Developments in Civil Engineering

Vol. 1 The Dynamics of Explosion and its Use (Henrych)

Vol. 2 The Dynamics of Arches and Frames (Henry ch)
Vol. 3 Concrete Strength and Strains (Avram et al.)
Vol. 4 Structural Safety and Reliability (Moan and Shinozuka, Editors)
Vol. 5 Plastics in Material and Structural Engineering (Bares, Editors)
Vol. 6 Autoclaved Aerated Concrete, Moisture and Properties (Wittmann, Editor)
Vol. 7 Fracture Mechanics of Concrete (Wittmann, Editor)
Vol. 8 Manual of Surface Drainage Engineering, Volume II (Kinori and Mevorach)
Vol. 9 Space Structures (Avram and Anastasescu)
Vol. 10 Analysis and Design of Space Frames by the Continuum Method (Kollr and Hegeds)
Vol. 11 Structural Dynamics (Vertes)
Vol. 12 The Selection of Load-Bearing Stuctures for Buildings (Horvth)
Vol. 13 Dynamic Behaviour of Concrete Structures (Tilly, Editor)
Vol. 14 Shells, Membranes and Space Frames (Heki, Editor)
Vol. 15 The Time Factor in Transportation Processes (Tarski)
Vol. 16 Analysis of Dynamic Effects on Engineering Structures (Bata and Plachy)
Vol. 17 Post-Buckling of Elastic Structures (Szabo, Gspr andTarnai, Editors)
Vol. 18 Fracture Toughness and Fracture Energy of Concrete (Wittmann, Editor)
Vol. 19 Pavement Analysis (Ullidtz)
Vol. 20 Analysis of Skeletal Structural Systems in the Elastic and Elastic-Plastic Range (Borkowski)
Vol. 21 Creep and Shrinkage of Concrete Elements and Structures (Smerda and Kfistek)
Vol. 22 Theory and Calculation of Frame Structures with Stiffening Walls (Pubal)
Vol. 23 Time Effects in Concrete Structures (Gilbert)
Vol. 24 Stresses in Layered Shells of Revolution (Kovarik)
Vol. 25 River Intakes and Diversion Dams (Razvan)
Vol. 26 Analysis of Dimensional Accuracy of Building Structures (Vorlicek and Holicky)
Vol. 27 Reinforced-Concrete Slab-Column Structures (Ajdukiewicz and Starosolski)
Vol. 28 Finite Models and Methods of Dynamics in Structures (Henrych)
Vol. 29 Endurance of Mechanical Structures (Nemec and Drexler)
Vol. 30 Shells of Revolution (Mazurkiewicz and Nagorski)
Vol. 31 Structural Load Modeling and Combination for Performance and Safety Evalution (Wen)
Vol. 32 Advanced Analysis and Design of Plated Structures (Kfistek and Skaloud)
Vol. 33 Regular Lattice Plates and Shells (Sumec)
Vol. 34 Combined Ultrasound Methods of Concrete Testing (Galan)
Vol. 35 Steel-Concrete Structures for Multistorey Buildings (Kozk)
Vol. 36 Analytical Methods in Bin-Load Analysis (Drescher)
Vol. 37 Design of Welded Tubular Connections - Basis and Use of AWS Code Provisions (Marshall)
Vol. 38 Fresh Concrete - Properties and Tests (Bartos)
Vol. 39 Stability, Bifurcation and Postcritical Behaviour of Elastic Structures (Pignataro, Rizzi and Luongo)
Vol. 40 Cable-Stayed Bridges (Ito, Fujino, Miyata and Narita, Editors)
Vol. 41 Numerical Analysis of Reinforced Concrete Structures (Avram, Bob, Friedrich and Stoian)
Vol. 42 Fiber-Reinforced-Plastic (FRP) Reinforcement for Concrete Structures: Properties and
Applications (Nanni, Editor)

Advisory Editor to this Series:

Professor Isaac Elishakoff, Center for Applied Stochastics Research, Department of Mechanical
Engineering, Florida Atlantic University, Boca Raton, FL, U.S.A.
FIBER-REINFORCED-PLASTIC
(FRP) REINFORCEMENT FOR
CONCRETE STRUCTURES
Properties and Applications
Edited by
ANTONIO NANNI
The Pennsylvania State University
University Parky PA, U.S.A.

ELSEVIER
Amsterdam - London - New York -Tokyo
1993
 ELSE VIER SCIENCE PUBLISHERS B.V.
Sara Burgerhartstraat 25
P.O. Box 211, 1000 A E Amsterdam, The Netherlands

Library of Congress Catalog1ng-1n-PublIcatIon Data

Fiber-reinforced-plastic (FRP) reinforcement for concrete structures :

properties and applications / edited by Antonio Nannl.
p. cm. (Developments 1n civil engineering ; v. 42)
Includes bibliographical references.
ISBN 0-444-89689-9
1. Reinforced concrete construction. 2. Fiber reinforced
plastics. I. Nannl, Antonio. II. Series.
TA683.F45 1993
624. 1 '8341dc20 93-25585
CIP

ISBN: 0 444 89689 9

1993 Elsevier Science Publishers B.V. All rights reserved

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or
by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written
permission of the publisher, Elsevier Science Publishers B.V., Copyright & Permissions Department,
P.O. Box 521, 1000 AM Amsterdam, The Netherlands.

Special regulations for readers in the U.S.A. -This publication has been registered with the Copyright
Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about
conditions under which photocopies of parts of this publication may be made in the U.S.A. All other
copyright questions, including photocopying outside of the U.S.A., should be referred to the copyright
owner, Elsevier Science Publishers B.V., unless otherwise specified.

No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a
matter of products liability, negligence or otherwise, or from any use or operation of any methods,
products, instructions or ideas contained in the material herein.

pp. 99-114, 423-434: Copyright not transferred

This book is printed on acid-free paper.

Printed in The Netherlands

 V

PREFACE
The idea of preparing this book originated in October of
1991 during the American Society of Civil Engineers (ASCE)
Convention in Orlando, Florida. At that time, it was recog-
nized that the use of fiber reinforced plastic (FRP) compos-
ites for prestressed and non-prestressed concrete reinforce-
ment had moved from the stage of an "exotic subject" to that
of a technology with serious and substantiatable claims for
the advancement of construction materials and methods.
Research and development (R&D) efforts on the subject were
being undertaken world-wide. This included several demonstra-
tion projects. A considerable number of publications was
already available in technical journals and conference pro-
ceedings. Two symposia specifically dedicated to the subject
were in the planning stage (JSCE, Tokyo, Japan in April '92;
and ACI, Vancouver, Canada, in March '93.) Two other sympo-
sia, one eight-month old (ASCE, Las Vegas, Nevada, February
'91) and the other in the planning stage (CSCE, Sherbrooke,
Quebec, October '92,) had FRP reinforcement for concrete as
the major thrust. With all of these activities taking place,
it appeared necessary to offer a comprehensive picture of the
international situation. The idea was that of a book intended
for engineers, researchers, and developers with the objective
of presenting a world-wide cross-section of initiatives,
representative products and significant applications.

Based on personal experience and contacts, I invited the

leaders in this field to contribute a paper. The response to
the invitation was enthusiastic. The book collects 20 contri-
butions subdivided into three parts. Part I (three papers)
introduces FRP reinforcement for concrete structures and
describes general material properties and manufacturing
methods. Part II (four papers) covers a three-continent
perspective of current R&D, design and code implementations,
and technical organizations' activities. Part III (13 papers)
presents an in-depth description of commercially-available
products, construction methods, and applications.
I am grateful to the authors and co-authors for their
collaboration, and to Elsevier Science Publishers B.V. for
publishing the book. It is my hope that this book make a
significant contribution in advancing knowledge and acceptance
of FRP composites for concrete reinforcement.

Antonio Nanni
State College, PA
March 1993
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 3
1993 Elsevier Science Publishers B.V. All rights reserved.

FRP reinforcement for prestressed and non-prestressed concrete

structures
A. Nanni, Ph.D., P.E.
Department of Architectural Engineering, The Pennsylvania
State University, University Park, PA 16802, U.S.A.

Abstract
This paper provides the overall introduction to the subject
of FRP reinforcement for concrete structures. It explains the
organization and contents of the book and outlines a vision
for future work.

1. INTRODUCTION
In the last decade, the use of f iber-reinf orced-plastic
(FRP) composites for reinforcement to concrete members has
emerged as one of the most exciting and promising technologies
in materials/structural engineering. There is a wide range of
potential applications of FRP reinforcement that covers new
construction as well as strengthening/rehabilitation, pre-
stressed as well as non-prestressed members, and prefabricated
as well as cast-in-place construction. The justification and
motivation for this interest in FRP reinforcement appears to
be a world-wide phenomenon with some peculiar geographical
connotations. For example, in Japan, the driving interest
appears to be in construction materials and methods that may
enhance prefabrication, automation, labor savings, and in
general, a cleaner, more efficient construction process. In
North America, the major interest is to find a solution to the
durability problems caused by steel reinforcement corrosion,
particularly in the infrastructure. Europe may have a combi-
nation of all the above, coupled with a keen interest in
strengthening/rehabilitation as a result of its large number
of invaluable historical structures in need of repair.
As different nations have different organizational/ econom-
ical structures, the approach to research and development
(R&D) has also a regional nature. In Japan, general contrac-
tors and fiber manufacturers are forming the alliances neces-
sary for the development of this new technology. In North
America, the construction industry has been totally absent
from the R&D process related to FRP reinforcement. Only
pultruders have had some involvement up to now. Their origi-
nal interest was in the production of FRP structural shapes
for construction applications. Subsequently, FRP
4

reinforcement to concrete has become an extension of that

initial interest. The need to convert U.S. military-oriented
industries to civilian-type applications may have some inter-
esting developments in the near future. In Europe, contrac-
tors were directly involved in the R&D of two of the earliest
and the most successful FRP products. These differences in
priorities and structure among nations (or blocks of nations)
has also resulted in specializations within the wide range of
FRP reinforcement types. With a certain degree of approxima-
tion, it can be said that Japan excels in prestressed rein-
forcement (pre-tensioned type) and multidimensional reinforce-
ment; North America excels in non-prestressed reinforcement
and gratings; and Europe excels in prestressed reinforcement
(post-tensioned type) and bonded plates. In any case, the
success and the adoption of FRP reinforcement in the market
place will depend on the creativity and resourcefulness of
scientists and engineers and in their ability to disengage
their thinking and approach from traditional construction
procedures and systems. In other words, the realization of
the full potential of FRP composites for construction is yet
to come.
As FRP reinforcement for concrete structures moves from the
R&D phase to the demonstration and commercialization phases,
the need for a comprehensive document addressing background,
approach, expectations, products, and uses has emerged. This
book is to serve this need and complement the body of litera-
ture already available on the subject. Presently, the techni-
cal literature consists primarily of papers collected in the
proceedings of specialty symposia [1-5] and conferences of
various professional societies (e.g., Architectural Institute
of Japan (AIJ,) Canadian Society for Civil Engineering (CSCE,)
Federation Internationale de la Precontrainte (FIP,) Interna-
tional Association for Bridge and Structural Engineering
(TABSE,) Japan Concrete Institute (JCI,) Japan Society of
Civil Engineers (JSCE,) Society for the Advancement of Materi-
al and Process Engineering (SAMPE,) Transportation Research
Board (TRB,) etc.) The aim of this paper, the first of 20
collected in the book, is to provide an overall introduction
to the subject, to explain the organization of the work, and
to highlight the specific topics dealt with in the three main
sections of the book.

2. BOOK ORGANIZATION
2.1 Part I - Introductory topics
Part I consists of three papers (including this one.) The
objective of this section is to familiarize the reader with
fundamental concepts, material forms and properties, and
manufacturing methods that relate to advanced composites and
FRP reinforcement, in particular. In terms of nomenclature,
advanced composites encompass all kinds of systems that are a
combination of two or more materials acting in concert and
exhibiting significant mechanical properties [6] For
 5

advanced composites made of continuous, non-metallic, rein-

forcing fibers, fillers, and a resin binder and used as the
reinforcing elements in concrete structures, the term fiber
reinforced plastic (FRP) reinforcement is generally adopted.
In the paper on "Materials and Manufacturing", Bakis
presents a summary with extended bibliography of various
fundamental subjects relative to FRP composites. This de-
scription reviews of the developments that have occurred in
the last half century. The intention is to educate the reader
on: a) material forms and associated terminology; b) composite
constituents classification, grades, and properties; and c)
manufacturing processes. It is likely that some of the
constituent materials (e.g., thermoplastic) and manufacturing
methods (e.g., compression molding) mentioned in this paper
may not find immediate application in the field of FRP rein-
forcement to concrete structures. However, there is a need to
draw from these experiences in order to address construction
industry-related issues without the constraints of traditional
forms and approaches.
The paper on "Properties of FRP Reinforcements" by Bank
reviews all physical and mechanical properties pertinent to
FRP composites for use as reinforcement to concrete. This
description is a necessary background to the reader as it
provides a key to understand what follows in the book relative
to testing procedures, design methods, and specific products
and applications. Some of the important facts that need
recognition are: anisotropy (i.e., directional dependency)
and the relationship (or lack of it) between longitudinal and
transverse properties; stress rupture (i.e., decrease in
static strength under constant stress over time); and stress
corrosion (i.e., accelerated deterioration caused by the
combined effect of load and adverse environment.) For exam-
ple, the longitudinal shear modulus of FRP bars cannot be
computed on the basis of the corresponding elastic modulus and
Poisson's ratio. Furthermore, in traditional reinforced
concrete members, the effects of thermal and moisture expan-
sion of steel reinforcement are negligible. This is certainly
not the case for FRP reinforcement.

2.2 Part II - International perspective

The objective of Part II is to present a collection of four
regional experiences (Canada, Europe, Japan, and U.S.A.)
These presentations include background, R&D efforts in acade-
mia, government and industry, work by professional organiza-
tions, code implementation, design guidelines, and research
needs.
The "Canadian Perspective" offered by Erki and Rizkalla
presents the highly coordinated efforts taking place in this
country. The fact-finding missions organized by CSCE to
Europe and Japan have produced two valuable publications and
set the stage for the Canadian approach to advanced materials
for construction. The creation of a network or umbrella
organization for industry, government, universities, and
6

professional societies interested in advanced materials for

construction is another novel approach being tried in Canada.
The overview of the R&D work in this country seems to indicate
that bridge-type applications have the highest priority.
"FRP Developments and Applications in Europe" by Taerwe
provides the historic perspective of the major accomplishments
in the area of FRP reinforcement to concrete occurred in
Europe since 1974. Some of the products and applications
introduced in this paper are presented in more detail in Part
III of the book (see papers by Wolff and Miesseler; Burgoyne;
and Meier et al.) It is worth noting that a European, coordi-
nated, 4-year project started in 1991. This effort involves
universities, contractors, and industry with the financial
support of Commission of the European Communities.
The "Overview of R&D in Japan" offered by Sonobe points out
the two fundamental events that have spurred the considerable
R&D work on FRP reinforcement undertaken in this country. The
first event is the establishment of the 5-year National
Research Project that, among other topics, includes FRP
reinforcement. The second event is the formation of the
39-member strong industry association known as "CCC Society"
(Association of Composite Materials Using Continuous Fiber for
Concrete Reinforcement) to promote FRP reinforcement for
concrete. Some of the results of the National Research
Project have been collected in a design guide published by
JSCE. In 1993, it is expected that additional conclusions
about this coordinated effort be made public through AIJ.
These results will certainly shape the R&D strategy of Japan
for the next decade.
Dolan summarizes U.S. activities in his paper entitled "FRP
Development in the United States". This fourth and last
regional review covers historical developments, university and
government research, and demonstration projects. In addition,
the paper presents a discussion on mechanics and design of FRP
reinforced concrete members with and without prestressing.
This review is based primarily on U.S. contributions even
though - the author notes - some works are influenced or
supported by international activities. In the U.S.A., the
first consensus-based document on this subject is expected to
be the State-of-the-Art Report being developed by the American
Concrete Institute (ACI) Committee 440 - FRP Reinforcement.
2.3 Part III - FRP Reinforcement Products
Part III is a collection of 13 papers describing FRP
reinforcement products available on the market or in advanced
state of R&D. All papers follow a similar format in that they
present: the primary constituent materials, the manufacturing
method and configuration of the FRP reinforcement, the
physico-chemical and mechanical properties of the reinforce-
ment, the performance in concrete members, the
constructability, and the demonstration projects/applications.
Part III is artificially subdivided into three sub-sections
based on the geometrical configuration of the reinforcement
and its application. The three sub-sections are:
 7

Part IIIa: 1-D Reinforcing Systems (eight papers) deals

with mono-dimensional systems that are suitable for non-
prestressed reinforced concrete (RC) applications and
prestressed concrete (PC) applications. Of the eight
papers, the first two address RC construction and the
remaining six, PC construction. The six papers describing
FRP tendons include pre-tensioned and post-tensioned
(bonded and un-bonded) applications, and, in some cases,
pre-bent shapes for use as shear reinforcement.
Part Ill.b: 2-D and 3-D Reinforcing Systems (three paper)
deals with two-dimensional (one paper) and three-dimension-
al systems (two papers.) At present, all applications
including these systems are RC-type.
Part III.c: External Reinforcing Systems (two papers)
addresses bonded plates (one paper) and wrapping (one
paper,) which are primarily repair procedures.
It may be added that the FRP products presented in Part III
are a significant sample rather than a complete list. In
terms of fiber systems, six papers deal mainly with carbon,
three with aramid, three with glass, and one with polyvinyl
alcohol (PVA.) Hybrid fiber systems are also proposed in one
paper. In terms of resin binders, all FRP products make use
of thermoset resins (i.e., epoxy, polyester, and vinylester)
with the exception of one that is not impregnated (and strict-
ly speaking not FRP.) Insufficient information is available
on the plastic used in FRP reinforcement, in terms of composi-
tion, additives, fillers, etc.

2.3.1 Part III.a: 1-D Reinforcing Systems

The paper on "Glass FRP Reinforcing Bars" by Faza and
GangaRao describes material properties and performance of
glass FRP bars produced in the U.S.A. by four different
manufacturers. These bars, intended for non-prestressed con-
crete members, are pultruded and have surface deformations
made with a strand helically wrapped around the bar. Three
manufacturers have agreed to follow the same common fabrica-
tion standards in order to facilitate product acceptance in
the market place.
The paper on "Vinylon FRP Rod (CLATEC Rod)" by Okazaki
introduces the use of a relatively new fiber material for FRP
reinforcement to concrete. The fiber of reference is a PVA
fiber known in Japan under the generic name of vinylon (in the
U.S.A. the generic term is vinal.) The paper describes
physico-mechanical characteristics that can make vinylon a
desirable choice, particularly in non-prestressed concrete
applications. Mechanical properties and chemical stability
compare favorably with those of other organic fibers and
E-glass. Temperature sensitivity may be a drawback with
respect to fire resistance.
Santoh, in his paper on "CFCC (Carbon FRP Cable)," presents
a detailed description of the constituent materials,
8

manufacturing, quality control, and FRP product properties

related to CFCC. CFCC is a carbon FRP prestressing cable for
primary use in stranded form (7, 19, and 37 wires.) The
properties of the cable presented in the paper include tensile
strength and modulus, conductivity, expansion, creep and
relaxation, fatigue, and durability. Properties relative to
the use of CFCC as concrete reinforcement include shear
capacity, bond to concrete, effect of temperature on bond,
flexibility, anchorage, and strength of a bent profile. Some
demonstration projects relative to pre-tensioned PC construc-
tion are described.
With reference to the same CFCC cable presented by Santoh,
Katou and Hayashida describe "Testing and Applications" of
post-tensioned, fully-grouted PC construction. The first part
of the paper is devoted to laboratory tests on beams with
straight and bent cables subjected to fatigue loading. The
second part of the paper addresses the application of post-
tensioned CFCC three-strands cables in a gateway building.
"Technora, an Aramid FRP Rod" by Noritake et al. outlines
the physico-mechanical characteristics of FRP elements made of
single or multiple pultruded FRP rods for primary use in
prestressing. The rods consist of aramid fibers impregnated
with vinylester resin. Surface deformation by means of a
spirally wound fiber strand is added to provide mechanical
bond with concrete. Several anchorage devices have been
developed for Technora rods, including wedge and bond types,
single and multi-tendon types, metallic and non-metallic
types. The authors recognize that the strength of the pre-
stressing system depends on anchorage type selection. The
paper describes the performance of this FRP cable in pre-
tensioned and post-tensioned PC construction, and introduces
some significant demonstration projects.
Tamura in his paper on "FiBRA" points out that this braided
epoxy-impregnated rod can be manufactured with different fiber
types depending on the intended application. The rod is also
available in a flexible and rigid form. Because the original
development work was based on the use of aramid fibers, FiBRA
is optimized for PC-type applications. In contrast with the
majority of the mono-dimensional FRP elements that are
pultruded, FiBRA is manufactured by braiding. This fabrica-
tion method offers two advantages: a deformed external
surface for mechanical bond with concrete, and efficient large
diameter sizes. Demonstration projects using FiBRA are dis-
cussed.
In the paper on "Glass Fiber Prestressing System", Wolff
and Miesseler describe the FRP prestressing system known under
the trade name of Polystal. This is a post-tensioned system
of glass FRP tendons with polyamide coating for chemical and
mechanical protection. It is stated that the resin matrix
cannot assure protection of the glass fibers from alkali
attack when the strain level in the tendon is above 0.2
percent. The paper describes the several field applications
that have been undertaken over a period of 13 years. It also
emphasizes the sensor technology that is made possible with
 9

the use of FRP reinforcement. This added value may become one
of the driving forces for acceptance in the market place.
"Parafil Ropes for Prestressing Applications" by Burgoyne
describes a prestressing system that is not exactly an FRP
composite material. In fact, fibers are not resin-impregnat-
ed, but simply confined within a polymeric sheath. Such
tendons can only be used for post-tensioned, un-bonded appli-
cations, internally or externally to the concrete member cross
section. The use of an un-bonded system may offer two advan-
tages. First, there is no stress concentration in the tendon
at the location of a crack in the concrete (i.e., no signifi-
cant increase in stress in the tendon due to live load appli-
cation.) Second, un-bonded tendons can easily be re-tensioned
or replaced. The system described in this paper may become
very attractive for repair solutions such as the application
of external post-tensioning to a deficient cooling tower.
2.3.2 Part Ill.b: 2-D and 3-D Reinforcing Systems
Presently, "NEFMAC - Grid Type Reinforcement" described in
the paper by Sugita is the FRP reinforcement type with the
largest use in concrete (over one million square meters.)
Interestingly enough, the major application has been as
replacement of steel wire-mesh in shotcreted tunnel linings,
with the justification of substantial savings in installation
time and labor. NEFMAC is produced by a modified filament
winding process and is available with several fiber types,
namely: glass, hybrid (glass and carbon,) carbon, and aramid.
The paper presents the mechanical characteristics of the FRP
grid as well as its durability. Performance of concrete
members (beams and panels) reinforced with NEFMAC and subject-
ed to various load configurations are then given.
The paper on "Three-Dimensional Fabric Reinforcement" by
Nakagawa et al. describes properties and applications of a
woven triaxial-3D fabric impregnated with epoxy resin. The
use of this reinforcement type is ideal for thin-section
concrete members such as panels for the building envelope.
The paper shows that the triaxial-3D fabric can be effectively
used with a concrete mixture containing short, randomly
distributed fibers. The significance of this paper is its
demonstration of the manufacturing freedom inherent to FRP
composites. This freedom can be used to conceive reinforcing
systems that are totally new with respect to steel reinforce-
ment.
The paper on "New Three-Dimensional FRP Reinforcement" by
Yonezawa et al. describes an FRP system that is composed of
pultruded rods (longitudinal reinforcement) trussed by tapes
(transverse reinforcement.) With the manufacturing method
presented in the paper, it is possible to tailor the shape of
the reinforcement in almost any three-dimensional config-
uration that is needed. This includes beam-type rectangular
and T sections, cylindrical section, and hollow slab. This
type of 3D FRP system targets members of relatively thick
cross section, such as slabs, beams, and columns. The driving
concept for this reinforcement system is the labor savings and
10

the ability to provide different reinforcement properties in

different directions.
2.3.3 Part III.c: External Reinforcing Systems
The paper on "FRP Bonded Sheets" by Meier et al. presents a
strengthening/rehabilitation technique that makes use of FRP
plates externally bonded to a concrete member face. The paper
describes the advantages in terms of constructability and
performance with respect to the use of conventional steel
plates. A further development is in the possibility of
prestressing the FRP plate before bonding. In this way, the
concrete elements have improved capacity both in flexure and
shear. Detailing for the plate-end anchorage appears to have
been resolved.
The paper by Kobatake et al. describes a "Retrofitting
Method" using a combination of carbon unidirectional prepregs
(adhered to the concrete surface) with overlapped epoxy-im-
pregnated carbon filament winding. This method has been
proven successful and has been used in several commercial
projects in the area of industrial chimney and smoke stack
repair. A natural extension of this retrofitting method is
column repair in order to increase shear resistance and
ductility of structures constructed in seismic zones without
altering their existing mass and rigidity. Column repair for
both infrastructure and buildings is a topic of world-wide
interest. It is worth noting that the authors are concerned
with the safe disposal of FRP products.

3. DISCUSSION AND SUMMARY

It is perhaps possible at this point to outline a vision
for the systematic continuation of the development of this new
class of reinforcing materials for concrete termed FRP compos-
ites. A four-dimensional approach is proposed as follows:
Materials/manufacturing. The effort in the formulation of new
or improved constituents for FRP composites has to continue
with concentration on resins. There is a need for resins
specifically designed for the alkaline environment of concrete
and for high strain rates during service conditions (say 60 to
80% up the ultimate elongation of the composite.) For exam-
ple , the utilization of relatively low-cost fibers, such as
E-glass, depends on the level of protection received by the
resin system. In parallel to material formulation, efforts
towards innovative manufacturing methods are needed to reduce
production costs, improve quality control, and realize config-
urations that are unique to FRP composites. The leaders of
this work in materials/manufacturing are outside the construc-
tion industry, namely: materials scientists, chemical engi-
neers , textile engineers, etc. A deeper level of collabo-
ration between teams of experts representing different indus-
tries has to be established.
 11

Physico-mechanical characterization/standardization. Insuffi-

cient understanding of the behavior of FRP composites (e.g.,
anisotropy, creep, stress rupture, stress corrosion) prevents
their full exploitation as reinforcement to concrete. For
example, the mechanism of mechanical bond between FRP rein-
forcement and concrete (which is the fundamental premise to
composite action between the two in both PC and RC type
members) needs to be re-addressed in more fundamental terms in
order to understand: stress concentration in the reinforce-
ment, reinforcement slippage and bursting stress in concrete,
concrete crack width and and spacing, concrete cover and
reinforcement development length requirements, bond deteriora-
tion under repeated environmental or mechanical loads, etc.
Parallel to physico-mechanical characterization, there is a
need for standardization, not necessarily in terms of prod-
ucts, but in terms of testing procedures in order to establish
a level of acceptable performance (and confidence) to satisfy
requirements from designers and owners. FRP composites are a
tremendous opportunity for the construction industry, but they
could create havoc if unscrupulous manufacturers were to
market unsafe products. This danger is real because of the
relative ease of FRP composite production which can be accom-
plished with relatively inexpensive equipment and off-the-
shelf constituents. Standardized testing should eventually
lead to product certification before distribution in the
market.
Design procedures/building code requirements. Design proce-
dures have to be established without altering the principles
of the building codes presently in place. This is not an
insurmountable problem provided that structural performance of
RC and PC members reinforced with FRP materials is understood.
For example, structures using conventional materials have
performance requirements in terms of plasticity (energy
absorption.) Even though FRP reinforcement is inherently
brittle (linear elastic up to failure,) a satisfactory plas-
ticity could be attained with the appropriate design at
several levels, namely: materials (i.e., hybrid FRP reinforce-
ment, concrete confinement, and debonding,) geometry (i.e.,
staggered reinforcement,) and construction (i.e., partial
prestressing.)
Construction methods/construction typologies. It is incon-
ceivable to expect to successfully replace steel reinforcement
while maintaining the existing construction methods and
typologies. The justification for FRP reinforcement cannot be
limited to corrosion resistance or magnetic permeability. In
a sense, it is a problem of revolution rather than evolution.
The practice and the configurations that were successful with
conventional materials become obsolete when new materials are
introduced. To understand this and the willingness to change
may be the real challenges to the construction industry.
12

The 20 papers collected in this book offer the state-of-

the-art on FRP reinforcement for concrete structures. They
point out accomplishments, limitations, and expectations.
This is the body of experience and knowledge that future
developments will be built on.

4. REFERENCES
1 Iyer, S.L. and R. Sen, Editors (1991), "Advanced Composite
Materials in Civil Engineering Structures," P r o c , American
Society of Civil Engineers, New York, NY., 443 pp.
2 JSCE Sub-Committee on Continuous Fiber Reinforcement
(1992), "Utilization of FRP-Rods for Concrete Reinforce-
ment," P r o c , Japan Society of Civil Engineers, Tokyo,
Japan, 314 pp.
3 White, T.D., Editor (1992), "Composite Materials and
Structural Plastics in Civil Engineering Construction," in
Proc. of The Materials Engineering Congress, American
Society of Civil Engineers, New York, NY., pp. 532-718.
4 Neale, K.W. and Labossiere, P., Editors (1992), "Advanced
Composite Materials in Bridges and Structures," P r o c ,
Canadian Society for Civil Engineering, Montreal, Canada,
705 pp.
5 Nanni, A. and Dolan, C.W., Editors (1993), "FRP Reinforce-
ment for Concrete Structures," P r o c , ACI SP-138, American
Concrete Institute, Detroit, MI. (in print)
6 ASM International (1989), "Composites," Engineering Materi-
als Handbook, Vol. 1, Metals Park, OH, 983 pp.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 13
1993 Elsevier Science Publishers B.V. All rights reserved.

FRP Reinforcement: Materials and Manufacturing

Charles E. Bakis
Composites Manufacturing Technology Center
Department of Engineering Science and Mechanics
The Pennsylvania State University
University Park, PA 16802

1. INTRODUCTION
Fiber reinforced polymer (FRP) materials have been the focus
of much attention in the engineering community since the
development of lightweight, high strength, high stiffness fibers
in the 1940's. Due to their initially high cost, these fibers
were designed for and used in high value-added applications
where weight was of primary concern, such as the aerospace and
high performance transportation industries. Based on the 30-50
years of design and service experience with FRPs in these
industries, it is the opinion of many observers that the
understanding and technology for good design with these fibers
has advanced to such a degree that the major issues preventing
the more widespread use of high performance fibers in many other
industries, including construction, are cost and long-term
durability. The durability concerns are not unique to FRPs,
however, and great progress has been made on reducing the cost
of manufacturing FRPs. It is therefore apparent that the
flexibility to tailor the constituents of a composite material
for particular global properties such as corrosion resistance,
stiffness, strength, dimensional stability, magnetic
permeability, electrical conductivity, etc. is a sufficiently
strong attraction to warrant closer examination of this class of
materials by the construction industry. With this added
flexibility comes the necessity of understanding what materials
and material forms are available, what their capabilities and
limitations are, and how they are manufactured. Additionally,
both cost and durability are issues that need to be examined and
understood by the construction industry before any changes in
traditional design procedures are suggested. To help build a
foundation for this process of understanding, this article
provides an overview of traditional FRP materials and
manufacturing techniques. Since most of the individual topics
covered in this article have been subjects of entire books, the
coverage here cannot possibly be comprehensive. However, an
effort has been made to annotate each subject with references
that provide additional details on each topic.
14

2. BACKGROUND
The traditional motivation for using FRP composite materials
has been to efficiently utilize the extraordinary strength and
stiffness properties of small diameter fibers by embedding the
fibers in a relatively ductile polymeric binder, or matrix.
Small diameter fibers achieve their remarkable properties by
virtue of their highly oriented and defect free microstructures.
The ductile matrix enables good transfer of load from one fiber
to the next an attribute beneficial for any type of stress,
but particularly for compressive and shearing stresses. Even in
tensile loading, a ductile matrix allows load to smoothly
transfer from a broken fiber to nearby intact fibers, thereby
decreasing local stress concentrations and allowing for higher
unidirectional composite strengths.
Polymer matrices used in FRPs can be classified as thermoset
or thermoplastic. A thermoset polymer, when cured by the
application of heat, undergoes a chemical change that results in
a substantially infusible and insoluble material. A
thermoplastic polymer, on the other hand, can be repeatedly
softened upon heating and hardened upon cooling [1]. Molecules
of a thermoset polymer are highly cross-linked, while those of
thermoplastic polymers are linear. In comparison with cross-
linked polymers, linear polymers tend to be more ductile and
tough, but more susceptible to degradation by solvents.
Layers of resin-impregnated fibers, sometimes called plies,
laminas or laminae, can be stacked and fused together to form a
laminate. The stacking arrangement of the laminate can be
tailored to achieve desired physical properties in specific
directions. As a general rule, unidirectionally reinforced FRPs
are seldom used because of their weakness in the transverse
direction. A notable exception to this rule is FRP rods, where
only longitudinal strength and stiffness may be needed. Fibers
of various orientations can be combined in a two- or three-
dimensional units by methods such as weaving and braiding.
What is clear from this brief introduction is that the final
properties of FRP composites are highly dependent upon the
"design" of the material. The design, in turn, is limited by
the available materials, material forms, and manufacturing
techniques. The objective of the following discussion is to
elaborate on these important design considerations.

3. MATERIAL FORMS
A summary of basic terminology [1,2] is a good starting point
for an introduction to FRP materials. The single filament forms
the basis for many forms of reinforcement available to the FRP
materials designer. The term "filament" is often used
synonymously with "fiber," but some consider a fiber to be a
filament with a length of at least 100 times the diameter. In
many instances, fibers are essentially continuous. A bundle is
a general term for a collection of essentially parallel
filaments or fibers, and a strand, tow, or end is an untwisted
bundle used as a unit. Although the cost per unit weight of tow
 15

decreases with the number of filaments in the tow, it is more

difficult to obtain good penetration of the matrix resin into
large tows containing several thousand filaments. In contrast
with a tow, a yarn is an assemblage of twisted filaments,
fibers, or strands that is suitable for weaving. A roving is a
number of strands or yarns collected into a parallel bundle with
little or no twist.
Fibers, yarns, rovings, and so forth may be formed into planar
structures by collimating them to form unidirectional tape,
weaving them to form a woven fabric, or simply pressing them
together to form a nonwoven fabric. Undirectional tape is
widely available in widths of 25 mm to 1.5 m [3]. It has the
most anisotropic (direction-dependent) properties of all
reinforcement forms due to the high degree of fiber alignment,
and is less expensive than woven fabric. Woven fabrics are
renown for their ability to be draped over complex contours
because of their shear-deformable fiber architecture. Two basic
weave patterns are shown in Figure 1. The plain weave is the
most highly interlaced, and is therefore the most resistant to
in-plane shear movement. Satin weaves have a minimum of
interlacing and are therefore more easily distorted in shear
[4]. Standard satin
weaves are the four-,
five-, and eight-harness
forms. In the
harness satin form, the
five
warp (lengthwise) tow
passes over four fill
(transverse) tows and
under the fifth,
repeatedly. Most woven

sraH^p ^D^SES
fabrics have slightly
different mechanical
properties in the two
principal directions due
to unequal fiber volume
fractions in those Figure 1. Basic weave patterns:
directions. Woven plain (left); 5-harness satin
fabrics, sometimes called (right). (After Ref. 4.)
broad goods, are available
in widths of up to 1.8 m.
Continuous fibers, yarns, tows, rovings, and chopped or short
fibers (aspect ratios of 100-15,000) can also be arranged
randomly in planar forms called mats. These mats are used
mostly for hand lay-up operations. When held together with a
binder, these nonwoven reinforcements are not as drapable as
woven fabrics. Short fibers less than about 12 mm long can be
injection molded along with a suitable polymer matrix.
When more than one type of fiber is used in the composite, the
composite is called hybrid-reinforced. If the different types
of fibers are contained within a given layer of a laminated
composite, the fibers are said to be commingled. Commingling
also refers to the practice of placing fibrous thermoplastic
matrix parallel to the reinforcing fibers prior to melting and
consolidation [5].
16

Unidirectional tow and the various planar forms of

reinforcement described above may be purchased dry or with resin
already impregnated (prepreg). If dry, the reinforcement must
be impregnated with resin when the final shape is fabricated, as
is done in the so-called wet lay-up process where a thermoset
resin is infiltrated in a liquid form [6]. Thermoset resin
infiltration must eventually be followed by volatile extraction,
consolidation and cure to form a rigid FRP. Volatiles are gases
emitted by prepregs at elevated temperatures and consolidation
is the process of compressing the fibers and matrix to reduce
voids and achieve a desired density. In thermoplastic matrix
composites, the matrix is already polymerized when it is first
introduced into the fibers. Thermoplastic matrix can be
introduced during prepreg fabrication via hot melt, solution,
slurry, dry powder, dry fiber (commingled), film, or surface
polymerization [5,7].
Thermoset prepreg material is preferred by some component
manufacturers because the proper mix of fiber and matrix has
already been ensured by the prepreg supplier and processing
temperatures are lower than those characteristic of high
performance thermoplastic prepregs. To fabricate a thermoset
prepreg, the resin is first infiltrated into the reinforcement
while it is still in the A-stage (soluble and fusible). Next,
the state of cure of the resin is advanced to the B-stage an
intermediate cure state in which the material softens with heat
and is partially soluble and fusible. Planar forms of prepreg
(i.e., thin sheets) are then stored for shipping by winding onto
rolls with interleaved silicone-coated paper to prevent self-
adhesion. Prepreg tow, or tow-preg [8], is simply wound onto a
cardboard core prior to shipping. A B-stage resin can be used
for a period of up to one year if it is stored at a temperature
of -20C. Before working with prepreg material, it must first
be brought to room temperature while still in an air-tight bag
so that moisture does not condense on the material. Extended
storage periods and repeated exposures to room temperatures
leads to loss of tack and fusibility. This attribute is a
notable disadvantage of thermoset prepregs in comparison with
the infinite room-temperature shelf life of thermoplastic
prepregs. Out time, which refers to the total amount of time
that the thermoset prepreg can be left at room temperature and
still be processable, is specified by the prepreg manufacturer.
The proper amount of prepreg tack is such that a layer lightly
sticks to an adjoining layer during the lay-up process. Too
much tack prevents one from repositioning layers, should the
need arise, and too little tack allows the layers to slide on
their own. To accomplish the final cure of a prepreg composite,
the resin is advanced to its final stage, or C-stage, so that it
is practically insoluble and infusible. Curing and
consolidation will be discussed in more detail shortly.
Thermoplastic FRP prepreg is attractive to some component
manufacturers because it has a fast processing time, an
indefinite shelf life, and no refrigeration requirement during
storage. These attributes reduce manufacturing costs by
permitting increased production rates and simplified handling
(there is no need to wait for the material to reach room
 17

temperature before using it and air-tight packing is not

required). Thermoplastic resins can be repeatedly reshaped by
simply reheating, provided the maximum temperature does not
exceed a critical value that permanently damages the material
[9].
Many advanced thermoset polymer matrix materials are semi-
crystalline; i.e., part of the material is amorphous and part is
crystalline upon cool-down from the melt temperature [10]. The
size of the crystallites and overall amount of crystallinity
decreases with higher cooling rates since the molecules do not
have sufficient time to orient themselves in an ordered
structure before the viscosity of the material is too high to
prevent further crystallization. Changes in crystallinity
during service or processing (due to thermal or chemical
exposure) have a dramatic effect on the solvent resistance and
mechanical properties of thermoplastic FRPs, and are therefore
the subject of much attention in the FRP community [9,11].
Thermoplastic polymers can be made tougher and, hence, more
damage resistant than typical thermoset polymers. Elongations
of high performance thermoplastic resins are of the order of 3 0-
100%, as opposed to less than 5% for many thermoset resins. In
composites usage, however, the gain in damage resistance in
thermoplastic FRPs over thermoset FRPs is not as dramatic as one
would expect because of the constraining effect caused by the
presence of relatively rigid fibers in the matrix [9,12].
Thermoplastic FRP materials are generally more expensive than
thermoset FRPs. The expendables and hardware needed to process
them are also relatively more expensive (typical costs for
thermoplastic resins range from several U.S. dollars per kg to
over $220/kg [5,9]). In addition, thermoplastics usually
require higher processing temperatures for consolidation of the
material, which results in higher residual thermal stresses upon
cool-down. The high viscosity and characteristic memory effect
in linear polymers makes the desired entanglement of molecules
within and between layers of FRP relatively difficult to achieve
during the infiltration and lamination processes [9]. The
boardy (stiff and non-tacky) nature of thermoplastic prepregs
complicates draping them over contoured molds, although spot-
welding and in-situ consolidation of thermoplastic tow-pregs or
tapes via localized heating during placement is a way around
this problem [13]. Since the application of thermoplastic
resins to FRPs is a relatively recent happening, many of the
processing disadvantages just mentioned are expected to be
mitigated eventually. In fact, some of the low cost
manufacturing methods to be discussed later are quite easily
adaptable to the use of thermoplastic FRPs.
Commercial unidirectional prepreg tapes are usually 0.08-0.25
mm thick once processed, and have fiber volume fractions of
0.55-0.72, where 0 and 1 correspond to the limiting cases of no
fibers and all fibers, respectively. Commercial woven fabrics
have final thicknesses of 0.17-0.34 mm and fiber volume
fractions of less than about 0.62. In fabrics, looser weave
patterns and less-dense packing of tows is advantageous because
of the added drapability, but the result is less fiber content
and lower strength. An additional performance penalty incurred
18

by the use of fabrics rather than unidirectional materials is

the reduction in compressive stability of fibers caused by the
crimps in the weave pattern. These crimps also act as sources
of out-of-plane stresses that could cause delaminations. A
potential performance advantage of FRP laminates fabricated of
woven fabrics rather than unidirectional tapes is better
compressive strength retention (on a percentage basis) after
low-velocity impact [14]. This is due to the interlocking
effect of the fibers and the associated capabilities of
dissipating impact energy via intralaminar cracks rather than
delaminations.
The thin layers (laminas, laminae, or plies) of woven or
unidirectional tape described above are typically stacked at
specific orientations such that the final layered structure,
known as a laminate, has the desired engineering properties. A
O-deg. ply orientation in a laminate is that in which the fibers
are aligned with some reference direction often the principal
loading direction. A (0)8T laminate has a total of eight plies
arranged in the O-deg. direction. Sometimes the "T" is omitted
in such laminate designations. A balanced laminate is an
arrangement of laminas where there is a lamina oriented at an
angle of +6-deg. for every lamina oriented at --deg. (0<<90).
In a balanced laminate, there is no shear deformation as a
result of a uniaxial stress applied in the O-deg. direction.
Conversely, an unbalanced laminate will undergo shear
deformation as a result of such loading. This phenomenon is
called shear-extension coupling. An example of a three-ply,
balanced laminate is (0/45/-45).
A symmetric laminate is an arrangement of laminas such that
lamina orientations are symmetric about the midplane, or the
plane midway through the thickness. An example of a symmetric
laminate is (45/0/0/45), or (45/0)s in short-hand notation, where
the subscript "s" denotes mirror symmetry. This laminate is
actually unbalanced and symmetric, and would therefore
demonstrate shear-extension and bending-torsion couplings. An
example of a balanced, symmetric laminate would be (0/45)s,
which would have only bending-torsion coupling due to the
different distances of the +45 and -45-deg. plies from the
midplane. A nonsymmetric, balanced laminate such as (0/90/0/90)
displays extension-bending coupling i.e. it bends as a result
of simple uniaxial stress along an edge. If this last example
had unbalanced plies, it would also display shear-extension
coupling.
An orthotropic material is one in which there exists an
orientation along which a uniaxial stress or bending moments can
be applied and no in-plane shear deformation or twisting will
occur. Individual unidirectional and orthogonally woven
laminas, as well as symmetric laminates of 0/90-type
construction, are therefore orthotropic. When no such symmetry
exists, the material is said to be anisotropic. A balanced,
symmetric laminate such as (0/30)s is sometimes called pseudo-
orthotropic since the material responds as an orthotropic
material to in-plane loads and as an anisotropic material to
out-of-plane bending loads. The elastic properties (stiffness,
Poisson's ratio) of this laminate differ in the two principal
 19

directions (0- and 90-deg.) A quasi-isotropic laminate has in-

plane elastic constants which are independent of direction in
the plane of the laminate. Examples of quasi-isotropic
laminates are the (0/45/90/-45)s and (0/60)s arrangements. It
is important to note that despite their isotropy of in-plane
elastic constants, quasi-isotropic laminates have neither in-
plane strength isotropy nor bending stiffness and strength
isotropies.
Multiple groupings of sublaminates are denoted by placing an
integer in the subscript, such as (0/45)2s and (0/45)s2. These
two laminates are distinct arrangements of the (0/45)
sublaminate, as elucidated by their respective long-hand
notations:
(0/45/-45/0/45/-45/-45/45/0/-45/45/0)
and
(0/45/-45/-45/45/0/0/45/-45/-45/45/0).
The in-plane elastic properties of these orthotropic laminates
are identical in each case, but the in-plane strengths, bending
strengths, and bending stiffnesses differ due to the different
stacking sequences. As with the quasi-isotropic laminates, the
reason for unequal strengths in the last example has much to do
with the presence of interlaminar stresses which develop at the
free edges of test coupons fabricated from multi-directional
arrangements of laminas [15,16]. Since the magnitudes of said
edge stresses are highly dependent on ply orientation, it
follows that the strength of specimens tested in the laboratory
depends on the orientation of the edge relative to the plies.
Further information on lamina and laminate stress-strain
relations, notations, and classifications can be found in a
number of general texts on laminated composite materials [17-
21] .
In addition to FRP laminas, other common forms of FRPs are bulk
molding compounds (BMCs), sheet molding compounds (SMCs), thick
molding compounds (TMCs), and stampable thermoplastic composites
[22]. A BMC is a type of premix prepared by kneading resin,
catalyst, powdered mineral filler, reinforcing fiber (chopped
strand), pigment, lubricants, and other additives in batch form,
resulting in a three-dimensionally reinforced material. BMCs
can be compression, transfer, or injection molded (see
discussion later). BMC injection materials are best used in
low-cost applications where durability, dimensional stability,
and good finish are important. Fiber lengths vary from 12 mm to
32 mm, although the longer fibers do not disperse well during
mixing. An SMC is fabricated in a continuous process by
depositing chopped glass fibers (about 12-mm-long) onto a film
of polyester or vinyl ester resin, followed by the deposition
and compaction of another film of resin on top of the fibers.
The sheet, which is usually about 8-mm-thick, is then covered
with a removable carrier film and stored in rolls until the
proper viscosity is reached, at which time it is tack-free and
leather-like in consistency. An SMC can then be compression
20

molded and cured to the desired shape. A structural SMC, or SMC-

C, is made by adding continuous strands during the sheet-making
process. Other ways to produce SMC are by adding randomly
oriented longer fibers (SMC-R) or directionally oriented fibers
(SMC-D). The finish and dimensional stability of SMCs can be
quite excellent, given the proper selection of fillers and
coatings. General purpose SMCs cost between $1.50 and $3/kg, as
of this writing. TMCs are essentially thicker SMCs (up to 50-
mm-thick) which can be compression or injection molded. Fibers
are better distributed in TMCs than in SMCs, and are damaged
less than in the BMC mixing process. Stampable thermoplastic
composites use long fibers uniformly dispersed in the resin.
The fibers may be unidirectional, biaxially woven, random mat,
or chopped. The woven fiber arrangements allow the least
formability, while the chopped fibers allow the most. Stampable
thermoplastic composites can be produced in continuous sheet
stock or in blanks that can be stacked, placed into compression
molds, heated, and formed into the final shape. Due to the
longer fibers, stampable thermoplastic composites are tougher,
stiff er, and stronger than BMC or SMC composites. There are
more than a dozen suppliers of stampable thermoplastic
composites worldwide.
The term "low profile" is applied to BMC and SMC systems that
contain a thermoplastic additive such as polymethyl methacrylate
which reduces shrinkage and surface roughness in the cured
component [23,24]. Low profile additives undergo a phase change
during processing which allows for the formation of microvoids
to compensate for the high shrinkage of polyester and vinyl
ester resins. "Low shrink" grades of BMC and SMC have low
shrinkages but do not have as good a surface finish as low
profile materials. Additional information on these moldable and
stampable FRP forms can be found in [23,25].

4. FUNDAMENTALS OF FRP PROCESSING

Processing of FRP composites requires the application of
specific temperatures and pressures to the material in order to
accomplish several goals: (1) correct fiber orientation; (2)
correct fiber to resin ratio; (3) correct fiber compaction; (4)
low void content; and (5) correct degree of cure [26]. Unless
these goals are met by proper manufacturing methods, FRP
composites can have property variations of several orders of
magnitude. The following discussion focuses on the items 2-5
since proper fiber orientation is initially determined during
one of the manufacturing steps to be discussed separately (such
as hand lay-up, filament winding, pultrusion, etc.). It should
be kept in mind, however, that once the fibers have been
properly oriented, large deformations during processing and high
resin viscosities can synergistically act to displace or "wash
out" fibers.
Low void content, optimal resin content, and good bonding
between matrix and reinforcement are desirable in FRPs because
they lead to better mechanical properties and better resistance
to moisture and elevated temperature environmental degradation.
 21

The sensitivity of strength and stiffness to the bond between

matrix and fibers, for example, is most evident in shear and
transverse tension of the FRP, although bonding is also a factor
in tensile and compressive longitudinal strength [27]. Voids
are most effectively eliminated during processing by applying
compaction pressure while raising the temperature of the FRP and
applying a vacuum. Under these conditions, the viscosity of the
polymer decreases and the mobility of the air pockets increases.
Most high molecular weight polymers used in FRPs are shear-
thinning, which means that their viscosity decreases with higher
rates of deformation [19]. Hence, as the polymer deforms to
accommodate shape changes during consolidation, it tends to flow
easier. Review articles related to consolidation and voids in
FRPs can be found in Refs. 28 and 29, respectively.
Raising the temperature of a resin is not always sufficient for
void elimination in FRPs. For example, since the viscosity of
thermoplastic resin melts are high 103-105 Pas at low shear
rates, as opposed to about 0.1 Pas for a typical liquid
thermoset resin [9], high pressures may also be needed to
infiltrate fiber tows during thermoplastic prepreg manufacture
[7]. In addition, increased temperatures accelerate the onset
of gelation in thermoset FRPs a result of reduced molecular
mobility associated with cross-linking. For reference, a gel
barely yields if probed with a sharp instrument. Hence, there
is a window of opportunity, so to speak, in the processing of
thermoset FRPs during which time all void extraction is to occur
before the matrix becomes excessively viscous. In fact, when
working with B-staged thermoset prepregs, it is common to impose
a moderate-temperature hold or dwell period with constant
pressure to promote proper consolidation before the temperature
is finally raised to a value which causes rapid cross-linking.
This dwell temperature and period coincides with the minimum
viscosity of the resin. The length of time that a catalyzed
thermoset resin system retains a viscosity low enough to be used
in processing is referred to as the pot life or working life.
A typical schedule of temperature and pressure versus time for
a thermoset prepreg is shown in Figure 2. In this case, the
first dwell period is for consolidation and removal of volatiles
released from the resin, and the second dwell period is for
curing. A typical thermoplastic prepreg consolidation schedule
is shown for comparison in Figure 3. The total processing time
for thermoplastic FRPs tends to be less than that for thermoset
FRPs because the former need to be only consolidated, whereas
the latter need to be consolidated and cured. Precise time,
pressure, and temperature relations such as those shown in
Figures 2 and 3 are optimized by the manufacturer for the best
quality FRP and should therefore be followed closely by the user
of the prepreg.
An occasional problem in the processing of thermoset FRPs is
good temperature control throughout the entire volume of the
component. Temperatures in the interior regions of thick
components tend to lag the outer regions during temperature
ramp-up, then tend to exceed the exterior as the exothermic
curing reaction accelerates. One novel way to deal with this
problem is to consolidate relatively thin sublaminates
22

Vacuum = 6 0 9 mm Hg
Pressure = 621 KPa

Resin Cure
177
Q:
Resin h \
< Flow / ! !\
Q: 121
LU
a.
UJ

21
TIME

Figure 2. Thermoset FRP processing schedule.

Consolidation
3 8 0
o

UJ *+/
o: UJ
3
1-
<
Q:
tu
LJ
Q.
2.0 tn
2 ^ L
J
1- 1
1
>^
Nk
1.4 OL
1 \^
0.5
21 1
*-
1
1
5-30 min.

TIME

Figure 3. Thermoplastic FRP processing schedule.

separately a procedure known as staged curing. Following a

cool-down to room temperature, the sublaminates are stacked to
form the thick laminate and the entire laminate is put through
the final stage of the cure schedule [30]. The simultaneous
 23

lay-up and cure method has also been shown to remedy the thermal
runaway problem in thick laminates [31].
Proper resin content or fiber volume fraction of the FRP is
assured by following the manufacturer's guidelines for
processing of prepregs. Such guidelines specify the amount of
resin to be squeezed, or bled, out of the prepreg by some
specific combination of temperature and pressure applied during
processing. Prepregs in which no resin bleed-out is specified
are sometimes called zero net flow material systems. All
thermoplastic prepregs and many thermoset prepregs require no
bleed-out. For structural FRPs, fiber volume fractions are
usually targeted to be between 0.55 and 0.72. The theoretical
maximum volume fraction of fibers with identical diameters
packed in a hexagonal array is 0.91, but this value is never
achieved in practice due to imperfect fiber compaction. Also,
good wetting of the fibers by the matrix is more important to
overall FRP performance than the attainment of the maximum fiber
volume fraction.
Proper degree of cure in thermoset FRPs and proper
crystallinity content in semi-crystalline thermoplastic FRPs can
be assured by using thermal analysis tools such as Differential
Scanning Calorimetry (DSC), Dynamic Mechanical Analysis (DMA),
and Fourier Transform Infrared (FTIR) Spectroscopy. DSC works
by measuring the amount of heat flow needed to raise the
temperature of the sample at a specified rate. As a thermoset
FRP cures, heat is released by the material due to the
exothermic nature of the cross-linking reaction. This energy is
known as the enthalpy or heat of reaction [32]. At higher
temperatures, more cross-linking occurs and therefore more heat
is released, to a certain limit. Direct measurements of the
heat of reaction by DSC can be used to quantify the degree of
cure of a material since the maximum amount of heat released is
a constant at a given cure temperature. Out-of-date prepreg
will have a lower heat of reaction than fresh prepreg. Degree
of crystallinity in semi-crystalline thermoplastics can be
measured via DSC by measuring the heat input required to melt
the crystalline and amorphous phases. These can be
distinguished as the temperature is scanned upward since the
crystallites melt at a lower temperature. By measuring the
energies required to melt each phase, the relative proportion of
each phase can be determined.
DMA is based on the measurement of the complex modulus and the
coefficient of thermal expansion (CTE) of the material as
functions of temperature. Complex modulus is measured by
cyclically loading a small sample of material with a glass probe
and measuring the cyclic strain response. The real part of the
complex modulus, called the storage modulus, corresponds to the
in-phase strain response of the material to the applied stress.
The imaginary part of the complex modulus, called the loss
modulus, relates the 90-deg. out-of-phase strain to the applied
stress. The tangent of the ratio of imaginary to complex
modulus, or tan S, changes in a characteristic manner when the
material changes from a rigid, glassy material to a soft,
rubbery material. This temperature, known as the glass
transition temperature (T ) , depends on the type of FRP, the
24

degree of cross-linking, or the degree of crystallinity.

Changes in the CTE with temperature can also be used to
determine T .
The FTIR technique works by measuring the spectrum of infrared
radiation emitted by an FRP as it cures. Due to the changing
molecular structure of the thermoset polymer during cure, the
spectrum of radiation is altered in a predictable manner, thus
enabling an in-situ determination of degree of cure at a
particular time and temperature. Differential scanning
calorimetry and FTIR spectroscopy can be used by a manufacturer
to determine the quality of prepreg before time is invested in
processing the material.
In glass FRPs, organofunctional alkoxysilanes, commonly known
as silanes, are often used as coupling agents to promote
adhesion between the fibers and matrix [33]. The mechanism of
adhesion promoted by silane is the formation of an interphase
which consists of a relatively open polymeric network with
occasional chemical attachments to the glass. It is thought
that the three-dimensional polysiloxane networks formed are
easily penetrated by the molecules of the matrix polymer so that
an entanglement of the polymer networks is formed in the
interphase region between matrix and reinforcement [34]. It is
not clear at this time whether or not the susceptibility to
moisture of a glass reinforced polymer composite is influenced
by silane couplant [35]. In commercial practice, silanes are
only 1-5% of the "sizing" applied to glass fibers [36]. Other
components of sizing include the handling aids such as carriers,
film-forming polymers, antistatic agents, and lubricants. It is
clear that there are many chemical species present in the
interphase, and much further research needs to be done to fully
understand the composition of the interphase and how it
influences composite properties [27,37].
In carbon FRPs, surface chemical treatments for encouraging
good bonding between fiber and matrix involve the removal of
weak, unordered graphitic layers from the fiber surface,
enhancement of fiber surface area, and implantation of oxygen or
nitrogen species on the surface. The result can be a two to
three fold improvement of lamina strengths [38,39].
Incorporation of tailored interphases to carbon FRPs via surface
treatments, sizings and processing is an idea that is gaining
wider acceptance as a means of optimizing performance for
diverse applications such as static or dynamic loads [40]. This
interphase has distinct properties from the fiber and matrix,
and can therefore be modified in terms of strength or stiffness
to suit the application at hand. Numerous methods for improving
the fiber/matrix bond in carbon FRPs, including liquid-phase and
gas-phase oxidative treatments and anodic oxidation, are
discussed in detail in Ref. 41.

5. CONSTITUENTS
The main issues in the selection of FRP reinforcements for
concrete structures are corrosion resistance, high tensile
strength, low mechanical relaxation, good toughness, high
 25

fatigue resistance, dimensional stability, and particular

electrical and magnetic properties which vary according to the
application. Continuous lengths of relatively stiff and strong
fibers have traditionally been the most effective means of
reinforcing polymeric materials for strength- and stiffness-
critical applications. For this reason, the following
discussion on constituents will focus on the various types of
continuous fibers and matrices that seem likely at this time to
be useful as FRP reinforcement for various applications in the
construction industry.

5.1. Reinforcements
As the principal source of strength, stiffness, and dimensional
stability in FRPs, fiber reinforcements often are the starting
point in the design of a composite material for a specific
application. Commercially available fibers vary widely in their
ability to meet particular design needs. For example, many
glasses tend to become weaker in the presence of alkaline and
humid environments, but they are quite tough. Carbon fibers can
have high stiffness and resistance to chemical attack, but they
tend to also have low toughness and low impact resistance.
Aramid fibers are less dense than carbon fibers, have a
toughness similar to that of glass fibers, and have a lower
Young's modulus than carbon fibers. In addition, the magnetic
permeability of aramid fibers, which may be of interest to those
concerned with the fabrication of guideways for magnetically
levitated and driven vehicles, is the smallest among the
aforementioned reinforcements. Other factors to consider in the
selection of fibers are electrical conductivity, coefficient of
thermal expansion, upper use temperature, radiation resistance,
and cost. The following discussion summarizes attributes of
several of the more widely used fibers for FRPs.

5.1.1. Carbon Fibers

Most structural carbon and graphite fibers are based on the
layered graphene (hexagonal) networks present in graphite [42].
The graphene layers are ordered in two dimensions in carbon and
three dimensions in graphite. The production of commercial
carbon fibers begins with organic precursors such as rayon,
polyacrylonitril (PAN), and isotropic and liquid crystalline
pitches. Depending on variables such as precursor, heat
treatment, and degree of stretch during processing, the strength
and elastic constants of carbon fibers can vary greatly [19].
Highly oriented microstructures of carbon fibers result in an
increased degree of anisotropy such that the stiffness and
strength of the fibers are higher in the longitudinal direction
than in the transverse direction. The general steps in
producing carbon fibers are as follows: (1) heat the fibers in
air to stabilize the precursor, which prevents melting or
fusion; (2) carbonize the fibers by heat treatment in an inert
atmosphere to eliminate the non-carbon elements; and (3)
"graphitize" the fibers at high temperatures (upwards of 3 000C)
to improve the microstructure of the fiber for enhanced
stiffness and strength. Actually, the fiber is not fully
"graphitized" since this would result in inferior properties.
26

The term "graphite fibers" has been used perhaps to describe the
turbostratic (2-dimensional) structure of high-modulus carbon
fibers [41].
Carbon fibers manufactured from rayon and pitch precursors are
more expensive than those manufactured from PAN precursor
because of the need to stretch the fibers during the high-
temperature graphitization procedure for optimal microstructure
and mechanical properties. Mesophase pitch fibers need not be
hot-stretched since they are spun under stress at lower
temperatures. These fibers maintain their orientation during
the subsequent stabilization and carbonization treatments [41].
Manufacturers fabricate smaller (4-10 ) diameter carbon
fibers in order to attain higher tensile strengths, but this is
done at the expense of lower buckling strengths. The principal
advantages of carbon fibers are their high strength- and
stiffness-to-weight ratios, low longitudinal and transverse
CTEs, low sensitivity to fatigue loads, and excellent moisture
and chemical resistance [43]. Their main disadvantage is low
impact resistance due to their low ultimate strain. Carbon
fibers can also be highly conductive to heat and electricity,
which may be an advantage or disadvantage, depending on the
designer's viewpoint.
Carbon fibers can be classified into four types based on
modulus: low, intermediate, high, and ultrahigh. In general,
lower modulus carbon fibers have lower density, lower cost,
higher tensile strength, and higher tensile strains to failure
than higher modulus fibers. PAN-based fibers tend to have
higher ultimate strain, whereas pitch-based fibers have higher
stiffness. Aside from a tendency to oxidize at temperatures
above 300-400C, carbon fibers with certain heat treatments show
negligible strength degradation to temperatures as high as
2000C. Because of the their negative longitudinal CTE, carbon
fibers can be used to manufacture low- or zero-CTE structures by
proper design with positive CTE matrix materials. Thermal
conductivity of carbon fibers can be near that of copper, and
electrical conductivity is about 1/50-th that of copper. Due to
the small diameter of individual carbon fibers, property
measurements such as those given in Table 1 are usually made on
tows. Transverse fiber properties are also difficult to measure
and are therefore usually found by back-calculation from
transverse property measurements made on unidirectional
composites with known resin properties [44]. Transverse and
shear stiffness and strength of carbon fibers are typically
quite low in comparison with their longitudinal properties.
Water, solvents, bases, and weak acids at room temperature do
not affect carbon fibers very much [45]. Couplants and organic
sizings are generally placed on the surface of fibers to
facilitate good bonding with polymer matrices and to prevent
damage during handling, respectively [37]. The type of sizing
can have an important effect on interfacial bonding and,
therefore, static [38,39] and fatigue performance [46] of the
composite. In 1992, prices of carbon fiber reinforcement ranged
from $13/kg for industrial grades to more than $2000/kg for
exotic types, with the more common aerospace grades about $25-
$65/kg [22,47]. The availability of carbon fibers is excellent
Table 1
Typical properties of several types of fiber.

High Mod. Int. Mod. Low Mod. UHMW

Carbon Carbon Carbon Aramid Polyethylene E-glass Quartz
(Pitch) (PAN) (PAN)

Tensile Strength(GPa) 2.4a 2.9a 5.7a 2.8-4.1a 2.6-3.0b 1.9a 3.7a

Tensile Modulus(GPa) 830 390 290 80-190 120-170 72 74-80
Ultimate Strain(%) 0.3 0.7 1.8 2.0-4.0 2.7-3.5 4.8 N/A

Density (g/cm^) 2.2 1.8 1.8 1.4-1.5 1.0 2.6 2.2

Diameter^m) 10 6.5 5.1 12 27-38 3-20 7-14
Longitudinal CTE(ppm/C) -1.45 -1.13 -.75 -2.0 -10 4.8-5.4 0.4-0.54
Dielectric Const.(@1 MHz) N/A N/A N/A 4.0 N/A 6.3-6.8 3-4

Resistivity^Q-m) 2.2 9.5 15 N/A N/A 402xl0 18 lxl0 2 0

a
Measured with impregnated strand.
^Measured with unimpregnated strand.
28

since there are at least six major manufacturers around the

world. Additional details on carbon fibers can be found in
texts on high performance fibers [41,48].
5.1.2. Aramid Fibers
Aromatic polyamide, or aramid, fiber is manufactured by
extruding polymer solution through a spinneret [49-53],
resulting in a fiber with high thermal stability, high strength,
and high stiffness due to the highly ordered domains of the
semi-crystalline polymer. The aligned polymer chains result in
higher strength and stiffness in the longitudinal direction than
in the transverse direction. Like carbon fibers, aramid fibers
have a negative CTE in the longitudinal direction and a positive
CTE in the radial direction. Aramid fiber is fibrillar in
microstructure, which results in poor flexural and compressive
properties. Maximum compressive creep strain before the onset
of kink bands is about 0.3-0.5%. The fibers are typically quite
tough in longitudinal tension and are therefore excellent for
ballistic impact resistance. Typical use temperatures for
aramid fibers range from -200 to 200C, but oxidation limits
their extended use above 150C. The dielectric constant of
aramid fiber is lower than that of glass and comparable to that
of quartz (see Table 1) . Exposure to moisture (especially
moisture transients) facilitates creep due to the interaction of
water with the molecular and crystallographic structures of
aramid fibers [54]. Since ultraviolet radiation degrades aramid
fibers, the fibers should be embedded in a protective matrix.
Plasma surface treatments may be used to improve the bond
between aramid fibers and polymer matrices [55]. Prices for
aramid fibers in 1992 ranged from about $18/kg for brake-lining
grades to $220/kg for exotic grades [22,47]. There are at least
three major manufacturers of aramid fibers in the world, making
for good availability.

5.1.3. Other Polymer Fibers

Another high-strength, high-stiffness polymer fiber, extended
chain polyethylene, can be used in applications where the
temperature is not to exceed about 100C [56,57]. This fiber is
spun from a solution of ultrahigh molecular weight (UHMW)
polyethylene, resulting in a highly crystalline, extended chain
morphology and a low density (see Table 1 for properties). The
resulting fiber is strong, possesses a low dielectric constant,
and is resistant to water, chemical attack, impact, and
abrasion. Like aramid and carbon, it has a negative CTE in the
longitudinal direction and a positive CTE in the radial
direction. As for disadvantages, polyethylene fibers are
susceptible to time dependent deformation (creep or relaxation),
do not bond well to most resin matrices, and have relatively low
upper temperature limits. The fibers can be cut with a hot
knife operating near the melt point (147C). For best results,
the polymer chosen as a matrix material for a polyethylene fiber
composite should be processable at temperatures of 93-104C.
Various fiber surface treatments have been shown to improve the
adhesion between polyethylene fibers and epoxy matrices [58,59],
although some of these treatments degrade the mechanical
 29

properties of the fibers and new treatments are currently being

sought.
High modulus polymer fibers have also been made recently by
drawing as-spun polyvinyl alcohol (PVA) fibers at temperatures
near the melting point (240C) [60,61]. In this process, the
crystallites are fully softened and the molecular chains are
easily drawn out by relatively low tension at temperatures of
80-260C. Young's modulus of these fibers varies from 20-70 GPa
and tensile strength varies from 0.6 to 2.4 GPa. Stiffness and
strength increase with higher drawing temperatures, but ultimate
strain decreases from about 35% to 4%. A typical glass
transition temperature of PVA fiber is about 60 C for any
drawing temperature. As is the case with most fibers, it is
clear that the properties of PVA fibers are strongly process-
dependent. PVA itself degrades quite rapidly in the presence of
moisture, so the fibers must be protected via additives or
barriers.
A review of several lesser known high strength polymer fibers,
such as polyacetate, polyimide, polyazomethine, aromatic
copolyester, and polyoxadiaozole hydrazide, can be found in
[48]. Attributes to consider in the selection of a polymer
fiber are its thermal, thermo-oxidative, and radiation
(including UV) stability, low-temperature elasticity, and
moisture resistance.
5.1.4. Glass Fibers
Glass fibers are the most common of all reinforcing fibers for
FRPs. Commercial glass fibers are formed by extruding molten
glass through an orifice that is usually 0.79-3.18 mm in
diameter, followed by drawing to a fine diameter of 3-20 .
Individual fibers are generally coated with a sizing agent to
minimize abrasion damage, to bind the fibers together, and to
aid coupling of the fiber with polymer matrices. The principal
advantages of glass fibers are low cost, high tensile strength,
and excellent insulating properties. The disadvantages are low
tensile modulus, high density, sensitivity to abrasion and
alkaline environments, and relatively low resistance to
moisture, sustained loads, and cyclic loads [62,63].
Two types of fibers most commonly used in the composites
industry are E-glass (calcium aluminoborosilicate), and S-glass
(magnesium aluminosilicate). E-glass is used as a general-
purpose fiber where strength, electrical resistance, acid-
resistance, and low-cost are important (Table 1 ) . S-glass has
higher strength, stiffness and ultimate strain than E-glass, but
costs more than E-glass and is more susceptible to degradation
in alkaline environments [64]. C-glass has a soda-lime-
borosilicate composition that is used for its chemical stability
in acidic environments. Alkali-resistant glass (AR) fibers have
been developed specifically to minimize weight and strength loss
in alkaline environments, but this problem has not been entirely
eliminated to date [62,65]. Softening points of glass fibers
are in the vicinity of 800-1000C [66]. There are at least five
principal sources of glass fiber, making their availability
excellent. Typical prices for glass fibers were $2-$13/kg in
1992 [22,47].
30

Quartz fibers are a high silica (99.95% by mass Si02) type of

glass fiber that have found use in applications requiring
special dielectric, thermophysical, acoustic, optical and
chemical properties [62,67]. Quartz fibers with oxide coatings
can resist temperatures as high as 1100C without considerable
loss of strength. They are also resistant to alkalies and all
acids except hydrofluoric and phosphoric [62]. High
resistivity, low dielectric loss and a low dielectric constant
make this fiber attractive for radio-transparent applications.
Like glass, quartz fibers are susceptible to degradation by
abrasion, long-term loading, and moisture.
5.1.5. Ceramic Fibers
Other types of fibers in less widespread use in the polymer
composites industry are boron, silicon carbide, and alumina
fibers [48]. In comparison with the previously mentioned fibers
for FRPs, these fibers are relatively stiff and strong at higher
temperatures. For example, tensile strengths of some ceramic
fibers can range from 0.5 to 2 GPa at temperatures as high as
1200C [68]. These types of fibers are quite costly, however,
and their full performance capabilities are typically not
approached in polymer matrix composites as they would be in
metal or ceramic matrix composites operating at higher
temperatures. Prices of ceramic fibers vary widely, with some
of the lower cost fibers in the $2-$13/kg range and the higher
cost fibers in the $400-$700/kg range [22]. Additional
information on these fibers and their composites can be found in
the literature on high temperature composites [69].

