You are on page 1of 512

Proceedings e report

90

ECOS 2012
The 25th International Conference on Efficiency, Cost,
Optimization and Simulation of Energy Conversion
Systems and Processes
(Perugia, June 26th-June 29th, 2012)

edited by
Umberto Desideri, Giampaolo Manfrida,
Enrico Sciubba

firenze university press

2012

ECOS 2012 : the 25th International Conference on Efficiency,


Cost, Optimization and Simulation of Energy Conversion
Systems and Processes (Perugia, June 26th-June 29th, 2012) /
edited by Umberto Desideri, Giampaolo Manfrida, Enrico
Sciubba. Firenze : Firenze University Press, 2012.
(Proceedings e report ; 90)
http://digital.casalini.it/9788866553229
ISBN 978-88-6655-322-9 (online)
Progetto grafico di copertina Alberto Pizarro, Pagina Maestra snc
Immagine di copertina: Kts | Dreamstime.com

Peer Review Process


All publications are submitted to an external refereeing process under the responsibility of the FUP
Editorial Board and the Scientific Committees of the individual series. The works published in the
FUP catalogue are evaluated and approved by the Editorial Board of the publishing house. For a
more detailed description of the refereeing process we refer to the official documents published on
the website and in the online catalogue of the FUP (http://www.fupress.com).
Firenze University Press Editorial Board
G. Nigro (Co-ordinator), M.T. Bartoli, M. Boddi, F. Cambi, R. Casalbuoni, C. Ciappei, R. Del Punta,
A. Dolfi, V. Fargion, S. Ferrone, M. Garzaniti, P. Guarnieri, G. Mari, M. Marini, M. Verga, A. Zorzi.
2012 Firenze University Press
Universit degli Studi di Firenze
Firenze University Press
Borgo Albizi, 28, 50122 Firenze, Italy
http://www.fupress.com/
Printed in Italy

ECOS 2012
The 25th International Conference on
Efficiency, Cost, Optimization and Simulation
of Energy Conversion Systems and Processes

Perugia, June 26th-June 29th, 2012


Book of Proceedings - Volume V
Edited by:
Umberto Desideri, Universit degli Studi di Perugia
Giampaolo Manfrida, Universit degli Studi di Firenze
Enrico Sciubba, Universit degli Studi di Roma Sapienza

ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY
EDITED BY UMBERTO DESIDERI, GIAMPAOLO MANFRIDA, ENRICO SCIUBBA
FIRENZE UNIVERSITY PRESS, 2012, ISBN ONLINE : 978-88-6655-322-9

ii

Advisory Committee (Track Organizers)


Building, Urban and Complex Energy Systems
V. Ismet Ugursal
Dalhousie University, Nova Scotia, Canada
Combustion, Chemical Reactors, Carbon Capture and Sequestration
Giuseppe Girardi
ENEA-Casaccia, Italy
Energy Systems: Environmental and Sustainability Issues
Christos A. Frangopoulos
National Technical University of Athens, Greece
Exergy Analysis and Second Law Analysis
Silvio de Oliveira Junior
Polytechnical University of Sao Paulo, Sao Paulo, Brazil
Fluid Dynamics and Power Plant Components
Sotirios Karellas
National Technical University of Athens, Athens, Greece
Fuel Cells
Umberto Desideri
University of Perugia, Perugia, Italy
Heat and Mass Transfer
Francesco Asdrubali, Cinzia Buratti
University of Perugia, Perugia, Italy
Industrial Ecology
Stefan Goessling-Reisemann
University of Bremen, Germany
Poster Session
Enrico Sciubba
University Roma 1 Sapienza, Italy
Process Integration and Heat Exchanger Networks
Francois Marechal
EPFL, Lausanne, Switzerland
Renewable Energy Conversion Systems
David Chiaramonti
University of Firenze, Firenze, Italy
Simulation of Energy Conversion Systems
Marcin Liszka
Polytechnica Slaska, Gliwice, Poland
System Operation, Control, Diagnosis and Prognosis
Vittorio Verda
Politecnico di Torino, Italy
Thermodynamics
A. zer Arnas
United States Military Academy at West Point, U.S.A.
Thermo-Economic Analysis and Optimisation
Andrea Lazzaretto
University of Padova, Padova, Italy
Water Desalination and Use of Water Resources
Corrado Sommariva
ILF Consulting M.E., U.K
iii

Scientific Committee
Riccardo Basosi, University of Siena, Italy
Gino Bella, University of Roma Tor Vergata, Italy
Asfaw Beyene, San Diego State University, United States
Ryszard Bialecki, Silesian Institute of Tecnology, Poland
Gianni Bidini, University of Perugia, Italy
Ana M. Blanco-Marigorta, University of Las Palmas de Gran Canaria, Spain
Olav Bolland, University of Science and Technology (NTNU), Norway
Ren Cornelissen, Cornelissen Consulting, The Netherlands
Franco Cotana, University of Perugia, Italy
Alexandru Dobrovicescu, Polytechnical University of Bucharest, Romania
Gheorghe Dumitrascu, Technical University of Iasi, Romania
Brian Elmegaard, Technical University of Denmark , Denmark
Daniel Favrat, EPFL, Switzerland
Michel Feidt, ENSEM - LEMTA University Henri Poincar, France
Daniele Fiaschi, University of Florence, Italy
Marco Frey, Scuola Superiore S. Anna, Italy
Richard A Gaggioli, Marquette University, USA
Carlo N. Grimaldi, University of Perugia, Italy
Simon Harvey, Chalmers University of Technology, Sweden
Hasan Heperkan, Yildiz Technical University, Turkey
Abel Abel Hernandez-Guerrero, University of Guanajuato, Mexico
Jiri Jaromir Kleme, University of Pannonia, Hungary
Zornitza V. Kirova-Yordanova, University "Prof. Assen Zlatarov", Bulgaria
Noam Lior, University of Pennsylvania, United States
Francesco Martelli, University of Florence, Italy
Aristide Massardo, University of Genova, Italy
Jim McGovern, Dublin Institute of Technology, Ireland
Alberto Mirandola, University of Padova, Italy
Michael J. Moran, The Ohio State University, United States
Tatiana Morosuk, Technical University of Berlin, Germany
Pericles Pilidis, University of Cranfield, United Kingdom
Constantine D. Rakopoulos, National Technical University of Athens, Greece
Predrag Raskovic, University of Nis, Serbia and Montenegro
Mauro Reini, University of Trieste, Italy
Gianfranco Rizzo, University of Salerno, Italy
Marc A. Rosen, University of Ontario, Canada
Luis M. Serra, University of Zaragoza, Spain
Gordana Stefanovic, University of Nis, Serbia and Montenegro
Andrea Toffolo, Lule University of Technology, Sweden
Wojciech Stanek, Silesian University of Technology, Poland
George Tsatsaronis, Technical University Berlin, Germany
Antonio Valero, University of Zaragoza, Spain
Michael R. von Spakovsky, Virginia Tech, USA
Stefano Ubertini, Parthenope University of Naples, Italy
Sergio Ulgiati, Parthenope University of Naples, Italy
Sergio Usn, Universidad de Zaragoza, Spain
Roman Weber, Clausthal University of Technology, Germany
Ryohei Yokoyama, Osaka Prefecture University, Japan
Na Zhang, Institute of Engineering Thermophysics, Chinese Academy of Sciences, China
iv

The 25th ECOS Conference 1987-2012: leaving a mark


The introduction to the ECOS series of Conferences states that ECOS is a series of
international conferences that focus on all aspects of Thermal Sciences, with particular
emphasis on Thermodynamics and its applications in energy conversion systems and
processes. Well, ECOS is much more than that, and its history proves it!
The idea of starting a series of such conferences was put forth at an informal meeting of the
Advanced Energy Systems Division of the American Society of Mechanical Engineers
(ASME) at the November 1985 Winter Annual Meeting (WAM), in Miami Beach, Florida,
then chaired by Richard Gaggioli. The resolution was to organize an annual Symposium on
the Analysis and Design of Thermal Systems at each ASME WAM, and to try to involve a
larger number of scientists and engineers worldwide by organizing conferences outside of the
United States. Besides Rich other participants were Ozer Arnas, Adrian Bejan, Yehia ElSayed, Robert Evans, Francis Huang, Mike Moran, Gordon Reistad, Enrico Sciubba and
George Tsatsaronis.
Ever since 1985, a Symposium of 8-16 sessions has been organized by the Systems Analysis
Technical Committee every year, at the ASME Winter Annual Meeting (now ASME-IMECE).
The first overseas conference took place in Rome, twenty-five years ago (in July 1987), with
the support of the U.S. National Science Foundation and of the Italian National Research
Council. In that occasion, Christos Frangopoulos, Yalcin Gogus, Elias Gyftopoulos, Dominick
Sama, Sergio Stecco, Antonio Valero, and many others, already active at the ASME meetings,
joined the core-group.
The name ECOS was used for the first time in Zaragoza, in 1992: it is an acronym for
Efficiency, Cost, Optimization and Simulation (of energy conversion systems and
processes), keywords that best describe the contents of the presentations and discussions
taking place in these conferences. Some years ago, Christos Frangopoulos inserted in the
official website the note that cos (
) means home in Greek and it ought to be
attributed the very same meaning as the prefix Eco- in environmental sciences.
The last 25 years have witnessed an almost incredible growth of the ECOS community: more
and more Colleagues are actively participating in our meetings, several international Journals
routinely publish selected papers from our Proceedings, fruitful interdisciplinary and
international cooperation projects have blossomed from our meetings. Meetings that have
spanned three continents (Africa and Australia ought to be our next targets, perhaps!) and
influenced in a way or another much of modern Engineering Thermodynamics.
After 25 years, if we do not want to become embalmed in our own success and lose
momentum, it is mandatory to aim our efforts in two directions: first, encourage the
participation of younger academicians to our meetings, and second, stimulate creative and
useful discussions in our sessions. Looking at this years registration roster (250 papers of
which 50 authored or co-authored by junior Authors), the first objective seems to have been
attained, and thus we have just to continue in that direction; the second one involves allowing
space to voices that sing out of the choir, fostering new methods and approaches, and
establishing or reinforcing connections to other scientific communities. It is important that our
technical sessions represent a place of active confrontation, rather than academic lecturing.
In this spirit, we welcome you in Perugia, and wish you a scientifically stimulating,
touristically interesting, and culinarily rewarding experience. In line with our 25 years old
scientific excellency and friendship!
Umberto Desideri, Giampaolo Manfrida, Enrico Sciubba
vi

CONTENT MANAGEMENT
The index lists all the papers contained all the eight volumes of the Proceedings of the
ECOS 2012 International Conference.
Page numbers are listed only for papers within the Volume you are looking at.
The ID code allows to trace back the identification number assigned to the paper within
the Conference submission, review and track organization processes.

vii

CONTENT
VOLUME V
V. 1 RENEWABLE ENERGY CONVERSION SYSTEMS
A co-powered concentrated solar power Rankine cycle concept for ....
small size combined heat and power (ID 276)
Alessandro Corsini, Domenico Borello, Franco Rispoli, Eileen Tortora

Pag. 1

A novel non-tracking solar collector for high temperature application ....


(ID 466)
Wattana Ratismith, Anusorn Inthongkhum

Pag. 17

Absorption heat transformers (AHT) as a way to enhance low enthalpy ....


geothermal resources (ID 311)
Daniele Fiaschi, Duccio Tempesti, Giampaolo Manfrida, Daniele Di Rosa

Pag. 26

Alternative feedstock for the biodiesel and energy production: the ....
OVEST project (ID 98)
Matteo Prussi, David Chiaramonti, Lucia Recchia, Francesco Martelli, Fabio
Guidotti

Pag. 38

Assessing repowering and update scenarios for wind energy ....


converters (ID 158)
Till Zimmermann

Pag. 47

Biogas from mechanical pulping industry potential improvement for ....


increased biomass vehicle fuels (ID 54)
Mimmi Magnusson, Per Alvfors

Pag. 56

Biogas or electricity as vehicle fuels derived from food waste - the


case of Stockholm (ID 27)
Martina Wikstrm, Per Alvfors

....

Pag. 68

of expansion ....

Pag. 78

Compressibility factor as evaluation parameter


processes in organic Rankine cycles (ID 292)
Giovanni Manente, Andrea Lazzaretto

Design of solar heating system for methane generation (ID 445)


Luca Mnica Gutirrez, P. Quinto Diez, L. R. Tovar Glvez

....

Pag. 94

Economic feasibility of PV systems in hotels in Mexico (ID 346)


Augusto Sanchez, Sergio Quezada

....

Pag. 106

Effect of a back surface roughness on annual performance of an air- ....


cooled PV module (ID 193)
Riccardo Secchi, Duccio Tempesti, Jacek Smolka

Pag. 114

Energy and exergy analysis of the first hybrid solar-gas power plant in ....
Algeria (ID 176)
Fouad Khaldi

Pag. 130

Energy recovery from MSW treatment by gasification and melting ....


technology (ID 393)
Fabrizio Strobino, Alessandro Pini Prato, Diego Ventura, Marco Damonte

Pag. 144

Ethanol production by enzymatic hydrolysis process from sugarcane


biomass - the integration with the conventional process (ID 189)
Reynaldo Palacios-Bereche, Adriano Ensinas, Marcelo Modesto, Silvia
Azucena Nebra

Pag. 159

....

-------------------------------------------------------------------------------------------------ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY
EDITED BY UMBERTO DESIDERI, GIAMPAOLO MANFRIDA, ENRICO SCIUBBA
FIRENZE UNIVERSITY PRESS, 2012, ISBN ONLINE : 978-88-6655-322-9

Evaluation of gas in an industrial anaerobic digester by means of ....


biochemical methane potential of organic municipal solid waste
components (ID 57)
Isabella Pecorini, Tommaso Olivieri, Donata Bacchi, Alessandro Paradisi,
Lidia Lombardi, Andrea Corti, Ennio Carnevale
Exergy analysis and genetic algorithms for the optimization of flatplate solar collectors (ID 423)
....
Soteris A. Kalogirou

Pag. 173

Pag. 185

Experimental study of tar and particles content of the produced gas in ....
a double stage downdraft gasifier (ID 487)
Ana Lisbeth Galindo Noguera, Sandra Yamile Giraldo, Rene Lesme-Jan,
Vladimir Melian Cobas, Rubenildo Viera Andrade, Electo Silva Lora

Pag. 197

Feasibility study to realize an anaerobic digester fed with vegetables ....


matrices in central Italy (ID 425)
Umberto Desideri, Francesco Zepparelli, Livia Arcioni, Ornella Calderini,
Francesco Panara, Matteo Todini

Pag. 209

Investigations on the use of biogas for small scale decentralized CHP ....
applications with a focus on stability and emissions (ID 140)
Steven MacLean, Eren Tali, Anne Giese, Jrg Leicher

Pag. 218

Kinetic energy recovery system for sailing yachts (ID 427)


Giuseppe Leo Guizzi, Michele Manno

....

Pag. 229

Mirrors in the sky: status and some supporting materials experiments ....
(ID 184)
Noam Lior

Pag. 253

Numerical parametric study for different cold storage designs and ....
strategies of a solar driven thermoacoustic cooler system (ID 284)
Maxime Perier-Muzet, Pascal Stouffs, Jean-Pierre Bedecarrats, Jean
Castaing-Lasvignottes

Pag. 274

Parabolic trough photovoltaic/thermal collectors. Part I: design and ....


simulation model (ID 102)
Francesco Calise, Laura Vanoli

Pag. 290

Parabolic trough photovoltaic/thermal collectors. Part II: dynamic


simulation of a solar trigeneration system (ID 488)
Francesco Calise, Laura Vanoli

....

Pag. 309

Performance analysis of downdraft gasifier - reciprocating engine


biomass fired small-scale cogeneration system (ID 368)
Jacek Kalina

....

Pag. 331

Proposing offshore photovoltaic (PV) technology to the energy mix of ....


the Maltese islands (ID 262)
Kim Trapani, Dean Lee Millar

Pag. 350

Research of integrated biomass gasification system with a piston ....


engine (ID 414)
Janusz Kotowicz, Aleksander Sobolewski, Tomasz Iluk
Start up of a pre-industrial scale solid state anaerobic digestion cell for ....
the co-treatment of animal and agricultural residues (ID 34)
Francesco Di Maria, Giovanni Gigliotti, Alessio Sordi, Caterina Micale, Luisa
Massaccesi

Pag. 363

The role of biomass in the renewable energy system (ID 390)


Ruben Laleman, Ludovico Balduccio, Johan Albrecht

Pag. 381

ix

....

Pag. 373

Vegetable oils of soybean, sunflower and tung as alternative fuels for ....
compression ignition engines (ID 500)
Ricardo Morel Hartmann, Nury Nieto Garzn, Eduardo Morel Hartmann, Amir
Antonio Martins Oliveira Jr, Edson Bazzo, Bruno Okuda, Joselia Piluski

Pag. 409

Wind energy conversion performance and atmosphere stability ....


(ID 283)
Francesco Castellani, Emanuele Piccioni, Lorenzo Biondi, Marcello Marconi

Pag. 427

V. 2 FUEL CELLS
Comparison study on different SOFC hybrid systems with zero-CO2 ....
emission (ID 196)
Liqiang Duan, Kexin Huang, Xiaoyuan Zhang and Yongping Yang

Pag. 440

Exergy analysis and optimisation of a steam methane pre-reforming ....


system (ID 62)
George G. Dimopoulos, Iason C. Stefanatos, Nikolaos M.P. Kakalis

Pag. 456

Modelling of a CHP SOFC power system fed with biogas from ....
anaerobic digestion of municipal wastes integrated with a solar collector
and storage units (ID 491)
Domenico Borello, Sara Evangelisti, Eileen Tortora

Pag. 472

----------------------------------------------------------------------CONTENTS OF ALL THE VOLUMES


-----------------------------------------------------------------------

VOLUME I
I . 1 - SIMULATION OF ENERGY CONVERSION SYSTEMS
A novel hybrid-fuel compressed air energy storage system for Chinas situation (ID 531)
Wenyi Liu, Yongping Yang, Weide Zhang, Gang Xu,and Ying Wu
A review of Stirling engine technologies applied to micro-cogeneration systems (ID 338)
Ana C Ferreira, Manuel L Nunes, Lus B Martins, Senhorinha F Teixeira
An organic Rankine cycle off-design model for the search of the optimal control strategy
(ID 295)
Andrea Toffolo, Andrea Lazzaretto, Giovanni Manente, Marco Paci
Automated superstructure generation and optimization of distributed energy supply
systems (ID 518)
Philip Voll, Carsten Klaffke, Maike Hennen, Andr Bardow
Characterisation and classification of solid recovered fuels (SRF) and model development
of a novel thermal utilization concept through air- gasification (ID 506)
Panagiotis Vounatsos, Konstantinos Atsonios, Mihalis Agraniotis, Kyriakos D. Panopoulos, George
Koufodimos,Panagiotis Grammelis, Emmanuel Kakaras
Design and modelling of a novel compact power cycle for low temperature heat sources
(ID 177)
Jorrit Wronski, Morten Juel Skovrup, Brian Elmegaard, Harald Nes Risl, Fredrik Haglind
Dynamic simulation of combined cycles operating in transient conditions: an innovative
approach to determine the steam drums life consumption (ID 439)
Stefano Bracco

Effect of auxiliary electrical power consumptions on organic Rankine cycle system with
low-temperature waste heat source (ID 235)
Samer Maalouf, Elias Boulawz Ksayer, Denis Clodic
Energetic and exergetic analysis of waste heat recovery systems in the cement industry
(ID 228)
Sotirios Karellas, Aris Dimitrios Leontaritis, Georgios Panousis, Evangelos Bellos, Emmanuel
Kakaras
Energy and exergy analysis of repowering options for Greek lignite-fired power plants (ID
230)
Sotirios Karellas, Aggelos Doukelis, Grammatiki Zanni, Emmanuel Kakaras
Energy saving by a simple solar collector with reflective panels and boiler (ID 366)
Anna Stoppato, Renzo Tosato
Exergetic analysis of biomass fired double-stage Organic Rankine Cycle (ORC) (ID 37)
Markus Preiinger, Florian Heberle, Dieter Brggemann
Experimental tests and modelization of a domestic-scale organic Rankine cycle (ID 156)
Roberto Bracco, Stefano Clemente, Diego Micheli, Mauro Reini
Model of a small steam engine for renewable domestic CHP system (ID 31 )
Giampaolo Manfrida, Giovanni Ferrara, Alessandro Pescioni
Model of vacuum glass heat pipe solar collectors (ID 312)
Daniele Fiaschi, Giampaolo Manfrida
Modelling and exergy analysis of a plasma furnace for aluminum melting process (ID 254)
Luis Enrique Acevedo, Sergio Usn, Javier Uche, Patxi Rodrguez
Modelling and experimental validation of a solar cooling installation (ID 296)
Guillaume Anies, Pascal Stouffs, Jean Castaing-Lasvignottes
The influence of operating parameters and occupancy rate of thermoelectric modules on
the electricity generation (ID 314)
Camille Favarel, Jean-Pierre Bdcarrats, Tarik Kousksou, Daniel Champier
Thermodynamic and heat transfer analysis of rice straw co-firing in a Brazilian pulverised
coal boiler (ID 236)
Raphael Miyake, Alvaro Restrepo, Fbio Kleveston Edson Bazzo, Marcelo Bzuneck
Thermophotovoltaic generation: A state of the art review (ID 88)
Matteo Bosi, Claudio Ferrari, Francesco Melino, Michele Pinelli, Pier Ruggero Spina, Mauro
Venturini
I . 2 HEAT AND MASS TRANSFER
A DNS method for particle motion to establish boundary conditions in coal gasifiers (ID
49)
Efstathios E Michaelides, Zhigang Feng
Effective thermal conductivity with convection and radiation in packed bed (ID 60)
Yusuke Asakuma
Experimental and CFD study of a single phase cone-shaped helical coiled heat exchanger:
an empirical correlation (ID 375)
Daniel Flrez-Orrego, Walter Arias, Diego Lpez, Hctor Velsquez
Thermofluiddynamic model for control analysis of latent heat thermal storage system (ID
207)
Adriano Sciacovelli, Vittorio Verda, Flavio Gagliardi
Towards the development of an efficient immersed particle heat exchanger: particle
transfer from low to high pressure (ID 202)
Luciano A. Catalano, Riccardo Amirante, Stefano Copertino, Paolo Tamburrano, Fabio De Bellis
xi

I . 3 INDUSTRIAL ECOLOGY
Anthropogenic heat and exergy balance of the atmosphere (ID 122)
Asfaw Beyene, David MacPhee, Ron Zevenhoven
Determination of environmental remediation cost of municipal waste in terms of extended
exergy (ID 63)
Candeniz Seckin, Ahmet R. Bayulken
Development of product category rules for the application of life cycle assessment to
carbon capture and storage (537)
Carlo Strazza, Adriana Del Borghi, Michela Gallo
Electricity production from renewable and non-renewable energy sources: a comparison
of environmental, economic and social sustainability indicators with exergy losses
throughout the supply chain (ID 247)
Lydia Stougie, Hedzer van der Kooi, Rob Stikkelman
Exergy analysis of the industrial symbiosis model in Kalundborg
(ID 218)
Alicia Valero Delgado, Sergio Usn, Jorge Costa
Global gold mining: is technological learning overcoming the declining in ore grades? (ID
277)
Adriana Domnguez, Alicia Valero
Personal transportation energy consumption (ID305)
Matteo Muratori, Emmanuele Serra, Vincenzo Marano, Michael Moran
Resource use evaluation of Turkish transportation sector via the extended exergy
accounting method (ID 43)
Candeniz Seckin, Enrico Sciubba, Ahmet R. Bayulken
The impact of higher energy prices on socio-economic inequalities of German social
groups (ID 80)
Holger Schlr, Wolfgang Fischer, Jrgen-Friedrich Hake

VOLUME II
II . 1 EXERGY ANALYSIS AND 2ND LAW ANALYSIS
A comparative analysis of cryogenic recuperative heat exchangers based on exergy
destruction (ID 129)
Adina Teodora Gheorghian, Alexandru Dobrovicescu, Lavinia Grosu, Bogdan Popescu, Claudia
Ionita
A critical exploration of the usefulness of rational efficiency as a performance parameter
for heat exchangers (ID 307)
Jim McGovern, Georgiana Tirca-Dragomirescu, Michel Feidt, Alexandru Dobrovicescu
A new procedure for the design of LNG processes by combining exergy and pinch
analyses (ID 238)
Danahe Marmolejo-Correa, Truls Gundersen
Advances in the distribution of environmental cost of water bodies through the exergy
concept in the Ebro river (ID 258)
Javier Uche Marcuello, Amaya Martnez Gracia, Beatriz Carrasquer lvarez, Antonio Valero Capilla
Application of the entropy generation minimization method to a solar heat exchanger: a
pseudo-optimization design process based on the analysis of the local entropy generation
maps (ID 357)
Giorgio Giangaspero, Enrico Sciubba
Comparative analysis of ammonia and carbon dioxide two-stage cycles for simultaneous
cooling and heating (ID 84)
Alexandru Dobrovicescu, Ciprian Filipoiu, Emilia Cerna Mladin, Valentin Apostol, Liviu Drughean
xii

Comparison between traditional methodologies and advanced exergy analyses for


evaluating efficiency and externalities of energy systems (ID 515)
Gabriele Cassetti, Emanuela Colombo
Comparison of entropy generation figures using entropy maps and entropy transport
equation for an air cooled gas turbine blade (ID 468)
Omer Emre Orhan, Oguz Uzol
Conventional and advanced exergetic evaluation of a supercritical coal-fired power plant
(ID 377)
Ligang Wang, Yongping Yang, Tatiana Morosuk, George Tsatsaronis
Energy and exergy analyses of the charging process in encapsulted ice thermal energy
storage (ID 164)
David MacPhee, Ibrahim Dincer, Asfaw Beyene
Energy integration and cogeneration in nitrogen fertilizers industry: thermodynamic
estimation of the efficiency, potentials, limitations and environmental impact. Part 1: energy
integration in ammonia production plants (ID 303)
Zornitza Vassileva Kirova-Yordanova
Evaluation of the oil and gas processing at a real production day on a North Sea oil
platform using exergy analysis (ID 260)
Mari Voldsund, Wei He, Audun Rsjorde, Ivar Stle Ertesvg, Signe Kjelstrup
Exergetic and economic analysis of Kalina cycle for low temperature geothermal sources
in Brazil (ID 345)
Carlos Eymel Campos Rodriguez, Jos Carlos Escobar Palacios, Cesar Adolfo Rodrguez
Sotomonte, Marcio Leme, Osvaldo Jos Venturini, Electo Eduardo Silva Lora, Vladimir Melin
Cobasa, Daniel Marques dos Santos, Fbio R. Lofrano Dotto, Vernei Gialluca
Exergy analysis and comparison of CO2 heat pumps (ID 242)
Argyro Papadaki, Athina Stegou - Sagia
Exergy analysis of a CO2 Recovery plant for a brewery (ID 72)
Daniel Rnne Nielsen, Brian Elmegaard, C. Bang-Mller
Exergy analysis of the silicon production process (ID 118)
Marit Takla, Leiv Kolbeinsen, Halvard Tveit, Signe Kjelstrup
Exergy based indicators for cardiopulmonary exercise test evaluation (ID 159)
Carlos Eduardo Keutenedjian Mady, Cyro Albuquerque Neto, Tiago Lazzaretti Fernandes, Arnaldo
Jose Hernandez, Paulo Hilrio Nascimento Saldiva, Jurandir Itizo Yanagihara, Silvio de Oliveira
Junior
Exergy disaggregation as an alternative for system disaggregation in thermoeconomics
(ID 483)
Jos Joaquim Conceio Soares Santos, Atilio Loureno, Julio Mendes da Silva, Joo Donatelli,
Jos Escobar Palacio
Exergy intensity of petroleum derived fuels (ID 117)
Julio Augusto Mendes da Silva, Maurcio Sugiyama, Claudio Rucker, Silvio de Oliveira Junior
Exergy-based sustainability evaluation of a wind power generation system (ID 542)
Jin Yang, B. Chen, Enrico Sciubba
Human body exergy metabolism (ID 160)
Carlos Eduardo Keutenedjian Mady, Silvio de Oliveira Junior
Integrating an ORC into a natural gas expansion plant supplied with a co-generation unit
(ID 273)
Sergio Usn, Wojciech Juliusz Kostowski
One-dimensional model of an optimal ejector and parametric study of ejector efficiency (ID
323)
Ronan Killian McGovern, Kartik Bulusu, Mohammed Antar, John H. Lienhard
xiii

Optimization and design of pin-fin heat sinks based on minimum entropy generation (ID 6)
Jose-Luis Zuniga-Cerroblanco, Abel Hernandez-Guerrero, Carlos A. Rubio-Jimenez, Cuauhtemoc
Rubio-Arana, Sosimo E. Diaz-Mendez
Performance analysis of a district heating system (ID 271)
Andrej Ljubenko, Alojz Poredo, Tatiana Morosuk, George Tsatsaronis
System analysis of exergy losses in an integrated oxy-fuel combustion power plant (ID 64)
Andrzej Zi bik, Pawe G adysz
What is the cost of losing irreversibly the mineral capital on Earth? (ID 220)
Alicia Valero Delgado, Antonio Valero
II . 2 THERMODYNAMICS
A new polygeneration system for methanol and power based on coke oven gas and coal
gas (ID 252)
Hu Lin, Hongguang Jin, Lin Gao, Rumou Li
Argon-Water closed gas cycle (ID 67)
Federico Fionelli, Giovanni Molinari
Binary alkane mixtures as fluids in Rankine cycles (ID 246)
M. Aslam Siddiqi, Burak Atakan
Excess enthalpies of second generation biofuels (ID 308)
Alejandro Moreau, Jos Juan Segovia, M. Carmen Martn, Miguel ngel Villaman, Csar R.
Chamorro, Rosa M. Villaman
Local stability analysis of a Curzon-Ahlborn engine considering the Van der Waals
equation state in the maximum ecological regime (ID 281)
Ricardo Richard Pez-Hernndez, Pedro Portillo-Daz, Delfino Ladino-Luna,
Marco Antonio Barranco-Jimnez
Some remarks on the Carnot's theorem (ID 325)
Julian Gonzalez Ayala, Fernando Angulo-Brown
The Dead State (ID 340)
Richard A. Gaggioli
The magnetocaloric energy conversion (ID 97)
Andrej Kitanovski, Jaka Tusek, Alojz Poredos

VOLUME III
THERMO-ECONOMIC ANALYSIS AND OPTIMIZATION
A comparison of optimal operation of residential energy systems using clustered demand
patterns based on Kullback-Leibler divergence (ID 142)
Akira Yoshida, Yoshiharu Amano, Noboru Murata, Koichi Ito, Takumi Hashizume
A Model for Simulation and Optimal Design of a Solar Heating System with Seasonal
Storage (ID 51)
Gianfranco Rizzo
A thermodynamic and economic comparative analysis of combined gas-steam and gas
turbine air bottoming cycle (ID 232)
Tadeusz Chmielniak, Daniel Czaja, Sebastian Lepszy
Application of an alternative thermoeconomic approach to a two-stage vapor compression
refrigeration cycle with intercooling (ID 135)
Atilio Barbosa Loureno, Jos Joaquim Conceio Soares Santos, Joo Luiz Marcon Donatelli
Comparative performance of advanced power cycles for low temperature heat sources
(ID 109)
Guillaume Becquin, Sebastian Freund
xiv

Comparison of nuclear steam power plant and conventional steam power plant through
energy level and thermoeconomic analysis (ID 251)
S. Khamis Abadi, Mohammad Hasan Khoshgoftar Manesh, M. Baghestani, H. Ghalami, Majid
Amidpour
Economic and exergoeconomic analysis of micro GT and ORC cogeneration systems
(ID 87)
Audrius Bagdanavicius, Robert Sansom, Nick Jenkins, Goran Strbac
Exergoeconomic comparison of wet and dry cooling technologies for the Rankine cycle of
a solar thermal power plant (ID 300)
Philipp Habl, Ana M. Blanco-Marigorta, Berit Erlach
Influence of renewable generators on the thermo-economic multi-level optimization of a
poly-generation smart grid (101)
Massimo Rivarolo, Andrea Greco, Francesca Travi, Aristide F. Massardo
Local stability analysis of a thermoeconomic model of an irreversible heat engine working
at different criteria of performance (ID 289)
Marco A. Barranco-Jimnez, Norma Snchez-Salas, Israel Reyes-Ramrez, Lev Guzmn-Vargas
Multicriteria optimization of a distributed trigeneration system in an industrial area (ID 154)
Dario Buoro, Melchiorre Casisi, Alberto de Nardi, Piero Pinamonti, Mauro Reini
On the effect of eco-indicator selection on the conclusions obtained from an
exergoenvironmental analysis (ID 275)
Tatiana Morosuk, George Tsatsaronis, Christopher Koroneos
Optimisation of supply temperature and mass flow rate for a district heating network
(ID 104)
Marouf Pirouti, Audrius Bagdanavicius, Jianzhong Wu, Janaka Ekanayake
Optimization of energy supply systems in consideration of hierarchical relationship
between design and operation (ID 389)
Ryohei Yokoyama, Shuhei Ose
The fuel impact formula revisited (ID 279)
Cesar Torres, Antonio Valero
The introduction of exergy analysis to the thermo-economic modelling and optimisation of
a marine combined cycle system (ID 61)
George G. Dimopoulos, Chariklia A. Georgopoulou, Nikolaos M.P. Kakalis
The relationship between costs and environmental impacts in power plants: an exergybased study (ID 272)
Fontina Petrakopoulou, Yolanda Lara, Tatiana Morosuk, Alicia Boyano, George Tsatsaronis
Thermo-ecological evaluation of biomass integrated gasification gas turbine based
cogeneration technology (ID 441)
Wojciech Stanek, Lucyna Czarnowska, Jacek Kalina
Thermo-ecological optimization of a heat exchanger through empirical modeling (ID 501)
Ireneusz Szczygie , Wojciech Stanek, Lucyna Czarnowska, Marek Rojczyk
Thermoeconomic analysis and optimization in a combined cycle power plant including a
heat transformer for energy saving (ID 399)
Elizabeth Corts Rodrguez, Jos Luis Castilla Carrillo, Claudia A. Ruiz Mercado, Wilfrido Rivera
Gmez-Franco
Thermoeconomic analysis and optimization of a hybrid solar-electric heating in a fluidized
bed dryer (ID 400)
Elizabeth Corts Rodrguez, Felipe de Jess Ojeda Cmara, Isaac Pilatowsky Figueroa
Thermoeconomic approach for the analysis of low temperature district heating systems
(ID 208)
Vittorio Verda, Albana Kona
xv

Thermo-economic assessment of a micro CHP systems fuelled by geothermal and solar


energy (ID 166)
Duccio Tempesti, Daniele Fiaschi, Filippo Gabuzzini
Thermo-economic evaluation and optimization of the thermo-chemical conversion of
biomass into methanol (ID 194)
Emanuela Peduzzi, Laurence Tock, Guillaume Boissonnet, Franois Marechal
Thermoeconomic fuel impact approach for assessing resources savings in industrial
symbiosis: application to Kalundborg Eco-industrial Park (ID 256)
Sergio Usn, Antonio Valero, Alicia Valero, Jorge Costa
Thermoeconomics of a ground-based CAES plant for peak-load energy production system
(ID 32)
Simon Kemble, Giampaolo Manfrida, Adriano Milazzo, Francesco Buffa

VOLUME IV
IV . 1 - FLUID DYNAMICS AND POWER PLANT COMPONENTS
A control oriented simulation model of a multistage axial compressor (ID 444)
Lorenzo Damiani, Giampaolo Crosa, Angela Trucco
A flexible and simple device for in-cylinder flow measurements: experimental and
numerical validation (ID 181)
Andrea Dai Zotti, Massimo Masi, Marco Antonello
CFD Simulation of Entropy Generation in Pipeline for Steam Transport in Real Industrial
Plant (ID 543)
Goran Vu kovi , Gradimir Ili , Mi a Vuki , Milan Bani , Gordana Stefanovi
Feasibility Study of Turbo expander Installation in City Gate Station (ID 168)
Navid Zehtabiyan Rezaie, Majid Saffar-Avval
GTL and RME combustion analysis in a transparent CI engine by means of IR digital
imaging (ID 460)
Ezio Mancaruso, Luigi Sequino, Bianca Maria Vaglieco
Some aspects concerning fluid flow and turbulence modeling in 4-valve engines (ID 116)
Zoran Stevan Jovanovic, Zoran Masonicic, Miroljub Tomic
IV . 2 - SYSTEM OPERATION CONTROL DIAGNOSIS AND PROGNOSIS
Adapting the operation regimes of trigeneration systems to renewable energy systems
integration (ID 188)
Liviu Ruieneanu, Mihai Paul Mircea
Advanced electromagnetic sensors for sustainable monitoring of industrial processes
(ID 145)
Uro Puc, Andreja Abina, Anton Jegli , Pavel Cevc, Aleksander Zidanek
Assessment of stresses and residual life of plant components in view of life-time
extension of power plants (ID 453)
Anna Stoppato, Alberto Benato and Alberto Mirandola
Control strategy for minimizing the electric power consumption of hybrid ground source
heat pump system (ID 244)
Zoi Sagia, Constantinos Rakopoulos
Exergetic evaluation of heat pump booster configurations in a low temperature district
heating network (ID 148)
Torben Ommen, Brian Elmegaard
xv i

Exergoeconomic diagnosis: a thermo-characterization method by using irreversibility


analysis (ID 523)
Abraham
Olivares-Arriaga,
Alejandro
Zaleta-Aguilar,
Rangel-Hernndez
V.
H,
Juan Manuel Belman-Flores
Optimal structural design of residential cogeneration systems considering their
operational restrictions (ID 224)
Tetsuya Wakui, Ryohei Yokoyama
Performance estimation and optimal operation of a CO2 heat pump water heating system
(ID 344)
Ryohei Yokoyama, Ryosuke Kato, Tetsuya Wakui, Kazuhisa Takemura
Performances of a common-rail Diesel engine fuelled with rapeseed and waste cooking
oils (ID 213)
Alessandro Corsini, Valerio Giovannoni, Stefano Nardecchia, Franco Rispoli, Fabrizio Sciulli,
Paolo Venturini
Reduced energy cost through the furnace pressure control in power plants (ID 367)
Vojislav Filipovi , Novak Nedi , Saa Prodanovi
Short-term scheduling model for a wind-hydro-thermal electricity system (ID 464)
Srgio Pereira, Paula Ferreira, A. Ismael Freitas Vaz

VOLUME VI
VI . 1 - CARBON CAPTURE AND SEQUESTRATION
A novel coal-based polygeneration system cogenerating power, natural gas and liquid fuel
with CO2 capture (ID 96)
Sheng Li, Hongguang Jin, Lin Gao
Analysis and optimization of CO2 capture in a Chinas existing coal-fired power plant
(ID 532)
Gang Xu, Yongping Yang, Shoucheng Li, Wenyi Liu and Ying Wu
Analysys of four-end high temperature membrane air separator in a supercritical power
plant with oxy-type pulverized fuel boiler (ID 442)
Janusz Kotowicz, Sebastian Stanis aw Michalski
Analysis of potential improvements to the lignite-fired oxy-fuel power unit (ID 413)
Marcin Liszka, Jakub Tuka, Grzegorz Nowak, Grzegorz Szapajko
Biogas Upgrading: Global Warming Potential of Conventional and Innovative
Technologies (ID 240)
Katherine Starr, Xavier Gabarrell Durany, Gara Villalba Mendez, Laura Talens Peiro, Lidia
Lombardi
Capture of carbon dioxide using gas hydrate technology (ID 103)
Beatrice Castellani, Mirko Filipponi, Sara Rinaldi, Federico Rossi
Carbon dioxide mineralisation and integration with flue gas desulphurisation applied to a
modern coal-fired power plant (ID 179)
Ron Zevenhoven, Johan Fagerlund, Thomas Bjrklf, Magdalena Mkel, Olav Eklund
Carbon dioxide storage by mineralisation applied to a lime kiln (ID 226)
Ins Sofia Soares Romo, Matias Eriksson, Experience Nduagu, Johan Fagerlund, Licnio Manuel
Gando-Ferreira, Ron Zevenhoven
Comparison of IGCC and CFB cogeneration plants equipped with CO2 removal (ID 380)
Marcin Liszka, Tomasz Malik, Micha Budnik, Andrzej Zi bik
Concept of a capture ready combined heat and power plant (ID 231)
Piotr Henryk Lukowicz, Lukasz Bartela
xv ii

Cryogenic method for H2 and CH4 recovery from a rich CO2 stream in pre-combustion
CCS schemes (ID 508)
Konstantinos Atsonios, Kyriakos D. Panopoulos, Angelos Doukelis, Antonis Koumanakos,
Emmanuel Kakaras
Design and optimization of ITM oxy-combustion power plant (ID 495)
Surekha Gunasekaran, Nicholas David Mancini, Alexander Mitsos
Implementation of a CCS technology: the ZECOMIX experimental platform (ID 222)
Antonio Calabr, Stefano Cassani, Leandro Pagliari, Stefano Stendardo
Influence of regeneration condition on cyclic CO2 capture using pre-treated dispersed
CaO as high temperature sorbent (ID 221)
Stefano Stendardo, Antonio Calabr
Investigation of an innovative process for biogas up-grading pilot plant preliminary
results (ID 56)
Lidia Lombardi, Renato Baciocchi, Ennio Antonio Carnevale, Andrea Corti, Giulia Costa, Tommaso
Olivieri, Alessandro Paradisi, Daniela Zingaretti
Method of increasing the efficiency of a supercritical lignite-fired oxy-type fluidized bed
boiler and high-temperature three - end membrane for air separation (ID 438)
Janusz Kotowicz, Adrian Balicki
Monitoring of carbon dioxide uptake in accelerated carbonation processes applied to air
pollution control residues (ID 539)
Felice Alfieri, Peter J Gunning, Michela Gallo, Adriana Del Borghi, Colin D Hills
Process efficiency and optimization of precipitated calcium carbonate (PCC) production
from steel converter slag (ID 114)
Hannu-Petteri Mattila, Inga Grigali nait , Arshe Said, Sami Filppula, Carl-Johan Fogelholm, Ron
Zevenhoven
Production of Mg(OH)2 for CO2 Emissions Removal Applications: Parametric and Process
Evaluation (ID 245)
Experience Ikechukwu Nduagu, Ins Romo, Ron Zevenhoven
Thermodynamic analysis of a supercritical power plant with oxy type pulverized fuel
boiler, carbon dioxide capture system (CC) and four-end high temperature membrane air
seprator (ID 411)
Janusz Kotowicz, Sebastian Stanis aw Michalski
VI . 2 - PROCESS INTEGRATION AND HEAT EXCHANGER NETWORKS
A multi-objective optimization technique for co- processing in the cement production (ID
42)
Maria Luiza Grillo Ren, Rogrio Jos da Silva, Mirian de Lourdes Noronha Motta Melo, Jos
Joaquim Conceio Soares Santos
Comparison of options for debottlenecking the recovery boiler at kraft pulp mills
Economic performance and CO2 emissions (ID 449)
Johanna Jnsson, Karin Pettersson, Simon Harvey, Thore Berntsson
Demonstrating an integral approach for industrial energy saving (ID 541)
Ren Cornelissen, Geert van Rens, Jos Sentjens, Henk Akse, Ton Backx, Arjan van der Weiden,
Jo Vandenbroucke
Maximising the use of renewables with variable availability (ID 494)
Andreja Nemet, Jiri Jaromr Kleme, Petar Sabev Varbanov, Zdravko Kravanja
Methodology for the improvement of large district heating networks (ID 46)
Anna Volkova, Vladislav Mashatin, Aleksander Hlebnikov, Andres Siirde
Optimal mine site energy supply (ID 306)
Monica Carvalho, Dean Lee Millar
xv iii

Simulation of synthesis gas production from steam oxygen gasification of Colombian


bituminous coal using Aspen Plus (ID 395)
John Jairo Ortiz, Juan Camilo Gonzlez, Jorge Enrique Preciado, Roco Sierra, Gerardo Gordillo

VOLUME VII
VII . 1 - BUILDING, URBAN AND COMPLEX ENERGY SYSTEMS
A linear programming model for the optimal assessment of sustainable energy action
plans (ID 398)
Gianfranco Rizzo, Giancarlo Savino
A natural gas fuelled 10 kW electric power unit based on a Diesel automotive internal
combustion engine and suitable for cogeneration (ID 477)
Pietro Capaldi
Adjustment of envelopes characteristics to climatic conditions for saving heating and
cooling energy in buildings (ID 430)
Christos Tzivanidis, Kimon Antonopoulos, Foteini Gioti
An exergy based method for the optimal integration of a building and its heating plant.
Part 1: comparison of domestic heating systems based on renewable sources (ID 81)
Marta Cianfrini, Enrico Sciubba, Claudia Toro
Analysis of different typologies of natural insulation materials with economic and
performances evaluation of the same buildings (ID 28)
Umberto Desideri, Daniela Leonardi, Livia Arcioni
Complex networks approach to the Italian photovoltaic energy distribution system (ID 470)
Luca Valori, Giovanni Luca Giannuzzi, Tiziano Squartini, Diego Garlaschelli, Riccardo Basosi
Design of a multi-purpose building "to zero energy consumption" according to European
Directive 2010/31/CE: Architectural and plant solutions (ID 29)
Umberto Desideri, Livia Arcioni, Daniela Leonardi, Luca Cesaretti ,Perla Perugini, Elena Agabitini,
Nicola Evangelisti
Effect of initial systems on the renewal planning of energy supply systems for a hospital
(ID 107)
Shu Yoshida, Koichi Ito, Yoshiharu Amano, Shintaro Ishikawa, Takahiro Sushi, Takumi Hashizume
Effects of insulation and phase change materials (PCM) combinations on the energy
consumption for buildings indoor thermal comfort (ID 387)
Christos Tzivanidis, Kimon Antonopoulos, Eleutherios Kravvaritis
Energetic evaluation of a smart controlled greenhouse for tomato cultivation (ID 150)
Nickey Van den Bulck, Mathias Coomans, Lieve Wittemans, Kris Goen, Jochen Hanssens, Kathy
Steppe, Herman Marien, Johan Desmedt
Energy networks in sustainable cities: temperature and energy consumption monitoring in
urban area (ID 190)
Luca Giaccone, Alessandra Guerrisi, Paolo Lazzeroni and Michele Tartaglia
Extended exergy analysis of the economy of Nova Scotia, Canada
David C Bligh, V.Ismet Ugursal

(ID 215)

Feasibility study and design of a low-energy residential unit in Sagarmatha Park (Nepal)
for envirnomental impact reduction of high altitude buildings (ID 223)
Umberto Desideri, Stefania Proietti, Paolo Sdringola, Elisa Vuillermoz
Fire and smoke spread in low-income housing in Mexico (ID 379)
Raul R. Flores-Rodriguez, Abel Hernandez-Guerrero, Cuauhtemoc Rubio-Arana, Consuelo A.
Caldera-Briseo

xix

Optimal lighting control strategies in supermarkets for energy efficiency applications via
digital dimmable technology (ID 136)
Salvador Acha, Nilay Shah, Jon Ashford, David Penfold
Optimising the arrangement of finance towards large scale refurbishment of housing stock
using mathematical programming and optimisationg (ID 127)
Mark Gerard Jennings, Nilay Shah, David Fisk
Optimization of thermal insulation to save energy in buildings (ID 174)
Milorad Boji , Marko Mileti , Vesna Marjanovi , Danijela Nikoli , Jasmina Skerli
Residential solar-based seasonal thermal storage system in cold climate: building
envelope and thermal storage (ID 342)
Alexandre Hugo and Radu Zmeureanu
Simultaneous production of domestic hot water and space cooling with a heat pump in a
Swedish Passive House (ID 55)
Johannes Persson, Mats Westermark
SOFC micro-CHP integration in residential buildings (ID 201)
Umberto Desideri, Giovanni Cinti, Gabriele Discepoli, Elena Sisani, Daniele Penchini
The effect of shading of building integrated photovoltaics on roof surface temperature and
heat transfer in buildings (ID 83)
Eftychios Vardoulakis, Dimitrios Karamanis
The influence of glazing systems on energy performance and thermal comfort in nonresidential buildings (ID 206)
Cinzia Buratti, Elisa Moretti, Elisa Belloni
Thermal analysis of a greenhouse heated by solar energy and seasonal thermal energy
storage in soil (ID 405)
Yong Li, Jin Xu, Ru-Zhu Wang
Thermodynamic analysis of a combined cooling, heating and power system under part
load condition (ID 476)
Qiang Chen, Jianjiao Zheng, Wei Han, Jun Sui, Hong-guang Jin
VII . 2 - COMBUSTION, CHEMICAL REACTORS
Baffle as a cost-effective design improvement for volatile combustion rate increase in
biomass boilers of simple construction (ID 233)
Borivoj Stepanov, Ivan Peenjanski, Biljana Miljkovi
Characterization of CH4-H2-air mixtures in the high-pressure DHARMA reactor (ID 287)
Vincenzo Moccia, Jacopo D'Alessio
Development of a concept for efficiency improvement and decreased NOx production for
natural gas-fired glass melting furnaces by switching to a propane exhaust gas fired
process (ID 146)
Jrn Benthin, Anne Giese
Experimental analysis of inhibition phenomenon management for Solid Anaerobic
Digestion Batch process (ID 348)
Francesco Di Maria, Giovanni Gigliotti, Alessio Sordi, Caterina Micale, Claudia Zadra, Luisa
Massaccesi
Experimental investigations of the combustion process of n-butanol/diesel blend in an
optical high swirl CI engine (ID 85)
Simona Silvia Merola, G. Valentino, C. Tornatore, L. Marchitto , F. E. Corcione
Flameless oxidation as a means to reduce NOx emissions in glass melting furnaces
(ID 141)
Jrg Leicher, Anne Giese
xx

Mechanism of damage by high temperature of the tubes, exposed to the atmosphere


characteristic of a furnace of pyrolysis of ethane for ethylene production in the
petrochemical industry (ID 65)
Jaqueline Saavedra Rueda, Francisco Javier Perez Trujillo, Lourdes Isabel Merio Stand, Harbey
Alexi Escobar, Luis Eduardo Navas, Juan Carlos Amezquita
Steam reforming of methane over Pt/Rh based wire mesh catalyst in single channel
reformer for small scale syngas production (ID 317)
Haftor Orn Sigurdsson, Sren Knudsen Kr

VOLUME VIII
VIII . 1 - ENERGY SYSTEMS : ENVIRONMENTAL AND SUSTAINABILITY ISSUES
A multi-criteria decision analysis tool to support electricity planning (ID 467)
Fernando Ribeiro, Paula Ferreira, Madalena Arajo
Comparison of sophisticated life cycle impact assessment methods for assessing
environmental impacts in a LCA study of electricity production (ID 259)
Jens Buchgeister
Defossilisation assessment of biodiesel life cycle production using the ExROI indicator
(ID 304)
Emilio Font de Mora, Csar Torres, Antonio Valero, David Zambrana
Design strategy of geothermal plants for water dominant medium-low temperature
reservoirs based on sustainability issues (ID 99)
Alessandro Franco, Maurizio Vaccaro
Energetic and environmental benefits from waste management: experimental analysis of
the sustainable landfill (ID 33)
Francesco Di Maria, Alessandro Canovai, Federico Valentini, Alessio Sordi, Caterina Micale
Environmental assessment of energy recovery technologies for the treatment and
disposal of municipal solid waste using life cycle assessment (LCA): a case study of Brazil
(ID 512)
Marcio Montagnana Vicente Leme, Mateus Henrique Rocha, Electo Eduardo Silva Lora,Osvaldo
Jos Venturini, Bruno Marciano Lopes, Claudio Homero Ferreira
How will renewable power generation be affected by climate change? The case of a
metropolitan region in Northwest Germany (ID 503)
Jakob Wachsmuth, Andrew Blohm, Stefan Gling-Reisemann, Tobias Eickemeier, Rebecca
Gasper, Matthias Ruth, Snke Sthrmann
Impact of nuclear power plant on Thailand power development plan (ID 474)
Raksanai Nidhiritdhikrai, Bundhit Eua-arporn
Improving sustainability of maritime transport through utilization of liquefied natural gas
(LNG) for propulsion (ID 496)
Fabio Burel, Rodolfo Taccani, Nicola Zuliani
Life cycle assessment of thin film non conventional photovoltaics: the case of dye
sensitized solar cells (ID 471)
Maria Laura Parisi, Adalgisa Sinicropi, Riccardo Basosi
Low CO2 emission hybrid solar CC power system (ID 175)
Yuanyuan Li, Na Zhang, Ruixian Cai
Low exergy solutions as a contribution to climate adapted and resilient power supply
(ID 489)
Stefan Goessling-Reisemann, Thomas Bloethe
On the use of MPT to derive optimal RES electricity generation mixes (ID 459)
Paula Ferreira, Jorge Cunha
xxi

Stability and limit cycles in an exergy-based model of population dynamics (ID 128)
Enrico Sciubba, Federico Zullo
The influence of primary measures for reducing NOx emissions on energy steam boiler
efficiency (ID 125)
Goran Stupar, Dragan Tucakovi , Titoslav ivanovi , Milo Banjac, Sr an Beloevi ,Vladimir
Beljanski, Ivan Tomanovi , Nenad Crnomarkovi , Miroslav Sijer
The Lethe city car of the University of Roma 1: final proposed configuration (ID 45)
Roberto Capata, Enrico Sciubba
VIII . 2 - POSTER SESSION
A variational optimization of a finite-time thermal cycle with a Stefan-Boltzmann heat
transfer law (ID 333)
Juan C.Chimal-Eguia, Norma Sanchez-Salas
Modeling and simulation of a boiler unit for steam power plants (ID 545)
Luca Moliterno, Claudia Toro
Numerical Modelling of straw combustion in a moving bed combustor (ID 412)
Biljana Miljkovi, Ivan Peenjanski, Borivoj Stepanov, Vladimir Milosavljevi, Vladimir Rajs
Physicochemical evaluation of the properties of the coke formed at radiation area of light
hydrocarbons pyrolysis furnace in petrochemical industry (ID 10)
Jaqueline Saavedra Rueda , Anglica Mara Carreo Parra, Mara del Rosario Prez Trejos,
Dionisio Laverde Catao, Diego Bonilla Duarte, Jorge Leonardo Rodrguez Jimnez, Laura Mara
Daz Burgos
Rotor TG cooled (ID 121)
Chiara Durastante, Paolo Petroni, Michela Spagnoli, Vincenzo Rizzica, Jrg Helge Wirfs
Study of the phase change in binary alloy (ID 534)
Aroussia Jaouahdou, Mohamed J. Safi, Herve Muhr
Technip initiatives in renewable energies and sustainable technologies (ID 527)
Pierfrancesco Palazzo, Corrado Pigna

xxii

ECOS 2012

VOLUME V

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCEON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

A co-powered Concentrated Solar Power Rankine


cycle concept for small size Combined Heat and
Power
Alessandro Corsinia , Domenico Borellob , Franco Rispoli c and Eileen Tortorad
a

Facolt di Ingegneria, Sapienza Universit di Roma, Latina, Italy, alessandro.corsini@uniroma1.it


b
Dipartimento di Meccanica e Aeronautica, Sapienza Universit di Roma, Roma, Italy,
domenico.borello@uniroma1.it
c
Dipartimento di Meccanica e Aeronautica, Sapienza Universit di Roma, Roma, Italy,
franco.rispoli@uniroma1.it
d
Dipartimento di Meccanica e Aeronautica, Sapienza Universit di Roma, Roma, Italy,
eileen.tortora@uniroma1.it (CA)

Abstract:
The present work investigates the matching of an advanced small scale combined heat and power Rankine
cycle plant to end-user thermal and electric load. The power plant consists of a conc entrated solar power
field co-powered by a biomass furnac e to produce steam in a Rankine cycle, with a Combined Heat and
Power configuration. A hotel was selected as the end user. The power plant design and its operation were
modelled and investigated by adopting transient simulations wit h a hourly distribution. The study of the load
matching of the proposed renewable power technology and the final user has been carried out by comparing
two different load tracking scenarios, i.e. the thermal and t he electric demands. As a result, the power output
follows fairly well the given load curves, supplying, on a selected winter day, about 50 GJ/d of thermal
energy and the 6 GJ/d of electric energy, with reduced energy dumps when matching the load. Furthermore,
for the same winter day, the system allows the reduction of about 4 103 kgCO2 of greenhous e gas
emissions.

Keywords:
Co-powered Concentrated Solar Power, Rankine Cycle, Transient Simulation, Load Matching.

1. Introduction
In recent years the use of Combined Heat and Power (CHP) was commonly considered to supply
energy to end users in the service or residential sectors. The basic argument in favour of CHP is the
possibility to obtain electric and thermal energy in situ, improving the power generation efficiency
and reducing the losses usually related to the energy distribution [1, 2]. Notably among the existing
CHP technologies, only some exceptions are based on the exploitation of different fuels from
natural gas, i.e. small-scale power plants based on biomass derived fuel exploitation, like wood or
biogas [3, 4].
In most applications the main factor which determines the economic viability of CHP schemes is
the high utilisation of heat and electric energy, which are produced simultaneously. Most of the
literature indicates that a CHP plant needs to be fully utilised providing heat and power for a
minimum duty of 4,500 h per annum to gain its breakeven point [5].
When designing renewable energy based CHP technologies, in a distributed generation concept, one
of the key factors is the capability of tracking the time-dependent end-user load. Renewable Energy
Sources (RES), intermittent by nature, produce inconsistently and somewhat unpredictably power
outputs uncorrelated with the end user power demands, typically variable according to predictable

daily load profiles. As a consequence of this mismatch the available RES energy may not meet the
energy demand, resulting in deficit and surplus energy situations.
Several solutions have been proposed to attenuate the RES- user matching inconsistency. The
conventional remedial strategy is to plug the supply gap providing alternative capacity, known as
spinning reserve [6]. Among the solutions devoted to RES electric grid integration, it is worth
mentioning the use of high capacity energy storage to save the produced energy surplus and
postponing the energy surplus delivery [7,8], or combining renewable energy sources with
complementary intermittencies [9].
In this respect, the present study investigates a CHP scheme combining a parabolic trough field for
concentrated solar power (CSP), a thermal energy storage and a biomass furnace as complementary
source. It is worth noting that the biomass source is a sui generis RES, in fact its storage simplicity
permits to customize the power production management, exactly like the fossil fuel sources.
Concerning the parabolic trough field, that device was selected for its high worldwide development
among the CSP systems [10]. Nonetheless, an important aspect of these plants is the size, which, is
usually large. In fact solar trough plants are characterised by multi-MW sizes, which range up to
about 50 MWel for parabolic trough systems. Also the biomass power plants are usually rated in the
range 5100 MW. Even so, while CSP plants size is still growing [11, 12], in the biomass field
there are several applications on small-scale biomass power plants [13, 14].
The aim to exploit CSP technology and limit the plant footprint led to the design of a small scale
plant, recently presented in [15, 16], composed by a 2,580 m2 parabolic trough field, a thermal
energy storage system (TES) and a 1,163 kW biomass furnace to face the solar source fluctuations.
A heat transfer fluid (HTF), i.e. diathermic oil, is heated by the parabolic through field and biomass
furnace and subsequently it is sent to a heat recovery steam generator where it produces low
enthalpy saturated steam that is sent to a 130 kW reciprocating steam engine for the electric energy
production. Moreover, the Rankine cycle (RC) economizer is fed by the exhaust gases derived from
the biomass combustion. A heat recovery for thermal energy production is obtained, using hot water
as heat carrier, in a back-pressure scheme at 134 C and 300 kPa.
The investigations on the proposed RES-based small- scale CHP Rankine cycle plant, when matched
to a typical hotelier end-user were carried out by transient model simulations. The selection of a
hotel as end-user was made for its high heat/electricity consumption ratio. The system matching
behaviour is analyzed for both thermal and electric load tracking with the aim to demonstrate its
capability to meet the end-users energy request on a 24 hour period in a winter day as more
challenging for the solar field performance.
The transient model and the simulations were performed in the TRNSYS environment [17]
supported by the in- house made types of the biomass furnace and reciprocating steam engine and
the STEC component model library [18]. The software TRNSYS was selected as it is a well-known
instrument to model complex energy systems, as demonstrated by several studies appeared in the
open literature which mostly deal on RES applications in a few fields like small- islands stand alone
power systems [8, 19], or, more related to the present paper, on CSP field simulations [20], TES
behaviour in solar trough plants [21] and matching to a hotel end-user [16].

2. Co-powered solar-biomass plant and model description


2.1. Component and system description
The proposed CHP concept, Figure 1, concerns of a solar-biomass Rankine cycle system. The basic
equipment of the power block consists of 1,200 kW solar trough field, 360 kW thermal energy
storage (TES) and 1,163 kW biomass furnace to feed the heat transfer fluid (HTF) loop and the
related RC. Although biofuel can be easily stored and is promptly available, the TES allows
avoiding the dump of surplus CSP energy occurring during the mismatch with respect to the load.
It is worth noting that the biomass furnace is constantly on duty at a minimum power that is the
35% of its maximum power (i.e. 407 kWth ), in order to ease its complementary source role avoiding
2

power output deficits and/or furnace start-up problems related to the Direct Normal Insulation
(DNI) sudden variations.

Figure 1.Power plant diagram.


Table 1. Main components description and nominal size.
Component description
Solar parabolic trough field (2,580 m2 )
TES
Biomass furnace
Reciprocating steam engine
Condenser
Diathermic oil circuit
Maximum/minimum temperature
Maximum/minimum specific heat
Operating pressure
Water/Steam circuit
Maximum/minimum pressure
Maximum/minimum temperature
Water/steam mass flow rate
Electric power
Thermal power

kWth
kWth
kWth
kWel
kWth

Size
1,200
360
1,163
130
1,240

C
kJ/kg K
kPa

300/240
2.36/ 2.19
800

kPa
C
kg/s
kW
kW

2,800/300
230/134
0.51
130
1,100

The HTF circuit supplies the thermal energy to the RC for the production of saturated steam to be
expanded in a 130 kW reciprocating steam engine fitted with an electric generator. According to a
bottomer CHP configuration, the expanded steam is condensed producing a thermal power output
3

available at a constant temperature of 80 C, i.e. the temperature demand of typical district heating
networks. Figure 2 illustrates the temperature-entropy diagram of the Rankine cycle and the
thermodynamic parameters in the reference points.The main components and system
thermodynamic parameters, subdivided in diathermic oil and water/steam circuit, are described in
Table 1. Additional details concerning the power system components are given in [15].
The temperature-heat diagram is shown in Figure 3. The exhaust gas, diathermic oil and watersteam temperatures with the Rankine cycle exchanged heat rate are shown. In particular, two lines
are plotted for the exhaust gas respectively showing the temperature evolution at CSP design
operation with the biomass furnace working at 35% duty rate (Gas-35%), and at 100% of the
biomass heat contribution (Gas-100%). In between these two limiting lines the solar contribution to
the Rankine cycle spans from maximum (Gas-35%) to zero (Gas-100%). Notably, the pinch point
temperature difference for the evaporator is set to 10 C.

406,73

407,15

503,29

503,29

406,73

3,00

28,00

28,00

28,00

[kg/m ]

931,78

932,75

827,10

13,99

1,83

u [kJ/kg]

561,27

562,06

987,39

2604,02

2345,94

h [kJ/kg]

561,60

565,06

990,78

2804,11

2509,69

1,67

1,67

2,61

6,21

6,46

0,0

0,0

0,0

1,0

0,9

T [K]
P [bar]

s [kJ/kg K]
quality

Figure 2. Temperature-Entropy diagram of power cycle.

Figure 3. Temperature-Heat diagram.

2.2. Transient model description


In order to evaluate the time-dependent behaviour and the performance of the proposed system a
transient model was developed in the TRNSYS framework [17] integrated with the STEC library
[18]. The RC transient model also includes in- house made types for the biomass furnace and for the
reciprocating steam engine [15]. The model subsets and their linkages are described by the flow
4

diagram in Figure 4. The present solar-biomass CHP plant is broadly based on a configuration
recently investigated and assessed [15].
The base-line model has been implemented by a control logic targeted to the tracking of different
loads, namely heat or power demands. The development of the load tracking strategy has been
based on the definition of algebraic correlations between the HTF flow rate, directly related to the
RES power input, and the system thermal power output (P th ) or the system electric output (Pel),
respectively. The HTF flow rate was selected as the reference parameter because it governs the
actual power outputs according to the instantaneous renewable energy availability. A sensitivity
analysis, was carried out on the power system configuration by varying
and recording Pel and
Pth values. Figure 5shows the values obtained with the sensitivity analysis (grey lines) and the
corresponding trend lines (black lines) and equations. The HTF control equations, accordingly
derived, read as
,
.

Figure4. Energy conversion system flow diagram.


The control logic was implemented, Figure 4, in order to match the requested HTF flow rate target
(
) at each time-step with the actual power demand according to the adopted load tracking law.
Hence, the HTF flow rate target tracks the load evolution following a two-level control strategy,
respectively driving the solar section and the whole system. In particular, the solar section control
verifies the state of charge of the TES, giving priority to the storage charging in case of emptiness
(
). The flow rate not needed to charge the TES can be can be directly supplied to the
Rankine cycle. The second control acquires the load data (
) and compares the HTF flow rate
target with the actual HTF flow rate achievable from the available solar field and the minimum
biomass furnace rate (
) at each time step, giving rise to three possible situations:
1. direct CSP contribution surplus, the exceeding HTF flow rate will be dumped;
2. direct CSP contribution deficit, the missing heat flux will be first requested to the TES (flow
rate
); and
3. in case of insufficient flux from the solar section and minimum biomass contributions, an
additional heat flux is requested to the biomass furnace (flow rate
).

Figure5. Thermal a) and electric b) output control equations.

3. End user description


3.1. End-user load profile
The behaviour of the proposed RES-based small-scale CHP Rankine cycle plant is investigated in
the matching of load curve of a typical hotelier end-user during a 24 hour time period. The hotel
was chosen, among tertiary sector end-users, for its high annual heat/electricity consumption ratio.
The end- user characteristics are summarized, in Table 2. The energy data gives a heat/electric
consumption ratio higher than five, Table 2 [22], which is typical of European hotelier end-user
figure, in contrast to the standard North-American hotel energy profile [23]. Furthermore, in order
to take into account the cooling load also, it is worthy referring to the equivalent thermal load
(obtained by the addition of the actual thermal load and the thermal load resulting if fulfilling the
cooling load with a absorption chiller) with a 0,7 COP. In this case the heat/electric rises to a value
of 7.44. The cooling load takes place only in the months from June to September, with a constant
distribution of about 600 GJ/month.
Table 2. End users characteristics [22].
Volume [m3 ]
Number of sleeping accommodations
Heat load [GJ/y]
Electric load [GJ/y]
Cooling load [GJ/y]
Equivalent thermal load
Heat/electric consumption ratio [GJth /GJel]
Equivalent heat/electric consumption ratio [GJth /GJel]

Hotel
43,000
350
8,640
1,656
2,580
12 326
5.23
7.44

Figure 6 shows the monthly distribution of the electric and equivalent thermal load for the selected
end-user; the average daily energy demand (dot sign) is represented in relationship with the daily
average power demand (x sign) and the power demand excursion (bar). It is evident that the electric
energy request has an almost constant behaviour with average daily energy demand always below
200 GJ/day. Whereas the thermal monthly profile has a seasonal connotation which entails a
thermal load range from 250 GJ on the summer period to 1,370 GJ on the winter one. It is worth
noting that generally the average power demand is positioned on the lower part of the power
demand excursion bars, indicating that the energy demand is composed by frequent low power
demand values and rare high power values. This behaviour is highlighted in the summer equivalent
6

thermal load curves (from June to September) of both the end users, when high peaks of cooling
energy are requested during the day.

Legend:
Left axis

Average daily energy demand

Right axis X Average daily power demand


Variation of daily power request on monthly basis

Figure 6. Hotel monthly electric and thermal load yearly behaviour.

3.2. RES data input


The RES input data are available on a hourly distribution over a year period. The direct normal
insulation data [24], are referred to Romes latitude, i.e. 4154'39"24 N, as indicative of a central
Italian location DNI data show a maximum value in the month of July, with 733.68 MJ/m2 and a
minimum value of 253.04 MJ/m2 in December, with an annual cumulative irradiation of 5,760
MJ/m2 . The DNI hourly distribution data on the selected winter day are provided in Table 3.
Table 3. Direct normal insulation data for the selected winter day [24].
Hour
DNI [W/m2 ]
Hour
1
0
13
2
0
14
3
0
15
4
0
16
5
0
17
6
0
18
7
0
19
8
18.33
20
9
560.83
21
10
848.06
22
11
918.06
23
12
938.06
24

DNI [W/m2 ]
938.06
918.06
848.06
560.83
18.33
0
0
0
0
0
0
0

As far as the biomass is concerned, the thermo-chemical characteristics are typical of short rotation
forestry derived woody pellet, with a lower heating value of about 17 MJ/kg and high carbon and
oxygen ratios.

4. Solar-biomass power plant performance


The analysis of solar-biomass plant is based on the comparison of transient and overall performance
under two power modulation scenarios. Namely, i. the tracking of the end-user thermal load in the
hypothesis of electric energy surplus sale to the grid, and ii. the tracking of the end-user electric
load with a dump of the thermal energy surplus.
In the following, the overall CHP plant performances are first discussed on a yearly and monthly
basis and then the time-dependent results on a winter day are shown and discussed. In particular, the
study focuses on a typical winter day in order to discuss the behaviour of the system in operating
conditions which are not favourable to the solar sub-system. The thermal and electric load curves
are shown in Figure 7. The thermal load ranges from 300 to 640 kW, with a sharp min- max
modulation. On the other hand, the electric load, always below 100 kW, achieves its peak level in
the morning and then it decreases during the day being nearly constant in the afternoon and evening
times.
100

800

Pel,d (kW)

Pth,d (kW)
90

700

80
600
70
500

60
50

400

40

300

30
200
20
100

10
0

12

16

20

0
24

Figure 7. End user electric and thermal load for a typical winter day [22].

4.1. Overall performance


In order to compare the performance of the solar-biomass CHP system under the two proposed
load-tracking logics, a number of indicators have been considered (Table 4). In particular the
indices concern the RES system performance, the output performance and the RC efficiency. The
surplus and deficit index for the output performance were calculated by adding the surplus or deficit
thermal and electric energy production which occurred hour per hour with respect to the
corresponding load energy request. The overall performances have been computed over a year
period.
The integration over the duty time showed that the parabolic trough field collect s 4,277.53 GJ/y of
solar energy. Furthermore, as the energy input need varies in the two scenarios in reason of the
different loads, the effective solar energy supply, which is a balance between the available solar
energy and the TES charge discharge rates, differs in the two cases with an amount of about 4,172
GJ/y in the electric tracking scenario and 4,093 GJ/y in the thermal tracking one. The biomass
energy supply varies for the same reason, leading to an effective solar supply fraction, calculated as
the percentage of the effective solar energy with respect to the sum of the effective solar energy and
the biomass furnace energy, of 18.71% in the electric tracking case and 19.20% in the thermal
tracking one. It is worth noting that the selected sizing of the solar collector field is made in
accordance to the Italian existing feed in tariff minimum size of 2,500 m2 for the concentrated solar
power.
8

Looking at the RC system performance Table 4, the value of 1.2 for the primary energy ratio
demonstrates that the presented solar-biomass Rankine cycle systems can effectively allow the
saving of conventional primary energy sources in each presented scenario. Looking at the electric
output, globally the system produces more electric energy than the need with a peak
production/request ratio of 124% for the electric tracking.

RC system

Thermal output Electric output

RES system

Table 4. Overall performance data.


Solar energy [GJ/y]
Effective solar energy supply [GJ/y]
Biomass energy [GJ/y]
Solar fraction
Biomass consumption [ton/y]
Global effective energy input Eg [GJ/y]
Plant electric energy output Eel [GJ/y]
Eel,d [GJ/y]
Eel/ Eel,d [%]
Surplus [%]
Deficit [%]
Plant thermal energy supply Eth [GJ/y]
Eth,d [GJ/y]
Eth / Eth,d [%]
Surplus [%]
Deficit [%]
Net electric efficiency = Eel/Eg [%]
Net thermal efficiency = Eth /Eg [%]
Electric index = Eel/Eth [-]
Primary energy ratio = (Eel el+Eth / th )/Eg [-]1

Electric Tracking
4,277.53
4,172.09
18,132.39
18.71
990.84
22,304.48
2,064.37
1,664.68
124.01
19.80
0.44
17,291.93
11,656.13
148.35
40.09
7.49
9.26
77.53
11.94

Thermal Tracking
4,277.53
4,092.72
17,221.31
19.20
941.06
21,314.03
2,017.10
1,664.46
121.19
25.80
8.32
16,895.49
11,653.99
144.98
33.27
2.26
9.46
79.27
11.94

1.21

1.24

4.2. Hourly power system performance


The global data in a RES based plant are not indicative of the effective load covering. As a matter
of fact, analyzing the hourly behaviour of the systems, there are both surplus and deficit situations.
It is worth noting that the hotel electric tracking scenario offers a completely absence of thermal
supply deficits, but shows a 132% of thermal energy surplus. Considering that the electric source is
easier to manage than the thermal one, as it can be sold or bought from the grid, the most suitable
configuration appears to be the thermal tracking one.
Figure 8 shows the surplus (values higher than zero) and deficits (values lower than zero) behaviour
of the electric and thermal power supply for both the electric and thermal tracking scenario. The
graphs, presented on a monthly basis, are based on hourly data, and show, on the left axis, the
minimum and maximum difference registered in the month between the load and the supplied
power. On the right axis the cumulative surplus and deficit energy is shown for each month. The
electric output of the electric tracking configuration, Figure 8 a), shows the smaller values variation.
Nevertheless, as this good result corresponds to the electric behaviour on the electric tracking

For the primary energy ratio evaluation, the values for the reference electric and thermal efficiencies are

el

= 0.38 and

th

= 0.8.

configuration, the thermal behaviour is worst, with a high rate of surplus distributed all over the
reference year and a deficit peak during the summer period, as the electric energy request is not
sufficiently high to let the system to produce the requested thermal energy too. The deficit and
surplus events have a quadruple explanation. The first one is that half of the results are loadindependent, e.g. when discussing the electric tracking configuration, the thermal output does not
follow any production law, but is dependent from the electric production trend, without any
correlation to the thermal load. Secondly, in most of the occasions the gaps with the requested load
are entailed to the used correlation among load energy and hot thermal fluid flowrate, which do not
perfectly fit the sensitivity analysis data, conducing to gaps between the desired output and the
obtained one. Nevertheless, those gaps are not particularly remarkable. The third reason, instead,
explains the high surplus peaks that occur, by observing that sometimes there are contemporarily an
elevated available solar supply and full thermal energy storage. In those cases the system, which has
to deliver the collected heat, sends all the hot flowrate directly to the Rankine cycle. The last reason
is that the biomass furnace is always on duty, even if on a minimum rate, supplying energy also in
extremely low energy request.

Figure 8. Hotel electric and thermal power surplus/deficit behaviour during a one year period
under electric and thermal load tracking conditions.

4.3. Matching through the load tracking


The thermal and electric load tracking are analysed by comparing hourly distribution of the
different power components. Figure 9 shows the thermal power inputs to the RC, respectively from
the solar field (PCSP ) and the biomass furnace (Pb ), the TES contribution during the charge/discharge
cycles (PT ES,c, PT ES,d), and the thermal power recovered from the exhaust gas (Peg).
As evident, the CSP power is available only between 9 a.m. and 4 p.m., with two peaks,
respectively ante- and post- meridian, of about 400 kW. It is worth noting that the PCSP reduction at
12 a.m. is caused by the reflection losses due to multiple reflections occurring for high solar
incidence angles [25].
In the thermal load tracking (Figure 9.a) the PCSP is not sufficient to meet the thermal load (Pth,d)
which rapidly rises to its peak value about 600 kW. For this reason the control system driven by the
thermal demand, activates the TES system to store fractions of the solar energy (PT ES,c) available in
10

the peak hours and to buffer it (PT ES,d) in the day time when the sun DNI falls below 3,000 kJ/h m2 .
The passage to the electric load tracking logic (Figure 9.b) appears to influence remarkably the RES
power inputs/outputs and the TES charge/discharge cycle. In particular, the TES charge cycle is no
more driven by solar radiation a.m. and p.m. peaks and it is shifted in the afternoon hours when the
overall electric power request reduces. This circumstance causes the shifting of the TES discharge
cycle to the evening time and unbalances the power input from the biomass furnace which is mainly
concentrated in the early morning hours. This finding confirms that the TES and the biomass
furnace have complementary behaviours by implementing an effective reserve to the solar source.
PCSP
PTES,c

800

kW
700

PCSP
PTES,c

800

Pb
Peg

kW
700

PTES,d

600

Pb
Peg

PTES,d

600

500

500

400

400

300

300
200

200
100

100
0

a)

12

16

20

24

b)

hours

12

16

20

24
hours

Figure 9. RES power contribution with the a) thermaland b) electric load tracking matching.
The matching of the power plant with the end-user demand, as driven respectively by the thermal
and electric profile, is described in Figure 10 and Figure 11, by plotting the thermal power output
(Pth ) against the thermal power request (Pth,d) (Figure 10.a and Figure 11.a), and the electric power
output (Pel) against the electric demand (Pel,d) (Figure 10.b and Figure 11.b).
Pth,d
Pth

8 00

kW

200

kW
180

7 00

Pel,d
Pel

160

6 00

140

5 00
120

4 00

100

3 00

80
60

2 00

40

1 00
20

a)

12

16

20

24

4
8
12
16
20
24
b)
hours
Figure 10. Thermal a) and electric b) behaviour with the thermal load tracking matching.
hours

In the thermal load tracking case, Figure 10.a, the thermal load (Pth,d) is completely satisfied by the
solar-biomass plant output (Pth ). The exceeding heat production during the periods of minimum
request is consequent to the control regime of the biomass furnace which is kept at a constant
minimum level. When looking at the electric matching, Figure 10.b, it is remarkable that the power
plant electric output (Pel) mimics the shape of the leading load component. As a result, the correct
11

sizing of the solar-biomass CHP system provides a fair matching in the period of peak electric
request, while the load tracking logic drives the system to an over-production of electricity during
the remaining duty time.
Looking at the electric load tracking case, as a matter of fact, the thermodynamic characteristics of
the solar-biomass CHP system determine the significant overproduction of the thermal power
output when the overall control is given to the electricity production. Figure 11.a shows the electric
peak request in the early morning which, giving rise to the intervention of the biomass, in absence
of any direct or stored solar contribution, results in a large surplus of heat availability. Moving to
the electric matching, Figure 11.b, it is shown that the delivered electric power (Pel) follows fairly
the load (Pel,d) between 4 a.m. and 12 p.m. while keeping it nearly constant in the remaining hours.
800

Pth,d

kW

200

Pel,d
Pel

kW

Pth

180

700

160

600
140

500

120
100

400

80

300
60

200

40
20

100
0

a)

12

16

20

12

16

20

24

24

b)
hours
hours
Figure 11. Thermal a) and electric b) behaviour with the electric load tracking matching.

5. Environomic issues
here the environmental and economic aspects will be analyzed. An effect of the application of this
system are the entailed Greenhouse Gases (GHG) emission savings related to the solar fraction,
estimated by means of emission factors related to the Italian thermoelectric power stations at
reference year 2003 [22]. The emissions savings are evaluated considering the entire electric energy
supply, in the hypothesis of grid transfer of the surplus, and the fraction of thermal energy supplied
to the end users, in the hypothesis of dump of the thermal energy surplus
The result is a higher emission saving in the thermal tracking scenario, which avoids the emissions
of 661 ton/y of carbon dioxide.
Table 5. Global emission savings for a typical winter day from solar fraction.
Electric tracking
Thermal tracking
CO2 [ton/y]
555.29
661.00
SOx [ton/y]
0.58
0.69
NOx [ton/y]
0.35
0.41
TSP [ton/y]
0.02
0.03
Another essential environmental aspect is the land use of the plant. Considering the net land use,
Table 6, the plant needs about 13,000 m2 , nevertheless, taking into account security distances and
the need of space for the power conversion block the needed surface amounts to 31,000 m2 . Even if
12

the solar field accounts for the 51.56% of the global footprintit is meaningful to reflect on the higher
specific power of the parabolic trough field, i.e. 0.46 kWp /m2 , when compared to the most
commercial photovoltaic power plants (0.17 kWp /m2 ) referring only to the devices surface area.
Concerning the plant costs, Table 7 indicates that the parabolic trough field with the thermal energy
storage are the most expensive devices of the proposed system. In particular, the capital cost of a
solar trough field with thermal storage has been evaluated in 4,820 $/kW for the reference year
2006 [10]. It is worth noting that these data refer to large CSP technologies and must be considered
only as a rough estimate of the present CSP device. Referring to the other technologies, the capital
costs have been obtained by private communications with producers. In the utilities heading, it
entails costs for electric panels, electric and hydraulic connections, civil works &c.
It is obvious that such high costs are constraining to the development of the proposed system when
thinking to the standard fossil fuel based power technologies. Nevertheless, in a fossil fuel free
power generation perspective, given from the exhaustion of fossil energy sources and from the need
to pull down the fossil sources related emissions, the current high costs become a side issue in
behalf of the sustainable development of the energy sector.
Table 6. Plant land use.
Solar field
TES
Biomass furnace, filter and stack
Biomass storage
Buildings (Rankine cycle elements, desalting units, offices)
Total
Table 7. Plant estimated capital costs.
Technology
CSP field with TES
Biomass furnace
Economizer
Evaporator
Steam engine
Condenser
Utilities
Total

Net land use [m2 ]


6,780
570
700
3,000
2,100
13,150

Cost []
7,870,000
130,000
15,000
45,000
220,000
15,000
300,000
8,595,000

6. Conclusions
A model of a combined solar-biomass CHP plant devoted to feed an hotelier end- user was
presented. The well-established TRNSYS software was adopted for transient simulation.
An analysis of thermal and electrical power production on a yearly basis demonstrated the
feasibility of the present configuration in satisfying the energy requirements of the hotel using a
fully renewable and sustainable approach.
Furthermore the model was matched with a thermal and an electric winter day load in transient
simulations. The results for the two different load tracking scenarios were compared in terms of
delivered power, matched load, RC system efficiencies and global GHG emission savings.
13

When looking at the output performance, the results show a most suitable behaviour for the thermal
load tracking scenario, as it delivers both electric and thermal energy with less gap from the enduser requested energy.
The Primary Energy Ratio values, in both electric and thermal tracking cases, indicate the capability
of the system to save energy in comparison of two separated plant for the single electric and thermal
energy production. Nevertheless, as the high plant capital costs, i.e. 7.2 k$/kW, are mostly related to
the solar section, the improvements must be oriented to the exploitation of low-tech solar field
entailing the passage from parabolic troughs to Compound Parabolic Concentrators (CPC) and the
adoption of Direct Steam Generators (DSG) systems with supercritical steam Rankine cycles.

Nomenclature
CHP
CSP
DNI
Eel
Eg
Eth
HTF

Combined Heat and Power


Concentrated Solar Power
Direct Normal Irradiation
Electric energy output
Global energy input from biomass and solar radiation
Thermal energy output
Heat Transfer Fluid
HTF flow rate
Solar field HTF delivered flow rate
Minimum biomass furnace HTF delivered flow rate
Additional biomass furnace HTF delivered flow rate
HTF demanded flow rate
Solar direct and TES delivered flow rate
TES HTF charge flow rate

Pb
Pb,min
PCSP
Peg
Pel
Pel,d
P
PT ES,d
Pth
Pth,d
RC
RES
TES
el
th

TES HTF discharge flow rate


Biomass derived power
Biomass furnace power at minimum duty
CSP derived thermal power
Exhaust gas power
Electric power output
Electric load power
Storage charge power
Storage discharge power
Thermal power output
Thermal load power
Rankine Cycle
Renewable Energy Source
Thermal Energy Storage
Reference electric efficiency
Reference thermal efficiency

14

References
[1] Maidment, G.G., Zhao, X., Riffat, S.B. and Prosser G., Application of combined heat-andpower and absorption cooling in supermarkets, Applied Energy 63 (1999) 169-190.
[2] Maidment, G.G. and Tozer, R.M., Combined cooling heat and power in supermarkets,
Applied Thermal Engineering 22 (2002) 653-665.
[3] Corsini, V. Naso, G. Mattei, P. Venturini. Biomass co-firing: analysis of the main technica l
problems in coal power plants. 15th European Biomass Conference, 7-11 May, Berlin,
Germany.
[4] Corsini, V. Naso, G. Mattei, P. Venturini. Biomass co-firing: estimation of fuel requirements
and land needed to feed some Italian coal power plants. 15th European Biomass Conference,
7-11 May, Berlin, Germany.
[5] Eastop T.D. and Croft D.R., Energy efficiency for engineers and technologists (1st ed.),
Longman Scientific and Technical, Halow, Essex, 1990, p. 335.
[6] Frhlke, K. and Haidn, O.J., Spinning reserve system based on H2 /O2 combustion. Energy
Conversion, S0196-8904(96)00128-8.
[7] Klouwani, S., Agbossou, K. and Chahine R., Model for energy conversion in renewable
energy system with hydrogen storage. Journal of Power Sources 140 (2005) 392-399.
[8] Corsini, A., Rispoli, F., Gamberale, M., and Tortora, E., 2009, Assessment of H2- and H2Obased renewable energy-buffering systems in minor islands, Renewable Energy 34 (2009)
279288.
[9] Sinden, G. Environmental Change Institute University of Oxford, The practicalities o f
developing renewable energy stand-by capacity and intermittency. Submission to The Science
and Technology Select Committee of the House of Lords, 2004.
[10] Sargent & Lundy LLC Consulting Group Chicago, Illinois, Assessment of Parabolic
Trough and Power Tower Solar Technology Cost and Performance Forecasts, NREL/SR-55034440, October 2003.
[11] Jun Li, Scaling up concentrating solar thermal technology in China, Renewable and
Sustainable Energy Reviews 13 (2009) 2051-2060.
[12] Al-Soud, M.S. and Hyayshat E.S., A 50 MW concentrating solar power plant for Jordan,
Journal of Cleaner production 17 (2008) 625-635.
[13] Dong, L., Liu, H. and Riffat, S., Development of small-scale and micro-scale biomass
fuelled CHP systems. A literature review, Applied Thermal Engineering 29(2009) 2119-2126.
[14] Badami, M. and Mura, M., Preliminary design and controlling strategy of small-scale wood
waste Rankine Cycle (RC) with a reciprocating steam engine (SE), Energy 34 (2009) 13151324.
[15] Borello, D., Corsini, A., Rispoli, F. and Tortora E., A combined solar-biomass Rankine
cycle concept for small-size cogeneration, Proceedings of ECOS 2009 Conference.
[16] Borello, D., Corsini, A., Rispoli, F. and Tortora E., Load matching for a combined solarbiomass Rankine cycle plant Proceedings of ASME-ATI-UIT 2010 Conference.
[17] Klein, S.A., Beckam, W.A., Mitchell, J.W., Braun, J.E., Evan,s B.L., Kummert J.P., et al.,
2000, TRNSYS a transient system simulation program. Version 15.1, Madison: Solar
Energy Laboratory, University of Wisconsin; 2000.
[18] Schwarzbzl, P., Eiden, U., Pitz-Paal, R., Jones, S., 2002, A TRNSYS model library for
solar thermal electric components (STEC). A reference manual. Release 2.2.
[19] Corsini, A., Gamberale, M., and Rispoli F., 2006, Assessment of renewable energy
solutions in an Italian small island energy system using a transient simulation model, ASME
Journal of Solar Energy Engineering 2006;128:23744.
15

[20] Jones, S.A., Pitz-Paal, R., Schwarzboezl, P., Blair, N. and Cable R., 2001, TRNSYS
modeling of the SEGS VI parabolic trough solar electric generating system, Proceedings o f
Solar Forum 2001: Solar Energy: The Power to Choose, April 21-25, 2001, Washington, DC.
[21] Kolb, G.J. and Hassani, V., 2006, Performance of thermocline energy storage proposed for
the 1 MW Saguaro solar trough plant, ISEC2006-99005.
[22] Macchi, E., Campanari, S., Silva, P., La micro-cogenerazione a gas naturale, Polipress Politecnico di Milano, 2005.
[23] National Action Plan for Energy Efficiency Sector Collaborative on Energy Efficiency,
Hotel Energy Use Profile, EPA Summer Workshop Report, 2007.
[24] SEL, 25 July 2003, Generated Hourly Weather Data. Solar Energy Laboratory, University of
Wisconsin-Madison. 12 November 2008 <http://sel.me.wisc.edu/trnsys/weather/generate.htm>.
[25] Ronnelid, M., Perers, B., Karlsson, B., On the factorization of incidence angle modifiers
for CPC collectors, S0038-092X(97)00016-9, Solar Energy, 1997.

16

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

A Novel Non-Tracking Solar Collector for High


Temperature Application
Wattana Ratismith and Anusorn Inthongkhum
Energy Research Institute, Chulalongkorn University,10330, Bangkok, Thailand,
wattana@eri.chula.ac.th and anusornaun@hotmail.com

Abstract:
A parabolic trough solar c ollector is improved the efficiency by a novel design of compound parabolic trough
solar collector where the aim is three-fold. Firstly, one aim is to achieve day-long collection efficiency without
the need for mechanical tracking of the sun. Secondly, the collector must be designed to operate efficiently
under diffuse solar irradiation as experienced for example in rainforest climate. Thirdly, one seeks to achieve
as a high an output temperature as possible. Newly developed system consists of multiple compound
parabolic troughs facing the sun at different angles. The s alient feature of this design is that the system can
collect the sunlight energy at every angle without any moving parts at the same time can receive the diffused
light, the maximum efficiency of the collector is 32% and has an ability to achieve high output temperature,
the maximum temperature at header of evacuated tube is 235 degrees Celsius, and is therefore suitable for
high temperature application such as industrial uses or cooling application.

