. A-Chord has several unique qualities compared to the prece-
dents. We took the contouring-folding method forward, by implementing additional
folding paths while challenging to create a self-standing structure. As a spatial qual-
ity, we intended to create an interior volume by folding two layers of contour lines
in opposite directions. This made the design suitable for multiple functions; such
as separators, space generators, or structures for climbing plants. Instead of a tradi-
tional representation and production workflow, we developed a parametric modeling
workflow, which can be used to generate variations of the structure along with the
production documents in real-time.
Today, design models are becoming more utilized in discovering, learning and
hence exploring beyond the means of representation (Sorguç and Selçuk 2013). The
Parametric Modeling of a Wooden Folding Structure 1411
possibility of managing complexity is strengthened by the fact that more than one
model can be used in parallel with the others, and the transition possibilities between
the models are increasing. In this article, we explain our design workflow through
several models; which are geometric, formal, structural and parametric. These models
should be considered as the generators of new parameters for our design process,
rather than different media for representation. In A-Chord project, these models
were organized to give feedback and complement each other. Geometric models
were used to reveal the mathematical basis of the design system. The feedback from
these models were helpful in creating a series of scale models. Scale models revealed
design parameters related to form-finding by simulating the folding motion. However,
it didn’t give feedback about the structural capacity of the system. This is why, a
structural model was developed to test the system under gravitational force. This
enabled us to determine the vital details for the self-standing performance. Finally,
a parametric model was developed by considering all of the feedback gathered from
other models. This parametric model could be used in the design and production of
a full-scale prototype of A-Chord. Below is the detailed information of these models
and their relationships with the design process.
2 Geometric Models
Geometric modeling method of the A-Chord system is called “abstract explication”

(after Yazar 2015). In this process, we aimed at constructing the basic relation-
ships of the system via Euclidean geometry. This was accomplished by defining and
labeling distances between critical points of the subject, instead of modeling singu-
lar outcomes. The initial mathematical basis of the A-Chord system was developed
through its section drawing (Fig. 2). After several experiments, the primary geometric
components of the design could be explained as:
• Contour line segments,
• Fixed points,
Fig. 2 Geometric model of A-Chord system

1412 T. Yazar
• Movable points.
The construction of the section was started by drawing AB, which is a single
contour line perpendicular to the ground line g. We split this line at point S, creating
two line segments, AS and SB. Point A was defined as a “fixed point”, and point
B was defined as a “movable point”. The basic folding motion was achieved by the
rotation of S around A without changing the lengths |AS| and |S B|. In order to solve
this problem, we used the Euclidean compass and ruler, which is an ancient method
of geometry (Martin 1998). First, we drew a circle centered at A with the radius of
|AS|. In order to determine the location of S on this circle, we changed the position
of point B by moving it on the y direction along the contour line. After that, we set the
compass to measure the length of |SB|, placed its needle at point B and drew another
circle (Fig. 2a). The intersection point of these two circles gave the new location of

our movable point, S . This
enabled
us to fold
the line segments AS and SB into AS
∼
and S B , where |AS| AS and |SB| S B . ∼
Although this geometric construction looked promising, it was not suitable for a
self-standing structure. In order to continue the design process, we elevated the fixed
point A from the ground line g, creating another movable point, C. After locating
points S and B as described above,
we located the point C by extending the new line
segment AS so that |AC| ∼

AC . A similar extension was repeated

at the other end
of the contour line, creating points D and D so that |BD| ∼ B D (Fig. 2b). Finally,
we finished the section drawing by adding a second set of points and segments on
←→
the other side of the AB axis (Fig. 2c). This led us to define point S2 . Then, a single
motion of movable point B would create points S, C and D at both sides of the axis.
The process explained above enabled us to create a three-dimensional digital model
of the A-Chord system. We multiplied all of the line segments and points (Fig. 2d) in
the direction perpendicular to the section plane. We defined fixed and movable point
sets of A, B, C, D, and S.
The feedback from the geometric modeling process revealed several rules regard-
←→
ing foldability, balance,
and
transportability
of A-Chord system. AB⊥g, therefore
line segments AC and AC2 should be extended to intersect g. Total segment
lengths (weights of the materials) should be similar at each side, meaning that,

AC + AS + S B + B D ∼
AC + AS + S B + B D
2 2 2 2

In its natural state, above equation is always true. However, when we extend AC

and AC2 , this would mean that we should also extend B D and B D2 with the
same amount. Moreover, the area centroid of AS B S2 quad would preferably be
← →
on the AB axis to guarantee that the structure is balanced in a self-standing state.
Another geometric rule is about the proportions of the sections. In order to keep the
system stable and safe, the distance between points C and C2 should be above 60 cm.
and the distance between points S and S2 should be below 90 cm. for a 200 cm-high
structure. At this stage of the research, we needed to use other types models to test
and verify these rules and assumptions. As a result of the geometric modeling, we
managed to understand the basic principles of contouring and folding through their
transformation in the section plane. It proved that the whole system can raise from
flat position to a three dimensional and volumetric state with the motion of a single
point set.
3 Scale Models
After studying the geometric properties of A-Chord, we developed several scale

models to obtain more information about its joinery detail and folding motion. The
scale models were made from foamboards, and needles and threads were used to
join the contours. These models revealed that the sets of fixed and movable points
could be organized to create continuous curves, instead of straight lines. This is why
we named those sets as “fixed curves” and “movable curves”. The geometric model
was updated according to this information (Fig. 2e). It was possible to experiment
the formal relationship between these curves and the overall shaping of the ruled
surfaces. Some of the preliminary scale models are shown in Fig. 3.
After first experiments, we observed that the system has a tendency to bend, turn,
and tilt when folded. This proved the necessity for extending the bottom ends C
and C2 to reach to the ground, keeping the structure in standing position. Another
feedback of the scale models was about the joinery detail of the system. In order to
hold the contour lines together, we needed a strong but flexible joinery detail that
passes through the fixed and movable curves. This enabled us to create continuous
surfaces along the multiple contour lines when folded. These models revealed that it
was possible to use steel wires in 1:1 product. However, scale models didn’t provide
enough information about the structural performance of the system.
Fig. 3 Some of the physical scale models created during the design process
1414 T. Yazar
Fig. 4 Structural simulation of the geometric model
4 Structural Model
The limitation of the scale model was the lack of any material feedback. Our choice
of scale and materials would mislead us about the real structural performance of
the system. This is why we developed a structural simulation. The geometric model
was revised by considering the feedback from the scale models. We transferred the
geometric model to Grasshopper, a plug-in for Rhino, by defining fixed and movable
points and placing loads on the segments. Then, we simulated the folding behavior
under constant gravitational force by using Kangaroo add-on for Grasshopper. This
simulation showed that the unlimited angle of rotation at the movable curves would
cause the structure to rotate excessively, and eventually causing to collapse on itself
(Fig. 4). As a vital feedback from the structural model, we realized that there should be
a limit on the amount of rotation, which could be determined by a maximum distance
between opposite points. We revised the geometric model and put additional tension
wires across some of the contour lines, between points S and S2 . The length of this
wire would determine the maximum distance between the contour lines it is attached
to, stopping the folding motion at a desired angle.
5 Parametric Model
After making a final revision of the geometric model, we developed a parametric

model in Grasshopper. This model was used to generate several instances of A-Chord
system, and test their validity according to the rules we have defined.
The parametric model of A-Chord consists of three clusters (Fig. 5). In Cluster A,
main input parameters of the system, such as the number and thickness of contour
lines, and fixed and movable curves were defined. We determined the maximum and
minimum values for these input parameters. Fixed and movable curves are defined
as function graphs in Grasshopper. This method is used to create smooth curves by
evaluating continuous functions. In Cluster B, the geometric components were drawn
according to the parameter inputs of Cluster A. We used the construction method
explained in the geometric modeling section. The aim of this cluster was to determine
the locations of all fixed and movable points and draw the segments with correct
thicknesses. This cluster was also used to simulate the folding motion by elevating the
Fig. 5 The parametric model of A-Chord
Fig. 6 Simulating the folding motion in the parametric model
movable curve B in the z-axis (Fig. 6). Combined with the input parameters in Cluster
A, this enabled us to generate instances of A-Chord quickly and precisely. In Cluster
C, the material cost and product documentation were calculated and generated. All
of the wooden rods are sorted, measured and labeled by this cluster. At the same
time, dimensions, weight, and the approximate cost were calculated. Lengths of the
steel wires attached to the fixed and movable curves are sorted, measured and labeled
for a seamless production process (Fig. 7). We used the output of this cluster directly
in the production of a full-scale prototype.
1416 T. Yazar
Fig. 7 A detail from the product documents generated by the parametric model
6 Full-Scale Prototype
After finalizing the parametric model, we generated several variations of A-Chord

(Fig. 8) in order to choose one to be build in full scale. Although we designed the
structure and generated production documents by using the parametric model, the
actual construction of the prototype is realized with the help of traditional carpentry
(Fig. 9).
First, we clarified the material and detail choices. Type and thickness of the wood,
hinges and steel cables are finalized. Total of 100 segments made of 40 mm × 40 mm.
beech rods were measured, cut, and drilled according to the measurements given in
the production document. Steel wires were inserted and stretched by a wire tensioner
at both ends. Small hinges were placed on the moving points (S). The assembly
was finished while the structure was lying flat-unfolded on the ground. After these
preparations, the structure was raised and suspended temporarily. Steel wires were
measured and attached at the appropriate locations to limit the folding motion at a
desired angle. The final product was 200 cm. in length, approximately 80 cm. in
width and 220 cm. in height. Two people were able to unfold, fold and carry the
Fig. 8 Variations generated by the parametric model

Fig. 9 Full scale production process
Fig. 10 Final installation of A-Chord at World Wood Day 2015 Festival in İstanbul
structure to its location at the festival area. As a result, the system could be raised
from a flat-unfolded position to a self-standing state in seconds (Fig. 10).
A-Chord performed as a foldable structure without a pre-defined function. After
its installation at the festival area, wood artists and park users enjoyed sitting nearby
its shadow, took photos with it, and the cats passing through the volumes rested under
1418 T. Yazar
it. The parametric model enabled us to extend the structure along an axis, while it
could be seen as a 3D morphing pattern. It showed potential multi-uses for gardening
facilities such as creating a 3D border and at the same time performing as a support
system for climbing plants.
7 Conclusions
A parameter is generally described as a piece of information that determines the

trends of a system. Sometimes it may appear irrelevant to the modeled system at first
glance. But when the system is reasoned in a wider context, a parameter is revealed
as a variable of a superstructure. In contrast to a popular belief, “parametric” design
is not limited to only generating variations from an associative model. In essence,
parametric design is not only a method, but an attitude that emphasizes the continuous
attempts of reasoning, with an aim to create and discover top-level relationships.
Parametric models encourage designers to create and manage geometric relationships
and search for top-level information that governs or explains these relationships.
This may seem to be a one-way process of clarifying ambiguity, but the attitude
explained above does not necessarily aim at finding an optimum solution. Instead,
it aims at extending the possible input and output sets in a design cycle. This is
generally realized by searching for new parameters such as structural, ecological,
social, historical, formal, etc., and developing a parametric model that associates
them.
Today, prototypes are used as a part of the research cycles in the field of archi-
tecture (Akipek 2010). The A-Chord prototype was an outcome of a multi-model
workflow. In this workflow, the parametric model served as a base, where the infor-
mation gathered from other models were processed and applied to the design. In
line with above ideas, A-Chord project revealed new potential parameters. Sun,
wind and topographic information could be implemented as potential inputs to the
system. Material and joinery details and the control over the folding motion could
be advanced. The scalability of the system offers various functions from industrial
design to architectural and urban design disciplines. The potential parameters from
those disciplines would further advance the design family. More researches are nec-
essary to investigate the various scales of folding-contouring method presented in
this article. For example, the structure could be tested in larger scales for architectural
spaces, or smaller scales for industrial design objects. The same parametric model
developed in this article could be used by all those disciplines.
Acknowledgements A-Chord is designed and produced as a part of an exhibition of International

World Wood Day Festival at Nişantaşı Sanat Park, İstanbul, Turkey in March 2015. This festival
was the 21st World Wood Day Event organized by Chung-Yun Hse, president of the International
Wood Culture Society (IWCS). A-Chord was chosen from among several design alternatives, which
are proposed by different designers in a two-day workshop. This workshop was organized together
with Simge Esin Orhun and Fulya Özsel Akipek.
More information about A-Chord and the other projects of POTplus can be found at www.potplus.
org.
References
Akipek FÖ (2010) Working with digital and physical prototypes in architectural design studio
with the search for innovation. In: Spiridonidis C, Voyatzaki M (eds) Proceeding of EAAE 2010
İstanbul: educating architects towards innovative architecture, pp 119–131
Akipek FÖ, Yazar T (2015) Prototypes in the recycling of design knowledge: A-CHORD a proto-
type experience for wooden folding surfaces. In: MSTAS 2015 conference proceedings: sustain-
able computational ecologies: computational design, material, performance, Özyeğin University,
pp 69–79
Calatrava S (1985) Ernstings warehouse. https://calatrava.com/projects/ernstings-warehouse-
coesfeld.html. Accessed 30 May 2018
Martin GE (1998) Geometric constructions (Undergraduate texts in mathematics). Springer
Sorguç GA, Selçuk AS (2013) Computational models in architecture: understanding multi-
dimensionality and mapping. Nexus Netw J: Archit Math 15(2):349–362
Van Embricqs R (2012) Rising chair. https://www.robertvanembricqs.com. Accessed 30 May 2018
Yazar T (2015) Design of dataflow. Nexus Netw J: Archit Math 17(1):311–325
Common-action Gardens: Performative
and Parametric Prototypes
Fulya Akipek
Abstract Common-action Gardens are wooden park structures that were featured
within a series of open-air exhibitions, the 3rd International Architecture Biennial
Antalya in 2015, and the Beşiktaş International Garden—Flower Festival İstanbul in
2016. These selected art works had been built in the city parks to be exhibited during
the festivals. They are still being used by the communities as urban gardens to grow,
harvest and share edible plants, rest and play, and sustain flora-fauna. The POTplus
Design Research Group has constructed these structures as architectural prototypes
in their research to integrate digital design and fabrication technologies with sustain-
able landscape issues. The specific realms within the scope of this research process
are parametric design, performance-based design, digital fabrication, permaculture
studies and material exploration. Common-action Gardens I and Common-action
Gardens II are constructed with the same composite material system, comprising
recyclable materials of beech marine plywood and ethylene-vinyl acetate. They are
designed to combine elements of permaculture and integrate these elements with
the needs of the park users within a complex structure. In the form-finding process,
contouring techniques facilitate the integration of various elements of permaculture
within one continuous structure and fast assembly, while nesting operations and CNC
milling enable minimum waste of materials. These two urban gardens are designed,
fabricated and assembled with the same design-built system based on contouring but
in order to adapt to different contexts and to enhance their performative issues they
have differentiations. This chapter covers the parametric design approach, digital
fabrication and assembly processes of these urban gardens to compare, and discuss
the performance-based design strategy behind these processes.
Keywords Performance-based design · Parametric design · Digital fabrication

Contouring · Urban gardens · Architectural prototypes
F. Akipek (B)
Faculty of Architecture, İstanbul Bilgi University, İstanbul, Turkey
e-mail: fulya.akipek@bilgi.edu.tr
URL: http://potplus.org

1422 F. Akipek
1 Introduction
We are witnessing an era in which we experience the ecological crises and advanced
technological opportunities simultaneously. While ecological issues such as climatic
change, energy crisis or waste production problems affect our everyday life, con-
currently with this, rapid technological advances reveal alternative and innovative
methods for any kind of production. Among these crises and opportunities, the neg-
ative effects of the building and construction industry on the environmental issues
are increasing rapidly, and this presents the necessity for a sustainable architecture.
As Menges claims (Menges et al. 2017), the building sector faces the very serious
need to substantially reduce its environmental impacts, as they contribute a major
part to global energy consumption, carbon dioxide emission and waste production.
Computational design and digital fabrication technologies can be used in a way
that enable architects to develop ecologically sound and economically viable strate-
gies intended for a sustainable architecture. Recent technological advances in compu-
tation, simulation and fabrication offer possibilities to expand the use of environmen-
tally appropriate and energy efficient building materials via developing alternative
design, manufacturing and assembly processes.
Together with Tuğrul Yazar we founded the POTplus Design Research Group, to
integrate digital design and fabrication strategies with issues of sustainable architec-
ture and in-specific with permaculture design principles.1 Within this research we
construct 1:1 scale architectural prototypes to experience the use of these technolo-
gies in design-built processes. These prototypes also help to get feedback data about
their real life uses, structural strength, material behavior and their long-term perfor-
mances on site. In our design research processes, we associate three basic knowledge
domains with each other: Parametric and performance-based design-digital fabrica-
tion; permaculture design and material performances.
Parametric design and digital fabrication enable setting a design and construction
process within a feedback relation with each other (Yalınay et al. 2008). In a paramet-
ric design approach designer defines the relationships between elements and assign
values or expressions to organize and control those definitions (Dunn 2012). In our
design research process, we compute the quantities in the early design phases, such
as amount of waste material, weights and material thicknesses, to control costs or
structural strengths. By using parametric design, we can include the quantities related
with the natural environment such as sun path, wind flow, water flow, and the life-
cycles of plants to associate these parameters with the anthropomorphic quantities
and material properties. We explore coding and simulation platforms to investigate
the consistency of the geometric decisions with the performance criteria. Digital fab-
rication technologies including 3D printing, CNC milling or robotic systems enable
geometric explorations in a relation with material behaviors and constructing issues.
Parametric design also helps to adapt the same design system to different contexts.
As Dunn (2012) claims, in a parametric system the designer may alter the values or
1 http://potplus.org/. Accessed on 29 July 2018.

Common-action Gardens: Performative … 1423
equations that form the relationships between elements and effect of these changes
will be incorporated across the entire system.
Permaculture is an integrated design system that uses natural systems as a model
for creating sustainable human environments in relation with flora and fauna. It is a list
of principles and a guide to support providing your own needs without destroying the
natural systems (Mollison 2012). Permaculture uses the inherent qualities of plants
and animals combined with the natural characteristics of landscapes and structures
to produce a sustainable life in the smallest practical area. We evaluate and use basic
principles of permaculture within a range of the site plan decisions, material choices,
and even decisions about the program and use.2 Permaculture design, in its narrow
sense as a guide for sustainable agriculture, defines basic structural elements for its
applications such as raised bed, compost, rainwater harvesting pot or herbal spirals
and inspires us to associate them with our design systems.
Material performances of traditional materials used within the built environment
and local construction techniques are expanded with the uses of digital fabrication
technologies. In the scope of our design research, we analyze local building materials
and systems to sustain their use, such as wood or rammed earth and manipulate these
techniques with digital design and fabrication technologies in an innovative way. The
material systems that we use also serve for an architectural position integrated with
ecological and natural cycles.
To associate these three knowledge domains, we work within an interdisciplinary
approach from the early phases of design and collaborate with people from various
disciplines according to the design brief that we have evaluated. Our team of con-
sultants involves people from the disciplines of permaculture and landscape design,
local building materials, building physics, sustainable infrastructures, industrial fab-
rication, cultural studies, and interaction design and in more recent works, renewable
energy systems, waste management and ecological urbanism. This diversity of dis-
ciplines, even on the scale of building a prototype, demonstrates that an ecological
architectural approach entails relational thinking and a working platform that feeds
from various disciplines. These profiles should be extended to consultants from the
social sciences and to each stake-holder, people from municipalities, or users, shifting
towards a more collaborative and common-action platform.
As full-time instructors, we also integrate our research inputs or outputs with the
graduate and undergraduate studios and courses in İstanbul Bilgi University, Depart-
ment of Architecture. İstanbul Bilgi University, Faculty of Architecture, has a unique
position among other faculties of architecture in İstanbul with its fabrication labora-
tory and with its curriculum that involve design, technology and material research.
This laboratory equipped with laser cutters, 3D printers, CNC routing and milling
machine, robotic fabrication and various crafting machines, facilitates our research
through computation and material studies. Within the summer school studios and
workshop series, our department has been collaborating with various International
Schools of Architecture including recent visits from AA Visiting Schools and ETH.
Our architecture department also took place in the 16th International Architecture
2 https://permacultureprinciples.com/principles/. Accessed on 29 July 2018.

1424 F. Akipek
Fig. 1 Common-action Walls, 4th International Biennial of Architecture, Antalya, October 2017
Exhibition in the Venice Biennial, with a student pavilion. It was one of the final
projects in our first year computation based basic design studio, coordinated by
Şebnem Yalınay Çinici, head of the department.3
As an example in this efficient feedback relation between our academic works and
the POTplus design research projects, our recent prototype Common-action Walls,
opened up the way to the Graduate Summer School Studio: Robotic Earth Crafts.
Common-action Walls is the recent prototype that we built for the 4th International
Architecture Biennial of Antalya in 2017. This prototype is another example of our
research about integrating ecological studies with digital design and fabrication.
This rammed-earth wall performs as a vertical planting-wall to grow edible plants
and share the harvest with park users. It has a drip irrigation system integrated into
its design-assembly system which enables a minimum use of water sources and earth
is obtained from construction excavation leaving no disturbing construction-waste at
the end of its use. The traditional rammed-earth building technique, generally used
for building massive walls, is shifted towards a modular permeable wall, having
continuous holes and surfaces for planting. Parametric models and performance
studies about Gyroid geometry, molding studies and digital fabrication processes
enabled these rammed-earth blocks to be produced in a facade factory and carried to
its place for the assembly phase, which took only two days (Fig. 1). This prototype
took place among the sustainable architecture conference small projects category
shortlist, and was awarded an Honourable Mention by S.ARCH 2018 Conference in
Venice.4
The outputs of this research guided our summer studio, Robotic Earth Crafts
in İstanbul Bilgi University (Fig. 2). The molding system shifts towards robotic
fabrication producing non-standard molds cut from expanded polystyrene (EPS)
blocks with a hot wire cutter integrated with the robot arm.5
Common-action Gardens I and II are initial prototypes of this design research
approach. Our research group had been invited to participate in an architectural bien-
3 https://www.bilgi.edu.tr/en/news/8002/bilgi-faculty-of-architecture-attends-16th-venice-
architecture-biennale/. Accessed on 29 July 2018.