5.2. Matrices
The primary role of the matrix in FRPs is to provide lateral
support to the fibers and to protect the fibers from physical
and chemical trauma due to the surroundings. The matrix may
also be used to impart desired physical properties to the FRP.
Some of the important material characteristics to consider in
selecting a matrix for a structural FRP are: stiffness,
strength, fracture toughness, thermal and electrical
conductivity, upper use temperature, CTE, processing
temperature, chemical shrinkage during processing, ability to
impregnate and bond to fibers, flame resistance, and sensitivity
to environmental factors such as moisture, chemicals, or
ultraviolet radiation. In addition, the selection of
thermoplastic or thermoset resins may be pre-determined by the
availability of proper storage and processing facilities.
Finally, perhaps the most important consideration for commercial
applications is cost. This discussion of matrices for FRPs
begins with a short review of common compatibility issues in
matrix selection, followed by a summary of the key attributes of
common resins for FRPs.
From the manufacturing standpoint, one important factor to
consider in the selection of a matrix for an FRP application is
the relative mismatch in shrinkage or expansion between the
fiber and matrix that can occur during processing. One source
of such a mismatch is unequal CTEs in the presence of a
temperature change, and another is chemical shrinkage of the
 31

matrix due to curing (in thermosets) or crystallite growth (in

semi-crystalline thermoplastics) during processing [70]. An
extreme example of thermally induced residual stresses arises in
carbon FRPs where the longitudinal CTE is about -1 //0 for
fibers and 10 to 100 //0 for the matrix. Considering that
processing temperatures can be anywhere from room temperature to
320C, it is clear that CTE mismatch during cool-down from the
processing temperature can cause significant residual stresses.
Coefficient of thermal expansion mismatch and chemical shrinkage
can both cause matrix cracks near the fibers, depending on the
degree of shrinkage mismatch and relative stiffnesses of the
fibers and matrix. Another manifestation of this same problem
is the development of ply-level residual stresses due to the
processing of multi-directionally reinforced laminates, which
can lead to strength degradation or matrix cracking [71].
Different cool-down rates can be used to control the development
of residual stresses in some cases [72-74]. Since the CTE of
the polymer matrix usually exceeds that of the fiber, another
approach to minimize residual stresses is to minimize chemical
and thermal shrinkage of the resin by the addition of relatively
rigid, dimensionally stable fillers such as silica particles
[22,75].
A second compatibility issue to consider in matrix selection
is processing temperature. It was noted earlier in the
discussion of polymer fibers that processing temperatures may be
limited to temperatures near 100C. Several low-temperature
cure thermoset resins meet this restriction, but have
correspondingly inferior elevated temperature performance and
shorter room-temperature pot life. Thermoplastic resins need to
have relatively high melt temperatures in order to have good
room temperature properties, and are therefore not suitable for
the manufacture of FRPs with low melting point fibers.
5.2.1. Polyester Resin
Commercial thermoset polyesters usually consist of an
unsaturated ester polymer dissolved in a crosslinking monomer
such as styrene [19,22,25,76]. The unsaturated polyester is the
condensation product of an unsaturated dibasic acid (usually
maleic anhydride) and one or more glycols such as propylene
glycol, ethylene glycol, diethylene glycol, dipropylene glycol,
or neopentyl glycol. Styrene reduces the viscosity of the resin
during handling and links neighboring polyester molecules at
their unsaturation points during cure. An inhibitor is usually
added to the styrene to prevent crosslinking until the addition
of a catalyst (for elevated cure) or promoter (for room
temperature cure) by the component manufacturer. The degree of
unsaturation of the polyester resin depends upon the presence of
a saturated dibasic acid such as phthalic anhydride, isophthalic
acid, or adipic acid. A larger frequency of unsaturation
results in a higher crosslink density which, in turn, improves
stiffness, glass transition temperature, and thermal stability,
but lowers ductility. Because styrene also contributes
unsaturation points, higher styrene contents lead to longer
curing times. Higher temperatures can lead to cure times of
less than five minutes, but there is an optimum curing
32

temperature for each catalyst-resin system that results in the

best final properties of the matrix [19].
Depending on the mix of ingredients, properties of polyesters
can vary widely. For example, the addition of adipic acid
rather than phthalic acids lowers the stiffness of polyester
molecules, whereas the addition of propylene glycol rigidizes
the resin by restricting molecular rotation. The use of
different catalysts enables polyesters to be cured at room
temperature, at elevated temperature (up to 163C), or by UV
radiation. Polyesters can be made quite resistant to fire,
moisture, acids, and alkalies, although they are degraded by
chlorinated solvents [76,77]. Ultraviolet radiation resistance
of polyester resins can be improved by the substitution of
styrene and other aromatic vinyl monomer derivatives with more
resistant acrylate or methacrylate monomers in cross-linked
composition and by the addition of UV-absorbing substances [77].
The upper use temperature of polyesters is about 120C [76].
The principal advantages of polyester for FRPs are low
viscosity, fast cure time, dimensional stability, excellent
chemical resistance, and moderate cost (about $2-3/kg [78]).
The principal disadvantage of unfilled polyesters is their high
volumetric shrinkage during processing a characteristic which
aids separation of the part from the mold but also promotes high
residual stresses in FRPs. The properties of cast polyester,
listed in Table 2, are generally lower than those of epoxy.
However, the combination of low cost, excellent properties, and
processability of polyester makes it the most widely used resin
for FRPs. The manufacturing methods for which polyesters are
well suited are compression molding, filament winding, hand lay-
up, mat molding, vacuum bag molding, pultrusion, and spray-up.

5.2.2. Vinyl Ester Resin

Vinyl ester resins, introduced in the late 1970s, are produced
by the reaction of a monofunctional unsaturated acid, such as
methacrylic or acrylic acid, and an epoxy resin [77]. As in
polyesters, styrene monomers are used to decrease viscosity
prior to cure and to crosslink unsaturated points in neighboring
vinyl ester molecules during cure.
Vinyl esters have advantages over polyesters in terms of
chemical resistance and high temperature resistance [25,77].
They are easier to handle during processing than either
polyester or epoxy, and they also have better resilience than
polyester due to relatively less crosslinking. The hydroxyl
(OH) groups occurring along the length of the vinyl ester
molecule form hydrogen bonds with similar groups on the surface
of glass fibers, resulting in excellent wet-out and high
interfacial strength. The characteristics of the epoxy resins
used in the process have a significant influence on the heat
resistance and thermal stability of vinyl ester, but do not
influence tensile strength significantly. Vinyl esters are
well-suited for the manufacture of FRPs due to their low
viscosity and short cure time, but they are at a disadvantage
relative to epoxies because of their high volumetric shrinkage
during cure. Table 2 lists some properties of a typical vinyl
ester. Vinyl esters are most often used in hand lay-up work
Table 2
Typical properties of several FRP matrix materials.

Polyester BMI Polyimide Epoxy PEEK

Tensile Strength (MPa) 20-100 40-100 40-190 55-130 103

Tensile Modulus (GPa) 2.1-4.1 2.7-4.2 3-5 2.5-4.1 1.1
Ultimate Strain (%) 1-6 1.2-6.6 1-60 1-9 30-150

Density (g/cm^) 1.0-1.45 1.2 1.3-1.4 1.1-1.3 1.3

TgCC) 100-140 220-320 210-340 50-260 144

CTE (/) 55-100 21-73 14-50 45-90 55

34

requiring superior chemical resistance, but some are used in

BMCs and SMCs. Vinyl ester resins cost about $3-4/kg [78].
5.2.3. Epoxy Resin
Epoxy resins, first commercially available in the late 1940s,
are the most versatile matrices for FRPs. They have an
exceptionally broad range of physical properties, mechanical
capabilities, and processing conditions. One can change such
characteristics as toughness, chemical and solvent resistance,
strength and stiffness, fatigue and creep resistance, fiber
adhesion, thermal stability, and electrical properties by
modifying reactants, conditions of cure, and the chemical
structures of the resin and the curing agent [79,80].
The term "epoxy" refers to a bridge consisting of an oxygen
atom bonded to two other atoms already united in some way. An
epoxy resin contains one or more epoxide groups, which are
three-membered rings wherein the other two atoms are carbon.
The primary starting material in epoxies is diglycidyl ether of
bisphenol A (DGEBA) and its higher homologs, which contain
epoxide groups at each end of the molecule. The chemical
structures of this and other epoxy resins are given elsewhere
[79,80]. The curing of epoxy begins by adding a reactive curing
agent (catalyst) which may contain amines, anhydrides, aldehyde
condensation products, and Lewis acid materials [81]. In the
case of amines, hydrogen atoms in the catalyst react with the
epoxide groups of DGEBA to enable DGEBA molecules to cross-link
with each other.
One of the major advantages of epoxies for the manufacture of
FRP laminates is that the exothermic polymerization process can
be slowed by lowering the temperature of the resin after the
fibers have been infiltrated. This so-called B-staged prepreg
can be fully cured at a later time by simply raising the
temperature of the resin to accelerate cross-linking. Higher
degrees of cross-linking lead to higher stiffness, higher glass
transition temperature, improved thermal stability, and lower
ductility. Additions of elastomer or thermoplastic particles
can significantly improve the toughness and impact resistance of
epoxy resin FRPs, although these improvements tend to reduce the
stiffness and glass transition temperature [80].
Epoxy resins are known for their excellent strength and creep
resistance, strong adhesion to fibers, chemical and solvent
resistance, good electrical properties, high glass transition
temperature, and low shrinkage and volatile emission during cure
(Table 2) . A low emission of volatiles results in better
laminate consolidation under a fixed amount of pressure. With
the proper hardeners, epoxies retain significant properties to
temperatures of 150-177C and sometimes as high as 200C [76].
Optimal curing temperatures are usually in the 107-177C range,
although room-temperature and ultraviolet cure systems can be
formulated [80]. Epoxies can be used in all common FRP
manufacturing processes. The principal disadvantages of epoxies
are high cost (about $4/kg [78]) and cure times as long as
several hours.
The choice of epoxy matrix is often dictated by compatibility
with the FRP manufacturing process rather than mechanical
 35

properties or other physical or chemical attributes. For

instance, to ensure good infiltration and consolidation, long
working times and low viscosities are necessary in wet lay-up
processes, while a tightly-controlled cure rate and temperature-
dependent viscosity are important in autoclave processing [79].
5.2.4. Polyimide Resin
Polyimides are polymers containing cyclic imide groups in the
main macromolecular chain [82]. Polyimide resins can be either
thermoset or thermoplastic polymers, but the thermoplastic types
degrade at temperatures near the melt temperature, making them
difficult to reprocess and giving rise to their
"pseudothermoplastic" label [22,76]. Thermoplastic polyimides
are classified as condensation-type and addition-type.
Condensation-type polyimides are made from low viscosity
starting materials that are well suited for the manufacture of
FRP prepregs. Full curing of condensation-type polyimides
entails imidization of the polymer at temperatures of 2 00-2 60C.
Maximum continuous use temperatures are as high as 316C
[22,76].
Bismaleimide (BMI) is a widely used addition-type polyimide.
Unlike condensation-type and other addition-type polyimides,
BMIs are capable of epoxy-like processing in terms of cure
temperature and handling characteristics. Usually, a post-cure
of 232C is needed to develop full high temperature capability.
BMIs can be cured to a B-stage to make FRP prepreg. Their
continuous-use temperature limitations are in the 200-230C
range.
The key advantages of polyimides are their outstanding
resistance to heat, thermal degradation, organic solvents, and
high energy radiation. They are slightly susceptible to attack
by dilute acids and are dissolved by strong mineral acids such
as nitric acid or concentrated sulfuric acid, particularly at
high temperatures. Also, they are hydrolized by alkalis [82].
Polyimides have low dielectric constants and high dielectric
strengths. Polyimide FRPs can be autoclave-processed at 200
psi, transfer molded at 3000 psi, and injection molded at 10,000
psi. In comparison with lower-temperature thermoset resins,
polyimides tend to be less ductile and more difficult to process
due to higher viscosities, greater volatile emissions, and
longer cure cycles [83]. Some typical properties of polyimides
are given in Table 2.

5.2.5. Thermoplastic Resins

Thermoplastic resins offer several potential benefits over
thermoset resins when used as matrix materials for FRPs. These
potential benefits can be categorized as either cost- or
performance-related. Cost savings may be realized in several
ways associated with the infinite shelf-life of thermoplastic
prepreg at room-temperature. For example, the need for large
storage freezers is eliminated, prepreg shipping costs are
reduced, the need for prepreg warm-up time is eliminated, waste
in the form of "old" prepreg is eliminated, and the need for
frequent prepreg quality checks is greatly reduced. Other cost
benefits accrue due to shorter processing cycles, reduction of
36

volatile emissions, and repairability of in-service FRP

structures. The main performance advantages of thermoplastic
FRPs over thermoset FRPs are improved toughness and, in some
instances, improved elevated temperature stability [9]. For
example, fracture toughnesses of thermoplastic resins and their
composites may be 10 times those of typical thermoset resins.
Glass transition temperatures range from 85 to 277C for
commercial thermoplastic resins used in high performance FRPs.
Processing temperatures for thermoplastics can be quite high,
however, with some as high as 300C. In general, the lower the
processing temperature of a thermoplastic FRP, the poorer the
creep and relaxation resistance of the structure at a given
temperature.
Some of the most common crystalline thermoplastic matrices for
high fiber volume applications are polypropylene, polyamide
(Nylon) 6, 6/6, and 12, polyurethane, polyphenylene sulfide
(PPS) and polyetheretherketone (PEEK). The latter two are at
the high end of the cost spectrum for thermoplastic materials,
and are therefore restricted to high value-added applications.
Other high performance thermoplastic resins are
polyetherketoneketone (PEKK), polyamideimide (PAI), polyether
sulfone (PES), and polyetherimide (PEI). PAI, like certain
polyimides, is a pseudothermoplastic resin, while PES and PEI
are amorphous resins [84]. Costs for aerospace grade materials
can reach $33 0/kg. The lower cost thermoplastic FRPs precursors
are typically made by hot-melt or solution impregnation of the
fibers rather than the hybridizing techniques such as commingled
yarn discussed earlier [5]. Properties of PEEK resin are given
in Table 2.

5.2.6. Resin Fillers and Additives

Fillers were first used to reduce the volume of polymer used
in an application and thereby reduce costs without excessively
degrading properties. Low cost fillers such as hollow glass
microspheres, peanut or walnut shells, fly ash, wood cellulose
serve this function well [85]. Functional fillers go beyond the
role of extending the matrix in FRPs by enabling the attainment
of specific physical properties such as improved strength,
higher viscosity, better surface finish, or a particular color.
Filler weight fractions of up to about 50% (equal parts filler
and resin) are commonly used, with upper limits often determined
by excessive resin viscosity [86]. A common filler for reducing
cost and shrinkage of polyester and vinyl ester resins is
calcium carbonate (CaC03) . Alumina silicate (clay) is used to
improve corrosion resistance, electrical properties, and surface
finish [86]. Other common fillers are kaolin, silica, talc,
mica, and wollastonite. The effects of these and other fillers
on physical properties and chemical stability in various
environments are available in the literature [75].
Numerous resin additives are available for enhancing the
resistance of matrices and FRPs to flames, smoke generation,
moisture, microbial degradation, oxidation, heat, chemicals,
shrinkage, surface roughness, and ultraviolet radiation [22,75].
For example, phosphorous-based flame retardants function by
developing a protective char which separates the unburned
 37

polymer from the flame and heat source. Hydrate-based flame

retardants, such as alumina tryhydrate, undergo endothermic
reactions and release water upon heating, thereby cooling and
quenching the combustion reactions. In the presence of oxygen,
polymers degrade through a complex sequence of chemical
reactions which usually result in the production of free
radicals and changes in molecular weight by chain scission or
crosslinking. Anti-oxidants inhibit atmospheric oxidation of a
polymer by one of the following two ways: free radical
scavenging or decomposition of peroxide molecules into non-
radical, stable products. Ultraviolet light stabilization is
achieved by the use of hindered amine light stabilizers and UV
radiation absorbers. The former works by trapping different
radicals, thereby inhibiting or retarding the photodegradation
process. Thermoplastic additives for reducing shrinkage and
improving surface finish were discussed in Section 3. Other
additives are available for coloration, to modify conductivity,
to aid processing (plasticizers, lubricants and mold releases),
and to aid coupling with fibers [75].

6. MANUFACTURING METHODS
There are perhaps a dozen or so widely-used methods for
orienting fibers and consolidating/curing an FRP component [87].
Some of these methods are used for the manufacture of low-cost,
lightly-loaded or non-structural FRPs, while others are used for
higher-cost, higher-performance structures. Since space
limitations forbid the detailed discussion of all of these
methods, a representative sampling was selected for review. The
discussion will focus on the manufacture of thermoset FRPs,
although comments on the applicability of the method to
thermoplastic FRPs will be made where appropriate.

6.1. Manual and Automated Lay-up

The manual (or hand) lay-up technique refers to the general
practice of placing FRP precursor by hand onto a flat tool plate
or contoured mold [88]. The precursor can be either dry fiber
or prepreg of any form described earlier although mat,
unidirectional tape, and fabric are the most common forms. The
reinforcement can be cut to the proper shape with a hand-held
knife or by machines equipped with die, laser, or ultrasonic
cutters [89,90]. In some applications, chopped fibers are
sprayed onto the wet resin (spray-up process). Dry
reinforcement needs to be impregnated with resin by brushing or
roller coating. The process of placing fibers and resin (or
layers of prepreg) is repeated until the specified thickness or
laminate stacking sequence is completed. For best consolidation
of thermoset FRPs, it is best to use a roller at regular
intervals to squeeze out as much entrapped air as possible.
In the automated lay-up process, computer controlled machines
are used to cut and lay unidirectional and woven prepreg tape
onto the tool plate or mold [91-93]. Although capital
expenditures for such machines are large, improvements in
productivity, quality, and scrap loss make automated tape laying
38

equipment cost-effective for high volume manufacturers.

Thermoplastic prepreg tape can be laid-up by hand or machine,
but it may be necessary to hold the layers in place by spot
welding as the stacking process continues particularly when
fabricating contoured laminates.
6.2. FRP Molding
Several molding techniques have been developed for
consolidating and curing laminates to desired shapes. The most
simple technique called contact molding or open-mold
fabricating is one of the oldest FRP forming techniques. The
contact mold is essentially a one-sided mold, treated with a
release agent, into which the reinforcement and matrix are
placed, consolidated, and cured. A typical contact mold process
might consist of applying a coat of resin to the release-coated
mold (for a good surface finish) , followed by manual or
automated lay-up of additional layers of FRP into the mold.
Special requirements for the materials are as follows: the
resin must have a viscosity that will enable impregnation and
wetting of the fibers without flowing down vertical surfaces of
the mold; the resin must have a sufficiently long gel time to
enable completion of the lay-up process; and the reinforcement
should be capable of being draped to conform to the contours of
the mold. Due to a lack of pressure for consolidation, this
method cannot be used for thermoplastic FRPs.
A slightly more complex version of the contact mold technique
is the vacuum bag molding technique where the FRP is covered by
an air-tight polymer sheet, or bag. Since the bag covers just
the side of the mold containing the FRP, it is necessary that
the mold be air-tight and that the border of the bag is sealed
with a putty of suitable temperature capability. The vacuum
inside the bag causes a pressure of about 101 KPa to be applied
to the laminate by the atmosphere and leads to the removal of
entrapped air and volatiles from the resin. Consequently, the
FRP tends to have a higher fiber content, better surface finish,
and improved mechanical properties. In FRP composites designed
to bleed excess resin during consolidation, the excess resin
needs to be absorbed by a so-called bleeder cloth. To prevent
the bleeder cloth from sticking to the FRP, it is physically
separated from the FRP by a porous release film or fabric. The
release film or fabric allows passage of excess resin without
any bonding taking place between itself and the resin. A
breather cloth (a coarse, porous fabric) may also be used on top
of the bleeder to ensure that pressure is uniformly applied to
all portions of the laminate. FRP material requirements for
vacuum bag molding are similar to those for contact molding.
Thermoplastic FRPs cannot be processed very effectively by the
vacuum bag molding technique since higher external pressures are
usually required for proper consolidation.
The pressure bag molding technique is similar to the previous
molding techniques except that positive pressures of about 3 50
KPa are applied to the outside surface of the bag i.e., on
one side of the mold. Consequently, the bag and mold must be
sturdier than in the previous cases. Pressure bag molding
results in better consolidation than either contact or vacuum
 39

bag molding. It is best suited for use with female molds since
the bag tends to wrinkle if used over a male mold. If the mold
is sufficiently strong to withstand pressures in the vicinity of
690 KPa, pressure bag molding can be used to process
thermoplastic FRPs.
The most capital intensive variation of the molding technique
to be discussed here is autoclave molding. An autoclave is
essentially a heated pressure vessel with ports for access by
vacuum hoses to the bagged laminate. Process temperature,
pressure, and vacuum are usually controlled by a programmable
computer. The autoclave technique is nearly the same as
pressure bag molding, except that the pressure is applied
uniformly to the entire surface of the FRP/mold assembly.
Hence, less sturdy molds are required than in pressure bag
molding, and higher consolidation pressures (up to 690 KPa, or
so) are more easily attained. Figure 4 is a schematic diagram
of a typical vacuum bag arrangement with a flat tool plate and
caul plate for autoclave processing of flat laminates. FRPs of
the highest quality can be made by autoclave molding. On the
downside, the autoclave is expensive to install and operate and
its internal dimensions limit the maximum part size. Since heat
transfer into and out of the FRP is limited by convection in an
autoclave, it may be slightly more difficult to control
temperature than in the case of compression molding where heat
conduction to and from platens is the dominant transfer
mechanism. Thermoplastic FRPs can be handled quite well by
autoclave molding.

Figure 4. Typical vacuum bag arrangement for processing of FRP

laminates.
40

Another common type of molding technique is the hot-pressing

technique. A hot press consists of two flat, parallel, heated
platens that are used to apply pressure and temperature to flat
or nearly-flat FRPs. Hot presses may be either manually
operated or computer controlled and are significantly less
expensive to purchase and operate than autoclaves. Because of
its unidirectional application of force, the hot-pressing
technique is best suited for the manufacture of flat FRPs. It
can also be used for mildly-contoured FRPs if sufficiently
sturdy molds with parallel outside surfaces are available.
Temperature control is usually quite good, and a vacuum bag
placed around the FRP/mold assemblage can be used to aid the
extraction of volatiles, if necessary. In this situation, a
bagging arrangement similar to that shown in Figure 4 may be
used. Hot pressing can be used for thermoplastic as well as
thermoset FRPs.

6.3. Tube Rolling

Tube rolling refers to the procedure of wrapping prepreg sheets
of unidirectional tape or woven fabric around the circumference
of a cylindrical or tapered mandrel [19,94]. The name comes
from the action of rolling the mandrel over the prepreg often
with a specially designed machine consisting of two long, flat
platens which grip the mandrel along its entire length and roll
the mandrel across the prepreg. Tube rolling is commonly used
in consumer applications where good material performance is
required at a low cost. The prepreg sheets can be oriented at
fiber angles between 90-deg. (circumferential) and 0-deg.
(axial). Bidirectional fabrics are most useful when a 0/90-type
fiber arrangement is desired, i.e. when torsional resistance is
not a major design issue. Intermediate angles are best made
with unidirectional layers. Any thermoset polymer that can be
cured to the B-stage is suitable for tube wrapping. If a
thermoplastic prepreg was to be used, it would need to be
continuously heated and consolidated as it was wrapped onto the
mandrel. Automatic tube rolling equipment is available for
rolling tubes up to 75-mm-thick [94]. Suitable mandrel
materials are aluminum, steel, silicone rubber, and
polytetraflouroethylene (PTFE).
Convolute wrapping the practice of continuously wrapping a
large sheet of prepreg around the mandrel in a radially outward
spiral is the fastest and simplest method of rolling uniform
diameter tubes. The length of the pattern, however, must be
equal to the length of the mandrel. This requirement
complicates the wrapping of hoop and angle plies since several
widths of undirectional tape must usually be spliced together to
enable full coverage of the mandrel for each revolution.
Convolute wrapping works best for thicker-walled tubes where the
variances in stiffness in the circumferential direction due to
ply overlaps or butts is negligible.
Spiral wrapping is another method of placing angle plies onto
a cylindrical mandrel. In this method, a thinner width of tape,
fabric, or mat is wrapped in spiral fashion along the length of
the mandrel, much like in cardboard tubes. The advantage of
spiral wrapping over convolute wrapping is that variations in
 41

circumferential stiffness due to overlaps and cut fibers can be

avoided if unidirectional tape is used. This attribute makes it
a better technique for wrapping thin-walled tubes. Reference 94
contains detailed information on how to derive patterns for tube
rolled sheets of prepreg, and also on how to wrap prepreg onto
tapered mandrels.
Thicker tube rolled structures need to be periodically debulked
(i.e., compacted) in order to minimize the wrinkling of the
innermost plies during cure. Curing can be accomplished by
wrapping heat-sensitive shrink tape around the wrapped mandrel
and placing the assembly into an oven for the recommended
heating schedule. Typical shrink tapes are cellophane,
polypropylene, or polyvinyl fluoride. Consolidation pressure
depends on the number of layers of shrink tape used, the tension
used in wrapping the shrink tape, the type of shrink tape used,
and its thickness.
6.4. Filament Winding
Filament winding refers to the general procedure of wrapping
a narrow band of resin-impregnated tow around a convex, rotating
mandrel, resulting in a hollow structure that can be
subsequently cured in an oven (Figure 5) [95]. Typical tow
widths are between 5 and 10 mm. Common filament-wound
structures are circular, elliptical, oval, and square cross-
sectioned tubes. Flat laminates can be made by slitting the
wound shape longitudinally. By adjusting the winding parameters
such as the rate of mandrel rotation and speed of traversal of
the payout eye, one can alter fiber angle between zero
(longitudinal or polar winding) to nearly 90 degrees (hoop
winding). At intermediate angles, the helical path of the
fibers forms an interlocked, 8-type layer, as shown in Figure
5. Winding parameters within a layer and from one layer to the
next can be varied to optimize the mechanical properties of the
structure [96]. Due to the complex relationships among the
various winding parameters, filament winders are generally
computer controlled.
Thermoset fiber tows can be impregnated either previously (B-
staged tow-preg) or just prior to winding by running them
through a spreader and a resin bath (wet winding) . In the
latter case, gel time and viscosity are major considerations in
resin selection [95,97]. Significant cost savings can be
realized by wet winding with a low cost resin rather than
winding with tow-preg. Another advantage of wet winding is that
the need for solvents commonly found in prepregs (for extending
shelf life) is avoided, thus eliminating a source of volatile
emission and void formation during cure. Curing can be
accomplished by wrapping the wound shape and mandrel assemblage
with shrink-tape and placing it in an oven for the prescribed
time.
Thermoplastic material forms such as tow-preg or commingled
fibers of reinforcement and resin can also be filament wound,
but best results are obtained when heating and compaction is
done as the fiber is placed. In this manner, relaxation of
fiber tension in the innermost layers during compaction a
common problem in thicker structures is minimized. Recent
42

/computer console

to tensioner cylinder

Figure 5. Filament winding process. (Courtesy: S.J. Claus,

Penn State University.)

research has pointed to the feasibility of combining aspects of

automated fiber placement and filament winding of thermoplastic
tow-pregs to enable the placement of fibers over non-geodesic
paths and on mandrels with reverse curvatures. To avoid
slippage of the fibers, the thermoplastic tow-preg must be
heated and bonded to the previous layer of material as it is
wound on the mandrel.
Common mandrel materials are aluminum, steel, plaster-over-
segmented/collapsible metal, and water-soluble sand. Mandrel
properties such as maximum use temperature and CTE should be
considered to ensure that the mandrel retains its shape and can
be separated from the FRP after curing and cool-down. The
segmented/collapsible and water-soluble mandrels are necessary
for bottle-like structures where small end openings prevent the
removal of other mandrels.

6.5. Braiding
Braiding, one of the oldest textile technologies known, has
been used to create fiber preforms for structural composites
only since the 1970's [98,99]. The process entails the
interlacing of two or more yarns to form an integrated structure
(Figure 6) . In comparison with FRPs made with unidirectional
tape, braided FRPs have better resistance to impact and
 43

Track Plate

Figure 6. Flat braiding process. (After Ref. 98.)

delamination because of the interlaced yarns. On the other

hand, they have slightly worse in-plane stiffness and strength
due to fiber curvature and lower fiber content. The overall
cost of producing tubular structures by braiding is said to be
about 56% of the cost of comparable filament wound structures
[98].
In two-dimensional braiding, fibers are placed at 6-deg.
relative to the braiding direction (Figure 6) . Additional
reinforcement in the O-deg. direction can be realized by laying-
in straight fibers in that direction. FRPs with 0/ fiber
orientations can have extremely high Poisson's ratios (greater
than 1) , although this attribute can be controlled by the
placement of fibers near the 90-deg. direction. Braided rope is
a good example of such a high Poisson's ratio material. Thick-
sectioned FRPs can be produced by repeated braiding over
previous layers or by braiding systems of yarns in three
directions (three-dimensional braiding). Three-dimensional
braiding can be used to create structures such as round and
square tubes, I,J, and U-shapes, hat stiffeners, and other
complicated cross-sections. Other advantages of braiding
include the capability of easily producing hybrid fiber
44

composites and the capability of producing complex shapes by

braiding over appropriately-shaped mandrels. After the fiber
preform has been manufactured, resin impregnation is typically
done by an injection technique such as resin transfer molding
(see discussion below). Hence, viscosity and pot life should be
considered carefully when selecting resins for braided FRPs.
6.6. Pultrusion
Pultrusion is a process for molding continuous lengths of
constant or nearly-constant profiles [81,86,25]. In most
commercial applications, reinforcement materials such as rovings
and mats are pulled through a bath of thermoset resin, through
forming guides, and through a heated die which is usually
tapered to achieve both compaction and cure of the FRP (Figure
7) . After travelling through the continuous traction pulling
device, the profiles are cut to the desired length with an

Mat

Forming Guide
Roving Heated Die

Resin Supply Finished Part

To Puller

Figure 7. Pultrusion process. (Courtesy: R. Allison, Creative

Pultrusions, Inc.)

abrasive saw or wheel. Longitudinal rovings are necessary to

provide sufficient strength for pulling the material through the
die, although mats and fabrics with fiber angles between 0- and
90-deg. are frequently added to obtain some transverse stiffness
and strength. If very smooth surfaces are required, veils
(ultra-thin mats composed of organic and/or glass fibers) can be
added to keep the reinforcing fibers away from the surface of
the part. Veils also protect the mold from abrasion by the
reinforcement fibers. Pultruded profiles can be made solid,
hollow, or with foam, wood, or wire cores. Hybrid fiber FRPs
can be easily made by the pultrusion process since any mix of
fibers can be selected at the time of manufacture.
In most instances, an unsaturated polyester matrix and glass
fibers are used in pultruded products because of the good
combination of cost, performance, and processability in these
materials. Other fibers discussed earlier are also quite
 45

suitable for pultrusion. An optimal matrix material for

pultruded FRPs must be able to gel and cure rapidly to
adequately release from the die wall. The high chemical
shrinkage of polyester also helps in this regard. Vinyl esters
are also good resins for pultrusion, although they are more
expensive than polyesters. Epoxies are not well-suited for
pultrusion because of slow reaction times, short pot lives,
higher scrap losses, and a tendency to stick to the wall of the
die [86]. Some thermoplastic resins have the potential to
become viable resins for pultruded FRPs [100]. Cost, shrinkage,
viscosity, and melt temperature are important considerations in
thermoplastic resin selection for the pultrusion process [101].
In the construction industry, much attention has been focused
on pultruded rods with "deformed" surfaces which promote better
mechanical bonding in reinforcement rod applications. For
example, a suitable deformation can be made by braiding or
wrapping the rod with additional strands of resin soaked fiber
in a helical pattern just before entering the heated die [102].
FRP cables have also been pultruded for use as tension elements
in large structures [103].

6.7. Compression Molding

Compression molding involves compressing the FRP precursor
materials in matched dies to eliminate entrapped air and form a
net-shape or near-net-shape part [19,25,104]. The precursor
material for this process can be either SMC or BMC, as described
earlier. In the case of SMC, stacks of flat sheets are cut,
stripped of carrier film, and stacked as needed to fill the
mold, whereas with BMC the putty-like material is simply added
to the mold in the appropriate quantity. Cycle times can be as
low as one minute for the fast-gelling thermoset resin systems
and thermoplastic FRPs one of the primary advantages of
compression molding. Other advantages of compression molding
are low material cost and excellent surface finish. Relative to
the fiber spray-up process, much tighter control over fiber
content can be achieved by using SMC and BMC precursors.
A disadvantage of compression molding is the need for high
compression pressures (eg., 21 MPa), which calls for strong
(i.e., expensive) presses and molds. Compression molded parts
are also likely to have variations of mechanical properties of
as much as a factor of two due to an uneven distribution of
fibers in the part. For larger, more complex-shaped parts, it
is more economical to use the lower pressure resin transfer
molding process (see below).
6.8. Resin Transfer Molding
Resin transfer molding (RTM) involves the low pressure
injection of catalyzed resin into a two-part, closed mold which
contains a fiber preform (Figure 8) . The resin is catalyzed
immediately prior to injection, making resin pot life less
important than in other manufacturing methods. After the resin
has filled the mold, a press is used to clamp the mold shut
while the resin undergoes gelation. Parts made by RTM can be
made with complex shapes, excellent surface finish, high
dimensional control, and can be quite cost competitive with
46

Curing
Agent Resin

Finished Part

Figure 8. Resin transfer molding process. (Courtesy: R.

Parnas, NIST, Gaithersburg, MD.)

compression molded parts by virtue of the low injection

pressure. This low pressure permits the use of low cost epoxy
molds for small production runs, although chrome plated steel
molds should be used for maximum durability. Another advantage
of RTM from the integrated manufacturing standpoint is that it
is easy to mold-in threaded fasteners and foam cores. It is
most advantageous to use RTM for large, highly integrated, high
performance structures produced in medium to high volumes [105].
Resin viscosity, fiber preform permeability, and mold design
should be carefully considered to ensure uniform resin
infiltration and fiber distribution [106,107].
Fiber preforms can be made of any fiber form that can fit into
the mold, such as unidirectional, random, braided and woven
rovings for planar shapes. Chopped fibers sprayed onto an
appropriately-shaped screen can also be used as a preform [19].
For thicker parts, preforms can be formed by stitching together
multiple layers of fibers or by using three-dimensionally
braided shapes.
Thermoset resins for RTM must have a low enough viscosity to
fully wet the fiber preform and should gel as fast as possible
to maximize production rates. Oftentimes the mold is heated and
a vacuum is applied in the mold to assist infiltration and
gelation of the resin. Following gelation, the FRP can be
pulled out of the mold and trimmed, if necessary, to final
dimensions. Polyester, vinyl ester, and epoxy are commonly used
RTM resins. Thermoplastic resins are not easily used in RTM
 47

since their high viscosity tends to deform the fiber preform and
prevent good infiltration of the fiber tows.
6.9. Other Manufacturing Topics
Factors to consider in the selection of molds for FRPs include
the temperature and pressure of the process and the thermal
conductivity, specific heat, abrasion resistance, surface
finish, and dimensional stability of the mold material [108].
For instance, nickel-faced steel molds provide good surface
finishes, high thermal conductivity, long life, and excellent
air-tightness. Carbon FRP have reasonably high thermal
conductivity, low specific heat and can be used to tailor the
CTE of the mold [109]. The thermal expansion characteristics of
the mold influences final part dimensions and the ability to
separate the part from the mold upon cool-down. Plaster, rubber
and wood are other low cost mold materials that have advantages
in certain situations [25,110,111].
The cutting and joining of FRP parts requires some special
tools and techniques [112]. Many of the fibers, such as carbon
and glass, tend to be quite abrasive and therefore necessitate
carbide-tipped or diamond-abrasive tools. To prevent damage to
the tools and the composite in the case of drilling, for
example, light feed rates, a firm backing material, and liquid
cooling work best. Other means of precision-machining specimens
include water-jet and laser cutting. Water-jet cutting (or
abrasive water-jet cutting) works by mechanically removing
material by a high-speed, thin jet of fluid. This technique is
advantageous for use on materials that are very tough (such as
aramid fiber), highly abrasive (such as carbon, glass, and
various ceramics), or which release harmful dust in conventional
machining operations, though it is not suitable for materials
that cannot tolerate moisture. The laser cutting process works
by thermally vaporizing, melting, or decomposing the target
material. Though quite versatile, laser cutting is not suitable
for materials which char or thermally degrade easily.
FRP components can be joined using mechanical fasteners and
adhesives, much like conventional materials. However, the need
for joining of FRPs is typically less than in conventional
materials due to the ability to form near-net-shape FRP
structures by a number of processing methods described
previously. Suitable mechanical fasteners for composites must
be selected based on strength and corrosion considerations.
Galvanic corrosion is a significant problem with many metals and
carbon FRPs, so titanium alloys, austenitic stainless steels,
and certain multiphase and Inconel alloys are commonly used in
those joining applications. Aramid and glass FRPs are not
susceptible to the galvanic corrosion problems seen in carbon
FRPs. Adhesives selected for joining operations should be
compatible with the FRP material system in terms of chemical
compatibility, strength, cure temperature, and service
environment. Adhesives may be either thermoset or thermoplastic
resins. The integrity of adhesively bonded joints is strongly
dependent upon the quality of the starting materials and the
following of proper bonding procedures. Welding by resistance
48

[113] or ultrasonic [114] heating is also an option for the

joining of thermoplastic FRPs.
Due to the heterogeneous nature of FRP composites, improper
manufacturing procedures can result in various types of
discontinuities within the material. Examples of such
discontinuities are fiber/matrix debonds, matrix cracks and
voids, broken fibers, and delaminations. Other types of
irregularities could be in the form of fiber-rich or -poor
areas, or contaminants such as carrier films or oil. If a flaw
or irregularity hinders the function of the material, it is
classified as a defect. Investigative techniques to determine
the severity of flaws without altering the serviceability of the
material are known as nondestructive testing (NDT) or
nondestructive inspection (NDI) techniques. Coupling these NDI
or NDT techniques with the engineering analysis and judgement
required to determine the severity of a flaw is the field of
nondestructive evaluation (NDE) [115]. NDT techniques applied
at the manufacturing stage do not differ dramatically from those
applied in the field except perhaps in terms of portability and
robustness.
Some of the more common testing techniques utilize ultrasonic
stress waves, X-ray radiography, eddy currents, and thermal
measurements [116]. In the ultrasonic testing technique, stress
waves of short wavelength about 1 mm or less are passed
through the material to enable measurements of internal
reflections caused by planar defects or attenuation and wave
speed variations due to cracks, porosity, or changes in
stiffness or density. The X-ray radiographic technique works by
determining the spatial variation in X-ray absorption, which
corresponds to density variations caused by cracks, porosity, or
heterogeneous distributions of fibers. The eddy current
technique is sensitive to variations in the magnetic or electric
properties of the material, and thermal techniques reveal
variations in the thermal properties of the material. When used
in conjunction with established NDE methods, these measurements
of material property variations enable a qualified inspector to
determine whether or not the material can function as designed.

7. CONCLUSION
The use of anisotropic, composite materials such as wood beams
and steel-reinforced concrete in the construction industry is an
idea that has paid great dividends over the years. Although FRP
composites have traditionally been used to great advantage in
the aerospace and transportation industries where improved
payload capacity and performance are worth the extra cost,
recent advances in FRP materials and manufacturing technologies
have sparked interest in FRPs in the construction industry as
well. This article has touched upon the characteristics,
limitations, and relative advantages and disadvantages of
various FRP materials that can be tailored for specific
applications. The modular nature of the FRP manufacturing
process makes the creation of new composites a relatively easy
proposition, provided certain materials compatibility issues are
 49

understood by the designer. As more experience is gained by

those involved in the manufacture and use of FRP composites for
the construction industry, new classes of materials optimized
for the unique requirements of that industry will undoubtedly
become available. This new generation of advanced FRP materials
may well enable significant gains in performance, durability,
and life-cycle costs.

8. ACKNOWLEDGEMENTS
The author would like to thank Mr. D.R. Campbell for
assisting with the preparation of this manuscript. Also,
helpful contributions by Drs. R.S. Engel and B.L. Lee are
greatly appreciated.

9. REFERENCES
1 Pebly, H.E., "Glossary of Terms," Section 1 of Engineered
Materials Handbook. Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 3-26.
2 Anonymous, "Glossary of Advanced Composites Terms," Advanced
Composites, 1993 Bluebook, pp. 16-26.
3 Dominguez, F.S., "Unidirectional Tape Prepregs," Section 3 of
Engineered Materials Handbook, Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 143-145.
4 Dominguez, F.S., "Woven Fabric Prepregs," Section 3 of
Engineered Materials Handbook, Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 148-150.
5 Rosenow, M.W.K., "A New Generation of Advanced Thermoplastic
Materials Confronts the Economic Challenge," Proc. 48th Annual
Conference of the Composites Institute, Society of the
Plastics Industry, New York, 1993, paper 11-D.
6 McCarvill, W.T., "Wet Lay-up Resins," Section 3 of Engineered
Materials Handbookf Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 132-134.
7 Muzzy, J.D., "Processing of Advanced Thermoplastic
Composites," The Manufacturing Science of Composites, Vol. 4,
ASME, New York, 1988, pp. 27-39.
8 Dominguez, F.S., "Prepreg Tow," Section 3 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 151-152.
9 Gosnell, R.B., "Thermoplastic Resins," Section 2 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 97-104.
50

10 Velisaris, C.N. and Seferis, J.C., "Crystallization

Kinetics of Polyetheretherketone (PEEK) Matrices," Polymer
Engineering and Science, 26:1574 (1986).
11 Lee, W.I., Talbott, M.F., Springer, G.S., and Berglund, L.A.,
"Effect of Cooling Rate on the Crystallinity and Mechanical
Properties of Thermoplastic Composites," J. Reinf. Plastics
and Composites, 6:2 (1987).
12 Bakis, C.E., Simonds, R.A., Vick, L.W., and Stinchcomb, W.W. ,
"Matrix Toughness, Long-Term Behavior, and Damage Tolerance
of Notched Graphite Fiber-Reinforced Composite Materials,"
Composite Materials: Testing and Design (9th Vol.). ASTM STP
1059. S.P. Garbo, ed., American Soc. for Testing and
Materials, Philadelphia, 1990, pp. 349-370.
13 Hauber, D.E. and Hill, L.A., "Fiber Placement of
Thermoplastic Composites on Complex Shapes," Structural
Composites Design and Processing Technologies, ASM
International, Materials Park, OH, 1990, pp. 351-358.
14 Bishop, S.M. and Dorey, G. , "The Effect of Damage on the
Tensile and Compressive Performance of Carbon Fibre
Laminates," Characterization. Analysis and Significance of
Defects in Composite Materials. AGARD-CP-355, Advisory Group
for Aerospace Research and Development, North Atlantic Treaty
Organization, 1983, pp. 10.1-10.10.
15 Pagano, N.J. and Pipes, R.B., "Some Observations on the
Interlaminar Strength of Composite Laminates," Int. J. Mech.
Sei., 15:679-688 (1973).
16 Herakovich, C.T, "Influence of Layer Thickness on the
Strength of Angle-Ply Laminates," J. Composite Materials,
16:216-227 (1982).
17 Jones, R.M., Mechanics of Composite Materials, Scripta,
Washington, D.C., 1975.
18 Tsai, S.W. and Hahn, H.T., Introduction to Composite
Materials. Technomic, Lancaster, PA, 1980.
19 Mallick, P.K., Fiber-Reinforced Composites. Marcel Dekker,
New York, 1988.
20 Christensen, R.M., Mechanics of Composite Materials. Krieger,
Malabar, FL, 1991.
21 Whitney, J.M., Structural Analysis of Laminated Anisotropie
Plates. Technomic, Lancaster, PA, 1987.
22 Modern Plastics Encyclopedia '92, Vol. 68, 1991, McGraw-Hill,
New York.
 51

23 Meyer, R.W., Handbook of Polyester Molding Compounds and

Molding Technologyf Chapman and Hall, New York, 1987.
24 Colley, B.R., Chang, K.K., and Tufts, T.A., "New Vinyl Ester
SMC Resins," Proc. 48th Annual Conference of the Composites
Institute, Society of the Plastics Industry, New York, 1993,
paper 18-D.
25 Schwartz, S.S. and Goodman, S.H., Plastic Materials and
Processes, Van Nostrand Reinhold, London, 1982.
26 Roberts, R.W., "Cure Quality Control," Section 10 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 745-760.
27 Drzal, L.T, Rich, M.J., Madhukar, M., and Herrera-Franco, P.,
"Characterization of Fiber-Matrix Adhesion in Composite
Materials," Structural Composites Design and Processing
Technologies, ASM International, Materials Park, OH, 1990,
pp. 141-154.
28 Gutowski, T.G. and Cai, Z., "The Consolidation of
Composites," The Manufacturing Science of Composites, Vol. 4,
ASME, New York, 1988, pp. 13-25.
29 Kardos, J.L., Dave, R. , and Dudukovic, M.P., "Voids in
Composites," The Manufacturing Science of Composites, Vol. 4,
ASME, New York, 1988, pp. 41-48.
30 White, S.R. and Kim, Y.K., "Effects of Staged Curing on
Mechanical Properties of Thermosetting Matrix Composites,"
Proc. American Soc. for Composites, 7th Tech. Conf..
Technomic, Lancaster, PA, 1992, pp. 69-79.
31 White, S.R. and Kim, C. , "Simultaneous Lay-Up and In-Situ
Cure Process for Thick Composites," Proc. American Soc. for
Composites, 7th Tech. Conf., Technomic, Lancaster, PA, 1992,
pp. 80-89.
32 Vergnaud, J.-M. and Bouzon, J. , Cure of Thermosetting Resins;
Modelling and Experiments, Springer-Verlag, London, 1992.
33 Plueddeman, E.P., Silane Coupling Agents, Plenum, New York,
1982.
34 Bascom, W.D., "Structure of Silane Adhesion Promoter Films on
Glass and Metal Surfaces," Macromolecules, 5:792 (1972).
35 Peterson, B.L., Pangborn, R.N., and Pantano, C.G.,
"Mechanical Behavior of Chopped Fiber Reinforced ABS at High
Strain Rates," Proc. Intl. Symp. for Testing and Failure
Analysis, ASM International, Materials Park, OH, 1989, pp.
419-427.
52

3 6 Dow Corning Corp., A Guide to Dow Corning Silane Coupling

Agents, 1985, p. 15.
37 Bascom, W.D., "Fiber Sizings," Section 3 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 122-124.
38 Madhukar, M.S. and Drzal, L.T., "Fiber-Matrix Adhesion and
its Effect on Composite Mechanical Properties, I: In-Plane
and Interlaminar Shear Behavior of Graphite/Epoxy
Composites," J. Composite Materials, 25:932-957 (1991).
39 Madhukar, M.S. and Drzal, L.T., "Fiber-Matrix Adhesion and
its Effect on Composite Mechanical Properties, II: Tensile
and Flexural Behavior of Graphite/Epoxy Composites," J.
Composite Materials, 25:958-991 (1991).
40 Swain, R.E., Reifsnider, K.L., Jayaraman, K. , and El-Zein,
M. , "Interface/Interphase Concepts in Composite Material
Systems," J. Thermoplastic Composite Materials, 3:13-23
(1990).
41 Donnet, J.-B. and Bansal, R.C., Carbon Fibers, Marcel Dekker,
New York, 1984.
42 Deifendorf, R.J., "Carbon/Graphite Fibers," Section 2 of
Engineered Materials Handbook, Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 49-53.
43 Konkin, A.A., "Properties of Carbon Fibres and Fields of
their Application," Handbook of Composites, Vol. 1 Strong
Fibres, W. Watt and B.V. Perov, Eds., Elsevier, Amsterdam,
1985, pp. 241-273.
44 Chamis, C.C., "Simplified Composite Micromechanics for
Predicting Microstresses," J. Reinf. Plastics and Composites,
6:268-289 (1987).
45 Judd, N.C.W., "The Chemical Resistance of Carbon Fires and a
Carbon Fibre/Polyester Composite," Proc. 1st Intl. Conf. on
Carbon Fibres, Plastics Inst., 1971, p. 258.
46 Swain, R.E., "The Role of the Fiber, Matrix, and Interphase
in the Compressive Static and Fatigue Behavior of Polymeric
Matrix Composite Laminates," to be published in Proc. Svmp.
on Compression Response of Composite Structures, held 16-17
Nov. 1992, Am. Soc. Testing and Materials, Philadelphia.
47 McDermott, J. , "The Structure of the Advanced Composites
Industry," Advanced Composites, 1993 Bluebook, pp. 6-15.
48 Watt, W. and Perov, B.V., Eds., Handbook of Composites. Vol.
1 Strong Fibres, Elsevier, Amsterdam, 1985.
 53

49 Pigliacampi, J.J., "Aramid Fibers", Section 3 of Engineered

Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 114-116.
50 Pigliacampi, J.J., "Organic Fibers", Section 2 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 54-57.
51 Dobb, M.G., "The Production, Properties and Structure of
High-Performance Poly(p-Phenylene Terephthalamide) Fibres,"
Handbook of Composites, Vol. 1 Strong Fibres, W. Watt and
B.V. Perov, Eds., Elsevier, Amsterdam, 1985, pp. 673-704.
52 Kakihara, R., Kamiyoshi, M., Kumagai, S., and Noritake, K.,
"A New Aramid Rod for the Reinforcement of Prestressed
Concrete Structures," Advanced Composites Materials in Civil
Engineering Structures. S.L. Iyer and R. Sen, Eds., Am. Soc.
Civil Engineers, New York, 1991, pp. 132-142.
53 Gerritse, A. and Werner, J., "ARAPREE, a Non-Metallic Tendon
Performance and Design Requirements," Advanced Composites
Materials in Civil Engineering Structures, S.L. Iyer and R.
Sen, Eds., Am. Soc. Civil Engineers, New York, 1991, pp. 143-
154.
54 Wang, J.Z., Dillard, D.A., and Ward, T.C., "Temperature and
Stress Effects in the Creep of Aramid Fibers Under Transient
Moisture Conditions and Discussions on the Mechanisms," J.
Polymer Sei.: Part B: Polymer Phys., 30:1391-1400 (1992).
55 Gaur, U. , "Interfacial Adhesion in Fiber Reinforced
Thermoplastic Composites," SPE Annual Technical Conf., Soc.
Plastics Engineers, Brookfield, CT, 1989, pp. 1513-1515.
56 Allied Signal Corp., Sales Literature for SPECTRA Fibers,
Morristown NJ, 1989.
57 Ward, I.M., "Preparation, Properties, and Structure of High-
Modulus Polyolefines and Polyoxymethylene," Handbook of
Composites, Vol. 1 Strong Fibres, W. Watt and B.V. Perov,
Eds., Elsevier, Amsterdam, 1985, pp. 705-740.
58 Nguyen, H.X., Merril, R.G., Schriver, A.K., and Leone, E.A.,
"New Method to Enhance Adhesion of Extended Chain
Polyethylene Fibers," Structural Composites Design and
Processing Technologies, ASM International, Materials Park,
OH, 1990, pp. 333-344.
59 Grummon, D.S., Schalek, R. , Ozzello, A., and Drzal, L. ,
"Modification of Fiber-Matrix Adhesion in Polyethylene
Reinforced Composites by Energetic Ion Irradiation,"
Structural Composites Design and Processing Technologies, ASM
International, Materials Park, OH, 1990, pp. 155-162.
54

60 Kunugi, T. , Kawasumi, ., and Ito, T. , "Preparation of Ultra-

High Modulus Polyvinyl Alcohol Fibers by the Zone-Drawing
Method," J. Applied Polymer Sei., 40:2101-2112 (1990).
61 Sakurada, I., Polyvinyl Alcohol Fibers, Marcel Dekker, New
York, 1985.
62 Aslanova, M.S., "Glass Fibers," Handbook of Composites, Vol.
1 -Strong Fibres, W. Watt and B.V. Perov, Eds., Elsevier,
Amsterdam, 1985, pp. 3-60.
63 Gupta, P.K., "Examination of the Tensile Strength of E-Glass
Fiber in the Context of Slow Crack Growth," Fracture
Mechanics of Ceramics, Vol. 5, 1983, p. 291.
64 Owens Corning Fiberglas Corp., Sales Literature for
Fiberglas, Toledo, OH, Pub. No. 5-ASP-10139-A, July 1983.
65 Majumdar, A.J., "Alkali-Resistant Fibres," Handbook of
Composites, Vol. 1 Strong Fibres, W. Watt and B.V. Perov,
Eds., Elsevier, Amsterdam, 1985, pp. 61-85.
66 Miller, D.M., "Glass Fibers," Section 2 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 45-48.
67 Quartz Products Corp., Sales Literature for Quartzel Fibers,
Louisville KY, April 1989.
68 DiCarlo, J.A., "High Temperature Structural Fibers Status
and Needs," NASA TM-105174, Nat. Aeronautics and Space
Admin., Sept. 1991.
69 Doychak, J., "Metal- and Intermetallic-Matrix Composites for
Aerospace Propulsion and Power Systems," JOM, 44:46-51
(1992) .
70 Yates, B., McCalla, B.A., Phillips, L.N., Kingston-Lee, D.M.,
and Rogers, K.F., "The Thermal Expansion of Carbon Fibre-
Reinforced Plastics. Part 5: The Influence of Matrix Curing
Characteristics," J. Materials Science, 14:1207 (1979).
71 Hahn, H.T., "Effects of Residual Stresses in Polymer Matrix
Composites," J. Astronautical Sciences, 32:253 (1984).
72 Coxon, B.R., Seferis, J.C., and Ilcewicz, L.B., "The Effects
of Process Variables on Transverse Matrix Cracking in High
Performance Composites," The Manufacturing Science of
Composites, Vol. 4, ASME, New York, 1988, pp. 129-139.
73 White, S.R. and Hahn, H.T., "Process Modeling of Composite
Materials: Residual Stress Development During Cure. Part I,
Model Formulation," J. Composite Materials, 26:2402 (1992).
 55

74 White, S.R. and Hahn, H.T., "Process Modeling of Composite

Materials: Residual Stress Development During Cure. Part
II, Experimental Validation," J. Composite Materials, 26:2423
(1992) .
75 Edenbaum, J. , ed., Plastics Additives and Modifiers Handbook.
Van Nostrand Reinhold, New York, 1992.
76 Harper, C A . , ed. in Chief, Handbook of Plastics. Elastomers,
and Composites. 2nd Ed., McGraw-Hill, New York, 1992.
77 Dudgeon, C D . , "Polyester Resins," Section 2 of Engineered
Materials Handbook. Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 90-96.
78 Dieter, G.E., Chair, Committee on Assessment of Research
Needs for Wind Turbine Rotor Materials Technology, Final
Report, NAT S-324, National Academy Press, Washington, D.C.,
1991.
79 May, C.A., "Epoxy Resins," Section 2 of Engineered Materials
Handbook. Vol. 1, Composites, ASM International, Metals Park,
OH, 1987, pp. 66-77.
80 May, C.A., ed. , Epoxy Resins Chemistry and Technology. 2nd
ed., Marcel Dekker, New York, 1988.
81 Meyer, R.W., Handbook of Pultrusion Technology. Chapman and
Hall, New York, 1985.
82 Bessonov, M.I., Koton, M.M., Kudryavtsev, V.V., and Laius,
L.A. , translated from Russian by L.V. Backinowsky and M.A.
Chlenov, Polyimides Thermally Stable Polymers.
Consultants Bureau, New York, 1987.
83 Meador, M.A., Cavano, P.J., and Malarik, D.C., "High
Temperature Polymer Matrix Composites for Extreme
Environments," Structural Composites Design and Processing
Technologies. ASM International, Materials Park, OH, 1990,
pp. 529-539.
84 Harvey, M.T., "Thermoplastic Matrix Processing," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 544-553.
85 Sekutowski, D. , "Fillers, Extenders, and Reinforcing Agents,"
Plastics Additives and Modifiers Handbook. J. Edenbaum, ed.,
Van Nostrand Reinhold, New York, 1992, pp. 493-502.
86 Martin, J.D. and Sumerak, J.E., "Pultrusion," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 533-543.
56

87 Strong, A.B., Fundamentals of Composites Manufacturing;

Materials. Methods and Applicationsf Soc. Manuf. Engineers,
Dearborn, MI, 1989.
88 Brand, R.A., "Manual Lay-up," Section 8 of Engineered
Materials Handbook. Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 602-604.
89 Singer, S., "Ply Die Cutting," Section 8 of Engineered
Materials Handbook. Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 608-614.
90 Cook, M.W. and McGill, P.J., "Ultrasonic Ply Cutting,"
Section 8 of Engineered Materials Handbook. Vol. 1,
Composites, ASM International, Metals Park, OH, 1987, pp.
615-618.
91 Pirrung, P.F., "Flat Tape Laying," Section 8 of Engineered
Materials Handbook, Vol. 1, Composites, ASM International,
Metals Park, OH, 1987, pp. 624-630.
92 Williams, L.A., III, "Contoured Tape Laying," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 631-635.
93 Goode, K. , "Automated Ply Lamination," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 639-641.
94 Roy, P.A., "Tube Rolling," Section 8 of Engineered Materials
Handbook. Vol. 1, Composites, ASM International, Metals Park,
OH, 1987, pp. 569-574.
95 Peters, S.T. and Humphrey, W.D., "Filament Winding," Section
8 of Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 503-518.
96 Claus, S.J., "The Effects of Winding Pattern on the
Compressive Behavior of Filament-Wound Cylinders," Proc.
American Soc. for Composites. 7th Tech. Conf.. Technomic,
Lancaster, PA, 1992, pp. 258-265.
97 McCarvill, W.T., "Filament-Winding Resins," Section 3 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 134-138.
98 Ko, F.K., "Braiding," Section 8 of Engineered Materials
Handbook. Vol. 1, Composites, ASM International, Metals Park,
OH, 1987, pp. 519-528.
99 Chou, T.-W. and Ko, F.K., Eds., Textile Structural
Composites. Vol. 3, Composite Materials Series, Elsevier,
Amsterdam, 1989.
 57

100 Wilson, M.L., Buckley, J.D., Dickerson, G.E., Johnson, G.S.,

Taylor, E.C., and Covington, E.W., III, "Pultrusion Process
Development of a Graphite Reinforced Polyetherimide
Thermoplastic Composite," Proc. 44th Annual Conference of
the Composites Institute, Society of the Plastics Industry,
New York, 1989, paper 8-D.
101 Wilson, M.L. and Buckley, J.D., "The Potential for Low Cost
Thermoplastic Pultrusion," Proc. 48th Annual Conference of
the Composites Institute, Society of the Plastics Industry,
New York, 1993, paper 2-A.
102 Pleimann, L.G., "Strength, Modulus of Elasticity, and Bond
of Deformed FRP Rods," Advanced Composites Materials in
Civil Engineering Structures. S.L. Iyer and R. Sen, Eds.,
Am. Soc. Civil Engineers, New York, 1991, pp. 99-110.
103 Kim, P. and Meier, U., "CFRP Cables for Large Structures,"
Advanced Composite Materials in Civil Engineering
Structures. S.L. Iyer and R. Sen, Eds., Am. Soc. Civil
Engineers, New York, 1991, pp. 233-244.
104 Johnson, C.F., "Compression Molding," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 559-563.
105 Johnson, C.F., "Resin Transfer Molding," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 564-568.
106 Tucker, C.L., ed., Fundamentals of Computer Modeling for
Polymer Processing. Hanser Publishers, New York, 1989.
107 Phelan, F.R., "Flow Simulation of the Resin Transfer Molding
Process," Proc. American Soc. for Composites. 7th Tech.
Conf., Technomic, Lancaster, PA, 1992, pp. 90-102.
108 Borstell, H. and Turner, K.T., "Tooling for Autoclave
Molding," Section 8 of Engineered Materials Handbook. Vol.
1, Composites, ASM International, Metals Park, OH, 1987, pp.
578-581.
109 Harmon, B.D., "Graphite-Epoxy Tooling," Section 8 of
Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 586-588.
110 Faston, M. and Adams, R.C., "Elastomeric Tooling," Section
8 of Engineered Materials Handbook. Vol. 1, Composites, ASM
International, Metals Park, OH, 1987, pp. 590-594.
111 Schneider, C.W. and Carroll, H.E., "Elastomeric Tooling
Application," Section 8 of Engineered Materials Handbook.
Vol. 1, Composites, ASM International, Metals Park, OH,
1987, pp. 595-601.
58

112 Meade, L.E., ed., "Machining, Assembly, and Assembly Forms,"

Section 9 of Engineered Materials Handbook, Vol. 1,
Composites, ASM International, Metals Park, OH, 1987, pp.
665-728.
113 Don, R.C., Bastein, L.J., Jakobsen, T. , and Gillespie, J.W.,
Jr., "Fusion Bonding of Thermoplastic Composites by
Resistance Heating," SAMPE J., 26:59-66 (1990).
114 Bongo, R.L., "Ultrasonic Assembly of Thermoplastics,"
Plastics Products Design Handbook, Part B, Processes and
Design for Processes, E. Miller, ed., Marcel Dekker, New
York, 1983, pp. 155-209.
115 Pangborn, R.N., Bakis, C.E., and Holt, A.E., "NDE
Engineering in the Materials Life Cycle," J. Pressure Vessel
Technology, 113:163-169 (1991).
116 Halmshaw, R. , Non-destructive Testing, 2nd ed. , Edward
Arnold, London, 1991.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 59
1993 Elsevier Science Publishers B.V. All rights reserved.

Properties of FRP Reinforcements for Concrete

Lawrence C. Bank, PhD, PE

Associate Professor, Department of Civil Engineering, The Catholic University of

America, Washington, DC 20064, USA, tel: (202) 319-5163, fax: (202) 319-4499,
e-mail: BANK@CUA.EDU.

Abstract
The physical and mechanical properties of polymer matrix composite materials,
often referred to as fiber-reinforced plastic or fiber-reinforced polymer (FRP)
materials, are discussed in this chapter. These properties are discussed with an
emphasis on the potential uses of composite materials as reinforcements for concrete.
The term FRP reinforcements is used to describe the composite material products
developed for this purpose. Properties of composite materials that are likely to be
significant for FRP reinforcements are identified. The influence of the properties of
the constituents, viz, the fibers, resins, fillers and additives, on the properties of FRP
reinforcements is discussed. Emphasis is placed on issues of heterogeneity and
anisotropy of composite materials and how they can be expected to affect the
properties FRP reinforcements.

1. INTRODUCTION

Physical and mechanical properties of composite materials are reviewed in this

chapter. An overview of the pertinent properties of composite materials for use as
FRP reinforcements for concrete is presented. Much of the data on composite
materials in the existing literature has been developed for the use of these materials
in the aerospace, mechanical and recreational industries and is not in a form readily
accessible or understandable to the construction industry and to civil engineers.

The performance requirements of the composite materials needed for FRP

reinforcements may be different from the performance requirements of composite
materials used in other industries. The properties of the composites used for FRP
reinforcements may therefore need to be different from other composites. For
60

example, the issue of the ability of the FRP reinforcements to bond to the
surrounding concrete medium is unique to the FRP reinforcement application and
may place demands on the properties of the composite materials that are not
generally considered significant, such as, the transverse hygrothermal expansion or the
chemical reactivity of the composite with concrete.

The development of specific FRP reinforcements is currently being pursued in

the marketplace. The emphasis of this chapter is therefore on identifying those
properties which may be of significance in the development of these products. The
actual forms of the FRP reinforcement products being considered may vary
depending on application. The primary concern of this chapter is with the properties
of those FRP reinforcement products which are of the rod or bar type. These include
FRP reinforcing bars (rebars), FRP tendons, FRP gratings, and FRP grids. Fibrous
cable tendons that are not embedded in resin matrices are not considered since they
are not composite materials. With regard to FRP grids and gratings (which are often
called two-dimensional reinforcements) in addition to the properties of the individual
bars that form these interconnected planar mesh-like structures, the structural
properties of the connections between the individual bars should also be considered.

Much of the data and theory that is reported for composite materials has been
developed for laminated composite materials that find widespread application in two-
dimensional thin plate and shell structures. FRP reinforcements are typically in the
one-dimensional form of bars, rods, tendons or cables and are relatively thick. The
data and theories developed for thin plate composites may not be applicable to FRP
reinforcements. Even FRP grids or gratings which may be connected in two-
dimensional or three dimensional arrays consist essentially of one-dimensional
reinforcing bars. Exceptions to this are FRP materials that are being considered for
concrete strengthening either as bonded plates or as column wraps which are thin
plate or shell composite structures.

The one-dimensional nature of FRP reinforcements has some technical

advantages. Since most of the one-dimensional FRP reinforcements, currently being
produced, are primarily unidirectionally reinforced or have radial symmetry (called
axisymmetric) as in the case of twisted tendons or braided bars, the abundant
literature on the behavior of unidirectional reinforced composite materials can be
consulted. Although many FRP reinforcing bars have surface "deformations" in the
form of helical wraps or braids that may have multidirectional fibers, this irregular
structure is usually a surface structure for bond development and does not penetrate
into the bar itself. The unidirectional nature of the reinforcing fibers, or the
axisymmetric structure, give FRP reinforcements a transversely isotropic elastic
symmetry. Fortunately, the transversely isotropic material has been one of the most
extensively studied materials in the composite materials literature. Mathematical
models that have been developed to predict the physical and mechanical properties of
transversely isotropic materials are reviewed in what follows. A discussion of the
accuracy of these models and how they can be used to assess the significant
properties of FRP reinforcements is provided.
 61

A discussion of processing and manufacturing methods for the composite

materials that may be used as FRP reinforcements is not included in this chapter.
With respect to the properties of composite materials it should, however, be noted
that the processing technique can have a significant influence on the properties of the
composite material. It will therefore become important in the future to determine
those properties of FRP reinforcements that are sensitive to processing methods and
to identify the processing method when reporting property data. It is important to
ensure that the processing technique produces a composite material of high quality
and of reproducible properties.

2. FIBERS, RESINS, FILLERS AND ADDITIVES

The three primary constituents currently used for production of FRP reinforce-
ments are inorganic and organic reinforcing fibers (e.g. glass, carbon, graphite, aramid
and vinal), thermosetting polymer resins (e.g. polyesters, vinylesters, and epoxies), and
inorganic fillers (e.g. calcium carbonate, clay, alumina trihydrate). The secondary
constituents used in the production of FRP reinforcements are additives (e.g. flame
retarders, UV absorbers, stabilizers, viscosity modifiers), fiber sizings, coupling agents
(silanes), catalysts, promoters, accelerators, and mold release agents [1],[2].

In what follows a distinction is made between the constituent materials and the
different phases that can be identified in the processed composite material. The term
fiber is used in its conventional sense to describe the continuous fibrous phase of the
composite material. The term matrix is used to describe the combination of materials
that bind the fibrous phase together. This phase includes the resin and the filler
constituents as well as the secondary constituents which are typically mixed into the
neat resin materials during the processing of the material. The matrix phase is often
referred to as the "binder" in the plastics industry. In fact, the term "plastic" is usually
used to describe a solid polymeric material that is a combination of resin, catalyst or
hardener, additives and fillers. In addition to the fiber and matrix phases, the
interface between the fiber and the matrix can be identified as a distinct phase with its
own physical and mechanical properties.

Actual properties of the individual constituents are not given in this chapter.
The reader is referred to the references cited and to the bibliography for specific
property data. However, the influence of the constituent properties on the properties
of the resulting composite material are discussed. Much of the composite materials
literature gives data in terms of the fiber and the matrix phases, which are related to
the major constituents of the composite, the fiber and the resin constituents. There is
not nearly as much information available on fillers and their influences. The use of
filler materials is increasing in the composites industry [1]. Inorganic fillers are used
for three primary reasons; cost reduction, property modification and processing
property control [2]. For the production of low-cost high-volume composite materials
62

for use as FRP reinforcements it is likely that significant amounts of filler (up to 20%
by weight of the resin) will be used. The influence that the fillers have on the
properties of FRP reinforcements may be important.

3. HETEROGENEITY AND ANISOTROPY

Two characteristics of composite materials that need to be considered are

those of heterogeneity and anisotropy. Both characteristics are due to the
microstructure of the composite material. Heterogeneity implies that the composite
material is composed of constituent materials having different properties. Anisotropy
implies that the composite material has properties that are directionally dependent.
The two characteristics are often interrelated. For example, an isotropic matrix
reinforced with isotropic and randomly distributed glass beads is heterogeneous but
isotropic while a piece of hot rolled steel may be anisotropic and homogeneous.

The heterogeneous nature of the composite material should be recognized

since the choice of constituent materials will often be directly related to one or more
of the desired properties of the composite. The advantageous or deleterious effects
of the constituents on the eventual properties of the composite need to be considered.
Discussion of the importance of the constituents needs to be within a framework of
the properties of interest for FRP reinforcements.

The anisotropic nature of the composite material should also be recognized.

The anisotropy makes it is necessary to account for many more physical and
mechanical properties in a composite material than need to be accounted for in an
isotropic material. As has been mentioned most FRP reinforcements are one-
dimensional and have transversely isotropic symmetry.

The transversely isotropic structure can be defined in a x, y, z coordinate

system. If the x coordinate defines the longitudinal axis of the bar, the y and z axes
define the transverse plane perpendicular to the longitudinal axis. It is customary to
identify properties associated with the x coordinate with the subscript L which
signifies the "longitudinal" property of the composite (if the composite consists of
unidirectional fibers this indicates that the fibers are oriented parallel to the
longitudinal axis). For a material that is transversely isotropic the properties are not
directionally dependent in the transverse plane (i.e. the material is isotropic in this
plane) and the material has a structure which can be thought of as being radially
symmetric about the longitudinal axis. The properties in this transverse plane are
identified with the subscript T.

The physical and mechanical properties of a transversely isotropic composite

material are identified with respect to the longitudinal and the transverse axes of the
material. For example, five independent elastic stiffness properties, or constants, are
 63

required to describe the constitutive relation of a transversely isotropic material. In

conventional engineering terminology these are usually given as the longitudinal
(axial) modulus, the transverse (radial) modulus, the longitudinal shear modulus, the
transverse shear modulus, and the major Poissons ratio. The other common
engineering properties, the minor Poissons ratio and the transverse Poissons ratio can
be obtained from these properties using appropriate relationships. Compared with
this, the elastic properties of an isotropic material can be specified in terms of two
independent constants, the Youngs modulus and the Poisson ratio. The shear
modulus is obtained from these properties using appropriate relationships between
the constants.

4. PROPERTIES

The properties of the composite materials discussed in this chapter have been
divided into two categories, physical and mechanical properties.

Physical properties are properties that can be, or are assumed to be, related to
the structure of the material at the molecular level (i.e. the scale at which the
individual constituents can be identified by their chemical compositions or physio-
chemical structures). Included in the term physical properties are the mass
properties, the geometric properties, the chemical properties, the thermal properties
and the transport properties of the material.

Mechanical properties are properties associated with the application of

mechanical forces to the material. These properties are usually defined in terms of
continuum mechanics concepts and cannot be directly related to the chemical
composition or the physio-chemical structure of the constituents. The mechanical
properties are typically obtained from mechanical tests; either on the constituents or
on the composite itself. The relationship between the mechanical properties and the
physical properties is not well understood. Often there is an interaction between the
physical and mechanical properties. For example, the temperature dependence of the
stiffness properties of a material is due to an interaction between the physical thermal
properties of the material and the mechanical stiffness properties.

The prediction of the properties of composite materials from the properties of

their constituents has been the subject of much research. Both physical properties
and mechanical properties have been considered. Although much has been
accomplished over the years, the mathematical models used to predict the physical
and mechanical properties of materials, and especially those of composite materials,
are still only capable of prediction of a limited number of properties and with varying
degrees of accuracy. A number of the simplified mathematical models that have been
developed to predict composite material properties will be reviewed in what follows.
64

Most properties of composite materials of interest in engineering applications

are determined from direct physical or mechanical tests on the materials. This fact
has added significance for FRP reinforcements for a number of reasons.

Firstly, the utility of developing mathematical models to predict properties of

composite materials from those of the constituents is brought into question if the
models are of limited accuracy and cover a limited number properties. As has been
argued in the uses of composites in the aerospace industry, the use of mathematical
models allows one to study the effects of the constituents on a parametric basis and
to identify the significant parameters. This argument should be equally valid for the
use of composite materials for FRP reinforcements and is the reason that a large
portion of this chapter is devoted to discussions of micro-structural aspects and
mathematical models.