Keywords:
solar energy, compound parabolic trough, non-tracking solar collector.

1. Introduction
A parabolic trough is a type of solar thermal energy collector which is generally used in solar power
plants. The solar collector is constructed as a long parabolic trough with a tube running its length at
the focal point. Sunlight is reflected by the trough and concentrated on the tube filled with synthetic
oil, which heats to 300-400 degrees Celsius [1-5]. The trough is usually aligned on a north-south
axis, and rotated to track the sun as it moves across the sky each day. Therefore it seems
unavoidable that there needs to be a tracking system that follows the position of the sun.
The disadvantage of the parabolic trough solar collector is that concentrating systems require sun
tracking to maintain sunlight focus at the collector. The tracking system increases the cost,
complexity and the maintenance cost due to the moving parts. This type of solar collector is not
preferred in a small residential house. Another problem is an inability to provide power in diffused
light conditions, which is due to the fact that the power output from concentrating systems drops in
cloudy conditions. As Thailand has a tropical rainforest climate, which causes the ratio of diffused
solar radiation to global solar radiation to be rather high (in the range of 31% to 58%) [8], one faces
a serious problem in utilizing such a solar collector to collect solar energy, especially in rainforest
climate.
A parabolic trough solar collector is improved the efficiency by a novel design of compound
parabolic trough solar collector which does not contain a solar tracking system and has an ability to
collect diffused sunlight by using compound parabolic troughs facing the sun at different angles [67]. The non-tracking parabolic trough solar collectors were presented in ref. [8-20]. The advantage
of this design is that there are no moving parts in the system, which leads to reductions in the cost
and maintenance. This collector yields higher temperatures than flat plate solar collector and could
be used in the residential house, the maximum temperature at header of evacuated tube is 235
degrees Celsius, and is therefore suitable for high temperature application such as industrial uses or
cooling application.
17

2. The Model
In order to design and develop the non-tracking solar collector, the mathematical model of
reflection of compound trough is calculated. Let the shape of a parabolic trough be described by the
curve y = f(x) on the x-y plane in Fig. 1. The law of reflection states that the angle of incidence is
equal to the angle of reflection relative to the tangent of the curve y = f(x) at any point (x,y). The
slope of this tangent line at point (x,y) is denote by mt = df(x)/dx, the slope of the incident ray by m0
and the slope of the reflected ray by m1.

Fig. 1. The reflection of a light ray by a curve y = f(x). is represented an angle of incidence and
an angle of reflection. m t, m0 and m1 are slope of a tangent line, an incident ray and a reflected ray
respectively.
From trigonometry [5], the relationship between the angle between two lines and their relative
slopes mt, m0 and m1 is given as
mt m0
m1 mt
tan
,
(1)
1 mt m0 1 mt m1
which yields a slope of the first reflected ray m1 as
m1

m0 mt 2 mt m0
.
mt2 2mt m0 1

Similarly, the ith reflected rays can be calculated by using the relation
mt mi 1
mi mt
,
1 mt mi 1 1 mt mi

(2)

(3)

where i are integers. From Eq. 1 and Eq. 2, the reflection of a parabolic trough can be simulated as
shown in Fig. 2.

Fig. 2. The reflection of parabolic trough solar collector at incident angle of 75 degrees where
blue and orange lines are incident and 1st reflected rays respectively. The circle is the position of
the focus point.

18

For the incident angle of 75 degrees, the conventional parabolic trough in Fig. 2 cannot receive the
reflected rays. Therefore it needs solar tracking system to maintain sunlight at the focus point. The
parabolic trough solar collector is designed to have an ability to achieve day- long collection
efficiency without the need for mechanical tracking of the sun by using 3 compound parabolic
troughs facing the sun at different angles. Using Eq.(1-3), the reflection of non-tracking solar
collector at various time are shown in Fig. 3.

12.00 am

1.00 p m

2.00 p m

3.00 p m

Fig. 3. The reflection of three-compound parabolic trough solar collector where blue, orange,
green and yellow lines are incident, 1st reflected, 2nd reflected and 3rd reflected rays respectively.
The circle in each trough is the position of evacuated tube.
The 3-compound parabolic trough shows that it has an ability to receive the sunlight at various time.
For 12.00 a.m., the solar collector can collect all reflected rays, the reflected rays in the middle
trough are concentrated at the lowest position of the tube and for both side of the middle trough, the
reflected rays are concentrated on the higher position inside the tube. When the time changes, the
reflected rays move up and down inside a tube. For this principle, this collector can collect the
sunlight in any time. However there are some ray losses when the time changes especially after 3.00
p.m. which could be ignored because of very low solar power.
The collector is designed to have an ability to collect diffused light. In Fig. 4, compound parabolic
trough can receive the incident rays in the period of 80 degrees. This implies that this collector has a
probability to collect incident rays from sunlight in both direct and diffused light in the period of 80
degrees at the same time while a conventional parabolic trough can collect the incident rays which
are nearly perpendicular to the trough. Although a parabolic trough could provide a high
concentration, the parabolic trough could not work effectively under diffused light conditions. The
experimental results have shown that the efficiency of the new design of solar collector is higher
than parabolic trough under diffuse solar irradiation as shown in Fig.10 and Fig. 11.
19

80

10

Fig. 4. The reflection of light rays at various angles of the incident rays. This design has an ability
to collect incident rays in the period of 80 degrees while the conventional parabolic trough can
receive the incident rays in the period of 10 degrees.
In this paper, SUNDA vacuum tubes (SEIDO1) are used to receive the concentrated light
from the trough. This tube is composed of flat plate absorber as shown in Fig. 5.

Fig. 5. The method to place an evacuated tube with flat plate absorber in compound parabolic
trough.
From Fig. 5, the flat plate absorber which is placed horizontally can receive reflected rays better
than the flat plate absorber which is placed vertically and cross shape absorber can collect all rays
but there are no cross shape absorber product at the moment. For this reason, flat plate absorber is
considered to place horizontally in each trough.
20

3. Experiment
The solar collector in Fig. 3 has been invented consisting of three compound parabolic troughs
made of stainless sheets, oriented at different angles. The solar collector has an overall width of 1 m
and a length of 1.9 m, and the evacuated tubes (SUNDA vacuum tube, (SEIDO1)) are placed along
its axis. These evacuated tubes are connected to a manifold header pipe and connected with the
pump to feed the oil. The flow rate is set at 5 lpm. The collectors are fixed on Earth and aligned
along the north-south direction as shown in fig (6-7).

Fig. 6. The novel non-tracking solar collector has an overall width of 1 m and a length of 1.9 m.

HE
AD
ER

W
S

N
E
Fig. 7. Diagram of test arrangement.

The experiment was performed in Bangkok, Thailand. The data was taken during the period of 9.00
a.m. to 4.00 p.m. on the 10th , 11th , 12th ,13th and 14th January 2012, The sky was not very clear
which lead the solar power is not smooth in any time. The diagram of test arrangement is shown in
fig. 7.
When the evacuated tubes absorb the sunlight from troughs, the heat from the tubes is transferred to
hot oil which flows in the system. The energy of the system can be calculated by [21]
QC t

mC Tout Tin
21

(4)

where t represents time, m and C are flow rate and the specific heat of the thermal oil respectively.
The efficiency of the system in any time is
t

QC
,
Qin

(5)

where Qin is the solar power. The evacuated tube is placed in the trough and measured the
temperature at the header. The maximum temperature at heat pipe is 235 degrees Celsius as shown
in Fig. 8 and the maximum temperature of hot oil is 180 degrees Celsius for 0.5 litres of oil as
shown in Fig. 9.

Fig. 8. The maximum temperature at the header of evacuated tube plotted against time from 8.00
a.m. to 5.00 p.m. on the 12th December 2011.

Fig. 9. The hot oil temperature plotted against time from 9.00 a.m. to 4.00 p.m. on the 14th
November 2011. The maximum temperature is 180 degrees Celsius for 0.5 litres of oil.

From the experiment, the solar power on the 11th ,12th ,13th and 14th of January 2012 in Bangkok
had been collected and its average is shown in Fig. 10. The results show that the efficiency of the
new-design solar collector at any time is fairly constant, which is similar to the parabolic trough
with solar tracking system, while the efficiency of a conventional parabolic trough at any time
distributes like a Gaussian curve having its maximum at around 11.30 a.m. as shown in Fig.11. The
three-compound parabolic trough solar collector yields higher temperature than flat plate or
evacuated tube solar collector. The average efficiency of solar collector is 25-32% .

22

50

800

40

600

30

Efficiency

Solar Power W m

1000

400

10

200

20

0
10 :00

12 :00

14 :00

16:00

10:00

12:00

14:00

16:00

Time

Time

Fig. 10. The average solar power and efficiency of 3-compound parabolic trough plotted against
times in the period of 9.00 a.m. to 4.00 p.m. on the 10th , 11th ,12th ,13th and 14th January 2012 in
Bangkok.

Fig. 11. The parabolic trough in Fig. 4 has been invented. The average efficiency of parabolic
trough plotted against time from 9.00 a.m. to 2.00 p.m. on the 4th, 6th and 8th January 2010[9]

4.Conclusions
The new-design of solar collector has an ability to collect the sunlight at every angle, similar to the
parabolic trough with a solar tracking system. This solar collector has an ability to receive the
diffused light, and this make it suitable for using in all kinds of climate. There are no moving parts
in the system, which results in the reductions in the cost of the system, the cost of maintenance and
complexity. This collector needs only 3 evacuated tubes while SUNDA collector (SEIDO1) needs 8
tubes at the same area. This collector yields higher temperatures than flat plate or evacuated tube
solar collector. The maximum temperature at heat pipe is 235 C and oil temperature is 180 C. It is,
therefore, suitable for high temperature application such as industrial uses or cooling application.

5.Acknowledgements

23

The authors would like to thank the National Research University Project of CHE, the
Ratchaphiseksomphot Endoment Fund (Project No. EN1180I), 2-V research program of National
Research Council of Thailand (NRCT) and Energy Research Institute of Chulalongkorn University
for the financial supports. We also would like to thank Mr. Narong Amornpitakpunt, AMP
METALWORKS [Thailand] Co.,Ltd for his help for inventing the 1st and 2nd prototype of solar
collector.

References
[1] Kearney, D., 1989. Solar Electric Generating Stations (SEGS). IEEE Power Engineering
Review (IEEE) 9, 4-8.
[2] Mills D., 2004. Advances in Solar Thermal Electricity Technology. Solar Energy 76, 19-31.
[3] Hodge B. K., 2010. Alternative Energy Systems and Application. John Wiley & Sons, Inc.
[4] Frank Kreith, D. Yogi Goswami, 2007. Handbook of Energy efficiency and Renewable
Energy, CRC Press.
[5] Milton Matos Rolim, Naum Fraidenraich and Chigueru Tiba, 2009. Analytic Modeling of Solar
Power Plant with Parabolic Linear Collectors. Solar Energy 83, 126-133.
[6] Wattana Ratismith, Parabolic Troughs Solar Collector with no Need of Solar Tracking System,
December, 9 2009 patent no.: 0901005526 ( patent pending)
[7] Wattana Ratismith, Compound Parabolic Troughs Solar Collector with no Need of Solar
Tracking System, July, 21 2011 patent no.: 1101001216 ( patent pending)
[8] Wattana Ratismith, Novel Parabolic Troughs without Solar Tracking System, Proc. Renewable
Energy 2010, Pacifico Yokohama, Japan, (2010)
[9] Wattana Ratismith and Urith Archakositt, Parabolic Troughs without Solar Tracking System,
Third International Conference on Applied Energy, Perugia, Italy (2011)
[10] Standard Mathematical Table, 25th Ed., CRC Press, Inc., 1978.
[11] Zambolin E., Del Col D., 2010. Experimental analysis of thermal performance of flat plate and
evacuated tube solar collectors in stationary standard and daily conditions, Solar Energy 84,
1382-1396.
[12] Grass C., Schoelkopf W., Staudacher L., Hacker Z., 2004. Comparison of the optics of nontracking and novel types of tracking solar thermal collectors for process heat applications up to
300 C. Solar Energy 76, 207-215.
[13] Winston R., and Hinterberger H., 1975. Principles of Cylindrical Concentrators for Solar
Energy. Solar Energy 17, 255-258.
[14] Winston R., 1974. Principles of Solar Concentrators of a Novel Design, Solar Energy 16, 89-95.
[15] Blanco J., Malato S. et al, 1999. Compound Parabolic Concentrator Technology Develop to
Commercial Solar Detoxification Applications. Solar Energy 67, 4-6.
[16] Rabl A., O'Gallagher J. and Winston R., 1980. Design and Test of Non-Evacuated solar
Collectors with Compound Parabolic Concentrators. Solar Energy 24, 335-351.
[17] Rabl A., Bendt P., and Gaul H. W., 1982. Optimization of Parabolic Trough Solar Collectors.
Solar Energy 29, 407-417.
[18] Pei Gang, Li Guiqiang, Zhou Xi, Ji Jie, Su Yuehong, 2012. Experimental study and exergetic analysis of a
CPC-type solar, Solar Energy 86, 12801286.
[19] M. Souliotis, P. Quinlan, M. Smyth, Y. Tripanagnostopoulos, A. Zacharopoulos, M. Ramirez, P. Yianoulis.
2011. Heat retaining integrated collector storage solar water heater with asymmetric CPC reflector, Solar
Energy 85, 24742487.
24

[20] O. Helal, B. Chaouachi, S. Gabsi, 2011, Design and thermal performance of an ICS solar water heater based
on three parabolic sections, Solar Energy 85, 24212432.
[21] Duffie J. A. and Beckmen W. A, 1991. Solar Engineering of Thermal Process, New York,
Wiley.

25

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Absorption Heat Transformers (AHT) as a way to


enhance low enthalpy geothermal resources
Daniele Fiaschi a, CA, Duccio Tempestib , Giampaolo Manfridac , Daniele Di Rosad
a,b,c,d

Universit degli Studi di Firenze, Dipartimento di Energetica Sergio Stecco


C/o Centro Didattico Morgagni Viale Morgagni 40/44 - 50134 Firenze Italy
a daniele.fiaschi@unifi.it
b duccio.tempesti@unifi.it
c manfrida@unifi.it
d daniele.dirosa@stud.unifi.it

Abstract:
Exploiting low enthalpy geot hermal resources (hot water at 45 to 90 C) is increasing its attractiveness, due
to their higher widespread compared to high enthalpy ones. Anyway, at these temperature levels, their
application is almost solely related to district heating, due to the marginal efficiency achievable when
eventually converted to electricity by a power plant. The possibility of applying an A bsorption Heat
Trans former (AHT) to enhance low enthalpy geothermal resources available in the range from 45 t o 90 C
and produce electricity with an acceptable efficiency was investigated. A thermodynamic model of an AHT
working with a Lithium Bromide (LiBr) solution was developed. It showed the possibility of enhancing low
temperature heat, ranging from 45 to 90 C, to the higher value heat at 130 140 C with a COP variable
within 40 to 50%. The enhanced heat can be used as the hot source of an Organic Rankine Cycle (ORC),
which exploits the geothermal heat (available at 80 90 C) in summer time. Generally the geothermal
resource supplies district heating in winter, but when heat demand resets, it becomes attractive to convert a
fraction to electricity instead of leaving it waste. Even if the primary heat lost in the A HT (having COP in the
40 to 45% range) is less than compensated by the inc reas e in the ORC efficiency, leading to a lower
system efficiency, the adoption of an A HT is wort h up to about 80 100 C, because geothermal heat
source temperatures below 100C are not suitable for the current ORC technology When the temperature
of the heat source is higher t han 100 C, the direct use of the geothermal res ource into the ORC bypassing
AHT is more performing and convenient.

Keywords:
Absorption Heat Transformer (A HT), Organic Rankine Cycle (ORC), Low Enthalpy Geothermal.

1. Heat transformers and their potential for upgrading


geothermal resources
Absorption heat transformers (AHT) are devices which transform a large heat resource, which is
available at temperature too low for correct thermal matching within an industrial process, in a
smaller amount of heat available at a higher temperature level. They differ from traditional heat
pumps, in that they use no (or a very limited amount of) electrical power or work. Basically, AHTs
work on the principle of an absorption inverse cycle: however the net effect is that of transferring an
amount of heat (smaller than the originally available) at a higher temperature level. This allows
recovering this heat into industrial processes.
Examples of heat transformers can be found in the petroleum refinery industry [1] and have been
proposed as a useful add-on to industrial waste heat recovery [2-4] and for applications related to
renewable energies [5-7]. Being based on absorption inverse cycles, AHTs are not very suitable for
small-scale, individual applications. However, they can prove to be very attractive when there is the
opportunity to deal with relatively large heat rates (>200 kW) available at low temperatures. In
these applications, they can be more competitive in comparison to large, distributed renewable
26

energy harvesting technologies (such as low-temperature, non-concentrating solar collectors; or


parabolic troughs with low concentration ratio). The commercial attractiveness of AHTs lies in the
possibility of building a relatively compact unit, which can effectively upgrade a large quantity of
heat without needing for example - a large solar field. In this sense, they can be an effective
alternative to hybrid geothermal-solar power plants [7], which are currently being proposed as a
promising technology.
With reference to geothermal applications, heat transformers can find applications for two purposes:
a) In geothermal heating networks, AHTs can be employed to convert a large amount of heat at
low temperature (50-90 C) into a smaller amount at higher levels (80-130C). This is
interesting as the district heating network is usually not used for a long period (e.g., March
to October, depending on the local climate). It can instead be used to distribute heat at
higher temperature, to be used locally with absorption cooling units during summer.
Preliminary studies demonstrate that the rate of return of the overall system (Heat
transformer/district heating network/local secondary heat exchangers/absorption cooling
units) is promising [8].
b) When large amounts of heat are available from the geothermal resource at medium- low
temperatures (80-100 C), the current technology and thermodynamic limits (Carnot
efficiency) do not allow to propose binary (ORC) energy conversion systems. However, as
small ORC units operating at levels between 120C and 150C with engineered fluids are
coming to the market [9,10], it makes sense to try to upgrade the resource using an AHT.
In both cases (a) and (b), the original geothermal resource would be unavailable for commercial
exploitation (heat and/or electricity): then, it should be considered as wasted.

2. Thermodynamic layout and model of AHT


Basically, an AHT system operates with a reverse cycle of the AHP. Then, the AHT consists of an
absorber (AB), an evaporator (EVA), a generator (GE), a condenser (CO) and an internal liquid
liquid solution heat exchanger (EX), as shown in Fig. 1, together with the related thermodynamic
behaviour of steam into the T s diagram. The working fluid considered in this article is waterLithium bromide solution (water LiBr) where water acts as a refrigerant in the cycle, and LiBr as
an absorbent. A geothermal energy resource at low temperature (45 90 C) feeds the Generator,
where water is vaporized and then separated from the liquid solution rich in LiBr. The steam from
generator (GE) flows into the condenser (CO), where is condensed, releasing heat to the ambient
(i.e. to cooling water). The saturated water is pumped to higher-pressure level to the evaporator
(EVA), where it is totally evaporated by heat still given by the low temperature geothermal
resource. It must be noticed that the EVA and the GE work at the same temperature (Fig. 1). The
steam evaporated then flows into the absorber (A), where the useful heat is recovered at higher
temperature. Hence, the main capability of the AHT is to raise the temperature of the solution above
the temperature of the geothermal resource. AHT thermal efficiency is improved using a counter
flow heat exchanger between the dilute and the concentrated solutions. Fig. 1 shows the temperature
and pressure levels of the AHT cycle. The Generator and the Condenser work at lower pressure,
which is the saturation pressure of the condenser, while the Evaporator and the Absorber operate at
higher pressure, which is the saturation pressure of the evaporator. Moreover, the Evaporator and
the Generator work at same temperature, while the Condenser works at the lower cycle temperature
and the Absorber works at the higher cycle temperature. Thus, there are three temperature levels in
the AHT system: the generator (T GE=TE VA), the absorber (TAB) and the condenser (TCO)
temperatures.
27

Several analyses and optimization studies of AHT can be found in the technical literature [1-6; 1113]. In the present study, the work of Horuz and Kurt [3] is taken as a fundamental reference, and a
complete thermodynamic model was written in a simulation environment which includes the link to
common absorption fluid mixtures [14]. According to the existing literature [2-3;12-13] the main
assumptions made are the following:
The system is analyzed assuming steady state and thermodynamic equilibrium conditions.
Pressure drops and heat losses (or gains) are neglected.
The mechanical work consumed by the pumps is neglected.
The evaporator and generator temperatures are the same.
The refrigerant at the condenser and evaporator outlets are all saturated.

Steam

Evap orator

T [C]

140
130
120
110
100
90
80
70
60
50
40 3
30
20
10 2
0
0
1

Condenser

10

s [kJ/kg-K]

Figure 1a-1b: Absorber Heat Transformer layout (1a, left) and thermodynamic behavior of steam
at evaporator and condenser on T s diagram (1b, right).

It must be noticed that, as shown in [3], evaporator temperature higher than generator temperature
leads to a very slight increase of COP, especially for temperatures higher than 80C.
Furthermore, the temperature difference between the strong solution at the Absorber inlet (T10 ) and
the weak solution at the absorber outlet T5 is fixed at 10 C, as in [3]. Finally, the refrigerant mass
flow rate (m1 ) is fixed at 0,1139 kg/s, as in [3].
The thermodynamic analysis was carried out calculating the energy and the mass balance for each
AHT component. For the mass balance, the balance at the Generator must be highlighted:
(1)
(2)
Where x7 and x8 are the lithium bromide concentration in the weak and strong solution respectively.
An important design parameter is the flow ratio (f), which is defined as the ratio between the strong

28

solution mass flow rate (m8 ) and the refrigerant mass flow rate (m1 ). Hence, considering (1) and (2),
the flow ratio can be defined as:
(3)
The performance of the AHT is measured by the COP, which is the ratio between the heat delivered
from the absorber and the heat supplied to the evaporator and generator [14]:
(4)
Table 1 shows the thermodynamic conditions of the AHT cycle points (as referred on figure 1a), in
case of TAB=130C, TGE=TEVA= 80C, TCO=25C, while Table 2 reports the heat exchanged at the
absorber, evaporator, condenser, generator, the weak and strong solution concentration and the COP
in various cases.
Table 1. AHT thermodynamic states if TAB=130C, TGE=TEVA= 80C, TCO=25C.
1
2
3
4
5
6
7
8
9
10

T [C]
100
0
-100
1
0.5926
0.5926
0.5926
0.6565
0.6565
0.6565

m [kg/s]
0.1139
0.1139
0.1139
0.1139
1.171
1.171
1.171
1.057
1.057
1.057

P [bar]
0.03169
0.03169
0.4737
0.4737
0.4737
0.4737
0.03169
0.03169
0.4737
0.4737

T [C]
80
25
25
80
130
98
98
80
80
120

h [kJ/kg]
2650
104.8
104.8
2643
291.5
226.9
226.9
227.6
227.6
299.2

s [kJ/kg-K]
8.875
0.367
0.3669
7.611
0.7226
0.5564
0.5564
0.4204
0.4204
0.6126

Table 2. AHT results for various cases.


TAB (Absorber Temperature)
TCO (Condenser Temperature)
TGE (Generator Temperature)
TEV (Evaporator Temperature)
T10 (Solution Heat Exchanger Outlet Temp.)
f (Flow Ratio)
COPAHT
COPCarnot
QAB (Absorber Heat Transfer)
QEV (Evaporator Heat Transfer)
QCO (Evaporator Heat Transfer)
QGE (Evaporator Heat Transfer)
QSOL EXCH (Solution Heat Exch. Heat Transfer)
Xw (Weak Solution Concentration)
Xs (Strong Solution Concentration)

29

Unit
C
C
C
C
C

kW
kW
kW
kW
kW

Case 1
65
25
45
45
55
30.25
0.4108
0.5314
196.9
282.2
282.4
197.1
77.84
0.4642
0.4795

Case 2
115
25
70
70
105
17.77
0.4667
0.5656
251.8
287.2
287.7
252.3
136.3
0.5779
0.6104

Case 3
155
25
90
90
145
11.5
0.4855
0.5895
275.6
291
292
276.7
120
0.6457
0.7018

3. AHT sensitivity analysis


The main performance parameters of the AHT, COP and QAB, depend on the three temperatures at
heat sources and releases: TCO, TGE and TAB. The results of the sensitivity analysis to these
parameters are presented in Figs 2-5. As shown in Fig. 2, when T GE and TAB are fixed, the COP and
the heat released at the absorber decrease as the condenser temperature increase. This result is
justified by the fact that as the condenser temperature increases, the minimum AHT pressure also
increases (as it works under saturated conditions), leading to a lower LiBr concentration in the
strong solution (x8 ), and finally to a lower enthalpy value at the absorber inlet (h10 ).
290

0,64

TGE=TEV= 80 [C], TAB=130 [C]

0,56

270

COP Carnot, present study


COP Carnot, Horuz

260

COP

COPAHT, present study


0,52

COPAHT, Horuz

QAB [ kW ]

280

0,6

TGE= TEV= 80 [C], TAB =130 [C]

250

Horuz
0,48

240
0,44
17

21

TCO [C] 25

29

33

present study
T CO [C]

230
18

21

24

27

30

Figure 2: COP(left) and Absorber heat (right) vs the variation of the condenser temperature.

In Fig. 2 the values for COP and Q AB obtained with the presented model are compared with the
values given in [3]. COP values show a good agreement, with an average relative error of 1%, while
for QAB the average relative error is slightly higher (3,5%). It is due to the strong influence that
small error on the Li- Br concentration in the strong solution value has on Q AB. However, taking into
account that our model is developed with a software having probably different (and unknown to the
reader) thermodynamic libraries of Li-Br than the one used in [3], the average error may be
considered acceptable.
On the other hand, Fig. 3 shows that if the evaporator - and thus the generator temperature raises,
also the COP and the absorber output heat increases. In fact, in this case the maximum pressure
increases and the LiBr concentration in the weak solution decreases, leading to a lower enthalpy
value at the absorber outlet (point 5, Fig. 1) and then to a higher absorption heat and COP.
290

0,5
0,495

TAB= 130 [C] , TCO=25 [C]

TAB=130 [C], TCO=25 [C]


280

QAB [kW]

0,49

COP

0,485

270

0,48

0,475

260

pr esent study

0,47

COPAHT, present study

0,465
0,46
0,455
74

Horuz
250

COPAHT, Horuz

76

78

80

82

84

86

88

90

92

T GE = T EV [C]

240
74

76

78

80

82

84

TGE = TEV [C]

86

88

90

Figure 3: COP(left) and Absorber heat (right) vs the variation of the generator temperature.

30

92

Also in this case our results are compared with [3]: COP values are approximately in agreement,
confirming an average error of 1%, while Q AB values show, once again, higher discrepancies , with
an average error of approximately 3,5%.
Finally, Figs 4 and 5 show, respectively, the variation of the COP and QAB with the absorber
temperature at different generator (i.e. evaporator) temperatures. The increase of the absorber
temperature leads, obviously, to decrease of COP and Absorber released heat. In fact, a higher TAB
implies a higher enthalpy value of the weak solution at absorber outlet (point 5) which means a
lower heat released at the absorber. At each fixed value of TGE, an absorber temperature which
optimizes COP and Q AB is found.
0,5

0,49
0,48

COPAHT

0,47
TCO=25 [C]

0,46
0,45

TGE=45 [C]

TGE=70 [C]

0,44

TGE=50 [C]

TGE=80 [C]

TGE=60 [C]

TGE=90 [C]

0,43
0,42
0,41
40

50

60

70

80

90

TAB [C]

100

110

120

130

140

150

160

Figure 4: COP vs absorber temperature at different generator (evaporator) temperatures.

290

QAB [kW]

270

250

TCO=25 [C]

T GE=45 [C]

230

T GE=50 [C]

T GE=70 [C]
T GE=80 [C]

210
T GE=60 [C]

190
40

50

60

70

80

90

100

110

120

130

T GE=90 [C]

140

150

160

T AB [C]
Figure 5: Absorber heat vs absorber temperature at different generator (evaporator) temperatures.
31

The results shown in Figs 4 and 5 are in a good agreement with those presented in [13]. It can also
be observed that the optimized values of COP and Q AB can be kept almost constant at relatively
high values, provided that the decrease of the heat source temperature is followed by an adequate
reduction of the temperature at the absorber.

4. ORC system

T [C]

The Organic Rankine Cycle (ORC) is a market proven technology [15-18]. In this work we
considered a standard ORC (Fig. 6), where a secondary, low boiling-point working uid (i.e.
R245fa) is vaporised (points 16-18), then it expands in a turbine, and it is subsequently condensed
in the condenser before being returned to the economiser and the evaporator by the cycle pump. An
internal heat exchanger (IHE) is added to recover the heat from the turbine outlet stream, in order to
improve the cycle efficiency. Thermodynamically, the power plant works with a pure Rankine
cycle, without superheater.

180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0

ORC thermodynamic cycle (R245fa)


25 bar
17

16

18

19

4,79 bar

15
14

1,1

20
21

1,2

1,3

1,4

1,5

1,6

1,7

1,8

1,9

s [kJ/kg-K]

Figure 6: layout of the proposed Organic Rankine Cycle (ORC, left) and the related thermodynamic
cycle (right)

The following assumptions for the ORC model are done:


The ORC operates in a steady-state condition.
The pressure drops throughout the heat exchangers and pipelines are neglected.
The turbine isentropic efficiency is fixed at 0.76.
The mechanical work consumed by the pump is neglected.
The organic fluid exits the vaporizer under saturated conditions (x=1).
The temperature at the condenser of the ORC is fixed at 25 C.
The main purpose of this work is to compare the efficiency of two systems for the exploitation of a
low enthalpy geothermal resource (45 to 90 C): in the first system (called Geo-ORC), the
geothermal resource is used to directly vaporize the organic fluid in the ORC vaporizer and
economizer (Fig. 6). In the second system (called ORC-AHT), the ORC vaporizer is fed by the heat
32

released at the absorber of the AHT previously described. In the Geo-ORC system, for a given
temperature of the geothermal resource (Tsource, i.e. T11 ), the temperature of the organic fluid at
point 18 is calculated by (5), where the Approach temperature difference ( Tapp;ORC) is fixed by
technical constraints at the heat exchanger at 10C:
(5)
Furthermore, since the geothermal flow rate is given as an input data, the temperature difference at
the pinch point ( Tpp;ORC) is fixed at 5 C by the following equation:
(6)
The energy balance at the vaporizer and at the economizer are then given by the following
equations, where (7) provides the organic fluid flow rate (mORC), while (8) calculates the reinjection
temperature of the geothermal resource (Tgeo;out ):
(7)
(8)
In the second system, i.e. the ORC-AHT, the geothermal resource is used in the evaporator and
generator, which work at the same temperature. Since Q GE (or QEV) is fixed to the operating value
obtained by the AHT pure model, mgeo and TGE (or TE V) are given by the following equations:
(9)
(10)
The geothermal resource temperature at the outlet of the generator (or evaporator) is calculated by:
(11)
From (9) it can be stated that, as the geothermal resource temperature increases, T GE, and then Q GE,
rises, leading to an augmentation of mgeo . The geothermal mass flow rate calculated by (9) is used
as input in the GEO-ORC system, i.e. in (7) and (8). Approach temperature difference at the
generator ( Tapp;GE) is fixed at 10C, while the temperature difference at the geothermal resource
outlet from generator ( Tout;GE) is fixed at 5C. Finally, in the ORC-AHT system, the heat Q AB is
coupled with the vaporizer and economizer. A heat carrier fluid is used between the two systems.
However, as the temperature of the heat source and the evaporation temperature of the organic fluid
are constant and respectively equal to TAB and T18 , the temperature difference ( Tapp;ORC) between
TAB and T18 is fixed by technical constraints at the heat exchanger at 10C. Hence, the energy
balance at the vaporizer and economizer is calculated by the equation (12), where the temperature at
point 18 is fixed using (13):
(12)
(13)

33

5. Model results of the ORC-AHT combined system


In Table 3, the thermodynamic parameters (pressure, temperature, mass flow rate) of the ORC cycle
in both AHT-ORC and GEO-ORC configurations for different temperatures of the heat source are
shown. As previously explained, the geothermal flow rate and power input are the same for ORCAHT and GEO-ORC at each source temperature. As the temperature of the resource is raised
through the AHT, the maximum ORC Temperature (i.e. the turbine inlet temperature) is higher for
ORC-AHT in every case, leading to higher ORC efficiency. The system efficiency is given by the
ORC efficiency for GEO-ORC, while for ORC-AHT it is the COP of the AHT times ORC
efficiency. Since the COP of the AHT is always very low, i.e. never higher than 0.5 for the simple
AHT configuration here considered, the ORC-AHT system efficiency is lower than the GEO-ORC
one. It must be noticed that, as both systems operate between the same temperature (i.e. source and
condenser temperature), the Carnot efficiency and the Carnot turbine power output are the same for
both systems. Since the geothermal power input and flow rate are the same for ORC-AHT and
GEO-ORC at the same heat source temperature, the lower system efficiency leads to a lower turbine
power output in case of ORC-AHT in comparison to GEO-ORC. In Table 3, columns for source
temperature of 60C and 80 C are shown in grey, because the current ORC technology does not
allow to exploit such low temperature resource. The minimum temperature at the evaporator of the
ORC, at current level technology, is about 120C, even though in a relatively short term it seems
reasonable to expect evaporator temperature of about 100C. Hence, the AHT-ORC system here
proposed can convert this geothermal resource into electricity at such low temperature, which,
otherwise, would be wasted.
Table 3. Geo-ORC and ORC-AHT results for various TSOURCE.
ORC-AHT
Heat Source Temperature [C]
Absorber temperature [C]
Geothermal flo w rate [kg/s]
Geothermal power
input [kW]
Maximu m ORC
temperature [C]
Maximu m pressure [bar]
Condenser pressure [bar]
Temperature at
turbine outlet [C]
Geothermal flu id
outlet temperature [C]
Organic flu id flow rate [kg/s]
ORC efficiency [% ]
ORC Carnot efficiency [%]
System efficiency [% ]
System Carnot efficiency [%]
COP A HT [-]
COP Carnot AHT [-]
Turbi ne power output [kW]
Turbine Carnot Power
output [kW]
Absorber heat transfer [kW]
Internal heat recovered [kW]
ORC Condenser heat [kW]
AHT Condenser heat [kW]
Specific power output [kW/kg ORC]

GEO-ORC

80
115
25.74

90
135
26.48

100
155
26.95

110
160
27.61

120
160
26.32

60
23.6

80
25.74

100
26.95

120
26.32

539

556.1

568

583.6

558

493

539

568

558

105

125

145

150

150

50

70

90

110

14.16
1.478

21.27
1.478

30.84
1.478

33.81
1.478

33.81
1.478

3.432
1.478

6.097
1.478

10.09
1.478

15.74
1.478

48.7

52.7

51.2

47.4

47.4

32.2

38.3

44.5

49.9

75

85

95

105

115

55

75

95

115

1.11
14
23.2
6.5
15.6
0.467
0.672
35

1.147
16
27
7.7
17.9
0.479
0.664
42.5

1.17
17.2
30.4
8.4
20.1
0.486
0.662
47

1.229
17.2
31.2
8.5
22.2
0.50
0.712
49.6

1.11
17.2
31.2
8
24.2
0.469
0.775
45

2.39
5.5
10.5
5.5
10.5
27

2.51
9.1
15.6
9.1
15.6
49

2.56
12.1
20
12.1
20
68

2.45
14.5
24.2
14.5
24.2
81

84

99.54

114

129.5

135

52

84

114

135

252
25
217
288
31.6

266.3
30.3
223.7
289.9
90

276
29
228
292
40.6

289.4
26.2
239.8
294.2
110

262
24
217
296
40.4

16
466
11.4

32
490
19.5

47
499
26.8

58
477
33.1

34

This analysis suggests that, with the proposed ORC, the adoption of AHT to upgrade the
geothermal heat is worth for low temperature of geothermal resource, up to 85 100 C.

6. Conclusions
The possibility of using an AHT system with a LiBr water solution to upgrade the low
temperature heat available from geothermal resources has been discussed. The upgraded heat is
used to feed an ORC working with R245fa in a pure Rankine recuperative power cycle.
Both systems are modeled by calculation codes developed with the same calculation tool, which
includes the thermodynamic libraries of LiBr solution and ORC working fluid. The results showed
that the COP of the AHT and the amount of upgraded heat (Q AB, for a fixed value of m1 ) are
strongly dependent on the three main design temperatures, i.e. TCO (close to the environmental
value), TGE and TEV (close to the available geothermal source), and TAB (the value of upgraded
heat). Specifically, for temperatures of low grade heat variable between 45 and 90 C and the
corresponding upgraded heat variable into the 55 155 C range, the achieved COP varies between
38 and 50%. The condenser (i.e. environmental) temperature has a slight influence on AHT
performance: with fixed upgraded heat from 80 C (TGE=TEV) to 130 C (TAB ), the COP is reduced
from about 48% to about 45% for an increase of condenser temperature from 15 to 32 C. The
results agree with those found on literature [3].
The enhanced heat released by the absorber of the AHT can be used into the boiler of an ORC,
which has the possibility to use the geothermal heat source, which otherwise would be unexploited,
and thus wasted. However, the convenience of AHT ORC combined system over the simple GEO
ORC is strongly related to the temperature of the geothermal heat source. Generally, coupling
AHT ORC is convenient until the geothermal resource temperature is below 120C, which is the
lowest temperature suitable for current ORC technology.

Nomenclature
COP
T
LiBr
m:
p:
Q:
T:
x:
W:
[1]...[n]:
Suffixes
app;GE
app;ORC
source:
geo;pp:
geo;out:
geo:
out;GE

Coefficient of Performance
Temperature difference [C]
Lithium Bromide
Mass flow rate [kg/s]
Pressure [bar]
Heat rate [kW]
Temperature [K]
Lithium bromide mass fraction in the solution
Power [kW]
Thermodynamic point of the cycle
Approach difference to GE
Approach difference to ORC
Geothermal inlet to the system
Geothermal at the pinch point
Geothermal reinjection into the well
Geothermal
Geothermal outlet from GE
35

p:
pp;ORC
sp
t:
Acronyms
AB
AHT
CO
EVA
EX
GE
GEO-ORC
ORC
ORC-AHT
IHE

Pump
Pinch point ORC
Specific
Turbine
AHT Absorber
Absorption Heat Transformer
AHT Condenser
AHT Evaporator
AHT Liquid to liquid heat exchanger
AHT Generator
ORC directly fed by the Geothermal resource
Organic Rankine Cycle
ORC coupled with AHT
Internal Heat Exchanger

References
[1] Tufano, V., Heat recovery in distillation by means of absorption heat pumps and heat
transformers, Applied Thermal Engineering 1997; II(2): 171-178.
[2] Zhao, Z., Ma, Y., Chen, J, Thermodynamic performance of a new type of double absorptio n
heat transformer, Applied Thermal Engineering 2003;23: 24072414
[3] Horuz,I., Kurt, B., Absorption heat transformers and an industrial application, Renewable
Energy 2010;35: 2175-2181
[4] Xuehu Ma, Jiabin Chen, Songping L., et alii, Application of absorption heat transformer to
recover waste heat from a synthetic rubber plant, Applied Thermal Engineering 2003;23: 797
806
[5] Rivera, W., Romero, R.J, Thermodynamic design data for absorption heat transformers: Part 7 :
Operating on an aqueous ternary hydroxide, Applied Thermal Engineering, 1998;18(3-4): 147156
[6] Best, Rivera, W. Thermodynamic design data for absorption heat transformers: Part 6 :
Operating on water-carrol, Heat Recovery Systems and CHP, 1994;14(4): 427-436
[7] Lentz, A., Almantza, R., Solargeothermal hybrid system, Applied Thermal Engineering
2006;26: 15371544
[8] Fangtian Sun, Lin Fu, Shigang Zhang, Jian Sun, New waste heat district heating system with
combined heat and power based on absorption heat exchange cycle in China, Applied Therma l
Engineering 37 (2012) 136 144
[9] Pratt and Whitney Corp., Model
280 kW PureCycle Power System,
http://www.pw.utc.com/media_center/assets/model_280_purecycle_power_system.pdf
[10] ElectraTherm's
Heat
to
Power
Generation
Systems
available
at
http://www.electratherm.com/products.html last access 10 February 2012
[11] Yin J, Ming-Shan Z, Li- Zhong H. Performance analysis of an absorption heat transformer
with different working uid combinations. Applied Energy 2000;67(3):281-92.
[12] Sozen A, Yucesu HS. Performance improvement of absorption heat transformer. Renewable
Energy 2007;32:267-84.

36

[13] Scott M., Jernqvist A., Olsson J., Aly G. Experimental and theoretical study of an open
multi-compartment absorption heat transformer for different steam temperatures. Part I :
hydrodynamic and heat transfer characteristics. Applied Thermal Engineering 1999;19: 279298.
[14] EES Engineering Equation Solver, http://www.fchart.com/.
[15] Di Pippo R. Geothermal Power Plants: Principles, Applications and Case Studies. London,
UK: Elsevier Advanced Technology; 2006.
[16] Turboden Combined Heat And Power Orc Units For The Pellet Industries, 2008, (see also
http://www.genewscenter.com/content/detail.aspx?releaseid=7229&newsareaid=2http://www.t
urboden.eu/en/public/press/Turboden_ORC_for_pellets_english.pdf)
[17] Obernberger I, Hammerschmid A, Biomass fired CHP plant based on an ORC cycle
Project ORC-STIA-Admont. Final Report, Bios-energy systems; 2001 (see also
http://www.bios-bioenergy.at/uploads/media/Report-ORCAdmont-Thermie-2001-03-26.pdf)
[18] GE Energy Announces Industrial Waste-Heat Recovery Innovation for Onsite Power Plants,
http://www.genewscenter.com/content/detail.aspx?releaseid=7229&newsareaid=2, 6 July, 2009
(last access 4 January 2012).

37

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, C OST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Alternative feedstock for the biodiesel and energy


production: the OVEST project
Matteo Prussi2, David Chiaramonti1, Lucia Recchia 3, Francesco Martelli2,
Fabio Guidotti4
1

RE-CORD c/o Dept.of Energetics S.Stecco, University of Florence (Italy);


2
CREAR, c/o Dept.of Energetics S.Stecco, University of Florence (Italy);
3
CREAR - DEISTAF, University of Florence (Italy);
4
Silo S.p.A.

Abstract:
A large share of the vegetable oil market is today shifting from food to energy sector. This trend led to
discussions about the chance of a confliction between these two chains and its possible social and
environmental consequences. Moreover the increasing interest in pure vegetable oils for energy and
biodiesel production promotes market volatilities and, in recent years, high and peak prices. The economics
of the biodiesel production is suffering the actual market situation.
In this framework, the present work presents the results of the OVEST project, which is supported by the
Tuscany region. Aim of the project is the investigation of the possibility to use waste oil streams of the food
industry to obtain an alternative feedstock for bioliquids and biofuels production. In the current food market, a
number of waste streams are available from the food industry: oleins and fatty acids, sludge oils and WCO
(Waste Cooking Oils) represent several examples of these waste streams from edible oil production. Most of
these feedstock oil streams are composed by distilled fractions obtained during vegetable oil refining, i.e.
free fatty acids, di- and mono-glyceride, etc. In the OVEST project, vegetable oil is obtained from waste
materials by the re-esterification process, in order to obtain a new feedstock for biodiesel production and for
power generation.
A market analysis has been carried out in OVEST, to estimate the economic viability and the potential of
each waste streams. PFAD (Palm Fatty Acid Distillate), fatty acids from vegetable oils and WCO are very
interesting feedstock to obtain re-esterified oils. The new vegetable oils obtained in OVEST show promising
chemical-physical characteristics for biodiesel production, such as the fatty acid composition and the
contamination from solids and inorganics.
Test on a modified Micro Gas Turbine (30 kWel) have been carried out. Performances and pollutant missions
have been monitored. The results of the experimentation confirmed the viability of these materials for the
energy production.
From the environmental point of view, the use of these feedstock allows to meet the requirements of the
European Directive EC/28/2009 (RED), that introduces minimum targets in terms of GHGs saving for
bioliquids and biofuels. According to the EC Directive, LCA must be carried out considering equal to zero the
wastes and residual products GHG emissions before processing.
The present work showed that most of the input materials to the esterification plant may be classified as
wastes or residues, and that assuring an accurate traceability for each treatment and transport phase is a
really complex issue. However the GHG savings indicated in the RED for bioliquids from wastes, was
confirmed by the work here carried out.
Therefore OVEST project demonstrates the viability of the use of re-esterified oils (reconstructed oils) for
biodiesel production.

Keywords:
Vegetable Oil, organic waste, conversion technology.

1.

INTRODUCTION

The European biofuels market is rapidly growing in recent years and it can be defined by three
main actors: biodiesel (78.5% on energy base), bioethanol (17.5%) and pure vegetable oil
(4.0%). Biogas is rapidly growing in EU market; the application to the transport sector is often
proposed but still represent a minor component.
38

Biodiesel is the most important biofuel in European region, both for consumption and for
production, representing about 78% of the total biofuels market in the transport sector. The EU27
production, in 2008, was 6,860 million liters, for a total consumption of 9,465 million liters
[USDA, 2008]. Germany, France and Italy are the leaders of this market and together accounted
for 80% of the EU biodiesel production.
The recent issued Renewable Energy Directive (RED [EU Parliament, 2009]) confirms the 10%
target, set by the EC, for energy from renewable sources in transport. This 10% is a minimum
target to be achieved by all Member States for the share of biofuels in transport petrol and diesel
consumption by 2020. Even if the EU directive poses a rigid framework for the fuel sustainability,
focusing on the GHGs reduction potential, the main reasons to set such ambitious target deals with
the EU fuels supply security.
From this point of view, it is necessary to underline that demand is not growing only in Europe.
China set a target of biodiesel penetration for 2020 and New Zealand, Australia, Japan, among the
others, have mandates for biofuels blends [New Zealand Parliament, 2008]. Biofuels worldwide
consumption accounted for 24.4 Mtoe in 2006, compared to only 10.3 in 2000. Despite this
considerable increase, they still represent only 1.5% of the total road-transport fuel demand in
2006 [IEA, 2008] and rooms are today present for an increase in the demand.
As regards the feedstock, in contrast with the United States, where about 50% of biodiesel is
produced from soybean oil, rapeseed oil and sunflower oils represent the major source of biodiesel
for the European industry (4.7 Mton [Pelkmans, 2009]).
The issue of the cost is today one of the critical issue for the biodiesel producers. During the entire
2007 and part of 2008 the price of the food oils rose up, constituting an example of the price
instability of this kind of market. Causes have been identified in the rapid growth of the Asian
market, in the temporary decrease in the yield of several crops, due to unfavorable climatic
conditions, in the increasing demand for biofuels and in the speculation on food and fuel
commodities, the crude oil price to 147 US$/barrel [US EIA, 2009]. Nevertheless price instabilities
still continue to appear cyclically and they can be still considered the major risk factor for transport
fuel market.
The total world vegetable oil production has increased in the last decades: the main growth has
been recorded for palm oil (+8% annually), followed by palm kernel oil (+7.6% annually) and
soybean oil (+5.7% annually) [Pelkmans, 2009]. Palm oil and soybean oil represent more than half
of the market, being 30% and 28% respectively. Rapeseed oil is also important, especially for
Europe [Lieberz, 2009].
The vegetable oil consumption is divided between food industry and other industrial applications,
among which, the biodiesel industry is rapidly rising in importance. Nonetheless more than 80% of
the vegetable oil produced worldwide is destined to the food market.
The sector of vegetable oil trade is today dramatically changing. In the framework of the
international scenario, European Union vegetable oil market is strongly affected by biodiesel
sector. Despite palm oil is often perceived as the major actor in the biofuel market, the production
of biodiesel from this feedstock is very limited due to the poor winter stability of the product. An
increasing use of palm oil is observed as bioliquid for co-generative systems (high displacement
reciprocating engine for power generation with heat recovery), or in substitution of rapeseed oil for
food industry.
Despite the intensive productions, European Union (EU-27) is a net importer of oil. The total
import of oil for food and industrial uses is estimated equal to 34% of the total domestic
consumption [USDA, 2009].
The OVEST project aims to investigate alternative feedstock for biodiesel and direct energy
production starting from food industry by-products and wastes. The main goal of the project are
related to the techno-economic demonstration of the viability of the use of this low quality
feedstock to produce energy in stationary systems. Moreover the lower costs of this materials
39

(figure 1), compared to the standard oils, are interesting as a potential answer for the biodiesel
sector.

2.

THE OVEST PROJECT

Many waste materials are today available from different markets. The possibility to consider these
waste streams as potential feedstock for energy production, it is very interesting topic. Feedstock
from waste materials are profitable in the actual context as an open discussion is on-going about
the land use and the competition between the food and the energy chains for bioenergy sector.
Supported by the Tuscany Region, the OVEST project focuses its attention on several possible
feedstock rich in fatty-acids content for oil production. There are several by-products of the
processing of edible vegetable oils; there are mainly fatty-acids and oleins, Waste Cooking Oils
(WCO) and sludge oils (the residues of tanks washings). The esterification process allows to obtain
tri-glycerides from these materials, creating a product useful for both energy of biodiesel
industries.

3.

NON-STANDARD FEEDSTOCK

Oleins and Fatty-Acids are by-products of the same food chain: the de-acidification of the edible
vegetable oils. This process is carried out to obtain a low acid-content edible oil, usually lower
than 1%. Oleins and Fatty-acids are the results of different techniques. Oleins are obtained by
treating a soapy paste, resulting from the chemical de-acidification of the oil, with sulfuric acid.
This kind of chemical de-acidification produces a material with over the 50% of FFA (Free FattyAcids) content [Dumont M-J et al. 2007]. Fatty-acids are secondary products of distillation
(physical de-acidification) of vegetable oils: their FFA content is very high (over 80%) [Hartman
L. 1978], with also a normally lower content of impurities and contaminant than oleins.
While the previous described feedstock may present a relatively little fluctuation of their FFA
content, different is the case of WCO and sludge oils. The properties of WCO are highly
influenced by the kind of heating treatment they undertook. In the case of the sladge oils, their
property are linked with the characteristics of the tanks they came from.
WCO can be a high quality feedstock if well pre-treated and in particular filtered for the removal
the solid impurities [Balat M. 2010].
The most interesting qualities of this kind of feedstock is related to their cost. This is generally
lower than the Pure Vegetable Oil. The price of this class of materials is subjected to volatility like
the price of the oil, nevertheless this it has been noticed that these fluctuations are smaller, which
make these substances interesting even in critical market situations.

Figure 1- Recent oil market price trend (CPO: crude palm oil; PFAD: palm fatty acids distilled).

40

In this framework, the sludge oils have to be considered for being very available and cheap,
compared to the other by-products. Nevertheless their utilization may be not easy in the actual
technologies, as their quality is poor. They can be used by a proper dilution with other purer
materials.
Particular is the case of WCOs; they represent a good-quality feedstock to start with, their cost is
similar to PFAD and they are available in large quantities even on regional scale. WCO has a high
potential which needs to be supported by a well-coordinated collection mechanism. An interesting
example is the experience carried out in Tuscany (Italy). In this Region, in the past years, the
collection of WCO was limited to the restaurants. Today it has been powered and WCO are also
available from a domestic collection system. The availability of this material has increased
strongly and it is still developing, so it may become possible the implementation of a highly
positive environmental short-chain embracing the attention to social and environmental
sustainability proposed by the European RED Directive (Renewable Energy Directive
EC/28/2009).
The actual market price is guiding the use of these non-standard feedstock. WCO is widely used as
feedstock for biodiesel production, even because the European directives allow to associate at this
material a double positive effect on the GHGs reduction. Sludge oils are more interested for the
use in engines for the direct power generation, as their quality is lower compared to the other
materials but also their costs.

4.

RE-ESTERIFICATION PROCESS

The esterification process allows to obtain a new vegetable oil from FFA and glycerol. This
process is well known in the biofuels area to be also a pre-treatment stage for the biodiesel
production.
Controlling the stoichiometric ratio between Free Fatty Acids and glycerol it is possible to re-built
an oil which is not edible (because of the reaction may alter the chemical residues from the
previous processing) [Bhosle B. M. 2004] that can be directed to the energy chain without
conflicting with the food sector. The feedstock required is largely available on international
markets (for example Palm Fatty Acid Distilled) and on regional scales.
The technology described in the present work is a non-catalyzed reaction, which will take place in
two separate reactors. In the first stage the glycerol and the FFA are introduced in the reactor and
then heated up at 200-250 C for two hours. The glycerol evaporates with the water produced by
the reaction itself, so it is necessary to discharge water outside while force the glycerol to
condensate back in the reactor, so to be re-circulated.
Depending on the kind of feedstock used, the esterification process needs some upstream
treatments on the feedstock, in order to obtain a high quality oil. To reduce the impurities and
contaminant: for example, on the WCOs, a severe filtration is required; de-phosphoration and
dehumidification are advised in any case.
The esterified oil could not immediately suitable for power generation or as feedstock for biodiesel
production. It is often necessary to provide further processes, like post-filtrations and
deodorization.

5.

POWER GENERATION FOR RE-ESTERIFIED OILS

Stationary power generation is an interesting market for the feedstock described. Standard
technology for energy conversion are the reciprocating engines, based on diesel cycle and the gas
turbine. The chemical and physical characteristics of the re-esterification process are different from
fossil diesel oil and rather far from the common technical specifications for engines. Major

41

differences are the kinematic viscosity and the Lower Heating Value, but also as regards
contaminant levels and composition, fuel cold properties, ignition behavior, etc.
A standard for the use of vegetable oils in engines is the DIN 51605 (table 1). It has been defined
for the use of rapeseed oil in transports engine. It is a reference to qualify the characteristics of the
vegetable oil for the power generation.
Table 1 DIN51605 technical specification of the use of vegetable oil in engines.

As previously stated, the sludge oils are an interesting class of materials for power generation. The
oil obtained by the sludge has presented the highest chemico-physical characteristics variation,
especially in terms of impurities and contaminants, because of the wide variation of the original
material. In general the quality of this oil is far from the technical specification defined in the
standard. In table 2 is reported the concentration of several important elements in the vegetable oil.
The high value shown confirm the difficulty to directly use these oils for engine application.
Their use is interesting when diluted to highest quality feedstock in order to reduce the cost of the
final product.
Table 2 - Metal content in oil from sludge .
Limits
max
Min
Mean

Max
Min
Mean

Sulfur
[ppm]
18,40
0,26
7,99
Calcium and magnesium
[ppm]
83,10
44,50
58,18

Sodium and potassium


[ppm]
114
86
104
Calcium
[ppm]
31,00
18,00
24,53

Vanadium
[ppm]
<1
Copper
[ppm]
1
0,1
-

For what concerns the other considered feedstock, the best results has been noticed for the fatty
acids and WCOs.

42

Table 3 - Examples of analysis resulted on samples of esterified oil from WCO


Sample
Impurities (ppm)

A
3

B
1.5

C
1.5

Humidity (%
d.b.)
Acidity (mg
KOH/g)

5.44

3.03

4.82

2.98

2.62

2.69

In OVEST project the re-esterified oil has been tested in a 30 kW Micro Gas Turbine. This
machine has been modified to be fed by this oils. In particular the fuel line has been adapted to the
characteristics of the oil in input. The functioning resulted very stable and the performances
substantially unmodified. After 250h of stable functioning many parts of the engine have been
examined. In particular analysis have been carried out on the deposits of the injectors. The
concentration of phosphorous in the oil appeared to be the most critical parameter as it was the
main constituent of the deposits.
The analysis of the pollutant concentration in the exhaust gases from the turbine confirmed the
good quality of the combustion. The concentration of NOx resulted to be higher when turbine
works with re-esterified vegetable oil compared with regular Diesel but CO concentration
appeared to be significantly lower. In figure 2 is shown a direct comparison with Diesel on the
pollutants emissions.

Figure 2: MGT C30 pollutant emission measured for pure vegetable oil feeding.

6.

LIFE CYCLE ASSESSMENT

The economical potential benefits in the use of these nonstandard products is not the only positive
effect. In the present work the LCA has been carried out starting with the definition and
classification of the feedstock. The classification as waste or by-product is a crucial point to assess
the environmental balance: according to the European RED Directive (EC/28/2009) definitions,
the emission of CO2 associated to the original chain are not taken into account if the product in
exam is a waste. This is an crucial starting point to meet the 35% minimum target of GHGs saving.
In order to verify the achieving of the RED target the CO 2eq calculation basing on the methodology
proposed by the directive has been implemented. Moreover, considering that the energetic
consumptions and therefore the fossil energy requirements are significant along the reesterification process also the CER parameter has been evaluated.

43

Besides the work provide to identify the production chain and the geographical origins of the
feedstock (especially when classified as by-product) and to define clearly the stages of
production and eventually the LCA methodology. The phases modeled (see, figure 3) in the LCA
carried out can be summarized in the following steps:
Transportation to the esterification reactor;
Re-esterification phase;
Use of oil for energy production.

Figure 3 - Scheme for the definition of the new oil in LCA analysis.

The LCA has been carried out through a spreadsheet built using the Biograce project assumptions:
particularly the Biograce emission factors have been considered (see table 4).
Table 4 - Emission factors used in the LCA (Source: Biograce).
GHG
g/g
CO2
1
CH4
23
N2O
296
Transport
Road transport
(liquid materials)
Ship transport
Train transport

MJ/tkm
1.01

gCO2/tkm MJ/tkm
88.34
1.1693

0.12
0.21

10.85
26.81

Re-esterification
Diesel
Natural gas
Grid electricity MV
Grid electricity LV

gCO2/MJ
87.64
67.58
127.65

MJ/MJ
1.16
1.13
2.70

129.19

2.73

0.1437
0.5660

The transport phase has been evaluated considering the geographical provenience of the current
feedstock treated in the Silo SpA, hypothesizing the means of transport and estimating the average
distance.
44

For the re-esterification process the energetic consumptions have been measured in the plant of the
Silo SpA (see table 5): the values measured do not highlight a difference due to the feedstocks
characteristics.
Finally, for the energy production a 250 kWel engine with 41% of electrical efficiency and 85% of
thermal efficiency has been considered.
As shown in table 6, the potential GHGs saving is near 90%. The important results allows to meet
the requirements of the directive EC/28/2009 on sustainability.
Moreover, the results obtained highlight that the transport phase have higher GHG emissions than
these expected by the RED because of the far distances to be covered for delivering the feedstocks
to the re-esterification plant. On the other hand the emissions are comparable to these indicated in
the directive for the biodiesel production from waste materials.
Table 5 - Energy consumption for the production phase
Power consumption
Methane consumption

47.81
48.72

kWh/t
kg/t

Than has been calculated the equivalent CO2 and the CER parameters, which consider the fossil
energy consumptions for the process.
Table 6 - Results of LCA
Phase
Trasport
Esterificaztion
Total
Saving for cogenation

gCO2/MJ_oil
2.54
4.53
7.08
92%

CER analysis
Process
Trasport
Esterification
Total

MJ/MJ_oil
0.0364
0.0783
0.11

As shown in table 5, the potential GHGs saving is near 90%. The important results allows to meet
the requirements of the directive EC/28/2009 on sustainability.

7.

CONCLUSIONS

B iodiesel

industry and power generation applications are today suffers for the market price and the
price volatility of the traditional feedstock. Alternative feedstock are available from food industry,
mainly from edible oils processing. By-products are oleins and the acid oils. Sludge oils and Waste
Cooking Oils (WCO) are also interesting materials, rich in free fatty acids, that can be converted
into re-esterified oils.
A new non-edible oil has been obtained thanks to the esterification process, adding glycerol to the
inlet materials. The characteristics of the re-esterified oil are variable, mainly based on the
characteristic of the feedstock. The oil could need further treatments but are mostly compatible
with the use in engines and for the biodiesel industry. The lower quality for engine feeding is an
aspect that can be improved, especially for very low cost feedstock, such as the sludge oils.
Test has been carried out in a 30 kW micro gas turbine. Good performances have been observed
for re-esterified feeding.
45

From technical and economical point of views the best results were given by Palm Fatty Acid
Distilled for the biofuel production. A good potential is for the Waste Cooking Oils, available also
at regional level.
For power production in engine the cost is a key factor and sludge oils are very interesting even if
the poor qua lity is today limiting their use.
The LCA resulted very interesting for the oils obtained from this process: the potential GHGs
saving resulted higher than 90% compared to the traditional fossil fuels, so the target (35%) set by
EU Directive can be reached.

ACKNOWLEDGMENT
The authors wish to acknowledge the Tuscany Region, for the support, project coordinator SILO
SpA and the project partner Il Trebbiolo.
The authors would also to acknowledge Eng. Lisa Bigozzi for her precious support.

BIBLIOGRAPHY
[1] Balat M, Potential alternatives to edible oils for biodiesel production A review
of current
work, Energy Conversion and Management vol. 52, Elsevier Journal, Ottobre 2010.
[2] Bhosle B. M, Subramanian R, New approaches in deacidification of edible oilsa review,
Journal of food engineering vol. 69, Elsevier Journal, November 2004
[3] Dumont M-J, Narine S, soapstock and deodorizer distillates from North American vegetable
oils: Review on their characterization, exrtation and utilization, Food Reserch International
n 40, Elsevier Journal, June 2007.
[4] EU Parliament. DIRECTIVE OF 2008/28/EC OF THE EUROPEAN PARLIAMENT AND
OF THE COUNCIL, 2009.
[5] Hartman L, Deacidification of edible oils by short path distillation, Rivista- Italiana-delleSostanze-Grasse Vol. 55, 1978.
[6] IEA, World Energy Outlook, 2008.
[7] Lieberz S.,EU27 Oilseeds annual, Global agricultural information network, NLE49039, 2009.
[8] New Zealand Parliament, Biofuel bill, www.parliament.nz, 2008.
[9] Pelkmans, F., Walter, F., Rosillo-Calle, C. (2009) A global overview on vegetable oils, with
reference to biodiesel, IEA Task 40, report.
[10] USDA. "United states department of agriculture foreign agricultural service".
www.fas.usda.gov, 2008.
[11] US Energy Information Administration, Policy research working pape, tonto.eia.doe.gov,
2009.
[12] USDA, EU27 biofuels annual, Global agricultural information network, NL9014, 2009.

46

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Assessing repowering and update scenarios for


wind energy converters
Till Zimmermanna
a

University of Bremen, Faculty of Production Engineering & artec | Research Center for Sustainability
Studies, Bremen, Germany, tillz@uni-bremen.de

Abstract:
Wind energy converters are a major technology for the generation of power with even increasing relevance
in the future. The shares of wind energy in the grid mix es of many countries are growing steadily. This
growth is not only due to new c onverters on new sites but also due to repowering of sites already in use as
well as updating of old wind energy converters. There is a huge pot ential for repowering and updating but
since wind energy is supposed to reduce anthropogenic impacts on the environment, it is important to
consider environment al aspects, for repowering as well as for updating. The methodological approach
described in t his paper allows identifying the environmentally preferable option from a variety of product
scenarios using the specific energy demand as an exemplary indic ator, while other indic ators can be
included, too. The method facilitates life cycle data regarding the converters energy balance or depending
on the desired indicators- other figures from the converters life cycle inventory. The described assessment
can also be performed c onsidering case (and site) specific factors. However, it has t o be noted that data
demand and availability might occur as a burden.

Keywords:
Wind energy converters, Wind turbines, Energy demand, Repowering, Updating, Sustainability.

1. Introduction
Wind energy is a technology of increasing importance the installed capacity and the total number
of installed wind energy converters (WEC) are steadily growing, in Europe as well as globally [1,2].
As a consequence thereof, repowering and updating of old converters have an increasing relevance
within the wind energy sector. While various LCA studies have been performed for new WEC,
environmental assessments of specific repowering scenarios or of updated second-hand WEC are
rarely found. Still, the question about the environmental benefits from updating or repowering old
WEC should not be discounted for. Also, the question whether updating or repowering is the
environmentally favorable option for a particular old WEC might be of increasing relevance in the
future. To address these aspects a method for comparing different product scenarios has been
developed based on a previous work that focused exclusively on assessing repowering scenarios [3].
This method uses the specific energy demand as an exemplary indicator; including other indicators
is possible, too. To demonstrate the method two exemplary assessments are performed illustrating
use and possible outcomes of an assessment.

2. Background: todays situation of wind power generation,


expected developments and the roles of repowering and
updating
Wind energy is a major technology within the grid mix of the future since it is supposed to provide
the world with clean and almost carbon neutral energy, reducing anthropogenic greenhouse gas
emissions and other environmental impacts.

47

Germany as the leading European country in terms of installed wind power capacities has an
installed capacity of more than 27,000 Megawatts [4] contributing already a significant share to the
German grid mix. This share is constantly growing in Germany as well as in other European
countries. The European Wind Energy Association, for example, has the announced target of 190
GW onshore and 40 GW offshore installed in 2020 and 250 GW onshore and 150 GW offshore in
2030 respectively [1]. This is accompanied by individual goals in different European countries. The
German government for example has the goal to increase the share of renewable energies in power
supply to 35% in 2020 [5]. A steady growth can also be observed globally [2]. Here, the installed
capacity grew within the past years already by 22.9% (mid-2010 to mid-2011) and 23.6% (mid2009 to mid-2010) respectively [2]. This growth is not only due to newly installed wind energy
converters but it is also caused by repowering or updating of already existing converters and the
installation of second hand converters, respectively. Repowering means the replacement of older
WEC by new state-of-the-art WEC, hereby improving the utilization of existing sites and reducing
the total number of installed converters. Since in most countries the most suitable sites for onshore
wind energy are already in use, this is the main motivation for repowering. Also potential land use
conflicts are hereby reduced. Updating of WEC includes refurbishing of an existing WEC,
replacement of obsolete parts and possibly a slight increase of its rated power.
Germany, for example, has a potential for repowering of about 25,000 MW, which equals
approximately the rated power of WEC already installed in 2010 [6,7]. The worldwide development
of wind energy indicates that also in many regions outside of Germany repowering is going to be an
issue as well as updating of WEC.
As besides economic interests- the reasons for using wind energy are to a large extent
environmental aspects, it is important also for repowering and updating, that they results in an
overall improved environmental performance. Choosing the (ecologically and environmentally)
preferable option from a variety of available repowering and updating scenarios as well as finding
the (ecologically and environmentally) right moment in time for the respective measures is a
challenging task. The methodological approach, described in the next section addresses this task.

3. Methodological approach
To assess the respective product scenarios we used a method comparing the specific energy yield of
the different product scenarios to each other. Here, a product scenario is regarded as the
combination of an old WEC that reaches the end of its life span and a repowering WEC or an
updated version of the old WEC. If, for example, a 0.6 MW WEC reaches the end of its life span
and there are two repowering options (A and B) available with updating/ refurbishing being another
option (C), this results in three different product scenarios. Additionally, the repowering options
might become available at different points in time, resulting in different repowering points:
a) 0.6 MW WEC + repowering WEC A at point A
b) 0.6 MW WEC + repowering WEC B at point B
c) 0.6 MW WEC + updated/refurbished 0.6 MW WEC at point C
For the identified (generic) product scenarios the specific energy demand is compared in the next
step. This can be done calculative as well as graphically, which will be demonstrated in the next
section. If required, the method can be used for additional impact categories like the carbon
footprint (as demonstrated in [3]) or aggregated measures of several impact categories depending on
the availability of the required data. Using several different impact categories might lead, of course,
to inconclusive results (or at least to results that might need intensive discussion), since there will
not always be an option that minimizes all assessed impact categories. Focusing on one impact
category on the other hand will provide more conclusive results but discounts other relevant
aspects. Still, this relevance always depends on the goal of the study that also might give indications
on which impact categories should be assessed with priority.
48

However, to demonstrate the methodological approach within an exemplary assessment, the


specific energy demand has been selected, since it can be regarded as one of the most discussed
impact categories for WEC besides the global warming potential and upcoming- the abiotic
depletion potential. It indicates the required energy input (consumed energy) per energy fed into the
grid (energy output) and can be calculated according to the following formula (cf. [3]); a description
of the variables is given in Table 1:
CEDold t r CEDx (t x )
ex t
Enet, old tr Enet, x t x
Table 1. Overview of variables
Scenario
year of installation
tr
repowering/updating point (EoL point of old WEC)
CED old(tr)
Cumulative energy demand (CED) of the old WEC up to the repowering
point tr
Enet,old(tr)
net energy production of old WEC up to repowering point tr
X
index of potential repowering WEC, or of the potential repowering
scenario, respectively
tx
expected life span of repowering WEC X
CEDx (tx )
CED of repowering WEC X subject to its life span
Enet,x (tx )
net energy production of repowering WEC X (full lifecycle)
ex (t)
specific energy yield in repowering scenario X at point t
As it can be seen in the formula ex is a time dependent variable. The specific energy demand
decreases over time with (usually) different gradients for the assessed scenarios which can lead to
different results depending on the analyzed time frame.
The data demand within the described methodology varies depending on the desired quality of the
results. To assess the specific energy demand as sole impact category, the minimum data demand
includes the net energy production for the respective converters as well as their cumulated energy
demand and data regarding boundary conditions like installation points or repowering points
respectively. Depending on the goal of the study, this data needs to be case specific meaning that it
needs to be based on the specific production, transport, site and end-of- life conditions, or it can be
generic allowing the use of aggregated or average data from other studies. If life cycle inventories
for the converters in question are available, it will provide the necessary data. If for example the
specific carbon footprint (i.e. the CO 2e-emissions per kWh) is to be included as an indicator, the
global warming potential (GWP) needs to be calculated based on the inventory data. The specific
carbon footprint (cfx ) is then calculated according to the following formula, analogously to the
specific energy demand with GWPold being the GWP of the old WEC subject to the repowering/
updating point tr and GWPx being the GWP of the product option X subject to its life span:

cf x t

GWPold tr
Enet, old tr

GWPx (tx )
Enet, x tx

Other impact categories or result indicators like the abiotic depletion potential which is of growing
relevance can be included accordingly.

49

4. Exemplary assessments
Within this section, the previously described method is applied to two examples. In the first
example different repowering scenarios are compared to each other, while two different update
scenarios are assessed in the second example.
The underlying data represents generic converters of different rated powers, i.e. it is based on other
studies (see [3,8]) and does not consider any site specific factors that would need to be considered
in a case or site specific assessments of specific converter types. In some cases, site specific factors
may have a critical influence on the results. Therefore, the following examples basically
demonstrate how the described method can be used in order to compare different product scenarios
to each other, rather than providing decisive conclusions or recommendations on how the life cycle
of particular WEC should be designed.
Generally, the maximum life span of WEC is considered to be 20 years. This needs to be taken into
consideration when analyzing the results of the assessment.

4.1. Repowering scenarios


For the exemplary assessment of repowering scenarios a 0.6 MW WEC installed in 1998 has been
selected as base scenario. For this converter three possible repowering scenarios have been
assessed: repowering in 2007 with a 1.8 MW converter, repowering in 2009 with a 2.0 MW
converter and repowering in 2014 with a 3.0 MW converter. An overview of these scenarios is
given in Table 2. Besides their increased rated power, the repowering WEC also have a higher
number of annual full load hours, due to their increased hub height, which is based on an expert
judgment from an industry partner (see [3]). Also, the 1.8 MW WEC has a rather high CED
compared to the 2.0 MW and especially the 3.0 MW WEC. As a first indication on the energetic
performance, the ratio of rated power to CED or the energetic payback time ([annual net energy
production / CED] * 12) could be calculated prior to the graphical or calculative assessment of the
different product scenarios. For a comparison of repowering scenario with identical repowering
points this figure even is sufficient to identify the preferable choice.
Table 2. Overview of assessed repowering scenarios
Scenario
year of
rated power [MW]
installation
base scenario:
1998
0.6
WECold
scenario 1
2007
1.8
scenario 2
2009
2.0
scenario 3
2014
3.0

annual full load


hours
2,500

CED [kWh]

2,800
3,100
3,100

3,600,000
3,053,000
3,100,000

1,880,000

Based on the method described in section 3 and the above data the comparison of the different
scenarios is carried out. The results of this comparison are shown in Fig. 1.

50

CED/Enet
[kWh/kWh]

1,00

0,10

0,01
1998

2008

WEC,old

2018

WEC, old

2028
Year
scenario 1

2038

scenario 2

2048

scenario 3

Fig. 1. Assessing the specific energy demand of different repowering scenarios.

In the figure, additionally to the different repowering scenarios, a reference scenario representing no
repowering is shown by the dashed black graph. This reference scenario functions basically as an
orientation; no repowering is not a real option due to the limited life span of WEC but may still
function as a benchmark (however, updating/refurbishing could be an option, which is assessed in
the following section).
In the diagram, the x-axis marks the installation of the original WEC which is associated with a
high energy consumption (including production, transport and other factors influencing the CED)
while energy production starts right after the installation. The resulting graph showing the specific
energy demand falls with proceeding time as the produced net energy grows. The installation points
of the different repowering scenarios are also associated with energy consumption appearing as
peaks within the figure, falling with proceeding time, too.
It can be observed, that the repowering scenario 1 does not fall below the reference scenario within
the assessed time scope, while scenarios 2 and 3 show a better performance. Scenario 2 falls below
the reference line in 2023, which is 14 years after its installation. Scenario 3 shows an even better
performance, falling below the reference line already in 2022, i.e. eight years after its installation.
So, the graphical assessment (and the calculative assessment, too) shows scenario 3 as the favorable
option of the assessed product scenarios; i.e., here, repowering of the 0.6 MW WEC in 2014 with
the 3 MW WEC is the best option in terms of the specific energy demand.

4.2. Update scenarios


For the exemplary assessment of different update scenarios a 2 MW WEC has been chosen as the
converter reaching its end-of- life. For this converter three different product scenarios are assessed:
1. Updating of the 2 MW WEC, increasing its rated power to 2.3 MW; installation at a different
site,
2. Updating of the 2 MW WEC, increasing its rated power to 2.3 MW; installation at the same site,
51

3. Repowering with a 3 MW WEC; installation at a different site.

For scenario 1 it has been assumed that rotor blades and tower cannot be reused and need to be
replaced. Rotor blades have a life span of maximum 20 years and usually after 12-15 years
refurbishing is not economically feasible anymore. Concrete towers are usually not suited for reuse,
either. So, it has been assumed that rotor blades and tower need to be re-produced, resulting in an
energy demand equal to the original primary production of these components. The same applies to
access roads, crane hard-standing and foundation. For the electrics it has been assumed that 90%
can be reused while 10% need replacement (cf. [9]). For the nacelle including the generator it has
been assumed that refurbishing and updating require an energy demand of 10% of the original CED
of these components. This can be considered as a rather conservative assumption based on [9] with
regard to the fact that around 80% of the production CED comes from material production, with
20% resulting from additional processing and transport (cf. [9]).
For scenario 2 with utilization of the same site it has been assumed that the existing infrastructure
(access roads, crane hard-standing, foundation) and the tower however requiring refurbishing
equal to 5% of the original CED can be used again; given that a life span of about 60 years can be
regarded as realistic for concrete towers (cf. [9]). For the other main components rotor blades,
generator, nacelle and electricity the same assumptions as for scenario 1 apply.
For both update scenarios 2,300 annual full hours have been assumed.
An overview of these parameters is shown in Table 3.
Table 3. Update scenarios
Component
Scenario 1
rotor blades
new
tower
new
generator
reuse, 10% of original CED
nacelle
reuse, 10% of original CED
electric components
reuse, 10% of original CED
infrastructure
new
full load hours
2,300

Scenario 2
new
reuse, 5% of original CED
reuse, 10% of original CED
reuse, 10% of original CED
reuse, 10% of original CED
reuse
2,300

The third scenario has been chosen as a benchmark for the update scenarios. It is identical with the
3 MW WEC that has been assessed in the previous section except for its earlier date of installation.
Within each product scenario, end-of- life credits are given 100% to the second product system. This
way, a consistent approach for each product system is assured and the comparability of the results is
increased. End-of- life credits are given according to the base end-of- life scenario described in [9]
using system expansion.
Using these assumptions and boundary conditions the graphical assessment of the update scenarios
has been carried out according to the described method, too, leading to the picture shown in Fig. 2.

52

CED/Enet
[kWh/kWh]

1,00

0,10

0,01
2005

2010

WEC, old

scenario 1

scenario 2

scenario 3

2015

2020

2025

2030

2035

2040

Year

Fig. 2. Assessing the specific energy demand of different update scenarios.

In the diagram, the dashed black line shows the original WEC. It is continued by the alternative
scenarios for updating and repowering, respectively. Contrary to the assessment in section 4.1 this
assessment is not focused on (or limited to) one particular site but assessed different possible life
cycle or product options resulting from the original WEC reaching its end-of- life.
As it can be seen in the figure, scenario 2 not requiring infrastructure shows the best performance of
the assessed scenarios. The assessment of scenarios 1 and 3 leads to rather similar results; here
considering site specific factors could give a better understanding of the environmental burdens.
This shows that, if the utilization of an existing infrastructure is possible, updating appears to be a
preferable option with regard to optimizing the energy yield, even more attractive than repowering
assuming that a new infrastructure is required, here. For the use of updated WEC as second- hand
WEC at a different site, the difference to new WEC is not that significant.

5. Conclusions and discussions


Repowering and updating of WEC are trends of increasing importance. It has been demonstrated
how both can be assessed under environmental aspects, using the specific energy demand as an
exemplary indicator. Within the exemplary assessment it has been shown how different repowering
or update scenarios can be compared in terms of their specific energy demand, a method that can be
adapted to other impact categories, too. In this exemplary assessment generic data has been used,
basically demonstrating how to use the described method and allowing only general conclusion on
the benefits of different repowering and update strategies.
Regarding repowering, the method allows assessing different repowering options at one particular
site in terms of their specific energy demand. Given a selection of available repowering scenarios
preferable options can be identified, also indicating the optimal point in time for repowering (cf.
[3]). General recommendations on repowering strategies cannot be deduced, here, except that
significant differences between the different repowering scenarios are possible and need to be
assessed, if the favorable option in terms of energy balance or other indicators shall be identified.
53

Based on the exemplary assessment of update scenarios, it can be said that the use of updated
converters as second-hand WEC is definitely associated with environmental benefits but since
components like tower and rotor blades that usually do not have second-hand potential are
responsible for a significant share of the CED, the advantages compared to state-of-the-art WEC are
reduced- at least regarding the assessed indicator. If however re-using tower and infrastructure is
possible as well, updating shows a supreme performance also compared to state-of-the-art
repowering WEC. Still, if the favorable option needs to be identified in a specific situation a case
specific assessment needs to be performed. In general, however, it has to be said, that second- hand
WEC and new WEC serve different markets and do not compete against each other. Also, re-use
can be generally recommended; this also applies to wind energy converters. Giving WEC a second
life makes sense in terms of energy balance as well as other environmental impact categories and
economically (since selling and reuse is more profitable than disposal). Especially from a resource
point of view re-use of WEC can be generally recommended. Even though recycling of WEC is
relatively unproblematic with the rotor blades being an exception re-use still is the most
resource saving option.
While an assessment using the described method can provide valuable information and help to
better understand the environmental implications of different product scenarios, it requires a
significant amount of data. To assess the specific energy demand, for example, knowledge of
several figures is required: rated power, time of availability, number of annual full load hours and
cumulated energy demand. While the first three are usually known, the latter requires some effort to
be identified, especially when site specific aspects shall be included in the assessment. Here,
conducting a life cycle wide assessment of the energy balance for all identified product scenarios
might be necessary, or if other indicators shall be included a full life cycle assessment (LCA).
These assessments require some effort and might not always be feasible due to data availability.
The development and use of customized LCA tools (cf. [8]) can be of help, here, and significantly
reduce the effort related with such assessments (a onetime extensive data collection is still required,
though). Performing LCAs and additional case studies using the described methodology would
allow creating a data basis that could be used for future assessment. This data basis might also allow
drawing more specific conclusions on the environmental benefits and chances of repowering and
updating of WEC.

References
[1] European Wind Energy Association, Statistics and Targets. Factsheet, 2011. Available at
<http://www.ewea.org/fileadmin/swf/factsheet/1_statisticsandtargets.pdf>. [accessed
13.12.2011].
[2] World Wind Energy Association, Half year report 2011, Bonn, 2011.
[3] Zimmermann T., Gling-Reisemann S., Optimal repowering of wind energy converters:
Energy demand and CO2 intensity as indicators. Life Cycle Management Conference 2011,
2011 Aug 28-31; Berlin, Germany.
[4] European Wind Energy Association, Wind in power: 2010 European statistics, 2011.
Available
at
<http://ewea.org/fileadmin/ewea_documents/documents/statistics/EWEA_Annual_Statistics_2010.pdf> [accessed 13.12.2011].
[5] Bundesregierung, Erneuerbare Energien ausbauen, 2011.
Available at <http://www.bundesregierung.de/Content/DE/Artikel/2011/06/2011-06-06-eegnovelle.html> [accessed 13.12.2011].
[6] Bundesverband WindEnergie e.V., Zukunftsmarkt Repowering, 2010. Available at
<http://www.wind-energie.de/de/themen/repowering/> [accessed 22.10.2011].

54

[7] Windenergie Agentur, Repowering von Windenergieanlagen in der Metropolregion BremenOldenburg: Handlungsempfehlungen zur Steigerung der Akzeptanz von Repowering,
Bremerhaven, Germany, 2010.
[8] Zimmermann T., Fully parameterized LCA tool for wind energy converters. Life Cycle
Management Conference 2011, 2011 Aug 28-31; Berlin, Germany. Available at
<www.lcm2011.org/papers.html?file=tl_files/pdf/poster/day2/ZimmermannFully_parameterized_LCA_tool_for_wind_energy_converters-567_b.pdf>, [accessed
23.12.2011].
[9] Zimmermann T., Entwicklung eines Life Cycle Assessment Tools fr Windenergieanlagen.
Thesis, Bremen, Germany, 2011.

55

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Biogas from mechanical pulping industry


potential improvement for increased biomass
vehicle fuels
Mimmi Magnussona and Per Alvforsa
a

Division of Energy Processes, School of Chemical Science and Engineering, KTH Royal Institute of
Technology, Stockholm, Sweden, mimmim@kth.se (CA), alvfors@kth.se.

Abstract:
Biogas is a vehicle fuel of the first generation of biofuels with great potential for reducing the climate impact
from the transport sector. Today biogas is mainly produced by digestion in Sweden and the total amounts to
1.4 TWhLHV /year (2010) of which about 0.6 TWh LHV is upgraded and used in the transport sector. Using
industrial wastewater, e.g. from a pulp and paper mill, as substrate for production of biogas, the amount of
renewable fuel to the transport sector could be increased. In the pulping industry, substantial amounts of
organic matter are generated; this is commonly treated aerobically to reduce the chemical oxygen demand
(COD) in the effluent streams before discharge to a recipient. Treating these effluent streams mainly
anaerobically instead could contribute to the transport sectors energy supply. The aim of this study is to
investigate the potential for using effluent streams from the Swedish mechanical pulp and paper industry to
produce biogas. A typical Swedish mechanical pulp mill is considered for anaerobic treatment of the
wastewaters. This type of pulp mill presently uses conventional methods for wastewater treatment to reduce
COD, but converting most of this to anaerobic treatment would increase the amount of biogas produced.
When considering this conversion in a larger context, supposing that anaerobic treatment would be applied
to all Swedish mechanical pulp mills, which stand for about 30% of the total Swedish pulp production, it is
shown that the production could amount to as much as 0.5 TWh LHV /year of biogas. This represents about
one third of the biogas produced in Sweden today. The main conclusion of this study is that if anaerobic
treatment of effluent streams from the pulping industry were introduced, the biogas production in Sweden
could be significantly increased, thus moving one step further in reducing the transport sectors climate
impact.

Keywords:
Anaerobic digestion, Biofuel, Biogas, Mechanical Pulping, Vehicle Fuel, Wastewater Treatment.

1. Introduction
The transport sector is seen as one of the most problematic areas when it comes to reduction of
climate impact, due to its heavy dependence on fossil fuels. Biogas is a vehicle fuel of the first
generation of biofuels, meaning that it is presently available on a commercial scale and has a welldeveloped infrastructure for distribution and filling, in some areas via a gas grid. In total, the biogas
production in Sweden amounts to 1.4 TWhLHV/year (2010) of which about 0.6 TWhLHV/year is
upgraded and used in the transport sector [1]. The production of biogas is commonly performed by
anaerobic digestion, a technology that has been seen as one of the oldest to stabilise waste and
sludge [2]. The production today is mainly based on sludge digestion and takes place in sewage
treatment plants, but co-digestion plants, for instance with food wastes or agricultural wastes as
substrates also exist, and there is further anaerobic treatment as a method for wastewater treatment
in other types of industries.
The pulp and paper industry has been one of the most important base industries in Sweden for a
long time, and still holds a strong position domestically as well as internationally. The annual
production of pulp in Sweden is 12 million tonnes of the world total which is 178 million tonnes
[3]. Pulp and paper production requires large amounts of water for the process and the effluent
waters are rich in dissolved organic matter, measured as chemical oxygen demand (COD). Great
56

progress has been made to reduce the water use at the mills by closure of the water circuits and also
to improve the wastewater treatment of the effluents. Since the late 1970s the environmental impact
from pulp and paper industries has decreased substantially, for instance the emissions of COD in
Sweden has decreased from 1.6 to 0.2 million tonnes of COD per year, while the pulp production
has increased by about 4 million tonnes per year [3]. In Sweden the most common way to treat
wastewaters from the pulp and paper industry has been by aerobic methods, which are rather
electricity demanding: anaerobic wastewater treatment has so far not reached a break-through.

1.1 Aim
The aim of this study is to investigate the potential for using effluent streams from the Swedish pulp
and paper industry to produce biogas. A case study is made on a typical Swedish mechanical pulp
mill, the Rottneros mill, where a combination of anaerobic and aerobic treatment is considered an
alternative to conventional aerobic methods for wastewater treatment. In the case study, design
calculations are performed based on present effluent data from the mill. Based on the results from
the case study on the Rottneros mill a potential biogas production for all Swedish mechanical pulp
mills (producing groundwood, thermomechanical and/or chemi-thermomechanical pulp) is
estimated as a possible way of increasing the share of renewable fuels in the transport sector.

2. Background
Biogas is a vehicle fuel that is well suited for both light and heavy vehicles. It is commercially
available and can be distributed either in a gas grid, or to filling stations by trucks. The most
common way of production is by anaerobic digestion, one of the oldest processes for treating
organic wastes [4, 5]. Although used to a great extent in sewage treatments plants (for sludge
digestion), the anaerobic methods for wastewater treatment have not reached a break-through in the
pulp and paper industry in Sweden, where other wastewater treatment methods are more common.
In the following sections, a short introduction to anaerobic digestion and to the pulp and paper
industry is given.

2.1. Anaerobic digestion


Digestion of organic material to produce biogas is a process used on a commercial scale. All types
of organic material may be subject to digestion; amongst the most commonly used substrates are
food waste, sewage sludge, manure and agricultural waste. It is one of the standard methods used in
wastewater treatment (WWT) technology, to reduce effluent treatment sludge from municipal plants
and various industries. The conditions for the digestion process are dependent on the type of
substrate used; the main difference between anaerobic treatment of wastewaters and digestion of
organic wastes concerns the reactor design and the retention times. The main stages are, however,
more or less the same, namely: collection of the substrate, pre-treatment, digestion, and, depending
on end-use, upgrading. The digestion itself is a complex series of reactions where the substrate is
degraded in four stages: hydrolysis (rate-limiting step); acidogenesis; acetogenesis and
methanogenesis, where the methane (CH 4) is formed [6, 7]. In this series of reactions different
groups of microorganisms work together in converting the substrates and intermediate products to
final products. Together with methane the main product from the digestion is carbon dioxide (CO 2).
The ratio between these varies depending on the substrate and process used, but the methane
content is typically in the region of 45-85% [6]. The by-products are, aside from possible traces of
other gases (e.g hydrogen sulphide and ammonia), water and the digestion residue (a solid fraction).
To be used as a vehicle fuel the raw gas has to be upgraded to a quality with high methane content
(97 +/- 2% according to Swedish standard, SS 15 54 38, European standards vary between 85% and
97% [8, 9, 10]) and thus most of the CO 2, and water have to be removed. Traces of other gases,
such as sulphur-containing gases, have to be removed to not cause disturbances in the engine. The
CO2 may be vented to the atmosphere, used in other industries, used for carbon capture and storage
57

(CCS) or, as described in a study by M ohseni et al [11], be used to produce additional methane by
reacting with hydrogen.
There are several parameters influencing the digestion and the rate of the different stages; among
these are, e.g. temperature, number of fermentation steps, pH, retention time, alkalinity and
moisture content of the substrate. These parameters are specific for each process and may be
adjusted to enhance the biogas production. The temperature for digestion, however, is often seen as
one of the most sensitive process parameters. There are three temperature intervals where the
digestion is performed optimally: psychrophilic/cryophilic (020 C), mesophilic (3042 C) and
thermophilic (4560 C): industrial digestion is most commonly performed with the two higher
intervals [6, 7]. Generally, a higher temperature gives a faster reaction rate, which is an advantage
since smaller reactor volumes are needed. However, the thermophilic digestion is more sensitive
when it comes to temperature and concentration variations and needs more process control, whereas
the mesophilic digestion is a more stable process [6]. Whether mesophilic or thermophilic, the
process requires heating to keep the reactor at the appropriate temperature.
The upgrading of the raw biogas, needed if the gas is to be used as vehicle fuel, is usually
performed adjacent to the digestion, mainly only at large facilities since upgrading is rather costly.
Common techniques for upgrading include water scrubbing, pressure swing adsorption (PSA) and
chemical absorption. After upgrading the biogas may be compressed or condensed to liquid biogas
(LBG) before being distributed to a gas grid, if available, or by trucks to filling stations [8].