4 http://s-arch.net/SARCH18-180606-versetzt.htm. Accessed on 29 July 2018.
5 https://mimarlikyl.bilgi.edu.tr/tr/mimarlik-tasarimi/etkinlikler/robotik-earth-crafts-57/. Accessed
on 29 July 2018.
Fig. 2 Robotic Earth Crafts, İstanbul Bilgi University Summer School Studio, İstanbul, 2018
nial in 2015 and a garden festival the following year. We used these two occasions to
create urban garden structures, as test prototypes within our research. Urban gardens
are shared gardens in the city to grow edible plants, share the harvest, preserve local
flora and fauna by creating an ecological area in the city and are powerful mediators
to create a common-public space via urban farming. These two urban garden struc-
ture prototypes are constructed onto formal urban parks to criticize the unsustainable
ways of landscape design and maintenance of these parks. The first prototype, Com-
mon-action Gardens I (Fig. 3) was constructed for the 3rd International Biennial of
Architecture in October 2015, which is still being used in Karaalioğlu Park, Antalya.
The second prototype, Common-action Gardens II (Fig. 4) was constructed in Etiler
Sanatçılar Park, for the Beşiktaş International Flower Festival in İstanbul. It has also
become part of everyday life.
The common method to construct these urban gardens is to use upcycled or
recycled materials such as wooden pallets or tire wheels and share design details by
open-source documents. This seems an economical way of constructing elements
such as raised beds for planting, or compost bins in community gardening thus most
of these urban gardens are neither a part of designed urban landscapes nor included
in land use strategies because of their non-design approach. The unique focus of
the Common-action Gardens projects was adding positive value to the urban garden
concept and collect attention through design, which is enabled by exploring form
and material in accordance with performance criteria. Parametric design helped to
set a design system to integrate elements of urban gardening into one continuous
structure and adapt this design system to different contexts by creating variations of
it.
1426 F. Akipek
Fig. 3 Common-action Gardens I, 3rd International Biennial of Architecture, Antalya, october

2015
Fig. 4 Common-action Gardens II, Beşiktaş International Garden-Flower Festival, İstanbul, May
Both of the prototypes are designed via the design and fabrication technique
called contouring (Fig. 5) that is used for various examples of design and architec-
ture around the world. In his book Digital Fabrication in Architecture, Dunn (2012)
claims that implementation of CNC milling and routing enables the designer to sys-
tematically remove material through a series of carvings or contours. CNC methods
are capable of quickly producing a greater number of either non-standard or repetitive
elements and contouring facilitates a clear dialogue between the digital design infor-
Fig. 5 Contouring as a design method used in Common-action Gardens I and II
Fig. 6 The composite material system comprised of wood and EVA in Common-action Gardens
mation and the making process. This design method-in our case- enables creating
a three-dimensional complex form out of sheet materials via slicing them into two-
dimensional sections. It also helps to gain control on the sections that are informed
by various input parameters mentioned above. This design technique has weaknesses
when the amount of the material used is compared with the usable volume it creates,
but has advantages about the issues of design such as the strength of the structure,
geometrical precision, and modular assembling opportunities.
Menges et al. (2017) claims that wood is experiencing renewed interest as a
construction material with unrivalled ecological virtues holding a very low level of
embodied energy, and having a positive carbon footprint. Recent advances in design
computation, simulation and fabrication offer the possibility to revisit the specific
nature of these materials and computation plays a key role as a decisive enabling
technology equipping both the designer and the maker. Common-action Gardens are
made of a recyclable material system that is composed of beech marine plywood and
ethylene-vinyl acetate (EVA) panels. These two layers line up to shape the complex
forms and also reveal various surface performances for park facilities, ventilation or
water flow. Wood panels enable structural rigidity while EVA panels as a waterproof
infill material are used to perform lightness and elasticity of the structure, and also
softness for body comfort (Fig. 6).
10 mm thick beech marine plywood panels and 30 and 40 mm thick EVA panels
are used as the basic materials of the construction system. The shape and size of
each panel within the complex forms is unique. They are cut by using CNC milling
1428 F. Akipek
technologies. These panels are assembled with long stainless steel anchorages from
appropriate points to enable parallel working on the assembly phase.
These two structures of contouring design technique and composite material sys-
tem differ in their geometric formation and material performances. The first garden
structure is formed as a solid that has holes in it to reserve space for permaculture
elements. The continuous free-form surface unites these holes of various dimensions,
involving edible-plants and compost and creating niches for animals. Common-action
Gardens II conceptualized as fibers is elongated along x axis with a strategy to have
more space for planting, enabling the diverting of fibers to adapt to the existing park
landscape. These concepts are also triggered by the various performances of EVA.
These two prototypes can be analyzed from various points such as their weight, ratios
of waste material, effectively used surfaces, time for fabrication and assembly (which
will be discussed in conclusions). Working with parametric software and digital fab-
rication, a designer can be aware of quantities to support qualitative factors, such
as sustainable landscapes in our case, and also explore latent properties inherent in
traditional materials such as wood.
2 Common-Action Gardens I
This first prototype demonstrated the integration of basic permaculture elements and
park furniture within a continuous structure in the 3rd Antalya Architecture Biennial,
with the theme “Thinking the Future” (Fig. 7). Common-action Gardens was built
on the concept that the future architects should approach design as an additive value
to serve for a sustainable environment in this crisis era This garden structure covered
holes for edible plants, flowers, a fruit tree, compost, gardening tools, and seed pots.
Besides this, it served as park furniture to rest on, and come together for farming
and harvesting in the city (Fig. 8). This structure performs as an ecologic system
using the output of the organic waste of the trees, plants and park becomes input for
compost to become soil again. Rainwater can be collected and directed to the plants
by the surface curvatures. Park users can share the harvest from edible plants, and
fruit tree and rest on the surfaces connecting these holes.
2.1 Design
The basic elements integrated in this structure required various dimensions and
heights according to permaculture principles. Raised beds as elevated planting areas
to be filled with appropriate soil require a height of 30–40 cm while compost requires
a minimum 90 cm. height to cover the organic waste that should be ventilated with
wind flows. The tree hole should have the appropriate dimensions and height for
its root development. In the first sketches, we agreed on creating ellipsoid holes,
Fig. 7 Planting of the Common-action Gardens I
Fig. 8 Elements of permaculture integrated within Common-action Gardens I and holes in plan
subtracted from a solid model like meta-balls or mercury drops blending within each
other. The first hand-sketch was the guide for the amount of these holes and geometric
explorations.
This first sketch helped for the overall geometry to create the mass model and the
solid-void relations. This model is developed via lofting main control curves which
organize the top, bottom and middle axes of the structure. This initial model is devel-
oped by using the contouring method in Rhinoceros software to enable constructing
this three dimensional free-form from two dimensional sheet materials. The material
system comprised of 10 mm bench wood and 30 mm EVA identified the contouring
distance. The digital model was sliced into 40 mm. pieces parallel to each other.
These control curves are defined and associated with each other and performance
criteria mentioned above by the Grasshopper plug-in and on-line control about the
1430 F. Akipek
Fig. 9 Parametric model built by Grasshopper plug-in to control the curves
Fig. 10 Various surface characters: 1. smooth surface, 2. elastic surface, 3. channeled surface
constructing issues such as weight of the structure, costs and dimensions affected
the final decisions at that phase (Fig. 9).
The two layers of the material system identified different surface characters in
relation with each other (Fig. 10). That was the basic parameter for the last phase,
fine-tuning of the model. When wood and EVA planes adapted to each other and
the height of the plane adjusted to the plane next to it, they form a smooth surface
appropriate for body comfort to sit, lie, rest on; when EVA is in the front of the wood
sections, there is a softer and elastic surface that is adjustable and comfortable for
animal niches and for adding plug in objects such as seed boxes; when wood is in
the front of the EVA sections, there are channels for water flows or a hard surface
character not to be used for bodily functions.
The elevated curves of EVA sheets, forming the bottom axis line of the second
layer, reserved channels for bottom ventilation and air flow necessary for farming
issues such as composting. These smooth, soft and hard surfaces created various
surface characters mapping the available uses and performances. Control curves
were set as inputs of the parametric model identifying the overall geometry. In parallel
with these form-find explorations, this parametric model enabled the computing of
the weight of the structure, usable surface and cost issues in real time.
Fig. 11 Assembly strategy
2.2 Production and Application
After this contouring operation of the final model, nesting and labeling of the plane
shapes are automated by Rhinoceros software. All of the unique shaped wood and
EVA planes were cut using CNC milling in two days. Each plane had three anchorage
holes on it to enable the assembly operation by using stainless steel anchorages. The
main strategy for application is based on parallel working on site by an application
simulation. Starting from a section taking place at the middle axis should let three
groups assemble the other two axes in opposite directions at the same time. All of
the assembly operation on site is completed in two full days. The production and
application process of Common-action Walls prototype 1 proved that we need to
develop the strategy by reducing the ratio of wood use, enlarging the planting holes
and reducing the waste of materials (Fig. 11).
2.3 Performance
Material weights are calculated as 83% wood, 17% EVA and stainless-steel anchor-
ages (Table 1). When we analyzed the facility areas, we observed that to create one
square meter of usable surface we used 15 kg EVA and 303 kg wood. This analysis
showed that an increase in EVA use should develop the system towards lightness and
an increase in usable areas. Usable surface area parameter is about the upper surface
created by the additive lining up of wood and EVA. We wrote a code which is based
on a draft angle analysis method to calculate this area of useable surfaces that is used
in molding technologies. Another parameter that can be analyzed through Table 1, is
the waste of materials, a performance criterion to be tested. The automated nesting
and labeling operations that enable the maximum condensed shape configuration of
the wood and EVA planes to be cut out from a sheet material for CNC milling opera-
tions, resulted in 58% effective use of EVA planes and 65% effective use of wooden
1432 F. Akipek
Table 1 Calculations about the basic parameters of Common-action Gardens I (Akipek et al. 2016)
Materials Wood EVA Anchorages Nuts Total
Dimensions 1700 × 1000 × 10 M10
2200 × 1500 ×
10 mm 30 mm
Number of 133 176 27 555
planes
Weights per 9.0 kg/m2 2.1 kg/m2 0.62 kg/m 11.4 g
square meter
Total area 43.77 m2 28.80 m2 21.22 m
Total weight 393.9 kg 60.5 kg 13.2 kg 6.3 kg 473.9 kg
Total 341.3 m 314.5 m 655.8 m
perimeters
Total area 3.41 m2 9.44 m2 12.85 m2
Usable 1.95 m2 5.82 m2 7.77 m2
surface area
planes. We identified that we should find out computational methods to maximize

the effective uses for a sustainable way of fabrication.
3 Common-Action Gardens II
This second prototype is realized for the Beşiktaş International Garden-Flower Fes-
tival İstanbul, in May 2016 (Fig. 12). Using the same design strategy and material
system, we aimed to enlarge the planting holes and reduce materials as well as the
ratio of wood use. We had a larger area in this new park site so we aimed to respond
to existing environmental factors such as existing bushes and trees. We also aimed to
integrate more elements of permaculture by adding a rainwater collecting strategy.
During the geometric explorations, we arranged surface angles in order to direct
water flow and collect excess irrigation water and rainwater into a small pool to be
used by animals for drinking water. All of these design objectives shifted the concept
of holes towards fibers. Common-action Gardens II is developed within the concept
of diverting and branching fibrous walls connected at some points to create enlarged
planting areas and surfaces to serve for park facilities as resting areas.
3.1 Design
Now that we decided to have a free linear form with this flow of fibers, we integrated
five elements of permaculture taking place within this continuous structure: Water
Fig. 12 Common-action Gardens II, built for Beşiktaş International Garden-Flower Festival, May
2016
collecting pool, herbal spiral, raised beds, compost and a tree hole (Fig. 13). We
first analyzed environmental issues such as existing landscape of the park, wind-sun
directions, and relations with the pedestrian ways to make a site plan sketch. With
this concept of fibers in mind, we planned to start with a bundle of fibers, diverting
into four branches in different directions and connecting at a node to create 3 planting
areas for herbals. Those four branches adjoin again to create resting surfaces. This
bundle diverts into three branches to create two areas of raised beds and, at the last
region, divert again into four branches to create two different areas separated from
each other. The first area is for tree planting and the other for the compost structure.
This branching algorithm of fibers helped decisions about site plan for, but to
model the required heights and surface curvatures we created another control surface
to integrate these heights (Fig. 14). Water harvesting strategy also gave input data
about these changing heights for minimum irrigation, sustaining water to flow to the
other planting areas and for directing excess water to the water harvesting pool. A
parametric model helped to get real time information about the amount of materials
used, usable space and cost issues.
3.2 Production and Application Process
Following the same strategy, the final model is contoured, nested and labeled. The
difference in the design method was shifting one-direction contouring that resulted
in parallel sections. In this second prototype, contouring of the branches was realized
1434 F. Akipek
Fig. 13 Elements of permaculture integrated in the structure of Common-action Gardens II
Fig. 14 Fibrous structure of the final model, Common-action Gardens II
all through their linear direction axis resulting in a more flexible system responding
to existing conditions (Fig. 15).
We used the same composite material system with 10 mm wood and 30 mm EVA.
The aim to decrease waste material shifted our method by using the EVA panels in
another way. We found that EVA has a great tolerance for bending and is an elastic
material that does not deform. We used EVA as the infill material again but its material
behavior identified the bending fillet radius to direct the branching of fibrous walls
into different directions (Fig. 16). The linear shapes of the materials also served
for minimum waste materials. A group of fibers bending in different directions is
Fig. 15 Fibrous structure of the final model and contouring of Common-action Gardens II
Fig. 16 Elastic behavior of EVA is used for connecting the wooden panels of Common-action
Gardens
nested and labeled within these groups with the help of a script. This production
and assembly strategy enabled fast production and assembly as the first prototype;
CNC milling and assembly took around two days each, although we had a larger
structure to be assembled. The only problem was the precision while bending the
EVA material according to the exact angles on-site. At the regions where the number
of connecting fiber walls increased, more effort and workmanship are required.
3.3 Performance
In this second prototype, when the ratio of wood use is reduced to 80%, ratio of EVA
use increased slightly. But the main advantage was about the increase in planting
1436 F. Akipek
Table 2 Calculations about the basic parameters of Common-action Gardens I (Akipek et al. 2016)
Materials Wood EVA Stainless steel Nuts Total
anchorages
Dimensions 1700 × 1000 × 10 M10
2200 × 1500 ×
10 mm 30 mm
Number of 32 67 126
planes
Weights per 9.0 kg/m2 2.1 kg/m2 0.62 kg/m 11.4 gr
square meter
Total area 81.3 m2 71.1 m2 33.9 m 224 pieces
Total weight 731.7 kg 149.3 kg 21.0 kg 2.5 kg 904.5 kg
Optimization 68% 71% 655.8 m
Usable 1.9 m2 5.8 m2 7.7 m2
surface area
Total area 13.0 m2
areas. In this prototype, planting areas increased from 29 to 40% of the structure.
The amount of steel anchorages decreased and thus workmanship time about the
joints and bolts decreased. Waste of materials decreased from 42% to 29% in EVA
CNC milling. This second prototype did not increase the usable surfaces for park
facilities such as resting or sitting but that was acceptable about the main use of the
structure as an urban garden. These structures still need strategies to use less material
because to have one square meter of usable surface you need to use 25 kg EVA and
385 kg wood, which is not sustainable in the desired way. This result guides us to
question the contouring method as a design strategy and compare its strengths and
weaknesses (Table 2).
4 Conclusion
In these two experiences we used parametric design and computation to gain control
over quantities and enhance design qualities to develop a design approach which is
more sustainable, ecological and economical. Performance-based design approach
integrates design development and construction phases in a feedback relation with
each other and challenges environmental factors to participate design in the early
phases (Oxman 2008). Grobman argues that a broader definition of performance
however contains three dimensions as empirical dimensions which focus on directly
measurable performances that usually relate to physical data, cognitive dimensions
which relate to mental functions and processes and a perceptual dimension in which
the senses play an important role (Grobman and Neuman 2012). Empirical dimen-
sion is immediately translatable into computer language but cognitive and perceptual
dimensions still constitute a complex problem. Hensel (2012) emphasizes that per-
formance entails first the indivisibility of formal and functional aspects, and second
the interaction between four domains of agency: the subject, the environment, and the
spatial and material organization complex. He also focuses on material behavior as
an interaction of the material with a specific environment. He focuses on the material
behavior to be utilized in a positive manner that reinforces feedback as a key char-
acteristic of performance–oriented design. These issues reveal that we need to think
of the natural materials within a process starting from where it grows or becomes to
where it is sited and is a guide for us to take our material and performance research
further.
Various parametric design software shift towards integrated design, to associate
form finding process with various performance criteria and with construction pro-
cesses. Beyond these technological developments in design-built processes, next-
generation architects should develop strategies and use these technologies within an
attempt to increase relations with our natural environment, local cultural issues and
reduce the environmental damages via ecological and economical solutions.
Common-action Gardens I and II had been prototype projects for our design-
research group to develop a strategy by integrating design-technology-nature-
material studies in a way to reach some kind of ecological architecture. These two
urban gardens are still being used and part of the everyday life of these city parks
with their growing edible plants. This contouring technique designed by paramet-
ric software and produced by CNC milling technologies has the strength for fast
production and assembly. It has disadvantages about the amount of materials used.
We observe that digital design enhances these urban gardens to be accepted within
landscape design elements and creates a potential public place for gathering around
production and common-actions in ordinary parks. This design process is a proof for
us to continue our research towards controlling quantities for enhancing qualities. We
still need to develop new strategies towards a sustainable architecture and broaden the
use of ecological materials in building industry, with the help of performance-based
computational design, digital fabrication and robotic assembly opportunities.
Acknowledgements The POTplus Design Research Group would like to thank to all of the consul-
tants, sponsors, teams and students who involved in Common-action Gardens and Common-action
Walls as our initial design research projects. A complete list can be seen at our website (see foot-
note 1). As full time lecturers in İstanbul Bilgi University Department of Architecture, our design
research projects are supported by all of our colleagues in our faculty, as the representative of all
we would like to thank to Prof. Şebnem Yalınay Çinici, Head of the Department of Architecture.
Copyright All of the images belongs to the POTplus Design Research Group and can be shared
with permission.
1438 F. Akipek
References
Akipek FÖ, Yazar T, Aydın A (2016) Performansa Dayalı Tasarım Araştırmaları için Prototipler:
Kent Bahçesi Strüktürleri. Yapı J 419:136–164
Dunn N (2012) Digital fabrication in architecture. Laurence Kıng Publishing, London
Hensel MU (2012) Performance-oriented design from a material perspective: domains of agency and
the spatial and material organization complex. In: Grobman YJ, Neuman E (2012) Performalism:
form and performance in digital architecture. Routledge, London, pp 43–48
Grobman YJ, Neuman E (2012) Performalism: form and performance in digital architecture. Rout-
ledge, London
Menges A, Schwinn T, Krieg OD (2017) Advancing wood architecture. Routledge, London
Mollison B (1991) Introduction to Permaculture. Ten Speed Press, California. Turkish edition:
Mollison B. (2012) Permakültüre Giriş, Sürdürülebilir Yaşam Kitapları (trans: Özkan E). Sinek
Sekiz Yayınevi, Ankara
Oxman R (2008) Performance-based design: current practices and research issues. Int J Arch Com-
put 6(1):1–17. https://doi.org/10.1260/147807708784640090
Pektaş ŞT (ed), Çinici Yalınay Ş, Özsel Akipek F, Yazar T (2008) Computational design, parametric
modeling and architectural education. In: Arkitekt J 4–5 (518): 16–24
Algorithmic Craftsmanship For Bespoke
Timber Architecture
Sille Pihlak and Siim Tuksam
Abstract Over the last ten years a lot of research has been put into bridging the
digital and the physical. It has been a time of rapid rise of robotics in architecture
and fabrication. Based on our research with Estonian wooden house manufacturers,
the academic research and the industrial reality still remain worlds apart. Even though
computer numerically controlled (CNC) technology is gaining more and more hold
in the timber industry, the vision how to use it is driven by economy and efficiency
rather than by design. Our research is looking at how architects can start to wield a
greater influence on the direction where the automation of construction industry is
heading and how to make bespoke design and craftsmanship affordable for the wider
public.
Keywords Algorithmic · Automation · Bespoke design · Industry collaboration

Parametric design · Modular architecture · Estonian architecture
1 Introduction
Much of the intellect and energy invested in architectural education over the last ten
years has gone into design and making. Most of the leading architecture universities
have the orange creature, a Kuka arm, set up somewhere in their workshop. How
to bring this discussion out of the laboratory? At the Estonian Academy of Arts
(EKA) the in-house fabrication facilities have been rather modest until recently.
For that reason, the practical training of our students in timber building has always
depended on industrial collaboration. We see this situation as an opportunity to bring
this relationship, which has already been set up, to another level by connecting
algorithmic design research with the capabilities of the factories. As competition in
S. Pihlak (B) · S. Tuksam

Faculty of Architecture, Estonian Academy of Arts, Põhja puiestee 7, 10412 Tallinn, Estonia
e-mail: sille.pihlak@artun.ee
S. Tuksam
e-mail: siim.tuksam@artun.ee

1440 S. Pihlak and S. Tuksam
this industry is growing fast, more and more manufacturers are looking for ways to
stand out through design and innovation.
For almost a century, in Soviet and ex-Soviet urban environments few if any new
timber structures have been built. Owing mostly to the scarcity and little variety of
building materials in Soviet times, generations have been raised without awareness
of any connection between renewable materials and architecture. Given the evolu-
tion of wood treatments, such as fire protection and moisture proofing, structural
improvements in glulam—all of this has reduced the material-related restrictions on
building and has opened up endless possibilities in construction.
Yet our renewed focus on the material should not be backward looking. Novel
prospects—advanced timber manufacturing machinery, new computational tools
(e.g. algorithmic modelling) for material optimisation and structural analysis, and
above all, ability to generate design strategies for new spatial qualities—can provide
data minded1 novel tectonics. Local wooden house manufacturers, digitally skilled
architects and engineers are aiming to collaborate around a common digital platform
in order to strive for more efficient, flexible and structurally intelligent architecture.
The research at PART2 is most and foremost focused on design methods. Experi-
ments so far should be understood as research into design methodology, rather than as
research into production and manufacturing as engineering tasks of manufacturing.
Repurposing industry machines towards creating complex, adaptive and structurally
intelligent architecture could help us overcome the modernist physical space of stan-
dardised precast concrete. Using algorithmic optimisation as a tool and the local
context as parametric criteria, we are provided crucial arguments for developing a
contemporary design language for algorithmic timber architecture.
2 Local Context and Points of Departure
For Estonians, wood and timber is the principal renewable natural resource, yet the
modern urban population has become estranged from this material. Luckily, larger
Estonian cities and towns still retain districts of century-old wooden architecture.
Despite the fact that new wooden buildings make up a marginal part of local archi-
tecture, logging has not decreased. In fact, 91% of timber not used for power or heat
generation is exported3 —for use in Alpine log homes or Central European backyard
saunas. In this regard, Estonia has become an exporter of a cheap commodity, rather
than value-added design.
1 Using computational strategies to provide new geometric outputs.

2 Practice for Architecture, Research and Theory.
3 In 2014, 90.8% of timber was exported. Source: Puuinfo.
Algorithmic Craftsmanship For Bespoke … 1441
2.1 Renewable Materials in Digital Fabrication
In the light of a worldwide shortage of energy resources and an increasing concern

about climate change, PART practice envisions an advanced application of timber as a
structural material, in particular in consideration of economic and ecologic criteria. In
this respect, we investigate a highly integrated digital design and fabrication process
for timber. The local concern about under-valued log export has brought greater
public attention to deforestation and diminishing wildlife habitats. Considering the
situation, PART research on timber investigates how architects could add more value
to the material through design, providing the Estonian wooden house manufacturers
with ideas for original products in an effort to overcome the cheap subcontractor
identity.
2.2 Industry-Connected Digital Fabrication
Our projects show various scales of investigation, varying from furniture scale to
infrastructure. Parametric design tools and a common design platform (Rhinoceros,
Grasshopper), which allow engineers and architects to work in the same model, have
helped create highly complex and performative architectural forms without com-
promising on speed and cost of production. The fourth dimension in architecture is
time—if you can make custom solutions more or less as time-efficient as standardised
ones, the industry is willing to adapt. In the examples presented in this research, the
Body Building Installation (built for Tallinn Architecture Biennale 2015) required
exactly the same amount of timber elements and production time as a small family
house (80m2), yet with the advantages of greater spans, height, cantilevers, etc. Also
the spatial qualities—fluid and curvaceous—created an understanding of various
possibilities. And above all—the fact that the Estonian forests would require only
one minute to produce all the material used (Keso 2015) certainly made this urban
installation a conversation starter.
2.3 Digital Material Fabrication
We argue that, compared to traditional craftsmanship, digital craftsmanship in the

techno-cultural environment of Estonia4 offers more opportunities in terms of speed,
4 President Toomas Hendrik Ilves claims to have in the mid-90 s “reverse-engineered” Jeremy
Rifkin’s The End of Work, the 1995 bestseller arguing that information technology would under-
mine large-scale industrial production. What he calls his “backward reading” of Rifkin led him to
recognise the importance of Estonia’s miniature physical size in creating a substantial post-industrial
economy, where a small, tightly knit hi-tech workforce of perpetually pivoting entrepreneurs could
reinvent Estonia as the original startup nation.
Fig. 1 Son of a Shingle—pedestrian bridge and tunnel in Tartu, Estonia. The infrastructural facility
smoothly blends in with the landscape and is notable for its use of wooden shingles. The scale-like
shingle surface creates a seamless connection between the bridge and the underpass. The project
combines traditional shingles with modern technology both in design and manufacturing. Copyright
PART
building scale and complexity of form. We should not regard the properties of building
materials in isolation, but rather embed them in the design process. Our investigations
vary from light transparency and bending directions in plywood to civil projects like
pedestrian tunnels (Fig. 1) and electricity pylon. For Elering’s high voltage corner
pylon (Fig. 2), we proposed the technology of acetylated wood whose structural
qualities (density, weight and strength) were immediately embedded in the simulation
process (Fig. 3) of the pylon’s dimensions and design. This kind of experimental
approach, without previous references, has put demonstrators—installations to test
certain methods—into an essential role in our daily practice. Therefore, through
the intermarriage of computational and physical space, we could create familiar yet
complex spatial encounters combining traditional materials with current technologies
and inherent aesthetics.
The algorithmic optimisation that allows all the previously mentioned possibilities
is especially useful in construction, where there is a constant need for collaboration,
communication and adaptation. Especially, when working with an organic material
like wood, efficient joinery requires millimetre precision, yet at the same time we
must consider that timber is constantly reacting to its surroundings (e.g. moisture).
By having a parametric 3D model, we can make timber architecture once again
approachable and feasible—from an idea to construction.
3 Workflows in Materialisation
With the timber industry being one of the largest in Estonia (Kobuszynska 2016),
employing diverse machinery and possessing a wide range of competences, PART’s
research has been seeking to map material and machine driven information for the
possible automated solutions in architecture.
Fig. 2 Bog Fox—high voltage design pylon. As the first of its kind, the timber design pylon would
have softened the visual impact on the area of natural beauty near Kuistlema Bog and become a
landmark not only for Lääne-Nigula municipality and Elering company, but also for the Estonian
timber industry, however, due to the absence of a similar reference project in Estonia, it was decided
to go with a different material and redesign it as a steel structure. The project is now developed
further in cooperation with Dutch steelworkers and Austrian engineers. Copyright PART
Wood processing companies differ in terms of scale from building construction to

manufacture of wood products.5 The range of wood industry products is extensive:
from the treatment of sawn timber to the manufacture of log cabins, windows, glulam,
plywood, etc.
As the knowledge of machines and their capabilities from CAD to CAM is an
essential input to our designs, we are setting up a mapping system for the machinery in
local context. Considering Estonia as the test laboratory for our work, our working
radius extends as far as 300 km. In collaboration with the Estonian Academy of
Arts Department of Architecture research lab on prefabricated housing,6 we are in
the process of mapping the capabilities of local timber house manufacturers, also
providing them with architectural insight of information needed for collaboration.
3.1 Material Specific Parameters
For the designs to be efficiently usable in timber fabrication, the data format must
contain definitions for material-specific features and take into account timber-specific
5 Puuinfo webpage for local wood companies http://puuinfo.ee/tootjad/.