Secondly, the use of statistical methods to develop properties for use in design
(design allowables) becomes mandatory since no precise methods are available for
prediction of the properties. Here the potential applications of composite materials
to FRP reinforcements may require methods of statistical analysis that are different
from those that have been used to develop allowable properties for other industries.
The methods developed for civil engineering applications must be compatible with
the accepted measures of reliability used in civil engineering.

Thirdly, when testing is performed on manufactured products it becomes

important to distinguish between the physical and the mechanical properties of the
materials and the structural properties of the product tested. The distinction between
a mechanical property and a structural property is not always clear. Mechanical
properties are defined in continuum mechanics concepts as those properties that can
be related to a "representative volume element" of the material that is "statistically
homogeneous" and over which average stress and strain fields can be postulated [3].
Structural properties are defined on a structural scale and may involve issues of
specimen geometry and loading conditions in addition to the physical and mechanical
properties of the material. Such properties may include properties from flexural
stiffness to fatigue life to damage tolerance. In the case of FRP reinforcements this
may also include, say, properties of bar "deformations", properties of bent bars, and
properties of grid intersections. The following discussion will concentrate primarily
on physical and mechanical properties, however, a number of structural properties are
also discussed where generic issues related to FRP reinforcement are involved.

A listing of pertinent physical and mechanical properties follows. Important

aspects of these properties are discussed. Where available, mathematical formulae
are given that can be used to predict the properties of the composite. It should be
noted that numerous methods have been used in the literature to predict the
properties of composite materials. These methods range from the simplistic to the
highly theoretical exact and bounding methods and are discussed in a number of texts
and review articles listed in the bibliography. The equations given in what follows are
taken from a number of sources and are typical of the semi-empirical type that fall
between the simplistic and the highly theoretical models. Wherever possible
 65

references are given for the more exact theoretical equations that have been
developed. The purpose of giving these equations is to enable the reader to get a
sense of the influence of the different constituent properties on the resultant
composite properties.

5. PHYSICAL PROPERTIES

GEOMETRIC PROPERTIES

Fiber Volume Fraction vf

Matrix Volume Fraction vm
Resin Volume Fraction vr
Filler Volume Fraction va
Void Volume Fraction vv
Interface Thickness tin

In the above the subscripts f, m, r, v and in, are self-explanatory. The subscript
a which is used for the filler signifies "additive".

The geometric properties, particularly the volumetric properties, are often used
in mathematical models. In most of the aerospace composite materials literature
volume fractions are used to describe the relative quantities of the constituents in the
composite. In most of the reinforced plastics literature weight fractions are used for
this purpose. When reporting quantities of constituents it is therefore, vital to
indicate whether the quantities are given "by weight" or "by volume". Equations for
relating weight and volume fractions are given in what follows.

The void content of the composite is an especially important quantity to

measure and monitor. The mechanical properties of the composite, particularly the
strengths of the composite, can be significantly influenced by the void content.
Microscopic voids can act as stress concentrators and crack initiators. The chemical
reactivity of the composite can depend on the void content of the composite. A high
percentage of voids in the composite can allow fluids to diffuse into the material and
affect its durability.

The volume fractions are defined for constituent Y as,

* > ! (1)
66

where Vj is the volume of the constituent and V is the total volume of the composite.
For V constituents,
n
(2)

MASS PROPERTIES

Fiber Weight Fraction wf

Matrix Weight Fraction wm
Resin Weight Fraction wr
Filler Weight Fraction wa
Density P

The weight fractions can be determined from tests. Often volume fractions of
the constituents, which are used in mathematical models to predict properties of the
composite, are determined from the measured weight fractions and the known
densities of the constituents. Weights of additives and fillers are often given in parts
per hundred parts (pph) resin by weight. For common reinforced plastics inorganic
fillers are used at between 20 pph and 60 pph [1].

The size of the inorganic particles, which may be spherical or elongated, is

given in terms of a mean particle size, MPS, normally expressed as an equivalent
spherical diameter in microns [2]. MPS sizes range from 0.1 to 30 . Particle
sizes are often graded, much like aggregates in reinforced concrete, to ensure good
spatial distribution of the filler. It should be recognized that the diameters of current
non-metallic reinforcing fibers which range from 6 to about 25 Mm, may be of the
same order of magnitude as diameters of the fillers. This could have an effect on the
resulting properties of the composite. In addition, many fillers have elongated
platelet shapes and may produce structural anisotropy in the matrix phase even
though the resin constituent may be nominally isotropic.

The relationship between the density of the composite, p, and the density of
the constituents is given [4] according to the "rule-of-mixtures" in the following
general form as,
 67

p = p(vi = - n,

,
where / indicates the constituent. Note that in the above form the equation can be
used to calculate the density of the matrix from the density of the filler and the resin.

The relationship between the volume fractions and the weight fractions are
given [4] in general forms as,

. (4)
P '

v{ = -i (5)
Pi

The above relationships are strictly valid only for a composite having zero void
fraction. To account for voids the relation must be modified slightly to account for
the fact that the voids have finite volume but zero weight [5]. The void ratio (or the
void volume fraction) can be obtained from the following equation [5] provided that
the density is first calculated from the volume fractions and not from the weight
fractions,

g{9f f

where g is the gravitational constant. Alternatively, the void fraction can be obtained
[4] by determining the relative difference between the measured and the predicted
densities (based on weight fractions),

_ Ppredicted "measured /\

^predicted

THERMAL PROPERTIES

Isotropie Material

Coefficient of Thermal Expansion

68

Transversely Isotropie Material

Longitudinal Coefficient of Thermal Expansion aL

Transverse Coefficient of Thermal Expansion

Knowledge of the coefficients of thermal expansion (CTE) of FRP reinforce-

ments is particularly important. In the case of concrete reinforced with steel
reinforcing bars the coefficients of thermal expansion of the reinforcing steel bars and
the concrete are very close in value. This fact has been used to suggest that steel
reinforced concrete can tolerate changes in temperature without generating internal
stresses at the interface between the reinforcing bars and the concrete which could
lead to damage and a reduction in bond capacity. In the case of FRP reinforcing
materials the same issue needs to be considered. However, in the case of FRP
reinforcements the issue is complicated by the fact that the FRP reinforcing bar has
two coefficients of thermal expansion; one in the longitudinal direction and one in the
transverse direction. For most unidirectional composite materials currently being
considered for FRP reinforcements the longitudinal coefficient of thermal expansion
is reasonably close to that of concrete, however, the transverse coefficient of thermal
expansion can be 5 to 10 times that of the concrete. This is due to the fact that the
longitudinal coefficient of thermal expansion is dominated by the fibers (which
typically have very small longitudinal CTE's) while the transverse coefficient of
thermal expansion is primarily a function of the matrix CTE which is significantly
larger that the fiber CTE.

The longitudinal and transverse coefficients of thermal expansion are given [4]
as,

a
L = 7( / + tt V
m ) (*)

aT = (1 + V/ )a,v f + (1 + v M )a w v m - a L v L r (9)

In the above expressions both the fiber and the matrix are assumed to have
isotopic coefficients of thermal expansion.

It should, however, be recognized that not only do the thermal expansion

coefficients for composite materials have different longitudinal and transverse values
but that the individual constituents may also be anisotropic and have different
longitudinal and transverse properties. It is well documented that both carbon and
aramid fibers are transversely isotropic and have different thermal expansion
coefficients in the longitudinal and transverse directions. In this case the equations
for predicting the composite thermal expansion coefficients become increasingly
complicated [6]. The above expressions depend on the elastic stiffness properties of
the constituents (subscripts / and m) and the composite (subscripts L and 7). These
properties are defined and discussed in what follows. It should be noted that the
 69

elastic properties can be functions of temperature and moisture and therefore may
lead to nonlinearity of the thermal expansion coefficients of the composite.

HYGROSCOPIC PROPERTIES

Isotropie Material

Coefficient of Moisture Expansion

Transversely Isotropie Material

Longitudinal Coefficient of Moisture Expansion L

Transverse Coefficient of Moisture Expansion

The coefficient of moisture expansion, also called the swelling coefficient, is a

physical property of composite materials that describes the dimensional changes that
develop as the material absorbs moisture. This phenomenon is not considered
relevant for steel reinforcements but may be significant in the case of FRP
reinforcements. The coefficient of moisture expansion is analogous to the coefficient
of thermal expansion. As in the case of the coefficient of thermal expansion the
coefficient of moisture expansion is different in the longitudinal and the transverse
directions. Thermal and moisture expansion may occur simultaneously. This is often
referred to as hydrothermal expansion of composite materials in the literature.

The phenomena of thermal expansion and moisture expansion are similar in

that both create strain fields in the composite and the constituents. However, some
important differences exists. In developing the expressions for the thermal expansion
it is assumed that the constituents and the composite all conduct heat and reach the
same equilibrium temperature. This is not the case for moisture expansion. In this
case the constituents may reach different equilibrium concentrations and this needs to
be accounted for in the analysis [5]. In addition, most fibers absorb significantly less
moisture than do polymeric resins (an exception to this may be the aramid fiber [6]).
Often the moisture absorption in the fiber is assumed to be negligible. For the
special case of unidirectionally reinforced composites it is therefore often assumed
the moisture expansion in the fiber direction is zero and that the moisture expansion
is the transverse direction is related to the moisture expansion of the matrix material
and is given [4] as,
Pt = o (10)
70

Expressions which account for the different concentrations in the constituents

are given in [5]. Often the analogy between thermal expansion and moisture
expansion is invoked and it is assumed that all expressions developed for thermal
expansion can be used for moisture expansion provided the a's are replaced with 's
[6].

CHEMICAL PROPERTIES

Glass Transition Temperature Tg

Chemical Reactivity
Chemical Stability
Flammability and Toxicity

The chemical properties of composite materials are perhaps the most difficult
properties to identify for FRP reinforcements. This is because the chemical
composition of composite materials plays a central role in two areas that are of great
concern to structural engineers, viz, loss of mechanical properties (especially stiffness)
at elevated temperature and durability.

The glass transition temperature is the temperature at which the amorphous

polymeric regions of a material undergo a reversible change from hard and brittle to
viscous and rubbery, and visa versa. The change is associated with changes in the
molecular mobility of the material. The glass transition temperature for polymeric
materials used as resins in FRP reinforcements is relatively low while those of organic
fibers such as carbon and aramid are high. When the glass transition occurs an
abrupt change in stiffness is observed. This change in stiffness is of great importance
FRP reinforcements since serviceability design criteria for structures often include
limitations on deflections of structural members.

Low glass transition temperatures may be acceptable for normal service condi-
tions but may produce unacceptably short fire rating times due to excessive
deflections in very short times when FRP reinforcements are exposed to extreme
temperatures. In addition, at elevated temperatures FRP reinforcements may be
flammable or produce toxic byproducts. For FRP reinforcements used as internal
reinforcements in concrete these concerns may be somewhat reduced due to the
"insulation" provided by the concrete. The precise nature of this insulation, which will
depend on the heat conduction capacity of the concrete, will need to be investigated.
For FRP reinforcements used for external prestressing, bonding or wrapping where
the composite material could be exposed directly to extreme temperatures or fire
 71

these concerns are amplified.

The issue of material durability is one that is related to both the chemical and
the mechanical properties of the material. The chemical reactivity or the chemical
stability of the material may provide some means of measuring the potential for the
material to retain its physical and mechanical properties when subjected to
mechanical and environmental loads over long periods of time. However, it is not an
easy task to identify all the potentially deleterious chemical environments and to
determine the rate of material degradation or deterioration in these environments. It
should, however, be assumed that chemical reactions will occur over time and that
these reactions could lead to material degradation. For example, glass fibers may
deteriorate due to a process called leaching. The reduction in the glass mass due to
this process may lead to changes in the mechanical properties of a composite material
reinforced with glass fibers.

Chemical reactions will occur between the reactive fluid and the constituent
materials of the composite. There is much test data available on the "corrosion resis-
tance" of common polymer resin materials. Most of this data is of an empirical
nature but it should be useful in determining which polymer resins will be suitable for
FRP reinforcements. There is also data on the chemical stability of most common
fiber materials and filler materials. It should, however, be noted that much of this
data has been obtained in stress free environments. There is much less data available
on "stress corrosion" of constituent materials for composite materials. Even the data
that does exist should be used with caution since the types of stress considered in
developing the data may not be applicable to FRP reinforcements. In addition, it is
important to note that the service lives contemplated for civil engineering structures
are much longer than those considered for industrial applications of composite
materials.

It is often assumed that the matrix plays the dominant role in determining the
chemical stability of the composite and hence its durability. While this is certainly
true, in many cases it is also important to consider the chemical properties of the
fibers, the fillers and the additives. In this regard it will be important to identify any
antagonistic combinations of constituents and fluid environments that may accelerate
degradation of FRP reinforcements. The effect of anisotropy on the chemical
properties of composite materials may also be of some interest. Besides the fact that
some fiber systems may have anisotropic structures on the microscale that may
influence the reactivity of the fiber, the anisotropic structure of the composite itself
may create the conditions for "directionally dependant" chemical reactions. For
example, the phenomenon of "wicking" whereby a liquid will travel along a fiber by
capillary action is well known. Consequently, regions of the composite along the fiber
may be more exposed to the reactive environment than other regions, thus creating
preferred locations for chemical activity.
72

TRANSPORT PROPERTIES

Isotropie Material

Moisture Diffusivity D
Thermal Conductivity k
Electrical Conductivity E
Magnetic Permeability M

Transversely Isotropie Material

Longitudinal Moisture Diffusivity DL

Transverse Moisture Diffusivity DT
Longitudinal Thermal Conductivity kL
Transverse Thermal Conductivity kT
Longitudinal Electrical Conductivity EL
Transverse Electrical Conductivity
Longitudinal Magnetic Permeability ML
Transverse Magnetic Permeability MT

The potential for chemical interaction between different environments and the
FRP reinforcements will depend to a large extent on the ability of the reactive fluid
(liquid or gas) to diffuse into the composite material. This will depend on the
diffusivity of the concrete and the diffusivity of the composite material itself. For
example, in the case of leaching of the glass fibers it is first required that the reactive
fluid diffuse through the matrix material and come into contact with the fibers.

The diffusivity of a material is one property of a group of properties called

transport properties [4]. When used to describe the passage of physical substances
(gases, liquids or solids) through the material these properties and called diffusivities.
When used to describe the properties of magnetic induction through the material
these properties are called permeabilities. When used to describe the passage of heat
or electrical current through the material they are called conductivities [6]. The
properties are often grouped together because they describe the same general
phenomenon. Since they describe similar phenomena the mathematical models that
have been developed to describe the dependence of a composite on the properties of
its constituents are similar. In fact, it is commonly assumed that the behavior of all
the transport properties can be described by the same basic mathematical model [6].
If the transport property of interest is identified by the letter , the longitudinal and
transverse transport properties are given [6] as,

VL = Hmvm +
*Vvf (12)
 73

= fi^L^ML^l) (13)

In the above expressions it is assumed that the fiber and the matrix are both
isotropic. In the case of anisotropic fibers these expressions should be modified as
described in [7]. In the case of moisture diffusion the diffusivity of the fiber is
generally neglected [7] since the moisture absorption of the fibers is negligible and
is set equal to zero. This leads to the following simple equations [7] for the
diffusivities,

L mm

DT =D (15)
1 +vf

For FRP reinforcements the most important transport properties are the mois-
ture diffusivities and the thermal conductivities. The durability of the FRP reinforce-
ment will be highly dependent on the moisture diffusivity of the concrete and of the
composite material itself. The duration for which FRP reinforcements will be able to
withstand unexpectedly high temperatures will depend on the thermal conductivity of
the concrete and of the composite itself. Use of FRP reinforcements may require
work on improving the thermal conductivity of the concrete in order to improve the
fire rating of FRP reinforced structures. For situations in which electrical currents or
magnetic fields are present the permeabilities and conductivities may need to be
considered. In fact, this is one of the areas for which FRP reinforcements are
currently considered competitive, since, relative to metallic materials composite
materials can be designed to be almost insensitive to electromagnetic fields.

6. MECHANICAL PROPERTIES

ELASTIC STIFFNESS PROPERTIES

Isotropic Material

Youngs Modulus E
Shear Modulus G
Poisson Ratio v
Bulk Modulus K
Plane Strain Bulk Modulus KP
74

To determine the elastic stiffness properties of an isotropic material only two

of the above quantities, often called constants, are needed. The basic relationships
are,
9KG
- (16)
3K + G

G = ^ (17)
2(1+v)

3K-2G (18)
V =
2(3K + G)

K = ^ (19)
3(1-2v)

Kp = -2 (20)
p
l-2v

As mentioned previously the matrix phase of the composite material consists of

the resin, the filler and the additives. In order to calculate the elastic properties of
the matrix which are used in the mathematical models to predict the properties of the
composite, it may be appropriate to consider the matrix itself as an isotropic
composite and to calculate its elastic properties from the properties of the neat resin
and the inorganic fillers. Assuming that the fillers can be modeled as spherical
particles and that they have isotropic elastic properties, the properties of the matrix
can be found from models developed for dilute suspensions of spherical particles [7].
The dilute suspension condition is given as,
< 1 (2D

where in this case it is assumed that the "composite" is the matrix phase and that,
v + v . =v + v =1 (22)
v
filler resin a r '

The elastic properties of the matrix phase (subscript, m) are then given [7] in terms of
the properties of the resin (subscript, r) and the filler (subscript, a, for additive)
properties as,
 75

~ - 7 - 5v, + 2(4 - 5v)(Ge/G.,)J

V
+ a(*, - g,) (24)
+
+ [(*. * * ) / ( * , |Gr)]J
The Youngs modulus and the Poissons ratio of the matrix phase can then be
calculated from the relationships for isotropic materials given above. The thermal,
hygroscopic and transport properties of the matrix phase can also be obtained in a
similar fashion by utilizing results obtained for composites with spherical inclusions.
For nondilute suspensions equations are given in [7]. For F R P reinforcements the
dilute suspension model should be appropriate since the volume fractions of the
fillers with respect to the matrix should be in the 0.2 (i.e. 20%) range. In non-
structural applications of reinforced plastics the volume fraction of the filler can
approach 0.6 and the non-dilute model should be used. It should be noted that the
presence of the filler tends to improve the stiffness of the resin and the assumption
that the matrix properties are equal to the resin properties is conservative with
respect to stiffness property prediction. However, this is not the case with all
properties. For example, the strength of the composite may be reduced by the
presence of fillers due to local stress concentrations.

Transversely Isotropic Material

Longitudinal Modulus EL
Major Longitudinal Poisson Ratio v LT
Longitudinal Shear Modulus G LT
Transverse Shear Modulus G^
Transverse Plane Strain Bulk Modulus KT
Transverse Modulus E^
Minor Longitudinal Poisson Ratio v^
Transverse Poisson Ratio v^

To determine the elastic stiffness properties of a transversely isotropic material

five of the above quantities are needed. As mentioned previously many methods
have been proposed to obtain the properties of a transversely isotropic composite
from those of its constituents. These methods assume that the composite is a
unidirectionally reinforced material. For the case of non-unidirectional materials,
such as braided materials, prediction of the elastic properties is much more complex.
In what follows, results for unidirectional composites are given. The semi-empirical
modified rule-of-mixtures equations found in [5] are given. Although not obtained by
exact analytical methods such as those found in [6] these equations have been shown
76

to be reasonably accurate when compared with experimental results and do not have
the critical limitations of the simplistic strength of materials models based solely on
the "rule-of-mixtures" that are often suggested. The basic relationships are,
+
EL = V f v (25)

V
LT = V V f + V
.V (26)

+
(<V?)(vf j ( l + GJGfl.r)^)
CJ, T = i- W)
'LT v +
fGm + \{l GJG^Gjn

, 3 - 4v m + Gm/GflT
( G /7rG m )k + r-= -v.
V
GjT - i ^ " ) L (28)
3 - 4v + GJGm.
yfGm + 2 5LJV Ger
m
4(1 - vm) >"

_ (
2(1 - vffi) m j (29)
1 + GJK
2(1 - vm) m "

The remaining constants can be found in terms of the above five constants as,

ET = -- r (30)
ELKT + GTJEL + \

ET
V
V
TL = LrTT (31>
E
L

(32)
2G
TT

In the above the matrix is assumed to be isotropic and the fiber is assumed to be
transversely isotropic. Consequently, all the fiber properties appear with L and T
subscripts in addition to the / subscript (for fiber). For the transversely isotropic fiber
the plane strain bulk modulus, Ka can be found in terms of ^, E^, G ^ , G ^ and
vLT from eq. (30) above. It should be noted that it is not physically possible to
 77

determine the transverse properties of individual fibers. The way the properties are
determined [6] is by performing tests on unidirectional laminates and back-calculating
the fiber properties from the above equations. Simplifications to the above equations
occur when the fiber is also isotropic. In that case the following substitutions are
made: E^ = E f , G ^ = G ^ = G f , K f r = K fP , ^ = ^ = v f .

With regard to the stiffness properties described above it should be recognized

that in the case of FRP reinforcements the transverse stiffness properties of a FRP
reinforcing bar are not going to be the same as its longitudinal properties. In the
case of isotropic steel reinforcing bars only the Youngs modulus and the Poisson ratio
need to be specified to characterize the bar. In typical manufacturers data for FRP
reinforcement products usually the longitudinal modulus E and perhaps the
longitudinal Poisson ratio v are given. In addition to creating the incorrect
impression that these bars are isotropic, this practice may also give the impression to
design engineers that the transverse properties of the bar are not going to be
significant or needed in design of concrete structures reinforced with FRP
reinforcements.

A potentially important property is the longitudinal shear modulus of FRP

reinforcements. Since it is an independent property it cannot be calculated from E
and v as for isotropic materials. It is expected to be low compared with steel rein-
forcements because it is highly dependent on the matrix of the composite. The
performance of the FRP reinforcements under torsional loads and transverse shearing
loads will be controlled by this property. With regard to bond it should be noted that
the longitudinal shear modulus plays a critical role in transferring the load into the
high stiffness fibers. LOW values of longitudinal shear modulus may necessitate
longer development lengths in order to develop the full strength of the bar. It has
been reported that thicker FRP bars show lower strengths than thinner FRP bars [8].
This may be due to the low longitudinal shear modulus of the bars since the interior
fibers in the thicker bars may not be fully stressed when the outer fibers reach their
failure strains. The transverse modulus of the bars could also be significant. In the
case of FRP tendons, the transverse and shear properties should be important in the
anchorages where all the load needs to be transferred into the tendon over a short
length (either in bonded tendons during the pretensioning or in unbonded tendons
over their lifetimes).

With regard to the properties described above it is often assumed that the
tensile and the compressive elastic properties of composite materials are the same.
For FRP reinforcing bars this may need to be investigated more thoroughly. Lower
compressive longitudinal stiffness than tensile longitudinal stiffness may need to be
accounted for in doubly-reinforced elements, particularly in columns. Since FRP bars
will typically be uniaxially stressed the flexural modulus of the bars, which is a
structural average of the tensile and the compressive modulus will not be that
significant. In FRP grating and grid products used for slab reinforcement, however,
the bars may be subjected to non-uniform stresses.

In addition to the above considerations there are a number of additional

78

factors that will have a significant impact on the stiffness of FRP reinforcements.
These factors are temperature, moisture, time and loading.

All of the stiffness properties discussed above are functions of temperature and
moisture. In FRP reinforcements the thermosetting polymeric resins are particularly
sensitive to temperature. The critical temperature will clearly be the glass transition
temperature, however, stiffness losses may be observed even before this temperature
is reached. Due to the anisotropy of FRP reinforcements the transverse properties
will be more influenced by temperature and moisture due to their dependence on the
matrix properties. In the design of structures utilizing FRP reinforcements the
serviceability limit state of deflection will often govern the design and therefore it is
critical to know the effect of temperature and moisture on the stiffness on the FRP
reinforcement. The development of appropriate test data for FRP reinforcements
under different hygrothermal conditions will be required for design. Although
elevated temperatures are usually of concern it may also be necessary to develop data
for low temperatures that structures could be exposed to colder climates.

The time dependence of the properties of FRP reinforcements are due to the
viscoelastic properties of composite materials and determine the creep, relaxation and
the damping properties. Polymeric resins are known to be viscoelastic. Therefore all
the stiffness properties described above will depend on the rate and duration of
loading and not just on the load itself. The long-term deflection under constant load
is called creep. The loss of force under constant displacement is called relaxation.
Creep compliances are used to describe the creep response and relaxation moduli are
used to describe the relaxation response. The general linear problem of obtaining the
viscoelastic properties of a composite material in terms of the viscoelastic properties
of its constituents is described in [6]. Often, empirical experimentally obtained
power-law viscoelastic models are used to obtain the viscoelastic properties of the
composite directly from tests on composite materials [9]. The term viscoelastic
modulus is often used to describe the time dependent modulus of the material.

With regard to the creep and relaxation properties of composite materials it

should be recognized that the viscoelastic behavior is highly dependent on
temperature. It should also be recognized that since the matrix material is the
constituent that governs the viscoelastic nature of the material the transverse and
shear stiffness properties of the materials will be most significantly influenced by time.
In terms of FRP reinforcements the comments made previously with respect to bond,
development length and anchorages should therefore be considered to be especially
important where long-term behavior is concerned.

The above viscoelastic phenomena are associated with static loads on

composite material structures. When dynamic loads are applied to structures the
ensuing mechanical vibrations will be damped, in part, due to the viscoelastic
properties of the material. For FRP reinforcements in concrete structures under
normal cyclic loading conditions where natural frequencies are low it is unlikely that
the damping capabilities of FRP bars and tendons will be significant. In the case of
 79

dynamic loadings due to earthquakes it is possible the beneficial damping attributes

of certain composite materials that have been utilized in the aerospace industry may
be of interest for FRP reinforcements.

In addition to the above time dependent phenomena the stiffness properties

can also be affected by loading, especially repeated loads. The degradation of
stiffness under repeated loading, called fatigue loading, can be attributed to the de-
velopment and accumulation of microscopic damage in the material. This
phenomenon is therefore associated with the failure of the material which is discussed
in the following section.

FAILURE PROPERTIES

Isotropie Material

Tensile Failure Stress oi

Compressive Failure Stress
Shear Failure Stress as

Tensile Failure Strain

Compressive Failure Strain
Shear Failure Strain es

The failure properties of materials, even isotropic materials, are far less well
understood than their stiffness properties. When describing the failure properties of
isotropic materials it is common to use the term "strength" to describe the stress at
which the material fails. Failure usually implies rupture of the material; the
phenomenon whereby adjoining portions of the material physically separate. This is a
very complex phenomenon, having its roots in the microstructure of the material and
its microscopic flaws and inhomogeneities. The term "strength" is used to signify the
ultimate stress that the material can carry on a macroscale.

Much of the study of the failure of isotropic materials has been devoted to the
study of ductile metallic materials that yield. The term failure is sometimes used in a
generalized sense to describe the yielding of these ductile materials. At yielding the
material usually "fails" to satisfy a serviceability criterion, hence the use of the term
failure. In addition to the three failure stresses listed above (that are often called
failure strengths) the three ultimate failure strains can also be used to determine the
failure of the material. In order to determine the failure of a material under
multiaxial (or combined) stresses, a failure (or yield) criterion must be used to
account for the interactions of the stress components on the failure of the material.
The failure stresses and the failure strains are related to each other by the
80

constitutive relations. For isotropic ductile materials for which the linear and the
nonlinear regimes of the multiaxial constitutive relations can be reasonably well
modeled in the continuum mechanics sense, not much emphasis is placed on the
failure strains. This is not the case for composite materials where the nonlinear
constitutive relations are not well established.

For homogeneous isotropic materials the failure stresses and the failure strains
that are used in engineering design, are obtained from standardized tests on material
coupons. As with the stiffness properties of materials the failure properties are
affected by temperature, humidity, time, repeated loads, loading rate and
environmental conditions. These factors will be discussed with respect to anisotropic
F R P reinforcements in what follows.

Transversely Isotropic Material

Longitudinal Tensile Failure Stress olL

Longitudinal Compressive Failure Stress ocL
Transverse Tensile Failure Stress \
Transverse Compressive Failure Stress
Longitudinal Shear Failure Stress o\
Transverse Shear Failure Stress asT

Longitudinal Tensile Failure Strain ^

Longitudinal Compressive Failure Strain ecL
Transverse Tensile Failure Strain *
Transverse Compressive Failure Strain ecT
Longitudinal Shear Failure Strain esL
Transverse Shear Failure Strain

As with isotropic materials it is most likely that the failure properties of F R P

reinforcements will be obtained from mechanical tests on coupons of the composite
material used to produce the F R P reinforcement. The failure properties of
composite materials are far less well (even less well than isotropic materials)
understood than their stiffness properties. It is necessary to identify the mechanisms
that will produce failure in FRP reinforcements and to identify the influence of the
constituent materials and phases on the failure properties of F R P reinforcements.

Knowledge of the failure mechanisms, or failure modes, is required to develop

an understanding of the effect of specimen geometry and size on the failure
properties. Testing of full-scale specimens taken from the actual F R P reinforcements,
such as is currently done, will require separate data bases for different geometries
(e.g. different bar diameters) as opposed to obtaining data for a particular material
system and applying this information to different geometries, as is traditionally done
with both conventional materials and composite materials.
 81

As with the stiffness properties, most work has been done for unidirectionally
reinforced composite materials. For these transversely isotropic materials the
longitudinal, transverse and the shear failure stresses and strains are required to
characterize the failure properties. As with isotropic materials failure criteria are
required to predict failure under states of combined stresses. For FRP
reinforcements it will be important to identify which of these failure properties and
failure criteria, will be required for structural design. For example, for steel
reinforcements only the yield stress is used in the ultimate strength design procedure
for reinforced concrete structures. For FRP reinforcements the situation is
complicated by the fact that in the longitudinal direction (assuming unidirectionally
reinforced materials) the tensile and compressive failure stresses may differ by an
order of magnitude and that under tensile load the failure mode is brittle as opposed
to ductile as in the case of steel reinforcements.

The failure modes of unidirectionally reinforced composites can be divided

into two distinct categories; the matrix dominated failure modes and the fiber
dominated failure modes. The failure properties are determined by the mode in
which the composite materials fails. Even though much progress has been made
toward identifying the dominant features of these failure modes the mathematical
models that have been developed to predict failure of unidirectional composites are
still not capable of reliable prediction of the failure stresses or strains. It is, however,
instructive to discuss the failure modes and to relate these to the failure properties of
FRP reinforcements.

For FRP reinforcements used as reinforcing bars and tendons the longitudinal
tensile failure stress (or strength) or the longitudinal tensile failure strain is sure to be
of interest. The longitudinal tensile failure stress of a unidirectionally reinforced
composite is dominated by the failure of the fiber and the composite fails in the fiber
mode. In typical unidirectionally reinforced composite materials considered for FRP
reinforcements, the failure strain of the fiber will be less than the failure strain of the
matrix. Under tensile loading the composite will fail when the fiber reaches its
failure (or ultimate) tensile strain. This condition is given as,

The longitudinal tensile failure stress (or strength) can then be estimated by the
"rule-of-mixtures" equation [4],[6] given as,

L = A V f +
OwVm (34)

where om is the longitudinal tensile stress in the matrix at the fiber tensile failure
strain (which equals the matrix tensile strain). It should be noted that the
longitudinal tensile failure stress is a composite property and by definition is related
to the force over a unit area of the entire composite and not just the area of the
fibers (as has occasionally been reported for FRP bars). Besides the fact that this
does not attempt to identify any of the physical mechanisms that cause the failure, the
equation assumes that all fibers have the same failure strain and that they are all
82

stressed uniformly. This is clearly not the case. Much research [6] has been done in
an attempt to use statistical models and physical models to account for the
distribution in fiber failure stresses (strengths) and the mechanisms whereby the
composite accumulates damage until eventual tensile failure occurs.

In situations where FRP reinforcements will be in the compression zone of

concrete elements, as in doubly reinforced beams or slabs and in columns, the FRP
reinforcement will be stressed in compression. It is therefore necessary to determine
the compressive failure stress or strain of FRP reinforcements. As mentioned above
the compressive failure properties may be significantly less [8] than the tensile failure
properties.

Compressive failure of unidirectionally reinforced composite material is

believed to be due to fiber microbuckling [6] or to transverse splitting [4] of the
matrix. Depending on the fiber volume fraction the fiber microbuckling theory
predicts two different buckling modes, called extensional and shear modes. For
volume fractions considered for FRP reinforcements the shear mode will dominate.
Under this condition the compressive strength of the composite material depends only
on the matrix shear modulus and the volume fractions and is given [6],[4] as,

ocL = = - (35)
(1 - vf) vm

In the transverse splitting theory the longitudinal compressive stress (strength) of the
composite is related to the transverse tensile failure strain and the longitudinal
stiffness properties of the composite and is given [4] as,

u
Or = -' (36)

where E ^ is the longitudinal modulus in compression. Empirical equations for

relating the transverse tensile failure strain in the composite to the transverse tensile
failure strain of the matrix are given in [4].

As can be seen, the longitudinal compressive strength of unidirectionally

reinforced composite materials is believed to be largely dependent on the properties
of the matrix material. Since the matrix properties are highly dependent on
temperature and environmental conditions the longitudinal compressive strength of
FRP reinforcements may also depend on these factors. The longitudinal compressive
failure mode is usually classified as a matrix dominated mode due to the significance
of the matrix, however, the fiber stiffness and the interfacial properties also play a
role.

The remaining failure properties of unidirectional composite materials are all

believed to be associated with failure of the matrix. These include transverse tensile
failure stresses and strains, transverse compressive failure stresses and strains,
longitudinal shear failure stresses and strains and transverse shear failure stresses and
 83

strains. Longitudinal shear, or in-plane shear, describes shearing in planes parallel to

the fibers, while transverse shear describes shearing in planes perpendicular to the
fibers [6]. Most attention has been paid to the problem of longitudinal shear which is
dominated by the matrix shear strength and is of much interest in plate-like structures
subjected to in-plane loads. In the case of FRP reinforcements that will be "thick"
and that will be subjected to axial and transverse shearing loads, the transverse shear
strength may be significant. It is expected that the fiber will contribute to the
transverse shear strength of the unidirectional composite. Mathematical models for
predicting these matrix dominated failure modes are described in [6].

For FRP reinforcements these matrix dominated failure properties will be

most important when the FRP bar or tendon is subjected to combined stresses. This
is likely to be at anchorages of FRP tendons, at bends in FRP stirrups, and at
intersections of FRP gratings and grids. Since these locations are frequently critical
in concrete structures it will be important to develop an understanding of these
properties in order to determine design procedures. It is especially important to
recognize, as has been mentioned previously, that the matrix properties are highly
environment dependent. It is also important to recognize that the matrix dominated
properties of the composite become nonlinear at elevated temperatures and at
increased stresses. Consequently, the knowledge of failure strains as well as failure
stresses is required.

As in the case of the stiffness properties it is now necessary to consider the

effects of temperature, moisture, time and loading on the failure properties of the
composite.

Both temperature and humidity will affect the failure properties of the
composite. Those properties that are matrix dominated will be more severely
affected than the properties that are fiber dominated. Most data on composite
materials has been obtained for elevated temperatures and it has been seen that the
failure properties (strengths) decrease with increasing temperature. Little has been
done with respect to low temperatures. There is evidence to suggest that some
properties, especially the matrix dominated properties, may increase slightly with
decreasing temperature due to the increased brittleness of the resins at depressed
temperatures.

With respect to moisture the phenomenon of stress corrosion is likely to be

important. The combined action of loads and adverse environment can cause
accelerated deterioration of the material and eventually lead to failure. In a
discussion of the failure of composites it should be recognized that a significant role
is played by the interface phase (or interfacial region) in the initiation and
accumulation of local damage which eventually leads to global failure of the
composite. This region, surrounding each fiber, is often the region in which the local
failure develops due to debonding of the fiber from the matrix. The properties of the
interfacial region are not well known but the presence of additives, coupling agents
and sizings may have an influence on these properties and hence have an influence
on the failure properties of the composite.
84

The failure properties of composite materials can be time dependent when the
material is subjected to constant stress or strain. Stress rupture, or static fatigue,
describes the decrease in static strength under constant stress over time. Glass fibers
are known to be especially susceptible to this phenomenon. For FRP reinforcements,
this phenomenon will be significant in FRP tendons since these products are expected
to be stressed to high percentages of their failure stresses. It will also be important to
determine if the matrix dominated failure modes are susceptible to stress rupture.
Recognizing that the matrix is composed of very stiff fillers in a flexible resin system
it is possible that over time failure could initiate at the filler interfaces. These
properties could be important in the anchorages, bends, and intersections, mentioned
previously.

The time dependent failure of FRP reinforcements under dynamic loads is

either due to repeated cyclic loads which is called fatigue failure or due to due to high
rates of loading which is called impact failure. For FRP reinforcements the fatigue
properties are expected to be important in structures such as bridges that are
subjected to large numbers of cycles of loading. The fatigue properties of composite
materials are known to be very good when the composite is loaded in the fiber
direction. However, when the composite is loaded in the transverse direction or in
shear the fatigue properties are not necessarily good. Fatigue properties are related
to the development and accumulation of damage on a microscale. Therefore, the
properties of the interfacial regions between the fibers and the matrix and between
the fillers and the resin can be expected to be contributors to the fatigue properties.
Mathematical models, however, have not been developed which attempt to relate
these microstructural defects to the fatigue properties.

The fatigue properties are typically determined by experiment. The fatigue

characteristics of a material are often given in the form of stress ratio versus number
of cycles curves, called S-N curves. Some data is available for unidirectional
composites loaded in the longitudinal direction but little data is available under
transverse or shear loading. In addition, very little is known on the effects of
combined stresses on fatigue life of composite materials. For FRP reinforcements the
longitudinal fatigue characteristics will be important, however, the transverse
properties may well affect the performance in those combined stress states described
previously that occur at anchorages, bends and intersections.

The impact properties of composite materials influence their damage tolerance

and their energy absorption capabilities. FRP reinforcements may offer advantages in
both these areas and an appropriate methodology will need to be developed to
quantify these properties and incorporate them into design procedures. The impact
properties are related to the constituents, especially their fracture toughnesses, and
the loading conditions. Impact properties in the longitudinal fiber direction are
expected to be higher than those in the transverse direction and shear directions that
depend primarily on the matrix.
 85

7. CONCLUSION

The physical and mechanical properties of FRP reinforcements for concrete have
been reviewed in this chapter. These properties have been related to the constituent
materials and microstructural phases of the composite materials that are used to
produce FRP reinforcements. The significant roles played by heterogeneity and
anisotropy of composite materials on the microscale have been emphasized. In order
to develop design procedures for structures reinforced with FRP reinforcements it is
necessary to recognize these factors and to incorporate them into these design
procedures in a rational manner.

CITED REFERENCES

[1] Introduction to Composites. 2nd edition, SPI Composites Institute, NY, 1992.

[2] Modern Plastics Encyclopedia '92. McGraw Hill, NY, 1991.

[3] Hashin, Z, "Analysis of Composite Materials - A Survey", Journal of Applied

Mechanics, Vol. 50, 1983, pp. 481-505.

[4] Agarwal, B.H., and Broutman, L.J., Analysis and Performance of Fiber
Composites. 2nd Edition, John Wiley & Sons, NY, 1990, Chapters 3 and 8.

[5] Tsai, S.W., and Hahn, H.T., Introduction to Composite Materials. Technomic
Publishers, Lancaster, PA, 1980, Chapter 9.

[6] Rosen, B.W. and Hashin, Z., "Analysis of Material Properties," in Engineered
Materials Handbook - Vol. 1. Composites. ASM International, Metals Park,
OH, 1987, pp. 185-205.

[7] Christensen, R.M., Mechanics of Composite Materials. John Wiley & Sons,
NY, 1979, Chapters 2 and 3.

[8] Prucz, J.C., and Wu, W-P., "Performance Simulation of Structural Composite
Rods," in Recent Developments in Composite Materials Structures, (eds. D.
Hui and C.T. Sun), American Society of Mechanical Engineers, AD-Vol.
19/AMD-Vol. 113, NY, pp. 45-50.

[9] Structural Plastics Design Manual. American Society of Civil Engineers, NY,
1984.
86

BIBLIOGRAPHY

Chawala, K.K., Composite Materials - Science and Engineering. Springer-

Verlag, NY, 1987.

Engineered Materials Handbook - Vol. 1. Composites. ASM International,

Metals Park, OH, 1987.

Engineered Materials Handbook - Vol. 2. Engineering Plastics. ASM

International, Metals Park, OH, 1988.

Grayson, M., (ed.), Encyclopedia of Composite Materials and Components.

John Wiley & Sons, NY, 1983.

Jones, R.M., Mechanics of Composite Materials. Hemisphere, NY, 1975.

Kelley, A. (ed.), Concise Encyclopedia of Composite Materials. Pergamon

Press, Oxford, UK, 1989.

Lubin, G. (ed.), Handbook of Composites. Van Nostrand Reinhold, NY, 1982.

Military Handbook MIL-HDBK-17-1C, Polymer Matrix Composites - Vol. 1.

Guidelines. US Department of Defense, 1992.

Structural Plastics Selection Manual. American Society of Civil Engineers, NY,

1985.

Vinson, J.R., and Sierakowski, R.L., The Behavior of Structures Composed of

Composite Materials. Martinus Nijhoff, Dordrecht, 1986.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 89
1993 Elsevier Science Publishers B.V.

A Canadian perspective on R&D, design/codes and technical

committees
M.A. Erkia and S.H. Rizkallab

department of Civil Engineering, Royal Military College of Canada, Kingston, Ontario,

Canada K7K 5L0
b
Dean,s Office, Faculty of Engineering, University of Manitoba, Winnipeg, Manitoba, Canada
R3T 2N2

Abstract
Activity in Canada regarding fiber-reinforced-plastics (FRP) for concrete began in earnest
in the late 1980,s, when the Canadian Society for Civil Engineering created a technical
committee on the use of advanced composite materials in bridges and structures. The major
accomplishments of the committee were the publishing of two state-of-the-art books on the
subject and the organizing of an international conference in 1992, which brought together
many of the leading engineers in this field. The efforts of the committee were supported in
large part by the Canadian Federal Government, which has a number of programs in place
to assist in the development of new technologies in Canada. Most recently, with support from
the Canadian Federal Government, a national network on advanced composite materials for
bridges and structures was established. Some construction ventures using FRP have begun in
Canada, notably two demonstration bridges are being planned, one in the City of Calgary and
the other in the Province of Nova Scotia. Although FRP production in Canada is still limited,
there are numerous research projects on FRP in concrete structures underway at government
laboratories and universities. Finally, the first structural design code in Canada which is
considering provisions for FRP in concrete structures is being prepared for publication in
1994.

1. ROLE OF THE CANADIAN SOCIETY FOR CIVIL ENGINEERING

In 1988 members of the Canadian Society for Civil Engineering (CSCE) visited the Swiss
Federal Laboratories for Materials Testing and Research (EMPA) in Zrich. Already,
researchers at EMPA had had 18 years of experience in the use of composite materials in
bridges. It was decided by the, then, Chairman of the Structural Division of the CSCE to
form a technical committee of the Structural Division on advanced composite materials in
civil engineering structures.
From its inception in 1989, the Technical Committee on Advanced Composite Materials
in Bridges and Structures has sought to maintain a balanced membership representing the
90

practising, manufacturing, and academic sectors. Presently there are over 60 members. The
mandate of the committee is to investigate and, where found appropriate following
investigation, to encourage the use of advanced composite materials in bridges and other
structures. To these ends, the first major effort of the committee was to prepare a
State-of-the-Art report on the use of advanced composite materials in bridges and other
structures. This resulted in the publication "Advanced Composite Materials with Application
to Bridges" [1], which was reviewed in the Canadian Journal of Civil Engineering [2]. The
book had nineteen contributing authors and was the first book of its kind anywhere. Much
of the information reported in the book was received firsthand following a visit to Europe by
members of the committee to meet with practitioners and researchers working with FRP for
structures.
As a result of the interest generated by the committee's first publication, and a lack of
information about Japanese developments with FRP, the committee turned its attention to
Japan. In 1992, with assistance from the Japan Science and Technology Fund [3], members
of the committee visited Japan and later wrote a report of their visit in "Advanced Composite
Materials in Bridges and Structures in Japan" [4].
The next effort of the committee was the organization of the First International
Conference on "Advanced Composite Materials in Bridges and Structures, ACMBS-1" held
in Sherbrooke, Quebec, 7-9 October 1992. Members of the committee at the University of
Sherbrooke took over the responsibility of the conference. The conference brought together
122 delegates from fourteen countries to discuss and exchange information on advanced
composite materials in construction applications. Sixty technical papers and three keynote
lectures were presented [5]. The success of the Conference prompted the establishment of the
ACMBS-1 as the first in a series of international conferences to be hosted in Canada every
four years. The ACMBS-2 is to be held in Winnipeg in 1996.
Since 1991, special sessions on the use of advanced composite materials in structures,
with papers on FRP reinforcements for concrete, have been organized at the Annual
Conferences of the Canadian Society for Civil Engineering. Presentations by members of the
committee have also been made at meetings of the 1992 Transportation Association of
Canada Annual Conference, the Industrial Research Assistance Program of Canada, the
Canadian Association for Composite Structures and Materials, and the Department of
Industry, Science and Technology Canada.
In early 1992, the committee identified several potential topics for a concentrated research
effort into the use of advanced composite materials in civil engineering structures. These
topics would have a number of components that could be undertaken at several universities
and research institutions. The research topics are (i) Parking Structures, (ii) Long-Term
Material Properties, (iii) Enclosure Systems, (iv) Rehabilitation and Repair, and (v) Bridges.
Work has started on these projects at research institutes, and the committee hopes to serve
as a vehicle for the exchange and distribution of the resulting information.
Finally, the committee is participating in the production and distribution of "FRP
International", which is a joint newsletter of the American Concrete Institute, American
Society of Civil Engineers, Canadian Society for Civil Engineering, and Japan Concrete
Institute. The first issue of the quarterly newsletter was published in January 1993.
 91

2. ROLE OF THE CANADIAN FEDERAL GOVERNMENT

The Canadian Federal Government has a number of programs in place to promote
research, development, and field applications of advanced industrial materials for civil
engineering construction. (The Federal Government uses the term "advanced industrial
materials (AIM)", because it is more encompassing, but FRP falls within the Government's
definition of AIM.) At the fore in providing support in the field of AIM is the Department
of Industry, Science, and Technology Canada (ISTC). External Affairs, Public Works Canada,
the National Research Council, the Natural Sciences and Engineering Research Council, the
Industrial Research Assistance Program, and National Defence Canada are some of the other
Federal Government departments and agencies from which support for AIM in civil
engineering construction can also be obtained.
ISTC provides funding for a Strategic Technologies Program (STP), with the objective
of enhancing the international competitiveness of Canadian industries. One of the target areas
in the program is Advanced Industrial Materials, under which the Canadian Society for Civil
Engineering Technical Committee on Advanced Composite Materials in Bridges and
Structures received funding for its work in providing information to the profession on the use
of new materials in civil engineering structures. STP has also funded a market study on the
potential of advanced industrial materials in civil engineering construction in Canada [6]. The
final report of the market study is due in late 1993. In addition, through the Advanced
Industrial Materials Networks component of ISTC's Technology Outreach Program, a number
of networks promoting advanced industrial materials in various engineering sectors have been
established. In the field of civil engineering construction, a new network for promoting
advanced composite materials in bridges and structures was formed in 1993. The network is
described in this chapter under the heading of "Advanced Composite Materials in Bridges and
Structures Network".
ISTC is primarily involved in promoting research and development in industry. For
example, in 1992, ISTC funded a major FRP production center, the Composite Materials
Centre (CMC) of Saint-Jerome Incorporated, near Montreal, Quebec. The funds are being
used to expand existing production and laboratory equipment, and to support CMC's ongoing
effort to identify and develop technologies strategic to the future of the composite materials
industry. The Centre occupies an area of 3716 sq.m. (40,000 sq.ft) with a total investment
in the plant of more than $10 (Cdn) million.
The Japan Science and Technology Fund (JSTF), initiated in 1989, is a five-year $25
(Cdn) million fund to promote joint scientific and technological research and development,
and to help create strategic partnerships in sectors of Canadian priority. It is managed jointly
by External Affairs and International Trade Canada, ISTC, and the Natural Sciences and
Engineering Research Council of Canada [3]. The JSTF funds research visits, research
exchanges, bilateral research and development projects. The Canadian Society for Civil
Engineering Technical Committee on the Use of Advanced Composite Materials in Bridges
and Structures was assisted in the planning and financing of their technical visit to Japan and
the publication of the related task force report [4] through the JSTF. At the time of writing,
with support from the JSTF, the Nova Scotia CAD/CAM Centre in Halifax is organizing a
Canada-Japan Workshop, to be held in Ottawa, July 1993. The objectives of the workshop
are to define research and business areas for enhanced science and technology in the use of
advanced composite materials in bridges and structures and to foster cooperation between
92

Canadian and Japanese universities, public and industrial organizations, and practising
engineers.
Public Works Canada and the National Research Council of Canada have joint research
programs into the use of FRP reinforcements, and these are discussed in this chapter under
the heading "Research and Development". The Canadian Federal Government is also
supporting FRP research and development at Canadian civilian universities through research
grants administered by the Natural Sciences and Engineering Council of Canada. The
Department of National Defence has similar research grants programs supporting FRP
research and development at Canadian Military Colleges. The Industrial Research Assistance
Program provides funding for research conducted by industry or by partnerships of industry
and universities.

3. COMPOSITE MATERIALS IN BRIDGES AND STRUCTURES NETWORK

With support from the Department of Industry, Science and Technology Canada through
the Advanced Industrial Materials Networks component of the Technology Outreach Program,
the Advanced Composite Materials in Bridges and Structures Network (ACMBSN) was
established in 1993 to focus activities in the field among Canadian and foreign members of
industry, government, universities, learned societies, and professional associations. The
objectives of the ACMBSN are to (i) identify products needed for the use of advanced
composite materials in bridges and structures; (ii) assist in forming industrial alliances to
develop these products where they do not already exist; (iii) provide research and
development advice and consulting services to the alliance members; and (iv) assist in
organizing conferences, publishing newsletters, and other information exchanges. The
Canadian Society for Civil Engineering and the Nova Scotia CAD/CAM Centre are providing
administrative support to the ACMBSN.

4. CONSTRUCTION PROJECTS

In Calgary, Alberta, eight precast concrete girders of the Centre Street/Beddington Trail
Bridge were tendered in February 1993 to be prestressed using two different types of carbon
fiber-based reinforcements. The project involves SCI Engineers, Contractors Incorporated
from Calgary, the University of Manitoba, and the University of Toronto. The bridge is a two
span, 23.83 and 19.23 meters, continuous skew bridge. The brand names of carbon fiber-
based reinforcements are LEADLINE [7], produced by Mitsubishi Kasei Corporation, Japan,
and Carbon Fiber Composite Cable (CFCC) [8], produced by Tokyo Rope Manufacturing
Company Limited and supplied by Tokyo Rayon through ITOCHU Canada Limited. The
remaining nine girders in each span will be prestressed using conventional prestressed steel
strands. Steel strands will also be used for post-tensioning to provide continuity over the
middle bridge pier. In addition, the bridge is tendered to use NEFMAC [9] to replace the
conventional steel reinforcements in one of the 250 mm thick approach slabs of the bridge.
NEFMAC is produced by Asahi Glass Matex, in cooperation with Shimizu Corporation,
Japan, and consists of continuous glass or carbon fibers wound to form a grid with
intersecting layers. Behavior of the bridge over its service life will be monitored using fiber
 93

optics. Laser optical filaments will be attached to each of the carbon fiber-based
reinforcements, and monitoring will use a multichannel fiber optic sensing systems with
sensors.
In Halifax, Nova Scotia, a group of companies led by Vaughan Engineering Associates
Limited, and including the Advanced Materials Engineering Centre, Deloitte Touche, the
Nova Scotia CAD/CAM Centre, and the Technical University of Nova Scotia, is proposing
to built an advanced composite materials demonstration bridge, which will be a fully
functioning three-span, two lane highway bridge. Each span will be constructed to showcase
one or more different uses of FRP components. One feature of the structure will be concrete
bridge decks which are devoid of any steel reinforcing bars and reinforced only by
incorporating chopped fiber filaments randomly distributed throughout the mix. FRP rods may
be used at the ends and edges of this deck system. Some of the spans will show how FRP
tendons are used for prestressing of concrete superstructures of bridges. Laser optical
filaments inside FRP tendons and other bridge components will provide "intelligent"
structures capable of indicating the presence and location of damage, as well as allowing
remote monitoring of the structure over its lifetime. Also, some of the spans will demonstrate
how pultruded FRP panels can enclose the superstructure of bridges, especially slab-on-girder
bridges, so as to minimize corrosion and maintenance work required throughout the life of
the bridge, as well as providing in-place working platforms for under-the-bridge rehabilitation
and improved aerodynamic stability.
In 1992, in Montreal, Quebec, the Ministry of Transportation of Quebec reinforced a
100 m length of barrier wall (New Jersey type) on a highway bridge with glass FRP
reinforcement. The bridge is part of Highway 15, crossing the Prairies River (Riviere des
Prairies). The reinforcements were produced by Pultrall Incorporated in Thedford Mines,
Quebec. The long term durability of the structure is being monitored, and the results will
determine the future uses of FRP by the Ministry.

5. RESEARCH AND DEVELOPMENT

A number of government and academic institutions are working on FRP reinforcements

for concrete. The following is a sample of the types of projects underway at some of these
institutions.
The Institute for Research in Construction at the National Research Council and Public
Works Canada are jointly conducting tests on fiber reinforced concrete and FRP
reinforcements. With financial support from Shimizu Corporation of Japan, Autocon
Equipment Incorporated (Ontario), and the Industrial Research Assistance Program (Canada),
one type of FRP reinforcement being evaluated in the study is a proprietary FRP grid,
NEFMAC [9], which could be used to reinforce highway bridge decks, barrier walls, and
parking garage slabs. Tensile strength and stiffness, endurance limits for a number of cyclic
load ranges, and creep under various levels of load sustained for 10,000 hours are being
determined. Tests are in progress to evaluate the effect on the strength of the grid from
ambient temperature, cyclic and sustained loads, and exposure to salt and alkali, freeze-thaw
cycles, and UV radiation. Tests are also being planned for bridge decks models and barrier
walls under simulated loads and environments in the laboratory. Two types of glass FRP rods,
one of which is ISOROD [10], are being tested to determine their fatigue strength and
94

durability.
Public Works Canada is also conducting field investigations in Hull, Quebec, where a
parking garage structure underwent major repair work using fiber reinforced concrete and
FRP bars in four soffit repair areas. The repaired areas were instrumented, and a data-logger
is monitoring service induced loads and loads induced by load tests. The objective is to
collect data on the performance of the conventional repair versus the fiber reinforced concrete
and FRP bars, the performance of the concrete application methods, and the load sharing
characteristics of the existing reinforcements with the replacement reinforcements.
At the University of Manitoba, researchers are investigating the flexural behavior of
concrete beams prestressed with aramid FRP tendons. The different modes of failure, as well
as the effect of varying levels of jacking stresses of the FRP tendons, are being studied. The
behavior is compared to beams prestressed with prestressing steel. The effect of adding non-
prestressed FRP rods on the behavior of the prestressed beams is also being investigated.
Independent research to study the behavior of prestressed concrete T-beams, prestressed with
a carbon FRP tendon, LEADLINE [7], is also in progress. Different cross sections with
different flange widths are used to achieve different prestressing indexes, and consequently
different modes of failure. The carbon FRP tendons are placed in successive layers to enhance
the ductility. The behavior will be compared to similar beams prestressed with steel tendons.
Researchers at Laval University have been studying the effects of strengthening reinforced
concrete beams using externally applied glass fiber reinforced composite sheets [11]. The
sheets are attached to the surface of the members using adhesive bonding, with and without
use of mechanical anchors (HILTI anchors). The beams have shown a significant increase in
flexural strength, leading the research group at Laval to conclude that this method of
strengthening could be feasible for short and medium span bridges (10 m to 30 m).
Similar to the research work at Laval University, an investigation into the use of carbon
and glass FRP sheets to externally strengthen reinforced concrete beams is being conducted
at the Royal Military College of Canada and Queen's University, both in Kingston, Ontario.
The effect on strength and stiffness of the beams subjected to monotonic static loading, cyclic
fatigue loading, and cold temperature testing (-30C) is being studied. Testing has also been
conducted on concrete beams prestressed with aramid FRP tendons [12].
At the Technical University of Nova Scotia, researchers have been investigating the
feasibility of using polypropylene fiber reinforced concrete (FRC) for bridge decks that are
entirely devoid of steel. A number of half-scale models have been tested, and the results
confirm that a FRC slab with inexpensive non-ferrous fibers is feasible, provided that the top
flanges of the steel girders are connected just below the deck by transverse steel straps and
the concrete deck is joined to the girders by shear connectors [13].
Following the encouraging results at the Technical University of Nova Scotia with
polypropylene FRC slabs, the Ministry of Transportation of the Province of Ontario is
conducting tests on polypropylene FRC skew slabs. Currently, the Ontario Highway Bridge
Design Code (1983) doubles the reinforcement in the end zones of skew slabs. The Ministry
is investigating whether end diaphragms, which have a high rigidity in the horizontal plane
and uses shear connectors, could be used with the polypropylene FRC slabs without the need
for steel reinforcement in the end zones. In addition to these skew slab trials, the Ministry
is studying a new barrier wall design to be used with the polypropylene FRC slabs. The
barrier walls will be reinforced with a non-ferrous grid reinforcement, NEFMAC [9], and
would be secured to the slab with a stainless steel, double-headed tension bar. Finally, the
 95

feasibility of using NEFMAC to reinforce the slab overhang is being studied.

At the University of British Columbia, researchers are investigating the impact resistance
of concrete and fiber reinforced concrete with FRP. Using an instrumented impact machine
of the drop weight of the swinging pendulum type, the behavior of beams reinforced with
FRP rods and subjected to very high rates of loading (or stressing) is being studied. It has
been found that these systems have a brittle response under rapidly applied impact loads, but
the use of fibers in concrete appears to be beneficial. As well, researchers are studying the
behavior of concrete beams reinforced with FRP at sub-zero temperatures. It is known that
FRP and concrete are temperature sensitive, and the investigation involves testing flexural
specimens in an environmental chamber at temperatures of about -60C. It has been found
that the brittle behavior of these systems is exacerbated at low temperatures. The next step
will be to investigate the use of fibers in the concrete to counterbalance the effect of low
temperatures.
Numerous projects are underway at the University of Sherbrooke on a variety of FRP
topics. Some of these are the behavior of glass FRP rods for reinforced and prestressed
concrete structures; the use of carbon and glass FRP for the rehabilitation and strengthening
of concrete beams and columns; and the reliability of structures reinforced with advanced
composite materials. The work on glass FRP reinforcements for concrete, including the
experimental work to determine the physical and mechanical characteristics of glass FRP
reinforcement used, is supported in part by Pultrall Incorporated of Thedford Mines, Quebec.
At Carleton University in Ottawa, there is work ongoing in the rehabilitation of concrete
beams using externally bonded carbon fiber and glass fiber sheets, the use of carbon fiber to
reinforce high-strength concrete, the use of polypropylene grids for shrinkage control
reinforcement and crack width control, and the use of polypropylene grids for increasing the
ductility of concrete beams and columns.

6. FRP PRODUCTION IN CANADA

The centers of FRP production in Canada are located primarily in the provinces of
Ontario and Quebec. This is not surprising since these are the provinces where the Canadian
aeronautical industry has been based. There are no carbon or aramid FRP being produced in
Canada, and there is only one Canadian company which commercially pultrudes glass FRP
reinforcements. This is Pultrall Incorporated in Thedford Mines, Quebec. Creative Pultrusions
North Incorporated (formerly EXCEL Incorporated) in Peterborough, Ontario is developing
a glass FRP reinforcement bar. Autocon Equipment Incorporated in Weston, Ontario, has
plans to start producing by the end of 1993 a continuous grid, NEFMAC [9] which is made
of glass, carbon, or hybrid glass/carbon FRP.

7. DESIGN CODES

The first structural design code in Canada which is considering provisions for FRP
reinforcements is the Canadian Highway Bridge Design Code (CHBDC) to be published in
1994. The CHBDC is a result of the success of the Ontario Highway Bridge Design Code
(OHBDC) [14], which was first published in 1979 and provides a state-of-the-art limit states
96

design code to bridge designers. It has been decided that the fourth edition of the OHBDC
will be adopted as the CHBDC 1994. A technical subcommittee on advanced composites has
been formed and is working to determine if there exists sufficient information and experience
to draft design provisions for the use of FRP in bridges. The technical subcommittee is
considering provisions for the use of FRP prestressing tendons for bridge girders and slab
bridges, as well as fiber reinforced concrete and FRP grid reinforcements for bridge decks
and barrier walls.

8. SELECTED ADDRESSES

Programs and Services Canadian Society for Civil Engineering

Advanced Industrial Materials Suite 700, 2050 Mansfield Street
Directorate Materials Branch Montreal, Quebec, Canada H3A 1Z2
Industry, Science and tel 514 842 5653
Technology Canada fax 514 842 8123
235 Queen Street
Ottawa, Ontario, Canada K1A 0H5 FRP International
tel 613 954 3114 c/o Dr. S.H. Rizkalla, Editor
fax 613 954 3079 Associate Dean and Professor
Faculty of Engineering
External Affairs and International Winnipeg, Manitoba, Canada R3T 2N2
Trade Canada tel 204 474 9809'
North Asia Relations Division (PNR) fax 204 275 3773
Japan Science and Technology Fund
125 Sussex Drive Nova Scotia CAD/CAM Centre
Ottawa, Ontario, Canada K1A 0G2 Technical University of Nova Scotia
tel 613 996 0980 P.O. Box 1000
fax 613 943 8167 Halifax, Nova Scotia, Canada, B3J 2X4
tel 902 420 7763
fax 902 422 8380

9. REFERENCES

1 Mufti, A.A., Erki, M.A., and Jaeger, L.G. (editors) "Advanced Composite Materials
with Application to Bridges". The Canadian Society for Civil Engineering, Montreal,
Quebec. 1991. 297 pp.
2 Buckland, P.G. "Advanced Composite Materials with Application to Bridges - Book
Review". Canadian Journal of Civil Engineering, Vol. 19, No. 2, April 1992. pp. 363.
3 Government of Canada. "The JapanrScience and Technology Fund". 1992.
4 Mufti, A.A., Erki, M.A., and Jaeger, L.G. (editors) "Advanced Composite Materials
in Bridges and Structures in Japan". The Canadian Society for Civil Engineering,
Montreal, Quebec. 1992. 172 pp.
5 Neale, K.W. and Labossiere, P. (editors) "Advanced Composite Materials in Bridges
and Structures". Proceedings of the First International Conference on Advanced
 97

Composite Materials in Bridges and Structures. The Canadian Society for Civil
Engineering, Montreal, Quebec. 1992. 705 pp.
6 John A. Bickley Associates Limited. "Market Study". Toronto, Ontario. 1993.
7 Mitsubishi Kasei Corporation. "Product Information of LEADLINE". Japan. 1991.
8 Tokyo Rope Manufacturing Company Limited. "Product information on CFCC
Ropes". Japan. 1991.
9 Shimizu Corporation. "Product information on NEFMAC". Japan. 1991.
10 Pultrall Incorporated. "Product information on ISOROD". Canada. 1991.
11 Deblois, M., Picard, A., and Beaulieu. "Renforcement de poutres en beton arme
l'aide de materiaux composites: etudes theorique et experimentale" (in French).
Proceedings of the 1st International Conference on Advanced Composite Materials in
Bridges and Structures, Canadian Society for Civil Engineering, 1992. pp. 265-275.
12 McKay, K.S. and Erki, M. A. "Flexural behaviour of concrete beams pretensioned with
aramid fibre reinforced plastic tendons". To be published in the August 1993 issue of
the Canadian Journal of Civil Engineering.
13 Mufti, A.A., Jaeger, L.G., Bakht, B., and Wegner, L.D. "The experimental
investigation of FRC deck slabs without internal steel reinforcement". Proceedings of
the 1992 Annual Conference of the Canadian Society for Civil Engineering, 27-29
May 1992. Volume , pp. 293-304.
14 Ontario Ministry of Transportation. "Ontario highway bridge design code". 1983.

10. ACKNOWLEDGEMENTS

The authors gratefully acknowledge the contributions of the following persons in

providing information used in this chapter: Dr. Baidar Bakht, Ontario Ministry of
Transportation; Dr. Nemy Banthia, University of British Columbia; Mr. John Crimi, Autocon;
Dr. Moe Cheung, Public Works Canada; Dr. Leslie G. Jaeger, Vaughan Engineering; Ms.
Kim McCall, Creative Pultrusions; Dr. Aftab A. Mufti, Nova Scotia CAD/CAM Centre; Dr.
K.W. Neale, the University of Sherbrooke; Mr. John Newhook, Vaughan Engineering; Dr.
Andre Picard, Laval University; Dr. Habib Rahman, National Research Council's Institute for
Research in Construction; Dr. A.G. Razaqpur, Carleton University; Mr. Guy Richard, Quebec
Ministry of Transportation; Ms. Leslie C. West, Canadian Association for Composite
Structures and Materials.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 99
1993 Elsevier Science Publishers B.V.

FRP Developments and Applications in Europe

Luc R. Taerwe, Prof. dr. ir.

Magnel Laboratory for Reinforced Concrete

Department of Structural Engineering
University of Gent, B-9052 Gent, Belgium

Abstract
This paper examines the development of non-metallic fibre based reinforcement for
concrete structures in Europe and gives a state-of-the-art of the applications until 1993.
Mainly applications in prestressed concrete members will be addressed. For post-tensio-
ning, glass fibre composite cables were developed in Germany whereas for pretensioning,
aramid fibre composite strips and bars found some applications in The Netherlands.

1. INTRODUCTION

The application of FRP reinforcement for concrete structures is related to a series of

more or less recent developments :
- the steadily increasing number of damaged concrete structures due to corrosion of the
reinforcing and prestressing steel.
- the considerable progress in the field of available composites for industrial applications
resulting in applications of composites in the aircraft and space industries, automobiles,
trains, industrial robots, vessels, piping, etc.
- the possibility of incorporating optical sensors in composites, which offers potential
perspectives for permanent monitoring of concrete structures.

In the 1970s, extensive research by Prof. Rehm at the University of Stuttgart showed
that composite glass fibre bars were ideally suited for prestressed concrete structures [1].
For reinforced concrete the application seemed less appropriate because the low modulus
of GFRP bars caused excessive deflections. The main problem that hampered wider
application was the lack of reliable anchorage devices. The further development of GFRP
is dealt with in the next section. Aramid fibre composites are discussed in section 3 and 4.
Finally, some applications in the field of repair and strengthening will be mentioned
briefly.

2. GFRP BARS FOR POST-TENSIONING

2.1. Development and applications

In 1978, Strabag Bau-AG and Bayer AG established a joint venture for the develop-
100

ment, testing and application of glass fibre bars. The scope of application included all
types and degrees of prestressing, with or without bond, temporary or permanent soil or
rock anchors, as well as tension cables for transmission masts.
In 1980, the latter field of application was put into practice at Munster (Germany)
where 150 tension cables were used with lengths up to 50 m and diameters varying
between 12 en 25 mm. Also in 1980, the experimental programme resulted in the
construction of a short span pedestrian bridge. The "Lnensche Gasse" bridge in
Dsseldorf has a span of 7 m and was prestressed with 12 non-bonded cables.
The first highway bridge, prestressed with cables consisting of glass fiber bars, was
opened for traffic in 1986 in Dsseldorf (Fig. 1) [2,3]. This bridge is a continuous
structure with spans of 21.3 and 25.6 m. A total of 59 tendons, each composed of
19 bars, 7.5 mm in diameter, provide forces of 600 kN per unit. The solid slab has a
width of 15 m. Permanent remote control of individual prestressing tendons, by means of
optical fibre sensors and copper wire sensors, confirmed a perfectly normal structural
behaviour [3]. This monitoring is still continuing.

<.6.90m
21.30 m 25.60 m

ffi B /
JJ.

r^ ^ r ^

15,00 m 1
.4.25 1 6 5 0 . 1. C 25

1 T I

Figure 1. Ulenbergstrasse Bridge in Dsseldorf (Germany).

For the sake of completeness we mention a number of heavy two-span concrete beams,
above a brine tank at the Bayer plant at Dormagen. The extremely aggressive atmosphere
was decisive for the choice of GFRP bars.
In 1988, the system was applied in Berlin in a two-span bridge with a pedestrian traffic
 101

lane and a bridle path (Fig. 2). The double tee cross section is about 5 m wide and the
spans equal 27.6 and 22.9 m. Prestressing was accomplished with seven external tendons,
consisting of 19 glass fibre bars. Also in this bridge, sensors were applied for long-term
monitoring of the bridge.

51.31m

27,61 m J 22,98 m 4

Figure 2. Marienfelde Bridge in Berlin (Germany).

The wide range of application of glass fibre tendons is further demonstrated by their
use in the rehabilitation of the Mairie d'lvry subway station in Paris, France. As a result
of one-sided excavation directly adjacent to the subway station, considerable cracking had
occured in the 70-year-old concrete vault over a length of about 110 m. Thirty-six glass
fibre prestressing ties were installed to strengthen the vault (Fig. 3). The service load per
tendon equals 650 kN. In this application, the electromagnetic neutrality of the tendons
proved to be a favourable property.
In 1991, a three-span highway bridge was completed at the Bayer plant in Leverkussen
(Fig. 4). Span lengths are 2 x 16.3 m and 20.4 m. The 1.10 m thick slab was prestressed
with 27 glass fibre tendons.
In early 1992, the Ntsch Bridge in Austria was opened for traffic (Fig. 5). Span
lengths are 2 x 13 m and 18 m. The thick slab was prestressed with 41 glass fibre
tendons. Both this bridge and the one in Leverkussen were designed according to the so-
called "limited prestress" principle, whereby small tensile stresses are allowed under full
service load.
Also the application in soil anchors (with incorporated sensors) has been investigated.

2.2. Technical properties [5]

The so-called "Polystal" bar consists of 68 percent glass fibres (by volume) and 32
percent resin (modified unsaturated polyester). The glass fibre content by weight is about
80 percent. The basic function of the resin matrix is to hold the fibres together so as to
prevent shear between them, to protect the fibres and to give them sufficient dimensional
stability over the required range of service temperatures. The bars have a diameter of
7.5 mm and are protected by a polyamide coating. In this way, sufficient resistance is
obtained against the alkaline concrete environment. Accelerated exposure tests showed that
the coating is necessary as otherwises rupture occurs after very short exposure times.
102

glass fiber ties

22.45 m

Figure 3. Strengthening of subway station in Paris (France) by glass fibre ties.

Elevation Cross section

Figure 4. Bridge at Leverkussen (Germany).

Plan view Cross section

Figure 5. Ntsch Bridge in Austria.

 103

In Fig. 6, the stress-strain curve of a Polystal bar is compared with those of prestres-
sing steel (tensile strength 1670 MPa) and other types of FRP bars. The plot of the glass
fibre bar is perfectly linear until the ultimate stress is reached, whereby sudden brittle
failure takes place. The tensile strength is comparable to that of conventional prestressing
steel. However, it is lower than the tensile strength of pure glass fibres.

1 2 3
tensile strain in %

Figure 6. Stress-strain curves of FRP-bars and prestressing steel.