2.2. Pulp and paper


The forestry industry is one of the largest industries in Sweden, and accounts for 10-12% of the
employment, export, sales and added value in the Swedish industry. On an international level, the
Swedish forestry industry also holds a strong position, being the third largest in Europe and second
largest combined exporter of pulp, paper and sawn wood products in the world. The Swedish
pulping industry produces a little less than 12 million tonnes of pulp each year, of which 3.7 million
tonnes are for the pulp market and the remaining portion is processed to paper. The largest share of
the pulp in Sweden is produced by chemical processes (mainly sulphate/kraft pulp), whereas the
mechanical pulps (groundwood, thermomechanical and/or chemi-thermomechanical pulp) account
for about 30%. On a global level, however, mechanical pulps only account for about 21% [3].
The purpose of the pulping process is to separate the cellulose fibres in the wood from each other in
order to use them as raw material i.e. in papermaking; this can be achieved either mechanically or
chemically. In mechanical pulping the cellulose fibres in the wood are physically torn from each
other by grinding. In the groundwood pulping (GWP) process the logs are pressed against a rotating
grinder stone while water is added; if the system is also pressurised the process is called pressure
groundwood pulping (PGWP). In the refiner mechanical pulping (RM P) processes wood chips are
ground between refiner discs; two variants of this process are thermomechanical (TM P) and chemithermomechnical pulping (CTM P), where the pressurised grinding is preceded by heat or chemical
treatment, respectively. Other variations of the above-mentioned processes also exist. M echanical
pulping has a very high fibre yield where 90-97% of the raw material is found in the end product,
compared to chemical pulping where about half of the wood is transformed to pulp and the rest (i.e.
lignin) is used for energy conversion. However, pure mechanical pulping is very energy demanding:
the higher the process temperature the higher the energy demand. M ost of the electric energy input
is transformed to heat, which is removed from the wood by cooling water. This energy can be
recovered either for pulp and paper drying or as hot process water [12].
The emissions from mechanical pulping include emissions to air, water and solid wastes. The water
emissions depend on the process used and increases if bleaching is applied, but they are mainly
comprised of organic compounds with a chemical oxygen demand (COD) that can cause oxygen
deficiency, acute or chronic toxicity, mutagenicity or eutrophication in the recipient water [4, 12].
The predominant methods for reducing COD from pulp and paper wastewaters are presently aerobic
effluent treatment methods, such as the activated sludge process and biofilm treatment; another type
58

of treatment method is chemical precipitation [12]. M ost WWT facilities at pulp and paper mills
today are external treatment plants handling effluent streams before discharge to the recipient.
Aerobically based WWT methods are generally rather energy demanding and also produce large
amounts of sludge. The solid wastes from the process include wood and bark residues, fibre rejects
(primary sludge) and biological sludge from the WWT [12]. Solids that are dry enough are usually
incinerated in the bark boiler or at external combustion plants, but the main fraction of solid wastes
primary sludge and biological sludge has to be dewatered, and possibly also dried, before
further treatment, for instance incineration, which is of course energy demanding [12, 13].

3. Anaerobic treatment of effluent streams in mechanical


pulping
Anaerobic treatment of wastewater to reduce COD has so far not been implemented to any large
extent in Swedish pulp and paper mills, however it is more common in European mills. A few of
the reasons for that are that the effluent volumes from the pulp and paper are very large (and thus
have been seen as difficult to handle anaerobically) and that the anaerobic system is very sensitive
to disturbances, which might be a problems since the wastewaters may contain large amounts of
sulphur or other substances that are toxic [12, 14]. However, TM P as well as GWP effluents are
well suited for anaerobic treatment. The carbohydrate content in the effluent is rather high (higher
in groundwood pulping than in TMP), as is the anaerobic degradability, while the methanogenic
toxicity is low. For CTMP the anaerobic degradability is not quite as high as for TM P and GWP,
mainly because of the high contents of lignin and wood extractives in the effluent [4], but the
potential for anaerobic digestion is still rather good. In a study by Sivard and Ericsson [15],
different process alternatives for effluent treatment of process flows from the pulping industry
(mechanical as well as chemical) have been suggested. The treatment alternatives included
combinations of anaerobic treatment, flocculation and membrane filtration and the aim was to
accomplish more energy- and resource-efficient wastewater treatment with low discharges and
more recirculation of process waters. The results from that study were promising: not only could the
process water be returned to the process to a greater extent, the anaerobic treatment also gave a
substantial biogas production. According to [12], the power demand for removal of 1 tonne of COD
by anaerobic treatment is about half of the power needed when using aerobic methods.
In following sections the case study and the potential for Sweden are presented. The present
situation for the WWT at the studied mills is described in Section 3.1, as well as the suggested
treatment with anaerobic digestion together with the relevant data for the calculations. As for the
Swedish potential, an overview of the mechanical pulp mills is given in Section 3.2, together with
the data and the assumptions made for the basis of calculations for the potential for biogas
production.

3.1. Case study The Rottneros mill


As stated previously this study uses a Swedish pulp mill to exemplify how anaerobic digestion
could be implemented for appropriate effluent streams, to produce biogas. The case study is based
on design calculations from a master thesis [16] performed at KTH the Royal Institute of
Technology in cooperation with F and Rottneros. The subject for the case study, is the Rottneros
mill, one of the two production sites in Sweden own by the Rottneros Group. The Rottneros mill
has two separate production lines where two grades of pulp are produced: GWP and CTMP. Their
yearly production capacity is about 170 000 tonnes of pulp; in 2010 the production amounted to 138
500 tonnes, of which 53% was groundwood pulp. [17].
Currently the Rottneros mill uses an aerobic wastewater treatment: a combination of activated
sludge treatment and chemical precipitation, which is a traditional and rather power-demanding
treatment (a simplified process diagram may be seen in Fig. 1). Simply described, the process
waters from the pulping process and the wood yard are taken to the primary clarifier, where solids
(primary sludge) are separated. The clarified water is then pumped to the activated sludge treatment,
59

including aeration basin and secondary clarifier. Nutrients (phosphorus and nitrogen) are added at
the activated sludge stage. Part of the sludge from the secondary clarifier is returned to the aeration
basin. There is no withdrawal of excess sludge from the secondary treatment stage. Water and the
remaining sludge are led to tertiary treatment comprising chemical addition (of polyaluminum
chloride) and flotation. The pH of the effluent water from the flotation is measured and adjusted
before discharge to the recipient, lake Fryken. The sludge from the flotation is dewatered and the
reject water is returned to the secondary clarifier, whereas the biologic and chemical sludge,
together with the primary sludge, is taken care of by an external company. The total degree of COD
reduction in the present wastewater treatment is 83% [16].

Fig. 1. Simplified process diagram over the present WWT for Rottneros.
In the suggested WWT for the Rotteros mill, anaerobic treatment in an upflow anaerobic sludge
blanket (UASB) reactor is combined with aerobic treatment, see Fig. 2 for a simplified process
diagram. Before entering the anaerobic reactor the flow is treated by chemical precipitation with
FeSO4 in the primary clarifier to reduce the amount of organic matter that is difficult to degrade and
to make the wastewater less toxic to the micro-organisms present in the anaerobic process. The
COD reduction in this primary precipitation stage is approximately 15%. The water from the
primary clarifier is pH-adjusted with acid, and then taken to the UASB reactor. The degree of COD
reduction in this stage is 70%. The produced biogas, with methane content of 70%, is separated and
compressed. After the anaerobic treatment some further biological degradation is performed in the
present aerobic treatment at Rottneros. An aerobic polishing is needed to avoid odours from the
effluent. The flotation stage is used for final removal of suspended solids and if needed some
chemicals may also be used for further reduction of COD [16]. The total degree of COD reduction
for the entire WWT (including the primary clarifier, aeration and secondary clarifier) is 84%. The
data for the present and for the suggested anaerobic treatment for Rottneros are presented in section
4.1.

Fig. 2. Simplified process diagram over the suggested WWT for Rottneros.

60

3.2. Potential for Sweden basis for calculations


There are presently 12 mills in Sweden producing GWP, TMP and/or CTMP. Their total pulp
production amounts to 3.6 million tonnes of pulp per year (2010) and, as mentioned, this stands for
30% of the total Swedish pulp production. The situation at these mills is of course variable; most of
them are integrated pulp and paper mills while three are market pulp mills. The pulp grades
produced also vary, where four mills produce only one grade (CTM P: two mills, TMP: one mill,
GWP: one mill) whereas the rest produce two or more grades. These mills include two mills
producing sulphate pulp and CTMP and two mills producing recycled fibre pulps as one of their
grades. [18] This means that the situation concerning the effluent WWT is different at these mills.
All mills, except two, use the activated sludge process, in some cases in combination with other
biological treatment stages and/or chemical precipitation. Because of their different conditions,
these mills have varying potential for using anaerobic digestion with biogas production as part of
their wastewater treatment system. However, as stated in the aim of this study, this potential is
roughly estimated based on the results of the case study. The pulp production and the present
emissions of organic compounds for the Swedish mechanical pulp mills are presented in Table 1.
Table 1. Total production of mechanical pulps and emissions of COD (based on [18]).
Groundwood pulp, t/year
TM P, t/year
CTM P, t/year
Total pulp production, t/year
COD emissions in effluent, t/year

400 000
2 400 000
740 000
3 600 000
26 000

In this study two scenarios concerning the potential for biogas production are presented: one
moderate and one optimistic. Three parameters are varied in the assumptions for the calculations for
these two scenarios; these parameters and other relevant data are summarised in Table 2, below.
The first one concerns the degree of COD reduction at the current wastewater treatments at the
mills, which is needed to be able to calculate the amount of COD from the process: the digestible
potential. As mentioned, the WWT methods are different at the Swedish mills and thus have
different degrees of COD reduction. According to [12], typical values for COD removal at
biological treatment plants are in the range of 80% to 90% of the COD. In this study, two levels of
reduction are assumed, the same for all mills: 85% in the moderate scenario and 90% in the
optimistic scenario. Secondly it is assumed that not all the effluent streams from the process are
suitable for anaerobic digestion, therefore it is assumed that 90% of the COD is possible to treat
anaerobically in the moderate scenario and in the optimistic scenario the corresponding figure is
95% [19]. The third parameter varied is the anaerobic degradability of COD in the digestion, which
depends on the material treated and the process conditions, such as the loading rate, for the
anaerobic system. According to [4] the anaerobic degradability is in the range of 60-87% for TMP
and somewhat lower for CTM P. In the case study, the anaerobic degradability was 70%, which is
reasonable for a mill with GWP and CTMP. The mechanical pulps produced in Sweden consist of a
large share of TMP though, and therefore the anaerobic degradability could be assumed to be
somewhat higher in average. In this study the degradability is assumed at two different levels, 75%
and 85% for the moderate and optimistic scenarios, respectively. The methane content in the
produced biogas is assumed to be 70% in both scenarios (common contents are in the range of 6575% according to [12]).

61

Table 2. Parameters used in calculations of the potential for biogas production from mechanical
pulp mills.
M oderate
Optimistic
scenario
scenario
a
Total pulp production, t/year
3 600 000
3 600 000a
COD emissions in effluent, t/year
26 000a
26 000a
Total degree of COD reduction in present WWT, %
85
90
Amount COD possible to treat anaerobically, %
90
95
Anaerobic degradability in digester, % of COD
75
85
Specific gas production, Nm3/kg COD removed
0.35b
0.35b
M ethane content of produced gas, %
70
70
3
c
Energy content of produced biogas, kWhLHV/Nm biogas
6.98
6.98c
a

From [18].
In [15, 20 and 21] speci fic gas productions in the range of 0.18 to 0.54 Nm3 /kg COD removed are reported.
c
The energy content of biogas varies with the methane content. The energy content for pure methane is 9.97
kWhLHV/Nm3 [8].
b

4. Results
In this section, the results from the design calculations are given, both for the case study on the
Rottneros mill and for the potential in the moderate and optimistic scenarios, respectively, if
implemented in all Swedish mechanical pulp mills.

4.1. Case study results


The present WWT and the suggested process for the anaerobic wastewater treatment for Rottneros
were described in section 3.1. Given that the pulp production and the amount of COD in the streams
from the process are the same as in the present situation, the biogas production could be calculated
from the COD reduction in digester. This would yield in a biogas production of 27 GWhLHV/year.
As shown in Table 3, further advantages are noted: both the total sludge from the treatment and the
electricity demand is lower in the anaerobic treatment compared to the present, aerobic, treatment.
The COD in the effluent discharge to the recipient is also somewhat lower for the anaerobic case
than for the present WWT.
Table 3. Data for the present situation and the suggested anaerobic digestion at the Rottneros
milla.
Present situation
With anaerobic digestion
Pulp production, t/year
140 000
140 000
COD emissions in effluent, t/year
3 100
3 000
Total degree of COD reduction in WWT, %
83
84
Anaerobic degradability in digester, % of COD
70
Total sludge, t DS/year
15 000
9 000
Electricity demand, GWhLHV/year
11
4.4
3
Biogas production, Nm /year
3 800 000
Biogas production, GWhLHV/year
27
a

All data from [16]

The Rottneros mill today has a sufficient and adequate WWT that is well suited for its production.
If it, however, aims to expand the production at the mill, an anaerobic digestion treatment combined
with the existing one is a promising alternative.

62

4.2. Assessment of the Swedish potential


When considering the entire Swedish mechanical pulping industry and the possible conversion to
anaerobic wastewater treatment, this study uses two scenarios to estimate the potential biogas
production. In the moderate scenario lower values were assumed considering degree of reduction in
the present WWT, amount of COD possible to treat anaerobically and anaerobic degradability in the
anaerobic reactor, than in the optimistic scenario.
To find the potential for producing biogas from the Swedish mechanical pulp mills, the amount of
digestible material in the process water was calculated from the present emissions of COD in the
effluents from the mills, using the assumed degree of COD reduction in the present WWT (see
Table 2). From this material, the amount of COD degraded in the anaerobic reactor and, further, the
amount of biogas produced from this was calculated using the parameters given in Table 2. The
biogas production from Swedish mechanical pulp mills was calculated to 285 GWhLHV/year for the
moderate scenario, and 511 GWhLHV/year for the optimistic scenario.
Table 4. Summary for biogas production from mechanical pulp mills in Sweden
M oderate scenario
Optimistic scenario
COD from process, t/year
173 000
259 000
Amount COD degraded in reactor, t/year
117 000
209 000
Biogas production, Nm3/year
40 800 000
73 300 000
Biogas production, GWhLHV/year
285
511
As was stated in section 3.2, the WWT used presently is mill specific, and thus the two scenarios
reflect the range of reduction degrees that actually exists. Concerning the amount of material
suitable for digestion, the situation will vary between mills, depending on the grades of pulp
produced, whether bleaching is applied or not and other types of process-specific reasons. This
means that some streams may be toxic to the anaerobic bacteria and thus must be treated separately
before taken to the anaerobic stage, if possible to treat anaerobically at all. The third variable that
differs between the two scenarios is the degree of COD reduction in the anaerobic reactor and here
it is assumed that the anaerobic system may be processed at better conditions in the optimistic
scenario, thus giving a higher production rate.

5. Discussion
This study has first investigated the change to anaerobic wastewater treatment in a case study mill
by design calculations, partly as a way of decreasing the energy demand in the WWT system, partly
as a way of producing biogas. Secondly, the potential for biogas production from all Swedish
mechanical pulp mills was estimated, assuming that anaerobic treatment was introduced at these
sites. This estimation was made for two different scenarios, a moderate and an optimistic. The
difference between these scenarios can be found in the assumptions regarding the degree of COD
reduction in the present WWT, amount of COD possible to treat anaerobically and the degree of
COD reduction in the anaerobic reactor. The figures for the moderate scenario reflect the fact that
there might be disturbances in the production during the year, that full load is not always possible
and is overall a more careful estimation. The optimistic scenario assumes that production is more
stable and without disturbances; that a larger share of the streams is possible (or made possible) to
treat anaerobically and that the process improvements that may increase the biogas production are
implemented. These factors would give the possibility for the higher biogas production shown in
the optimistic scenario, however it should not be seen as easily achievable.
The focus of this study is biogas used as vehicle fuel in the transport sector. To be used for this
purpose the raw biogas has to be upgraded to vehicle standard. An upgrading facility might not be
an alternative for a smaller mill because of the rather large investment this would require. If not
being upgraded on-site, a possibility would be to transport the biogas to an external upgrading
63

facility, even though the raw biogas requires large volumes. The biogas could of course also be used
for other purposes, for instance internally in the pulp mill, replacing fossil and/or biomass-based
fuels needed for the production. In [16], the biogas was assumed to replace either oil or biomass
internally in the mill and, as expected, it was shown that it was more economic to replace the oil.
Biogas could also be used for heating, power or combined heat and power, especially use for
heating purposes is common for biogas produced in Sweden today. In 2010, about the same amount
of biogas was used for heating as for upgrading in the Swedish energy system (0.6 TWhLHV) [1].
The size of the anaerobic facility, producing 27 GWhLHV/year, is comparable to other digestion
facilities. The average production for industrial digestion plants and co-digestion plants in Sweden
is 19-23 GWhLHV/year. The average size for a wastewater treatment plant in Sweden is lower, about
5 GWhLHV/year, but the number of plants as well as range of their production is large, for instance
three of the wastewater treatment plants in the Stockholm area produce around 50 GWhLHV/year
respectively [1,22, 23]. It would thus be reasonable to believe that investing in anaerobic
wastewater treatment could be economic for the mill. Considering the production, as well as COD
emissions in effluent, of the Rottneros mill in comparison to other Swedish mechanical pulp mills,
it may be seen that the mill is within the range of an average sized mill [18] and thus possible
anaerobic wastewater treatment facilities at several of the mechanical pulp mills in Sweden should
also be in the same size.
The process for anaerobic treatment assumed to be used in this study is, as in most industrial
digestion facilities in Sweden, of the mesophilic type. However, as has been discussed, the higher
the temperature used, the faster the reaction rate and therefore it is possible to digest at thermophilic
conditions, which even may be preferable at TM P or CTM P mills because of the high temperature
(close to the thermophilic working interval) of the effluent streams from the process [14]. This was
also investigated in [15], where digestion in the thermophilic interval was assumed to be more
energetically favourable, compared to mesophilic digestion, since most process streams from the
pulp mill have rather high temperatures and cooling of the process streams would thus be avoided.
The WWT processes investigated in [15] have similarities to the one used in this study, however
these processes also use different set-ups of membrane filtration in combination with the anaerobic
digestion, as mentioned. Another difference between the processes in [15] and the one here is that
the sludge from the aerobic treatment following the anaerobic reactor is returned to this reactor to
be treated together with the rest of the effluent material. This is an efficient way of reducing the
sludge from the WWT and thus also the need for further sludge treatment. Depending on the
amount of sludge a separate sludge digestion stage might be needed.
The Swedish mechanical pulp mills all have their mill specific conditions and thus the biogas
potential estimated in this study should be considered carefully. To implement a change in the
wastewater treatment such as suggested in this study a thorough evaluation of the mill, and the
specific process used there, would be necessary. Depending on the process type, i.e. whether
bleaching is applied and, if so, which type of bleaching, only few of the effluent streams might be
suitable for anaerobic treatment. However, for a mill that currently has a satisfactory WWT but is
looking to increase their production, investing in an anaerobic treatment could be a promising
option for coping with the increased effluents.
This study only investigated pulp mills producing mechanical pulps. It is also possible to treat some
effluent streams from chemical pulp mills, thus investigating the potential for these mills is also
relevant. The production of chemical pulp in Sweden is about three times as large as the production
of mechanical pulps. Moreover, these types of mills give higher amounts of COD per tonne
produced pulp. However, the situation at the chemical mills is more complex than at mechanical
mills, for instance, the streams from the chemical mills are even more varying and may also contain
other types of chemicals that might disturb the anaerobic process. A potential assessment for biogas
production from the whole pulp and paper industry in Sweden will be made in a coming study.

64

6. Conclusions
Using effluent streams from the mechanical pulping industry in anaerobic digestion to produce
biogas could increase the share of biofuels to the transport sector. The main conclusions of this
study are summarised below:
Using anaerobic wastewater treatment combined with aerobic polishing stage as an alternative
to conventional aerobic wastewater treatment methods at mechanical pulp mills gives a
satisfactory wastewater treatment as well as a substantial contribution of biogas.
When using anaerobic digestion both electricity consumption and sludge production at the mill
are decreased, giving additional environmental advantages.
The production size of an anaerobic wastewater treatment facility at a typical Swedish
mechanical pulp mill is about the same as other industrial digestion plants and anaerobic
wastewater treatment plants and such an investment should therefore have the potential of being
be economic for a mill.
If implemented in the whole Swedish mechanical pulping industry the biogas produced could
amount to 0.3 TWhLHV (moderate scenario) or as much as 0.5 TWhLHV (optimistic scenario)
annually.
With a biogas production from the mechanical pulping industry according to the optimistic
scenario (0.5 TWhLHV/year), the total Swedish biogas production would increase by one third
compared to todays production.

Acknowledgments
This work has been carried out under the auspices of The Energy Systems Programme, which is
primarily financed by the Swedish Energy Agency. The authors would like to thank M ats
Westermark (prof. at the division of Energy Processes) for valuable comments, sa Sivard, Tomas
Ericsson (F) and Nippe Hylander (F and Adj. prof. at the division of Energy Processes) for
contributing with the basis for the potential estimation and for valuable comments throughout the
work and Yi Liu, author of the master thesis on which the case study in this paper is based.

Nomenclature
CCS Carbon capture and storage
CH4 M ethane
CO2 Carbon dioxide
COD Chemical oxygen demand
CTM P Chemi-thermomechanical pulp
DS
Dried substance
GWh Gigawatt hour
GWP Groundwood pulp
kWh Kilowatt hour
LBG Liquid biogas
LHV Lower heating value
Nm3 Normal cubic meter
PGWP Pressurises groundwood pulp
PSA Pressure swing adsorption
t
Tonne
TM P Thermomechanical pulp
TWh Terawatt hour
65

UASB Upflow anaerobic sludge blanket


WWT Wastewater treatment

References
[1]

[2]

[3]

[4]
[5]

[6]

[7]
[8]

[9]

[10]

[11]
[12]

[13]

[14]

[15]

Swedish Energy Agency, Production and use of biogas 2010 (Produktion och anvndning av
biogas 2010) [in Swedish]. Eskilstuna, Sweden: Swedish Energy Agency; 2011 Nov. Report
No: ES 2011:07.
van Lier, J.B., Tilche, A., Ahring, B.K., M acarie, R., M oletta, M ., Dohanyos, L.W., Hulshoff
Pol, L.W., Lens, P., Verstraete, W., New perspectives in anaerobic digestion. Water Science
and Technology 2001;43(1):1-18.
Swedish Forestry Industry Federation. The Swedish Forest Industries 2010. Facts and figures.
Available at <http://www.skogsindustrierna.se/web/Utgivningsar_2011.aspx> [accessed
2011-10-03]
Rintala, J.A., Puhakka, J.A., Anaerobic treatment in pulp and paper-mill waste management: a
review. Bioresource Technology 1993;47:1-18.
Yadvika, Santosh, S., Sreekrishnan, T.R., Kohli, S., Rana, V., Enhancement of biogas
production from solid substrates using different techniques a review. Bioresource
Technology 2004;95:1-10.
Appels, L., Baeyens, J., Degrve, J., Dewil, R., Principles and potential of the anaerobic
digestion of waste-activated sludge. Progress in Energy and Combustion Science
2008;34:755-781.
Weiland, P., Biogas production: current state and perspectives. Applied M icrobiology and
Biotechnology 2010;85:849-860.
Swedish Gas Centre. Biogas folder 2011 Basic data on biogas (Biogas Basdata om biogas
2011) [in Swedish]. Available at
<http://www.sgc.se/uk/display.asp?ID=1352&Typ=Rapport&M enu=Rapporter> [accessed
2011-09-29].
Svensson, M ., Evaluation of Swedish biogas standard Basis for a future revision
(Utvrdering av svensk biogassatandard Underlag fr en framtida revision) [in Swedish].
M alm, Sweden: Swedish Gas Centre; 2011 April. Report No: SGC 229.
Danish Gas Technology Center. Adding gas from biomass to the gas grid. Final report of
Altener, Contract No: XVII/4.1030/Z/99-412. Available at <http://www.dgc.eu/publi2.htm>
[accessed 2012-04-02].
M ohseni, F., M agnusson, M ., Grling, M ., Alvfors, P., Biogas from renewable electricity
Increasing a climate neutral fuel supply. Applied Energy 2012;90:11-16.
EUROPA The official website of the European Union. European Commission, Joint
Research Center, Institute for Prospective Technological studies, Integrated Pollution
Prevention and Control. Draft reference document on best available techniques in the pulp
and paper industry, Draft April 2010. Available at: <http://eippcb.jrc.es/reference/> [accessed
2011-12-20]
Stoica, A., Sandberg, M ., Holby, O., Energy use and recovery strategies within wastewater
treatment and sludge handling at pulp and paper mills. Bioresource Technology 2009;100:
3497-3505.
Salkinoja-Salonen, M ., Apajalahti, J:, Silakoski, L:, Hakulinen, R., Anaerobic fluidised bed
for the purification of effluents from chemical and mechanical pulping. Biotechnology
Advances 1984;2:357-375.
Sivard, ., Ericsson, T., Process integration of wastewater treatment technique decrease of
the energy and resource use in the forestry industry (Processintegration av vattenreningsteknik
66

[16]

[17]
[18]

[19]
[20]
[21]
[22]

[23]

minskning av energi- och resursanvndningen i skogsindustrin) [in Swedish]. Stockholm,


Sweden: F-Industry AB; 2011 Jan. Report for Swedish Energy Agency Project No 32218-1.
Liu, Y., Combination of the membrane method, aerobic and anaerobic wastewater treatment
for mechanical pulping (Kombination av membranmetoden, aerob och anaerob vattenrening
vid mekanisk massaproduktion) [master thesis in Swedish], Stockholm, Sweden: Royal
Institute of Technology (KTH); 2011.
Rottneros. Annual report 2010. Available at: <http://www.rottneros.com/en/ir/annual-reports>
[accessed 2011-12-13].
Swedish Forestry Industry Federation. The environmental database for the Swedish Forestry
Industry Federation Compiled data from all mills in the same table, year 2010. Available at:
<http://miljodatabas.skogsindustrierna.org/si/main/main.aspx?l1=home> [accessed 2012-0109].
Sivard, . Process Consultant, Water and Effluent treatment, F. Private Communication.
Habets, L.H.A, de Vegt, A.L, Anaerobic treatment of bleached TMP and CTMP effluent in
the BIOPAQ UASB system. Water Science and Technology 1991;24(3/4);331-345.
Elliott, A., M ahmood, T., Pretreatment technologies for advancing anaerobic digestion of
pulp and paper biotreatment residues. Water research 2007;41;4273-4286.
Kppala Wastewater Treatment Plant. Kppala Association and the Kppala Wastewater
Treatment Plant, information brochure. Available at:
http://www.kappala.se/default.asp?show=1&varuid=&lid=1&ulid=82 [accessed 2012-02-07].
Stockholm Vatten. Key figures 2001-2010 (Nyckeltal 2001-2010) [in Swedish]. Available at:
http://www.stockholmvatten.se/sv/Om-oss/Broschyrer-rapporter-och-filmer/Rapporter/OmStockholm-Vatten/ [accessed 2012-02-08].

67

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Biogas or Electricity as Vehicle Fuels


Derived from Food Waste
- the Case of Stockholm
Martina Wikstrma, Per Alvforsb
a KTH Royal Institute of Technology, Department of Chemical Engineering and Technology,
Division of Energy Processes, Stockholm, Sweden, marbjor@kth.se
b KTH Royal Institute of Technology, Department of Chemical Engineering and Technology,
Division of Energy Processes, Stockholm, Sweden, alvfors@kth.se
Abstract:
The demand for renewable energy is increasing in Stockholm as well as t he rest of the world. Imperative
factors, such as the need to reduce anthropogenic green house gas emissions and s ecurity of supply, force
this development. The European Commission distinguishes the organic compound in municipal solid waste
as food waste. Food waste may be digested, form biogas and after upgrading, the biogas may be used as
fuel in automotive applications.
This study is based on the food waste pot ential estimations performed by the Stockholm County
Administration Board in the County of Stockholm, both in 2009 and in 2030. The County Administration
Board aim for this food waste to be converted to t he vehicle fuel biogas since this would improve t he share
of renewable trans port fuels and, simultaneously, decrease the green house gas emissions coupled with the
degradation of organic material. In 2009, Stockholm generated 122 000 tonnes of food waste which could
have been converted to 130 GWh biogas. This amount of biogas corresponds to approximately 15 million
litres of petrol. In 2030, the County Administration Board estimates the food waste has increased t o 152 000
tonnes, which c onverted would c orrespond t o 170 GWh biogas. This study will expand the analysis and will
consider the option where the biogas from the food waste is use t o generate electricity to fuel electric
vehicles in Stockholm. In 2009, no large-scale introduction of electric vehicles in Stockholm had begun but
it is vital for decision-makers to assess this option for 2030 in order to obtain a resource and energy efficient
Stockholm.
When considering electricity as vehicle fuel, converting the energy carrier will include additional steps such
as electricity generation, distribution, charging of the vehicle as well as the electric powert rain. The overall
energy efficiency, from biogas to electric propulsion, is in the order of 40 %. E ven though when adding
process steps, which imply losses, the more energy efficient energy carrier is electricity. Converting the
biogas from the food waste to electricity adds approximately another 10 % of driving distance. Assuming an
annual driving distance of 15 000 kilometres, in 2030 this would imply either 27 450 biogas or 27 200 electric
passenger cars in the county of Stockholm. The most resource and energy efficient usage of the biogas from
food waste would be to convert it to electricity for electric vehicles.
Keywords:
Automotive applications, Biogas, Electricity, Food waste, Stockholm

1. Introduction
Today, the global transport sector is heavily dependent on fossil fuels and fossil-originated fuels
accounts for 98 % worldwide [1] and 92 % in Sweden [2]. The demand for renewable energy
carriers is increasing in Stockholm as well as the rest of the world. A transition towards a
renewable-fuelled road transport system is desirable for a number of imperative factors, such as the
need to reduce anthropogenic green house gas emissions and security of supply. One measure to
increase the share of renewable fuels in a region, is to benefit from its generated organic waste,
which may be digested, form biogas and when purified, be used as fuel in automotive applications.
68

The European Commission distinguishes the organic compound in municipal solid waste as food
waste [3].
To investigate the biogas potential derived from food waste in the region, the Stockholm County
Administration Board carried out a regional inventory of available food waste in the Stockholm
County and the results was published in February 2011. The final estimations from this report
constitute the foundations for the calculations carried out in this paper. The analysis in this paper is
expanded to also consider utilisation of produced biogas as fuel for renewable electricity generation
aimed for automotive applications and tank-to-wheel (TTW) analysis including the actual vehicles,
in order to assess the overall energy efficiency of the two vehicle fuels.

2. Objectives and methodology


To assess the potential of biogas production from food waste, the Stockholm County
Administration Board initiated a study in 2010 to evaluate the amount of petrol that could be
replaced by more resource efficiently utilising food waste as an energy carrier. The final report, [4],
considers the option of biogas production utilizing available food waste as raw material and
estimates the production potential in its administrated territory. The biogas is considered a
substitute for fossil fuels in the transport sector. The usage of biogas may be differentiated, and in
the context of increasing the share of renewable fuels in the road transport system, the biogas could
also function as fuel in order to generate renewable electricity for an electrified vehicle fleet.
This paper will consider biogas both as a transportation fuel and also as a renewable resource for
producing electricity to fuel an electrified transport system. This paper will assess each conversion
systems overall energy efficiency and will exemplify this by the means of absolute driving distance
produced by each fuel and the corresponding number of units by each vehicle technology. The aim
with this paper is to determine the most energy efficient way of utilising the biogas that The
Stockholm County Administration Board has acknowledged as possible to produce from the
Countys food waste potential.

1.1 Input data


Input data for this paper will consist of both direct results from the report [4] and additional
estimations and calculations applied onto these reported results. To distinguish the two separate
sources of input data, it has been characterized into two categories:
Direct data reported results from the report [4]
Generated data further work carried out and applied upon the direct data
1.1.1 Direct data
Estimating the available amount of food waste in the Stockholm County, four sources were
considered:
Households
Restaurants
Large-scale catering establishments
Grocery stores
From the theoretical total amount of available food waste, 60 % is considered possible to convert to
biogas [5]. The relationship between the available and the food waste with biogas potential is
illustrated in Figure 1.

69

Figure 1. Estimated amount of food waste available in the Stockholm County and the
corresponding share with biogas potential accounted as thousand tonnes.
The report [4] does not include any considerations regarding the biogas production process. All
process steps are therefore assumed to been accounted for as well as the produced biogas is
assumed to maintain Swedish automotive quality standard, SS 15 54 38.
The report [4] expresses its main result as litres of petrol possible to substitute with the food waste
originated biogas. The fossil fuel substitution potential concluded in the report [4] is complied in
Table 1. To reach these conclusions, assumptions such as energy content and conversion efficiency
has been made but is not included in the report. Relevant thermodynamic assumptions fundamental
for this paper will be presented in section 1.1.2.
Table 1. Fossil fuel substitution potential corresponding the available food waste
2009
2030
Food waste [tonnes]
122 000
152 000
Petrol substituted [litre]
15 000 000
19 000 000
1.1.2 Generated data
According to the EU Directive 2009/30/EC, also referred to as the EU Fuel Quality Directive, the
petrol sold in Sweden constitutes of 5 % ethanol [6]. The E5 petrol has the lower heating value
(LHV) of 8,94 kWh/l [7]. The production capacity of biogas is calculated from the reported amount
of petrol substituted. The produced amount of biogas corresponding to the substituted petrol, see
Table 1, is found in Table 2:
Table 2. Potential biogas production (LHV) originated from food waste in Stockholm County
2009
2030
Biogas [GWh]
134,1
169,9
To illustrate the production of biogas from food waste, Figure 2 is a simplified schematic
illustration of the conversion system.

70

Figure 2. System description adopted from the report [4]

1.2 Description of the expanded system


The expanded system include, in addition to the biogas production in Figure 2, the two considered
options of utilising the renewable resource food waste as vehicle fuel. The produced amount of
biogas (Table 2) is the shared input parameter as energy content. Produced biogas is assumed to
maintain automotive quality. This assumption implies the initial system would include upgrading
steps not necessary prior combustion of biogas. The implications of this assumption are discussed in
section 4 Discussion. The available amount of produced biogas may be utilised as vehicle fuel in
two ways:
Directly as transportation fuel to a biogas vehicle.
As a renewable fuel for electricity production - to an electric vehicle.

Figure 3. System description of two possible vehicle fuel options utilising biogas as a renewable
resource

3. Analysis of conversion routes


This paper will assess the absolute driving distance inflicted by the available biogas when
considering two vehicle technologies, either a biogas or an electric vehicle. As Figure 3 show, there

71

are different parameters that determine each routes energy efficiency, which influence the absolute
driving distance of the vehicle.
To ensure an accurate comparison, a model vehicle is designed using the shared classification
system of the European Commission and Euro NCAP [8] and the vehicle size is assumed to
correspond to a C-segment medium car/ small family car, for example Ford Focus or Opel Astra.
To obtain an additional measure of the total energy efficiency for each transportation fuel, an
annual driving distance for such model vehicle is determined. This paper will assume an annual
driving distance of 15 000 kilometres, which is the statistic distances a Swedish C-segment vehicle
owner drive every year [9]. The total number of each vehicle, respectively, function as an additional
comparison of the overall energy efficiency of the systems.

1.3 Biogas as road transportation fuel


Since the produced biogas is assumed to maintain automotive quality, the fuel economy of the
biogas vehicle alone will determine the driving distance. The losses of biogas while fuelling are
negligible. The energy efficiency of a biogas vehicle powertrain is approximately 20 % [10],
influenced foremost by energy losses in the internal combustion engine (ICE), friction losses
throughout the mechanical powertrain and the total running resistance. The C-segment biogas
vehicle utilise approximately 55 kWh/100 km [11], when operating in urban/semi- urban areas.
In 2030, the overall fuel economy of passenger vehicles is expected to improve in the order of 2530 % [12], [13]. This paper will pessimistically assume a 25 % fuel economy improvement when
assessing the vehicle technologies in year 2030. Table 2 presents the available amount of biogas.
Table 3 show the results of the total number of kilometres generated by the produced amount of
biogas and its corresponding numbers of biogas vehicles with an annual driving distance of 15 000
kilometres.
Table 3. Biogas total propulsion and corresponding numbers of vehicles
2009
2030
Promoted propulsion [100 km]
2 438 181
4 117 818
Numbers of vehicles
16 254
27 452

1.4 Electricity as road transport fuel


The other option considered in this paper is to utilise the produced biogas to generate renewable
electricity for an electrified vehicle fleet, which may also increase the share of renewable fuels in
the road transport system. As seen in Figure 3, there are additional steps (generation, distribution
and fuelling) with coupled energy losses, see Table 4. Biogas may fuel either a gas turbine or an
Otto engine, generating electricity. During assumed prevalent working conditions, approximately
10-14 MWh of biogas, these two energy conversion components acquire similar energy efficiency,
as seen in Table 4. Energy losses during transmission and charging are small but not negligible.
Table 4. Energy efficiency of electricity generation, distribution and charging of vehicle
Energy efficiency [%]
Reference:
Electricity generation
44 %
[14], [15]
Transmission and distribution
92 %
[16]
Conductive charging
99 %
[17]
Converting the energy carrier from biogas to electricity implies energy losses. By converting the
energy input of biogas shown in Table 2, a conductive Electric Vehicle Supply Equipment (EVSE)
may deliver the amount of electricity corresponding to Table 5.

72

Table 5. Potential renewable electricity originated from food waste in Stockholm County
2009
2030
Electricity [GWh]
53,4
67,9
Within the category electric vehicles (EVs), the degree of powertrain electrification may vary from
a conventional mechanical powertrain equipped with a powerful machine working as start engine,
to a fully electrified powertrain. To utilise the generated electricity and charge from the grid, this
paper distinguish between two different types of EVs:
Battery Electric Vehicles (BEV) that have an electric engine and a battery as an energy
storage unit. No internal combustion engine (ICE).
Plug- in Hybrid Electric Vehicles (PHEV) that have both an ICE and an electric engine.
The ICE generates electricity, via a generator mounted upon the ICE, to the electric
powertrain. The energy storage in this case is also a battery.
This paper will only consider BEV since its sole fuel source is electricity, since the degree of
electrical mileage may vary for a PHEV.
The electrical powertrain, compared to a conventional powertrain, is characterised by a high overall
energy efficiency, about 80 % and includes charge/discharge cycle losses of the battery. The
improved energy efficiency is predominantly the result from replacing the ICE by an electric engine
(90 % energy efficiency compare to 30 % as for an ICE) and improvements due to electrical power
transmission instead of mechanical [18]. A C-segment electric vehicle utilise approximately 20
kWh/100 km [19]. Table 6 show the results of the total number of kilometres fuelled by generated
electricity and its corresponding numbers of electric vehicles with an annual driving distance of
15 000 kilometres.
Table 6. Electricity total propulsion and corresponding numbers of vehicles
2009
2030
Promoted propulsion [100 km]
2 682 000
4 529 600
Numbers of vehicles
16 111
27 209

1.5 Food waste as a renewable resource for road transportation fuels


Comparing the promoted driving distance in Table 3 and Table 6, it is shown that the overall energy
efficiency in the case of electricity is 10 % higher then for utilising the biogas directly as a road
transportation fuel.
The results may also be visualised using Sankey diagrams, see Figure 4 and Figure 5. The Sankey
diagram presents an average overview of the energy flows, and is not representative for a specific
time during vehicle operation. The total energy input in both Sankey diagrams is the available
amount of automotive quality biogas presented in Table 2.
The Sankey diagram illustrating the option of biogas as a road transport fuel, Figure 4, uses the
input information given in section 1.3, which describes the energy efficiency of a biogas vehicle.

73

Figure 4. Sankey diagram of energy flows if utilising available biogas as road transport fuel
The Sankey diagram illustrating the option of utilising renewable electricity as road transport fuel,
visualise the additional process steps associated with converting the energy carrier to electricity.
Data given in Table 4 is applied to each step, respectively, to express the coupled energy loss.

Figure 5. Sankey diagram of the energy flows if utilising the available biogas for power generation
in order to obtain renewable electricity for an electrified vehicle fleet.
This paper have made the comparison of the corresponding number of vehicles each fuel option
promote, found in Table 3 and Table 6.In 2009, the number of passenger vehicles in the County of
Stockholm was 800 534 [20]. Assessing the potential impact in fleet composition, the share of
electric vehicles would run up to 2,2 % compared to the corresponding share of biogas vehicles that
may be fuelled by the same resource that would be 2,0 %. Forecasting the vehicle fleet in the
Stockholm County by 2030, it may be assumed that the potential impact might increase hence to
intended governmental efforts to reduce the transport demand in the region [21], [22].

4. Discussion
The measure of energy efficiency used in this paper is total accumulated kilometres and
corresponding number of vehicles caused by each fuel option derived from the equal amount of
food waste originated biogas. In the case of biogas, produced amount of biogas is shown in Table 3.
This amount of biogas is available for the end-consumer. In the case of electricity, the same amount
74

of biogas is combusted, generating electricity, which may be distributed and transmitted to a


charging point. For end-consumer available electricity, in [GWh], is shown in Table 5. Converting
the energy carrier from biogas to electricity results in energy losses, hence the reduced amount of
energy input. When extending the analysis to include the vehicles it is notable how much more
energy efficient the electric vehicle is compared to the biogas vehicle.
Using biogas for automotive applications demands additional purification steps to upgrade the raw
gas obtained when digesting the food waste. Biogas that may be used for electricity generation does
demand the same quality measures in order to operate efficient. This paper assumes automotive
biogas, which in reality would not be necessary for power generation purposes, which implies
additional benefits both in terms of energy efficiency and reduction in physical process step for
electricity as an energy carrier.
To recognise other driving forces then the desire to optimise the resource efficiency of an energy
carrier, it is much more profitable today in Sweden to produce, upgrade and sell automotive quality
biogas compared to retail electricity. In a Swedish context, the most significant incentives,
influencing the overall turnover for the producers are:
Biogas: tax-exempted vehicle fuel [23] and investment grants [24]
Renewable electricity: no energy or carbon dioxide tax1 [23] and production plant may be
rewarded green certificates [25]
Today, the relative benefit of selling automotive biogas is on average approximate 30 % higher,
compared to selling electricity for a biogas producer [26]. This might contribute to decisions,
comprising the energy- and material efficiency of the system, maximising monetary return.
To further extend the energy system analysis of the electricity production, to facilitate heat recovery
may contribute to an even higher overall energy efficiency but also a sales possibility if integrated
with, for example, a district heating system network.

5. Conclusion
Many factors may influence the utilisation of the limited renewable resources available, in this case
food waste. The two options considered in this paper was to digest the food waste to biogas, then
utilising the biogas either as a road transportation fuel or as fuel for renewable electricity
generation. Including the vehicle and its powertrain in the analysis showed that the most energy
efficient way of utilising the renewable resource was to fuel an electrified vehicle fleet. Converting
the biogas from the food waste to electricity adds approximately another 10 % of driving distance.
Assuming an annual driving distance of 15 000 kilometres, in 2030 this would imply either 27 450
biogas or 27 200 electric passenger cars in the county of Stockholm. Regarding food waste as a
renewable resource, of course most desirable would be to minimise this source and instead enjoy
the food before it becomes waste.

6. Acknowledgements
The work has been carried out under the auspices of The Energy Systems Programme primarily
financed by The Swedish Energy Agency.

In 2011, the Swed ish energy and carbon dio xide tax was 0,85 eurocents/kWh and 11,70 eurocents/kg CO2 ,
respectively [27].

75

7. References
[1] Kahn Ribeiro, S., S. Kobayashi, M. Beuthe, J. Gasca, D. Greene, D. S. Lee, Y. Muromachi, P. J.
Newton, S. Plotkin, D. Sperling, R. Wit, P. J. Zhou, 2007: Transport and its infrastructure. In
Climate Change 2007: Mitigation. Contribution of Working Group III to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA
[2] Swedish Energy Agency. Energy Outlook 2010 (Energilget 2010). Report no. ER 2010:45.
[3] European Parliament. Directive 2006/12/EC of the European Parliament and of the Council of 5
April 2006 on waste
[4] Stockholm County Administrative Board. Present and future 2011:2 Food waste provides new
energy (Nutid och framtid 2011:2.Matrester ger ny energi).
Available at:
http://www.lansstyrelsen.se/stockholm/Sv/publikationer/2011/Pages/nof- matrester- ger-nyenergi.aspx accessed: 2011-09-15
[5] Swedish Waste Management Association. Report 2008:02. The Swedish biogas potential from
domestic raw materials (Den svenska biogaspotentialen frn inhemska rvaror). ISSN 1103-4092.
[6] European Parliament. Directive 2009/30/EC of the European Parliament and of the Council of
23 April 2009, amending Directive 98/70/EC as regards the specification of petrol, diesel and gasoil and introducing a mechanism to monitor and reduce greenhouse gas emissions and amending
Council Directive 1999/32/EC as regards the specification of fuel used by inland waterway vessels
and repealing Directive 93/12/EEC.
[7] SPBI. Swedish Petroleum and Biofuel Institute. Fact sheet available at:
http://spbi.se/faktadatabas/artiklar/berakningsmodeller accessed: 2011-09-28
[8] Euro NCAP. Technical report Development of the European new car assessment programme
(EURO NCAP). Available at: http://www.euroncap.com/files/Development-of-Euro-NCAP-1998--0-7498ffa2-36b6-4328-8b6e-f44fe5d35563.pdf accessed: 2011-12-27
[9] Sweden Statistics. Transports and Communications. Driving distance and fuel consumption
(Transporter och Kommunikationer. Krstrckor och brnslefrbrukning). 2009
[10] H Engerer, M Horn. Natural gas vehicles: An option for Europe. Energy Policy 38 (2010)
1017-1029
[11] Joint Research Centre of the European Commission (JRC). Tank-to-Wheel analysis of future
automotive fuels and powertrains in a European context. Version 2c. JRC, EUCAR, CONCAWE.
2006.
available
at:
http://iet.jrc.ec.europa.eu/sites/aboutjec/files/documents/TTW_Report_010307.pdf accessed: 2011-12-30
[12] Duval M. Advanced Batteries for Electric Drive Vehicles A technology and cost-effective
assessment for battery electric vehicles, power assist hybrid electric vehicles and plug- in hybrid
electric vehicles. EPRI. May 2004
[13] S E Plotkin. Examining fuel economy and carbon standards for light vehicles. Energy Policy.
37 (2009) 3843-3853
[14] A Poullikkas. An overview of current and future sustainable gas turbine technologies.
Renewable and Sustainable Energy Reviews 9 (2005) 409-443
[15] M Edstrm, Nordberg. JTI Swedish Institute of Agricultural and Environmental
Engineering. Report no. 107. Producing biogas on the farm (Producera biogas p garden). 2004.
[16] Swedish Energy Agency. Energy in Sweden facts and figures 2010. Report no. ET 2010.46.
2010
[17] Kgeson P. Nature Associates. The climate impact of electrified road transport. Klimateffekten
av elektrifierad vgtrafik. 2010-06-29.

76

[18] Helms H, Pehnt M, Lambrecht U, Liebich A. Electric vehicle and plug- in hybrid energy
efficiency and life cycle emissions. 18th International Symposium Transport and Air Pollution.
Dbendorf. Switzerland. 18-19 May 2010
[19] Duval, M. Advanced Batteries for Electric Drive Vehicles A technology and cost-effective
assessment for battery electric vehicles, power assist hybrid electric vehicles, and plug- in hybrid
electric vehicles. EPRI. May 2004
[20] Sweden Statistics. Vehicle statistics. Vehicles in counties and municipalities 2010/2011.
(Fordonsstatistik, Fordon i ln och kommuner vid rsskiftet 2010/2011). 2011
[21] Office of Regional Planning, Stockholm County Council. Report no. 5:2010. Regional
development plan for the Stockholm region (RUFS 2010 Regional Utvecklingsplan Fr
Stockholmsregionen). 2010
[22] Office of Regional Planning, Stockholm County Council. Energy future of the Stockholm
region 2010-2050: The way to reduce climate impact. 2010. Available at:
http://www.tmr.sll.se/Global/Dokument/publ/2010/2010_r_energy_future_of_the_stockholm_regio
n_2010-2050.pdf accessed: 2011-12-30
[23] Regulation (2010:178) on the taxation of energy (Frordning (2010:178) om skatt p energi).
Stockholm. 2010. (SFS 2010:178)
[24] Regulation (2003:564) on public grants to promote efficient and environmental- friendly energy
supply (Frordning (2003:564) om bidrag till tgrder fr en effektiv och miljanpassad
energifrsrjning). Stockholm. 2003. (SFS 2003:564)
[25] Regulation (2011:1480) on the electricity certificate system (Frordning 2011:1480 om
elcertifikat). Stockholm. 2011. (SFS 2011:1480)
[26] M. Gunnarsson, T Lygnegrd. Swedish Biogas International. stergtland biogas production
(stergtland
biogasring).
2008.
Available
at:
http://www.energiost.se/content/images/energiost/pages/biogasring_ostergotland_081201.pdf
accessed: 2011-12-30
[27] Swedenergy. Available at: http://www.svenskenergi.se/sv/Om-el/Miljo-och-klimat/Mal-ochstyrmedel/ Accessed: 2011-12-29

77

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Compressibility factor as evaluation parameter of


expansion processes in Organic Rankine Cycles
Giovanni Manentea and Andrea Lazzarettob
a

University of Padova, Padova, Italy, giovanni.manente@unipd.it, CA


b
University of Padova, Padova, Italy, andrea.lazzaretto@unipd.it

Abstract:
In the conversion of low temperature heat sources into electricity using an Organic Rankine Cycle system the
working fluid selection is a major design choice to maximize the overall performance. The placement of the
power cycle on a T-s diagram depends on the fluid critical temperature. Several studies have shown that the
power output can be maximized by using fluids with critical temperatures similar or lower than the inlet
temperature of the heat source, which allow a better temperature profile match between the heat source and
the working fluid. However, the choice of a fluid having a specific critical temperature also influences the fluid
properties in the expansion process over a given temperature interval, as shown by the generalized
compressibility chart. The aim of this paper is providing a better insight into the results of optimized ORCs
through the analysis of the compressibility factor in the expansion process.
To this purpose the real enthalpy change in the expansion process i s regarded as two separate terms
associated with temperature and pressure drops, re spectively. Starting from the analysis of different
expansion processes in optimized cycles a correlation is obtained between the compressibility factor and the
ratio between real enthalpy change and the enthalpy change term associated with temperature drop. Thus,
the ratio between the former and the latter can be directly evaluated from the simple knowledge of pressure
and temperature values along the expansion process and the observation of the compressibility chart.

Keywords:
Organic Rankine Cycle, Expansion Process, Real Gas, Compressibility Factor.

1. Introduction
In Organic Rankine Cycle systems low/medium temperature heat sources are used to evaporate a
secondary working fluid with a low boiling point that expands through a turbine to generate power.
The fluid is then discharged to the condenser where the condensing heat is transferred to a cooling
water or cooling air flow. The liquid condensate is pumped into the evaporator completing the
cycle. These systems, also known as binary cycle power plants [1], have found many applications in
the recovery of waste heat from industrial plants and are playing a very important role in the
modern geothermal electricity market [2].
The thermodynamic properties of the working fluid strongly influence the performance of these
systems. In particular, the critical temperature plays a primary role in fixing the placement of the
thermodynamic cycle on a T-s diagram [3]. The results of power-based performance comparisons
show that the best fluids have critical temperatures similar or lower than the inlet temperature of the
heat source ([4], [5]) so that the evaporation occurs at high subcritical reduced pressures (pR=p/pCR)
or even at supercritical pressures that imply a good match between thermal profiles of the heat
source and working fluid. At these conditions the ideal gas equation of state does not provide an
acceptable representation of the p-v-T relationship, since the compressibility factor (Z= pv/RT)
could assume values much lower than 1. An accurate knowledge of the p-v-T behavior is important
in the evaluation of the specific enthalpy change that can be expressed in terms of p,v,T and cP.
The favourable role of gas compressibility effects in reducing the work absorbed by the
compression process in vapor refrigeration and heat pump cycles was studied in [6]. The same real
gas effect is used to improve the conversion efficiency, through the minimization of the
78

compression work, in supercritical closed CO 2 Brayton cycles operating at moderate temperatures


[7]. However, to the authors knowledge, no studies are available about the relationship between the
compressibility factor and the work obtained by the expansion process in organic Rankine cycles,
with the exception of a very recent study [8] which only remarks that working fluids with higher
compressibility factors at the turbine inlet result in higher turbine isentropic work.
In this paper the expansion process of four working fluids (CO 2, R125, R134a and isobutane) in
optimized Organic Rankine Cycles ([9], [10]) is analyzed. The enthalpy change is divided into
many intervals. At each interval the two terms that form the real enthalpy change are calculated. It
is shown that the ratio between real enthalpy change and temperature related enthalpy change
correlates well with the value of the compressibility factor. In particular, at low values of the
compressibility factor (around 0.5-0.6) the real enthalpy change is only a small fraction of the
enthalpy change calculated as the product of the fluid specific heat and the temperature drop. On the
other hand, the generalized compressibility chart links the compressibility factor with reduced
pressure (pR) and reduced temperature (TR=T/TCR). So, it immediately shows thermodynamic states
in the expansion process associated with low values of the compressibility factor, which imply a
marked abatement of the real enthalpy change compared with that calculated from the temperature
drop.
These findings suggest that the working fluid should be selected so that the expansion process
occurs at temperatures and pressures associated with high values of the compressibility factor,
which allow the real enthalpy change to approach the enthalpy change calculated from the
temperature drop.

2. Evaluation of the specific enthalpy change in single-phase


regions
2.1. Ideal gas
For a gas obeying the ideal gas model the specific enthalpy depends only on temperature, so the
specific heat cP, is also a function of temperature alone. That is:
dh
dT

cP T

(1)

Separating variables in (1):


(2)

dh c P T dT

On integration:
h T2

T2

h T1

T1

cP T dT

(3)

When the specific heat is taken as constant (3) reduces to:


h T2
h T1 cP T2 T1
(4)
The constant value of cP is a mean value calculated as follows:
T2

cP

T1

c P T dT

T2

T1

(5)

However, when the variation of cP over a given interval is slight, little error in normally introduced
by taking the specific heat required by (4) as the arithmetic average of the specific heat values at the
two end temperatures. Alternatively the specific heat at the average temperature over the interval
can be used.
79

2.2. Real gas


For a real gas the specific enthalpy can be regarded as a function of temperature and pressure. That
is:
h h T, p
(6)
The differential of this function is:
h
T

dh

h
p

dT
p

(7)

dp
T

It can be shown [11] that (7) can be expressed in the following form:
dh cP dT

v
T

(8)

dp
p

where the right side of (8) depends solely on p,v,T and cP.
Changes in specific enthalpy between two states are found by integrating (8):
h2

h1

cP dT

v T

v
T

(9)

dp
p

Integration of the second term on the right of (9) requires knowledge of the p-v-T behavior at the
states of interest.

2.3. Compressibility factor


The relationship among pressure, specific volume and temperature is expressed by using the
compressibility factor Z:
pv
RT

(10)

The compressibility factor Z tends to unity as pressure tends to zero at fixed temperature. From (10)
the specific volume can be expressed as:
v

ZRT
p

(11)

On differentiation:
v
T

RZ
p

RT
p

Z
T

(12)
p

Using the previous two expressions, the second integrand on the right of (9) becomes:
v T

v
T

ZRT
p

RZ
p

RT
p

Z
T

RT 2
p

Z
T

(13)
p

Thus the change in specific enthalpy between two states can be evaluated by:
h2

h1

2
1

c P dT

2
1

RT 2
p

Z
T

(14)

dp
p

And assuming an infinitesimal change in enthalpy:


80

dh cP dT

RT 2
p

Z
T

dp

(15)

That can be expressed as:


dh dhdT

dhdp

(16)

i.e., the real enthalpy change dh can be regarded as difference of two terms associated with the
temperature drop ( dhdT ) and pressure drop ( dhdp ), respectively. The first term depends on the
specific heat, the second one on pressure and temperature, and on the particular gas constant and
partial derivative of the compressibility factor Z with respect to temperature. Note that the term
dhdT is not the enthalpy change calculated under the ideal gas model, being cP dependent on both
pressure and temperature.

3. Analysis of the expansion process in optimized ORCs


In this section the results of two optimization studies ([9], [10]) of ORCs are analyzed through the
evaluation of the fluid properties along the expansion process. According to (15) the real enthalpy
drop is composed by two components: the first related to the temperature change, the second to the
pressure change. The calculation of these two terms shows that the real enthalpy drop is only a
small fraction of the enthalpy drop related to the temperature change. Different cases are considered
to understand how this fraction depends on the value of the compressibility factor, which in turn
depends on the values of reduced pressure and temperature of the working fluids, as shown by the
generalized compressibility chart. The examples presented in the following sections clearly show
that the working fluid choice fixes the placement of the thermodynamic cycle in the T-s diagram
and consequently the operating region in the compressibility chart.

3.1. Expansion processes analyzed


A brief summary of the thermodynamic optimization studies performed in ([9], [10]) is given to
introduce the expansion processes that are analyzed in the following. Both optimization studies
consider a thermodynamic objective function: the net power output in [9], the exergy recovery
efficiency in [10], the latter being defined as the ratio between net power output and exergy
available from the geothermal fluid. The mass flow rate of the heat source in [10] is 4000 times the
mass flow rate in [9], which leads to net power outputs of around 3-3.5 M W in the first case and
less than 1 kW in the second case. The isentropic efficiency of the turbine is 0.85 in the larger scale
system and it was assumed equal to 0.80 in the smaller scale system. The larger system employs a
dry cooling system with air cooled condensers (the specific consumption of which was assumed to
be 0.15 kW per kg/s of air), whereas the smaller system includes a wet cooling system having a
negligible power absorption.
In [9] the working fluids (CO 2 and R125) operating in transcritical cycles are compared assuming
an inlet temperature of the heat source of 100C. Given the low critical temperatures of both
working fluids, supercritical cycles were obtained. Although the CO 2 turbine generates 28% more
power than the R125 turbine, the R125 cycle results in a net power output 14% higher than the CO 2
cycle due to the high power absorbed by the CO 2 feed pump. In [10] isobutane and R134a are
compared at geothermal fluid inlet temperatures between 130 and 180C. The optimal R134a cycles
are mainly supercritical whereas the isobutane ones are mainly subcritical. The R134a cycles
provide more net power output than isobutane cycles in the whole temperature range considered for
the heat source. At 150C the R134a turbine generates 25% more power than the isobutane turbine,
and the net power generated by the R134a cycle is 17% higher than the isobutane cycle. The 150C
case is analyzed here as representative of all cases presented in [10].
81

Table 1 shows the main working fluid properties and summarizes the optimal cycle parameters for
the four expansion processes analyzed in this study. Note that the fluid critical temperatures vary in
a large interval between 31C and 135C. The critical pressure of CO 2 is almost double the critical
pressure of the other working fluids. In addition, the reduced pressure (pR) at turbine inlet for CO 2 is
the highest (around 1.7) among the fluids considered, which implies much higher evaporating
pressure. The reduced temperature (TR) at turbine inlet varies in the interval between 0.9 for
isobutane and 1.2 for CO2. The reduced pressure at turbine outlet is still high for CO 2 (just below
1), whereas it is much lower for isobutane and R134a. Only CO 2 has a turbine outlet temperature
higher than the critical temperature.
Table 1. Fluid properties and optimal cycle parameters in the considered optimized ORCs.
1
2
3
4
Theat_in (C)
100
100
150
150
mheat_in (kg/s)
0.025
0.025
100
100
Working fluid
CO2
R125
R134
Isobutane
TCR (C)
30.98
66.02
101.06
134.66
pCR (bar)
73.77
36.18
40.59
36.29
Optimal turbine inlet and outlet states
pturb_in (bar)
123.32
44.56
47.97
18.85
tturb_in (C)
92.5
85.0
129.4
98.5
pturb_out (bar)
69.04
16.37
8.323
4.376
tturb_out (C)
47.72
35.95
50.8
49.2
Optimal turbine inlet and outlet reduced states
pR_turb_in
1.67
1.23
1.18
0.52
TR_turb_in
1.20
1.06
1.08
0.91
pR_turb_out
0.94
0.45
0.21
0.12
TR_turb_out
1.06
0.91
0.87
0.79
Power generated and absorbed
Pturb (kW)
0.5863
0.4587
4822.9
3863.1
Ppump (kW)
0.2965
0.128
757.0
311.2
PACC (kW)
/
/
567.8
556.6
Pnet (kW)
0.2898
0.3308
3498.1
2995.3

3.2. Calculation of the real enthalpy change


The EES software [12] is used to implement the equations of the expansion process. EES provides
built-in thermophysical property data for many fluids, included those analyzed in this paper. The
values of the thermodynamic properties calculated using the equations of state implemented in EES
may slightly differ from those calculated using the REFPROP database in the original studies [9]
and [10], shown in Table 1.
The expansion process between turbine inlet and outlet is divided into 100 intervals to approach the
limiting condition of an infinitesimal enthalpy change. The elementary isentropic enthalpy change
is therefore given by:
hIS _ el

hIS
100

(17)

The real elementary enthalpy change is calculated using the isentropic efficiency of the overall
expansion process ( IS = 0.80 for CO2 and R125; IS = 0.85 for isobutane and R134a):
hel
hIS _ el IS
(18)

82

Equation (18) assumes that the polytropic efficiency ( p) is equal to the isentropic efficiency, that
is:
IS

hIS _ el

h IS

(19)

which is quite well satisfied in the expansion processes considered.


For each interval the following parameters are evaluated using EES built-in fluid properties:
Real enthalpy change, calculated from the temperature and pressure of the real gas at the inlet
and outlet of each interval:
h hin hout
(20)
Enthalpy change associated with the temperature drop, where the specific heat is calculated as a
function of temperature and pressure at the inlet of each interval:
h T cP _ in Tin Tout
(21)
Ratio between real enthalpy change and enthalpy change term associated with temperature drop:
R

h
h T

(22)

Enthalpy change associated with pressure drop:


h

RTin2
pin

Z in
T

(23)

p in

Real enthalpy change calculated integrating (15) over an elementary expansion process. This
value is an approximation of the real enthalpy change (20) because finite (and not infinitesimal),
although small, enthalpy changes are considered:
h approx
h T
h p
(24)
Percentage error between the enthalpy change calculated using (24) and the enthalpy change
calculated from (20):
Err %

happrox
h

h
100

(25)

3.3. Evaluation of the enthalpy changes in the considered expansion


processes
The fluid properties in the expansion process and the values of the parameters presented in the
Section 3.2 are shown in Tables 2 to 5. For each fluid two Tables are included. Table a shows the
fluid thermodynamic properties p, T, h, cP, Z and the derivative

Z
T

at inlet and outlet of intervals


p

1, 30, 60, 90 and 100, respectively called I, II, III, IV, V. The thermodynamic conditions at the
outlet from the last interval (turbine outlet) may slightly differ from those in the original works
shown in Table 1 due to the different property methods used and the assumption that the polytropic
efficiency is equal to the isentropic efficiency. Table b shows, in the first three columns, the real
enthalpy change (20), the enthalpy change associated with the temperature drop (21) and their ratio
(22); in the last three columns the enthalpy change associated with the pressure drop calculated
using (23), the approximated real enthalpy change (24) and the percentage error (25).

83

3.3.1. CO2 expansion


In the optimized power cycle [9] CO 2 expands from 123.3 bar to about 69 bar. As shown in Table
2a, the specific heat of CO 2 remains almost constant along the expansion process. The
compressibility factor varies within a narrow range between 0.655 and 0.663, whereas the
derivative Z T p increases from 4.7E-3 to 5.9E-3 from turbine inlet to outlet.
From Table 2b we see that the real enthalpy change over an elementary interval is 0.195 kJ/kg,
whereas the enthalpy change associated with the temperature drop is about 4.5 times higher. The
enthalpy change associated with pressure drop slightly increases from inlet to outlet, which leads to
a slight reduction of the ratio R h from 0.237 to 0.217. The relative error introduced by using the
approximate expression (24) is below 1%.
Figure 1 shows on a T-s diagram the expansion process of CO 2 in the optimized cycle. The
expansion is divided into 10 intervals, having the same enthalpy drop. The values of the
compressibility factor are shown near points 1, 4, 7 and 10. These points correspond to points 1, 31,
61 and 91 using a 100 intervals discretization (i.e. the inlet or outlet states of intervals I, II, III and
IV). The temperature difference between two consecutive points (i.e. the vertical distance) is about
the same along the whole expansion process since the specific heat is almost constant and the ratio
between real enthalpy change and enthalpy change associated with temperature drop varies only
slightly.
Table 2a. Thermodynamic conditions of CO 2 at inlet and outlet of intervals I, II, III, IV and V.
Interval
Points
p (bar)
T (C) h (kJ/kg)
cP (kJ/kg-K)
Z
Z T p (1/K)
I
1
123.3
92.50
-36.13
1.954
0.6630
0.004671
2
122.7
92.08
-36.33
1.955
0.6628
0.004686
II
30
105.3
80.06
-41.79
1.972
0.6576
0.005080
31
104.7
79.62
-41.98
1.973
0.6574
0.005093
III
60
88.83
66.79
-47.64
1.979
0.6557
0.005468
61
88.32
66.34
-47.83
1.979
0.6557
0.005480
IV
90
74.41
53.23
-53.49
1.973
0.6573
0.005807
91
73.97
52.78
-53.68
1.973
0.6574
0.005817
V
100
70.04
48.67
-55.44
1.969
0.6586
0.005906
101
69.62
48.22
-55.63
1.968
0.6588
0.005916
Table 2b. Enthalpy changes of CO 2 calculated using Equations 21 to 26.
Interval
h (kJ/kg)
h T (kJ/kg)
h h T
h p (kJ/kg)
h
I
0.1950
0.8237
0.2368
0.6269
II
0.1950
0.8608
0.2266
0.6639
III
0.1950
0.8865
0.2200
0.6896
IV
0.1950
0.8983
0.2171
0.7014
V
0.1950
0.8989
0.2170
0.7020

84

- h p (kJ/kg)
0.1968
0.1969
0.1969
0.1969
0.1969

Err%
0.91
0.94
0.96
0.97
0.96

CarbonDio xide

125

140 bar

80 bar

100
Z=0.6630
Z=0.6574

T [C]

75

Z=0.6557

120 bar

Z=0.6575

100 bar

50

25

60 bar
40 bar

0
-1.75

-1.50

-1.25

-1.00

-0.75

-0.50

s [kJ/ kg-K]

Fig. 1. Turbine expansion of CO 2 in the optimized cycle represented on a T-s diagram. Note that
the values of the compressibility factor Z are almost constant along the expansion process.

3.3.2. R125 expansion


In the optimized power cycle [9] R125 expands from 44.5 to 16.5 bar. As shown in Table 3a the
specific heat markedly changes during the expansion process, almost halving from turbine inlet to
outlet. The compressibility factor is low, around 0.50, at turbine inlet and gradually increases up to
0.72. The derivative Z T p decreases from 1.26E-2 to 5.3E-3.
The fourth column in Table 3b shows that at the beginning of the expansion process the real
enthalpy change is less than 10% of the enthalpy change calculated from the temperature drop
( h T). As the expansion progresses the enthalpy term associated with pressure drop ( h p)
gradually decreases so that the enthalpy drop ratio ( h h T) reaches the value of 0.265. The
relative error introduced by using (24) is variable, however in any case below 8%. This error is the
highest among the cases considered and could be reduced using a discretization in a higher number
of intervals, however it was considered acceptable for this analysis.
Figure 2 shows on a T-s diagram the expansion process of R125 in the optimized cycle. As already
done for CO2 in Fig. 1, the expansion line is divided into 10 intervals having the same enthalpy
drop. The temperature difference between two consecutive points (i.e. the vertical distance) is lower
at the end of the expansion than at the beginning. This is easily explained since the effect due to the
decrease of the specific heat from turbine inlet to outlet is overcome by the increase of the ratio
between real enthalpy change and enthalpy change associated with temperature drop.
Table 3a. Thermodynamic conditions of R125 at inlet and outlet of intervals I, II, III, IV and V.
Interval Points p (bar)
t (C)
h (kJ/kg)
cP (kJ/kg-K)
Z
Z T p (1/K)
I
II
III
IV
V

1
2
30
31
60
61
90
91
100
101

44.56
44.06
32.56
32.22
24.16
23.93
18.28
18.12
16.71
16.56

85.0
84.39
67.84
67.28
52.66
52.20
40.17
39.79
36.47
36.12

360.7
360.6
357.6
357.4
354.3
354.2
351.0
350.9
349.9
349.8
85

2.231
2.223
1.817
1.801
1.431
1.421
1.207
1.202
1.156
1.151

0.4959
0.4978
0.5618
0.5642
0.6346
0.6370
0.6983
0.7002
0.7170
0.7188

0.01265
0.01267
0.01089
0.01078
0.007881
0.007795
0.005799
0.005744
0.005282
0.005234

Table 3b. Enthalpy changes of R125 calculated using Equations 21 to 26.


Interval
h (kJ/kg)
h T (kJ/kg)
h h T
h p (kJ/kg)
h
I
0.1097
1.363
0.08047
1.251
II
0.1097
1.007
0.1089
0.9037
III
0.1097
0.6526
0.1680
0.5515
IV
0.1097
0.4576
0.2396
0.3555
V
0.1097
0.4140
0.2649
0.3114

- h p (kJ/kg)
0.1116
0.1030
0.1011
0.1021
0.1025

Err%
1.8
6.1
7.8
6.9
6.5

R125

125

50 bar

100

40 bar

Z=0.4959

T [C]

75
Z=0.5642
30 bar

Z=0.6370

50
20 bar

Z=0.7004

15 bar

25

10 bar

0
1.00

1.20

1.40

1.60

1.80

s [kJ/kg-K]

Fig. 2. Turbine expansion of R125 in the optimized cycle represented on a T-s diagram. Note the
low value of the compressibility factor Z at turbine inlet.

3.3.3. R134a expansion


In the optimized power cycle [10] R134a expands from 48.0 to 8.5 bar. As in the previous case, the
specific heat (Table 4a) decreases along the expansion line. At the beginning of the expansion
process the compressibility factor is low, below 0.6, and gradually increases to around 0.85. The
derivative Z T p decreases by a factor 3.7 from inlet to outlet.
The results reported in Table 4b show that at the beginning of the expansion the real enthalpy
change is only 0.15 times the enthalpy change calculated as the product of the specific heat and the
temperature drop. Then it rapidly increases up to around 0.54 times in the final stage. The relative
error introduced by using (24) decreases from about 5% to values well below 1%, and it was
considered acceptable for this analysis.
As already seen for the other working fluids, Fig. 3 shows the expansion process of R134a on a T-s
diagram where the expansion line is divided into 10 intervals of equal enthalpy change. At the
beginning of the expansion the real enthalpy drop is only a small fraction of the enthalpy change
calculated from the temperature drop, due to the high values of the enthalpy change associated with
pressure drop shown in the fifth column of Table 4b. Thus, despite the higher specific heat values, a
higher temperature change is needed in the first stages compared to the final stages, as clearly
shown by the vertical distance between consecutive points in Fig. 3. The points get closer as the real
gas approaches the ideal gas conditions.

86

Table 4a. Thermodynamic conditions of R134a at inlet and outlet of intervals I, II, III, IV and V.
Interval
Points
p (bar)
t (C)
h (kJ/kg)
cP (kJ/kg-K)
Z
Z T p (1/K)
I
II
III
IV
V

1
2
30
31
60
61
90
91
100
101

47.97
47.05
28.05
27.56
16.76
16.48
10.21
10.04
8.669
8.529

129.4
128.3
100.2
99.28
76.62
75.92
57.44
56.85
51.67
51.11

318.3
318.0
309.2
308.9
299.8
299.4
290.3
290.0
287.2
286.8

1.868
1.852
1.440
1.429
1.197
1.192
1.064
1.061
1.032
1.029

0.5837
0.5871
0.6875
0.6909
0.7766
0.7791
0.8407
0.8425
0.8577
0.8593

Table 4b. Enthalpy changes of R134a calculated using Equations 21 to 26.


Interval
h (kJ/kg)
h T (kJ/kg)
h h T
h p (kJ/kg)
h
I
0.315
2.084
0.1511
1.755
II
0.315
1.273
0.2475
0.9538
III
0.315
0.8339
0.3777
0.5171
IV
0.315
0.6262
0.5030
0.3100
V
0.315
0.5817
0.5415
0.2657

0.006943
0.006891
0.004788
0.004717
0.003125
0.003084
0.002145
0.002120
0.001908
0.001886

- h p (kJ/kg)
0.3295
0.3190
0.3168
0.3161
0.3160

Err%
4.6%
1.3%
0.6%
0.4%
0.3%

R134a

150

55 bar

Z=0.5837

45 ba r

100

Z=0.6911

T [C]

35 bar
25 bar

15 bar

50

Z=0.7794
Z=0.8428

8 bar

0
0.25

0.50

0.75

1.00

1.25

1.50

s [kJ/kg-K]

Fig. 3. Turbine expansion of R134a in the optimized cycle represented on a T-s diagram. Note the
substantial increase of the compressibility factor Z from turbine inlet to turbine outlet.

3.3.4. Isobutane expansion


In the optimized power cycle [10] isobutane expands at subcritical pressures from 18.9 to 4.4 bar.
The fluid thermodynamic properties are shown in Tables 5a and 5b. The specific heat markedly
decreases from inlet to outlet conditions. As the expansion progresses, the compressibility factor
increases from 0.675 to 0.90, while the derivative Z T p becomes around 5 times smaller. At the
beginning of the expansion the real enthalpy change is 0.23 times the enthalpy change associated
with the temperature drop, then it rapidly increases reaching the value of 0.68 in the final stage. The
87

percentage error introduced by using (24) is in all cases well below 1%. According to the procedure
followed in the previous cases, Fig. 4 shows on a T-s diagram the expansion process of isobutane
divided into 10 identical enthalpy intervals. The values of the compressibility factor are reported
near the inlet points of intervals 1, 4, 7 and 10. The points get closer toward the end of the
expansion process: the effect due to the increase of the ratio h h T is stronger than the decrease of
the specific heat. In the final stages the real enthalpy drop is more than 65% the enthalpy drop at
ideal gas conditions, which is the highest ratio among all cases considered in this study.
Table 5a. Thermodynamic conditions of isobutane at inlet and outlet of intervals I, II, III, IV and V.
Interval
Points
p (bar)
t (C)
h (kJ/kg)
cP (kJ/kg-K)
Z
Z T p (1/K)
I
II
III
IV
V

1
2
30
31
60
61
90
91
100
101

18.85
18.55
12.09
11.92
7.864
7.754
5.169
5.098
4.499
4.437

98.50
97.75
80.29
79.76
65.84
65.40
53.48
53.09
49.64
49.26

679.0
678.5
664.7
664.2
650.0
649.5
635.2
634.7
630.3
629.8

2.842
2.812
2.324
2.314
2.090
2.084
1.948
1.944
1.911
1.907

0.6754
0.6799
0.7790
0.7817
0.8461
0.8479
0.8909
0.8921
0.9025
0.9035

0.005170
0.005043
0.002874
0.002827
0.001839
0.001815
0.001260
0.001245
0.001121
0.001108

Table 5b. Enthalpy changes of isobutane calculated using Equations 21 to 26.


Interval
h (kJ/kg)
h T (kJ/kg)
h h T
h p (kJ/kg)
h T - h p (kJ/kg)
I
0.4917
2.121
0.2318
1.629
0.4923
II
0.4917
1.241
0.3961
0.7485
0.4927
III
0.4917
0.9156
0.5370
0.4231
0.4925
IV
0.4917
0.7579
0.6488
0.2655
0.4924
V
0.4917
0.7227
0.6803
0.2303
0.4924

Err%
0.12
0.20
0.17
0.15
0.14

Isobutane

150

40 bar
30 bar
20 bar

100

T [C]

15 bar
10 bar

Z=0.6754
Z=0.7818
Z=0.8480
Z=0.8922

50

3 bar

0
1.0

1.5

2.0

2.5

3.0

3.5

s [kJ/kg-K]

Fig. 4. Turbine expansion of isobutane in the optimized cycle represented on a T-s diagram. The
values of the compressibility factor Z show that ideal gas conditions are nearly approached at
turbine outlet.
88

4. Influence of the compressibility factor in the evaluation of


the enthalpy change
The results presented in Section 3 have shown that the real enthalpy change is much lower than the
enthalpy change calculated from the product of the specific heat and the temperature drop. While
this difference is almost constant for CO2, it gradually decreases from turbine inlet to turbine outlet
for the other working fluids. The compressibility factor accounts for real gas effects and shows a
similar trend of variation: it is roughly constant for CO 2 whereas it gradually increases from turbine
inlet to turbine outlet conditions for the other working fluids. Starting from these findings, the ratio
R h between real enthalpy change and enthalpy change associated with the temperature drop is
correlated in the following sections with the compressibility factor.

4.1. Correlation between enthalpy change and compressibility factor

h/ h

Figure 5 shows the ratio (22) between real enthalpy change and the term of the enthalpy change
associated with temperature drop as a function of the compressibility factor Z. A good correlation is
found for each working fluid. As the compressibility factor increases the real gas approaches the
ideal gas behavior and the real enthalpy change approaches the enthalpy change calculated as the
product of the specific heat and temperature drop (15, 16). Thus, the real gas effects associated with
the second term on the right side of (15) can be easily correlated with the compressibility factor Z.
These effects result in a decrease of the enthalpy change from the value calculated as the product of
the specific heat and the temperature drop. Higher values of Z result in higher fractions of
temperature related enthalpy change converted into real enthalpy change.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0

CO2
R125
R134a
Isobutane

0.4

0.5

0.6

0.7

0.8

0.9

Compressibility factor Z

Fig. 5. Correlation obtained between compressibility factor Z and the ratio between real enthalpy
change and enthalpy change evaluated from the temperature drop.
According to these findings, the real enthalpy change in an elementary expansion interval can be
expressed as:
h cP _ in Tin Tout f ( Z )
(26)
where cP_in is the specific heat of the working fluid at the inlet of the temperature interval (Tin-Tout)
and f(Z) is a function that depends on the compressibility factor Z.
The functions h / h T obtained from the analysis of the considered expansion processes (see
Section 3) undergone by different fluids, are shown in Fig. 5. These functions are interpolated using
89

the second order polynomials shown in Table 6. The expression of f(Z) for CO2 is not shown since
the expansion process of CO 2 spans a narrow range (almost constant) of Z. It is noteworthy that
h / h T is a second order function of Z.
Table 6. Correlations between real enthalpy change and enthalpy change calculated from the
temperature drop.
Working fluid
f(Z)= h h T
R125
f ( Z ) 2 .4454 Z 2 2.1233 Z 0.5304
R134a
f ( Z ) 2 .9650 Z 2 2.8602 Z 0 .8115
Isobutane
f ( Z ) 3.2149 Z 2 3 .1110 Z 0.8682
The turbine power output is:
Wt

mwf

h
n

mwf

cP _ in Tin

Tout

f (Z )

(27)

where h is given by (26) and the sum is over all the temperature intervals (n).
The options to maximize the turbine power output consist in increasing all factors included in (27),
that is: working fluid flow rate, temperature drop between turbine inlet and outlet, specific heat and
Z along the expansion process. Accordingly:
A good match between the thermal profiles of the heat source and the working fluid leads to both
higher working fluid mass flow rates and to higher turbine inlet temperatures (close to the inlet
temperature of the heat source).
The use of a more effective cooling system allows the expansion to progress to lower
temperature levels, which are also more effective in generating power due to the higher values of
Z.
The reduced pressures (pR) and reduced temperatures (TR) involved in the expansion process
depend on the working fluid selection and fix the value of Z and f(Z).
Higher values of the specific heat increase the enthalpy drop ( h T ), however they might result
in lower working fluid mass flow rates evaporated by heat transfer with the heat source as well.
All these strategies contribute to the increase of turbine power output, but they do not affect the
above conclusion that operating conditions involving high values of the compressibility factor are
always suitable to maximize the enthalpy drop and, in turn, the turbine power output, according to
(27).

4.2. Use of the generalized compressibility chart in the working fluid


selection
Figure 6 shows the generalized compressibility chart in which the turbine inlet and outlet
thermodynamic conditions of the considered expansion processes are superimposed. The choice of
the working fluid and in turn the choice of the critical temperature fixes the operating conditions in
terms of reduced pressure and reduced temperature in the expansion process. In the 100C
application the expansion process involves only supercritical states using CO 2, whereas it
encompasses both supercritical and subcritical states in the case of R125. In the 150C application
the expansion process encompasses both supercritical and subcritical pressures using R134a,
whereas it only involves subcritical pressures in the case of isobutane. These operating regions are
associated with different values of the compressibility factor which, according to Fig. 5, imply
different fractions between real enthalpy change and enthalpy change associated with the
temperature drop. The compressibility factor becomes therefore an evaluation parameter of the
quality of the expansion process.
90

Fig. 6. Generalized compressibility chart with superimposed the turbine inlet and outlet conditions
of the analyzed expansion processes.

5. Conclusions
This study analyzes the influence of the real gas effects in the calculation of the enthalpy change in
an expansion process. The enthalpy change is regarded as the difference between two terms related
with temperature and pressure drops, respectively. The study shows that the ratio between the real
enthalpy change and the enthalpy change calculated as the product of the specific heat and the
temperature drop can be easily correlated with the compressibility factor Z using a second order
polynomial. As the compressibility factor increases, the real gas approaches the ideal gas behavior
and the enthalpy term related with pressure drop gradually decreases.
The choice of the working fluid in an Organic Rankine cycle fixes the thermodynamic states in the
expansion process in terms of reduced pressure and reduced temperature and, in turn, the values of
the compressibility factor from the generalized compressibility chart. Thus, working fluids and
operating conditions which lead to high values of the compressibility factor are always suitable to
maximize the turbine power output.
This condition does not guarantee by itself the achievement of the highest turbine power output,
which is the result of many concurring factors. Among them the optimal match between thermal
profiles of heat source and working fluid, the work minimization in the compression process and
features of the cooling system are the most important ones, and may force the expansion process
where the compressibility factor is relatively low. Nevertheless, the information that can be inferred
from this study may help both in the preliminary choice of the operating fluid and in the subsequent
analysis of the ORC system performance.

Nomenclature
cP
Err%
h
heat
m

ORC
p

specific heat, kJ/kg-K


percentage error
specific enthalpy, kJ/kg
heat source
mass flow rate, kg/s
Organic Rankine Cycle
pressure, bar
91

R
R
s
t
T
v
Z

specific gas constant, kJ/kg-K


ratio between enthalpy changes
specific entropy, kJ/kg-K
temperature, C
temperature, K
specific volume, m3/kg
compressibility factor

Greek symbols
efficiency
Subscripts and superscripts
ACC
air cooled condenser
approx approximated
CR
critical
dep
departure
t
related to temperature drop
p
related to pressure drop
el
elementary
idg
ideal gas
in
inlet
IS
isentropic
net
net
out
outlet
p
polytropic
pump
feed pump
R
reduced
rg
real gas
turb
turbine
wf
working fluid

References
[1] DiPippo R., Geothermal power plants: principles, applications, case studies and environmental
[2]
[3]
[4]

[5]

[6]

impact. Butterworth-Heinemann, 2008, Burlington, MA, USA, Chap. 8.


Bertani R., Geothermal power generation in the world 2005-2010 update report. Geothermics
2012; 41:1-29.
Angelino G., Colonna di Paliano P., M ulticomponent working fluids for Organic Rankine
Cycles (ORCs). Energy 1998; 23:449-463.
Shengjun Z., Huaixin W., Tao G., Performance comparison and parametric optimization of
subcritical Organic Rankine Cycle (ORC) and transcritical power cycle system for lowtemperature geothermal power generation. Applied Energy 2011;88:2740-2754.
Dai Y., Wang J., Gao L., Parametric optimization and comparative study of organic Rankine
cycle (ORC) for low grade waste heat recovery. Energy Conversion and M anagement
2009;50:576-582.
De Monte F., Analysis of the gas compressibility effects on the constant-entropy reversible
processes of refrigerants and refrigerant mixtures. International Journal of Refrigeration
2002;25:765-779.
92

[7] Utamura M ., Tamaura Y., A solar gas turbine cycle with super-critical carbon dioxide as a

working fluid. Proceedings of ASM E Turbo Expo 2006.


[8] Stijepovic M .Z., Linke P., Papadopoulos A.I., Grujic A.S., On the role of working fluid
properties in Organic Rankine Cycle performance. Applied Thermal Engineering 2012;36:406413.
[9] Baik Y.-J., Kim M ., Chang K.C., Kim S.J., Power-based performance comparison between
carbon dioxide and R125 transcritical cycles for a low-grade heat source. Applied Energy
2011;88: 892-898.
[10] Toffolo A., Lazzaretto A., M anente G., Rossi N., 2010, Synthesis/Design Optimization of
Organic Rankine Cycles for Low Temperature Geothermal Sources with the HEATSEP
M ethod. ECOS2010: Proceedings of the 23rd International Conference on Efficiency, Cost,
Optimization, Simulation and Environmental Impact of Energy Systems; 2010 June 14-17;
Lausanne, Switzerland.
[11] M oran M.J., Shapiro H.N., Fundamentals of engineering thermodynamics. Wiley 2006, Fifth
Edition.
[12] S.A.Klein, Engineering Equation Solver, F-Chart Software, M iddleton, WI, 2010.

93

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Design of Solar Heating System for Methane Generation


L. M. Gutirrez Castroa, P. Quinto Diez a, L. R. Tovar Glvezb
aSeccin de Estudios de Posgrado e Investigacin, Escuela Superior de Ingeniera Mecnica y Elctrica,
Av. IPN s/n, UPALM Edif.5, 3er.Piso, C.P. 07738 Mxico, D.F. Instituto Politcnico Nacional, moni-80lgc@hotmail.com, pqd510@hotmail.com
bCentro Interdisciplinario de Investigaciones y Estudios sobre Medio Ambiente y Desarrollo, Calle 30 de
Junio de 1520 s/n, Barrio. la Laguna Ticomn, C.P. 07340 Mxico, D.F. Instituto Politcnico Nacional.
luisraulto@gmail.com

Abstract:
In this paper the methodology carried out to design a solar heating system (SHS) is shown, it applies in a
thermophilic process of anaerobic digestion for a new prototype of digester. The thermophilic process
required to keep the mix into the digester between temperatures of 40 to 60C. The optimal temperature to
this process is 55C. The SHS was designed with the f method and with the help of the selection program
Solar carried out at the Instituto Politcnico Nacional. To meet the application needed, the SHS is
integrated by flat plate solar collectors of parallel flow with a total collecting area of 12 m2. In this project a
new design of digester was needed. The most important parameters of design are rate and length of stirring
and hydraulic retention time (HRT). Taking in account these parameters, the digester designed has a volume
of 2 m3, the feedstock, which is the organic fraction (OF) sorted at the source of the municipal solid waste
(MSW) of Mexico City, is fed with centrifugal pumps and a heating exchange system was designed to
achieve a more homogenous heat transmission in the digester. As a calculation tool, a computer program
was developed in order to obtain the amount of energy required to several volumes of mix to heat into the
thermophilic range. Improving the design and operation of an anaerobic digester can be result of an organic
waste management more controlled and obtain the best design of anaerobic digester to applying it in rural
zones. The SHS reduces fuel consumption in the process of methane generation. The advantages of this
design are: reducing the time of methane generation and, thereby, improving the efficiency of the process,
also reducing the solid retention time inside the digester and reducing its volume. Among the important
achievements of this new design are: the deviation from landfilling of the OF since Mexico City, nowadays
generates about 5,000 ton/day of it.

Keywords:
Anaerobic digestion, Digester design, Methane generation, Solar heating system, Thermophilic
process.

1. Introduction
Among the most pressing global environmental challenges that faces humankind are the increase in
emission of greenhouse gases causing the climatic change and the exponential generation of waste.
In M exico the generation of solid municipal solid waste (M SW) is calculated in about 85,000 tons
per day. Nowadays, the problematic has advanced, overcoming the common methods of
recollection and its final disposal, in the specific case of M xico City (with a population of 8.9
millions of people), it generates about 11,000 tons per day of M SW [16]. It is known that 40% of
M SW correspond to the organic fraction (OF) of this, therefore only in M exico City, about 5,000
tons are generated on a daily basis, with a moisture content between 70% and 80%, which
potentially could produces about 385,000 Nm3 of methane (CH4) per day. For the above, and since
there is no treatment or method to avoid this methane to be released into the atmosphere, it is
contributing to pollution, as methane is a powerful greenhouse gas [17].
The trouble of handling the M SW in M exico has considerably increased due to the closure of an
important landfill (Bordo Poniente stage IV). The Bordo Poniente stage IV received 4872,300 tons
of M SW in 2009 and a slightly higher quantity in 2010. At the time of the closure (December 31,
94

2011), it ended up with 69 million tonnes buried under an area of 375 ha, its emissions to the
atmosphere are of about 1068,384 m3 of biogas per day [17].
Due to the dimensions of the problem, in this work an improvement to the method of anaerobic
digestion is proposed using solar energy to work under the thermophilic range. The anaerobic
digestion is a process to generate energy, which is carried out by managing the organic fraction of
the M SW to obtain a stabilized product and thus, produce biogas. The biogas is made of methane
and carbon dioxide (50 70% v/v), which can be used as fuel. The effectiveness in destructing
pathogens, as well as higher production in biogas production, depends on the temperature range, in
this case the thermophilic range (55 C) that is more efficient than the mesophilic range (35 C),
[14],[7].
The thermophilic range needs an external energy source that in many cases is a conventional fuel
which results in higher operation costs and does not avoid the problem of greenhouse gases emitted
to the atmosphere, then using solar energy to heat the digester is a good alternative, because these
sources are environmental friendly [18]. Therefore, the contribution of this work, is the design of a
solar heating system (SHS) to provide the energy needed to maintain the thermophilic range in the
process and to reduce costs. The obtained result was a new model of digester adapted with a heat
transfer system to use the SHS and the methane produced. It was designed as well; a computer
program to design digesters that calculates the energy needed to the process depending on the M SW
quantity to convert it in methane.