6 Research lab led by Renee Puusepp, Estonian Academy of Arts Department of Architecture.
Fig. 3 Body Building Installation. The freeform structure is a contemporary take on timber post
and beam construction, manufactured on a fully automatic timber house production line where
traditional log houses are produced on a daily basis. The computer-assisted 5-axes milling machines
allow the production of customised solutions at the speed of traditional industrial manufacturing
while the algorithmic design methods provide flexibility in planning. Photo Tõnu Tunnel
material parameters, such as the notion of fibre direction and varying dimensions.
Material-driven changes could have serious effect on production and assembly.
In the case study Body Building Installation (2015), the parametric design method
adopted for the project aimed to build upon parametric changes coming from a
CNC machine (depending on fabricator) and material measurements (moisture level,
uneven planing). The installation contains 236 unique pieces, all with at least two
CNC-milled lap joints. Every varying millimetre could create loose connections and
structural complications. Just a day before the commencement of the fabrication, the
new measurements of logs were received. While using parametric setup, necessary
adjustments could be made quickly and sent to the factory.
Besides the fast configuration in CAD files, the project looked into “cutting the
middleman”.7 The intention was to drive production geometries and information
to the fullest extent possible, quite literally from Grasshopper to Hundegger.8 As
opposed to the common practice of handing over design geometry to production
7 Excerpt from the 18.10.2017 lecture by architect Enric Ruiz Geli in the course of the Estonian
Academy of Arts Faculty of Architecture open lecture programme.
8 Joinery machine K2i—Hans Hundegger AG.
specialists for shop drawings and post-rationalisation (Pena 2012), it saved time and
provided a better and more detailed overview of design. Body Building Installation
(Fig. 3) offered an insight into troublesome workflows, revealing where the infor-
mation flow, tools and knowledge did not quite match.
3.2 Cumbersome Workflows
The timber sector in Estonia has been pioneering digital fabrication in the local
building industry for the last decade. Today, CNC manufacturing on three- and five-
axis machines is very common, even in medium-scale carpentries, and specialised
joinery machines are used to process beams. Due to their precision and efficiency,
these machines have even managed to bring back traditional timber detailing like
pegs, lap joints and dovetails, which had almost completely been replaced by metal
connectors.
Yet, the understanding of the need for collaboration between architects, engi-
neers and fabricators is not there. Industry uses domain-specific CAD programmes
that are tailored to support the traditional workflows and provide the necessary
data for machining, establishing the connection between contemporary design tools
(Rhinoceros, Grasshopper) and CNC machines remains bumpy when it comes to
non-standard applications.
3.3 Setting up a Common Platform
We propose a workflow that is ever more common nowadays, where construction in

general is moving away from the narrative (architect to engineer to builder) design
process.9 The process where a concept design created by the architect is passed on
to the engineer who solves the technical problem of producing drawings and an
operational design is far from ideal. The architect is left out of most of the process,
and any later changes due to engineering or construction-related limitations may no
longer be in sync with the original concept design. The inclusion of manufacturer at
an early stage of the design process provides much better opportunities for developing
novel wooden architecture solutions.
Fortunately, the practice of having all parties, from the architect to the builder,
work on a single 3D model is becoming a more widespread phenomenon.10 This
replaces the above linear process with an iterative system, which is a prerequisite
for designing effective human environments. Modifications made to the CAD design
reach all parties simultaneously, which helps them to arrive at an optimally economi-
9 Excerpt from the 03.12.2015 lecture by engineer Jan Knippers in the course of the Estonian
Academy of Arts Faculty of Architecture open lecture programme.
10 E.g. CAD and BIM programmes.
cal and energy-efficient solution. The parametric 3D model, where changing any one
parameter results in the automatic adjustment of the entire solution, makes managing
ever more complicated geometries faster and easier.
In times when investments in software and hardware are of high interest locally,
one might hope that a collaborative platform for the building industry would catch
on quickly. Currently, the testing phase with international collaborators is already
underway. Case study Bog Fox (Fig. 4) is PART’s first and still ongoing fully para-
metric project, where all collaborating sides, from design to structural engineering
to shop drawings, have worked together in Grasshopper. PART design on finding the
most optimum form of a high voltage pylon, with a height of 42 m and a footprint
of 10*10 m, was further developed and rationalised in collaboration with Bollinger
and Grohmann Ingenieure, with all 3 possible material solutions—timber, steel and
centrifuged concrete—in one algorithmic model. Unfortunately, the client lacked
the faith in timber for such a high risk project and timber was left aside in favour
of steel. The process continued with CIG Architecture complex metal structures
manufacturer, who were able to extract their information for shop drawings and
manufacturing. The project is still in the design development stage, but it is a helpful
example of fast scaling in the algorithmic design platform, from installation scale to
infrastructural scale.
3.4 Part-Based Digital Workflow
The advent of parametric modelling in timber design has opened up a whole new
world to material based designers. By defining geometric dependencies instead of
final shapes, highly complex structures can be described with justifiable effort. Addi-
tionally, breaking them down into thousands of individual components, a necessary
step in timber structures, has become manageable in regards of industry capabilities.
Constructed demonstrators in Estonia, where CNC machines produce customised
parts at the speed and cost of serial production, have raised interest in automated
designs among wooden house manufacturers. However, even in the digital age, the
transition from manufacturing to assembly, still remains the weakest link, at least
when more complex operations and large quantities of individual components are
required.
3.5 Assembly Difficulties
Where large quantities of individually shaped non-standard components are required,

feature recognition is not reliable and manual job preparation is too laborious. How-
ever, when the components are the result of a parametric CAD model, chances are
high that they all follow the same geometric rules and require similar sequences of
machining operations for their production, even though they all look different. In
Fig. 4 Bog Fox—design high voltage pylon. Using Karamba3D, in collaboration with Bollinger +
Grohmann Ingenieure, we set up structural and geometric checks to be able to run a genetic algorithm
for material weight minimisation. Manual adjustments were made afterwards to balance additional
weight versus aesthetic preference. Copyright PART and Bollinger + Grohmann Ingenieure
regards of assembly, those rules do not apply. For onsite assembly, when workflow is
highly dependent on labels assigned to parts, sorting and separating pieces remains
time-consuming. To bypass the laborious task of searching and sorting, we have
looked more into opportunities offered by modular and repetitional elements.
The PART research has applied the least automation in the construction phase.
Scale and weight-wise we have been investigating smaller elements, which allow
manual construction, often realised by volunteers with no previous knowledge. Parts
of the most time-consuming activities, labeling and sorting, have been moved to the
preassembly stage in factory. In the case of space-filling systems, such as Digital
Thicket11 and Here and Elsewhere,12 the construction logic was built into the prefab-
11 Digital
Thicket. Estonian EU presidency Opening Ceremony Installation at Freedom Square.
12 Withinthe context of the Estonian presidency, “BEL:EST. A Laboratory for Europe in Brussels”
brought together five Estonian and five Belgian architecture offices to think about the capital of
Europe and the problems associated with its physical look and image.
ricated details. This left the possibility to work with onsite volunteers who, without
drawings, could approach elements as playful and engaging building blocks.
4 Automation for Design
Automation is the process of eliminating the human from a task or process. It has
an inherent relation to economy—which is quantitative. At the same time, it relates
to culture—automatons were already present in Greek mythology.13 As any techno-
logical advancement, automation has a cultural relevance.
We are not just automating calculation and making, but more and more also sensing
and thinking—expanding the realm of sensible. These developments are changing
the way we build, but also the way we think about buildings and our environment
(Picon 2013). Using CNC log milling lines on an industrial scale has brought back the
dovetail joint as a more economic alternative to metal connectors (Schindler 2009).
The complex joinery facilitated by automated tooling is a way of bringing back long
lost constructive details, stimulating the imagination for novel ornamental uses of
these details.
Automation depends on control loops. There can be either open or closed feedback
loops. Construction of buildings is most often the first—an open loop where we set
a machine to produce a part or assemble the parts. This, for instance, is how most
3D printers work and why quite often when using FDM,14 you can receive a fuzzy
filament ball instead of your design. Constructing complex structures would be just as
hazardous without proper simulation, analysis and optimisation. A closed feedback
loop is much more easily implemented in the digital design phase. Automation of
these processes, like finite element analysis, was born because of the necessity to
speed up the evaluation of large building projects. The result is a fast process for
design evaluation which can be manipulated by any member of the design team,
as it comes with a built-in specialised expertise. Therefore, it allows architectural
design decisions informed by physical constraints from the very beginning, while also
bringing computational engineering language (e.g. force flow lines) into architectural
design.
As most of these automated systems are in part also developed and widely adapted
by architects, structural and environmental simulation, analysis and optimisation
using genetic algorithms have become tools for design—automation for design, for
bespoke solutions.
13 The Greek god Hephaestus created automata for his workshop.

14 Fused deposition modeling.
4.1 Organising Chaos
One of the underlying causes ushering the advent of modernism was the poor quality
of the first industrially produced products. The principles of efficiency, minimal-
ism and form following function were set to bring order to a world full of clutter.
A century later, our cities are cluttered with what the market economy has turned
modernism into. The means of industrial production proved so efficient that they
became ubiquitous and are used for absolutely all purposes, not just for achieving
the minimalist utopia. Even early digital architecture used the same means of con-
struction (and still is). So ever more complex software solutions have been developed
to satisfy the desire for complexity, while the construction methods and materials
have not really changed—we are still producing linear profiles and plate materials
that are cut into standard measurements.
Our hypothesis is that by looking at the current means of production and algorith-
mically embedding them into the design process, we can develop a more sustainable
second (Carpo 2017) digital architecture—architecture that through functional orna-
mentation embodies the re-emergence of joinery, algorithmic structural systems and
adaptations to environmental forces.
Ornament was considered a crime, because it was wasteful, labour-intensive and
expensive. With automated fabrication and renewable materials, this is no longer the
case. Furthermore, a century of anonymous concrete, steel and glass surfaces has
created a demand for adornment and ornamentation.
4.2 Digital Tools of Production
When calculus entered architecture in the early nineties, the first instinct for most was
to go for maximum variability. After the 2008 financial crisis, more emphasis has been
put on the development of automated construction processes of complex and variable
systems. Over the last 10 years or so, we have witnessed a plethora of fascinating
demonstrator pavilions from numerous research institutions around the world and
many novel techniques that have also been demonstrated in larger scale. Most of them,
as ingenious as they may be, are yet to be applied in the industrial market situation.
Most innovation in industry is happening in prefabrication, where automation is
easiest to achieve. This means the outcome of innovation in construction is often a
product. From this perspective, the algorithmic tool or adaptable joint could be seen
as the product entering the market model. What is often missing, is demand.
4.2.1 Body Building Installation
With our first project we aimed to demonstrate the capabilities of the local wooden
house manufacturers and by that, hopefully, create demand. We found a local factory
Fig. 5 Body Building Installation. The 236 unique elements of the Body Building installation
were generated using algorithmic design and engineering software (Rhinoceros, Grasshopper,
Karmaba3D). Lumber elements with 464 different joints were milled on a 5-axes Hundegger 2Ki
in 10 h and assembled by volunteers and students with 2320 screws in five days. Copyright PART
in South of Estonia who owned a Hundegger K2i with a 5-axis universal tool. Using
their machine, we could demonstrate a parametric approach to joinery.
The Body Building installation was an urban extension of Tallinn Architecture
Biennale (TAB) main exhibition “Body Building” at the Museum of Estonian Archi-
tecture. The exhibition explored “hybrid forms of construction where cutting-edge
technology and science meet the self-driven variability of material systems and
degrees of freedom and control define an outcome of multiplicity within tolerance,
trying to find a balance between the unruly and the predictable—body and building”
(Pihlak and Tuksam 2015). The installation is an attempt to contextualise these ideas
by using local resources and capabilities.
The installation, located on the main artery of the city of Tallinn (Fig. 5), is a
conceptual beacon and guide for the 3 main venues of TAB—the Estonian Centre of
Architecture, Viru Square and the Museum of Estonian Architecture. The freeform
structure (body) is the result of algorithmic negotiations between ideal geometries,
e.g. lines, planes, circles and cuboids (building). By using computational methods,
the installation brings softer forms and materials into the otherwise rigid city centre.
The structure was intended to show a transition between geometric shapes: circle,
horizontal plane, vertical plane and arch. The base geometry was generated using
magnetic field simulation, to create a twisted blend between the four edge conditions
(Fig. 6). The resulting field lines were divided into segments that would become
the axis for the timber pieces. As the curves were not planar, all the elements had
to be individually adjusted to create a smooth transition between elements. The
negotiations between the rectilinear lumber and smooth geometric transitions formed
a gradient field of varying joints.
The pavilion is a contemporary take on wooden post and beam construction,
manufactured on a fully automatic wooden house production line, where traditional
log houses are produced on a daily basis. The project aims to promote the use of
Fig. 6 Body Building Installation. Using an algorithmic 3D model, changes may be made until the
very last moment and all the structural changes are constantly recalculated. This allows us to make
sure that the construction of the system will also hold up after each change. Copyright PART
algorithmic design in industrial production and spark a discussion on the future of

wooden architecture between the local wood industry, engineers and architects.
The 236 unique elements of the installation were generated using algorithmic
design and engineering software (Rhinoceros, Grasshopper, Karmaba3D, Kangaroo).
The 95*95 mm lumber elements with 464 different joints were 5-axes CNC-milled
on a Hundegger 2Ki in 10 h and assembled with 2320 screws in 5 days by volunteers.
The fully algorithmic 3D-model was developed over a period of 6 months, resulting in
a design tool where the base geometry is interchangeable within a few moments and
various parameters of the design are controlled by numeric input. Using the Karamba
structural analysis plugin, all the structural changes are constantly recalculated and
optimised.
With the project we developed a strategy for constructing freeform structures
out of off-the-shelf lumber using variable half lap joints that were optimised within
the limits of production and structural needs. Using Kangaroo physics engine, the
angles were constrained within tolerances to allow for a sufficient contact area to
insert the minimum required amount of screws. The design had to be adjusted to
hsbcad restrictions as we did not have access to a software that would have bridged
Rhinoceros and hsbcad. Many of the conversions failed to work properly, meaning
we had to delete and add treatments by hand by also exporting the boolean geometry.
A few weeks after the opening, we were contacted by the city officials who
asked whether we agreed to leave the temporary structure on its location for another
year. The project was a discussion starteramong the local architects, engineers and
manufacturers, hoping to start working together on more experimental projects and to
develop more productive, innovative workflows by incorporating academic research.
4.2.2 Digital Thicket Installation
With the Body Building installation and other15 fully variable pieces, the weak link
was construction. With the Digital Thicket installation for the opening ceremony
of Estonian Presidency of the Council of the EU, the scale increased substantially,
requiring disproportionate funds from the budget. Therefore, we needed to find a
way to automate parts of the assembly process.
We started looking at Lindenmayer systems,16 particularly at regular tri-stars
(Fig. 7) connected end to end and adding rotation. Setting up a parametric model, we
realised, first, that a branching structure would not be structurally viable, and second,
that by rotating the elements by about 70.5° or the tetrahedral angle, they interlock
and form a modular structure. In further analysis, we have found the geometry to
follow the surface and edges of truncated octahedra, which is a space-filling system
(Fig. 8).
The geometric system also follows the logic of the Steiner tree problem,17 i.e. it
forms a minimum weight network. This is exactly what we were looking for: a system
that uses a minimum amount of material to fill a maximum volume of space. The
event happened on one night only, so the structure needed to be easily transported
and assembled (as well as disassembled) right before and right after.
The tri-star consisting of three identical pieces coming together in a triangle shape
with half lap joints formed structurally stable stiff elements. The other end was cut
at an angle allowing for the 70.5° twist. The tri-stars were connected by two bolts at
the centre of geometric axes (Fig. 9), where the momentums are the smallest.
The whole structure consists of identical elements of 95*95 mm lumber that have
a single way of assembly therefore eliminating construction mistakes. The system
does not remove the human from the assembly process, but most tasks, like search-
ing, sorting and fitting, are eliminated, making it a quasi-automated process. The
whole installation was assembled by non-specialist workers through GoWorkaBit,
an Estonian startup for flexible working options.
15 2016 SoundWaves for Tallinn Music Week, 2017 Rheological Formation for Into the Valley music
festival.
16 https://en.wikipedia.org/wiki/L-system.
17 https://en.wikipedia.org/wiki/Steiner_tree_problem.
Fig. 7 95*95 mm lumber, milled on a Hundegger 2Ki into repetitive 1 m long pieces that form
tri-stars, which are connected with bolts at an angle of 70.5°. Copyright PART
Fig. 8 Truncated octahedron stacking in relation to the Digital Thicket structure. Copyright PART
Designing with the Digital Thicket elements also has similarities with gardening.
The script is connected to structural evaluation, therefore, the designer receives imme-
diate feedback when adding or removing elements, indicating whether the structure
got stronger or weaker.
Fig. 9 Digital Thicket. Estonian EU presidency Opening Ceremony Installation at Freedom Square.
An open system of seven solid cubic metres and around 700 identical wooden elements calling for
imagination and various functions. Copyright PART
The same structure has been re-erected several times at various places in the city
of Tallinn for shorter and longer periods of time.
4.2.3 Here and Elsewhere Installation
After the initial project, we have applied this geometric system on various scales
from that of furniture to interior to urban. The Here and Elsewhere18 project, con-
ceptualised in collaboration with LASSA Architects, looked at the interior scale and
created spaces for work, study, leisure and interaction. The module was realised in a
scale with 40 cm steps to create seating, tables and surfaces to be used for standing
(Fig. 10). The installation was exhibited displayed at the Tallinn Architecture Bien-
nale 2017 and at BOZAR in Brussels from November 2017 until January 2018. This
also implied that multiple assembly and transportation cycles had to be taken into
account. The structure filled a space measuring 64 m3 in assembled state and about
half a cubic metre when disassembled for transport.
Its roughly 250 joints in 4 different sizes were CNC-milled out of 12 mm plywood.
The round pine profiles of 35 mm diameter with notches creating the 70.5° rotation
were also 3-axis CNC-milled using jigs. The failure to automate the rotation part of
the fabrication process was the biggest weakness of the project.
18 As part of the exhibition BEL:EST—A Laboratory for Europe in Brussels.

Fig. 10 Here and Elsewhere/Ici et ailleurs. Installation for the Exhibition BEL:EST—A Laboratory
for Europe in Brussels. Assambly diagram. Copyright PART
Design-wise the 4 different sizes of joints allowed for various usable surfaces like
chairs, tables and information carriers. It also created a variation within the structure,
allowing for orientation, light and shade effects, etc.
In the third, the most recent installment of the Digital Thicket geometric system,
we created a shelving system as an exhibition design for PART.icular—Bespoke
Timber Architecture as part of Time Space Existence organised by the GAA Foun-
dation and European Cultural Centre at Palazzo Bembo in Venice (Fig. 11). In this
case, the angle was not fixed by using round profiles and joints that allow rotation,
meaning the form is self-organising by fixing the wall connections.
A further addition was the introduction of a surface. We have been working with
nets and stretch fabric to create minimal surfaces filling the cells (Fig. 12). In this case,
we needed to print on the surfaces, so we decided to go with paper and developable
surfaces—triangulation with fillets. The curved folds give an impression of a doubly
curved surface still. Furthermore, as it is made of paper, it means that while scaling
the system up, the same strategy could be used with other bendable sheet materials
to fill the cells in a similar manner.
4.3 Product and Service
Working together with industry partners to develop new projects of architecture has
proven that a new model for architecture could be somewhere between product and
service. In the process of creating our demonstrators, we have created products that
include geometric systems, modular designs, or tools for creating joinery. In order
Fig. 11 PART.icular—Bespoke Timber Architecture as part of Time Space Existence organised by

the GAA Foundation and European Cultural Centre at Palazzo Bembo in Venice. Shelving system
for models and drawings, consisting of 3D printed joints, various length of wood elements and
paper surfaces. Photo Tõnu Tunnel
to use these tools, our services are still needed. We consider our practice a research
project in constant development. With each project, we build on the previous one and
develop new parts. Our product quite often is not the end result, but an automation
of a task, be it the design of ornamental joinery for free form plate structures or
development of modular geometric systems with high degrees of freedom for spatial
configuration.
5 Conclusion
Estonia, due to its very small size and flexible enterprises, is a perfect testing ground
for industry collaboration. The market retains a rapid growth, while the competition
is intense. Consequently, wooden house manufacturers are looking for ways to stand
out, innovate and be sustainable.
Unfortunately, until now, not too many novel design solutions and innovations
have been created to match the capabilities of the available machinery. The initial
aim of the manufacturers is to create efficiency, not novelty—sell more, not for more.
Fig. 12 PART.icular—Bespoke Timber Architecture. The round joints allow for rotation, meaning
the structure locks into shape by the precisely positioned wall mounts. The connections are held
together by friction alone. Photo Tõnu Tunnel
Considering the current polemics surrounding deforestation and export of raw

timber, we are looking for innovative solutions to increase the value of exported
timber products, while also creating the architectural language for the Industry 4.0.
Estonia is one of the pioneers of digitalisation of government services (Reynolds
2016). However, the same cannot be said of the construction industry, yet. Still,
digitalisation and automation are spreading fast.
All of this makes Estonia a great testing ground for automation in architecture.
Our research specialises in the ways of applying algorithmic tools and automation
for advancing architectural design resulting in an architecture that would hopefully
reflect the digitally optimistic Estonia.
References
Carpo M (2017) The second digital turn: design beyond intelligence. The MIT Press, Cambridge,
MA
Keso K (2015) http://kes-kus.ee/mets-numbrites/
Kobuszynska M, approved by Nicely R (2016) Wood Sector in Estonia, report https://gain.fas.usda.
gov/Recent%20GAIN%20Publications/Wood%20Sector%20in%20Estonia_Warsaw_Estonia_
12-20-2016.pdf
Pena A (2012) Rationalisation of freeform façades: a technique for uniform hexagonal panelling.
pp 243–252
Picon A (2013) Ornament: the politics of architecture and subjectivity. Wiley, Chichester
Pihlak S, Tuksam S (2015) Body Building. TAB main exhibition 2015. Estonian Centre of Archi-
tecture, Tallinn. p 3
Reynolds M (2016) Welcome to E-stonia, the world’s most digitally advanced society, Wired Mag-
azine
Schindler C (2009) Ein architektonisches Periodisierungsmodell anhand fertigungstechnischer Kri-
terien, dargestellt am Beispiel des Holzbaus. ETH, Zürich p 208 https://doi.org/10.3929/ethz-a-
005956976
FracShell: From Fractal Surface
to a Lattice Shell Structure
Iasef Md Rian
Abstract ‘FracShell’ was a digital design workshop as part of a computational mor-

phogenesis project. This chapter outlines the workflow of the workshop highlighting
the transformation of the fractal geometry, more specifically, the Takagi-Landsberg’s
fractal surface, into a lattice shell structure. This workflow discusses how the mathe-
matical model was transformed into the generative digital model, then the variables-
based parametric model followed by the finite element model for the finite element
structural analyses, and finally constructional model in order to construct the real-
scale lattice shell structure. This workshop illustrates the easy manual construction
of a complex digital design taking the advantage of the rule-based shape generations
as well as the self-similar modular properties of fractal geometry. The versatility of
wood in transforming the complex digital design into its real-scale physical structure
has further been mentioned.
Keywords Fractal geometry · IFS · Parametric design · Lattice shell · FEM

Wooden structure
1
1 Introduction
In the current trend of digital design and digital fabrication, we see a growing prac-
tice of free-form and complex designs in architecture. The technological and digi-
tal advancement has pushed the boundary of experimentations with new forms for
architecture. Mathematical expressions and almost all type geometries are now eas-
ily possible to visualize in a digital interface as digital models. Almost any kind of
1 The content of this chapter was published in 2018 by the same author (Rian et al. 2018). However,
in this book chapter, the workflow of the FracShell construction as a design workshop is the main
focus. The construction part has been more emphasized while the use of wood in realizing the digital
design has been further highlighted.
I. M. Rian (B)
Department of Architectural Engineering, University of Sharjah, Sharjah, UAE
e-mail: iasefrian@gmail.com