The ultimate strain is only 3.3 percent, which is significantly lower than for steel.
However, loading tests on beams show that sufficient strain capacity is generally available
to provide the required flexural ductility. The modulus of elasticity is about one-fourth of
that of steel. An advantage of the lower modulus of elasticity is that losses of prestress
due to creep and shrinkage are significantly reduced. Final relaxation losses remain
limited to 3.2 percent at an initial stress level of 0.5 fpt (extrapolation to 57 years).
104

Long-term sustained loading results in a strength loss of 30 percent, i.e., the long-term
strength is 70 percent of the short-term strength (Fig. 7). As for fatigue, it was found that,
for a mean stress level am = 736 MPa, a stress variation 2 = 55 MPa corresponded to
a fatigue life of 2 x 106 cycles [4]. The coefficient of thermal expansion equals about 7 x
10"6 per deg C.

Relative
Tensile Strength ft ( ) / ft ( 0 )

JTn
1 *^v
I
1^^l T " I
^ XT

0,01 0,1 1 10 100 1000 10000

Time under load (h)

Figure 7. Influence of time under load on tensile strength.

At increased temperatures, the stress-strain curve of glass fibre bars is only slightly
modified, which results in a fire resistance which appears to be sufficient for most
applications.
The bars or tendons are placed in the usual steel sheet ducts, which are grouted with a
resin having appropriate properties for this purpose. The greatest difficulty to overcome
was the development of reliable anchorage devices for cables. Fig. 8 shows the anchorage
for the 19 bar cables. Particular aspects of this type of anchorage are discussed in [6].

2.3. Laboratory tests on structural elements

In 1985, loading tests were performed on two T-beams prestressed with the Polystal
bars, at the University of Stuttgart. The beams with a span of 9 m, failed by crushing of
the concrete. In 1986, a fire test was performed on a concrete member prestressed with
Polystal bars, at the University of Braunschweig. The thickness of the concrete cover was
in accordance with the current codes of pratice. The element failed after 105 min exposure
to 1000 C.
Loading tests on several beams were performed in the Magnel Laboratory of Gent
University in the framework of the BRITE programme, "Monitoring of Prestressing Force
by Integrated Sensors". "BRITE" stands for "Basic Research in Industrial Technologies
for Europe" and is sponsored by the Commission of the European Communities. Different
 105

Figure 8. Anchorage of GFRP cable consisting of Polystal bars.

German research institutes were also involved in the tests, especially in the application of
different nondestructive measuring techniques. The major aim of the research programme
was to test the reliability of these techniques and to compare the results with conventional
deformation measurements.
In the first part of the programme, three beams, with a span of 2 m were loaded up to
failure. In this case, the sensor techniques could be tested for short concrete members. In
the second stage, a long beam with a span of 20 m was tested. The dimensions approach
those of a full sized prestressed concrete girder and thus scale effects are eliminated to a
large extent. In both cases, also conventional bonded reinforcing steel was present, besides
the GFRP prestressing bars.
Fig. 9 shows the test set-up and characteristics of the long beam, which is shown after
failure in Fig. 10. Further details of the tests can be found in [5] and [7].
Although the glass fibre bars by themselves show a brittle behaviour, the tests indicated
that the failure of concrete members prestressed with these bars is accompanied by
extensive deformations, similarly to members prestressed with steel wires or strands.
With respect to the design method, the usual approach can be applied on the condition
that the appropriate initial stress and stress-strain diagram for the tendons are introduced.
This is confirmed by the good agreement between calculated and observed structural
characteristics and deformations [7].

2.4. Sensors
Another important feature of glass fibre bars is that they allow integration of optical
sensors and copper wire sensors. In this way, it is possible to attain permanent monitoring
of the integrity of prestressed concrete structures.
Optical fibres have already been in use for almost a decade in the telecommunications
sector as signal transmitters. The light transmitted by an optical fibre is subject to losses
106

TEST SET - UP CROSS - SECTION

800mm
r" ^

PI IP 150mmJ_ |[ '

1
W77
1 1000 mm
2.4n
1 1 i
A
20 rr] Cable 19 Polystal bars
V d 7.5 mm L
-2dU
200 mm

TENDON PROFILE

1m

QAm 0.4m
0.089 m

777777" 777777
I
20 m

Figure 9. Test setup, dimensions and tendon layout of long beam.

Figure 10. Long beam after failure.

 107

due to absorption and scattering caused by impurities present in the glass. For optical
fibres, light attenuation has been reduced to the lowest possible value. However, when
optical fibres have to be applied as a strain sensor, light attenuation will only be used as a
test signal and, thus, a certain dependency between this attenuation and mechanical
changes in the sensor should exist. Hence, the problem is to select the most appropriate
optical fibre type for the intended purpose. More details on this technique may be found in
[5] and [7].
A monitoring technique which makes use of copper wire sensors was developed by L.
Franke of the Technical University Hamburg-Harburg [8]. Three copper wires are applied
along the outer part of the glass fibre bar and one copper wire is embedded along the
centreline. The electric capacity between the outer wires and the central wire is measured.
Elongation of the bar results in increased capacity and rupture of the bar gives rise to a
sudden drop in capacity which is proportional to the length of the wires and the dielectric
constant of the glass fibre composite.

3. ARAMID FRP ELEMENTS

3.1. Characteristics [9, 10, 11]

In 1983, AKZO and HBG (Hollandsche Beton Groep) decided to cooperate in the
development of aramid fibre based prestressing elements. The composite elements were
fabricated with the so-called Twaron aramid fibres and received the commercial name
Arapree (Aramid Prestressing Element). Arapree consists of bundles parallel Twaron
fibres (mean diameter of 12 ) embedded in an epoxy resin. The composition by volume
is 35 - 45 % Twaron HM and 65 - 55 % resin. The different types are designated with the
number of filaments or fibres. For example, f 100 000 means that the element consists of
100 000 filaments that have a total cross section of 11.1 mm2. Both strips and round bars
have been produced. The surface of the strips show a waffle structure in order to improve
bond with concrete. The surface of the bars is sanded for the same reason. Table 1 gives a
survey of the current Arapree prestressing elements.

Table 1
Arapree-types
Type Shape Dimensions Fibre cross Characteristic
(mm) section (mm2) tensile strength (kN)
f 100 000 bar 0 5.7 11.1 31.1
f 200 000 bar 0 7.9 22.2 62.2

f 100 000 strip 20 x 1.5 11.1 31.1

f 200 000 strip 20 x 2.6 22.2 62.2
f 400 000 strip 20 x 5.0 44.4 124.4
108

Product specifications mention a tensile strength of 3000 MPa (2800 MPa as characte-
ristic value i.e. the 5 % fractile), modulus of elasticy of 125 000 MPa and an ultimate
strain of 2.3 %. A stress-strain diagram is included in Fig. 6. The long-term relaxation is
estimated at 15 % in dry environment and amounts to 20 % in an alkaline solution. The
fatigue behaviour is very favourable. Creep rupture or strength under long-term sustained
stress is shown in Fig. 11 both for air as ambient medium as for an alkaline solution at
20 C.

SUSTAINED STRESS
CHAR. STRENGTH
(%)
-

AI R;2 0C

AU<ALir IES OLLITIOI 1,2 DC

40-

50 100 YEARS
TIME

Figure 11. Stress-rupture behaviour of Arapree.

Arapree elements are not sensitive to the usual types of corrosion as they are resistant
to most aggressive environments. Moreover they are not sensitive to electro-magnetic
currents.
Based on the established mechanical and physical properties and considering a design
service life of 100 years, the initial stress after transfer is limited to 0.55 fptk, were fptk
stands for the short-term characteristic tensile strength.

3.2. Applications
Commercial applications remained fairly limited up to now, and mainly focus on thin
elements where a reduction in cross section can be obtained due to the smaller concrete
cover that is needed compared to similar elements with steel reinforcing bars. However,
due to the fairly short development length, high circumferential tensile stresses develop in
the surrounding concrete, which in turn determine the thickness of the concrete cover.
Moreover it was recognized that the transverse coefficient of thermal expansion of the
Arapree-bars is several times higher than that of concrete. The differential thermal
expansion may cause additional tensile stresses which, combined with those resulting from
 109

the Hoyer-effect may cause severe longitudinal cracks. For this purpose, a compressible
coating was developed that serves as outer layer of the bars.
Arapree found application in the posts of a noise barrier along a highway in Schiedam
(The Netherlands) in 1989. In total 90 posts were prestressed with Arapree strips of the
type f 200 000. The particularly aggressive environment due to deicing salts and exhaust
gasses of cars, was one of the main reasons why Arapree reinforcement was chosen. The
piles have an overall cross-section of 320 x 320 mm or 220 x 220 mm (Fig. 12). The total
length equals 4.2 m, whereby the lower 0.5 m is embedded in a grout-filled steel tube.
Fig. 13 shows the load-displacement curve obtained during a loading test on a barrier
post. The tested element behaved linear until the cracking load of 3.2 kN/m, which is
higher than the design wind load.

Arapree

I I

'

deflection ef point B [mm]

Figure 12. Cross-section of Figure 13. Load-displacement curve

noise barrier post. of noise barrier post.

Thin elements prestressed with Arapree strips were used in a fish-passage along a
hydroelectric power plant on a river in Alphen (The Netherlands - 1990). Both the piles
and shutters were prestressed with Arapree elements (Fig. 14). Weight needed to be
limited because one person should be able to handle the shutters. Previously these
elements were made of tropical hardwood.
Railway sleepers are considered as a potential field of application. A tentative research
programme was performed at the Technical University of Eindhoven [12] (Fig. 15). Use
was made of sanded bars of the type f 200 000.
Next to the so-called "House of the Future" in Rosmalen (The Netherlands), a 4.6 m
high brick wall was axially prestressed with Arapree tendons in order to resist wind
loading.
Near Tokyo, Kajima Corporation erected a stressed ribbon pedestrian bridge with a
54 m span. In the bridge deck, 8 Arapree cables, with a tensile strength of 960 kN each,
were incorporated either bonded or unbonded. Fig. 16 gives a longitudinal section of the
bridge.
110

Figure 14. Piles and shutters used in a fish-passage.

~"Vn r*f~ ..
220
2200 mm

250 o o o o

40 40
prestress H prestress
40 i
lil
i4o ' i
i
'4.73

'
l\ 140 80 I
2.04 (1,1-A) 250 250 1-3
2.07 (III) I 1
Type 1,1-A & III Type II & Il-A

Figure 15. Railway sleepers prestressed with Arapree bars.

Figure 16. Stressed ribbon bridge with Arapree cables.

 Ill

4. VARIOUS DEVELOPMENTS AND APPLICATIONS

"Parafil" ropes (developed by the ICI Company) contain a core of parallel filaments of
yarn with a thermoplastic sheath. A variety of core yarns can be used. In the case of the
Type G Parafil, the yarn used is a stiff polyaramid yarn. The combination of high yarn
strength (2760 N/mm2) and stiffness (126 kN/mm2) makes this version of the rope suitable
for structural applications, particularly prestressing tendons for concrete [12]. The
terminals used for the Parafil ropes have been designed by the manufacturer. Figure 17
shows the terminal for a 600 kN rope, modified for use as a prestressing tendon. The
geometry basically consists of an internal spike which grips the fibres against an external
conical barrel. In this system, every fibre is subjected to an evenly distributed gripping
force, which allows friction to develop the full strength of each yarn. To modify the
terminals for prestressing operations, two threads are placed on the end of the terminals.
The inner thread is used to connect to a pull-rod which is used to apply the prestressing
force, while the external thread is used for a back-nut which transmits the force to the
concrete itself. It is normal pratice, according to the manufacturer's instructions, to
pretension ropes to 60 % of the nominal breaking load prior to use, whenever possible.
This will have the effect that the terminal spike is properly bedded, and will also give a
check on the tendon length before being placed in the structure. Two concrete beams,
prestressed with Type G Parafil tendons, were tested up to failure at Imperial College in
London. Fig. 18 shows the general arrangement of the test beams. More details of the
tests are given in [12].

Figure 17. Terminator of Parafil cable.

In Europe, the external strengthening of reinforced and prestressed concrete members

by epoxy bonded steel plates has become a well-proven technology. To cope with the
danger of corrosion, FRP plates are a promising alternative to steel. In Switzerland, the
use of CFRP plates was investigated [14]. Loading and fatigue tests on concrete girders
were performed in order to validate this strengthening technique. The Ibach Bridge, in the
County of Lucerne, was repaired with this technique. The bridge was designed as a
112

continuous multispan box beam with a total length of 228 m. During core drilling in one
of the webs, a prestressing cable was accidently damaged. The bridge was strengthened
with two CFRP sheets with dimensions of 150 x 5000 x 1.75 mm and one sheet with
dimensions of 150 x 5000 x 2.00 mm. Also a wooden bridge at Sins (Switzerland) was
strengthened with CFRP sheets.
In Germany, research was performed on GFRP plates for bridge strengthening [15].
The technique was applied to the Kattenbusch Bridge, which is a continuous post-
tensioned concrete road bridge with 11 spans (1 = 36.5 m) consisting of two separate
hollow box girders. After several years, wide cracks were observed at the working joints.
These cracks ran through the bottom slab and extended in the webs of the girders. The
main cause of the cracks was restrained deformation due to temperature gradients in
summertime, which was not considered in the original design. Due to the cracks, an
abrupt increase in the stress of the prestressing steel resulted, which could endanger
fatigue strength. Eight joints were strengthened by means of steel plates and two joints
with GFRP plates.
The application of FRP cables for tension-lag off-shore platforms was investigated in
France [16].

^^

A L

(mm)

500

T
A

JL

Figure 18. General arrangement of test beams.

5. EUROPEAN RESEARCH PROJECTS

At several locations in Europe, research is performed on different types of FRP reinforce-

ment for concrete structures. We only mention here the BRITE/EURAM Project "Fibre
composite elements and techniques as non-metallic reinforcement of concrete", that started
in November 1991 and lasts for 4 years. The following partners participate in this joint
 113

research effort, with financial support from the Commission of the European Communi-
ties : SICOM Gmbh, AKZO Fibres and Polymers Division (Enka bv), Hollandsche Beton
Groep nv, the IBMB Institute of The Technical Unversity Braunschweig, The Magnel
Laboratory for Concrete Research of the University of Gent. The following tasks will be
performed :
- Task 1 : Evaluation of the potentials and production technologies of FRP
- Task 2 : Materials research on FRP tensile units
- Task 3 : Materials selection, elaboration of typology and performance profiles for FRP
tensile elements
- Task 4 : Development, modification, and production of FRP tensile elements. Investiga-
tion of force transfer to concrete
- Task 5 : Development of post-tensioning tendons
- Task 6 : Investigation of load bearing behaviour of concrete members prestressed and/or
reinforced with FRP
- Task 7 : Development of criteria for design, detailing, and execution for concrete
members prestressed and/or reinforced with FRP elements
- Task 8 : Tentative plan for a model project

The second "International Symposium on Non-Metallic Reinforcement for Concrete

Structures" is scheduled for August 1995 in Gent (Belgium).

6. PROSPECTS

FRP reinforcement may offer a practical and economical alternative to conventional steel
reinforcement.
Sensor techniques have the potential of becoming useful monitors of long-term behavi-
our both for new and older structures.
It must be appreciated that the application of FRP reinforcement is still in an experimen-
tal stage and that different aspects of this new technology will be the subject of more
detailed investigations. Research is going to increase the knowledge of material characte-
ristics, interaction of non-metallic reinforcement and surrounding concrete, and structural
behaviour of concrete members. Also flexural ductility and long-term behaviour require
further attention.

7. REFERENCES

1 Rehm G. and Franke L., Kunstharzgebundene Glasfaserstbe als Bewehrung im

Betonbau, Die Bautechnik, Vol. 4, 1974, pp. 115-120.
2 Wolff R., Heavy Duty Composite Material for Prestressing of Concrete Structures,
IABSE Paris-Versailles Symposium, IABSE Reports, Vol. 55, 1987, pp. 419-424.
3 Wolff R. and Miesseler HJ., New Materials for Prestressing and Monitoring Heavy
Structures, Concrete International, Vol. 11, No. 9, September 1989, pp. 86-89.
114

4 Miesseler H J . and Levacher F., Monitoring Stressing Behaviour with Integrated

Optical Fibre Sensors, 13th IABSE Congress, Congress Report, Helsinki, June 1988,
pp. 313-318.
5 Taerwe L., Lambotte H., Miesseler H J . , Loading Tests on Concrete Beams Prestres-
sed with Glass Fibre Tendons, PCIJournal, July-August 1992, Vol. 37, No. 4,
pp. 84-97.
6 Faoro M., Development and Conversion of the Properties of ACM into Prestressing
Tendons for Bridges and Structures, in "Advanced Composite Materials in Bridges
and Structures" (Proceedings ACMBS 1), eds. K. Neal and P. Labossi&re, Canadian
Society for Civil Engineering, 1992, pp. 415-423.
7 Taerwe L., Miesseler H J . , Structural Behaviour of Concrete Beams Prestressed with
Glass Fibre Tendons, in "Advanced Composite Materials in Bridges and Structures"
(Proceedings ACMBS 1), eds. K. Neal and P. Labossikre, Canadian Society for Civil
Engineering, 1992, pp. 211-220.
8 Franke L. and Wolff R., Glass Fibre Tendons for Prestressed Concrete Bridges, 13th
IABSE Congress, Congress Report, Helsinki, June 1988, pp. 51-56.
9 Gerritse A., Maatjes E., Schrhoff H., Prestressed Concrete Structures with High
Strength Fibres, IABSE Reports, Vol. 55, Zrich 1987, pp. 425-432.
10 Gerritse A., Werner J., First application of Arapree, in "Fibre Reinforced Cements
and Concretes : Recent Developments", eds. R.N. Swamy and B. Barr, Elsevier
Applied Sciences, London, 1989, pp. 50-59.
11 Gerritse A., Schrhoff H., Prestressing with Aramid Tendons, Technical contribution
to FIP 10th Congress, New Delhi, 1986.
12 De Sitter W., Slooten W., Tests on Railway Sleepers Prestressed with Arapree and
Epoxy Coated Steel, Report No. BREU 4-92, 1992.
13 Burgoyne C , Tests on beams prestressed with polyaramid tendons, in "Advanced
Composite Materials in Bridges and Structures" (Proceedings ACMBS 1), eds. K.
Neal and P. Labossi&re, Canadian Society for Civil Engineering, 1992, pp. 231-239.
14 Meier U., Deuring M., Meier H., Schwegler G., Strengthening of Structures with
CFRP laminates : Research and Applications in Switzerland, in "Advanced Composite
Materials in Bridges and Structures" (Proceedings ACMBS 1), eds. K. Neal and P.
Labossiere, Canadian Society for Civil Engineering, 1992, pp. 243-251.
15 Rostasy F., Hankers C , Ranisch E.H., Strengthening of R/C and P/C - Structures
with bonded FRP plates, in "Advanced Composite Materials in Bridges and Structu-
res" (Proceedings ACMBS 1), eds. K. Neal and P. Labossiere, Canadian Society for
Civil Engineering, 1992, pp. 253-263.
16 Chabert A., Creton B., Jartoux P., Des Materiaux Nouveaux pour la Precontrainte et
le Renforcement des Ouvrages d'Art, Annales de l'Institut Technique du Btiment et
des Travaux Publics, No. 496, Septembre 1991, pp. 85-111.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 115
1993 Elsevier Science Publishers B.V. All rights reserved.

An Overview of R&D in Japan

Yasuhisa Sonobe

Institute of Engineering Mechanics, University of Tsukuba, 1-1-1, Tennohdai,

Tsukuba-city, Ibaraki, 305, Japan

Abstract
In this paper, history, R&D efforts in university, Government and industry, commer-
cially available products, institutional efforts, national research project and research
needs in Japan are described.
R&D of FRP concrete structures in Japan started in the 1980fs. Although some
bridges and other structures with FRP reinforcement have been constructed, FRP
concrete structures are not yet popular in Japan. Since 1988, the comprehensive national
research project has been carried out for the use of new materials in the construction
field sponsored by the Ministry of Construction with 5 year duration. In the project,
researches for P.C. bridges with FRP cables as tendons and concrete building structures
with FRPR are included to draw the guidelines of design and construction.
In the section of the national research project in Japan, the research organization,
research items, test results and code acceptance are introduced.

1. HISTORY

It could be said that R&D of FRP concrete structures in Japan started in the 1980's,
though a paper [1] on the retrofit of reinforced concrete columns with FRP was
presented in 1978.
The first research paper [2] on the FRP reinforcement as tendons for prestressing in
Japan was presented by Prof. Kobayashi et al. in 1984 at the 6th annual meeting of J O .
The first application of CFRP as tendons for the real bridges, called Jingu-bashi, 7.0m
wide and 5.6m span, was constructed in 1988. The first and still only one approval of
the Ministry of Construction for main building structural members, i.e., beams which
were used on the ground floor of a 3 story apartment house isolated at the base from
earthquake, has been given in 1991. Since 1988, the Society for Research of Composite
Materials for Reinforcing Concrete using Continuous Fibers, which consists now of 39
general contractors, material makers and related companies and is called the CCC
Society, has encouraged activities of Architectural Institute of Japan (AIJ) and Japan
Society of Civil Engineers (JSCE) by entrusting research items (see section 4).
116

Since 1984, the Ministry of Construction has investigated the application of new
advanced materials and began in 1988 a comprehensive technological development
project, which is hereafter named as the National Research Project, to examine the
possibility of the use of new materials in the construction field and to establish
technological guidelines of their use. The project contains the study of FRP concrete
structures and is to last for 5 years (see section 5).
A list of references of FRP concrete construction [3] is shown in Concrete Journal,
JCI, where a total of 300 papers are referenced. However, the 20 dated before 1986 are
selected ones. Time history distribution of references is shown in Table 1. The number
of presented papers in Japan is increasing year by year, that is, 25 in 1987, 53 in 1988,
65 in 1989, 80 in 1990, 30 in 1991 (till June).
In parallel with the project, another national research project, which is called New RC
Project and deals with high strength concrete of 60 - 120MPa and high strength steel
reinforcement of yield point of 500 - lOOOMPa, is running for the same term. By the
fruit due to the New RC Project highrise reinforced concrete buildings of concrete
strength of 60MPa and of steel reinforcement of yield point of 700MPa are becoming
popular in Japan.

Table 1 Distribution of references

Items 1986 1987 ~ Jun. 1991

Japan Foreign C. Japan Foreign C.
General - - 4 2 +1*
Material 3 - 79 6 +1*
Structure 7 9 142 11 +3*
Durability - - 16 -
Others - 1 13 1 +1*
Total 10 10 254 20 +6*
* presented abroac by Japanese auth<3rs

2. R&D EFFORTS IN UNIVERSITY, GOVERNMENT AND INDUSTRY

There are about 95 universities and institutes of technology in Japan which have the
civil and/or the building engineering departments. 20 in the civil engineering and 10
in the building engineering are involved in R&D of FRP concrete construction.
As for main governmental research institutes, Public Works Research Institute and
Building Research Institute, the Ministry of Construction, Port and Harbor Research
Institute, the Ministry of Transportation, Japan Highway Public Cooperation and
Railway Technical Research Institute are engaged in R&D.
 117

Some leading general contractors which are strong developers in this field, such as
Shimizu and Mitsui, keeping the close relation to a certain fiber or fiber rod maker, has
proceeded their R&D, sometimes in cooperation with universities and prestressed
concrete companies.

3. COMMERCIALLY AVAILABLE PRODUCTS

There are 15 makers of continuous fiber reinforcement at present in Japan as shown

in Table 2 [4]. If limiting kinds of fibers and binders to carbon, aramid and glass and
plastics, respectively, the number is reduced to 13.

Table 2 Continuous fiber reinforcement in Japan

Developer and Maker (*) Fiber Type Binder Configuration

1 Obayashi-gumi(G.C) Carbon(Pitch) Epoxy Round,
Mitsubishi-kasei(F.M.) Deformed
2 Kajima(G.C) Carbon(Pan,Pitch), Epoxy 3 Dimension
Arisawa(B.M.) Aramid, Vinylon
Akzo(B.M.+F.M.) Aramid Epoxy Plate
3 Shimizu(G.C) Glass(E), Vinyl-ester 2 & 3 Dimension
Dainihon-glass(B. M.) Carbon(Pan) Grid(NEFMAC)
4 Kumagai-gumi(G.C) Carbon(Pan), Epoxy, PPS Deformed,
Arisawa(B.M.) Glass(E) Spirally wound
5 Taisei(G.C) Carbon(Pan) Special Cement Plate, Shell Plate,
Toyo-Rayon(F.M.) and Silica Fume Rod
6 Mitsui(G.C) Aramid, Epoxy Braided(FiBRA)
Dupon, Tore Kepler(F.M.) Carbon(Pan)
7 Tokyo-Seiko(B.M.) Carbon(Pan) Epoxy, Twist(CFCC)
Toyo-Rayon(F.M.) Bismareimid
8 Sumitomo(G.C), Teijin(F.M.) Aramid Vinyl-ester Deformed
9 Mitsubishi-Rayon(F,B.M.) Carbon(Pan) Epoxy Round, Deformed
10 Tore(F,B.M.) Carbon(Pan) Epoxy Twilled
11 New Nippon Steel(S.M.) Carbon(Pitch), Epoxy Mesh
Kanebo(F,B.M.) Aramid
Carbon Epoxy Twist
12 Kurare(F,B.M.) Vinylon Epoxy Twilled, Braided
13 Okumura-gumi(G.C) Carbon(Pitch), Vinyl-ester Spirally wound
Showa-kohbunshi(F,B.M.) Aramid, Glass(E)
14 Takenaka-komuten(G. C.) Carbon(Pitch,Pan), Epoxy 3 Dimension
Aramid
15 Osaka-Gas(F.M.) Carbon(Pitch) Epoxy Brd.,Round,Mesh
* G.C: general contractor, F.M.: fiber maker, B.M.: bar maker, S.M.: steel maker
118

4. INSTITUTIONAL EFFORTS

Trends of committees on the application of the continuous fiber bars as reinforcing

materials in Japan are summarized in Table 3. Some details of comprehensive techno-
logical development project (the National Research Project) are described at section 5.
A report for entrusting from the CCC Society was completed by the Research
Committee of AD, 1991, and was entitled "Items and Methods of Assessment of New
Reinforcing Materials and New Reinforced Concrete with the Materials".
A report for entrusting from the Society was published by the Research Sub-
Committee of JSCE in a book form, Concrete Library 72, entitled "Application of
Continuous Fiber Bar Reinforcement to Concrete Structures", and consisting of 4 parts
as follows: 1) Concept of structural design, 2) Testing method, 3) Durability and
4) State of the art report. At the same time of its publication, April 1992, a symposium
on application of FRP reinforcement for concrete structures sponsored by JSCE was
held. A total of 54 papers, including 3 invited ones, were presented at the symposium.
Two of the invited papers were of the National Research Project, that is, infrastructure
and building engineering, and the rest was presented by Prof. A. Nanni. The papers
were composed of ones related to the trust by the CCC Society and of general
researchers. Unfortunately the papers in building engineering field were hardly
presented. JSCE will organize a new research committee to establish the guideline of
structural design by entrusting again from the CCC Society.

Table 3 Trends of committees on the application of the continuous fiber bars for
structures in Japan
Promoters Committees

Building Research Institute Comprehensive Technological Development Project (1988-

The Ministry of Construction
1992)
Entrusted Research: Working Group for New Structural
Design Procedure on Continuous Fiber Bar Reinforced
Concrete (Chief examiner; Y. Sonobe, secretariat; Japan
Institute of Construction Engineering)
Joint Research: R&D for Reinforcing Substitutes for Steel
Bars to Concrete with High Ability; Person to person
contract, eight general contractors in the core (1989-1990)
and the CCC Society (1991-1992)
For Research of Common Problems: Researching Commit-
tee of Continuous Fiber Bar Reinforced Concrete (Chairman
Y. Sonobe, secretariat; Japan Association for Building
Research Promotion) with 16 companies (1988-1990), and the
CCC Society (1991-1992)

Public Works Research Institute Comprehensive Technological Development Project (1989-

The Ministry of Construction 1992)
Joint Research: Development for Technological Application
of New Material to Bridge Cable; Person to person contract,
10 companies with general contractors in the core
 119

Table 3 Trends of committees on the application of the continuous fiber bars for
structures in Japan (continued)
Promoters Committees

Society for Research of Composite Material for Reinforcing

Concrete using Continuous Fibers (the CCC Society, 1988-,
Chairman; M. Sugita)
Entrust
Architectural Institute of Japan |-Research Committee for Continuous Reinforced Concrete
(1988-1990, Chairman; K. Kishitani)
Structural W.G. (Chief Examiner; S. Mochizuki)
Material W.G. (Chief Examiner; F. Tomozawa)
Japan Society of Civil Engineering -Research Sub-committee on Continuous Fiber Reinforce-
ment (1989-1991, Chairman; T. Okamura)

ACC Club

Basic Industries Bureau Round-Table Conference of New Material for Large Scale
The Ministry of International Structure (1988-1989, Chairman; T. Ojima)
Trades and Industries

Japanese Society of Steel Con- Research Sub-committee for Application of New Material
struction to Structural Members (1988- , Chairman; T. Okukawa)

Kyusyu Association for Bridge and Section Meeting for Research of Application Material to
Structural Engineering Highly Technological Fiber (1990-1993, Chairman;Y. Sakamoto)

Japan Ocean Industries Association R&D Committee for the Materials for Off-shore structure
(1991-1996, Chairman; F. Tomozawa)

Japan Technology Transfer Asso- Technological Committee to Improve the Durability of Off-
ciation shore Structure (1991- , Chairman; M. Sakuta)

5. NATIONAL RESEARCH PROJECT OF JAPAN

5.1. General
The organization of the National Research Project on the use of new materials in the
construction field, sponsored by the Ministry of Construction, is shown in Figure 1.

National Research Project

Infrastructure | Building Engineering |

TT X -L
Bridge(PC) Metallic Materials Non Metallic Materials

The Use of Short Fiber The Use of Continuous Fiber New Functional Materials
in Concrete Construction in Concrete Construction
(RC and PC)

Figure 1 Organization of the National Research Project

120

The project is separated into 2 sections, that is, one is infrastructure and the other is
building engineering, which is further divided into 2 divisions, i.e., the use of metallic
materials and non metallic materials. The latter division has 3 working groups, which
are: the use of short fiber in cemented boards, the use of continuous FRPR in concrete
construction and new functional materials which contain durable paint finishing. The
metallic division has also 3 working groups, which are: the use of high strength steel
(585MPa), the structural use of stainless steel and the use of aluminum alloys and
vibration isolators.

5.2. General of infrastructure engineering concerning FRPR

In the infrastructure section, the prestressed concrete bridges using FRP cables as
tendons are now being studied by Public Works Research Institute in cooperation with
10 companies which are general contractors in the core. A schedule of research items
is shown in Table 4. Establishment of guidelines of design and construction of
prestressed concrete bridges using FRP cables as tendons is being programmed.

5.3. General of building engineering concerning FRPR

Objectives of the working group for the use of continuous FRPR in concrete con-
struction are to establish standard testing methods for materials and to draw guidelines
of structural design. But the project was not sufficient in funds for experimental works
which would give back up data to produce the standards and the guidelines. Then an
executive committee for experiments was to run parallel with the working group, rather
together, in cooperation with universities, the ministry and related 16 companies at the
beginning of the project and the CCC Society instead of the companies for its last 2
years, which have taken financial support for common experimental works. As a matter
of fact the main theme given to the project is the research on concrete structures
reinforced with continuous fiber reinforcement substituting for steel bars and tendons.
We have recognized that concrete structures prestressed by FRP tendons are more
promising than the ones without prestressing, and that all of them would be effective
under extraordinary environment such as deicing salt, corrosive action of sea or attack
by chemicals. As prestressed concrete bridges are to be studied in the section of infra-
structure program in the project, it is expected that its results can be interpreted with
building applications.
We have used FRPRs which are available in Japan, and limited fiber types to carbon,
aramid and glass in consideration of their mechanical properties. In the experiments,
all the materials were tested in the same conditions and structural members had the
same cross sections and were loaded in the same way. However, the purpose is to
characterize the feature of every type of FRPR, but not to compare them each other.
It was expected that there would be many difficult problems with the application of
FRPR in concrete construction. Fortunately, because of the National Research Project
and the infancy of the technology, we were to research the subject academically in
detail, expecting further research and development in the future, while somewhat
disregarding the feasibility issues.
 Table 4 Schedule of research items (infrastructure)

1989 1990 1991 1992 Final Output

Investigation of Function of
FRP Cable to be Required

Examination of Quality Characteristics of

FRP Cable

Examination of Sheath for FRP Cable by

Loading Test

Examination of Anchoring Method of

FRP Cable by Loading Test

Examination of Durability of Prestressed Concrete Beam

using FRP Cables as Tendons

Guideline of
Examination of Load Bearing Capacity of Pre- Design and
stressed Concrete Beam using FRP Cables as Tendons Construction
of Prestressed
Guideline of Design Concrete Bridges
and Construction using FRP Cables
as Tendons
122

The executive committee consists of 4 sub working groups. They are:

1) Structural design and structural experiments (chaired by the author)
2) Prestressed concrete (chaired by Prof. Y. Matsuzaki)
3) 2 and 3 dimensional reinforcement (chaired by Prof. S. Mochizuki)
4) Material, durability and fire resistance (chaired by Prof. A. Shimizu)

5.4. Research items, test results, code acceptance, etc. of building engineering
A tentative guideline of structural design is written in the form of limit state design.
Carried out and scheduled research items are shown in Table 5 with related contents
of the guideline and in Table 6 showing classification into 5 categories.
Representative test results of the National Research Project are to be presented at the
International Symposium on FRPR of ACI, 1993, spring convention.
As for FRPR bar, general and ordinary material tests are carried out at the laboratory
base. Surveys for wall, column and beam-column joint are not adopted in the project
because of our reasons. To cover the latter 2 items mentioned above, a lateral loading
test of a frame of 3 stories and 2 bays is now being conducted.
To establish the concept of structural design of building of FRP reinforced concrete,
great obstacles in the way for us who are accustomed to ordinary RC structures would
be low fire resistance and lack of yielding or plastic flow of FRPR bars. Structural
design of usual building frame of concrete construction in Japan emphasizes on earth-
quake resistance design. Therefore, if FRP reinforced concrete will not be able to have
earthquake resistance, it can not play a role in primary. Otherwise an earthquake
isolation system will be required.
The general concept of earthquake resistance design in Japan is as follows. Highrise
building frames are designed to have sufficient ductility of translational angle per story
of about 1/100 - 1/50 radians against strong earthquake having maximum velocity of
50cm/sec. For lowerrise building frames, moderate lateral strength and ductility, which
is derived from yielding of steel reinforcement, are required. For lowrise and rigid
building frames sufficiently large lateral strength such as base shear coefficient of 0.50
is required.
Building members which are supposed to have little relation to earthquake action are,
1) Slabs and beams which are supported by girders
2) Non structural members such as curtain wall
3) Foundation slabs
If the scope of applications of FRP reinforced concrete is limited within building
members just mentioned above, drawing of guideline of structural design would be
much easier, though problems of serviceability and fire resistance remain.
New type structures of infrastructures can be realized under approval of related
institutions, such as JSCE, PWRI and JHPC, and MITI in certain cases, according to
provided design procedures. On the other hand, design and practice of buildings are
regulated by the law, i.e., standard building code, called Kenchiku Kijun Ho, provided
by Japanese Government. It is necessary to get the approval of the Minister of
Construction to use new material and new construction method according to the article
 123

Table 5 Research items versus guideline (building engineering)

Content of Tentative Guideline Research Items

1. General
1.1 Scope
1.2 Nomenclature
1.3 Symbol
2. Design Method
2.1 Design Philosophy
2.2 Safety Factors and Design
3. Material Test of Tensile Strength
3.1 General Bond Strength by Simple Pull Out Test
3.2 Material, Kind, Quality and Coefficients Chemical Resistance
of Mechanical Property of Concrete Creep under Tension
3.3 Quality, Configuration, Dimension and Compressive Strength and Behavior under
Coefficients of Mechanical Property of Tension and Compression Reversals and
FRPR Cyclic Tension
3.4 Other Materials Tensile Strength at High and Low
Temperature
4. Loads and Their Combination
4.1 Loads
4.2 Combination of Loads
5. Stress and Deformation
5.1 Fundamentals of Analysis of Stress and Strain
5.2 Analysis of Stress and Deformation of Frames
5.3 Analysis of Stress and Deformation of
Slabs and Beams
6. Ultimate State Design |- Compressive Strength and Behavior under
6.1 General Tension and Compression Reversals and
6.2 Examination of Members under Ultimate Cyclic Tension
State Effect of Confinement by Lateral Rein-
(1) Ultimate Strength for Axial Force forcement
Flexural Behavior of Simple Beams
Flexural Behavior of Beams Subjected to
(2) Ultimate Flexural Strength- Antisymmetrical and Cyclic Loading
Evaluation of Crack Width
L
* Tension Stiffening Effect
Shear Behavior of Beams Subjected to
Antisymmetrical Loading
(3) Ultimate Shearing Strength- Tensile Strength of Bent Corner
L
Behavior of Bars under Tension Combined
Shear
124

Table 5 Research items versus guideline (building engineering) (continued)

Content of Tentative Guideline Research Items

Bond Splitting Behavior by Simplified

Method
(4) Ultimate Bond Splitting Strength- Bond Splitting Behavior by Cantilever
Method
Bond Splitting Behavior of Beams
Subjected to Antisymmetricai and
(5) Ultimate Torsional Strength Cyclic Loading

6.3 Examination of Frames under Ultimate- Structural Behavior of Frame

State
Flexural Behavior of Simple Beams
Flexural Behavior of Beams Subjected to
6.4 Examination of Floor Slabs and Beams Antisymmetricai and Cyclic Loading
under Ultimate State Shear Behavior of Beams Subjected to
Antisymmetricai Loading
7. Serviceability State Design Tensile Strength of Bent Corner
7.1 General Evaluation of Crack Width
7.2 Examination of Crack Flexural Behavior of Simple Beams
7.3 Examination of Deflection Behavior of RC and PC Beams Subjected
7.4 Examination of Vibration to Long-Term Loading
Flexural Behavior of Simple Beams
8. Structural Detail

Appendix
1. Standard Test Method-
2. Example of Structural Design-
3. Evaluation Standard of Durability-
4. Evaluation Standard of Fire Resistance-
Examination of Testing Method of
Tensile Strength
Structural Behavior of Frame
Heat Cycle at Dry and Wet State by Bond
Test
Chemical Resistance
Behavior of RC and PC Beams Subjected
to Long-Term Loading
Tensile Strength at High Temperature
Flexural Behavior of Beams under Heating
Flexural Behavior of Beams after Heating
 Table 6 Schedule of research items (building engineering)

Category 1988 1989 1990 1991 1992 Final Output

Fundamental Test Methods
Properties of Investigation Examination of Testing Method of Tensile Strength of FRPR Bar
FRPR Bar by Documents

Bond Strength by Compressive Strength and Behavior

Simple Pull Out under Tension and Compression
Test Reversals and Cyclic Tension

Creep under Tension

Chemical Resistance

Tensile Strength at High Temperature Tensile Strength at High Temperature

without Concrete Covering with Concrete Covering

Tensile Strength at
Low Temperature

Performance of Technical Data

Structural Investigation Bond Splitting Behavior by for Structural
Members of by Documents Simplified Method Performance on
FRP RC and FRP RC and
PC Bond Splitting Behavior by PC
Cantilever Method
 Table 6 Schedule of research items (building engineering) (continued)

Category 1988 1989 1990 1991 1992 Final Output

Performance of Technical Data
Structural Flexural and Bond Splitting Behavior of Beams Subjected to for Structural
Members of Antisymmetrical and Cyclic Loading Performance on
FRP RC and FRP RC and PC
PC Tension Stiffening Evaluation of
Effect Crack Width

Flexural Behavior of Simple Beams Flexural Behavior of

Subjected to Cyclic Loading Prestressed Beams

Shear Behavior of Beams Subjected to

Antisymmetrical Loading

Tensile Strength Behavior of Bars

of Bent Corner under Tension and
Combined Shear

Effect of Confine- Evaluation of Effective Prestress:

ment by Lateral Test of Friction Coefficient and
Reinforcement Relaxation

Evaluation of 2 and
3 Dimensional FRPR

Structural Behavior
of Frame
 Table 6 Schedule of research items (building engineering) (continued)

Category 1988 1989 1990 1991 1992 Final Output

Durability of Technical Data for
Structural Heat Cycle at Dry Heat Cycle at Dry and Wet State by Durability on
Members and Wet State by Bond Test FRP Reinforced
Flexural Test Concrete

Behavior of RC and PC Beams Subjected to Long-Term Loading

Fire Resistance of Technical Data for

Structural Flexural Behavior of Beams after Heating Fire Resistance on
Members FRP Reinforced
Flexural Behavior of Beams Subjected to Concrete
Vertical Load under Heating

Structural Design Guideline of

Design Procedure for FRP Reinforced Concrete Structural Design
128

38 of the code, to which the use of FRP reinforcement and its reinforced concrete are
applied. Actual works of the approval are done at appraisal committees concerned
which are organized at Building Center of Japan in the Minister's behalf. Technical
data, standard testing methods and the guideline of structural design, which are to be
derived from the National Research Project, are to be referred at the appraisal
committee concerned, though the persuasion of the leading staff of the Ministry of
Construction about important and serious items is inevitable.

6. RESEARCH NEEDS

After finishing the National Research Project, the fruit of research of FRP reinforced
concrete of building engineering is to be published and following institutional activity
research will again begin at AIJ.
To develop FRP concrete constructions, to get its social approval and to overcome
problems of cost performance, it is very important to increase the number of realized
structures step by step. Even in case of ordinary reinforced concrete, which has history
of 130 years, research works such as the New RC Project are still being carried out.
It has only taken a decade or so from very beginning of this art in Japan. We have just
traced out the way of RC. It might need more than a decade to make FRP concrete
structures popular in the construction field with much R&D. Therefore, prospective
objectives must be looked for and realized in consideration of the merits of FRP
reinforcement such as excellent durability, high tensile strength density ratio, low
density, etc. by appropriate and reliable research.

7. REFERENCES

[1] K. Yamamoto, M. Kawakubo: Test of shear strengthening effect of existing RC

columns by FRP, Summaries of Technical Papers of Annual Meeting, AIJ, No.
2620, pp. 1749-1750, 1978
[2] K. Kobayashi: Developmental research of FRP reinforcement as tendon of
prestressed concrete, Proceedings of JCI, vol. 6, pp. 369-372, 1984
[3] C. Hashimoto: References for FRPR and FRP Reinforced Concrete, Concrete
Journal, JCI, vol. 29, No. 11, pp. 104-116, 1991
[4] Japan Society of Civil Engineering: Application of Continuous Fiber Reinforcement
to Concrete Structures, Concrete Library 72, 1992
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 129
1993 Elsevier Science Publishers B.V. All rights reserved.

FRP Development in the United States

Charles W. Dolan, P.E., Ph.D.

Associate Professor
University of Wyoming
Department of Civil and Architectural Engineering
Laramie, WY, 82070, United States of America

Abstract
This report examines the development of nonmetallic fiber based reinforcement in the
United States. Much of the research is original and some is influenced or supported by
international activities. The current lack of use of these materials in the United States belies
the fact that considerable knowledge exists regarding FRP characterization and performance.
To put the combined experience into perspective, FRP applications are grouped into
functional categories. These categories include; characterization of FRP reinforcement,
reinforced concrete structures, prestressed concrete beams, and the mechanics and design of
FRP reinforced members. Research needs are identified based on the current development
of FRP reinforcement both in the United States and worldwide research.

1. INTRODUCTION AND HISTORICAL DEVELOPMENT

The United States has a long and continuous interest in fiber based reinforcement for
concrete structures. The concept of short glass fiber reinforcement in cements was first
introduced in the 1930's By the 1950's and early 1960's the U.S. Army Corps of Engineers
was sufficiently interested in long glass fibers for reinforcement that a series of
comprehensive reports was complied under the direction of Bryant Mather [Mather, 1955,
Pepper, 1959, Wines 1966, & 1966]. While these reports existed, research and applications
were limited. In the 1970's corrosion of concrete structures, particularly bridge decks, led
to a renewed interest into design strategies that reduces structural susceptibility to corrosive
environments. By the 1980's the corrosion resistant properties of nonmetallic bars and
tendons took on a renewed interest. The National Bureau of Standards (now the National
Institute of Standards and Technology) examined non-metallic rods for antenna guys and in
130

the process conducted some of the first research into anchorage of fiber reinforced plastic
(FRP)composite rods [NBS, 1976].
The development of FRP reinforcement in the United States has been the result of
individual investigators and two national initiatives. Unlike Japan, there is no coordinated
national research program, and unlike the United Kingdom, Germany, and Holland, there
have been no major corporations sponsoring extensive FRP product development. Even
with a lack of coordinated effort, significant research and development has been completed.
Drs. Jack Scalzi and Ken Chong at the National Science Foundation have coordinated and
sponsored several independently submitted proposals from institutions such as such the
University of Arizona, University of Michigan, Massachusetts Institute of Technology,
Catholic University of American, Pennsylvania State University, University of California,
Long Beach, West Virginia University and the University of Wyoming [FRP, 1993]. These
projects have led to a comprehensive understanding of FRP material behavior and structural
mechanics of concrete reinforcement. The U. S. Department of Transportation's Federal
Highway Administration (FHWA) is supporting research into FRP performance as their
Research and Development budgets will allow. The FHWA research into composite
materials is under the overall direction of Mr. Eric Munley. Major projects at the University
of California, Long Beach, University of California, San Diego and the University of West
Virginia have examined FRP reinforcement and FRP composite construction. In 1993,
FHWA took a major step forward by initiating research into accelerated aging and
standardized testing of FRP materials. Florida State Department Of Transportation is
sponsoring as series of durability studies and the California Department of Transportation is
using FRP materials to strengthen bridge columns. Both of these projects are discussed in
Section 9, Applications.
In January 1991, the American Society of Civil Engineers sponsored a conference on
Advanced Composites Materials in Civil Engineering Structures [Iyer and Sen, 1991]. The
conference examined many active research projects using FRP reinforcement. The Third
Bridge Engineering Conference, Sponsored by the Transportation Research Board included
presentations of FRP reinforcement and prestressing research [TRB, 1991]. The
Transportation Research Board formally established a committee to examine the use of FRP
in bridge structures. This committee meets annually at the TRB Conference in Washington
DC and has serves as valuable conduit for researchers, designers and interested parties to
follow the development of FRP research in bridge structures. The American Concrete
Institute (ACI) formed Committee 440, FRP Reinforcement, to investigate and promulgate
knowledge regarding the use of FRP in concrete structures. In March 1993, ACI sponsored
an International Symposium on FRP in Concrete Structures [Nanni and Dolan, 1993]. Over
60 technical papers were presented or published. ACI Committee 440 is developing a state-
of-the-art report on the use of FRP reinforcement and the American Society of Civil
Engineers has established a committee to prepare standards for FRP products.
This report addresses seven significant functional area. The report begins with the
theory and behavior of individual fibers, then examines the assembly of these fibers into fiber
reinforce plastics to be used as reinforcement. The behavior and mechanics of FRP
reinforcement in concrete is examined next followed by the behavior and mechanics of FRP
reinforced and prestressed concrete. Design guidelines for FRP reinforced concrete
 131

structures is reviewed, prototypical structures are examined and research needs are
identified.

2. CHARACTERIZATION OF FRP REINFORCEMENT

FRP reinforcement is typically a composite rod made of a high modulus fiber and a
resin matrix. Fibers include aramids, carbon and glass. The resins include polyester,
vinylester and epoxy. A singular exception to this FRP fabrication is the Parafil tendon,
which is a bundle of unbonded Kevlar1 fibers encased in a polyethylene sheath. The FRP
rods are produced by a pultrusion method. That is the fibers are bundled together and drawn
through a resin mix then pulled through a shaping die. As the rod emerges from the shaping
die, it passes through a curing chamber where the resin is allowed to harden, Figure 1. The
pultrusion process allows considerable latitude in the selection of a structural shape. While
circular rods are most common, flat bars and other shapes have been manufactured.

Pulling device
fiber spool mmmmm e-
die curing chamber

a a takeup spool
Resin bath
Figure 1. Schematic of pultrusion process

The pultrusion process is compatible with all the high modulus fibers and with a wide
variety of resins. Reinforcement rods of fiberglass, aramid, and carbon have all been
fabricated. It is the variety and possible combination of fibers and resins that make the
characterization of the FRP reinforcement difficult. Since the bonding and durability
characteristics of the resins differ, reinforcement with identical fibers may respond quite
differently with changes in the resins. An examination of the bare fiber properties and test
procedures provides an appreciation of how the FRP performance can vary.

2.1 Individual Fiber Properties

Three high modulus fibers are generally considered for FRP reinforcement; aramid,
carbon and glass. The desirable mechanical characteristics of these fibers are high elastic
modulus, low creep and relaxation, stability at normal ambient temperature and high tensile

Kevlar is a registered trademark of the E.I. DuPont company and describes their aramid based fibers.
132

strength. These properties are obtained by drawing the fibers to very small diameters.
Individual fiber diameters are on the order of 5 to 10 microns. The fibers often have
relatively low shear strength. While steel has a shear capacity of approximately one third of
its tensile capacity, some fibers have shear strength of only five percent of the tensile
strength. Representative individual fiber properties are given in Table 1.
High modulus fibers exhibit a nearly linear stress strain relationship up to their
ultimate tensile capacity, Figure 2. Within any spool of fiber there is a variation in the fiber
diameter and the fiber strength. This leads to statistical variation of strength within the fiber
bundles. Since there is no fiber yielding, the strength is reflected as a variation in ultimate
tensile strength of the individual fibers, the fiber bundle and the FRP reinforcement. Two
other factors affect the strength of the fibers; defects in the individual fibers and sustained
load capacity. Defects in the fiber will create weak points. Therefore, as the fiber length
grows longer, the probability of a defect occurring occurs and the strength of an individual
fiber diminishes. Simultaneously, as the number of fibers in the bundle increases, the
opportunity for an individual fiber failure to initiate a progressive failure of the bundle
decreases. Phoenix [Phoenix, 1979] has been a leading developer of the Weibul distribution
theory for multiple fiber bundles. Secondly, high modulus fibers cannot sustain their ultimate
tensile over long periods of time. This phenomenon is called creep-rupture and prevents
high modulus fibers from being used for long periods at much more than 80 percent of their
static strength. Thus, the sustained load capacity of the fiber bundle will be less than the
theoretical summation of the individual fiber strength. This strength reduction is
demonstrated in tests and in commercial literature [DuPont, 1986].

Table 1
REPRESENTATIVE PROPERTIES OF HIGH MODULUS FIBERS

Fiber Specific Tensile Tensile Coefficient Strain

Gravity Modulus Strength of Thermal to
Expansion Failure

E-Glass1 2.58 10,500 ksi 500 ksi 9.0xlO"6/OF 2.4%

Aramid^ 1.44 19,000 ksi 525 ksi -3.6xlO"6/F 2.8%
Carbon^ 1.80 34,000 ksi 526 ksi 1.6%

E-Glass1 2.58 72.5 GPa 3.45 GPa 5.0xlO- 6 /K 2.4%

Aramid^ 1.44 131 GPa 3.6 GPa -2.0xlO"6/K 2.8%
Carbon^ 1.80 235 GPa 3.6 GPa 1.6%

SOURCES:
1. Watson, 1989 2. Pigliacampi, 1989 3. Tokyo Rope, 1990
 133

Figure 2. Typical Fiber Stress Strain Curves

2.2 FRP Reinforcement

When placed in a pultruded rod, the properties of the fibers become commingled with
the properties of the resins. The properties of the composite may be determined by the rule
of mixture; i.e., each contributes in proportion to its volume in the composite. If the
percentage volume of fibers is denoted vf, then the first approximation for composite
modulus of elasticity can be computed as follows.

E c r = E f v f +Er(l-v f ) (1)

Where E c r is the modulus of elasticity for the composite rod, Ef is the fiber modulus of
elasticity, and E r is the resin modulus of elasticity.
The tensile capacity of the rod is based almost exclusively on the fiber content since
the resin strength is an order of magnitude lower than the fibers. Therefore, the rod capacity
can be estimated as

f
u = v f f uf (2)

where fuf is the available tensile capacity of the individual fiber and fu is the ultimate stress
capacity of the composite reinforcement.
Considering all of the above constraints, the ultimate tensile capacities of
representative parallel fiber reinforcement rods are listed in Table 2. Other sources of non-
metallic reinforcement are available. The list of FRP reinforcement is presented to be
134

indicative of overall materials properties. Even with the same materials and resins, there are
variations in strength with the reinforcement size.
Three other factors influence the strength of reinforcement rods; shear transfer, fiber
alignment within the resin matrix, and reinforcement surface finishing. The diameter of an
FRP rod affects it strength as reinforcement since the strength of the reinforcement can
decrease as the diameter of the rod increases. This is discussed later in the chapter Fiber
Reinforced Plastic Rebars for Concrete Reinforcement. The principle reason for the
strength reduction with bar size is the shear lag as the surface bond stresses are transferred
to the core of the rod. Excessive shear forces can initiate a progressive failure of the
individual fibers. For this reason a #8 reinforcement bar will have a lower ultimate strength
than a #3 bar. This behavior has generally led to the development of small diameter
prestressing reinforcement. Prestressing tendons may be solid rods up to 3/8 inch (9 mm) in
diameter or they may be assembled for multiple 1/8 inch (3 mm) diameter rods.

Table 2
ULTIMATE TENSILE CAPACITY OF REPRESENTATIVE REINFORCEMENT
RODS1

Material Diameter (in) Strength, fu (ksi) Resin Source

Kevlar .125 220 Vinylester E.I. DuPont2

1.57 279 none Parafil3
Fiberglass .080 174 epoxy Glastic Corporation^
E-glass
.120 198 epoxy Glastic Corporation
.140 199 epoxy Glastic Corporation

Carbon .120 260 epoxy Leadline^ smooth rod

.310 270 epoxy Leadline
.670 250 epoxy Leadline
Notes on FRP Rods and sources:
1. Parafil tendons are made in Great Britain and Leadline reinforcement comes from
Japan. US distributors are listed below.
2. E. I. DuPont, Company, Wilmington, DE
3. Parafil tendons are distributed by VSL Corporation, Vienna, VA
4. Glastic Corporation, Cleveland, OH
5. Leadline is distributed by Mitsubishi/Kasei Corporation, Melno Park, CA

While the pultrusion process is intended to provide a highly uniform product, some
variation in the fiber alignment and initial stress occurs. Research by Whiteside found that
these natural misalignments can lead to a variation in the strength of the rods [Whiteside,
1983]. Research in progress at the University of Wyoming indicates that under static loading
 135

AFRP and CFRP rods will display an increase in the modulus of elasticity. This stiffening
occurs as the fibers align to a straight configuration within the composite material.
The data in Table 2 is provided for smooth parallel fiber rods. In order to improve
bond, the rods are sometimes modified during or after initial fabrication. For example, some
rods will have an additional wrap of fiber epoxied to the outside of the rod. Other
reinforcement manufactures will pull a tight fiber around the rod before it is cured, thereby
inducing a deformed surface. This deformed surface improves bond but increases the shear
stresses in the fiber. Sand may be applied to the outer surface of the rod during the final
pultrusion process to artificially roughen the surface. Other manufactures braid their rods
for maximum bond capacity. Representative samples of reinforcement are shown in Figure
3.
These considerations lead to an industry need to establish testing methods to provide
strength and modulus information on rods. Manufacturers of glass fiber reinforced rods have
a basic standard for manufacture. As alternative resins and fibers are introduced; guaranteed
ultimate strength, strain capacity, and modulus of elasticity data must be provided.

Figure 3. Representative FRP reinforcement surface finishes including smooth, braided,

deformed and external spiral.
136

3. FRP REINFORCEMENT FOR CONCRETE

Use of FRP reinforcement in concrete members raises issues of bond, transfer length,
durability and sustained load capacity. These issues have been addressed by several
researchers and some guidance is available.

3.1 Bond And Transfer Length

Smooth reinforcement rods have insufficient bond to the concrete to develop their
full tensile capacity. Reports of bond failure of smooth rods in prestressed concrete flexural
tests have been reported [Dolan, 1990]. Deformed reinforcement, as shown in Figure 3,
significantly aids in the development of surface bond capacity. The addition of sand, epoxied
to the external surface of the rod, also increases the bond capacity. Later in this book, Faza
and GangaRoa suggest values for basic development lengths for glass fiber reinforced plastic
(GFRP) rods. Their equations follow the format of the ACI Building Code [ACI-318] and
apply to deformed GFRP rods similar to Figure 3c. Faza and GangaRoa suggest that the
average development length for a GFRP rod is 50% longer than a corresponding deformed
steel bar. Their recommendation for development length is given as

b
, d= Abfyf (3).~x
vrf
To use this ACI format for development length, the authors have to define a yield stress for
the rods. Since FRP has no yield, a pseudo yield stress for the FRP reinforcement is defined
by applying a factor of 0.75 to the ultimate tensile capacity of the rod.
Ehsani conducted a series of bond tests for #3, #6, and #9 GFRP reinforcing bars
[Ehsani, 1993]. The bars were tested in a flexural mode pullout experiment and were not
prestressed. Failure was defined as rupture of the rod, 0.01 inch extension or bond failure.
The #3 and #6 bars failed by the pullout length criteria while the #9 bar failed by splitting the
concrete. Ehsani also examined the effect of reinforcement placed at the top of a concrete
beam. Tests indicated that some strength reduction occurs in "top" bars.
Ehsani's tested bends in GFRP reinforcement used as stirrups. Since the GFRP may
be made with virtually a zero radius, the tests compared small radius rods with ACI bends.
Tests indicated that the long radius rods failed in tension on the outside of the bend, while
the short radius rods broke at the bend. Further, the short radius rods broke at much lower
loads. Preliminary recommendations suggest that ACI radii be used. Longer tails on the
bends did not help the development.
The transfer length for FRP rods used as prestressing tendons also vary with the
surface treatment. The transfer length of smooth rods having a sanded coating was on the
order of 60 bar diameters for small diameter rod bundles [Dolan, 1989]. Research in
progress at the University of Wyoming and the University of Michigan, suggest that the
transfer length for "seven wire" carbon composite cables provided by Tokyo Rope may be as
low as 25 to 30 rod diameters that corresponds to research reported by Nanni. The high
splitting tensile stresses generated by these short transfer lengths have led to splitting of test
specimens.
 137

Nanni summarizes bond transfer and development length for braided AFRP rods
[Nanni, July 1992]. These tests used flexural specimens and induced bond, splitting and
tension failures into the reinforcement. The research concluded that, for braided rods, the
transfer length for prestress tendons ranges from 20 to 57 centimeters and that development
length ranged from 85 to 120 cm. These lengths correspond to transfer lengths of 18 to 37
rod diameters and development lengths of 70 to 120 rod diameters. Corresponding steel
tests had transfer and development lengths of 120 and 183 diameters respectively. Table 3
summarizes the data from Nanni's paper and compares the development length results with
the Faza - GangaRoa predictions.

Table 3
SUMMARY OF TRANSFER AND DEVELOPMENT LENGTH FOR BRAIDED AFRP
REINFORCEMENT

Tendon Diameter dr Id it lt/dr Id Id !db (in>

(MPa)
K64 8mm/.31M 1619 30cm/11.8M 37.50 97cm/38" 121 12.22
K128 12mm/.47" 1511 38cm/15" 31.67 98cm/38" 82 24.45
K128 12mm/.47H 1511 40cm/15.7" 33.33 100cm/39" 83 24.45
K256 16mm/.63" 1444 57cm/22.4H 35.63 114cm/45M 71 46.74
K128s 12mm/.47" 1511 22cm/8.6" 18.33 82cm/32" 68 24.45
Steel 12.7mm/.5H 1731 120cm/47" 94.49 183cm/72" 144 28.95

Notes: 1. Concrete Strength is 30 MPa (4320) psi

2. l^b values computed using Faza and GangaRoa recommendations
3. lt and 1^ numbers required averaging data with different prestress levels
4. Source: Nanni, July 1993

The difference between the 1^^ predictions and the actual development lengths may
be attributed to the method of testing, the fact that the AFRP rods were prestressed, the
surface of the AFRP rods differs from the GFRP rods, or those aramids have different
moduli of elasticity than glass. The comparison of different FRP materials further indicates
that reconciliation of test methods and data is needed before final design recommendations
are complete.

3.2 Durability
The durability of FRP reinforcement in concrete is more difficult to assess.
Durability includes the chemical reactivity and fatigue performance of the FRP
reinforcement. The primary attractiveness of FRP is its chemical stability in environments
where steel may be readily corroded. This includes high chloride solutions that occur in
structures near sea water or where deicing salts are used.
138

3.2.1 Chemical Reactivity

Fiberglass has a known reactivity to alkali environments and aramids have strength
loss associated with moisture absorption. The presence of a resin around the fibers
influences the rate at which any deterioration can occur. Most FRP reinforcement is
designed to minimize the resin surrounding the fibers, so the protective benefit of the resin is
limited. Furthermore, the resin may crack under load, leaving moisture migration paths to
the fibers. Resins have deterioration characteristics in moist environments and they may act
as a filter or a conduit for chemical agents reaching the fibers. FRP reinforcement used in
prestressing applications is particularly susceptible to chemical reactivity because of its initial
high energy state due to the prestressing.
Exposure studies for GFRP prestressed piles at the University of South Florida
reported significant deterioration of the GFRP prestressing tendons [Sen, 1992]. In these
tests, the wet dry cycles of tidal action transported an alkali solution from the concrete into
the epoxy resin. The resulting alkali concentration initiated a stress rupture failure in the
fiberglass and led to nearly complete deterioration of the tendons after 18 months. In
contrast the Ulanbergstrasse Bridge in Germany used fiberglass tendons in a polyester resin
with a polyamide coating [Pries, no date]. The tendons were post-tensioned and then
grouted with an epoxy grout. No wet dry cyclic testing was reported. Unpublished research
at the University of West Virginia suggests that non-prestressed GFRP reinforcement can be
expected to lose fifty percent of its strength over the life of the structure. No deterioration
tests have been reported for GFRP reinforcement with multiple levels of moisture protection
systems. The Florida exposure tests suggest that wet-dry long-term durability tests are
needed for all FRP prestressing tendons.
Nanni has reported on durability studies conducted in Japan [Nanni, 1992]. Braided
AFRP rods with an 8 millimeter (.31 inch) diameter and a vf of 65 percent were placed in an
alkali and chloride solutions. The rods were tested at elevated temperature and reserve
strength capacity recorded at the end of the exposure period. The alkali solution had a pH of
13. At the end of 2160 hours at 80 C, the rods exhibited a tensile capacity of 97 percent of
the static rod strength. The research concluded that after 500,000 hours, the residual rod
strength would be 94.8 % for the highest test temperatures, 80 C, and 98.2 % for 20 C
exposure.
Doxsee has examined the strength of aramid- epoxy composites subjected to
moisture absorption [Doxsee, 1991]. This research concludes that there is approximately a
2% moisture absorption after 40 hours when samples were submitted to 95% relative
humidity at 76C. They tested a short beam specimen (21 mm long) in single point bending
and found a 60% loss in shear capacity. Examination of the tests suggest that the shear loss
occurred because of deterioration of the epoxy resin or bond between the epoxy and the
resin. This deterioration led to a loss if horizontal shear capacity. The tests are not directly
applicable to uniaxial FRP rods because a similar shear condition does not exist. The test
does raise concerns regarding epoxy socketed anchorages where high shear stresses are
required to assure development and transfer of the tensile stresses to the anchor.
 139

3.2.3 Fatigue Performance

The fatigue properties of FRP rods are generally quite good. Tests with CFRP and
AFRP pultruded rods indicate fatigue performance that is superior to prestressing steel
materials when tested at prestressing stress levels [Walton, 86]. GFRP tests on pultruded
rods indicate that the glass performance is not as good as steel when tested at elevated stress
levels. Since the initial stress in a reinforcing bar is low, and the stresses at service level are
also low, GFRP should perform well at working stress fatigue loading. Prestressing
applications that may be subjected to cyclic loading with a high base stress level will require
a comprehensive review of fatigue performance.
Fatigue tests on fiberglass plates indicated a deterioration of stiffness, especially at
low cycle testing [Ellyin, 1992]. This deterioration may be due to creep in the resin matrix
more than deterioration of the fiber structure. If creep in the matrix is the cause of the loss
of stiffness, then the long term effects of plate bonding systems must be investigated.
Anchorage tests at the University of Wyoming indicated that Carbon, Kevlar, and E-
Glass rods could easily sustain 100,000 cycles of load between 60 - 70 percent of their
ultimate tensile capacity. These fatigue tests further indicated that the fiberglass rods lost
stiffness as the number of cycles increased, but that the aramid and the carbon rods gained
stiffness at the number of cycles increased. This behavior is consistent with other test data.
Fibers in the GFRP rods break in fatigue. This leads to a deterioration of stiffness. Fibers in
the AFRP and CFRP rods did not break, but may have straightened out during testing. This
would lead to the higher stiffness.

3.3 Sustained Load Capacity

There is little active research on the sustained load capacity of FRP prestressing
tendons. The durability studies reported by Nanni included prestressed reinforcement
[Nanni, 1992]. Rods with an initial prestress of 0.7 to 0.8 fu failed between 300 and 2000
hours. Since these rods were in an alkali bath, it is not possible to separate the sustained
load capacity from any exposure effects. European tests suggest limiting the prestressing to
60% of fu to limit sustained load effects. The cyclic load tests of prestressing strands at the
University of Wyoming were conducted at 5 hertz, and consequently the test duration was
not long enough to record sustained load behavior.

3.4 Relaxation
Relaxation is a concern for FRP reinforcement used as prestressing tendons.
Relaxation tests for 5000 hours extrapolated that the 100 year relaxation for aramid FRP is
in the range of 10 to 15 percent [Dolan, 1989]. These tests are in agreement with tests done
in Europe for aramid fiber tendons.
The prestressed pile tests in Florida examined the total losses in the tendons for
square rectangular piles. For members subjected to primarily axial load, the losses were
comparable to steel strands. The losses were not separated into relaxation, creep and
shrinkage components. The modulus of elasticity of the GFRP tendons is much lower than
steel. Consequently, larger relaxation losses may be masked by lower losses due to axial
shortening of the member.
140

4. FRP REINFORCED CONCRETE

Concrete may be reinforced with FRP rods similar to conventional steel

reinforcement, it may be reinforced with two and three dimensional grids, it may be
strengthened with bonded FRP plates, or it may be confined with external FRP wrapping.
GFRP reinforcement has been commercially used in several small applications in the United
States. For example, a concrete pad for a Nuclear Magnetic Resonance (NMR) imaging
system at Tompkins Community Hospital in Ithaca, NY is reinforced with GFRP rebar. An
upper story room in a hospital in San Antonio Texas is similarly reinforced. These are
typical of industrial and commercial installations where corrosive resistance or non-magnetic
properties are needed. FRP column wrapping is being implemented in California as a
method to provide seismic toughness for columns.

4.1 Conventional Reinforced Concrete

The performance of concrete beams reinforced with GFRP bars is discussed in length
in the chapter Fiber Reinforced Plastic Rebars for Concrete Reinforcement. The general
conclusions that can be derived from that chapter are:

The strength of a GFRP reinforced beam may be predicted using strain compatibility
relationships.

The crack widths in GFRP reinforced beams are greater than steel reinforced beams
for the same reinforcing ratios.

The deflections in GFRP beams are greater than steel reinforced beams for the same
reinforcing ratios.

In general, the results for GFRP may be extended to AFRP and CFRP. For the same
reinforcing ratios, designers may expect smaller crack widths and less deflection with CFRP
and AFRP than with GFRP since the effective modulus of the reinforcement is higher.
A typical moment defection curve for a FRP reinforced concrete beam is shown in
Figure 4. The behavior pattern can be idealized as two straight lines. The initial line
represents the moment deflection behavior of the uncracked section. For reinforced sections
this will be comparatively short and will increases as prestressing is applied. The second line
represents the cracked section stiffness. Since the FRP reinforcement remains elastic, this
line does not reduce in slope to the horizontal as does a moment deflection curve for steel
reinforcement.
 141

1.5
Deflection (in)

Figure 4. Typical Moment Defection Curve for FRP Reinforced Beam

4.2 Grid System Reinforcement

Goodspeed, at the University of New Hampshire, and Banks, at Catholic University,
have both examined reinforced concrete using a two dimensional grid as reinforcement
[Goodspeed, 1991, Bank, 1991]. Both report excellent results and the data would suggest
that the cracking and deflection are slightly less than that obtained using conventional type
reinforcement. The grid type reinforcement appears particularly well adapted for slabs
where two way action may be present. The large area of reinforcement in the grid creates an
over-reinforced section. The resulting design is very stiff and the researchers report that the
performance in bridge decks should be very good. The tie between the longitudinal and
transverse direction stiffens the reinforcement system.
The flexural behavior of the grid reinforced structures are similar to Figure 4. Since
the members have no prestress, the cracking moment will be lower. The slab and beam
elements have no shear reinforcement. Bank's test members were slab type elements and had
the grid reinforcement extending up to one quarter of the member depth. The test results
indicated that approximately one third of the specimens developed the full compressive
capacity of the concrete. The remaining members failed in horizontal shear. Examination of
the structural failure indicates that a shear plane is generated at the top of the grid system.
Decks reinforced with two layers of grid, one on the top face and one on the bottom face,
showed extensive shear cracking.

4.3 Bonded Plate Systems

The discussion of reinforced concrete behavior generally addresses new construction.
In may cases, strengthening existing structures require the addition of external
reinforcement. FRP plates, bonded to the tension side of flexural members, can provide a
142

corrosion resistant strengthening method. Numerous investigators have examined the

possibility of bonding FRP plates to existing structures [Klaiber, 1987, Saadatmanesh,
1990,1991, Ehsani, 1990, Ritchie, 1990, Meier, 1991, Triantafillou, 1991, Karem 1992].
These studies vary in complexity and detail. For example, studies at Lehigh compared the
performance of AFRP, GFRP, and CFRP plate systems to strengthen existing reinforced
concrete beams. Saadatmanesh, Klaiber and other investigators examined applications of
bonded plates for strengthening bridges. Triantafillou's bonded plate work specifically
examined the interface stresses along the epoxy bond line and reported that the interface
shear stresses rise significantly near the end of the plate. Dolan examined the effect of
external shear reinforcement to complement the increased flexural strength and found that
Kevlar fabric bonded to the exterior of a T-beam significantly increased the shear capacity of
the member [Dolan, 1993]. While all of the studies reported success of the experimental
prototypes none addresses the possible shear deterioration along the bond line due to
sustained environmental exposure. In these cases, the relaxation phenomenon reported by
Doxsee and by Ellyin may become significant [Doxsee, 1991, Ellyin, 1992].
Failure of bonded plate elements may occur in one of three modes. The tensile
capacity of the plate may be exceeded, the plate may fracture the cover concrete off the
member, and the plate may delaminate. Delimitation of the plates was reported by Ritchie
[Ritchie, 1990]. Ritchie examined end details for terminating the plates on the beam. End
fasteners were not especially effective, but continuation of the plate to the support of the
beam did significantly help. For heavier plates, run the full length of the beam, the concrete
failed by fracture. The fracture often occurs at the plane of the original steel reinforcement
and near the center of the beam. This experimental research and Triantafillou's analytic work
are in general agreement [Triantafillou, 1991].
The failure mode may limit the total plate strengthening available. Concrete fracture
or bond failure between the plate and the concrete may limit the size of the plate that can be
used. Conventional reinforced concrete uses stirrups to limit the horizontal shear in the
concrete. To be effective, the stirrups must be anchored in the compression zone of the
concrete. Most plate research has not included continuous shear reinforcement or anchorage
into the compression zone. Dolan's external shear wrap suggests that supplemental shear
reinforcement may be effective.