2. Proceeding
According to El-M ashad et al. (2003) the use of solar energy for heating anaerobic reactor represent
a kind of solar energy storage in the form of biogas. Hamed M . El-M ashad et al.(2002), present a
design of a stirred tank reactor for anaerobic treatment of liquid cow manure under thermophilic
conditions (50 C), using a solar heating system mounted on the reactor roof, in this case a
simulation models for two systems were developed.
In the present study, the solar thermophilic anaerobic digester (STAD) designed has been evaluated
for different configurations the dimensions used a computer program. The STAD designed to obtain
variable conditions which was used in the program is compound by 6 m2 of solar collector area. The
digester designed has a volume of 2 m3; the feedstock is the organic fraction (OF) at thermophilic
conditions (55 C). The main objective was to design a STAD whit high energy efficiency
including lower constructions costs, as well as a simple control system.
First, the design of solar heat transfer system was made. Second, the design of the digester was
modified to maintain the temperature of thermophilic process about 55 C. In the design all
parameter was took in a count, such as the feed, the stirring and heating. Therefore, a new design of
the digester was the result of analysis. Details of the procedure are below.

3. Design Methodology
3.1. Solar Heat Transfer System
The importance of the design is maintaining the mixture into the digester at 55 C, so that several
compounds are involved in the process to achieve it, [7]
Based on the energys requirements of the thermophilic process to generate biogas, the solar heating
system (SHS), with flat plate solar collectors was designed to apply to the Solar program [1]. This
program was designed for flat plate solar collectors; it gives the solar collection area and storage of
water per area of collection [10]. The solar collection area depends on the monthly radiation, on the
installation zone of the SHS and also on the monthly heating loads needed.
95

The available measured data for solar energy were the monthly average daily solar radiation on a
horizontal surface and the monthly average daily hours of bright sunshine [4].
The total solar radiation incident on a horizontal plane at the surface is the sum of direct and diffuse
radiation. Also, it is necessary to determine the solar radiation on tilted surfaces such as flat solar
collectors. Studies on the availability of solar resources in these FPSC have shown that the most
appropriate orientation is north - south, which means that, if the FPSC are to be installed in the
northern hemisphere, the collector should be oriented towards the geographical south and vice
versa, [12]. The radiation is calculated by (1):
cos
cos

Rb

(1)
Z

To determine the total solar radiation on a tilted surface, based on solar radiation data on horizontal
plane, it is necessary to define the coefficient R, which is the ratio of total radiation on the tilted
surfaces in relation to radiation total surfaces in the horizontal plane it is shown in (2):
R

HT
H

(2)

Therefore:
HT

R H

(3)

It can also be expressed as follows:


R

Hb
R
H b

Hd
R
H d

(4)

The total solar radiation on a horizontal plane (H) was measured for this study. Data measured are
shown in the table 1 (see Appendix A, table A.1). These values were plotted as it shows in the Fig.
1.
The method used to both design and selecting is the f design method.
The f chart-method or f design method is used to estimate the annual thermic performance of
the heating solar system activity by using a working fluid, by means of the calculus of the f energy
fraction, supplied by the sun to meet the required heating load. The main design variable is the
uptake area of solar energy. Among the secondary variables are: the type of collector, the storage
capacity, the flow rate and the size of the heat exchangers. The f fraction is the relation between the
useful energy delivered to the solar system (Qt), which is the difference between the energy of the
system only using conventional fuel (Laux ) and the heating load required by the system (L) [12]. For
a given month, the reduction of the f fraction of the supplied solar energy is shown in (5):
f=

Laux
L

Qt
L

(5)

The f fraction is the correlation obtained from results of hundreds of experiments in this area as
well as simulations with solar heating systems operating under different circumstances. This
fraction depends of two dimensionless parameters. The first one is related to the index of losses of
heating load, UL, of the collector and the second one with the solar energy absorbed, G, and the
heating load. M eteorological data of the zone are also required on a monthly basis.
The calculus of the f fraction is in function of two dimensionless parameters: X and Y. The
procedure which Klein describes to determine these two variables is shown by the equations (6) and
(7).

96

Fig. 1. Solar radiation on the study area from 2008 to 2010 [CINVESTAV, 2010]

FR AC
L
FR AC
L

IT

dt

FR AC
H
L

U L Tref

Ta dt

FR AC
U L Tref
L

(6)

Ta

(7)

The value f was obtained as a function of X and Y as it shows in the Fig. 2, [12].
The annual fraction of energy is provided for the solar heating system. This is the sum of
contribution monthly that brings the system, divided by the annual thermal load of heating as it
shown in equation (8):
12

f annual

i 1

Qu

(8)

Lannual

By means of experimental assays, design ranges of the variables used in the development of the f
method for air and liquid systems have been obtained, which are the transmittance and absorbtance
product (0.6 to 0.9), the collector-heat exchanger efficiency factor, (5 to 120 m2), and the overall
coefficient of heat losses, UL (2.1 to 8.3 W/m2-C) [12].
The collectors heat transference efficiency factor or removed heat factor, FR, is defined as the
relation between the actual energy available for the collector and the energy that would be available
if the whole absorber plate of the flat solar collectors, FSC, was at the same temperature as that of
the fluids entrance to the flat plate solar collectors. [4].

97

Fig. 2. The f chart to design of solar heating fluids systems

3.2. Application of the Solar Program


When the data that the SOLAR program requires are introduced [1], this determinates the
provided amount for the sun at the system. The introduced data in the program are the next:
Latitude of the place:
19.8
Product (FR)(UL):
3.75
Product (FR)( ):
0.68
Collectors efficiency factor (F):
1
Number of collectors covers:
2
Temperature of the supplied water:
18 C
Hot water temperature:
60 C
Volume of the water to be heated:
550 litters per day
Collectors inclination:
20
The total solar radiation in the zone and the average temperature are also introduced in the program.
The Fig. 3 shows the results of the energy fraction to different radiation values.
For this design, the space available was 30 m2, hence we selected the storage capacity of 75 l / m 2.
The storage capacity is also selected, taking into account the costs of installing the system. The
value of the fraction of energy for the SSCP is approximately 43%, indicating the percentage of
energy used by the solar system.

3.3. Determination of Net Thermal Energy Production Including Solar


Energy
The quantity of overall energy needed to raise the temperature of water is shown in (9) [2]:
QT mCp T
(9)
When the heat flow is obtained, the properties of the system are determined in order to dimension
the heat transfer system of the mixture. In this case the heating is released by flat plate solar
collectors. The exit temperature of the collectors is 60 C. For the calculations, the convection and
conduction losses of the digester were taken into account. The plus of the quantities of heat flows
give to result the overall heating load (overall heat flow Q ), which is provided to the solar heating
system by flat plate collectors.

98

Fig. 3. Annual behaviour to the f energy fraction which is provided by the sun
The heat transfer in the heat exchanger involves convection in the fluids side and conduction
through the wall which separates it. The overall heat transfer coefficient (U), takes an account the
contribution of this effects, as follows [2]:
U=

1
1

int

(10)

ext

The overall heat transfer coefficient is used as follows:


QT

(11)

UAs Tlm

The logarithmic mean temperature difference is calculated as depicted in (12):


Tlm=

T1
ln

T2
T1
T2

(12)

In this evaluation the overall heat balance including all temperatures and heat fluxes, such as Tamb
ambient, Tpch air in pump chamber, Tb biogas and Td digester temperatures. The heat fluxes that was
considered are solar flux on a tilted surface IT, top losses from the solar collector Qtlc , sides losses
from the solar collector Qlc , bottom heat loss from the solar collector Qblc , electrical energy
consumption in agitation Qecp, heat losses from the liquid via biogas bubbles Qlbb, heat losses by
convection from the liquid to biogas Qlcb, useful heat gain rate from the collector Qu, auxiliary heat
add to the digester Qaux , rate of heat gain from the pumps Qrhp, heat losses from the pump chamber
to environment Qlche , heat losses from the pump chamber to the reactor Qlcd, heat recovered from
the effluent Qre .
The calculates was made more facility whit the development of a computer program in M ath lab, in
this program was considered all temperatures and heat fluxes, as well as the production of biogas.
(see Apendix B, Table B.1). This program is easier tool to release the calculus.
99

4. Results
4.1. Digester Designed with the Solar Heating System
The amount of energy needed to heat the digester mixture to 55 C was previously determined, a
value that is meet by means of solar energy. Therefore, the heating load for the SSCP is 7.2 x 109 J
/ month. The temperature of the exhaust manifold is 60 C. The digester is 1.8 m3 with a height of
1.45 m. The heat exchanger is composed by 22 aluminium tubes, an arrangement of 6 flat plat solar
collectors, a 150 l/m2 storage tank, a 7 m2 collection area. The design of this exchanger was made
thinking about temperature uniform and more heat transfer of a fluid to another. The mixed is for
two centrifugal pumps. The bottom on the digester has a conical shape to encourage the decantation
of the sludge and avoid the passage of these through of the mixtures recirculation system. The heat
exchanger is collocated in vertical position with its fixed outboard.
Into the digester the mixture is very corrosive due the corrosive components that are naturally
formed in the process.
Control system: the control system is responsible for monitoring of systems functions; it controls
the temperature into of the digester, the temperature of solar collectors, opening and closure of
valves to solar and auxiliary heating system and on-off pumps system.
This new digester design was designed to build and industrial scale, below is shows the integral
system to energy generation in Fig. 4 y Fig. 5.

Fig. 4. Collector designed that shows heat exchanger

100

Fig. 5. Biogas generation system with the solar and auxiliary heating system
The result more important of computer program that was calculated is the generation of methane
about the retention time. These values were determined taking in account standard conditions. In the
Fig. 6 is shown a chart of generation of methane.

Fig. 6. This chart shows the generating methane from the digester design
The digesters of M SW are designed below the base to minimum retention time from 20 to 30 days.

6. Conclusions
The calculation methodology to the design of integral system called Digester-SHS was developed.
Also the calculation tool was made which allows analysing of behaviour to volumes and loads
different. The contribution of this study was an economic and environmentally friendly digester
101

model, as well as this digester can be used in the industry. The list following show the advantages
of the digester designed:
Higher speed of reaction to generate methane
Lower solids retention time
Lower volume of the digester
Contribution to the final dispose of the SM W in the country
Contribution to the reduction of greenhouse effect to the atmosphere

Appendix A
This table show the data of solar radiation that were plotted of 2008 to 2010 to obtain the behaviour
of the radiation solar and can be consider whit the better option.
Table A.1. Data measurement of the total solar radiation on the zone
Year

2008

2009

2010

M onth

H
[J/m2-day]

H
[J/m2-day]

H
[J/m2-day]

M onthly average
temperature (C)

January
February
M archs
April
M ay
June
July
Augusth
September
October
November
December

1.55E+07
1.87E+07
2.00E+07
2.05E+07
2.02E+07
1.84E+07
1.72E+07
1.85E+07
1.47E+07
1.64E+07
1.63E+07
1.55E+07

1.54E+07
1.88E+07
2.21E+07
2.26E+07
2.21E+07
1.86E+07
1.82E+07
1.80E+07
1.42E+07
1.60E+07
1.58E+07
1.43E+07

1.38E+07
1.88E+07
2.42E+07
2.46E+07
2.02E+07
1.84E+07
1.72E+07
1.75E+07
1.47E+07
1.64E+07
1.63E+07
1.31E+07

14.3
15.7
17.6
19.8
20.1
19.8
18.5
19.3
18.3
18.4
15.6
15.4

102

Appendix B
In this appendix an example of data calculated for the compute program is shown.

103

Nomenclature
.

c
h
T

mass flow rate, kg/s


specific heat, J/(kg K)
heat transfer coefficient, W/(m2 K)
temperature difference within the digester, C

Q
U

X
load
Y
load

heat flow rate, W


overall heat transfer coefficient, W/(m2 K)
overall heat transfer coefficient of solar collectors; W/(m2 K)
refence temperature; 100C
monthly overall temperature of the zone to evaluate; C
solar collector rea; m2
Solar overall radiation on the solar collector area; W/m2
monthly overall of transmitance absortance product of the collector
collector-heat removed factor
number of the day of the months
relation between total energy loss for the solar collector to the environment and total heat
relation between absorbed energy for the absorbed plat of the solar collector and total heat

References
[1]
Barbosa J. G. M todo de diseo de sistemas solares para calentamiento de agua. Tesis de
M aestra, 1999. Instituto Politcnico Nacional, SEPI. ESIM E. 151p.
[2]
Cengel Y. A. Transferencia de calo y masa. Tercera edicin, 2007, 970p.
[3]
Couper J. R., Fair J. R., Penney W. R. (2010). Chemical process equipment. ButterworthHeinemann is an imprint of Elsevier, 166-164, 121, 134. 730 p.
[4]
Duffie J. A., Beckman W. A. (2006). Solar engineering of thermal processes. Ed. Wiley
Intersciencie publication, USA. 908 p.
[5]
El-M ashad H. M ., Van Loon WKP, Zeeman G.. A model of solar energy utilization in the
anaerobic digestin of cattle manure. Biosystems Engineering, 2003. 84(2), 231-238.
[6]
Foster C. T. (2005). Digestin anaerobia termoflica seca de residuos slidos urbanos:
estudio de las variables del proceso en el arranque y estabilizacin del bio-reactor. Tesis Doctoral.
Universidad de Cdiz. 421 p.
[7]
Gutin S., M arinek-Logar R. Effect of PH, temperature and air flow rate on the continuos
ammonia stripping of the anaerobic digestion effluent. Process Safety and Enviromental Protection
89 (2011) 61-66.
[8]
Hamed M . El-M ashad; Wilko K.P. van Loon; Grietje Zeeman (2002). A model of solar
energy utilization in the anaerobic digestion of cattle manure. Science Direct, 84 (2), 231238.
[10] Hottel H. C. (1976). A simple model for estimating the transmitance of direct solar radiation
trough clear atmospheres. Solar energy, 18, 129-134.
[11] Kaparaju P., Buenda I., Ellegoard L., Angelidakia I. Effects of mixing on methane
production during thermophilic anaerobic digestion of manure: Lab-sacale and pilot-scale studies.
Bioresurce Technology, 2008. 4919-4928.
104

[12] Klein S. A., Beckman W. A., Duffie J. A (1977). Solar heating design by the f chart method.
Ed., Willey Intersciencie Publication, USA. 636 p.
[13] M ata-lvarez, J., M ac, S., Llabrs, P. (2000). Anaerobic of organic solid wastes. An
overview of research achievements and perspectives. Biores. Techn., 74:3-16.
[14] Romero L. I., Sales D., Galn M . A. Thermophilic anaerobic digestin of winery waste
(vinasses): Kinetics and process optimization. Process Biochem 1988. 23 (4), 119-125.
[15] Soteris A. K. (2009). Solar energy engineering; processes and systems. Elsevier, USA,
760p.
[16] SEDESOL-INE. Informe de la situacin general en materia de equilibrio ecolgico y
proteccin al ambiente. Mxico, 1994. 220p.
[17] Tovar L. R., Castro R. A., Gutirrez M . E., Estrada R. What if the organic fraction of the
wholesale central market of Mexico city (Central de abastos del D.F.) is anaerobically digested?.
Proceedings Venice 2010, Third International Symposium on Energy from Biomass and Waste
Venice, Italy.
[18] Van Lier J. B. Thermophilic anaerobic waster water treatment; temperature aspects and
process stability. PhD Thesis, 1995. Wageningen A gricultural University. Wageningen, The
Netherlands.

105

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Economic feasibility of PV systems in hotels in Mexico


Augusto Snchez Cifuentesa, Sergio Quezada Garcab
a

UNAM, Mexico, D. F., augsan@servidor.unam.mx, CA


b
UNAM, Mexico, D. F., sequga@gmail.com

Abstract:
In this study, the hotels located in Mexico, with energy demand over 100 kW are analyzed to determine the
conditions for which it is profitable to produce their own electricity through photovoltaic panels. The electric
tariff paid by hotels, has different prices for electricity during the day, i.e., is consider the hour of day in which
electrical energy is consumed and the maximum demand that is taken in the month.
By using HOMER software, economic profitability is analyzed under different levels of solar radiation and
demand curve to energy electricity in hotel.
The result shows that only for the hotels located in regions with higher solar radiation in Mexico the use of
photovoltaic panels is profitable. During the study, one of the most important factors to profitability is the time
of day that peak demand comes. However, when making a future projection of the electricity tariff, if they
keep raising the price of electricity as in recent years, in early 2014 photovoltaic panels could be the best
choice for hotels in most parts of the country.

Keywords:
Hotel, photovoltaic, Mexico

1. Introduction
M exico has very favorable natural conditions for the application of photovoltaic systems. In many
parts of its vast territory, the average solar radiation is twice that of European countries like
Germany [1], which is now one of the largest PV markets in the world.
Given the limited information currently available, regarding the financial feasibility of PV systems
connected to the grid in M exico, in this paper examines whether the use of photovoltaic systems in
hotels, with energy demand greater than 100 kW (kilo-watt), enables cost savings from the
perspective of an investor, compared with the purchase of all electricity to the national grid.
To perform the analysis is used the Homer software.In this study, is limited to analysis of
photovoltaic systems as an alternative to the purchase of electricity to the national grid.Other
measures and technologies, such as energy efficiency or other renewable energy technologies are
not considered.

2. Regulatory framework
In 2001 is approved by the Regulatory Commission for Energy (CRE - ComisinReguladora de
Energa) the "interconnection agreement for renewable energy source or system on a small scale
cogeneration"that allows users to general use, install a maximum power generation of 30 kW
produced by renewable energy source under the regime of Net Metering energy. In 2004 comes into
force amendments to the Law on Income Tax (Ley del ImpuestoSobre la Renta), which states that
taxpayers who invest in machinery and equipment for power generation from renewable sources,
can deduct 100% of the investment in a single year [2].

3. Electric tariff H-M


106

The Tariff H-M (Horaria-M edia Tensin, Time-M V), applies to energy services intended for any
use, supplied at medium voltage, with a demand of 100 kW or greater [3]. This tariff is applied to
hotels that are studied in this work.
The tariff has a charge by energy maximum demand (refers to the maximum power demanded,
measured in kilowatt, it is considered from keeping it for 15 minutes) and charges for energy
consumed in time: peak, intermediate and base; therefore, the energy does not have the same cost
during the day.Table 1 shows the different fees for this tariff.

Table 1. Charges averages for tariff H-M for January 2012.


Charges for energy
Charges for energy
Charges for energy
consumption in
maximum demand
consumption in
intermediate
($/kW 1)
peak($/kWh2)
($/kWh)
191.79
2.0206
1.2876

Charges for energy


consumption in base
($/kWh)
1.0332

The annual increase in maximum demand charges in the last three years is on average 4.12%,for
energy in top is 8.33%, for energy in middle 12.61%, for energy in base 12.14%.
Figure 1 show the different periods of peak, intermediate and base for a long of the year, summer
and not summer. The base period is from 00:00 hours until 06:00 throughout the year; the peak
period, in not summer, is from 18:00 to 22:00 hours, while for summer season from 20:00 to 22:00;
the rest are intermediate period.

Fig. 1. Periods: peak, intermediate, and base, during the year.

1
2

Mexican pesos/kilo-watt
Mexican pesos/kilo-watt-hour

107

4. Methodological considerations and scenarios


Homer is energymodeling software for hybrid renewable energy systems, i.e., is a powerful tool for
designing and analyzing hybrid power systems. In this work, Homer is used to analyze the financial
viability of PV systems installed in hotels with electrical tariff H-M . Homer allows changing values
such as: solar radiation, energy consumption, equipment price and equipment capacity, to name a
few, is also possible to establish a minimum rate of return for consider financially feasible the
installation of the photovoltaic system.
For the analysis is performed simulation in Homer by considering the following aspects:
In the present study is considered financially feasible a PV system, if get an internal rate of
return equal to or greater than 6%.
Costs due to machinery and equipment in a PV system is about 65% of the total investment,
therefore, this is the percentage that can be deducting with the Income Tax (Impuestosobre la
renta ISR).
Radiation values are taken between 4 and 6 kw/m2 /day, since this radiation in the M exican
territory [4].
PV system is 30 kW and is connected to the national grid.
Interconnection is regulated by the principle of Net Metering energy, allows offset the cost of the
electricity used by the energy supplied to the national grid.
The price of installing a photovoltaic system is average 40 $/ W, identified in a market survey
for the summer of 2011, if 65% is deducted from this price, the real price of the installation is 14
$/W.
Lifetime of a solar panel is 25 years and the converter is 15 years.
When these values are introduced in Homer are obtained graphs showing the conditions under
which it is profitable to install P Vsystems.It is noteworthy that this study is valid only for buildings
that have energy demand curves similar to those discussed below.
Some studies [1,5] shown that the financial viability of PV systems depend mainly on the energy
consumption, solar radiation and PV system price; therefore the graphs made in this paper illustrate
the behavior of these three variables.
We study two hotels with energy demand over 100 kW, for which two scenarios are taken:
S cenario 1 present tariff:At current prices of PV systems,prices peak demand and energy by
January 2012 for tariff H-M .
S cenario 2 future tariff:With the price of photovoltaic systems for summer 2011 and prices for
the tariff expected for January 2014, in Table 2 are reported these prices.
These scenarios are considered for study the economic viability of photovoltaic systems in the
present (scenario 1) and for projection of economic viability with expected prices of electrical
energy in near future (scenario 2).

108

Table 2. Tariff prices for the two scenarios


Charges for energy
demand ($/kW 3)

Charges for energy


peak ($/kWh4)

191.79
207.92

2.0206
2.3711

Scenario 1
Scenario 2

Charges for energy


intermediate
($/kWh)
1.2876
1.6327

Charges for energy


base ($/kWh)
1.0332
1.2993

5. Analysis of PV systems
We analyze two hotels which have different demand curves of electricity.Figure 2 shows the energy
demand curves for each hotel, in the property a the peak demand of energy occurs from 16:00 to
18:00 hours, while for the property b the maximum demand of energy occurs from 17:00 to 19:00
hours, therefore the hotel a has no maximum demand in the peak hours, while for bthere is
maximum demand in peak hours.
The energy demand curves for each hotel are formed from data measured on the property for a
month every 20 minutes.The hotels are located in M exico and were measured by the National
Autonomous University of M exico.

5.1. Scenario 1
Figure 3 shows the results of the simulation carried out in Homer,for hotela,the right axis shows the
maximum price that can be paid by the photovoltaic system to make it profitable to install,as seen
with increasing solar radiation the price you can pay for the system is greater.
Figure 4 shows the results of the simulation carried out for the hotelb, the right axis shows the
maximum price you can pay for the system for this is even profitable. As with the hotela, with
increasing solar radiation can be paid a higher price for the system, can also see that both hotels the
consumption level of energy is not relevant to the study, because, for different energy consumption
and the same solar radiation, the maximum price arrangement is the same.
Contrary to expectations, for these two hotels the level of energy consumption (x-axis) does not
affect the financial viability of PV systems, i.e., the financial viability of PV systems is not a
function of energy consumption, is depends mainly on solar radiation and the total price of the
array.

Fig. 2. Daily profile for each hotel.

3
4

Mexican pesos/kilo-watt
Mexacan pesos/kilo-watt-hour

109

The results for the two hotels are very similar, although it is a bit more profitable system in the
hotela, is due the peak demand of energy occurs earlierand may cover some of this demand with the
PV system.
At current prices of PV systems,installation is profitable, if you subtract part of initial investment of
the Income Tax (ISR) and has a solar radiation in the area greater than 4.65 kWh/m2/day.

Fig. 3.Analysis of the photovoltaic system for the hotel a.

Fig. 4.Analysis of the photovoltaic system for the hotel b.

110

5.1. Scenario 2
M aking a future projection is repeated the study for both hotelswith priced electrical tariffs that are
expected for 2014.
Figures 5 and 6 show the results obtained in the simulation to the hotel aandb respectively.As in the
previous scenario the results are similar for both hotels, does not influence the energy consumption
in the project's profitabilityand as solar radiation increases the maximum price you can pay for the
system increases.
If there is no change in the prices of the system and have no change in the Law of Income Tax,for
hotels that subtract part of the investment with the Income Tax, the installation of systems
photovoltaic will be profitable, area regardless of the country where they are located.
Comparing the two scenarios is clear that the financial viability of PV systems is affected by the
price of energy in the electricity tariff, i.e., with increasing prices of energy also increases the
maximum price of the PV system with which is profitable yet.

Fig. 5.Analysis of the photovoltaic system for the hotel a with the expected electricity tariff in
January 2014.

111

Figure 5.Analysis of the photovoltaic system for the hotel b with the expected electricity tariff in
January 2014.

6. Conclusions
The financial viability of PV systems is not affected by the level of energy consumption in hotels,
i.e., the financial viability of PV systems depends mainly on solar radiation in the area and the total
system price.As seen in Figures 3 to 5, a given solar radiation has a maximum price of the system
which remains constant for all levels of power consumption.This is true only for hotels with tariff
H-M .
Comparing the two scenarios can be seen that as the price of energy from the grid grows the
maximum price of the photovoltaic system can be greater and therefore becomes more viable
economically, i.e., the price of electricity from the grid Electricity is directly proportional to the
financial viability of PV systems.
At current prices of PV systemsis profitable installation in hotels that cover the tariff H-Mthat found
in areas where average solar radiation in the year is higher than 4.65 kWh/m2/day, regardless of the
level of energy consumption, is a little more profitablefor those hotels that can meet peak demand
of energy with the photovoltaic system.
If you continue the current trend in the electrical tariff increase,by January 2014, will be
economically feasible to install photovoltaic systems in hotels located in any area of M exico.

References
[1] CONUEE, Nichos de mercado para sistemas fotovoltaicos en la conexin con la red elctrica
en Mxico, 2009
[2] Diario Oficial de la Federacin(2004). Fracc. XII, Art. 40, Ley del Impuesto sobre la Renta,
M xico.
112

[3] CFE - The official web site of the ComisinNacional de Energa. Available at:
<http://www.cfe.gob.mx/Paginas/Home.aspx>.[accessed 4.1.2012].
[4] ECOTEC ECOTECSA S.A. de C.V. Available at: <http://www.ecotec2000.de/espanol/sun7.htm> . [accessed 4.1.2012].
[5] Quezada S. Anlisis de la rentabilidad econmica de producir energa elctrica en el sector
residencial y comercial por medio de paneles solares y aerogeneradores en M xico. Tesis, Facultad
de ingeniera. 2011.

113

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Effect of a back surface roughness on annual


performance of an air-cooled PV module
Riccardo Secchia , Duccio Tempesti a , Jacek Smolkab
a

Universit degli Studi di Firenze, Dipartimento di Energetica Sergio Stecco, Florence, Italy,
riccardo.secchi@unifi.it
b
Institute of Thermal Technology, Silesian University of Technology, Gliwice, Poland,
jacek.smolka@polsl.pl

Abstract:
In this paper, a study of increasing a photovoltaic (PV ) module efficiency via natural/ forced cooling of t he PV
cell is presented. The PV module is cooled by the air flowing in a duct placed under a back surface of the PV
panel. The cooling air is moved either by fan or naturally by the temperature difference generated from the
heat trans fer with a panel. The system could be applied on t he roof of industrial facilities, with electrical and
thermal energy needs. The pres ented 1-D model examines an already published thermal and electrical PV/ T
approach supplement ed with a radiative heat transfer in the panel duct. In addition, an effect of the
roughness of the panel back surface is also added to the model. The model coded in an Engineering
Equation Software (EES) is capable of evaluating the PV module efficiency and other thermal parameters as
the outlet air temperature and back surface temperature. As a result, the performance of the system in a
typical winter and summer day is discussed. Furthermore, an annual simulation of the system is also

analysed coupling the EES model with the Trnsys software. The results of the annual simulation show that
the raise of the relative roughness of t he panel back surface from 0 to 0,05 leads to an increase of 0.25% of
the average efficiency of the PV module in cas e of air moved by a fan. In terms of overall efficiency, this
result means a relative improvement of 1.73%. Annual average thermal efficiency of the PV/ T system of
28.6% is achieved for relative roughness of the panel back surface set to 0.05.

Keywords:
Solar photovoltaic thermal (PV/ T) air collector, heat trans fer, system optimization.

1. Introduction
In Europe, the solar photovoltaic (PV) electricity generation capacity has increased 160 times from
185 MW in 2000 to 29.5 GW in 2010 [1]. This is mainly due to the public support programs for
renewable energies, which usually consist of feed- in tariffs for the new installed PV systems. Since
only a small fraction of the incident solar radiation is used by the PV cells to generate electricity,
the PV module efficiency is in the range of 4-17% depending on the type of solar cells and the
working conditions. This means that at least 50% of the incident solar energy is mainly turned into
heat in the cells and substrate, leading to two consequences: (a) a drop in cell efficiency (usually
0.4% per C increase for c-Si cells) and (b) a permanent structural damage of the module if the
thermal stress persists for extended period [2]. In solar photovoltaic thermal (PV/T) collector, the
PV module is cooled by air or water, producing thermal and electrical energy simultaneously. The
higher efficiency leads to a higher production of the system, which causes an increase in the
economic incomes from the feed- in tariff for the PV system holder. In addition, the heated cooling
flow rate (air or water) can be used both for domestic heating and production of hot water. For
these reasons, in the last 40 years a lot of studies (theoretical as well as experimental) on the PV/T
systems has been carried out. Kern and Russel [4] presented the design and the performance of
water and air cooled PV/T systems, while Florschuetz [5] included the PV/T modelling in his
114

works. Garg and Adhikari [6] presented a variety of results regarding the effect of design and
operational parameters on the performance of the air type PV/T systems. Hegazy [7] investigated
glazed photovoltaic/thermal air system for a single and a double pass air heater for the space heating
and drying purposes. Kalogirou [8] modelled the hybrid photovoltaic/thermal (water) solar energy
system with TRNSYS and the simulations carried out for Nicosia (Cyprus) showed an increase of
the mean annual efficiency of the PV solar system from 2.8% to 7.7% with the thermal efficiency of
49%. Zondag et al. [9] developed a range of steadystate and dynamic simulation models for the
PV/T (water) energy performance analysis that included 1-D, 2-D and 3-D models of the serpentine
PV/T collector (their accuracy was verified by experimental data). Tiwari et al. [10] validated the
theoretical and experimental results for the photovoltaic (PV) module integrated with an air duct for
a composite climate of India and concluded that the overall thermal efficiency of PV/T system is
significantly increased due to utilisation of thermal energy from the PV module. Dubey et al. [14]
studied different configurations of the glass-to-glass and the glass-to-tedlar PV modules. Analytical
expressions for electrical efficiency with and without airflow were developed as a function of
climatic and collector design parameters. Experiments that were performed at the Indian Institute of
Technology, Delhi showed that the glass-to- glass type achieves higher supply air temperature and
electrical efficiency. By the use of validated theoretical models, Tonui and Tripanagnostopoulos
[16] studied a degree of improvement by adding suspended metal sheet at the middle of the air
channel and the finned arrangements at the opposite wall of the air channel. It was found that these
low-cost improvements are more effective at small collector length, and can be readily applied to
PV/T (air) installations. On the other hand, an effect of the channel depth, mass flow rate or system
length on the fan power consumption was found small.
In this study, on the basis of the PV models developed by Tiwari et al [10,15], and
Tripanagnostopoulos and Tonui [16], a theoretical steady state 1-D model including a convective
and radiative heat transfer in a panel duct is developed with the EES software [19]. In addition, the
annual gain in energy production of the investigated system is modeled integrating the EES model
with TRNSYS that is a transient simulation program, and using typical meteorological year (TMY)
conditions data for the central Italy. Furthermore, an effect of the roughness of the back surface on
the PV module effectiveness is also investigated. Finally, the results of the cooling of the PV
module through air moved by fan or naturally flowing are presented.

Air outlet

PV module

Insulating
panel

LD

Air inlet
b
Fig.1 Perspective view of the hybrid PV/T system studied.

115

2. Configuration of PV/T hybrid system


The specific configuration of the hybrid PV/T system that has been studied consists of a
photovoltaic module installed on industrial shed roof. A typical configuration of the system is
shown in Fig.1. For such system, a 1-D thermal model was formulated based on a models reported
in literature [10,11,13,14,15]. The energy balance equations have been modified taking into account
the irradiation between the back surface of the PV module and the opposite wall of the duct.
Fig.2 shows a simplified cross-sectional view of the PV/T air collector that is composed of three
layers. Furthermore, the equivalent thermal resistant circuit of the system discussed in Section 2.1 is
also illustrated in Figure 2.
T AMB

T SKY
h rad

h conv,upp
Ut

Glass
T CELL

Silicon Cell

UT

Tedlar
Airflow m

TBS
hf
TAI R

Airflow m
h rad_duct

hf

Insulation

Ti
Ub0
h conv, und
T AMB

Fig.2 Thermal resistance circuit diagram for PV/T system.

2.1. Energy balance for PV/T hybrid system


In order to formulate the energy balance equations for each component of PV/T hybrid system, the
following assumptions have been made:
The heat conduction is one-dimensional (only along the height of the module).
The temperature of the model layers is uniform.
The system is in quasi-steady state.
The ohmic losses in the solar cell are negligible.
Air flow in the duct between a tedlar layer and insulating structure is one-dimensional.
To determine the temperatures of solar cells and then the efficiency of the considered system, four
energy balance equations need to be written.
According to Fig.2, one can write the equation describing the overall heat transfer from the top of
the module to the ambient below the insulation in a form:

116

CELL

p I

1 p

U t TCELL TAMB

I b dx

U T TCELL TBS

b dx

CELL

CELL

(1)

I b dx

where the terms in Eq. (1) are the following:


1 - The rate of solar
energy available
on solar cell

2 - An overall heat
loss from top
surface of cell
to ambient

3 - An overall heat
transfer from cell
to backsurfac e of tedlar

4 - The rate of
electrical energy
produced

The second energy balance equation describes the heat transfer for back surface of tedlar:
UT

TCELL

TBS

b dx

hf

TBS

TAIR b dx

hRAD _ DUCT

TBS

TI

b dx

(2)

where the terms in Eq. (2) are the following:

3 - An overall heat

5 - The rate of heat trans fer

6 - The rate of heat

transfer from cell to


back surface of tedlar

from back surface of the


tedlar to flowing fluid

transfer from back surface of


tedlar to insulation upper surface

The next equation is the overall duct energy balance:


hf

TBS

m AIR C AIR

T AIR b dx h RAD _ DUCT TBS


dTAIR
dx U b0 TI
dx

TI b dx

(3)
TAMB b dx

where the terms in Eq. (3) are the following:

5 - The rate of heat trans fer


from back surface of
the tedlar to flowing fluid

6 - The rate of heat


transfer from back
surface of tedlar
to insulation
upper surface

7 The mass
flowrate of
flowing fluid

8 An overall heat
transfer from upper
surface of insulation
to ambient

Finally, the fourth energy balance equation for the upper surface of insulation can be written as:
h RAD _ DUCT TBS

TI

b dx

U b 0 TI

TAMB b dx

hf

TI

T AIR b dx

(4)

where the terms in Eq. (4) are the following:

6 - The rate of heat trans fer

8 An overall heat

9 - The rate of heat trans fer

from back surface of tedlar


to insulation upper surface

transfer from upper surface


of insulation to ambient

from upper surface of insulation


to flowing fluid

The mathematical transformations to obtain the explicit formulation for all temperatures are
described in Appendix A. In addition, the relations used in Eqs (1)-(4) for the heat transfer
117

coefficients, and the assumptions adopted for the calculation of the radiative heat transfer
coefficient are given in Appendix B.
In order to calculate the temperature-dependent electrical efficiency of the PV module, the
following expression has been used [17,24]:
TCELL TREF
(5)
CELL
REF 1
The rate of thermal energy obtained from the hybrid system is:

QU

mAIR C AIR TAIR LD

TAIR _ IN

(6)

The thermal efficiency of the PV/T system has been calculated using the following equation:
TH

QU
b LD I

(7)

2.2 Roughness of back surface of PV panel


In order to obtain a larger temperature drop of photovoltaic cells, the heat transfer between the back
surface of the PV module and the air flowing in the duct should be improved.
The convective heat transfer coefficient (hf), which identifies the amount of heat exchanged
between the module and the air flow rate, has been calculated by means of an internal function of
the software EES called ductflow [19,21]. This function calculates the Nusselt number and the
friction factor, and then uses these values to evaluate the pressure drop along the duct and the
mentioned convective heat transfer coefficient. The input parameters of that function are: the size of
the duct (width, height, length ), the average airflow temperature, pressure, the air mass flow rate
and Relative Roughness (RR).
The RR (roughness related to hydraulic diameter of the duct) can be between 0 and 0.05. The
boundary of RR range (i.e. 0 and 0.05) have been considered to test a possible gain in terms of the
net electrical efficiency of PV module, taking into account that the energy consumption to move the
air flow rate raises at RR=0.05.

3. Results and discussion


3.1 Roughness influence
The governing equations have been initially computed with EES software to evaluate the behavior
of the system in two steady-state weather reference conditions:
1. Italian Winter : I = 600 W/m2 , TAMB = 283 K.
2. Italian Summer : I =1000 W/m2 , TAMB = 308 K.
In order to simulate less advantageous heat removal conditions from the upper surface of PV/T
system, in both cases the wind speed, that directly affects the convective heat transfer coefficient on
upper surface of PV module, according to Equation B.2, has been considered to be zero (Vw = 0
m/s).
The aim of these simulations was to find the value of airflow velocity that would maximises the net
power gain between RR=0 and RR=0.05.
The operating and design parameters of the PV/T air system used in the model simulations are
described in Table 1. Fig.3 shows the electrical efficiency variation as a function of the air flow
velocity in a duct for all the four cases considered (summer, winter, roughness considered or
neglected).
118

Table 1. Design parameters of the PV/T air system [11].


PV/T air hybrid system parameters
Length of PV module and air duct, LD
Width of PV module, b
Thickness of glass cover, LG
Conductivity of glass cover, K G
Transmittivity of glass cover, G
Emissivity of glass cover, G
Absorptivity of solar cell, CELL
Thickness of solar cells, LCELL
Conductivity of solar cells, KCELL
Absorptivity of tedlar, T
Thickness of tedlar, LT
Conductivity of tedlar, K T
Emissivity of tedlar, T
Emissivity of insulation upper surface, I
Thickness of tedlar, LI
Conductivity of insulating panel, K I
Equivalent duct depth
Packing factor of solar cells, p
Electrical efficiency at Standard Conditions, REF
Solar cells temperature at Standard Conditions, TREF
Solar radiation intensity at Standard Conditions, IREF

Value
2m
1.143 m
0.003 m
1 W/(mK)
0.95
0.88
0.85
300 X 10-6 m
0.036 W/(mK)
0.5
0.5 X 10-3 m
0.033 W/(mK)
0.87
0.1
0.05 m
0.035 W/(mK)
0.0782 m
0.83
0.1446
298 K
1000 W/m2

Electrical efficiency (-)

In winter reference conditions, the lower ambient temperature leads to a higher electrical efficiency.
Furthermore, for RR=0.05, an air flow speed of 2 m/s is sufficient to obtain the maximum
efficiency.
0.15
0.145
0.14
0.135
0.13
0.125
0.12
0.115
0.11
0.105
0.1
0.095

Summer,
RR=0
Summer,
RR=0.05
Winter,
RR=0
Winter,
RR=0.05

0.2 0.4 0.6 0.8

1.2 1.4 1.6 1.8

2.2 2.4 2.6 2.8

Airflow velocity (m/s)

Fig. 3. Electrical efficiency for RR=0 and RR=0.05, for summer and winter weather reference
conditions.
119

The effect of the roughness cannot be determined for average air flow speed lower than 0.2 m/s,
because the Reynolds number is too low at that speed (see overlapped curves in Fig.3).
In Fig.4, the power gains between the configuration with RR=0 and RR=0.05 for both reference
conditions are shown. The gap of power gains between summer and winter shown in Fig.4 include
the power costs for ventilation, given by the total pressure drop in the duct, in accordance with the
following equation:
1 m AIR
0.6 AIR

W FAN

(8)

Power (W)

20
18
16

Summer

14

Winter

12
10
8
6
4
2
0
0

0.2

0.4 0.6 0.8

1.2 1.4 1.6

1.8

2.2 2.4

2.6 2.8

Airflow velocity (m/s)

Fig. 4. Net power gains (in Watt) between RR=0 and RR=0.05 for summer and winter weather
reference conditions.

An airflow velocity equal to 1.5 m/s, which is a mean value between the optimum of the single
reference conditions, is considered the optimum value for the annual simulation.
The convective (hf) and radiative (hRAD_DUCT) heat transfer coefficients are the key parameters that
affect the heat transfer inside the duct. For this reason, they are extracted from the results and
shown in Table 2.
Table 2. Convective (hf) and radiative (hRAD_DUCT) heat transfer coefficients for RR=0 and RR=0.05
obtained for summer and winter weather reference conditions.

hf , W/(m2 K)
hRAD_DUCT , W/(m2 K)

Summer
Winter
RR=0 RR=0.05 RR=0 RR=0.05
7.797 22.170 8.190 24.110
0.790
0.733
0.578
0.546

3.2 Comparison with the literature models


The differences between the proposed model and the two literature models used as a basis for this
study are the following:
Tripanagnostopoulos and Tonui - The upper layer of the model is considered to be composed of
only glass and PV cells, which cover the entire surface. Thus, there is no distinction between the
120

fraction of solar radiation, which impacts the PV cells and the one that strikes the spaces between
the cells. Furthermore, a consequence of this assumption concerns the calculation of
temperatures in the different layers of the model: no distinction is made between the temperature
of the PV cells and that of the back surface of Tedlar. As a result, the model one equation.
Another important assumption concerns the airflow temperature, which is assumed to vary
linearly in the flow direction. Tan and Charters [18] correlation is used to compute Nusselt
number and the resulting the forced convection heat transfer coefficient in the air channel. The
friction factor that is needed for the calculation of the pressure drop along the duct, is calculated
from the equations given by Incropera and DeWitt [23].
Tiwari at al. - Radiative heat transfer in the air duct is neglected. In consequence, the temperature
of the insulation upper surface is not calculated and the model is reduced by one equation. The
convection heat transfer coefficient inside the duct has been originally assumed to be a constant
value, but in recent works [11] is calculated according to flow regime and its Nusselt number (no
information is given in [11] about the formulas used).
The three models have been tested with the design parameters of the present study in order to
compare the values obtained for layer temperatures and other basic parameters of the PV/T system.
The comparison is shown in Tables 4 and 5 for winter and summer reference conditions,
respectively.
Table 4. Comparison of simulations results obtained with the three different models, for winter
reference condition (TAMB=TAIR IN=283 K, I=600 W/m2 , Vw=0 m/s), Relative Roughness=0, and
forced circulation with the airflow velocity of 1.5 m/s.

MODEL
Tripanagnostopoulos
Tiwari et al.
Present work

TAIR OUT

T AIR

TCELL

TBS

TI

hf

EL

TH

(K)

(K)

(K)

(K)

(K)

(W/(m2 K))

285.5
285.5
285.5

284.2 302.4
285.5
284.2 310.8 306.5
284.3 310.1 305.7 285.1

9.49
8.19*

P
(Pa)

0.1417 0.3013 0.54


0.1361 0.3023
0.62*
0.1365 0.31

Table 5. Comparison of simulations results obtained with the three different models, for summer
reference condition (TAMB=TAIR IN=308 K, I=1000 W/m 2 , Vw=0 m/s), Relative Roughness=0, and
forced circulation with the airflow velocity of 1.5 m/s.
TAIR OUT
MODEL
Tripanagnostopoulos
Tiwari et al.

(K)

312.7
312.0
312.1

T AIR
(K)
310.4
310.0
310.1

TCELL

TBS

TI

hf

EL

TH

(K)

(K)

(K)

(W/(m2 K))

343.1
312.6
350.9 344.6
349.8 343.1 312.8

9.08
7.79*

P
(Pa)

0.1146 0.3178 0.52


0.1094 0.2685
0.59*
0.1102 0.2792

Present work
* Since Tiwari et al. [11] gives no information about the formulas used for convective heat transfer coefficient and
pressure drop inside the duct, these values have been calculated by means of the EES function ductflow like in the
present work.

3.3 Annual performance


The annual performance of the system have been simulated with the software TRNSYS, which
allows the authors to test the system on the hourly averaged values of a solar radiation, ambient
temperature and wind speed that characterize the annual weather conditions of one of many sets in

121

the software database. As an example, the weather data for Rome have been chosen in the present
study.
In an annual simulation, the incident radiation on the PV module varies considerably depending on
the position of the panel (azimuth, tilt to the horizontal) and the incoming radiation direction.
In this simulation, the available radiation on the panel was calculated as a function of the angle of
incidence of the three radiation components (direct, diffuse and reflected from the ground) in each
time step. Incidence Angle Modifiers (IAM) of each radiation component was determined as a
function of the Incidence Angle of the component [20,25].
The PV/T system has been considered oriented to the south with a slope of 30.
Annual simulations have been performed for the following cases:
1. Natural circulation: the model has been modified so that the flow of air is generated by the

temperature difference between the air inside the duct and the surrounding environment. For this
configuration, only the simulation with RR=0 is considered.
2. Forced circulation (air average velocity of 1.5 m/s) with RR=0 and RR=0.05.
It is important to point out that in the annual simulations the wind speed values from the database
have been used, while in the simulations in steady-state reference conditions the wind speed was
fixed to zero. This leads to higher values of electrical efficiency due to the increased heat removal
on the upper surface of a PV module. Moreover, the effects of wind are accentuated by the fact that
the incoming direction is not accounted for.
The monthly net electrical energy generated by the PV module for all three cases considered is
shown in Fig.5. According to this figure, it can be observed that the effects of the forced circulation
and non- zero roughness lead to the maximum increase in energy produced that occurs in the months
from May to September. Furthermore, Fig.5 shows a difference of the energy produced above 2
kWh between natural convection and convection with RR=0.05, and above 1 kWh between natural
convection and forced convection with RR=0.

Forced Circulation,
RR=0

65.00
60.00

Forced Circulation,
RR=0.05

55.00
50.00
45.00

Natural Circulation,
RR=0

40.00
35.00

O
CT
N
O
V
D
EC

SE
P

L
A
U
G

JU

A
PR
M
A
Y
JU
N

R
M
A

JA

FE
B

30.00
25.00
N

Net Energy generated (kWh)

70.00

Month

Fig. 5. Monthly Net Energy in kWh generated for natural circulation (RR=0) and forced circulation
(RR=0 and RR=0.05).

Fig.6 presents the monthly averaged electrical energy efficiency for all three cases simulated. In this
figure, it can be observed that an increase in the electrical efficiency obtained with a forced
122

0.1450
0.1430
0.1410
0.1390
0.1370
0.1350
0.1330
0.1310
0.1290
0.1270
0.1250

Forced Circulation,
RR=0
Forced Circulation,
RR=0.05

JU
L
A
UG
SE
P
OC
T
NO
V
D
EC

AP
R
M
A
Y
JU
N

M
AR

FE
B

Natural Circulation,
RR=0

JA
N

Average Electrical Efficiency (-)

circulation is higher in the months from May to September. The difference in the monthly averaged
electrical efficiency between the forced convection cases and the natural convection case is shown
in Table 3.

M onth

Fig. 6. Monthly averaged electrical efficiency for natural circulation (RR=0) and forced circulation
(RR=0 and RR=0.05).

As already shown for the steady-state weather reference conditions, in winter months the increase in
electrical efficiency of the PV module is limited, due to the low ambient temperature, which enables
the module to naturally work close to its maximum efficiency.
The resulting annual average electrical efficiency are the following:
1. 13.35% for Natural Circulation
2. 13.60% for Forced Circulation and RR=0
3. 13.85% for Forced Circulation and RR=0.05
Table 3. Difference in monthly averaged electrical efficiency between the forced circulation (FC)
cases with RR=0 and RR=0.05, and natural circulation (NC) case.

(FC,
RR=0)
- (NC)
(FC,
RR=0.05)
-(NC)

JAN

FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

OCT NOV

DEC

0.16

0.18

0.25

0.25

0.29

0.28

0.24

0.28

0.31

0.26

0.21

0.22

0.29

0.33

0.46

0.49

0.57

0.56

0.50

0.57

0.61

0.52

0.40

0.39

The monthly averaged thermal efficiency for all simulated cases is shown in Fig.7. The air flow
velocity under natural circulation reaches the maximum value of 0.4 m/s. Then the thermal energy
removed by the air flow results in the annual average thermal efficiency of about 8.3%. The annual
average thermal efficiency of about 18.5% is achieved for the forced circulation simulations with
123

RR=0, due to the air flow velocity of 1.5 m/s. The third configuration (forced circulation,
RR=0.05), leads to the maximum increase in annual average thermal efficiency of 28.6%.

Forced Circulation,
RR=0

0.2500

Forced Circulation,
RR=0.05

0.2000
0.1500

Natural Circulation,
RR=0

0.1000
0.0500

JA

FE
B
M
A
R
AP
R
M
A
Y
JU
N
JU
L
A
U
G
SE
P
O
CT
N
O
V
D
EC

0.0000
N

Average Thermal Efficiency (-)

0.3000

Month

Fig. 7. Monthly averaged thermal efficiency for natural circulation (RR=0) and forced circulation
(RR=0 and RR=0.05).

4. Conclusion
In this paper, the electrical and thermal performance of three different configurations of a hybrid
PV/T system has been presented. A detailed one-dimensional thermal model was developed to
calculate the thermal parameters of the system. The photovoltaic electrical efficiency was calculated
using a linear correlation of the solar cell temperature. Numerical simulations concerning the
evaluation of the optimal air flow velocity varying the roughness of the back surface of the module
were carried out.
From the study the following conclusions can be drawn:
The airflow velocity that maximizes the net power generated gains between RR=0 and RR=0.05
results is 1.5 m/s;
For annual simulation, the raise of the relative roughness of the panel back surface from 0 to 0.05
leads to an increase of 0.25% of the annual average efficiency of the PV module in case of air
moved by fan. In terms of overall efficiency, this result corresponds to a relative improvement of
1.73%.
For RR=0, there is a difference of 0.25% in terms of the annual average electrical efficiency
between forced and natural circulation configurations.
The annual average thermal efficiency of the PV/T system increases by 10%, while a natural
circulation is replaced with a forced circulation (both with RR=0). With a roughness of 0.05,
there is an additional increase of 10%, which leads to an overall annual average thermal
efficiency of 28.6%.
This preliminary assessment of the performance of the PV/T system here proposed will be used as a
reference solution for the next planned experimental evaluation. However, in terms of power gain
(and, finally, in terms of cash earnings from the feed in tariff) the PV/T system proposed is very
interesting for industrial facilities with large roof surface, where it can easily be installed.
124

Appendix A
The mathematical steps for the explicit formulation of all model parameters are reported in this
Appendix.
From Eq. (1), the expression for solar cell temperature is:

EFF

TCELL

I U t T AMB U T TBS
Ut UT

(A.1)

where
EFF

CELL

1 p

CELL

(A.2)

Using Eq. (1) and Eq. (2), the expression for the temperature of tedlar back surface is:
h P1

TBS

EFF

I U tT T AMB
U tT

hf

h f T AIR

hRAD _ DUCT TI

(A.3)

hRAD _ DUCT

where

UT
U
U t U T , tT

h P1

U t UT
U t UT

Combining Eqs (1), (2) and (3), the following linear differential equation is obtained:

dTAIR
dx

TAIR

(A.4)

where
hP 3

b,

TAIR U tf

U b0

h P1 hP 2

I TI

EFF

hP 4

mAIR C AIR
h P2

U tf

hf
hf

hRAD _ DUCT

hf

, h P3

U tT

h RAD _ DUCT

h RAD _ DUCT

U tT

U tT h f
hf

hRAD _ DUCT

U tT

, h P4

TAIR _ IN

U tT hRAD _ DUCT
hf

hRAD _ DUCT

U tT

Thus, solving Eq. (7) , with boundary conditions at x = 0, TAIR x


T AIR x

m AIR C AIR
hRAD _ DUCT

U tT h f

U b0

T AIR _ IN

T AMB , leads to:

(A.5)

The temperature of the air leaving the duct (at a distance LD form the entrance) can be obtained
from Eq. (A.5) for x=LD.
A fluid temperature can be averaged over the considered length and can be calculated as:
L

TAIR

1 D
T x dx
LD 0 AIR

1
LD

T AIR _ IN

1 e

125

LD

(A.6)

The expressions obtained in Eqs (A.5) and (A.6) for air temperature involves the value of TI
(temperature of the insulation upper surface), therefore, they can be solved using Eq. (4).

Appendix B
The heat transfer coefficients used in the modeling equations, are defined as follows:
hCONV ,UPP

2.8 3 Vw

(B.1)

hCONV ,UND

2.8 3 VA _ UNDER _ I

(B.2)

h RAD

TSKY

TCELL

TSKY

TCELL

(B.3)

The radiative heat transfer coefficient inside the duct was calculated considering the formula for
two infinitely long, gray, opaque, directly opposed parallel plates of the same finite width [22]. The
emissivity of the back surface of PV module was set at 0.87 [26] and that of the insulation upper
surface made of aluminum sheet was defined to be 0.1 [26].
1

h RAD _ DUCT

BS

TSKY
Ub0

Ut

UT

TAMB

TBS

TI

TBS

(B.4)

(B.5)
1

(B.6)
1

hCONV ,UND
1

LG
KG

(B.7)
1

hCONV ,UPP

hRAD

1
LCELL
K CELL

(B.8)

LT
KT

Nomenclature
b
LD
C
I
L

LI
KI

TI

width of PV module, m
length of PV module, m
specific heat, J/(kgK)
incident solar intensity, W/m2
thickness, m
126

hCONV_UPP
hCONV_UND
hf
hRAD
hRAD_DUCT
K

QU

mass flow rate, kg/s


convective heat transfer coefficient on upper surface of PV module, W/(m2 K)
convective heat transfer coefficient on lower surface of insulating panel, W/(m2 K)
convective heat transfer coefficient inside the air duct, W/(m2 K)
radiative heat transfer coefficient from solar cells to ambient, W/(m2 K)
radiative heat transfer coefficient inside the duct, W/(m2 K)
thermal conductivity, W/(mK)
rate of useful energy transfer, W
temperature, K
packing factor
pressure drop, Pa
overall back loss coefficient from upper surface of insulation to ambient, W/(m2 K)
overall heat transfer coefficient from solar cell to ambient ,W/(m2 K)
conductive heat transfer coefficient through solar cell and tedlar, W/(m2 K)
overall heat transfer coefficient from glass to tedlar through solar cell, W/(m2 K)
velocity, m/s

T
p
P
Ub0
Ut
UT
UtT
V
Greek symbols
)EFF

absorptivity
product of effective absorptivity and transmittivity
efficiency reduction coefficient
emissivity
StefanBoltzmann constant, W/(m2 K4 )
efficiency
density, kg/m3
transmittivity

Subscripts and superscripts


AIR
air
AMB
ambient
BS
back surface of PV module
CELL
solar cell
G
glass
I
insulating panel
REF
reference condition
T
tedlar
TH
thermal
w
wind

References
Journals :
[1] Jger-Waldau A. The Photovoltaic Market Appendix D. In: McEvoy A., Markvart T. and
Castaner L., Practical Handbook of Photovoltaics (Second Edition) Fundamentals and
Applications. Elsevier 2012, p. 1145-1151.
127

[2] Chow T.T. A review on photovoltaic/thermal hybrid solar technology. Applied Energy
2010;87:36579.
[3] Wolf M. Performance analysis of combined heating and photovoltaic power systems for
residences. Energy Conversion Management 1976;16:7990.
[4] Kern Jr. E.C., Russell M.C. Combined photovoltaic and thermal hybrid collector systems. In:
Proceedings of 13th IEEE photovoltaic specialists, 1978, Washington, DC,USA; 11537.
[5] Florschuetz L.W. Extension of the HottelWhiller model to the analysis of combined
photovoltaic/thermal at plate collectors. Solar Energy 1979;22:3616.
[6] Garg H.P., Adhikari R.S. Conventional hybrid photovoltaic/thermal (PV/T) air heating
collectors: steady-state simulation. Renewable Energy 1997;11:36385.
[7] Hegazy A.A. Comparative study of the performances of four photovoltaic/thermal solar air
collectors. Energy Conversion Management 2000;41:86181.
[8] Kalogirou S.A. Use of TRYNSYS for modelling and simulation of a hybrid PV thermal solar
system for Cyprus. Renewable Energy 2001;23:24760.
[9] Zondag H.A., De Vries D.W., Van Helden W.G.J., Van Zolengen R.J.C., Van Steenhoven A.A.
The thermal and electrical yield of a PV-thermal collector. Solar Energy 2002;72(2):11328.
[10] Tiwari A., Sodha M.S., Chandra A., Joshi J.C. Performance evaluation of photovoltaic
thermal solar air collector for composite climate of India. Solar Energy Material Solar Cells
2006;90(2):17589.
[11] Sarhaddi F., Farahat S., Ajam H., Behzadmehr A., Adeli M.M. An improved thermal and
electrical model for a solar photovoltaic thermal (PV/T) air collector. Applied Energy
2010;87:232839.
[12] Kumar R., Rosen M.A. A critical review of photovoltaicthermal solar collectors for air
heating, Applied Energy 2011;88:3603-3614.
[13] Dubey S., Tiwari G.N. Thermal modeling of a combined system of photovoltaic thermal
(PV/T) solar water heater. Solar Energy 2008;82:60212.
[14] Dubey S., Sandhu G.S., Tiwari G.N. Analytical expression for electrical efficiency of PV/T
hybrid air collector. Applied Energy 2009;86(5):697705.
[15] Joshi A.S., Tiwari A., Tiwari G.N., Dincer I., Reddy B.V. Performance evaluation of a
hybrid photovoltaic thermal (PV/T) (glass-to- glass) system. Int J Therm Sci 2009;48:15464.
[16] Tripanagnostopoulos Y., Tonui J.K. Air-cooled PV/T solar collectors with low cost
performance improvements Solar Energy 2007;81:498511.
[17] Evans D.L. Simplified method for predicting PV array output. Solar Energy 1981;27:555
60.
[18] Tan H.M., Charters W.W.S. Effect of thermal entrance region on turbulent forcedconvective heat transfer for an asymmetrically rectangular duct with uniform heat flux. Solar
Energy 1969;12,513516.
Books:
[19] Klein S.A., Engineering Equation Solver, Academic Version V8.603, www.fchart.com
[20] Duf e J.A., Beckman W.A. Solar engineering of thermal processes. 3rd ed. New York,
USA: Wiley; 2006.
[21] Nellis G.F, Klein S.A. Heat Transfer. Cambridge (UK): Cambridge University Press; 2009
[22] Cengel Y.A., Termodinamica e trasmissione del calore, McGraw-Hill,1998.
[23] Incropera F.B., DeWitt D.P.. Fundamentals of Heat and Mass Transfer. John Wiley & Sons,
New York, 1996.
Conference Papers :

128

[24] Schott T. Operational temperatures of PV modules. In: Proceedings of 6th PV solar energy
conference; 1985. p. 3926.
[25] King, David L., Jay A. Kratochvil, and William E. Boyson. Measuring the Solar Spectral
and Angle-of-Incidence Effects on Photovoltaic Modules and Irradiance Sensors. Proceedings
of the 1994 IEEE Photovoltaics Specialists Conference. Sept 30-Oct 3, 1997. pp. 1113-1116.
Web references:
[26] Material Emissivity Properties : http://snap.fnal.gov/crshield/crs- mech/emissivity-eoi.htm

129

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCEON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Energy and exergy analysis of the first hybrid


solar-gas power plant in Algeria
Fouad Khaldi
Department of Physics, University of Batna, Batna, Algeria, fouadkhaldi@gmail.com

Abstract:
The first concentrating solar power plant in Algeria began to produce electricty in the middle of the last year.
The hybridiz ation solar-gas is realized by the integrating of parabolic trough collectors into a c ombined cycle
power plant. The thermodynaic evaluation of the power plant performance at design running conditions is
based on the exergy analysis. In this analysis, the exergy flow diagram is aideed by the value diagram which
is used as a visualisation tool for identifying ex ergy degradation in heat trans fer processes involved in the
HRS G and the solar steam generator. The exergy destruction and the functional exergy efficiency are the
key paramet ers in assessing the performance of every power plant component. In regard to the electrical
power capacity of t he power plant, 160 MW, the solar energy s hare is 14% (22 MW), while the solar exergy
share is 12% (18.4 MW). The combustors and the solar field are the less efficient systems.The combustors
destruct about one third of natural gas exergy and the solar field consumes about three quarters of solar
exergy.

Keywords:
Integrated solar combined cycle, Concentrating solar power, Parabolic
Thermodynamic performance, Exergy analysis, Value diagram, Cycle-Tempo.

trough

collector,

1. Introduction
In Algeria the total installed power generating capacity is over 9 GW, 98% of which is provided by
gas- fired plants [1]. Algeria, located in the MENA region, has impressive solar resources [2-4]. The
Algerian desert is exposed yearly to a direct sun irradiation higher than 2000 kWh/m2 gained from
3500 hours of sunshine. These solar potential associated to huge land resources are suitable for the
implementation of concentrating solar power plants (CSPPs) [5]. The Algerian authorities are
planning to produce 6% of electricity from CSPPs by 2020 [1].This target should be reached
through the building of four CSPPs totaling a solar installed capacity of 240 MW [1]. The first
CSPP, at Hassi RMel, is already running since 2011. Since Algeria is a major producer and
exporter of natural gas, hybrid solar-gas power plants are more appropriate; besides this option
avoids the issues of storage systems. This choice is supported by the observation that in the last
years the peak of electrical consumption in Algeria was recorded in summer and in day times [1].
That meets with the times of the day and the year when solar energy has the advantage to produce
energy. The planned hybrid plants will contribute to save about 2.7 billion m3 of natural gas.
Similar projects are in progress in Morocco, Tunisia, Egypt, Jordan, and Iran [6-11] and elsewhere
of MENA region [12].
The solar thermal power generated by CSPPs can be supplied at different temperature levels. Puresolar power tower plants work at high temperature level (>500C) [13-16]. CSP systems at low or
medium temperature levels (250 to 400C), based on trough or Fresnel collectors, are more suitable
as options for solar repowering of new or existing fossil- fired power plants [17-22]. The parabolic
trough collector (PTC) is the most common medium-temperature solar technology. It represents the
most mature technology, from both commercial and technical viewpoints, for mid-to- large scale
130

grid connected power plants [23, 25]. Based on this technology a total installed electric capacity of
354 MW is generated from nine steam power plants in the California Mojave Desert [23].
In terms of cost and efficiency, integrated solar combined cycle systems (ISCCS) is the more
appropriate cycle design for MENAs countries [26]. An ISCC consists on the integrating of the
parabolic trough into a gas fired combined cycle composed of a gas topping cycle (GTC) and a
steam bottoming cycle (SBC). The PTCs can supply a hot heat transfer fluid (HTF) at a temperature
of about 400 C. Thus, the thermal energy carried by the HTF becomes the hot source of generating
additional steam to be sending to the SBC.
In last years, many studies had demonstrated the usefulness of exergy analysis in assessing the
performance of power plants [27-30]. The methodology has been proven in analyzing gas turbine
[31-33], steam turbine [34, 35] and combined cycle power plants [36-38].
In regard to CSPPs subjected to exergy analysis, the study [17] showed the benefits of a scenario
proposed for the solar repowering of a coal fired steam turbine. In another work [19], the best
options for solar boosting and fuel saving of two existing coal- fired power plants were determined.
In analyzing the hybridization solar-gas in ISCCs [22], a detailed examination was paid to the solar
collector, i.e. the linear Fresnel reflecting solar concentrator. The global optimal design of an
ISCCS was carried out based on the exergoeconomic approach [20]. Reference [18] performed an
exergy analysis of the first ISCCS in Iran. The plant is with full capacity 467 MW and the solar
power is about 17 MW. The main results are that the solar collectors are the least efficient
components in the plant and the combustors are the major destructors of exergy.
The present study presents the exergy analysis of the first ISCCS in Algeria. The Hassi RMels
plant has been lunched in operation few months ago. The analysis is based on the exergy flow
diagram and the value diagram, the later is an unusual visualization tool for identifying exergy loss
in heat transfer processes.
The thermodynamic simulations are performed by the flow-sheet program, Cycle-Tempo. This
software is a freeware advanced tool for the analysis and optimization of energy systems, developed
at the Delft University of Technology [39].

2. Power plant description and operation


The hybrid power plant is located at the Algerians largest natural gas field, Hassi RMel, province
of Laghouat, in middle of Algeria, at about 500 km from Algiers. The site is at 337 latitude and
321longittude, and its elevation above sea level is 750 m. The ambient temperature ranges
between 21 C and 50C in summer and ranges between -10 C and 20C in winter. In summer the
Direct Normal Irradiation (DNI) can reach 930 W/m2 . Based on yearly average value, the site is
blessed daily by 9.5 sunny hours offering DNI estimated at 7,138 Wh/m2 /day. The design of the
power plant considered air ambient at 0.928 bars and 35C with relative humidity at 24%. The
design solar output power is based on the average value of DNI considered as 751 W/m2 . This solar
irradiation intensity is intended to produce some 50 MW of thermal energy. At design running
conditions the full capacity of the plant is 160 MW.
The flow diagram of the plant is shown in Fig. 1. The plant is composed of a power block and a
solar field. The power block is a conventional combined cycle power plant with two 40 MW SGT800 gas turbine (GTs) [40] and an 80 MW SST-900 steam turbine (STs) [41]. The power block
contains also two identical single-pressure HRSGs with supplementary firing and no reheats. The
flow diagram shows the power block with only 1 GTPP and 1 HRSG. The HRSG is equipped by
low pressures economizer and evaporator (DECO and DEVA), two super heaters (SHE1 and SHE2)
and two duct burners (DBs). The first DB is integrated downstream of the GT exit to increase the
temperature of the exhaust gas that passes through the SHE1. The second DB is integrated into the
evaporator to compensate the deficit of solar steam in low sunny times. This DB is considered off
131

because the analysis is performed with the power plant running in design mode, with favourable
sunny conditions.
The GTPP is fuelled with Hassi RMels field natural gas. The gas is rich in CH4 by about 85%, it
has a LHV=45778 kJ/kg. The GTPP is supported by an inlet air cooling system to counter the
adverse effect of air temperature during hot times on its performance. Therefore, 4 Sadinter chiller
(CH) unites contribute to boost the GTPP by reducing intake air temperature to 15C. Although,
higher is the air temperature, higher is the power consumption of the chiller, this is usually
synchronous with higher solar irradiation (higher solar power output). An air cooled condenser
(ACC) is adopted as an option for condensing steam. This technology is preferable to the traditional
water-cooled condenser when water availability is limited, notably in Hassi Rmel, characterized by
an arid climate. The use of ACCs can reduce the plant water consumption by 90%; however, the
plant efficiency suffers with the higher condensing pressure. The SPX cooling system with heat
transfer duty of 147600 kW guaranteed by regrouping 15 fans, provides to the Hassi RMel power
plant statured water at 52 C and at about 0.14 bars. The design parameters of the power block are
summarized in Table 1.
At 183,120 m2 mirrors, the solar field comprises 224 parabolic collectors assembled in 56 loops, 4
collectors per loop. The collector is of ET-150 technology [24], aligned on a north-south line, it
tracks sun from east to west by a single axis tracking system. The HTF circulating in loop in the
solar field is synthetic oil; Therminol PV-1, its thermophysical properties versus temperature can be
found in Reference [42]. Cycle-Tempo assumes 3rd degree polynomial approximation for
calculating heat capacity of the HTF at any temperature. The supplying of thermal energy from the
solar field to the power block is performed when water/steam recovers the thermal energy from the
HTF through the SSG.
The SSG is the assembling of an economizer (ECO), an evaporator (EVA) with a drum (DR) and a
super heater (SHE).
The net output power of the plant is proportional to steam flow rate expanding in the ST, it is the
sum of the flow rate of steam generated in the HRSG and that generated in the SSG. The operation
of the hybrid plant is under the compulsory condition that the HTF circulates across the SSG at
constant inlet temperature, 393C, and at constant outlet temperature, 293C, but it can be with
variable mass flow rate. The high limit of temperature is imposed because the long-term exposure
of the organic HTF beyond 400C temperatures can lead to thermal decomposition of the fluid.
Thermal fluid decomposition occurs when enough heat is applied to the fluid to cause the breaking
of molecular bonds, which results degradation in the HTFs physical properties [43].
The solar steam flow rate is proportional to the HTF mass flow rate. The HTF mass flow rate varies
following the DNI intensity, in other words, according to time during day and to climate conditions.
The design value of the HTF mass flow rate is 200 kg/s, it is the resultant of DNI assumed at 751
W/m2 . This value of DNI is able to generate 22.6 kg/s of solar steam. Then, at the exit of the
economizer, 22.6 kg/s of pressurized water is withdrawn from the HRSG and is sent to the SSG.
After preheating and evaporating, the resultant saturated steam is superheated and resent to the
HRSG at 372C. The solar steam mixes with the steam exiting the SHE1, the whole steam passes
through the SHE2 before expanding in the ST. At design load the ST, with about 70 kg/s of
superheated steam at 560C and 80 bars, delivers 80 MW of electric power.

3. Exergy analysis
The exergy analysis of the power plant is based on the results of the thermodynamic simulations
performed by the flow-sheet program, Cycle-Tempo. The calculations considered the
simultaneous resolutions of mass equation and energy equation applied for each power plant
component. Cycle- Tempo also calculates exergy values of all fluid flows of the cycle by using the
ambient air (35C and 0.928 bars) as the reference state. Furthermore, the program calculates
exergy losses and efficiencies of all apparatuses. In the present study is used the functional exergy
132

definition where the exergy efficiency of any power plant component is determined as the ratio of
the exergy flow rate considered to be the product of the power plant component and the exergy flow
rate considered necessary for making this product [36].
Cycle-Tempo is able to draw Q-T diagrams and value diagrams.
Table 1. Design parameters of the power block.
Value
GT
Model
SGT-800
Ambient pressure
0.928
Ambient temperature
35C
Intake compressor air temperature
15 C
Compressor pressure ratio
20.2
Compressor isentropic efficiency
0.88
Inlet turbine temperature
1200C
Turbine isentropic efficiency
0.88
Exhaust mass flow rate
120,20 Kg/S
Exhaust temperature
550 C
LHV of natural gas
45778 kJ/kg
Net output power
40 MW
Thermal efficiency
35%
HRSG
Type
Fuel mass flow rate in the DBs
Approach temperature
Pinch temperature
Pressure losses in flue gas side
Pressure losses in water/steam side
Inlet water temperature
Exit stack temperature
Thermal efficiency
ST
Model
Inlet steam temperature
Inlet steam pressure
Steam mass flow rate
Condensate temperature
Isentropic efficiency
Full output capacity
SSG
Inlet water temperature
Inlet water pressure
Exit steam temperature
Water/steam mass flow rate
Inlet HTF temperature
Exit HTF temperature
HTF mass flow rate
Pressure losses in water/steam side
Pressure losses in HTF side
Thermal efficiency

Single pressure without reheat


0.66 kg/s
25 C
25 C
0.025
16 bars
60 C
100 C
98.50%

SST-900
560C
83 bars
70 kg/s
52C
0.9
80 MW
195C
93 bars
372C
22.60 kg/s
392C
292C
205 kg/s
5.8 bars
2 bars
98 %
133

3.1. Value and exergy diagrams


In order to illustrate and to present heat transfer evolutions, for example, in the HRSG, it is common
to use the so-called Q-T diagram. It shows profiles for the heat transfer process between exhaust gas
and water/steam, using temperature on the ordinate axis and heat transferred on the abscissa axis.
A better insight into the exergy losses due to heat transfer in the HRSG can be derived from the
value diagram [36].It is a graph that allows a simple presentation of how much is the recovering of
the exergy of the flue gas by water/steam flow?
In the value diagram the temperature of the flows are also given as a function of the heat transferred
to the water/steam flow; but the temperature at the vertical axis is replaced by the term (1 T0 /T). As
this axis begins at T0 and goes up to , the values on this axis can go from 0 to 1.
The term (1 T0 /T) indicates which part of the considered heat can in principle be converted into
work and can be seen as the exergy fraction of this amount of heat.
The total of exergy absorbed by water/steam flow is smaller than the exergy transferred from the
flue gas flow. The difference is exergy that is lost due to the temperature difference necessary to
transfer heat from the flue gas to water/steam flow.
The exergy flow diagram or Grassmann diagram is a convenient visualization tool to provide briefly
an overview of the most important information with regard to magnitude and location of
thermodynamic losses occurring through the power plant. It consists to plot out exergy flow, in the
conversion of natural resources (coal, gas, solar irradiation) into electricity, on a line diagram of the
flow system. This illustrates exactly the consumption or destruction of exergy in the power plant
components.

3.2. Solar field efficiencies


The solar field receives the incident solar irradiation energy:
,

(1)

where A is the total mirrors area.


The exergy input through this insolation is determined by the formula [19]:

(2)

the symbols T0 and Ts are, respectively, the ambient temperature and the temperature of the Sun
(5777 K), and f is the dilution factor ( 1.3 10 5 ).
The useful heat and the associated exergy transported by the HTF to the SSG are, respectively:
(3)
and
.

(4)

In these equations the symbols


,
, and
denote, respectively, the mass flow rate of the
HTF and its mass enthalpy difference and exergy difference through the SSG.
134

The global energy and global exergy efficiencies of the solar field are defined, respectively, as
follows:

(5)

and

(6)

The global efficiency includes the optical and thermal efficiencies of the solar field, in other words,
the efficiency of the global conversion of solar irradiation to heat to be injected into the SSG. The
optical efficiency depends upon incident angle effects, solar field availability, collector tracking
error and twist, geometric accuracy of the mirrors, mirror reflectivity, cleanliness of the mirrors,
shadowing of the receiver, transmittance of the receiver glass envelope, cleanliness of the glass
envelope, absorption of solar energy by the receiver, end losses, and row-to-row shadowing [44].
The thermal efficiency is function of receiver thermal losses and piping thermal losses. Receiver
thermal losses are caused mainly by the thermal radiation of the receivers selective coating. Piping
thermal losses corresponds to the thermal losses from the solar field header piping and the HTF
system piping.

4. Results and discussion


In Fig. 1, the flow diagram of the plant displays some important thermodynamic properties
(pressure, temperature, mass flow rate and mass enthalpy) at both entry and exit of every apparatus
at design running conditions. Cycle-Tempo calculations of absolute mass enthalpies of air, the flue
gas and the HTF present some shifts relatively to the exact values, however the enthalpy differences
are accurate because it is just a matter of the selection of the reference point. At design load, the
Hassi RMels power plant is able to deliver 160 MW of net output power, 80 MW from the GTPPs
and 80 MW from the ST. The SSG supplies 49,906 kW of the total superheated steam, 179,758 kW,
delivered by the HRSGs to the ST. The result is that the solar electric power is about 22 MW, and
thus the solar share is about 14%. The thermal efficiency of the power plant is 56%, while,
separately, the thermal efficiency of the GTPPS is 35 %. In reference to the conventional combined
cycle (138 MW and 48%) the solar contribution increases the output electric power by about 16%
and increases the thermal efficiency by 8 points. This fact is without computing the solar energy as
input energy in the definition of the thermal efficiency, thanks to the fact that solar irradiation is
free and inexhaustible; else, the thermal efficiency of the hybrid power plant is about 38%. That
decrease in efficiency associated to solar boosting is related to the global efficiency of the SF
limited at 78% (5), and to the thermal efficiency of the SSG, 98%.
For simplifying the analysis in the following, all of the apparatuses are categorized in four
technology families, i.e. turbomachines (ACs, GTs and ST), heat exchangers (SSG and HRSGs),
combustors (CCs and DBs) and the SF. The SSG and the HRSGs together have 17 heat exchangers
(HXs), 7 HXs per 1 HRSG and 3 HXs in the SSG.
Based on the exergy flow diagram depicted in Fig. 2; it appears that the exergy efficiency of the
plant is 53%; the efficiency of the GTPPs is about 34%. The solar exergy share in producing the
160 MW of electric power is about 12% (18.4 MW). Thus the solar exergy share is lower than the
energy share. The explanation is that the HRSG supplies thermal power to the ST at 560C while
the SSG sends thermal power to the HRSG at lower temperature 372C. From the exergy flow
135

diagram, the hybrid power plant is able to converting 397.4 MW of natural resources exergy, of gas
and sun, in electrical power, 160 MW. The rest, 234.5 MW (60%), is lost.
The percentage of the solar exergy input is 24% (96.9 MW), that of natural gas is 76% (300.5 MW),
from which 60% (237MW) are injected through the CCs and 16% (63.5 MW) are injected through
the DBs.
Gas

5
33 .00

17 .74

453 .44

1 17. 735

320 .65

232 .21

P el = 400 00. 32 kW

AC
0. 928 0

GT

35. 00
48

-1 16.5 6

0 .87 80

AIR
Ai r

1 200 .00

CC

117 .735

17 .54
120 .200

25. 00

2. 465 -3 990 .96

0.9 530

15 .00

5 50. 00

12 0.2 00 - 568 .78

117 .73 5 -1 36. 81

FLU EGAS

CH
5

Gas

33 .00

32

H TF

14. 00

392 .00

1 02.6 61

499 .59

DB

25.0 0

39

92. 10

3 04. 99

11. 310

27 39.3 6

SHE
31

51

87. 20

372 .00

11 .310

30 39.7 5

0.95 30

36

0.6 60 -39 90. 96

PMP
35

41

GAS

2 80

GTP P1

87.2 0

3 72. 00

34. 704

303 9.7 5

H
44

37

37

11. 310

13 16.1 5

92.1 0

3 04.9 9

11.7 82

273 9.36

379 .71

49

1 02. 661

466 .49

37 2.00

1 02. 661

309 .72

94. 10

19 5.00

11. 310

83 3.47

47

88 .20

301 .97

2 92.0 0

ECO

27

19 5.0 5

11

V alve

26

52

8 8.2 0

27 6.9 7

2 3.3 94

121 9.68

0 .95 30

25 9.4 1

12

16

h
= Ma ssf low [kg/ s]

94.1 0
34. 704

14 7.91

0.3 21

62 3.28

34. 704

63 4.66

0 .953 0

DEVA

13

P MP

27
2 01.9 1

7.0 00

60 .00

37 .880

25 1.7 3

CD

34

14 7.91

0.3 21

62 3.22

18

DECO
9

22
4 .500

4.5 00

14 0.00

34. 485

58 9.26

12
1 40. 00

3 7.8 80 5 89. 26

21

24

0.9 280

13

145 36. 204

7.0 00

14 0.0 3

3.3 95

58 9.5 6

PM P
20

11

HRSG#1

38

23

39

19

35. 00
-1 16.5 6

ACC

31

WATER

P MP
0. 928 0

16

-1 11.4 9
22

FAN

28

14

- 116 .56

40 .00

14 536 .204

AIR

4.5 00

35 .00

14 536 .20 4
0.9 280

25

Stack

0. 928 0

24

33

147 .91
623 .22

20 7.12

22 0
0.95 30

30

4.50 0
35.0 25

120 .86 0 -1 215 .91

12 0.8 60 - 122 1.5 5

DEA

15

26

P MP

29

WATER

el

14 9.23

25

83 3.47

5.0 00

STEAM

23

96. 10

19 5.00

2 743 .39

12 0.86 0 - 115 8.82

55

ECO1

14 7.91

0 .321
17

20

ECO2

4 .500
46

28

12 0.86 0 - 108 4.07

23 .39 4 83 3.4 7

h = En tha p
l y [kJ/kg ]
P = Electr ical Pow er [kW]

P MP

54

32 6.9 7

SSG

p = Pr es sur e [ bar ]
T = Te mpe rat ure [C]

EV A

19

89 .20

52. 00

DR

0 .95 30

43

0 .136 3

301 .90

10

SF

2 862 .38

3 4.48 5 2 352 .83

18

2 51.6 7

20 1.85

0 .219

1 28. 391 16 33. 54

1 28. 391 1 355 .70

H
12.0 0

4 .500

14

53

10 2.66 1

17

FLU EGAS
35

89 .70
29

ST

301 .90

58 3.4 9

12 0.86 0 - 788 .91


317 .06

36

33

88 .20

23 .39 4 27 45. 84

10
0 .95 30

12 .50

42

P MP

21

23 .39 4 3 039 .75

DR
45

87 .20

E VA

30
34

13 .50

40

6 31. 20

S HE1

9
2 95. 06

Pe l = 4 000 0.04 kW

50

STEAM
0.9 530

560 .00

12 0.8 60 - 732 .02

H
92. 10

83 .00

34 .70 4 3 543 .16

S HE2

38

7 50.3 8

12 0.8 60 - 587 .47

10 0.00

120 .860 -1 331 .01

32

7 .000

FLU EGAS

5 2.04

3 4.48 5 2 18. 47

0 .13 63

52.0 0

3 4.4 85

217 .69

Fig.1. Flow diagram of the ISCCS.


The loss of exergy is caused by both internal thermodynamic irreversibilities occurring in the
apparatuses and by exergy escaping into the environment.
The combustors are the major loser of exergy in the power block; they destruct 32% (95.7 MW) of
natural gas exergy input (300.5 MW). The CCs are responsible of about 24% (73 MW) and the DBs
of some 8% (22.7 MW). In regard to the total exergy input (397.4 MW), the turbomachines
(ACs+GTs+ST) cause the destruction of around 9% (33.7 MW) of exergy. All of the heat
exchangers (SSG+HRSG) consume 5% (21.9 MW) of exergy. The rest of exergy lost, 14. 5 MW
(3%), is rejected as heat into environment through the ACC (7.82 MW) and the stack (6.71 MW).
The lost in solar exergy is caused mainly by the SF which consumes about 74% (71.5 MW) of input
solar exergy (96.9 MW), the SSG is responsible only of less than 3% (2.7 MW). Thus, in the whole
77% of solar exergy is lost before arriving at the HRSG. That explains why the solar energy
increases the output electric power of the hybrid power plant but at lower efficiency (40%) when
taken in consideration the solar exergy as an exergy input in the definition of the exergy efficiency.
This definition of exergy efficiency is more pertinent when discussing the performance of the SF
and the SSG.
136

Since the ACC has only the function to discharge heat to the environment, similarly to the stack,
evaluating its performance does not have a significant importance.

Fig. 2. Exergy flow diagram of the ISCCS


In terms of exergy efficiency, as shown in Fig. 3, the turbomachines are the most efficient
components, theirs efficiencies are almost at the same level; they are between 90%, for the ST, and
94%, for the GTs, the ACs have 93%. In the second order appear the heat exchangers, but relatively
at different levels, at all the SSG operates at 89 % and the HRSGs operate at 79%. The combustors
are the least efficient systems in the power block, the CCs run at 68% and the DBs run at 64%. Far
behind, the global exergy efficiency of the SF is about 25% (6). Although its reactants are hotter,
the DBs are less efficient than the CCs. The explanation is that the combustion processes within the
DBs results flue gases at low temperature (750C), the consequence of being with air factor of 7.5,
well far of the stochiometric mixture air/fuel where the air factor is around 1.On the other hand,
with air factor estimated at 3 the flue gases leave the CCs at higher temperature (1200 C).
Close look to the performance of the HXs is given in the following. The analysis is based on both
the Q-T diagram and the value diagram which are combined in one graph. Figs. 4 and 5 show,
respectively, the Q-T/value diagram for the SSG and the HRSG.
Considering the SSG, the sizes of the shaded areas reveal that the EVA is the most efficient HX
while the ECO is the least efficient one, and the SHE is between both. The calculations confirm this
order. The exergy efficiencies of the EVA, the SHE and the ECO are about, respectively, 92%, 88%
and 79%. Any effort intending to enhance the performance of the SSG should be oriented in first
priority to the ECO, then in less degree to the SHE. The value diagram helps to identify that the
137

temperature difference at the cold end of the ECO should be the pertinent location for increasing the
efficiency. The potential of increasing the performance of the EVA is limited because its efficiency
is yet relatively high and because it is common to heat transfers associated to phase change to be at
lower efficiency.

Fig. 3. Exergy efficiencies of ISCCS components.


Examining the HRSG, from Fig.5, supported by calculations, it comes out that the DECO has the
smallest efficiency (56%). For the rest of HXs the efficiency ranges between 75 %, for the DEVA,
and 87%, for the ECO2. A look at water/steam and the flue gas temperature profiles shows that the
distance between both profiles varies slightly along the HRSG. Except for the EVA (phase change),
increasing the efficiency of the HXs requires the narrowing of this distance. The magnitude of this
distance can be related to the mass flow rates of both water/steam and the flue gas. In accordance to
sunny conditions, the HRSG runs at different regimes, then at different mass flow rates. In
consequence the shaded areas in the value diagram vary and the same for the exergy efficiency.
1

SSG

SHE

300

EVA
ECO

200
100
35

0
0

3.4

Transmitted heat [MW]

19.5

Fig. 4. Q-T/value diagram of the SSG.


138

25

Temperature [C]

1 - T0 / T [-]

600
500
400

Fig. 5. Q-T/value diagram of the HRSG.

Conclusion
The purpose of the present study is the thermodynamic analysis of the first hybrid solar- gas power
plant in Algeria. The Hassi RMels power plant has been lunched in operation few months ago.
The power block is composed of two SGT-800 gas turbines and a SST-900 steam turbine. The
HRSG is of single-pressure with supplementary firing and no reheat. The solar field gathers ET-150
parabolic trough collectors, in which circulates the heat transfer fluid Therminol PV-1.
The exergy analysis is adopted as a methodology for evaluating the performance of every power
plant component. The thermodynamic modelling and simulation, and also the post-processing of the
results are performed by the flow-sheet program Cycle-Tempo.
At design load the capacity of the plant is 160 MW of electrical power, 80 MW from the two gas
turbines and 80 MW from the steam turbine. Of this capacity the exergy solar share is lower than
the solar energy one, 12% (18.4 MW) against 14% (22 MW). The reason is that the solar steam
generator supplies about 50 MW of thermal power to the HRSG at 372 C, which is lower the
temperature of 560 C, at which the HRSG sends superheated steam to the steam turbine.
The thermal efficiency of the plant is 56% whereas the exergy efficiency is 53%. The combustors,
with efficiencies lower than 68%, are the least efficient components in the power block. They are
the major destructors of exergy; they consume about 32% of gas natural exergy. The turbomachines
(compressors, gas turbines and steam turbine) are the most efficient; their efficiencies are in the
range of 90% to 94%. The HRSG and the solar steam generator, equipped together by 17 heat
exchangers, have, receptively, the efficiencies 79% and 89%. With global exergy efficiency (optical
and thermal) limited to 25%, the solar field loses 74% of solar exergy.

Acknowledgments
The author is grateful to the management and staff of New Energy Algeria (NEAL SpA) and he
thanks especially Mr. Nordine Hamrene, Design Engineer, for providing the Hassi RMels power
plant data and for fruitful discussions. The author also thanks Mr. Teuss van der Stelt, from Delft
University of Technology, for giving help about Cycle-Tempo.