1460 I. M. Rian
geometric shape or mathematical model as a digital model can now be turned into
its structural model using Finite Element Method (FEM) for the quick structural
analyses in order to check its architectural applicability, structural viability, and opti-
mality. The variables-based parametric modeling has further expanded the frontier of
the design of desirable, adaptable, optimal, and efficient structural forms. Materiality
has advanced this level a step ahead. The combination of mathematics, geometry, dig-
ital tools, materiality and structural forces manipulation (or optimization) has offered
a unique opportunity to designers for unique form-making and efficient form-finding
in architecture (Burry and Burry 2012; Menges and Ahlquist 2011).
FracShell is one of such digital design ventures where the mathematical concept
of fractal geometry was used in finding a new possibility of lattice shell design where
the wood was used as the main material to show its architectural, structural and con-
structional versatility. A design workshop was organized at the Politecnico di Torino
as a part of the Computational Morphogenesis course. The whole objective was two-
fold: one, making a parametric model using the concept of fractal geometry; two,
construct the complex model manually without using any digital fabrication tools.
The idea was to present that the complex designs created in computational modeling
method do not necessarily be fabricated with the help of digital fabrication tools.
The right choice of a rule-based complex geometry and a strategic constructional
approach of modular fabrication can make it easy to manually construct a complex
structure. The workflow followed four major stages (Fig. 1).
a. Geometric Modelling: In this stage, mathematical formulation of the fractal sur-
face has been transformed into a visual model on a digital interface. This digital
model, i.e., the generative model was transformed into a parametric model which
can be morphed by changing the geometric variable of the fractal surface. In this
stage, an architectural model was developed from the parametric model based
on the desired design concept/requirement.
b. Structural Modelling: In this stage, the parametric model was transformed into
a finite element model for performing structural analysis in order to check the
Fig. 1 Workflow of the FracShell design workshop

FracShell: From Fractal Surface to a Lattice Shell Structure 1461
structural viability and optimality of the architectural model. One suitable form
was selected for the real-world construction after the structural analyses.
c. Constructional Modelling: In this stage, the selected shape of the architectural
model was further modified for the real-world construction by emphasizing the
joints and connections. In this stage, structural elements were labeled for the
modular construction.
d. Physical Construction: This is the final stage in which the final structure was
physically constructed. All the elements were made manually by following the
details taken from the constructional model.
2 FracShell: Geometric Modelling
2.1 Fractal Geometry
Fractal geometry is a concept in which a geometric shape replicates itself multiple

times over different scales resulting in the self-similarity and the reflection of the
whole in its parts. In other words, fractal shapes are iteratively generated self-similar
figures (Mandelbrot 1983; Bovill 1996). In mathematical definition, fractal shapes
are fractional or non-integer dimensional shapes, which means, it can be, for example,
in between one-dimensional and two-dimensional shapes, or two-dimensional and
three-dimensional shapes (Barnsley 2014). They are always irregular and broken.
There are different types of fractals. One type is purely ‘self-similar fractals’, which
means it’s all parts are exactly identical but smaller in scales. In mathematics, we
have some well-known self-similar fractals such as the Sierpinski triangle, Koch
curve, etc. Another type is known as ‘approximate fractals’. All the smaller parts of
this type of fractals are not exactly the scaled copies of the main shape, instead, few
or none of them are exactly same but many of them are nearly self-similar, or slightly
distorted form of the main shape. Many natural shapes, such as tree-branches and
river basins can be considered as the examples of approximate fractals. There are
different ways of constructing fractal shapes in geometry. Iterated Function System
(IFS) is one the most popular methods in which one shape is replicated, translated
and scaled to create a self-similar pattern by using a set of function (or geometric
rules), and the resulted pattern is again passed through the same set of functions
(geometric rules). This process is continued recursively multiple times (Fig. 2), thus
resulting in fractal shapes.
The iterative process of IFS can be easily demonstrated by using the vector method.
Figure 3 shows the vector-based IFS method in constructing a Koch curve. In pro-
ducing a Koch curve, a straight line has been divided into three equal segments, and
the midpoint of the middle segment has been moved up along the local Z-direction in
such a distance that the new moving point of the middle segment with its endpoints
create two new lines and the length of the new lines are exactly the same with other
1462 I. M. Rian
Fig. 2 A recursive mechanism of the iterated function system. An is a previous shape which is
transformed into a self-similar or a self-affine but contracted shape An+1 by a set of geometric rules
(f i ; i 1, 2, 3, …, m). Each new shape goes through same rule multiple times
Fig. 3 IFS-based repetition of midpoint displacements on each newly-born one-third segment of

a straight line that eventually produces a Koch curve which is a canonical example of a fractal
geometry
Fig. 4 Integer and fractal dimensional geometric transformations of a 2-dimensional surface
segments. This geometric operation has been repeated iteratively on each segment
by taking its endpoints and midpoint as the new vectors.
Fractal geometries can be identified by their geometric dimensions, more pre-
cisely, by their fractal dimensions. A fractal dimension is a unit which quantifies the
fractality of a fractal shape, which means, it measures the degree of detailing, irregu-
larity or unsmoothness of the fractal shapes. (Bovill 1996) Higher the fractal dimen-
sion, higher is the detailing or irregularity of the shape. One shape can be changed
into a fractal shape from a regular shape if its fractal dimension can be changed from
its integer value to its higher or lower but non-integer values. Figure 4 is an example
which explains that how a surface as a two-dimensional shape can be transformed
into both a one-dimensional grid-like shape as well as a three-dimensional fold-like
shape. The intermediate shapes in between the two integer-dimensional shapes are
fractal versions of the original surface and their geometric dimensions are called
‘fractal dimension’.
2.2 Fractal Surface: Generative and Parametric Modelling
Koch curve is a fractal curve which has been produced from a straight line using
IFS by displacing the midpoint of its middle segment. Different operations of itera-
tive midpoint displacement, thus, can generate different types of fractal curve. This
process of iterative midpoint-displacement was first used by Archimedes (Mandel-
brot 2002) in calculating the area of a parabolic segment. By using his approach a
parabolic curve can be constructed. First, the midpoint of a straight line is moved
up along the global Z-axis by h height. Connecting the endpoints with this new dis-
placed point produces two new straight lines. The midpoints of these two new lines
are further moved up along the global Z-axis by height h1 in such a way that h1 w.h.
This process of midpoint displacement of each newly generated lines is repeated in
such a way that hn w.hn−1 , which eventually reaches to a parabolic curve (Fig. 5,
Top). Here, w is a scaling factor, known as relative size value and its value is 0.25.
In the early 1910s, Teiji Takagi changed the w value and replaced it by 0.5, and a
remarkable change was noticed. The parabolic curve became an irregular curve hav-
ing some fractal-like self-similar quality (Takagi 1903) (Fig. 5, Bottom). This new
curve became familiar as Takagi curve or Blancmange curve, a unique example of a
fractal curve. Later, George Landsberg made this w as a variable value ranging from
0.25 to 1.0, thus transforming the Takagi curve as a parametric fractal curve which
is sensitive to w value (Mandelbrot 2002). This parametric curve is known as the
Takagi-Landsberg curve. When the w 0.25, then this curve is a smooth parabola,
but if w > 0.25, then it becomes irregular. Higher the w value, higher is the irregularity
and roughness of the curve.
This same process can be applied on a flat surface whose midpoint of each side can
be iteratively moved up by the factor w value. Figure 6 shows the two-dimensional
counterpart of the Takagi-Landsberg curve and can be called the Takagi-Landsberg
surface. This surface is a smooth paraboloid when w 0.25, but it becomes unsmooth
when w > 0.25. The irregularity and roughness of the surface increase with the
increasing of w value (Fig. 7). This roughness can be quantified by fractal dimension
and measured by the relation (Dubuc 1989).
Fig. 5 Iterated midpoint displacement that produces the parabolic curve (top) when w 0.25, and
the Takagi curve when w 0.5 (bottom)
1464 I. M. Rian
Fig. 6 The construction of a Takagi surface after several iterations when w 0.5
Fig. 7 The Takagi-Landsberg fractal surface as a parametric surface. With the changing of the w
value (ranging from 0.25 to 1.0), the surface texture of the paraboloid shape also changes and so
their fractal dimensions
DH log(8w)/ log(2); for 0.5 < w < 1.0 (1)

Fig. 8 A schematic steps of translation of the mathematical function of Takagi-Landsberg surface

to its generative and parametric models
Fig. 9 A Grasshopper definition of the parametric modeling translated from the mathematical
model of Takagi-Landsberg’s fractal surface and the corresponding fractal dimension
Till w 0.5, the fractal dimension of the surface is always 2.0. But, after w
0.5, the fractal dimension of the surface gets higher in between 2.0 and 3.0 with the
increasing of w value till 1.0.
This whole process of midpoint-displacements to create a paraboloid, Takagi
surface and the Takagi-Landsberg parametric surface was executed here by using
the vector-based IFS method. For this, first, an IFS function was prepared using a
programming language which translated this operation to create a digital model of
it. The resulting model was a surface which was transformed into a mesh geometry.
It was, therefore, a generative model created by some input values that are geometric
variables such as the base size, relative size value, the scale of the base length for
the height. These variables were made changeable within a given or valid range,
thus transformed the generative model into a parametric model by taking the mesh
edges as a lattice of straight lines only. This schematic steps of this transformation
from a mathematical function to a parametric model has been illustrated in Fig. 8.
1466 I. M. Rian
Rhinoceros, a NURBS-based 3D modeling software was used as a visual interface for

this digital modeling, while Grasshopper, a parametric modeling program embedded
in Rhinoceros was used for generative and parametric modeling. Figure 9 is the
Grasshopper definition of the above operation.
2.3 FracShell: Architectural Model
The geometric model of the Takagi-Landsberg fractal surface resembles the shape
of shell-like structures. Architecturally, the appearance of the rough yet self-similar
texture makes this fractal surface unique than other typical smooth shell-like struc-
tures. This unique appearance of the fractal surface motivated us to explore it for
designing a space structure. The impact of roughness as an irregularly folded roof on
achieving the self-stiffness quality (Rian and Asayama 2016) was another reason to
verify and assess the structural application of the Takagi-Landsberg fractal surface.
In this experiment, the shell structure using the Takagi-Landsberg surface has been
named as ‘FracShell’.
For making the architectural model of the FracShell, we took the fractal surface
having w 0.6 value which is neither too rough nor smooth, and which possess a
non-integer fractional dimension that is 2.26. It means the shape of the FracShell is a
pure ‘fractal’ with reference to the definition of the fractal dimension. We decided a
triangular base having each side 4-m length. Only the third iterated model was chosen
to avoid the high number of elements. Initially, we wanted to make a continuous
concrete shell structure so that the appearance of the self-similar roughness can
be better expressed. But, for the lightness and easy construction to avoid complex
formwork for casting concrete, we preferred to make a lattice shell structure instead
of a continuous shell structure. Wood was chosen as the main construction material
for its lightness, strength, and appearance. Putting light-weight wooden panels on
top of the lattice shell was planned to achieve the self-similar multi-fold appearance
of the FracShell structure. The architectural model was prepared from the parametric
model which is a lattice of straight lines. All the lines were replaced by wooden bars
having 2.4 cm2 × 2.4 cm2 cross-section and all the connecting points were replaced
by wooden balls. Figure 10 shows an artistic impression of the FracShell structure
before the construction.
3 FracShell: Structural Modelling
It was important to check the structural feasibility, strength and stability of the archi-
tectural model of the FracShell, and for this reason, its parametric model was trans-
formed into its structural model for doing the structural analyses. In the beginning, the
structural analyses were done on the fractal model having w 0.6 and the paraboloid
model having w 0.25 separately. However, later, the structural analyses were per-
formed on the parametric structure having variable values of w within the range
of 0.25–1.0. This analyses had allowed us to find the optimal (or sub-optimal) and
suitable structure in terms of less deformation under vertical and horizontal forces.
The structural analyses were intended to be done using the finite element method and
therefore the parametric model was transformed into a finite element model (Fig. 11).
Fig. 10 An early stage architectural representation of FracShell
Fig. 11 Pre-processing scheme: transforming the parametric model into a Finite element model
and preparing for the ‘pre-processing’ for the finite element analyses
1468 I. M. Rian
Fig. 12 Structural representation of the FracShell; a main lattice, b supporting frame, c lattice with
the supporting frame, and d planar cladding on the top
3.1 Finite Element Modeling
For the analysis, we adopted the finite element analyses method, and therefore, the
parametric model was transformed into a finite element model. For the analyses,
Karamba was used which is a finite element solver, a plug-in for Grasshopper inter-
actively does the calculation with the changing of the shapes of the parametric model.
For making the finite element analyses of the structure, we prepared the digital model
as a finite element model using Karamba components at the pre-processing stage.
Followed by the architectural model of the FracShell, an equilateral triangle with
each side of 4 m was taken as a base. Only the third iterated models were taken for
the calculation. All the overlapping lines were removed and made them bar elements.
All the overlapping points were also removed and assigned them as structural nodes.
The uniform cross-section of the bar elements was 2.4 cm. Teak wood was selected
to make the bar elements. All these bars were connected at the joints by wooden
spheres having 9.0 cm diameter and considered them as the hinged joints. Three feet
of the final structure was vertically and horizontally restrained on the ground. For
the analyses and for the prototype construction, only the self-weight (1 kN/m2 ) with
uniformly projected vertical loads (1 kN/m2 ) were considered.
At the beginning of the analyses, we noticed that the FracShell as a lattice structure
(Fig. 12a) is weak because of multiple inward depressions of the shape. To strengthen
the structure, few supporting members were needed. However, our approach was to
have a minimal number of supports, and at the same time, the supports should blend
with the appearance of the lattice structure. Based on the maximum deflections at
certain nodes, we developed a structural frame (Fig. 12b) and added to the main
lattice layer to support those weak nodes (Fig. 12c). For the comparative analyses,
this additional supporting frame was added to all the structures of different w values
(Fig. 9).
Fig. 13 Post-processing scheme: nonlinear finite element analyses (large deflection) and buckling
analyses of the finite element model of the structure
3.2 Structural Analyses
After this pre-processing stage, a nonlinear static and buckling analyses were per-
formed on the finite element model of the parametric structure which was always
added to the supporting frame (Fig. 13). Karamba, the parametric FEM solver, gave
prompt feedback of structural behavior with the changing of w value. The main target
of the analyses was limited to observe the maximum displacements, axial forces and
the buckling behavior of the parametric variations of the lattice shell structure under
the gravity as well as the uniformly distributed vertical loads. With this arrangement,
first, we performed the nonlinear static analysis on the paraboloid-based lattice shell
structure having a smooth outer profile when w 0.25 and then on the fractal-based
lattice shell structure when w 0.6 having irregular or unsmooth outer profile. After
these analyses, we observed their results to compare the structural behavior between
the smooth form and the fractal form of both the lattice shell structures (Fig. 14).
Later, the maximum displacements δ of the different variations of the parametric
lattice shell structure with respect to w value have been recorded under three different
load conditions in order to find a relation between the w versus δ. Buckling is a major
issue in lattice shell structures. Hence, we also performed a buckling analysis on this
structure just to check whether the structure is structurally stable and constructible
or not (Fig. 15).
3.3 Finding the Suitable Structural Form for Construction
Because of the interactive setting of the parametric model with the finite element
solver, we were able to do an optimization taking the minimum deflection as the
main objective and considering the gravity as well as the uniformly distributed verti-
cal loads. After doing the optimization till the 100th generation, we found the lowest
1470 I. M. Rian
Fig. 14 Nonlinear finite element analyses of the maximum deflections and axial stresses (Von
Mises stress)
displacements at 0.37 < w < 0.45. The interactive setting also provided us a w-δ rela-
tion (Fig. 16). In this relation, we can notice the zone of optimal models that validate
the applied optimization process (Fig. 16). According to Mandelbrot (Mandelbrot
2002), mathematically, the outer profiles as an imaginary surface of the lattice shell
structure that lie in the ‘optimal zone’ follow Lipschitz function, which implies that
these surfaces (0.25 < w < 0.5) are rectifiable, with the fractal dimension saturated at
the value 2.0. That means, mathematically the shapes of the optimal structures are
not fractals, whereas the shapes with 0.5 < w < 1.0 have non-integer fractal dimen-
sion values (2.0 > DH > 3.0) and they are mathematically fractals (finitely iterated).
Therefore, for the real-scale prototyping, we have selected the model having w 0.6
value (1 decimal value) which has the less deflections as compared to other structures
having w > 0.5. Besides, this model offers a fancy outcome that expresses a typical
complex yet elegant appearance of the Takagi-Landsberg’s fractal surface.
Fig. 15 The first three buckling modes of the Paraboloid-based (w 0.25) lattice shell structure
(top) and the fractal-based (w 0.6) lattice shell structure (bottom)
Fig. 16 A relation between the relative size value (w) versus the maximum displacement (δ)
4 FracShell: Constructional Modelling
4.1 Constructability: Modularity of Fractal-Based Structure
After the finite element analysis of the digital model of the fractal-based lattice shell
structure, we planned to verify its practical constructability and structural viability
by making it’s real scale physical prototype. The key advantage of making a fractal-
1472 I. M. Rian
based complex structure is its scheme of self-similar repetitions which enabled us to

make a complex assembly of a single module first and then repeat the same assembly
for making other self-similar modules.
In Takagi-Landsberg surface, self-similar copies exist at the humps of the surface
shown in Fig. 17. These humps are not limited to the central part of the surface; it
happens everywhere where each copy (Ti ) is the 1/4th-scaled copy of its immediate
larger copy (Ti+1 ) (Allaart and Kawamura 2011). Generally, in pure self-similar
fractals, each part is a scaled copy of the overall appearance, while in the Takagi-
Landsberg surface, only the humps are scaled copy of the whole, but remaining parts
are the scaled copy of the non-hump parts of the Takagi-Landsberg surface. This
is due to the self-affine nature of Takagi-Landsberg surface (Allaart and Kawamura
2011) which, in fact, displays a modular scheme.
For the prototype construction, as we mentioned earlier, we selected the fractal-
based lattice shell structure whose relative size value (w) is 0.6. In this model, we
had three sets of self-similar modules (‘1’, ‘2’ and ‘3’) as three legs and one unique
module (‘4’) in the center (Fig. 18, Right). The additional supporting frame (Fig. 12b)
had some bars which occupied the usable space inside the structure (Fig. 18, Left).
4.2 Constructional Model
The generative model and the structural model created for structural analysis was
not enough to give us the detail information for the real construction. Therefore,
before the real construction, the foremost task was to transform the parametric
model into a constructional model (Fig. 19). The key element of the model was
the joints. The design workshop team came up with two different joint types using
different materials. One was circular/cylindrical joint and another was spherical joint
(Fig. 20). Because of the flexibility in three-directional connections, we preferred
Fig. 17 Self-similar repetition of the Takagi-Landsberg surface (T0) into its parts (T1, T2, T3, …
and Ta, Tb, Tc, …) in different scales
Fig. 18 Left—A simplified assemblage of the additional frame structure; right—four structural
modules, three of them (1, 2 and 3) are perfectly self-similar
Fig. 19 Schematic steps of transforming the chosen parametric model into a constructional model
for getting the detail dimensions and references for the real construction
sphere (Fig. 20, Right). The next task was to find the solution for connecting the
linear members with the spherical joints. Structurally, the size of each spherical joint
had to be light yet large enough so that all the spheres could not make the whole
structure unnecessarily heavy and all the wooden bars could be easily connected at
one sphere without congestion and overlapping of the bar ends. We found a 9.0 cm
diameter wooden ball was a suitable one. Wood was the good choice as it is light-
weight and easy to cut for connections. For the easy connection to make a lattice
(fitting/connecting one bar with already connected two joints), the length of each
wooden bar was reduced such a way that a gap of 2.0 cm was created in between
its both ends and their corresponding connecting point of the wooden ball surfaces.
This gap not only allowed us to insert a bar easily in between two wooden balls but
also allowed us to expose the balls as joints architecturally to showcase an abstract
image of the mathematical and structural elements (Fig. 27).
Another challenge was how to connect the bars with the joints at their 2.0 cm
gaps. For this, we came up with the idea of connecting those using metal screws. For
this, one end of the metal screw was attached with the connecting ball and the other
1474 I. M. Rian
Fig. 20 Two different trials for making joints; left—cylindrical ring joints; right—spherical joints
Fig. 21 Schematic steps: from constructional model to physical construction
end with the wooden bar end. With this strategy, the digital model was improved to
get the exact length of each bar. Same length bars were grouped and labeled so that
they could be easily selected for making the modules.
5 FracShell: Physical Construction
The whole fabrication process was intended to be done completely by manual

process made of mainly wood. As it was mentioned in the beginning, the purpose
of such manual construction was to demonstrate that a rule-based complex-shaped
structure can be realized without any digital fabrication tools which consume con-
siderable energy. A startegic scheme was prepared to realize the physical structure
of the model from its digital-made construction model (Fig. 21). To begin with,
the joints, i.e., the wooden balls with 9 cm diameter were made from the pieces of
wooden laths using simple hand tools such as the small wood-turner, router, and
sandpaper. The next challenge was to locate the points for holes on each ball surface
so that the bars coming from different directions can be fixed at assigned holes by
maintaining their directions. From the top, the angle between two adjacent bars near
to the ball was always 30° (Fig. 20, Right), but vertically their angles were different
and that was the main reason of difficulty to locate each hole on the ball surface.
For this, we developed our own coordinate measuring system in the wood workshop
(Fig. 22a) that could locate the points on each ball using both the vertical and
horizontal coordinates taken from the digital constructional model. Metal screws
Fig. 22 a Custom-made coordinate system for inserting the screws to the wooden ball followed by
the calculated directions; b a collection of wooden balls with the screws that define the directions
of attached wooden bars
were inserted into all balls to ensure right direction of the bars when connected as
well as to create 2.0 cm gap between bar ends and the ball surface (Figs. 22a, b).
The next item was to make the wooden bars. All the bars based on their lengths,
taken from the constructional model, were cut from wooden laths having 2.4 cm2
cross-section (Fig. 23a). All the same length bars were grouped first (Fig. 23b), and
later all the bars were regrouped based on the modular units and the corresponding
joints. Each bar was labeled to identify its two ends labeled joints, neighboring bars
and its module (Fig. 23c). Lastly, both ends of each bar are grooved so that the screws
attached to the balls can be easily inserted (Fig. 23d).
The metal screws of each wooden ball were attached to the grooves of all the
adjacent bars (Fig. 24a) and then capped with small wooden pieces using high-
strength glue (Fig. 24b). This joinery system is flexible which enabled us to remove
or add any bar from the lattice without disturbing other bar connections. Each ball
was connected first making a single unit (Fig. 25a) and then each unit was assembled
to make a single module (Fig. 25b). This way, three identical modules (‘1’, ‘2’ and ‘3’
of Fig. 21 Right) were prepared (Fig. 25c), and finally, all modules were connected
to get the final structure (Fig. 25d). After the final assemblage, the lattice shell
structure was brought to the site (Fig. 26). Because of the light weight of the wood,
the lattice structure was easily moved to the newly placed by 5–6 people only. The
structures fixed at its supports (Fig. 27 Left), and finally, lightweight transparent PVC
panels were capped on the top to express the fractal character as an unsmooth but an
approximately self-similar surface (Fig. 27 Right). The structure was standing there
successfully for couples of months without any visible deformations and buckling.
1476 I. M. Rian
Fig. 23 a Preparing the bars from the wooden laths taking the lengths from the pre-construction
digital model; b same length bars are grouped together; c bars were regrouped and labelled such
a way that their ends can match with the ends of the adjacent bars at the joints as per the digital
constructional model; d a groove was created at each end so that the metal screws fixed with the
wooden ball can be inserted easily during assembling
Fig. 24 a Bars are connected with the balls using the grooves which are filled with glue; b finally
a small wooden piece was capped to fix the steel member with each bar
Fig. 25 a Connecting the bars to make the modules; b the main part of one module was ready; c
three self-similar modules are prepared for the final assembling; d the last part (legs) were connected
to all the self-similar modules and the all these modules were finally connected to the central unique
module to complete the final structure
1478 I. M. Rian
Fig. 26 After the final assemblage, the whole structure was moved to the designated site
Fig. 27 The final structure was erected on the ground by fixing its legs. Later, light-weight trans-
parent panels were fixed in such a way that the joints, i.e., the wooden balls were exposed to exhibit
the FracShell’s mathematical and structural expressions
6 Conclusion
This workshop was designed to showcase the link between mathematics and archi-
tecture through digital design while materiality in construction was also an important
focus. This chapter has outlined the workflow of the design workshop emphasizing
the translation of the mathematical fractal surface into an architectural structure. The
FracShell project was also intended to showcase the easy manual construction of
a complex digital design. Fractal geometry shows the advantage for the rule-based
complex shape generation as well as the efficiency for the self-similar modular con-
struction.
The FracShell workshop provided us an experience that how a strategic approach

of using geometry can offer not only a unique design appearance and structural
forms but also an easy construction in order to avoid the heavy dependence on digital
fabrication. The choice of material, in fact, plays the important role in actualizing
the desirable design and facilitating the flexibility for the preferred constructional
approach. In this workshop, wood was the most suitable material which showed its
versatility in easy realizing of the complex-shaped structure. Because of the lightness
of wood, the structure was so light-weight that three people could carry it easily.
The FracShell workshop, hence, can be considered as a benchmark example that
encourages to explore fractal geometry for complex digital designs and the easy
modular construction of light-weight wooden structures.
Acknowledgements The author is highly thankful to Dr. Mario Sassone for organizing the Com-
putational Morphogenesis design workshop in which FracShell was one of its project led by the
author himself. Special thanks to Bruno Iorio, Leonardo Ramondetti, Elisa Pitassi, Samuele Marino,
Leonardo Ramondetti and Gabriele Fusaro for their active involvement in this workshop right from
the digital modeling to the physical construction of FracShell. This workshop was funded by the
Politecnico di Torino’s departmental research grant and the final manuscript development was
assisted by the University of Sharjah.
References
Allaart PC, Kawamura K (2011) The Takagi function: a survey. Real Anal Exch 37(1):1–54
Barnsley MF (2014) Fractals everywhere. Academic Press, New York
Bovill C (1996) Fractal geometry in architecture and design. Birkhäuser, Boston
Burry J, Burry M (2012) The new mathematics of architecture. Thames & Hudson, London
Dubuc B (1989) On Takagi fractal surfaces. Can Math Bull 32(3):377–384
Mandelbrot BB (1983) The fractal geometry of nature. W. H. Freeman and Co., New York
Mandelbrot B (2002) Gaussian self-affinity and fractals: globality, the earth, 1/f noise, and R/S, vol
8. Springer Science & Business Media, New York
Menges A, Ahlquist S (2011) Computational design thinking. Wiley, New Jersey
Rian IM, Asayama S (2016) Computational design of a nature-inspired architectural structure using
the concepts of self-similar and random fractals. Autom Constr 66:43–58
Rian IM, Sassone M, Asayama S (2018) From fractal geometry to architecture: designing a grid-
shell-like structure using the Takagi-Landsberg surface. Comput-Aided Des 98:40–53
Takagi T (1903) A simple example of the continuous function without derivative. Phys Math Soc
Jpn 1:176–177
Developable Wooden Surfaces
for Lightweight Architecture: Bio-Dune
Pavilion
Andrés Martín-Pastor and Rodrigo García-Alvarado
Abstract Within the broad context of teaching and research, the results of a project
involving a pavilion based on developable wooden surfaces, parametric design, and
digital fabrication are presented through the collaboration of Ibero-American uni-
versities. The project addresses applications of wood to architecture that involves
the CNC cutting of plates, thanks to the properties of developable surfaces. The in-
depth knowledge of the geometrical properties of these surfaces opens up a wide
range of morphological exploration and new constructive solutions. The proposed
system provides a creative response to the criteria of limited material resources (a
thin sheet of material) and low production and assembly costs as an analogy with nat-
ural construction. This high-tech and low-cost system, together with the Bio-mimetic
design, offers an eco-compatible proposal to the three processes under study: design,
manufacture, and assembly.
Keywords Wood design · Flexible wood · Developable surfaces