4.4 Concrete Confinement

The compressive strength and ductility of concrete may be increased by providing a
confining pressure. This pressure may be active or passive. One method to provide
confinement is to wrap a concrete cylinder or column with a composite material. Harmon
reports that wrapping a concrete cylinder with carbon fibers, then subjecting the cylinder to a
compressive loading allows an increase of the ultimate axial strain capacity of the concrete
from 0.003 to 0.04, Figure 5 [Harmon, 1992].
 143

STRESS (MPa)
276 I 1
241 h -^
^^^p-0.054
207 h .^^^<^.

^ ^ ^ ^ ^ p -0.027
138h .^^^^^>^><^.

L ^ x ^ ^ ^ ^ 5 ^ ^ P-0.014
p
/ ^ * ^ ^ ^ -0.007

w/ X^P-O.OOO

0 0.01 0.02 0.03 0.04

AXIAL STRAIN (mm/mm)

Figure 5. Stress vs. Axial Strain [Source: Harmon, 1992]

Priestly and Seible recognized that the development length of reinforcement may be
shortened if the concrete is confined and have adapted a FRP wrapping system to retrofit
columns [Priestly, 1992]. In these applications, GFRP mats are wrapped around columns
having insufficient splice lengths between the foundation and the column. A cement grout is
pressure injected under the mat to prestress the mat to the column. The resulting pressure
reduces the required development length of the splice bars and effectively toughens the
column's response to seismic loadings. In addition to the test results on three large scale
columns, the authors report the analytic models for confinement strength gains.

5. FRP PRESTRESSED CONCRETE

In many ways, FRP reinforcement is ideally suited to function as a prestressing

tendon. The corrosion resistance of the material makes them less susceptible to stress
corrosion failure. The relatively high strains to failure allow for adequate warning prior to
failure and the low modulus of elasticity reduces the losses due to creep and shrinkage. In
addition to the FRP properties listed earlier, it is necessary to address tendon anchorage and
ductility of the FRP/concrete structure. This section reviews anchorage, moment deflection
behavior, member flexural strength and prototypical structures.

5.1 Anchorage
The low shear strength of the composites limits anchorage options to hold tendons.
Two philosophies for functional applications exist regarding anchor development. In the
first application, the anchors are needed only to pretension the FRP reinforcement. Thus, the
144

anchors are temporary and need only develop about half of the tendon capacity. In the
second application, the anchors are needed to develop the long term ultimate tensile capacity
of the FRP reinforcement. Each of these philosophies is explored in the context of possible
anchor solutions. Anchors can take one of several forms. The most common configurations
are; split wedges in cones, wedge in sockets, epoxy socketed, epoxy sleeves, and metal
sleeve overlays. Each of these anchors is shown in Figure 6.
The split wedge anchor is used by several investigators as a means of pretensioning
tendons [Iyer, 1988]. A variation of the split wedge uses two plates clamped together over
the tendon [Sippel, 1992]. These anchors will often use a grit on the rod or on the anchor to
improve the frictional forces between the anchor and the rod. Plastic split wedge grips,
using Teflon grips, have been reported to develop about 50 percent of the rod capacity. The
split wedge grips tend to have a shallower taper than conventional strand chucks, with 7.5
being close to an optimum taper [Holte-Dolan, 1993]. The split wedge chucks also tend to
be longer than conventional chucks, a typical length being about 4 inches (100 mm).
Epoxy sleeve anchors are a common method for development of tendon ultimate
capacities. The epoxy sleeve is used for individual rods and multiple rod anchors. The
individual rods are anchored in small diameter metal sleeves that may be gripped with
conventional strand chucks [Porter, 1991]. Porter reported some slip in a nine inch long
(228 mm) anchor for a 3/8 inch (9.5 mm) diameter GFRP rod. Extending the anchor length
to 11 inches (280 mm) solved the slippage problem. Porter used a Sikadur 33 epoxy as an
adhesive. Tests on multiple rod GFRP anchors were conducted at the University of
Washington [Mattock, 1989]. These tests used a 76 mm (3 inch) diameter threaded metal
sleeve and 8-7.5 mm (.3 inch) diameter Polystal GFRP rods, Figure 7. The rods were run
straight through the anchor and secured with a sand filled epoxy matrix. Mattock reported
development of 93 percent of the tensile capacity and occasional pull out of individual rods.
Some rods pulled completely out of the anchor while other pulled out 3 inches (76 mm). No
warning was evident prior to the bond failures. Tokyo Rope uses an epoxy sleeve anchor for
its CFCC carbon tendons. Tests at the University of Wyoming developed the full tensile
strength of the tendons in all tests and only a 3 percent variation between individual tests
[Holte, 1993]. In addition to the internal threads of the anchor sleeve, the "seven wire"
tendon is splayed open in the center of the anchor to improve the bond area and to provide a
mechanical interlock.
 145

Conical Socket

Wedges

a. Split Wedge

Multiple rods Conical Socket

Cone

b. Plug and Cone

Rod / Sleeve
\ r y ^ f " " ^ ^ " ^ ^ " ^ " ^ - - ^ Resin

c. Resin Sleeve

Rod
Resin
4= r"t\
-* I,
" I, , K, X
! "'
X X , ,71
" V ' "''"' """ '

d. Resin Socketed

Rod y Sleeve
^ ^ ^ "i

e. Soft Metal Overlay

Figure 6. Typical Anchorage Configurations

146

Figure 7. Multiple Rod Anchorage Test [Source: Mattock, 1989]

Epoxy socketed anchors use a resin potting compound to secure the tendons similar
to a steel cable spelter socket. Tests at Cornell University indicated that the reliability of the
anchors was questionable since numerous bond failures occurred [Dolan, 1989, 1990], This
experimental and analytical work found that unbonding the epoxy from the metal sleeve
improved the reliability of the anchor performance. Varying the internal geometry of the
socket further improved the anchor reliability. Using a parabolic surface and a bond release
agent allowed comparison with conic anchors and allowed a variation of epoxy resins Figure
8. The parabolic anchor was 100 percent successful in developing the full tensile strength of
AFRP CFRP and GFRP small diameter tendons with a Sikadur 35 resin [Holte, 1993]. The
success of the parabolic anchor is attributed its ability to control the shear concentrations at
the entrance of the anchor. Figure 9 compares the shear stresses for bonded and unbonded
linear anchors and an unbonded parabolicly tapered anchor. An eight fold reduction in shear
stress concentration is possible with the parabolic anchor design compared to a conventional
anchor design.
 147

i
7.90

1.10 h
1 1
Bonded 7.5 Linear Taper
Unbonded 7.5 Linear Taper
0.90
0)
Unbonded Segmental Taper (l-8)
J-H
0.70 Resin Modulus: 200 ksi
J-i Applied Load: 10.0 ksi
0.50 P
00
0.30 F

0.10 E
-0.10 r ' I i I I I 1 I I I I I I I

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Distance Along Rod From Tail End, in.

Figure 9. Shear Stress Distribution in Resin Socketed Anchors [Source: Holte-Dolan 1993]

5.2 Flexural Performance

The flexural response of FRP prestressed beams have been reported for AFRP,
CFRP and GFRP tendons [Dolan, 1991, Naaman, 1993, Iyer, 1988]. All researchers report
satisfactory short term performance of the prestressed members. Naaman expressed concern
about the ductility of carbon fiber tendons and conducted his tests using the carbon tendons
as partial prestress. By adding reinforcing steel to the member for ductility, he obtained a
three stage moment deflection curve similar to Figure 4. The three step curve occurs
because the yielding of the steel provides a separate slope in the moment deflection curve.
Comparative tests between bonded and unbonded carbon fiber tendons in prestressed
concrete slabs are being conducted at the University of Wyoming. In these tests, the tendons
are pretensioned to 50 percent of their tensile capacity and the beams are subjected to one
million cycles of loading between 40 and 60 percent of their ultimate flexural capacity.
Preliminary results suggest that the bonded tendons perform excellently. The full flexural
capacity was obtained following the million cycles. The members lost stiffness with
increased cyclic loading because new cracks formed. However, the initial crack widths were
small and crack growth terminated after a few hundred cycles. Beams with unbonded
tendons had very few cracks. The crack width grew with the cyclic loading and one test
nearly resulted in fatigue failure of the concrete. There was no fatigue damage to the
tendons. One of the unbonded tendons was accidentally damaged during construction. A
small notch was cut into the side of the tendon. This tendon was still intact after one million
cycles. When the beam was monotonically tested to failure, it failed in concrete
compression. The beam was then deflected to the full extension capacity of the jack, 8
inches (200 mm), and no damage to the tendon could be found.
148

6. DESIGN GUIDELINES

There are three approaches to the design of FRP reinforced flexural members. They
are working stress methods, strength methods and limit states methods. Each approach has
been suggested by some of the researchers, but no clearly preferred design approach has
emerged.

6.1 Working Stress Method

The working stress method concrete structures by examining the stresses in the
structure by looking at elastic behavior and service level stresses [Roll, 91, Stetson, 92,
Triantafillou, 91, Karam, 92]. The rationale for using working stress design is that the FRP
reinforcement has a low elastic modulus and limiting the tensile stresses will control both the
cracking and the deflections. For flexural design, this requires establishing a neutral axis
based on the transformed area, Figure 10, and limiting the concrete and reinforcement
stresses to allowable levels. The cracked transformed area is used for deflection control.

kd

nAr

SECTION STRESSES

Figure 10, Working Stress Section and Stress Distribution

The section properties are based on the depth to the neutral axis. For linearly elastic
materials, the value for k, the ratio of the compression block to the overall structural depth,
is

k = ^ ( p n ^ + p n - pn (4)
Since the moment of inertia is based on the cracked section, and the lower modulus
of elasticity of the FRP is reflected in the value for n, this approach is reported to provide
acceptable results. Because service level stresses are associated with cracked sections, a
working stress approach does not readily address the behavior of prestressed concrete
members.
 149

At the present time there is not a coherent design rationale for shear design at
allowable stress levels. The larger cracks resulting from the lower modulus of elasticity and
the lower modulus of the reinforcement imply that the "concrete contribution" to shear
resistance may be lower and that premature shear failure may occur.

6.2 Strength Design

Strength design has been advocated by several researchers [Challal, 92, Dolan, 91,
and Faza and GangaRoa]. Strength design assumes that failure occurs when either the
tensile capacity, fu, of the tendon is reached, or when a compressive strain of 0.003 is
obtained in the concrete, Figure 11. By finding equilibrium on the cross section, then
computing the nominal bending moment capacity, researchers have demonstrated that use of
the rectangular stress block can predict the strength of the flexural member. Predications are
accurate for both reinforced and prestressed members, however, three caveats must be
observed. First, the concrete must develop it full compressive capacity. Second, if the
tendon strength is reached before the concrete failure, catastrophic failure can result. Third,
the distribution of strain in the reinforcement must be considered in the calculations,
b 0.85 fcl
H concrete strain

->fr

Ar I I reinforcement strain

SECTION STRAIN STRESS

Figure 11, Strength Relationship for FRP reinforced Concrete

Figure 11 assumes that the concrete reaches it ultimate compressive strain. Since
the FRP reinforcement does not yield, the possibility of an over-reinforced section exists.
Dolan has postulated a "brittle reinforcement" ratio, Pbr [Dolan, 1991] equation 2. For p
less than pb r failure of the tendon is assured and the concrete may not reach its ultimate
strain. For p greater than p^ r , tendon integrity is assured and a concrete compressive failure
will occur. The tensile strain that occurs at failure is much larger than the tensile strain at
yield of ordinary steel reinforcement. Therefore, some "ductility" of the structure is assured.
This is discussed further in the "Ductility" section of the limit states design approach.
150

pbr = 0.85 l (5)

fu cu + u - pi

Faza and GangaRoa have modified the basic ACI flexural strength prediction
equation so it is applicable for FRP reinforcement. For their equation to work, the authors
have defined a pseudo yield point of fyf equal to 0.80 fu for the FRP reinforcement.

M n = Arfyf d( l-0.585p fyf/f c ) (6)

The equation is limited by the fact that in some instances the strain required to
develop the FRP reinforcement is not available in the flexural member. High strength
concrete, f c > 6000 psi (41 MPa), improves the opportunity to develop the full tensile
strength of the reinforcement. Equation 6 further assumes that all of the reinforcement is in
a single layer and includes a capacity reduction factor in the fyf term. To assure that the
nominal moment capacity can be attained, a strain compatibility check of the tensile
reinforcement is required. Besides the establishment of a pseudo yield stress, this method is
similar to those outlined in any text describing reinforced or prestressed concrete behavior
[Nilson, 1992].
If p is greater than pbr, the concrete does not reach its ultimate strain, the underlying
assumptions of the rectangular stress block are no longer valid, and equation 6 will not
predict the strength of the section. For this case, the nominal moment capacity must be
computed on the actual stress distribution in the concrete. Significantly over reinforced
sections may lead the engineer back to a working stress approach to the design.
Because either failure mode will be brittle, some allowance must be made beyond the
existing ACI guidelines for capacity reduction, , factors. Faza and GangaRoa have
suggested applying a pseudo yield stress for the FRP reinforcement by using a = 0.80 on
the reinforcement and then using a = 0.9 to obtain ultimate strength from the nominal
moment capacity. Dolan has suggested a single lower factor to account for either the
tendon failure or the concrete compression failure. A factor of 0.70 is consistent with
compression failure in structural members such as piles and columns. The factor of 0.70
for columns includes tied reinforcement and stability considerations in addition to material
strength issues. Use of a factor of 0.70 would be a reasonable starting point while a
reliable data base of experience is being acquired.
Shear design is dependent on the strength of the stirrups and the shear crack width.
The larger crack widths in FRP reinforced members reduces the shear strength of both the
concrete contribution and the dowel action. This leads to lower overall shear strength.
Some introductory work with prestressed members suggests that the crack width is smaller
and that ACI predictions of concrete contribution may be adequate [Dolan, 1993].
Substantial over capacity with FRP shear reinforcement allowed the full flexural capacity of
the beam to be developed.
Shear provisions for strength design are not yet developed. One method for shear
design is to detail a member for shear using the provisions of the ACI building Code using
steel reinforcement. Substitute FRP reinforcement for the steel to provide equivalent
 151

stiffness; area times modulus of elasticity, as the steel design. This will lead to more strength
than is required, but should control the crack width to allow the member to function as
designed. Ehsani further recommends that stirrups should be detailed using ACI bend radii
[Ehsani, 1993]
The discussion of crack width in Fiber Reinforced Plastic Rebars for Concrete
Reinforcement provides insight into the width of cracks in reinforced FRP beams. Crack
widths in prestressed beams are significantly smaller. One of the important issues associated
with crack width is the availability of deleterious material to reach the reinforcement. Since
the FRP reinforcement is selected for its inherent corrosion resistance, crack width may be
more a function of appearance and deflection than a corrosion control criterion.

6.3 Limit State Design

The ACI Building Code [ACI-318, 1993] contains both strength and serviceability
criteria for the design of concrete members. A Limit State approach identifies both strength
and service level conditions which should be controlled or limited by the designer. Head
[Head, 1986 and 1991] has proposed limit states approaches for FRP structures and this
work may be reasonably applied to FRP reinforced concrete structures. Design
methodologies must identify the appropriate limit states for design. Some suggested limit
states are flexural strength, shear strength, deflection, ductility, and bond and development.

6.3.1 Flexural and Shear Strength

Both flexural and shear strength are discussed in the Strength Design section and are
not repeated. Eventually, sustained load limits or retained strain capacity limits must be
established to support the strength calculations.

6.3.2 Deflection
Building codes currently set limits on deflections in members. The mechanics of
computing deflection in FRP reinforced member is not well defined. Comparison of the
predicted and actual deflections of members prestressed with aramid and carbon fibers
suggest that the effective moment of inertia equations in the ACI Code may serve as a basis
for predicting deflections [ACI-318, 1993]. The effective moment of inertia is given as

I e = Ig(Mcr/Ma)3 +I cr 0-(M cr /Ma)3) (7)

Deflections of prestressed concrete planks prestressed with carbon tendons were

accurately predicted by equation 7. The cracked moment of inertia was computed using
guidelines in the working stress section and the appropriate value for n with the carbon
tendon. The deflection for aramid prestressing tendons were accurate if the exponents in
equation 7 were increased from 3 to 4 [Dolan, 1989] This suggests that the greater cracking
associated with the lower modulus tendons may be compensated using a "softer" effective
moment of inertia.
152

6.3.3 Ductility
FRP reinforcement does not yield. Therefore, if the reinforcement is placed in more
than one tensile level, failure of the most highly strained reinforcement will precipitate
failure. For this reason, the structural depth must be defined to the centroid of the
reinforcement farthest from the compressive face rather than to the centroid of the tensile
force. This minor modification is consistent with recommendations currently being
considered by the ACI Building Code Committee 318 [Mast, 1992]. Mast also suggests that
a net tensile strain be used as a deliminator for ductile behavior He recommends a net tensile
strain of 0.005 be used to define the transition to full tensile behavior. This value is selected
to include both conventional reinforcement and prestressing steel. If adopted, it may also
serve as a deliminator for elastic behavior in FRP reinforced members. Strains in excess of
0.005 would correspond to the yield strains in conventional members. This allows the
designer to compute a ductility ratio. Using this 0.005 net tensile strain criteria, the ductility
ratio would be the area beneath the moment curvature diagram for strains greater than 0.005
divided by the comparable area for strains less than 0.005. Using these criteria, ductility will
be a function of the reinforcement ratio. FRP reinforced sections may be designed to meet
specified ductility criteria.

6.3.4 Bond and Development

The research results for bond and development still need some refinement. Faza and
GangaRoa's work may be adequate for non-prestressed members. Development lengths for
prestressed members may come from work reported by Nanni, however, additional testing
and data evaluation are needed to define the effects of transverse thermal stresses.

7. DEMONSTRATION TESTS AND PROGRAMS

There are several significant demonstration programs in the United States. These
include a bridge deck in South Dakota, a prestressed pile project in Florida, a bridge deck
evaluation project in New Hampshire, and a series of column wrapping projects in
California.

7.1 Bridge Deck

A transversely post-tensioned bridge deck was installed at the entrance to a
concrete factory near Rapid City South Dakota. The bridge deck was designed,
fabricated and installed in 1991 by the South Dakota School of Mines and Technology.
The bridge consists of steel girders and an 8 inch (200 mm) thick concrete deck. The
bridge is about 20 feet wide and 30 feet long (6.0 x 9.1 m). The entire structure was
prefabricated at the University and shipped to the installation site.
 153

The deck is transversely post-tensioned with glass, carbon and steel tendons.
There are two transverse tendons every 8 inches (200 mm) along the length of the bridge.
Each 10 foot (3.0 m) section contains a different tendon type. The tendons were
prestressed to 55% of their ultimate strength, grouted with an epoxy mortar and then the
anchors were cut off. The bridge is being monitored. Technical reports after one year of
operation suggest that the deflection due to live load has increased approximately 20
percent since the deck was first installed.

7.2 Prestressed Piles

Florida Southern University in conjunction with the Florida Department of
Transportation fabricated a series of prestressed concrete piles using fiberglass and steel
tendons. The piles used steel and fiberglass spiral ties. The piles were driven into the
sandy Florida soil with a 3 kip (13 kN) drop hammer using an 8 foot (2.4 m) drop. The
piles were all driven successfully. After successful completion of driving, an increased
driving force was delivered to the piles. During this overdriving, the piles with the
fiberglass spirals fractured the spirals. The fiberglass ties were designed for the same
strength requirements as steel, yet were unable to take the driving overloads.
Pile sections were immersed in a tidal pool and exposed to alternate wetting and
drying. Substantial deterioration of the tendons resulted. The durability test results are
reported in the durability discussion.

7.3 Bridge Deck Evaluation

Researchers at the University of New Hampshire have constructed a small bridge
deck simulation and evaluation test facility at the entrance to a sanitary landfill near the
University. FRP reinforced deck panels will be installed at the entrance to the facility
and will be subjected to daily truck traffic. The facility was opened in the Spring of 1993
and test results are not yet available.

7.4 Column Wrapping Projects

As a part of their general seismic upgrading program, Caltrans, the California
Department of Transportation, is placing confining jackets around bridge columns in
seismic sensitive areas. The confinement is intended to improve the ductility of the
reinforced concrete and improve the bond between the column and the foundation splice
bars.
The initial Caltrans Design used metal confinement. Recently, two major freeway
bridge columns have been wrapped with a fiberglass mat. After the glass was applied,
epoxy was spread over the mats to provide a lap bond splice of the fiberglass. Expansive
grout was injected beneath the mat to assure contact with the original concrete.

8. RESEARCH NEEDS

To develop a complete Limit States Design Recommendation for FRP reinforced

members the following research is needed.
154

Long Term Sustained Load Capacity - The upper bound for sustained load capacity
is needed for each tendon type. This information is critical to establishing initial
and final prestress levels and to qualify fatigue performance.

Bond and Temperature - The transverse effect of thermal expansion and contraction
are needed to assure that bond does not deteriorate over the life of the structure.
Correlation of bond length and transfer length and the surface preparation of the
FRP tendons is needed.

Bond and moisture - FRP reinforcement may deteriorate if the resin deteriorates with
moisture. Sustained load bond tests must be conducted to assure that the long
term bond can be maintained.

Durability - Evaluation of tendon performance at elevated load and in wet dry

exposure to chloride and alkali solutions is needed. A standard test method must
be defined which can replicate the results obtained by Sen's tidal pool test.

Ductility - Analytic studies must be conducted to provide guidelines for ductility in

FRP reinforced members. This will allow establishment of ductility ratios for safe
design and performance.

Shear - The world wide data of experimental work must be reviewed and firm
guidelines for design of shear reinforcement prepared. The review of existing test
data will provide the background necessary to determine if additional testing is
required.

NOTATION
a depth of the equivalent rectangular stress block
Afc Area of FRP reinforcement for bond calculations
Ar Area of FRP reinforcement
AFRP Aramid Fiber Reinforced Plastic
b width of concrete compression section
c depth from the compression face to the neutral axis
CFRP Carbon Fiber Reinforced Plastic
d distance from the compression face to the centroid of the FRP reinforcement
dr diameter of FRP reinforcement
Ec Modulus of elasticity of concrete
Ecr Modulus of the uniaxial composite reinforcement along its principal longitudinal
axis
Ef Modulus of elasticity of the fiber in the longitudinal direction
Er Modulus of Elasticity of resin
FRP Fiber Reinforced Plastic
fc compressive strength of the concrete
fr stress in the FRP reinforcement
 155

fu Ultimate tensile capacity of the composite reinforcement

fyf Pseudo yield stress of FRP reinforcement = 0.80 fu
fuf Ultimate available tensile strength of an individual fibers
fyf pseudo yield stress in FRP reinforcement = <|)fu
GFRP Glass Fiber Reinforced Plastic
Icr Moment of Inertia of a cracked concrete section
Ig Moment of Inertial of an uncracked concrete section
k ratio of the compression block depth to the structural depth for a cracked
elastic section
Id development length of FRP reinforcement
lbd basic development length of reinforcement
It transfer length of prestressing reinforcement
Ma Service load moment used for deflection calculations
Mcr Service load moment that causes cracking in the section
Mn Nominal flexural moment capacity
n modular ratio = E c r /E c
vf volume fraction of the fiber in the reinforcement cross section
concrete strength reduction factor = 0.85 less 0.05 for every 1000 psi for f c >
4000 psi, but not less than 0.65
8CU Ultimate concrete strain
Spj Prestressing initial strain less any identical strain losses
8U Ultimate reinforcement strain
p reinforcement ratio = A r /bd
Pb r brittle reinforcement ratio = reinforcement ration which simultaneously causes
compression failure and rupture of the FRP tendon

REFERENCES

ACI 318 (1993). BUILDING CODE REQUIREMENTS FOR REINFORCED

CONCRETE (ACI-318), American Concrete Institute, Detroit, MI

Bank, L. C. and Mosallam, A. S. (1991). "Experimental Study o f FRP Grating Reinforced

Concrete Slabs". A D V A N C E D COMPOSITES MATERIALS IN CIVIL ENGINEERING
STRUCTURES, Iyer and Sen, eds., American Society o f Civil Engineers, N Y , N Y ,
Specialty Conf. Jan 31 - Feb 1 p. 111.

Chaallal, O., Benmokrane, B., and Masmoudi, R. (1992). "An Innovative Glass Fiber
Composite Rebar for Concrete Structures". A D V A N C E D COMPOSITE MATERIALS IN
BRIDGES A N D STRUCTURES, Neale and Labossiere, eds., Canadian Society o f Civil
Engineers, Montreal, Que. p. 169.
156

Daniali, S. (1995). "Development Length for Fibre-Reinforced Plastic Bars". ADVANCED

COMPOSITE MATERIALS IN BRIDGES AND STRUCTURES, Neale and Labossiere,
eds., Canadian Society of Civil Engineers, Montreal Que. p. 179.

Dolan, C. W. (1990). "Developments in Non-Metallic Prestressing Tendons". PCI

JOURNAL, Precast/Prestressed Concrete Institute, Chicago, IL, Vol. 35, No. 5, September
p. 80.

Dolan, C. W. (1991). "Kevlar Reinforced Prestressing for Bridge Decks". THIRD BRIDGE
CONFERENCE, Transportation Research Record No. 1290, Vol. 1, Transportation
Research Board, Washington, DC p. 68.

Dolan, C. W. (1989). Prestressed Concrete Using Kevlar Reinforced Tendons. Ph.D.

Thesis, Cornell University

Doxsee, L. E., Janssens, W., Verpoest, I., and DeMeester, P. (1991). "Strength of Aramid-
Epoxy Composites during Moisture Absorption". Journal of REINFORCED PLASTICS
AND COMPOSITES, Vol. 10., December p. 645.

DuPont (1987). DATA MANUAL FOR KEVLAR 49 ARAMIDS, Material Sciences

Corporation, Spring Hill, PA

Ehsani, M. and Saadatmanesh, H. (1990). "Fiber Composite Plates for Strengthening Bridge
Beams". COMPOSITE STRUCTURES, Vol. 15, No. 4 p. 343.

Ehsani, M. R, Saadatmanesh, H. H. and Tao, S. (1993). "Bond of GFRP Rebars to

Ordinary-Strength Concrete". INTERNATIONAL SYMPOSIUM ON FRP
REINFORCEMENT FOR CONCRETE STRUCTURES, American Concrete Institute,
Detroit, MI

Ellyin, F. and Kujawski, D. (1992). "Fatigue Testing and Life Predictions of Fibreglass-
Reinforecd Composites". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES," Neale and Labossiere, eds., Canadian Society of Civil Engineers,
Montreal, Que. p. 111.

Faza, S. S. and GangaRoa, H. V. S. (1991). "Bending Response of Beams Reinforced with

FRP Rebars for Varying Concrete Strengths". ADVANCED COMPOSITES MATERIALS
IN CIVIL ENGINEERING STRUCTURES, Iyer and Sen, eds., American Society of Civil
Engineers, NY, NY, Specialty Conf. Jan 31 - Feb 1 p. 262.

Faza, S. S. and GangaRoa, H. V. S. (1992). "Pre and Post-Cracking Deflection Behavior of

Concrete Beams Reinforced with Fibre-Reinforced Plastic Rebars". ADVANCED
COMPOSITE MATERIALS IN BRIDGES AND STRUCTURES, Neale and Labossiere,
eds., Canadian Society of Civil Engineers, Montreal, Que. p. 151.
 157

Faza, S. S. and GangaRoa, H. V. S. (1991). "Bending and Bond Behavior of Concrete

Beams Reinforced with Plastic Rebars". TRANSPORTATION RESEARCH RECORD, No.
1290, Volume 2, Third Bridge Engineering Conference, Transportation Research Board,
Washington, DC p. 185.

FRP Research (1993). FRP INTERNATIONAL, S. Rizkalla, Ed., Winnipeg, Canada, Vol.
l,No. 1, January, p. 5.

GangaRoa, H. V. S., and Faza, S. S. (1992). "Fiber Reinforced Plastic (FRP) Bars for
Concrete Reinforcement". PROCEEDINGS OF THE 1992 NSF STRUCTURES,
GEOMECHANICS AND BUILDING SYSTEMS GRANTEES CONFERENCE, National
Science Foundation, San Juan Puerto Rico, June 10-12 p. 139.

Gerritse, A. (1992). "Durability Criteria for Non-Metallic Tendons in an Alkaline

Environment". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil Engineers, Montreal,
Que. p. 129.

Gerstle, F. P., and Kunz, S. C. (1983). "Prediction of Long-Term Failure in Kevlar 49

Composites". LONG-TERM BEHAVIOR OF COMPOSITES, ASTM STP-813, T. K.
O'Brien, ed., American Society for Testing and Materials, Philadelphia p. 263.

Goodspeed, C. (1991). "FRP Grids for Concrete". THE CONSTRUCTION SPECIFIER,

Vol. 44, No. 12, December p. 102.

Goodspeed, C, Schmeckpeper, E., Gross, T., Yost, j., and Zhang, M. (1991). "Cyclical
Testing of Concrete Beams Reinforced with Fiber Reinforced Plastic (FRP) Grids".
ADVANCED COMPOSITES MATERIALS IN CIVIL ENGINEERING STRUCTURES,
Iyer and Sen, eds., American Society of Civil Engineers, NY, NY, Specialty Conf. Jan 31 -
Feb p. 278.

Harmon, T. G. and Slattery, K. T. (1992). "Advanced Composite Confinement of

Concrete". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds, Canadian Society of Civil Engineers, Montreal,
Que. p. 299.

Head, P. R. (1992). "Design Methods and Bridge Forms for the Cost Effective use of
Advanced Composite Materials". ADVANCED COMPOSITE MATERIALS IN BRIDGES
AND STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil Engineers,
Montreal, Que 15.

Head, P. R. and Templeman, R. B. (1986). "The Application of Limit State Design

Principles to Fibre Reinforced Plastics". British Plastics Federation, Reinforced Plastics
Group Congress p. 69.
158

Holte, L. E. (1993). "Anchorage of Non-Metallic Prestressing Tendons". M.S. Thesis,

University of Wyoming, Laramie, WY, May.

Holte, L. E., Dolan, C. W. and Schmidt, R. S. (1993). "Epoxy Socketed Anchors for
Nonmetallic Prestressing Tendons". INTERNATIONAL SYMPOSIUM ON FRP
REINFORCEMENT FOR CONCRETE STRUCTURES, American Concrete Institute,
Detroit, MI

Iyer, S. L. and Khubchandani, A. (1992). "Evaluation of Graphite Composite Cables for

Prestressing Concrete". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, 1st International Conf., Neale and Labossiere, eds., Canadian Society of
Civil Engineers, Montreal, Que p. 73.

Iyer, S. L. and Sen, R., eds. (1991). ADVANCED COMPOSITES MATERIALS IN CIVIL
ENGINEERING STRUCTURES. American Society of Civil Engineers, NY, NY, Specialty
Conference, Jan 31 - Feb 1

Iyer, S. L., and Anigol, M. (1991). "Testing and Evaluating Fiberglass, Graphite and Steel
Prestressing Cables for Pretensioned Beams". ADVANCED COMPOSITE MATERIALS
IN CIVIL ENGINEERING STRUCTURES, Iyer and Sen, eds., American Society of Civil
Engineers, NY, NY, Specialty Conf. Jan 31 - Feb 1. p. 44.

Iyer, S. L., and Kumarswamy, C. (1988). "Performance Evaluation of Glass Fiber

Composite Cable For Prestressing Concrete Units". 33rd International SAMPE Symposium,
Anaheim, CA, March 7-10 pp. 12.

Iyer, S., Khubchandani, A., and Feng, J. (1991). "Fiberglass and Graphite Cables for Bridge
Decks". ADVANCED COMPOSITES MATERIALS IN CIVIL ENGINEERING
STRUCTURES, Iyer and Sen, eds., American Society of Civil Engineers, NY, NY,
Specialty Conf. Jan 31 - Feb p. 371.

Karam, G. N. (1992). "Optimal Design for Prestressing with FRP Sheets in Structural
Members". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil Engineers, Montreal,
Que. p. 277.

Klaiber, F. W., Dunker, K. F. Wiptf, T. J. and Sanders, W. W. (1987). "Methods of

Strengthening Existing Highway Bridges". National Cooperative Highway Research Council
REPORT NO. 293, Transportation Research Board, Washington, DC

Larralde, J. and Silva-Rodreguez, R. (1993). "Bond and Slip of FRP Rebars in Concrete".
JOURNAL OF MATERIALS IN CIVIL ENGINEERING, American Society of Civil
Engineers, NY, NY, Vol. 5, No.l, February p. 30.
 159

Mandell, J. F. and Meier, U. (1983). "Effects of Stress Ratio, Frequency, and Loading Time
on the Tensile Fatigue of Glass-Reinforced Epoxy". LONG TERM BEHAVIOR OF
COMPOSITES, ASTM STP-813, T. K. O'Brien, ed., American Society for Testing and
Materials, Philadelphia p. 55.

Mast, R. F. (1992). "Unified Design Provisions for Reinforced and Prestressed Flexural
Members". ACI STRUCTURAL JOURNAL, American Concrete Institute, Detroit, MI,
Vol. 89, No. 2, March/April p. 185.

Mather, B. and Tye, R. V. (1955). "Plastic-Glass Fiber Reinforcement for Reinforced and
Prestressed Concrete; Summary of Information available as of July 1, 1955". Technical
Memorandum No. 6-421, Report 1, Corps of Engineers, Waterways Experiment Station,
Vicksburg, MS.

Mattock, A. H. and Babaei, K. (1989). "Application of Fiber Reinforced Plastic Rods as

Prestressing Tendons in Concrete Structures". Washington State Transportation Center
(TRAC), University of Washington, Seattle, WA pp. 45.

Matzeg, R. and Santrach, D. (1991). "Effects of the Glass Fiber Type on Corrosion of FRP
Laminates". SAMPE QUARTERLY, October p. 12.

McKay, K. S. and Erki, M. A. (1992). "Aramid Tendons in Prestressed Concrete

Applications". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Lavossiere, eds., Canadian Society of Civil Engineers, Montreal,
Que. p. 221.

STRUCTURES, Iyer and Sen, eds., American Society of Civil Engineers, NY, NY,
Specialty Conf. Jan 31- Feb 1 p. 224.

Mitchell, R. A., Wooley, R. M., and Halsey, N. (1974). "High Strength End Fittings for
FRP Rod and Rope". JOURNAL OF THE ENGINEERING MECHANICS DIVISION,
American Society of Civil Engineers, Vol. 100, No. EM4, August p. 687.

Mittelman, A. and Roman, I. (1991). "Tensile Properties of real Unidirectional

Kevlar/Epoxy Composites". COMPOSITES, Vol. 21, No. 1, January p. 63.

Naaman, A. E., Jeong, S. M., Tan, K. H. and Alkhairi, F. M. (1993). "Partially Prestressed
Beams with Carbon Fiber Composite Strands: Preliminary Test Evaluation".
INTERNATIONAL SYMPOSIUM ON FRP REINFORCEMENT FOR CONCRETE
STRUCTURES, American Concrete Institute, Detroit, MI

Nanni, A and Dolan, C.W., eds. (1993). INTERNATIONAL SYMPOSIUM ON FRP

REINFORCEMENT FOR CONCRETE STRUCTURES. American Concrete Institute,
Detroit, MI
160

Nanni, A. and Tanigaki, M. (1992). "Pretensioned PC Members with Bonded FRP Tendons:
Development and Flexural Bond Length (Static)". ACI STRUCTURAL JOURNAL,
American Concrete Institute, Detroit, MI, Vol. 89, No. 4, July-August p. 433.

Nanni, A., Henneke, M. J., and Norris, M. S. (1992). "Research on FRP Reinforcement at
Penn State". PROCEEDINGS OF THE 1992 NSF STRUCTURES, GEOMECHANICS
AND BUILDING SYSTEMS GRANTEES' CONFERENCE, National Science Foundation,
San Juan, Puerto Rico, June 10 - 1 2 p. 152.

Nanni, A., Matsubara, S. and Hasuo, K. (1992). "Durability of Braided Epoxy-Impregnated

Aramid FRP Rods". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil
Engineers, Montreal, QUE p. 101.

Nanni, A., Tangigaki, M., and Hasuo, K. (1992). "Bond Anchorage of Pretensioned Aramid
FRP Tendon". JOURNAL OF STRUCTURAL ENGINEERING, American Society of Civil
Engineers, NY, NY, Vol. 118, No. 10, October p. 2837.

Nanni, A., Utsunomiya, T., Yonekura, H., and Tangaki, M. (1992). "Transmission of
Prestressing Force to Concrete by Bonded FRP Tendons". ACI STRUCTURAL JOURNAL,
American Concrete Institute, Detroit, MI, Vol. 89, No. 3, May-June p. 335.

NBS (1976). "Non-Metallic Antenna Support Materials Pultruded Rods for Antenna Guys,
Catenaries and Communications Structures". Technical Report AFML-TR-76-42, National
Bureau of Standards, Washington, DC pp. 126.

Nilson, AH. (1992). DESIGN OF CONCRETE STRUCTURES, 11th Ed. McGraw Hill,
Highstown, NJ

Oplinger, D. W., Parker, B. S., Gandhi, K. R., Lamothe, R., and Foley, G. (1985). "On the
Streamline Specimen for Tension Testing of Composite Materials". RECENT ADVANCES
IN COMPOSITES IN THE UNITED STATES AND JAPAN, ASTM STP-864, Vinson and
Taya, eds., American Society for Testing and Materials, Philadelphia p. 532.

Pepper, L., and Mather, B. (1959). "Plastic-Glass Fiber Reinforcement for Reinforced
Prestressed Concrete; Summary of Information from 1 July 1955 to 1 January 1959".
Technical Memorandum No. 6-421 Report 2, Corps of Engineers, Waterways Experiment
Station, Vicksburg, MS.

Phoenix, S. L. (1979). "Statistical Theory for the Strength of Twisted Fiber Bundles with
Applications to Yarns and Cables". TEXTILE RESEARCH JOURNAL, Vol. 49, No. 7, July
p. 407.

Pigliacampi, J. J. (1989). "Organic Fibers". ENGINEERING MATERIALS HANDBOOK,

VOLUME 1, COMPOSITES, ASM International, Ed., Metals Park, OH p. 54 & 114.
 161

Porter, Max and Barnes, Bruce A. (1991). "Tensile Testing of Glass Fiber Composite Rod".
ADVANCED COMPOSITES MATERIALS IN CIVIL ENGINEERING STRUCTURES,
Iyer and Sen, eds., American Society of Civil Engineers, NY, NY, Specialty Conf. Jan 31 -
Feb 1 p. 123.

Pries, L. and Bell, T. A. "Fiberglass Tendons for Post-Tensioning Concrete Bridges".

Mobay Corporation, Pittsburgh, PA pp. 20.

Priestly, M. J. N., Seible, F. and Fyfe, E. (1992). "Column Seismic Retrofitting Using
Fibreglass/Epoxy Jackets". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil Engineers, Montreal,
Que. p. 287.

Riefschneider, K. L., Schulte, K., and Duke, J. C. (1983). "Long-Term Fatigue Behavior of
Composite Materials". LONG-TERM BEHAVIOR OF COMPOSITE MATERIALS,
ASTM STP-813, T. K. O'Brien, ed., American Society for Testing and Materials,
Philadelphia p. 136.

Ritchie, P. A., Thomas, D. A., Lu, L-W, and Connelly, G.M. (1990). "External
Reinforcement of Concrete Beams Using Fiber-Reinforced Plastics". ATLSS Report No. 90-
06, Advanced Technology For Large Structural Systems, Lehigh University, Lehigh, PA pp.
32.

Roll, Richard (1991). "Use of GFRP Rebar in Concrete Structures". ADVANCED

COMPOSITES MATERIALS IN CIVIL ENGINEERING STRUCTURES, Iyer and Sen,
eds., American Society of Civil Engineers, NY, NY, Specialty Conf. Jan 31 -Feb 1 p. 93.

Saadatmanesh, H. and Ehsani, M. (1991). "Fiber Composite Bar for Reinforced Concrete
Construction". JOURNAL OF COMPOSITE MATERIALS, Vol. 25, No. 2, Feb p. 188.

Saadatmanesh, H. H., Ehsani, M. R. and Li, M-W (1993). "Behavior of Externally Confined
Concrete Columns". INTERNATIONAL SYMPOSIUM ON FRP REINFORCEMENT
FOR CONCRETE STRUCTURES, American Concrete Institute, Detroit, MI

Saadatmanesh, H., and Ehsani, M. (1990). "Fiber Composite Plates can Strengthen Beams".
CONCRETE INTERNATIONAL, American Concrete Institute, Detroit, MI, Vol. 12, No.
3, March p. 65.

Sadatmanesh, H. and Ehsani, M. R. (1991). "RC Beams Strengthened with GFRP Plates: I
Experimental Study". JOURNAL OF STRUCTURAL ENGINEERING, American Society
of Civil Engineers, NY, NY, Vol. 117, No. 11, November p. 3417.
162

Sen, R., Issa, M. and Mariscal, D. (1992). "Feasibility of Fiberglass Pretensioned Piles in a
Marine Environment". RESEARCH REPORT CEM/ST/92/1, University of South Florida,
Tampa, Fl, August

Sen, R., Iyer, S., Issa, M., and Shahway, M. (1991). "Fiberglass Pretensioned Piles for
Marine Environment". ADVANCED COMPOSITES MATERIALS IN CIVIL
ENGINEERING STRUCTURES, Iyer and Sen, eds., American Society of Civil Engineers,
NY, NY, Specialty Conf. Jan 31- Feb 1 p. 348.

Sippel, T. M. (1992). "Design, Testing and Modeling of an Anchorage System for Resin
Bonded Fibreglass Rods used as Prestressing Tendons". ADVANCED COMPOSITE
MATERIALS IN BRIDGES AND STRUCTURES, Neale and Labossiere, eds., Canadian
Society of Civil Engineers, Montreal, Que. p. 363.

Stetson, M. B. and Goodspeed, C. H. (1992). "Fundamental Considerations in the Design of

Flexural Concrete with Non-Prestressed FRP Reinforcement". ADVANCED COMPOSITE
MATERIALS IN BRIDGES AND STRUCTURES Neale and Labossiere, eds., Canadian
Society of Civil Engineers, Montreal, Que. p. 141.

Tokyo Rope Manufacturing Co., Ltd. (1990). "Carbon Fiber Composite Cable". Corporate
Report

TRB (1991). THIRD BRIDGE ENGINEERING CONFERENCE. Transportation Research

Record, No. 1290, volumes 1 and 2, Transportation Research Board, Washington, DC

Triantafillou, T. C. and Plevris, N. (1991). "Post-strengthening of R/C Beams with Epoxy-

Bonded Fiber Composite Materials". ADVANCED COMPOSITES MATERIALS IN
CIVIL ENGINEERING STRUCTURES" Iyer and Sen, eds., American Society of Civil
Engineers, NY, NY, Specialty Conf Jan 31- Feb 1 p. 245.

Trintafillou, T. C. and Meier, U. (1992). "Innovative Design of FRP Combined with

Concrete". ADVANCED COMPOSITE MATERIALS IN BRIDGES AND
STRUCTURES, Neale and Labossiere, eds., Canadian Society of Civil Engineers, Montreal,
Que p. 491.

Walton, J. M. and Yeung, Y. T. C. (1986). "The Fatigue Performance of Structural Strands

of Pultruded Composite Rods". JOURNAL OF THE INSTITUTE OF MECHANICAL
ENGINEERS, London, C286/86 p. 315.

Watson and Ragluipathi (1989). "Glass Fibers". ENGINEERING MATERIALS

HANDBOOK, VOLUME I, COMPOSITES, ASM International, Eds, Metals Park, OH p.
107.

Whiteside, J. B., Delasi, R.J., and Schulte, R. L. (1983). "Distribution of Absorbed

Moisture in Graphite/Epoxy Laminates After Real-Time Environmental Cycling". LONG-
 163

TERM BEHAVIOR OF COMPOSITES, ASTM STP-813, T. K. O'Brien, ed., American

Society for Testing and Materials, Philadelphia p. 192.

Wines, J. C , Dietz, R. J., and Hawley, J.L. (1966). "Laboratory Investigation of Plastic-
Glass Fiber Reinforcement for Reinforced and Prestressed Concrete," Report 2. Corps of
Engineers, Waterways Experiment Station, Vicksburg, MS.

Wines, J.C. and HofF, G. C. (1966). "Laboratory Investigation of Plastic-Glass Fiber

Reinforcement for Reinforced and Prestressed Concrete", Report 1. Corps of Engineers,
Waterways Experiment Station, Vicksburg, MS

Wu, A., Saadatmanesh, H., and Eshani, M. R. (1991). "RC Beams Strengthened with FRP
Plates II: Analysis and Parametric Study". JOURNAL OF STRUCTURAL
ENGINEERING, American Society of Civil Engineers, NY, NY, Vol. 117, No. 11,
November p. 3434.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 167
1993 Elsevier Science Publishers B.V. All rights reserved.

Glass FRP Reinforcing Bars for Concrete

S. S. Fazaa and H.V.S. GangaRaob
a
Research Assistant Professor, Constructed Facilities Center, P.O.Box 6101, West Virginia
University, Morgantown, WV 26506-6101, United States of America
b
Professor, Director, Constructed Facilities Center, P.O.Box 6101, West Virginia University,
Morgantown, WV 26506-6101, United States of America

1. INTRODUCTION AND SCOPE

America's constructed facilities which include bridges, buildings, parking garages,

treatment plants and many other structures are typically built of some combination of
concrete and steel. The steel reinforcing bars in these facilities are exposed to a highly
corrosive environment leading to structural degradation and consequent costly repairs and
user inconveniences.
Despite extensive research to reduce the corrosion damage caused by steel reinforcement,
little progress has been made in improving the longevity of our constructed facilities. Non-
corroding materials, such as fiber reinforced plastics (FRP) possess many characteristics
which can be an excellent alternative to steel reinforcement. FRPs are among the many
advanced composite materials originally developed for use in the aerospace industry.
Adoption of advanced composites technology by the construction industry has been a slow
process, however. Despite the many favorable characteristics of advanced composites, there
are many barriers which restrain their widespread application.
The use of FRP rebars in lieu of mild steel as reinforcing material in concrete is expected
to extend the service life of constructed facilities because of better resistance to corrosion and
higher damping properties than steel. FRP rebars can also result in lower maintenance costs
by eliminating corrosion induced damage. Furthermore, FRPs can non-destructively reveal
their stress state and remaining life under service conditions [1], thus offering the possibility
of monitoring the rate of degradation of structures and providing a basis for developing
intelligent structures.
The objective of this paper is to characterize the material properties of glass FRP rebars
that are produced and used in the United States. In addition, mathematical models, including
micromechanics and strength of materials approaches, are discussed briefly herein to establish
material properties. These models are used to predict the mechanical properties (Young's
modulus, Poisson's ratio and ultimate strength) of FRP rebars individually and also their
performance as reinforcing elements of concrete.
168

Another objective of this presentation is to discuss the influence of various parameters on

pre- and post-cracking bending and bond resistance, crack propagation, ductility, and
deflection limits of concrete beams reinforced with FRP rebars.
Theoretical correlations are presented in terms of elastic and ultimate bending moment,
crack width, and bond and development length. Emphasis is placed on the beam bending
analysis and design using regular as well as high strength (20-69 MPa) concrete reinforced
with FRP rebars by modifying the state-of-the-art design as per the ACI 318-89 code
provisions [2] that are applicable for steel reinforced beams. However, modifications from
current ACI code for FRP reinforced beams in terms of ultimate moment capacity, crack
pattern, and development length are made without deviating significantly from the design
philosophy given in the ACI 318-89 code.
Design equations for bending resistance, bond and development lengths, and crack widths
are developed in a simplified form for practical applications. Similarities and parallels of these
design equations with current ACI 318-89 equations are maintained when possible.

2. OVERVIEW

Since the 1930's, glass has been considered as a possible substitute for steel in the
reinforcing or prestressing of a concrete section. Surface protection of glass from attacks by
the environment was accomplished with plastics. The resultant system consisted of fibers of
glass embedded in a matrix of plastic, generally referred to as a glass fiber reinforced plastic
(GFRP) system. Various forms of glass fibers were produced. These included fiber glass
yarns and cords composed of twisted strands with several surface treatments, as well as glass
fiber rods and tapes, some of them made up of unidirectional fibers bonded together with
plastics or resins. In order to develop a good bond between concrete and glass fiber
reinforced plastic rods different surface conditions have been developed, such as coating the
rods with a layer of epoxy, rolling them in a bed of sand, or sanding the rod with an emery
cloth to create a roughened surface. Surface protection of certain types of continuous glass
fibers, particularly one common fiber type, E-glass, from attacks by the environment or
alkaline reaction has been accomplished by coatings, including resins. For example, a typical
continuous E-glass fiber reinforced plastic rebar has 55 percent glass volume fraction
embedded in a matrix of vinyl ester or isophthalic resin. These thermoset resin systems have
excellent resistance to corrosion and impact, are good electrical and thermal insulators, are
easy to manufacture, and are cost effective.
Manufacturers of composites have long recognized that a continuous manufacturing
process is essential in order to produce a reliable and high quality fiber reinforced plastic
rebar. It is obvious as well that continuous manufacturing processes can lead to higher
production rates (=5 meters per minute) and lower unit costs. One such manufacturing
process analogous to metallic extrusion is known as the pultrusion process. Originally, simple
solid uniform sections reinforced with unidirectional fibers were developed. However, since
the pultrusion process has matured significantly, this method can produce a variety of cross
sectional shapes, including hollow sections with cross-plied reinforcing strips.
International Grating, Inc., Vega Technologies, Inc., and Creative Pultrusions, Inc.
recently introduced to the market a fiberglass reinforced rod that has a deformed surface
 169

(ribbed). The deformation is made by wrapping the rod with an additional strand of resin
soaked fiberglass in a 45 degree helical pattern. The shape of the resulting rod resembles
the shape of deformed steel rebars used in the concrete construction industry. The principal
constituents which influence the strength and stiffness of composites are the reinforcing
fibers, the matrix, and the interface between the fibers and the matrix. An overview of the
constituent materials and the manufacturing process is outlined herein.

2.1. Glass Fibers

Glass fibers are manufactured by the continuous drawing of molten glass from an
electrically heated furnace through platinum bushings at high speed. The filaments cool from
the liquid state, at a temperature of about 1200 degrees C, to room temperature in
approximately 10-5 seconds. On emerging from the bushings, 204 filaments are bundled
together and bonded to one another by a lubricant. This process is known as sizing and has
three main functions:

a) To reduce the abrasive effect offilamentsrubbing against one another

b) To reduce the damage to fibers during mechanical handling
c) To facilitate the molding process

The most common glass fibers manufactured and used in the construction market are E-
glass fibers with low alkali content. It was not until the advent of polyester resin in 1942 that
this glass was used successfully [3]. Now it is employed widely, especially with polyester and
epoxy resins. Other types of glass include Z-glass, A-glass, and S-glass.
The functional requirements of the glass fibers in afiber/matrixcomposite are:
a) A high modulus of elasticity to give stiffness to the composite.
b) A high ultimate strength.
c) A low variation of strength between individual fibers.
d) Stability during handling.
e) A uniform diameter.

2.2. Matrix
Polymers are the products of combining a large number of small molecular units called
monomers by the chemical process known as polymerization to form long-chain molecules.
There are two main types of polymerization. In the first type, a substance consisting of a
series of long-chain polymerized molecules, called thermoplastics, is produced. All the chains
of the molecules are separate and can slide over one another. In the second type, the chains
become cross-linked so that a solid material is produced which cannot be softened and which
will not flow. Such solids are called thermosetting polymers. The thermosetting polymers
which are currently used in the construction industry are polyesters and epoxies.
Thermoplastics show a good potential for use in the construction industry, but manufacturing
processes to incorporate thermoplastics are still under development.
The functions the matrix must fulfill are:
a) To bind the fibers together and to protect their surface from damage during the
service life of the composite
b) To transfer stresses to the fibers efficiently by adhesion and/or friction
c) To distribute the fibers and to separate them
170

d) To be chemically and thermally compatible with fibers

2.3. Interface
The interface between the fiber and the matrix is an anisotropic transition region exhibiting
a gradation of properties through which the transfer of stresses between fiber and matrix
occurs. It must provide adequate chemically and physically stable bonding between the fibers
and the matrix. Its functional requirements vary considerably according to the performance
requirements of the composite during its various stages under service conditions.

2.4. Pultrusion Process

The pultrusion process has been used for making profile shapes (rods, beams, channel,
plates, etc.) since 1948. The detailed schematic diagram is shown in Figure 1. In the
pultrusion process, reinforcements are pulled continuously through a thermosetting resin bath,
shaped into a specific constant cross-section and cured to hardness while being held in the
desired shape. The term pultrusion was coined to differentiate the process from "extrusion"
where plastics or metals are pushed through a die opening. In pultrusion, line speeds are
generally low and the average production output is one to five linear feet per minute.

Figure 1 Continuous pultrusion process

 171

2.5. Minimum Manufacturing Controls For FRP rebars

U.S. Manufacturers (International Grating, Inc, Vega Technologies, Inc., and Creative
Pultrusions, Inc.) have agreed to the following minimum manufacturing controls towards
future standardization of FRP rebar manufacturing.

Rein/./Matrix: 75/25% by weight

Profile: Pitch - 1/2M + 1/8"
Depth - 1/8" + 1/32"
Reinforcement: Type E Glass fiber, 113 yds/lb
Single end roving
Matrix Resin: Dow 411-45 Vinylester
Ashland 922 Isophthalic Polyester
Filler: 20% (of resin by wt.) CaC0 3
Viscosity: To be achieved by the addition of 20% CaCC>3 to "off the shelf resin
Catalyst: 1 % Benzoyal Peroxide (BPO)
Pull Speed: A function of manufacturing method and, therefore, not definable
Die Temp.: 300 degree F maximum
Tension: Each Strand tensioned at creel. Also a function of line speed
Grit Surface: #3 Silica Sand

2.6. Gripping Mechanisms

The design and development of suitable grips for FRP bars in tension tests and in pre- and
post-tensioned concrete have presented major difficulties to researchers. Typically, the FRP
rebar is found to fail in the grip itself due to the combined effects of shear and crushing
coupled with the tensile stresses. An ideal grip must be easy to use in the laboratory for
tension tests and in the field for pre- and post-tensioning applications. The grip must grasp
the rebar in a manner as to avoid failure of the rebar at the grips; on the contrary, in
experimental tension tests the failure must take place in the rebar away from the grip region.
A review of literature on gripping mechanisms does not reveal any availability of a simple
reusable grip that is suitable to test the rebars in the laboratory [4]. Several methods for
anchoring the FRP rebars have been investigated. The sand grips, 7 x 3 in. and 0.75 in. in
thickness, as illustrated in Figure 2, were developed by the Constructed Facilities Center
(CFC) at West Virginia University, and made out of steel. A semi-circular groove is cut out
of each plate. The groove diameter is increased by 1/8 in. over the size of the rebar size to be
tested to allow for variations in the size of the rebar. Fine wet sand on top of an epoxy-sand
coating is used to fill the groove. A top piece is attached with two screws or can be welded
to the plates in order to seat them properly inside the jaws of the Baldwin Testing Machine as
shown in Figure 3. With slight modification to the original set of grips used in the testing
machine, a set of grips have been successfully used outside the testing machines as in the case
of pre-tensioning of the FRP rebars. Six high-strength bolts are used to tie the two plates
together, as shown in Figure 4.
172

Figure 2 Sand grip for FRP rebar testing

Figure 3 FRP rebar test set-up

 173

Figure 4 Bolted sand grips

3. MECHANICAL PROPERTIES

3.1. Tension properties

Static tensile strength and stiffness measured by subjecting FRP rebars to uniaxial tension
are presented in this section. All tension results have a linear stress-strain relation up to 95-
98% of the ultimate strength. For smooth FRP rebar, failure is governed by the tensile
strength of fibers; whereas, matrix cracking is noted for wrapped or ribbed rebars as an initial
failure before fiber breakage. Various researchers have experienced difficulties with gripping
methods. Sand grips, shown in Figure 2, can be successfully used in tension tests [5] to
achieve a gradual and uniform load transfer over the whole gripping surface through friction
and to make certain that failure does not take place in the grips. The sand layer is used to
prevent slippage and to protect a specimen surface from damage that could be caused by
direct contact with the steel jaws of the testing machine.
The tensile stress-strain curves show a linear behavior up to 95-98% of the ultimate
failure load. The average tensile stiffness depends on the fiber type and volume fraction,
and is virtually independent of the manufacturing company, bar size, bar type (with or
without ribs), test procedure, and type of resin. Glass fiber reinforced rebars with
approximately 55 percent fiber volume fraction indicate a mean tensile stiffness of
49.6 x 10 3 MPa (7.2 xl()6 psi). The ultimate tensile strength values determined for FRP
rebars show a strong sensitivity to various factors including bar diameter, quality control
in manufacturing, type of matrix system, and fiber type. The ultimate tensile strength
values are very sensitive to the gripping mechanism. In ribbed rebars, outer fibers break
first, then peel off in the gage section. As the diameter of the E-glass continuous fiber
rebar increases from 0.375 in. (#3) to 0.875 in. (#7), matrix cracking becomes more
pronounced and clearly noticed at about eighty percent of the rebar* s ultimate tensile
strength. Once the critical load is reached, a brittle catastrophic failure can be observed in
174

the gage section. Since the ultimate tensile strength depends on several factors, average
experimental tensile strengths of FRP rebars, with 55% fiber volume fraction, obtained
from different manufacturers are given in Table 1, which is representative of the state-of-
the-art FRP rebar quality. Results in Table 1 reveal that as the bar diameter increases,
the ultimate tensile strength of continuous glass fiber reinforced rebars with vinylester
resins decreases rapidly. The strength reduction is attributed to the resin dependent shear
lag phenomenon.

Table 1
Ultimate tensile strength (MPa)
FRP Rebar Manufacturer Rebar Size
#3 #4 #7
International Grating, Inc. 19.92 15.51 12.50
Vega Technologies, Inc. 13.47 N/A 8.05
MMFG 18.3 N/A 11.64
Creative Pultrusions, Inc. 21.74 N/A N/A

3.2. Compression properties

Static compression stiffness and strength properties are presented in this section. FRP
rebars compression behavior shows a linear stress-strain relation for smooth rebars, while
a slight deviation from linearity is observed forribbedrebars. The combination of matrix
cracking and fiber kinking can be observed for ribbed rebars, while fiber-matrix splitting
failure can be observed for smooth rebars. Unlike the tensile stiffness, the compressive
stiffness values vary with rebar size, type, quality control in manufacturing, and length to
diameter ratio of the test specimen. A summary of ultimate compressive strength values
for 55 percent volume fraction of continuous E-glass fiber and vinylester or isophthalic
resin obtained by Wu [5] is given in Table 2.

Table 2
Compressive strength of FRP rebars (MPa)
Rebar Manufacturer Deformed Smooth
Rebar Rebar
International Grating, Inc. 6.65 N/A
Vega Technologies, Inc. 5.25 N/A
MMFG N/A 9.91

3.3. Torsion properties

FRP rebars show a linear torque versus twist relationship. The failure mode of ribbed
rebars can be observed as a combination of rib breakage and matrix fracture. Matrix
cracking can be observed in smooth rebars. Shear stiffness does not significantly vary
with the manufacturing quality or rebar type and a mean value of 4.55 x 10^ MPa (0.66 x
 175

16 psi) can be obtained as the mean value. However, torsional strength can be found to
decrease with increasing diameter, from 73 MPa for 0.5 in. diameter bar (#4) to 61 MPa
for a 0.875 in. diameter bar (#7). Torsional strength also depends on manufacturing
quality. Once again, the torsional properties are based on 55 percent volume fraction of
continuous E-glass fibers and vinylester resin.

3.4. Bending properties

Static flexural stiffness and strength were be obtained from three point bending tests
[5]. Ultimate bending strength shows a variation with diameter as in the case of ultimate
tensile strength of FRP rebars. The dominant failure mode for both smooth and ribbed
rebars is illustrated by matrix cracking with no buckling of fibers in the compressive zone.
An average value for flexural stiffness in tension can be taken as
46.9 x 103 MPa, whereas the average value for flexural strength in compression can be
taken as 41.4 103 MPa. It is typical for composite materials to have different stiffnesses
in tension and compression. Bending stiffness values are virtually independent of rebar
type or manufacturing quality. However, ultimate bending strength values varied greatly
depending upon the specimen diameter, as it can be seen in Table 3.

Table 3
Bending strength of FRP rebars (MPa)
FRP Rebar Manufacturer Rebar Size
#3 #4 #7 #8
International Grating, Inc. 19.00 14.36 8.84 8.22
Vega Technologies, Inc. 10.94 N/A 8.16 7.26
MMFG N/A N/A 12.09 N/A

4. THEORETICAL MODELING OF MECHANICAL PROPERTIES

Theoretical modeling of the mechanical properties of a rebar, subjected to a variety of

static loads, has been attempted through micromechanical modeling, macromechanical
modeling, and three-dimensional finite element modeling (3D FEM) [5].
The objective of micromechanical modeling is to predict the material properties of the
rebar as a function of the properties of the constituent material. A unidirectional fiber
reinforced rebar can be analyzed as a transversely isotropic material. Fibers are usually
assumed to be isotropic, which is true for glass or Kevlar but not for carbon.
In the macromechanical modeling, FRP rebars are treated as homogeneous but
anisotropic straight rods of circular cross-section. The theory of elasticity solution for
circularly laminated rods was used by Wu [5]. The rebar is assumed to be axisymmetric,
with a number of thin layers of transversely isotropic material comprising the cylinder
wall. A monoclinic material description is used since each layer may have arbitrary fiber
orientation.
A three dimensional finite element analysis using isoparametric elements and
constitutive equations of monoclinic materials was also employed by Wu [5]. Simulations
176

of actual tensile test conditions of FRP rebars are performed assuming a linear distribution
of shear transfer between the gripping mechanism and the rebar. First ply failure along
with the maximum stress failure criterion is employed in these models [6].
The ultimate tensile strength predicted by the 3D FEM is 25% higher than the
experimental value and the theory of elasticity solution [7] results in large discrepancies
with the experimental values. To overcome the limitations of both the 3D FEM and the
elasticity solutions GangaRao, Wu, and Prucz developed a mathematical model using the
strength of materials approach, including the shear lag between the fibers [8]. The
maximum failure strain of the glass fibers (approximately 3%) is considered as the only
governing criterion for failure. The proposed model uses a circular cross section to
compute tensile or bending strength. The major assumption in developing this model is
that the strain distribution across the section is parabolic and axisymmetric. The parabolic
strain distribution is assumed to result from the radial stresses induced by the gripping
mechanism. The model predicts tensile forces in the core fibers lower than those forces at
the surface of the bar. Similarly, the strain distribution under bending is assumed to be
parabolic across the cross section.

4.1. Proposed Theory for Tensile Strength

The theoretical model proposed uses a mechanics of materials approach to correlate the
measured ultimate tensile strength as a function of the bar diameter. The model uses a
circular cross-section to compute bar resistance in tension. Prior to the model development,
three assumptions are made. They are (i) FRP bar fails when the tensile strain reaches the
ultimate fiber strain; (ii) first fiber failure precipitates global failure; and (iii) strain distribution
is assumed to be a combination of two parts - one is uniform throughout the cross-section
and the other is parabolic but axisymmetric as shown in Figure 5. Variable Mc" (thickness of
boundary layer zone) in this approach results mainly from different curing rates in FRP bars
as the size changes. Failure behavior of FRP bars in tension leads to assumptions (i) and (ii).
The third assumption is based on the curing related problem of pultruded FRP bars. This
effect results in the shear lag phenomenon associated with the tensile force resistance
differential of core fibers and those at the outer surface of the bar.

^-<ar 2 +b)E u

Figure 5 Strain distribution across cross-section of FRP rebar in tension

 177

From assumption (iii), the strain distribution becomes

= e u 0<r<R (la)

e t r = (ar2 + b) R-c < r < R (lb)

where "a" and "b" are constants that define the strain distribution, is directly proportional
to degree of cure of resin and specimen size. Vaughan's numerical results showed a 40%
resin cure at the center (relative to full core at outer surface) for a 2 inch diameter bar.
Therefore, this number is used in this approach, i.e., = .4. However, it should be noted ""
is a manufacturing parameter, depending on pulling speed and curing temperature.
By multiplying Equations (la) and (lb) with the effective Young's modulus, Ej, of the FRP
bar, and performing integration over the cross-section, the total force, F t , acting on the
composite bar can be expressed as a superposition of the two strain distributions.

R* R
F t = 2 [ dr + f(ar2+b)rdr] (2)
0 R-c

From the rule-of-mixture micromechanical model for a single, unidirectional E-

glass/polyester composite with 55% fiber volume fraction, the ultimate tensile strength, Xj-, is
165 ksi, if the ultimate strain is , , is 3%. Dividing Equation (2) by the cross-sectional area,
( r 2 ), and using "", Constants "a" and "b" are expressed as a function of the thickness , c,
of the boundary layer zone, the constant strain distribution effect, , and the specimen
diameter, R.

By substituting Equations (3) into (2) and performing the integration, the ultimate tensile
stresses for FRP bars with varying diameters are predicted by the relation

at = 0 . 5 X t [ l + 2 X - ( l - c / R ) 2 ]

The thickness of the boundary layer zone "c" in Equation (4) is chosen to provide the best
correlation with the test results. According to this principle, the constant "c" for International
Grating, Inc., Vega Technologies, Inc., and MMFG specimens has values of (3/32), (3/40)
and (1/32) inches, respectively. The experimental values of the ultimate tensile stresses and
the corresponding theoretical results are given in Table 4. Good correlations are obtained for
all specimens except for #7 Vega Technologies, Inc. bar.
178

Table 4
Comparison of theoretical predictions and experimental stresses under uniaxial tension
Specimen Experimental (MPa) Theoretical (MPa)
size KODIAK VEGA MMFG KODIAK VEGA MMFG
#3 19.92 13.47 18.30 18.53 13.20 17.22
#4 15.51 N/A N/A 16.86 12.36 15.67
#7 12.50 8.05 11.64 14.14 11.22 13.30
KODIAK = International Grating, Inc., VEGA = Vega Technologies, Inc.

4.2. Proposed Theory for Bending Strength

Applying the same concept of the previous section, a new approach for predicting the
bending strength as a function of specimen diameter is derived. Superposition of the two
strain distributions acting on the FRP bar in bending is zero. Assuming that the bars fail at
the outer surface when ultimate bending strain, , is reached, constants "a", "b" and " are
expressed as a function of thickness of boundary layer zone, "c" and the specimen diameter,
"R".

2R2 - (R-c) 2
a- " cR2 (2R-c) (5a)

c 2 (2R-c) 2 -R 4
b - 2cR2 (2R-c) (5b)

c c
= <5c>

The lever arm, , of the cross-section is computed as

R R
CXrdA + f(ar 2 + b ) r d A

r -ME _ 0 R^

" FB - R R ^
+ f(ar 2 + b)dA
0 R-c

Using Equation (5) in (6) and performing integration, we can show that the of FRP bars
under bending depends on both radius, R, and boundary layer zone thickness, c. Once, the
lever arm is determined, the internal moment, M ^ , resulting from the induced bending stress
is calculated as

M F
int = B , (7)
 179

Using different values of c in Equation (7), the best regression values of KODIAK and VEGA
bars under bending are (5/32) inches and (9/64) inches respectively. A comparison of
theoretical and test results is given in Table 5. From Table 5, the proposed model results in
good correlation with the experiments for KODIAK bars but provides a poor prediction for
VEGA specimens, especially for the #3 specimens.

Table 5
Comparison of theoretical prediction and experimental bending moment (three-point bending)
Specimen Experimental (N. m) Theoretical (N. m)
size KODIAK VEGA KODIAK VEGA
#3 76.34 44.18 78.00 75.72
#4 137.44 N/A 157.5 143.78
#7 453.76 418.54 476.30 420.00
#8 630.25 555.80 605.80 531.66
KODIAK = International Grating, Inc., VEGA = Vega Technologies, Inc.

5. FRP REBAR PERFORMANCE AS CONCRETE REINFORCEMENT

5.1. Ultimate moment capacity

The phenomenon of developing moment resistance in FRP reinforced beams is identical
to that of beams reinforced with steel rods, provided that adequate bond between FRP and
concrete is developed. The ultimate moment capacity of FRP reinforced concrete beams
can be obtained by satisfying the internal force and moment equilibrium equation (Eq. 8).

M n = A s fy d (l-0.59p fy/fc') (8)

where Mn = nominal moment capacity of a section, in-lb.

As = area of tension reinforcement, in^.
= specific yield strength of reinforcement, psi.
7 = distance from extreme comp. fiber to centroid of tension, in.
p = ratio of tension reinforcement =A s /bd.
fc = specified compressive strength of concrete, psi.
b = width of concrete section, in.

The ultimate resisting moment based on the equilibrium Equation 8 can be applied for
FRP reinforced concrete beams as shown in Equation 9, by substituting the ultimate tensile
strength of FRP rebars for the yield strength of steel. The validity of this equation is
substantiated through experimental results outlined by GangaRao and Faza [4]. However,
a designer must check for bond capacity of the FRP reinforced beams as well as sudden
crack formation and growth. It should be noted, that to attain the full bending mode
failure, bond, shear and compression failure must be avoided.
180

M n = A s f y f d(l-0.59p fyf/fc1) (9)

where Af = Area of FRP rebars in tension, in^.

fyf = ultimate tensile strength of FRP rebars, psi.

In order to account for the current variation in manufacturing and quality control of the
FRP rebars , the ultimate tensile strength of FRP rebars should be computed using,
fyf = 0.80 ff (rupture failure) as described by GangaRao and Faza [4].
Previous research on FRP rebars as reinforcement for concrete by Nawy and Neuwerth
[9] showed that in most cases it was not possible for the FRP rebars to develop their full
strength due to high tensile strength associated with these rebars. In order to take
advantage of the high tensile strength of FRP rebars, high strength concretes 41-69 MPa
(6000-10000 psi) should be used to maximize the bending resistance of the structure [4].
The cracking moment of the high strength concrete structures will increase and a
substantial decrease in the crack width will be noted due to good bond.

5.2. Flexural cracking

Various limitations on crack widths for steel reinforced beams have been proposed by
different investigators [10]. Excessive cracking is undesirable because it reduces stiffness,
enhances the possibility of deterioration, and adversely affects the appearance of the
beams. The ACI 318-89 code prescribes rules for the distribution of flexural
reinforcement to control flexural cracking in concrete beams. Good detailing practice is
also required to establish adequate crack control.
Cracking occurs when the induced tensile stress in the beam reaches the ultimate
concrete tensile stress. The tensile stress in concrete is transferred to the reinforcing bar
through bond forces developed between the concrete and reinforcing bar. The tensile stress
in concrete at the cracked section is relieved, becoming zero at the time the crack occurs;
however, the reinforcing bar must carry the tensile forces at that cracked section. The
neutral axis position must shift upward at the cracked section in order to maintain
equilibrium of forces at that section. Cracking will continue to take place between old
cracks until the concrete stresses do not exceed the concrete tensile strength. New cracks
will cease to occur because of one of the following reasons:

1) Excessive slip between the rebar and concrete.

2) Reduction in the distance between cracks to transfer sufficient stress to the
concrete.

A number of equations have been proposed [10] for the prediction of crack widths in
flexural members reinforced with steel. The ACI 224 committee report [11] on control of
cracking of steel reinforced concrete beams and the ACI 318-89 code reached the
following conclusions:

1) The reinforcement stress is the most important variable.