139

Nomenclature
A
e
h

mirror area
mass exergy, J/kg
mass enthalpy, J/kg

LHV

exergy flow rate, W


Lower Heating Value, J/kg
mass flow rate, kg/s

heat flow rate, W


temperature, C

Greek symbols
efficiency
Subscripts and superscripts
0
Ambient
s
Solar
Acronyms
AC
Air Compressor
ACC Air Cooled Condenser
CC
Combustion Chamber
CH
Chiller
CSPP Concentrating Solar Power Plant
DB
Duct Burner
DEA De-aerator
DECO low pressure Economizer
DEVA low pressure Evaporator
DNI Direct Normal Irradiation
DR
Drum
ECO Economizer
EVA Evaporator
G
Generator
GT
Gas Turbine
GTC Gas Topping Cycle
GTPP Gas Turbine Power Plant
HRSG Heat Recovery Steam Generator
HTF Heat Transfer Fluid
HX
Heat Exchanger
ISCCS Integrated Solar Combined Cycle System
MENA Middle East and North Africa
PMP Pump
PTC Parabolic Trough Collector
SBC Steam Bottoming Cycle
SF
Solar Field
SHE Super Heater
SSG Solar Steam Generator
ST
Steam Turbine

140

References
[1] Commission de Regulation de lElectricit et du Gaz (Electricity and gas authority of Algeria).
Annual Report 2010 (in French). Available at:<http://www.creg.org.dz> [accessed 12.6.2011].
[2] Himri Y., Arif S. Malik, Boudghene Stambouli A., Himri S., Draoui B., Review and use of the
Algerian renewable energy for sustainable development, Renewable and Sustainable Energy
Reviews 2009;13:1584-591.
[3] Boudghene Stambouli A., Promotion of renewable energies in Algeria: Strategies and
perspectives. Renewable and Sustainable Energy Reviews 2011;15:1169181.
[4] Boudghene Stambouli A., Algerian renewable energy assessment: The challenge of
sustainability. Energy Policy 2011;39:4507-519.
[5] Quaschning V., Technical and economical system comparison of photovoltaic and
concentrating solar thermal power systems depending on annual global irradiation. Solar
Energy 2004;44:171-78.
[6] Kane M., Favrat D., Ziegler K., Allani Y., Thermoeconomic analysis of advanced solar-fossil
combined power plants. Int. J. Applied Thermodynamics 2000;3(4):191-98.
[7] Horn M., Fhring H., Rheinlnder J., Economic analysis of integrated solar combined cycle
power plants. A sample case: The economic feasibility of an ISCCS power plant in Egypt.
Energy 2004;29:935-45.
[8] Dersch J., Geyer M., Herrmann U., Jones S. A., Kelly B., Kistner R., Ortmanns W., Pitz-Paal
R., Price H., Trough integration into power plants-a study on the performance and economy of
integrated solar combined cycle systems. Energy 2004;29:947-59.
[9] Hosseini R., Soltani M., Valizadeh G., Technical and economic assessment of the integrated
solar combined cycle power plants in Iran. Renewable Energy 2005;30:1551-555.
[10] Al-Soud M. S., Hrayshat E. S., A 50 MW concentrating solar power plant for Jordan. Journal of
Cleaner Production 2009;14:625-35.
[11] Tsikalakis A., Tomtsi T., Hatziargyriou N.D., Poullikkas A., Malamatenios Ch., Giakoumelos
E., Cherkaoui Jaouad O., Chenak A., Fayek A., Matar T., Yasin A., Review of best practices of
solar electricity resources applications in selected Middle East and North Africa (MENA)
countries. Renewable and Sustainable Energy Reviews 2011;15:2838849.
[12] Assessment of the World Bank/GEF strategy for the market development of concentrating solar
thermal Power. Global Environment Facility Program. The World Bank; 2006. Available at
<http://www.worldbank.org/gef > [accessed 12.6.2011].
[13] Heller P., Pfnder M., Denk T., Tellez F., Valverde A., Fernandez J., Ring A., Test and
evaluation of a solar powered gas turbine system. Solar Energy 2006;80:1225-230.
[14] Schwarzbzl P., Buck R., Sugarmen C., Ring A., Crespo, M. J. M., Altwegg P., Enrile J., Solar
gas turbine systems: Design, cost and perspectives. Solar Energy 2006;80:1231-240.
[15] Xu C., Wang Z., Li X., Sun F., Energy and exergy analysis of solar power plants. Applied
Thermal Engineering 2011;31:3904-913.
[16] Spelling J., Favrat D., Martin A., Augsburger G., Thermoeconomic optimization of a
combined-cycle solar tower power plant. Energy; Available online 29 April 2011.
[17] Hu E., Yang Y., Nishimura A., Yilmaz F., Kouzani A., Solar thermal aided power generation.
Applied Energy 2009;87 (9):2881-885.
[18] Baghernejad A., Yaghoubi M., Exergy analysis of an integrated solar combined cycle system.
Renewable Energy 2010;35:2157-164.
[19] Suresh M. V. J. J., Reddy K. S., Kolar A. K., 4-E (Energy, Exergy, Environment, and
Economic) analysis of solar thermal aided coal- fired power plants. Energy for Sustainable
Development, 2010;14 (4):267-79.
141

[20] Baghernejad A., Yaghoubi M., Exergoeconomic analysis and optimization of an Integrated
Solar Combined Cycle System (ISCCS) using genetic algorithm. Energy Conversion and
Management 2011;52:2193-203.
[21] Popov D., An option for solar thermal repowering of fossil fuel fired power plants. Solar
Energy 2011;85:344-49.
[22] Siva Reddy V., Kaushik S. C., Tyagi S. K., Exergetic analysis of solar concentrator aided
natural gas fired combined cycle power plant. Renewable Energy 2012;39:114-25.
[23] Price H., Lpfert E., Zarza E., Cohen G., Gee R., Mahoney R., Advances in parabolic trough
solar power technology. Journal of Solar Energy Engineering, 2002;124:109-25.
[24] Geyer M., Lpfert E., Osuna R., Esteban A., Schiel W., Schweitzer A., Zarza E., Nava P.,
Langenkamp J., Mandelberg E., EUROTROUGH - Parabolic trough collector developed for
cost efficient solar power generation. In: Steinfeld A., editor. Proceeding of 11th Int.
Symposium on Concentrating Solar Power and Chemical Energy Technologies; 2002
September 4-6; Zurich, Switzerland. Published by Paul Scherrer Institut:1-7.
[25] Fernndez-Garca A., Zarza E., Valenzuela L., Prez M., Parabolic-trough solar collectors and
their applications. Renewable and Sustainable Energy Reviews 2010;14:1695-721.
[26] Dersch J., Geyer M., Herrmann U., Jones S. A., Kelly B., Kistner R., Ortmanns W., Pitz-Paal
R., Price H., Trough integration into power plantsa study on the performance and economy
of integrated solar combined cycle systems. Energy 2004;29:947-59.
[27] Szargut J., Morris D. R., Steward F. R., Exergy analysis of thermal, chemical, and
metallurgical processes. Springer ; 1988.
[28] Kotas T.J., The exergy method of thermal plant analysis. Reprint edition, Krieger, Malabar, FL,
1995.
[29] Bejan A, Tsatsaronis G, Moran M, Thermal design and optimization. New York: Wiley; 1996.
[30] Kaushik S.C., Siva Reddy V., Tyagi S.K., Energy and exergy analyses of thermal power plants:
A review. Renewable and Sustainable Energy Reviews 2011;15:1857-872.
[31] Reddy B. V., Butcher C., Second law analysis of a natural gas- fired gas turbine cogeneration
system. International Journal of Energy 2009;33(8):728-36.
[32] Ehyaei M.A., Mozafari A., Alibiglou M.H., Exergy, economic & environmental (3E) analysis
of inlet fogging for gas turbine, power plant. Energy 2011;36:6851-861.
[33] Khaldi F., Adouane B., Energy and exergy analysis of a gas turbine power plant in Algeria. Int.
J. Exergy, 2011;9(4):399-13.
[34] Al jundi I. H., Energy and exergy analysis of a steam power plant in Jordan. Applied Thermal
Engineering 2009;29:324-28.
[35] Regulagadda P., Dincer I., Naterer G.F., Exergy analysis of a thermal power plant with
measured boiler and turbine losses. Applied Thermal Engineering 2010;30:970-76.
[36] Woudstra N., Woudstra T., Pirone A., van der Stelt T., Thermodynamic evaluation of combined
cycle plants. Energy Conversion and Management 2010;51:1099-110.
[37] Ahmadi P., Dincer I., Rosen M. A., Exergy, exergoeconomic and environmental analyses and
evolutionary algorithm based multi-objective optimization of combined cycle power plants.
Energy 2011;36:5886-898.
[38] Srinivas T., Reddy B. V., Gupta A. V. S. S. K. S., Parametric simulation of combined cycle
power plant: A Case Study. Int. J. of Thermodynamics 2011;14 (1):29-36.
[39] Cycle-Tempo, Release 5.0., Delft University of Technology. 2007. Available at
<http://www.cycle-tempo.nl.>

142

[40] Shukin S., Annerfeldt M., Bjorkman M., Siemens SGT-800 industrial gas turbine enhanced to
47MW. Design modifications and operation experience. ASME Turbo Expo 2008: Power for
Land, Sea and Air GT2008; 2008 Jun 9-13; Berlin, Germany. Proceedings of ASME:65-70.
[41] Steam turbines for solar thermal power plants. Siemens AG; 2008. Order No. E50001-W410A105-V1-4A00. Available at <http://www.siemens.com/ energy> [accessed 12.6.2011].
[42] Therminol-VP1. Technical Bulletin. Available at <http://www.therminol.com> [accessed
12.6.2011].
[43] Moens L., Blake D., Mechanism of hydrogen formation in solar parabolic trough receivers.
National Renewable Energy Laboratory; 2008 Feb. Technical Report NREL/TP-510-42468.
Available at<http://www.osti.gov/bridge> [accessed 12.4.2012].
[44] Assessment of parabolic trough and power tower solar technology cost and performance
forecasts. National Renewable Energy Laboratory; 2003 Oct. Subcontractor Report No.
NREL/SR-550-34440. Available at<http://www.osti.gov/bridge> [accessed 12.6.2011].

143

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

ENERGY RECOVERY FROM MSW TREATMENT


BY GASIFICATION AND MELTING TECHNOLOGY
F. Strobinoa, A. Pini Pratob, D. Venturac and M. Damonted
a Paul Wurth Italia S.p.A. Via di Francia 1, I-16149, Genova, Italy,
fabrizio.strobino@paulwurth.com,(CA)
b,c,d University of Genova, Department of Thermal Machines, Energy Systems and Transportation
Via Montallegro 1, I-16145, Genova, Italy, salabi@unige.it

Abstract:
The increase of waste production, joined to the difficulties concerning both the identification of new disposal
sites and the construction of big conventional incinerators, hardly accepted by the communities, led in recent
years to the development of new technologies for waste management.
The waste gasification and melting treatments, if compared with conventional incinerating methods, allow
reducing significantly the burdens on final disposal sites. Therefore gasifying and melting technologies are
attracting more and more the attention of academia and market operators.
Consequently, the possibility to introduce in the Italian context the Direct Melting System (DMS)
technology, designed and manufactured by Nippon Steel Engineering Co. Ltd., has been taken into account
for the scope of proposed work.
DMS technology consists in MSW gasification, slags melting and combustion of the syngas produced, with
the consequent generation of electric energy through a steam cycle. The system does not simply treat
wastes; in fact it minimizes environmental impact because there is an effective recycling of useful resources
since the melted slags are vitrified and inert and then still reusable, particularly for the most demanding tasks
in terms of leachability such as road pavement. DMS technology has also the real advantage of owning a
modular design. This aspect allows greater degree of flexibility in terms of matching communities needs,
bringing therefore a significant advantage during the phase of site identification.
The aim of this article is to consider different plant configurations in order to optimize the energy recovery
downstream the DMS module.
As a case study also landfill gas exploitation integrated in the DMS plant will be considered as a typical
situation that could occur in the Italian scenario. The energetic input provided by the biogas is generally
interesting because it allows improving the thermo-economic performances also thanks to market incentives.

Keywords:
DMS, gasification, landfill gas, melting, municipal solid waste, recycling, slag.

1. Introduction
Disposal of municipal solid waste is a problem that every local government has to face. Waste
volume annually produced has been increasing almost constantly in the various European States [1].
In densely populated areas proper disposal of a considerable mass of waste and the need of large
areas dedicated to the disposal are two important problems.
The EU has ruled issuing directives under which the legislative authorities of each M ember State
should act with regard to waste management [2]. These principles address to the so-called
integrated waste management that can be pursued by: reducing the amount of waste produced,
increasing the recycling and reuse, recovering energy from the not recyclable fraction and reducing
the volume of waste conferred to landfill without having undergone any treatment [3, 4]. In fact
landfilling of unsorted waste represents both a waste of resources, and a potential source of
environmental contamination.

144

On the one hand exists the energy recovery maximisation goal, which leads to the hypothesis of a
few large size plants network able to reach an highly efficient electric conversion, on the other hand
the ethical-economic-environmental principle of zero kilometer waste leads to the hypothesis of a
smaller size distributed waste-treatment plants network.
In the field of the innovative technologies for the municipal waste treatment, DM S technology has
attracted strong interest. The technology has been designed by Japanese Nippon Steel Engineering
Co. Ltd. and DM S means Direct M elting System.
DM S technology is very competitive because, due to modularization, allows small and medium size
treatment plants construction, performing energy recovery with overall efficiencies in line with
similar size conventional technology plants [5].
This article is reporting the preliminary study carried out on some technical improvements, starting
from the DM S original solution designed by Nippon Steel Engineering, aiming to the increase of
electricity production in order to suit better on a specific Italian case, suggested by Paul Wurth Italia
S.p.A.
The case study represents a potentially typical plant scheme for the Italian scenario and therefore
proposed solutions are expected to be generally implementable in order to match Italian electric
market requirements.

2. Italian regulatory framework


The current Italian legislation [6] provides an incentive scheme for plants that produce electricity
from renewable sources (IAFR) based on the awarding of green certificates (CV). The weighed
average value of such CV for the year 2011 was of 80 /M Wh [7].
DM S technology is among the ones that can receive this incentive for the share of electricity
produced by the biodegradable fraction of M SW and the said share currently is conventionally set
by the law at 51% [6].
The Italian law also recognizes a so called all-inclusive tariff, currently amounting to 180 /M Wh
[6] and paid directly by the GSE, for electricity produced from renewable sources that do not
exceed the annual average power of 1 MW with the exclusion of all plants that use the source of
solar and wind energy for which the limit is 200 kW. This category includes many systems that
exploit the landfill biogas.
Electricity production in the Italian context turns out to be an important item for the cash flow of
waste to energy plants.
In this sense, the study focused on increasing the overall electric efficiency and therefore on the
increase of electric power production.

3. DMS technology
The technology has been derived from a thermal process used so far in the metallurgical industry,
which has been extended to the urban waste treatment. The process (Fig.1) consists in M SW
gasification and slags melting in a shaft furnace. The system provides a high degree of flexibility in
treating different types of wastes both combustible and incombustible. This includes wastes that
cannot be recycled at the collection level and those unsuitable for conventional technology based
waste to energy plants, such as incineration residues, sludge, automobile shredded residues, landfill
reclamation waste, CFC gas and asbestos [8, 9]. This technology can be defined safe and reliable
because 35 plants are nowadays in commercial operation (since the first installation in 1979) [10].
The process consists in a combined treatment of waste gasification, combustion of the syngas
produced and slags melting. The gasification and melting occur within the same equipment that
consists in a shaft furnace where waste is introduced from the top (after size reduction) together
145

with coke and limestone in percentage of around 4-5% of the waste input quantity (depending on
the characteristics of the waste) [11]. The slags are discharged through the bottom of the furnace
and granulated by a cooling system.
The syngas generated in the DM S is emitted from the top of the furnace and completely oxidized in
the post-combustion chamber, by controlling the temperature, the turbulence and the residence time.
Thanks to the homogeneous conditions of combustion which is taking place between gaseous
phases, the pollutants generation is limited in comparison to conventional mass burn incineration
and the plant, equipped with state-of-the-art flue gas treatment line, can achieve emissions into
atmosphere that are by far lower than regulatory limits and fully in compliance with BREF both in
terms of pollutants concentration and volume of flue gases [12].
The exhaust flue gas heat can be used to produce electricity by a recovery boiler and a steam cycle
[13].
The system does not simply treat wastes; in fact it minimizes environmental impact because there is
an effective recycling of useful resources since the melted slags are vitrified and inert and then still
reusable, particularly for the most demanding tasks in terms of leachability such as road pavement.
Even metal produced as final output of DM S can be reused as a resource [14]. At the end of the
process, the only residue to landfill is made up from the fly ashes caught by the filters and generally
it amounts at about the 2-3% of the waste input [14].
The system has generally a modular design, with multiple melting furnaces working in parallel,
which allows flexibility and reliability, while the auxiliary systems can be grouped as a single
module. Such an aspect allows greater degree of flexibility also in terms of matching communities
needs, bringing therefore a significant advantage during the phase of site identification.

Fig. 1. Process diagram.

146

4. Case of study
The plant will have to treat 60,000 t/y and the average composition of wastes is reported in
Table 1. The mean lower calorific value (LCV) of the feeding mixture is 10,691 kJ/kg and the heat
flow diagram for the baseline solution is reported in Fig. 2. The area identified for the construction
of the plant is located near to a landfill that provides a flow of biogas which should not be dispersed
into the atmosphere and therefore has to be used for power generation. The average biogas mass
composition consists of approximately 49.9% of methane, 27% carbon dioxide, 10.8% nitrogen,
9.5% moisture, 1.86% oxygen and argon, plus traces of other compounds.
Table 1. MSW average composition.

Fig. 2. Heat flow diagram.


147

It can be noticed that the availability of the biogas will decrease from a value of 1,000 Nm3 /h till
406 Nm3/h (Fig. 3). Both mean biogas and waste composition values are the result of experimental
and statistical data processing.
In order to use the energy released by the combustion of biogas at high temperature, it was chosen
to apply only external components outside DM S to avoid the redesign of the module and to
maintain independent from the actual biogas availability the superheated steam output parameters
outgoing the module.

Fig. 3. Estimated biogas production.

5. Plant solutions
The steam cycle shown in Fig. 4 is the basic configuration, designed for heat recovery from the flue
gas coming out from the DM S. The boiler consists of an evaporator (EVA) a superheater (SH1) and
an economizer (ECO). The steam generation does not occur only in the recovery boiler, but also in
the bottom of the combustion chamber covered by boiler pipes.
The flow rate coming from the DM S is about 48,400 Nm3/h and it is available at a temperature of
1,100C and it exchanges heat with steam till 200C.
In the superheater the steam reaches a maximum temperature of 428C. The steam cycles subjected
to DM S technology in Japan, reach a maximum temperature of 400C [15] and this choice is mainly
due to two reasons:
in Japan there is only moderate interest in producing electricity from waste to energy plants
because there are not incentives, hence the lack of need to reach high performances;
steam temperatures above 430C may cause problems of corrosion of pipe bundles [16].
In Italy electric generation efficiency must be maximized to exploit the beneficial contribution of
market incentives.
Considering on the one hand minimum steam quality (in terms of water to vapour ratio) accepted by
steam turbine manufacturer at the end of the expansion and on the other hand maximum
temperature allowed by material resistance to corrosion, the implemented thermodynamic cycle
optimization process has led to the selection of the steam condition equal to 428C and 65 bar at
turbine inlet.
The search for a further performance increase would lead to reach higher values of pressure, but this
would necessarily require to adopt a reheater which is not justified given the small size of the plant
[17].
148

In the gas treatment line, there are two heat exchangers which aim to preheat the feed water and at
the same time to cool the flue gases coming from the DM S, before they are ejected from the stack.
With regard to the condenser, the choice falls on the air-cooled one for two main reasons:
the use of water as coolant needs complicated path authorization;
the air-cooled condenser is not tied to the availability of water.
M oreover the benefits, in terms of cycle efficiency, obtained by using water instead of air are
negligible because, in each of the two cases, it is necessary to opt for a different cycle pressure. The
maximum steam cycle pressure allowed when using water as coolant is 45 bar because the increase
of this pressure would compromise the steam quality at the end of the expansion.

Fig. 4. Steam cycle Basic configuration.

The second plant configuration (Fig. 5) aims to exploit the biogas produced in the landfill. The fluegases generated by the biogas combustion are not critical with regard to pipes bundle corrosion as
DM S ones, so it is possible to realize an external superheater (SH2) which can further increase the
steam temperature above 428C. Thus in an external component, i.e. separated from the main fluegas stream proceeding from the DM S, it is possible to reach higher temperatures without any
problems due to the pipe bundles corrosion.
Flue gases generated by biogas combustion are available at a temperature of about 680C and they
are cooled to 460C. With this energetic input it is possible to reach a temperature of about 470C
(with an average biogas flow rate of 700 Nm3/h); this value is far from the typical temperature of
about 540C which characterizes the steam cycle of fossil fuelled power plants, but that is because
the biogas flue-gas flow rate is small if compared to the flow rate coming out from the DM S.
149

Heat remaining in the biogas flue gases is then used in a heat exchanger to preheat the feed water.
The increase in temperature, both of the feed water and of the superheated steam, can increase the
overall electric efficiency and the power production.
The possibility to exploit the biogas directly into the DM S module was discarded in principle
because the post-combustion chamber and the boiler redesign and optimization would not be
justified given the uncertainty about the availability of biogas, hence the need for the use of a
configuration with an external component.

Fig. 5. Steam cycle Configuration with external superheater.

Another plant solution with an external component (Fig. 6) has been taken into account. The biogas
can be used in an external evaporator, to produce more steam. The flue gases coming from the DM S
module are cooled till about 310C and the maximum temperature reached from the steam is 428C
like the base configuration. Also with this configuration, is possible to use the heat remaining in the
biogas flue gas coming from the external evaporator, to preheat the feed water.

150

Fig. 6. Steam cycle Configuration with external evaporator.

The last plant solution (Fig. 7), considered in present work, consists in coupling DM S plant with a
biogas internal combustion engine. Part of the biogas (about 430 Nm3/h) is burned in the engine
with an electric efficiency of 33% and it is possible to produce about 700kW. The biogas flue gases,
at a temperature of 500C are still usable in a heat exchanger with the task to further preheat the
feed water. The remaining part (about 280 Nm3/h) can be recovered in the external component,
evaporator or superheater. In this case, the plant solution is very flexible and also competitive from
an economic point of view, because the electric power generated from the engine awards
considerable incentives.
In most plants, the feed water degassing is made by means of a steam bled from the turbine. On the
contrary, the plant here described proposes a solution that does not require to bleed any steam from
the turbine, in order not to penalize the work obtained from the expansion with the aim to maximize
the energy production. In this solution (Fig. 8), a portion of the feed water flow rate, is bled at a
pressure of 65 bar (cycle pressure) and sent to a heat exchanger where it can exploit the heat of the
flue gases coming from the recovery boiler. The water reaches a temperature T x with an enthalpy
hx=hL(65 bar, T x); then it expands maintaining the same enthalpy. The steam produced in this way,
at an enthalpy condition hx can degas and preheat the water coming from the condenser.

151

Fig. 7. Steam cycle Configuration with external superheater coupled with ICE

Fig. 8. Degassing system


152

6. Results
Results obtained from analysis are reported in Table 2. All results are referred only to the DM S
plant, so in the case of presence of the internal combustion engine (ICE) (with a biogas flow rate of
430 Nm3/h), the electrical power output has to be increased by 700 kWhel.
Table 2.
Configurations

Biogas mass flow


rate recovered by
DM S [Nm3/h]

Basic configuration
Cycle with external SH
Cycle with external SH and ICE not coupled
Cycle with external SH and ICE coupled
Cycle with external EVA
Cycle with external EVA and ICE not coupled
Cycle with external EVA and ICE coupled

700
270
270
700
270
270

DM S net
electric
power output
[kWhel/t MSW ]
626
714
663
686
701
657
679

DM S net
overall
electric
efficiency 1[%]
18.52
18.78
18.66
18.91
18.43
18.49
18.71

In the configurations with DM S and ICE coupled, the DM S net overall electric efficiency is better
than the case of all biogas recovered in the external SH/EVA. In fact the largest part of biogas
power input is exploited in a higher efficiency converter like ICE and DM S receives only as direct
energy input the ICE waste heat and a little part of biogas flow.
Basic configuration results are included in the typical performance range of 400700 kWhel/t MSW
for commercial gasification technologies [18]. Other solutions overcome this range up to +10% due
to the landfill biogas combustion contribute (directly or indirectly). To note that steam cycles
subjected to the DM S technology of Japanese operating plants, with steam at a pressure of 40 bar
and a temperature of 400C reach a net overall electric efficiency of approximately 16% (194,000
t MSW /y) [12].

7. CHP option
Coupled DM S and biogas engine solutions can achieve the best electric performances, so it was
interesting for these solutions to analyze the performance sensitivity versus the heat cogeneration. It
was examined a solution using a steam bleed from the turbine at a pressure of 4 bar and a
temperature of 146C to heat a pressurized water flow from 80C to 120C which can feed a district
heating network.
These coupled solutions allow to maximise the electric production even exploiting a thermal load.
In fact, owning to the electric production incentive frame that induce not to regulate, if the ICE
works independently, it will dissipate part of the exhaust gas heat when the district-heating load is
under the effective ICE production capacity. In this sense coupled solutions offer a more flexible
ratio between electrical power output and thermal power output than not coupled ones by reason of
the variable steam bleed.

DMS net overall electric effi ciency is obtained by the following ratio:

153

Table 3.
Configurations
Cycle with external SH
and ICE coupled
Cycle with external EVA
and ICE coupled

DM S electric
production
[%]
100
90
80
100
90
80

Net electric
power
[kWhel/t MSW ]

Thermal power
[kWhel/t MSW ]

686
617
549
679
612
544

0
278
552
0
274
548

DM S net
overall electric
efficiency [%]
18.91
17.01
15.14
18.71
16.86
14.98

20%
660

[kWh /t

18%

620

16%

580

14%

540

12%

500

10%
0

100

200

300

40 0

500

[kWh th/t M SW ]
DMS Net Electric Power vs. Thermal Powe r
DMS Net Overall Electric Efficiency vs. Thermal Power

Fig. 9. DMS net electric power production and DMS net overall electric efficiency vs. Thermal
power generation diagram for Cycle with external SH and ICE coupled

20%
660

18%

[kWh /t

620

16%

580

14%

540

12%

500

10%
0

100

200

30 0

400

500

[kWhth /t M S W ]
DMS Net E lec tric P ower vs. Th ermal Power
DMS Net Overall E le ctric E fficiency vs. Thermal P ower

Fig. 10. DMS net electric power production and DMS net overall electric efficiency vs. Thermal
power generation diagram for Cycle with external EVA and ICE coupled
According to the steam bleed progressive increase, electric performances decrease linearly with the
increase of the heat production.
DM S CHP option is particularly suitable because if necessary it achieves up to 550 kWh/t MSW of
thermal power in respect of a reduction of overall electrical efficiency of only 4 percentage points.
154

CHP option is generally very suitable to improve environmental and economical performance of
waste to energy plants compensating for their lower overall electric efficiency, caused by technical
constraints, and their higher specific capital costs [19].
Furthermore, a significant aspect is that with a bleed of steam corresponding to a thermal power of
about 440 kWh/t MSW it is possible to produce approx. 26 GWhth/y. This thermal energy corresponds
to the heating requirements of 5,000 inhabitants in climatic zone E2 with 2850 degree-days [20]. It
thus demonstrates the versatility of the DM S in particular to serve small communities needs on the
territory, while maintaining the electric standard efficiency usually accepted in Japan.

8. Conclusions
Italian and European laws prescribe to increase recycling and reuse and to decrease waste
production and landfilling. The share of non-recyclable waste must be treated in plants for energy
recovery in order to reduce volumes landfilled and reduce, as much as possible, danger of
environmental contamination.
In Italy, the construction of these facilities is supported by an incentive that rewards the production
of electricity.
The techniques to reach high conversion efficiencies to electricity require the realization of large
size plants. These need during their life cycle to be fed with a waste flow that often comes from
areas very far from the plant. This factor, together with the largest impact on the surrounding
environment, often causes opposition of local people to their implementation.
M oreover recycling and reducing policy should ensure that the amount of unsorted waste decreases
over the years, making difficult to operate the plant at design treatment capacity.
The present study considered the application of DM S technology in a particular Italian scenario
requiring landfill biogas exploitation.
The results lead to say:
DM S technology reaches high value of overall net efficiency (all solutions exceed 18%),
considering a small plant size (60,000 t MSW /y);
coupling DM S with external SH or EVA and ICE can maximize revenue from electric energy
production incentives even in CHP configuration.
M ore generally DM S technology energy recovery performances result aligned with other non
conventional waste to energy technologies, while at the same time DM S features the peculiar
advantage of minimizing the volume of material landfilled thanks to the slag and metal reuse. This
fact represents a significant improvement in economical and in environmental terms. If the site
chosen for construction offers a biogas flow rate, DM S technology can improve its performances by
external components and if coupled with an ICE it can guarantee a very flexible electrical power vs.
thermal power ratio in CHP configuration.
In particular, concerning environmental performances, the basic configuration examined in present
work generates emissions into atmosphere that are by far lower than regulatory limits and analogous
to those of modern conventional incinerators equipped with efficient flue gas cleaning systems
according to BREF, both in terms of pollutants concentration and volume of flue gases [12].
At this stage of the study, as regards both the external evaporator configuration and the external
superheater configuration, the pollutants incremental contribution to the DM S exhaust flow gases
was not deemed significant, considering both biogas combustion gases small volume and low
2

According to Law Decree D.P.R. N. 412 (1993) the Italian territory is divided into 6 climatic zones with the purpos e
of referen ce for the estimation of space heating systems average en ergy consumption. These climatic zones are named
with letters from A to F in function of number of degree-days. Zone A includes the municipalities featuring a number
of degree-days up to 600 whilst Zone F includes the municipalities featuring a number of degree-days greater than
3,000.

155

temperature due to cooling effect provided by high combustion air excess. In addition, these two
conditions of combustion and heat exchange seem to have no more critical elements than ICE
combustion that is deemed fully environmental acceptable in similar applications.
Finally, concerning the investment, previous studies have shown how, although the insertion of
components such as external SH and EVA in addition to DM S baseline configuration or CHP
option increase the initial investment amount, the Net Present Value and the Internal Rate of Return
of the project are improved by the increase of revenue items [21].
Nevertheless economically best solution, among the ones discussed in this preliminary study,
depends strongly on specific scenario and in particular on available biogas flow amount, heat
demand and site characteristics.

Nomenclature
BREF
CHP
CV
DMS
d.b.
ECO
EGR
ESP
EVA
EVA1
EVA2
GSE
h
hL
IAFR
ICE
LCV
mi
mr
ms
MSW
mv
SCR
SH1
SH2
Tamb
Tcond
Tfb in
Tfb out
Tsat
TSH1

Reference Document on the Best Available Techniques for Waste Incineration (European
Commission Integrated Pollution Prevention and Control)
Combined Heat and Power
Certificato verde (Green Certificate)
Direct M elting System
dry basis
Internal economizer
Exhaust gas recirculation
Electrostatic Precipitator
Internal evaporator
Internal evaporator
External evaporator
Gestore Servizi Energetici, Italian state owned energy service company
enthalpy, kJ/kg
enthalpy in liquid phase conditions, kJ/kg
Impianto Alimentato da Fonti Rinnovabili (Plant Powered by Renewable Sources)
Internal Combustion Engine
Lower Calorific Value
incondensable mass flow rate, equal to mr, kg/s
reintegration mass flow rate, 2% of mv, kg/s
feed water mass flow rate bled, kg/s
M unicipal Solid Waste
mass flow rate at turbine inlet, kg/s
Selective Catalytic Reactor
Internal superheater
External superheater
ambient temperature, C
steam condensing temperature, C
biogas flue gases temperature at external superheater inlet, C
biogas flue gases temperature at external superheater outlet, C
steam saturation temperature, C
superheated steam temperature at boiler outlet, C
156

TSH2
Tx
Ty
w.b.

superheated steam temperature at external superheater outlet, C


degassing system inlet temperature
ECO inlet steam temperature, C
wet basis

References
[1]
[2]
[3]
[4]

[5]
[6]

[7]
[8]

[9]

[10]
[11]

[12]

[13]

[14]

[15]

ISPRA, Rapporto rifiuti urbani edizione 2009, Roma, 2010 Available


at:<http://www.isprambiente.gov.it/site/it-IT/Pubblicazioni/Rapporti/> (in Italian).
European Directive 2006/12/CE, 5/04/2006.
Arena U., Gassificazione dei rifiuti solidi urbani: stato dellarte dei principali processi e
tecnologie. RS Rifiuti Solidi vol. XXVI n. 1 Gennaio-Febbraio 2012 (in Italian).
Arena U., Advantage aspects of combustion and gasification processes for thermal treatment
of municipal solid wastes, International waste management and landfill symposium, Sardinia,
Italy; 2011.
Belgiorno V., De Feo G., Della Rocca C., Napoli R.M .A., Energy from gasification of solid
wastes. Waste M anagement 2003; 23: 1-15.
M inisterial Decree, 18/12/2008, Incentivazione della produzione di energia elettrica da fonti
rinnovabili, ai sensi dellarticolo 2, comma 150, della legge 24/12/2007 n. 244, Gazzetta
ufficiale del 02/01/2009 (in Italian).
Gestore dei mercati energetici, M arket data available at < www.mercatoelettrico.org > (in
Italian).
Tanigaki, N., M otoyama K., Osada M ., Jung B., Lee Y, Operating performance of gasification
and melting facility in Korea, Proceedings of International conference on Combustion,
Incineration/Pyrolysis and Emission Control (i-CIPEC 2008), Chiang M ai, Thailand; 16-19
December 2008: 416-421.
Tanigaki N., M anako K., Osada M ., Co-gasification of municipal solid waste and material
recovery in a large-scale gasification and melting system. Waste M anagement 2012; 32: 667675.
<http://www.nsc-eng.co.jp>.
Osada M., M urahashi K., Shibaike H., Characteristics of gasifying combustion in M SW
Direct M elting System, Proceedings of International conference on Combustion,
Incineration/Pyrolysis and Emission Control (i-CIPEC 2000), Seoul, Korea; 8-10 June 2000:
101-107.
Genon G., Tedesco V., Urso P., Assessment of the feasibility of an innovative technology
plant aimed at the energetic valorization of municipal waste in the province of Turin, Third
Symposium on Energy from Biomass and Waste, Venice, Italy; November 2010.
Chiappero A., La tecnologia DM S per la gassificazione dei rifiuti solidi urbani. Le tecnologie
per il recupero di energia da rifiuti a confronto. Politecnico di M ilano sede di Piacenza, 18
June 2008 (in Italian).
Takamiya K., Osada M ., Shibaike H., Kajiyama H., Reclamation waste treatment using a
gasification and melting system, International waste management and landfill symposium,
Sardinia, Italy; 2007.
M anako K., Kashiwabara T., Kobata H., Osada M ., Takeuti S., M ishima T., Dioxins control
and high-efficiency power generation in a large-scale gasification and melting facility,
Proceedings of 27th International Symposium on Halogenated Persistent Organic Pollutants
(DIOXIN 2007), Tokyo, Japan; 2007.

157

[16] Berlo M . V., Simoes P, High efficiency waste-to-energy: Amsterdams experiences after 4
years of operation, International waste management and landfill symposium, Sardinia, Italy;
2011.
[17] Genon G., Tedesco V., Urso P., Verifica della fattibilit di un impianto di trattamento termico
dei rifiuti a tecnologia innovativa nella Provincia di Torino. M ilan: CIPA Editore, 2010 (in
Italian).
[18] Arena U., A critical assessment of municipal solid waste gasification, International waste
management and landfill symposium, Sardinia, Italy; 2011.
[19] Consonni S., Giugliano M ., Grosso M ., Alternative strategies for energy recovery from
municipal solid waste Part A: M ass and energy balances. Waste Management 2005; 25: 123135.
[20] Pini Prato A., Strobino F., Broccardo M ., Parodi Giusino L., Integrated management of
cogeneration plants and district heating networks. Applied Energy 2012, article in press doi:
10.1016/j.apenergy.2012.02.038.
[21] Damonte M ., Ventura D., Ottimizzazione del recupero energetico da RSU mediante
tecnologia DM S. M aster thesis. Universit di Genova, Italy, M arch 2011 (in Italian).

158

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Ethanol production by enzymatic hydrolysis from


sugarcane biomass the integration with the
conventional process
Reynaldo Palacios-Berechea , Adriano Ensinasb , Marcelo Modestoc, and Silvia Nebrad
a

Centre of Engineering, Modelling and Social Sciences, Federal University of ABC (CECS/UFABC),
Santo Andr, SP, Brazil, reynaldo.palacios@ufabc.edu.br
b
Centre of Engineering, Modelling and Social Sciences, Federal University of ABC (CECS/UFABC),
Santo Andr, SP, Brazil, adriano.ensinas@ufabc.edu.br
c
Centre of Engineering, Modelling and Social Sciences, Federal University of ABC (CECS/UFABC),
Santo Andr, SP, Brazil, marcelo.modesto@ufabc.edu.br
d
Centre of Engineering, Modelling and Social Sciences, Federal University of ABC (CECS/UFABC),
Santo Andr, SP, Brazil, and Interdisciplinary Centre of Energy Planning, University of Campinas
(NIPE/UNICAMP), Campinas, SP, Brazil, silvia.nebra@pq.cnpq.br

Abstract:
The aim of this study is to make an evaluation of the possibilities of ethanol production increas e through the
introduction of bagasse hydrolysis process in conventional distilleries, considering the limiting situation of
bagasse use: it is the major by-product in sugar and ethanol production and is burnt in boilers t o satisfy the
steam and power requirements of the process. Simulations in ASPEN PLUS soft ware were performed, in
order to evaluate the mass and energy balances, for the integrated process, considering the pre-treatment of
sugarcane bagasse by steam explosion. The cogeneration system was also modelled and integrated with
the ethanol production process. It consists of a steam cycle with backpressure steam turbines and
parameters of live steam of 67 bar and 480C. In all the cases studied it was considered that the steam flow
used in the system was just that necessary to fulfil the process thermal needs, so, it was assumed that the
surplus of bagasse was used to produce ethanol. The use of sugarcane trash was considered in order to
accomplish the energetic needs of the overall process as well as lignin cake, which is a hydrolysis process
residue. Several cases were evaluat ed, which include: the conventional ethanol production plant without
hydrolysis (Case I), the conventional plant joint with hydrolysis process without thermal integration
considering different solid contents in the hydrolysis reactor (Cases II, III and IV ), and the conventional plant
joint with the hydrolysis process considering t hermal int egration through Pinch method (Case V). The res ults
shown a modest ethanol production increase of 9.7% for the situation without thermal integration and low
solid content in the hydrolysis reactor, on the other hand, the c ase where t hermal integration was applied
presented an ethanol production inc rease of 17.4%.

Keywords:
Ethanol, sugarcane, enzymatic hydrolysis, thermal int egration.

1. Introduction
Ethanol is produced in Brazil in large scale using sugarcane as raw material by fermentation of
sugars and distillation. World consumption of ethanol tends to grow in the next years because of the
growing interest of many countries by biofuels use, due to factors such as: environmental damage
(avoided emissions of greenhouse gases) energy security (diversification of energy sources and
reducing dependence on oil) and support to farmers [1]. The sugarcane bagasse is the major byproduct in sugar and ethanol production and it is burnt in boilers to satisfy the steam and power
requirements of the process. Sugarcane bagasse, as well as other lignocellulosic materials, can be
also used for ethanol production but, the introduction of the bagasse hydrolysis process in the
current ethanol production system is a real challenge, being bagasse the fuel of the current process
and at the same time, raw material for the new one. Thus, the aim of this study is to make an
evaluation of the possibilities of ethanol production increase through the bagasse hydrolysis
159

process, considering the limiting situation of bagasse use. Simulations in ASPEN PLUS software
were performed, in order to evaluate the mass and energy balances and Pinch Analysis was used to
determine the minimum hot and cold utilities required by the integrated process in order to increase
the ethanol production. The characteristics of the cogeneration system were adopted considering
devices currently used in the new industrial plants, and prioritising the increase in ethanol
production, and not electricity cogeneration.

2. Description and modelling of ethanol and electricity


production process from sugarcane
2.1. Conventional ethanol and electricity production process from
sugarcane
In this study, a plant producing anhydrous ethanol and electricity was considered, using sugarcane
as raw material.
Sugarcane is composed essentially of juice and fibres. The juice is an aqueous solution of sugars
(meanly sucrose and small fractions of fructose and glucose monosaccharaides) and other organic
and inorganic substances (minerals and impurities). Fibre is defined as all insoluble material in
cane. The composition of sugarcane stalks depends on a large number of factors, including the age
of the cane, growing conditions, disease, while the composition of sugarcane delivered to the
factory depends on stalk composition, cane variety, age, amount of tops and leaves and other
exogenous matter carried in harvesting operation [2].
The sugarcane composition adopted in this study is shown in Table 1. Due to the fibre components
cellulose, hemicellulose and lignin, are not in database of ASPEN PLUS simulator, these
components were created and their properties were specified using parameters reported in [3].
Table 1. Sugarcane composition specified in simulation [4]
Component
Sucrose
Fibres
Reducing sugars
Minerals
Others non saccharides
Water
Soil

% Mass
13.85
13.15
0.59
0.20
1.79
69.35
1.07

In sugarcane factory the process begins with the cleaning operation: the sugarcane that arrives to the
factory contains some amount of soil that is carried in the harvesting operation; so, it must be
cleaned upon reception. Dry cleaning was considered in the simulation. After that, cleaned
sugarcane goes to the extraction system where sugarcane juice and bagasse are obtained; an
extraction system with mills was considered. From specifications in the extraction system, the
composition of bagasse obtained was (dry basis, wt.): cellulose 36.8% wt., hemicelluloses 35% wt.,
lignin 20.3% wt. and ashes 2.3%. The moisture content of bagasse (wet basis, wt.) was 50%.
Raw sugarcane juice goes to the physical-chemical treatment while bagasse goes to the
cogeneration system. For juice treatment, the following operations were considered: screening,
heating, liming, decantation and mud filtration. After that, the clarified juice goes to the
concentration stage.
In conventional autonomous distilleries, the must for ethanol production is prepared from sugarcane
juice, its concentration is necessary in order to obtain an adequate must sugar concentration for the
fermentation process. Thus, a part of the clarified juice, with its original sucrose content of 11.8%,
160

is concentrated in an evaporation system of multiple effects to achieve a sucrose content of 55.4%.


Then, the original clarified juice is mixed with the concentrated one, resulting in a must with a
sucrose content of 17%.
In the evaporation system, part of the vapours, resulting from the concentration process, are used to
satisfy heat duties in the plant and other part is used in the next evaporator body as heating source.
The must sterilization is done using a treatment type HTST (High Temperature Short Time). In this
study it was adopted that must is heated until 130C. After that, there is a fast cooling until a
fermentation temperature of 32C [5].
In this study, fermentation was based on the Melle Boinot process (batch fed fermentation with cell
recycle), the most common system used in the local industry. Other fermentation by-products are
considered in the simulation, such as glycerol, succinic acid, acetic acid, isoamyl alcohol and yeast,
according to conversion data reported by [6]. At the end of the fermentation, the wine is centrifuged
to recover most of the yeast. The yeast cream obtained is submitted to an acid treatment with H2 SO4
to decrease the pH [7, 8]. An absorption column is considered in order to recovery the ethanol
carried in fermentation gases.
After that, wine goes to distillation stage. A conventional distillation system was simulated
considering the distillation and rectification columns, according to [8]. Ethanol content of hydrous
ethanol is 93.7% while ethanol content of vinasse and phlegmasse is approximately 0.02%,
according to [8].
Ethanol dehydration was simulated adopting the process of extractive distillation with
monoethylene glycol (MEG), which considers the extractive column operating at atmospheric
pressure and the recovery column at 0.2 bar [5,9]. Anhydrous ethanol is obtained with ethanol
content of 99.4% mass basis.
The cogeneration system adopted in this simulation consists of a steam cycle with backpressure
steam turbines and parameters of live steam of 67 bar and 480C. The steam cycle with
backpressure steam turbines was adopted in order to generate only the steam necessary for the
process, therefore producing surplus bagasse that can be used in enzymatic hydrolysis process.
Table 2. Parameters adopted for the simulation of conventional ethanol production process [4]
Parameter
Value
Sugarcane crushing rate, t/h
500
Efficiency of soil removal in cleaning operation, %
70
Efficiency of sugar extraction in extraction system, %
97
Conversion yield from sugars to ethanol, %
89
Ethanol content in vinasse and phlegmasse, %
0.02
Ethanol content in anhydrous ethanol, wt %
99.4
Cogeneration system
Pressure of boiler live steam, bar
67
Temperature of boiler live steam, C
480
Isentropic efficiency of electricity generation in steam turbines, %
80
Alternator efficiency of turbine generator, %
97.6
Turbine mechanical efficiency, %
98.2
Isentropic efficiency of direct drive steam turbines, %
50
Pump isentropic efficiency, %
70
Boiler thermal efficiency, % (LHV base)
85
Mechanical power demand of cane preparation and extraction system, kWh/t of
16
cane
Electric power demand of sugar and ethanol process, kWh/t of cane
12
Process steam pressure, bar
2.5
Process steam temperature, C
127.4
161

Steam turbines have a bleed at 22 bar for mills direct drive turbines and at 6 bar for must
sterilization process and ethanol dehydration requirements. The boiler was modelled according to
previous studies [10] and [11].
Table 2 shows the mean parameters adopted in the simulation of conventional ethanol production
process.

2.2. Ethanol production through enzymatic hydrolysis


The ethanol production through enzymatic hydrolysis begins with the pre-treatment of
lignocellulosic material for the solubilisation of the hemicelluloses and for release the lignin in
order to increase the porosity and contact area of the materials to let the cellulase enzymes gain
access to the cellulose molecules [12].
There are several different pre-treatment methods proposed by researchers and engineering
companies [13]: physical, physical-chemical, chemical and biological. In this study the steam
explosion process was adopted because of its efficiency and low cost in comparison with other
chemical pre-treatments.
The second step is the hydrolysis of cellulose, where cellulose chains are broken down in order to
produce glucose for the fermentation step. Figure 1 shows the flow sheet of the ethanol production
process by enzymatic hydrolysis considered in this study.

Fig. 1. Flow sheet of enzymatic hydrolysis process.


Sugarcane bagasse was considered as raw material for the hydrolysis process. Some authors [14]
indicate the need of cleaning (washing) the raw material, before sending it to the pre-treatment
162

reactor, in order to remove impurities. This operation would reduce the amount of reactants in
subsequent stages. As this study is an initial assessment, previous washing of bagasse and trash was
not considered.
Thus, stream of bagasse B2 is sent to the pre-treatment reactor PRE-TRAT. Some studies indicate
that it is not necessary the addition of a catalyst in the steam explosion reactor because it is an
autocatalytic process: there is a dissociation of water molecules in their ions H+ and OH-, which
hydrolyse hemicelluloses. The formation of acetic acid happens also, which catalyses the
subsequent reactions. On the other hand, some researchers report that the addition of acid catalyst,
such as SO2 or H2 SO4, is necessary in order to achieve higher yields of hemicellulose conversion. In
this study, it was adopted the addition of SO 2 catalyst in the pre-treatment reactor in a rate of 2%
w/w, according to [14].
In relation to the steam consumption by steam explosion pre-treatment, experimental data are
reported in the range of 0.55 to 0.65 kg of steam /kg moist bagasse for moisture bagasse contents in
the range of 38.6 to 65.5% [15].
In the pre-treatment reactor, the formation of xylose (C5 H10 O5 ), acetic acid (C2 H4 O2 ), furfural and
glucose (C6 H12 O6 ) was considered, according to [5]. Table 3 shows the reactions and conversion
yields considered in pre-treatment reactor, according to [14].
Table 3. Yields considered for the reactions in pre-treatment reactor
Equation
Product
Yield (%)
C5 H8O4 + H2O
C5 H10O5
Xylose
61.4
C5 H8O4 + H2O
2.5 C2 H4 O2
Acetic acid
9.2
C5 H10 O5
FURFURAL + 3 H2 O
Furfural
5.1
C6 H10 O5 + H2 O C6 H12O6
Glucose
4.1

From
Hemicellulose
Hemicellulose
Xylose
Cellulose

The decompression in pre-treatment tank was represented by the expansion valve VE2 and the
unitary block FLASH-3. Hence, there are the steam flashed V-FL-PR and the pre-treated material
B-PRE1.
In order to remove xylose and other components that could inhibit the posterior processes of
enzymatic hydrolysis and fermentation, pre-treated bagasse is washed in unit block SEPA-PE. After
that, two fractions are present: the liquid fraction L-PE (xylose liquor) and the solid fraction CELLIG. In this first study, the use of xylose liquor was not considered.
For the next stage, water is added to the process (stream H2O-HIDR) in tank T-MIST, in order to
achieve an appropriate concentration of water insoluble solids in the hydrolysis reactor. Three cases
were considered in this study: solid contents of 5%, 8% and 10% in the hydrolysis reactor. In the
next step, stream CEL-LIG4 goes to the hydrolysis reactor R-HIDROL, where enzymes (stream
ENZIMA) are added to catalyze the hydrolysis reactions. After hydrolysis stage, the hydrolysate
goes to a filter in order to separate the lignin cake (LI-CAKE) of the glucose hydrolysate (LI-GLI).
Table 4 shows the yields adopted in the hydrolysis reactor calculated from [14] data
Table 4. Yields considered for the reactions in the hydrolysis reactor for different solid contents in
the reactor (5%, 8% and 10%)
Reaction
From
Product
Yield (%)
5%
8%
10%
C6 H10 O5 + H2 O C6 H12O6
Cellulose
Glucose
69.2
60.6
55.8
C5 H8O4 + H2O
C5 H10O5
Hemicellulose
Xylose
46.9
44.4
40.6
163

2.3. Insertion of enzymatic hydrolysis plant in the conventional distillery


To integrate the enzymatic hydrolysis process with the conventional distillery, it is necessary that
the hydrolysate concentration (LI-GLI current) achieves appropriate glucose content for the
fermentation process. After the concentration, the hydrolysate can be mixed with must of sugarcane
juice to go to the fermentation process. Moreover, steam for pre-treatment (steam explosion) should
be supplied from the cogeneration system. Lignin cake (LI-CAKE) should be prepared to be burnt
in order to satisfy the heat requirements for steam generation.
Figure 2 shows the block diagram of the hydrolysis plant inserted in the conventional ethanol
production process proposed.

2.3.1. Steam explosion pre-treatment


In this study, steam for pre-treatment is taken at the desuperheater outlet of direct drive turbines, in
the cogeneration system. Then, part of the steam at 22 bar, 300C is sent to the hydrolysis process,
while the other part goes for direct drive turbines. The steam for hydrolysis pre-treatment passes
through an expansion valve before to get into the pre-treatment reactor, to reduce its pressure until
12.5 bar, value adopted in this study [16].

2.3.2. Concentration of glucose hydrolysate


Glucose hydrolysate obtained at the outlet of the separator SEPA-L-T has very low glucose content,
in the range of 1.8% to 3.4%, depending of the solid content adopted in the hydrolysis reactor. To
match the concentration values needed in the fermentation step, this study assumed that the
hydrolysate was concentrated in an appropriate plant, showed in Fig 3.

Fig. 2. Ethanol production process-Conventional process integrated with hydrolysis process.


Table 5 shows the mean parameters adopted for simulation of enzymatic hydrolysis process.

164

Table 5. Parameters adopted for the simulation of ethanol production through enzymatic
hydrolysis
Parameter
Value
a
Pre-treatment reactor temperature, C
190
Pre-treatment reactor pressure, bar
12.5
Pre-treatment reactor steam consumption, kg of steam/kg of raw materialb
0.55
Pressure at unitary block FLASH-3, bar
1.01
Efficiency of solid in solution removal in unit block SEPA-PE, %a
90
a
Loss of soluble lignin in unit block SEPA-PE, %
6.3
Moisture content of solid fraction CEL- LIG, %c
60
c
Water for xylose washing, l/kg of dry material
15
d
Hydrolysis reactor temperature, C
50
Enzymatic load cellulose, FPU/g dry biomassa
53
a
Enzymatic load b glucosidase, IU/g dry biomass
83
Moisture content in solid fraction TORTA-LI0, %
70
Solid content in concentrate hydrolysate, %
19
e
Moisture content in sugarcane trash, %
10.05
e
Trash lower heating value, MJ/kg
13.9
Energy consumption in trash shredder, kWh/t of trashe
82.03
Energy consumption in cleaner station, kWh/t of trashf
13.6
Energy consumption in bagasse feeder, kWh/t of bagasse
0.459
Energy consumption in xylose separator SEPA-PE, kWh/t of material
2.3
3
Energy consumption in separator SEPA-L-T, kWh/m
0.4
Energy consumption in dewatering press of lignin cake, kWh/kg of dry matter
56.09
a

Carrasco et al. (2010) [14]; b Kling et al. (1987) [15]; c Palacios-Bereche, (2011) [4];
Hassuani et al. (2005) [20]; f Cella (2010) [26]

Galbe and Zacchi (2010) [25];

Fig. 3. Concentration of glucose hydrolysate.


According to the Fig. 3, glucose hydrolysate is preheated in the heat exchanger PHX with steam
flash recuperated from the pre-treatment decompression (Fig. 1) before get into the evaporation
system, which operates with exhaust steam at 2.5 bar. An evaporation system of five stages was
considered in order to reduce the steam consumption. For the simulation in Aspen Plus software,
each stage of the evaporation system was considered compounded by two unit operations: a heat
exchanger and a flash separator, [9, 17]. It was assumed that the condensate of exhaust steam from
this evaporation system C-S-EV-L returns to the cogeneration system.
165

Due to the possible presence of soluble lignin in glucose hydrolysate, as well as phenolic groups
(they were not considered in this simulation), the detoxification of glucose hydrolysate is
recommended before its mixture with must of sugarcane juice [18, 19]. In this simulation the
removal of the components acetic acid, furfural and sulphurous acid was considered, before the
mixture with must.

2.3.3. Use of lignin cake


In order to recuperate the heat value of the wet lignin cake, stream CAKE-LIG is sent to the
dewatering press, where its moisture content is reduced to 50%. After that, the lignin cake is sent to
the burner (stoichiometric reactor) where it is burned. In real installations, the lignin cake burning
will happen in the boilers of the cogeneration plant. In this case, an independent burning was
assumed with the unique purpose of simplifying the calculations. So, the gases stream of this burner
block is conducted to the boiler. As no other date is available, the efficiency adopted for lignin
boiler was the same than bagasse boilers.

2.3.4. Sugarcane trash utilization


The amount of residues from sugarcane harvesting depends on many factors such as: harvesting
system, topping height, cane variety, age of crop, climate, soil and others. In this study, an average
trash potential of 140 kg of dry residues per tonne of cane stalks was considered [1,20] and the
amount of trash left in field was assumed as 50% of total, according to [21]. Thus 50 % of total
trash was assumed available for use as fuel in the cogeneration system.
Sugarcane trash and lignin cake, are considered in order to satisfy the energetic requirements of the
integrated process. In this way, it is possible to send a higher amount of bagasse to the hydrolysis
process. Nevertheless, it is still necessary to burn part of the bagasse in the boilers to satisfy the
energy requirements. The amount of bagasse for hydrolysis is defined after an iterative process,
because the increase of raw material for hydrolysis increases the steam consumption of the plant;
thus the amount of bagasse, trash and lignin cake for steam generation should be enough to satisfy
the increase in energy requirements of the plant, and so on.
For the simulations, it was considered that trash and lignin cake are burned in the boiler with an
efficiency of 86% (LHV base) [4].

3. Thermal integration applying Pinch-Point method


The Pinch Point method was used to analyse the streams of the process that are available for
thermal integration. The minimum approach temperature difference ( Tmin) adopted in this study
was 10C for the process and 4C for the evaporation systems. As vapour bleedings of the
concentration step are used to fulfil the process heating requirements, and these bleedings change
depending on the process characteristics, the thermal integration procedure was accomplished
following the next sequence:
Step 1. Calculation of the amounts of trash and bagasse to be burnt in the boiler, from an initial
assumption of the steam consumption of the overall process.
Step 2. Thermal integration of the available process streams, excluding the evaporation systems
(sugarcane juice and glucose liquor).
Step 3. Integration of both evaporation systems and calculation of the appropriate vapour bleeding
demand according to the procedure of [22] and Grand Composite Curve construction.
Step 4. Re-calculation of the steam consumption of the overall process until the convergence is
obtained.
In this study the thermal integration procedure was applied to the case of enzymatic hydrolysis with
solid content 10% in the hydrolysis reactor. This value was selected because it presents the best
results in terms of ethanol production [4].

166

From the iterative procedure explained before, the amount of bagasse for hydrolysis can be
determined, moreover, thermal integration procedure permits to determine the targets of minimal
energy consumption, maximizing, in this way, the production of ethanol from bagasse hydrolysis.
Table 6 shows the streams considered for thermal integration, as well as the exchange heat and the
initial and final temperature. Figure 6 shows the final Grand Composite Curve for the analyzed
case.
Table 6. Streams considered for thermal integration, 10% of solid content in the hydrolysis reactor
(Case V)
Heat streams
Sterilized juice
Fermented wine
Phlegmasse
Vinasse
Anhydrous ethanol
Vapour Condensates
Condenser column B
Condenser extractive
column
Condenser column D
Vapour steam
explosion

Ti
C
130.0
32.0
103.8
109.3
78.3
83.2

Tf
C
32
28
35
35
35
35

H
MW
50.8
13.2
3.7
45.9
10.2
16.8

81.6

81.6

30.5

78.3

78.3

8.7

85.1

35

34.6

100.7

100

17.9

Cold streams
Treatment juice
Pre-heating juice
Juice for sterilization
Final wine
Reboiler column A
Reboiler column B
Reboiler extractive
column
Reboiler recovery
column
Hydrolysis water
Pre-heating glucose
liquor
Imbibition water

Ti
Tf
C
C
34.2 105.0
98.1 115.0
89
130.0
31.2 90.0
109.3 109.3
103.4 103.8

H
MW
44.9
2.8
22.5
41.4
53
35.4

112.1

137.1

8.0

149.6

149.6

2.9

25

50

9.7

50

115

27.2

25

50

4.7

Fig. 4. Grand Composite Curve including the evaporation systems GCC (Case V).

4. Results and discussion


Table 7 shows the results of the anhydrous ethanol production, electricity surplus, bagasse sent to
the hydrolysis process and steam consumption for the evaluated cases. Case I corresponds to the
167

conventional distillery without hydrolysis, while cases II, III and IV correspond to the hydrolysis
process coupled with the conventional distillery without thermal integration. Case V corresponds to
the hydrolysis process integrated with the conventional process according to the method presented
in item 3. A higher solids concentration in the hydrolysis process means lower steam consumption
in the glucose liquor concentration step. This fact permits to assign higher amounts of bagasse for
the hydrolysis process and consequently, to obtain higher increases in the ethanol production.
About electricity surplus, there is an increase of 47.9%, 29.3%, 21.3% in Cases II, III and IV
respectively in comparison with Case I, these increases happen owing to an increase in steam
consumption in the mentioned cases. On the other hand, in Case V there is a reduction of 27.9% in
comparison with Case I, fact that is explained because in Case V, a reduction of steam consumption
happens. In general, these results are due to the turbine configuration adopted in the cogeneration
system (backpressure steam turbine) and the reduction of steam consumption for higher solid
contents in the hydrolysis step.
From the thermal integration procedure, it can be observed a significant decrease in steam
consumption in Case V. Steam consumption in Case V is 11.1% lower than the consumption in
Case IV, while ethanol production in Case V is 4.6% higher than the production obtained in Case
IV.
Table 7. Results of the simulation: Anhydrous ethanol production l/t of cane, electricity surplus
kWh/t of cane and steam consumption kg/t of cane.
Parameter
Solid content in hydrolysis reactor
Anhydrous ethanol, (l/t of cane)
Electricity surplus, (kWh/t of cane)
Bagasse for hydrolysis, (kg/t of cane)
Steam consumption, (kg/t of cane)
Increase in ethanol production, (%)
Increase in ethanol production, (l/t of cane)

Case I
-79
42.3
0
489.6

Case II
5
86.7
62.5
110.3
794.2
9.7
7.7

Case III Case IV


8
10
88.1
88.7
54.6
51.3
149.5
172.7
756.9
745.1
11.5
12.3
9.1
9.7

Case V
10
92.8
30.5
239.7
661.8
17.4
13.7

In relation to vinasse production, there is an increase of 11.6%, 14.1%, 15.1% and 20.8% for Cases
II, III, IV and V in comparison with Case I. It is due to the higher amount of must processed in
distillation stage in comparison with Case I. In Brazilian current practice, the vinasse is used to
irrigate the fields, where the nutrients are recuperated.
Figure 5 shows the bagasse balance for each case. Bagasse for hydrolysis increases from Case II to
V, while bagasse for boiler decreases. The amount of trash for boiler is constant for all cases, 38.9
t/h considering a moisture content of 10%. The amount of lignin cake also increases from Case II to
V, being a percentage of the bagasse for hydrolysis. Thus, the amount of lignin cake at 50% of
moisture content was 0.43 and 0.46 kg/kg of bagasse for hydrolysis for Cases II and III respectively
and 0.48 kg/kg of bagasse for Cases IV and V. The lower heating value of lignin cake was
calculated through the simulator from their composition data. The lower heating values obtained for
lignin cake were: 8563 kJ/kg for Case II, 8441 kJ/kg for Case III and 8373 kJ/kg for Cases IV and
V. These differences occurred because of the conversion yields adopted in the hydrolysis reactor.

168

Fig. 5. Bagasse balance (kg/t of cane) for Cases I, II, III, IV and V.
Figure 6 shows the balance of electricity generated. A decrease in electricity surplus appear from
Cases II to V due to the reduction in electricity produced in the cogeneration system, as a
consequence of the decrease in the steam generated in boiler, and due to the consumption increase
in the hydrolysis process.

Fig. 6. Balance of electricity generated (kWh/t of cane) for Cases II, III, IV and V
The electricity consumed in the hydrolysis process was estimated from the data shown in Table 5. It
was identified a significant electricity consumption in the agitators of the hydrolysis reactors, due to
the large volume of the hydrolysis reactors. The reactor volume was calculated assuming reactors
CSRT (Continuous Stirred Tank Reactors) and residence time of 48h, in agreement with data of
Table 4. The total volume of reactors resulted in 20520 m3 , 17165 m3 , 15731 m3 and 21835 m3
respectively for the cases II, III, IV and V. The electricity consumption in the reactor agitators was
determined from a scale up procedure based on the data reported in [23].
Results of the present study can be compared with others of the literature. Walter and Ensinas [19]
indicated an increase in ethanol production of 25.6%, but in a future scenario and considering
glucose and xylose fermentation. Dias et al. [24] indicated ethanol production increases of 18.3%
and 20.3%, but considering an Organosolv hydrolysis process with diluted acid, while Leite [1]
indicated an ethanol production increase, through enzymatic hydrolysis process, of 12.2% and
25.6%, for scenarios corresponding to the years of 2015 and 2025 respectively.
169

Dias et al. [27] also accomplished a simulation study considering the ethanol production by
enzymatic hydrolysis and pre-treatment by steam explosion, and a reduction of 30.8% in steam
consumption was obtained from thermal integration. These authors indicate an ethanol production
in the integrated process in the range of 107.5 to 120.6 l/t of cane; these values represent an increase
of 15.9% and 30% in comparison to their base case.

5. Conclusions
An evaluation of the mass balance and energy consumption for the ethanol production process by
enzymatic hydrolysis was accomplished. Moreover, this study showed the potential ethanol
production increase due to the introduction of the bagasse enzymatic hydrolysis in the conventional
ethanol production process.
The results obtained for cases II, III an IV were modest because the study considers a conservative
conventional factory with average technology, thus, in order to increase the ethanol production the
optimization of the conventional process would be recommended, aiming an energy consumption
reduction. Hence, the ethanol production increase for cases II, III and IV was 9.7%, 11.5% and
12.3%, respectively. These results showed that a higher ethanol production is obtained for higher
solids concentrations in the hydrolysis process. They show that thermal energy consumption in the
hydrolysate concentration stage is very significant, thus, other separation technologies with lower
energy consumption should be studied. Reverse osmosis membranes in combination with
evaporation systems can be tested for this application.
Case V showed the highest ethanol production of 92.8 l/t of cane, which represents an increase of
17.4%; however, this is a prospective study and pilot plant/industrial data would be necessary in
order to adjust the modelling of the plant. The Pinch Analysis applied to this case showed to be a
useful tool to evaluate the thermal integration potential.
There is an intrinsic vinasse increase due to the higher amount of must being processed with the
introduction of the hydrolysis process; the highest vinasse production was 430.8 t/h for Case V
which represents an increase of 20.8% in comparison with Case I. Currently, vinasse is used as
fertirrigation in Brazilian sugarcane fields, however this practice is being regulated. Thus, the
introduction of technologies to reduce the vinasse production becomes necessary.
The energy balance in the cogeneration system shows that the energy supplied by lignin cake is
significant to operate the integrated process. For Case V, energy supplied by lignin cake represented
44.2% of total energy supplied to the cogeneration system. Hence, the study and characterization of
lignin cake is important in order to enable the ethanol production by enzymatic hydrolysis process.

Acknowledgments
The authors wish to thank to CNPq (Process 304820/2009-1), FAPESP (Process 11/05718-1) and
FINEP (Contract FINEP FUNCAMP Nr. 01/06/004700) for the financial support.

References
[1] Leite R.C.C., Fuel bioethanol an opportunity to Brazil. Brasilia: Centro de Gesto e Estudos
Estratgicos; 2009, (In Portuguese).
[2] Rein P., Cane Sugar Engineering. Berlin: Verlag Dr. Albert Bartens K. G.; 2007.
[3] Wooley R.J., Putsche V., Development of an ASPEN PLUS Physical Property Database for
Biofuels Components, National Renewable Energy Laboratory, 1996. Available at
<www.p2pays.org/ref/22/21210.pdf> [accessed 12.11. 2007].

170

[4] Palacios-Bereche R., Modeling and energy integration of the ethanol production process fro m
sugarcane biomass [PhD Thesis]. Campinas, Brazil: University of Campinas; 2011 (I n
Portuguese).
[5] Dias M.O.S., Simulation of ethanol production processes from sugar and sugarcane bagasse,
aiming process integration and maximization of energy and bagasse surplus. [dissertation].
Campinas, Brazil: University of Campinas; 2008 (In Portuguese).
[6] Eijsberg R., The design and economic analysis of a modern bioethanol factory located in Brazil.
[dissertation]. Delft, The Netherlands: University Delft; 2006.
[7] Rossell C.E.V., Sugarcane processing to ethanol for fuel purposes, In Chemistry and Processing
of Sugarbeet and Sugarcane, edited by Clarke M. A. and Godshall M.A., Elsevier Science
Publishers B. V., Amsterdam, 1988.
[8] Finguerut, et al., Fermentation, Hydrolysis and Distillation. In: Cortes et al. editors. Biomass for
Energy. Campinas, Brazil: Ed. Unicamp. 2008. p. 436-474.
[9] Dias M.O.S., Modesto M., Ensinas A.V., Nebra S.A., Maciel Filho R., Rossell C.E.V.,
Improving bioethanol production from sugarcane: evaluation of distillation, thermal integratio n
and cogeneration systems. Energy 2010; 36: 3691-3703.
[10] Palacios-Bereche R., Nebra S.A. Thermodynamic modeling of a cogeneration system for a
sugarcane mill using ASPEN PLUS, difficulties and challenges. COBEM 2009: Proceedings o f
the 20th International Congress of Mechanical Engineering, 2009 Nov 15-20, Gramado, Brazil.
[11] Magnusson H., Process simulation in Aspen Plus of an integrated ethanol and CHP plant.
[dissertation]. Sweden: Umea University. 2005.
[12] Sun Y., Cheng J., Hydrolysis of lignocellulosic materials for ethanol production: a review.
Bioresource Technology 2002; 83: 111.
[13] Efe C., Technical and economical feasibility of production of ethanol from sugar cane and
sugarcane bagasse. [dissertation], Delft, The Netherlands: TU-Delft; 2005.
[14] Carrasco C., Baudel H.M., Sendelius J., Modig T., Roslander C., Galbe M., Hahn-Hgerdal
B., Zacchi G., Lidn G. SO2 -catalyzed steam pretreatment and fermentation of enzymatically
hydrolyzed sugarcane bagasse. Enzyme and Microbial Technology 2010; 46: 6473.
[15] Kling S.H., Carvalho Neto C., Ferrara M.A., Torres J.C.R., Magalhaes D.B., Ryu D.D.Y.,
Enhancements of enzymatic hydrolysis of sugar cane bagasse by steam explosion pretreatment.
Biotechnology and Bioengineering 1987; 29: 1035 1039.
[16] Sanchez O.J., Cardona C.A., Trends in biotechnological production of fuel ethanol from
different feedstocks. Bioresource Technology 2008; 99: 52705295.
[17] Jorge L.M.M, Righetto A.R., Polli P.A., Santos O.A.A., Maciel Filho R., Simulation and
analysis of a sugarcane juice evaporation system. Journal of Food Engineering 2010; 99: 351
359.
[18] Cardona C.A., Sanchez O.J., Energy consumption analysis of integrated flowsheets for
production of fuel ethanol from lignocellulosic biomass. Energy 2006; 31 (13): 2447-2559.
[19] Walter A., Ensinas A.V., Combined production of second generation biofuels and
electricity from sugar-cane residues. Energy 2010; 35: 874-879.
[20] Hassuani S.J., Leal M.R.L.V., Macedo I.C., Biomass power generation: Sugarcane bagasse
and trash, Piracicaba, Brazil: Ed. PNUD and CTC, 2005.
[21] Michelazzo M.B., Sensitivity analysis of six systems for collection of sugarcane trash
(Saccharum spp.). [dissertation]. Campinas, Brazil: University of Campinas, 2005 (I n
Portuguese).
[22] Westphalen D.L., Wolf Maciel M.R., Pinch Analysis of evaporation systems, Brazilian
Journal of Chemical Engineering 2000; 17: 4-7.

171

[23] Pereira L.T.C., Teixeira R.S.S.T., Bom E.P.S., Freitas S.P. Sugarcane bagasse enzymatic
hydrolysis: rheological data as criteria for impeller selection. Journal of Industrial Microbiology
& Biotechnology 2010; DOI 10.1007/s10295-010-0857-8.
[24] Dias M.O.S., Ensinas A.V., Nebra S.A., Maciel Filho R., Rossell C.E.V., Production of
bioethanol and other bio-based materials from sugarcane bagasse: Integration to conventiona l
bioethanol production process. Chemical Engineering Research and Design 2009, 87: 12061216.
[25] Galbe, M., Zacchi, G. Ethanol production from lignocellulosic materials. In: CORTEZ,
L.A.B. (Coord.). Bioethanol from sugarcane: R&D for productivity and sustainability. Sao
Paulo: Blucher, 2010. Part. 4, Chap. 12, p. 697-716.
[26] Cella, N. Trash use as fuel in biomass boilers. Course Boilers, Environment and Renewable
Energy. Ribeirao Preto, 23-24 de june 2010.
[27] Dias, M.O.S.; Cunha, M.P., Maciel Filho, R., Bonomi, A., Jesus, C.D. F., Rossell, C.E.V.
Simulation of integrated first and second generation bioethanol production from sugarcane :
comparison between different biomass pre-treatment methods. Journal of Industria l
Microbiology & Biotechnology, 2011, 38: 955-966.

172

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Evaluation of gas production in a industrial


anaerobic digester by means of Biochemical
Methane Potential of Organic Municipal Solid
Waste Components
Isabella Pecorini a , Tommaso Olivierib , Donata Bacchi c, Alessandro Paradisid , Lidia
Lombardi e, Andrea Corti f and Ennio Carnevaleg
a

Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze (CA), Florence, Italy,
isabella.pecorini@unifi.it
b
Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze, Florence, Italy,
tommaso.olivieri@unifi.it
c
Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze, Florence, Italy,
donata.bacchi@unifi.it
d
Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze, Florence, Italy,
alessandro.paradisi@unifi.it
e
Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze, Florence, Italy,
lidia.lombardi@unifi.it
f
Dipartimento di Ingegneria dellInformazione - Universit degli Studi di Siena, Florence, Italy,
corti@dii.unisi.it
g
Dipartimento di Energetica Sergio Stecco - Universit degli Studi di Firenze, Florence, Italy,
ennio.carnevale@unifi.it

Abstract:
The Biochemical Methane Potential (BMP) of several components of the Organic Fraction of the Municipal
Solid Waste (OFMSW) were tested in order to assess the possibility to obtain a good estimate of the biogas
production of a real scale anaerobic digestion plant. In particular, five different fractions and a mixed food
waste sample were t ested with batch anaerobic digesters at 37C and both the BMP after 21 days (BMP21)
and final BMP (BMP f) were measured. Regarding the mixed food waste substrate it was found an average
BMP21 of about 405 NL/kgVS and a B MPf of 484 NL/kgVS with an average methane content of 57%. From
the experimental results, some industrial potential biogas production were defined to compare them with
data from real anaerobic digestion plants. In particular two different plants were considered: one located in a
rural area that treats the source selected OFMSW from a public collection point, another locat ed in a city
area with a curbside collection system. Furthermore, studying the BMP of the pre-treatment reject of these
plants, it was possible to study the pre-treatment efficiency and the difference performance of the two real
plants.

Keywords:
Anaerobic Digestion, Biochemical Methane Potential, Biogas, Organic Fraction Municipal Solid Waste.

1. Introduction
In 2009, more than 10 million tons of waste, corresponding at the 33,6% of the whole amount of
municipal solid waste (MSW) produced in Italy, were collected as source separated fractions. About
the 35% of these fractions were organic fraction from kitchen and yard and garden waste and, since
2005, a constant increasing of 11% every year have been recorded. Moreover ISPRA [1] shows that
most of them were treated in composting plants (about 281 facilities were registered in the 2009)
and about 540'000 t where instead stabilized in anaerobic digestion plants, in particular 18 plants of
which 15 working.

173

New strategies in MSW management, i.e., source-separate collection of the OFMSW and the need
to reduce the biodegradable-MSW allocated in landfill, have favoured the development of
composting and anaerobic digestion as useful biotechnologies for transforming organic waste into
suitable agricultural products [2]. Moreover, given that the amount of OFMSW is still increasing
and the attention to the environmental impacts is becoming all the time more important, the
possibility to recover not only compost form waste but also energy could enhance the anaerobic
digestion of OFMSW as way to provide a clean fuel from renewable energy [3]. In this way the
quality of the OFMSW in terms of potential methane production becomes important in order to
assess the biogas production expectation from the anaerobic treatment.
In the last years, several researches have been carried out for the analysis of biomethane potential of
several waste substrates. In particular, most of these utilize BMP analysis as a possible way to
characterize the biodegradability of the organic matter in order to assess its stability and the waste
treatments efficiencies such as composting or anaerobic digestion. Others are instead interested in
determining waste BMP as relevant in the context of treatment by anaerobic digestion and useful to
determine the amount of organic carbon that can be anaerobic converted to methane. This research
focus on this last purpose with the aim to understand how, by measuring the BMP of the main
component of the OFMSW is possible to estimate the potential biogas production of real digestion
plant. Moreover, to obtain more realistic results, this research focuses on how the pre-treatments
and the operating environment could affect the biogas production of real plants.

2. Materials and methods


To study the BMP of the source-selected organic fraction, essentially kitchen and garden waste,
several substrates were tested; in particular: proteins from meat and dairy products, carbohydrates
from bread and pasta, fruit and vegetable, dirty paper from kitchen and other organic materials from
yard and garden waste. Moreover, to assess the efficiency of a typical anaerobic digestion plant, it
was also necessary to measure the BMP of the fractions rejected by the pre-treatment. In particular
it has been possible to test the light fraction and the small heavy removed from the OFMSW by a
specific treatment in two real industrial anaerobic digestion plant with different location: a rural and
a urban area.
To estimate the biogas potential production of each fraction, the BMP analysis were carried out in
duplicate and both the BMP21(biogas produced at 21 days) and the BMPf (when no significant
biogas production is detected) were measured. For analysis a modified method of Ponsa et al. [4]
was used and in the following, according with Angelidaki, Alves and Bolzonella et al. [5], the
materials and the method used will be described.

2.1. Inoculum and substrates tested


Active inoculum from a mesophilic anaerobic digestion plant, that primarily treats organic fraction
from MWS, was used. In order to deplete the residual biodegradable organic material present [5],
the inoculum was pre-incubated for three day in a water bath at 37C. Total Solid (TS) and Volatile
Total Solid (VS) contents were about 3,9% on wet weight basis (w/w) and 64,1% on TS basis,
respectively.
Table 1. Characterization of the substrates used
Substrate
Experimental ID
Proteins
Proteins
Carbohydrates
Carbohydrates
Fruit and vegetables
Fruit and Veget
Leafs
Yard and Garden waste
Cellulose
Dirty paper
174

TS, %FM
33,10,24
94,50,04
14,00,05
6,50,61
40,30,02

VS, %TS
89,90,01
97,50,01
98,10,06
81,81,48
90,40,04

The source selected OFMSW from curbside collection was used for the substrates tested in the
batch assay. For each fractions water content and VS were measured in triplicate, in Table 1 only
the mean value are shown.
Furthermore, a mixing of those fractions (Mixed Food Waste MFW) was employed for the tests
too; its composition is shown in Table 2. TS and VS contents, as the sum of each fractions, were
about 42,40,38% (w/w) and 93,30,01% on TS basis respectively.
Table 2. Mixed food waste composition
Fraction
Fruit and vegetable
Proteins
Carbohydrates
Leafs
Cellulose

Weight (g)
8
8
8
3
3

Percentage %(w/w)
26,7
26,7
26,7
10
10

2.2. Set-up of measurement


The BMP was determined using 1L stainless steel bottles, incubated in a water bath at 37,5C,
tightly closed by special cap provided with a ball valve to enable the gas sampling. To ensure
anaerobic conditions, the bottles were flushed with inert gas. All the equipment, 2 bar proof
pressure, was specifically design and developed.

Fig 1. Batches and cap used


With the fractions described above, six sample and twelve batch reactors (the test was performed in
duplicate) were prepared. Each reactor was loaded with different quantity of substrate, depending
on the characteristics of the materials, to achieve a concentration of substrate in each batch of about
2gVS/100 mL solution, given that this concentration is a compromise of, one hand, the need to use
a large sample to have good representativity and to get a high easy-to- measure gas production, and,
on the other hand, to avoid too large and impractical volumes of reactors and gas production and
keep the solution dilute to avoid inhibition from accumulation of volatile fatty acid (VFA) and
ammonia [6].
Moreover, in every case, the inoculum to sample ratio was kept under 10:1 weight ratio, according
with Ponsa et al. [4] for fresh feed-in substrate, and because it was demonstrated that the amount of
inoculums should be enough to prevent the accumulation of volatile fatty acids and acid
conditions[5].
To determine the background methane production, a blank assay with the only inoculum was done
in duplicate.
According with the authors [7], biogas production was estimated by measuring the pressure in the
head space of each reactor and then converting to volume by application of the ideal gas law.
175

Pressure was measured using a membrane pressure gauge (Model HD2304.0, Delta Ohm S.r.L.,
Italy). The values of pressure measured were converted into biogas volume as:
Pmeasured TNTP
Vbiogas
Vr
PNTP Tr
where:
Vbiogas, volume of daily biogas production, expressed in Normal litre (NL);
Pmeasured, headspace pressure before the gas sampling (atm);
Tr and Vr, temperature (K) and volume (L) of the reactor;
TNTP and PNTP , Normal temperature and pressure, 273,15K and 1 atm respectively.
The headspace volume, calculated as the difference between the total volume of the batch and the
volume occupied by the sample considering a sample density of 1g/mL, was about 600ml for each
bottle.
The gas produced was routinely analyzed using an IR gas analyzer (ECOPROBE 5 - RS Dynamics).
After every measurement the bottles were shaken to guarantee homogeneous conditions in the assay
vessels [5].

Fig 2. Laboratory equipments


The BMP was determined as the cumulate biogas production, calculated as the sum of the daily
volumes, divided by the TS and the VS present in each batch. The results, reported in the Normal
Temperature and Pressure (NTP), were obtained after 21 and about 90 days. After this period, in
fact, the quantity of biogas produced by every sample was found to be lower than the blank
production and no significant biogas volumes can be considered.

3. Results and Discussions


3.1. BMP assay
The results obtained for the biochemical methane potential at 21 days are shown in Table 3. In
particular the quantity of biogas produced is referred to the TS and the VS and, in order to consider
the inoculum biogas production, a percentage error is also shown. It was calculated as the ratio
between the quantity of biogas produced by the inoculum and the total biogas produced from each
fraction.
Table 3. BMP21
176

Experimental ID
MFW (a)
MFW (b)
Proteins
Carbohydrates
Fruit and Veget
Yard and Garden waste
Dirty paper
Inoculum

NL/kgTS
413
396
386
91
250
115
315
47

NL/kgVS
542
520
528
109
353
175
422
74

Error, %
14
14
14
67
21
42
17
-

Preliminary experiments on similar waste showed that 90 days of incubation at 35C, after the lag
period, was sufficient to insure the total gas production expression [8], therefore the values of BMP
measured after 93 days has been considered as the final biogas produced by each substrate. The
results obtained, deducted the inoculum yield, seem to be comparable to that obtained in other
similar studies [9] and [10].
Table 4. BMPf
Experimental ID
MFW (a)
MFW (b)
Proteins
Carbohydrates
Fruit and Veget
Yard and Garden waste
Dirty paper
Inoculum

NL/kgTS
501
466
452
101
327
205
377
130

NL/kgVS
657
611
617
120
462
312
504
203

Error, %
31
33
33
168
44
65
40
-

The highest BMP value was obtained for the MFW, while, observing the calculated errors, it is
clear that for charbohydrates some problem of acidification occurred.

Fig. 3: Cumulative Biochemical methane production of the tested samples


177

This can be also noticed in the temporal plot of the average cumulative biochemical methane yields
(Fig. 3): for the soon after few days, no significant methane production was detected.
With reference to Fig. 4, the correlation between the BMP21 and the BMPf has been studied [4]. As
results from the chart, the biogas obtained at 21 days corresponds to the 89% of ultimate potential
methane. In fact, as supported by the comparison between the cumulative and the daily MFW
average BMP, the main part of the biogas totally produced was released during the first 20 days
(Fig. 5).
Focusing on the MFW results, the peak value of daily biogas production was about 51 NL/d*kg VS,
while the cumulative biogas were 542 and 520 NL/kgVS for the MFW (a) and (b) sample.

Fig. 4: BMP21 and BMPf correlation

Fig. 5: Daily and cumulative BMP, Mixed Food Waste


As said before, the composition of the sampled biogas has been routinely analysed. As shown in
Fig. 6, the methane content in the biogas produced by the MFW, increases from the 20% to 65%
until the 20th day of digestion and from the 30th remains constant around the 60%.
178

Also for the other samples, the biogas composition was analysed in order to understand if the
methanogenesis phase was correctly taking place during the digestion process.
From the biogas analysis it has also been possible to esteem the quantity of methane totally
produced, simply as the product between the biogas released and the methane percentage of its
composition (Fig. 7). Considering an average methane percentage of about the 57%, it was
esteemed a methane yield of 309 and 296 NLCH4 /kgVS for the sample MFW (a) and MFW (b)
respectively.

Fig. 6: MWF Sample biogas analysis

Fig. 7: MWF, BMP and CH4 production Comparison


As said before, the results obtained for the carbohydrates assay shows that the inhibition of the
inoculum occurred. This is evident also looking at the errors behaviour (calculated as the ratio
between the inoculums BMP and the BMP of each fraction) shown in Fig. 8. After 5 days, the
biogas produced by the inoculums was the 20 % of the quantity produced by the sample, to become
higher than the 120% at the 93rd day of digestion, which means a negligible biogas production from
the substrate comparing with the one from the inoculums.
179

This is probably due to an accumulation of volatile fatty acid and acid conditions. In fact, in the
preparation of the carbohydrates batches, too high concentrations of substrate were probably used,
as well as the substrate to inoculums ratio was higher than 1,2, value suggested by the authors to
avoid acidification[7] given that this ratio is recognised as one of the major parameter affecting the
results of anaerobic assay[11].

Fig. 8: Errors behaviour


Regarding the other substrates tested, the estimated errors were high too. After the 20th day of
digestion, the BMP inoculum to BMP substrates ratio starts to increase until the 50 th when all of
them become higher than the 50%. This is probably due to a really high activity of the inoculum
and an high VS content, as it is possible to observe in the average biomethane produced measured
daily (Fig. 9): after 50 days the daily biogas production of all the substrates became comparable
with the inoculum BMP.

Fig. 9: Avarage Daily BMP comparison

180

As Fig. 5, the graphic in Fig. 9 shows that, comparing with the background methane production
from the inoculum determined in blank assays, the biogas production after 40th day of digestion can
be neglect and the most of the biogas was produced during the first 20th day of digestion.
Furthermore, looking at the daily biogas pruduction,it can be noticed that, according with Zhu B.,
Gikas P. and Zhang R. et al. [7], the biogas production duration of food waste was prolunged, with
initial daily biogas yields lower compare to others.

Fig. 10: MFW BMP schematization


With reference to the BMP measured for the mixed food waste samples, it is possible to find a good
schematization of the biogas production. As show in Fig. 10, the curve could be well approximated
by two lines with different slope. At the beginning the anaerobic digestion is faster and high biogas
producing, but, after the meeting point of the two lines, anaerobic kinetics seems to be no more
convenient given that the biogas production decreases and the biodegradation became really slow.
As suggested this raw model, for the first 20 days (first line) the anaerobic digestion seems to be a
convenient treatment, after which (second line) aerobic stabilization process seems to be more
suitable.

3.2. Industrial biogas production comparison


Moving from the results obtained from the tests, in order to compare the results with the biogas
production from a real scale anaerobic digestion plant, it was calculated a specific BMP for a ton of
input waste fraction (Table 5).
Table 5: BMP21 and BMPf for a ton of input waste
Experimental ID
BMP21, Nm3 /t
MFW (a)
175
MFW (b)
168
Proteins
128
Carbohydrates
86
Fruit and Veget
35
Yard and Garden waste
7
Dirty paper
127
Inoculum
2

BMPf, Nm3 /t
213
198
150
95
46
13
152
5

Adding up the BMP21 of each waste component and considering a typicall source selected
OFMSW composition shown in Fig. 11, three potential industrial biogas production are define
(Table 6):
181

a maximum biogas production (GPmax), calculated assuming that the pre-treatment before the
anaerobic digestion is able to remove only the undesired waste fractions with an efficiency of the
100%, i.e. wood packaging, plastic film and plastic packaging, other plastic, rubber, leather,
ferrous and non ferrous metals, inert and hazardous waste;
a potential production, in which is consider the 80% of pre-treatment efficiency and the 20% of
biodegradable fraction removed wrongly (GP80%);
another potential production defined as above but assuming an 75% removal efficiency and the
25% of organic fractions separate erroneously (GP75%).

Fig. 11: OFMSW composition


Table 6: GPmax, GP80% and GP75% calculated values
Potential Industrial
Removal efficiency, % OF removed, %
biogas production
GPmax
100
0
GP80%
80
20
GP75%
75
25

Calculated value, Nm3 /t


134
107
100

These estimates could be compared with the biogas production of real anaerobic digestion plants
(GPreal). In particular, two different anaerobic digestion plants are considered: one located in a
rural area treating source selected OFMSW from a public collection point with a GP of 98 Nm3 /t,
another located in a city area with a curbside collection system characterized by a higher GP (Table
7). It is possible to notice that the estimated biogas production (GP80% and GP75%) and the GPreal
have similar values and that the rural area digester is characterized by the lowest biogas production.
Table 7: Biogas production, real industrial anaerobic digestion plant
Anaerobic digestion
Collection system
GPreal, Nm 3 /t
plant
Rural Area
Public collection point
98
City Area
Curbside collection
108

Data Source
Management data
Bozano Gandolfi P., [12]

As said before, during the experimental assay, it has been possible to measure also the BMP of the
pre-treatment rejects. In particular the light fraction (LF) and the small heavy fraction (SHF) from
the pre-treatment of the two anaerobic digesters were tested (Table 8).
182

Table 8: BMP21 of pre- treatment reject


Pre-treatment reject
Provenience
LF
Rural Area
SHF
LF
City Area
SHF

BMP21, NL/kgVS
240
261
508
214

BMP21, Nm3 /t
57
8
99
33

As before, also the BMP21 of both LF and SHF form the rural area digester were lower than the
ones from the city area. This is probably due to an high presence of yard and garden waste in the
rural waste composition, given that the yard and garden waste has a low BMP21 (with reference to
Table 5, 7 Nm3 /t). On the other side, the rejects from the city area digestion plat are characterized
by a high BMP21, probably because a considerable part of biodegradable fraction of input waste,
with the highest BMP21, is wrongly removed by the pre-treatment.

5. Conclusions
The results of this study show that the biochemical methane potential assay could provide useful
data to study the pre-treatment efficiency and the performance of real anaerobic digestion plant.
Focusing on the experimental work, the MFW had the highest biochemical methane potential, with
an average BMP21 of 405 NL/kgTS and a BMPf of about 484 NL/kgTS and an average methane
content of 57%. The laboratory equipments developed prove to be suitable to this kind of
experimentation, given that no airtight problem occurred. However, the results obtained for the
carbohydrates sample and the errors behaviour show that the measurement protocol and the sample
preparation have to be improved in some parts. In particular more attention and specific evaluation
have to be done for the substrate to inoculum ratio as well as for the inoculum characteristics.
Actually, the estimates obtained with the experimental assay were comparable with the biogas
production of real anaerobic digestion plant and it has been possible to assess the pre-treatment
efficiency and the performance of some real cases. From the comparison results that: the GP of the
rural plant is affected by the presence of yard and garden waste, as it is also supported by the BMP
of the pre-treatment rejects; the pre-treatment reject of the city area plant, that treats OFMSW
collected by a curbside system, has a high BMP probably because a considerable biodegradable
fraction is removed by the pre-treatment.