Wood complex surface · Parametric design · Digital fabrication · Pavilion
Woodworking
1 Introduction
The project forms part of an open line of research on the geometry of developable
surfaces and its application in architecture. These collaborative experiences have been
developed among Ibero-American universities, in intensive conferences on teaching
and research. The general proposal consists of approaching the phases of design,
A. Martín-Pastor (B)
ETSIE, Department of Graphic Engineering, University of Seville, Seville, Spain
e-mail: archiamp@us.es
R. García-Alvarado
Department of Design and Theory of Architecture, University of Bio-Bio, Avda. Collao,
1202 Concepcion, Chile
e-mail: rgarcia@ubiobio.cl

1482 A. Martín-Pastor and R. García-Alvarado
Fig. 1 Bio-Dune Pavilion. Hilario Hernández Gurruchaga Library, Bío-Bío University, Concep-
ción, Chile, January 2017
digital fabrication, and assembly of a wooden pavilion by means of Descriptive

Geometry and with digital tools. A low-cost and reusable construction is proposed
that is designed to be cut in a digital fabrication workshop while armed with a simple
instruction manual. The academic project is articulated in a workshop format by
applying a theoretical-experimental model that is verified through the realization of
small-scale prototypes and, finally, with the assembly of the full-scale architectural
installation.
The experience carried out in Chile (Fig. 1) is that of a hollow dune, called the
Bio-Dune Pavilion, consisting of a self-supporting skin of MDF (Medium Density
Fibreboard) of 5 mm thickness. This ephemeral installation was mounted in the access
gallery of the Hilario Hernández Gurruchaga Library, in the Concepción Campus of
the Bío-Bío University in January 2017. The pavilion was created through a workshop
within the framework of the Doctorate in Architecture and Urbanism (DAU) and the
Master’s degree in Wooden Construction of the Faculty of Architecture, Construction
and Design of the Universidad del Bío-Bío, Concepción, Chile. The work was a
collaboration with the Higher Technical School of Building Engineering (ETSIE),
of the University of Seville and it has been supported by a grant of The Iberoamerican
Association of Postgraduate Universities (AUIP).
The proposed design, fabrication and assembly system strives to provide an opti-
mized architectural response to the requirement for minimum material resources and
reduced production costs. This design is unlike other international experiences that
have applied similar strategies and equipment, but with rigid pieces and/or using addi-
tional supports (Jabi and Woodbury 2013). Digital architecture in wood is proposed
that is based on the geometric properties of developable surfaces, and on enabling
a favourable strength-to-weight ratio of the wood pieces themselves, plus the power
of digital tools and digital fabrication laboratories.
Developable Wooden Surfaces for Lightweight Architecture … 1483
2 Contextualization: Previous Experiences
The facilities designed and built in this research strategy are based on the geometrical
properties of thin wooden sheets and their ability to be bent while cold from a flat
initial state. Through the use of CAD/CAM tools, parametric generation laws have
been defined that enable different types of developable surfaces to be handled and,
subsequently, their deployment in smaller flat pieces. These pieces have been digitally
cut by a numerically controlled CNC milling machine, and then manually assembled
as a large puzzle on the ground. Each of the proposed prototypes is a self-supporting
skin made up of a single layer of material of a thickness ranging between 3 and 6 mm,
where final stability depends on the geometry and folding.
The experimental journey began with the ‘quadratic surfaces intersection theo-
rems’ put into practice with the installation ‘The Caterpillar Gallery’ (Fig. 2a) at the
University of Seville together with Fablab-Sevilla (Narváez-Rodríguez et al. 2014),
and ‘The Cocoon’ (Fig. 2b) at the National University of Colombia with Fablab Unal
Medellin (Martín-Pastor et al. 2014). We have also addressed the study of developable
surfaces of equal slope, with the pavilion ‘SSFS Pavilion-Sante Fe’ (Fig. 2c), in the
Master’s in Architecture, Digital Projection and Construction Module, of the FADU,
University of Litoral, Argentina (Chiarella and Martín-Pastor 2015). An improved
and complete adaptation of the same design was produced with the pavilion ‘SSFS
Pavilion—Fablab Sevilla’ (Fig. 2e) built in Plaza Nueva in Seville for the event
European Researchers’ Night. This installation was awarded the 2nd Ephemeral
Architecture National Award (Emporia) in the category ‘Innovation in Ephemeral
Architecture’.
The developable helicoidal surfaces of a cylindrical helix were subjected to formal
exploration with the ‘Butterfly Gallery-Helicoidal Surfaces’ (Fig. 2d) at the Federal
University of Rio de Janeiro, in the academic context of the Post-graduate program
in Architecture, FAU (Martín-Pastor 2015). This same type of helicoidal surface was
also the geometrical argument of the bio-mimetic installation titled ‘Mollusc Pavil-
ion’ (Fig. 2f), inspired by the form of Bolinus Brandaris, made at the University of the
North with the collaboration of the Fablab Unal Medellín. The expansible properties
of these helicoidal surfaces were developed in depth in (Martín-Pastor and López-
Martínez 2017). After the ‘Bio-Dune Pavilion’, the resulting surface of tangentially
convoluted linking with the installation ‘Dieste-Pavilion’ (Fig. 2g) was explored at
the Faculty of Architecture, Design and Urban Planning, UdelaR, Montevideo.
3 Lightweight Architecture in Wood as a Digital

Fabrication Process
Logic similar to that used in the furniture industry has been followed: create an opti-
mized and efficient design, and a decentralized production that minimizes materials
the cutting process and transport, and then designate the final assembly of the product
Fig. 2 a Caterpillar Gallery, Seville 2014; b Cocoon Gallery, Medellin 2014; c SSFS Pavilion-
Santa Fe, 2015; d Butterfly Gallery, Rio de Janeiro 2015; e SSFS Pavilion-Fablab Sevilla, 2015; f
Molusco-Pavilion, Barranquilla 2016; g Dieste-Pavilion, Montevideo 2017
to the user. Three main phases have been differentiated in the process: Design, Fabri-
cation, and Assembly (Fig. 3). Each of these phases is sufficiently autonomous to be
carried out by different people. As in the furniture industry, the key is in optimizing
resources, energy, and labour, which ultimately have repercussions on the cost of the
architectonic work.
Fig. 3 This diagram summarizes the three main phases of the process: design, manufacture, and
assembly, showing the lines of feedback that occur between said phases (Source Authors’ own)
The process includes these three phases in addition to tests, analyses and feedback
carried out until a final product is attained. In this project, we consider the ‘design
phase’ to be that which ends with the creation of the CNC cutting files to produce
the full-scale prototype. This phase includes the conception, the generation of the
parametric algorithm, the exploration of the form, the development of the constructive
design, the structural calculation, the flat development of the surface, the cutting of
said piece, the estimation of the cost (human resources, materials, and time factors),
and finally the preparation of the cutting files (Meredith 2008; Oxman and Oxman
2014).
The project is verified by means of the creation of a small-scale model (scale
1:8), where the possible errors in the cutting are checked, as well as the feasibility of
the assembly sequence (Sass and Oxman 2006). The fabrication phase in the digital
manufacturing laboratory is then started, which not only comprises the cutting of
the contour of each piece, but also the machining for the passage of all the bolts and
connecting elements, as well as the engravings (alphanumeric code writing) that guar-
antee the correct positioning of the pieces. This phase contemplates the refinement
of the cutting files and the optimization and planning of work within the workshop,
where strict security measures are required. Likewise, the way of packaging, storing,
and transporting said pieces to the place of assembly is anticipated.
The assembly phase requires a small number of generally non-specialized per-
sonnel for the assembly process whose protocol must be included in an annexed
document that forms part of the project. The creation of this manual with graphic
Fig. 4 Assembly manual of the Bio-Dune Pavilion (Source Authors’ own)
and precise instructions for assembly forms part of the digital manufacturing process.
In this case, it is the design team who are responsible for producing the instructions
of said assembly manual, which should allow the raising of the pavilion without any
margin of error and for any work team (Fig. 4). At the same time, safety measures
are contemplated in the work and ergonomic recommendations in these tasks.
Bringing ideas from the field of industrial furniture to the world of architecture
carries advantages from the point of view of sustainability. The concept of digital
craftsmanship emphasizes the figure of the architect, or designer, as holding the
maximum responsibility for the interaction of all digital processes related to design
and production. The classic model of the construction process is altered, where
the architect is the author of two-dimensional graphic information and a group of
specialized craftsmen/craftswomen/builders execute, with more or less expertise,
that which has been previously defined in the project. Kohler and Gramazio (2008)
place digital materiality among the complex interweaving of computer programming,
3D construction (or digital modelling), data and matter, in the various stages of the
architectonic project.
In the manufacturing process, the robotic cutting machine (CNC) is responsi-
ble for the fabrication and machining of the work. Human error is considerably
reduced, whereby humans focus on only the supervision and optimization of pro-
cesses (Iwamoto 2009). All aspects of the architecture are reduced to a set of elements
where there is no shortage or surplus of any piece of wood or any metal fixing element.
In this way, only that necessary is produced, thereby minimizing the waste generated
in the manufacturing process. Likewise, the process of packaging the material guar-
antees that only the material that is truly necessary is transferred to the assembly site,
since this is a type of joinery construction that generates no waste (Martín-Pastor
et al. 2017).
4 Geometry in the Post-digital Era, Digital Fabrication

and Construction in Wood
Approaching a project of this nature requires a triple study; one must address the
‘fundamentals of geometry’, the ‘fundamentals of software’, and the ‘fundamentals
of digital fabrication’. The network of knowledge, skills and abilities of students
needs to grow in complexity, by assuming conceptual competences (theory and fun-
damentals), instrumental competences (design tools), and productive competences
(manufacturing tools) (Celani 2012).
The ‘fundamentals of geometry’ provide the grammar that enables the operation of
higher geometric reasoning which serves in the articulation of problems and specific
solutions in architecture. Without entering into debate on the current importance of
Descriptive Geometry (Tafteberg Jakobsen and Matthiasen 2014), this discipline has
been especially prolific in treatises, and the purpose of its teaching has been to put
these skills into practice. The field of traditional woodwork (structural carpentry) as
well as other crafts, such as stonework, plastering, and tiling, is a common territory
where this knowledge can be put into practice. These geometric foundations as the
basis of digital design in wood and digital manufacturing are therefore not unusual
nor innovative.
Five phases of the architectonic process have been synthesized where knowledge of geometry
is strongly present.
‘Geometry and creativity’. The geometric idea of the project. In the tradition of architectonic
drawing, it could be related to sketching. In parametric design, we talk about the form-
generating algorithm.
‘Geometry, graphic definition, and control of the form’. The drawing of the creation, or the
3D model as a substitute for the object, which could be related to the definition that a basic
project demands.
‘Geometry of precision and of details’. The geometric definition of each of the parts of an
architectonic object and the relations of conjunction or assembly of the parts with each other.
It is a level of precision necessary for the creation of the cutting file.
‘Geometry, measurements, and economic cost’. Geometry as a measurement tool.
‘Geometry, construction, and assembly’. This is the geometry necessary in the graphic defi-
nition of the assembly systems, which explains the assembly processes and the geometry of
the control systems.
The five phases discussed above, in relation to the basic processes of digital fabrication,
principally coincide with the ‘design’, ‘manufacture’, and ‘assembly’ of industrialized pro-
duction. Each of these processes can be conceived in autonomous actions, while maintaining
close similarity with the processes of planning and industrial production.
Fig. 5 Assembly of the model reduced to a scale of 1:8, and where to check for possible errors and
to optimize processes (Source Authors’ own)
The experience of design, manufacture, and assembly of a pavilion from the geo-
metric foundations also becomes an exercise of intergroup communication and coor-
dination, whereby students can work on the necessary skills of teamwork (Fig. 5).
Likewise, the final scenario of the assembly and display of the installation is also
revealed as fundamental in carrying out these relations and optimizing the program-
ming of the assembly (Alvarado et al. 2009).
Another major aspect incorporated into this working formula should be borne
in mind: the digital tool is conceived as a laboratory of geometric research where,
empirically, the geometric and structural properties of the surfaces to be put into
practice are investigated. In this way, a variable that transcends that of purely teaching
is added to the initial approach to convert the experience into an experimental research
workshop (Casale et al. 2013).
Fig. 6 The dune is generated as the enveloping surface of a series of cones of the same slope
(Source Authors’ own)
5 Dunes as a Source of Inspiration: Bio-Mimesis,

Morphogenesis, and the Project
Dunes are dynamic elements of topography and include a natural adaptation of the
terrain to climatic variations with minimal surfaces. The incorporation of bio-mimetic
or morphogenetic logic with respect to formal search techniques constitutes a field
of exploration that has been addressed by numerous researchers over the last decade
(Hense et al. 2010; Weinstock 2010). The shape of a dune in its natural state is
defined by various parameters: the angle of internal friction of the granular material
that forms it, the prevailing winds of the area, the natural obstacles, the presence of
vegetation, humidity, the pockets of material in its lowest area, etc. However, from
an ideal point of view, any dune is, geometrically, a surface of equal slope defined by
the natural angle of the terrain that forms it (Fig. 6). A dune is always a developable
surface in that it can be unfolded on a plane without pleats, bulges or folds, and this
provides enormous constructive advantages. The dune could therefore be understood
as a hollow structure formed by a thin sheet of material that works in its entirety as
a self-supporting skin.
Under the assumption of this law of generation of a dune, the composition of the
project can be determined by a very precise geometric and mathematical logic. This
exerts consequences on the level of pure design, since this parameterized architecture
offers strong tension between the formal creation and the margin of freedom of
the generation law chosen. The formal, constructive, and structural solutions are
articulated within the infinite variants and/or mutations that the initial genetic code
allows, which is assumed as part of the project (Whitehead 2005).
The morphological research and the study of the material possibilities of the
project are therefore strongly linked to the tools, which in this case constitute the
parameterization software. The designer makes use of these tools by exploring their
possibilities, anticipating the final result without the modifications of their parameters
dissociating the relationships of the parts from each other, and by maintaining the
integrity towards its final manufacture (Ramsgard Thomsen and Tamke 2013).
Therefore, and paradoxically, the artisanal quality of the process, ‘digital craft’,
will tend to be increasingly present in the initial phases of design, and necessarily
tend to disappear in the final processes, since the ultimate goal is the automation of all
reiterative actions, including those of Self-Detailing and Self-Documenting Systems
(Meredith and Kotronis 2013). A previously analysed concept, called “constructed
digital materiality”, is thereby connected (Chiarella et al. 2018).
5.1 Developable Surfaces of the Same Slope: Algorithms

of Generation of the Form
Developable surfaces are well-known mathematically: they belong to the category

of ruled surfaces, as constituted by a set of straight lines arranged according to
a type of law, and are usually classified into three groups: cylinders, cones, and
tangential surfaces (Glaeser and Gruber 2007). Their study is strongly marked by
the mathematical progress of the eighteenth century and the figure of Monge (1769,
1780, 1785, 1795a, 1795b), Jules de la Gournerie (1860–64), among others. The
morphological approach has been carried out from graphic thinking, following the
conceptual premises of Descriptive Geometry, the potential of digital tools, and
parametric design. These surfaces can be studied from two different directions: on
the one hand, the approximate processes can be studied in order to convert any
type of surface into a discrete series of developable surfaces (González-Quintial
et al. 2013; Barrallo et al. 2017), or, on the other hand, they can be studied in order
to morphologically enter into greater depth into the different ways of generating
surfaces that are strictly developable in a continuous way.
The dune is included within the developable surfaces ‘of equal slope’, this being
a specific case of the tangential surfaces. It is not a usual surface in architecture
although its generation does appear in certain Descriptive Geometry texts, such
as that by Izquierdo Asensi (1985), Gentil Baldrich (1990), and others close to
Architectural Geometry (Pottmann 2010; Glaeser 2012). Despite being a tangential
developable surface, its edge of regression (the curve where all the straight lines are
supported tangentially, and the one which separates the two leaves from the surface)
is not known a priori, unlike the helicoid or other developable surfaces. Therefore,
the generation of the surface must be formulated from a different curve on which
to impose the angular conditions. Two possibilities of parametrically generating
the surface are considered, which correspond to the two ways of its interpretation
(Martín-Pastor 2019).
The first consists of conceiving the generation from the bottom up. In this way,
a flat directrix guideline is determined, considered as the ‘contour of the dune’, on
which a set of straight lines are supported perpendicularly, which maintain a constant
angle β with respect to the plane of the directrix. Using this formulation, the crest
of the dune, which is a three-dimensional curve, is entirely determined by the two-
dimensional contour of the guideline and the support angle β.
The second formulation consists of conceiving the generation of the surface from
the crest of the dune downwards. This formulation strives to roughly determine the
enveloping surface of the displacement of a cone along the upper curve or crest. The
generatrices of the cone, since this is an equiangular surface with respect to a horizon-
tal plane of reference, determine the slope of the surface. With this formulation, the
contour of the base of the dune can be determined by means of the three-dimensional
geometry of the crest.
The two algorithms that control the shape of the dune from the two conditions
presented above (the base and the crest), must be sufficiently debugged and be suffi-
ciently operative for the design. Only then can they become an appropriate tool for
the exploration of architectonic form and, consequently, useful for design purposes.
It should be considered that the software used (Rhinoceros-Grasshopper) does not
always handle these developable surfaces precisely with their usual “development”
commands. Therefore, the developability (ability to move from space to the plane)
becomes an algorithmic problem in itself. In order to solve this new problem, a third
processing algorithm has been defined.
5.2 The Constructive System
The construction system of the Bio-Dune Pavilion has been conceived by taking into
account the feedback between the phases of design-manufacture, design-assembly,
and fabrication-assembly, from previous experiences. In accordance with the objec-
tives, the proposed architecture must have no improvisation in its manufacturing and
assembly phases (Hauschild and Karzel 2011).
Its structure and materiality constitute a self-supporting skin formed, in principle,
by any thin laminar element, capable of being curved when cold. The use of wood is
supported by its aesthetic and constructive properties, such as lightness, flexibility,
ease of cutting and assembly, structural strength of the resulting final assembly and
the safe handling and assembly of the pieces, as well as its low cost (compared to
other thin laminars such as sheet metal) (Solomon et al. 2012). The tensile strength
of the fibres of the wood around the fixing points also offers a suitable response.
The assembly process of the pavilion is planned to take place on-site (Fig. 7).
Once the whole pavilion has been development as a flat surface, it must be broken
down into smaller parts, no larger than the dimensions of the panels of the local
industry. The dimensions of the MDF panels in Concepción, Chile, were 1200 ×
2440 mm and the cutting machine was a CNC with a bed of 1300 × 2500 mm.
The self-supporting skin of the pavilion rests on a 30 mm-thick base beam that
is fixed to the pavement. In this beam, a series of brackets are placed vertically and
fixed by screws. The function of these brackets is to fasten the skin along the base-
beam, thereby forcing the surface of the dune to rest on said beam, and hence the
final shape is acquired (Fig. 8). In turn, the beam absorbs the horizontal stresses of
the surface lamina of the dune, which, through active tension, strives to recover its
flat shape (Figs. 9 and 10).
In order to guarantee the rigidity of the free edges of the accesses to the pavilion,
two curved beams of triangular section were created along the edges. This beam is
made by folding the skin of the pavilion inwards twice. As with the skins, the beam
was cut with respect to the maximum dimensions established by the local industry.
The geometrical continuity of all the pieces that form the large surface of the
pavilion is achieved by bolting a 160 mm-wide overlapping strip behind each cutting
line. The bolts are staggered along four lines parallel to the cut. These bolts are
separated by a distance of 250–350 mm, which guarantees the mechanical behaviour
of the material. This system of connection is created by means of a T-nut, a bolt, and
a washer. The perforations for the passage of the bolt and T-nut are made by CNC
machining.
The system of fastening the skin of the dune with the beam, and that of fastening
the edges of the triangular beam that reinforce the accesses, is carried out by means
of nylon flanges. These flanges sew the different skins along the edge of the crest
of the dune with the help of mechanized holes, arranged specifically in the design
(Fig. 11).
Fig. 7 Reorganization on site. Bio-Dune Pavilion (Source Authors’ own)

Fig. 8 Assembly of the Bio-Dune Pavilion as a single skin (Source Authors’ own)
Fig. 9 Curvature and fastening of the skin along the crest of the dune and along the base beam
(Source Authors’ own)
Fig. 10 Bio-Dune Pavilion, final phase (Source Authors’ own)
5.3 Panelization of the Final Surface and Generation

of the Cutting Files
From a quantitative point of view, the surface that makes up the pavilion is constituted
by a series of pieces extracted from standard panels; this leads us to consider several
ways of proposing optimization. A first criterion contemplates taking advantage of
the maximum amount of material. Another optimization criterion prioritizes the use
of the smallest number of machined end pieces for its construction (leaving open
the possibility of not taking full advantage of each panel). A third criterion seeks
to reduce the number of joining elements (T-nuts and bolts) and the extra material
needed for the overlap, as well as the assembly time.
For the minimization of the material resources, the most effective criterion is that
which needs the smallest number of panels and the smallest number of overlaps
between pieces. However, other factors unrelated to this geometric optimization
intervene in the way that the cutting is broken down. In the first place, it is necessary
to guarantee the correct mechanical behaviour over time of the structure. To this end,
it is important to prevent the orientation of the overlaps from coinciding with the
orientation of the straight lines which conform the surfaces, by maintaining an angle
preferably close to 45° between said surfaces. Secondly, it is important to avoid long
seam lines that travel across the surface from component to component. To this end,
the pieces are staggered. Finally, it is important to reinforce the free edges of the
surface, either with edge ribbing, or with reinforcements made specifically for that
Fig. 11 Bio-Dune Pavilion, interior image and triangular access beam detail, with overlaps and
fastening system using nylon flanges (Source Authors’ own)
purpose. It is also necessary to ensure that no hole or machining is located between

the seam lines or over a bolt.
A more complete optimization criterion should take into account other issues,
apart from the purely quantitative structural criteria and the criterion of material:
these include the cutting time, ease of packaging of the pieces, ease of transport, and
ease of assembly. The work of cutting is therefore a complex task to automate and
remains far from the precise incorporation of each and every one of the parameters that
intervene therein into the algorithmic programming (Bohnenberger et al. 2011; Burry
2013). This demonstrates what was previously stated regarding the feedback between
the design, manufacturing and assembly processes, in that a constant revision and
refinement between phases is crucial for the development of this approach (Mesnil
et al. 2016; Austern et al. 2018).
6 Results, Constructability and Lines of Further