 181

2) The thickness of the concrete cover is an important variable, but not the
only geometric consideration.
3) The area of concrete surrounding each reinforcing bar is also an important
geometric variable.
4) The bar diameter is not a major variable.
5) The size of the bottom crack width is influenced by the amount of strain
gradient from the level of the reinforcement to the tension face of the beam.
The committee concluded that the most probable maximum crack width can be
predicted by the Gergely-Lutz expression. The maximum flexural crack width, W m a x ,
according to Gergley-Lutz, can be expressed as
w
max = 0 7 6 f s >/<A X 10"3 ( 10 )

where = Ratio of distances to the neutral axis from the extreme tension fiber and
from the centroid of the main reinforcement. A value of = 1.20 may
be used to estimate the crack widths obtained in flexure.
fs = Maximum stress in the reinforcement at service load level with 0.6 fv to
be used if no computations are available, ksi.

dc = Thickness of cover measured to the center of the first layer of bars, inch.

A = The effective tension area of concrete surrounding the principal

reinforcement divided by the number of rebars. It is defined as having
the same centroid as the reinforcement as defined per the ACI 318-89,
in^.

While the above expression can be used to predict the maximum crack width, the ACI
318-89 code prescribes rules for the distribution of flexural reinforcement to control
flexural cracking in beams [2]. Good detailing practice is required to lead to adequate
crack control. The ACI 318 code specifies:

" When design yield strength fv for tension reinforcement (steel) exceeds 40,000
psi, cross sections of maximum positive and negative moment shall be so
proportioned that the quality z is given by

z = fs ^ (11)

does not exceed 175 kips per inch for interior exposure and 145 kips per inch for
exterior exposure."

Equation (11) will provide a distribution of the reinforcement bars that will reasonably
control flexural cracking. The equation is written in a form emphasizing reinforcement
details rather than crack width; yet it is based on the Gergely-Lutz expression (Equation
182

10). The numerical limitations of z are 175 kips per inch and 145 kips per inch which
correspond to limiting crack widths to 0.016 in. and 0.013 in respectively.

5.3. Crack Width Analysis of FRP Reinforced Beams

The ACI 224.2R-86 committee report on cracking of steel reinforced concrete beams
recognizes that the expected value of the maximum crack spacing is about twice that of the
average crack spacing [11]. Experimental results on steel reinforced beams outlined by
Bresler [12] have shown that the average crack spacing value is about twice the cover
thickness as measured to the center of the reinforcing rebar. Therefore, the maximum
crack spacing is equal to about four times the concrete cover thickness. It is recognized by
the committee that crack width may be estimated by multiplying the maximum crack
spacing (four times the concrete cover) with the average strain in the reinforcement.
The current ACI 224.2R-86 mathematical expressions for predicting crack widths
cannot be directly used to predict the crack widths in high strength concrete beams
reinforced with FRP rebars because of different material properties associated with higher
strength concretes and with FRP rebars [11]. The effects of the tensile strength of higher
strength concrete, the bond strength between concrete and FRP rebar, and the low
modulus of elasticity of FPR rebars are investigated to establish crack spacing and crack
widths.
The low modulus of elasticity of the FRP rebar alone would alter the Gergely-Lutz
expression (Equation 10) for predicting crack widths for beams reinforced with FRP rebars
because the current expressions for maximum flexural crack widths are proportional to the
strain in steel reinforcement. The strains in FRP rebars are expected to be four times
those of steel because the FRP modulus of elasticity is 49.7x 103 MPa compared to 200 x
103 MPa for steel. By substituting the steel stress with strain, the Gergely-Lutz expression
for beams reinforced with steel is rewritten as

w
max = 0 7 6 E s 8 s vdcA x 10"3 ( 12 )

where es = strain in the steel rebars.

Es = modulus of elasticity of steel = 29 x 10^ psi.

In order to incorporate FRP rebar properties in the above expression, where Es/Ef = 4,
by substituting FRP strains, the Gergely-Lutz expression can be modified to

Wmax = 0.076 (4 Ef) 8 f /d^ x 10-3 (13)

= 0.30 (Ef) e f Vd^A x lO-3 (14)

= 0.30 ff yfd^ x 10"3 (15)

where Ef = Modulus of elasticity of FRP rebar = 7.2 x 103 ksi.

ff = FRP stress, ksi.
 183

8f = Strain in FRP rebar.

Equation (15) would be valid under the assumption that the same crack spacing is
expected to occur when using FRP rebars in place of steel rebars. As mentioned earlier,
the crack spacing is governed by the tensile strength of concrete and bond strength of the
reinforcing bar. Therefore, an investigation of the validity of equation (15), which was
developed for regular concrete with steel reinforcement, is carried out for cases when
higher strength concretes are used with the FRP rebars.
Watstein and Bresler have investigated the relationship between the tensile strength of
concrete, the bond strength, and the crack spacing, leading to the crack width calculation
[13]. Their study prescribed the distribution of bond stresses between adjacent cracks to
calculate crack spacing and crack widths. Assuming that bond stress is constant, when the
concrete reaches its ultimate tensile strength, the crack spacing, 1 is defined by Watstein
and Bresler as:

1 = (2 ft' A) / ( D) (16)

where fy' = tensile strength of concrete, psi.

= maximum bond stress, psi.
D = rebar diameter, in.

The crack width may be approximated by an average strain in a FRP rebar multiplied
by the crack spacing , 1.

W m a x = (ff/E f )l (17)

Since crack spacing is governed by the bond stress between the FRP rebar and concrete
(Equation 17), it is expected that a higher bond strength will lead to smaller crack spacing
and in turn will result in finer cracks. Therefore, it will be expected that the crack spacing
multiplied with strain, would be half of the crack width of steel reinforced beams.
Better bond behavior of sand coated FRP rebars reduces crack spacing to half the
expected crack spacing in a steel reinforced beam. Equation (15) which is based on
multiplying crack spacing with stress overestimates the crack width when high bond
strength is developed with sand coated rebars. The quality of the surface condition
(deformations) and the quality of the sand coating may vary without a set of standards. In
order to accommodate all types of FRP rebar surface conditions and to achieve a more
realistic crack width calculation, a crack width equation that incorporates the actual bond
strength of FRP rebars is needed. The equation must be obtained from the experimental
data until the FRP rebar manufacturing process is standardized.
Based on the assumption that the maximum crack width (Equation 15) may be
approximated by an average strain in FRP rebar multiplied by the expected crack spacing,
and by substituting expected crack spacing in equation (16), the resulting expression for
maximum crack width is shown in equation (18).
184

W m a x = (ff / Ef) (2 ft' A) / ( , D) (18)

Substituting for Ef as 7.2 xlO^ psi in equation (18), it can be rewritten as

W
- = 0 , 4 ^ ^ x 1 0 - 3 (.9,

where f/ = 7.5 ^jfc

ff = Maximum stress in FRP reinforcement at service load level
with 0.5 fyf to be used if no experimental results are
available, ksi.
A = The effective tension area of concrete surrounding the
principal reinforcement divided by the number of rebars. It
is defined as having the same centroid as the reinforcement,
in2.

As shown in equation 15, the effects of high strains (that develop as a result of low
modulus of elasticity of FRP rebars together with higher bond strength when sand coated
FRP rebars are used) will produce crack widths that are two times larger than in the steel
reinforced beams. Use of sand coated FRP rebars should permit higher tolerable crack
widths, since the corrosion problem using noncorrosive FRP reinforcement is not a major
concern. As a result, the larger crack widths that might develop in structures should be
considered acceptable and new crack width limits may be required.

5.4. Bond strength and development length

Bond strength is a function of:
1. Adhesion between the concrete and FRP rebars which is controlled partly
by the concrete strength and quality of the rebars.
2. Gripping forces resulting from the drying shrinkage of the surrounding
concrete and the rebars.
3. Frictional resistance to sliding.
4. Mechanical anchorage of the rebar through development length, splicing,
hooks, and transverse rebars.
5. Diameter, shape (with or without wrapping, surface coating) and spacing
of the rebars.
6. Moment stress gradient in a given zone.

5.5. Anchorage and Adhesion Bond

Assume 1^ to be the length of a rebar embedded in concrete and subjected to net
pulling force dT. If dfo is the rebar diameter, is the average bond stress and ff is the
stress in FRP rebar due to bending in a beam which leads to a pull-out force; the
anchorage pulling force dT(= d^ lj) is equal to the tensile force dT of the rebar cross
section which can be expressed as
 185

tfr-^ff (20)
hence, from equilibrium conditions,
f
^ =4 f (21)
where, = Bond strength, psi.
dfc = Rebar diameter, in.
Id = Embedment length, in.
ff = Rebar stress, psi.

from which the bond strength, , can be derived as:

or written in terms of the development length, 1^

d
Id = ^ b (23)

5.6. Basic Development Length

It has been verified by earlier tests on steel rebars that bond strength, , is a function of
the compressive strength of concrete such that

= / (24)

where K = constant.

If the bond strength equals or exceeds the yield strength of a rebar with cross sectional
7C d h
area of A5 = 1 1 - , then
dfc Id > A b fyf (25)

The basic development length, 1^5 can be written as

ldb = k b ^ (26)
where 1% is a function of the geometric properties of the reinforcing bar and the
relationship between bond strength and compressive strength of concrete.

5.7. ACI Approach

The development length of rebars in tension is computed as function of rebar size, yield
strength and concrete compressive strength. The basic development length is modified
186

according to the requirements of the ACI 318-89 code subsections 12.2.3.1, 12.2.3.2 or
12.2.3.3 to reflect the influence of cover, spacing, transverse reinforcement, casting
position, type of aggregate and epoxy coating [2].
The basic development length, 1^5, as specified by the ACI 318-89 code for rebar size
of #11 and smaller is

l d b =0.04^^ (27)

where, 1^5 must not be less than 12 in. and ^ must be less than or equal to 100 which
corresponds to concrete compressive strength of 10,000 psi.

5.8. Basic development length for FRP rebars

For FRP rebars deformed with a helical wrap of 45 degree and sand coated, the
experimental bond stress ( = / dfo 1^), are based on the average of the experimental
results of the cantilever test setup and straight pullout results outlined by GangaRao and
Faza ( 290 KPa for #3, and 203 KPa for #4 rebars) [4]. In addition a reduction factor =
0.75, is used. Using the reduced bond strength values, theoretical development lengths
are obtained using the following bond strength and development length relationship given
in Equation 23 which results in a minimum development length of l^b = 6.5 in. for a #3
rebar and 10.38 in. for a #4 rebar. Using the ACI basic development length stated in
AK f f
Equation 26, (1^5 = K5 r~ ) and applying the theoretical embedment length
calculated above, with fyf taken as the effective yield strength of FRP rebar; fc' taken as
69 MPa (10,000 psi); and A5 as the rebar cross sectional area (in^).

K = (6.5 x 100)/(0.11 x 104,000) = 0.057 for #3 rebar

K = (10.38 x 100)/(0.196 x 85,600) = 0.062 for #4 rebar

From the above calculation of the constant K, the ACI basic development equation can
be modified to account for the use of FRP rebars to

ldb = 0 . 0 6 ^ f (28)
In the above calculations of the basic development length equation, a reduced bond
strength value was assumed. If the maximum experimental bond stress value is used
without a reduction factor, , the constant K will become 0.0426 for #3 rebar and 0.0465
for #4 rebar.

6. CONCLUSIONS

1. Stiffness properties of FRP bars are dominated by the fiber properties and their
volume fraction. They are insensitive to the specimen size and type, matrix, test methods,
 187

and manufacturers. The Young's modulus in bending is lower than in tension by about 8
to 10 percent, whereas the compression modulus value is close to 38x 10 MPa.
Similarly, the mean value of compressive strength is about 50% of that in tension. This
clearly indicates that local buckling of fibers is more critical in governing the bar strength
in compression than the ultimate fiber strength. Shear stiffness of different size bars from
different manufacturers did not vary much; however shear strength variations appear to be
significant. The strength variation are attributed to curing rate of the bars during
manufacturing.
2. The stiffness properties are strongly affected by fiber properties and its volume
fraction. However, tensile, compressive, shear and bending strengths of FRP bars depend
on the specimen curing rate during manufacturing leading to shear lag across the cross-
section. Such phenomenon can not be explained by either the FE model or by elasticity
solutions. However, a new approach presented herein yields theoretical results that are in
good agreement with experimental data.
3. A reusable gripping mechanism for testing FRP rebars was developed by the
Constructed Facilities Center, West Virginia University.
4. Both the ultimate strength deign method and the working stress (elastic) design
method for flexural design of concrete beams reinforced with FRP rebars are acceptable.
Design equations are developed to establish balanced reinforcement for both the ultimate
design theory and working stress theory, so that engineers can design concrete beams
reinforced with FRP rebars as required by the ACI 318-89 provisions.
5. In order to estimate the maximum crack width of FRP reinforced concrete beams,
knowledge of the rebar bond strength with concrete is essential before utilizing the
proposed crack width equation (19). Otherwise, as a conservative design practice, the
crack width may be estimated to be four times that of steel reinforced concrete beams
(equation (15)).
6. The basic development length of FRP rebars should be computed using equation
(28), and utilizing the current ACI 318-89 code modifications for development lengths
without any changes.

7. RECOMMENDATIONS

To improve the applicability of FRP rebars as reinforcing or prestressing elements in

construction, the following work is recommended to be undertaken in the future.
1. Development of standardized manufacturing techniques including surface conditions.
2. Development of standardized testing procedure.
3. Improvement of the strength and stiffness characteristic and their reliability.
4. Development of long term degradation / aging behavior.
5. Understanding fatigue behavior.
6. Creep and relaxation behavior.
7. Thermal properties.
8. Long-term durability under alkaline conditions.
188

8. REFERENCES

1 H.L. Chen, Z. Sami, and H.V.S. GangaRao, NCA-Vol. 14, Vibro-Acoustic

Characterization of Materials and Structures, ASME 1992.
2 ACI Building Code Requirements for Reinforced Concrete (ACI 318-89), American
Concrete Institute, Detroit, MI, 1989.
3 L.C. Hollaway, "Polymers and Polymer Composites in Construction" Thomas
Telford, Ltd, London, (1990).
4 H.V.S. GangaRao and S.S. Faza, WVDOH-RP-83 Phase I Report, 1992.
5 W.P. Wu, H.V.S. GangaRao, and J. Prucz, ASCE Journal of the Structures
Division (to be published, 1993).
6 J.N. Reddy and A.K. Pandey, Computers and Structures, 25(3), 371 (1987).
7 R. Sherrer, Journal of Composites Materials, 1, 344 (1967).
9 E.G. Nawy and G.E. Newerth, ASCE Journal of the Structures Division, 103(2),
421 (1977).
10 Halvorsen, G., Proceedings of the ACI Fall Convention, Lewis H. Tuthill
International Symposium on Concrete and Concrete Construction, pp. 104, 1987
11 ACI 224.2R-86, State-of-the-Art report on High Strength Concrete, American
Concrete Institute, Detroit, MI, 1986.
12 B. Bresler, "Reinforced Concrete Engineering", Vol.1, John Wiley and Sons,
(1972).
13 D. Watstein and B. Bresler, "Reinforced Concrete Engineering", Vol. 1, John Wiley
and Sons, (1972).
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor)
1993 Elsevier Science Publishers B.V. \ go.

Properties and Applications of Vinylon FRP

Rod (CLATEC ROD)
Masakt Okazaki

Industrial and Living Goods Division, Marketing and Development

Department, Kuraray Co., Ltd.
82, 3Chome, Nihonbashi, Chuoku, Tokyo, 103, Japan

Abstract
This paper reports the results of studies to produce polyvinyl
alcohol fiber (hereinafter referred to as Vinylon or PVA fiber) and
the evaluation of Vinylon FRP rod as a replacement of steel
reinforcing bar for construction material.

1. INTRODUCTION
In recent years, FRP products utilizing what is known as Advanced
Composite Fiber, (ACF) e.g. carbon fiber and aramid fibers, have been
used in the fields of civil engineering and construction. [1.3,5]
They are used to overcome the drawbacks of rusting, heavy weight,
electric conductivity, magnetism and the like associated with
conventional iron bars. This paper reports the results of a study to
make hightenacity polyvinyl alcohol fiber ( hereinafter referred to
as hightenacity Vinylon or hightenacity PVA fiber), the evaluation
of hightenacity Vinylon FRP rod (hereinafter referred to as CLATEC
ROD)[2.4] and applications as a replacement of steel reinforcing bar
for construction material: Some possible applications include
stirrups for beams, reinforcing bars for pontoons, floors, and walls
and embankments.
190

2. PVA FIBER FOR CLATEC ROD

2.1. Outline of PVA fiber
Polyvinyl alcohol is obtained from the saponifiation of polyvinyl
acetate. Polyvinyl acetate is obtained by polymerizing vinylacetate
monomer. Vinylacetate was discovered in 1912. In 1927, W.O. Herrman
and co-workers reported the synthesis of vinylacetate. In 1931, he
applied for first patent on polyvinyl alcohol fiber.
In Japan, studies to produce a textile fiber from polyvinyl
alcohol started in 1938 by Sakurada and his collaborators at Kyoto
University.
Kurashiki Rayon Co., Ltd. ( now Kuraray Co., Ltd. ) established
an industrial process for the production of polyvinyl alcohol resin
as well as for the fiber. In 1948, the fiber was given in Japan
general name" Vinylon" . [6] Commercial production of PVA fiber
started in 1950.

(1) Process of production of polyvinyl alcohol resin

Acetic acid-j catalyst

Ethylene A >* Vinyl acetate ^ Polyvinyl acetate * PVA
Oxygen J synthesis polymerization saponification

(2) Process of production of polyvinyl alcohol fiber

PVA aqueous solution + Coagulated fiber >>

wet spinning washing, drying
_ _ ,~ Winding
drawing, finish application

(3) Usage
Originally, the main objective of the production of the fiber was
to obtain a textile fiber for general use. Now, vinylon has little
use for textiles. The fiber is mainly used for industrial
applications.
The following properties of vinylon are especially valuable for
industrial applications :
a) High strength and high modulus.
b) Good resistance to weathering.
c) Good resistance to chemicals, especially to alkali.
Vinylon fiber is applied in the field of rubber hoses, conveyer
 191

belts for reinforcement, and fishery material, such as fishing nets,

long lines and ropes, fishery threads and laver cultivating nets.
Vinylon filament yarn is used for sewing threads, and meshcloths
for agricultural use in the field. Heavy fabric made from Vinylon
yarn is used for cover sheets and hoods for cars and trucks, sheets
for public works and constructions, covering sheets for outdoor goods
tents, and flexible containers, etc. Vinylon cut fibers, the
majority of which is processed into spun yarns and then into various
woven fabrics, are also used for nonwovens, paper and reinforcing
plastics.
Recently, cement asbestos composites are being withdrawn from the
market due to the health hazard attributed to asbestos fibers.
Vinylon replaces asbestos in cement boards. Vinylon is now used
as reinforcing material for cement and concrete composites. Vinylon
production is approximately 40 thousand tons per year in Japan.

2.2. High-tenacity Vinylon for CLATEC ROD

Vinylon is used for reinforcing cement as a replacement for
asbestos. However, with the recent trend of requiring industrial
materials to exhibit better performance, there has also been an
increasing demand for PVA fiber with higher strength and elastic
modulus.

(1) The process of production

Hightenacity Vinylon (type 7901) is comprised of polyvinyl
alcohol with a higher polymerization degree than regular and a
tensile strength of at least 15g/denier (1720MPa). The hightenacity
Vinylon [7] is produced as follows. PVA with a polymerization degree
of 3,500 and a saponification degree of 99 mol percent is dissolved
in water. Boric acid is added to the solution in several percent on
the weight of PVA. A dope solution is prepared by adding to the
solution obtained above nonylphenolethylene oxide in an amount of 5
percent by weight, based on the weight of PVA. The dope thus
prepared is wet spun through a spinneret with 1000 circular holes of
0.08 mm diameter into an aqueous coagulating bath containing sodium
hydroxide and sodium sulfate at70C. The dope leaves the bath at a
rate of 6 m/min. The fiber is, then, in the usual manner,
successively rollerdrawn, neutralized, wet heatdrawn, washed, dried,
heat stretched at 240 C and taken up onto a bobbin to give a filament
yarn of 1,800 deniers/1,000 filaments.
192

The properties of this hightenacity Vinylon are shown in Table1.

The fiber has a higher tensile strength and modulus than
conventional vinylon (Table2).

Table1. Properties of high-tenacity Vinylon (type 7901)

Thickness Tenacity Modulus Elongation

Denier Dia- (MPa) (GPa) (%)
meter(um)
Yarn 1800/1000 2009 44.7 4.9

Single- 1.8 14 2250 59.8 5.0

fiber

(2) Comparison with other fibers

1) Model of molecular structure of fiber
Figure1 schematically shows how the molecular configuration
changes in the fiber during the manufacturing process. Table2 shows
the properties of the fiber compared with those of other fibers and
in Figure2 the representative loadelongation curve of the fibers
are shown.

Aqueous polyvinyl alcohol solution Hightenacity Vinylon

CH - C H - Type 7901
I
OH High crystallinity

Figure1 Molecular configulation of PVA in hightenacity Vinyion

 193

Table-2 Properties of high performance fibers [1]

Aramid fiber Other organic fiber

Unit PPTA Co-PPTA PE Poly- PVA
Kevlar Technora Techmilon arylate Regular High-
49 29 Vectran Vinylon tenacity

Tensile
strength MPa 2744 2744 3332 1470-3430 2626 686-1470 2254

Initial
modulus GPa 127.4 61.7 75.5 78. 4-98. 0 86.2 10. 8-36. 3 59.8

Elongation % 2.3 4.0 4.4 3. 0-6. 0 3.9 7.0 5.0

Density g/cm 3 1.45 1.44 1.39 0.96 1.41 1. 26-1.30 1.3
Diameter m 12 12 12 50-100 23 14 14
Decomposition
temperature V 600 600 500 140(c) 500 240 240
Price(/kg) /Kg 10000 6000 6000 8000 5500 700-1300 2000
(a) (b)

a) 700 / k g for wet formation purpose b) 1300 A g for reinforcement

of concrete c) Melting temperature

(3) Price/performance
In the fields of civil engineering and construction, high
performance fibers (Table-2) for construction fulfill the require-
ments. When we use these materials, we are subject to the
restriction of material price.
It's believed that the high-tenacity Vinylon is excellent in price
/performance. The tensile strength of high-tenacity Vinylon is 20
percent lower than tensile strength the other high performance
organic fibers. However, the price of high-tenacity Vinylon is 1/3
to 1/5 cheaper than the others.
194

(Single fiber)

4000
Carbon fiber
PPTA VECTRAN
{
(HM) / ppTA

Hightenacity
Vinylon
7901

Regular Vinyion
' 5501

Steel bar SD40

n
2 4 6 8 16 18

Elongation (%)

Figure2 StressStrain curves of various fibers

 195

2.3.Chemical resistance
(a) Vinylon fiber is unaffected by animal, vegetable and mineral
oils, and also has good resistance to acids, alkalis and salts.
Table-3 shows vinylon's chemical resistance compared with that of
other fibers.

Table-3 Chemical resistance of fibers

Condition Fiber*
Chemicals
Cone Temp .Time Vinylon Rayon Nylon Polyester
(%> ( e c) (hrs) -6

Hydrochloric acid 10 20 0.1 100 79 77 91

10 20 1 100 83 76 100
10 20 10 100 69 77 95

Sodium hydroxide 1 20 10 100 88 101 99

1 100 100 93 71 75 29
40 20 10 100 0 82 96

Sodium hydrosulfite 1 70 10 96 87 96 97

Benzene 100 20 100 90 88 88 88

^Retention ratio of fiber tenacity after soaking.

For reinforcement of cement materials, alkali resistance of fiber

is very important. From the above Table3 showing general resistance
to chemicals, we find Figure3 and 4 showing the alkali resistance
of Vinylon in reinfocement of cement products. We also find its
excellent property on this application.
196

1.0 O Vinylon
Strength 0.8
Polyester
retention 0.6
Alkali resistance glass fiber:
of fiber 0.4 unmeasurable after 14 days
0.2 E-glass fiber:
(index)
i unmeasurable after 7 days
untreated 7days 14days
Soaking days in cement extract at 80 C
Figure3 Alkali resistance of various fibers

Tenacity

Elongation

Young's Modulus

3 6 9 12 15 18 21 24
Soaking period(months)
Longperiod exposure to aqueous solution extracted from cement slurry
(pH=13). Soaked at room temperature. Measurements of fiber were
conducted in wet state. Mechanical properties of fiber are each
expressed in terms of the ratio to that obtained after 24 hours
soaking.
Figure4 Alkali resistance of Vinylon
 197

It is seen that alkali resistance of vinylon is superior to other

fibers. Vinylon fiber is suitable for fiber reinforced cement with
respect to alkali resistance.
(b) Weatherability of Vinylon
Vinylon is widely used in agricultural and marine industries,
because of its excellent weatherability as shown in Figure5.
Vinyion's durability to ultraviolet rays varies depending on its
chemical structure.
Test: Apparatus : Sunshine Weatherometer
Condition Rainfall for 18 min. during 120 min.
environmental temperature.
of black panel: 63C
Specimen Fibers tested in the form of a net.
100

80

Strength
retention ^0 A
of fiber
Polyethylenp
(%) fiber
40

20

200 400 600 800

Irradiation time (hrs.)
Figure5 Weatherability of Vinylon
198

(c) Durability of Vinylon used as a reinforcing material for asphalt

roofing
In Japan, Vinylon has been used in asphalt roofing sheets. The
following data proves the excellent durability of Vinylon in outdoors
use and especially when contained in certain matrices. In the
following test, roofing sheet samples were exposed outdoors in Tokyo.
The fiber samples for measurement were obtained from the
corresponding roofing sheets by washing off the asphalt. Mechanical
properties of fibers are each expressed in terms of the ratio to that
of the fiber before exposure.

1.1
Tenacity 1.0 K > --0- -o
0.9
0.8

1.1
Elongation 1.0 hO

0.9
0.8

1.1
Young's 1.0 K > -Q
Modulus
0.9
0.8
10
Exposure period (years)

Figure6 Durability of Vinylon

 199

(d) Influence of Temperature on Vinylon in Air

Long time exposure at elevated temperatures deteriorates the
strength of Vinylon. Fiber strength decreasese to a large extent
within a short time at temperatures over 180 C.
120*C

o --..
140C
~"--o
a ^-
80 --a 160C

Strength
retention
60
of fiber
(%)
40

20

_L _L _L _L
10 20 30 40 50 60 70
Heating time (hrs.)
Figure-7 Influence of temperature on Vinylon in Air
200

3. PREPARATION OF CLATEC ROD

(a) Single-Type Rod
A bundle of the hightenacity Vinylon fiber was impregnated with
an epoxy resin and then drawn to a fiber content of at least 60%. The
rod thus drawn was, while the resin was being cured, wrapped
diagonally with the bundle and heat treated. Further, a spiral
formed stirrup was obtained by winding the bundle with the resin
around a die having a designated shape in a heating oven.

Photo-1 Single-type CLATEC ROD

(b) Multi-Type Rod

Multitype rod is made by bonding several rods together.

Photo-2 Multi-type CLATEC ROD

 201

4. PROPERTIES OF CLATEC ROD

4.1. Tensile breaking load
Table4 and-5 summarize the tensile breaking loads of
representative samples of the Single CLATEC ROD. The property was
measured by a method as illustrated in Figure8 and with a tensile
tester (AUTOGRAPH, made by Shimadzu Corporation) at an extension rate
of 25 mm/min.

Table-4 Single-type ROD

Type Nominal Surface shape Tensile breaking

<iiameter(mm ) load (KN)

3S 3 Straight 5.9
3D 3 Diagonal, irregular 5.9
6S 6 Straight 20.6
6D 6 Diagonal, irregular 22.5
7.5D 7.5 ditto 34.3
10D 10 ditto 61.7

Table-5 Multi-Type Rod

Steel bar
CLATEC ROD
Item Unit (SD-30^
7-6D 7-8D 7-10D D-90 D-32

Cross section
oO
Diameter
(outer) mm 23 28 35 19.1 31.8
(minimum) mm 20 25 31 - -
Cross sectional
area cm2 2.0 3.52 5.5 2.87 7.94
Circumference mm 86 105 120 60 100
Unit weight kg/m 0.32 0.52 0.82 2.25 6.23
Breaking load KN 123 225 364 140 383
Breaking strength 1 MPa , 1 632J1 63zl, 657J 42oJ1 42QJ
Rods shown in photos-1,2 (singletype Rod and Multitype Rod).
202

sleeve for fixintg the chuck

CLATEC ROD specimen

Figure-8 How to test CLATEC ROD for tensile properties

4.2. Comparison with steel bars for reinforcing

concrete
Table6 shows the characteristics of CLATEC RODs and steel bars
(SD295A) as specified in JIS G3112 with the same tensile breaking
loads.

Table6 Comparison between steel bars and CLATEC RODs

Item Unit Steel CLATEC Steel CLATEC

bar ROD bar ROD
D10 7.5D D13 10D
Specification
Diameter mm 9.53 7.5 12.7 10.0
Outer dia.* mm - 8.5 - 11.2
Crosssecti- cm* 0.7133 0.442 1.267 0.785
onal area
Density g/cm3 7.84 1.25 7.84 1.25
Unit weight kg/m 0.56 0.070 0.995 0.112

Mechanical
properties
Breaking KN 34.3 34.3 61.7 61.7
load
Strength MPa 490 784 490 784
Elongation % 16 3.4 16 3.5
Initial GPa 211 29 211 29
modulus

*Diameter including diagonally wound part.

 203

Table6 shows that, when compared on the basis of the same

breaking load, CLATEC RODs are lighter than steel bars, having a
weight 1/8 to 1/9 that of steel bars, and having smaller cross-
sectional areas ( about 60 percent of those of steel bars). The
strength (per unit area) of CLATEC RODs is 1.6 times that of steel
bars, while their elongations and initial moduli are about 1/5 and
about 1/7, respectively.
Microscopic observation of the plane of fracture of a CLATEC ROD
upon tensile test according to the method shown in Figure8 revealed
that the fracture had occurred at one place (Photo3). In the cross-
section before rupture (Photo4), the multiplicity of fibers are
uniform. It is observed that the fiber and the matrix resin are
firmly bonded to each other.

Photo3 Fracture plane

(xl60)

Photo4 Crosssection of CLATEC ROD with an optical microscope

204

4.3. Resistance to alkali of CLATEC ROD

In using CLATEC RODs, it is necessary to know their durability
in cement or mortar and in an alkaline soil utilizing quicklime or
coagulating agents. Both the vinylon fiber itself and CLATEC ROD
were tested for alkali resistance under relaxed conditions and under
tension.

(a) Alkali resistance under relaxed conditions.

Samples of CLATEC ROD 6D were used. They were
immersed under no tension in an alkaline solution of aqueous
synthetic cement solution at both 60 C and 100 C. Figure9 shows the
change of strength retention with time. In this acceleration test,
the samples showed about 90 percent strength retention at both 60 C
and 100C.

Strength Aqueous synthetic

60t
cement solution
retention 100t

Ca(OH) 2 : 2 g/1
8ffl-
of rod
NaOH: 10 g/1
(%) 60-
KOH : 14 g/1

x^.
0 7 14 28
pH=12.7

Immersion time (days)

Figure9 The change in strength retention with time

(b) Alkali resistance under tension

CLATEC ROD 6D samples were immersed under tension in
an alkaline solution of an aqueous synthetic cement solution at both
20 eC and at 60C for 1,000 hours and tested for the strength and
elastic modulus. The testing conditions and the results obtained are
shown in Table7.
 205

Table7 Testing conditions and results of alkali resistance under

tension

Temp. Ratio Tensile Elastic Ratio to that of

of Medium of strength modulus original (%}
1
medium 1 tension* Tensile Elastic
! (Q (MPa) (GPa) strength modulus

Control(untreated, dry) 777 33.8 100 100

Aqueous none 688 33.8 89 100

20 alkaline 0.1 665 33.6 86 100
solution 0.3 712 34.1 92 87

Aqueous none 662 29.5 85 87

60 alkaline 0.1 717 32.5 92 96
solution 0.3 770 32.5 99 96

*Ratio of the load applied, to the breaking load.

The above results show that at 20 C and 60 C CLATEC RODs have
an alkali resistance of nearly 90 percent in terms of retention of
strength or elastic modulus, under both relaxed conditions and under
tension.

4.4. Other mechanical and physical properties

(a) Shear strength : Each shear strength was measured by ASTM method.
Tensile and bending shear strength of CLATEC ROD are higher than
glass FRP rod, but inplane shear strength is lower.
Table-8 Shear strength ( Unit : MPa )

Glass CLATEC Method of

FRP rod ROD Measurement
(commercial) 6D

Short beam method 30.6 31.6 ASTM D 4475

Test method for
inplane shear strength 35.4 18.5 ASTM D 3914
Test method for
tensile shear strength 14.9 20.8 ASTM D 3916
206

(b) Adhesive strength

CLATEC ROD has high adhesive strength with concrete (bonding force
required to pull out from concrete) approximate as same as deformed
steel bar.

Table-9 Adhesive strength ( Unit : MPa )

Adhesive Maximum
strength adhesive
Strength
CLATEC ROD 6D 14.8 rod broke

Deformed steel bar D16 10.0 12.8

(c) Electric resistance and relative magnetic permeability

Table10 Electric resistance and relative magnetic permeability

Volume specific Relative magnetic

resistance (fl-cm) permeability

CLATEC ROD 6D 1.2xl0 1 2 0.7

Steel bar 11 x 10~6 7800

(d) Thermal expansion coefficient

Thermal expansion coefficient of CLATEC ROD is 4.4K 10"6 T:,"1 smaller
than steel bar and concrete (7-13 x 10~6 C~l).
 207

4.5. Reinforcement
Bending load was measured respectively reinforcing concrete with
CLATEC RODs and steel bars. Testing beams shown in Figure10 were
prepared. From bending loaddeflection curves, ultimate deformed
loads of reinforcing concrete with CLATEC RODs are approximately same
as the steel bars. But, deflection of reinforcing concrete with
CLATEC ROD is smaller than the steel bars.

(1) Composition of Test beam

Cement : 400kg/m3 Water : 200kg/m 3
Fine aggregate : 775kg/m3 A.E.agent : 0.8 wt%/Cement
Coarse aggregate : 824kg/m5
(2) Reinforcement drawing (Unit : mm)

Steel 3
-#ti =4
f*l
-i 1 1 r
I Stirrup
lf 50
-1000- 4IOO V
CLATEC ROD 6 D

Figure-10 Test beam

50]Bending test

Deflection (mm)
Figure11 Bending LoadDeflection curve
208

4.6. Processability of CLATEC ROD

(a) Bending
CLATEC RODs can be readily bent on the spot by using a specific
jig after heating with a simple hot air heating device (Photo5,6).
Photo7 shows several examples of bent rods.

Photo-5 Photo-6 Photo-7

Heating divice Specific jig Bent rods

The tensile strength of the bent part (A) of CLATEC ROD was
tested. A CLATEC ROD was bent at a right angle with a bending radius
of 3 times the nominal diameter and the other sample with a radius of
5 times the nominal diameter. The bent samples were each buried in
concrete as shown in Figure12 and subjected to tensile test. The
bent parts showed tensile strengths of about 50% and 65% of those
of straight rods, respectively. The results show that the decrease
in tensile strength is smaller with increasing bending redius.
P
A
71
Composition of concrete:
W/C/G=200/420/1692
W/C=48, S/a=48

Figure12 Embedment conditions in concrete and testing direction

 209

The influence of temperature on CLATEC ROD upon heating for

bending was also studied. Figure13 shows the results, with no
appreciable influence upon heating up to 200 C.

100 -O-
~cr

80
Strength
retention
60 Heating conditions
Oven: hot air oven
40 Heating time: 20 min.
Measurement: room temperature.
20

0
140 160 180 200 220
Heating temperature (C)

Figure-13 Effect of temperature on strength of CLATEC ROD

210

(b) Cutting
The usual saws can readily cut CLATEC RODs, because the rods
consist of an organic fiber and a synthetic resin. Photo-8 shows
how a rod is cut.

Photo8 Shows how a rod is cut

4.7. Summary of Properties

The properties of CLATEC ROD utilizing a high-tenacity vinylon
are summarized as follows.
(1) The tensile strength is as high as 784 MPa and the elongation is
also relatively large. When compared with a steel bar that has the
same tensile strength, CLATEC ROD has a weight of 1/8 to 1/9 that of
the steel bar and an elastic modulus of about 1/7.
(2) CLATEC ROD has an alkali resistance of at least 90 percent in
terms of retention of strength and elastic modulus after 1,000 hours'
immersion in a strong alkaline solution of pH=12.7, both under
relaxed conditions and under tension.
(3) Some properties of CLATEC ROD were studied : shear strength,
adhesive strength with cement, coefficient of thermal expansion and
strength of bends.
(4) CLATEC ROD can be bent with a light, simple jig on the spot and
be readily cut with conventional saws.
 211

5. APPLICATIONS

5.1. Shear test of beams utilizing CLATEC ROD

as stirrups
Presented are the results of a study made by J.Kondou, M.Tezuka
of Oriental Construction and M.Okazaki of Kuraray, on the shear
behavior of RC or PC beams utilizing stirrups obtained by
continuously forming into spiralshape an FRP rod comprising the high
-tenacity Vinylon.
(1) Materials used
The CLATEC ROD used in this test was obtained as described in
chapter 3. To obtain a spiralshaped stirrup, the composite rod was
wound around a die having a designated shape under a heated
atmosphere (100 to 110C). See photo9. Table11 shows the properties
of the CLATEC R O D .
Table-11 Properties of the CLATEC
ROD
Nominal diameter: 7 mm
Fiber content: 66.5%
Cross sectional area (Av.): 40.7mm
Breaking load: 24.5 KN
Breaking strength: 599 MPa
Initial modulus: 36.3 GPa
Elongation: 3.0%
Photo9 Spiralshaped stirrup

A reinforcement steel of SD30A was used as the main reinforcement.

For PC steel bars, a 7ply twisted uncoated stress relieved steel
wire was used. For FRP rod utilizing aramid fiber (hereinafter
referred to as AFRP rod ), a rod having a guaranteed breaking load of
125,4KN/piece and elastic modulus of 64.7GPa was used.
The strength of the concrete was 29.4MPa for RC beams and 39.2MPa
for PC beams. Highearlystrength Portland cement and a coarse
aggregate having a maximum size of 20 mm were used.

(2) Shapes and types of beam samples

Tests were made, with the variable parameters being shear span
ratio(a/d), pitch(s) of stirrups (i. e. spiralshaped CLATEC ROD )
and amount of prestress(P).
212

Twenty three samples included RC beams and PC beams. Figure-14

shows representative shape and bar arrangement.
a = 90 3 0 90
2-D6
D6

ra B\4JZ=
I6-D2 2
mm D II 3
3 0 2 1 0 I 3 0 3JU 3.5
27 0 2 01

(a) RC beam sample

a- 1 05 4 0 105
2-D 1 0

BffiMMMMI
Dl 0

VFRP
2 5 0 IE PC I 3 0
31Q Steel bar'. 20

(b) PC beam sample

Figure-14 Bar arrangment of samples (Unitrcm)
(3) Cracking
Experimental results of the 23 beam samples were fully shown in
reference[4]. Comparison of the crack patterns leads to the
following conclusion.
1) Cracks generate more dispersedly with smaller pitch of stirrup.
2) With samples utilizing a main reinforcement of Aramid FRP, high
dispersion of bending cracks and shear cracks is observed. This is
believed to be due to the fact that the surface of Aramid rod is like
braidedrope were sand adhered to the surface, thereby exhibiting
high adhesiveness. Aramid FRP has low elastic modulus compared to
steel. Figure15 shows the relationship between the width of shear
cracks and the shearing unit stress, with the pitch of stirrup taken
as a parameter.
pitch of stirrup
2.0
Shearing ^ r L.^ _ * d/2
o 5d/7
unit 1.0

stress
(MPa) 0 .0.4 0.8 1.2 1.6
Width of shear crack (mm)
Figure15 Width of shear crack vs shear stress
 213

(4) Deflection
Figure16 shows the relationship between the applied load and the
deflection at center span, with the parameter being the stirrup pitch.
Representative series of RC beams and PC beams are shown.

Load 2 0 0 Load 2 0 0
-d/4 2dA
15 15
(KN) (KN) ^^d/2 j
" 5d/7
100 100 ^
fco
50 - 1 50
1 1 I t 1 1 1 1
0 5 10 15 20 0 5 10 15 2C
Deflection (mm) Deflection (mm)

(a)I RC beam (b) PC beam

Figure16 Applied Load vs Center Span Deflection

Deflections are similar until the generation of a crack. After

crack formation, shear failure occurs at smaller deflection for
larger stirrup pitch.

(5) Shear strength

We compared calculated shear strength with observed values. From
this comparison, the following facts have been found.
(a) The type of main reinforcement has only a small influence on the
effect of the stirrup.
(b) With large pitches of stirrup, the observed strength values tend
to be, more or less, smaller than the calculated ones.
(c) The shear span ratio influences the effect of the stirrup only to
a small extent.
(d) Although test data was insufficient for drawing a definite
conclusion with respect to the influence of effective prestress
forces, it is found that the observed values tend to be smaller than
the calculated values with larger effective prestress forces. This
is considered to be due to the fact that with larger effective
prestress forces, the dispersion is smaller. This is also confirmed
from the shape of cracks.
214

(6) Summary
The results of the present experiment are summarized as follows.
1) The effect on shear strength, produced by a stirrup of spiral-
shaped CLATEC ROD can be evaluated by a theoretical equation which
gives approximate conservative values. Separate consideration should,
however, be given for stirrups with large pitch and for large
effective prestress forces.
2) Shear cracks have higher dispersion and bettercontrolled width
with decreasing stirrup pitch.
3) The width of shear crack has nearly linear relationship with the
deformation of stirrup.
4) Spiral-shaped CLATEC ROD gave good results as a replacement of
steel stirrups for concrete RC and PC beams.

5.2. Application of FRP rod to floating bridge

This is to report a PC pontoon utilizing FRP material instead of
steel wires for prestressed concrete or reinforcing steel a first
attempt in Japan.
A prestressed concrete floating bridge has been built over a pond
in a golf course. It has a length of 56.37 m and a width of 4.0 m.
The bridge consists of a 6 unit PC casing formed by precasting,
castinplace concrete abutments, mooring anchor blocks and bearers.
Each precast PC unit has a length of about 7.6 m, a width of 3.0 m
and a height of 1.0 m. The weight is about 14.0 tons.
Figure17 shows a schematic view of the casing structure. The PC
casing is comprised by a box having partitions but no bottom and
formed of concrete cast around a styrofoam core.

Figure17 Schematic view of the casing structure

 215

The reinforcement for concrete consiste of : (a) 8 pieces in all

of aramid FRP tendons (FiBRAK192) and FRP stirrup of continuous
spiral shape (CLATEC ROD) for the side walls : and (b) aramid FRP
(FiBRAK64) in a mesh shape for both the deck board and walls
positioned perpendicularly to the side walls.
The 6 units of PC casing are joined by unbonded aramid FRP tendons
joining through located in the styroforam. The ducts are not
grouted. A neoprene rubber plat is provided between adjacent casings
to make the joint flexible.
The pontoon uses two types of FRP rods : " FiBRA" obtained by
knitting an aramid fiber into braid and molding the braid into a rod
and CLATEC ROD obtained by molding vinylon fiber into a rod. FiBRA
has been jointly developed by Mitsui Construction Co., Ltd., Kobe
Steel Co., Ltd., and Shinko Steel Wire Industry Co., Ltd. and CLATEC
ROD has been developed by Kuraray Co., Ltd. The properties of the
reinforcing FRP's are shown in Table12. Properly specifying the
shape of the FRP to be delivered will improve the executability and
structural performance.

Table12 The properties of the reinforcing FRP

Type Brand Nominal Cross- Tensile Weight

dia. sectional strength
(nun) area (nim^ ) (KN) (g/m)

FiBRA K64 8.0 42.0 62.7 60

FiBRA K192 14.0 135.0 188.2 180

FiBRA K256 16.0 180.0 250.9 240

CLATEC 10D 10.0 78.5 61.7 112

ROD

This floating bridge was designed and constructed by Oriental

Construction Co., Ltd. and this application example is taken from
the report made by M. Tezuka et al., of Oriental Construction
Co., Ltd. [9]
216

Photo10 Straining with simple abutments

Photo11 Framework with reinforcing FRP's assembling

Photo12 After completion

 217

5.3. Application of CLATEC ROD

as reinforcing bar for floor and wall
When an iron mill built an openhearth electric furnace, CLATEC
ROD 10D was used for all of the reinforcing bars to be provided near
hightension cable ducts to protect them from large highvoltage
currents and for heating the electric furnace. The induced current
would cause the floor and wall to be heated, thereby creating cracks
of concrete and a large power loss. The rods were assembled at 10 cm
intervals in 2 stages as shown in the following photographs and the
concrete was poured.
Precast panels having an area of 13 m 2 molded in a frame work
were assembled for the floor construction and those with an area of
5 m 2 for 3 walls. This is also shown in the following photograph.
[10]

Photo-13 Reinforcing floor with CLATEC ROD

Photo14 Assembled at 10 cm intervals in 2 stages

218

5.4. Reinforcing core with CLATEC ROD for

reconstruction of existing slopes by cutting method

This construction method is called Reinforced Railroad with

Rigid Facing, (hereinafter referred to as RRRProcess Method) [11],
and was developed by the Railway Technical Research institute. Tokyu
Construction Co., Ltd., Tenox Co., Ltd. and Kuraray Co., Ltd. are
collaborating. Test results of basic and applied experiments are
reported in the 47th Annual Conference of the Japan Society of Civil
Engineers. [12]

(1) Outline of processing method

This method is to construct permanent walls having a near-
vertical wall face by reinforcing an existing gentle slope of
embankment using steel bars, or CLATEC ROD substituting for steel
bars, and a continuous rigid facing.

(2) Advantages
a) The deformation of slope can be kept very small by
reinforcement of CLATEC ROD and by the use of a continuous rigid
facing. When necessary, columns of improved soil are constructed by
insitu mixing at the shoulder of the slope in advance of cutting
work.
b) A new free area is created next to the slope by changing a
gentle slope of embankment to a nearvertical wall.
c) Safety, good workability and easy handling.

Figure18 Reconstruction of existing slope by cutting

 219

(3) Construction procedure

a) Construct column of improved soil by insitu mixing

b) Excavate the first layer, place CLATEC ROD anchor within a slope
and place a layer of shotcrete on the slope surface.
c) Complete the cutting work by repeating above procedure.
d) Place a castinplace concrete layer directly on the wall surface

Shotcrete

Columns of Anchor bolt or

improved soil CLATEC ROD

Anchor bolt or
CLATEC ROD Castin place
> \ concrete layer

u-^->.. W k s , ,

Figure19 Construction procedure

(4) Application for CLATEC ROD instead of steel bars

The cutting method of existing embankment has only been studied
with respect to reinforced earth with steel bars. CLATEC ROD was
used as a substitute for steel bar. In comparison with steel, CLATEC
ROD features light weight, high tenacity, no rusting and easy
handling. Multi-type CLATEC ROD ( 7-6D, 7-8D, 7-10D) were used.
Two different processes were proposed to anchor the rod end.
CLATEC ROD which are useful for various construction fields. One
comprises covering the rod end with a steel casing and fixing with
epoxy resin. The other comprises bending the end while heating it
with hot air and embedding the bent end into the concrete. See the
following photographs.
220

Photo-15 CLATEC ROD Photo-16 Bending the end

Photo17 Under construction

 221

Photo18 Under construction

Photo19 Completed wall

222

Acknowledgements
The author expresses his deep gratitude to Messers. J. Kondo,
and M. Tezuka of Technical Research Institute of Oriental
Construction Co., Ltd. for their collaboration, and H. Wakui and
S. Tottori of Concrete Structural Laboratory of Railway Technical
Reseach Institute, and T. Tateyma of Soil Mechnics and Foundation
Engineering Laboratory of Railway Technical Reseach Institute for
their guidance through the course of this study.

References

[I] Kobayashi, et al. J. of Prestress Concrete, Japan Vol.35 No.5.21

(1988)
[2] M. Okazaki, The 44th Annual Conference of the Japan Society of
Civil Engineers, (Oct., 1988, at Nagoya Institute of
Technology), Symposium " Construction and New MaterialsUse and
Problems of HighPerformance Fibers"
[3] K. Kobayashi, Proceedings of Japan Society of Civil Engineers
No. 420/V-13 1990-8
[4] M. Tezuka, J. Kondo and M. Okazaki, Proceeding of the Japan
Concrete Institute, Vol. 12, Dec. 1990
[5] H. Wakui, Concrete Journal, Vol. 28, No. 12 (1990)
[6] I. Sakurada, Polyvinyl Alcohol Fibers. (International fiber
science and technology series ; 6) 1985 by Marcel Dekker. Inc.
[7] U.S.Patent, Patent Number : 5110678, Date of Patent : May 5, 1992
Narukawa et al.
" SYNTHETIC POLYVINYL ALCOHOL FIBER AND PROCESS
FOR ITS PRODUCTION"
[8] M. Okazaki and M. Tezuka, " Symposium on application of continuous
fiber reinforcing materials on concrete structures" , in April,
1992, in Tokyo
[9] M. Tezuka, et al. Civil Engineering Journal 1992. Sep. (The
DOBOKU SEKO) Vol. 33 No.9
[10] K. Shiroyama, Plastic Industry (GOSEI JYUSHI KOGYOU) Vol.39,
No.3 (1992)
[II] The Catalogue of ASSOCIATION OF RRR METHOD
" REINFORCED RAILROAD WITH RIGID FACING METHOD
[12] M. Okazaki, M. Tateyama, Y. Tamura and T. Kami, The 47th Annual
Conference of Japan Society of Civil Engineers, (Sep. 1992, at
Tohoku University)
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 223
1993 Elsevier Science Publishers B.V. All rights reserved.

CFCC (Carbon Fiber Composite Cable)

Norihiko Santoh, Deputy General Manager
Tokyo Rope Mfg. Co., Ltd., 5707 Shisikura, Dejima-Mura, Japan

Abstract
CFCC is the trademark for a carbon FRP cable. Because of its composition and
manufacturing it possesses attractive qualities such as high strength, light weight, excellent
corrosion resistance, lower relaxation, and non-magnetic properties. Its configuration as a
cable allows CFCC excellent flexibility and adhesion to concrete, as well as ease of
preparation.

The advantages of CFCC cable flexibility are:

1. Ease of transportation due to the ability to reel the cable

2. Fabrication of products with long spans
3. Easy handling
4. Ability to bend

Presently, the diameter of the products which can be manufactured is within the range of 1.3
40.0 (breaking load: 1.43 kN1080 kN). However, depending on demand of
the customers it is possible to widen this range.

1. MATERIALS

CFCC is composed of a prepreg in which PAN (polyacrylonitrile) based carbon fibers are
impregnated with epoxy resin.

A prepreg of carbon fiber is fabricated through the process as shown in Fig. 1-1.

Aery Iic Ni tr iIe I

Polymerization Wet Spinning Precursor
Special
Comononer
Prepreg Conversion into
Carbonization
Flame Resistance
Solvent
X
Resin Impregnation Surface Treatment
Catalyst

Fig. 1-1 Manufacturing process of carbon fiber prepreg

224

At the time of wet spinning carbon fiber becomes 12,000 pieces of filament, and in the
process of carbonization the diameter of spun yarn becomes 7 . The basic physical
properties of carbon fiber in use are as shown in Table 1-1.

Table 1-1 Basic physical properties of carbon fiber

Raw material Tensile strength Tensile modulus Elongation at break
(kN/mm2) (kN/mm2) (%)
Carbon fiber 3.63 235 1.5

Resins are divided by their heat-resisting properties. The basic physical properties thereof are
as shown in Table 1-2. Each heat-resisting grade of resins has approximately the same basic
properties and unless the type of resin is specified, 130C type resin is used.

Table 1-2 Basic physical properties of resins

Heat resisting Resins Tensile strength Tensile modulus Elongation at break
grade (N/mm2) (N/mm2) (%)
130C type Denatured epoxy 88 3,530 4.2
resin

180C type Heat-resisting 69 3,920 1.7

epoxy resin

240C type Bismaleic amid 78 3,530 2.5

resin
 225

2. MANUFACTURING AND QUALITY CONTROL

2.1 Manufacturing Method

The CFCC manufacturing process is shown in Fig. 2-1. First, plural pieces of prepreg are
made into a prepreg strand. The surface of this strand prepreg is treated by special coating
and turned into a line-shaped body. Then plural pieces of prepreg strands are stranded and
turned into a composite stranded body. This composite stranded body is finally heated,
hardened, and finished as a product.

PREPREG Layer Special Coat- Stranding Heating and

ing Resin Machine Hardning Line
CFCC
Prepreg i t
Strand prepreg Composite linear Resin is
turned into
strand
is turned into body is turned hardened
prepreg
conposite linear into a composite
body. strand body.

Fig. 2-1 CFCC manufacturing process

By means of combining 12,000 pieces of prepreg toe, products can be done easily. Presently,
the diameter of the products which can be manufactured is within the range of 1.3
mm<|>40.0 . However, depending on requirements, it is possible to widen this range.
In addition, it is also possible to manufacture on untwisted cable, if required.

2.2 Quality Control

Operation and inspection data are recorded and monitored throughout the manufacturing
process.

Specimens are sampled from both ends of the cable for monitoring product quality. Diameter,
twisting pitch and unit weight of each specimen are measured. Void is monitored by a
photographic picture or image analysis technique linked with optical microscopy. A terminal
fixing is formed on each specimen after dimensional measurement. This is followed by
destructive tests to determine breaking strength and tensile modulus. The volume of one
product lot consists of one cable which is normally 600 m or 1200 m in length. Sampling
is usually 3 pieces at each end, and by this means testing is carried out with 6 specimens
from each cable. Specimen size is usually 1.6 m for 12.5 diameter 1 x 7 CFCC structure
including terminal fixing parts. Data are treated statistically and serve as process and
production quality control.
226

3. FRP PROPERTIES

3.1 CFCC Standard Specification

CFCC properties used on the market are shown in Table 3-1.

Table 3-1 CFCC standard specification

Construction Diameter Effective section area Guaranteed breaking load Unit weight
(mm) (mm2) (kN) (g/m)
Single 3.0 3.2 ( 5.1) 9.8 10 |

Single 5.0 9.7 ( 15.2) 28.4 30 |

1x7 5.0 6.5 ( 10.1) 17.7 24 |

1x7 7.5 19.4 ( 30.4) 56.9 64 |

1x7 10.5 35.6 ( 55.7) 96.1 114 1

1x7 12.5 48.5 ( 76.0) 142 151

1x7 15.2 72.5(113.6) 199 226

1x7 17.8 98.8 (154.9) 272 310

1 1 x 19 19.3 115.0(180.2) 277 361

1 1 x 19 20.3 123.8 (193.9) 297 389

1 1 x 19 21.8 141.8 (222.2) 340 445

| 1 x 19 25.0 185.6 (290.9) 446 583

| l x 19 28.0 238.8 (374.1) 574 750

1 1 x37 31.5 291.9 (457.3) 651 916

1 1x37 35.5 377.3 (591.2) 841 1,185

1 1 x37 40.0 480.3 (752.6) 1,072 1,508

The effective section area shows the area of carbon fiber only.
Each figure in parenthesis shows the section area of carbon fiber and resin.

When the specification is within the range of the above table, the products with any adequate size can be
manufactured.

Although such CFCC as turned into non-rotation on its axis can be supplied, special specification is applied.

Density of CFCC is approximately 1.5 grams/cm3.

 227

3.2 Electrical and Magnetic Conductivity

(1) Magnetic Permeability

Test results of magnetic permeability are shown in Table 3-2.

Table 3-2 Magnetic permeability

Magnetic permeability at 1 kOe.
CFCC Less than 1.000 (precision limit of the apparatus)
Conventional steel strand 1.003

(2) Magnetic Inductive Heating

Changes of the surface temperature during the period from start to 13 minutes lapsed
time are given in Table 3-3. The test configuration is shown in Fig. 3-1.

Table 3-3 Magnetic inductive heating

Output power of generator 2.5 kW 14.0 kW
CFCC 1 x 7 12.5 Not observed Not observed
Aramid rod 12.0 Not observed Not observed
Non-magnetic PC 1 x 7 12.4 +95C +432
Non-magnetic PC wire 4.2 +22C
Steel wire 4.2 (+700800C at 5 sec.)

Room temperature
Frequency: 9.1 kHz
Coil Heating tube
(Length=l. Om Inner dia. =16mm0)

mmimmi \
Test piece

o- High frequency generator

Fig. 3-1 Testing method

228

3.3 Tensile Modulus

There is almost no difference in the elastic modulus of CFCC due to difference of the
construction, diameter and heat-resisting grade as shown in Table 3-4.

Table 3-4 Elastic modulus classified by construction, diameter

and heat-resisting grade of CFCC
Construction Diameter (mm) Heat-resisting grade Elastic modulus (kN/mm2)

Single wire 3.0 130C type 226 (147)

Single wire 5.0 130C type 225 (137)

1 1x7 5.0 130C type 215 (137)

1 1x7 7.5 130C type 215 (137)

| 1x7 10.5 130C type 215 (137)

1x7 10.5 240C type 219 (139)

1x7 12.5 130C type 215 (137)

1 x7 12.5 180C type 216(138)

1 x7 17.5 130C type 215 (137) |

1 x 19 22.4 130C type 215 (137)

1 x 19 25.0 130C type 209 (133)

1 x37 40.0 130C type 200 (127)

(Note) In figuring out an elastic modulus, the section area is based on carbon fiber only. Each figure in
parenthesis shows the case of calculation with the section area of carbon and resin.

3.4 Thermal Expansion Coefficient

The measured results of linear expansion coefficient by raising temperature at the rate of 2C
per minute in the air are as shown in Fig. 3-5. There is almost no difference in the results
due to difference of the heat-resisting grade resin. The values were approximately (0.50.6)
x 10"6/C. These values are approximately 1/20 of that of steel. In this case, the range of
measuring temperature was 20C200C.

Table 3-5 Linear expansion coefficient of CFCC

Test piece The coefficient of linear expansion
CFCC 1 x 7 12.5 0.6 x 10"7C
CFCC 1 x 7 240 10.5 0.5 x 10'7C
PC 1 x 7 12.4 12.0 x 107C
 229

3.5 Static Tensile Characteristics

The mean value of the breaking load is about 90% of the collective breaking value of carbon
fiber in use. Taking dispersion of the mean value into account, guaranteed breaking load is
set at 80% of the collective breaking load.

Fig. 3-2 shows an example of the load-elongation curve. It shows deformation of an elastic
body which has no yield point until the body fails under load. It is the characteristic of this
elastic body that there is almost no generation of residual warp in the elastic body even if
load is added to it repeatedly. The elongation at break is lowered because there is no plastic
region.

0. 5 1 1. 5
ELONGATION {%)

Fig. 3-2 Load-elongation curve

230

3.6 Creep and Relaxation

(1) Creep

Result of creep tests with CFCC is shown in Fig. 3-3. Creep tests were conducted
under the following conditions:

Specimen: CFCC 1 x 7 12.5

Condition: Room temperature (22C)
Testing load: 92.4 kN (65% of nominal breaking force 142 kN)

0.008

I J
0.006

0.004 I j-
\^^ l l
0.002
1 i^tf^ j
T ^^"^
1 J I I
1 2 5 10 20 50 100 200 500 1,000
Tiie (Hrt)

Fig. 3-3 Creep of CFCC

 231

(2) Relaxation

Results of relaxation tests which were conducted under different loads or temperatures
are shown in Fig. 3-43-7. In order to compare with steel strand, the tests are
conducted with CFCC and PC steel strands.

0. 96S
0.81X
0.48X

Fig. 3-4 Results of room temperature relaxation tests

(CFCC 1 x 7 12.5 mm)

Fig. 3-5 Results of room temperature relaxation tests

(steel strand 1 x 7 12.4 mm)
232

10
BtftJi
8h
Tempejratu \y
DO. 2?G
6
0.65Pi
22C !
4 O 0.50Pi
I B i
2
2 i }

Of
2 3 5 10 20 30 50
Time (Hrs)

Fig. 3-6 Results of elevated temperature relaxation tests

(CFCC 1 x 7 12.5 mm)

20
jBonc
Teiperature,
15 Q 0a1n...f$..
-1 v Q D
D
D flBPii \
22C \ i
10

A
* ; 0.65PU j

O o
5l ? 0.50PU j
0&
2
_
i3

O
5 10 20 30 50
Time (Hrs)

Fig. 3-7 Results of elevated temperature relaxation tests

(steel strand 1 x 7 12.4 mm)
 233

3.7 Fatigue

In case of CFCC, mean stress is considered as the governing (effective) factor against fatigue.
The rapid progress in fatigue occurs when the mean stress is beyond 1470 N/mm2. The
CFCC, however, has better tensile fatigue characteristics when compared with PC stranded
steel wire.

Table 3-6 Legend

CFCC 1 x 7 12.5 PC 1 x 7 12.4
Not failed at 2 million cycles O

Failed before 2 million cycles A

600
0.65Pu
500 i
o
400
CO
LU
GC
- (
- L
CO 300 c
1
LT*-
o
LU
i
200 " \

Q_
ft < \ < ) OOf

100 "
o o i o
0
0 500 1,000 1,500 2,000
MEAN STRESS (MPa)
Fig. 3-8 Tension fatigue of CFCC 1 x 7 12.5 and PC 1 x 7 12.4
234

3.8 Heat-Cycle Test

In the case of a concrete structure being put into practical use, it is inevitably subject to
change in ambient temperature. Because the coefficient of linear expansion of CFCC is
different from that of concrete, this results in extra stress to the CFCC due to the difference
of the coefficient of linear expansion. It is feared that the bond of CFCC to concrete at the
interface is deteriorated by this extra stress. In view of such a situation as this, the simulation
test was conducted by making a test piece of concrete using CFCC for tension material.
Assuming that the change in ambient temperature during a year is -10C+50C and that the
concrete structure is used for 50 years, the temperature cycle was applied 50 times.

150 2500

WA>##A>A>AtA>AtA>A>A>A>^A>MA>A>MMA>A>A>A>A>A>A>A>A>A>#A<A-
^ CONCRETE

CFCC 1 X 7 12.5

Fig. 3-9 Shape of the test piece

Table 3-7 Characteristics of concrete member

Design standard strength of concrete ack = 39.2 MPa
Initial tensile force of CFCC Pi = 82.3 kN/pcs
Effective prestress ace = 2.89 MPa

OB

|j 20

0
Q_
3E
LU
- -20
5 10 15 20 25
TIME (Hrs)

Fig. 3-10 Change in temperature during 1 cycle

 235

Table 3-8 Results

Design Normal Temperature
value temperature cycle (N = 50)
At the time of Load (kN) 11.9 15.7 15.0
crack initiation
Displacement of 0.9 2.0 1.5
member (mm)
At the time of Load (kgf) 21.4 24.3 27.3
destruction
Displacement of 1.53 1.92 1.90
member (mm)

The loads at the time of crack initiation are approximately the same for all specimens but the
loads at the time of destruction vary more or less probably due to dispersion of the
characteristics of the members. From the above results, a decrease of strength of the
members was not recognized even after the temperature cycle was carried out. In addition,
the reason why the measured values were detected in general as slightly greater than the
design values, it is conceivable that the strength of the concrete members had greater values
than the design standard strength (5ck = 39.2 MPa). Whereas the design values were
calculated by using effective prestress, it is presumed that prestress of the tested members had
not yet reached the effective values.

3.9 Alkaline Exposure

In consideration of the necessity of long time use of CFCC, we are at present running a long-
time exposure test in three places in Japan. Some of these samples have been under test for
more than two years and hardly show any deterioration.

(1) Test site : Our company's laboratory

(2) Sample : CFCC 1 x 7 10.5

Table 3-9 Surface state after exposure for 600 days

Sample Exposure Tension (kN) State after exposure for 600 days
No. conditions
1 20% NaOH aq 47.3 (0.52 Pu) Although the cover material in the
neighborhood of interface eluted, no
irregularity occurred in the dipped part
and part exposed to the air.

12 47.3 (0.52 Pu) No irregularity

3 61.1 (0.68 Pu) No irregularity
236

4. PERFORMANCE AS REINFORCEMENT

4.1 Bond Characteristics to Concrete

In the case of using CFCC as the tension material of prestressed concrete, bond characteristics
of CFCC to concrete become important. The following bond test was conducted in view of
such a situation. The method of test and test results are shown in Fig. 4-1 and Table 4-1
respectively. In addition, in order to confirm stability of bond, a load equivalent to 60% of
simple pulling out load was applied to the test piece 10 times and bond strength after 10
loadings was measured, as well.

Bond strength and the area of the bond surface were defined with the following expressions:
Pulling out force (kN)
Bond strength (kgf/mm2) =
Bond surface area (cm2)
Diameter 240
Surface area of stranded cable = [ xx x 6] x bond length
3 360

Fig. 4-1 Method of bond test

4.2 Flexibility

By forming a beam supported at 2 points as shown in the following figure, concentrating a

load at one point located in the center of the beam and applying the measured values of load
and deflection of the beam to the official expression, the El value was calculated in the
reverse way. In both cases of 1 x 7 and 1 x 37 construction, the El value was approximately
1/4 of the value of steel respectively.

Fig. 4-2 Measuring method of El value of bending stiffness

 237

Table 4-1 CFCC to concrete bond characteristics

Construction 1 x7 Single wire
Test piece CFCC 1 x 7 PC 1 x 7 CFCCU SD deformed
12.5 12.4 6.0 reinforcing bar
6.0

Pull out force 57.1 22.2 23.6 11.8

(kN)

Bond strength 7.22 2.85 7.83 4.17

(MPa)

Repeating frequency 3,500 kgf 1,360 kgf

X X
10 times 10 times Specification of concrete
Water cement ratio: 40%
Pull out force 54.3 24.0
Sand percentage: 10%
(kN)
Age: 4750 days
Bond strength 6.86 3.08 Compressive strength: 47.6 MPa
(MPa)

Table 4-2 Measured results of El of bending stiffness

Type Bending stiffness El (MPa)
CFCC 1 x 7 12.5 568
Stranded steel cable 2080
1 x 7 12.4
CFCC 1 x 37 40.0 10100
Stranded steel cable 39500
1 x 37 40.0
238

4.3 Shearing Characteristics

Compared with other materials, shearing characteristics of CFCC are shown in the following.

(1) Specimen and tensile properties

Table 4-3
Cross section area Tensile strength Elongation at Break
(mm2) (GPa) (%)
CFCC 1 x 7 12.5 76.0 2.12 1.6
Aramid rod 12.0 90.5 1.42 2.0
Glass rod 12.0 113.0 1.37 3.0
PC 1 x 7 12.4 92.9 1.72 3.5
Steel rod 12.7 126.7 0.49 18.0 1

(2) Testing setup

Load

Fig. 4-3 Testing method of shearing

(3) Results
Table 4-4
Perpendicular shearing strength (MPa)
CFCC 1 x 7 12.5 236
Aramid rod 12.0 219
Glass rod 12.0 151
PC 1 x 7 12.4 801
Steel rod 12.7 523
Maximum pushing load x 0.5/Cross sectioned area

Ratio of perpendicular shearing energy normalized by that of CFCC is shown in Table

4-5.
 239

Table 4-5
Shearing energy ratio
CFCC 1 x 7 12.5 1
Aramid rod 12.0 1.20
Glass rod 12.0 0.66

4.4 Method of Anchorage

Generally, a CFCC fixing method is possible in the "Resin filling method" and "Diecast
Method."

The resin used is newly developed by us and the type of resin must be changed for each
fixing method. Also, the performance of resin varies greatly according to how it is handled.
For all these reasons, end fixers must be specially manufactured.

Table 4-6 shows examples of end fixing methods suitable to various usage.

Table 4-6 CFCC end fixing method

Fixing Applicable size Fixing material Resin type Fixing Use example
method (mm) efficiency
Metallic Non- A B c (%)
metallic

Up to 12.5 O O O 100 Out-cable

Post-tension method
Resin
filling 12.5 and over O Being O 100 Out-cable
but up to 40 studied Post-tension method

Concurrent O O 100 Post-tension method

fixing of a
number of cables

Diecast Up to 17.8 O Min. 90 Pre-tension method

Post-tension method
240

(1) Resin Filling Socket

The length of the fixing metal piece is determined based on the bonding strength of
the CFCC and resin but the standard length is 13.5 times the CFCC diameter. The
outer diameter of the fixing metal varies by the material used.

The processing procedure is as follows:

CFCC

Fixing Socket Resin Charging Heating Completion

Resin

Fig. 4-4 Process in resin fixing method

(2) Diecast Method

In the diecast method, a special alloy is diecasted onto the CFCC end and the diecast
part is fixed with a wedge. In this case, a wedge that is commonly used for PC steel
strand wires can be used. At present, this method is applicable to CFCC 1 x 7 7.5
and 12.5 and the wedge for PC 1 x 7 12.7 and PC 1 x 19 19.3 are used,
respectively. The setting loss is minor. Also, the work of pushing in the bottom,
needed in the wedge method, is not necessary. Therefore, a wedge can be mounted in
a one-touch motion and the tension can be immediately introduced. The wedge can
be re-used and the special alloy can also be used again by re-melting, reducing the
cost comparatively. Thus, this method is excellent in the applicability and workability.
 241

4.5 Bending Tensile Test

(1) Specimen CFCC 1 x 7 12.5

(2) Installation

A metal sheave was fixed (no rotating) on one corner of a 500 kN tensile testing
machine. The specimen was put on the sheave and fixed by a M-S-R terminal.
Different diameter sheaves were used for different D/d ratios. The groove coming in
contact with CFCC had three different shapes; no groove (plain), R6.5 groove or R6.5
groove + Polymer sheath.

(3) Results

Sheave No Groove Groove

shape groove (R6.5)

Sheave 520 500 500

bottom 650 625 750
diameter 770 750 1,000
D (mm<|>) 895 875
1,020 1,000
160 I
O No groove
Groove
D Groove and
Fixed sheave sheath 90
material grade
in JIS S45C 140
o <& g
fco
80

120
& 70

J L
100 100 70 50
CFCC 1 X 7 12.50 D/d 60
Breaking strength = 161kN
_L _1_
Tension d=12.5 1 2
d/D(XlO"2)

Fig. 4-5 Testing method Fig. 4-6 Bending tensile test results
242

4.6 Small Angle Bending Tensile Test

Specimen CFCC 1 x 7 12.5

Fixing box
Fixed saddle for bending

Tensioning and load measuring

\
op Specimen
/
3. 7 m

Fig. 4-7 Fixing setup

Saddle radius R: 150, 200, 300 mm

Bending ratio R/r: 24, 32, 48 (2r = 12.5)
Bending angle a: 5, 10, 15, 25, 30
With or without Polyethylene sheath between the cable and saddle

2.2 10
<)
? 2 ()
< '
==1.8 1i -10
)
li.e ii {i -20
I
CO [ \\ !
1

S1.4 1
-30 g

CO

NO; MAL WITH PE -40

Si. 2 2>.... A

. 1
. -50
10 5 10 15 20 25 30 35
BENDING ANGLE (DEGREE)

Fig. 4-8 Relationship between tensile strength and bending angle

for R/r = 48 with and without sheath
 243

5. APPLICATIONS

5.1 Shinmiya Bridge

Strongly affected by the seasonal wind in winter, Soto-Noto of Ishikawa Prefecture is counted
as a prominent salt damage area in Japan. PC steel wire used for a PC bridge is very much
corroded and breakage of the PC steel has been recognized on two bridges. In such an area,
the development of an anticorrosive material to replace the PC steel is very much wanted.