Nomenclature
BMP Biochemical Methane Potential, NL/kgTS or NL/kgVS
BMP Biochemical Methane Potential after 21 days, NL/kgTS or NL/kgVS
BMPf Final Biochemical Methane Potential, NL/kgTS or NL/kgVS
GP80%
Industrial biogas production, considering 80% of pre-treatment efficiency, Nm3 /t
GP75%
Industrial biogas production, considering 75% of pre-treatment efficiency, Nm3 /t
GPmax
Industrial biogas production, considering maximum pre-treatment efficiency, Nm3 /t
GPreal Industrial biogas production, considering existing real plants, Nm3 /t
LF
Light Fraction, anaerobic digestion plant rejects
MFW Mixed Food Waste
MSW Municipal Solid Waste
NL
Normal litre
NTP Normal Temperature and Pressure conditions, 273,15K and 1 atm respectively
183

OFMSW
Organic Fraction from Municipal Solid Waste
SHF Small Heavy Fraction, anaerobic digestion plant rejects
TS
Total Solid, % v/v on wet weight basis
VS
Volatile Total solid, %v/v on TS basis

References
[1] Istituto Superiore per la Protezione e la Ricerca Ambientale. Rapporto Rifiuti Urbani Edizione
2009

Available
at:
<
http://www.apat.gov.it/site/itit/apat/pubblicazioni/rapporti/documento/rapporto_108_2010_rifiuti.html>.
[accessed
17.1.2012]
[2] Pognani M., BarrenaR., Font X., Scaglia B:, Adani F., and Snchez A., Monitoring the organic
matter properties in a combined anaerobic/aerobic full-scale municipal source-separated waste
treatment plant. Bioresource Technol, 2010; 101:6873-6877
[3] Schievano A., Pognani M., DImporzano G., and Adani F., Predicting anaerobic
biogassification potential of ingestates and digestates of a full- scale biogas plant using chemica l
and biological parameters. Bioresource Technol, 2008; 99:8112-8117
[4] Ponsa S., Gea T., Alerm L., Cerezo J., Sanchez A., Comparison of aerobic and anaerobic
stability indices through a MSW biological treatment process. Waste Manage, 2008; 28:27352742
[5] Angelidaki I., Alves M., Bolzonella D., Borzacconi L., Campos J.L., Guwy A.J., Kalyuzhnyi
S., Jenicek P and van Lier J.B., Defining the biomethane potential (BMP) of solid organic
wastes and energy crops: a proposed protocol for batch assays. Water Sci Technol, 2009;
59:927-934
[6] Hansen L.T., Schmidt E.J., Angelidaki I., Marca E., Jansen J.L.C., Mosb k H. and Christense n
T.H., Method for determination of methane potentials of solid organic waste. Waste Manage,
2004; 24:393-400
[7] Zhu B., Gikas P., Zhang R., Lord J., Jenkins B. and Li X., Characteristics and productio n
potential of municipal solid waste pretreated with a rotary drum reactor. Bioresource Technol,
2009; 100:1122-1129
[8] Bayard R., de Arajo Morais J., Ducom G., Achour F., Rouez M. and Gourdon R., Assessment
of the effectiveness of an industrial unit of mechanical-biological treatment of municipal solid
waste, 2010; 175:23-32
[9] Lesteur M., Bellon-Maurel V., Gonzalez C., Latrille E., Roger J.M., Junqua G. and Steyer J.P.,
Alternative methods for determining anaerobic biodegradabilit: A review. Process Biochem,
2010; 45: 431-440
[10] Cho J.K. and Park S.C., Biochemical methane potential and solid state anaerobic digestion
of Korean food waste. Bioresource Technol, 1995; 52: 245-253
[11] Neves L., Oliveira R. and Alves M. M., Influence of inoculum activity on the biomethanization of a kitchen waste under different waste/inoculum ratios. Process Biochem,
2004, 39:2019-2024
[12] Bozano Gandolfi P., Pratical experiences in the production of biogas and energy from
wastes Available at: < http://bta-international.de/downloads.html?&lang=3>. [accessed
24.1.2012]

184

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Exergy Analysis and Genetic Algorithms for the


Optimization of Flat-Plate Solar Collectors
Soteris A. Kalogirou
Department of Mechanical Engineering and Materials Science and Engineering, Cyprus University of
Technology, P. O. Box 50329, 3603, Limassol, Cyprus, email: Soteris.kalogirou@cut.ac.cy

Abstract:
This paper employs exergy analysis to derive a general equation for the exergy efficiency of flat plate
collectors and thus optimize its design and operation. Exergy analysis of a flat-plate solar collector is a more
effective method of finding the optimum relationship between flow rate and the collector area. The fixed
parameters in this optimization are the collector inlet temperature, available solar radiation, the collector
transmittance-absorptance product and the ambient temperature. The collector heat loss coefficient is
estimated according to the collector plate temperature, wind convection loss, number of glass covers,
collector inclination, ambient temperature and emittance of collector plate and glass cover. In order to find
the optimum value of this multivariable problem genetic algorithms are used which are based on the
principles of genetics and survival of the fittest. The optimization parameter is exergy efficiency and the
objective is to maximize this parameter. Genetic algorithms proved suitable and very quick in obtaining the
required results. These results prove that the exergy efficiency of a flat-plate solar collector is maximized for
small distances between the riser tubes and for very small diameter of these tubes. By using a more
practical distance of 10 centimetres between the tubes, and excluding this parameter from the optimization
procedure, very small differences are observed in maximum exergy efficiency and if the cost of the
materials is accounted this is a more cost-effective solution. Other findings are that the exergy efficiency
increases considerably at higher solar radiation and that the transmittance absorptance product affects to a
great extent the exergy efficiency.

Keywords:
Exergy analysis, Genetic algorithms, Flat-plate collectors, Optimisation, Efficiency.

1. Introduction
Solar energy collectors are special kind of heat exchangers that transform solar radiation energy to
internal energy of a transport fluid. Flat plate collectors are the most popular type of solar devises
for low temperature applications. The main use of these collectors is in solar water heating systems
operating at maximum temperatures of 80-90C. M ost of these systems are operating
thermosiphonically at very small flow rate created by the small density differences between the hot
and cold water. A number of researchers have used exergy analysis to design flat plate collectors.
Badescu [1] optimized the width and thickness of the fins of a flat-plate collector by minimizing the
cost per unit useful heat flux. The proposed procedure allows computation of the necessary
collection surface area. A rather involved, but still simple, flat plate solar collector model is used in
the calculations. M odel implementation requires a specific geographical location with a detailed
meteorological database available. Fins of both uniform and variable thickness were considered.
The optimum fin cross section is very close to an isosceles triangle. The fin width is shorter and the
seasonal influence is weaker at lower operation temperatures. Fin width and thickness at the base
depend on the season. The optimum distance between the tubes is increased by increasing the inlet
fluid temperature, and it is larger in the cold season than in the warm season.
Badescu [2] also considered the best operation strategies for open loop flat-plate solar collector
systems. A direct optimal control method (the TOM P algorithm) is implemented. A detailed
collector model and realistic meteorological data from both cold and warm seasons are used. The
maximum exergetic efficiency is low (usually less than 3%), in good agreement with experimental
185

measurements reported in literature. The optimum mass-flow rate increases near sunrise and sunset
and by increasing the fluid inlet temperature. The optimum mass-flow rate is well correlated with
global solar irradiance during the warm season. Also, operation at a properly defined constant massflow rate may be close to the optimal operation.
Torres-Reyes et al. [3] presented a procedure to establish the optimal performance parameters for
the minimum entropy generation during the collection of solar energy. The Entropy Generation
Number, Ns, and the criterion for the optimal thermodynamic operation of a collector under nonisothermally, finite-time conditions, are reviewed. The M ass Flow Number, M , corresponding to the
optimum flow of working fluid as a function of the solar collection area, is also considered. A
general method for the preliminary solar collector design, based on Ns, M and the SunAir or
stagnation temperature, is developed. This last concept is defined as the maximum temperature that
the collector reaches at non-flow conditions for a given geographic location, geometry and
construction materials. The thermodynamic optimization procedure was used to determine the
optimal performance parameters of an experimental solar collector.
Torres-Reyes et al. [4] also presented the thermodynamic optimization of flat plate collectors based
on the first and the second law, developed to determine the optimal performance parameters and to
design a solar to thermal energy conversion system. An exergy analysis is presented to determine
the optimum outlet temperature of the working fluid and the optimum path flow length of solar
collectors with various configurations. The collectors used to heat the air flow during solar-tothermal energy conversion are internally arranged in different ways with respect to the absorber
plates and heat transfer elements. The exergy balance and the dimensionless exergy relationships
are derived by taking into account the irreversibilities produced by the pressure drop in the flow of
the working fluid through the collector. Design formulas for different air duct and absorber plate
arrangements are obtained.
The use of genetic algorithms (GAs) for the optimal design of solar collectors is well known [5, 6].
Genetic algorithms have been used as a design support tool by Loomans and Visser [7] for the
optimization of large hot water systems. The tool calculates the yield and the costs of solar hot
water systems based on technical and financial data of the system components. The genetic
algorithm allows for the optimization of separate variables as the collector type, the number of
collectors, the heat storage capacity and the collector heat exchanger area.
Kalogirou [8] used also genetic algorithms together with a neural network for the optimization of
the design of solar energy systems. The method is presented by means of an example referring to an
industrial process heat system. The genetic algorithm is used to determine the optimum values of
collector area and the storage tank size of the system which minimize the solar energy price.
According to the author the solution reached is more accurate than the traditional trial and error
method and the design time is reduced substantially.
Krause et al. [9] presented a study in which two solar domestic hot water systems in Germany have
been optimized by employing validated TRNSYS models in combination with genetic algorithms.
Three different optimization procedures are presented. The first concerns the planning phase. The
second one concerns the operation of the systems and should be carried out after about one year of
data is collected. The third one examines the daily performance by considering predictions of
weather and hot water consumption and actual temperature level in the storage tank.
The objective of the present work is to maximize the exergy efficiency of flat-plate collectors. The
fixed parameters in this optimization are the collector inlet temperature, available solar radiation,
the collector transmittance-absorptance product and the ambient temperature. The collector heat
loss coefficient is estimated according to the collector plate temperature, wind convection loss,
number of glass covers, collector inclination, ambient temperature and emittance of collector plate
and glass cover. For this purpose an evolution strategy based on genetic algorithms is used to
determine the optimum solution.

2. Energy Analysis
186

In this section various relations that are required in order to determine the useful energy collected
and the interaction of the various constructional parameters on the performance of a collector are
presented.
The useful energy collected from a collector can be obtained from the following formula [10]:
Qu

Ac FR Gt (

) UL Tf,i

Ta

mc p Tf,o Tf,i

(1)

where FR is the heat removal factor given by [10]:


FR

mc p
U F ' Ac
1 Exp L
A c UL
mc p

(2)

In (2) F is the collector efficiency factor which is calculated by considering the temperature
distribution between two pipes of the collector absorber and by assuming that the temperature
gradient in the flow direction is negligible [10]. This analysis can be performed by considering the
sheet-tube configuration shown in Fig. 1, where the distance between the tubes is W, the tube
diameter is D, and the sheet thickness is .

Fig. 1. Schematic diagram of a flat-plate sheet and tube configuration


As the sheet metal is usually made from copper or aluminum which are good conductors of heat, the
temperature gradient through the sheet is negligible, therefore the region between the centerline
separating the tubes and the tube base can be considered as a classical fin problem. By following
this analysis the equation to estimate F is given by [10]:

1
UL

F'
1
W
U L D (W

D) F

1
Cb

(3)

1
Di hfi

In (3), Cb is the bond conductance which can be estimated from knowledge of the bond thermal
conductivity, the average bond thickness, and the bond width. The bond conductance can be very
important in accurately describing the collector performance and generally it is necessary to have
good metal-to-metal contact so that the bond conductance is greater that 30 W/mK and preferably
the tube should be welded to the fin.
Factor F in (3) is the standard fin efficiency for straight fins with rectangular profile, obtained from:

tanh n(W D) / 2
n(W D) / 2

(4)

where n is given by:


187

UL
k

(5)

The collector efficiency factor is essentially a constant factor for any collector design and fluid flow
rate. The ratio of U L to Cb, the ratio of UL to hfi, and the fin efficiency F are the only variables
appearing in (3) that may be functions of temperature. For most collector designs F is the most
important of these variables in determining F . The factor F is a function of U L and hfi, each of
which has some temperature dependence, but it is not a strong function of temperature.
Additionally, the collector efficiency factor decreases with increased tube center-to-center distances
and increases with increases in both material thicknesses and thermal conductivity. Increasing the
overall loss coefficient decreases F while increasing the fluid-tube heat transfer coefficient
increases F .
Therefore it is obvious from the above analysis that by increasing F more energy can be intercepted
by the collector. By keeping all other factors constant increase of F can be obtained by decreasing
W. However, decrease in W means increased number of tubes and therefore extra cost would be
required for the construction of the collector.
The collector efficiency is found by dividing Q u given in (1), by the incident radiation A c Gt. By
doing so the following Equation is obtained:
FR (

) FR U L

Tf ,i Ta
Gt

mc p Tf ,o Tf,i

(6)

AcG t

By plotting against
/Gt a straight line is obtained with the slope equal to F RUL, called the loss
coefficient and the intercept on the y-axis equal to F R ), called optical efficiency.
Generally the efficiency of a solar thermal system increases by increasing the flow rate. This
increase is asymptotic, i.e., the increase is rapid at small flow rates and becomes almost asymptotic
to a certain maximum value at higher values of flow rate. In a similar way the collector outlet
temperature increases with the collector area and decreases with increasing flow rate and vice versa.
This creates difficulties in selecting appropriate values of flow rate and collector area.
In a real system the variation of the fluid inlet temperature depends to a great extent on the storage
tank configuration, thermal load demand and the consequent make-up water to the storage tank
which affects the fluid inlet temperature to the collector.

3. Exergy Analysis
According to Kalogirou [10] the temperature at any position y at a fluid inlet temperature T f,i is
given by:

Tf

Ta

S
UL

Tf ,i

Ta

S
UL

Exp

U L F' NWy
mc p

(7)

For y=L the collector area is given by Ac = NWL. If T f,i = T a the fluid temperature variation at the
exit from the solar collector is:

Tf ,o Ta

S
UL

1 Exp

U L F' A c
cp m

(8)

The flow rate of exergy transferred from the sun to the fluid that is heated while crossing the riser
pipe is:

Ef

mef

m h f ,o

h f ,i

Ta s f ,o sf ,i

(9)
188

Where hf,o - hf,i = cp(T f,o - T f,i) is the variation of specific enthalpy. The variation of the specific
entropy is:

sf ,o s f ,i

c p ln

Tf ,o

(10)

Tf ,i

Therefore:

Ef

mcp

T Ta ln

Tf ,o

(11)

Tf ,i

And the exergy efficiency is given by dividing E f by the available solar radiation Qs,in which is
equal to Ac Gt:

ex

Ef
Q S,in

Replacing

mc p
ex

A cG t

T Ta ln
mc p

Tf ,o
Tf ,i

(12)

AcGt

given by (8) in (12), the following relation for the exergy efficiency can be obtained:

S
1 Exp
UL

U L F ' Ac
mcp

Ta ln 1

S
1 Exp
Ta U L

U L F ' Ac
mc p

(13)

As can be seen from the exergy analysis, a general equation for the exergy efficiency of flat plate
collectors is derived. This efficiency depends on the values of flow rate, collector area and collector
efficiency factor. The latter depends on the distance between consequent riser tubes, the diameter of
the riser tubes, the collector fin efficiency, the internal riser tube diameter (depends on outside tube
diameter) and the convection coefficient inside the tube - which depends on the mass flow rate and
tube inside diameter, which affect the Reynolds number and thus the Nusselt number.

4. Method Description
The objective of this work is to find the parameters that maximize the exergy efficiency. In order to
do this a numbers of parameters need to be considered as constants. These are shown in Table 1.
Table 1. Constant parameters used in exergy optimization
Parameter
Fluid specific heat, cp
Available solar radiation, G t
Transmittance-absorptance product, ( )
Ambient temperature, Ta
Bond conductance, Cb
Absorbing plate thickness,
Thermal conductivity of fluid, k

Value
4185 J/kg-K
500-1000 W/m2
0.60-0.90
25C
100 W/m2-K
0.5 mm
410 W/m-K

From these constant parameters a number of other parameters are evaluated. These are:
1. Heat loss coefficient, U L, estimated from [10]:

189

Tp

1
Ng

UL

C Tp Ta
Tp N g f

0.33

1
hw

1
0.05N g 1

Ta
p

Tp Ta
2N g f 1

(14)
Ng

Where:
f

1 0.04h w

0.0005h w2 1 0.091N g

C 365.9 1 0.00883

hw

0.0001298

(15)
(16)

8.6V 0.6
L0.4

(17)

This estimation ignores the bottom and edge losses and considers the wind velocity to be 1 m/s,
collector slope, , equal to 45, one glass cover and emittance values of 0.8. The estimation is also
done at a plate temperature T p to be 20C above the mean collector fluid temperature.
2. Factor n estimated from (5)
3. Fin efficiency F, estimated from (4)
4. The convection heat transfer coefficient inside the pipe, hf,i, estimated from the principles of
heat transfer and according to the type of flow, turbulent or laminar according to the mass flow
rate and the riser tube diameter.
5. The collector efficiency factor F , estimated from (3)
For the above parameters the inputs required are the distance between successive riser tubes, W, the
riser tube diameter, D, and the collector mass flow rate. These are varied during the optimization
process until the exergy efficiency is maximized using a genetic algorithm described in the
following section.
It should be noted that a proper exergy optimization should include optimization of both the
process parameters and the size of equipment by considering two key parameters, i.e., exergy
efficiency and cost, which include the capital investment (proportional to the collector area) and
operating cost due to the pressure drop. However, in this work the pressure drop is not included as
the collector is assumed to operate at very small flow rate (similar to the thermosyphonic one),
therefore, no exergy consumption for the fluid pumping is included in the analysis and this should
shift the optimum to the smaller diameters of tubes and distances between the tubes. This
assumption will be proved by the obtained results. Additionally, no other cost-connected parameter
is included in the optimization, such as exergy cost of equipment, so the optimization related to the
collector area only affects the amount of radiation collected by the system.

5. Genetic Algorithm
The genetic algorithm (GA) is a model of machine learning, which derives its behavior from a
representation of the processes of evolution in nature. This is done by the creation within a
machine/computer of a population of individuals represented by chromosomes. Essentially these are
a set of character strings that are analogous to the chromosomes that we see in the DNA of human
beings. The individuals in the population then go through a process of evolution.
It should be noted that evolution as occurring in nature or elsewhere is not a purposive or directed
process, i.e., there is no evidence to support the assertion that the goal of evolution is to produce
M ankind. Indeed, the processes of nature seem to end to different individuals competing for
resources in the environment. Some are better than others are, those that are better are more likely
to survive and propagate their genetic material.
190

In nature, the encoding for the genetic information is done in a way that admits asexual
reproduction, which typically results in offspring that are genetically identical to the parent. Sexual
reproduction allows the creation of genetically radically different offspring that are still of the same
general species.
In an over simplified consideration, at the molecular level what happens is that a pair of
chromosomes bump into one another, exchange chunks of genetic information and drift apart. This
is the recombination operation, which in GAs is generally referred to as crossover because of the
way that genetic material crosses over from one chromosome to another.
The crossover operation happens in an environment where the selection of who gets to mate is a
function of the fitness of the individual, i.e., how good the individual is at competing in its
environment. Some GAs use a simple function of the fitness measure to select individuals
(probabilistically) to undergo genetic operations such as crossover or asexual reproduction, i.e., the
propagation of genetic material remains unaltered. This is fitness - proportionate selection. Other
implementations use a model in which certain randomly selected individuals in a subgroup compete
and the fittest is selected. This is called tournament selection. The two processes that most
contribute to evolution are crossover and fitness based selection/reproduction. M utation also plays a
role in this process.
GAs are used for a number of different application areas. An example of this would be
multidimensional optimization problems in which the character string of the chromosome can be
used to encode the values for the different parameters being optimized.
In practice, this genetic model of computation can be implemented by having arrays of bits or
characters to represent the chromosomes. Simple bit manipulation operations allow the
implementation of crossover, mutation and other operations.
When the GA is executed, it is usually done in a manner that involves the following cycle [11]:
Evaluate the fitness of all of the individuals in the population.
Create a new population by performing operations such as crossover, fitness-proportionate
reproduction and mutation on the individuals whose fitness has just been measured.
Discard the old population and iterate using the new population.
One iteration of this loop is referred to as a generation. M ore details on genetic algorithms can be
found in Goldberg [12].
The first generation of this process operates on a population of randomly generated individuals.
From there on, the genetic operations, in concert with the fitness measure, operate to improve the
population. Genetic algorithms (GA) are suitable for finding the optimum solution in problems
where a fitness function is present. Genetic algorithms use a fitness measure to determine which
of the individuals in the population survive and reproduce. Thus, survival of the fittest causes good
solutions to progress. A genetic algorithm works by selective breeding of a population of
individuals, each of which could be a potential solution to the problem. The genetic algorithm is
seeking to breed an individual, which either maximizes, minimizes or it is focused on a particular
solution of a problem. In this case, the genetic algorithm is seeking to breed an individual that
maximizes the exergy efficiency of the solar collector.
The larger the breeding pool size, the greater the potential of it producing a better individual.
However, the fitness value produced by every individual must be compared with all other fitness
values of all the other individuals on every reproductive cycle, so larger breeding pools take longer
time. After testing all of the individuals in the pool, a new generation of individuals is produced
for testing.
A genetic algorithm is not gradient based, and uses an implicitly parallel sampling of the solutions
space. The population approach and multiple sampling means that it is less subject to becoming
trapped to local minima than traditional direct approaches, and can navigate a large solution space
with a highly efficient number of samples. Although not guaranteed to provide a globally optimum
191

solution, GAs have been shown to be highly efficient at reaching a very near optimum solution in a
computationally efficient manner.
During the setting up of the GA the user has to specify the adjustable chromosomes, i.e. the
parameters that would be modified during evolution to obtain the maximum or minimum values of
the fitness functions. In this work, the fitness function is exergy efficiency given by (13).
Additionally the user has to specify the range of the input parameters called constraints.
The genetic algorithm parameters used in the present work are:
Population size=50
Population size is the size of the genetic breeding pool, i.e., the number of individuals contained in
the pool. If this parameter is set to a low value, there would not be enough different kinds of
individuals to solve the problem satisfactorily. On the other hand, if there are too many in the
population, a good solution will take longer to be found because the fitness function must be
calculated for every individual in every generation.
Crossover rate=90%
Crossover rate determines the probability that the crossover operator will be applied to a particular
chromosome during a generation.
M utation rate=1%
M utation rate determines the probability that the mutation operator will be applied to a particular
chromosome during a generation.
Generation gap=96%
Generation gap determines the fraction of those individuals that do not go into the next generation.
It is sometimes desirable that individuals in the population be allowed to go into next generation.
This is especially important if individuals selected are the most fit ones in the population.
Chromosome type=continuous
Populations are composed of individuals, and individuals are composed of chromosomes, which are
equivalent to variables. Chromosomes are composed of smaller units called genes. There are two
types of chromosomes, continuous and enumerated. Continuous are implemented in the computer as
binary bits. The two distinct values of a gene, 0 and 1, are called alleles. M ultiple chromosomes
make up the individual. Each partition is one chromosome, each binary bit is a gene, and the value
of each bit (1, 0, 0, 1, 1, 0) is an allele. Enumerated chromosomes consist of genes, which can have
more allele values than just 0 and 1.
The genetic algorithm is usually stopped after best fitness remained unchanged for a number of
generations or when the optimum solution is reached. In this work the genetic algorithm was
stopped after best fitness remained unchanged for 75 generations.

6. Results
The input parameters (adjustable chromosomes) were used in a genetic algorithm program to find
the values that maximize collector exergy efficiency. The whole model was set up in a
spreadsheet program in which the various parameters are entered into different cells. The optimum
values of the various parameters were used in (13) to estimate the collector exergy efficiency which
is the fitness function that needs to be maximized. The various input parameters are constrained to
vary between certain values. The ones used in this work are shown in Table 2.

192

Table 2. Constrains of the input parameters


Parameter
M ass flow rate
Collector area
Distance between riser tubes
Riser tube diameter

Symbol
m

Ac
W
D

Range
0.001-0.08 kg/s
1-10 m2
0.03-0.15 m
0.004-0.022 m

The optimum results can be presented graphically or in tables. Here the table presentation is
preferred so as to show the exact values obtained. The results are shown in Tables 3-5. It should be
noted that for each run of the program the optimum solution was reached in less than 5 seconds on a
Pentium 3.2 GHz machine, which is very fast.
Table 3 presents the results of the optimization process with respect to the effect of inlet
temperature and radiation.
Table 3. Results of the optimization process-effect of inlet temperature and radiation
Gt (W/m2)
( ) T in (C)
W (m) D (m)
T out (C)
Ac (m2)
ex (%)
25
0.03
0.004
3.07
66.02
7.49
27
0.03
0.004
4.85
59.86
9.98
500
0.8
30
0.03
0.004
7.90
52.24
9.61
35
0.03
0.013
13.77
43.27
8.76

m (kg/s)

0.0106
0.0196
0.0304
0.0797

According to the results shown in Table 3 by increasing the collector inlet temperature the collector
outlet temperature reduces, the optimum flow rate increases and the exergy efficiency increases. As
can be seen the distance between the riser tubes remains to the minimum value, same as the riser
tube diameter except the last value which differs from the minimum value possible, determined by
the constrains. It is apparent that the optimum collector area reaches a maximum value at the inlet
temperature of 27C and then drops for bigger values. The effect of solar radiation is shown in
Table 4.
Table 4. Results of the optimization process-effect of solar radiation
Gt (W/m2)
( ) T in (C)
W (m) D (m)
T out (C)
ex (%)
500
25
0.03
0.004
3.07
66.02
500
27
0.03
0.004
4.85
59.86
0.8
1000
25
0.03
0.004
5.19
95.88
1000
27
0.03
0.004
6.21
92.22

Ac (m2)
7.49
9.98
8.24
9.93

m (kg/s)

0.0106
0.0196
0.0140
0.0190

As can be seen from Table 4 by increasing the solar radiation and keeping the other parameters
constant, the collector outlet temperature increases, as expected, and the same applies to the
optimum collector area and the optimum mass flow rate. Both the riser tube spacing, W, and
diameter, D, remain at their minimum values. The effect of solar radiation and the value of the
transmittance-absorptance product are presented in Table 5.
Table 5. Results of the optimization process-effect of radiation and transmittance absorptance
product
Gt (W/m2)
( ) T in (C)
W (m) D (m)
T out (C)
Ac (m2)
m (kg/s)
ex (%)
500
0.6
0.03
0.004
3.57
49.56
10.0
0.0224
500
0.9
0.03
0.004
5.60
64.56
9.96
0.0190
27
1000
0.6
0.03
0.004
4.10
77.32
9.68
0.0180
1000
0.9
0.03
0.004
7.39
99.03
9.26
0.0180
The results presented in Table 5 reveal that the increase of the transmittance-absorptance product
causes increase in the collector outlet temperature, and a slight decrease in the optimum collector
193

area and the optimum mass flow rate. The exergy efficiency however almost doubles. It is thus
required to achieve a high value of ( ) as possible in flat-plate collectors. Again both, the riser tube
spacing, W, and diameter, D, remain at their minimum values.
As seen from the above tables the distance between the riser tubes, W, is very small. In fact it could
be zero but was stuck to the minimum value specified as a constrain. For this reason a more
practical distance of 10 centimetres between the tubes is used and this parameter is excluded from
the optimization procedure. By doing so the results shown in Table 6 are obtained.
Table 6. Optimization results for a fixed distance between riser tubes of 10cm
Gt (W/m2)
)
T in (C)
D (m)
T out (C)
Ac (m2)
ex (%)
500
0.8
27
0.004
4.72
59.6
9.98
1000
0.8
27
0.004
6.02
92.1
8.38

m (kg/s)

0.0193
0.0156

As can be seen for this more practical case, and by comparing the values shown with the values
presented in previous tables, the exergy efficiency is not affected too much. This case however is
much more cost effective because the minimum the distance between the riser tubes the maximum
would be the number of tubes and the collector will cost more.
Another practical case that needs to be investigated is the effect of riser pipe diameter on the exergy
efficiency. For this exercise the distance between the riser pipes is kept to 10 cm and the other
parameters, like optimum collector area and mass flow rate, as the ones presented in Table 6. The
results of this exercise are shown in Table 7.
Table 7. Effect of pipe diameter for a fixed distance between riser tubes of 10 cm
Gt (W/m2)
)
T in (C)
Ac (m2)
m (kg/s)
D (m)
500
0.8
27
9.98
0.0193
0.004
0.008
0.012
0.015
1000
0.8
27
8.38
0.0156
0.004
0.008
0.012
0.015

ex (%)

4.72
4.71
4.70
4.69
6.02
6.01
5.99
5.99

As can be seen the pipe diameter can be increased from the optimum small diameter without any
problem as the exergy efficiency is marginally affected for the bigger sizes. So this needs to be
decided solely on cost, which increases for bigger diameter pipes and the friction factor imposed by
the smaller diameter pipe, which needs to be kept as small as possible in order not to block the
thermosiphonic flow.
It should be noted that in all results presented in Tables 3-7 a very small flow rate is given as the
optimum and the actual value approaches the thermosiphonically created one which is the most
frequently used operation mode for flat-plate collectors. So the assumption made earlier about the
small flow rate and the exclusion of fluid pumping from the analysis is proved.

7. Conclusions
It is proved in this paper that exergy analysis and genetic algorithms are very suitable for obtaining
the required results quickly. These results prove that the exergy efficiency of a flat-plate solar
collector is maximized for small distances between the riser tubes and for very small diameter of
these tubes. By using a more practical distance of 10 centimeters between the tubes, and excluding
this parameter from the optimization procedure, very small differences are observed in maximum
exergy efficiency and if the cost of the materials is accounted this is a more cost-effective solution.
The exergy efficiency it is also insensitive to the size of the riser pipe diameter. Other findings
194

prove that exergy efficiency increases considerably at higher solar radiation and that the
transmittance absorptance product affects to a great extent the exergy efficiency.

Nomenclature
Ac
B
C
Cb
cp
D
Di

collector area, m2
bond width , m
factor given by (16)
bond conductance, W/(mK)
specific heat capacity, J/(kgK)
riser tube outside diameter, m
riser tube inside diameter, m

Ef

exergy flow rate, W

specific exergy, J/kg


collector efficiency factor
fin efficiency
factor given by (15)
heat removal factor
solar radiation, W/m2
specific enthalpy, J/kg
heat transfer coefficient inside absorber tube, W/(m2K)
wind loss coefficient, W/(m2K)
solar radiation, W/m2
absorber plate thermal conductivity, W/(mK)
bond thermal conductivity, W/(mK)
collector length, m
mass flow rate, kg/s
factor given by (5)
number of riser tubes
number of glass covers
available solar radiation, W/m2
rate of useful energy collected, W
specific entropy, J/(kgK)
power absorbed per unit area of collector, W/m2
ambient temperature, K
fluid temperature, K
collector inlet temperature, K
collector outlet temperature, K
plate temperature, K
overall heat loss coefficient, W/(m2K)
wind speed, m/s
distance between riser tubes, m

F
f
FR
Gt
h
hfi
hw
I
k
kb
L
m
n
N
Ng
Qs,in
Qu
s
S
Ta
Tf
T f,i
T f,o
Tp
UL
V
W

Greek sym bols


collector slope, degrees
195

p
g

absorber (fin) thickness, m


plate emittance
glass cover emittance
temperature difference, K
transmittance-absorptance product

References
[1] Badescu V., Optimum fin geometry in flat plate solar collector systems, Energy Conversion and
M anagement 2006; 47(15-16): 2397-413.
[2] Badescu V., Optimal control of flow in solar collectors for maximum exergy extraction,
International Journal of Heat and M ass Transfer 2007; 50: 4311-22.
[3] Torres-Reyes E., Cervantes-de Gortari J.G., Ibarra-Salazar B.A., Picon-Nuez M ., A design
method of flat-plate solar collectors based on minimum entropy generation, Exergy the
International Journal 2001; 1(1): 46-52.
[4] Torres-Reyes E., Navarrete-Gonzalez J.J., Zaleta-Aguilar A., Cervantes-de Gortari J.G.,
Optimal process of solar to thermal energy conversion and design of irreversible flat-plate solar
collectors, Energy 2003; 28(2): 99-113.
[5] Kalogirou S.A., Use of genetic algorithms for the optimal design of flat plate solar collectors,
Proceedings of the ISES 2003 Solar World Congress on Solar energy for a sustainable future,
2003 June 14-19; Goteborg, Sweden, on CD ROM .
[6] Kalogirou S.A., Artificial neural networks and genetic algorithms for the optimisation of solar
thermal systems, In: Sayigh, A.S., Editor, WREC IX: Proceedings of the IX World Renewable
Energy Congress on CD-ROM , 2006; Florence, Italy.
[7] Loomans M ., Visser H., Application of the genetic algorithm for the optimisation of large solar
hot water systems, Solar Energy 2002; 72: 427-39.
[8] Kalogirou, S.A., Use of artificial intelligence for the optimisation of solar systems, International
Journal of Renewable Energy Engineering 2002; 4: 499-505.
[9] Krause M ., Valen K., Wiese F., Ackermann H., Investigations on optimizing large solar thermal
systems, Solar Energy 2002; 73: 217-25.
[10] Kalogirou S.A., Solar Energy Engineering: Processes and Systems, New York: Academic
Press, Elsevier Science; 2009.
[11] Zalzala A., Fleming P., Genetic Algorithms in Engineering Systems, London, UK: The
Institution of Electrical Engineers; 1997.
[12] Goldberg D.E., Genetic Algorithms in Search Optimisation and M achine Learning, Reading,
MA: Addison-Wesley; 1989.

196

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Experimental study of tar and particles content of


the produced gas in a double stage downdraft
gasifier
Ana Lisbeth Galindoa , Sandra Yamile Giraldob , Rene Lesme-Janc, Vladimir Melian
Cobasd , Rubenildo Viera Andradee and Electo Silva Loraf
a

Federal university of Itajub (UNIFEI), Itajuba, Brazil, lisbethgn37@hotmail.com


b
Federal university of Itajub (UNIFEI), Itajuba, Brazil, sayagire@yahoo.com
c
University of Oriente, Cuba, lesme3258@yahoocom
d
Federal university of Itajub (UNIFEI), Itajub, Brazil, vlad@unifei.edu.br
e
Federal university of Itajub (UNIFEI), Itajub, Brazil, ruben@unifei.edu.br
e
Federal university of Itajub (UNIFEI), Itajub, Brazil, esl43@yahoo.com

Abstract:
Biomass gasification is not a new technology, but there is a renewed interest in its further development,
mainly to produce power and heat as part of locally based combined heat and power plants. The produced
gas mainly consists of H2, CO, CO2, CH4, H2O and some trace impurities such as H2S, COS, NH3, HCN, HCl,
alkali, tar and particulate matter. These im purities are the responsible for clogging, corrosing, poisoning and
carbon deposition on different elements of the power systems, like alternative internal combustion engines,
gas turbines, also in fuel cells and auxiliary equipments, being necessary their removal or the adjusting of its
concentration level, depending on the final application.
This work presents the experimental evaluation of the tar and particle content in t he produced gas from fixed
bed downdraft gasifier with two stages of air supply. A very widely considered technology for t he biomass
gasification is the downdraft fixed-bed reactor, because had shown to produce a gas with lower tar level,
compared with other gasifiers. The ex periments were carried out varying the amount of air supplied to the
reactor and the air ratio between the two stages (AR). The temperatures in different points along of gasifier
and the gas compositions were also measured. The results show that there is an operational point where
there is a coincidence of the highest conversion of the gasifier and the better quality of the gas, (higher
calorific and lower tar content) and also that the use of a s econd stage can reduce the gas tar content up to
87%.

Keywords:
Biomass gasification, Double stage downdraft gasifier, Produced gas, Tar content, particle content.

1. Introduction
Biomass has been considered as a promising source of energy for the partial substitution of fossil
fuels, not only because its global potential, but as a neutral source of CO 2 ; its thermal conversion
generates low emissions of SO2 and NOx, it is cheap and relatively fast-growing, can be grown on
marginal land, without affecting food crops and helps the retention of water and fertilizers in the
soil. The possibility to produce fuels from biomass on a large scale reduces the greenhouse effect,
the environmental pollution and also increases the security and energetic independence of the
countries, but its challenging complexity lies on the need to implement all the stages of the
technological biomass production chain, for the conversion into biofuels and chemicals or in
electrical generation. There are several processes to transform chemical energy from biomass into
thermal energy (combustion, gasification and pyrolysis), of which, gasification is the one with
greatest perspectives, because it offers advantages as: a higher efficiency of conversion, compared
with combustion and pyrolysis. In the practice, the gasification can convert from 60 to 90% of the
197

biomass energy conversion in gas energy [1,2]. The produced gas by biomass gasification is
composed mainly of H2 , CO, CO2 , H2O, CH4 , light hydrocarbons (C x Hy ) and some impurities, as
heavy hydrocarbons (tar), nitrogen compounds (NH3 , HCN), sulphur compounds (H2 S, COS) and
solid particles. This gas can be used for energy production in internal combustion engines, gas
turbines and fuel cells or for the production of synthetic natural gas (SNG) and liquid fuels or
chemicals through the Fischer Tropsch synthetic path [3-7]. Each application needs a specific
quality of the gas used, especially in terms of impurities concentration, which depend on the
characteristics of biomass, the type of the reactor used and the operational conditions [8-12].
Tar formation are one of the biggest problems that occurs during the gasification of biomass, it can
cause blockade, fouling of pipes and valves and operation problems in the equipments where the
gas is used, increasing the maintenance cost of equipments. Thus, it is necessary a gas cleaning
system for the removal of tar, whose complexity depends on the concentration limits for these
pollutants according to the final application [13,14].
Double stage gasification is being used as an economical primary method to reduce the tar content
in the produced gas from biomass [15-16]. The basic concept of a two-stage gasification process is
the separation of the two zones the pyrolysis zone and the combustion zone, with the air injection in
each stage, which may occur in a same reactor or in separate reactors. Recently, experimental
evaluations of the tar content in two and three stage downdraft gasifiers had been reported. These
gasifiers are distinguished by their constructive characteristics, where different biomasses had been
used (composition and shape) and different experimental conditions had been evaluated
(temperature, ratio of the air supplied between the stages), getting different results in terms of tar
and particle content of the gas [17-20], as it is shown in Table 1.
Table 1. Tar content in gas from downdraft gasifiers.
Operation
Tar
Temperature
(mg/
Biomass
(o C)
Nm3 )
Eucalyptus Pelets
(2 cm)

800

Air injection

43.2

Two inlets. Zone of


pyrolysis, mixed air gas.

Yang and Par


wood 10-15 mm
in length

950

10

Primary and
secondary air
preheated up to 210
0
C

Coconut husk

700-900

28

Three air inlets.

Pine wood chips of


5x5x25mm

780-850

100

Two air inlet, internal


pyrolysis gas
recirculation.

Reactor
size
25 cm
diameter,
100 cm
height
25cm
diameter,
190 cm
height,
20 cm
diameter,

Ref.

[17]

[18]
[19]
[20]

Jaojaruek et al [17] made an experimental study and evaluated the relation between the tar content
of the gas and the gasifier operational parameters, conducted in three conditions: a) the reactor
working with a single stage, b) with two stages and c) two stages, with an air-gas mixture fided at
the second stage. The results showed that the second stage reduces the gas tar content in the gas.
Additionally, using an air-gas mixture for the second stage an increase in the calorific value of the
gas was achieved. Bhattacharya and Dutta [18], tested a two stage gasifier and with a layer of coal
for the start up of the reactor. The influence of the level of the coal layer, type of biomass and
moisture content over the tar content and gas composition were checked. It was possible to find out
198

that during the start-up and warming of the gasifier, due to low temperatures, there was a large tar
production; to reduce this period its recommended to use a layer of charcoal, with a height of the
layer slightly above the primary air inlet and adjusting combustion conditions at the second air
stage. They also concluded that, for a constant airflow at the primary stage, increasing the airflow
through the secondary stage leads to a decrease of the tar content and increase the CO and H2
content. Regardless of the results achieved, there is not always a charcoal availability for the
operation of the gasifiers and warming of the air up to 210 o C requires an additional heat source.
Bhattacharya et al [19] tested a three stage gasifier. During the start up of the reactor a layer of
charcoal was also used and was possible observe that the airflow distribution of 40% at the first
stage, 28 % at the second, and 30% at 3rd stage, getting a significant decrease of tar content.
As a conclusion, the decreasing of tar formation is achieved when: a) preheating the air is made, b)
use dual stages of air injection, c) air-gas mixture in the second stage, d) optimal parameter of
operation as ER of 0.3-0.4 and moisture content of 6-20%. The tar content in these gasifiers is in the
range of 10-100 mg/Nm3 , values lower than those obtained in other types of gasifiers.
The objective of this work is to evaluate the tar and particulate content from the produced gas in a
downdraft gasifier with two-stage air injection, analyzing the influence of the total air inlet and the
ratio between stage one and stage two in these pollutants. The tested biomass was eucalyptus wood,
charcoal was not used during the start up of the gasifier and the air supplied was not heated.

2. Materials and methods


2.1. Biomass used in the tests
Eucalyptus wood with a size smaller than 5 cm in length and diameter was used; its proximate
composition (ash, volatile and fixed carbon content), ultimate composition (carbon, nitrogen,
hydrogen, sulphur and oxygen content), low heating value and moisture are shown in Table 2.
Table 2. Analyses of the biomass (dry basis)
Parameter
Value
Proximate analysis (wt %)
Ash
1.34
Volatile
83.27
Fixed carbon
15.66
Ultimate analysis (wt %)
Carbon
45.78
Nitrogen
5.92
Hydrogen
3.21
Sulphur
0.093
Oxygen
42.83
Moisture (%)
12.23
LHV (kJ/kg)
18432.67

2.2. Downdraft gasifier


A double stage downdraft gasifier, manufactured by the Brazilian Company Termoquip Energy
Alternative Ltda, was used. It has an internal diameter of 0.3 m and a height (from the reactor top
to the grate) of 1.06 m. The gasifier is built of carbon steel with an internal coating of refractory
material. Six K-type thermocouples were installed along the reactor that recorded the temperature at
different points. Two thermocouples measure the inlet air temperature and other thermocouple
199

measures the temperature of the gas at the gasifier exhaust. The air is supplied by a blower (1900
mmH2 O). The gases generated in the reactor go out by the lower section, after crossing the
gasification zone and the grid and pass finally by a cyclone where the larger solid particles are
removed.
The biomass gasification starts by burning the wood layer by and external heat source, by this way
the reactor is heated up to 150 o C (temperature at the combustion zone in the inner wall of the
reactor). Once this temperature is attained, air is fed into the gasifier. Quasi- stoichiometric
combustion conditions are adjusted reaching a bed temperature (in the combustion zone) of around
600 o C. Air flow values are measured through an orifice plate, according to the methodology given
in ISO 5167-2 [21]. After reaching the steady state (no significant variations in temperature and gas
concentrations are observed) the operating conditions are fixed, the sampling of tar and particles are
made.
Previous experimental tests [22] showed that the condition of the higher gasifier cold efficiency ( )
is 68%. That corresponds to 20 Nm3 /h of total airflow, ratio air/biomass (ER) of 0.4, with air flow
ratio between the first and second stage (AR) of 80%. Under these conditions the concentrations of
CO and H2 was 19.04% and 16.83% respectively, and low heating value of the gas was 4.53
MJ/Nm3 . These conditions were take for the experimental planning. The details of the experimental
installation and termocouples position in the gasifier are shown in Figure 1.

Figure 1. Diagram of the experimental installation and termocouples in the gasifier.

2.3. Experimental planning


Gasification variables that influences tar and particles content in the produced gas and were
considered in this study are the total air flow and the ratio between the primary and secondary stage
air flow (AR) according to the Equation (1).
AR

Va

1 est

Va

2 est

.100%

(1)

Where:
(Va) 1est : Air flow through stage 1 (Nm3 /h)
200

(Va) 2est : Air flow through stage 2 (Nm3 /h)


During the tests the temperatures along the gasifier height, the CO, CH4 , H2 and CO2 concentrations
of the produced gas were measured using a BINOS 100, HYDROS 100 and MaMos gas analyzer
system. The precission of the equipment is 0.1 for the BINOS 100, 0.01 for the HYDROS 100
and 0.01 for the MaMos. The concentrations were measured at the cyclone outlet. Nine
experimental runs with three variations of the total air flow and three values of AR, were carried
out. Values are showed in Table 3.
Table 3. Experimental Conditions for the gasification tests
Airflow (Nm3 /h)
No.
AR (%)
(Va) 1est
(Va) 2est
1
0
18
0
2
0
20
3
0
22
4
5,14
12,86
5
40
5,71
14,29
6
6,28
15,72
7
8
10
8
80
8,88
11,12
9
9,77
12,23

Total air
18
20
22
18
20
22
18
20
22

2.4. Tar and particles sampling and measurement


2.4.1. Sampling set and analytical procedure
The tar and particles sampling were carried out based on European Fifth Framework Programme
report [23], that consist in an isokinetic sampling of gas, solid filtration and tar absorption in a
solvent contained in bottles that are kept at low temperature. The diagram of the sampling line is
shown in Figure 2.

Figure 2. Diagram of tar and particle sampling line [23].


201

The module one is the pre-conditioning stage of gas and particle collection, consists of an isokinetic
tube connected to the gas line, heated with an electrical resistance to prevent condensation of the tar
inside of the tube and the vessel that contains the filter for the particles collection. Thimble quartz
filter Advantec No. 86R was used. The tube and the port filter are heated up to 250 o C. The
module two, six impingers bottles for tar dilution. The first impinger that is empty, acts as moisture
collector; the others four impinger filled with Isopropanol, in which water and tar are condensed
from the produced gas and the last impinger with silica to dry the gas. Salt/ice/water mixture to
keep the impingers at low temperature was used. The module three contains the vacuum pump to
extract the gas, flow meter and temperature indicator.
The sample is taken during an hour after established operational conditions. After the sampling, a
gravimetric analysis to determine the tar and particles content is made. The filter is weight before of
sampling. After the sampling, the filter that not only retains particle but also a part of the tar is
placed in a soxhlet extraction system, where the residual tar from the particles is separated using
isopropanol. Them, the filter is dried and the particles mass is obtained for weight difference. The
bottles content and the liquid produced in the soxhlet extraction are evaporated by rotavaporation to
remove the solvent and the tar mass is obtained. An analytical balance model BL 210S to determine
the mass of particles and tar in the sample was used. Finally, the tar and particles content per
volume of gas sampled is reported (mg/Nm3 ).

3. Results and discussions


3.1 Temperatures behaviour inside the gasifier
During the development of the experimental runs, a data acquisition system was used to register the
temperatures at different section of the reactor. The Figure 2 and 3 shows the typical variation of
temperature over time for different sections of the reactor for total air of 20Nm3 /h. Figure 2 for AR
of 0% (unique stage air inlet) and Figure 3 for AR of 80% (two stage of air inlet).

Figure 2. Temperature profile of the gasifier for AR of 0% and total airflow of 20 Nm 3 /h

202

Figure 3. Temperature profile of the gasifier for AR of 80% and total airflow of 20 Nm 3 /h
In Figures 2 and 3 its shown that 20 minutes after ignition begins a sudden rise in gasifier
temperatures, is possible to observe how the use of a second stage of air supply increases the
pyrolysis zone temperature, bringing it closer to the temperature of the combustion zone in the
reactor. According of Figure 2 and 3, the average temperature in the pyrolysis zone was 539 o C and
661 o C, and the difference with respect to average temperature in the combustion zone was 174 o C
and 40 o C for one stage and two stages respectively. This behavior reduces the amount of tar
because a high temperature inside the reactor induces the tar destruction. Figure 4 shows the
average temperature of different zone in the gasifier depending of the total air flow.

Figure 4. Average temperature inside of the gasifier for AR of 80% and different airflow.
As its shown in the Figure 4, its possible to confirm that in a two stage downdraft gasifier, the
temperature in the pyrolysis zone is closer to the temperature in the combustion zone, being highly
differentiated at high air flows.
203

3.2. Composition and low heating value of the produced gas


During the development of the experimental runs, a data acquisition system was used to register
CO, CO2 , H2 , CH4 concentrations. The low heating value (LHV) of gas is obtained from the fuel
gases concentration of CO, CH4 and H2 , and their energy contents as shown in Equation (2). Those
experimental results for AR of 80% are presented in Figure 5.

LHV

10790. % H 2

12630. % CO

35800. %CH4

(2)

Figure 5. Composition and LHV of gas for AR=80%.

Figure 5 shows the CO, CO 2 , CH4 , H2 concentration of the produced gas for AR of 80% for runs 7
to 9. It also shows LHV, as a function of the total airflow fed to the gasifier. At 20 Nm3 /h1 of total
air flow (run 8) the highest H2 concentration was 17.14 %v and the LHV of 4738 kJ/Nm3 in the
produced gas is reached. At 22 Nm3 /h (run 9), the H2 concentration and the LHV showed a slight
decrease reaching 16.56 %v and 4714 kJ/Nm3 respectively. From this total airflow, the process
begins to be favored by combustion given the temperature increase along the gasifier. The higher
concentrations at these conditions are also attributed to the good stability and performance of the
combustion and pyrolysis zones.

3.3. Tar and particles content in the produced gas


The tar and particles content of the produced gas are presented in Table 4 and the influence of
different parameters on the results are shown in Figures 6-7.
Table 4. Tar and particles content of the gas.
Results
Run 1 Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9
AR
(%)
0
0
0
40
40
40
80
80
80
3
Total air flow Nm /h
18
20
22
18
20
22
18
20
22
3
Tar
mg/Nm 1269.70 418.95 179.86 76.09 104.99 78.57 171.49 54.09 99.61
Particles
mg/Nm3 216.45 146.04 176.04 142.39 107.16 164.99 97.19 22.69 292.2
Parameters

Units

204

Figure 6. Influence of AR over the tar content at different total airflows..

Figure 7. Influence of AR on the gas particles content at different total airflow.

According to Figure 6, at the point of higher efficiency of the gasifier when AR is 80% and total air
flow is 20 Nm3 /h, the tar content of 54.09 mg/Nm3 is within the range and even lower than values
reported in Table 1 for other two stage downdraft gasifiers, and without the use of an air- gas
mixture through the second stage. As it is observed in Figure 7, at same conditions, the particles
content presents its lower value of 22.69 mg/Nm3 . The increase of the total airflow and of the AR
values decrease the tar and particles contents in the gas as its observed in Table 4 and confirmed in
Figures 6 and 7. Specifically, this suggests that when an 80% of the total air is entering into the first
stage, the biomass volatization is favored, allowing the formation of lighter compounds that are
more easily cracked when the gas stream pass through the combustion zone, and in this way
reducing the tar content of the produced gas.

205

4. Conclusions
For the gasifier used in this study with operational conditions of 20 Nm3 /h of total air flow, ER of
0.4 and AR of 80%, a fuel gas with levels of CO, CH4 , CO2 , H2 contents of 19.2, 1.34, 14.21 and
17.14 %v was obtained, at these conditions the low heating value of the gas was 4738.4 kJ/Nm3 and
the tar content decrease to 54.09 mg/Nm3 , compared with CO, CH4 , CO2 , H2 contents and LHV of
17.90, 1.72, 13.90, 15.71 %v and 5103 kJ/Nm3 respectively, and 418.95 mg/Nm3 tar content
obtained at AR of 0% and 20 Nm3 /h. These results showed a decreasing in the tar content of 87%.
The results suggest a relationship between the CH4 concentration with the tar content in the
produced gas. In this way, increasing the air supply and use of a second stage decrease the tar
content and the CH4 concentration. This suggest advantages for the biomass devolatilization in the
pyrolysis zone, by production of lighter compounds that are easily cracked when the gas passes
through the combustion zone.
The use of a second stage of air supply increases the temperature at the pyrolysis zone. This allows
to reducing from four to three temperature zones inside the reactor: drying, the first zone, pyrolysiscombustion, the second one and gasification, the third one. This behavior leads to a decrease of the
tar content in the produced gas, so this method is considered an efficient and economic primary
method for tar conversion without air/gas mixture o preheated air as was reported in literature.

Acknowledgments
The authors would like thank the Energy Company of Minas Gerais (CEMIG) and the Secretary of
Science and Technology (SECTES) for the financial support received for the development of this
project. Also to thank the Committee on Coordination of Improvements in Higher Education
(CAPES), the National Research Council of Brazil (CNPq) and the Foundation for Research
Support of Minas Gerais State (FAPEMIG) for the graduate and research productivity grants
allocations.

References
[1] Marsh R, Hewlett S, Grif ths T, Williams K. Advanced thermal treatment for solid waste a
waste managers guide. Proceeding of the 22nd international conference on solid waste
management and technology. Philadelphia (USA); 2007
[2] Kristina Difs, Elisabeth Wetterlund, Louise Trygg, Mats Sderstrm. Biomass gasi catio n
opportunities in a district heating system. Biomass and Bioenergy. 2010: 34: 637651.
[3] Kurkela E, Simell P, mckeough P, Kurkela M. Production of synthesis gas and clean fuel gas
[synteesikaasun ja puhtaan polttokaasun valmistus]. In: VTT Publications, vol. 682. Technica l
Research Centre of Finland, VTT; 2008.
[4] Goyal HB, Seal D, Saxena RC. Bio- fuels from thermochemical conversion of renewable
resources: a review. Renew Sust Energy Rev 2008: 12: 50417.

206

[5] K.D. Panopoulos, L. Fryda, J. Karl, S. Poulou, E. Kakaras. High temperature solid oxide fue l
cell integrated with novel allothermal biomass gasi cation Part II: Exergy analysis. Journal o f
Power Sources 2006: 159: 586594
[6] H. Moritaa, F. Yoshibaa, N. Woudstrab, K. Hemmesc, H. Spliethoff. Feasibility study of wood
biomass gasi cation/molten carbonate fuel cell power systemcomparative characterization o f
fuel cell and gas turbine systems. Journal of Power Sources 2004: 138: 3140.
[7] Anastasia A. Zabaniotou, Vasiliki K. Skoulou, Dimitris P. Mertzis, George S. Koufodimos, and
Zissis C. Samaras. Mobile Gasi cation Units for Sustainable Electricity Production in Rura l
Areas: The smart-CHP Project. Ind. Eng. Chem. Res. 2011: 50: 602608.
[8] Jos Corella, Jos Manuel Toledo and Gregorio Molina. Biomass gasification with pure stea m
in fluidised bed: 12 variables that affect the effectiveness of the biomass gasifier. Int. J. Oil, Gas
and Coal Technology 2008: 1: 194-207.
[9] Li, C.-Z. Some recent advances in the understanding of the pyrolysis and gasi cation behavior
of Victorian brown coal. Fuel 2007: 86: 16641683.
[10] Hong Cui, Scott Q. Turn, Vheissu Keffer, Donald Evans, Thai Tran, and Michael Foley.
Contaminant Estimates and Removal in Product Gas from Biomass Steam Gasification. Energy
Fuels 2010: 24: 12221233.
[11] S. Monteiro Nunes, N. Paterson, D. R. Dugwell, and R. Kandiyoti. Tar Formation and
Destruction in a Simulated Downdraft, Fixed-Bed Gasifier: Reactor Design and Initial Results.
Energy & Fuels 2007: 21: 3028-3035.
[12] Takashi Yamazaki, Hirokazu Kozu, Sadamu Yamagata, Naoto Murao, Sachio Ohta, Satoru
Shiya and Tatsuo Ohba. Effect of Superficial Velocity on Tar from Downdraft Gasification o f
Biomass. Energy & Fuels 2005: 19: 1186-1191
[13] Yan Cao, Yang Wang b, John T. Riley, Wei-Ping Pan. A novel biomass air gasification
process for producing tar- free higher heating value fuel gas. Fuel Processing Technology 2006 :
87: 343 353.
[14] Devi L, Ptasinski KJ, Janssen FJJG. A review of the primary measures for tar elimination in
biomass gasi cation processes. Biomass Bioenergy 2003: 24: 125-1240.
[15] Abu El-Rub Z, Bramer EA, Brem G. Review of catalysts for tar elimination in biomass
gasi cation processes. Ind Eng Chem Res 2004: 43: 6911-6919
[16] Han J, Kim H. The reduction and control technology of tar during biomass
gasi cation/pyrolysis: an overview. Renew Sust Energy Rev 2008: 12: 397-416.
[17] Kitipong Jaojaruek, Sompop Jarungthammachote, Maria Kathrina B. Gratuito , Hataitep
Wongsuwan, Suwan Homhual. Experimental study of wood downdraft gasi cation for a n
improved produced gas quality through an innovative two-stage air and premixed air/gas supply
approach. Bioresource Technology 2011: 102: 48344840.
[18] S. C. Bhattacharya and Animesh Dutta. Two-stage gasification of wood with preheated air
supply: a promising technique for producing gas of low tar content. Energy Program, Asia n
Institute of technology, P O Box 4, Klong Luang, Pathumthani 12120, Thailand.
[19] S.C. Bhattacharya, San Shwe Hla, Hoang-Luang phamm. A study on a multi-stage hybrid
gasifer-engine system. Biomass and Bioenergy 2001: 21: 445-460
[20] Susanto H, Beenackers AACM. A moving-bed gasifier with internal recycle of pyrolysis
gas. Fuel 1996: 75: 1339-47.
[21] ISO 5167-2. Measurement of fluid flow by means of pressure differential devices inserted in
circular cross-section conduits running full. Part 2: Orifice plates. European standard. 2003.
[22] Martnez J. D, Silva L. E. E, Viera A. R, Lesme J. R. Experimental study on biomass
gasification in a double air stage downdraft reactor. Biomass and Bioenergy. 2011: 35: 34653480.
207

[23] Neeft J, Knoef H, Zielke U, Sjstrm K, Hasler, P, Simell P. Guideline for sampling and
anaysis of Tar and particles in Biomass produced gases (Tar protocol). Energy Project ERKGCT1999-2002

208

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

FEASIBILITY STUDY TO REALIZE AN


ANAEROBC DIGESTER FED WITH VEGETABLES
MATRICES IN CENTRAL ITALY
Umberto Desideri a, Francesco Zepparelli a (CA), Livia Arcioni a, Ornella Calderini b,
Francesco Panarab and Matteo Todini
a

University of Perugia, Dept. of Industrial Engineering, Perugia, Italy, zepparelli@tre-eng.com


b
CNR IGV Perugia, Italy, ornella.calderini@igv.cnr.it

Abstract
In the present paper we have analysed the possibility to realize an anaerobic digester in a bio-Energy Park
located in Citt della Pieve, a small town in Central Italy. The use of anaerobic digesters is quite common in
Europe for reducing the environmental impact of manure in a co-digestion procedure with vegetables
materials. In addition, for several areas of Central Italy there is the need to find alternative productions to
improve farmers incomes, as traditional cropping systems are loosing convenience. An interesting
alternative seems to be cultivation of energy crops because of the favourable conditions of the electric
energy market. We are suggesting a low input cropping system to be implemented in areas where low input
food/feed crops are no more profitable.
In particular our case-study is an example based on the use of a forage legume, alfalfa (Medicago sativa L.),
together with other crops, like sorghum, to realize small-size bio-digesters plants.
Alfalfa: is a highly sustainable crop as it is able to fix nitrogen and therefore it does not require this
fertilization with the consequence of avoiding underground water pollution. Moreover alfalfa residual products
are nitrogen rich thus improving soil structure and fertility more than popular graminaceous energy crops
such as corn. Beside, alfalfa mostly does not need irrigation in the typical Central Italy environment, all these
traits make it one of the species with the lowest energy needs for growing.
The aims of this feasibility study are: i) optimization of plant materials feeding the bio-digester, ii) typology of
bio-digester, iii) size of bio-digester in relation with land availability for growing energetic cultures, iv) the
utilization of bio-gas produced by bio-digester plant to produce electric and thermal energy using
cogeneration engines, vi) disposal of waste-water produced according to regional and national laws.
The final aim of this study is to verify the possibility to develop an alternative economical use of marginal
soils in relatively dry areas of Central Italy that would be replicable in other European areas with a similar
climatic situation.

Keywords:
Alfalfa, Anaerobic digestion, Digestate, Silage, Sorghum.

1. Introduction
Anaerobic digestion is an appropriate technique for converting biomass such as ensiled energy
crops into renewable energy. In addition, since the digested residue can be used as a fertilizer, a
cropping system based on energy crops has favourable traits of sustainability.
The interest in using ensiled crops for anaerobic digestion is increasing. In Europe the development
of anaerobic digestion began in the sector of civil sewage treatment plants for the stabilization of
sludge and currently it is estimated that there are more than 1600 operational digesters.
At the moment this technique is considered to be one of the best for the treatment of the wastewater
from agro-industrial complexes with high organic content. As early as 1994 there were about 400
business and consortium biogas units while now there are more than 3500 anaerobic digesters
operating on livestock effluent in all countries of the European Union. The highest number is in
Germany followed by United Kingdom and Italy. There are currently about 450 active plants for the
recovery of biogas from MSW landfills with a high concentration in Great Britain. This type of
209

treatment is being increasingly supplemented in recent years by the treatment of the organic fraction
deriving from the differentiated collection of municipal waste (bio waste), digested with other
organic industrial waste and livestock slurries. In Denmark alone there are now 21 centralized codigestion plants of this type, treating about 1,750,000 tons of livestock slurry and 450,000 tons of
organic industrial waste and bio waste.
In 2010, primary energy production from biogas had an impressive increase of 31,3%. Biogas
produced more than 10.9 M toe in 2010, which is an additional 2.6 M toe in just twelve months, and
energy was primarily channeled into electricity production. The power output from this source
should be as much as 30.3 TWh in 2010, which is 20.9% more than 2009 [1].
The country that developed anaerobic digestion at the highest degree in the last ten years is
Germany, particularly in the livestock sector: the German biogas association reports that the country
had 7,100 methanation plants in 2010 with 2,780 M W of electrical capacity. [1]
In 2010 Italy should become Europes number three biogas producer, with primary energy
production estimated at 478.5 ktoe. This is the result of the policy of incentives adopted by the
national government which, in addition to providing a contribution for the investment, pays a price
for electrical energy from biogas which may reach 0.28 /kWh over a period of 15 years.
Central Italy is characterized by a great percentage of farmland localized in marginal areas. These
kinds of areas are suffering to a greater extent of the general crisis of the primary sector due to
lowering of incomes, abandoning of farms and the scarce appeal to the new generations. Eurostat
reports a reduction of farmers income of 3.3% in Italy. Energy from biomasses has become an
interesting alternative to food/fodder crop production in the last years. The majority of power plants
settled in Italy are based on biogas production in medium-large scale farms with animal husbandry.
Because of the general Italian condition for farming this occurs mostly in Northern Italy. Anyway
due to constant loss of income for traditional crops, farmers from marginal areas in central Italy are
seeking an alternative to improve the profitability of their land. The present study analyse the
possibility to suggest a model for biogas production in typical medium-small size farm in marginal
areas where animal husbandry is not as common as it used to be. We are suggesting the use of two
low/modest input crops with a special emphasis on alfalfa. Calculations are based on literature data
but experimental analysis is in progress to test crop yield and biogas profitability in our conditions.
The present work takes into account the Regolamento regionale 4 maggio 2011, n. 4 of Regione
Umbria, concerning the management of facilities for the treatment of livestock manure and biomass
for biogas production and utilization [2]. The new regional framework, in twenty articles,
establishes: 1) the requirements for operation and management of anaerobic digestion plants and
company, 2) inter-treating livestock manure and/or biomass to produce electricity and heat from
biogas power up to 1 M W, 3) the terms of the agronomic use of the resulting digestate from
anaerobic digestion.
In particular, the regulation sets that the materials to be treated must not come from more than thirty
kilometres from the plant and that the same distance should be respected for the transport of
digestate from the plant to the land of the company (art. 9). With articles 10 and 11 regulates the
management and agronomic use of digestate and possible further treatment.

2. The anaerobic digestion process


Anaerobic digestion is an appropriate technique for converting biomass such as ensiled energy
crops into renewable energy. The system proves its sustainability also because the digested residue
can be used as a fertilizer.
M ethane can be produced from biomass by either thermal gasification or biological gasification:
biological gasification is commonly referred to as anaerobic digestion. A consortium of several
different anaerobic bacteria carries out the process using a wide range of temperatures from 10C to
210

over 100C and at a variety of moisture contents from around 50% to more than 99%.. Bacteria
living optimally at temperatures between 35 40 C are called mesophiles, those surviving warmer
and more hostile conditions at 5560 C are called thermophiles.
In the absence of oxygen, anaerobic bacteria ferment biodegradable matter into methane and carbon
dioxide, a mixture called biogas. Biogas contains 6070% methane and 3040% carbon dioxide
depending on the feedstock type [3]. Trace amounts of hydrogen sulfide, ammonia, hydrogen,
nitrogen, carbon monoxide, oxygen and siloxanes are occasionally present in the biogas. Usually,
the mixed gas is saturated with water vapor.
Anaerobic digestion takes place in basically three stages. In the first stage, complex organic
macromolecules are hydrolyzed into simpler soluble molecules. In the second stage these molecules
are converted by acid forming bacteria to simple organic acids, carbon dioxide and hydrogen; the
principal acids produced are acetic acid, propionic acid, butyric acid and ethanol. In the third stage,
methanogen bacteria form methane, either by breaking down the acids to methane and carbon
dioxide, or by reducing carbon dioxide with hydrogen [4].
The biogas produced in an anaerobic digestion energy plant consists of 5580% CH 4, 2045% CO2,
0.01.0% H2S, and 0.00.05% NH3, and it is saturated with water [5].

2.1. Digester Technology in Europe


In Europe digesters are mainly made of concrete with a steel skeleton or just steel. Their sizes vary
between 500 and 3,000 m, although there are still smaller units for small users. The digesters have
usually a cylindrical form standing upright in most cases. Not only because of the climatic
conditions in Europe but also in order to control temperature conditions inside the digester tanks are
equipped with an insulation and a heating system. Digesters are also equipped with a system to
agitate or to stir the digesting slurry. There are many systems available to stir the system: some with
slow moving propellers stirring for longer periods or such with fast turning propellers switched on
only for short periods. The biogas is collected either in an external plastic bag or in the space above
the slurry covered with a plastic membrane [6].
The digesters are flow through systems, which are fed several times per day. In the case of
agricultural biogas plants the slurry comes directly from the stables or is collected in small storages
before entering the digester. There is often a premixing pit where other feedstock can be added to
the slurry. Sometimes the bulk feedstock can be added directly to the digester through an extra input
system. The outlet works in parallel to the inlet. The digested slurry is often pumped to a post
digester and/or to a storage tank. These storage tanks must have the capacity to store the slurry for
several, often six to nine, months [6].
The average retention time in the main digester is usually approximately 28 days, but it can be
easily demonstrated that, especially if crops and crop residues were added, biogas production can be
detected still after 90 days. Therefore many biogas plants work with a post digester and/or the slurry
storage tank is also covered with a foil, which works as gas storage. During post-digesting process
and storage, approximately 30% of the total biogas evaluation is captured [6].
In addition to the described technology of wet anaerobic digestion there is a growing interest in dry
anaerobic digestion [7]. The wet technology works with slurries of less than 12% dry matter content
whereas the dry process can handle dry matter contents of 30% and more which would enable the
user to use mainly crops and crop residues as feedstock. In the past dry anaerobic digestion was
limited to waste processing biogas plants. Dry continuous-flow systems are very expensive and the
income from waste disposal fees was necessary for an economic business.

211

2.2. Crops characteristics


Different crops are being used for bioenergy production; in our system we want to stress the
presence of low input crops to be able to enhance production of energy from biomasses in
agricultural areas where cash crops are not profitable anymore. This will avoid competition with
food/fodder cultivation and at the same time it will sustain farmer's income. We are focusing on
alfalfa and sorghum, two commonly cultivated crops in Italy.
Alfalfa is a polyannual (2-5 years) plant species used as forage crop. In the last year work from
different laboratories suggested its alternative use as a source of biomass for biofuel production [5].
On average its dry matter yield is 8-10 tons/ha, which is rather competitive with other crops more
widely used for bioenergy such as giant cane (Arundo donax) or M ischantus. In fact the input
needed by an alfalfa field is very limited in terms of irrigation and fertilization, in particular, due to
its ability to fix nitrogen, it has a widely acknowledged ability to improve the organic matter
content and structure of the soil. Therefore alfalfa can be suggested as a highly sustainable
alternative to commonly used grasses for bioenergy production, in particular in Italy where it is a
well-adapted crop more than the mentioned grasses.
The initial establishment of the field that depends on the soil preparation and the availability of deep
ground water, enhance alfalfa productivity. Anyway according to literature, a satisfactory
productivity can be obtained also in low rainfall regimes. It has to be considered that maximum
productivity generally occurs in the second and third year of establishment. An average dry matter
yield that is considered in our system for this period is 8-10 tons/ha. To provide a continuous supply
to the biodigestor we assume the storage of plant biomass as silage. Data from literature suggest
that alfalfa silage provides a dry matter yield of 11.5 t/ha [9], with 69.5% moisture, 5.7g protein, 1g
fat, 8.8g fiber, 2.4g ash.
The study conducted by Heiermann et al. [3, 10], investigates the suitability of various field crops
for anaerobic digestion included alfalfa, thanks to laboratory scale batch anaerobic digestion tests
under mesophilic condition.
Production on a continuous basis and an almost homogeneous feedstock is indispensable to enable
an uninterrupted supply of crops for anaerobic digestion. Focusing on biogas production ensiling is
the favorable and common method of whole crop preservation.
Legumes such as alfalfa have been ensiled but ensiling has relatively recently become a common
means of conservation (Albrecht and Beauchemin, 2003).
Considering that chemical composition and structure of crops change during their growth, harvest
time also plays a major role with regard to silage quality and maximum yield per hectare.
Fiber and sugar sorghum has been suggested as well suited for energy production. Sorghum itself is
another low requirement crop and it is particularly interesting for its high resistance to drought and
parasites. It is characterized by a high level of rusticity, growing well in different types of soil with
a vegetative cycle of 95 - 120 days. Its dry matter yield on average is around 12-18 t/ha. Sorghum
silage DM yields is 18 t/ha [2].
Energy outcome of the two crops is considered to be on average 530 l/kg biogas/organic DM for
alfalfa silage ([5, 6], P. Weiland pers. comm) and 610 l/kg biogas/organic DM for sorghum silage
(P. Weiland pers. comm).
We have considered cultural cost for the two crops as in the following table (sorghum data are from
Contagraf, University of Padova, 2010, alfalfa data referred to a 4 years cycle, are from Rinaldi,
2005).

212

Table 1. Sorghum and alfalfa cultural costs


sorghum, /ha
Seed bed preparation
230.00 250.00
Seed and sowing
160.00 170.00
Fertilization
130.00 150.00
Weed control
50.00 60.00
Harvest/ensiling
350.00 400.00
Total
930.00 1,030.00

alfalfa, /ha
90.00 100.00
50.00 60.00
50.00 60.00
60.00 70.00
250.00 300.00
500.00 600.00

As equipment for harvesting of sorghum, a combined forage harvester (mowing, chopping, loading)
was accounted, for alfalfa swathing, raking, chopping of windrows and loading operations were
considered.
We have used a combination of data from different literature sources to proceed with the
calculations reported in the following parts of the paper.

4. Feasibility study
The chain for producing methane through anaerobic digestion from energy crops is presented in
Fig. 1, from the production and harvest of crop biomass, to storage and pre-treatment of the
biomass, production and utilization of biogas, storage, post-methanation and post-treatment of the
digestate, and finally returning the digestate back to the crop production areas as fertilizer and soilimprovement medium [12].

Fig. 1. Biogas production chain


Energy crops and crop residues can be digested either alone or in co-digestion with other materials,
employing either wet or dry processes. In the agricultural sector one possible solution to processing
crop biomass is co-digestion together with animal manures, the largest agricultural waste stream. In
addition to the production of renewable energy, controlled anaerobic digestion of animal manures
reduces emissions of greenhouse gases, nitrogen and odors from manure management, and
intensifies the recycling of nutrients within agriculture. Animal manures typically have low solids
content (<10% TS), and thus, the anaerobic digestion technology applied in manure processing is
mostly based on wet processes, mainly on the use of continuously stirred tank reactors (CSTRs).
In co-digestion of vegetable matrices and manures, manures provide buffering capacity and a wide
range of nutrients, while the addition of plant material with high carbon content balances the carbon
to nitrogen (C/N) ratio of the feedstock, thereby decreasing the risk of ammonia inhibition. The
positive synergy effects often observed in codigestion, due to the balancing of several parameters in
the co-substrate mixture, have offered potential for higher methane yields.
213

The most important parameter in choosing crops for methane production is the net energy yield per
hectare, which is defined mainly by biomass yield and convertibility of the biomass to methane, as
well as cultivation inputs. Energy crops should be easy to cultivate, harvest and store, tolerant to
weeds, pests, diseases, drought and frost, have good winter hardiness and be able to grow on soil of
poor quality with low nutrient input.
Due to the problems related to the animal manure management, the present feasibility study
concerns a co-digestion of vegetable matrices, in particular alfalfa and sorghum.
The co-digestion of two vegetable matrices may create problems in regard to the activation of the
anaerobic digestion process. Because of this it may be necessary to make an inoculum of slurry to
help the activation process.

Table 2. Methane and gross energy potentials of energy crops and crops residues
Substrate
Methane potential
Gross energy
potential
Forage beet

Alfalfa

Potato
Maize
Wheat
Barley
Rape
Grass

Clover

Marrow
kale
Jerusalem
artichoke
Sugar beet
tops
Straw

(m3 CH4 ha-1 a-1 )

Ref.

(m3CH4 kg-1 VSadded )

(m3CH4 kg-1TSadded )

(m3CH4 t-1 ww)

0.46
0.36
0.41
0.32
0.28
0.41
0.39
0.36
0.34
0.41
0.27
0.27-0.35
0.35
0.14-0.21
0.26
0.32
0.27

n.r.
0.32 c
n.r.
0.28 c
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
0.24 c
0.25-0.32
n.r.
0.12-0.19
n.r.
0.28 c
0.24 c

n.r.
55 c
n.r.
56 c
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
46 c
64-83
n.r.
24-36
n.r.
42 c
49 c

5800 a
3240 b
3965 a
2304 b
2280 a
5780 a
2960 a
2030 a
1190 a
4060 a
1908 b
n.r.
2530 a
n.r.
1680 a
2304 b
2862 b

56 ac
34 b
38 ac
24 b
22 ac
56 ac
28 ac
20 ac
12 ac
39 ac
20 b
n.r.
25 ac
n.r.
16 ac
24 b
30 b

1
2
1
2
1
1
1
1
1
1
2
3
1
3
1
2
2

0.23
0.36-0.38
0.25-0.26
0.30 c

0.19 c
0.29-0.31 c
0.23-0.24
0.25 c

n.r.
36-38 c
139-145
n.r.

n.r.
n.r.
n.r.
n.r.

n.r.
n.r.
n.r.
n.r.

4
5
3
6

in Germany, b in Sweden, c Values calculated from the data report ed, a = year, n.r. = not reported. 1: Weiland 2003, 2:
Brolin et al. 1988, 3: Kaparaju et al. 2002, 4: Gunaseelan 2004, 5: Zubr 1986, 6: Badger et al. 1979.

This study assesses the sizing of the anaerobic digestion dimension plant, the quantification of the
necessary agricultural area for the crops cultivation and for the agronomic utilization of the digested
produced by the anaerobic digestion process.

214

We have considered a size of the CHP engine of 250 kW, which is a power that fits the needs of a
typical medium-small size farm. In addition the following technical parameters are assumed:
M esophilic temperature range;
Hydraulic retention time (HRT) equal to 28 days;
Co-digestion of alfalfa and sorghum.
Following these assumptions, we have determined the amount of alfalfa and sorghum necessary to
feed the 250 kW CHP engine. Considering a lower heating value of the biogas equal to 6.8 kWh/m3,
an efficiency of the engine equal to 36% and 7,500 operating hours for year, we obtain an amount
of biogas equal to 766,544 m3 per year.
In order to evaluate the amount of alfalfa and sorghum necessary for the production of biogas
amount stated above, we have considered the parameters reported in table 3 (regarding alfalfa) and
table 4 (regarding sorghum).
Table 3. Alfalfa silage parameters
Alfalfa silage
Dry matter, DM
40%
Organic dry matter, ODM
85%
Biogas yield
0,55 m3 biogas/kg ODM

Table 4. Sorghum silage parameters


Sorghum silage
Dry matter, DM
30%
Organic dry matter, ODM
85%
Biogas yield
0,60 m3 biogas/kg ODM
We have also assumed that anaerobic digestion of alfalfa produces 60% of the biogas, while 40% is
produced by sorghum digestion.
The volume of the digester is calculated based on the following equation, given the amounts of the
vegetable matrices and the hydraulic retention time:
V = HRT * Q
The necessary volume of the digester resulted of 407 m3.
Given the vegetable yield/ha of the vegetable matrices it is possible to evaluate the amount of areas
necessary to produce enough supply for the digester. Assuming alfalfa silage yield equal to 10.92
t/ha*year, and sorghum silage yield equal to 16.91 t/ha*year, and we have obtained that 225.23 ha
are required for the production of alfalfa silage, and 118.51 for production of sorghum to feed the
digester.

4.1. Digestate management


Digestate is a solid material remaining after the anaerobic digestion of a biodegradable feedstock
and it is produced both by acidogenesis and methanogenesis with different characteristics.
Digestate is an easy product to handle and to apply and it can be used successfully as a substitute of
mineral fertilizers. The fertilizer value of digestate depends on the nutrients present in the feedstock.
However, digestate is the result of a living process and therefore has characteristics that are specific
to each digester tank.
In the present study we have assumed that the digestate produced by the digester is stored in a
specific waterproof lagoon, that as to be dimensioned to contain the amount of digestate produced
in 150 days. The storage device of the digestate also envisages a frank minimum safety of at least
215

fifty centimeters. The design must include all the necessary measures to minimize odorous
emissions [2].
In fact, the application of digestate at times of the year when there is little plant uptake, for instance
autumn and winter, can result in nutrient leaching and runoff into ground and surface waters (e.g. of
N and P). Digestate must therefore be stored until the correct time of application.
Digestate applications should be matched with crop nutrient requirements; this will minimize any
unintended negative impact to the environment and also maximize farmers profits. Table 5 reports
application rates (especially for nitrogen), length of storage periods, and timings for applications
that must also comply with national limits.
Table 5. National limits regulating nitrogen loading on farmland, required storage capacity and its
spreading season
Maximum nutrient load
Required storage
Compulsory season for
capacity
spreading
Austria
170 kg N/ha/year
6 months
28 feb 5 oct
Denmark
170 kg N/ha/year (cattle)
9 months
1 feb harvest
140 kg N/ha/year (pig)
Italy
170 - 340 kg N/ha/year
150 days
depends on the weather
conditions
Sweden
170 kg N/ha/year
6 10 months
1 feb 1 dec
Northern Ireland
170 kg N/ha/year
4 months
1 feb 14 oct
Germany
170 kg N/ha/year
6 months
1 feb 31 oct arable land
1 feb 14 nov grassland

5. Conclusions
This study has assessed the opportunity to realize an anaerobic digester in a bio-Energy Park
located in Citt della Pieve, a small town in Central Italy.
We have chose to evaluate the anaerobic co-digestion of vegetable matrices such as alfalfa and
sorghum silage and we have evaluated the principal process parameters.
In particular, assuming a 250 kW CHP engine, we have determined the biogas necessary, the
volume of the digester, the amount of alfalfa silage and sorghum, and also the areas required to their
production.
The area requested seems to be rather relevant, depending on the lower productivity of the crops
chosen as alternative to the largely used energy crops for biogas production, such as maize. The
economic feasibility is being currently investigated also taking into account smaller digester size
and that marginal lands should be used. Further analysis including other low-input energy crops as
suggested by latest literature in the field will be carried out.
Finally we have reported some important consideration about the digestate management, that could
be a great opportunity but also a critical point of the system.

Nomenclature
CSTRs
HRT
ODM
TS
TVFA
VFA

Continuously Stirred Tank Reactors


Hydraulic Retention Time
Organic Dry M atter
Total Solids
Total Volatile Fatty Acids
Volatile Fatty Acids
216

VS
ww

Volatile Solids
wet weight

References
[1] Eurobserver, The state of renewable energies in Europe. 2011 edition.
[2] Regolamento regionale 4 maggio 2011, n. 4 Norme di attuazione dell'articolo 4, comma 1,
lettera e) della legge regionale 10 dicembre 2009, n. 25 concernente la gestione degli impianti
per il trattamento degli effluenti di allevamento e delle biomasse per la produzione di biogas e
l'utilizzazione agronomica delle frazioni palabili e non palabili Bollettino Ufficiale della
Regione Umbria n. 21 del 11 M aggio 2011.
[3] M . Plchl, M . Heiermann, B. Linke, H. Schelle. Biogas Crops Part II: Balance of Greenhouse
Gas Emissions and Energy from Using Field Crops for Anaerobic Digestion. A gricultural
Engineering International: the CIGR Ejournal 2009. M anuscript number 1086. Vol. XI. M onth.
[4] R. M . Jingura, R. M atengaifa, Optimization of biogas production by anaerobic digestion for
sustainable energy development in Zimbabwe, Renewable and Sustainable Energy Reviews 13
(2009) 11161120.
[5] L. Xiaohui, N. Sugiura, F. Chuanping, M. Takaaki, Pretreatment of anaerobic digestion effluent
with ammonia stripping and biogas purification, Bioresource Technology 100 (2009) 5478
5484.
[6] Plch M ., Heiermann M . Biogas Farming in Central and Northern Europe: A Strategy for
Developing Countries?. A gricultural Engineering International: the CIGR Ejournal. Invited
Overview No. 8. Vol. VIII. March, 2006.
[7] Kttner M ., Dry fermentation a new method for biological treatment in ecological sanitation
systems (ECOSAN) for biogas and fertilizer production from stackable biomass suitable for
semiarid climates. In Proc. 3rd International Conference on Environmental M anagement, pp 16.
Johannesburg, South Africa 2002.
[8] Chen F., Dixon RA Lignin modification improves fermentable sugar yields for biofuel
production Nat Biotechnol 2007. 25(7):759-61.
[9] Borreani G, Tabacco E. Insilamento dellerba medica per valorizzare le proteine. Terra e vita
2008, 14: 80-82.
[10] M . Heiermann, M . Plchl, B. Linke, H. Schelle, C. Herrmann. Biogas Crops Part I:
Specifications and Suitability of Field Crops for Anaerobic Digestion. Agricultural
Engineering International: the CIGR Ejournal. M anuscript 1087 2009. Vol. XI. June.
[11] Candolo G. Energia dalle biomasse vegetali: le opportunit per le aziende agricole
Agronomica 2006, 4:26-35.
[12] A. Nordberg, A. Jarvis, B. Stenberg, B. M athisen, B. H. Svensson. Anaerobic digestion of
alfalfa silage with recirculation of process liquid. 2006.

217

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Investigations on the Use of Biogas for Small


Scale Decentralized CHP Applications with a
Focus on Stability and Emissions
Steven MacLeana , Eren Talia , Anne Giesea , Jrg Leichera
a

Gaswrme-Institut e.V, Essen, Germany, maclean@gwi-essen.de

Abstract:
For both environment al and economic reasons, the use of biogas for heat and power generation (CHP),
especially on a small and decentralized scale, is predicted to increase dramatically in the years to come.
However, these unconventional fuel gases present new challenges to manufacturers of combustion systems
as their properties differ from natural gas. In general, their calorific values (LCV) are much lower than those
of natural gas, as they contain large amounts of inert species such as CO 2 or N2. Also, their chemical
compositions may vary significantly over time and they may contain species such as HCN or NH3, leading to
increased NOx formation during combustion due to fuel-bound nitrogen. While NOx formation due to fuelbound nitrogen is common in coal combustion, NOx reduction measures for the combustion of gaseous fuels
are usually aimed towards the reduction of thermal NOx formation and are thus not able to prevent the
conversion of chemically bound nitrogen in biogas into nit rogen oxides.
In the course of several res earch projects, Gaswrme-Institut e.V. Essen (GWI) investigated on how to best
make use of these renewable fuels in future combustion systems. Using both numerical and experimental
techniques, several burner systems were developed which can achieve a stable combustion of different
types of biogases with a minimum of NOx formation. Using CFD simulations, burners based on the COS TAIR

and flameless oxidation (FLOX ) principles were modified to operate with low calorific value fuel gases. The
performance of these burners was then further investigated by experiment al investigations in GWIs semiindustrial test rigs where a satisfactory agreement between numerical and experimental data was observed.
In a further step, the COSTA IR burner was then mounted into a commercially available 100 kW micro gas
turbine (MGT) and tested under real operating conditions. It was shown that the combustion system was
able t o operate in a stable manner while producing only a minimum of NOx -emissions, making the
combination of a MGT and a burner system optimized for low calorific value gases an ideal choice for small
scale decentralized combined heat and power applications.

Keywords:
Low Calorific Gases, Alternative Fuels, Simulation, Fuel-Bound Nitrogen, COSTA IR, Flameless Oxidation,
Product Gases formed from Biomass, NOx -Emissions

1. Introduction
Product gases formed in biomass gasification plants usually have a fluctuating lower heating value
and contain several different chemical species, depending on the type of process and biomass being
used. The gasified biomass produces a low calorific value gas that can be used in a similar manner
as natural gas. Generally some of the more common chemical species of low calorific value product
gas are e.g. methane (CH4 ), carbon monoxide (CO), hydrogen (H2 ), carbon dioxide (CO2 ), nitrogen
(N2 ) and traces of benzol, toluol and tars. Low calorific product gases are defined as gases with
lower heating values (LCV) less then < 3 kWh/mN3 . The large percentage of N 2 and CO2 and the
low CH4 contents of many (LCV) product gases are the reason why the lower heating value
decreases significantly compared to natural gas. However the term low calorific value gas is not
solely restricted to product gases formed in gasification processes, but in a broader sense sums up
various types of product gases produced in steel mills, chemical processing facilities or landfills. An

218

overview of typical chemical compositions and properties of some common low calorific gases is
listed in table 1.
Table 1. Compositions and Properties of various Low Calorific Value Gas

At present, the majority of biogas is produced in biogas fermentation plants. In the


future,gasification of biomass may pose an alternative to increase the generation of low calorific
product biogas and pave the way for sustainable use of biomass in numerous industrial heat and
power applications.
Burning low calorific product gases with conventional burners is difficult because of the large
amounts of inert species (mainly N 2 and/or CO2 ), tar particles and low LHVs of these gases.
However, another challenge with burning these fuel gases is fuel-bound nitrogen, an issue that is
normally not a problem with gaseous fuels. Fuel-bound nitrogen in fuel gases leads to a
considerable increase of NO x -emissions during combustion. Nitrous oxides are toxic greenhouse
gases which cause nitric acid to form in the atmosphere and the deplete Earths ozone layer.
Furthermore, NO x -emissions have a 310 times more environmentally damaging impact than CO 2
with regards to global warming. Therefore, advanced burner systems specifically designed to
prevent the formation of NO x during combustion are needed. Fuel-bound nitrogen in gases normally
consists up to 95 % out of ammonia compounds, for example when saw dust, chicken manure and
various types of biomass are gasified at 1000C.
The intermediate species formed from the fuel-bound nitrogen tend to convert to NO molecules
when a surplus of oxygen due to over-stochiometric conditions is available in the combustion zone .
A simplified diagram such a NO-formation mechanism is shown in Fig. 1. Several research
activities have focused on exploiting this NO-mechanism to develop a low NO X combustion system
for low calorific biogases containing fuel-bound nitrogen [1].

Fig. 1. Simplified NO-Mechanism of Fuel-Bound Nitrogen


219

In a recent research project conducted by the Gaswrme-Institut e. V. Essen, two burners based on
the different concepts of FLameless OXidation (FLOXT M) and COntinuously STaged Air with
Internal Recirculation (COSTAIR) were subject to extensive numerical and experimental
investigations. The main goal of the project was to develop a method to reduce the NO x -emission
levels during combustion of biogases containing fuel-bound nitrogen. In past research projects, both
burner concepts have proven to be low NO x -emission technologies capable of handling fluctuating
fuel gas qualities and low calorific values [2]. The prototype burners were developed in three steps.
As a first step, the unmodified burners were simulated with CFD FLUENT with regards to their
predicted NO x-emission levels while combusting low calorific product gas containing fuel-bound
nitrogen. In the next step of the project several geometric modifications and improvements to the
burners were simulated, again applying the same gases and simulation parameters. Finally, both
burners were built and experimentally tested at GWI. The aim of this paper is to give readers a
greater insight into the numerical and experimental results as well as the approaches taken to
develop a low NO x combustion system for low calorific gases containing fuel-bound nitrogen.

2. Numerical Development

Fig. 2. COSTAIR Burner Concept

The COSTAIR burner concept applies continuously staged air with internal recirculation of flue
gases to achieve a stable combustion with low NO X- and CO- emission levels. Air is injected
perpendicularly to the main flow by means over a large number of small nozzles on the surface of
the air distributor, thus continuously varying the local air ratio. The internal recirculation of the flue
gas helps to produce a stable, uniform combustion that minimizes peak flame temperatures and
pockets of oxygen in the combustion reaction zone [3]. The fuel gas nozzles are equally spaced
around the central air distributor to ensure that the combustion products mix appropriately with the
continuously staged air. In Fig. 2 a simplified sketch of the COSTAIR burner is given.
Throughout the course of this project, CFD simulations played an important role in the development
and optimization of COSTAIR burner s. The main idea behind reducing the high NO x -emission
levels of product gases with fuel-bound nitrogen was to stage the air in such a manner that the
combustion reaction zone is globally split into a sub-stochiometric ( < 1) reduction zone followed
by over-stochiometric ( < 1) burnout region. The gases simulated in the numerical models
consisted of the chemical species CO, H2 , CH4 , CO2 and N2 . By injecting small amounts of
ammonia gas (NH3 ) into the product gas,the presence of fuel-bound nitrogen was emulated. The
simulation results show that adding small amounts of NH3 to the fuel gas, caused the NO x emission
levels to increase drastically during combustion since the fuel NOx formation pathway yields much
220

more nitric oxides than the thermal NOx production pathway which is usually dominant in the
combustion of gaseous fuels.. During the development of the COSTAIR burner a strong emphasis
was focused on the design of the air distributor. The initial CFD results showed that the positioning
and size of the air jets on the air distributor had a tremendous impact on the stability of the
combustion zone and and the increase of NO x emissions levels. The results also confirm that
sectioning the air jets on the air distributor causes the reaction zone to shift within the combustion
chamber. If for instance the reaction zone is shifted behind the air distributor a considerable rise in
NOx -emission levels is calculated. After determining the basic design of the modified COSTAIR
burner, further optimization of the burner and air distributor geometries was required in order to
effectively burn low calorific value gas containing fuel-bound nitrogen. A promising approach
taken was to extend the length of the air distributor in order to divide the reaction zone more clearly
into a sub-stochiometric and over-stochiometric reaction zone. Another important aspect during the
development of the prototype burners was the size and positioning of the fuel gas nozzles.
Therefore the optimum distance and angle of the fuel jets to the air distributor were numerically
determined as well. In Fig. 3 the simulated velocity and temperature distributions of both
COSTAIR burner variants are shown. The temperature and velocity distributions confirm that a
stable combustion is possible with both COSTAIR variants. By extending the length of the air
distributor the combustion reaction zone is clearly divided into two regions with different air ratios.
Most importantly, the CFD results confirm that a reduction of NO x emissions from fuel-bound
nitrogen can be achieved with the optimized air distributor.