Improvement
The theoretical and practical nature of the proposal and its vocation to be a complete
exercise of digital construction, raises the obligation to specifically define the tools
or machines (software, computers, cutting systems, etc.) for the control and produc-
tion of an effective human-machine interface. From this theoretical reflection, it has
become sufficiently clear that Fundamentals in Geometry provide the structure that
is ‘common’ to these linguistic processes.
As a result, we have taken an experimental approach to the production of
lightweight architecture in high-tech and low-cost wood. This occurs within an aca-
demic context thanks to optimized cutting processes, and a sustainable production
model with its own self-assembly system. Making the leap from the academic world
to the industrial world means redefining several of these premises, and accepting a
higher level of automation and rationalization.
In previous experiences and also in this participation, the assembly of the instal-
lation has required a large number of people due to the weight of the structure. It
is possible to investigate the possibility of new materials that enable this type of
difficulty to be reduced, such as employing recycled wooden products, wood foam
and cork chipboard in an effort to reduce the density of the material.
It is important to study the curvature at each point in order to verify that there are
no areas with a higher curvature than that which sheet of material can support. This is
a direct contribution that digital technologies and parametric design can incorporate
into the process. In the case of the Bio-Dune Pavilion, excessive curves can be seen
in the vicinity of the end point of the crest. This led to an optimization of the design
in accordance with strict geometric conditions, which involved the elimination of
material by applying the curvature map of the surface.
7 Conclusions
Geometry is understood as knowledge prior to the use of parametric tools. With

this knowledge, the condition of ‘developability’ of a surface has been understood
and refined, and its morphological variants have thereby been explored. Low-cost
wood laminars have been applied and subjected to variable efforts, which enable
the space to be extensively covered with an ingenious execution. This experience
therefore demonstrates new design paths for wood design and construction, which
in turn open the door to new opportunities.
The Bio-dune pavilion experience held in Concepción, Chile, expresses an innova-
tive application of bio-mimetic geometry for wood construction; it employs advanced
design and manufacturing methods, and a major international collaboration in both
research and teaching. In this context of digital manufacturing in wood, ‘Geometry’
appears as the conceptual support capable of giving a specific response to the require-
ments for the generation of the different forms of the project, their execution, and
optimization. Not only does this system aim to optimize the use of material and struc-
tural efforts, but also to optimize the conception and production. Hence, inspired by
the furniture industry, the combination of digital technologies and natural resources
with modest local capacities not only provides a unique and significant product, but
also presents an operational method that allows the generation of new architectural
forms.
Acknowledgements
CREDITS
BIO-DUNA PAVILION
Project and Design: Andrés Martín Pastor
Collaborator Design: Alicia López Martínez
Academic Coordination: Rodrigo García Alvarado
Manufacturing coordination: Alexis Pérez Fargallo. Manufacturing team: Prototype
Laboratory, Escuela de Diseño Industrial, Fernando Figueroa Ramírez, Víctor Urrea
Founet.
Assembly: Collaborative between teachers and students of the Workshop “Geometría
y Fabricación Digital en Madera”: Andrés Vizcaíno, Juan Pablo Navarrete, Pablo
Matus, Helmut Martes, Alex Hernández, Elvert Duran, Danitza Pereira, Francisca
León Solís, Hugo Cáceres Bravo, Raúl Henriquez Reyex, Francisca Lavoz Fue-
nealba, Rodrigo Ceballos, Vicente Valenzuela. External collaborators: Stefánia Jara,
Francisca Rozas.
Funding: Magister en Construcción en Madera, Doctorado en Arquitectura y Urban-
ismo, Centro de Investigación en Tecnologías de la Construcción (CITECUBB),
Asociación Universitaria Iberoamericana de Postgrado. AUIP.
Collaborations and acknowledgments: Cecilia Poblete Arredondo (UBB). Ricardo
Hempel Holzafphel (UBB). Víctor Mora Padrón (HUBiobio). Ariel Bobadilla
Moreno (CITECUBB). Verónica Otarola Sanzana (UBB). Carlos Campos Saez
(UBB). Pamela Sierra Villalobos (UBB). Pedro Concha Márquez (CITECUBB).
Department of Graphic Engineering, Universidad de Sevilla. Escuela Técnica Supe-
rior de Ingeniería de la Edificación, ETSIE, Sevilla. Research group “HUM976:
Expregráfica. Lugar, Arquitectura y Dibujo”. Instituto Universitario de Arquitectura
y Ciencias de la Construcción (IUACC. Universidad de Sevilla). Research group
“Arquitectura y Construcción Sustentable. GACS-UBB”.
PREVIOUS WORK CREDITS
The Caterpillar Gallery
Project and Design: Roberto Narváez-Rodríguez, Andrés Martín-Pastor
In collaboration: Margarita Infante-Pereda, María Aguilar-Alejandre
Manufacture: Fablab Sevilla
Assembly: Teachers and students of ETSIE and ETSA, Universidad de Sevilla

Collaboration: Department of Graphic Engineering, ETSIE, Universidad de Sevilla
The Coccon Gallery
Project and Design: Andrés Martín-Pastor, Roberto Narváez-Rodríguez
Manufacture: Fablab UNAL Medellín team
Execution and Assembly: students of UNAL Medellín and UNAL Manizales
Coordination: Facultad de Arquitectura. Unal Medellín, Colombia
SSFS Pavilion-Santa Fe
Project and Design: Andrés Martín-Pastor, Roberto Narváez-Rodríguez
Parametric co-design: Juan Expósito Bejarano
Academic Coordination and management: Mauro Chiarella
Programa Maestría en Arquitectura, FADU, Universidad Nacional del Litoral,
Argentina
Manufacture and Assembly: Teachers and students of FADU
The Butterfly Gallery
Project and Design: Andrés Martín-Pastor, Juan Expósito Bejarano
Academic Coordination: María Angela Dias. Programa de Pós-graduação em
Arquitetura, FAU, Universidade Federal do Rio de Janeiro (UFRJ), Brasil
Manufacture: LAMO3D, Foco Design
Assembly: Teachers and students of UFRJ
Collaboration: Department of Graphic Engineering, Universidad de Sevilla, ETSIE,
Fablab Sevilla
SSFS Pavilion-Fablab Sevilla
Project and Design: Andrés Martín-Pastor, Juan Expósito Bejarano y Roberto
Narváez Rodríguez
Manufacture: Fablab Sevilla
Production Manager: Daniel del Valle. Safety director: Alejandro Flores Pina
Assembly: Teachers and students of ETSIE, ETSA and Edificalab
Collaboration: European Researchers’ Night, Department of Graphic Engineering,
Universidad de Sevilla, ETSIE
Molusco-Pavilion
Project and Design: Andrés Martín-Pastor, Alicia López Martínez.
Academic Coordination: Antonio Olmos and Johanna Ebrahaim. Fablab Uninorte.
Universidad del Norte, Barranquilla, Colombia.
Manufacture: Fablab UNAL Medellín. Universidad Nacional de Colombia, sede
Medellín.
Manufacturing team: Edgar Alonso Meneses, Juan Pablo Gutiérrez, Luis Daniel
Asprilla, Camilo Andrés Idárraga.
Assembly: Teachers and students of Universidad del Norte, Colombia.
Collaboration: Department of Graphic Engineering, ETSIE, Universidad de Sevilla,
Fablab Sevilla.
References
Alvarado RG, Lagos R, Salcedo P, Ramos M, Labarca-Montoya C, Bruscato U (2009) Emociones

precisas: fabricación digital en la enseñanza de la arquitectura; en Arquitetura Revista. Unisi-
nos—Universidade do Vale do Rio dos Sinos 5(2):122–136
Austern G, Guedi Capeluto I, Jacob Grobman Y (2018) Rationalization methods in computer aided
fabrication: a critical review. Autom Constr 90:281–293. https://doi.org/10.1016/j.autcon.2017.
12.027
Barrallo J et al (2017) Laminar constructions and reciprocal structures. Nexus Netw J 19:723.
https://doi.org/10.1007/s00004-017-0346-8
Bohnenberger S, Rycke K, Weilandt A (2011) Lattice spaces: form optimisation through customiza-
tion of non-developable 3D wood surfaces. In: 29th eCAADe conference proceedings. eCAADe:
Conferences. Ljubljana, Slovenia: University of Ljubljana. Faculty of Architecture, pp 751–758
Burry M (2013) From descriptive geometry to smartgeometry. In: Peters B, Peters (eds) Inside
smartgeometry: expanding the architectural possibilities of computational design. Wiley, Londres
Casale A et al (2013) Surfaces: concept, design, parametric modeling and prototyping. Nexus Netw
J 15:271. https://doi.org/10.1007/s00004-013-0146-8
Celani G (2012) Digital fabrication laboratories: pedagogy and impacts on architectural education.
Nexus Netw J 14:469. https://doi.org/10.1007/s00004-012-0120-x
Chiarella M, Martín-Pastor A (2015) Thinking graphic and design collaborative. Developable
geometries for folded architectural compositions, In: Proceedings of the XIX conference of the
iberoamerican society of digital graphics 2015, Blucher, São Paulo, pp 702–707. https://doi.org/
10.5151/despro-sigradi2015-110107
Chiarella M, Martín-Pastor A, Sáez N (2018) Graphic thinking and digital processes: three built
case studies of digital materiality. In: Castaño-Perea E, Echeverria E (eds) Architectural draughts-
manship, from analog to digital narratives. Springer, pp 1033–1044. https://doi.org/10.1007/978-
3-319-58856-8_81
Gentil Baldrich JM (1990) Método y aplicación de la Representación Acotada. Cuadernos EGA 2,
Sevilla
Glaeser G (2012) Geometry and its applications in arts, nature and technology. Springer, New York
Glaeser G, Gruber F (2007) Developable surfaces in contemporary architecture. J Math Arts
1(1):59–71. https://doi.org/10.1080/17513470701230004
González-Quintial F et al (2013) Freeform surfaces adaptation through developable surfaces using
apparent contours. In: Zhang J, Sun C (eds) Global design and local materialization. CAAD
futures 2013. Communications in computer and information science, vol 369. Springer, Berlin
Hauschild M, Karzel R (2011) Detail practice: digital processes. Birkhäuser Architecture
Hense M, Menges A, Weinstock M (2010) Emergent technologies and design: toward a biological
paradigm for architecture. Routledge, Oxon
Iwamoto L (2009) Digital fabrications: architectural and material techniques (architecture briefs).
Princeton Architectural Press
Izquierdo Asensi F (1985) Geometría Descriptiva Superior y Aplicada. Dossat
Jabi W, Woodbury R (2013) Parametric design architecture. Laurence King, London
Kohler M, Gramazio F (2008) Digital materiality in architecture. Lars Müller Publishers
La Gournerie JM (1860–64) Traité de Géométrie descriptive. Mallet-Bachelier, Paris (3 vols)
Martín-Pastor A (2015) Um retorno a os fundamentos da geometria. The Butterfly Gallery-
Superfícies Helicoidais. Estratégias Geométricas para a Fabricação Digital. In: Cadernos
PROARQ. Revista de Arquitetura e Urbanismo 25, UFRJ Post-graduation Program in Archi-
tecture, Rio de Janeiro, pp 2–30
Martín-Pastor A (2019) Augmented graphic thinking in geometry. Developable architectural sur-
faces in experimental pavilions. In: Marcos C (ed) Graphic imprints, Springer, Berlin, pp
1065–1075. https://doi.org/10.1007/978-3-319-93749-6_87
Martín-Pastor A, López-Martínez A (2017) Helicoides desarrollables de hélice cilíndrica y su

aplicación como superficie arquitectónica. In: Proceedings of Geometrías17, Coimbra, 16–18
June 2017, pp 73–77
Martín-Pastor A, Torres-Holguín J, Narváez-Rodríguez R, Galindo-Díaz J (2014) Los workshops
de geometría en Cad3D y prefabricación digital como estrategia docente en la enseñanza de la
geometría para la arquitectura. Geometría y proyecto. In: Proceedings of the XVIII conference
of the iberoamerican society of digital graphics: design in freedom 2014, Blucher, São Paulo, pp
213–216. https://doi.org/10.5151/despro-sigradi2014-0040
Martín-Pastor A, Martín-Mariscal A, López-Martínez A (2017) Rethinking ephemeral architecture.
advanced geometry for citizen-managed spaces. In: Mercader-Moyano P (ed) Sustainable devel-
opment and renovation in architecture, urbanism and engineering. Springer, Berlin, pp 301–310.
https://doi.org/10.1007/978-3-319-51442-0_25
Meredith M (2008) From control to design. Parametric/algorithmic architecture. Actar, Barcelona
Meredith N, Kotronis J (2013) Self-detailing and self-documenting systems for wood fabrication:
the burj khalifa. In: Hesselgren L, Sharma S, Wallner J, Baldassini N, Bompas P, Raynaud J (eds)
Advances in architectural geometry 2012. Springer, Vienna
Mesnil R, Douthe C, Baverel O, Léger B (2016) Marionette mesh: from descriptive geometry to
fabrication-aware design, advances in architectural geometry 2016, vdf Hochschulverlag AG,
Zurich. http://dx.doi.org/10.3218/3778-4_7
Monge G (1769) Sur les développées des courbes à double courbure et leurs inflexions. J Encycl
284–287
Monge G (1780) Mémoire sur les propriétés de plusieurs genres de surfaces courbes, particulièrte-
ment sur celles des surfaces développables, avec une application à la theorie des ombres et des
pénombres. Mémoires de divers sçavans 9, 9: 593–624 (written 1775)
Monge G (1785) Mémoire sur les développées, les rayons de courbure, et les différents genres
d’inflexions des courbes à double courbure. Mémoires de divers sçavans 1, 10: 511–50 (written
1771)
Monge G (1795a) Feuilles d’analyse appliquée à la géométrie, Paris (Second ed. with the title e
Applications de l’analyse à la géométrie. Paris: Baudouin, 1811)
Monge G (1795b) Géométrie Descriptive. First ed. in e Les Séances des écoles normales recueillies
par des sténographes et revues par des professeurs, Paris
Narváez-Rodríguez R, Martín-Pastor A, Aguilar-Alejandre M (2014) The caterpillar gallery: quadric
surface theorems, parametric design and digital fabrication. In: Block P, Knippers J, Mitra N,
Wang W (eds) Advances in architectural geometry 2014. Springer, Berlin, pp 309–322. https://
doi.org/10.1007/978-3-319-11418-7_20
Oxman R, Oxman R (2014) From composition to generation. In: Theories of the digital in archi-
tecture. Routledge, New York
Pottmann H (2010) Architectural geometry as design knowledge. Archit Des 80:72–77. https://doi.
org/10.1002/ad.1109
Ramsgard Thomsen M, Tamke M (2013) Digital crafting: performative thinking for material design.
In: Peters T, Peters B (eds) Inside smartgeometry: expanding the architectural possibilities of
computational design. Wiley, London
Sass L, Oxman R (2006) Materializing design: the implications of rapid prototyping in digital
design. Des Stud 27(3):325–355
Solomon J, Vouga E, Wardetzky M, Grinspun E (2012) Flexible developable surfaces. Comput
Graph Forum 31:1567–1576
Tafteberg Jakobsen I, Matthiasen J (2014) Descriptive geometry and/or computer technology? What
mathematics is required for doing and understanding architecture? Nexus Netw J 16:505. https://
doi.org/10.1007/s00004-014-0199-3
Weinstock M (2010) The architecture of emergence: the evolution of form in nature and civilisation.
Wiley, Chichester
Whitehead H (2005) Laws of form. In: Kolarevic B (ed) Architecture in the digital age: design and
manufacturing. Taylor & Francis, New York
SMALL IS MORE. Wooden Pavilion As
a Path of Research
Marco Imperadori, Graziano Salvalai, Andrea Vanossi
and Federica Brunone
Sense experience and necessary demonstration

Galileo Galilei
Abstract Several shelters have been recently developed by Politecnico di Milano

for different purposes but with a common character: smallness. Small is More shows
how workshops with students can be extremely efficient and productive on real tar-
gets and how materiality is important to learn and step over any parametric approach.
Of course, parametric tools can be very useful to manage the design phase and option-
eering strategies but, in the end, only real built structure unveils the power of archi-
tecture. The paper will show some of this examples where Smallness can be of great
effect: Akragashelter in Agrigento—Valle dei Templi (UNESCO Workshop), with a
structure in gabion-wall, the roof in fir wood, and parametric design; C-ASA, Active
House Shelter in Lecco (Politecnico di Milano ASA Workshop), with CLT structure,
aerospace thermo-reflective insulations, and zinc-titanium outer skin; Equilatera,
the social Shelter in Favara (FARM Cultural Park), digitally designed, with recycled
CLT wasted studs for the structure; Teagloo, in Tokyo with Kengo Kuma Lab (The
University of Tokyo), digital designed structure in sushi bamboo baskets, with FRP
reinforcements; Veneer House, in VIS (Croatia) with Prof. Hiroto Kobayashi and
KOBAYASHI MAKI DESIGN WORKSHOP (Keio University, Tokyo). The learn-
ing process has been thought with real targets for social purpose and students have
participated in designing real case studies, which are now still in full function.
Keywords Wooden shelters · Smallness · Materiality · Design optioneering

Real prototypes
M. Imperadori (B) · G. Salvalai · A. Vanossi · F. Brunone


1502 M. Imperadori et al.
1 Introduction
Design can only find its reliable check by real construction. After modeling, checking,
optioneering, and prototyping through scaled mock-ups, the only real way to see if
it works or not is to build. How is it possible involving students and respecting low
budgets and safety? Small pavilions are an opportunity of endeavor and research
about functions, structure, volumes, and aesthetics.
The use of small pavilions for experimentation in new technologies and new
approaches in construction is very evident in a series of small pavilions presented in
a set of conferences that, starting from Advanced in Architectural Geometry 2008, has
been held in all over the world in the last ten years. Smart Geometry or Computational
Design Symposium are other examples of this process. Parametric is the driver,
students from many universities, like Politecnico di Milano, TU University of Vienna,
ETH Zurich ecc.. and professionals from companies like Zaha Hadid, Norman Foster,
Arup… present every year innovation in a small scale pavilion or mock-up with the
aim to test it for a new path in buildings. To design and to build pavilions is also
a way of innovation from the process point of view, indeed this approach allows
comparison between student’s ideas and professional approach.
Pavilions have been used also in the past, before the parametric era, from famous
architects to test their design. The use of pavilions to investigate the design thinking
in order to extend it from a small pavilion to real building is a methodology very
common in the approach of Kengo Kuma. In this paper, several examples from Kengo
Kuma And Associate (KKAA) have been used as references (Kuma 2018; Kuma and
Daniell 2015).
Wood is clearly one of the materials that offer many possible chances to tackle this
field, and the digital/parametric approach, added by CNC-based production, helps
to check materials and volumes in real 1:1 scale (Sass 2007; Sheil 2012). But all
this endeavor must be done with a purpose, a real, social and long-lasting purpose.
Without it, without a specific use and a reliable application, all the experiments are
ok, but not enough.
This article presents the design process of some small wooden pavilions, modeled
and built by and with the students of Politecnico di Milano, selected during the
Building Design and Innovations workshop, led by Prof. Marco Imperadori.
The first case is Akragashelter (realised in 2013), a canopy built for the Val-
ley of the Temples—UNESCO area near Agrigento—with the aim to protect some
archaeological artifacts with great value. Close to Demetra Temple, some old Greek
columns have been protected by a structure in gabion walls, bearing a wooden roof.
Digital modeling has optimized the solution, which was sketched as a pitched roof
with the same slope of the Concordia Temple, adapting it also to the different ground
configuration. Two glulam beams, connected with steel bars on pillar line, bear three
hinges lattice massive pine studs which give to the roof, covered by FRP undulated
panels, the feeling to “fly in the forest”: An automatic CNC-based (Computer Numer-
ical Control) cutting process has been applied to all wood elements to speed up the
process. The structure still functions for the same purpose.
SMALL IS MORE. Wooden Pavilion As a Path of Research 1503
The second experiment is C-ASA (realised in 2016), the new Sport Office of
Politecnico di Milano—Lecco Campus (https://www.activehouse.info/cases/c-asa-
polimi-sport-office/). After a competition among students, a concept has been
selected and implemented through digital design methods, both for architectural
and engineering purposes. The office is a minimum space of less than 15 m2 . All the
structure has been conceived as a wood box, with cantilever roof, made in CLT (Cross
Laminated Timber). These wooden elements have been cut by CNC-based processes
in factory and delivered in big elements on site in order to be quickly assembled.
After that, all envelopes, inside and outside, have been realized with thermo-reflective
multi-layered insulation (ACTIS). Therefore, the small office is insulated like an air-
craft and fully equipped with sensors. Final skin outside is zinc-titanium (Zintek ®)
and inside is special gypsum board with nanotech zeolite (KNAUF Cleaneo panels).
Equilatera (realised in 2017), the third small project, is the output of a process
dedicated to recycling CLT waste cuts. CLT technology is based on planar elements
(walls or slabs) which are produced in standard dimensions because of industrial
and transport reasons. Transformers and turn-key house builders cut these structural
panels to optimize their dimensions, causing, therefore, a certain amount of CLT
waste, normally in long studs with some lattice oriented in length and the cross lattice
glue to give a certain structural collaboration. Hence, the idea beyond Equilatera is
to use only one element in length and connect it in a triangular shape (self-bracing)
with a dense structure, which is connected by wood screws (Rothoblaas). As CLT is
normally fir or pine with low resistance to water, the structure has been covered by
FRP undulated panels with a topping in galvanized steel which allows and enhances
ventilation. The pavilion was firstly assembled by the team of Politecnico di Milano
and ESPE—the training institute for the construction sector, in Lecco. Then it was
delivered in Favara at FARM Cultural Park (https://www.farmculturalpark.com/) and
became the core of this art center, hosting and protecting people from rain or sun,
thanks to the fiberglass roof, and functioning as a social attractive point.
Teagloo (realised in 2015) is the result of a workshop held in Tokyo between the
students of Politecnico di Milano and the University of Tokyo—Kengo Kuma Lab
(http://kuma-lab.arch.t.u-tokyo.ac.jp/). The goal was to design a small ephemeral tea-
house (which represents the core of Japanese aesthetics) in the park of the University
of Tokyo, close to the famous and ancient AKA MON (the Red Gate). The starting
point was the bamboo baskets, with a clear structural pattern, which normally are
used for sushi. Parametrical design has been performed both by the design teams
of Milan and Tokyo, to assess the shape after the first concept based on sketches.
Thanks to parametrical design, it was possible to assess the correct number of baskets
and also to verify with Prof. Jun Sato, a famous Japanese structural designer, that
instead of a dome shape the surface should have been more undulated (like a shell) to
achieve the greater rigidity. On the floor, a tatami mat was ready for the tea ceremony
and many kids enjoyed playing with this funny space, looking like an animal hut,
which was magic, porous and collecting the beautiful yellow gingko leaves falling
in Tokyo autumn.
The Veneer House (realised in 2016) was born using the concept of Prof. Hiroto
Kobayashi (Keio University of Tokyo) (https://www.kmdw.com/veneer-en), who,
jointly with Prof. Marco Imperadori, decided to erect this small pavilion in the island
of Vis (Croatia) in the framework of the international workshop Anatomy of Islands.
The idea was to interpret the Veneer system, developed by Kobayashi for emergency
and post-disaster reasons, to create a public small pavilion on the Adriatic Sea in Vis.
Veneer system is conceived with no nails and all the panels and fixing elements are
the results of parametrical design and CNC-based cutting processes, which optimizes
the use of Veneer panels. These panels have been supplied by Federlegno Arredo,
Italian wood association, shipped in Slovenia, and processed by the team of Prof.
Tadaj Glazar of the University of Ljubljana. Then, all the codified pieces have been
delivered to Croatia and by boat from Split to Vis. Once there the team of students
and tutors assembled the small pavilion, like a toy, in a beautiful olive garden, facing
the sea and very close to the Venetian fortress.
All these experiences represent examples of a possible pedagogic path, for students
and designers to learn through real case studies the logic of materials, connections and
proportion, approaching the concreteness of architecture (Imperadori and Senatore
2011).
2 Small Is More: An Applied Method on the Design Process
Small is more is a design concept that collects a set of different design experiences
on small pavilions. According to a crossing perspective, all of them could be defined
as real case studies of an applied methodology of designing, which starts from prime
conceptual sketches (Imperadori and Senatore 2011) and ends with the realization of
small architectural objects and buildings, passing through an integrated path between
the virtuality and materiality of constructions (Parvin 2013; Sass 2007; Sheil 2012).
This paper describes, therefore, the methodological approach that has brought to
the realization of five constructions, through the account of their direct applications
and enhancing the role of wooden elements and products that have been used.
All the experiences, indeed, could be encoded into a precise path, an intense design
process split up into three different phases:
• At first, the design development on paper, driven by sketches that intensively dig
the prime architectonical concept, the size, and shape, the general geometries,
but also the technical details that could involve the following realization phase
(Imperadori and Senatore 2011);
• At second, the project is integrated by the experimentation, for students, on virtual
tools, from the parametric and design optioneering methodology to deal with the
geometries and structural issues, and/or the software for digital simulations of the
physical and thermal behavior of the building (Parvin 2013; Sass 2007; Sheil 2012;
Vanossi et al. 2014a, b; Vanossi and Imperadori 2013);
• Finally, the construction phase, whose technical aspects and issues could be
directly experimented by the teams of students, coordinated by Professors and
Researchers, thanks to the small scale of the objects.
These three moments have not been just consequential, but have been a part of
an iterative process, that has moved back and forth between them, in order to dig all
the different aspects and issues of the design, digitalization, and construction.
Akragashelter, C-ASA, Equilatera, Teagloo, and Veneer house projects have fol-
lowed this breakdown structure, in different ways and using different type of wood-
based and derived products (structural grid of beams and pillar, CLT panels, CLT
studs, bamboo components or veneer elements), attesting how this construction mate-
rial, in its different forms, could fit the purpose of this research and pedagogic path.
3 Set of Case Studies
3.1 Akragashelter
The Akragarshelter project was developed within the “Architecture X Archaeol-