The Civil Engineering Department of Ishikawa Prefecture, P.S. Co. and Tokyo Rope Mfg. Co.
have jointly constructed a bridge using carbon fiber for the first time in the world, under the
guidance of the Public Works Research Institute of Ministry of Construction. This bridge is
explained as follows:

Bridge length: 6.1 m

Width: 7.0 m
Concrete strength of main beams: 49.0 MPa
Installation tensile strength:
Tension during operation: 0.6 Pu (Pu: guaranteed breakage load)
Tensile strength immediately
after installation: 0.55 Pu
Design tensile strength: 0.45 Pu
Reinforcement: Epoxy coated steel bars
(JISG3112 SD295)
Safety:
Bridge actual breakage load is higher than
1.3 x (dead weight) + 2.5 x (live weight)
and also is higher than 1.7 x (dead weight) + (live weight)

Fig. 5-1 Cross section of the beam

244

8 200

Fig. 5-2 Cross section of the bridge

Fig. 5-3 Completion of the bridge

 245

5.2 Hakui Bicycle Road #15 Bridge

The Hakui Bicycle Road #15 Bridge is a bicycle bridge. Construction was completed in
Ishikawa Prefecture in March of 1991.

Bridge length: 10.5 m

Width: 4.3 m
Concrete strength of main beams: 49.0 MPa
Installation tensile strength:
Tension during operation: 0.6 Pu
Tensile strength immediately
after installation: 0.55 Pu
Design tensile strength: 0.45 Pu
Reinforcement: CFCC and Epoxy coated steel bars

Fig. 5-4 Cross section of the beam

4300
ml 350Q 400

ml
c
m 4^
3

J3$0 ' 3W
-20114^800-3200

Fig. 5-5 Cross section of the bridge Fig. 5-6 Completion of the bridge
246

5.3 Hakui Bicycle Road #16 Bridge

The Hakui Bicycle Road #16 Bridge is a bicycle bridge. Construction was completed in
Ishikawa Prefecture in March of 1992.

Bridge length: 7.3 m

Width: 4.3 m
Concrete strength of main beams: 49.0 MPa
Installation tensile strength:
Tension during operation: 0.8 Pu
Tensile strength immediately
after installation: 0.7 Pu
Design tensile strength: 0.6 Pu
Reinforcement: All CFCC

700
640
H
CFCC 6=1. 5mm

CFCC d=12. 5mm ~

kL

lea @ @ ro ta fa a &

3@60 130 3@60

Fig. 5-7 Cross section of the beam

4 300
3500 4Qd

CO
J 2 3 L K 22L_8f
U j U

Im 4(9 800=3 ?00 J55Q

Fig. 5-8 Cross section of the bridge

 247

5.4 Tea-House Beams

Although wood is a major material for building in Japan, it has disadvantages like weakness
or low elasticity coefficient and the use is limited. Accordingly, reinforcement of wood with
iron plate or stainless steel wire has been attempted. However, the first case bears the
problems of rust and increase of the weight and the second case bears the problems of poor
bending with resin and increase of the weight.

Fig. 5-7 shows a whole view of a tea-house. In a tea-house, the number of pillars must be
minimized since appreciating the scenery around the house is an important issue in the tea
ceremony and the view must not be disturbed by the pillars. However, if the beam is made
of wood only, the span between pillars cannot be expanded. This is an example of using a
CFCC-reinforced wood for the beam. Fig. 5-8 is a magnified photograph of the beam.

Fig. 5-9 Tea-house

Fig. 5-10 Beam

Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 249
1993 Elsevier Science Publishers B.V.

Testing and Applications of

Prestressed Concrete Beams w i t h CFRP
Tendons

T. Katoua and N. Hayashidaa

a
Structural Engineer,Technical Research & Development, Kumagai Gumi Co
.,Ltd. Tsukuba-shi, Ibaragi Pref. Japan

A b s t r a c t
Flexural fatigue characteristics and applications of prestressed
concrete beams with Carbon Fiber Reinforced Plastic (CFRP) tendons
prove promising. Test results and analysis show CFRP to be consistent
with Prestressed Concrete (PC) Standards and good in fatigue. A detail
of an application CFRP in a post-tensioning system outlines the process
and highlights several advantages of CFRP tendons over conventional
steel.

1. I N T R O D U C T I O N
Since CFRP tendons are intended for use in PC beams, subject to
repetitive loading, it is necessary to ascertain their fatigue
characteristics. The results of such testing are presented here along
with the first example of CFRP in a post-tensioning prestressed
concrete application.
Several samples of CFRP tendons in PC beams were tested under static
flexural and repetitive loading. Experimental results corresponded to
values calculated according to PC Standards and indicated good fatigue
properties ; effects of repetitive loading were noted as relatively
insignificant.
Prestressed concrete beams with CFRP tendons were used in a two-story
reinforced concrete application. A review of the procedure revealed some
definite advantages of CFRP over traditional steel reinforcing rods in
construction. For example, the lighter CFRP tendons, for example, proved
to be easier to work with than conventional steel, and their low moduli
of elasticity allowed for better control of prestressing.

2. B E N D I N G F A T I G U E T E S T S
2.1 S p e c i m e n s
A list of specimens is given in Table 1. The specimens consisted of
three for static bending and three for bending fatigue, a total of six
specimens with a variety of CFRP tendons (strand type, rod type) and
250

Table 1
L i s t of specimen of bending fatigue tests
w.
Specimen Variety of
Tend(
tendons Tendions
Arramgement
of Tendons
Variety
of Test
CFRP NO. 1 bar type 2-CFRP 120 straight line
CFRP. NO. 2 strand type 2-CFRP 12.50 straight line static
CFRPP NO. 3 strand type 2-CFRP 12.50 curve test
CFRP NO. 4 bar type 2"( :FRIP120 straight line
CFRP. NO. 5 strand type 2" FRLP12.50 straight line fatigue
CFRPP NO. 6 strand type :FRP12.50 curve test
2-(

Z able 2
haracteristic of materials
Typ Diameter Area Tensile Tensile Elastic Extension
2 Load N Modulus3
Tendon (mm) (mm ) ( KN ) N/mm .
2
CKN/nim )
Strand type
tendon 12.5 76.0 137 1804 137 1.6
Bar type
tendon 13.0 113.1 172 1520 127 1.3

P/2 f P/2

iiuJJUli

Figure 1. Dimensions of specimens

 251

forms of tendon arrangement (straight line, curve) as parameters. The

specimen configurations are shown in Fig. 1 and material characteristics
are given in Table 2.
Specimens had cross sections of 200mm x 250mm and lengths of 2500mm ,
with two cables of CFRP tendons placed in the cross section. Prestress
induction force was 85.3 KN per cable.

2.2Method of t e s t i n g
The method of testing was by two-point loading of simple-beam type
with bearing span of 2200 mm and equal bending moment section of 400
mm. The static flexural loading method is shown in Table 3, and the
repetitive loading method of the bending fatigue tests in Table 4.
Repetitive loading was done with loading level divided into two
stages. At the first stage, static loading was done until occurrence of
cracking, after which repetitive loading of 500,000 cycles was done
with upper limit load the cracking load and lower limit load as 2 tons,
following which, in the second stage, repetitive loading of 550,000
cycles was done with upper limit load 50 to 70% of the failure load
obtained from the results of static flexural loading tests and the
lower limit load 2 tons, for a total of 1,050,000 cycles, after which
one-way static loading was carried out to failure. The excitation
waveform was that of sine waves and the excitation frequency 3 Hz.

Table 3
Loading method ( s t a t i c tests)
Load Stage Control
Method Cycle
NO. 1 step cracking point load 2
NO. 2 step = 3 mm deformation 2

m its
NO. 3 step = 6 mm deformation 2
= 18 mm deformation 2
= 36 mm deformation 2
crushing point deformation 1
NO. 6 step

Table 4
Loading method (fatigue tests)
Specimen Loading Lower Upper Frequ- j c l e of
Stage e repetitive
NO. hfl) [% b .oads
C
T0.4 NO.1 step
NO. 2 step
19.6
19.6
49.0
68.6 3
500, 000
550, 000
CFR
r NO. 1 step 19.6 53.9 500, 000
NO.2 step 19.6 73.5 3 550, 000
NO. 5
CFRP NO. 1 step 19.6 44.1 500, 000
NO. 6 NO. 2 step 19.6 58.8 3 550, 000
252

Photo. 1. A view of bending fatigue test

Measurement were obtained at the designated number of cycles,

suspending excitation at the specified number of cycles on carrying out
static loading. The items of measurement were displacement at beam
middle, strains of tendons and concrete, and condition of cracking. A
view of bending fatigue test is shown in photo. 1.

3. T E S T RESULTS
3. l U l t i m a t e f lexural loads
The flexural failure loads of specimensCCFRP NO.4, CFRP NO.5, CFRP

Table 5
Results of bending static and fatigue tests
Specimen Ultimate Flexural Load ( KN)
Variety Failure Type
NO. of Test Experimental Analytical
Value Value <D/@ of Beam
CFRP NO. 1 142.7 139.7 1.02
static rupture of
CFRP NO.2 120.4 123,4 0.98
test CFRP tendons
CFRP NO. 3 86.7 123.4 0.71
CFRP NO. 4 137.6 139.7 0.99
fatigue rupture of
CFRP NO. 5 121.5 123.4 0.98
test CFRP tendons
CFRP NO. 6 81.8 123.4 0.67
 253

NO.6) subjected to repetitive loading compared with results of static

bending testsCCFRP NO.1, CFRP NO.2, CFRP NO.3) are as shown in Table 5,
where values of approximately the same degree are indicated for each
specimen, and reductions in maximum load due to repetitive loading were
not seen. Maximum loads were determined for all specimens by rupturing
of CFRP tendons. On comparison of experimental values and calculated
values according to the PC Standards, they were more or less the same.
3.2 C o n d i t i o n of c r a c k i n g
The condition of cracking is shown in Fig.2. On comparison of the
condition of crack occurrence by static bending and flexural fatigue
results, crack widths and numbers of cracks were roughly the same for
each specimen. Cracks that followed the shapes in which tendons were
arrayed had been formed immediately after failure of PC beams and may
be considered to have occurred due to the effect of the rupture of CFRP
tendons and subsequent release of the prestressing loads.

CFRP
fJX
N0.1 tf A A 7 ? frrV
if.lB-
CFRP
N0.2

CFRP
NO.3
.*j
/ Jf
JT
-\ 7TW
static Bending tests

JL
&
CFRP
NO.4

X^ft&fr-
CFRP
N0.5

CFRP
NO. 6 W
Bending fatigue tests

Figure 2. Final state of cracking

3. 3 L o a d - d i s p l a c e m e n t curve
The load-displacement curve of the PC beam subjected to repetitive
flexural loading and static bending test results are shown in Fig.3.
In repetitive loading at the first and second loading stage, each
specimen (CFRP N0.4, CFRP NO.5, CFRP NO. 6 ) did not show changes in
curve gradients, and hysteresis curves were roughly the same as static
bending test results. However, as the number of cycles of loading
increased, an increasing trend was seen in amount of displacement at
action of upperlimit load and in residual displacement.
254

0 4 8 12 16 20 0 4 8 12 16 20 0 4 8 12 16 20
d e f o r m a t i o n (mm) d e f o r m a t i o n (mm) d e f o r m a t i o n (ram)
CFRP- NO. 1 CFRP- NO.2 CFRP- N O . 3
(1) static Bending tests

0 4 8 12 16 20 0 4 8 12 16 20 ^ 0 4 8 12 16 20
deformation (mm) d e f o r m a t i o n (mm) d e f o r m a t i o n (mm)
CFRP- NO. 4 CFRP- NO. 5 CFRP- N O . 6
(2) Bending fatigue tests

Figure 3 . Load - deformation curve

3.4 C h a n g e in e q u i v a l e n t r i g i d i t y
Defining the gradient (load/displacement) of the s t r a i g h t line
connecting any point on the plane at various loads and the origin of
the load-displacement curve as equivalent rigidity, the relationship of
equivalent rigidity and displacement is shown in Fig. 4.
The equivalent rigidities of specimens subjected to repetitive loading
and static bending specimens indicated more or less the same changes
with all specimens. Influences of repetitive loading were not seen.
Further more, changes in equivalent rigidities were constant up to
occurrence of cracking, but showed sudden declines in the range of
displacement of 2 to 5 mm, after which the trend was a gentle decline
with increases in displacement up to ultimate failure.

3.5 C o n c l u s i o n s
The conclusions below were drawn regarding flexural fatigue properties
of PC beams using FRP tendons.
The cracking load and ultimate flexural loads can be roughly
estimated using the PC Standards equations.
The maximum flexural strength, load-displacement curve, and variation
in equivalent rigidity of a PC beam subjected to repetitive flexural
loading in excess of one million cycles are roughly the same as results
of static bending, with effects of repetitive loading not recognizable.
It may be ascertained that fatigue properties are good.
 255

i 1 1 r- ^ b 0 r ^ 50 ! 1 1
N O s t a t i c Bending
N O s t a t i c Bending
O s t a t i c Bending 40 - test (NO. 2) < test (NO. 3)
test (NO. 1) 40
ABending fatigue fau ABending fatigue tad ABending fatigue
test (NO. 4) . - test (NO. 5) >,30 test (NO. 6)
*** nrt %

OO 20
~A
~
oo
20

~ 10
V-.
^k u-

C3
a> V^_ ^*te*w*~
^^
- 10
ca>
0 Li 1 1 1 1 I 1 1 I 1
5 o
> 0 10 20 30 40 0 5 10 15 20 25 30 35 40 45 50 0 10 20 30 40 50
g" deformation (mm) deformation (mm) deformation (mm)
C F R P NO. 1, NO. C F R P NO. 2 , NO. 5 C F R P NO. 3 , NO. 6

Figure 4. Relationship of equivalent rigidity and displacement

4. O U T L I N E OF BUILDING
The building in which CFRP tendons were applied was a two-story
reinforced concrete structure of plane dimensions 21 x 5 m. Because
there was an opening of 15 x 5 m at the first story , girders of 16-m
span were installed at the second story and roof. Since this building
was to be finished with plaster and cracks would greatly affect
aesthetic appearance , prestress was induced using CFRP tendons with the
aim of preventing cracks due to long-term deflection and drying
shrinkage.
A general view of the building is shown in photo. 2 and an outline of
it in Fig. 5. The building was completed in September 1989.

Photo. 2. A general view of the structure

256

fe
RF-Girder irf*
rr

RAH
i^

,
1 1L

77777777777777
10,625

Figure 5. A plane of the structure Figure 6. FEM analysis model

5. O U T L I N E O F D E SIG N
The girders were designed to be reinforced concrete for design load ,
but were made PRC ( prestressed reinforced concrete ) in order that
cracking would not occur under condition of normal use.
5. I D e s i g n of P R C G i r d e r
In design of PRC girders, FEM analyses were made and prestressing
forces to be used were selected so that tensile stresses produced in the
girders under design load would be less than the allowable tensile
stress intensity of the concrete. The model for FEM analysis was made a
two-dimensional continuous plane element with the second-story girders,
the roof girders, columns, and walls integrated. The analysis model is
shown in Fig. 6 and the results of analyses are given in Table 6.
According to the analytical results , the elements would each be below
allowable tensile stress intensity of concrete (Ft=0. 07 x Fc = 1.64 N/
mm2 ,Fc = specified concrete strength ) with prestresses of 1729 kn
induced in second-story girders and 493.9 KN induced in roof girders,
respectively.
The tendons used were CFRP three-strand multicable ( 3-12.5 0 ,
rupture load Pu= 412 KN, prestressing force 0.6 Pu= 247 KN ). there were
7 cables for each the second-story girders, and 2 cables for each of
roof girders, a total of 18 cables.
The tendons type and prestressing forces induced are given in Table 7,
the CFRP tendon arrays are shown in Fig. 7, and the girder cross
section list in Fig. 8. The outline and standard characteristics of the
CFRP tendons used in this case are given in Table 8 and 9.
 257

Table 6
Results of FEM analysis
Load n CASE 1 CASE 2 CASE 3 Displacement(mm)
Condition Dead Load Prestress- CASE 1 or
+Live Load ing Load + CASE 2 Stress(N/mm2)
Center of vertical
2F-Girder - 3.39 -0.47 - 3.80 displacement(mm)
Top of horizontal
RF-Column 1.06 0.75 1.81 displacement(mm)
Upper edge of
ends of 2F-Girder 2.04 -1.96 0.08
Lower edge of
ends of 2F-Girder -3.83 -1.50 -5.33 : Stress 2
N/mm ;
Upper edge of
center of 2F-Girder -3.54 -1.92 -5.46 + : compression
Lower edge of - : tension
center of 2F-Girder 2.41 -1.44 0.97
Center of edge of
RF-Girder -0.25 -0.73 -0.98
Center of middle
of RF-Girder -2.08 -0.73 -2.80

able 7
FRP tendon type and prestressing forces
Girder CFRP tendon type
ntnrrm Sectionalstress
(N/mm2)
Strip A of c
2F-Girder 7 -3 CFRP 12.5 0 1729 1.96
Strip B of c
2F-Girder 7 -3 CFRP 12.50 1729 1.96
Strip A of
RF-Girder 2 c
-3 CFRP 12.50 494 1.03
Strip B of
RF-Girder 2 c
-3 CFRP 12.50 494 1.03
258

2 e -3CFRP

( D <S>

Figure 7. CFRP tendon arrays

ends center
p?i **!

8 i^l ends center

hr hfe-fl
3sr s
|
--- ! s
"H
-fL -.
-~l
^J JJ
LOAJ
- 1
)0 od
hana\
,200,2001

(1) a section of 2F-Girder (2) a section of RF-Girder

Figure 8. A list of beams cross section

Table 8
T h e outline of CFRP tendons
Type of CFRP tendons
Length of CFRP tendons
three-strand multicable
( 3-CFRP 12.5 0 )
20, 950 mm
per one cable

Number of cables g|*r

..rder
14 cable
4 cable
total 18 cable
 259

able 9
tandard characteristics of CFRP tendons

a>tio
Rupture Elastic Extension Specific Relaxation Cofficient of
strength
f
(N/mm )
modulus
(KN/W)
gravity
a>
, expansion N
( 0-* / c )
0.65
1764 137.2 1.6 2.10 2.503 I] 0.6
1) 20 C, 0.8Pu, 10 hour ,2) 20C^ 60C^ 20 "C, 0. 8Pu. 16 hour

Table 10
Results of compressive strength tests of concrete
one week four weeks strengths at time
strengths of prestressing
Jvmnr (N/mm2) (N/mm2)
2F-Girder 19.8 28.6 25.9
2F-Girder 20.2 27.7 23.9
RF-Girder 23.7 33.6 23.7

6. P R E S T R E S S I N G WORK
The prestress application was done twice, for the second story and for
the roof. The timing of application was after placing concrete, field-
curing test pieces, performing compressive strength tests, and
confirming that the specified strength of 23.5 N/mm2 had been attained.
Injection of grout was done immediately after each prestressing
operation , and the grout was of cement base , the same as
conventionally used. The results of compressive strength tests are
given in Table 10.
6.1 O u t 1 i n e of anchorage zone
The anchorage device used employed a combination of a grout system
,where grout is injected between tendon and sleeve and compressive
force is applied from outside the sleeve after hardening of the grout,
and a compression system ( photo. 3). At the anchorage zone ,
arrangement of sheaths and other objects was done fixing anchoring
hardware for performed holes in columns at the two ends to serve as
anchorage zones to accommodate the anchorage devices in the concrete
frame. Details of a tendon anchorage zone are shown in Fig. 9. Views of
fixing of forms and anchorage zone , arrangement of sheathing, and
arrangement of CFRP tendons are shown in photo. 4 to 8.
260

Photo. 3. Anchorage of three multi type

nut ( D=98.1mm, H=45mm)

270 | 20,710 270

21,250 '

Figure 9. Details of a tendon anchorage zone

Placing of CFRP tendons in sheaths when performed during reinforcement

work or formwork construction may possibly cause damage to the CFRP
tendons , and therefore , when placing concrete, protective pipes were
inserted into the sheaths, and these protective pipes were removed
immediately before applying prestress. It was possible for placing of
CFRP tendons to be done manually since the tendons were light compared
with conventional prestressing steel.
 261

Photo. 4. View of fixing forms Photo. 5. Anchorage zone

Photo. 6. Arrangement sheaths Photo. 7. View of CFRP tendons

262

Photo. 8 .Arrangement of CFRP tendons Photo. 9 .A view of applying prestress

Figure 10. A method of applying prestress

6.2 A p p 1 i c a t i o n of p r e s t r e s s
Application of prestress was done by adjusting elongation by setting
jacks at both ends since elongation of CFRP tendons would be great
because the PRC girders were as long as 21 m. The method of applying
prestress was that of attaching special couplers to anchorage device
sleeves and employing tension rods and 686 KN center hole jacks. The
method of applying prestress in shown in Fig. 10, and a view of
prestress being applied in shown in Photo. 9.
6.3 C o n t r o l of p r e s t r e s s i n g
Control of prestressing was done by control of elongation consisting
of measuring loads of prestressing and elongations of tendons. The
theoretical elongations of the various CFRP tendons were calculated
considering friction losses due to tendon arrangement configurations and
length changes, and control of prestressing was carried out with b%
 263

of the calculated elongations as the allowable range of control.

The results of group control of prestressing are shown in Fig. 11. The
results of control of prestressing indicated that all tendons were
within allowable limits of control. The elongations measured were
scattered in a range of 163 mm to 167 mm , and this is thought to have
been influenced by precision of sheath placement and measurement
errors.

1 2 3 4 5 6

nuiuuei 5 CM
1 1 1 1
tO
1 T 1
CM
1 ? CMT in
1
tO
1 1 T 1 T 1
of cable CM CM Oi, ct
1
< <1 < <
CM
1 1 1 1 1
CM CM 1 7 1 1 1 1
CM
1 1
< <
1 1 1
< < <
03 CQ CQ CQ CQ CQ CQ CQ CQ

Al(mm) tO
o o
5 ID
CO
to
8 to
CO
CO tO to
CO to
to
to
to 8
to s
T(mm) 163.8 163.3 165.5 165.0 167.0 165.0

Set loss(mm) 2 3 2 3 3 2 3 2 1 1 1 2 2 2 1 1 1 1

g UC1
J^
e

/
\A -^ ,' \
1
2. <u
"
/ *
y
(%) o i V
o
\ / ' V j V
<u
I \/ \ J
-
C
o
L. V
LCL
<_>
+5
<u

... UCL
o
o ^ ^ k
^ - -
co-o (%) 0
. _ - _
^-*^
o
=3 C3
CL-
5

Figure 11. Results of group control of prestressing

RP-Girder

G
-69
-54
(unit : x 10 ")
L/-36 I 2F-Girder

(-):compression strain

2,5001 5,000

Figure 12. T h e strain distribution of beam at Strip

A of second floor in prestressing
264

Table 11
Comparison with stress of design and measurement in prestressing
position ends of strip center ends of strip
measured design measured design measured design
Girder value value value value value value
2F upper edge - 1.89 -1.96 - 1.94 -1.92 - 2.03 -1.96
Girder lower edge - 1.48 -1.50 - 1.06 -1.44 - 1.50 -1.50
RF
Girder middle - 0.24 -0.73 - 0.17 -0.73 - 0.25 -0.73
2
stress unit : N/mm
notes C - ) : compressiye stress of concrete, 02
2F-Girder of elastic modulus :E= 27.6 KN/mm2
RF-Girder of elastic modulus :E= 25.7 KN/mm

U Strain m e a s u r e m e n t s of P R C g i r d e r s
Concrete gauges were embedded in PRC girders at the second story and
roof in order to grasp the stressed stages of various parts of the
concrete frame accompanying transfer of prestress, and strains at
transfer of prestress were measured. The strain distribution of the
girders at Strip A of second floor is shown in Fig. 12.
With regard to stresses of girders at time of prestressing, measured
strains converted to stresses were compared with design values, and the
results are given in Table 11. Measured and design values were in rough
agreement at the two ends of second-floor girders, but at midspans
measured values were slightly smaller than design values. This is
thought to have been because deformation was restrained due to the
influences of the slab and restraining walls at the two ends, but
nevertheless, it was confirmed that the amount of prestress as designed
had been about secured.

7. C O N C L U S I O N S
CFRP tendons were used with the post-tensioning system for the first
time in a prestressed concrete structure, and it was ascertained by the
results of measurements made at the time of applying prestress that
prestressing force as designed had been induced. Furthermore, the
following findings were obtained regarding construction execution
methods for CFR tendons.
(1) Since CFRP tendons are light weight compared with conventional steel
prestressing tendons, placing operations can be amply carried out
manually without using equipment such as cranes.
(2) Since CFRP tendons have surfaces which are easily damaged it will be
necessary for protection of surfaces to be provided when transporting
and placing.
 265

(3) Since CFRP tendons have low moduli of e l a s t i c i t y and elongations are
l a r g e , c o n t r o l of p r e s t r e s s i n g can be r e a d i l y provided, w h i l e
fluctuations in load after prestressing are small.
(4) Cement-based material can be used for grout as for conventional
post-tensioning.

REFERENCES

1 Architectual Institute of Japan, Standards for Design and Construction

of Prestressed Concrete , with Commentary, 1987.
2 Architectural Institute of Japan, Recommended Practice for Structural
Design and Construction of prestressed Reinforced ConcreteCTypel PC)
, with Commentary, 1986
3 Kato, Ishibasi, and Kawaguchi : Development of CFRP Tendon Anchoring
d e v i c e and PC Girder Flexural loading Experiment, P r e s t r e s s e d
Concrete Engineering Association, Sept. 1988
4 Maruyama, Ito, et al. :Flexural Fatigue Characteristics of PC Beams
Using CFRP and AFRP Rods, Concrete Research and Technology, Vol.12,
No. 1, 1990.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 267
1993 Elsevier Science Publishers B.V.

Technora, an Aramid FRP Rod

Kunitomo Noritake*
Ryuichi Kakihara**
Shin'ichiro Kumagai*
Jun Mizutani*
* Sumitomo Construction Co. Ltd.
13-4, Araki-cho, Shinjuku-ku, Tokyo 160, Japan
* *Teij in Corporation
1-6-7, Minamihon-machi, Chuo-ku, Osaka 541, Japan
(Abstract)
The new FRP materials have mainly been used up to now in the aeronautical and motor
vehicle industries. There has recently been a movement to try to adapt these materials to the
field of construction, and research on this topic has been actively pursued in several countries.
Sumitomo Construction Co. Ltd. and Teijin Corporation, noting the high tensile strength of
Aramid fibres, have conducted research aimed at developing Aramid FRP rods as a substitute
for existing types of high tensile strength steels. In addition to possessing high tensile strength,
Aramid FRP rods have superior durability, and do not deteriorate even in harshly corrosive
environments. The present paper examines the physico-mechanical and strength characteristics
of Aramid FRP rods, and also discusses the conditions for their use as tendons in prestressed
concrete. Examples are also given of practical applications of Aramid FRP rods in practical
applications.

1. MATERIALS
Aramid (Technora) FRP consists of Technora fibres strengthened with vinylester resin
and formed into rods (Technora rod'). This is a material with high tenacity and durability,
suitable for use as tensioning material in prestressed concrete.
Technora fibre is made from PPODTA (co-poly-paraphenylene/3,4-oxydiphenylen
terephthalamide), of which the chemical structure is shown below. i)

4HN -Q-NHOC-Q-CO)- -CHN - Q - O S ^ N H O C - O - C O ) -

m n
Compared with other high-tenacity fibres, Technora fibre presents excellent fatigue
resistance to abrasion, flexure and stretching. It also has a breaking strength which is one of
the highest of all commercially available fibres (8 times the strength of steel, and 3 times that
of glass or polyester of the same weight.)
Because of its high thermal stability, Technora fibre attains its full tensile strength after
10 hours in hot air at 250C (482F). Technora fibre exhibits high resistance to acids, bases,
oxidising and reducing agents, as well as organic solvents. The physical properties of Technora
fibre are listed in Table 1.
268

Table 1 Physical Properties of Technora Fibre

Tensile strength (GPa) 3.5

Tensile modulus (GPa) 74
Elongation (%) 4.6
Density (g/cm ) 1.39
Diameter () 12
Thermal decomposing temperature (C) 500

Vinylester was selected as the matrix material because, in addition to enhancing the
elongation properties of Technora fibre, it also has superior moldability, adhesion and
resistance to corrosion. Its physical properties are shown below.

Table 2 Physical Properties of Vinylester Resin

Tensile strength (GPa) 0.09

Tensile modulus (GPa) 3.40
Compressive strength (GPa) 0.16
(Upper yield point)
Poisson's ratio 0.373

2. MANUFACTURE
Technora rods are manufactured from PPODTA fibre and vinylester resin by a pultrusion
process as shown in figure 1. A straight bundle of aramid fibres is impregnated with a resin
and passed through a die to squeeze the resin. An identical fibre is wound spirally around the
bundle of straight fibres to provide roughness. The resin is then cured in an oven. Vinylester
resin is employed because of its excellent moldability, adhesive properties and resistance to
chemicals. The volume ratio of the fibre and the resin was kept to 2:1.

Pultrusion Method

I1
t
^ ^

T T T
t
Fibres Resin impregnation Winding Heat Curing Take-up Product

Figure 1. Manufacture of Technora rods

Technora rods were developed as prestressed concrete tendons, and as such must have a
high bond performance with grout or concrete to transmit the stress. In order to improve the
bonding performance, a spiral wound rod as shown in Figure 2 was developed. As a result,
Technora spiral wound rods are now open to practical application.
 269

Winding Fibre

Round Securing Fibre

o
o
VO H
%
00
V>

-I

Figure 2. Technora spiral wound rod (<j)6mm)

As shown in Photo 1, a range of different types of rod is manufactured by the pultrusion

method. A twisted strand using Technora fibre is currently being developed.

Table 3 Type and Size of AFRP rods

Type Mark Nominal Nominal Cross Unit Specified
Diameter Sectional Area Weight Tensile Strength
(mm) (mm2) (g/m) (kN)
Round rod ARPR1 3.0 7.07 9 13.0
4.0 12.56 16 22.7
6.0 28.26 37 49.9
8.0 50.24 66 86.1
Spiral wound ARPD1 3.0 7.07 10 13.0
rod 4.0 12.56 18 22.7
6.0 28.26 41 49.9
8.0 50.24 73 86.1

6mm dia. Round Rod

6mm dia. Spiral Wound Rod
8mm dia. Round Rod
8mm dia. Spiral Wound Rod
Rectangular Rod

12.5mm dia. Twisted Strand

(under development)

Photo 1 Technora rods

270

3. PROPERTIES

3.1 Physico-mechanical Properties

The physico-mechanical properties of Technora rods are influenced by the fibre volume
fraction (Vf) and the matrix resin. Technora fibres have an extremely high tensile strength of
3.4 GPa, and the theoretical strength of Technora rods is obtained from the following
formula, when Vf = 65%:
3.4 GPa x 0.65 = 2.21 GPa

However, when the rods are formed using multiple bundles of fibre impregnated with
resin, the tensile strength is generally slightly less than the theoretical strength. Also, the
tensile strength of the rod changes depending on the anchorage system used. A <|>6mm rod,
when used with a wedge anchorage, has an average tensile strength of 2.14 GPa, while with a
bond anchorage, the tensile strength is 1.88 GPa. The physico-mechanical properties of
Technora rods are shown in Table 4, below.

Table 4 The Physico-mechanical properties of Technora rods

Matrix resin Vinylester

Volume of fibre (%) 65
Density (g/cm3) 1.3
Tensile strength (GPa) 1.9 (bond anchorage)
Tensile modulus (GPa) 54
Elongation (%) 3.7
Relaxation (%) 7-14
Poisson's ratio 0.35 (axial)
Coefficient of thermal exp. (/C) -3 X 10"6 (axial)
Electrical resistance (/cm) 5 X 1014

The load-strain relationship of <(>6mm Technora rods as shown in figure 3 is straight up

to the breaking point. The tensile modulus, 54 GPa, is one-fourth that of ordinary steel
tendons, and the specific density, 1.30, is one-sixth that of steel tendons.

60
i i i ! :
55
, .;..i....;..4 ! i / i i
50
1 v 1 i I
45 : : : i A S j |
40 ;-...{4--~
Load 35 Break
(kN) 30
: : : 7\ I ! I I Elongation
25 ! I 1i I 3.7%

20
I \ \ \ \
15
i \ i.IiiiIJ
10
5 \A I I i i i i
! ! j ; ; j

Rod Elongation (%)

Figure 3. Load-Strain Curve
 271

3.2 Tensile Strength

The most important property when Technora rods are used as tensioning material is the
tensile strength. The data shown below on tensile strength was obtained experimentally as
follows. A 160mm long section at each end of a rod was coated with partially saturated
polyester resin, forming a round rod of 8mm diameter, which was held in a wedge anchorage.
The length of the test sample was 600mm, and the tensioning speed was 20mm/min. The data
obtained from sampling a 28,000m rod at 50m intervals has been collected in statistical form.
This testing showed the average tensile strength of the rods to be 2.14 GPa, or 96% of the
tensile strength of Technora fibre. Test results are shown in table 5, and the frequency
distribution curve obtained from these is shown in figure 4. 2)

Table 5 Tensile properties of Technora rods (<|>6mm)

Property Average Value Standard Deviation No. of samples

Tensile Strength (kN) 60.5 1.57 560
(GPa) 2.14 0.056
Tensile modulus (GPa) 53.9 0.69 50
Elongation at break (%) 3.8 0.15 30

26 52 78 104 130
Tensile 1 1
55.8
strength (kN) -3 No. of samples: 560
56.5
-2 57.3 Average tensile strength: 60.5
58.1 Standard deviation: 1.57
- 58.9
59.7
X 60.5
61.3
+ 62.1
62.8
+2 63.6
64.4
+3 65.2

Cumulative ratio (%)

Figure 4. Frequency distribution of test results

Photo 2 is a microscopic photograph of a rod after breaking during the test described
above. The type of breaking shown here is characteristic of Aramid fibres, and is known as
'fibril breaking'; this occurs when the proper tensile strength is achieved.
In order to investigate the relationship between tensile strength and tensioning speed, a
tensioning test was performed on specimens 1000mm long, varying the tensioning speed
within the range 5-50mm/min. Figure 5 shows the results of this test. A slight tendency was
observed for tensile strength to increase as the speed of tensioning is increased. Technora
fibres themselves have a tendency to increased breaking strength as the tensioning speed is
increased, and the results of this test are thought to reflect this tendency.
272

Photo 2 The state of the fibres after breaking of the rod

Rod strength
(GPa) 2.40

2.20 L

2.00 |_

1.80
1 2 3
Tensioning Speed (%/min)

Figure 5. The relationship between tensile strength and tensioning speed

In order to investigate the relationship between tensile strength and length of sample,
tensioning tests were performed on samples of varying lengths between 100 and 4000 mm.
Tensioning speed in this case was 3.3%/min. As figure 6 illustrates, the length of sample was
found to have no effect on the tensile strength. In the case of FRP materials, there is an
increased probability of defects occurring along the length of longer samples, and these
materials thus exhibit a tendency for the tensile strength to fall off with longer samples.
Technora fibres have a tendency not to show any decrease in strength with longer samples,
and this tendency is repeated in rods using these fibres.
 273

Rod strength 2.40 |-

(GPa)

2.20 \-

2.00 U

1.80
1 2 3

Length of sample (m)

Figure 6. The relationship between tensile strength and sample length

The dependency of tensile strength on temperature was then investigated, by performing

tensioning tests at a range of temperatures from -20C to +60C. Sample length was 600mm,
and tensioning speed 20mm/min (stretch speed 3.3%/min). The results of this test are shown
in figure 7. Tensile strength showed a tendency to fall off as the temperature rose, dropping
around 8% at 60C from the value observed at 20C. This result shows a close correspondence
with the temperature dependency characteristics of Technora fibres, and suggests that the
matrix material employed is not a factor in the reduced strength observed.

Rod strength
(GPa) 2.40

2.20 r-

-20 0 20
Sample temperature (C)

Figure 7. Temperature dependency of tensile strength

274

3.3 Relaxation
The question of relaxation of the tensioning material is an important consideration in
prestressed concrete. FRP materials are said in general to have a greater relaxation than steel
tendons. Figure 8 shows the results of relaxation tests performed on Technora rods. These
results show that, under the stresses experienced in practical applications, a relaxation of
approximately 14% can be expected after 30 years. A relaxation of several percentage points
is observed during the first few minutes after tensioning a Technora rod. Several methods
have been proposed to eliminate this initial relaxation, including the brief application of
tensioning slightly greater than the anchorage load, anchoring the tendons after maintaining
the anchorage load for a few minutes. 4)

o
5
Relaxation
(%) -0.80P max
-0.60P max
10 ^0.*40P max

15
1 10 10' 103 104 10s
Time (minutes)
Figure 8. The Relaxation of Technora Rods

3.4 Fatigue
Figure 9 shows the results of fatigue testing of Technora rods. Fatigue testing was
performed varying the lower limit of stress intensity ( a j , and the amplitude. The results of
this testing reveal clearly that the fatigue characteristics of Technora rods are superior to
those of steel.
1.2

H J A A A / , I Amplitude a,
1.0
Number of loading
Amplitude
(GPa) o.8
OL =0.18GPa

0.6

0.4

0.2 -p-0 p=**

<7L =1.4GPa (7 L =1.2GPa
0 i i
r
10 10 ' 10
Number of Loadings
Figure 9. The fatigue characteristics of Technora rods
 275

The relationship between the tension load and the static fatigue breaking time is shown in
figure 10. With static fatigue breaking time plotted on a logarithmic scale, the graph is a
straight line, with the tension load after 100 years estimated at 1.23 GPa. This value corresponds
to around 57% of the average strength (2.14 GPa) of the rods.

2.20
200
1.80
Tension 1 6 0
Load 1.40
(GPa) 1.20 (100 years)
1.00
0.80
060
0.40
10"2 10"1 1 101 10" 10" 10" 10'
Static Fatigue Breaking Time (hrs)

Figure 10. The relationship between tension load and static fatigue breaking time in Technora
rods

3.5 Durability
One of the characteristics of Technora rods is that they are unaffected by a wide range of
chemical actions. Rods immersed untensioned in water at 60C, sea water and a pH 13
alkaline solution for 400 days were then tested for changes in tensile strength and weight, and
tensioned rods immersed in an alkaline solution for 200 days were tested for changes in
tensile strength. The results of these tests are shown in table 6. None of the conditions
described caused a decrease in the tensile strength of the rods, and neither is any loss of
strength observed in the tensioned rods immersed in alkaline solution. These results show
that Technora fibres, the vinylester resin used as matrix material, and the interface between

Table 6 Chemical resistance of Technora rods to a variety of chemical environments

Immersion Tension Immersion period (Days)

Liquid Force (GPa) 40 90 200 400

Fresh water 0 92 102 100 100

(+0.6) (+0.9) (+1.0) (+1.4)
Sea water 0 100 103 100 100
(+0.3) (+0.7) (+0.4) (+1.5)
Alkali (pH13) 0 100 98 100 98
NaOH 2g/l (+0.6) (+0.8) (+1.1) (+1.4)
Ca(OH) 2 2g/l
1.10 100 99 101

Upper row of figures: % tensile strength retained;

Figures in parentheses: % weight change
276

the two, are all adequately resistant to alkaline hydrolysis. Measurements of the weight of the
rods are used chiefly as a measure of the solidity of the interface, and in this case also, almost
no changes were observed in the rod weight. In addition to the high chemical resistance of
the fibres themselves in Technora rods, the resin selected also has a high corrosion resistance
to match the aramid fibres.
The weathering properties of Technora rods were investigated by exposing rods in a
well-lit, south-facing position at a 45 angle for 360 days, after which the tensile strength,
surface hardness, and flexural strength were measured. The results of these tests are shown in
figure 11. The tensile strength of the rods exhibits a tendency to fall off as the number of
days of exposure increases, but the drop in strength tends to level out, with the tensile
strength after 360 days dropping to around 10% of the average strength of 2.14GPa, or
1.93GPa. Aramid fibres have a tendency to deteriorate in ultra-violet light, but since the
effects of this deterioration are limited to the fibres located on the surface of the rods, no
further decrease in tensile strength would be expected as a result of longer exposure. The
surface hardness of the rods (JIS K7060) is a measure of the stability (durability) of the resin,
but no deterioration in surface hardness was noted as a result of the exposure. The triaxial
bending strength (JIS K6911) of the rods is a measure of the overall durability of the
composite material, including the reinforcing fibres, resin and the boundary layer between
these, but this measure also showed no deterioration as a result of exposure. 3)

100

Tensile 80
Strength Exposure conditions: Rooftop, Osaka;
60 South-facing, angle of 45

110 [
Surface (JIS K7060)
Hardness ioo

90

110
Flexural Triaxial bending strength (JIS K6911)
100
Strength
90
(%)
100 200 300 400
No. of days of exposure

Figure 11. Weatherability of Technora rods

 277

4. PERFORMANCE AS REINFORCEMENT
Technora rods were developed chiefly as a tensioning material for prestressed concrete.
One important consideration when using FRP materials as prestressed concrete tendons is the
anchoring. Also discussed here are various matters related to the practical application of the
rods, such as the bond performance of the rods, the frictional coefficient, and their structural
properties when used as pre- and post-tensioning members.
4.1 Anchorage
It is important when employing FRP materials as tensioning materials to develop anchorages
suited to each material. As illustrated in the photographs below, wedge-type anchorages and
bond-type anchorages have been developed as anchorages for Technora rods. These anchorages
have been developed for use with the spiral wound rods previously referred to, and are
constructed so that the rod is inserted into a housing and then fixed with injected mortar,
anchoring being achieved through the bonding of the rod and the mortar. A screw thread is
cut into the outer surface of the housing, and the anchoring is secured with a nut. A range of
bond anchorages has been developed, to accommodate any number of rods from 1 to 19.
Bond anchorages are also being manufactured from FRP, to provide non-corroding anchorages.
The guaranteed tensile strength of tendons in practical application has to be the tensile
strength of rods used in conjunction with anchorages. The tensile strength when used with
bond anchorages is in general lower than that when the rods are used with wedge-type
anchorages; the guaranteed tensile strengths obtained from testing of tendons used with bond
anchorages are shown in table 7.

Multi-wedge Type Anchorage Single-wedge Type Anchorage

19<|>6mm Bond Anchorage 7<|)6mm Bond Anchorage

Photo 3 Various types of anchorage

278

Table 7 Types of Bond Anchorage

T
YPe Tensile Strength Cross-sectional Area (cm)
6 51 0.283
8 90 0.503
36 153 0.848
46 204 1.131
76 357 1.979
126 612 3.393
196 800 5.372

4.2 Bond Performance

Prestressed concrete tendons must have a high bond performance with grout or concrete
to transmit the stress. In order to improve the bonding performance, a spiral wound rod as
shown in Fig. 2 was developed. As a result, Technora spiral wound rods can be used
practically. In order to examine the bonding performance of Technora spiral wound rods, a
pull-out test was carried out and similar tests of the existing prestressed concrete tendons
were carried out for comparison. Table 8 shows the bond stress of each type of tendon. 5)

Table 8 Bond Stress of Prestressed Concrete Tendons

Tendon Steel AFRP AFRP rod AFRP rod

strand single rod (bundled) (dispersed)
Maximum average
bond stress (N/mm2) 5.8 15.5 16.5 15.6
Shape
O
cftn
. 12 ,
Specimen length: (300m) (150mm) (150mm) (150mm)
Surface area
(300mm length) 15400 mm2 2800mm2 7100mm2 8500mm2
Average bond stress = Tensile load (N/mm2)
Surface area

4.3 Pre-tension Beam Test

The following tests were carried out to determine whether Technora rods are suitable as
tendons for pre-tensioned concrete beams:
1) Bonding Length Test
2) Bending Fatigue Test
3) Static Loading Test
The beams used in the tests were 2000mm long, 200mm high and 100mm wide. Six spiral
wound rods of 6mm diameter were used as tendons in the beams, two rows of three rods,
each located in the upper and lower sections of the beams.
 279

Two different types of beam were tested, as follows:

Series (A): Using a prestressing force of 157 kN
Series (B): Using a prestressing force of 225 kN

2000

20@90 = 1800 2@50 = 100

Stirrup 6

AFRP Spiral Wound Rod

100

Figure 12. Illustration of Test Beam

The distribution of the prestressing force is observed by the measurement of concrete

surface strain after pre-tensioning. Strain gauges were set on the test beams of series (A) and
(B), at 100mm intervals. The bond strength of spiral wound rods was obtained by measuring
the strain of the concrete at the time of prestressing. As the compressive strains in the
concrete were constant at points 250mm~350mm from both edges of the test beams in both
series (A) and (B), the bond length of Technora rods was determined to be 250mm~350mm.
The strain of the concrete caused by prestressing is shown in Figure 13.

Strain of
Concrete

0 1
Specimen Length (m)
Figure 13. Strain Distribution of Pre-tensioned Specimen
Bending fatigue tests were carried out on both series (A) and (B) test beams. Based on the
assumption that the effective prestressing be 80% of the initial prestressing, the compressive
stress sustained on the test beam is 6.4Mpa for series (A), and 8.8Mpa for series (B). In this
bending fatigue test, the upper load was set so that the compression stress at the bottom edge
fibre of the specimen becomes zero. The lower load was set at 25% of the upper load.
Fatigue in the beams was tested over four million load cycles. The test results are shown in
280

Figure 14. There was no change in the observed deflection of the test beams after the
application of four million cyclic loadings.

0.8
(Applied Prestressing Force 225 kN)
7
Deflection at [ o o o 0 o o o o
max. loading
a
(mm) (Applied Prestressing Force 157 kN) ffheoreticid DeTTection)

0.5

0.4 * (Theoretical Deflection)

1

0.3
i

0.2
i

0.1
1

0 4.0
4 X10
6
1.0 2.0 3.0

Figure 14. Results of the fatigue test

The static loading tests were carried out on the series (A) test beam after the fatigue test
loadings, and on a normal test beam in order to observe the change in the effective prestressing
force after prestressing. The beams were broken due to compressive failure of the concrete.
The ultimate bending moment was 1.1 times the calculated value. The loading-deflection
curve is shown in Figure 15. The measured value and the calculative value match closely,
suggesting that the ordinary design methods for prestressed concrete girders using steel
tendons are also applicable to the design of pre-tensioned girders using Technora rods.

r~
100 31.25 (X103)
A-l Mca.ured Value
*/i- I M easured Val JC
80 25.00
1 ^K

-
* "*'
m

,^^*
60 '. 18.75 Bending
Loading ^^^ -*-*V
*"
CaJculativeval ue
Moment
(kN)
40 12.50
(kN.m)

20 6.25

0 2 4 6 8 10 12 14 16 18 20
Deflection (mm)
Figure 15. Loading - Deflection curve
 281

4.4 Post-tension Beam Test

The static loading test and bending fatigue test were carried out on large-scale post-tensioned
test beams in order to determine whether Technora rods can be used as tendons for post-
tensioning. The post-tensioned test beam was a simple beam of rectangular cross-section
60cm x 25cm x 10.5m. A bundle of 19 6 Technora rods with a combined tensile strength
of 800 kN was used as a tendon. A 65mm dia. hard polyethylene sheath was used.

&
5250 (Beam Length = 10500) 250
50 ,250 2 X 220 15X300 = 4500 50 5|
440 D13 150

D13>1

| DIP AFRP
4 D10

5000 ( Span =10000)

250 Side Elevation Section

Figure 16. Dimensions of testing beam

The static loading test was carried out by applying a load in four steps until the final
breaking point. Each loading cycle is shown in Table 10, below. The load steps were function
of the calculated bending moment capacity at various stages (see Table 9).

Table 9 Calculated Bending Moment

Loading Stage Symbol Bending Moment Percentage of Ultimate

(kN.m)
At dead load Md 101 24%
At live load Ml 46 11%
At design load Md+Ml 147 35%
At cracking Mc 192
At breaking Mu 420 100%

Table 10 Loading Cycle

Loading Steps Load Bending Moment Symbol
1st step
1 o * o * * >
147 Md+Ml
2nd step 192 Mc
3rd step 285 Md+4M1
4th step Ultimate Mu
282

500

400
Bending
Moment
300
(kN.m)

200
CalculativeValue
I
Measured Value
100

0 100 200 300

Deflection (mm)
Figure 17. Bending moment - deflection curve

In the graph of the results of the static loading test, the calculated and measured values for
the bending moment show a good coincidence, which suggests that existing design techniques
can be employed on post-tensioned girders using Technora rods.
A bending fatigue test was performed using a large-scale post-tension girder testing bed.
For the bending fatigue test, the lower moment was set at 0.25 Mu, and three tests were
performed with 2 million loading cycles, with the upper moments of 0.65 Mu, 0.55 Mu and
0.45 Mu (the moment where cracking appears). Here, Mu means ultimate moment. A static
loading test was performed after the fatigue test, where the load bearing capability up to
breaking point was tested. The results of the test where the upper moment was set at 0.45 Mu
(the moment where cracking appears) are shown in figure 18. The graph shows that some
decrease in load bearing capability is evident when static loading is performed after the
bending fatigue test at 0.45 Mu, but it is thought unlikely that any loss of load-bearing
capability would occur due to fatigue in girders at the design moment (0.35 Mu.)

Bending fatigue test

 283

upper moment CalculaveValud

Measured Value!
j ^

100 100 200

Deflection (mm) Deflection (mm)

CalculativeValue
Measured Value

100 200

Deflection (mm)
Figure 18. Bending moment - deflection curve

5. PRACTICAL APPLICATIONS
5.1 Prestressed Concrete Road Bridge
A prestressed concrete road bridge using Technora rods was constructed within the
grounds of the Oyama factory of Sumiken Concrete Industry Corporation in Tochigi Prefecture,
Japan, for the transport of raw materials and finished products. As shown in figure 19, one
traffic lane was constructed as a pretensioned bridge, and the other was constructed as a
post-tensioned bridge. Photo 4 shows the completion of the Aramid FRP prestressed concrete
road bridge. The pretensioned bridge was a composite girder bridge 12.5m long, with three
main girders. The post-tensioned bridge was a box girder bridge 25.0m long, with girder
height 1.90m. External cables were also used. Technora <|>6mm spiral wound rods were
used in all tendons, and particularly to be noted is the use of Technora rods (<|>8mm spiral
wound rods), in the stirrups and reinforcement of the pretensioned bridge. In the main cable
of the pretensioned bridge, a bundle of 3 Technora rods of <j)6mm was used, and the same
type of cable was used for the prestressing of the cross-beams and slabs. The internal cable of
the post-tensioned bridge consisted of a bundle of 19 Technora rods of <|>6mm, while the
external cable consisted of a bundle of 7 Technora rods of . The layout of the cables
in the pre- and post-tensioned bridges is shown in figure 30. 4 cables of 3 6 rods at the top,
and 12 similar cables at the bottom, a total of 16 cables, were positioned in the pretensioned
bridge. 10 internal cables (19 ) were positioned in the web of the post-tensioned
284

bridge, and 6 external cables (7 ) in the box girder. Bond anchorages were used to
anchor the Technora cables in the post-tensioned bridge. Particularly to be noted is the use
of non-corrodible FRP anchorings for the external cables. 6)

300 Span length 11790 ,300

0
cr> | P retens i oned br i dge iM

o
cr>
cr> Post-tensioned bri.dge
CO

ioo|
400 Span length 24100 B80
P I a n
9200
600 3990 3990 600
Pretensioned bridge Post-tensioned bridge

,f
CO

i 1 i
1695 1695
Cross section
Figure 19. Outline of Aramid FRP prestressed concrete road bridge
 285

Pretension
Post-tension
79V *--
3<*6 AFRP
At support At mid-span
Rod

AFRP
Reinforcing
Rod

3 d 6 AFRP'
IRo

Figure 20. Cross-section showing layout of Technora tendons

The bridge under discussion was designed in accordance with the Japanese Specifications
for Highway Bridges. The general design conditions for the bridge are shown in table 11. A
live load of TL-20 was assumed as a design load, but the conditions for trailers laden with
concrete (axle weight 300 kN) are also being investigated. The specifications for the Technora
tendons are shown in table 12. Since it had already been established experimentally that
Technora rods have a high tensile strength similar to or greater than that of steel strands for
prestressed concrete, the values for allowable tensile stress as given in the current Japanese
Specifications for Highway Bridges were applied.

Table 11 Design Conditions

Type of structure Pretensioned composite Post-tensioned box

girder bridge girder bridge
Bridge length 12.500 m 25.000 m
Span length 12.390 m 24.880 m
Girder length 11.790 m 24.100 m
Width of 0.6 m + 3.99 m 0.6 m + 3.99 m
carriageway
Longitudinal slope 2.90% 2.90%
Lateral slope 2.0% 2.0%
Live load TL-20 TL-20
Coefficient of :_ 10 20 :_ 10 20
Impact 254-L, 50+L 25+L, 50f L
Horizontal seismic Kho = 0.2 Kho = 0.2
coefficient:
286

Table 12 Specifications for AFRP tendons

Tendon type Pretension bridge Post-tension bridge

36 76 19 6
Guaranteed tensile 150 kN 350 kN 790 kN
strength (Pu)
Allowable tensile strength
During prestressing 0.8 Pu

Immediately after 0.7 Pu

prestressing
With design load 0.6 Pu
applied

Photo 4 Completion of the Aramid FRP prestressed concrete road bridges

5.2 Aramid FRP Prestressed Concrete Berth

One outstanding characteristic of Technora rods is that they do not rust, and thus are
suitable as a reinforcing material in the highly corrosive environment of a marine structure.
The case introduced here is that of a prestressed concrete berth using Technora tendons. A
cargo handling berth within the Ehime plant of Sumitomo Chemical Co.,Ltd. at Niihama
City, Japan, constructed over 30 years ago, had deteriorated and become due for reconstruction.
The new berth was planned as a prestressed concrete slab construction with simple hollow
girders, consisting of five spans and a total length of 61m. In one 9m span, prestressed
concrete girders using AFRP rods as the tensioning material were employed (Fig. 21). The
other spans were constructed using girders tensioned with steel strands (T12.4). The portion
of the berth using girders with aramid tendons consisted of a total of 17 pre-tensioned hollow
girders, with the general design conditions as set out in table 13 below. Figure 22 shows the
cross-section of the berth, the dimensions of the beams, and the layout of the Technora
 287

tendons. In addition to the design-load TL-20, a 500 kN cargo-handling truck-crane and a

350 kN power shovel were also used. Photo 5 shows the completion of the prestressed
concrete berth. 7)

Photo 5 Completion of the prestressed concrete berth

Control room Pit

PC slab.

J
AFRP girder section
500 14 000 14 000 9 000 14 000 9 000 500
Berth length L = 6I,OOO

Figure 21 Outline of prestressed concrete landing berth

288

Deck width 13800

Cross-section of Aramid-PC slab

3 ^
AFRP tendon en

Yi
bo n n o o o no'

100 62.5 XS =500 100

Cross-section of girder

Girder length 8720

CZ 1 1 1 1 1

200 L Soan length 8320 -^ 200

Dimensions of Aramid-PC girder (side elevation)
Figure 22. Outline of prestressed concrete berth using AFRP tendons

Table 13 General design conditions

Type of structure: Pretensioned PC simple hollow girders

Girder length: 8.760m
Effective span: 8.320m
Effective width: 13.800m
Live load: TL-20,Truck crane,Power shovel
10 20
Impact coefficient: 25fL , 50+-L
Design seismic
coefficient: Kh=0.24
 289

5.3 Aramid FRP Ground Anchor

Technora rods have also been used in ground anchors, taking advantage of their rust-free
property. Since AFRP tendons have a smaller elastic modulus than steel, the loss of stress
due to deformation and creep of the natural ground is also small. The material is also
comparatively light, making it easier to work with on steep slopes.
The AFRP ground anchor in this experimental construction was installed as an earth
retaining anchor in work on an underpass beneath a railway line in Kagawa Prefecture,
Japan, as shown in Photo 6. The excavation during this work was 10m deep, and the length
of the anchor was approximately 12m. The experimental construction consisted of an AFRP
ground anchor and a ground anchor using steel strands installed together, so that tension tests
and long-term observations of the two types could be compared.

Photo 6 AFRP ground anchor (Tensioning)

REFERENCES

1 R. Kakihara et al., "A New Aramid Rod for the Reinforcement of Prestressed Concrete
Structures", Proc. of Speciality Conference on Advanced Composites in Civil Engineering
Structures, MT Div. ASCE, Las Vegas, 1991, pp. 132-142
2 R. Kakihara, M. Kamiyoshi and T. Kawasaki, "Mechanical Characteristics of AFRP
Rod", Proc. of JSCE Symposium on Applications of Continuous Fiber Reinforced Plastics
Materials for Concrete Structures, 1992, pp.79~82
3 R. Kakihara, M. Kamiyoshi and T. Kawasaki, "Durability of AFRP Rods", Proc. of JSCE
Symposium on Applications of Continuous Fiber Reinforced Plastics Materials for Concrete
Structures, 1992, pp. 305-308
290

4 H. Ashai et al., "Stress Relaxation Characteristics of Aramid FRP Rods", Proc. of 46th
JSCE Conf. (V), 1991, pp.230~231
5 J. Mizutani et al., "Study on Bond Performance AFRP Rods", Proc. of 45th JSCE Conf.
(V), 1990, pp.606~607
6 J. Mizutani et al,. "Application of AFRP Tendons for Prestressed Concrete Road Bridge",
Proc. of JCI Symposium on Extensive Utilization of Prestressing Technology, 1991,
pp. 187-192
7 Y. Komori et al., "Construction of Prestressed Concrete Using AFRP Rods", Proc. of
JSCE Symposium on Applications of Continuous Fiber Reinforced Plastics Materials for
Concrete Structures, 1992, pp.33~38
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 291
1993 Elsevier Science Publishers B.V.

FiBRA
Tomio Tamura, Research Engineer,

Technical Research Institute, Mitsui Construction Co., Ltd.,

518 Komagi Nagareyama City, Chiba Prefecture 270-01, Japan

Abstract
FiBRA is a continuous fiber rod formed by braiding high-strength fibers
such as carbon, aramid, glass, nylon, or vinylon impregnated with epoxy
resin. FiBRA has found various applications in the construction industry,
where it is used as a concrete reinforcing material, a prestressing tendon,
and soil or rock reinforcement. The main features of FiBRA are as
follows.
Excellent bonding to concrete thanks to the irregular texture inherent to
braided fibers
o A high tensile strength of braided fibers developed by impregnating them
with resin
The ability to manufacture large rods.
Hard, flexible, and styrene foam-cored structural members have been
manufactured since the development which commenced in 1984.

1. MATERIALS USED

1.1. Fibers
Carbon, aramid, glass, nylon, and vinylon fibers can be used to make
FiBRA. Table 1 shows the physical properties of these fibers. FiBRA is
produced by braiding either one of the fibers and impregnating the braid
with resin. Initially, FiBRA based on aramid fibers was developed. Aramid,
and the KEVLAR brand developed by Du Pont, are para-aromatic
polyamide synthetic fibers offering high strength and excellent resistance to
chemicals and heat.

Table 1
Properties of fibers.
KEVLAR-49 NYLON E-GLASS CARBON VINYLON STEEL
Specific 1.45 1.14 2.54 1.75 1.30 7.85
gravity
Tensile gpd 23 10
strength N/mm2 3,000 1,300 2,450 3,600 940 2,000
292

Table 1
Properties of fibers.
KEVLAR-49 NYLON E-GLASS CARBON VINYLON STEEL
Tensile gpd 2 850 48
elastisity N/mm 111,000 6,300 70,000 235,000 26,000 200,000
Elongation % 2.4 19 4.0 1.5 8 1.1

1.2. Resins
Resins act as binding material which helps to fully bring out the
strength of the fiber. Epoxy resins with good resistance to water and heat
are used in the manufacture of FiBRA.

2. MANUFACTURING AND QUALITY CONTROL METHODS

FiBRA is manufactured using a continuous production system consisting
of the processes shown in Figure 1: braiding, resin impregnation,
preheating, hardening, cooling, and cutting.

Figure 1. Production processes.

 293

The processes are identical for all the different types of fibers, since only
the type of resin is varied. Quality control of the FiBRA is carried out as
follows.

1) Diameter m e a s u r e m e n t s
The diameter of a FiBRA is measured in two directions continuously by
the use of laser beam instruments.

2) Visual inspection
Every FiBRA is checked visually for surface irregularities and over-thick
resin.

3) Tensile s t r e n g t h
Five rods out of every batch are tested for tensile strength to verify that
they meet the specifications.

4) Others
Specific gravity and bending tests are also carried out periodically.

3. P R O P E R T I E S OF FIBRA (1>

As illustrated in Figure 2, FiBRA is a braided product, therefore it

features a variety of outstanding properties. As previously mentioned,
FiBRA is used to make rods composed of various fibers. This section
specifically elaborates the properties of a rod in which aramid fiber is used.

Figure 2. Shapes of FiBRAs.

3.1. Tensile strength

FiBRA includes two types of rods; rigid and flexible types. The rigid
type of rod is mainly used for the reinforcement of concrete while the
flexible type of rod is used as a prestressing tendon. Table 2 shows the
properties of both types of rods.
294

Table 2
Properties of FiBRA (Rigid type)
Diameter Sectional area Tensile strength Weight Young's modulus Elongation
Name (mm) (mm2) (KN) (g/m) (MPa) (%)
RA7 7.3 42 62.7 54
RA9 9.0 63 94.1 82
RAH 10.4 85 125.4 108 68,600 2.0
RA13 12.7 127 188.2 162
RA15 14.7 170 250.9 216

Properties of FiBRA (Flexible type)

XT Diameter Sectional area Tensile strength Weight Young's modulus Elongation
Name , ,
(mm) (mm 2 ) (KN) (g/m) (MPa) (%)
RAH 10.4 85 117.6 108
RA13 12.7 127 176.4 162 68,600 2.0
RA15 14.7 170 235.2 216

3.2. Tensile fatigue

A partial pulsating tensile fatigue test of FiBRA RA7S was performed at
room temperature, using a 10-ton electrohydraulic servo type fatigue testing
machine. This was done in order to attain a fatigue limit at which RA7S
would not rupture even after two million cycles of a test. This was carried
out under the condition that the range of stress changes in response to each
cycle of a test, and had the lower-limit load fixed at 50% of the actual
rupture load. Table 3 shows the test results. The RA7S did not rupture
following two million cycles of a test which was conducted with the range of
stress less than 423.4MPa, but did rupture at the 305,000th cycle at the
stress range of 500.8MPa.

Table 3
Tensile fatigue test results.
Lower Upper Stress Frequency
Test Anchorage
Load Stress Load Stress range until break Remarks
No. method
KN MPa KN MPa MPa X10 3

1 32.2 645.3 47.0 939.5 294.2 > 2,000 Did not break
2 32.2 645.3 42.5 990.5 345.2 > 3,557 Did not break Resin
3 32.2 645.3 51.0 1019.9 374.6 > 2,063 Did not break anchorage
4 32.2 645.3 54.4 1088.5 443.3 306 Broke

5 32.2 645.3 46.1 921.8 276.5 > 2,000 Did not break ^chorage
 295

3.3. Relaxation
Relaxation tests were performed at room temperature. The initial load
was 0.6 times the tensile strength. The test results are as plotted in Figure
3. After the lapse of 1,000 hours, the relaxation loss of FiBRA was 12%.

I 10 100 1000
Time (hrs)

Figure 3. Relaxation loss

3.4. Bond strength

The bond test of the interface betv/een concrete and FiBRA was
performed using a testing device illustrated in Figure 4. The test results are
as plotted in Figure 5. Although the initial value of FiBRA varied with
RA9S (the FiBRA with a sanded surface) and RA9 (the FiBRA without a
sanded surface), it was evident that as a whole, both had the same level of
bond strength as that of a deformed steel bar.

150
n ^ H L - DiSDlacement meter
4 \Reinforced bar

^-t Load cell

I I
(mm) Pull out Displacement (mm)
I Pull

Figure 4. Bond test apparatus. Figure 5. Bond strength

296

4. DURABILITY

4.1. Test methods

Table 4 shows the chemical resistance test items and the test conditions.
On the assumption that a rod with a smaller diameter is prone to be
affected by heat and chemicals, RA3 was selected as a test sample. Heat
resistance tests were performed using dry heat and moist heat. Chemical
resistance tests used acid, alkali, artificial seawater, organic solvent and
other similar solutions, for a total of nine kinds of solutions. Taking into
account the affect of the interface of artificial seawater, both a whole rod
and a half of a rod were immersed in the artificial seawater. The
evaluation of the durability of the rods is represented by using the
percentage of the tensile strength retained by each rod and the percentage of
change in weight. Tensile tests were performed using the procedure referred
to in the preceding paragraph.

Table 4
Chemical resistance test results.
Percentage Percentage of Percentage Percentage of
Chemical of change in residual tensile Chemical of change residual tensile
substance weight strength substance in weight strength
(%) (%) (%) (%)
NaOH (30%) -0.25 92 H2SO4 (30%) 1.46 80
NaOH (5%) -0.27 90 H2SO4 (10%) 1.38 85
NaCIO (5%) -0.73 95 Toluene -0.36 96
Ca(OH)2
-0.26 96 Trichlene 0.61 93
(saturated)
Seawater -0.10 100 Dimethy-
1.53 99
formamide
Seawater -0.20 93 Kerosene -0.09 93

4.2. Test results

Figure 6 shows the results of dry moist heat resistance tests, both in
terms of the percentage of the strength retained at each of the temperatures
with the strength of untreated FiBRA taken as 100. Each of the figures
shows the average value of six samples. The results of these tests indicated
that the percentage of the residual tensile strength maintained up to the
treatment temperature of 220, was high at approximately 95%. This
represents a considerably high level of heat resistance for an organic fiber.
Table 4 shows the chemical resistance test results of FiBRA attained on the
30th day following its immersion in each of the chemicals used. In the
table, both the percentage of change in weight and the percentage of
residual tensile strength each represent the average of five samples. With
the exception of sulfuric acid (H2SO4), all of the residual tensile strengths
exceeded 90%, and there were no specifically noticeable changes observed
with regards to the 30-day immersion.
 297

1
*~5
0
'' I
lOOf
Q)

0
Excosed
odryte! Teratre

noist heat 40
D 1 (W) arc
rai
140 150 200 NO 260 <80 rot ^ r
ABffC
50 100 200 500 I0GO 2G00 5000
exccsue
exposed temperature ('C Exposed Time (Hours)

Figure 7. Alkali resistance test

Figure 6. Heat resistance test result result

Figure 7 shows the results of tests conducted on the alkali resistance of

aramid fiber (Kevlar 49). Although there has been an indication that
aramid fiber is less resistant to alkali, recently the fiber material of aramid
fiber itself has been increasingly improved, providing for a considerable
degree of compensation for the weakness. As well, because the strength of
aramid fiber is little affected by Ca(OHK the saturated solution given in
Table 4, and from the fact that aramid fibers were gathered to create a rod
and were immersed in epoxy resin, the alkali resistance of FiBRA has been
considerably improved as compared to the resistance of aramid fiber itself.

5. APPLICATION EXAMPLES OF FIBRA

Since 1984 when FiBRA was developed, it has been utilized extensively
in various areas, such as in the reinforcement of concrete and rock. The
following are some detailed application examples.

5.1. A prestressed concrete (PC) b r i d g e

In 1990 the construction of a bridge in which FiBRA was incorporated as
a prestressing tendon, was completed. This bridge is a 36.02m-long, 2.4m-
wide, three-span pretensioned girder slab bridge of I shape spanning over a
pond at a golf course. Figure 8 provides a general illustration of the bridge.
Seven pretensioned I-shaped girders, which are the PC girders specified for
light-load slab bridges by JIS A 5319, constitute each span. Of the total of
the 21 girders (7 girders X 3 spans), three were made with FiBRA
tendons.
298

C r o s s Section of Main Girder

Cross Section 0rd inarr Girder AFRP Girder

1.800
AFRP Girder
300,

ESi s^
V V
I OU

\ \ Main Girder
Deforced Bar for Transversal Rainforeeeent D2S
I
P.C. Strand B r a i d e d AFRP ..
l- 10. 8 Rod 8 - 0 1 4 fc\

P r c s l r c s s c d C o n c r e t e Bridge Girder for Light Weight Slab Blidgc

Figure 8. General view of P.C. bridge.

As in the case of an ordinary steel tendon girder, the FiBRA girders

were manufactured in shop, followed by delivery to the construction site for
erection. Figure 9 shows the girder being erected.

Figure 9. Erection of FiBRA girders is proceeding.

 299

5.2. A floating bridge

FiBRA was employed in the form of prestressing tendons and a concrete
reinforcing material, for incorporation into a floating bridge which must be
light and highly durable. The bridge is 56.37m long and 4m wide crossing
over a pond at a golf course in Ibaraki Prefecture, Japan. Figure 10 is a
general view of the bridge. It consists of six PC pontoons, cast-in-place
concrete abutments, and anchor blocks for mooring.

Figure 10. View of a floating bridge.

Each PC pontoon is about 7.6m long, 3m wide, l m high, and weighs 14

tons. It is a bottomless, partitioned concrete box, and has the hollow inside
space filled with polystyrene foam.
The concrete reinforcing material of the pontoons includes four FRP
tendons (for a total of eight) (FiBRA FA13). There are continuous, spiral
FRP stirrups (CLATEC rod-10D) installed in each side wall, and FRP
reinforcing mesh (FiBRA-RA7) in the floor and the end walls. To connect
these PC pontoons, four FRP tendons (FiBRA-FA15) were installed through
ducts, formed using polystyrene foam in advance, and tensioned at the outer
ends of the six-pontooned structure. The inside of the ducts were not
grouted, so the tendons remained unbonded. Sandwiched between the end
walls are layers of neoprene rubber, each 10cm thick, so that the pontoons
were of a flexible construction.
300

Figure 11 shows the floating bridge as completed.

Figure 11. A floating bridge as completed.

5.3. Building floors

In 1991, with the aim of reducing the deflection of building floors and
protecting the concrete from cracking, FiBRA (type FA15) was used as post-
tensioning tendons. Figure 12 shows the FiBRA cable being installed.

Figure 12. FiBRA tendons for building Figure 13. Installation of a beam is
floor. proceeding.

5.4. Building beams (2)

In 1992 FiBRA was employed for use as prestressing tendons (FiBRA-
FA13) and stirrups (FiBRA-RA7) for incorporation in building beams. This
was the first case in which FiBRA was used, with the approval of the
 301

Construction Minister, as a main structural member of a building. Figure

13 is a photograph showing the beam being installed.

5.5. FiBRA bolts (for rock reinforcement)

While a tunnel is being excavated through the bedrock, the natural
ground surrounding the tunnel may become loose or collapse. To prevent
this, steel rods known as rock bolts are inserted to reinforce the bedrock.
We have developed an FRP bolt, which has excellent durability, for use as
reinforcing material (FiBRA bolt).

A FiBRA bolt is structured so that the outside is comprised of braided

fiber, and the inside, foam plastics, thereby increasing the bond strength.
The features of the FiBRA bolt include excellent durability, and permit ease
of cutting when excavation is resumed. These advantages have further
expanded its use as a means to stabilize the tunnel face or to reinforce the
tunnel wall to be re-excavated. Figure 14 shows a FiBRA bolt and an
anchor.

Figure 14. FiBRA-bolt with an anchor. Figure 15. Floor panel for raised
floor.

5.6. Floor panels

Taking advantage of the lightness and nonmagnetism of FiBRA, we have
developed a panel for an accessible type floor, as illustrated in Figure 15.