Fig. 3. Temperature and Velocity Distributions of the COSTAIR Burner Variants


FLOXT M burners are low NO x -emission combustion systems used in various clean energy
processes. The term FLOXT M implies that during combustion the flame appears invisible. This is
achieved by injecting both fuel gas and oxidizer at high velocities into the combustion reaction
zone. The high momentum injection causes the flue gases to be entrained into the reaction zone,
thus reducing peak flame temperatures and hence NOx emissions [4]. The internal recirculation of
the flue gases also dilutes and stabilizes the combustion reaction zone. In Fig. 4 a schematic
drawing of the FLOX concept is presented. The FLOX burner investigated in this project was
originally designed to burn natural gas with pre heated air. Since the LCVs of product gases are
considerably lower than that of natural gas, a larger gas flow is required to deliver an equal amount
of thermal energy. Therefore, several modifications to the air and fuel nozzles were made in order
to burn low calorific product gas with an optimized FLOX burner variant. Prior to the
221

experimental tests performed at GWIs testing facilities, the same CFD models and boundary
conditions used to investigate the COSTAIR burner were applied to adapt a FLOX burner system
to low calorific product gas. The cross sections of the fuel and air nozzles were increased in order to
generate the required gas velocities for flameless oxidation.

Fig 4. Flameless Oxidation Burner Concept

Further evaluation of the results showed that placing the secondary air inlets behind the combustion
reaction zone leads to a considerable decrease in NO x -emission levels. The temperature and velocity
distributions shown in Fig. 5., illustrates that the injection of secondary air does not influence the
recirculation of flue gases in the combustion reaction zone. The maximum peak flame temperature
of the FLOX burner is 250 C lower compared to the COSTAIR burner. The largest reduction of
NOx -emissions was achieved when the primary and secondary air inlets were spaced 1000 mm
apart. Injecting the secondary air behind the main combustion reaction zone causes the maximum
flame temperatures to decrease while simultaneously the primary reaction zone is divided into two
different reaction regions with different air ratios, thus leading to a decline of NO x - and CO
emission during the combustion of low calorific product gas containing gaseous ammonia.

Fig. 5. Temperature and Velocity Distributions of the FLOX Burner Variants

222

3. Experimental Investigations

Fig. 6. Layout of the Test Rig (bottom left COSTAIR Burner, bottom right FLOX Burner)
A schematic drawing of the test rig along with images of the FLOX and COSTAIR burners are
presented in Fig. 6. The experimental investigations of the COSTAIR and FLOX prototype
burners were conducted in two testing stages. During the first phase the original COSTAIR and
FLOX burners were tested without modifying the original burner geometries and secondary air
staging inlets. During the second stage of testing the air staging concepts developed for the both
burner prototypes were investigated and evaluated. The testing conditions for both burners were
kept constant throughout the course of the experimental investigations in order to directly compare
the testing results. The low calorific gases used in the experimental tests at GWI were synthetically
produced with a gas mixing station, while a gas pre-heater warmed up the gas mixtures to 350C in
order to simulate the actual operating temperatures of a variety of product gases formed in landfills
or gasification processes. One of the fuel gases used was seeded with small amounts of ammonia
gas (NH3 ) in order to investigate the impact of fuel-bound nitrogen on NO x-emissions. The
ammonia gas was added to the fuel gas in increments of 1000, 3000 and 5000 ppm. The
composition of the flue gas was measured with an exhaust probe to determine concentrations of CO,
CO2 , O2 , NO and NO2 . Moreover, the pressures of the fuel gas and air along with the temperature of
the flue gas were recorded.
223

3.1 COSTAIR Burner Tests


After the final simulations of the COSTAIR variants were completed, experimental testing of both
COSTAIR burner variants began. In Fig. 7 the bar graphs of the recorded NO x -emission levels of
the unmodified and optimized burner designs prove that air staging is an effective method to
decrease NOx -emissions. The graph also illustrates that adding 1000, 3000 and 5000 ppm of
ammonia to the product gas causes the NO X-emission levels to almost double in value when air
staging is not applied, as in the case of the unmodified COSTAIR burner variant.

Fig. 7. Experimental Results of Both COSTAIR Burner Variants (NOx-Emissions)


The results in table 2 show that nearly a 90 % reduction of NOx-emission levels was possible with
the optimized COSTAIR air distributor.
Table 2. Overview of NOx Emissions for Both COSTAIR Burner Variants
Unmodified
Optimized
Air Ratio

NH3 in Fuel Gas

COSTAIR

COSTAIR

[ppm]

NOx [ppm]

NOx [ppm]

@ 3 Vol.-% O2

@ 3 Vol.-% O2

NOx
Reduction
[%]

1.2

9.77

7.51

23.13

1.2

1000

384.45

37.45

90.26

1.2

3000

735.44

78.93

89.27

1.2

5000

1275.33

151.42

88.13

224

The data collected during the first set of burner tests served as points of reference in order to
determine the total reduction of NOx -emissions and the effectiveness of the optimized air distributor
on NOx -emissions during the second stage of testing. A tremendous decline in NO x -emissions was
demonstrated by imposing a strong gradient of the local air ratio along the distributor length of the
COSTAIR burner while burning product gases containing fuel-bound nitrogen.

3.2 FLOX Burner Tests


The results of the experimental tests reveal that the NO x -emissions from fuel-bound nitrogen are
slashed by nearly 50% with the optimized FLOX burner. The bar graph of the testing results in
Fig. 8. confirm that in general, the NO x-emissions produced by the FLOX system are remarkably
low.

Fig. 8. Experimental Results of both FLOX Burner Variants (NOx-Emissions)


An overview of the NO x reduction of the FLOX combustion system is provided in table 3.
Table 3. Overview of NOx Emissions for Both FLOX Burner Variants
Air Ratio
NH3 in fuel gas
Unmodified
optimized FLOX
[ppm]
FLOX
NOx [ppm]
NOx [ppm]
@ 3 Vol.-% O 2
@ 3 Vol.-% O 2
1.2
1.2
1.2
1.2

0
1000
3000
5000

6.91
56.85
118.66
193.74

6.38
36.18
64.62
96.88

NOx
Reduction
[%]
7.67
36.36
45.54
50.01

The measurements of the basic FLOX burner variant reveal that less NOx-emissions are formed
when compared to the basic COSTAIR burner. However the use of optimized air distributor had
225

greater impact on reducing NO x -emissions compared to the secondary air used with the FLOX
system. A direct comparison of the optimized burners generally shows that a reduction of NO x emissions is achieved equally well with both combustion concepts. The results confirm that nearly a
50% reduction of NO x -emissions from product gases containing fuel-bound nitrogen is possible
with the FLOX burner when the combustion air is staged.

4. CHP Applications for Low Calorific Gases


Micro gas turbines are a beneficial and profitable alternative to utilize low calorific gases in CHP
processes, due to their simple design, low CO and NO x -emissions, operating and maintenance costs
as well as noise emissions and adaptability to changing gas qualities compared to gas-powered
engines. Within the scope of a prior research project [5], GWI in collaboration with several research
and industrial partners optimized and experimentally investigated the burner concepts of flameless
oxidation FLOX and continuously staged air COSTAIR for low calorific gases under micro
gas turbine conditions. Furthermore, the COSTAIR burner was validated on an actual landfill
during continuous operation, using a gas with a CH4 content below 30 vol-%. The first set of results
describing the methodology, burner design, burner optimization as well as experimental burner tests
were already published in [5].
The initial experimental tests under atmospheric conditions at GWI indicate that the optimized
COSTAIR burner operates steadily without causing high emissions. To assure a stable operation
mode during testing, the positioning of the pilot burner and the spark ignition were determined
numerically with the CFD-program FLUENT. These fluid flow investigations were carried out by
the Department of Energy Plant Technology of the Ruhr-University Bochum. The pilot gas nozzle
is aligned in between two of the main gas nozzles as shown in Fig. 9. In Fig. 10 a side view of the
pilot burners exit velocities is shown. The recirculation zone (highlighted red circle) formed above
the gas jet, improves the stability of the flame once the pilot gas ignites. The radially injected air
stream collides with the axially flowing pilot gas jet improving the overall mixture and ignition of
the pilot burner. A front view of the burner in Fig. 11 clearly illustrates the radial recirculation zone
located above the fuel gas nozzle.

Fig. 9. Close Up View of the MGT Fig. 10. Simulation of the Exit Velocities of the
COSTAIR Burner
Pilot Burner Nozzle

Further evaluation of the numerical results concluded that a small area circa 30 mm above the gas
jet forms a low velocity field containing an ignitable gas/air mixture. Corresponding to the
numerical results a new burner flange plate was constructed and installed in the MGT T100 for
226

validation testing. In Fig. 12 the optimized COSTAIR burner is shown (please note the position of
the pilot gas nozzle and spark ignition). In the adjacent images the entire COSTAIR burner
including the head pipe of the MGT T100 is displayed.

Fig. 11. Radial Velocities 30 mm above the Fig. 12. Adjusted Burner Geometry(left) and Head
Gas Nozzle
Pipe Mounted to the Burner (right)

5. Conclusion
Throughout the course of this project, the suitability and optimization of the COSTAIR and FLOX
combustion systems using low calorific product gases containing fuel-bound nitrogen were
investigated. One of the first steps taken in this project was to develop two burner systems for low
calorific product gases. The influence of different fuel gas compositions and properties was
simulated with CFD FLUENT under realistic burner operating conditions. During the second stage
of development, both burner designs were optimized to efficiently burn low calorific value gases
and reduce NO x -emissions of product gas containing fuel bound nitrogen. A variety of NO x reduction methods were considered and simulated in numerous simulation approaches. Finally, both
prototype burners were installed and tested in a GWI test rig. The results show that secondary air
staging is an effective method to reduce NO x- emissions due to fuel-bound nitrogen. Also both
burners responded stably to fluctuations of the fuel gas composition and maintained low emissions
over a wide range of operating conditions.
The outcome of these projects shows that applying secondary air staging to FLOX and COSTAIR
burner systems leads to a considerable decline in NO x -emission levels. Yet a further conclusion is
that small traces of fuel bound nitrogen in product gases causes the NO x -emission levels to increase
immensely. The developments and combustion concepts investigated in this project, may soon pave
the way for new and innovative CHP applications such as micro gas turbines, fuel cells, gas
powered engines to utilize low calorific product gases containing fuel-bound nitrogen in an
economically feasible manner. The developed burner is able to combust low calorific fuel gases
with LHVs as low as 1.25 kWh/Nm3 (corresponding to a CH4 concentration of about 12.5 vol- %)
without releasing high NO x -emissions. Furthermore, the necessity to treat product gases containing
fuel-bound nitrogen can be reduced considerably using both combustion concepts as primary
measures to reduce NO X formation in the combustion space. The use of alternative fuels combined
with innovative burner systems may give industrial plant operators a technical advantage to react in
a more competitive and flexible manner to fuel price fluctuations. On behalf of all the
corresponding authors, we thank the Department of Energy Plant Technology, Ruhr-University of
Bochum, Germany.

227

References
[1] Brink, A., Huppa, M., Kurkela, E., Suomalainen, M.: Nitric oxide yield from combustion of a
low calorific gasification product gas: numerical and experimental results, Progress in
Computational Fluid Dynamics, Vol. 6, No. 4/5, 2006
[2] AiF-Projekt MGT: Neue Brennersysteme zur dezentralen Nutzung von schwachkalorige n
Gasen in Mikro-Gasturbinen. AiF-Vorhaben Nr.: 13216 N, Projektdauer: 01.03.2002 bis
31.08.2004
[3] Al-Halbouni, A.: Entwicklung NO x-emissionsminimierter Heizkesselfeuerungen. Habilitation,
Otto-von-Guericke-Universitt Magdeburg, Shaker Verlag 2001
[4] Maclean, S., Tali, E., Giese, A., Leicher, J., Grner, K.: Investigations on the Reduction of NO x
Emissions of Product Gases containing Fuel Bound Nitrogen in Thermal Processing Plants, 9th
European Conference on Industrial Furnaces and Boilers, Porto, Portugal, 2011
[5] Leicher, J., Giese, A., Grner, K., Scherer, V., Schulzke, T.: Development of a Burner Syste m
for Use of Low Calorific Fuel Gases in Micro Gas Turbines, Proceedings of the Europea n
Combustion Meeting, Cardiff, England 2011 .

228

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Kinetic energy recovery system for sailing yachts


G.L. Guizzia and M. Mannob
a

Universit degli Studi di Roma Tor Vergata, Roma, Italia, guizzi@ing.uniroma2.it


Universit degli Studi di Roma Tor Vergata, Roma, Italia, michele.manno@uniroma2.it

Abstract:
SEAKERS (SEA Kinetic Energy Recovery System) is a research project, funded wit hin t he 7th EU
Framework Programme and officially started on January 1st, 2011, whose goal is to develop an innovative
devic e consisting in a kinetic energy recovery system for sailing yachts based on the conversion of boat
oscillations (heave, pitch and roll) caus ed by the sea into electric energy by means of a linear generator.
Therefore, SEAKERS addresses a well known unsatisfied requirement of yacht owners, since energy is a
resource of primary importance in a boat, especially in a sailing one: it is well known that during a one day
cruise, electricity consumption has to be carefully managed (for instance the refrigerator is switched off), so
as not to be short of energy at night. It often happens that, after one day of sail cruise, it is necessary to
recharge the batteries through the onboard generator, which means keeping it on for hours, producing very
annoying noise, smoke and pollution.
The device that is going to be developed aims at recovering as much kinetic energy as possible from the
natural movements of a sailing yacht on the sea, therefore taking the view of a boat as a moving wave
energy converter with energy harvesting capacity. The boats motions can be vertical oscillations due to the
buoyancy in the presence of sea waves, both when the boat is still or sailing, and rolling and pitching
motions originated both by sailing in wavy waters and by the normal boat dynamics due t o the sails
propulsion. Linear generators will convert kinetic energy into electrical energy to be used as green
electricity for any possible application on board.
Preliminary calculations show that a properly configured system could be able to recover 100-400 W under
most sea conditions, which can be an extremely attractive result since an electric energy availability of 1-2
kWh on a sailing yacht is of signific ant interest.

Keywords:
Wave Energy Recovery, Linear Generator, Sail Yacht.

1. Introduction
This paper presents some preliminary results obtained in the SEAKERS project, whose aim is to
design and test a kinetic energy recovery system to be used on board of sail yachts in order to
recover energy from the wave-induced boats vertical motion.
Such a system is able to recover actual free energy, as opposed to other devices, already
commercially available, that subtracts energy from the propulsion offered by the winds lift on the
sails, as in the case of micro-wind turbines installed on the boat, which are set into motion by the
apparent wind originating from the yachts motion.
In practical terms, the SEAKERS device is intended to be a linear oscillator, with a mass oscillating
vertically inside a prismatic guide and gaining kinetic energy; if the mass is the moving element of
a linear generator, the resulting mechanical energy can be extracted and converted into electricity.
The oscillating mass incorporates permanent magnets which, moving in proximity of stator
windings, generate electric power due to electromagnetic induction.
The SEAKERS device addresses a well known unsatisfied requirement of yacht owners, since
energy is a resource of primary importance in a boat, especially in a sailing one: it is well known
that during a one day cruise, electricity consumption has to be carefully managed (for instance the
refrigerator is switched off), so as not to be short of energy at night. It often happens that, after one

229

day of sail cruise, it is necessary to recharge the batteries through the onboard generator, which
means keeping it on for hours, producing very annoying noise, smoke and pollution.
The idea of a linear generator originates from work carried out at the University of Uppsala [1-4],
where such devices have been designed and tested in order to recover wave energy from a buoy,
oscillating on the sea surface, connected to a rope that makes a piston move inside a generator
placed on the seafloor. In the SEAKERS project, the oscillating mass is set into motion not directly
by the sea waves but by its inertia as the yacht is subject to heave, pitch and roll motions.
In order to design the test-bed for testing the generator, it is necessary to set up a reliable model of
different sea conditions that could be of practical interest for a normal cruise on a sail yacht (thus
there is no need to consider extreme, stormy waves) and of the ship motion due to such sea states.
Furthermore, it is interesting to find out, by means of a very simple mechanical model of the linear
generator, how much power could be extracted under these simplifying assumptions, in order to
decide whether the projects outcome could in principle be commercially viable, and quickly to
provide data against which results from more detailed analytical models and experimental tests
could later be compared.
This paper presents the results obtained in this first stage of the project, detailing first the model of
sea conditions (section 2), then the outcome of simulations on the yachts motion carried out by
means a commercial software (section 3), and finally the results of the analysis carried out on a
linear mechanical system located on the boat (section 4).

2. Wave excitation
2.1.

Wave spectra

The main characteristic of sea waves is randomness. Indeed, by checking even a short a time series,
two characteristics arise: height and period of a wave are different from height and period of
another wave. For this reason, the free surface elevation of sea waves is modelled as a stochastic
process and is assumed to be a random, Gaussian, ergodic process in the time domain [5-8].
Mathematically, sea elevation can be reconstructed in one dimension as a Fourier series as follows:
n

x, t

Z j cos

t kjx

(1)

j 1

In this equation, Zj is the wave amplitude for the j-th wave form of circular frequency j, k j its wave
number (dependent on j through the dispersion relation) and j its phase shift. The dispersion
relation defines the relationship between wave frequency and wave number; in deep water it is
expressed as [5-8]:
2

kg

(2)

where g is the acceleration of gravity. (It may be useful to recall that wave number and wave length
are mutually dependent: k 2 / ).
Given a sea elevation time pattern for a given spatial coordinate x, the amplitudes Zj of its Fourier
series may be evaluated as Fourier transforms of :
1
T

Zj

T
2
T
2

t exp

i2

(3)

t dt

The most meaningful representation from a statistical point of view of a particular sea state is given
in the frequency domain by means of the wave spectrum S
, which is defined as:
S

Z j Z *j

Zj

(4)

230

Z *j being the complex conjugate of Zj. Therefore, the spectrum S j is proportional to the energy
content of the j-th wave form of circular frequency j, while the area under the spectrum s
proportional to the overall energy content of the sea state described by sea elevation t :
n

Zj

(5)

j 1

Since S( ) is an even function, that is, S


) = S( ), and taking also into account that negative
angular frequencies have no physical meaning beyond that of the corresponding positive values, it
is frequently adopted an alternative definition of the energy spectrum (S ), which is defined for
positive angular frequencies only:

(6)

By virtue of equation (5), the energy spectrum is correlated to the overall energy content of the sea
state, because the energy content of a single sinusoidal wave is proportional to the square of its
height. Furthermore, statistical data that can be gleaned from the energy spectrum correspond to
important parameters for the description of a sea state. Of particular importance is the 0-th spectrum
moment, which is equivalent to the area under the wave spectrum curve:

m0

(7)

For a narrow band spectrum, it can be demonstrated that the root mean square (RMS) wave
amplitude is given by m0 , and the RMS value of wave height (crest to trough) is therefore:
H RMS

2 m0

(8)

One of the most useful parameter to represent the sea state is the significant wave height, which is
the mean of highest third wave heights, and for narrow band spectrum it is given by [5-8]:
Hs

4 m0

(9)

Significant wave amplitude is by definition half the corresponding wave height:


0s

2.2.

2 m0

(10)

Simulation assumptions

The simulations that will be presented in the following sections were carried out taking into account
statistical wave data for the Mediterranean Sea, with particular reference to the measurements taken
at Capo Linaro (Civitavecchia, Italy) 1 .
In the case of random waves, it is possible to find a particular set of parameters that make the
JONSWAP spectrum suitable to represent sea conditions in the location of interest (the above
mentioned Capo Linaro near Civitavecchia).
The JONSWAP spectrum was developed from extensive field measurements in the context of the
Joint North Sea Wave Project [5-8]. This formulation is suitable for wind-generated waves in fetch
limited locations. The inputs are the wind speed and the fetch length. The mathematical formulation
is given by equation:

Personal co mmunications with Prof. Felice Arena, University of Reggio Calab ria, 2011.

231

S ( )

5
p

5
4

exp

In the above equation,

exp ln( ) exp


p

p
2

)2

2
p

= 2 /Tp is the peak circular frequency,

(11)

is the Phillips parameter given

by
0.0076( gx / U ) 0. 22 , where x is the fetch length and U the mean wind speed, and is the
peak-shape parameter. For practical applications, can be assumed equal to 0.08 in the whole
frequency domain.
In deep water, wave period and length are correlated by the dispersion relation (2), which may be
rewritten as follows:
gT 2 2

(12)

Wave velocity is given by:


c
/ T gT 2

(13)

Furthermore, wave period is also related to the significant wave height through parameters
Tp

Hs g

and :

(14)

Thus, higher waves are longer (12), propagate faster (13) and are less frequent (14). The JONSWAP
spectrum is completely defined when the significant wave height Hs and parameters and are
specified.
In order to represent correctly sea conditions at Capo Linaro, values of , , f and Tp are chosen
according to the following table. The corresponding wave spectra are illustrated in fig. 1.
Table 1. Parameters used to represent random sea waves at Capo Linaro near Civitavecchia, Italy.
Hs [m]
Tp [s]
f ,
0.5
0.016
1.0
13.2
2.98
1.0
0.008
2.0
14.9
4.75
1.5
0.010
0.5
15.5
6.06
2.0
0.008
0.5
16.4
7.40

Fig. 1. Wave spectra representing sea conditions at Capo Linaro near Civitavecchia, Italy.
232

3. Yachts response
3.1.

Encounter frequency

Due to its forward speed V, the wave spectrum for the ship is different than for a fixed observer.
When studying the ships response it is therefore necessary to take into account the frequency at
which it actually encounters the waves (encounter frequency). The encounter frequency depends on
wave velocity and ship speed and relative direction with respect to waves . Angle is defined
between the forward directions of wave and ship: thus for bow waves = , for transverse waves
= /2, and for aft waves = 0.

Fig. 2. Definition of angle of encounter [5].


Encounter frequency2 must be evaluated taking into account the velocity component of the ship in
the direction of the waves, subtracting wave velocity c. The relative velocity is given by:
Vrel

c V cos

(15)

Thus the encounter period is:


Te

Vrel

(16)

c V cos

The encounter frequency is thus given by:


2
c V cos
e

(17)

For seakeeping purposes, the assumption of deep water may be applied; in this case, taking into
account the dispersion relation (2), the encounter frequency can be finally derived as:
2

V
cos
g

(18)

The wave energy spectrum must also be modified according to the encounter frequency (it is
practically a Doppler shift of the spectrum). Since the energy content of a spectrum must be the
same for any observer, fixed or moving with the ship, the 0-th momentum must be the same:

m0

Se

(19)

Therefore, the relation between wave spectrum and encounter spectrum is the following:
Se e d e S d
(20)
2

In this paper, the term frequency will be used indifferently to identify both frequency f, measured in Hz, or angular
(circular) frequency , measured in rad/s.

233

which becomes, taking into account that d

1 2 V cos

Se

(21)

V
1 2
cos
g

3.2.

gd :

Response amplitude operators

The ship response is usually described in terms of transfer functions (RAO, Response Amplitude
Operator), which give the normalised amplitude of the resulting ships motion for a sinusoidal
excitation of frequency e, the normalization factor being the wave amplitude 0 for linear motions,
2
the wave slope k 0 2 0 /
for angular motions and the wave acceleration
e 0 for
accelerations:
z0
RAO z e
(22)
0

RAO

(23)

a0

RAO a

(24)

2
e

Obviously, equally important are the phase shifts


of each motion with respect to the wave
excitation. With the knowledge of RAOs and phase shifts, it is possible to reconstruct heave ( z),
pitch ( ) and roll ( ) motions from a sinusoidal wave excitation t
0 exp i e t as follows:
zt

z 0 exp i

exp i

exp i

(25)

(26)

(27)

Figure 3. Coordinate system and definition of motion [9].


Therefore, vertical oscillations for any point on the ship may be calculated as follows (fig. 3):
y t z t L sin t
B sin t
(28)
where L and B are the longitudinal and lateral distance of the point of interest from the center of
gravity. Since angular motions (pitch and roll) are usually small, it is possible to approximate the
above expression:
234

yt

zt

L t

B t

(29)

Therefore, being the sum of harmonic motions (phasors), the vertical oscillation y t
represented by a harmonic oscillation:
yt
y 0 exp i e t
(30)
y

is also

and it is possible to define a RAO for the particular point of interest on the ship:
y0
RAO y e
(31)
0

In case of a random wave excitation, with the assumption that the response is a linear function of
wave amplitude and applying the superposition principle, vertical motion can be reconstructed as:
n

yt

y 0, j cos

e, j

(32)

y, j

j 1

where each oscillation amplitude y0,j is a function of frequency and amplitude of the j-th harmonic,
according to (31).
Furthermore, it is possible to demonstrate that the ships response energy spectrum is given by the
product of the square of the RAO and the wave energy spectrum. Thus, heave motions energy
spectrum is:

Sz

RAO 2z

(33)

and analogous equations hold for the other motions, while for any point on the ship the energy
spectrum associated to its wave- induced motion is:

Sy

3.3.

RAO 2y

(34)

Simulation results

The foundation for the commercial software package Seakeeper 3 , which was used to carry out the
computation of the yachts motions under different wave conditions, is the linear strip theory based
on the work of Salvesen [10], which is used to calculate the coupled heave and pitch response of the
vessel; the roll response is calculated using linear roll damping theory [11].
The main purpose of the kinematic model presented is to provide reasonable data about the
response of a generic yacht to different sea conditions, in order to have reliable information on the
motion which the SEAKERS device is subjected to. Since the project does not address a particular
yacht model, nor even a specific size of boat, there was no point in developing a focused in-house
software: hence the choice of adopting a commercial software that has a proven record of reliability,
using it to simulate the response of a yacht of adequate length included in the extensive library
provided.
The yachts model used in the numerical simulations is one of the library models that can be found
in Seakeepers library, since it has geometric and mass properties comparable to those of
commercial sail yachts of interest for the SEAKERS project.
The most relevant hydrostatic properties of this yacht are given in table 2. The generator considered
in the simulations presented is placed at bow on the longitudinal axis (B = 0) at a distance L =
5.17 m from the center of gravity.

Seakeeper is a software by Formation Design Systems Pty Ltd (trad ing as FormSys); website:
http://www.formsys.com/maxsurf/ msproducts/seakeeper.

235

Figure 4 gives an overview of 21 two-dimensional sections used in the Seakeeper software to


evaluate sectional hydrodynamic masses, damping coefficients, and all other data needed in the
context of the strip theory [5-9].
Table 2. Yachts hydrostatic properties.
Value
6.531
6.372
11.5
2.475
3.054
10.64
2.866
1.213
21.21
0.494
0.068

Displacement
Volume (displaced)
Overall length
Draft amidships
Immersed depth
Waterline length
Max beam on waterline
Max section area
Waterplane area
Prismatic coefficient (Cp)
Block coefficient (Cb)

UoM
t
t
m
m
m
m
m
m2
m2

Fig. 4. Yachts mapped sections used to evaluate hydrodynamic coefficients in the equations of
motion by the Seakeeper software.

As illustrated in section 3.2, the ships response is defined by means of Response Amplitude
Operators (RAO) and phase shifts, with reference to a sinusoidal wave excitation. Figures 5 and 7
show values of RAO for each motion (heave, pitch and roll) for two different speeds (V = 5 knt and
V = 8 knt respectively), while figs. 6 and 8 show the phase shifts, as obtained by means of the
Seakeeper software. For the roll motion, the default value of non-dimensional damping factor
proposed by the software has been taken into account.
The response to random waves is illustrated in figs. 9 and 10 in terms of energy spectra of the
vertical oscillations (34). The significant oscillation amplitudes are obtained from these spectra in
the same way as the significant wave amplitude (10) is calculated from the wave spectrum:

y 0s

2 m0 y

Sy

(35)

Values of significant vertical oscillation amplitudes, corresponding to the energy spectra of figs. 9
and 10, are given in table 3.
236

Fig. 5. Yachts response: response amplitude operators (RAO) at speed V = 5 knt.

237

Fig. 6. Yachts response: phase shifts at speed V = 5 knt.

238

Fig. 7. Yachts response: response amplitude operators (RAOs) at speed V = 8 knt.

239

Fig. 8. Yachts response: phase shifts at speed V = 8 knt.

240

Fig. 9. Energy spectrum of vertical oscillations at the generators location at speed V = 5 knt.

241

Fig. 10. Energy spectrum of vertical oscillations at the generators location at speed V = 8 knt.

Table 3. Yachts significant vertical oscillation amplitudes at the bow generators location.
y0s [m]
V = 5 knt
V = 8 knt
Hs [m]
= 90 deg
= 135 deg
= 180 deg
= 90 deg
= 135 deg
= 180 deg
0.5
0.433
0.383
0.322
0.405
0.366
0.318
1.0
0.721
0.684
0.688
0.680
0.697
0.732
1.5
0.980
0.943
0.967
0.937
0.964
1.024
2.0
1.147
1.127
1.175
1.120
1.152
1.233

242

4. Linear oscillator
4.1.

General remarks

The linear generator that will be used in the SEAKERS device to recover energy from the waveinduced motions of the yacht is analysed and approximated in this paper as a simple linear
mechanical oscillator, where the damping element represents a linear approximation of the effect of
the electromagnetic force exerted by the generator as it provides a voltage difference proportional to
the square of its relative velocity with respect to its basement, and the spring represents the stiffness
of the generators support. It is further assumed that the damping coefficient can be dynamically
varied depending on sea conditions: this could be achieved in the final system by means of a
variation of some electrical parameters in the associated circuit. The equation of motion is thus:
mx c x

K x

mg

Fs

(36)

where m is the generators mass, x its position in an inertial frame of reference, y is the basements
position, c is the damping coefficient, K the springs stiffness, g the acceleration of gravity, Fs a
static force provided by the support in order to balance the weight mg such that Fs = mg.
It is assumed here that the support can exert such a static force in order to balance the mass weight;
it can be seen that mechanical springs alone cannot play such a role, because the resulting stiffness
would be too high for the typical forcing frequencies. Indeed, if the spring were to counterbalance
the weight with a limited elongation at rest l = 0.05 m, the resulting natural frequency would be n
= g/l 2.2 Hz, which is much larger than the forcing frequency of sea waves: as the following
section explains, this would make the system too stiff, i.e. the mass would move rigidly with the
basement, with no relative motion between the two and, thus, no power extracted.
Equation (36) can thus be rewritten eliminating all static forces and introducing the relative position
s = x-y of the mass in a frame of reference moving with the basement:
m s cs Ks
my
(37)
which becomes the well-known second order ordinary differential equation for an oscillating body:
s

2
n

(38)

with the introduction of the natural frequency of the oscillator:

K m

(39)

and of the damping ratio:


c 2 Km

4.2.

c 2m

(40)

Response to sinusoidal waves

The steady-state response of the linear mechanical system to a harmonic forcing of the type
yt
y 0 exp i e t is itself harmonic:
st

exp i

with a complex amplitude


0

y0

n2

(41)
0

given by:

n2
1 i2 n

(42)

where n is the ratio of forcing and natural frequency:


n
(43)
e
n
243

The magnitude of 0 gives the amplitude s0 of the harmonic motion of the generator (fig. 11, top),
and its ratio with the ships oscillation amplitude is the response amplitude operator for the
generators relative motion:
s0
y0

RAO s

n,

n2
e

1 n

while its argument


(fig. 11, bottom):
2 n
arctan 2
n 1

2 2

(44)
2 n

gives the phase of the generators motion with respect to the forcing oscillation
(45)

The resulting harmonic motion can therefore be expressed as:


s t s 0 exp i e t
(46)

Fig. 11. Frequency response of the harmonic oscillator: amplitude (top) and phase (bottom).
244

In this model, damping the oscillations in the linear mechanical system provides the mean to extract
energy from the wave excitation; thus, it is interesting to identify optimal values for the damping
coefficient c in order to extract the maximum power. The power absorbed is given by:
cs 2

Pt

(47)

and its average value over one cycle (which will be indicated as ) is:
1 T
1 2 2
P t dt
c e s0
(48)
T 0
2
the above expression, taking into account (40) and (44), becomes:

n3

3 2
e 0

1 n2

2 n

(49)

Since the oscillations are constrained by the size of the generator, two different scenarios must be
considered. In the first one, let us imagine that the undamped oscillations do not reach the
maximum range allowed smax : in this case, increasing the damping coefficient from 0 initially yields
higher values of even though s0 decreases according to (44), until a maximum for is reached,
beyond which it decreases. The optimum value of can thus be found when d /d = 0:
opt

2
11 n
2 n

(50)

with corresponding optimum damping coefficient, maximum average power and oscillation
amplitude given by:
1
(51)
c opt m 1
n2
1
m
4

max

s0

y0

3
e

1
y 1
n2
2
0

1
n2

(52)

(53)

In the second case, the undamped oscillations would be larger than the maximum allowed range
smax : then it is possible to extract more power by increasing the damping coefficient while the
oscillation amplitude y0 remains at its maximum permissible level smax . It is possible to show that
the maximum power is obtained when the damping coefficient is such that the oscillation given by
(44) is exactly equal to the stroke (s0 = smax ):
opt

c opt

n
2

y0
s max

y0
s max

1
n2

(54)

1
n2

(55)

and the corresponding maximum average power is:


max

1
m
4

3 2
e max

y0
s max

1
1
n2

(56)
245

If the mechanical system is tuned to the forcing wave condition (n


can be simplified as follows:
1
m e3 s max y 0 if n 1
(57)
max
4

1), then the above expression

It is possible to find out which wave excitations make the system reach its maximum stroke smax by
setting the oscillation amplitude given by (53) equal to smax , yielding:
y0
s max

21

1
n2

(58)

Thus, for wave amplitudes originating boat oscillations lower than the limit set by the above
equation, the system oscillates freely and equations (51)-(53) apply, while for higher waves more
damping, and thus more power, is available, in order to constrain the system within the maximum
stroke allowed, and (55)-(56) apply.
It is worth to point out that in both cases the optimum value for the damping coefficient is directly
proportional to the oscillators mass and to the forcing frequency: the average power absorbed is
therefore proportional to mass m and to the third power of forcing frequency ( e3 ), as shown by
(52) and (56).
In particular, the linear dependence on the oscillating mass m is, on the one hand, almost obvious
because energy recovery depends on inertia and kinetic energy, but on the other hand it is an
important property to be taken into account because it allows to design, test and prototype modular
systems of relatively low mass, with the overall power extracted given by the sum of power
available from different modules. For this reason, the results discussed in section 4.4 will be given
with reference to a unit mass m = 1 kg.

4.3.

Response to random waves

The frequency response of the harmonic oscillator can also be used when the external forcing is not
harmonic (as in the case of real wave excitation): if Sy( e) is the energy spectrum associated to the
vertical oscillations of a particular point of interest, which can be evaluated from the wave energy
spectrum by means of (34), then the energy spectrum associated to the relative motion s of a linear
system such as the one described in the previous section is given by:
Ss

RAO s

2
e

Sy

(59)

In the following considerations the dependence of relative motion and its spectrum on natural
frequency n will be implicitly assumed, so that Ss ( , e) Ss ( , n , e).
The spectrum of relative motion allows the evaluation of the significant oscillation amplitude as
follows:

s0 s

2 m0 s

Ss

(60)

Since power generation depends on the square of the generators velocity (47), the energy spectrum
related to relative velocity must be introduced. This is simply given by:
Ss

2
e

Ss

2
e

RAO s

2
e

Sy

(61)

The 0-th moment of the velocity spectrum gives velocitys root mean square (RMS), which is
related to average power generation:

s RMS

m0s

Ss

(62)

246

c
s RMS
m

RMS

(63)

n RMS

As in the case of sinusoidal waves, for a given natural frequency the generators motion depends on
the choice of the damping coefficient , for which an optimum value is found by maximizing power
output (63) with the constraint that the significant oscillation amplitude is lower than the maximum
allowed range smax (for this non- linear optimization procedure the MATLAB function fmincon
has been used):
d

RMS

0 with s 0s

opt

(64)

s max

opt

4.4.

Simulation results

In this section, results obtained with the mechanical model of the linear generator subject to random
wave excitations for two different speeds (V = 5 knt and V = 8 knt) are discussed. As detailed in
section 3.3, the yacht's oscillation is described by energy spectra represented in figs. 9 and 10.
The maximum stroke taken into consideration in the simulations is smax = 0.5 m, since this value is
about the highest possible on sail yachts of length from 10 to 14 m, which are the main target for
the SEAKERS project. The natural frequency n is taken as 0.25 Hz, in order to make the
mechanical system almost resonant with most sea conditions that may be encountered.
Figures 12 and 13 show the energy spectra associated to the generators relative motion and RMS
power generation for different speeds and directions. Power generation values are also given in
table 4, while table 5 gives calculated values of significant oscillation amplitudes.
From table 5 it is possible to see that for wave heights higher than 0.5 m, the generators oscillation
is always limited to the maximum range smax : in order to analyze the results, it is thus useful to
consider the simplified equation for average power generation (57), which shows that, if the
mechanical system is tuned to the forcing wave condition (n 1), average power is proportional
to the third power of the encounter frequency, and to the product of boats vertical oscillation y0 and
maximum range smax .
Therefore, an increase in significant wave height gives rise to two opposite effects on power
generation: on the one hand it increases due to its dependence on y0 , but on the other hand the wave
energy spectrum shifts towards lower frequencies (fig. 1), resulting in a decrease in power
generation. Clearly, this gives rise to a maximum power generation for a particular sea state, that
under the assumptions taken into account in this paper correspond to a significant wave height of
1.5 m, as figs. 12 and 13, along with table 4 show.
In other words, even with high values of significant wave height, which correspond to rather low
values of peak frequencies (table 1), if the full spectrum is taken into account significant
contributions to the excitation can be found also at frequencies higher than the peak one, and these
contributions increase average velocities and, consequently, power generation. Nonetheless, it is
still possible to find that increasing wave heights beyond a certain threshold decreases the power
output, because in this case significant contributions can indeed be found only at low frequencies.
Tables 6 and 7 report values of optimum damping coefficients and damping ratios as defined by the
optimization procedure (64).

247

Table 4. Average power generation.


RMS/m

[W/kg]
V = 8 knt
= 135 deg
0.352
0.644
0.852
0.757

= 180 deg
0.230
0.630
0.848
0.789

Table 5. Significant oscillation amplitudes for the linear mechanical system.


s0s [m]
V = 5 knt
V = 8 knt
Hs [m]
= 90 deg
= 135 deg
= 180 deg
= 90 deg
= 135 deg
0.5
0.500
0.332
0.281
0.477
0.303
1.0
0.500
0.500
0.500
0.500
0.500
1.5
0.500
0.500
0.500
0.500
0.500
2.0
0.500
0.500
0.500
0.500
0.500

= 180 deg
0.263
0.500
0.500
0.500

Hs [m]
0.5
1.0
1.5
2.0

= 90 deg
0.552
0.617
0.753
0.664

V = 5 knt
= 135 deg
0.401
0.620
0.807
0.715

= 180 deg
0.247
0.546
0.721
0.656

= 90 deg
0.500
0.572
0.704
0.619

Table 6. Optimum damping coefficients.


c/m [N s/(m kg)]
Hs [m]
0.5
1.0
1.5
2.0

= 90 deg
1.522
2.410
3.096
3.107

V = 5 knt
= 135 deg
2.709
2.729
3.665
3.732

= 180 deg
2.724
2.816
3.851
4.017

= 90 deg
1.469
2.232
2.894
2.934

V = 8 knt
= 135 deg
3.142
3.036
4.139
4.223

= 180 deg
3.228
3.413
4.725
4.922

= 90 deg
0.468
0.710
0.921
0.934

V = 8 knt
= 135 deg
1.000
0.967
1.317
1.344

= 180 deg
1.027
1.086
1.504
1.567

Table 7. Optimum damping ratios.

Hs [m]
0.5
1.0
1.5
2.0

= 90 deg
0.484
0.767
0.986
0.989

V = 5 knt
= 135 deg
0.862
0.869
1.167
1.188

= 180 deg
0.867
0.896
1.226
1.279

248

Fig. 12. Energy spectra associated to the linear mechanical systems motion at speed V = 5 knt.

249

Fig. 13. Energy spectra associated to the linear mechanical systems motion at speed V = 8 knt.

250

5. Conclusions
This paper has provided an overview of the results obtained with the kinematic and mechanical
models of the yachts response to different wave excitations, and of the linear generator taken as a
simple mechanical linear system which extracts power from the wave- induced motion by means of
an ideal linear damping.
These preliminary evaluations, even though based on a rather simplified model of the generator,
have produced some important insights on system dynamics and on the range of values to be
assigned to several significant parameters, such as mechanical stiffness and damping ratio.
In particular, given the particular range of forcing frequencies, the mechanical stiffness must be
chosen so as to obtain a natural frequency within the range of most forcing frequencies: a value of
0.25 Hz has been considered in this paper, with a resulting stiffness around 2.5 N/(m kg).
The choice of damping ratios is based on the maximisation of power output for a given natural
frequency for different wave excitation conditions. Values of damping coefficients in the range 1.55.0 N s/(m kg) have been found, and this result will be useful in the definition of the electric
circuits physical parameters associated to the linear generator.
Power generation of up to 0.85 W/kg have been obtained in the most favourable sea conditions, and
anyway values higher than 0.5 W/kg are available in most cases, which represent an interesting
result for this particular application. Indeed, a total weight for the SEAKERS device of up to 200 kg
can be considered acceptable on sail yachts with length in the range 12-14 m, especially in the case
of a first equipment (as opposed to retrofitting an already existing yacht, because in this case many
more design constraints should be addressed). With an optimized design, it is conceivable from
preliminary evaluations that up to 50% of this weight (100 kg) could be allocated to the oscillating
masses; in this case, an average power generation of almost 100 W is feasible, which could make
possible to recover at least 1 kWh at the end of a day- long cruise. This amount of energy generation
could indeed be interesting for this particular application.
Obviously, the issue related to the influence of this moving mass on sailing performance should be
addressed: but specific calculations, which have not been reported here for brevitys sake, show that
the inertial forces generated by the mass motion are at least two orders of magnitude lower than the
forces exerted by the sea on the boat. After all, this must be the case because the energy absorbed
by the linear system is but a small fraction of the total energy of the incoming waves. Therefore, in
all probability the impact of the added mass due to the generator on sailing performance can be
safely deemed negligible.
It should be observed that the concept of power availability (or availability factor) routinely
used to appreciate the performance of a system based on renewable energy, is much less useful for
this particular application, because the final goal is not the generation of electricity per se on a
continuous basis but, rather, only on the particular occasion when the yacht is used for a cruise. An
availability factor should therefore be considered only with reference to a single cruise, and it would
take into account the time frequency of encountered wave height during a typical cruise of a sailing
boat, because this is the main parameter influencing average power output. However, it is rather
difficult to make any prediction about this probability, other than saying that sail cruises are most
common when wind (and, consequently, wave) conditions are not extreme (i.e. with moderate
winds and waves, while it is safe to assume that calm or stormy seas are avoided); furthermore, a
leisure cruise usually requires stable weather conditions. In the end, it is reasonable to infer that in
most cases sea conditions encountered on a leisure cruise are reasonably constant and marked by a
significant wave height within the range 0.5-1.5 m. Under these assumptions, average power
generation is expected to be almost constant during the whole cruise.
The next activities in the SEAKERS project will be focused on the implementation of a non- linear
electro-mechanical model of the generator, in order to define suitable ranges for the most important

251

electrical parameters in the system, and to narrow down the set of possible runs of the more detailed
2D and 3D models that have been set up.

Acknowledgments
The research leading to these results has received funding from the European Union Seventh
Framework Programme (FP7/2007-2011) under grant agreement n 262591. The authors would
also like to acknowledge the invaluable effort provided by all the staff of Labor srl, the projects
coordinator, as well as the contribution of the SMEs and RTD performers involved in the project.

References
[1] Ivanova I.A., Bernhoff H., gren O., Leijon M., Simulated generator for wave energy
extracction in deep water. Ocean Engineering 2005;32(14-15):1664-1678.
[2] Thorburn K., Leijon M., Farm size comparison with analytical model of linear generator wave
energy converters. Ocean Engineering 2007;34(5-6):908-916.
[3] Eriksson M., Isberg J., Leijon M., Hydrodynamic modelling of a direct drive wave energy
converter. International Journal of Engineering Science 2005;43(17-18):1377-1387.
[4] Leijon M., Danielsson O., Eriksson M., Thorburn K., Bernhoff H., Isberg J., Sundberg J.,
Ivanova I., Sjstedt E., gren O., Karlsson K.E., Wolfbrandt A., An electrical approach to
wave energy conversion. Renewable Energy 2006;31(9):1309-1319.
[5] Bertram V., Practical ship hydrodynamics, Oxford, UK: Butterworth-Heinemann; 1999.
[6] Biran A., Ship hydrostatics and stability, Oxford, UK: Butterworth-Heinemann; 2003.
[7] Jensen J.J., Load and global response of ships, Oxford, UK: Elsevier; 2001.
[8] Rawson K.J., Tupper E.C., Basic ship theory, Oxford, UK: Butterworth-Heinemann; 2001.
[9] Perez N.A., Sanguinetti C.F.O., Scale model tests of a fishing vessel in roll motion parametric
resonance. Sntesis Tecnolgica 2006;3(1):33-37. ISSN 0718-025X. Available at
<http://mingaonline.uach.cl/scielo.php?pid=S0718-025X2006000100004&script=sci_arttext>.
[10] Salvesen N., Tuck O.E., Faltinsen O., Ship motions and sea loads. Transactions, Society of
Naval Architects and Marine Engineers 1970;78:250-287.
[11] Formation Design Systems Ltd. Seakeeper Manual Available at:
<www.formsys.com/extras/FDS/webhelp/seakeeper/skmanual.htm> [accessed 3.2.2012].

252

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Mirrors in the sky:


Status and some supporting materials experiments
Noam Lior
University of Pennsylvania, Philadelphia, USA, lior@seas.upenn.edu

Abstract:
This paper critically reviews the state of the art of an approach to supply energy to earth from space mirrors
that would be placed in orbit with angle control to reflect solar radiation to specific sites on earth for
illumination. These mirrors would be of t he order of a square kilometer or more, planned to be made of thin
plastic reflective films, which are launched to some optimal orbit around the Earth. One could, for example,
thereby provide night or emergency illumination to cities and other locations, or illuminate agricultural
production areas to lengthen the growing season, or to illuminate photovoltaic or thermal collectors on earth
for producing electricity or heat. Proposals were also made for using such mirrors for weather modification,
and we added here the possibility of using t he spac e mirrors for shading the earth to reduce global warming.
Experiments with spac e mirrors were conducted in the past by the former Soviet Union. In addition, thin film
aluminized Kapton mirrors were manufactured and optically an mechanically tested to examine their property
changes when exposed to a cryogenic temperature, an economic analysis related to several applications
was performed, and leading issues that must be taken int o account in the sustainability analysis of the
concept were described. Our experiments with thin film mirror have shown that the reflectance at 77K is
always higher (about 1.7 fold) than that at 300K, and the ultimate tensile strength and modulus increased at
the cryogenic temperature. Without (yet) consideration of environmental and social impact externalities, our
economic analysis agrees with past studies that if transportation costs to mirror orbit are reduced to a few
hundred $/kg, as planned, the use of orbiting space mirrors for providing energy to earth is an investment
with a good rate of return and a cost effective alternative to other power sources. This energy concept is very
appealing relative to other options for addressing the severe energy and global warming problems that we
face, and des erves much and urgent R&D attention.

Keywords:
Space mirrors, Space energy, Solar mirrors, Mirrors, Cryogenic mirrors, Illumination

1. Objectives and general background


This paper critically reviews the status and potential of space-based solar mirrors that reflect solar
energy (light) for use on earth, as well as our experiments with manufacturing and testing a
prototype thin- film mirror intended for that purpose.
Escalating problems of energy, environment and increased and more demanding population make it
increasingly difficult to generate power, heat, light and food on earth [1,2]. As described in the
publications by Glaser and co-workers (e.g. [3,4]), Mankins [5,6], Criswel and co-workers [7],
Brown [8], Woodcock [9], NASA [10], Lior and co-workers [11-13] and many others. Space has
many desirable attributes for serving as the location for supplying energy to earth by constructing
space satellite or moon-based power generation stations where the power is beamed to earth by
microwave or laser for use. This topic has received significant support by the U.S. NASA during
the late 1970-s till the early 1990-s, and beginning somewhat later but to some extent still
continuing by several European countries and Japan. A web site of the National Space Society [14]
keeps at this time a record of developments, an important recent update report Space Solar Power:
The First International Assessment of Space Solar Power: Opportunities, Issues and Potential
Pathways Forward edited by Mankins and Kaya having been published in 2011 [6].
The other space energy approach, which is the subject of this paper, is the construction and
deployment of space mirrors that would be placed in orbit with orbit, angle and altitude control to
reflect solar radiation to specific sites on earth for illumination. These mirrors would be very large,
253

typically of the order of a square kilometer or more each, highly reflective, planned to be made of
thin plastic reflective films to minimize weight and cost, mounted in an appropriate light frame.
They would be launched to one or more orbits around the earth. One could, for example, thereby
provide night illumination to cities and other locations and for emergency lighting, or provide
sunshine for agricultural production in some areas to enable or lengthen the growing season, or for
applications such as crop drying ad water desalination, or to illuminate photovoltaic (PV) and
thermal collectors (including salt-gradient solar ponds) on earth for producing electricity or for
heating. This approach was originally briefly proposed by the space science pioneer H. Oberth in
1928 [15]. This concept seems to have laid dormant and was then advanced most notably by
Buckingham and Watson (in 1968, 60 years later [16]), NASA (Billman, Gilbreath and Bowen)
[17-19], Ehricke [20-22], and others [23-25]. The review portion in this paper relies strongly on
[16-21] in recognition of the pioneering work of these authors.
When considering space power generation, the major advantages of the space mirrors approach are:
(1) instead of PV collectors on the energy source spacecraft there are only mirrors (optical
reflectors), planned to be made from very thin film coated polymers (microns thick); (2) the energy
is transmitted directly to earth in the form of solar light, without need for conversion of the
collected solar energy to microwave or laser beams and their transmission through the atmosphere
to earth; (3) sunlight is less threatening environmentally than the transmission of microwave or laser
radiation; (4) no requirement for power management and distribution or thermal management
systems on the spacecraft; (5) constructed of light thin ( m order) films, mirrors are easier to bring
to orbit and deploy than the equivalent PV cells; (6) if used for power generation, it would probably
need smaller collector and energy conversion fields on earth because of the safety-dictated need to
make microwave beams diffuse when PV satellites are used; (7) no need for technical energy
conversion systems on earth when the reflected sunlight is used for lighting, agriculture or bioenrichment. There are, however, also a number of significant technical challenges: (1) the reflected
sunlight arriving at the earth surface is more subject to the effects of weather, such as overcast, haze
and atmospheric refraction, than microwave or laser beams; (2) amounts of sunshine reflected to
earth that are sufficient to help supply significant fractions of needed global energy, and to be
commercially viable, would require very large (order of 1 km2 or more) mirrors, that must be
optically flat (to a fraction of a wavelength of light) over these huge areas, and durable both
mechanically and optically; (3) environmental effects, such as associated glitter and other light
pollution. These challenges have contributed to the fact that significantly less has been done so far,
or planned to be done, on space mirrors, when compared with PV solar satellites.
A significant albeit brief step in the development of space mirrors was the Russian Space Mirror
Project Znamya (banner) developed by the Space Regatta Consortium (SRC) [24] established
in 1990 by the Russian space agency and the corporation Energia [26] (which specializes in space
and launch vehicles and rocket boosters). The purpose of SRC in project Znamya, according to their
official website, was the development of thin sheet technologies for solar reflection and solar sails
and then for illuminating high latitude earth regions during winter months.
Detailed information about the Znamya experiments (Fig. 1) is somewhat sparse [24,26-28] , and
the following is available. The first SRC to be tested in space was Znamya-2, on February 4,
1993. The mirror was a 20 m diameter circular 5 m-thick aluminized PETF (Mylar) film, with an
estimated aerial density of 22 g/cm2 that was composed of 8 sections with radial gaps between
them. It was installed together with the unfolding mechanism inside the docking compartment of
the cargo space vehicle Progress M-15 which disengaged from the MIR space station. The
deployment test was successful. According to the SRC website, the spot of light produced by the
mirror was about 5 km in diameter and moved across the earths surface (starting in France and
through Eastern Europe and Asia) at a speed of around 8 km/s. The brightness of the mirror as seen
from the earth was reported to have been similar to that of a single full moon (<1 lux).
Znamya 2.5 was the second attempt to launch a space mirror, as a continuation of SRC's space
254

reflector experiments that hopefully will lead to the deployment of 200- m-diameter reflectors. The
reflector was 25- m-diameter and was constructed of materials and design similar to Znamya 2.
Deployment of Znamya 2.5 was attempted on February 4, 1999. As the sail unfurled it collided with
and wrapped around the docking antenna, and the whole apparatus crashed into the ocean. Since
then, there have been no attempts to launch a solar mirror. The Znamya experiments received much
attention from the media, including criticisms about light pollution that such space mirrors may
create [29-31].

a.

b.

Figure 1. Artists illustrations of the Znamya Solar Reflector a. [24], b. [27]


It appears that NASA has dedicated very small resources to research and development of spacebased solar mirrors. It did perform some fundamental studies describe in Section 3 below and
performed some slightly related experiments [32,33].
In the rest of the paper we show the basic equations for molar light reflection to earth, the different
space mirror system concepts proposed, reflector configurations, energy considerations, our work
on mirror materials and coatings, system economics and system sustainability considerations.

2. A few key equations for space mirrors [16-19,21]


The basic reflection optics are shown in Fig. 2.

Fig. 2 The illuminated ground area as a function of the space mirror orbit altitude h [17].
For mirrors of diameters considered in the studies so far, the illumination Ie at the earth's surface
by a space reflector in terms of solar illuminance Is. reflectance , cloudiness factor C (C = 1 for
cloudless sky), reflecting area Ar (of single or cluster of reflectors), angle of incidence between the

255

reflected beam and the illuminated earth spot, and angle at the reflector between the incident and
reflected light beam, is given by
Ie
A
(1)
C r f ( ) cos cos
Is
Ae
2
where f ( ) is a function representing the intensity extinction, due to haze and zenith distance (that
increases the beams path length through the atmosphere) and is the elevation angle of the
reflector above the horizon; for = 00 , f ( ) =1.
Equation (1) shows that he earth spot illumination intensity (Ie) increases in proportion to the
reflector area (Ar).
If the solar reflector is above the atmosphere, as typically planned for such space mirrors, the solar
radiation intensity at the reflector is at the atmospheres edge, Isc the Luminous Solar Constant is
133,334 lx or 134,108 lx and the Solar Constant = 1,366.1 kW/m. If it is within the atmosphere,
the illuminance Is at the reflector is diminished by effects of air molecules, dust and water vapor
along the beam path, with this diminution expressed by Co , the overall coefficient of absorption ad
reflection in a cloudless atmosphere, so
Is
Co
(2)
I sc
The image area on the earth of an orbiting mirror of area Ar positioned at a height h above that
image is expressed by
Ae
Ae

Ar

( h) 2 for a non-focusing reflector, and

(3)
2

( h) for a focusing reflector or point sources

where is the angular subtense of the Sun, = 1.39xl06 km/l.5xl08 km = 9.27 mrad, h = r - rearth , r
is the radial distance between the orbiting mirror and the center of the earth, and rearth is the earth
radius. The area illuminated on the ground is an ellipse with major axis (Dm + h cos ) and minor
axis (Dm + h). Equation (3) shows that the illuminated earth area becomes gradually independent
of the reflector area as the orbit altitude increases.
Very significantly, the overall reflectance of the mirror depends not only the surface specular
reflectance but also on its flatness to within a small fraction of the sun's angular diameter a. This
means that all parts of the mirror's surface must point in the same direction to within 1 or 2 mrad.
The reflector diameter influences the sharpness of the image. For a reflector of diameter Dr,
the earth spot image will have a penumbra region of shadow of the same diameter, which thus does
not practically affect the spot size.
For a synchronous orbit of h = 22,400 miles (36,049 km), the diameter of the illuminated spot on
earth is then 208 miles (~335 km). Obviously, if a smaller illumination area is needed the satellite
can be placed in lower orbits but then, as shown by Eq. (4), the illumination will take place for
shorter periods of time. This can be remedied by using a number of mirrors in the same or similar
orbits.
The period of a satellite (T) and the mean distance from the earth (h) are related by the
equation: T

2
h3
3600 K

1/2

(4)

where K GM earth , G is the Universal Gravitational Constant, G = 6.673 x 10-11 N m2 /kg2 , and
Mearth = 5.9742 x 1024 kg, so K = 398,659 km2 /s2 .

256

3. Space mirror concepts and proposed applications


Buckingham and Watson have in 1968 published a paper in which they described a system, shown
in Fig. 3, in which a synchronous altitude satellite with a large reflecting surface is used to reflect
the suns rays to earth [16]. The reflector is required to continuously change its angle of incidence
with respect to the sun- line to illuminate continuously a given spot on the earth. They provided
equations to calculate the illuminated area and illuminance with effects of cloud cover, and
proposed structural methods for frames to construct such thin film mirrors, as that shown in Fig. 3.
They concluded that reflector satellites are technically feasible but quite expensive for illumination
levels of 0.1 lx and higher, but may be economical for low levels of illumination of the order of 10-3
to 10-2 lx (less than 1/10 of brilliant moon light), useful for low- light- level sensors and could thus
roughly double their utility for night use.

0o

REFLECTOR SATELLITES

45o
SUN'S
RAYS
NORTH

90o

POLE

SYNCHRONOUS
ORBIT

180o

Fig. 3 Buckingham and Watsons basic concept of a reflector satellite and structure for supporting
the thin-film mirror [16].
A very comprehensive conceptual, technical and socio economic study and exposition of space
mirrors as conducted by the space visionary Krafft Ehricke [20-22]. He proposed and analyzed in
some detail a number of generic applications for providing lunar-type night illumination service
(Lunetta). solar type light energy services (Soletta), insolation for bio-production enhancement
(Biosoletta) to produce food and biomass, insolation for agricultural weather stabilization,
precipitation management, crop drying and desalination (Agrisoletta), and insolation for
generating electricity on earth (Powersoletta). He made an economic feasibility study and
predicted that very competitive electricity generation costs can be obtained; for example he
predicted that Powersoletta with a PV energy conversion on earth can produce electricity at
4.8c/kWh.
He added a number of new concepts [21] beyond past considerations:
use of a variety of sub-geosynchronous orbits, particularly, sun-synchronous ones,
splitting" of large single reflectors into a number of smaller reflectors operating in clusters and,
to reduce the size of the individual reflector, lower cost, increase system robustness. The
illumination pattern in this configuration is determined by the number of co-orbits, the time
position of their maximum latitude passage, and the lighting power (number of reflector units)
assigned to each co-orbit.
257

splitting of one orbit into several co-orbits (Fig. 4) which is particularly advantageous for urban
night illumination where multi-directional illumination creates a more diffuse and uniform
lighting effect;
More possible applications;
the concept of retro-reflection (called by some others relay mirrors), i.e. reflecting light from a
mirror that does not have direct optical sight line to an area on earth that needs to be illuminated,
to one in orbit that does (Fig. 5; thus enabling day and night operation raises the system's
utilization factor.

Fig. 4 The NASA SOLARES multiple orbiting mirror concept. Note 2 co-orbits and that several
mirrors exposed to the sun at the same time are reflecting to the same earth spot [19].

Fig. 5. Retro-reflection technique for daytime use of space light (adapted from [21]). Primary
reflectors (marked 1) are those that face the Sun and reflect the light to secondary (2) reflectors
that beam to the ground service area. Reflectors numbered with a prime beam solar radiation
directly to the ground service area without retro-reflection.
We add, without further analysis, the possibility of using the space mirrors for shading the earth, a
possibly useful application that may be locally useful for a number of obvious reasons, or as a geoengineering way to reduce global warming, a concern that did not exist when the early researchers
did their studies of solar mirrors in the 1970-s. In the extreme case, the mirrors, or some of them,
could be turned towards the sun and thus prevent the solar radiation reaching the atmosphere and
earth area that are that is now in the mirrors shadow.

258

NASA performed some detailed preliminary feasibility studies of the design, deployment and use of
space mirrors, published in the late 1970-s [17-19] 1 , and concluded that The use of orbiting mirrors
for providing energy to ground conversion stations to produce electrical power is shown to be a
viable, cost effective and environmentally sound alternative to satellite solar power stations and
conventional power sources.
Their proposal, which they called SOLARES, was to use a cluster of free-flying very lightweight
(10 g/m2 ) metal-coated polymeric film mirrors, optimally 1 km2 each which, after deployment at
altitude of 800 km, are placed in operational orbit and controlled by solar radiation pressure, to
almost continuously illuminate a chosen surface on earth an intensity of Ie ~ 1kW/m2 (at a fairly
constant level, which, however, must take into account atmospheric variability with time). This
would increase the available insolation at the earth energy collection and conversion station about
4-fold, and, if the insolation is uniform enough over time would also eliminate or reduce the need
for energy storage.
They developed equations showing the influence of a number of parameters - mirror altitude, orbit
inclination, period, mirror size and number, and atmospheric effects - on the reflected insolation
that may be received by a round spot as a function of location. In their economic analysis they
found that generated electricity costs range may be as low as about 1.6 c/kWh (in 1977/8 US cents).
At the same time, as discussed in more detail in Section 7 below, they used extremely low costs for
transportation into orbit, which make the costs of electricity and heat generated they determined
much too low when considering current technology. As the environmental issues of principal
concern they identified the perpetual twilight that neighboring communities might experience and
the land area required, and felt that atmospheric effects are minimal and to their opinion perhaps
beneficial. More details about their economic and environmental study can be found in Sections 7
and 8. They expressed the opinion that SOLARES could supply the entire global energy
requirement.
Other authors have proposed mirror deployment at geostationary orbits (GEO), but Eq. (3) shows
that at this altitude of h = 35,800 km the area illuminated on earth would have the huge diameter of
about 3,329 km. At the chosen ground intensity of 1 kW/m2 the mirror area would then have to have
an area of about 150,000 km2 . The annual energy generated at one such location with 15% ground
conversion efficiency would be, if atmospheric solar radiation transmission effects are ignored, up
to about 41,200 EJ, 82 fold of the current world usage of 500 EJ. To achieve a practical ground area
size with realistic capital investment and energy output, to provide energy to more than a single
ground station, and to be able to employ the enhanced insolation for nonelectrical applications if
desired, they postulated the use of a large number of flat1-km diameter reflectors in lower orbits
(Fig. 4). Such configuration would allow each selected ground site could be insolated at all times
(excluding eclipse and inclement periods), and any given mirror could be used for other tasks,
including the insolation of other sites. The use of many and small reflectors clearly also allows the
desirable feature capability of incremental implementation and easier repair and replacement.
Smaller reflector areas also require much lower torque for their control, since their moment of
inertia scales as Ii ~ ARi2 t where is the average areal mass density, A is the mirror area and Ri is
the characteristic radial dimension along the ith rotational axis.
These NASA studies also calculated the daily and annual variation in the solar flux, both the natural
one and that supplied by the orbital mirror system, and it is shown in Fig. 6. The most impressive
feature is that although the direct solar input varies seasonally by more than a factor of two, the
mirror input is constant to about 10%, making the system suitable for baseload electricity
generation use.

It is noteworthy that while these NASA publications all had the same objective and focus, the assumptions and results
changed with the publication time, probably indicating an evolution of the concept.

259

Thorough techno-economic analysis is required to find the optimal system, so they only considered
an example of the mirrors at an altitude of 4,146 km in a 3-hr periodic orbit with inclination of 45
relative to the equatorial plane (or several inclined orbits separated by latitude). The mirrors would
be deployed or erected at an altitude of approximately 800 km. From this altitude, where the solar
radiation pressure is much larger than the drag force, it is possible to "solar sail the mirrors to their
operational orbit (i.e. 4,146 km), requiring about a 3- mo transit time.

Fig. 6 The annual variation of the solar flux at the area illuminated on earth by the sun alone, by
the proposed mirror system, and by their combination [19]
Using Eqs. (1) and (3), and as best it can be concluded from [17-19] assuming 23% losses and
geometric spreading due to the sun- mirror-site angle, eclipsing, non- zenith mean apparent reflector
location and atmospheric effects, 62,800 km2 of mirror area was stated to be required to deliver an
average 1.25 kW/m2 (0.25 kW/m2 from regular solar incidence + 1 kW/m2 from the mirrors) to
ground stations at 30 latitude. With their proposal to build individual mirrors of 1 km diameter
each, this translates to the need for 80,000 orbiting mirrors having a total mass of about 6.3 l08 kg.
This mirror system was estimated to be able to supply this flux to at least 5 (of a theoretical 13)
ground sites around the world. For each site, about 70% of the incident insolation falls within area
of diameter
(38.4 km, Ae = 1,167 km2 ) and 99% within 2
(76.8 km, Ae = 2,334 km2 ). We note
that a conclusion from [19] appears to be that the ratio of the mirror area and the ground area
steadily insolated with the added 1 kW/m2 from the mirrors is 62,800/(5x1,167) = 10.76. With 15%
conversion of just the
insolation to the five ground stations, up to 27.6 EJ of electricity can be
generated, amounting to about the electricity generated globally in 1977 (at the time of the NASA
studies) and 36% of the electricity generated in 2010. After deducting the energy converted in the
PV system, the remaining 85% of the energy at , as well as the energy in the annulus between
and 2 could additionally constitute a large usable energy resource.

4. Some reflector configurations


The reflector may be of any geometry, usually dictated by structural and weight considerations, and
the mirrored surface may be flat, as proposed in most of the studies, or curved for concentration. If
the reflector is so large that its size may no longer be regarded as a negligible fraction of its image
size a curved concentrating surface is needed. It is noteworthy that it may be possible to vary the
concentration (curvature) as needed during the orbital motion of the reflector.

260

Based on such a concept, the NASA [20] study proposed that the additional insolation from the
space mirrors can be significantly greater than average ambient sunlight, and they have chosen to
supply an earth surface and additional solar intensity of Ie ~ 1kW/m2 (total of ~1.25 kW/m2
including the natural, unreflected, insolation) for sizing the space system and ground stations and
for deriving costs. They point out that the maximal average U.S. value of normal sunlight is about
0.25 kW/m2 , and therefore this increase intensity should reduce the area-related terrestrial solar
converter system costs (for collectors, converters, land preparation, etc.) fivefold [(1 + 0.25)/0.25 5] from that of a solar power generation system of equal output that operates without the space
mirrors.
The NASA study [18-20] has shown that for a given orbital inclination the number of mirrors
needed to provide continuous insolation at a given ground site increases with decreasing altitude,
and thus the total mirror area for a fixed ground site intensity decreases. However, several factors
place a limit on the lowest usable altitude. First, atmospheric drag necessitates an altitude above
1,500 km for the 15 g/m2 structure they proposed to allow a system life of 30 years. A remedy is to
employ solar sailing for countering drag, thus perhaps providing the desire system life down to
altitudes of 1,000 km. Second, the angular acceleration needed for the mirror to insolate a given
spot during its transit varies approximately inversely with the third power of the altitude, thus
creating significantly tougher demands on structural characteristics and control at lower altitudes.
Third, lower orbits increase the fraction of time the mirror is eclipsed by the Earth. They thus
conclude that the lower bound for an operational reflector system is probably not less than 1000 km.

5. The energies: generation and embodied


As stated above, the reflected insolation to a ground area can be augmented by using the space
mirrors as a Fresnel field. For various environmental and social reasons it is safer to limit the Ie,m
reflected from the mirrors to approximately maximal natural levels and in the NASA study [17-19]
it was proposed to make it 1 kW/m2 . This would be suitable for agricultural as well as heating and
power generation purposes. Atmospheric radiation- loss effects have been considered in the
estimation of the required mirrors area, and to avoid the important losses due to persistent
cloudiness (scattered clouds were indicated to have minimal effect), it is recommended to install the
mirror- illuminated ground stations in sunny regions that experience least cloudiness. It is important
to keep in mind that the space mirrors eliminate the diurnal and seasonal periodicities due to the
rotation of the earth and the sun. Based on global insolation data, it was assumed that the timeaveraged insolation without the mirrors is 0.25 kW/m2 , for a total ground insolation of 1.25 kW/m2 .
On this assumption The NASA study found that the total energy used to produce the SOLARES
space system (mining through turn-on) was about 1.5x1012 kWh. With the above assumption of Ie =
1.25 W/m2 and a 15% ground conversion efficiency, this embodied energy is equivalent to only 10
weeks of energy production at the five ground sites. Inclusion of the ground system added from 4 to
15 weeks to this number, depending on the conversion technique.

6. The space mirrors


6.1 Materials and optics
The design of the mirror needs to provide maximal specific power reflected with very low weight
and payload volume. Its surface must have a reflectance ( ) that is as close as possible to 1.0, it
must be accurate enough to ensure that most of the solar radiation reflected from its surface arrives
at the earth site area dictated by the fully-planar mirror optic, it must be durable and easily deployed
and the material needs to withstand years of solar winds, radiation, and the extreme temperatures
and their variations in space. Importantly, the mirror is exposed to micro- meteorites/ space debris,
electromagnetic radiation from sunlight [34], including solar wind, comprised of streams of
261

particles originating from the sun and propelled by the Earths magnetic field. The composition of
solar wind includes approximately 2x108 H+ /cm2 s protons (96%), of about 6x106 alpha
particles/cm2 s (3-4%) and a few 105 ions/cm2 s of higher mass with average velocities of about 400
km/h (corresponding to energies of 0.85, 3.4, and 10 keV, respectively) as well as 3-30 keV
electrons. There are also some occasionally emitted particles as the result of solar flares or storms.
Studies in [35] have shown a rapid worsening in the optical properties, including loss of reflectivity
and defocusing due to blistering, due to the effects of solar wind, affecting the very beginning of a
solar mirrors operation. They suggest the mirrors film thickness should be at least 0.1 m, which
is the maximum penetration depth of the solar particles. As discussed below, this is also the same
minimal thickness required for the metallic thin film to remain opaque to visible light and to other
low frequencies of electromagnetic radiation. Another type of radiation is the ultraviolet (UV) part
of sunlight with wavelengths between 4 and 400nm. Two UV bands are particularly relevant to
materials degradation, the near UV range (200 400nm) and vacuum UV range (100 200nm)
[36]. This type of radiation causes the greatest material degradation to polymers and most of this
damage is sustained by the first 0.3 microns from the surface.
A commonly used mirror substrate is polyethyleneglycolterepthalat (PETP) with the net
composition (C10 H8O4 )n , known as Mylar, Hostaphan, etc., and are produced by companies like
Bayer, and Du Pont, and Kapton, which is poly(4,4'-oxydiphenylene-pyromellitimide) made by
DuPont. They are coated with a thin film metallic surface by chemical vapor deposition (CVD) to
provide high reflection. The most common metal is aluminum due to its good reflectance and low
cost, but gold and silver were also used for small mirrors due to their stability or higher reflectance.
It is generally agreed that the film mirrors should be maintained periodically, say once in 10 years,
by applying a thin fresh Al (or other reflective material) layer in situ, which could be accomplished
by flying a furnace with evaporating Al along the foil surface at a certain distance [34]. This was
expected to restore the initial reflectivity (smoothing of the blistered flaked areas), and sintering
together the eventual brittle foil surface areas (flakes) by the freshly evaporated material. to
prolong the mirror reflectivity to the order of a 100 years. Such maintenance would also be needed
to repair possible holes due to meteorite impact, or for replacement of part or all of the reflector
film. It was estimated, however, that meteoroid damage would be very small, 3% for 30 years in
orbit [17].
The solar mirror must have mechanical properties that can withstand its temperature in space, that
goes down to 3K (with the associated embrittlement) but also rises to much higher values especially
on its non-reflecting parts which are intermittently exposed to the sun. The mirrored surface
substrate has the leading role in the mirrors overall structural integrity.

6.2 Our reflective thin film mirrors construction and experiments


We have investigated the fabrication of thin film mirrors akin to those that were considered suitable
for space mirror application (Sections 1, 3, 6.1), and then examined their microscopic surface
quality and measured their reflectance, tensile strength, creep (fatigue) at both room and cryogenic
temperatures, i.e. 300K and 77K, respectively.
Based on a list of the most common space materials used by NASA, the polymer substrate was
chosen to be the polyimide Kapton HN (Dupont) because of its much greater tolerance by as
much as three orders of magnitude to radiation, such as UV and soft X-Ray, in comparison with
another commonly used polymer, Mylar (PET), and because of its proven performance in
temperatures near 0 K and greater resistance to high temperatures, and its higher tensile strength
[37]. Aluminum was chosen as the reflective thin film because it is a commonly used reflective
material for thin film mirrors, primarily because its density is an order of magnitude and its cost is
two orders of magnitude lower than silver, which is often used for mirrors.
The thickness of the Kapton film substrate was 7.6 m, as thin as Dupont had commercially
available. The thickness of the aluminum coating was chosen to be 100 nm, again as done by
262

NASA. The size of the available vacuum evaporator dictated the diameter of the mirror film to be
134.6 mm. The pure aluminum coating was done by vapor deposition, and the manufactured film
mirror supported by a ring structure is shown In Fig. 7.

Fig. 7: Our manufactured reflective thin film


The film reflectivity was measured as a function of wavelength in the visible light range (300-1000
nm), at ambient and cryogenic temperatures, by using a spectrophotometer, using as reflector
reference a commercially available reflective mirror film (Mylar with Ag coating) made by 3M.The
cryogenic temperature was obtained by immersion in liquid nitrogen (~77K) for 60 min. While the
absolute values of the measure reflectance are subject to large experimental error, the trends are
consistent: the reflectance at 77K is always higher than that at 300K. For example, at the median
wavelength of 450 nm, where solar radiation has maximal intensity, the reflectance at 77K is about
1.68 fold higher than at 300K for our sample and 7% higher for the 3M film. Also consistently, the
reflectance of our film is about 16% lower than that of the reference 3M one. The film transmittance
was measured too, in the same wavelength range and was found to be zero for all films tested, at
both temperatures.
While surface contraction at the cryogenic temperature, resulting in contraction of pinhole and other
defects may partially explain the increased reflectance, much more detailed examination of the
films is needed for full understanding of the reasons for that reflectance-temperature relationship.
Samples from the test, of both the mirror film we manufactured and the commercial 3M reference
sample, were examined by an optical microscope. The first set of these two samples were kept at
ambient temperature (approximately 300 K), and the second set was exposed to at least 60 minutes
in an isothermal container with liquid nitrogen (~77K). Once the samples were removed from the
nitrogen bath, they were rapidly cleaned and observed and photographed at a magnification of 500.
The surface photographs are shown in Fig. 8.
Film mirror
temperature

300K

77K

Our film
mirror

Reference
film mirror

Fig. 8. 500X microscope surface photos of our thin film mirror and of the reference one, at 300K
and 77K. Note: the bean-shaped dark area in the center of each photograph is a smudge on the
microscope lens, not a defect.
Figure 8 shows that the number of defects/imperfections per unit area (i.e. surface defect density) is
higher for our film mirror than for the reference one. These defects include mostly pitting for the
samples and spits in the standard, typical of using a vapor deposition. This may therefore be the
reason for the measured lower reflectance of our mirror. Since the vapor deposition process we
used can be improved significantly, there is no doubt that better reflectance can be obtained, as
shown by many.

263

An interesting observation is that for both films the imperfection surface density is lower for the
cryogenic temperature, perhaps thus justifying the higher reflectance at this temperature. Another
interesting observation is that nevertheless, the reference film experience significantly higher
degradation when exposed to the low temperature, as shown by lines the along the vertical direction
of the image that appear to be micro-cracks that formed after exposure to liquid nitrogen (~77 K).
In fact, when the reference film was placed into the liquid nitrogen it made a loud crackling noise
that lasted for a few seconds, while no sound was heard for our film. In contrast, the sample did not
exhibit any sounds at all when placed inside the container. No micro-cracks developed in our film.
The likely explanation is that the thermal expansion coefficients of silver and Mylar (the reference
film) differ by 6-fold, while those between Aluminum and Kapton (our film) differ by only ~15%.
Table 2 below shows the experimental values for ultimate tensile strength and Youngs modulus
from our mechanical tests.
Table 2: Measured Mechanical Properties for our film mirrors exposed to 300K and 77K
Film temperature, K Ultimate Tensile Strength, MPa
Young's Modulus, MPa
300

9.9

129

77

13.3

448

The observed trends are consistent with the behavior of thermoset polymers, like polyamides, which
are characterized by a high tensile strength and modulus with a small total elongation, prior to
failure, brittle behavior, and an increase of the ultimate tensile strength and tensile modulus at lower
temperatures.

6.3 Mirror Mounting Structures


The structure must of course support the reflective mirror film, but must also provide the necessary
tension for maintain its shape, as well as the control mechanism for adjusting its orientation during
flight as needed. It was proposed that the control devices (thrusters) be momentum flywheel pairs
(Fig. 9), which can change their relative speeds to provide turning torques and steering corrections,
and which could be charged with small motors powered by solar cells or by radiation pressure.
The huge size of these mirrors mandates assembly in space, with the components ferried to the
location by a cargo space vehicle, and then constructing the frame from its component and unfurling
the reflective film and attaching it to the frame. As described in Section 3, the NASA plan was to do
the assembly at an altitude of 800 km and then ferry the mirror into its final orbit by solar sailing.
For the SOLARES concept, NASAs study [17,19] proposed mirror structures, shown in Fig. 9,
which, although not yet optimized, can work. The aluminized film was assumed to have an areal
density of 4.0 g/m2 . The film is tensioned onto a supportive structure consisting of an outer torus
with radial-segmented spokes and concentric rings. Such tensioning is necessary to maintain the
reflector planarity at 0.01 rad (deemed as adequate to maintain a flatness that will not increase the
spot size by more than 5% over that of a perfectly planar mirror) despite the perturbing radiation
pressure, gravitational gradient, and angular acceleration forces acting on it. The rim torus must
have sufficient buckling strength to provide this tension. They proposed that the support structure
would be fabricated from a high- modulus carbon-fiber-epoxy (or polyimide) matrix composite.

264

Fig. 9 Proposed structural configuration of a space mirror and its flywheel control system [17,19].
The design draws on interest in constructing and using solar sailing for transporting payloads in
deep space exploration, in one case of which an ultralight truss mast that can be deployed to
kilometer lengths was designed, and a means to fabricate a large quadrant of sail material and stow
it without material creasing or trapped air were proposed [28,38,39]. Sailcraft areal densities
ranging from 8.9g/m2 down to 4.6g/m2 were proposed by NASA as a goal, and since the sail
material used in NASA experiments had an areal density of 3g/m2 , this leaves 1.6g/m2 for the
supporting structure, bus and payload.
Parallel applications of more immediate interest are high-precision space reflectors for
communications, earth observation, or radio-astronomy. At a few tens of meters in diameter, the
ones considered are orders of magnitude smaller than space mirrors and they can afford an order of
magnitude higher areal mass of typically 0.5 kg/m2 , but some of the goals and technology are
synergistic. Here we mention the work by Datashvilli, Baier and co-workers [40,41], with a small
model of their design shown folded and then deployed in Fig. 10.

Fig. 10. Deployment of membrane reflector model (0.6m diameter), a) stowed, b) deploying, c)
(scaled) deployed [40].

7. System Economics
While it is obviously extremely difficult to predict the costs of generated energy by space mirrors,
because of the novelty of the concept, uncertainties about developments that will arise during the
265

decades needed for its materialization, as well as about all of the externalities that will accompany
it, several economics analyses were conducted to at least provide an initial prediction.
A study by Ehricke in 1979 [21] estimated that the investment for Biosoletta is of the order of
$1,200 billion, that is about $80 billion/year for the 15 yr construction time he predicted.
The NASA economic analysis of their SOLARES space mirror concept [19] see also Section 3) to
supply approximately the global electricity demand at that time (32 EJ, which in 2010 is 77 EJ [2])
based on a desired 15% capital return, 30-year system life, and a load factor which takes into
account eclipse and inclement periods, found that generated electricity costs range from about 2.5
c/kWh to less than 16 c/kWh (in 1977 US cents), and that the ground station for converting the solar
radiation received from the mirrors to electricity by using PV is the major component of the total
system investment, since the cost of reflectors in space is much lower. We note that this was based
on PV system costs of $5/Wpeak that have since then dropped in some favorable cases up to less than
$2/Wpeak [42], but that is likely to have also been true for other components of the space mirror
system. The same study states that if the solar radiation incident on earth from SOLARES was used
just as heat, the cost would be about 1 c/100MJ thermal.
In that analysis, performed on the assumption of making 80,000 space mirrors of 1 km diameter
each to be placed in a 4,146 km altitude orbit, they assumed the use of a thin film mirror made of a
0.1 m Al reflector on 2.5 m polymer substrate weighing 4 gr/m2 , and the structure made from
HM graphite-epoxy. The total mirror weight, including also controls, instrumentation and growth
allowance, was estimated to be 10.01 gr/m2 , i.e. 7,860 kg/mirror, and the total cost of the mirror
including prorated R&D costs excluding transportation costs into orbit was $1,654,000 per mirror,
i.e. $2.11/m2 . Including the transportation they came up with a total price of $2 million per mirror,
i.e. $2.55/m2 . They estimated the cost of the total system to supply approximately the entire world
1977 electricity demand of about 32 EJ to be about $500 billion or an average of nearly $40
billion/year with their assumption of a 15-year implementation period.
A major problem with their economic analysis is that they used a cost of $44/kg for transportation
into the needed orbit, assuming the existence of the planned Heavy Lift Vehicle. This is an areal
mirror cost of $0.44/m2 . A vehicle that can transport at such a cost has never materialized, in fact
the current space transportation costs acan at best be around $3,000/kg and are more likely around
$10,000/kg [43]. Conservatively assuming the latter cost and keeping all the NASA assumptions the
same, the price of a 1 km diameter orbiting mirror would rise to $80.3 million, i.e. $102.30/m2 (40.1
times higher) and would thus raise the price of generated electricity in similar proportion and make
it highly uncompetitive. It is noteworthy that the cost of transportation to orbit is also the major
obstacle for economic deployment and use of the SPS, and the only way to have competitive space
power generation is to reduce the transportation costs to about $200/kg, i.e. by nearly 2 orders of
magnitude, achievable by frequently planned but never commercially produced reusable launch
vehicles (RLV) [10,11,14] and other methods.
We have also conducted an economic feasibility analysis of the use of a space solar mirror system
by approximating life time costs, profits and resulting revenues for three separate applications: (1)
24hr farm lighting, (2) Night-time illumination in Polar Regions, and (3) Greenhouse produce
growing in Polar Regions.
The mirror aerial weight we determined was 11 gr/m2 (an aluminized 7.62 m Kapton mirror
reflector),which turned out to be close to the 10 gr/m2 used in the NASA study [19], and
considering all of the reflector components the reflector cost was estimated at $2.05/m2 . Using the
NASA data that show that reflector cost is about 48% of the total mirror cost without transportation,
our estimate of the mirror area cost was thus $4.27/m2 .
For transportation costs, we used those for the Falcon 9 rocket in [44], at $7,143/kg. i.e $78.57/m2
of mirror. Like in NASAs analysis we assumed a functional life of the mirror as 30 years and a
discount rate of 15% but also calculated for a 5% discount rate to provide a range of values that
266

come closer to present conditions. A lifetime maintenance cost of $1.27/m2 (4.25 c/(m2 yr) was
added. Summing the mirror and its launch costs, the total areal investment cost comes to $82.84/m2
($65.03 million per mirror) and the 4.25 c/(m2 yr) recurring maintenance costs.
From Eqs. (1) and (3) and as following NASAs estimates of the mirror area shown above and in
Section 3, each m2 of the mirror illuminates 0.093 m2 of ground site area, at the planed insolation of
1.25 kW/m2 . This amounts to 116.2 W per m2 of the space mirror. The annual solar energy
incidences are thus 39.4 GJ per m2 ground area, or 3.67 GJ per m2 of the space mirror.
If the energy is used for PV electricity generation at 15% conversion efficiency and normal
incidence, this would steadily generate 187.5 W electricity per m2 of the illuminated ground area,
and thus 17.4 W electricity per m2 of the space mirror, i.e. 150 kWh/yr (0.55 GJ/yr) per m2 of the
space mirror. At the typical US electricity price of $0.012/kWh, this should generate an annual
revenue of $18 per m2 of the space mirror. The cost of the PV system is estimated to be $4/Wpeak
[42,45], and considering the above result that the power generation is 17.4 W per m2 of the space
mirror, this would be i.e. $69.60 per m2 of the space mirror and is added to the $82.84/m2 space
mirror cost for a total areal investment cost of $152.44 per m2 of the space mirror.
The present value of the initial investment into the solar space mirror system for a discount rate of
5% is $276.05 and the NPV = $276.05 $152.44 = $123.61, and the internal rate of return (IRR) is
11%, indicating a financially rather viable investment. The NPV based payback period is 14
years. If the discount rate was 15%, as assumed in the prior NASA study, then the NPV becomes
negative, -$34.53, i.e. not financially viable. Discount rates up to 11% produce positive NPV.
Since transportation into orbit is in this analysis 52% of the needed capital investment, and since it
is the cost item most likely to decrease significantly in the future, the same economic analysis was
performed for the transportation costs of $200/kg, amounting to $2.22/m2 and resulting in an overall
mirror system cost of $6.51/m2 and to total mirror+PV system cost of $76.11/m2 . Now the
investment achieves an IRR of 24% and becomes very viable even under the 15% discount rate, and
we note that the mirror system becomes only 8.6% of the total cost.
Using the solar radiation from the mirror to grow agricultural products in Polar Regions allows an
extension of the growing season from the current 4 months [46] to 12 months per year. The benefit
is obviously avoidance of the need for importation of expensive produce in the off seasons, and the
amount saving by growing locally is US$4.50/lb or $45.52 per m2 of the space mirror. This creates
annual revenues that are 2.5- fold higher than those estimated above for electricity generation.
Furthermore, while not calculated here, the investment into a tomato growing greenhouse system
may also be lower than the installation of a PV energy conversion system, with all of this pointing
to the recommendation that such agricultural use is financially most viable.
Nighttime municipal illumination failed in this analysis to break even over the lifetime of the
mirror, because of the relatively low density of needed streetlights in the selected northern regions,
but it may be a rather viable application for both civilian and limitary purposes, when the demand
justifies it.
It is noteworthy that the space mirror designs considered in the NASA analysis and also used here
were not optimized. Furthermore, the costs of materials, of space deployment and of PV electricity
generation are dropping. The economics can thus only improve, unless some unknown technical or
environmental problem arises during the more detailed system development and testing,
The estimated very high needed investments, of the order of more than $600 billion (about $40
billion per year for about 15 years), for the space mirror system become more acceptable and
appealing when compared with the expected accomplishment of providing renewable and relatively
clean energy for satisfying energy of the order of the entire global demand, with relatively minimal
global warming effects. These investments should also be compared with some other global
financial values: in 2010 the annual world and OECD GDPs were about $63,000 billion [48] and
$42,000 billion [49], respectively, and the world defense budgets were $1,437 billion (2.3% of
267

GDP) [50]. The estimated annual expenditure for the space mirrors project are thus 0.06% of the
World GDP, of the order of the World Bank subscribed capital of $44 billion for 2010.
Some other proposed high magnitude global renewable energy projects were estimated for the
Space Power Satellite (SPS) [5,9] at $908 to $15,000/kWe, which for generating the current global
power capacity of 4.4 TWe [2] would require an investment of $4,000 to $65,000 billion f, and for
the DESERTEC project at close to $600 billion to supply by 2050 only 700 TWh per year of
electricity from the Saharan deserts [51]. The space mirror concept is predicted to incur much
lower costs.

8. System Sustainability
A huge and basically untested project like this one requires a very careful formal scientific
sustainability analysis from the very start. Founded on the commonly used economic,
environmental, and social pillars of sustainability, such an analysis in quantitative form is beyond
the scope of this paper, but some major issues are identified and discussed, as follows.

8.1 The environmental pillar


On the positive side:
the concept promises the satisfaction of a good part of the global energy demand from
renewable energy,
alleviation by orders of magnitude of emissions and global warming, and
preservation of the remaining fossil fuels/hydrocarbons for other uses.
Negative impacts are not minor and must be carefully considered and alleviated:
Emissions and noise from the launch vehicles; a current space launch typically produces 28
tons of CO2 , and 23 tons of toxic particulate matter[
Embodied emissions in the space mirror materials and construction
Effects on the atmosphere from the passage of the launch vehicles
Effects on the atmosphere from the added sunlight reflected to earth, including possible
photochemical effects; these may be accelerate global warming if greenhouse gas
concentrations in the atmosphere continue rising, but may ultimately reduce global warming
as fossil fuels are replaced by solar space mirror system.
The associated light glint (global) and scattering (near conversion sites) may add to the
ecological light pollution", a phenomenon well know among ecological and health hazards,
is one that alters natural light regimes in terrestrial and aquatic ecosystems, and some of the
catastrophic consequences of light for certain taxonomic groups are well known, such as the
deaths of migratory birds around tall lighted structures, and those of hatchling sea turtles
disoriented by lights on their natal beaches. The more subtle influences of artificial night
lighting on the behavior and community ecology of species are still less well recognized and
should be studied [52]. Light pollution also has detrimental effects on human lives possibly
impairing vision [53] and altering the production of melatonin, the hormone that makes us
sleep as it is released during darkness, and thus altering human circadian clock whose
disruption has been linked to depression, insomnia, cardiovascular disease, and cancer. Also,
according to the First World Atlas of the Artificial Night Sky Brightness, two thirds of the
US population and about half of the European Population can no longer see the Milky Way,
thus destroying the aesthetic value of the night sky and impeding human connection to nature
[54].
The additional light at night is an especially serious problem for astronomy since it prevents,
or at the least makes very difficult to conduct astronomical observations and studies. At the
268

same time, the associated launch capabilities developed for the space mirror system should
advances possibilities for space astronomical laboratories [19]. A response to the strong
criticisms about the effects of light that projects like Novey Svet ([24-26], Section1) may
create was posted by the Russian Space Regatta [24]; they indicate that the problem must
indeed be studied and any projects were and should be implemented with minimal damaging
impact, but also list the large advantages that such lighting can provide.
Generation of a large amount of space debris, and possible risks of their fall to earth

8.2 The economic pillar


As with any energy endeavor, the economics of the system metrics focused on the magnitude of
the investment, and on the cost of the generated energy for use, all in comparison with
alternatives methods for meeting the same objectives, are key. These were discussed in Section
7, and show some promise that the system can provide much of the global energy needs at a
reasonable cost, especially if the cost of transportation to space is reduced to a few hundred $/kg.
This reduction is included in the planning by several national space organizations and private
businesses and is synergistic with the need for low-cost space transport for other commercial
applications..
To move towards sustainable development of the concept, the economic analysis must include
monetization of all externalities, some of them negative and some positive, many of which are
mentioned under the discussion of the environmental and social pillars in this Section.
The magnitude of the needed capital investment, of above $600 billion over about 15 years,
should naturally be considered relative to other alternatives and to global economic conditions.
In such a novel and large project the risks play a key role. Gradual, incremental, introduction of
the system, with careful preparation and monitoring would be necessary.