ogy” UNESCO workshop, held by the Park Agency of the Valley of the Temples
—Agrigento, together with Tokyo University, Politecnico di Milano, and the Uni-
versity of Palermo (Vanossi et al. 2014a, b). The construction of coverings for the
ancient archaeological ruins of Greece temples aimed to carry out an experimental
action, which can produce innovation in the context of lightweight and temporary
structures. The “Architecture X Archaeology” workshop (Vanossi et al. 2014a, b)
explored the state of the art in terms of works protecting archaeological landmarks,
with the goal of identifying new modular and flexible structures, which are adapt-
able to the diverse needs of archaeological sites. The use of wood and the parametric
and computational strategies are the bases used for the design of the shelters. The
first preliminary part of the workshop introduced issues relating to the complex-
ity of designing within archaeological sites, applications of computational design,
structural stability, and practical solutions for quick physical materialization within
limited time and budget. The second part explored culture and nature of which served
as a pretext for further exploring design for archaeological sites. Moreover, the shelter
theme represents a complex theme because induces a substantial modification of the
site aspect. In case of construction—besides technical problems such as anchoring
the structure to the ground—the site will be inescapably changed by the addition of
a new volume on the site. In this sense, the teams were asked to see their respective
project as a shape that merges and connects to the landscape, and not as an opposing
element.
Among the different projects realized, Akragarshelter design process started from
the definition of a modular and flexible structure adjustable to different scenarios and
needs (Fig. 1). The prime concept referred to the cultural and architectonical heritage
of the site, Akragas (old name for Agrigento). Akragashelter was defined as a light
wooden canopy, supported and anchored to the ground by a structure in gabion walls,
and whose pitched roof had the same slope of the Concordia Temple, near there. Due
to the sacredness of the place and the impressive archaeological site, the team opted,
Fig. 1 Akragashelter project started from ideas, thoughts and considerations, directly impressed
on paper through sketches; those have followed through the entire design process, with the aim to
focus on the details as well as to give the impressions of the general volume (Credit Prof. Marco
Imperadori—Politecnico di Milano)
indeed, for a simple design solution, which used of simple materials such as wood,
corrugated fiberglass, and rocks collected on site. Moreover, as the design work was
shared with the Kengo Kuma Lab (http://kuma-lab.arch.t.u-tokyo.ac.jp/)—which was
developing its own pavilion in the same site – the team referred to some aspects of
Japanese culture—Zen mysticism and Temple of Ise—as well as local one—Greek
classicism (of Concordia Temple). The reference to the Greek temples of Agrigento
is expressed in the project through the use of massive columns realized in galvanized
steel gabions filled with local stones which serve for anchoring the shelter to the
ground. This reference is highlighted by the slope of the roof, with an angle of 22.5°,
identical to that of Concordia Temple. Moreover, the wooden roof evokes the typical
gable roof of the Greek temples. Concerning the reference to Japanese culture, the
shelter gets inspiration from Ise Shrine, the most important Japanese shrine, made
in Inoki (cypress) wood. In particular, it looks at the temporality and reversibility
of Ise—which is disassembled and reassembled every 20 years—and at the shape
of the roof summit, surmounted by a V-shaped ventilation duct for water collection.
The roof structure is covered by a white corrugated fiberglass which gives a soft
atmosphere to the shelter when lit up at night, with an effect similar to a Japanese
lamp made of rice paper.
Akragasheleter was designed by a parametric and associative model (Fig. 2). The
shelter –which is modular and flexible–was conceived as a module that can be built
in different areas of the park. Four variables have been taken into consideration for
making the model: (1) the variety of sites with ruins in the park, (2) the different
orientations of the sites, (3) the different size of the ruins to be covered by the shelter,
(4) the uneven level of the ground. Each variable was considered as a parameter in the
digital model. The digital model designed can adapt its global geometry and that of
its components to different variables. The height of every pillar is adjusted to ground
Fig. 2 Parametric model and construction information database of Akragarshelter (Credit Politec-
nico di Milano)
conditions. The ground was modeled as a cloud of points built by importing data
through a laser scanner survey. Length and width of the roof came as a result of the
dimensions of the ruins. Every component of the shelter is hierarchically associated
to the geometry of other components in a tree of associative relationships. Once the
shelter is digitally modeled, it is possible to simulate the real condition of the site
and associate i.e. its geometry to a database from which it is consequently possible
to extract: amount, dimension and specifications of construction elements.
Finally, in just one week, the team of professors and students built the small archi-
tecture (Fig. 3), simple and respectful of the surroundings, functional and elegant,
perfectly integrated into the natural environment of the Valley of the Temples. A
Fig. 3 The construction phase of Akragarshelter has involved the students of the international
workshop “Architecture x Archeology”, held at the Parco Archeologico Valle dei Templi of Agri-
gento, together with Tokyo University, Politecnico di Milano, and University of Palermo (Credit
Politecnico di Milano)
special and holy place, where architecture and nature combine and unite themselves
in an empathetic way (Fig. 4).
Conceived and finished in a very condensed period of less than 45 days, both
computational design and digital fabrication enabled the speed and the challenges
undertaken in each of the shelters. The Japanese and Italian students interacted prof-
itably, both with the local authorities and with the park functionaries and architects.
The architectural design meeting took advantage of the scientific supervision of
Kengo Kuma and Jun Sato as well as the contributions of subject matter experts like
Giuseppe Guerrera, former director of a master course about Archaeology held at the
University of Palermo. The synergies among universities, public functionaries and
craftsmanship represents a unique moment of exchange and growth which opens new
paths to explore. The experimentation shows the potentiality in the use of Parametric
and BIM-based process, especially with reference to a flexible and adaptable con-
Fig. 4 Akragarshelter is still standing, after its construction in 2013, at the Valley of the Tem-
ples—UNESCO site near Agrigento, Sicily, where it still protect the ancient archeological ruins of
the Greece temples (Credit Politecnico di Milano)
struction. The possibility offered by this approach to adapt the shelters to different
morphologic situations and the possibility to extract updated information according
to the different shelters modules have been identified as a strategic concept for the
construction of futures shelters.
3.2 C-ASA
C-ASA, is the name of the new Sport Office of Politecnico di Milano—in Lecco
Campus (Salvalai 2017). It stands for Construction—ASA, representing the design,
engineering and realization processes that follow the Advanced School of Archi-
tecture (ASA) international workshop and competition, held in Lecco Campus of
Politecnico di Milano, among selected students. After the competition, hence, two
of the awarded concepts have been joint and implemented with digital design, both
for architectural and engineering purposes (Fig. 5).
The resulting architectural concept has been simple and inspired by the cube
geometry, with a useful floor surface area of 9 m2 (height of 3.00 m) for a total volume
of 27 m3 . The roof is characterized by projecting overhang that protects the facades
from the direct summer sunlight. The dimensions and positioning of the windows and
the skylight have been designed to favors a superior natural light and thermal comfort
conditions. During the design phase and the engineering process of the project, the
possible layering of the building envelope has been deeply investigated, in order to
guarantee high levels of performances and a faster construction phase, almost based
on prefabricated components. CLT panels actually define the load-bearing structure:
this kind of components are the result of an assembly of wooden layers, that are
cut with CNC-based machines in fabric, and then delivered on site for the quick
assembly procedures. The wooden structure is covered inwards with a multilayer
thermo-reflective insulation (ACTIS triso super 10+), bonded together and sealed by
Fig. 5 C-ASA was born as the integration and re-calibration of two awarded concepts at the
Advanced School of Architecture competition, held among international students at Politecnico
di Milano—Lecco Campus (Credit Prof. Marco Imperadori—Politecnico di Milano)
Table 1 The material layering of the facade

n° layer Thickness (m) Description Thermal Specific Density (kg/m3 ) Thermal
conductivity heat c (J/kg K) resistance
λ (W/m K) R (m2 K/W)
1 (Inside) 0.0125 Gypsum 0.200 840 760
panel
2 0.0125 Gypsum 0.200 840 760
panel
3 0.0620 Actis Triso − 1.183
Super 10
4 (Outside) 0.0830 CLT 0.130 1380 460
PANELS
means of a special airtight tape. The 54 m2 of the vertical wall has been finished with
zinc-titanium slab (Zintek) and inside the roof has been finished with special gypsum
board with nanotech zeolite (Knauf Cleaneo). The effect of the Knauf Cleaneo slabs
is based on the combination of gypsum and Zeolite able to decompose several air
pollutants in basic substances not harmful to the building users. A decentralized heat
recovery ventilation unit with high efficiency has been installed in order to guarantee
the required air change rate. All these features lead the project to be awarded the
Active House Label, which attests the high performances of the building according
to the principles of comfort, energy savings, and low environmental impacts.
Tables 1, 2, 3, and 4 list the detailed features of the designed Structure/Envelope
solutions, whose performances have been validated throughout a campaign of a
virtual simulation of the building’s behavior, according to its location, orientation,
materials and user interaction during the operation phase. In particular, the façade is
composed, from outer to inner side, by zinc-titanium slab on OSB wood board, CLT
panel, multi-layer insulation, double gypsum board. The door and the two windows
are composed of a triple glass with 90% of Argon with a total thickness of 36 mm
(Solar factor of 47%). The frames are composed by 85 mm of lamellar wood, and
30 mm of XPS covered by a thin aluminum sheet. The total U value of the windows
is equal to 0.860 [W/m2 K]. A skylight has been designed in the center of the roof to
increase the daylight and the natural ventilation. A mechanical opening and shading
system operated by remote control allows the users to adapt the internal comfort
condition. Finally, the thermal resistance of the envelope has been guaranteed by
innovative multi-layer insulator of 19 different foils composed of several materials
among which foam, batting and thin reflecting layer. The properties of the material
have been analyzed in another investigation, obtaining an R-value of 1.183 m2 K/W
(Salvalai et al. 2015, 2017; Salvalai 2016).
Once the design strategies have been validated, the construction phase could
start. It has been deeply affected by the prefabrication processes and all the related
advantages. The pre-fab and pre-cut wooden panels have been installed in an easy
and quick way. The overall wood structure has been implemented in one day starting
from the floor to the vertical walls and the roof. The building envelope has been first
Table 2 Material layering of the roof

n° layer Thickness (m) Description Thermal Specific Density (kg/m3 ) Thermal
1 (Inside) 0.0125 Gypsum 0.200 840 760
panel
2 0.0400 Mineral 0.040 670 40
Wood
3 0.0370 Actis Triso 0.042 1600 1000
Super 10
4 0.0830 CLT 0.130 1380 460
PANELS
5 0.030 Actis Triso 1.183
Super 10
6 0.025 OSB 0.036 1700 650
7 (Outside) 0.0008 Zintek 109 401 7140
Table 3 Material layering of the floor

n° Layer Thickness (m) Description Thermal Specific Density (kg/m3 ) Thermal
1 (Inside) 0.0120 Parquet 0.200 1800 1300
2 0.0330 Brio floor 0.380 100 84.84
3 0.030 Actis Triso 1.183
Super 10
4 (Outside) 0.0830 CLT 0.130 1380 460
PANELS
finished in the inner site using gypsum board and then in the outer side positioning
the two color zinc titanium slabs. Moreover, as a further guarantee of airtightness, the
joints between the wood panels as well as those between wall and floor and between
wall and ceiling have also been carefully sealed. The window holes, both in the wall
and on the roof, have been sealed to perfection using self-expanding membranes and
special sealing tapes. The quality of the building has been tested using the Blower
Door Test procedure; considering a pressure of 50 Pa the air infiltrations have been
registered equal to 2 ACH (corresponding to 67 m3 /h) verifying the high quality
of the building envelope. At last, the gap between building floor and the external
pedestrian surface has been managed through a wooden walking ramp covered in
vertical parts with high pressure laminated panels (HPL). The sheets consisting of
layers of wood-based fibres (paper and/or wood) impregnated with thermosetting
resins and surface layer(s) on both sides, having decorative colours. A transparent
topcoat is added to the surface to enhance weather and light protecting properties. All
the prefabricate element has been codified and assembled in the building site. The
decorative panels high-pressure compact laminates (HPL) with thicknesses of 8 mm
Table 4 Thermal properties of the building envelope, calculated according to UNI EN ISO
6946:2018
Façade
Total thickness (m) 0.17
Internal thermal capacity (KJ/m2 K) 15.96
Thermal resistance R (m2 K/W) 2.11
Thermal transmittance U (W/m2 K) 0.472
Roof
Floor
have been installed on vertical wood elements by means of 6 mm screw coloured as

the panels finishing (Fig. 6).
After its construction and inauguration, the building has continued to be a proto-
type for research projects, being the object of several in situ measurement, to attest
the reliability of the dynamic simulation performed at the design stage. A continuous
monitoring campaign had been carried out for 5 days, from 4th to 9th of March 2016.
The sets of sensors had been installed according to the scheme explained in Fig. 7a.
The set-up is composed of one data logger and 5 temperature sensors positioned on
the north facing the wall and in the center of the room.
The wall performance measurement is based on two temperature level measure-
ments, on both wall side, and heat flux meter positioned on the inner surface. Fur-
thermore, one temperature sensor took measurements of the external air temperature
and a temperature and humidity sensor measured the interior conditions. All the
sensors were connected to a data logger with wireless communication. The data
analysis had been analyzed according to the Standard ISO 9869 (UNI 2015), which
gives the measurement methodologies, the equipment to be used and the data pro-
cessing procedures, considering the variability of the measured phenomenon. The
measure consists of the acquisition of the values of heat flux density and temperature
levels through the wall with a frequency of 6 min. The measured surface tempera-
V
tures and heat flux had been used for estimating the AM—the value of the wall
before and after retrofit with the progressive “Average Method” according to the
above-mentioned standard by means of the equation below:
Fig. 6 The construction phase of C-ASA involved several prefabricated components, in order to
make the assembly easier and faster: e.g., the CLT structure, whose massive panels have been
precut in factory with CNC machines, by Ri-Legno (Credit Chiara Mariska Chiodero, Daria Pantò
Mancuso, and Cristiana Topo, for Politecnico di Milano)
Fig. 7 The digitalization processes involved different stage of the project development, from the
dynamic simulations of the building’s behavior, to the definition of technical details for the structural
components in wood (CLT panels) and the post-construction campaign of measurement in situ. This
figure shows, in the upper part, the technical drawings realized by Ri-Legno to manage the execution
procedures (Credit Ri-Legno), and, below, the results of the monitoring campaign through sensors:
in the left graph, Temperature level analysis (data collected with a frequency of 6 min), in the right
one, Temperature and humidity sensor: in red the value of thermal transmittance; in black the value
of heat flux (Credit Politecnico di Milano)

n
qj
j1
Λ n

Tsij − Tsej
J 1
The U-value has been calculated indirectly, using the reference values of inside
and outside surface heat resistance (Rsi 0.125 m2 /K W and Rse 0.04 m2 /K W).
The following formulas have been used for the thermal transmittance calculation:
1
U 1
Rsi + Λ
+ Rse
The method needs a minimum temperature difference of 10 °C between the inside

and outside to work and it is assumed that the thermal boundary resistance is constant.
The graph of the Fig. 7b shows the temperature level of the internal and external
ambient. The Fig. 7c shows the U value calculated according to the ISO standard
showing an average thermal transmittance of the wall equal to 0.46 W/m2 k.
Fig. 8 C-ASA, the new sport office of Politecnico di Milano—in Lecco Campus, is also a built
prototype used by students and researchers to investigate different aspects of buildings’ behavior
and performances (Ph. credit Irene Giunta for Politecnico di Milano)
In conclusion, C-ASA has been the object of experimentation for students on

wood potential as construction material, since the first conceptual idea. This small
shelter has been conceived, developed in its architectonical and engineering aspects,
realized, and tested by students, during a complex and long endeavor (Fig. 8).
3.3 Equilatera
Equilatera is an experimental project based on the reuse of the wood waste from the
CLT industry. Its name derived from its triangular shape, which was conceived during
the prime design stage. The brainstorming and collection of ideas and project propos-
als have been fixed on paper through an intense endeavor on sketches and technical
drawings that defined the first architectural concept and the following technological
details (Fig. 9).
The starting point has been the recovering of 66 laminated studs, derived as co-
products from the cutting processes of CLT panels. This structural system, indeed,
is under an ongoing diffusion in the construction market, for its mechanical per-
formances, and its high feasibility and fastness of the assembly procedures on site.
CLT panels result from the cross gluing of linear elements in factory, where the final
planar elements are pre-cut to fit the dimensions needed for specific projects, before
the delivery on site. This part of the production process generates a considerable
amount of wasted material, CLT linear elements, which still presents high levels of
mechanical performances and could be re-used as structural members.
The project development, therefore, consisted of the assembly of those 66 tim-
ber studs, into a modular construction system: three linear elements—4500 mm’
length, 100 mm’ thickness and with variable widths—compose a triangle that is
repeated to define a space with a flexible length, and, therefore, use and function
(Fig. 10). The intersection of the wooden elements creates a particular configura-
tion of the upper joint, letting the natural ventilation effect across the shaped space,
and defines a cantilever geometry on the bottom, that has been used as the load-
Fig. 9 Equilatera project has been conceived with the aim to recycle the wooden waste derived
from the CLT panels cutting process (Credit Prof. Marco Imperadori—Politecnico di Milano)
bearing structure of an external bench. The so-defined modular structure has been,
then, validated in its mechanical performances, according to the technical specifi-
cations of Italian standards (Circolare n. 617 del 02/04/2011; CNR-DT 2007, 2008;
D.M. 14/01/2008; D.M. 06/05/2008; UNI 2014, 2005, 2010a, b, c) the calculations
have been confirmed by the structural simulations on the software MyProject, by
Rothoblaas, which attested the structure’s resistance to its own weight (0,60 kN per
element), and wind loads (variable values according to the structure location).
Set the basic configuration of the wooden elements, different design solution
about the spatial scheme have been analyzed, with the support of the parametric
design (Fig. 10) and the optioneering approach (Vanossi and Imperadori 2013). The
use of the well knows visual programming tool Grasshopper sustained the designer
in the creation of an algorithm that combines different solutions in order to optimize
the use of the 66 elements in the definition of the space under the shelter. Combining
different options, indeed, a simple triangular solution resulted in the best for max-
imizing the surface for using the pavilion. The use of the model assumed different
characteristics according to the phase of analysis. Indeed, one of the advantages of
working with small pavilions is the possibility to follow all the phases of the life, and
eventually the new life, of the designed object. Starting from the concept and moving
to the realization, the working group have had the chance to design and think about
all the aspect of a new construction entity. The model has been considered as pow-
erful tools for analyzing different options in the form finding process. Optioneering
is a methodology based exactly on the evolution of the design idea starting from
parametric constraints and parametric opportunities. In this case, several constraints
like the number and the dimensions of the components, and the area of the pavilion,
have been compared with opportunities like an evaluation of different material on the
covering of the pavilions, or like displacements of the components in different ways.
The model has been developed and detailed as a model for construction, analyzing
the sequences of assembling, the connection between the components and the details.
Finally, in the last phase, the model becomes a repository of a database containing
all the information needed for the use of the small pavilion, before becoming again
new material for a new cycle of life. This small pavilion with its simple geometry has
been generated by a very advanced “design engine”. Indeed, behind the realization,
there is a complex process based on a continue workflow, from concept to realization.
The small dimension in term of building mass and the short time that passed from
the idea to the delivery of the pavilion allows to the group to test a methodology
based on a new way in design, based on continuing fast revision and analysis of the
results. The advantage is that more options can be fast evaluated, more solutions can
be considered in the design process.
Equilatera resulted, therefore, into a unique space, with a maximum height of
4.40 m and a planar width of 4.80 m. It has been conceived as an open shelter,
located outside and, therefore, directly exposed to natural elements. Because of that,
the CLT frames needed to be covered by protective layers, without nullifying the
contrasting effect of lights and shadows created by the modular repetition of the
wooden beams, along with the space development. Thus, the CLT structure has been
finished with a series of translucent corrugated panels that do not hide the underlying
Fig. 10 The project has been studied also according the parametric design processes, in defining
the wooden structure and the covering layers (Credit Vito Iovino, Luca Scandella and Cristina
Savoldelli for Politecnico di Milano)
rhythm of the supporting wooden bones and, at the same time, avoid its otherwise
inevitable rapid degradation. Due to its wooden composition, moreover, the small
architecture has been conceived to be placed detached from the ground, thanks to
a longitudinal substructure made of ties and fixed through steel plates. Besides,
in order to ensure the ventilation effect across the side walls through a small air
chamber between the supporting structure and the translucent façade, the fiberglass
panels have been fixed on a series of wooden slats, whose heads have been treated
with a white impregnating varnish to prevent any infiltration of water damaging
the substructure. A folded steel profile protected the joint at the top of the beams,
allowing, moreover, the activation of the natural ventilation chimney. This effect
has been verified through both fluid dynamics simulations and in situ measurements,
performed after the construction. The interior flooring was covered with light-colored
wooden beads, which maintained the wooden matrix of the project and highlight the
contrast between lights and shadows created by the structure when it is hit by the
sun’s rays. Outside, the bench has been thought to be covered with a waterproof and
transpiring membrane, to protect the beams from possible water infiltrations, and
finished with a double layer of concrete slabs and Caltagirone majolica.
The so-detailed design phase allowed to enter into a deep definition of the object
and report all those technical specifications into a sort of Guide manual for the con-
struction process. The instructions follow three principal stages, described through
3D and 2D drawings of the technical details:
1. the construction of the basic module, three CLT beams with the same size fixed
together with the aid of self-piercing screws;
2. the vertical positioning of the first triangular module, and the second one, to be
anchored together;
3. the repetition of the second phase, for each module, following the same previous
directions.
These operations have been carried out, indeed, during the first assembly at
ESPE—the training institute for the construction sector, in Lecco, where the design
team of Politecnico di Milano and the students of the professional institute tested the
feasibility of the construction processes (Fig. 11). The prototype has been then disas-
sembled, to be transferred and re-located in Favara, at FARM Cultural Park. Once the
material arrived on the site, the design team and local workers started with leveling
the ground floor on which the small pavilion would rise up. Being it a temporary
structure, the basement should be easily removable in the future, quickly achievable
and, at the same time, endowed with high resistance. It was defined, hence, as a
perimeter of tufa blocks filled with gravels, to allow rainwater to flow and avoid the
stagnation below the structure. After that, the CLT modular structure has been easily
and quickly assembled and covered with the selected finishing materials (Fig. 11).
After the construction in May 2017, the inauguration took place concurrently
with the Opp Festival, of which Equilatera was one of the protagonists. The small
temporary shelter has been, indeed, a covered space to organize social and commu-
nity lunches, according to the first conceptual idea of the project and following the
philosophy of FARM Cultural Park (Fig. 12).
Fig. 11 The construction phase of Equilatera involved the teams of students from Politecnico di
Milano and ESPE, which assembly the structure at first in Lecco, to attest its feasibility, and finally
in Favara, with the help of local workers (Credit Vito Iovino, Luca Scandella and Cristina Savoldelli
for Politecnico di Milano)
Fig. 12 Equilatera resulted in a small pavilion in the core of i 7 Cortili of FARM Cultural Park,
functioning as a social attractive point for thousands of visitors (Credit Vito Iovino, Luca Scandella
and Cristina Savoldelli for Politecnico di Milano)
3.4 Teagloo
The idea of creating a tea house based on the use of bamboo baskets for sushi came
from the principle of reusing processed materials after their “traditional” lifecycle.
This upcycle process consists, indeed, to reinvent a new life for materials that are
normally at the end of their proper use.
The design process of Teagloo started exactly from the circular components, the
bamboo baskets, positioning and moving them on a 3D surface, in order to create a
covered space with a tatami area for a traditional Japanese tea ceremony. Since the
first sketches and considerations (Fig. 13), the ideal surface has been identified into a
semi-sphere, especially in consideration of the geometrical and structural properties
of the geodetic geometry.
The first sketches and design concepts, then, were transferred into a parametric
environment. Here, the shape of the pavilion has been based on a geodetic semi-
sphere, whose surface has been approximated by triangles (containing the baskets).
The maximum number of baskets (1,000) and their fixed radius have been consid-
ered as a constraint for the geodetic algorithmic definition (Fig. 14). Several options
(Vanossi and Imperadori 2013) have been investigated in order to fix the right bal-
ance between the tatami proportion, the dome, the baskets and the entrance to the
pavilion. An important aspect that needed a special focus was the definition of the
real mechanical properties of the bamboo baskets since this kind of objects was not
originally conceived as construction components. The rigidity of the small containers
collected by the team of Kengo Kuma Lab—University of Tokyo did not coincide
Fig. 13 The Teagloo pavilion design process started with the sketches about its shape and the
position of the small bamboo baskets, before any kind of parametric approach (Credit Prof. Marco
Imperadori—Politecnico di Milano)
Fig. 14 Geodesic hypothesis (left) for the assembly of the bamboo baskets and real model based
on the rigidity of the bamboo baskets (right) (Credit Politecnico di Milano)
with the values previously considered as constraints, in the first parametric analysis
and derived solution for the project. With the help of Professor Jun Sato, the structural
approach changed, from considering the gravity of the material as the driver aspect
for the assembly to suspension: the pavilion had to be hung like a sheet or a wooden
tent. Following these instructions, after its definition into a parametric environment,
the small temporary structure has been assembled and built up in the garden of the
University of Tokyo, by the students of Premio Compasso Volante—Ettore Zambelli
workshop, together with the team of Kengo Kuma Lab (Fig. 15).
Even though this temporary and ephemeral structure seems very simple, it encodes
in itself a lot of complexity. Materials and algorithms have been combined together,
showing how complex is to manage an even-small construction. Very often the design,
especially for young designer rest in their mind without matching the reality of the
material and of the construction. Small pavilions allow overcoming this limit, testing
the real property of the material, their rigidity, their connection each other and their
relation with the design. This extreme, small example has shown very well how
complex is the translation between design and realization enhancing the value of the
methodology “small is more” (Fig. 16).
3.5 Veneer House
The Veneer House pavilion is a small teahouse located in Vis (Croatia), realized in
2016 for the XVII edition of the international design competition for university stu-
dents Premio Compasso Volante—Ettore Zambelli, led by Prof. Marco Imperadori.
Fig. 15 The assemblage of the small pavilion, made of bamboo baskets, in the courtyard of the
University of Tokyo, has involved the students of the Kengo Kuma Lab and Politecnico di Milano,
with the supervision of Kengo Kuma, Prof. Jun Sato, Proff. Marco Imperadori, Manuela Grecchi,
and Matteo Brasca (Credit Politecnico di Milano)
Fig. 16 The Teagloo pavilion after the realization, used as innovative tearoom, with the tatami
carpet, typical of the Japanese culture (Credit Politecnico di Milano)
This experience took place in the framework of the international workshop Anatomy
of Islands 2012–2016, held in Vis, and which has involved the academic institutes
of Keio University of Tokyo, Politecnico di Milano, and the University of Ljubljana
into the realization of a small pavilion as a first glimpse of the future development
of the island.
The project started since the end of 2015 as a shared experience of collaboration
between the academics of Politecnico di Milano and Keio University of Tokyo. Their
meetings in several events, workshops, and symposiums gave the birth to the prime
idea: to use the design concept of the Veneer system by Prof. Hiroto Kobayashi to
build up a small wooden construction, and host a place of gathering and inclusion.
The system has been developed by the KMDW—Kobayashi Maki Design Work-
shop as a solution to quickly address the need for housing, in post-disaster scenarios.
The first experience, indeed, was moved by the catastrophic event of March 11th,
2011, when a magnitude-9.0 earthquake—officially known as the Great East Japan
Earthquake—hit the Tohoku region of Japan, causing a crushing tsunami that dev-
astated the villages, with more than 17 thousand victims, 340 thousand refugees.
Beyond the first responses to the immediate needs, long-term strategies to host the
surviving communities needed to be defined.
Hence, as several other Japanese architects, the Kobayashi’s team developed a
new-engineered building methodology, starting from the idea of prefabrication as a
strategy to answer the issues of an emergency. The Veneer House system, indeed,
exploits the great potential of wood to be used into prefabrication processes and
introduces a method of structural plywood assembly, based on Japanese notch-cut
joinery (Hiroto Kobayashi+VVAA 2017) (Fig. 17). Plywood panels are cut and easily
assembled, without any steel screw or nails, thereof defining the main characteristics
of this method as:
• Speed—since time is one of the most essential factors into a post-disaster response
in recovering community, the system consists of a quick and rapid assembly of
defined panels, pre-cut offsite;
• Economy—involving wooden elements with a high recycled content, the system
represents a solution in terms of low-cost constructions, affordable both as emer-
gency shelters and new buildings for the communities under regeneration;
• Simplicity—the single connection and the optimization of the panels into few
different geometries make the system as simple as possible, enabling local civilians
to build by themselves their own buildings and home, disseminating a feeling of
involvement and community;
• Innovation—the system uses the most advanced computer-oriented technology,
such as the CNC routers and laser cutters, according to the idea of prefabrication;
• Flexibility—the notch-based joint method perpendicularly connects two panels
without any additional elements, making the structure firm and defining an assem-
blage that could be used of the entire construction, at different scales (until the
children’s toys one).
The flexibility aspect, moreover, could be perceived also at a higher level, since
the construction system has been adopted in several different occasions and loca-
Fig. 17 The Veneer House project in Vis started from the analysis and deep study of the Japanese
Veneer House System by Prof. Hiroto Kobayashi, a shell construction (c) made in plywood precut
panels (b), interconnected by notch-cut joinery (a). Below, the scaled mockup, by Nicolò Panzeri,
Enrico Sterle, and Akitaka Suzuki for their master thesis project “Insight Vision”, supervised by
Prof. Grabiele Masera (Credit Politecnico di Milano)
tions. The Minami-sanriku Community House and the Maeami-hama Community

House in Miyagi Prefecture, the Ma Naw Heri Village Learning Center in Myanmar
represented the first trials. Then, other experiences in Philippines, Nepal, Japan, dig
into the experimentation, towards the continuous optimization of the components,
joints solutions, and dimensions of the panels.
All these features portray the Veneer House construction method as a very innova-
tive one, a perfect object of studying and testing for students, through direct experi-
mentation. Once the site in the island of Vis has been defined—a school country-yard
facing an olive grove yard near the seaside—, therefore, the international team of
professors, researchers and students started to work on the small pavilion project,
analysing and adapting the construction system to its new place and function, and
checking and discussing the outcomes. The resulting building consists of a unique
space, 4.133 × 3.654 m, whose dimensions derived from the repetition of four iden-
tical frames that encode the three structural modules. The external ones have a span
of 0.91 m, while 1.81 m for the internal one (Fig. 17c). The chosen position makes
the small building facing the courtyard from North, the olive garden from East,
while the southern and western façades remain close, from trees and bushes. This
disposition influenced the treatment of the different sides: the southern one hosts the
entrance, leading people to turn around the pavilion and watch it from the outside
before entering. Besides, the international nature of the entire project guided the
definition of the function and details. The Croatian Veneer House would be used,
indeed, as a traditional Japanese teahouse, to enhance the cultural exchange between
the two countries; the northern side is, therefore, enriched by a traditional round
window, designed to be at the precise height to see the bay while sitting inside.
Afterward, a double sliding door assures the permeability to the olive garden, on the
eastern façade, to whom an external wooden deck is connected, recalling again the
traditional aesthetics of a Japanese teahouse.
One of the most interesting and advanced aspects of these experience consisted on
the digitalization of the design and production process, following the so-called Open
Source Architecture approach. It is an inclusive paradigm to the design process, based
on the radical changes that the advent of technology has caused – the Computer Aided
era for design and manufacturing, with CAD and CAM technologies. The “design
fabrication” workflow has been redefined, thanks to digital fabrication tools, enabling
designers to directly communicate with fabricating machines, into a cohesive process.
This approach found a fertile application field within the Veneer House project,
thanks to both the high potential of the wooden materials to be treated throughout
this kind of processes, and the advanced skills of all the actors involved (Fig. 18).
Therefore, once the geometry of the pavilion has been set (Fig. 18b), according also
to the structural verifications, the digital data has been shared within the widespread
design team members (Fig. 18a, c), and the production process started. 80 plywood
panels, Russian pine, 18 mm’ thickness, 2440 × 1220 mm’ dimensions, supplied
by Federlegno Arredo were delivered to the High School near Ljubljana, where
Professor Tadej Glazar guided the cutting procedures by a CNC machine, checking
the device tolerances and all the possible incongruities between the CAD files and the
realized components (Fig. 18d). Then, the pre-cut pieces arrived at the construction
site in Vis, where the foundations of the pavilion were already set, made by the local
company. Over the concrete foundations, the structure arose from the assembly of
all the pre-cut plywood panels (Fig. 18e, f), completed in just one day of work. After
that, the essence of the pavilion was already fulfilling the site, but, despite its raw
beauty, the construction process had to continue with the protective layers: the roof
was realized with polycarbonate sheets, fixed on a wooden substructure between the
plywood frames, while spruce wood slats made the finishing materials of the façades
(Fig. 19). Meanwhile the international workshop Anatomy of Island was held, several
students have been involved into the construction phases, from Politecnico di Milano,
Keio University of Tokyo and University of Ljubljana, as design team members, and
also the University of Zagreb, University of Zadar, University of Split, Université de
Limoges, and Tokyo University, as collaborators and supports.
Fig. 18 The scheme resume the phases followed during the entire design process of Vis Veneer
House, according to the so-called Open Source Architecture approach, an inclusive paradigm, based
on the radically changes that the advent of technology has caused—the Computer Aided era for
design and manufacturing, with CAD and CAM technologies (Credit Nicolò Panzeri, Enrico Sterle,
and Akitaka Suzuki for their master thesis project “Insight Vision”, supervised by Prof. Grabiele
Masera, Politecnico di Milano)
Fig. 19 Veneer House construction in Vis has involved not only the selected students from Politec-
nico di Milano; it has been an international experience, involving both the teams of several uni-
versities and local people, making the community to participate and enjoy the small pavilion, even
before its realization (Credit Politecnico di Milano)
Fig. 20 Finally, the small pavilion unveils a strong relationship and balance between the interna-
tional nature of the project, and its local definition to be an object for the community, even in its
architectural details (Credit Politecnico di Milano)
The pavilion arose habitants and visitors’ curiosity, engaging locals into both the
realization process and the definition of the future possible uses as a public space.
Finally, due to its proximity to the Middle and Elementary school, the small pavilion
was gifted to the Vis community as a children’s theatre, enhancing once again the
social impact of the experience on local people, and their attachment to the small
construction.
The Vis Veneer House unveils a strong relationship and balance between the
international nature of the project, and its local definition to be an object for the
community, even in its architectural details (Fig. 20). The exteriors are characterized
by defined Japanese elements, while the interiors celebrate the raw beauty of the
structural wooden system, without any decoration, letting users customize the inner
space according to different functions.
4 Conclusions
Small is More concept collects, at a first glance, a set of real and realized case studies,
small wooden pavilions realized at Politecnico di Milano in several workshops with
students. Going into a deeper analysis, it represents more, though. Small is More
is a research and pedagogic method actualized into a collection of real experiences
on construction processes, which use the most innovative designing approaches but
maintaining the concreteness of the realization issues, at the same time.
These experiences have also enhanced how its feasibility and reliability is strongly
related to the great potential of wood, as a construction material, in terms of structures,
functions and aesthetics, as Vitruvius’s firmitas, utilitas and venustas, and its easily
adaptation and use across the digitalized processes, such as the parametric design
and the CNC-based production.
At last, all the projects have represented the occasion to involve the students
into several formative experiences, enriching both their cultural and professional
knowledge and their personal fascination to act as members of a bigger community,
beyond individuality.
Acknowledgements The authors would like to credit and thank all the professors, researchers,
collaborators that have been involved in these intensive design experiences. A special acknowledge
goes to all the industries and companies that have sponsored the realization of these small archi-
tectures, supporting since several years all the activities of the Building Design and Innovations
workshop, led by Prof. Marco Imperadori at Politecnico di Milano: Knauf, Ri-Legno, Rothoblaas,
Rubner, Vanoncini, Velux, Zintek. For the supply of the construction materials and the technical
assistance, the different design teams would like to thank: Alma, Actis, Cerlogne Pavimenti, Fed-
erlegno Arredo, Ing s.r.l., Internorm. At last, all the project won’t be possible without the intensive
collaboration between the team of Politecnico di Milano, led by Prof. Imperadori and composed
by Prof. Gabriele Masera, Prof. Pietro Tagliabue, Matteo Brasca, Roberto Francieri, and the other
academic institutes, experts, designers, architects, engineers, and pros: Atelier2—Arch Valentina
Gallotti, Kengo Kuma, Salvator-John A. Liotta, the Kengo Kuma Lab team, and Prof. Jun Sato
from University of Tokyo, Prof. Hiroto Kobayashi, Prof. Darko Radovic and the team of the KMD-
W—Kobayashi Maki Design Workshop, Keio University of Tokyo.
In the end, the authors would like to credit the design team for each experience, to thank them for
their passionate and essential contribution to each realization:
For Akragarshelter, Kengo Kuma, Salvator-John A. Liotta, Vito D’anna, Carmelo Bennardo,
Giuseppe Carmelo Parrello, Pietro Giamei, Antonella Colistra, Giorgio Ratti;
For C-ASA, Manuela Grecchi, Giulio Franceschini, Giuliana Iannaccone, Walter Klasz, Salvator-
John Liotta, Gabriele Masera, Barbara Piga, Carlo Poggi, Pierpaolo Avanzi, Francesco Pittau, Elena
Seghezzi, Valentina Gallotti, Paola Trivini, Sergio Brambilla, Fabio Roncoroni, Sergio Colombo,
Manuela Ghielmetti, Andrea Nulli, Francesco Salvetti, Chiara Mariska Chiodero, Daria Pantò Man-
cuso, Cristiana Topo, Tommaso Maserati, Matilde Valagussa, Danica Raznatovic, Stephanie Bart-
Mensah, Chiara Sala, Enrico Pinto, Mohamed Ahmad Hilal Ahmad Kristina Azaric, Elena Sacca-
perni, Mattia Di Carlo, Reine El Kabbani, Sezer Turkmen, Sara Parrinello, Francesca De Ponte,
Sirppika Muthusamy Sellvabarathe, Alessandra Peña, Clara Donati, Luigi Savio, Sahar Dous-
tar, Tamara Akhrameeva, Yehan Wang, Alberto Giacopelli, Veronica Rigonat, Jiali Hu, Simone
Marchetti, Lucrezia Maria De Capitani, Yasmin Baagil Mariyam;
For Equilatera, ESPE, Andrea Bratoli and Florinda Saieva, FARM Cultural Park, Vito Iovino,
Cristina Savoldelli, Luca Scandella;
For Teagloo, Kengo Kuma, Jun Sato, Manuela Grecchi, Matteo Brasca, Kazuya Katagiri, Domenico
Arcadi, Piergiacomo Acerboni, Chiara Mariska Chiodero, Marco Cucuzza, C. Nogara, Daria Pantò
Mancuso, Roberto Pezzutto, Eric Pitalieri, Andrea Redaelli, Benedetta Rota, Danilo Tomasoni,
Cristiana Topo, Sofia Fondelli, Michele Amdori, Martino D’Alberto + Kengo Kuma Lab;
For Veneer House, Gabriele Masera, Darko Radovic, Hiroto Kobayashi, Tadej Glažar, Akira Suzuki,
Don O’Keefe, Kaz Yoneda, Akitaka Suzuki, Ayumu Magome, Masako Sugawara, Hiraki Rihoko,
Hanako Kato Hisamu Yamamori, Enrico Sterle, Mayuko Oizum, i Michika Nakamura, Miyu Oka,
Nicolo’ Panzeri, Paula Jaen, Taichi Kubo, Testuya Yamamoto, Shota Kusano, Xing Tong, Alberto
Tettamanti, Andrea Mazzoleni, Anna Stellini, Anna Casati, Arianna Di Paola, Daisuke Kobayashi,
Hisamu Yamamori, Keitaro Onishi, Marco Fornoni, Martina Scieghi, Mayuko Mikogami, Mei
Morimoto, Omar Ferloni, Simone Conti, Soyoka Tsuji, Sotaro Miyatake, Shota Kusano, Yukie
Takasu.
References
D.M. 06/05/2008—Ministero delle Infrastrutture—Integrazione al decreto 14 gennaio 2008 di

approvazione delle nuove “Norme tecniche per le costruzioni” (Gazzetta ufficiale 02/07/2008
n. 153)
D.M. 14/01/2008—Ministero delle Infrastrutture—Approvazione delle nuove norme tecniche per
le costruzioni (Gazzetta ufficiale 04/02/2008 n. 29)
Circolare n. 617 del 02/04/2011 del Ministero degli Interni—Istruzioni per l’applicazione delle
CNR-DT 206/2007 Istruzioni per la Progettazione, l’Esecuzione ed il Controllo delle Strutture di
Legno
CNR-DT 207/2008 Istruzioni per la valutazione delle azioni e degli effetti del vento sulle costruzioni
Hiroto Kobayashi + VVAA (2017) Veneer House Experience, Keio University Graduate School of
Media and Governance, KMDW. Retrieved from https://www.kmdw.com/veneer-en
Imperadori M, Senatore A (2011) Schematic Design: tecniche ed esempi di comunicazione del
progetto, Collection I libri di Arketipo, Il Sole24Ore, Milano
Kuma K (2018) Ja 109 Kengo Kuma: a lab for materials. Shinkenchiku-sha Co Ltd., Tokyo
Kuma K, Daniell T (2015) Small architecture/natural architecture. Architec Assoc Publ, London
Parvin A (2013) Architecture (and the other 99%): open source architecture and design commons.
Architec Des 83(6):90–95. https://doi.org/10.1002/ad.1680
Salvalai G (2016) Smart campus as urban open lab. Modulo 399:27
Salvalai G (2017) Comparative measurement of the insulation properties of different materials and
facade system coupled with Cross Laminated Timber construction, Colloqui.AT.e 2017 Demoli-
tion or reconstruction, Ancona, pp 652–662
Salvalai G, Imperadori M, Scaccabarozzi D, Pusceddu C (2015) Thermal performance measure-

ment and application of a multilayer insulator for emergency architecture. Appl Thermal Eng
82:110–119
Salvalai G, Imperadori M, Pusceddu C, Blumberg G (2017) The use of building technology to
support disaster resilience: the case study of air shelter house. Int J Disaster Resilience Built Env
8(2):10–18
Sass L (2007) Synthesis of design production with integrated digital fabrication. Autom Constr
16:298–310. https://doi.org/10.1016/j.autcon.2006.06.002
Sheil B (2012) Manufacturing Bespoke Architecture. Nexus Netw J 14(3):441–456. https://doi.org/
10.1007/s00004-012-0126-4
UNI 11035-1 (2010a) Structural timber—Visual strength grading for structural timbers—Part 1:
terminology and measurements of features
UNI 11035-2 (2010b) Structural timber—Visual strength grading for structural timbers—Part 2:
visual strength grading rules and characteristics values for structural timber population
UNI 11035-3 (2010c) Structural timber—visual strength grading—Part 3: Timber Uso Fiume and
Uso Trieste
UNI EN 1995-1-1:2014 Eurocode 5—Design of timber structures—Part 1-1: general—common
UNI EN 1995-1-2:2005 Eurocode 5—design of timber structures—Part 1-2: General—Structural
fire design
UNI EN ISO 6946:2018 Building components and building elements—thermal resistance and
thermal transmittance—calculation methods
UNI ISO 9869-1:2015 Thermal insulation—building elements—In-situ measurement of thermal
resistance and thermal transmittance—Part 1: heat flow meter method
Vanossi A, Imperadori M (2013) BIM and optioneering in dry technology small-scale building. In:
ICT, Automation and the Industry of the Built Environment, Maggioli, Milano, pp 53–65 March
2013
Vanossi A, Parello G, Imperadori M, Bernardo C, Liotta SJA, Ito Y (2014a) BIM for archaeology,
Use of BIM process and parametric model in a temporary shelter adaptable to the diverse needs of
archaeological sites. In: PPC 2014 Conference proceedings. Monza, Mantova (Italy), pp 191–200
May 2014
Vanossi A, Parello G, Imperadori M, Bernardo C, Liotta SJA, Ito Y, Occhipinti F (2014b) Architec-
ture for archeology: identifying new modular and flexible types of shelter adaptable to the diverse
needs of archaeological sites. In: XII Forum Internazionale Le Vie dei Mercanti—proceedings.
La scuola di Pitagora, Anversa-Capri, pp 457–466 June 2015
https://www.activehouse.info/cases/c-asa-polimi-sport-office/
https://www.farmculturalpark.com/
http://kuma-lab.arch.t.u-tokyo.ac.jp/
https://www.kmdw.com/veneer-en

(Lecture Notes in Civil Engineering 24) Fabio Bianconi, Marco Filippucci - Digital Wood Design - Innovative Techniques of Representation in Architectural Design-Springer International Publishing (2019

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

(Lecture Notes in Civil Engineering 24) Fabio Bianconi, Marco Filippucci - Digital Wood Design - Innovative Techniques of Representation in Architectural Design-Springer International Publishing (2019

Hochgeladen von

Copyright:

Verfügbare Formate

Lecture Notes in Civil Engineering

More information about this series at http://www.springer.com/series/15087

Digital Wood Design

Contributions are published under double blind review mode

ISSN 2366-2557 ISSN 2366-2565 (electronic)

Library of Congress Control Number: 2018963030

© Springer Nature Switzerland AG 2019

Part I Integrated Approach and Generative Design

Reinterpretation of Traditional Wood Structures with Digital

Part II Centrality of Representation As a Synthesis Model

Shape and Design of Ancestral Fishing Machines Between

Part III Connection Between Nature Lessons and Material

Bamboo’s Bio-inspired Material Design Through Additive

Part IV Constructive Wisdom and Realization Challenges

Empathic Architecture: Digital Fabrication and Community

Part V Parametric Transﬁgurations and Morphological

Digital Design Thinking in Architectural Education Testing

Grandi invero sono le cose che in questo breve trattato io

Digital Wood Design. Innovative Techniques of Representation in Architectural

and digital manufacturing are decisively contributing to renew a spinneret to look

Hybridization as synthesis. To conclude his digression, Alberti proposes a

The Structure of the Volume

The volume is composed of 61 chapters, written by 153 authors from 5 continents,

Ramsgaard Thomsen, Runberger, Strehlke, Antemann). IAAC and at the Bartlett

(Bufﬁ, Angelini). This section continues analyzing the knowledge-based engineering

Fabio Bianconi and Marco Filippucci

Abstract The contemporary development and digital culture in architecture, from

Keywords Representation · Generative design · Artificial Intelligent

F. Bianconi · M. Filippucci (B)

© Springer Nature Switzerland AG 2019 3

decision-making. The frontiers of the computational design see the organization of

Fig. 3 Wood design: didactic modelling to materialize the space

Defined these coordinates, it is possible to present how AI is changing our research

In this relation, actually contemporary principles of the topological approach of the

3 Methods and Materials

4.1 Responsive Architecture: Study, Characterization

Fig. 6 Modelling as result of observation of phenomena: curvature in function of the humidity in

to the diagonal of the specimen (called “45 configuration”).

4.2 Innovations and Experimentations for Double-Curved

In contemporary architecture, the language used to draw forms is connected to the

4.3 Redrawing Descriptive and Projective Geometry:

Fig. 14 Parametrizing gemetrical forms: three-dimensional development and parallel projections

4.4 Wooden Tensegrity Structures and Morphological

4.5 AI for Mass-Customized Housing. Multi-objective

Building energy performance assessments are complex multi-criteria problems

4.6 The Ames Room Pavilion. The Reinterpretation

Fig. 28 Generative model: variation in the whole family of distorted rooms

Bergmann E, Hildebrand S (2015) Form-finding, form-shaping, designing architecture. Mendrisio

Burkhardt RW (2008) A practical guide to tensegrity design. Cambridge

de Rubertis R (2012) Piramide visiva. Retrieved 28 Aug 2018 from http://www.wikitecnica.com/

Pillet J (1887) Traité de stéréotomie (C. Delagra). Paris

Saint Aubin JP (1994) es enjeux architecturaux de la didactique stéréotomique de Desargues. In:

David Correa, Oliver David Krieg and AnnaLisa Meyboom

Abstract This paper introduces a series of prototypes investigating a new

Keywords Robotic fabrication · Computational design · Material computation

© Springer Nature Switzerland AG 2019 61

Innovation in material results in innovation in architecture. Technology both supports

2 Patterns of Innovation in Materials and Their

By looking at patterns of innovation in materials in the past, it is possible to strengthen

and compression. As the knowledge of iron-related construction was developing,

Compared to steel, concrete maintains the monolithic and massive characteristics of

3 Current Innovation Relating to Wood

3.1 Innovation—Customized Timber Products