8.3 The social pillar


Some human impacts
Those related to health and aesthetics are described under the above-discussion of the
Environmental Pillar
Public anxiety due to large number of satellites in orbit; self destruction of failing mirrors to
safe levels would be a way to alleviate this problem somewhat
It is very likely that the project would make significantly positive contributions to
employment, education, and creativity with associated beneficial spinoffs
Generation of adequate energy to about 1/6 of the world population which lacks it will
certainly improve their health and education and improve chances for reducing poverty.
International space stewardship: There is considerable and very justified concern about assuring
internationally fair and safe use of space. This certainly would apply to the massive proposed
space mirrors project, which in the case of NASAs SOLARES proposes to place 80,000 1-km
diameter mirrors in the sky, with all the associated impacts. The Outer Space Treaty ratified by
the UN in 1967 [55] provides the basic framework on international space law, including the
following principles:
the exploration and use of outer space shall be carried out for the benefit and in the interests
of all countries and shall be the province of all mankind;
outer space shall be free for exploration and use by all States;
outer space is not subject to national appropriation by claim of sovereignty, by means of use
or occupation, or by any other means;
States shall not place nuclear weapons or other weapons of mass destruction in orbit or on
celestial bodies or station them in outer space in any other manner;
269

the Moon and other celestial bodies shall be used exclusively for peaceful purposes;
astronauts shall be regarded as the envoys of mankind;
States shall be responsible for national space activities whether carried out by governmental
or non- governmental entities;
States shall be liable for damage caused by their space objects; and
States shall avoid harmful contamination of space and celestial bodies.
As of October 2011, 100 countries and states are parties to the treaty, while another 26 have
signed the treaty but have not completed ratification. The treaty remains, however, very
incomplete and lacks some essential detail and clearly effective enforcement measures.
Amongst the obvious omissions are intellectual property of space research, and space pollution
[56,57]. A much more solid treaty must be developed to ensure internationally fair and safe
deployment and use of the space mirrors project.

9. Conclusions and recommendation


A critical review of the current status of the space mirrors concept was conducted, thin film
aluminized Kapton mirrors were manufactured and optically an mechanically tested to examine
their property changes when exposed to a cryogenic temperature, an economic analysis related to
several applications was performed, and leading issues that must be taken into account in the
sustainability analysis of the concept were described. We add, without analysis, the possibility of
using the space mirrors for shading the earth: if needed, the mirrors, or some of them, could be
turned towards the sun and thus prevent the solar radiation from reaching the atmosphere and earth
area that is then placed in the mirrors shadow.
Our experiments with thin film mirror have shown that the reflectance at 77 K is higher than that at
300 K, about 1.68 fold at the peak solar intensity wavelength of 450 nm. The imperfection surface
density is lower for the cryogenic temperature. The ultimate tensile strength and tensile modulus
increased at the cryogenic temperature, consistent with the behavior of polymers like Kapton.
As in any large energy development endeavor, it is impossible to eliminate all negative impacts, just
to render them tolerable and sustainable, especially relative to other available options. The overall
concept sustainability, especially taking into account the environmental and social impacts and their
associated costs, must be analyzed carefully and quantitatively, and all externalities must be
included in its future evaluations.
Without consideration of these externalities, our economic analysis agrees with NASAs Ehrickes,
published in the late 1970-s [18-22], that if transportation costs to mirror orbit are reduced to a few
hundred $/kg, as planned, the use of orbiting space mirrors for providing energy to earth is an
investment with a good rate of return and a cost effective alternative to satellite solar power stations
(SPS) and terrestrial renewable and conventional power sources.
This energy concept is very appealing relative to other options for addressing the severe energy and
global warming problems that we face, and deserves much and urgent R&D attention.

Acknowledgments
My former students Kamal Shair, Betsy Rosenblatt, and Travis Schlegel have conducted the
experiments presented in Section 6.2 and contributed to other parts of this paper. I am grateful to
William P. Gilbreath and Stuart W. Bowen, authors of some the seminal NASA papers on space
mirrors, for their answers to my questions. Iris Chu and Eri Mizukane have been very helpful in the
literature search for this study.

270

References
[1] Lior, N. The ECOS 2009 World Energy Panel: An introduction to the Panel and to the present
(2009) situation in sustainable energy development, Energy 36 (2011) 3620-3628.
[2] Lior, N. Sustainable energy development: A brief introduction to the present (May 2011)
situation, accepted for publication in Energy, 2012,
[3] Glaser, P.E. Power from the Sun: Its Future, Science, 162 (1968) 857-861.
[4] Glaser, PE, Davidson, FP, Csigi, KI, editors. Solar power satellites, the emerging energy option.
New York: Ellis Horwood, 1993.
[5] Mankins JC. Space solar power: a major new energy option? Journal of Aerospace Engineering
2001;14(2):38-45.
[6] Mankins, J.C. (USA), Kaya, N. (Japan), Editors. Space Solar Power: The First International
Assessment of Space Solar Power: Opportunities, Issues and Potential Pathways Forward.
August 2011, International Academy of Astronautics, 6 rue Galile, BP 1268-16, 75766 Paris
Cedex 16, France, ISBN/EAN 978-2-917761-11-3
[7] Criswell DR, Thompson RG. Data envelopment analysis of space and terrestrially-based large
scale commercial power systems for earth: a prototype analysis of their relative economic
advantages. Solar Energy 1996;56:119-31.
[8] Brown WC. The history of wireless power transmission. Solar Energy 1996;56:4-21.
[9] Woodcock, G.R. Solar Satellites - space key to our power future. Astronautics and Aeronautics,
Vol. 15, July- Aug 1977, 30-43
[10] NASA, Solar Power Satellite Concept Evaluation. vol. 1, NASA JSC-12973, July 1977
[11] Lior, N. Power from space. Energy Conversion & Management J. 2001;42(15-17):1769-1805.
[12] Tarlecki Jason, Lior Noam, Zhang Na. Analysis of thermal cycles and working fluids for
power generation in space. Energy Conversion & Management2007;48:2864-78.
[13] Zidanek, A., Ambroi , M. Milfelner, M. Blinc, R. Lior, N. Solar orbital power: Sustainability
analysis. Energy 36 (2011) 1986-1995.
[14] National Space Society, http://www.nss.org/settlement/ssp/index.htm [accessed 2/19/2012].
[15] Oberth, H. Wege zur Raumschiffahrt. Berlin: Oldenburg Verlag, 1929 (Translation, NASA TTF-662).
[16] Buckingham, A.G, Watson, H.M. Basic Concepts of Orbiting Reflectors," J. Spacecraft, vol. 5,
May 1968, pp. 852-853.
[17] Billman, K.W., Gilbreath, W. P., Bowen, S.W. Introductory Assessment of Orbiting Reflectors
for Terrestrial Power Generation. NASA TM X-73,230, 1977.
[18] Billman, K.W., Gilbreath, W. P., Bowen, S.W. Satellite Mirror Systems for Providing
Terrestrial Power: System Concept. Paper AAS-77-240, 23rd Meeting of the AAS, San
Francisco, Calif., Oct. 1977.
[19] Billman, K.W., Gilbreath, W. P., Bowen, S.W. SOLARES-A New Hope for Solar Energy?
Alternative energy sources; Proceedings of the Miami International Conference, Miami Beach,
Fla., December 5-7, 1977. Volume 1. (A79-33984 13-44) Washington, D.C., Hemisphere
Publishing Corp., 1978, p. 233-255. IISSN: 0278-1662
[20] Gilbreath, W. P, Billman, K.W. Bowen, S.W. Enhanced solar energy options using earthorbiting mirrors. Proc. 13th Intersociety Energy Conversion Engineering Conf. San Diego,
California August 20 - 25, 1978, Paper 789186, Vol. 2, 1528-1534 Published by the Society of
Automotive Engineers, Inc. 400 Commonwealth Drive, Warrendale, PA 15096.
[21] Ehricke, Krafft A. Kraftsoletta - eine Industrie-Sonne fur Europa - Moglichkeiten
kostenrealistischer Solarenergie. Astronautik, vol. 3, 1977, 85-87.
[22] Ehricke, Krafft A. Space light: space industrial enhancement of the solar option. Acta
Astronautica Vol. 6. pp. 1515_1633, 1979.
[23]
Krafftehricke.com
Krafft
A.
Ehricke
and
America
in
Space.
[http://www.krafftaehricke.com/ehricke_power_relay_satellite_2.php] [accessed 2/19/2012].
271

[24] Sistach, L. A. The moon orbital mirror.The International Academy of Astronautics, and the
International Institute of Space Law, 54th International Astronautical Congress of the
International Astronautical Federation, 29 Sept. 2003, Bremen, Germany.
[25] Space Regatta Consortium. An open letter to astronomers professional and amateur.
http://src.space.ru/page_30e.htm [accessed 2/19/2012].
[26] Ashurly, M.Z. Space-based mirror reflector of solar radiation. J. Opt. Technol. 67 (6), June
2000 577-579.
[27] S.P. Korolev Rocket and Space Corporation Energia. http://www.energia.ru/english/
[accessed 2/19/2012].
[28] Shpakovsky, N. "Space Mirror." Space Mirror. The TRIZ Journal. <http://www.trizjournal.com/archives/2002/06/e/index.htm>. [accessed 2/19/2012].
[29] Garner, C., Diedrich, D., Leipold, M. A Summary of Solar Sail Technology Developments and
Proposed Demonstration Missions, Jet Propulsion Laboratory report JPC992697.
[30] Freiman, Chana. "Mirror, Mirror, up in Space? - Space Mirror for Providing Light." CBS
Interactive,
3
Dec.
1993.
Web.
10
Dec.
2011.
<http://findarticles.com/p/articles/mi_m1590/is_n7_v50/ai_14719235/>. [accessed 2/19/2012].
[31] Steen, M. (1999, August 2). UK Astronomers Attack Russian Space Mirror Plan.
http://www.planetark.com/avantgo/dailynewsstory.cfm?newsid=1073 [accessed 2/19/2012].
[32] NASA. History: Glenn Contributions to Deep Space 1. May 21, 2008 update, NASA.
<http://www.nasa.gov/centers/glenn/about/history/ds1.html>. [accessed 2/19/2012].
[33] Wong, Wayne. "SBIR Inflatable Concentrator Tested for Space." Aerospace Technology
Innovation. Sept.-Oct. 2001. NASA. http://nctn.hq.nasa.gov/innovation/innovation95/6smallbiz3.html [accessed 2/19/2011]..
[34] Fink, D. Biersack, J. P., Stdele, M. On the performance and lifetime of solar mirror foils in
space. Space Solar Power Review. Vol. 5, pp. 91-100, 1985.
[35] Gulino, D.A. The Survivability of Large Space-Borne Reflectors Under Atomic Oxygen and
Micrometeoroid Impact. NASA Technical Memorandum 88914, Prepared for the 25th
Aerospace Sciences Meeting sponsored by the American Institute of Aeronautics and
Astronautics, Reno, Nevada, January 12-15, 1987
[36] Kutz, Myer. Handbook of Environmental Degradation of Materials. (pp. 479-487). William
Andrew Publishing. 2005.
[37] Griffin, Michael D.; French, J. R. Space Vehicle Design (2nd Edition). (pp. 80, Table 3.5 on p.
87,
p.431).
American
Institute
of
Aeronautics
and
Astronautics.
<http://knovel.com/web/portal/browse/display?_EXT_KNOVEL_DISPLAY_bookid=1597&V
erticalID=0>. [accessed 2/19/2012].
[38] Lichodziejewski, D., Derbs, B., West, J., Reinert, R., Belvin, K., Pappa, R. Bringing an
Effective Solar Sail Design Toward TRL 6. Paper AIAA 2003-4659, 39th AIAA
Joint Propulsion Conference, 2003.
[39] Brown, M.A. Mechanical Design of a Kilometer-Sized Solar Sail AIAA 2010-2586 12 15
April 2010, Orlando, Florida 1010-1023
[40] Datashvili, L., Baier, H.
Building blocks of large deployable precision membrane
reflectors", Proceedings of the 42nd AIAA/ASME Structures, Structural Dynamics and
Materials Conference, AIAA paper no. 2001-1478, April 2001, Seattle
[41] Datashvili, L., Baier, H. Wehrle, E., Kuhn, T., Hoffmann, J. Large Shell- Membrane Space
Reflectors. 51st AIAAlASME/ASCE/AHS/ASC Structures, Structural Dynamics, and
Materials Conference, 18th AIAA 2010-2504 12 15 April 2010, Orlando, Florida.
[42] Solarbuzz, Module prices - Retail Price Summary - February 2012 Update.
http://solarbuzz.com/facts-and-figures/retail-price-environment/module-prices [accessed 2/19/2012].

[43] Koelle, D.E. Cost efficiency as design and selection criterion for future launch vehicles Acta
Astronautica Volume 57, Issues 28, JulyOctober 2005, Pages 623629
[44] Spacex Space Exploration Technologies. Falcon 9 Overview. March 2009.
http://www.spacex.com/falcon9.php [accessed 2/19/2012].
272

[45] REN 21 (Renewable Energy Policy Network for the 21st Century). Renewables Global Status
Report 2011 Update, http://www.ren21.net/default.aspx?tabid=5434 [accessed 2/19/2011].
[46] Nield, Jeff. Growing Food in the Arctic Circle. August 29, 2008.
http://www.treehugger.com/files/2008/08/growing- food- in-the-arctic-circle.php
[accessed
2/19/2012].
[47] Cook, Roberta, Calvin, Linda. North American Greenhouse Tomatoes Emerge as a Major
Market
Force.
Amber
Waves.
April
2005.
http://www.ers.usda.gov/Amberwaves/April05/Features/GreenhouseTomatoes.htm [accessed
2/19/2012].
[48]
World
Bank.
World
GDP.
http://data.worldbank.org/indicator/NY.GDP.MKTP.CD/countries/1W?display=graph
[accessed 2/19/2011].
[49] OECD. Gross domestic product in US dollars. OECD 2012.
http://www.oecdilibrary.org/sites/gdp-cusd-table-2012-1en/index.html;jsessionid=cld88aof9i8.delta?contentType=&itemId=/content/table/gdp-cusdtable-2012-1-en&containerItemId=/content/table/gdp-cusd-table-2012-1en&accessItemIds=&mimeType=text/html [accessed 2/19/2012].
[50] Stockholm International Peace Research Institute (SIPRI). http://milexdata.sipri.org/ [accessed
2/19/2012].
[51] DESERTEC Foundation, Clean power from deserts. http://www.desertec.org/ [accessed
2/19/2012].
[52] Longcore, T., Rich, C. Ecological light pollution. Frontiers in Ecology and the Environment,
Vol. 2, No. 4, May, 2004, 191-198.
[53] Laframboise, J. G., Chou, B. R.. Space Mirror Experiments: A Potential Threat to Human
Eyes. Journal of the Royal Astronomical Society of Canada, 94, 237-240, 2000, June 27.
[54] Chepesiuk R 2009. Missing the Dark: Health Effects of Light Pollution. Environ Health
Perspect 117:A20-A27. http://dx.doi.org/10.1289/ehp.117-a20
[55] U.N. The outer space treaty, http://www.oosa.unvienna.org/oosa/SpaceLaw/outerspt.html
[accessed 2/19/2012].
[56] European Centre for Space Law (ECSL). Space policy documents and useful readings on
regional
and
national
space
legislations.
http://www.esa.int/SPECIALS/ECSL/SEMDTNLTRJG_0.html [accessed 2/19/2012].
[57] Oosterlinck, Rene <Who owns and regulates space and why are space laws necessary?.
European Space Agency (ESA) http://www.youtube.com/watch?v=ADTg5CfNNUA [accessed
2/19/2012].

273

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Numerical parametric study for different cold


storage designs and strategies of a solar driven
thermoacoustic cooler system
Maxime Perier-Muzeta, Pascal Stouffsb, Jean-Pierre Bedecarratsc
and Jean Castaing-Lasvignottesd
a

Universit de Pau et des Pays de lAdour, LaTEP, Pau, France, maxime.perier-muzet@etud.univ-pau.fr (CA)
b
Universit de Pau et des Pays de lAdour, LaTEP, Pau, France, pascal.stouffs@univ-pau.fr
c
Universit de Pau et des Pays de lAdour, LaTEP, Pau, France, jean-pierre.bedecarrats@univ-pau.fr
d
Laboratoire de Physique Et Ingnierie Mathmatique pour lEnergie et lenvironnemeNT (PIMENT),
Ile de la Runion, France, jean.castaing-lasvignottes@univ-reunion.fr

Abstract:
A heat driven thermoacoustic cooler consists of a thermoacoustic engine that converts heat into acoustic
waves, coupled to a thermoacoustic cooler that converts this acoustic energy into cooling effect. These
machines have simple structures without moving parts. The coupling of a solar concentrator and a heat
driven thermoacoustic cooler seems to be an interesting alternative to the electrically driven compression
vapour cycle. As the other solar refrigeration systems, even if the cooling demand generally increases with
the intensity of the solar radiation, one of the major difficulties is to insure a frigorific power supply when
there is no or low solar radiation. In our prototype, in order to guarantee a sufficient cooling capacity to face
to refrigeration loads in spite of the production fluctuations, a latent cold storage has been considered. The
aim of the work presented here is to investigate the behaviour of this key element under several design and
operative conditions. A description of the future prototype is done insisting on the thermoacoustic
refrigeration and the cold storage system. A modelling of the main elements of the prototype is developed.
The results of simulations under real solar radiation as well as a parametric study considering the main
design and operative parameters of the cold thermal storage system are presented.

Keywords:
Solar energy, Solar refrigeration, Thermoacoustic refrigerator, Cold latent thermal storage.

1. Introduction
Energy use for refrigeration has risen sharply in recent years. Nowadays, the major part of this
production is provided by electrically driven vapour compression machines. Globally, the
refrigeration devices consumed roughly 15% of the world electricity production. The forecasts for
the refrigeration indicate an increase in the number of units in operation over the coming years. To
ensure the refrigerating production in the next years while responding to environmental challenges
(emissions of greenhouse gases and ozone layer) the future cooling machines should not be primary
energy intensive and should use environmentally friendly refrigerants [1, 2].
Thanks to numerous possible combinations between solar thermal collectors and heat driven
cooling machines, solar cooling equipment seems to be an interesting alternative [3]. Among these
possibilities, the coupling of a solar concentrator with a heat driven thermoacoustic refrigerator is a
promising technology. Compared to the other refrigeration technologies, heat driven thermoacoustic
machines have several advantages with no moving part and environmentally friendly gas as
working fluid like helium, argon or air [4]. During the last decades, experimental investigations
have been carried out on these machines. They have proven their ability to reach very low
temperatures compatible with the liquefaction of gases such as, natural gas, nitrogen or hydrogen
[5-7]. Other studies have demonstrated the possibility to uses thermoacoustic refrigeration for
higher temperature applications like for food refrigeration [8]. The couplings of thermodriven

274

thermoacoustic refrigerators and solar concentrators have also already been experimentally studied
thanks to low power prototypes [9, 10].
Although the cooling demand is generally higher when the sun is shining, the intermittent nature of
solar energy is one of the major constraints for using solar cooling systems. To meet the timedependency of the primary energy supply and end-use requirement, a thermal energy storage has to
be used. Various configurations of the energy storage can be achieved; storing the hot energy for
continuously supply the refrigeration system or storing the produced cool energy. This energy can
be stored in the form of sensible heat (in liquid or in solid), latent heat of a Phase Change Material
(PCM ), or by chemical reaction. The choice of the storage system depends on various criteria like
the amount of energy that has to be stored or the storage temperature. However for cooling
application, the latent storages have numerous advantages: the technology is well known with a
high energy storage density; the stored and retrieved energy is at a quasi constant temperature
which corresponds to the phase change transition of the PCM [11].
The studied prototype consists in a one kW scale solar thermodriven thermoacoustic refrigerator
supplied in primary energy by a solar dish concentrator. To ensure a low variability of the
availability of refrigerating capacity, a cold latent thermal storage is coupled to the refrigerator.
In the first part of this paper the project and the main parts of its development are presented.
Secondly, the lumped model that has been built is described. Then, the first simulation results for a
long period including days with various solar radiation conditions are presented and discussed.
Finally, the results of the parametric study are presented, considering the above-mentioned different
cold storage design characteristics and storage strategies.

2. Prototype description
The device considered here consists in a solar concentrator, a solar flux modulator, a solar receiver,
a thermoacoustic machine composed firstly of a thermoacoustic prime mover linked to an acoustic
resonator and secondly of a thermoacoustic refrigerator linked to a cool thermal energy storage
(Fig. 1).

Thermoacoustic
refrigerator

Solar
T
receiver cg

Tfr

Loads
Thermoacoustic
prime mover
Cool thermal
storage
Solar
flux modulator

Solar concentrator

Fig. 1. Solar driven thermoacoustic refrigerator experimental plant


275

The solar driven thermoacoustic refrigerator heat fluxes represented in Fig. 2 are described in the
following sections. The direct solar radiation is collected, reflected and concentrated by a parabolic
mirror. M ore details on this element can be found in [12]. A solar flux modulator is placed between
the concentrator and the receiver cavity. This element regulates the solar power entering in the
absorbing cavity; it thus allows controlling the temperature of this latter. The concentrated solar
radiation is collected by an absorber situated in the receiver cavity. This latter transfers
approximately 4 kW to the working fluid (helium at about 4 M Pa), while being at a temperature
close to 500C. M ore details on this component can be found in [13, 14].
Concentrator optical losses
and
modulator losses
Solar receiver
heat losses

Collected
solar radiation

C oncentrated
solar radiation

Generator
heat losses

Heat rejected
at Tam b

Resonat or
l osses

Heat absorbed
at T cold stor age

Refrigerator
heat losses

Aborbed solar
radiation

Heat rejected
at T amb

Acoustic work

Solar concentrator
and
solar flux modulator

Thermoacoust ic
prime mover

Solar recei ver

Thermoacoustic
resonator

Thermoacoustic
refrigerator

Fig. 2. Solar driven thermoacoustic refrigerator heat fluxes


The thermoacoustic cooler is composed of a wave generator, an acoustic resonator and a
thermoacoustic refrigerator. Thank to the solar heat absorbed, the thermoacoustic prime mover
generates a traveling wave with large acoustic power. This power is used into the thermoacoustic
refrigerator to pump approximately 800 W of heat from the cold heat exchanger and reject it at the
intermediate temperature exchanger of the refrigerator. A detailed explanation of the way these
coolers work is given in [15].
To ensure a constant power supply at the cold exchanger, a latent storage is intended to be used.
This later uses encapsulated nodules filled with an eutectic Phase Change M aterial (PCM ). The
storage tank is filled with these nodules and the cold transfer fluid which circulates in the tank
ensures the heat exchange between the PCM and the cold production unit [16]. During the time
where the tank is being cooled, the crystallization of nodules allows the storage of the energy.
To produce a controlled thermal load, an electrical fluid circulation heater is placed between the
storage tank and the cold heat exchanger. The power regulation of this latter allows generating
different consumption and storage strategies.

3. Modelling
To predict the future performance of the prototype a lumped model of the entire plant has been
developed. The time variation of the energy source and the frigorific power consumption imposes a
transient approach. The solar part and the cold storage are treated by simplified transient models.
For the thermodriven thermoacoustic refrigerator, a quasi-stationary approach it used. We describe
the equations and the principal assumptions of this model in the following sections.

3.1. The solar concentrator and the solar flux modulator


The solar concentrator consists in a 8.5 m diameter parabolic mirror. For this prototype, the aperture
of the dish is reduced at 13.5 m by using non reflective bands at the concentrator periphery. Thanks
to the sun tracking system, solar rays are always perpendicular to the collector aperture. The
collecting solar power is calculated as following:
Qsol _ coll ect

Acol lector _ aperture . DNI ,

(1)

276

Where DNI is the Direct Normal Irradiation and Acollector_aperture is the collector aperture area.
Non ref lective band

Useful area

Qsol modul

Qs ol refl
Amodul_opened

Amodul_aperture

Fig. 3. Concentrator aperture

Fig. 4. Modulator geometry

The Thermoacoustic machine, the flow pipes, the frame of the concentrator, etc shadow a part of
the dish (Fig. 3). The sun tracking system is considered to be accurate so that the useful area of the
concentrator remains constant all along the day. A constant and uniform mirror reflectivity is
assumed. According these hypotheses, the reflected solar power is expressed by:
Qsol _ ref lect

Where

Acollector _ us eful . .DNI ,

(2)

Acollector_useful is

Acollector _ useful

the useful area of the concentrator aperture calculated by


Acol lector _ aperture Ashadow , and is the mirror reflectivity.

A uniform solar concentrated flux is assumed in the modulator plan (Fig. 4). It is also considered
that it has no impact if it is totally opened. The solar modulated power is determined by:
Qsol _ mod ul

Qs ol _ ref lect

Amod ul _ apert ur e
Amod ul _ opened

(3)

Where Amodul_aperture is the area of the modulator aperture and Amodul_opened is the area of the
modulator aperture when it is totally opened.

3.2. The solar driven thermoacoustic refrigerator


We consider in this modelling, a uniform wall and fluid temperature in all heat exchangers of the
thermoacoustic device. The temperature of the hot exchanger of the thermoacoustic generator is
also assumed to be equal to the temperature of the solar receiver cavity wall.
Lint_cavity
Qloss_wall

modul

Qloss_or
Qsol_abs

Dint_cavity

Q sol

Dext_cavity

Qoverflow

Qech hel

Lext_cavity

Fig. 5. Absorber cavity and hot exchanger heat


fluxes

Fig. 6. Absorber cavity geometry

The cavity absorbs the solar concentrated flux. Thanks to the hot exchanger, a large part of this heat
flux is transferred to the working fluid, the rest is exchanged with the surroundings (Fig. 5). The

277

cavity walls are considered as an equivalent mass of the hot exchanger. The energy balance of these
elements can be written as:
dU cavity

hg

Qsol _ abs

dt

Ql oss _ cavi ty

Qhel

wall _ hg ,

(4)

Where Qsol _ abs is the absorbed solar power. The absorber cavity absorbs the major part of the solar
power which passes through the solar flux modulator. The rest is the solar loss by overflow. This
loss is due to the diameter difference between the solar focus point and the cavity orifice. If the
diameter of the focal point is larger than the one of the cavity aperture (as it is the case for the
studied prototype, in order to limit the thermal losses through the orifice) a part of the solar
radiation cannot enter in this later. In this model, the solar overflow losses are evaluated considering
a solar overflow rate ( sol_overflow). This rate is equal to a constant value when the solar flux
modulator is totally opened and equal to zero when the modulator is partially closed. The absorbed
solar power is calculated by:
Qsol _ abs

Qsol _ mod ul . 1

overflow

(5)

The cavity exchanges with the surrounding by the orifice, the lateral wall and the back wall. The
cavity thermal losses Qloss _ cavity can be determined by:
Qloss _ cavity

Qloss _ or

Qloss _ lateral _ w all

Qloss _ back _ wal l , (6)

Qloss_ lateral _ wall and Qloss_ back _ wall are calculated considering that radiative heat exchanges are
negligible compared to convective ones from the insulating material surface. The convective heat
exchange coefficient between the orifice and the surrounding is estimated from [17].
The two others cavity losses are determined using the various thermal resistances except the thermal
resistance of contact between the cavity wall and the insulation which is neglected.
Q evac

Qech hel

Qloss_wall
Fig. 7. Heat fluxes in the ambient and cold heat exchangers
The cold exchanger of the refrigerator as well as the heat exchangers at intermediate temperature,
(Fig. 7) transfer on one side with the helium and on the other with a heat transfer fluid. The cold one
is submitted to heat transfer with surroundings while such an exchange is neglected for the other
ones because of the low temperature difference between these exchangers and ambient air.
Applying the first law to the wall and to the heat transfer fluid of these elements gives:
dU wall _ exch
dt
dU fluid _ exch
dt

Q fluid

wall _ exch

Qhel

wall _ exch

Qloss _ exch

,
Qwall

fluid _ exch

m fluid _ exch h fluid _ input _ exch

h fluid _ output _ exch

(7)

Qvis co_ exch

The convective heat transfers between the heat exchanger walls and the fluids are given by:
Q fluid

wall _ exch

fluid wal l .Afluid wall

T fluid

Twall _ exch ,

(8)

Where the heat exchange coefficients ( fluid-wall) are calculated thanks to a correlation for a fluid
flowing across a bank of tubes proposed in [18].
278

The viscous dissipation inside the exchanger is calculated by:


Qvis co _ exch

Pexch

m flui d _ exch

, (9)

flui d

Where Pexch are the pressure drops inside the exchangers. They are calculated thanks to a
correlation for a fluid flowing across a bank of tubes determined by [18] and fluid is the density of
the fluid.
Each equation relative to the exchanger walls has a term of transfer with helium. A stationary model
is used to determine these heat fluxes at each calculation time step of the quasi-stationary model
[19].

3.3. The cool loop and cool thermal energy storage


The cool loop is mainly composed of flow pipes, an electrical load, a pump and a cool thermal
storage.

3.3.1. Flow pipes


The flow pipes insure the connection between the elements. For this model, only the two longer
pipes are considered (between the pump and the cold exchanger and between the cold exchanger
and the cool storage). The temperatures of the fluid and the wall are assumed uniform inside the
pipes. They are also considered perfectly insulated. The pipe walls are considered as an equivalent
mass of fluid. With these hypotheses the energy balance on these elements is:
dU f luid

wall _ pipe

m f luid _ cool _ circuit h fluid _ i nput _ pipe

dt

h flui d _ output _ pipe

Qvisco _ pipe ,

(10)

The viscous dissipations Qvisco _ pi pe , are calculated by (11) with the pressure drop determined by:
Ppipe

V fluid 2
fluid

L pipe
Dint_ pipe

(11)

With K the local coefficients of pressure drop for valves, changes of direction, and
coefficient of pressure drop.

the linear

3.3.3. Circulating pump and electrical fluid circulation heater


We assumed that there is no variation of the internal energy of these elements. They are also
considered adiabatic.
Applying the first law to the heat transfer fluid of these elements under these hypothesis gives:
Qelec _ heater

m fluid _ cool _ circuit h fluid _ output _ heater

h fluid _ input _ heater ,

(12)

Where Qel ec _ heater is the electrical power which feeds the heater.
A constant efficiency ( pump) is considered for the pump. Considering these limitations, the energy
balance of the fluid inside the circulating pump is expressed by:
hfluid _ output _ pump hfluid _ input _ pump

Pcircuit

pump

(13)

fluid

Where Pcircuit are the total pressure drops of the circuit.

3.3.3. The cool thermal energy storage


The model which has been already described in detail in a previous paper [20], considers the
aspects of both the surrounding heat transfer fluid and the phase change material packed inside the
nodules in the charge mode as well as in the discharge mode. Only the main hypotheses and
equations are presented here with a special adaptation to the selected configuration.

279

C
ry

n
tio
sa
l li

mfluid
Tfluid int

So lid

ng

Liqui d

Tnod

C ry st a llisa t io n a f te r a
inc o mpl et e m elt ing

M el tin g a f t er a
inc o mp let e c ry st a llisa t io n

So lid

mfluid
Tfluid out

Cr
ys
tal
lis
a

Liquid

io

sta

In order to simplify the physical model of the latent heat storage, the following assumptions were
made:
The tank is vertical with flow from the bottom to the top for the charge mode and the discharge
mode;
The flow in the tank is axial and incompressible;
Variation of temperature of the heat transfer fluid occurs only along the axial direction, i.e. as
checked experimentally [16], the temperature is independent of the radial position;
The insulation of the tank is considered perfect;
Heat transfer by conduction is neglected in the heat transfer fluid;
Kinetic and potential energy changes are negligible;
The tank is divided in several control volumes according to the height (Fig. 9);
The nodules are considered as exchangers. The energy flux exchanged is proportional to the
difference of temperature between the fluid and the interior of the spherical nodule;
The supercooling phenomenon is taken into account.
The variation of the internal energy of the tank elements (metal wall pipe, flow diffusers, etc..) is
neglected.
The pressure losses are considered for the flow inlet and outlet, the flow diffusers and the nodule
bed thank to local coefficients of pressure drops.

elt
in

Solid

el

ti

Liquid

Fig. 9. Tank control volume.

Fig. 10. State of nodule and way of


crystallisation and melting
Based on these assumptions the energy balance equations in transient state in each control volume
for the heat transfer fluid and each spherical nodule i are:
dU fluid
dt
dU n od ,i
dt

N n od

m fluid h fluid in

h fluid out

Q nod

fluid ,i

Q visc o _ storag e (14)

i 1

mn od

du n od ,i
dt

Q no d

fluid ,i

(15)

With

280

u nod ,i

uref

Cv PCM solid Tnod ,i

Tref

uref

Cv PCM solid Tmelt

Tref

uref

Cv PCM solid Tmelt

Tref

if

xnod ,i

xsolid L

if

L CvP CM liq Tnod ,i Tref

if

xnod ,i

x nod ,i

(16)
Even when the heat transfer fluid temperature is considered uniform in each layer, all the nodules of
each layer do not simultaneously pass through the phase change at the melting temperature Tmelt
during cooling because of the supercooling and the erratic character of the crystallisation. The
nodules can be in different states (non-crystallised, entirely crystallised or partly crystallised)
according to their own value of the beginning of the crystallization (Fig. 10).
Applying the nucleation laws, the number of new crystallisations and the corresponding fluxes can
be calculated at each time t [20].
Considering a nodule of inner radius rint (Fig. 11). Uniform cooling of its surface will result in a
spherically symmetric crystallisation-front, r = rinterface (t) the inner radius of solid PCM , propagating
inwards from r = rint with liquid at Tmelt for 0 r rinterface (t) and solid for rinterface (t) r rint.
Assuming constant thermal properties in each phase, the steady-state solution of the heat conduction
in the solid phase has the form ( is the temperature of the solid PCM ):
r
(t )
1 interface
r
(17)
(r , t ) Tm elt T fluid (t ) Tmelt
rinterface (t )
k PCMsolid
kP CMsolid k PCMsolid rinterface
1
1
kenv
rint
k env
rext
ex trext
The interface conditions here have the standard form:
kP CMsolid

( r, t )
r

r ri nt erface ( t )

Qnod fluid ,i
2
4 rinterface
( t)

(18)

r ext

Liquid
Solid

r int
S ol id

r interface

Liquid

Fig. 11. Crystallization inside a nodule (left) and melting inside the nodule (right).
The determination of Q nod fluid before crystallisation starts and after crystallisation is finished is
done considering the uniform PCM temperature and it is possible to write that the internal energy
variation in the PCM is equal to the flux that leaves the nodule.
The heat transfer coefficient between the nodule and the fluid is determined by a correlation [20]
and so depends on the flow rate and on the fluid temperature.
Supercooling occurs only upon crystallisation but never upon melting. So, all the nodules from each
layer simultaneously pass through the phase change at the melting temperature Tmelt.
According to a simplifying assumption, the melting-front is considered to be concentric (Fig. 11)
and equations for Q nod fluid are the same than for crystallisation. During the melting process, heat is
transfered by natural convection and conduction. Only the heat conduction equation is kept into
consideration but an apparent thermal conductivity [21] is used in order to take the natural
convection into account.
281

4. Results
From the model described in the previous section, numerical simulations of the experimental plant
have been carried out assuming a constant ambient temperature of 20C. The solar radiation
conditions that have been used are the DNI measured between the 1st and the 7th of July 2006 by
PROM ES laboratory at Odeillo (South of France) (Fig. 12). The main parameters that have been
considered for this simulation are summarised in Table 1.
-2

DNI (W.m )

1000
800
600
400
200

Time (h)
0
0

24

48

72

96

120

144

168

Fig 12. DNI


Table 1. Main simulation parameters
Solar concentrator aperture area (m)
Shadow area on the concentrator (m)
Reflectivity of the concentrator
Cavity orifice diameter (m)
Solar overflow rate

13.5
3.5
0.92
0.1
0.1

Hot heat exchanger mass (kg)


Ambient heat exchanger mass (kg)
Cold heat exchanger mass (kg)
Electrical power of the heater (W)
M CP nodule diameter (mm)

30
1
1
350
77

4.1. Standard case


A standard case for design and operating parameter of the cool storage has been considered (Table
2). The simulation results of this case are presented in this section.
Table 2. Main storage tank design and operative parameters
Standard case Variation for parametric study
Tank volume (m3)
0.2
0.15 and 0.25
-1
M asse flow rate (kg.s )
0.1
0.05 and 0.2
PCM melting temperature (C)
-21.3
-18.8 and -26.2
Temperature (C)

500
400
300
200
100

Time (h)
0
0

24

48

72

96

120

144

Fig 13. Temperature of the hot exchanger of the generator


282

168

As shown in Fig. 13, when the sun is shining, the temperature of the hot exchanger of the generator
increases extremely rapidly to reach the temperature threshold for wave generation. Actually the
machine starts only if a sufficient temperature gradient exists between the two heat exchangers of
the engine cycle. According to previous experiences, the starting hot temperature (and also stop
temperature when the exchanger temperature decreases) of the generator has been set at 600 K. The
solar power available is very large compared to the one consumed by the acoustic wave generation
and the thermal losses so that the temperature of the hot exchanger continues to grow rapidly to
reach the regulating temperature. This regulation is ensured by the control of the absorbed solar
radiation thank to the solar flux modulator. When the DNI is important, it can shut up to 45% of the
reflected solar power. At the end of the day or during cloudy periods, the hot exchanger is cooled
due to the power consumed by the acoustic wave generation (if its temperature is higher than 600
K) and due to the radiative and convective heat losses. The hot exchanger behaviour is roughly the
same for each simulated days with variations in function of the solar radiation fluctuations.
Qelec_h eater (W)

400

300

200

100
Time (h)
0
0

24

48

72

96

120

144

168

Fig 14. Electrical power of the fluid heater

Solid fraction

Temperature (C)
So lid f raction
Tfr

0.8
0.6

20
15
10
5
0
-5
-10
-15
-20

0.4
0.2
0
0

24

48

72

96

120

144

Time (h)

-25
-30
168

Fig 15. Solid fraction in the storage tank and temperature of the cold exchanger of the refrigerator

The operative conditions corresponding to the frigorific loads have been set as follows: a constant
heating power if the fluid temperature at its outlet stays lower than -18C and no heating power if
this temperature is higher than this limit. To ensure the cooldown of the storage and the loop
elements the heater is turned off during the first 30 hours (Fig. 14 and 15). During this period, when
the refrigerator works, the temperature of the cold exchanger of the refrigerator decreases to reach a
temperature close to the melting temperature of the PCM (Fig. 15). Due to the adiabatic assumption
and no heating power, the temperature of the storage remains constant during the first night. At the
beginning of the second day (close to 32 hours after the beginning of the simulated
experimentation) the tank temperature continues to decrease below the melting temperature of the
PCM . Due to thermal resistances between the fluid and the nodule and due to the supercooling
283

effect, the temperature of the fluid has to be a few degrees lower than the melting temperature to
initiate the crystallisation. Even if the refrigerator continues to produce a frigorific power, the
temperature of the cold exchanger remains quasi constant after the initiation of the crystallisation
thank to the isothermal behaviour of the nodules during this phase. At the end of the second day, the
majority of the PCM is solid. During the following night, a large part of the solid PCM is melting to
supply the loads. During the next four days, the system reaches a quasi periodically steady state
condition, the storage tank solid fraction evolves between the same levels from one day to the other
and the temperature of the cold loop stays close to the melting temperature of PCM . Due to a low
insolation during the two last days, the solid fraction of the tank decreases and there are periods
where the storage cannot provide the cooling capacity within the imposed temperature condition.

4.2. Parametric study


From the standard case described previously, we consider the effects of the variation of the
principal parameters of the cold thermal energy storage on the temperature of this storage and on
the global performance of the experimental plant. The effect of the supercooling is firstly analyzed.
Then the effect of the storage tank volume, the melting temperature of the PCM and the fluid mass
flow rate, are successively considered (Table 2).
Supercooling effect: A simulation has been made considering no supercooling effect. As expected
the lack of supercooling allows the initiation of crystallization of the PCM at a temperature higher
than when this effect is considered. This temperature difference can be observed (Fig. 16) during
the first crystallization stage. During the next four days, no difference can be observed because a
large number of nodules remain constantly partially crystallized. Because of the total discharge of
the stock during the night preceding, this phenomenon is again perceptible on the last day.
Temperature (C)

20
10

Ti me (h)
0

Standard case
0

24

48

72

96

-10

120

144

168

Without supercooling

-20
-30

Fig. 16. Temperature of the cold exchanger of the refrigerator with and without PCM supercooling

Volume of the storage tank: The variations of the tank volume have significant impact on the cold
exchanger temperature (Fig 17) and the PCM solid fraction (Fig 18). With the smallest tank there
are periods where all the PCM into the tank is solid, thus the residual energy have to be stored as
sensible heat. This generates very low temperatures in the cold circuit. Considering the largest tank,
due to higher mass of fluid and PCM , a larger part of the energy is stored as sensible heat. The solid
fractions are thus in this case lower than the other one. Due to this, with this configuration, there is a
large period of time between the two last days where the storage cannot provide the cooling
capacity.

284

Temperature (C)
20
10

CS
Standard case

0
0

24

48

72

96

120

144

Ti me (h)

= 0.15 m
VVtank
= 0.15m3

168

= 0.25 m
VVtank
= 0.25m3

-10

-20
-30
-40

Fig. 17. Temperature of the cold exchanger of the refrigerator for different tank volumes

Sol id fracti on

1
0.8
0.6

St
Srie4
andard case
3

0.4

VSrie2
ta nk = 0.15 m

0.2

VSrie5
ta nk = 0.25 m

Ti me (h)

0
0

24

48

72

96

120

144

168

Fig. 18. Solid fraction in the storage tank for different tank volumes
Heat transfer fluid mass flow rate: Simulations have been made for different heat transfer fluid
mass flow rates. The obtained results show that in the simulation conditions, this parameter has only
a slight effect on the operation of the cold circuit (Fig. 19).
Temperature (C)

20
10

Ti me (h)
0

Standard case
CS
0

24

48

72

96

-10

120

144

168

-1

mfluid
= 0.05
kg.s
m
= 0.05
kg.s-1

-1

mfluid
= 0.2
kg.s
m
= 0.2
kg.s-1
-20
-30

Fig 19. Temperature of the cold exchanger of the refrigerator for different mass flow rates
Melting temperature of the PCM: For these simulations the thermophysical properties of the
PCM (except the melting temperature) are considered identical. Because of the low temperature
difference between the expected temperature at the heater and the melting temperature of the
highest PCM melting temperature, the major part of the energy is stored as sensible heat (Fig. 20.
and Fig. 21). Due to the difficulty of the refrigerator to reach a low temperature to initiate the
crystallisation of the lowest melting temperature PCM , the variation of the solid fraction in this case
is low. In these two cases the temperature stabilisation which is expected with latent energy storage
is not assured (Fig 21).
285

Solid fraction

1
0.8
0.6

Standar
Standardcas
case
e
0.4

Tm elt==-18C
-18.3C
Tm
Tm elt==-26C
-26.2C
Tm

0.2

Time (h)

0
0

24

48

72

96

120

144

168

Fig 20. Solid fraction in the storage tank for different PCM melting temperatures

Temperature (C)

20
10

Ti me (h)

0
0

24

48

72

96

-10

120

144

168

CSandard case
St
Tm
T
-18.3C
mel =
t =-18C
T
-26.2C
Tm
mel =
t =-26C

-20
-30
-40

Fig 21. Temperature of the cold exchanger of the refrigerator for different PCM melting
temperatures

4. Conclusion
A solar driven thermoacoustic cooler which consists of a thermoacoustic engine that converts solar
energy into acoustic waves, coupled to a thermoacoustic cooler that converts this acoustic energy
into cooling effect, has been studied using a numerical model. The designed prototype would be
able to deliver a refrigerating capacity of about 1 kW at -30C.
In order to guarantee a sufficient cooling capacity to face to refrigeration loads in spite of the
production fluctuations, a cold storage using encapsulated PCM has been considered. It stores the
excess cooling power when the sun is highly shining and restitutes it at night or during cloudy
hours.
The developed model permits to determine the future performances of the prototype but also to
choose the best configuration of the storage tank which has to be adapted to the loads. The main
design parameters for the cool storage are the tank volume, the crystallization and melting
temperature of the PCM and the mass flow rate of the heat transfer fluid.

Acknowledgments
This work was carried out within the framework of the TACSOL project. It is funded by the ANR
PRECODD. We would like to thank our project partners, S. Cordillet (PROM ES), P. Duthil
(IPNO), M.X. Franois (Hekyom), T. Le Polles (Hekyom), G. Olalde (PROM ES), M . Pierens
(IPNO), J.P. Thermeau (IPNO), for their help.

Nomenclature
A
C

area, m
specific heat, J kg-1 K-1
286

DNI
h
k
K
L
L
.

N
P

Direct Normal Irradiation, W m-2


enthalpy, J kg-1
thermal conductivity, W m-1 K-1
local coefficient of pressure drop
length, m
latent heat of the PCM , J kg-1
mass flow rate, kg s-1
number of nodules in the tank layer
pressure drop, Pa

Q
heat flux, W
r
radius, m
t
time, s
T
temperature, K
u
specific internal energy, J kg-1
U
internal energy, J
x
solid fraction of PCM
Greek symbols
convective heat transfer coefficient, W m-2 K-1
emissivity
reflectivity of the concentrator
linear coefficient of pressure drop
density, kg m-3
StefanBoltzmann constant, W m-2 K-4
rate
efficiency
Subscripts and superscripts
abs
absorbed
amb ambient
collect collected
collector
solar collector
env
nodules envelope
exch exchanger
ext
external
fluid heat transfer fluid
hg
hot exchanger of the generator
hel-wall
exchange between the helium and the exchanger wall
int
internal
interface
interface liquid/solid of the PCM
liq
liquid phase of the PCM
loss yhermal losses
melt melting of the PCM
modul modulated or solar flux modulator
287

nod
or
PCM
ref
reflect
sol
solid
visco

nodule
orifice
Phase Change M aterial
reference
reflected
solar
solid phase of the PCM
viscous

References
[1]

J. M . Calm, Emissions and environmental impacts from air-conditioning and refrigeration


systems , International Journal of Refrigeration 2002; 25 (3): 293-305.
[2] C. James M ., The next generation of refrigerants Historical review, considerations, and
outlook , International Journal of Refrigeration 2008; 31(7): 1123-1133.
[3] D. S. Kim et C. A. Infante Ferreira, Solar refrigeration options - a state-of-the-art review ,
International Journal of Refrigeration 2008;31(1): 3-15.
[4] F. Zink, J. S. Vipperman, et L. A. Schaefer, Environmental motivation to switch to
thermoacoustic refrigeration , Applied Thermal Engineering 2010; 30(23): 119-126.
[5] J. J. Wollan, G. W. Swift, S. Backhaus, et D. L. Gardner, Development of a thermoacoustic
natural gas liquefier , presented at the AIChE New Orleans M eeting, New Orleans, 2002.
[6] G. Yu, E. Luo, et W. Dai, Advances in a 300 Hz thermoacoustic cooler system working
within liquid nitrogen temperature range , Cryogenics, vol. 50, no. 8, p. 472-475, Aug. 2010.
[7] J. Y. Hu, E. C. Luo, S. F. Li, B. Yu, et W. Dai, Heat-driven thermoacoustic cryocooler
operating at liquid hydrogen temperature with a unique coupler , Journal of Applied Physics,
vol. 103, p. 104906, 2008.
[8] M . E. Poese, R. W. M . Smith, S. L. Garrett, R. van Gerwen, et P. Gosselin, Thermoacoustic
refrigeration for ice cream sales , in Proceedings of 6th IIR Gustav Lorentzen Conference,
Glasgow, UK, 2004.
[9] J. A. Adeff et T. J. Hofler, Design and construction of a solar-powdered, thermoacoustically
driven, thermoacoustic refrigerator , The Journal of the Acoustical Society of America, vol.
107, no. 6, p. L37-L42, 2000.
[10] C. Shen, Y. He, Y. Li, H. Ke, D. Zhang, et Y. Liu, Performance of solar powered
thermoacoustic engine at different tilted angles , Applied Thermal Engineering, vol. 29, no.
13, p. 2745-2756, Sept. 2009.
[11] A. Sharma, V. V. Tyagi, C. R. Chen, et D. Buddhi, Review on thermal energy storage with
phase change materials and applications , Renewable and Sustainable Energy Reviews, vol.
13, no. 2, p. 318-345, Febr. 2009.
[12] F. Nepveu, A. Ferrire, et F. Bataille, Thermal model of a dish/Stirling systems , Solar
Energy, no. 83, p. 81-89, 2009.
[13] S. Cordillet, P. Duthil, F. Nepveu, T. Le Polles, G. Olalde, A. Salome, et J.-P. Thermeau,
Theoretical proof of concept of an optimal solar receiver to produce low-temperature (40C) cooling using a thermoacoustic tri-thermal machine , in Proceedings of SolarPACES
2010, Perpignan, 2010.
[14] S. Cordillet, P. Duthil, E. Guillot, G. Olalde, et C. Gueymard, Potential effects of rapid
incident flux variations on the design of a solar thermoacoustic receiver , in Proceedings of
SolarPACES 2010, Perpignan, 2010.
[15] G. Swift et Acoustical Society of America., Thermoacoustics: a unifying perspective for some
engines and refrigerators. M elville NY: Acoustical Society of America through the American
Institute of Physics, 2002.
288

[16] J. P. Bdcarrats, J. Castaing-Lasvignottes, F. Strub, et J. P. Dumas, Study of a phase


change energy storage using spherical capsules. Part I: Experimental results , Energy
Conversion and M anagement, vol. 50, no. 10, 2527-2536, oct. 2009.
[17] F. Nepveu, Production dcentralise dlectricit et de chaleur par systme
Parabole/Stirling: Application au systme EURODISH , Universit de Perpignan, Odeillo,
France, 2008.
[18] Y. A. engel, Heat transfer: a practical approach. Boston: M cGraw-Hill, 2003.
[19] M . Perier-M uzet, J. Castaing-Lasvignottes, P. Stouffs, P. Duthil, J.-P. Thermeau, G. Olalde,
S. Cordillet, et M.-X. Franois, M odlisation et simulation transitoire dun refrigerateur
thermoacoustique, in Actes Congrs Franais de Thermique 2010, Le Touquet, (2): 851-856.
[20] J. P. Bdcarrats, J. Castaing-Lasvignottes, F. Strub, et J. P. Dumas, Study of a phase
change energy storage using spherical capsules. Part II: Numerical modelling , Energy
Conversion and M anagement, vol. 50, no. 10, p. 2537-2546, Oct. 2009.
[21] S. Fukusako et M. Yamada, Recent advances in research on melting heat transfer
problems , presented at the 10th Int Heat Transfer Conf, Brighton, Brighton, 1994, 313-331.

289

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON


EFFICIENCY, COST, O PTIMIZATION, S IMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS
JUNE 26-29, 2012, PERUGIA, ITALY

Parabolic Trough Photovoltaic/ Thermal


Collectors. Part 1: Design and Simulation model
Francesco Calisea, Laura Vanoli b
a

DETEC- Univ. of Naples Federico II, P.le Tecchio 80, 80125 Naples, Italy, frcalise@unina.it
DIT Univ. of Naples Parthenope C. D. IS.5, 80143 Naples, Italy, laura.vanoli@uniparthenope.it

Abstract:
Photovoltaic/thermal (PVT) solar collectors are based on a combination of solar thermal and solar
photovoltaic collectors. PVT systems allow one to produce simultaneously electrical energy and thermal
energy by solar irradiation. Different PVT arrangements are presently under investigation. The most common
configuration is the "tube and sheet" one in which a photovoltaic layer is encapsulated in the absorber of a
conventional flat plate solar thermal collector. PVT electrical efficiency may be even higher than the one of a
similar PV system when the fluid average temperature is relatively low. Therefore, the majority of PVT
systems presently under development produce hot streams at temperatures lower than 45 C. However, in
the last few years different types of high temperature PVT systems are also under investigation. In fact, rising
PVT fluid outlet temperature would dramatically increase the range of possible thermal applications. This
paper is focused on this specific technology and presents a design procedure and a simulation model of a
novel concentrating PVT collector. The layout of the PVT system under investigation was derived from a
prototype recently presented in literature and commercially available in order to improve its electrical
performance. The prototype consisted in a parabolic trough concentrator and a linear triangular receiver. The
bottom surfaces of the receiver are equipped with mono-crystalline silicon cells whereas the top surface is
covered by an absorbing surface. The aperture area of the parabola was covered by a glass in order to
improve the thermal efficiency of the system. In the modified version of the collector considered in this paper,
two changes are implemented: the cover glass was eliminated and the mono-crystalline silicon cells were
replaced by triple-junction cells. These modifications allow one to increase significantly the electrical
efficiency of the system especially in case of high operating temperatures. In order to analyze the
performance of the modified Concentrating PVT (CPVT) collector a detailed mathematical model was
implemented. This model is based on zero-dimensional energy balances on the control volumes of the
system. The simulation model allows one to calculate in detail the temperatures of the main components of
the system (PV layer, concentrator, fluid inlet and outlet and metallic substrate) and the main energy flows
(electrical energy, useful thermal energy, radiative losses, convective losses). Results showed that the
performance of the system is excellent even when the fluid temperature is very high (>100 C). Conversely,
both electrical and thermal efficiencies dramatically decrease when the incident beam radiation decreases

Keywords:
PVT, triple-junction, solar energy

1. Introduction
In the last few years, special attention has been paid to the renewable energy sources as a
consequence of the dramatic decrease of the availability of conventional fossil fuels and the related
increase of their costs. Among the available renewable technologies (solar, wind, hydro,
geothermal, etc.), solar energy is commonly considered one of the most viable options. This is due
to large availability of solar radiation for the all over the world and to the recent development of its
conversion technologies. On the other hand, solar energy technology is still suffering of high capital
costs and low power density. However, during the last few years the capital cost of solar collectors
significantly decreased - specially in case of electrical ones - and their cost is expected to further
decrease during the next few years. Conversely, dramatic improvements of the power density are
not realistically expected since this parameter could be increased only by raising the conversion
efficiency of the solar collectors. The solar energy can be exploited for producing both electricity
290

(by photovoltaic collectors, PV) and heat (by thermal solar collectors, SC). From this point of view,
different commercial devices are available since several decades [1-2]. However, a possible
improvement of both PV and SC technologies consists in a combination of their effects. This occurs
in photovoltaic/thermal collectors (PVT) which simultaneously provide electricity and heat. The
basic principle of a PVT collector is simple, since it can be obtained by a conventional thermal
collector whose absorber is covered by a suitable PV layer [3]. The absorbed thermal energy is
distributed to a fluid (typically air or water), whereas the PV produces electricity [1-2]. The final
result of this arrangement is the combined production of electricity and heat and a possible
improvement of PV efficiency. In fact, the PV electrical efficiency is strongly dependent on the
system operating temperature, linearly decreasing with high values of such parameter [4].
Therefore, if the PV layer operating temperature is reduced by a cooling fluid, the system efficiency
will be higher than the one of conventional PV [1-2, 4]. In order to achieve this result, the outlet
temperature of the cooling fluid should be sufficiently low (usually < 40 C). For this reason, the
heat available from PVT systems can be used only for low-temperature heat demand (e.g., domestic
hot water, floor heating, etc.). From this point of view, an interesting application is the desiccant
cooling [5-6]. The selection of the PVT operating temperature is an important key-point in the
system design. In fact, while higher operating temperatures increase the potential use of the
cogenerative heat, they decrease the electricity production [1-2, 4]. As a consequence, researchers
are performing a special effort seeking to realise a PVT collector providing medium-temperature
heat (60 80 C) at high electrical efficiency [7]. A possible alternative for increasing fluid P VT
outlet temperature without decreasing PV electrical efficiency, may consist in the use of a heat
pump (driven by PV electricity) [1-2, 7]. Although the basic idea of the PVT was developed about
40 years ago, this product is still far from a mature commercialization [8]. Thus, several
researchers are investigating several novel P VT arrangements [8-11]. For example, Zhao et al.
investigated a novel PVT where thermal and electrical sections are separated [12]. One of the key
points in the design of a P VT system is the eventual selection of a transparent cover. In fact, the
covering can improve the thermal performance (higher insulation) but reduces the electrical one
(higher reflection). This topic is still under investigation by a number of researchers analyzing
different types of covers (tedlar, DEA, glass-to-glass, etc) [13]. Similarly, different studies
investigated the optimal design of fluid channels. Different options are under investigation: the
cooling fluid may flow between the PV and the absorber or between the absorber and the insulation
(as usual in SC) [1-2]. Usually, the adopted PVT cooling fluid is water [10]. However, several
studies are focused on the use of air [11] or in a combination of air and water [9]. The use of air as
cooling fluid is very attractive in case of building integration. In fact, some companies and
researchers are developing new prototypes of PVT facades or Building Integrated PVT (BIPVT)
[14] which can: i) produce electricity; ii) provide space heating during the winter; iii) shade from
the solar radiation in summer; iv) act as a ventilated faade in summer for reducing the cooling load
[1-2]. In this case, the appropriate selection of PV technology is crucial for maximizing system
overall performance [14]. Anyhow, the majority of the studies regarding P VT are focused on the
development of devices that reduce capital costs and improve the system reliability. In this
framework, a possible system configuration consists in the adoption of Concentrating P VT
collectors (CPVT). Basically, they are simple PVT collectors placed in the focus of some reflectors
(Fresnel, parabolic, dish, etc.) [1-2, 7, 15]. Obviously, the specific cost of this system is
dramatically lower than the flat plate PVT one; this is due to the lower amount of PV employed per
unit area. On the other hand, it must be considered that concentrating solar radiation devices
determine an increase of radiative flux on PV, increasing its operating temperature and therefore
decreasing its electrical efficiency. Usually, this drop off is not too high (typically, 0.45%/K for
silicon cells), if the increase of temperature is fair (a silicon cell having an efficiency of 15 % at 25
C, will show an efficiency of 11.6 % at 75 C). Conversely, for high concentration ratios, the PVT
operating temperature and the corresponding inefficiencies may significantly increase (for silicon
cells, the voltage drops to zero at 270 C) [4]. Therefore, the use of CP VT may be improved
adopting novel P V materials, such as multi-junction solar cells which can approach a nominal
291

efficiency of 40% [15-16]. The adoption of such materials in CPVT may lead to a system operating
up to 240 C at reasonable conversion efficiency (slightly lower than 20%) [15]. The perspective of
using high temperature PVT is very interesting since it extends the number of possible applications.
An example consists in the use of the high-temperature heat provided by the PVT to drive a heat
engine [17] or an Organic Rankine Cycle (ORC) [18] or a Solar Heating and Cooling system [15].
However, commercially or pre-commercially available CPVT systems are typically a small amount
of the PVT under development [8]. Similarly, the availability of theoretical and experimental
studies investigating CP VT performance is scarce. In particular, M ittelman et al [15, 19-20]
performed some experimental and theoretical works dealing with CPVT systems. In reference [20]
presented a novel miniature CPVT based on a dish concentrator (0.95 m2) and a silicon PV cell. A
thermal model for that system was developed in order to predict its performance. The system
provides 140-180 W of electricity and 400-500 W of heat. The cost of the system was evaluated at
2.5 $ per peak electric Watt. The operation at high temperature of similar CPVT systems was also
analyzed both for solar cooling [15] and water desalination [19] applications. A Parabolic Trough
CPVT prototype was experimentally investigated by Coventry [21]. In this work the author pointed
out that one of the major challenges in designing CPVT systems is to achieve an acceptable
radiation flux distribution. The concentration ratio of the system under investigation was 37,
thermal and electrical efficiencies were rated respectively at 58 % and 11 %. In this study the author
also concluded that the thermal efficiency of CPVT is higher than the one of flat plate solar
collectors only when radiation is scarce and/or ambient temperature is low. In fact, flat plate
collectors convert both beam and diffuse radiation, whereas CPVT can only convert the beam one.
Parabolic Trough CPVT were also investigated by Li et al. [22-24] comparing the electrical and
thermal efficiencies of the system varying the PV technology, for different concentration ratio.
Authors concluded that GeAs cells increase electrical efficiency with respect to silicon cells.
However, the thermal efficiency of GeAs results lower than the one of silicon cells. Authors also
pointed out that the cost of unit area of the GeAs is 3067.16 $/m2 versus the 131.34 $/m2 of the
silicon cell. A similar work was performed by Bernardo et al. analyzing a low concentrating
parabolic trough CPVT for the Swedish climate [25]. This study presents a thermal model of the
PVT subsequently validated by experimental data. The system is based on a combination of a one
axis tracking parabolic concentrator and silicon cells. This prototype is distributed by the Swedish
company Absolicon. Additional specific applications of CPVT technology are also investigated in
literature. Xu et al. [26] analyzed the integration of CP VT in the evaporator of an heat pump
system. Al-Alilli et al. [27] investigated the use of a CPVT in a desiccant cooling system. Finally,
Rosell [28] et al. investigated a low concentrating P VT system based on a linear Fresnel receiver.
the system was analyzed by a mathematical model subsequently validated by experimental data.
The rated thermal efficiency was about 60 %.
This paper is focused on the technology of concentrating parabolic trough PVT solar collectors and
it presents a design procedure and a simulation model of a novel concentrating PVT collector. The
layout of the PVT system under investigation was derived from a prototype recently presented in
literature and commercially available [25]. The design of this prototype was modified with the
scope to improve its electrical performance and it consists in a parabolic trough concentrator and a
linear triangular receiver. The bottom surfaces of the receiver are equipped with mono-crystalline
silicon cells whereas the top surface is covered by an absorbing surface. The aperture area of the
parabola was covered by a glass in order to improve the thermal efficiency of the system. In the
modified version of the collector considered in this paper, two modifications are implemented: the
cover glass was eliminated and the mono-crystalline silicon cells were replaced by triple-junction
cells allowing one to increase significantly the electrical efficiency of the system especially in case
of high operating temperature. In order to analyze the performance of the modified Concentrating
PVT (CPVT) collector a detailed mathematical model based on zero-dimensional energy balances
on the control volumes of the system was implemented. The simulation model calculates in detail
the temperatures of the main components of the system (PV layer, concentrator, fluid inlet and
outlet and metallic substrate) and the main energy flows (electrical energy, useful thermal energy,
292

radiative losses, convective losses). The input parameters of the model include all the weather
conditions (temperature, insolation, wind velocity, etc) and the geometrical/material parameters of
the systems (lengths, thermal resistances, thicknesses, etc.). The model was used in order to
evaluate both electrical and thermal efficiency curves related to the beam incident radiation, fluid
inlet temperature and external temperature.

2. CPVT simulation model


The idea of CPVT considered in this study is based on the work performed by Bernardo et al [25,
29] and on the prototype recently commercialized by Absolicon [8, 25, 29]. The CPVT (Fig. 1)
consists in a parabolic trough concentrator, equipped with a one-axis tracking system using the
same operating principle of the solar thermal Parabolic Trough Collectors (PTC) [30-32]. The
collector is horizontal and its axis is North-South oriented, whereas the tracking system follows the
solar azimuth angle. However, while in solar thermal PTC at the focus of the parabola is installed
an evacuated tube for heating the fluid, in the considered CPVT system the focus of the parabola is
equipped with a triangular receiver (Fig. 1). A metallic substrate is used between the circular fluid
channel and external surfaces (PV layer and top surface) in order to promote conductive heat
transfer.

Absorber

Fluid Channel
Substrate
Concentrator
Receiver

PV layer

PV layer

Fig. 1 - CPVT layout

The two sides of the triangle facing the parabolic concentrator are equipped with triple-junction PV
layers, whereas the top side of the receiver is equipped with a thermal absorber. The triangular
receiver includes an inner channel where the fluid to be heated flows. Therefore, the concentrated
solar irradiation is converted simultaneously in electricity by the PV layer and in thermal energy by
the cooling fluid. Note also that the top side of the triangular receiver is capable to absorb both
beam and diffuse radiation, whereas PV layers can only convert the concentrated beam radiation. In
summary, the system is basically the same as the one shown in references [25, 29] differing from
that for two reason: i) there is no covering glass; ii) the P V layer is based on InGaP/InGaAs/Ge
triple-junction solar cells [16]. These two modifications allow one to increase significantly the
electrical efficiency of the system with respect to the values rated in references [25, 29]. In fact, the
covering glass is used to increase the thermal efficiency of the system since it reduces convection
and radiation losses. However, the glass also reduces the radiation incident on the PV layer,
determining a decrease of the electrical efficiency of the system. Then, the triple-junction cells are
significantly more efficient than silicon ones and they are also less sensitive to the variation of the
operating temperature.
293

Although simplified models for the calculation of CPVT performance are available in literature
[33], they cannot be applied to the system under investigation due to the use of concentrating
systems and triple junction cells. Therefore, an appropriate model, based on energy balances, has
been developed in order to design and simulate the CP VT under investigation. This is a 0-D model
since the final scope of this work is to create a new Type to integrate in TRNSYS environment.
Therefore, the model should be sufficiently fast for being used in a quasi-stationary yearly
simulation. Therefore, 1-D models have not been considered since they are too computationalintensive for the scope of that work.
The general assumptions adopted for the model are: thermodynamic equilibrium, steady state,
kinetic and gravitational terms negligible in the energy balances, radiation uniformly concentrated
along PV area. In addition the small thickness of the PV layer and the high conductivity in the metal
substrate allow one to assume negligible temperature gradients in the PV film and in the substrate.
In other words PV and substrate temperature are assumed uniform.
The system was assumed to operate below 100 C, since it is safer for the reliability of PV cells,
although the system can theoretically operate up to 240 C [15]. In this case, the CP VT could drive
a double effect ACH, significantly increasing the overall efficiency of the system. However, this
possibility must still be explored by experimental tests. Therefore, water was assumed as cooling
fluid. Nevertheless, several types of cooling fluids can be implemented in the model. The
thermodynamic and thermo-physical properties of the fluids, namely air and water, were calculated
using the appropriate routine included in TRNSYS.
The concentration ratio is defined as the ratio between the area of the receiver, A PVT, namely the
two PV triangular sides, and the aperture area, A ap, of the concentrator:
C PVT

APVT
Aap

(1)

The optical efficiency ( opt) of the concentrator is assumed being constant [15]. Therefore, the
radiation indent on the PV surface is:
GPVT

APVT Ib CPVT

opt

IAMth

(2)

As usual in concentrating systems, in the previous equation only the beam incident radiation (Ib) is
considered. Such radiation is also corrected considering both the optical efficiency of the receiver
and the Incidence Angle M odifier (IAM ) [34], the last considering that the radiation decreases when
the angle of incidence increases. The IAM , related to the thermal production is evaluated on the
basis of the data experimentally calculated by Bernardo et al.[25, 29]:
60

IAM th 1 b0,th

60 IAM th

1 b0,th

1
cos

1
cos
1

1
1

60
30

(3)

Simultaneously, additional thermal energy is absorbed by the top thermal absorber.


Qtop

Atop I tot

(4)

top

294

In this case, the top surface area, A top, can convert both beam and diffuse radiation, i.e. the total
radiation (Itot) since the insolation incident on that surface is not concentrated.
Assuming the top surface area as gray surface and considering that the area of the top surface is
much lower than the one of the sky, the radiative heat transfer between the top absorber and the sky
can be calculated as follows [34]:
Qtop

sky

Atop

top

Ttop4 Tsky4

(5)

Here, the sky equivalent temperature (T sky ) is calculated using TRNSYS routine. T top is the
temperature of the top surface.
Similarly, assuming the area of the concentrator much larger than the one of the PVT receiver and
assuming both PVT and concentrator as gray surfaces, the radiative heat transfer between the PVT
and the concentrator [34]:
QPVT

conc

APVT

4
TPVT

PVT

4
Tconc

(6)

T PVT and T conc are respectively PVT and concentrator surfaces temperatures.
The convective heat transfer between the PVT and the air is calculated as follows [35]:
Qconv,P VT

AP VT hc,P VT TP VT

Ta

(7)

The convective heat transfer coefficient, hc,PVT, is calculated taking into account that the wind
velocity is typically around 4-5 m/s. Therefore, the convection mechanism is definitively a forced
convection. Therefore, the corresponding heat transfer coefficient is calculated using the following
correlation, relating the Nusselt, Reynolds and Prandtl numbers [35]:
1

Nu

0.664Pr 3 Re 2

(8)

In this equation the characteristic length is the length of the surface in the wind direction, assumed
parallel to the CPVT longitudinal axis, i.e. L tube . The same correlation is used to calculate the heat
transfer coefficient for the forced convection between the top absorber and the air, hc,top. The
corresponding heat flow is [35]:
Qconv,top

Atop hc,top Ttop Ta

(9)

The gross electrical power produced by the PV layer is:


PPVT , gross

CPV T AP VT Ib

opt

PV

IAM el

(10)

Note that this energy is calculated considering the concentrated beam radiation (corrected by the
concentrator optical efficiency and by the IAM coefficient) incident on the PV layer, corrected by
the electrical efficiency of the PV, PV. The electrical efficiency of the triple-junction PV ( PV) is
experimentally related to the concentration ratio and to the temperature [15].
PV

0.298 0.0142ln CPVT

0.000715 0.0000697 ln CPV T


295

TPVT

298

(11)

Note that this equation returns ultra-high values of electrical efficiency, also approaching 40 %, as
usual in III-V PV cells. The IAM el is also evaluated on the basis of the experimental data provided
by Bernardo et al.[25, 29]:
60

IAM el

60 IAM el

1 b0,el

1 b0,el

1
cos

1
cos
1

1
1

(12)

60
30

The net power produced by the system is reduced of the amount of electricity lost in the module
connections and in the inverter, considering the corresponding conversion efficiencies ( mod and
inv) [15].
PPVT ,net

PPVT ,gross

(13)

mod inv

Finally, the heat absorbed by the cooling fluid is:


Qf

mf hout

hin

(14)

In the previous equation, the enthalpies of the inlet and outlet cooling fluid (hin and hout) are
calculated by the thermo-physical property subroutine discussed above.
Therefore, the overall energy balance on a control volume including the entire triangular receiver is:
APV T I bC PVT
APV T I bC PVT

opt

IAMth

opt

IAMth

APV T hc ,PVT TPVT

Ta

Atop Itot
PVT

mf hout

top

Atop

top

hin

4
Ttop
Tsky4

CPV T APV T I b
APVT

PVT

opt

TP4VT

PV

IAM el

4
Tconc

(15)

Atop hc,top Ttop Ta

Note that in this energy balance the left side is representative of the energy flows entering the
control volume, whereas the terms at the right side of the equation are the energy flows exiting the
control volume. Among these terms at the right side, the first one is the useful thermal energy, the
second one is the electrical power produced and all the remaining terms are losses. As discussed in
the following section, this energy balance is dominated by the radiative terms being the convection
losses low, due to the low receiver area, as typical in concentrating solar collectors
A second energy balance considers the control volume including the metallic substrate and the fluid
channel (also including the fluid flowing inside). In this study, this control volume can be
considered as a heat exchanger. In particular, it is here assumed that the temperature of the metallic
substrate is homogeneous along both radial and circumferential directions. In this case, the primary
side of the heat exchanger is at constant temperature equal to the temperature of the metallic
substrate. This assumption can be considered acceptable as a consequence of the high thermal
conductivity of the metallic substrate [15, 19-20, 25, 29] and allows one to develop a 0-D model of
the CPVT. According to the 0-D approach here implemented, the performance of the heat
exchanger can be calculated using the well-known -NTU technique [36]. For the case under
consideration, the NTU number is:

296

1
1

rsub

h fluid

NTU

AHEX
(16)

mf c f

The heat exchange area, A HEX, is the lateral area of the fluid channel. The thermal resistance of the
metallic substrate, rsub, is typically orders of magnitude lower than the one of the fluid. The fluid
heat transfer coefficient, hfluid, is calculated using the following correlation [35]:
4

0.023Re 5f Pr f5

Nu f

(17)

Finally, the heat transfer effectiveness is [36]:


1 e

NTU

(18)

Defined T sub the temperature of the metallic substrate, the energy balance for the considered heat
exchanger is:
mf hout

hin

mf c f Tsub Tin

(19)

Note that, for the given boundary conditions (inlet temperature and mass flow rate, beam and total
radiations and relative angle of incidence, ambient and sky temperature, ambient pressure and wind
velocity), the unknowns are five, namely: PVT temperature, substrate temperature, fluid outlet
temperature, temperature of top receiver surface (facing the sky) and temperature of the
concentrator. Therefore, three further equations, in addition to eqs. (15) and (19), must be
considered.
The third of the required five equations is derived from an energy balance on a control volume
including the PVT layer, and the metallic substrate.
APV T

TPVT Tsub
rPVT sub

m f hout

hin

Atop

Tsub Ttop
(20)
rtop

In other words, the previous equation is showing that the conductive thermal flow coming from the
PVT layer is partly used to increase the outlet temperature of the cooling fluid and partly is
conductively exchanged with the top side of the triangular receiver. Note that the top thermal
resistance, rtop, is the conductive resistance calculated considering both the metallic substrate and
the top absorbing surface included between the fluid channel and the top surface.
A fourth energy balance can be considered with respect to the control volume including the top side
of the substrate and the top surface of the triangular receiver:
Atop

Tsub Ttop
rtop

Atop I top

Atop I top

top

Atop

top

Ttop4 Tsky4

Atop hc,top Ttop Ta

(21)

Finally, the last energy balance considers the control volume including only the parabolic
concentrator.

297

AP VT

PVT

4
TPVT

4
Tconc

Aconc hc ,conc, front Tconc Ta

I tot Aconc

c onc

Aconc

conc ,back

4
4
Tconc
Tsky

(22)

Aconc hc ,conc,back Tconc Ta

In this case, the left side of the energy balance includes the radiative heat transfer with the PVT and
the radiative energy absorbed by the concentrator surface. Conversely, the terms at the right side of
the equations are respectively: the radiative heat transfer of the back surface (external side of the
parabola) of the concentrator, the convective losses at the front surface (facing the sun) of the
concentrator, and the convective heat loss at the back surface.
Eqs (15), (19), (20), (21), (22) are a system of five equations in the above mentioned five
unknowns. This system of equations is highly non linear as a consequence of the radiative terms
included in the energy balances and of the correlations for the calculations of heat transfer
coefficients. This system must be solved by conventional numerical iterative techniques.
Note also that the model discussed above lies on the assumption of steady state. However, this
assumption can be easily removed, simply adding the capacitive terms in the five energy balances
described above. In that case, the algebraic system of equations turns in a system of differential
equations that can be easily solved using the tool included in TRNSYS package.
Table 1- CPVT design parameters
Parameter

S ymbol

Value

Unit

CPVT aperture area

Aap

12

m2

Top absorber area

Atop

0.60

m2

PV layer area
Fluid channel diameter
Fluid specific heat
Rated fluid flow rate
Top surface absorptance
Concentrator absorptance
Back surface concentrator emissivity
Top surface emissivity
PV reflectance
PV emissivity
IAM electrical coefficient
IAM thermal coefficient

APVT
d
cf
mf

0.12
0.03
4.1877
0.15
0.90
0.03
0.30
0.20
0.03
0.20
0.28
0.14

m2
m
kJ/ kg K
kg/s

top
conc
conc
top
PVT
PVT

b0el
b0th

The overall performance of the CPVT is often evaluated using the well-known thermal and
electrical efficiencies, which are conventionally related to the incident beam radiation and to the
collector aperture area:

CP VT , th

CP VT , el

mf hout hin
Aap I b

CP VT APVT Ib opt
Aap Ib

PV

(23)

IAM el

(24)

298

CPVT design parameters are reported in Table 1[15, 19-20, 25, 29]. For the design parameters
assumed in this table, the concentration ratio is 10.

3. Results and Discussion


On the basis of the model of the CPVT discussed in the previous section, the curves of the electrical
and thermal efficiencies were analyzed as a function of the well-known parameter [34].

Tin Tout
2
Ib

Ta

Tf ,avg

Ta

(25)

Ib

The result of this analysis is shown in Fig. 2. Here, it is clearly shown that the correlation between
the thermal efficiency and the above mentioned parameter is very good. In fact, the value of the
correlation index is very high even using a linear interpolating curve. However, it is also clear that
the markers are not perfectly aligned on the interpolating curve. In fact, the thermal efficiency curve
of a solar thermal collector is a second-order polynomial when constant heat transfer coefficients
are considered [34]. Therefore, the deviation of the performance calculated by the model discussed
in the previous section and the interpolating curve is basically due to the fact that, in the proposed
model, the overall heat transfer coefficients is not constant and varies with the CPVT operating
conditions.

el,PVT,

etat,PVT =0.6 3660 4 - 0 .7164 08ics - 0.0381 838ics

0.25

0.6

0.24

0.5

0.23

0.4
0.3

el,PVT

t,PVT

0.7

el,PVT
t,PVT

0.21
0.2

0.1

0.19

0.1

0.2 0.3 0.4 0.5 0.6 0.7


2
(Tf,mean -Ta)/I b [m C/W]

Ib=200 W /m2, eta el,PVT=0.250269 - 0.000399788 Tmean, f

0.22

0.2

0
0

Ib =900 W/m , eta el, PVT=0.247561 - 0.000401822Tmean,f

0.8

0.18
20

40

60

80 100
T mean ,f [C]

120

140

160

Fig. 3 - CPVT electrical efficiency vs fluid


mean temperature

Fig. 2- CPVT electrical and thermal


efficiency

Conversely, there is not any good correlation between the electrical efficiency and the parameter
shown in eq. (25). In fact, the electrical efficiency of the CPVT basically depends on the operating
temperature of the PV layer and on the concentration ratio, as shown in eq. (11). In the case of the
electrical efficiency it is a common use to show this parameter as a function of fluid average
temperature [21, 25, 29]. This graph is shown in Fig. 3 where the linear interpolating curve is also
displayed. Such Figure shows that the electrical efficiency basically linearly depends on the fluid
mean temperature since such temperature directly affects the PV layer temperature. Note also that
the electrical efficiency scarcely depends on the incident beam radiation. In fact, a reduction of the
beam radiation, for a given fluid average temperature, only determines a slight decrease of PV layer
temperature, consequently causing a slight increase in CPVT electrical efficiency. The plots of Fig.
2 and Fig. 3 also show that the efficiency curve of the proposed CPVT is slightly different from the
original model of the Bernardo et al. [25, 29] due to: the use of triple-junction PV and to the
absence of the covering glass. As a consequence, the optical and the electrical efficiencies are
higher. In fact, this last efficiency ranges between 20 % and 25 % for the considered operating
points. The use of a covering glass reduces the amount of radiation available for both electrical and
299

thermal conversions but also reduce thermal losses. Note also that a higher electrical efficiency is
often counterbalanced by a lower value of the thermal efficiency. However, the thermal
performance of the CPVT considered in this work is also very good due to the excellent radiative
properties of the considered surfaces. Therefore, it can be concluded that the modifications in the
layout of the CPVT considered in this work, allow one to increase both thermal and electrical
efficiency with respect to the values published by Bernardo et al. [25, 29]. Obviously, this increase
in efficiency is counterbalanced by a significantly higher capital cost due to the use of triplejunction PV layer. Therefore, the system considered in this work may be considered a suitable
option for the next future when the cost of triple-junction PV is expected to dramatically decrease.
In the following paragraphs a brief parametric study is performed, with the scope to analyze the
performance of the CPVT under different operating condition and varying some of its main design
parameters. The study was performed considering the set of boundary conditions shown in Table 2
This parametric analysis aims at evaluating the effect of the variation of the design parameters on
the CPVT thermal and electrical performance.
Table 2- CPVT boundary conditions
Parameter Value Unit Parameter Value Unit
T in
70
C
wa
5
m/s
T sky
25
C
Itot
1000 W/m2
Ta
25
C
Ib
800 W/m2
pa
101
kPa
0
deg
As mentioned above, the layout of the system under investigation was derived by some prototypes
previously developed and discussed in literature [8, 21, 25, 29] and has been here modified in order
to improve the electrical performance of the system and to achieve a good thermal performance at
operating temperatures up to 90-100 C. To this scope, in the following figures some of the main
geometrical parameters were varied from the initial configuration, in order to assess their impact on
the overall performance of the system. First, the axial length - Ltube - of the CPVT (including the
axial lengths of the triangular receiver and of the parabolic concentrator) was varied in very large
range. Fig. 4 shows both thermal and electrical efficiencies as a function of this parameter. Here, it
is clearly displayed that both efficiencies are very slightly affected by the increase of CPVT length,
showing a slight decrease for higher CP VT length. However, thermal and electrical efficiencies
drops are lower than 1%. Obviously, an increase of CP VT length also determines a proportional
increase of CPVT aperture area and of PV layer area. This linear relationship is clearly shown in
Fig. 5, where both thermal and electrical powers are plotted as a function of the CP VT length. Note
also that the slight decrease of efficiencies shown in Fig. 4 also determines a deviation of the plots
of Fig. 5 from an ideal proportional relationship. In fact, for a CP VT length of 4 m, thermal and
electrical powers are respectively 8320 kJ/h and 4531 kJ/h, whereas at 16 m these values increase
up to 33403 kJ/h and 15178 kJ/h, which are values slightly lower than the ones expected in case of
constant thermal and electrical efficiencies. The reason of this decrease may be explained by the
temperature profiles shown in Fig. 6. Here, it is shown that higher lengths of the CPVT system also
determine a slight increase of all the temperatures. In fact, an increase in the length also causes an
increase of fluid outlet temperature that consequently determines an overall increase of the
temperatures of the system. In particular, the increase in P V temperature determines the decrease of
the electrical efficiency. M oreover, the increase of PV and top side temperatures also determines an
increase of both convective and radiative losses causing the slight reduction of thermal efficiency
shown in the previous figure. Therefore, from this analysis it could be concluded that, for
maximizing the efficiency of the CPVT the length of the system should be small and the
temperature increase of the fluid should be also small. This result is in accordance with the expected
behaviour of the system since an increase of the fluid mean temperature generally determines a
decrease of both thermal and electrical performances.

300

e l,PVT

t,PVT

0.218
0.217
0.216
0.215
0.214
4

12 [m]
Ltube

16

x 10
4.5

16

4
3.5
3

P el
Qt

12

2.5
2
1.5

8
4

Fig. 4 - CPVT thermal and electrical


efficiency vs L tube

8
12
Ltube [m]

16

Q t [kJ/h]

el,PVT

0.219

x 10
20

Pel [kJ/h]

0.22

0.609
0.608
0.607
0.606
0.605
0.604
0.603
0.602
0.601
0.6
20

t,PVT

0.221

1
0.5
0
20

83. 5

Ttop

29.35

TPVT

To

29.3

Tconc

29.25
29.2

79

29.15

12
Ltube [m]

16

0. 2182
0.218
0. 2178

0. 6053
0. 605
0. 6048

0.217
0. 2168
0.01

29.05
70
0

0. 6058
0. 6055

0. 2176
0. 2174
0. 2172

29.1

74. 5

0. 2186
0. 2184

29
20

el,PVT

88

29.4
Tsu b

Tconc [C]

Ttop, To, Tsub, TPVT [C]

92. 5

e l,PVT

0. 6045
0. 6043
0. 604

t,PVT

0.02

0.03
0.04
d [m]

0.05

Fig. 7- CPVT thermal and


efficiency vs channel diameter

Fig. 6 - CPVT Temperatures vs L tube

t,PVT

Fig. 5 - CPVT Electrical and Thermal Power


vs Ltube

0. 6038
0. 6035
0.06

electrical

The variation of the diameter of the fluid channel does not significantly affect the overall
performance of the system, being important only for the internal balance of the system. In fact, Fig.
7 and Fig. 8 clearly show that an increase in channel diameter determines a slight decrease of
electrical and thermal efficiencies. A similar decrease is also observed for both thermal and
electrical powers produced by the CPVT. Conversely a variation in the diameter shows a significant
impact on the temperatures of the layers of the CP VT, as shown in Fig. 9. Here, it is clearly shown
that an increase of the diameter determines a general increase of the temperature of all the solid
layers of the CPVT, whereas the outlet temperature of the fluid does not significantly vary. In fact,
the lower the diameter, the higher the velocity of the fluid also determining a corresponding
increase of its heat transfer coefficient. Simultaneously, the higher the diameter, the higher the heat
exchange area. The two contrasting effects determine the temperature plot shown in Fig. 9.

2.093
2.092

9.3

2.091

9. 29
9. 28
9. 27

2.09

P el

2.089

Qt

2.088

9. 26
9. 25
0.01

Qt [kJ/h]

Pel [kJ/h]

9. 31

86

2.087
0.02

0. 03

0.04
d [m]

0.05

85
84

Fig. 8 - CPVT Electrical and Thermal Power


vs channel diameter
301

To

TPVT
Tc onc

29.22

Tsub

29.2

83
29.18
82
29.16

81

29.14

80
79
0.01

2.086
0.06

29.24
Ttop

Tconc [C]

x 10
2.094

Ttop, To, Tsub, TPVT [C]

x 10
9. 32

0.02

0.03
0.04
d [m]

0.05

29.12
0.06

Fig. 9 - CPVT Temperatures vs channel


diameter

e l,PVT

0. 22

0.215
0. 21
0.205
0.2
0.04

0.08

0.12
Lrec [m]

0.16

9.8

Pel

2.16

9.6

Qt

2.14
2.12

9.4

2.1

9.2

2.08

2.06

8.8

2.04

8.6
0.04

Fig. 10- CPVT thermal and electrical


efficiency vs receiver length

31.5
30.5

82

30

81

29.5

80

29

79
78
0.04

0.08

0.12
L rec [m]

0.63

0.22
0.218

0.16

0.216

el,PVT

0.625

t,P VT

0.62
0.615
0.61

0.214

0.605

0.212

28.5

0.21
0.6

28
0.2

Fig. 12 - CPVT Temperatures vs receiver


length

2.02
0.2

Fig. 11 - CPVT Electrical and Thermal


Power vs receiver length

31

83

0.16

el,PVT

84

0.12
Lrec [m]

0.222

32
Ttop
T pl
TP VT
To
Tc onc

85

T conc [C]

Ttop, To, Tsub, TPV T [C]

86

0.08

Qt [kJ/h]

t,PVT

x 104
2.18

t,PVT

el ,PVT

0.225

x 103
10

P el [kJ/h]

0. 23

0.63
0.625
0.62
0.615
0.61
0.605
0.6
0.595
0.59
0.585
0. 2

t,PVT

0.235

0.6
0.8

1.2 1.4 1.6


Lap [m]

1.8

0.595
2

Fig. 13- CPVT thermal and electrical


efficiency vs aperture length

A further design parameter is the length of each of the three sides of the triangular receiver, Lrec.
The variation of such parameter is very important since affects: the concentration ratio, PV area and
top absorber area. The length of the receiver significantly affects both thermal and electrical
efficiencies, as shown in Fig. 10 displaying that an increase in receiver length determines a decrease
of the electrical efficiency and a simultaneous increase of the thermal efficiency. Similar trends are
also shown for the thermal and electrical powers (Fig. 11). An increase of the length of the receiver
determines a proportional decrease of the concentration ratio, reducing the radiative flow incident
on the PV layer. Such phenomenon determines, as expected, a decrease of P V layer temperature
(Fig. 12). A reduction of the PVT temperature would suggest an increase of the electrical efficiency,
in contrast to the trend shown in Fig. 10. This is due to the fact that the electrical efficiency of the
triple-junction PV layer also depends on the concentration ratio, as shown in eq. (11). In particular,
a reduction of the concentration ratio also determines a decrease of the of the PV electrical
efficiency. This effect is dominant over the decrease of PV temperature, determining the overall
result of a significant decrease of P V electrical efficiency, as shown in Fig. 10. Conversely, the
thermal efficiency increases faster than the electrical efficiency decreases. In fact, the thermal
efficiency increases for two simultaneous effects: i) a lower electrical efficiency makes more heat
available for thermal conversion; ii) an increase of top absorber area improves the utilization of the
total radiation. This is also clear by Fig. 12, showing that an increase of the temperature length
determines an increase of the top surface temperature, causing also an increase of the outlet
temperature of the fluid. Therefore, it may be concluded that the length of the receivers plays an
important role in the design of the CPVT system. In case the goal is the maximization of the
electrical efficiency this parameter should be as low as possible. Conversely, it should very high if
the goal is the increase of the thermal efficiency.

302

92.5

Qt

2.5
10
2

1.5

6
4
0.6

Q t [kJ/h]

Pel [kJ/h]

12

0.8

1.2 1.4
L ap [m]

1.6

1.8

1
2

0.2181

el,PVT

el,PVT

0.2181

t,PVT

0.218
0.218
0.2179
0

2.2

4.4
6.6
rtop [(m 2*C)/W]

8.8

TPVT

To

Tsub

92
70
0

4.4
6.6
2
rtop [(m *C)/W]

8.8

1.6

28.8
2

1.8

2.08
Pe l
Qt

2.06
2.04

9304

9298
0

2.02
2
1.98
2.2

4.4
rtop

6.6
[(m *C)/W]
2

1.96
11

8.8

Fig. 17 - CPVT Electrical and Thermal


Power
vs
top
thermal
resistance

0.5 4
0.5 2
0.5

29.14
2.2

1.2 1.4
Lap [m]

9308

0.5 6</