5.7. Plantable concrete panel

A plantable concrete panel which we have developed, allows for the
growth of the roots of plants through the voids of the porous concrete
reinforced with FiBRA. This came into being through the utilization of the
rust free properties of FiBRA rods.
302

6. T H E M E S UNDER D E V E L O P M E N T

In order to use FiBRA in various other areas, we are proceeding with the
developments as discussed in the following paragraphs.

6.1. E x t e r n a l cable
Taking advantage of FiBRA's excellent durability and flexibility, we are
testing the applicability of FiBRA for use as an external cable for
prestressed concrete. Figure 16 is a photograph of a beam being tested.
The beam has a tee cross section and is 12 meters in length. With an
external cable arranged on either side of the beam, a load was applied to
the center of the beam, followed by a bending test. This test proved the
applicability of FiBRA as an external cable.

Figure 16. FiBRA for external cable. Figure 17. Falling weight test.

6.2. Shock-absorbing concrete structure ( 3 )

We have developed a concrete structure which takes advantage of the
excellent shock-absorbing ability of FiBRA. The use of this structure
includes a rock shed, and a road concrete barrier. Figure 17 shows a falling
weight test in progress.

7. REFERENCES
1. M. Tanigaki et al., Study of Braided Aramid Fiber for Reinforcing
Concrete, Proceedings of IABSE Symposium, Helsinki, Vol. 13, 1988, pp.
15-20

2. S. Matsubara et al., Practical Application and Performance of PPC

Beams Reinforced with Braided FRP Bars, ACI Spring Convention, 1993
 303

3. H. Mikami et al., Impact Resistance of Concrete Slab Reinforced with

Braided AFRP Rods, Proceedings of the ISEE, Sendai, Vol. 1, 1992
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 305
1993 Elsevier Science Publishers B.V. All rights reserved.

GLASS FIBRE PRESTRESSING SYSTEM

Dr.-Ing. R. Wolff
Dipl.-Ing. H.-J. Miesseler

SICOM GmbH
Gremberger Str. 151a
D-5000 Cologne 91, Germany

1. INTRODUCTION

High performance glass fibre composite materials - till now nearly exclusively used
in the field of aerospace or car industry - are now also applied in the construction industry.
The choice of different kinds of fibres, such as carbon, glass or aramid fibres, and
of different kinds of resin, e.g. polyester, epoxy or vinyl - offers the possibility to produce
fibre composite materials for a large range of applications.
Glass fibre composite materials - especially used in the construction industry as
prestressing elements, distinguish themselves by their tensile strength which is equal or
higher than that of high tensile steel. They have a high elasticity, an excellent corrosion
resistance and there is the possibility to integrate optical fibre sensors for a permanent
monitoring of fibre composite materials and the concrete structures themselves.

2. MATERIAL CHARACTERISTICS

2.1 Available dimensions and strengths

The glass fibre composite bars employed to date comprise 68 vol.-% glass fibres
and 32 vol.-% unsaturated polyester resin or epoxy resin. A total of 64 000 strictly uni-
oriented E-glass fibres with a diameter of approx. 25 \ form a 7.5 mm diameter bar. The
longitudinal tensile strength of this new composite material is similar to that of high-tensile
prestressing steels [1]. A comparison of the stress/strain curve of glass fibre materials with
that of comparable conventional ST. 1470/1670 prestressing steel is shown in Fig. 1. Table
1 gives the main characteristic values of the material in comparison to the steels and
composite bars comprising other fibres. These different materials are also classified in areas
of application. The main distinctions between the new developed glass fibre prestressing
tendons (HLV - tendons) in comparison to prestressing steel are:
306

- the modulus of elasticity for HLV - tendons amounts at 51.000 N/mm2 to only one fourth
of that of steel prestressing tendons
- the prestressing tendons show linear stress/strain behaviour until failure
- low specific weight of 2.0 g/cm3 compared with 7.85 g/cm3 for prestressing steel
- possibility to integrate sensors to create intelligent prestressing systems.
Whilst a linear and a plastic range can be differentiated with steel, the characteristic curve
for glass fibre composite material up to failure point complies precisely with Hooke's Law.
Other important properties of the glass fibre composite material are:
- the long-term strength amounts to 70% of the short-term strength
- good resistance in aggressive environments
- shows the same behaviour at high temperatures as prestressing steel
- electro-magnetic neutrality.
The glass fibre bar is provided with a purpose-developed polyamide coating for the glass
fibre bars with polyester resin and a special epoxy resin powder coating for the glass fibre
bars with epoxy resin as a protection against chemical environments, for example against
chloride and alkaline, or mechanical damage.

2000 1 1
SI" 1470/1670
1800

h
/ 1 |

1600
^ HLV-Bar
<J 1400
E = >1525 / im2
1200 - E = 51000 N/rTim2
e u = 3,3%
n 1000
s
CO

800 /
a
600 1

400 BSt500 S ^

200

4 6 8 10 12 14
Strain [%]

Fig. 1: Stress/strain diagram of a glass fibre bar in comparison with steel bars
 307

Reinforcing Prestressing (R) Polystal (R)Arapree Carbonfiber

steel steel ( 6 8 % Glass- (Aramid- composite

BSt 5 0 0 St 1470/1670 fibers) fibers) material

Tensile strength (N/mm 2 ) >550 >1670 1670 1610 1700

Yield strength (N/mm 2 ) >500 >1470

Ultimate strain (%) 10 6 3,3 2,5 1,1

Modulus of elasticity
(N/mm 2 ) 210.000 205.000 51.000 64.000 146.000

Specific weight (g/cm3) 7,85 7,85 2,0 1,3 1,5

Reinforced Stay cables

Fields of application concrete Prestressed structures bracings

structures

Table 1: Materialcharacteristics and -comparisons

2.2 Long-term behaviour

Long-term behaviour under permanent static load is illustrated in Fig. 2. Due to the
relatively, but for this type of test not unusually large dispersion of individual values, only a
calculated value for the expected time-dependent creep strength can be estimated. At the
present, a time-dependent tensile strength of 70 % of the short-term strength is assumed.
This value is covered by a safety factor of 1.42 in the dimensioning.
Long term strength
Short term strength

1.0

0.9

0.8

0.7

0.6

0.01 0.05 0.1 0.5 1 5 10 50 100 5001000 5000 104 5x10* 105 5x105 10*

Lifetime t, log(hours)
Fig. 2: Time-dependent tensile strength of glass fibre composite bars
308

The relaxation processes observed with composite glass fibre bars are traceable to
irregularities caused in manufacture, e.g. deviations from the strict axial orientation of the
fibres within a bar. As with prestressing steel, stress losses resulting from relaxation show
linear behaviour if plotted along a logarithmic time axis (Fig. 3). It is thus possible to
extrapolate data gained in short term tests over longer periods. A characteristic of z t =
3,2 % is derivable therefrom for an application as prestressing reinforcement for a time
period of 5 * 105 hours. A significant change in this behaviour is not to be expected even at
higher temperatures (e.g. 50C).

II H U I
Testing Temperature ]?oc
,=790 N/mm2
evidenced by tests
- - extrapolated
II II II I
i 2 1 1 1
*** -J - ^
"--i 1 1
1
1
1
1 1 + - _J
1 1 1
1 1
100 200 500 1000 10000 100000 Hours
1 5 10 25 50 Years

Stress duration

Fig. 3: Relaxation of the glass fibre composite bars

2.3 Dynamic behaviour

The influence of non-static loads was assessed in fatigue tests whereby, with respect
to utilisation as prestressed reinforcement, only the range of swelling tensile stresses which
includes stress generated by the prestressing force is of significance.' The results of such
tests, which were carried out at a constant maximum stress' limit to the extent of the
working load are shown in Fig. 4. While in the case of prestressing steel it is permitted to
define the resultant 2 * 106 variable stress component as the fatigue strength, in the case of
composite fibre materials failures still have to be expected even with higher load cycles.
A possible consequence thereof is that the load cycles to be expected for a structure
must be forecast and investigated in each individual case to determine whether the tolerable
stress ranges are sufficient.
 309

120
F
E
\ 100
-z.
_c
b"
<( 80
OJ
cn
c
60

\


40
^~~~~
3 70
<

-
J4 10 5 10| 10 7 1C

Ng=2x106 Alternativ stress rate N

Fig. 4: Fatigue behaviour (10 % failure probability)

2.4 Influence of temperature

In fire tests on structural elements carried out at the Brunswick University, an
exposure time of over 100 mins. at 1000 C in the fire chamber could be achieved using the
minimum concrete covering for the prestressing tendon, as laid down by ZTVK
(Supplementary Technical Regulations for Bridges and Tunnels).
Proof of fire classification F 90 demonstrates the glass fibre material's applicability at high
temperatures.
Glass fibres themselves are relatively insensitive to high or low temperatures.
However there is partial loss of strength in the matrix sheath at approx. 250 C.
The glass fibre bars therefore have to be protected against overheating wherever the
composite properties of the matrix are statically required, hence, particularly in the
anchorage zone, by structural means, e.g. thicker concrete cover.
The dependency on temperature of the tensile strength of glass fibre composite material
sections is shown in Fig. 5.
310

Ultimate tensile strength DZ(T) at elevated temperatures

Ultimate tensile strength D2(T0) at ambient temperature

| |
1.0
^ ^ Prestressing steel
0.9
*^^. r ^ St. 1450/1600 |

0.8 '^v
Prestressing steelj-L-\ \
St. 1600/1800 I \ \
0.7
Glass fiber-^Tv
composites v \ "f
0.6 \
\
0.5 \
\
25 100 150 200 250 300 350 400

Temperature T(C)

Fig. 5: Tensile strength retention at elevated temperatures

2.5 Behaviour under the effects of aggressive media

Reinforcement tendons for applications in prestressed concrete structures must be
permanently resistant to the strongly alkaline reacting medium of concrete. Moreover,
during prestressing they must withstand high mechanical loads without damage. Both
requirements place high demands on the bar material. Although glass fibres are sensitive to
many media due to their exceptionally large surface area relative to their volume, glass
fibre composites are employed successfully and comprehensively for the manufacture of
pipes, tanks and installations in chemical plant construction, because these composite
materials are more resistant to many corrosive influences than other materials. In a
composite material the glass fibres are protected by the resin matrix. However a
prerequisite for permanent protection is that strain remains low (0.1 - 0.2 %) under
operating conditions.
Prestressing reinforcement comprising composite glass fibre bars, however, reach
strains of > 1.5 % in their working state. These high strain values fundamentally change
the behaviour of media influences. Under these conditions the matrix material is no longer
able to assure permanent protection of the glass fibres. During development of the glass
fibre bars it was therefore not possible to refer back to the experience gained in the
construction of tanks. New concepts had to be put forward, both with respect to the basic
materials and to the methods of testing. In this conjunction particular reference should be
made to test procedures for the investigation of media influences subjected to simultaneous
 311

mechanical loads in compliance with practical conditions. These procedures substantially

supported the specific development of utilizable composite glass fibre bars for prestressing
applications. While time accelerated tests, which were performed under exaggerated
conditions (temperatures 50 - 70 C, use of aqueous cement solution pH = 13), provided
relatively rapid conclusions with regard to the fundamental utility of a bar material,
investigation of long-term media behaviour was effected practice-orientated with the aid of
so-called standard tests. The result of a sophisticated and arduous series of tests is a bar
material with a coated consisting of a high-fill Polyamide 6, which on the one hand ensures
media resistance and on the other withstands extreme frictional loads during prestressing.

Thermostat with pump i

Buna N" hoses

#:
i - POLYSTAL - bar prestressed to 50%
1
Aqueous cement solution pH - 13,0
Outer steel tube as abutment

Fig. 6: Investigation into media resistance

Prestressing anchorage
Outer steel tube as support

POLYSTAL- bar
prestressed to 50%

#=8-
L
~ PE perforated tube
{
Cement mortar

Fig. 7: Investigation of resistance to media by means of standard tests

312

3. THE PRESTRESSING TENDON AND ITS ANCHORAGE

Being composite anisotropic material, these composite glass fibre materials only
tolerate transverse pressure up to 10 % of their longitudinal tensile strength. For this
reason, completely new solutions had to be found in the field of anchorage engineering.
Thus, the absence of cold workability prevents the utilization of upset heads, rolled-on
threads or even the utilization of steel wedges which "bite" directly into the "soft"
composite glass fibre bar material. The relatively low interlaminar shear strength of the
resin matrix requires a comparatively larger anchor length than would be the case with steel
prestressing tendons.
On the basis of experience gained at Stuttgart University [2, 3] where clamping plate
anchors had been used with batch-produced small rectangular bars, the round 7.5 mm
diameter glass fibre bars were tested by using modified clamping plate anchors [11]. The
bars had been produced by a continuous manufacturing process since 1978. These
experiments, in which the round bars were secured between semi-circular grooves in the
clamping plates, did indeed show satisfied results but also clearly demonstrated the limits of
economical application using round bars. The development of a tubular grouted anchor at
Strabag Bau-AG heralded a breakthrough for the anchorage of high performance glass fibre
composite tendons.
The composite tendon is grouted in a profiled steel tube with a synthetic resin
specially developed for this purpose. The use of these grouted prestressing tendons covers
the entire spectrum of light- and medium-weight prestressing tendons and the entire sphere
of soil and rock anchoring.

Fig 8: Grouted anchorage of a high performance composite tendon

 313

The types of anchors developed up to the point of application maturity will be employed for
the anchorage of prestressing tendons in prestressed concrete structures, and also for airside
anchorage with soil and rock anchors. With soil and rock anchors, the anchorage
technology of conventional steel soil anchors on the soilside can be adopted due to the
relatively good bonding property of the bar material with the surrounding concrete medium
(comparable with profiled prestressing steels).

4. APPLICATIONS

4.1 "Lnen'sche Gasse" Bridge

Already in the year 1980 the small bridge, "Lnen'sche Gasse" was erected by the
joint venture "HLV Elements" (Strabag Bau-AG, Bayer AG) as a trial structure within the
scope of the Road, Bridge and Tunnel Department of the City of Dsseldorf. It was
provided with limited prestressing without bond by means of 12 composite glass fibre
prestressing tendons and monitored by way of continual stress force measurements. Having
finally been dismantled after 5 years for materials investigations it was replaced by a new
prestressed tendon generation. The positive measuring results confirmed the preceding
laboratory tests as expected. The grouted anchorage for the round bars was then further
developed up to application maturity from the four anchorages employed there.

Fig. 9: "Lnen'sche Gasse" Bridge

314

4.2 Ulenbergstrasse Bridge

Fig. 10: Overall View of the Ulenbergstrasse Road Bridge

Owing to the successful experience with the Lnen'sche Bridge pilot project, the
Road, Bridge and Tunnel Department of the City of Dsseldorf which had already foreseen
the new development once again decided that this bridge, which is part of a heavily
frequented road, should be built with the innovative prestressing system. Opened to traffic
in 1986, the Ulenbergstrasse Bridge represents a milestone for the continuous progress in
development of the glass fibre prestressing technology. Countless initial trials with the new
bar material were carried out both in the laboratories of the joint venture partners and at
notable German universities, so that finally all the knowledge gained could be concentrated
in the bridge structure.
The Ulenbergstrasse road bridge (load classification 60/30 as per DIN 1072) is a two
span, solid slab bridge with span widths of 21.30 m and 25.60 m. The 1.44 m high and
15.0 m wide superstructure received limited longitudinal prestressing with a total of 59
HLV prestressing tendons (each with a working load of 600 kN) and was subsequently
grouted with a synthetic resin mortar specially developed for this purpose [4, 5].
Measurements taken during the works, particularly tensioning paths during prestressing and
deflection after handover to traffic, conformed well with the previously calculated values.
 315

46.90

Longitudinal Section Technical Data

15.00 Spans: L1:L2 - 21,30 : 25.60 m
Slab width: 15.00 m

Slab thickness: 1.44 m

Clear height: 4.75 m

Load class (DIN 1072): 60/30
Degree of prestressing: limited

Nature of the post-tensionlng with

Cross Section composite action: subsequent bond

Fig. 11: Technical Data of the Ulenbergstrasse Road Bridge

4.3 Brine Tank at Bayer, Dormagen

Brine tank covers at Bayer's Dormagen plant were heavily affected or already
destroyed by vapours with a high chloride content as a result of the highly corrosive
environment. The two-span beam structures, which must withstand a load in compliance
with bridge classification 60 according to DIN 1072, were replaced by a prestressed,
prefabricated structure. Ulenbergstrasse type HLV-prestressing tendons with a working load
of 600 kN were also used in this instance, whereby the excellent resistance of the glass
fibres to an aggressive environment was decisive for the choice of this material. Previous
observations have shown that use of this corrosion retardant prestressing reinforcement can
considerably increase the life of building components subjected to such heavy demands.
316

4.4 Pedestrian Bridge at Berlin-Marienfelde

Fig. 12: Marienfelde Bridge

The range of the prestressed bridge structures was extended by an innovation in the
year 1988. The Berlin-Marienfelde pedestrian bridge was not only distinguished by the
novel prestressing method but was also the first building structure in Germany since the war
to be executed with external prestressing. The superstructure consists of a two span, 1.10 m
high, TT-beam with span widths of 22.98 m and 27.61 m and is furnished with partial
prestressing with 7 HLV-prestressing tendons. The tendons run externally between the two
main bridge beams, are diverted at two points on each span and upwards along the central
pier with the aid of guide saddles, and anchored through the two cross-beams at the
beginning and end of the bridge.
After completion, the bridge was loaded by the placing of 5 layers of 250 concrete
slabs with the weight of 1 t each (corresponding to twice the total traffic load). This trial
came within the scope of a research project and was accompanied by a comprehensive
measuring programme in which changes of the prestressing force in the tendons and the
deflection in the midspan were constantly controlled [6].
 317

Longitudinal section

l 5.00*|
Technical data
Spans 27,61 m + 22,98 m
Width of TT - beam 5,00 m
Thickness of TT - beam 1,10 m
Clear height 4,95 m
Degree of prestresslng Partial
Nature of the composite External
Cross section action prestresslng

Fig. 13: Technical Data of the Marienfelde Bridge

4.5 Mairie d'lvry Metro-Station (Paris)

Fig. 14: Entrance to the Station Platform

318

Rehabilitation of the Mairie d'lvry metro-station in Paris impressively establishes the

wide range of application possibilities for HLV-prestressing tendons. As a result of one-
sided excavation directly adjacent to the metro-station considerable cracking over a length
of ca. 110 m had occured in the approximately 70 years old concrete vault arch owing to
lateral load release. The client envisaged the installation of a prestressed tie-rod as a
rehabilitation and safety measure. The required prestressing of the arch walls was achieved
by installation of 36 HLV-prestressing tendons and the rehabilitation measures were thereby
executed as required. The electro-magnetic neutrality of the bar material proved to be most
advantageous here. The tendons were prefabricated in the R&D Department's laboratory
and the prestressing works executed by the same team.

|- 22.45

Fig. 15: Technical Data of the Mairie d'lvry Metro-Station

 319

4.6 Schiessbergstrasse Bridge

Fig. 16: Schiessbergstrasse Bridge

In the course of services connection to the Bayer multi-storey car park, the
Schiessbergstrasse to the north of the Bayer plant in Leverkusen was elevated to the second
level. The three-span road bridge (load class 60/30) with span widths of 2 x 16.30 m and 1
x 20.40 m and a slab thickness of 1.10 m has limited prestressing with 27 post-bonded glass
fibre prestressing tendons with a working load of 600 kN each. The handover to traffic was
in November 1992.

Plan view

Technical data
1.35 1.35 Spans L1=L3:L2--16.30:20.40
Slabs width b70m
Slabs thickness 1.12m
Clear height 3.00m
Load class ( DIN 1072 ) 60/30
Degree of prestressing Limited
Nature of the post-tensloning with
Cross section composites action subsequent bond

Fig. 17: Technical Data of the Schiessberstrasse Bridge

320

4.7 Ntsch Bridge in Krnten, Austria

Fig. 18: Ntsch Bridge

The Ntsch Bridge is the first bridge in Austria with glass fibre prestressing tendons.
The triple-span road bridge (bridge class I, Austrian Standard B 4002) with two span widths
of 13.00 m and one of 18.00 m and a slab thickness of 0.75 m, will be furnished with
limited prestressing comprising 41 glass fibre prestressing tendons, with post-bond also with
a working load of 600 kN each. The handover to traffic was in May 1992.

Ansicht

Technische Daten
Spannweiten L 1 - L2:L3==13.00:18.00
Plattenbreite 12.00m
Plattendicke 0.75m
Lichte Hhe 4.70m
Bryckenklasse (-Norm B 4 0 0 2 ) 60t
Grad der Vorspannung beschrnkt
Art der Vorspannung nachtrglicher
Verbund

Fig. 19: Technical Data of the Ntsch Bridge

 321

4.8 Monastery Heydau

Fig. 20: Monastery Heydau

During the rehabilitation of the Monastery Heydau in Altmorschen piers have been
anchored with masonry bolts with the respective masonry walls. A prestressing of the
masonry anchors was required. Moreover the anchoring with masonry bolts in the masonry
itself was requested because optical impairements of the historical front sides should be
prevented. The chemical composition and the moisture conditions of the masonry did
complicate the use of conventional anchorings with steel bolts. Therefore a safety concept
on the basis of glass fibre masonry anchors grouted with resin mortar has been elaborated.
In case of these safeguards the tensile force of the prestressed masonry anchor is transfered
to the sandstone blocks over a length of approx. 120 mm. In case of applied prestressing
forces of 15 kN per anchor there is a safety margin of 2,5 times between the state of
working and the state of failure.
The experiences made during the described project allow to draw the conclusion that
with these discussed new masonry bolts a promising possibility for a durable strengthening
of masonry buildings has been found.

4.9 Horrem Community Centre

In order to attain as elegant a design as possible, the 4 timber beams with span
widths of 16.84 m were trussed. Each beam was furnished with 5 prestressing tendons
having 4, 5, 9 and 10 glass fibre composite bars. The tendons were prestressed with 200
kN. The structure is to remain exposed, thus achieving an interesting and somewhat
technical effect. The material's resistance to corrosion was a further point in favour of this
choice of design.
322

Fig. 21: View of timber beams and tendons

5. SENSOR TECHNOLOGY

In order to monitor buildings in a useful way it is necessary to develop appropriate

sensors which are able to guarantee reliable measured values during a long period of time.
Nowadays the intelligent processing of the high quantity of measured values is no problem
due to the available and efficient personal computers.
The application of strain gauges on the prestressing steel or the measurement of the
prestressing forces with the aid of load cells is not possible in case of prestressing with post-
bond. Moreover it is not a durable solution in case of prestressing without bond.
Only the application of fibre composite materials facilitates a permanent control of the
prestressing element over its entire length due to the integration of copper wire sensors or
optical fibre sensors. Even the monitoring of each individual bar is possible. The sensors
are already integrated into the tendon during its fabrication. The sensors indicate the
integrity of the tendon or they locate the damage.
Besides the monitoring of the prestressing elements the observation of the
stress/strain behaviour of the concrete in the zone subject to tensile forces is very important.
Therefor the tensile zone above the piers and the spans are monitored permanently with
integrated optical fibre sensors with a measurement accuracy of _ 0,2 mm. With special
developed optical strain sensors it is possible to measure the influence of loadings with a
short term accuracy in the -range.
 323

5.1 The crack detection sensor

In the case of this type of sensor several optical fibre sensors or one optical fibre
sensor are stranded with each other with one or several steel wires. If this sensor - which is
connected normally to the structural element which shall be monitored in a distance of 1 m
through a bearing material - is exposed to tensile load, the sensor effect is created by a loss
of light intensity as a result of micro-bending. This light attenuation allows an integral
measurement of the changes of the length of the sensor with a long term accuracy of _ 0,2
mm. The localization of these extensions changes (e.g. cracks) is carried out with an
Optical Time Domaine Reflector (OTDR) by reflection measurements and a local
dissolution of i 0,5 m.

wire strand protective sheath

Fig. 22: Crack detection sensor

5.2 The crack width sensor

The crack width sensor is a loop bended optical fibre sensor which is fixed at two
bars. These bars are connected to the structure on the right and on the left of the joints or
cracks which shall be monitored. The sensor monitors and measures permanently changes in
the joints or crack width with a measuring sensitivity of 0,02 mm and a measurement ranee 6
of0,l-10mm.

recording range for crack width monitoring load transfer rail

Fig. 23: Crack width sensor

324

5.3 The reflector sensor

A light signal is passed through the optical fibre sensor. The reflectors reflect part of
the light while non-reflected light travels to the next reflector. In this way up to 30
measuring points can be monitored with a single sensor. Velocity measurements provide
data on the deformations.
An other design is the parallel arrangement of several optical fibre sensors. In this
case a light signal is passed one after the other through each optical fibre sensor with the aid
of a multiplexer. With velocity measurements the length of the respective optical fibre
sensor is measured and then compared with the sensor length of the parallel arrangement.

1. Analysis electronic
2. Bore hole
3. Glass fibre composite pipe
4. n optical fibre sensors
5. Reflective end of sensor
6. Synthetic resin
7. Basic length
as required
8. Semi-permeable fe^
reflector

Fig. 24: Reflector sensor

5.4 Corrosion monitoring system

Due to the increase of environmental damages (acid rain and de-icing salts) the
influence of chemicals becomes bigger and bigger. For that reason the Institut fr
Bauforschung (Institute for Construction Research) in Aachen developed so-called corrosion
cells. In the Schiessbergstrasse Bridge in total 9 corrosion cells have been integrated into the
superstructure and in the side walk. In the Bridge Ntsch in total 5 corrosion cells are
integrated in the side walk. These cells indicate the progression of carbonation and an
eventual penetration of chlorides. Here all parameters having a decisive influence on the
chemical changes of the concrete and consequently on the reinforcing steel are measured.
 325

Fig. 25: Chemical sensors for the bridge

5.5 Data processing

The measured values are processed by a personal computer in a measuring chamber
with a special developed software. A phone line from the measurement chamber to the
office of the client allows an inquiry and a storage of the measured values. Comparing the
measured values with the theoretical values it is possible to indicate immediately
considerable divergences by so-called alert indications. This automatic processing and
control of all data which - important for all changes in the building structures - facilitates a
significant reduction of the maintenance costs of the buildings. This permanent observation
allows at any time to recognize changes in the building structures and to start at an early
stage with countermeasures. It is known that the costs for maintenance can be reduced
considerably when they are done in good time.
326

6. TRIAL LOADINGS

6.1 Ulenbergstrasse Bridge, Dsseldorf

The first application of the optical fibre sensors was the Ulenbergstrasse Bridge in
Dsseldorf. Both measurement systems - optical fibre sensors in the prestressing tendons
and also directly embedded in the concrete - were installed. The measuring results of this
optical fibre monitoring system were stored on tape and than transferred to a computer. The
attenuation curve of the monitoring results recorded in one week in August 1987 showed,
that no changes had taken place on the structure. Also the measurements taken in the last 5
years do not show any changes in the load bearing structure.

Control of prestressing
Control device

H *&
1HLV prestressing tendon withl
integrated optical fiber sensorj

)ata
processing

Fig. 26: Measurement principle

temperature

tension

attenuation

17.8. 18.8 19.8. 20.8. 21.8. 22.8.

MONTAG DIENSTAG MITTWOCH DONNERSTAG FREITAG SAMSTAG

Fig. 27: Measurement results

 327

6.2 Marienfelde Bridge, Berlin

The Marienfelde Bridge in Berlin, a two span double T-beam construction, is
designed as having external prestressing with glass fibre tendons without bond. Each of the
seven glass fibre tendons has integrated sensors. A series of optical fibre sensors has also
been placed directly in the concrete or retro-mounted. The arrangement of the sensors in the
bridge is shown in Figure 28. During a trial loading carried out in November 1989, 250
concrete slabs in five layers, with a weight of one metric ton per slab, were used. Twice the
live load was applied in this trial loading. The cracks generated in this way - there was
partial prestressing in this case - could be accurately recorded with the aid of optical fibre
sensors. A crack width sensor, provided with a 8 m extension on the underside of the
bridge in the long span, detects the position of the individual slabs on the surface of the
deck slab.

cross section
Fig. 28: Measurement layout
SLAB 20 SLAB 21 SLAB 22
LAYER 4 LAYER 4 LAYER 4

*
i L/v
f
I
-*

REACTION OF THE CRACK - WIDTH SENSORS ON REMOVAL OF

THE CONCRETE SLABS FROM THE CENTRE OF THE LONG SPAN

Fig. 29: Monitoring of the test load by means of the

optical fibre sensors embedded in the concrete
328

6.3 Schiessbergstrasse Bridge, Leverkusen

This bridge (bridge classification 60/30), is designed with limited prestressing
comprising 27 glass fibre prestressing tendons. Three glass fibre bars per tendon are
provided with sensors and there are to be four additional optical fibre sensors integrated
directly into the concrete on the upper and four on the lower side of the slab. The trial
loading of the Schiessbergstrasse Bridge which has been realized on 31st of March 1992,
using two trucks of a total load of 50 tonnes, showed in an impressive way the efficiency of
the optical fibre sensors embedded in the concrete construction. By connecting two highly
sensitive sensors for this trial load it was possible to measure additional elongations in the
-sector. These high-sensitivity sensors are normally only used for structural monitoring in
special cases, and were particularly applied for demonstration purposes in this instance.
The sensors, which are now permanently connected in the bridge, only register additional
extensions in the magnitude of 0,2 mm, thus enabling the crack behaviour of the concrete to
be monitored.
The bridge has a highly sophisticated permanent monitoring system for the bridge structure,
with the possibility of on-line diagnosis.
The main benefits are:
prestressing with intelligent glass fibre tendons monitored by optical fibres
crack control of the concrete by optical fibre sensors
embedded directly in the concrete
integrated chemical sensors in the concrete for
measuring the carbonation depth and chloride penetration
all data checkable from a personal computer via a
telephone line from the bridge to the client's office.

Fig. 30: Sensor layout

 329

Longitudinal section

|
rB
f hWiyfeJ^Wil 111J1111II111 It! 111111II111II1111111111111111111111111111 M 111
rA
nihiimiiiiiiii l l l l i l t l i l l l l i l l l l i l ;

C Lb
I ,'
1 1
LB
i -*
LA

L=> d
TJ

3
'

computer control

Cross section:

A-A B_J

Fig. 31: Trial loading

6.4 Ntsch Bridge, Karaten, Austria

The Ntsch Bridge got also a highly sophisticated permanent monitoring system with
the possibility of on-line diagnosis with the aid of telephone line to transfer the datas
directly to the client's office.
Similar to the Bridge Schiessbergstrasse the suitability of the sensors has been proved by a
trial loading. It was carried out on May 1992 using two trucks with a weight of 22 metric
tonnes each. As expected the integral additional elongation of approxemately 50 /xm in the
second span corresponded to a deflection of approxemately 2-3 mm.
330

4 optical fiber sensors for

4 optical fiber sensors for

monitoring of cracks tendon integrated with
optical fiber sensors
longitudinal section
11.00 >
chemical sensors
4 optical fiber sensors for (corrosion cell)
monitoring of cracks

tendon integrated with

optical fiber sensors
4 optical fiber sensors for
chemical sensors
monitoring of cracks
(corrosion cell)
cross section

Fig. 32: Sensor layout

extension in //m

sensor section 2
/ V
sensor pier 1 /
10
o.o \/y \v.
-10 measuring -
-20 time / sec.
-30

Fig. 33: Trial loading

 331

7. CONCLUSION

The first applications, in particular the Bridges of the Ulenbergstrasse, Marienfelde,

Schiessbergstrasse and Ntsch are only the beginning of a whole series of other possible
applications. On the 7 th of December 1992 the "Institut fr Bautechnik" in Berlin granted
the General Approval for the glass fibre composite prestressing system, which is now the
first nonmetallic system worldwide based on fibre composite material. The scope of
application includes all types and degrees of prestressing, with or without bond, temporary
or permanent soil or rock anchors, as well as tension cables for transmission masts or
bracings of any kind. Work is being done to continue to develop fibre composite materials
into an intelligent high-tech product with tensile strength and media resistance. The
possibility of monitoring the integrity of prestressed concrete structures and earth anchor
systems is also being researched. This development will give prestressed concrete design a
decisive impulse in the future, as concrete structures could have integrated sensitive nerve
fibres just like organic bodies and be more durable. Any change in their stress-strain
behaviour can be immediately detected at any location within the structure. A
comprehensive control of concrete structures will considerably reduce the maintenance costs
in the long run, because even minimal changes in the load bearing bahaviour are recorded
and the necessary repair and reconstruction works can be undertaken at a very early stage.

I N S T I T U T FR BAUTECHNIK
Anstalt des ffentlichen Rechts

1000 B e r l i n 30, 7 . Dezember 1992

R e i c h p l e t s c h u f e r 74-76
T e l e f o n : (030) 264 87-266
T e l e t e x : 308258
T e l e f a x : (030) 264 87-320
GeschZ.: I 1-1.13.1-67

ZULASSUNGSBESCHEID

Der

Zu1assungsqeqenstand: Spannverfahren HLV

wird hiermit allgemein bauaufs1chtl1ch/baurechtl1ch zugelassen.

Antragsteller: Strabag Bau-AG SIC0M, Gesellschaft

Siegburger Str. 241 fr Sensor- und Vor-
5000 Kln 21 spanntechnik mbH
Gremberger Str. 151a
5000 Kln 91
Bayer AG
5090 Leverkusen

Geltungsdauer bis: 30. November 1997

Zulassungsnummer:

Der zugelassene Gegenstand darf nur verwendet werden, wenn seine Herstellung
berwacht ist und dies am Verwendungsort geprft werden kann.

Dieser Zulassungsbescheid umfat 23 Seiten und 12 Anlagen.

Fig. 34: General approval

332

8. REFERENCES

[I] Miesseler, H.-J.; Preis, L.: Heavy-Duty Composite Bars Made of Glass Fibre as
Reinforcement in Concrete and Foundation Constructions. In: Bauen mit
Kunststoffen. Vol. 2/1988. Pages 4-14
[2] Rehm, G.; Franke, L.: Synthetic Resin Bonded Glass Fibre Bars as Reinforcement
in Concrete Construction. In: Die Bautechnik. 1974. No. 4. Pages 115-120
[3] Rehm, G.; Franke, L.; Patzak, M.: Investigations Regarding the Question of
Introduction of Forces in the Synthetic Resin Bonded Glass Fibre Bars. Issue 304.
The German Committee for Reinforced Concrete - Berlin:
Published by Wilhelm Ernst & Son 1979
[4] Waaser, E.; Wolff, R.: A New Material for Prestressed Concrete, HLV -Heavy
Duty Composite Bars Comprising Glass Fibres. In: "Beton" 36. 1986. Issue 7.
Pages 245-250
[5] Vollrath, F.; Miesseler, H.-J.: Glass Fibre Prestressing for Concrete Bridges -
Experience Gained During Construction of the Ulenbergstrasse Bridge.
In: TIEFBAU - BG 4/1987, Pages 206-211
[6] Franz, A.; Miesseler, H.-J.: Bridge at Berlin-Marienfelde- A Research Project -
External Prestressing and Permanent Monotoring by Means of Integrated
Sensors. In: Paper at the German Concrete Congress. Hamburg. April 1989.
[7] Wolff, R.; Miesseler, H.-J.: Application and Experience with
Intelligent Prestressing Systems Based on Fibre Composite Materials. In: Paper at
the XI. FIP-Congress, Hamburg, June 1990. Pages R63-R67
[8] Miesseler, H.-J.; Levacher, K.: Monotoring Stressing Behaviour with Integrated
Optical Sensors. 13th IABSE Congress. Helsinki. June 1988. (German)
[9] Miesseler, H.-J.; Lessing, R.: Monotoring of Load Bearing Structures with Optical
Fibre Sensors. IABSE. Lissbon 1989. (English)
[10] Miesseler, H.-J.; Wolff, R.: Experience with the Monotoring of Structures Using
Optical Fibre Sensors. In: Paper at the XI. FIP- Congress, Hamburg. June 1990.
Pages Q12-Q17. (English)
[II] Faoro, M.: The loadbearing behaviour of resin bonded glass fibre tendons in the
field of end-anchoring and fractures in concrete. University of Stuttgart,
Institute for materials in the field of civil engineering (Institut fr Werkstoffe im
Bauwesen). Stuttgart: University of Stuttgart, dissertation 1988.
IWB-news 1988/1
[12] Schiessl, P.; Raupach,M.: Chloride-induced corrosion of steel in concrete-
investigations with a concrete corrosion cell. Institute for Building Research,
Technical University of Aachen. The life of structures 24 to 26 April 1989,
Brigthon, England.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 333
1993 Elsevier Science Publishers B.V. All rights reserved.

Parafil ropes for prestressing applications

CJ.Burgoyne
Engineering Department, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ,
United Kingdom.

Abstract
Parafil ropes, which contain a core of parallel filaments of aramid yarns within a polymeric
sheath, have been used for prestressing concrete and for a variety of other structural
applications. The construction, properties and methods of using the ropes are described, as are
the results of tests of beams prestressed with the ropes. A number of real applications of the
ropes are also given.
1. INTRODUCTION
Parafil ropes have several features which distinguish them from most other prestressing
systems; they cannot be bonded to concrete; they contain no resin, and they were not initially
developed for prestressing. Nevertheless, they have been used for prestressing concrete on a
number of occasions, and have recently been adopted by one of the largest manufacturers of
prestressing systems (VSL) as an alternative to steel tendons when corrosion is likely to be a
problem.

2. DESCRIPTION
Parafil ropes are manufactured by Linear Composites Ltd in Yorkshire, England. They
contain a core of parallelfilamentsof a high strength yarn within a polymeric sheath. A variety
of core yarns are used, the most common being polyester (known as Type A), Kevlar 29 (Type
F) and Kevlar 49 (Type G). Kevlar was the first of the aramid fibres to be developed, by El
Du Pont de Nemours, in 1973. The basic properties of ropes manufactured with these yarns
are shown in Table 1. Those of primary interest to prestressing engineers are the Type G
ropes, which have the highest stiffness and lowest creep properties, although the lower
modulus versions could be used in applications where the prestressed structure itself tended to
creep under the influence of the prestress. In this case, the lower modulus of the fibres would
require larger jack extensions at the time of prestressing, but would mean less loss of prestress
due to creep.

Table 1
Tensile properties of Parafil Ropes
Designation Material Strength Stiffness
N/mm2 kN/mm2
Parafil Type A Polyester 617 12.0
Parafil Type F Kevlar 29 1926 77.7
Parafil Type G Kevlar 49 1926 126.5
(Manufacturer's data)
334

Other fibres could also be used, including the alternative aramid fibres such as Technora
(made by Teijin), or possibly Vectran (made by Hoechst). Both claim to have certain properties
better than Kevlar, and there is no reason why they could not be used in these ropes.

3. TERMINATION SYSTEM
The most important component of any system carrying tension is the anchorage, where the
forces are transmitted to the rope. In Parafil, the ropes are anchored by means of a barrel and
spike fitting, which grips the fibres in an annulus between a central tapered spike and an
external matching barrel (Figure 1).
To attach the termination, the end of the rope is passed through the terminal body, and the
sheath is removed over the length of the spike; the yarns are then spread out evenly around the
terminal body before the spike is introduced. The rope is drawn back into the terminal and the
rope pretensioned to a load in excess of that to be applied in practice. During pretensioning, the
spike is drawn fully into the termination, applying an outward force on the fibre pad and
gripping the yarns. Subsequent changes in force cause only tiny movements in the spike and
can be ignored for practical purposes. For prestressing operations, where the largest load
applied to the rope is the act of prestressing, the pretensioning of the rope can take place at the
same time as the prestress. Allowance then has to be made for the bed-down of the spike to
ensure that the rope is the correct length, but this is easily done.

Figure 1. Barrel and spike

termination for Parafil.

Figure 2. Gripping forces

within termination.

This system has a number of advantages over wedge systems which grip the outside of a
tension member.
1. The gripping force between the spike and the barrel has to pass through every fibre
(Figure 2), which means that each fibre can develop an equal friction force against its
neighbours or the fitting. Thus, there is no tendency for some of the fibres to carry a
 335

disproportionate amount of the load, which would cause early failure of those fibres, and hence
the rope. Systems which rely on external wedges have a tendency to develop hoop
compression around the outside of the tension member, leaving the inner fibres less well
gripped.
2. There is no resin in the system, which means that the effectiveness of the termination is
not affected by temperature or creep.
3. The system is easy to fit, on site if necessary, simply by removing the sheath and
splaying out the fibres. If possible a pretension in excess of that expected in the service life of
the ropes should be applied.
4. There are no size effects; terminations for large ropes are linearly scaled versions of the
terminations for small ropes. The mechanics of operation remain the same.
5. The terminations can develop the full strength of the parent rope; when used for tension
tests, the rope breaks away from the termination.

4. DEVELOPMENT OF ROPES
The ropes were first developed in the early 1960s to meet a requirement for mooring
navigation platforms in the North Atlantic. These would have required mooring lines several
kilometres long, and the weight of steel ropes would have been prohibitive. At the same time,
accuracy of position of these platforms meant that the lines had to be stiff and to have good
axial fatigue performance. Conventional structured ropes, where the individual fibres follow
tortuous paths along the rope, could not be used since they lose a significant proportion of the
fibre's inherent stiffness. The large number of points where fibres cross also causes a loss of
fatigue strength.
Linear Composites Ltd (then part of Imperial Chemical Industries) developed the idea of
keeping the fibres straight, and giving the rope some structure (normally provided by braiding
or twisting) by enclosing the fibres in an extruded sheath. The fibres used at that time were
polyester, the aramids not yet being available.

Figure 3. Typical mooring

application, in 6000m water depth,
showing Parafil combined with
floats and other equipment.

~) SINKER
/
/ // / ' ' / / / // ' / ' / / / / /
336

In the event, the requirement for aircraft navigation systems was met by satellites, so the
North Atlantic platforms were never needed, but methods of producing the ropes and their
properties were well established, and this led to their adoption in a variety of applications. The
earliest of these were as guys for radio antennae where the non-conducting nature of the ropes
did away with the need for conventional insulators. The first installations used polyester ropes,
but since the development of aramid fibres, and with the communications industry using arrays
of masts which have to be placed accurately in relation to one another, the suffer aramid ropes
are now being used more extensively.
Moorings for floating systems, such as buoys, have been extensively used (Figure 3).
These often have floats, weights, and other equipment attached at the top and bottom, with the
bulk of the length of the mooring being provided by Parafil.
Other early uses of Parafil as replacements for steel wire followed. Such uses have included
standing rigging in ships, where the smooth sheath has been found to provide the added
advantage that ice can easily be shaken free; supports for overhead wires in trolley-bus systems
(again making use of the electrical insulation), and in safety rails around the deck of ships.
Many of these systems are in use by military authorities around the world.
In the 1970s Kevlar, the first of the aramid fibres, became available. Experiments showed
that the techniques used for making ropes from polyester could easily be adopted for aramids,
resulting in a stronger and suffer rope. The strength of the rope is about 20% higher than a
normal prestressing steel, while the stiffness is about 2/3 that of steel. These properties make
the ropes very attractive as structural elements in their own right, and a programme of research
was undertaken to give practising engineers confidence in both the short and long term
properties of the ropes.
Prestressing tendons for concrete were soon identified as a very suitable application. These
tendons are the most heavily stressed elements in normal use; no other structural component is
regularly loaded to a permanent force of 70% of its break load. For this use, there are clear
advantages in using the suffer Type G Parafil, incorporating Kevlar 49, and in all that follows it
is this type of Parafil that is being considered, unless otherwise stated.

5. TESTING
Much of the early testing on Kevlar concentrated on the short term properties of the fibre,
but for structural engineering applications, the long term properties are just as important. A lot
of testing has thus been done, on the ropes rather than the fibres, to establish their properties.
Strength and size effects.
The stress-strain curve of the ropes (Figure 4) matches quite closely that of the constituent
fibres. The Young's modulus is about 120 kN/mm2, and the strength is about 1930 N/mm2.
There is a slight stiffening at about 1000 N/mm2; that is a property of the fibre and is not
significant in most cases. Once the rope has been fitted with terminals, and these have been
bedded down properly in accordance with the manufacturer's instructions, the ropes have their
full stiffness from zero load.
When a number of fibres are used together, it is not possible to use the full strength of all the
fibres, or even to achieve the average strength of the fibres. This is because the weaker fibres
fail at a lower load than the stronger ones, leaving the total load-carrying capacity reduced.
This process has to be applied twice in Parafil ropes; they are made as a bundle of parallel
yarns, which are in turn made from about 1000 individual filaments. The filaments themselves
have strengths of about 3500 N/mm2; the yarns have a strength of about 2900 N/mm2, and the
ropes have a minimum strength of about 1930 N/mm2.
These effects are described by bundle theory, which accounts for these size effects. It is
also possible to account for length effects in a similar way, by using weakest link theory. Both
 337

Figure 4. Typical stress-strain curve from tests

on 60 Tonne Type G Parafil ropes.
2200

2000 ./+
1800

_1600
/E=121kN/mm2
1400
\
-1200

2 1000
To

800

600 /
E=111kN/mm2
400

200

0 0.2 0 0-6 0.8 1.0 1.2 U 1.6 1-8

strain (/o)

theories rely on the variability of strength of yarns and fibres and predict that the strength will
reach an asymptotic value as the rope size gets bigger, and also as it gets longer. Figure 5
shows the variation in strength with rope size as measured in Type G Parafil ropes. Other tests
have been carried out on much larger ropes (up to 1500 tonnes break load) and these lie very
close to the asymptotic value. Thus, for all practical rope sizes and lengths used in
prestressing, the strength of the ropes can be taken as 1930 N/mm2, measured over the cross
sectional area of the yarns.

Figure 5. Measured strengths of ropes of

different sizes, showing asymptotic
strength for large ropes.
338

Creep, relaxation and loss of prestress

Aramid fibres offer significantly lower creep than most other fibres used in rope making;
indeed, for many rope applications the creep is negligible. But when used for prestressing
concrete, engineers are interested in the creep as a proportion of the initial extension, since this
governs the amount of prestressing force lost.
Total creep strains are of the order of 0.13%, which can be compared with a rope extension
(when stressed to about 50% of its initial break load), of about 0.8%. Thus, we can expect to
lose something like 16% of the initial prestress force in a Parafil tendon.
Figure 6 shows predicted stress-relaxation figures for different ages and for different initial
stresses (expressed as percentages of the nominal break load NBL), based on a series of rope
tests. It will be seen that they agree quite closely with the limiting value given above.

^ /
c
o

x --- Initial load (f) 70% NBL
Initial load (f) 50% NBL
( Initial load ( f ) 30% NBL
cr
30 days 1 year 10 years 100 years
0 1 L I1
i ? 7 n n
n
10 2 10 4 10 s 10 e
10 3 2 5
log scale
Time (t) hours

Figure 6. Stress relaxation predictions for Type G Parafil, from different levels of initial
prestress, expressed as a percentage of the nominal break load (NBL) (Chambers data).
The total loss of prestress force in a member prestressed with Parafil is very similar to that in
a beam prestressed with steel. Although the losses due to the relaxation of the tendon are
higher, those due to the elastic shortening of the concrete are lower, due to the lower elastic
modulus of the Kevlar yarns. The amount depends on details of the design, which will differ
for structures designed with steel or Parafil tendons, but for most cases these two effects cancel
one another fairly closely.
Stress rupture
Stress rupture, or creep rupture as it is sometimes known, is the name associated with failure
caused by a material creeping until it breaks. This is not normally a problem in steels, except at
 339

1.0

en
c

X Tests on 1.5 tonne ropes

+ Tests on 3 tonne ropes
Tests on 60 tonne ropes
95% prediction
Mean prediction
L_ . 5% prediction

days years
100 100
10
I 1
10'
i2
10 5
101
IT
:
10 10 3 4
10 ~ 10 e
10 e 2

Lifetime (hours) log i

Figure 7. Sfress rupture test results and predicted lifetimes, based on tests by Guimaraes and
Chambers.
high stresses or high temperatures, but it is likely to be a governing criterion for the long term
use of most systems that rely on new materials.
Stress rupture is clearly related to creep and relaxation. At higher stresses, materials creep
more, and fail in a shorter period of time than at low stresses. There are strong theoretical
arguments, related to the activation energy of the creep process, why there should be a linear
relationship between the applied stress and the logarithm of the lifetime of the material. This is
indeed observed in tests on both Parafil ropes and on Kevlar yarns. Figure 7 shows values of
lifetimes as measured in tests on Parafil ropes, and compares them with theoretical predictions
based on tests performed on Kevlar 49 and epoxy bars. The results have been normalised with
respect to the short term strength, because of the bundle theory effects described above.
Statistical analyses have been carried out on this data, and it is predicted that a rope loaded to
50% of its short term strength will have a 1.4% chance of failing if the load is maintained
continuously for a period of 100 years. This prediction is based on extrapolation of tests
carried out at ambient temperature for periods of about 4 years, and from consideration of tests
caried out for shorter periods at elevated temperatures (which can be related to those at ambient
temperature via the activation energy). At the moment, tests are underway with ropes loaded by
dead weights to produce failures in ropes after periods in the 5-10 year range. These will give
engineers more confidence in the extrapolation up to structural lifetimes.
Work is currently underway identifying the cumulative damage rule that must be applied if a
rope is subject to varying loads. The most likely rule appears to be one where the rope sustains
stress rupture damage as a linear proportion of the lifetime that it spends at that load. This
allows the stress rupture lifetime to be calculated where the load is reducing as the prestress
force drops off because of concrete creep and tendon relaxation.
One point needs to be made about these results. The stress rupture lifetime relates to loads
applied continuously; it does not mean that the short term strength is reduced by the same
extent. The strength retention observed in a rope that has been subjected to a load for half of its
340

stress rupture lifetime would be virtually unchanged from the short term strength. This has
important implications for prestressing concrete. The initial prestressing force can be chosen on
the basis of the long term stress rupture of the tendon, taking due account of the relatively short
period of time the rope spends at a higher force before creep and relaxation have occurred. But
the force in the tendon then changes very little due to live load effects, other than very
occasional excursions when the structure is overloaded. On these occasions, the tendon will
still have virtually its full strength.
These ideas can all be combined. Figure 8 shows allowable initial prestressing forces in a
beam to give a 10-6 chance of failure due to stress rupture for a given design life. Curve D
shows the maximum allowable force for a constant load, but curves A, B and C show the
situation for minimum, typical and maximum prestress losses respectively. For the case of
typical losses, an initial prestressing force that is about 10% higher can be allowed, since the
force will subsequently reduce, thus reducing the stress rupture damage that is taking place.

10 20 30 40 50 60 70 80 90 100
structural design life (years )

Figure 8. Suggested initial prestress, based on assumed losses of force due to stress relaxation
and creep of concrete, to give a 106 probability of failure due to stress-rupture (based on
Chambers test results).

6. STRUCTURAL APPLICATIONS OTHER THAN PRESTRESSING

Because the ropes can be made in almost any size, and with efficient terminals at the ends, it
is possible to use them in a variety of ways. A number of novel applications have been made,
where the various properties of Parafil have contributed significantly to the success of the
scheme.
 341

Bicentennial tent
During the Australian Bicentennial celebrations in 1988, a touring exhibition was mounted,
which consisted of a series of tented exhibition stands mounted on trucks (Figure 9). These
were moved around the country and erected many times at different locations. Parafil ropes
were used as the main supporting cables for the tents, and also as the tensioning elements
around the edges of the tents. They were lighter than the equivalent steel cables, and stood up
well to the rigours of repeated assembly and disassembly.

Figure 9. Tent for Australian bicentenniel touring exhibition.

Tank Bridge

Parafil
CaHt

Figure 10. Scissoring action of tank-launched bridge, controlled by Type G Parafil (Defence
Research Agency)
The British Army uses an armoured deployable bridge system, mounted on a tank chassis.
This uses a rope to deploy the bridge in a scissoring action (Figure 10); the available lever arm
is small, so the forces that have to be carried are high. The rope must be stiff as well as strong,
342

as it controls the accuracy with which the bridge can be placed. Conventional steel rope is very
heavy, and is awkward to both carry and stow. To overcome these problems the Defence
Research Agency carried out tests using Parafil ropes. They are much easier to handle and are
just as effective in operating the bridge; future versions of the bridge are likely to incorporate
Parafil scissoring ropes.
Bus Station Roof, Cambridge
A small bus station was completed in Cambridge in 1991; this has a roof supported by
Parafil ropes with a 7 m cantilever (Figure 11). There are four masts, each supporting a pair of
forestays and a pair of backstays. Although designed primarily to resist snow loading, the
stays are permanently stressed to ensure that the roof remains stiff even when wind loads cause
uplift. The structure was designed by Cambridgeshire County Council, and the ropes were
fitted with terminals and preloaded in the Engineering Laboratories at the University of
Cambridge.

Figure 11. Cable stayed roof on bus station at Cambridge (Cambridgeshire County Council).
Aberfeldy Bridge, Scotland
The western world's first all-plastic bridge has just been completed at Aberfeldy in Scotland.
It combines a deck and towers made from lightweight glass reinforced plastic pultrusions
(developed by Maunsell Structural Plastics Ltd) with stay cables made from Parafil. The
bridge carries a footpath linking two halves of a golf course across the River Tay, with a clear
span of 64m (Figure 12). The only non-plastic components are concrete in the foundation and
 343

some connecting pieces between the deck and the cable terminations to distribute the
concentrated load.
The bridge was built by students from Dundee University; the pultrusions were assembled
on scaffolding on shore, and then launched across the river, supported on a cats-cradle of
cables made up from the permanent Parafil stay cables and some temporary cables.

Figure 12. Aberfeldy bridge, Scotland. The deck and towers are glass-reinforced plastic; the
cables are Type G Parafil ropes. (Maunsell Structural Plastics)
7. PRESTRESSING APPLICATIONS
Thorpe Marsh Power Station.
Thorpe Marsh electricity generating station in the north of England was one of a series built
in the 1960s to use coal mined locally. It has six large cooling towers, of which three were
recently found to have large cracks at the top; this left them in a very unstable condition.
Demolition would have kept the station out of service for a considerable period, but it was
decided that the towers could be repaired by circumferential prestressing after injecting the
cracks with resin; Parafil ropes were used for this application (Figure 13). The prime benefits
were the resistance to corrosion and the light weight, which meant that the prestressing could be
344

carried out by steeplejacks carrying coils of cable up the towers. They could work their way
round the towers, installing the cables as they went, before stressing the cables one-by-one.

Figure 13. Cooling towers at Thorpe Marsh electricity generating station, Doncaster,
circumferentially prestressed with Parafil ropes.
Beam tests at Imperial College
Tests have been carried out on two beams prestressed with Parafil to demonstrate the
feasibility of producing structural elements in this way. Two designs were produced; the first
had a single, straight unbonded tendon, contained within a duct on the centreline of a simple I-
beam, while the second had two external deflected tendons, one on each side of a T-shaped
cross-section (Figure 14). In each case the tendons were Type G Parafil with a nominal break
load of 60 tonnes, prestressed to about 50% of their short term strength.
For an internal tendon the terminals have to be fitted before the rope is placed in position in
the beam; since the terminals are too large to pass through the duct, this is built up around the
tendons. In the second beam, the tendons were to be placed outside the concrete, so there was
no need to assemble the rope in a duct prior to casting. Holes were formed in the thickened end
blocks to receive the rope terminations, by casting-in plastic pipes.
The principles of the stressing procedure are shown in Figure 15. The tendon is placed in
the structure, and a pull-rod fitted to the internal thread of the termination. The pull-rod is then
passed through the centre hole of a hydraulic jack, and secured by means of a nut. The jack is
held away from the beam by means of a trestle, which allows access to the terminal to secure
the back-nut. Force is applied by the jack, which brings the terminal just outside the face of the
concrete; the back-nut can then be fitted to lock the tendon in position in its stressed state. The
jack, trestle and pull-rod are removed, and a security cap fitted to prevent dirt and debris getting
 345

into the termination. This would also serve to contain the anchorage in the unlikely event of a
rope failure.

Figure 14. 8m long beam, prestressed externally with two 60 tonne Type G Parafil ropes.
Measurements of the forces in the second beam showed that the coefficient of friction was
about 0.32, which is slightly higher than would be expected with steel tendons, but could be
brought down by a better selection of sheath and deflector material. Measurement of the force
in the tendons, after the force had been transferred from the jack to the permanent back-nut,
indicated that no loss of prestress occurred at this stage.

Centre
10 *l 0' hole
> A y
-': .'?:'?.'-'.- .q'-.V ?'-, I'.-V. ?.'-6:'*''* 1 iack\
- Terminal
H 1 &
..'..'.'. i>'.'0'' 3.'
r
J-Pullrod
a) Tendon installation b) Jack attachment

Trestle
r&.fiy .$ - . ?v Tendon
03
. ''f-p:^" *'-V.'
Backnut

c) Prestressing force application d) Final arrangement

Figure 15. Stressing procedure for Parafil ropes.

346

Load deflection behaviour

Both beams were tested in four point bending rigs with loads being applied by hydraulic
jacks. The beams were taken through several elastic loading cycles; the second beam was kept
under sustained load for 42 days to monitor the effects of creep and relaxation.
The relationship between the applied load and the deflection at the centre of the second beam
is shown in Figure 16. On the application of the load, the response is almost linear, with the
portions of the curves corresponding to loading and unloading being parallel. The
instantaneous camber produced by the prestressing is indicated by the horizontal part of the
curve at zero load. The increase of deflection due to the effects of shrinkage and creep of
concrete after 42 days was 59% of the instantaneous deflection caused by the applied load.
This figure is not affected by relaxation of the tendon, the increased deflection being due to loss
of stiffness of the concrete.

200

180 Upworda Downwords time - 42

|
days

160

140

120
Load cycle 1

fw
100 Load cycle 2
I
..I

o
0 /
60

40

20 - Prestressing _

-<1 J I mfri _,.-^~^:j

u
J:
rf
, 1
' -6 -4 2 ' S z 4 6 8
Deflection (mm)

Figure 16. Applied load versus mid-span deflection for the second (8m )beam, showing
camber due to prestress, elastic response due to load, and extra deflection due to creep of
concrete.
The total loss of prestress in both tendons of the second beam, due to shrinkage and creep of
concrete and due to stress relaxation in the tendons, is shown in Figure 17. The tendons were
tensioned initially to approximately 22% of their tensile strength, when the age of concrete was
10 days. Losses of 13% and 14% of the initial force were observed in the two tendons after 23
days, when the full prestressing force was applied. Over this period of time the beam was
subjected only to its own weight. 43 days after the application of the full prestressing, the
losses of prestress under service load were 12% in tendon 1 and 11% in tendon 2. It can be
seen in the figure that most of the losses occurred within the first day after prestressing. From
then on the curves show a very low rate of loss. These figures are very similar to losses to be
expected in steel tendons, in accordance with the comments made earlier.
 347

Ultimate load tests were carried out on both beams, which responded as expected. After
passing the cracking load, the stiffness reduced considerably; when unloaded from the cracked
(but still elastic) state, the stiffness remained lower until the cracks had closed up but the full
elastic stiffness was recovered and there was virtually no permanent set.

PD? = 306 kN
P = 297 kN
Tendon 2
808((!)()000 80 0^^ 0 0 0 ( ( ! ) ( 1 ) ) 8) | 1fl Q QQ p^
Tendon 1

- Loading, t 24 days
End of load cycle 1

o P.', = 137 kN
c Partial prestressing, t =0
0.86 PJ2
o.87 p;, Full prestressing. t = 23 days

30 40
Time (days)

Figure 17. Variation of prestressing force with time in 8 m beam.

When loaded until failure, both beams showed considerable curvature at virtually constant
load, with large cracks forming in the bottom of the beam. Failure occurred in both beams by
crushing of the top flange. Figure 18 shows the load deflection curves for the second beam; the
results for the first are similar.
There were slight differences in the final failure mode of the two beams which cast important
light on the behaviour of unbonded and external tendons. In both cases, the top flange failed
by crushing, but in the first beam, as the tendon was constrained in the bottom flange, the beam
did not completely collapse. The compression zone passed down through the web, and into the
top of the bottom flange, with a consequent reduction in load. However, the bottom flange did
not fail, remaining axially prestressed. After the test, the tendon was found still to be carrying a
significant force.
In the second beam, the tendon was outside the bottom flange, which could thus deflect
while leaving the tendon in its original position relative to the ends of the beam. The beam thus
failed suddenly and completely, with a total loss of prestress.
Reasons for not bonding to concrete
Aramids, like carbon fibres and glasses, exhibit brittle behaviour when tensile loads are
applied to them, which means that they are very sensitive to applied strains which exceed the
design values. This is in contrast to steel, where the plateau on the stress-strain curve means
that the tendon can absorb high strains locally with no significant problems other than a
permanent set.
348

300
"
270 - ^ ^=
240

Cracking y^r
210
/
y y i / /

/ CD
O
Load cycle 3
Load cycle 4
Load cycle 5

\
1 L 1 i
40 60 60
Deflection (mm)

Figure 18. Load-deflection curve for 8m beam, when loaded to failure. Note the plateau
corresponding to crack opening.

Load cycle 3
<D Load cycle 4
Load cycle 5

Tendon force (kN)

Figure 19. Change in tendon force in 8m beam when loaded to failure. Note the relatively
small change in force as the tendon was unbonded.
 349

We must therefore be very careful when deciding whether to bond these materials to
concrete. In the vicinity of cracks, the local strains are very high; indeed if we have perfect
bond between steel and concrete, they are infinite. Thus we might expect that beams
prestressed with any of these materials, if they are bonded to the concrete, will fail by local
snapping of the tendon, with no possibility of redistribution of load, or of plastic deformation
to accommodate the strain.
Thus, beams prestressed with new materials should be designed with unbonded tendons.
This has some implications for design procedures, as there will be relatively little increase in
tendon force as the beam is loaded.
This effect was observed in the beam tests; Figure 19 shows the tendon force during the
ultimate load cycles on the second beam, from which it can be seen that fairly small changes in
tendon force occurred, even though the concrete was significantly cracked, because the tendon
could slide relative to the concrete. There is thus very little chance of snapping the tendon. The
change in resistance to external bending moment is almost exclusively due to an increase of the
lever arm between the internal compression and tension forces whose magnitudes remain
relatively unchanged.
Parafil ropes, with their polyethylene sheath, cannot be bonded to concrete. Even if they are
cast in place, there will be slip between the tendon and the sheath and creep of the sheath itself.

8. LINK WITH VSL INTERNATIONAL

Parafil ropes have now moved from the development and prototype stage to fully fledged
prestressing tendons with an agreement between Linear Composites Ltd and the VSL
International group, who are one of the world's leading suppliers of prestressing systems.
Engineers will now be able to design structures with Parafil prestressing, in the knowledge that
jacking systems and specialist assistance will be available at the time of installation. The
jacking system used will follow the principles described above with special fittings to allow the
use of existing VSL jacks.
VSL have recently introduced polyethylene duct systems for conventional prestressing
tendons; combining these with Parafil will completely remove the danger of corrosion of the
prestressing system.

9. PREDICTIONS FOR FUTURE PRESTRESSING SYSTEMS

It is expected that beams, both in bridges and buildings, will be prestressed with Parafil.
The tendons will either lie outside the concrete, or unbonded in ducts within the concrete. No
account will be taken of increased forces in the tendon due to live load.
The results of the Imperial College tests show that basic design principles for prestressed
concrete do not need altering radically; the following are points which a designer should take
into account when designing a beam with Parafil tendons.
1. The tendon should be pretensioned, with the terminals in place, to a load level in excess
of that expected during both the initial stressing operation, or the service life of the structure.
This will have the effect of ensuring that the terminal spike is properly bedded, and will also
give a check on the tendon length before being placed in the structure.
It is normal practice, according to the manufacturer's instructions, to pretension ropes to
60% of the nominal breaking load prior to use, whenever possible. These ropes, when used in
conventional rigging arrangements, are normally stressed to much lower load levels than those
in use in prestressing tendons; in these cases, 60% is perfectly adequate as a pretensioning
load. However, in prestressing tendons, where high force levels are normal, a higher
pretensioning level may be needed to ensure adequate bedding of the termination.
350

2. Any deflector points should be properly flared to ensure no damage to the sheath during
stressing operations; this should not be difficult to arrange if taken into account at the design
stage.
3. The coefficient of friction between the tendon and the duct (or the deflector) should be
reduced wherever possible. This may mean undertaking some studies of friction coefficients
between various possible sheathing materials and alternative duct materials. Alternatively,
coating materials, such as PTFE or nylon tapes, might be considered.
4. The working load design of prestressed concrete beams should be based on allowable
stress limits taking account of the design prestressing force, after allowing for losses, and the
ultimate strength of the section should be based on the assumption that only minimal increases
of force take place due to geometry changes as the beam deflects.
5. The compression zone of the concrete should be provided with confining reinforcement
to increase the ductility of the concrete in that area.
6. If the tendons are external to the concrete, they should pass through loose rings so that,
in the event of failure, the tendons are forced to deflect with the beam. This will ensure that
failure occurs in the more controlled manner of the first beam.

10. CONCLUSION
It is clear that Parafil will start to find more widespread use as a non-corroding prestressing
tendon. Repair of structures by the use of external tendons will become more common, and the
use of unbonded, replaceable tendons is likely to become the norm for all structures in the near
future. Structures, such as water towers, which often have poorly protected steel prestressing
and a high incidence of corrosion, are currently being studied with a view to their repair with
external Parafil tendons. A number of bridges with suspect prestressing tendons are also being
identified by the current bridge assessment programmes; these would make useful
demonstration sites for new materials as the new tendons would be adding an extra margin of
safety, rather than providing the primary stressing.
Offshore, as exploration for oil and other minerals moves into ever deeper water, the
arguments for using mooring lines with almost neutral buoyancy become more persuasive.
Structures can be moored in 300 m of water using steel, but not in 3000 m. Virtually all the
major oil companies have conducted studies into the use of lightweight mooring lines; when
economics dictate that such structures be built, Parafil ropes, or similar systems, will
undoubtedly be used.
Similarly, as bridge spans increase, the use of lightweight stiff materials becomes more
economic. The excellent fatigue behaviour will also be seen to be important. The Eurobridge
proposal to cross the English Channel with seven spans of 4.5 km was probably 20 years
ahead of its time, and had some conceptual flaws. Nevertheless, such large spans are only
going to be possible if new materials are used.
Other applications will make use of the non-magnetic nature of the material; applications
such as de-Gaussing facilities for ships, or as strength elements in members carrying important
communications, (such as railway signalling and control equipment), can also be envisaged.

11. BIBLIOGRAPHY
The following bibliography is not exhaustive, but includes major works relating to the
properties of Parafil, and all works from which data given here has been drawn.
1. Kingston D., Development of parallel fibre tensile members, Symposium on
Engineering Applications of Parafil Ropes, Imperial College, London, 1988.
 351

2. Chambers J J., Parallel-lay aramid ropes for use as tendons in prestressed concrete,
Ph.D Thesis, University of London, 1986.
3. Guimaraes G.B., Parallel-lay aramid ropes for use in structural engineering, Ph.D
Thesis, University of London, 1988.
4. Burgoyne CJ. and Chambers J.J., Prestressing with Parafil Tendons, Concrete,
19/10, 12-16, 1985.
5. Burgoyne C.J., Structural uses of polyaramid ropes, Construction and Building
Materials, 1, 3-13, 1987.
6. Chambers JJ. and Burgoyne C.J., An experimental investigation of the stress-rupture
behaviour of a parallel-lay aramid rope, Journ. Mat. Sei., 25, 3723-3730,1990.
7. Burgoyne C.J., Properties of polyaramid ropes and implications for their use as
external prestressing tendons, External Prestressing in Bridges, ed. A.E.Naaman and
J.E.Breen, American Concrete Institute, SP-120, 107-124, Detroit, 1990.
8. Hobbs R.E. & Burgoyne C.J. Bending fatigue in high-strength fibre ropes, Int. Journ
Fatigue, 13, 174-180, 1991.
9. Burgoyne C.J., Guimaraes G.B. and Chambers JJ. Tests on Beams Prestressed with
Unbonded Polyaramid Tendons, Cambridge Univ. Eng. Dept Tech Report CUED/D -
Struct/TR. 132, 1991
10. Guimaraes G.B. and Burgoyne C.J., The Creep Behaviour of a Parallel-lay Aramid
Rope, Journ. Mat. Sei., 27, 2473-2489, 1992.
11. Guimaraes G.B. and Burgoyne C.J., Repair of Concrete Bridges using Parafil Ropes,
U.S European Workshop - Rehabilitation of Bridges,
12. Burgoyne C.J., (Editor) Proceedings of Symposium on Engineering 1988.
13. Burgoyne C.J., Structural applications of Type G Parafil, Symposium on Engineering
Applications of Parafil Ropes, Imperial College, London, 1988.
14. Burgoyne C.J., Polyaramid ropes for tension structures, 1st Int. Oleg Kerensky
Memorial Conference on Tension Structures, London, 1988.
15. Burgoyne C.J., Laboratory testing of Parafil Ropes, Les Materiaux nouveaux pour la
precontrainte et le renforcement d'ouvrages d'art, LCPC Paris, 1988.
16. Burgoyne CJ. and Flory J.F., Length effects due to Yarn Variability in Parallel-lay
Ropes, MTS-90, Washington DC, 1990.
17. Snell M.B. and Taylor R.M., The use of Parafil Ropes in tank launched bridges,
RARDE Div. Note EE/2/89.
18. Dean B.K. and Wynhoven J.H., Australian Bicentennial travelling exhibition, 1st Int.
Oleg Kerensky Memorial Conference on Tension Structures, London, 1988.
19. Richmond B. and Head P.R., Alternative materials in long-span bridge structures, 1st
Int. Oleg Kerensky Memorial Conference on Tension Structures, London, 1988.
20. Burgoyne CJ., Tests on beams prestressed with polyaramid ropes, Procs 1st Int.
Conf. on Advanced Composite Materials in Bridges and Structures, Sherbrooke, Quebec, 1992
21. Amaniampong G. Variability and visco-elasticity of parallel-lay ropes, Ph.D. Thesis,
University of Cambridge, 1992.
22. Burgoyne C.J., Should FRP tendons be bonded to concrete? Int. Symp. on Non-
metallic reinforcement and prestressing, American Concrete Institute, 1993.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 355
1993 Elsevier Science Publishers B.V.

NEFMAC - Grid Type Reinforcement

Minoru Sugita
Technology Division, Shimizu Corporation
Seavans South, 1-2-3, Shibaura, Minato-ku, Tokyo 105-07, Japan

ABSTRACT
NEFMAC, grid type innovative reinforcement has been developed in Japan.
This chapter mentions about fundamental material properties, structural
characters and actual applications of grid type reinforcement NEFMAC.

1. INTRODUCTION
New Fiber Composite Material for Reinforced Concrete (NEFMAC) is
an innovative concrete reinforcement consisting of high strength continu-
ous fibers such as carbon fiber (CF), glass fiber (GF) and aramid fiber (AF).
All of them are impregnated with resin and formed into two or three dimen-
sional grid shapes.
NEFMAC is very lightweight, stronger than steel reinforcement, free
from rust or corrosion and highly resistant to salt.
The characteristics of NEFMAC are shown in Table 1. In addition, Photo 1
is a photograph showing examples of the form of NEFMAC and Photo 2
shows the intersection of a grid.

2. MANUFACTURING PROCESS [1]

NEFMAC is formed into flat, curved and three-dimensional grid shape
by the newly developed pin-winding process, which is a kind of filament
winding process. In a batch process, to form large cross-sectional or three-
dimensional grid shape, fibers are impregnated with resin by a peroxide
curing system, and formed into grid shape repeatedly at the room tempera-
ture as shown in Fig. 1. In this process it takes a few hours to cure. In case
of a continuous process to form a small cross-sectional grid shape, fibers are
impregnated with resin through an ultraviolet curing system, and formed
into flat grid shape as shown in Fig. 2. Since the line speed is 0.5 to 1.5m/
min., 60 to 180m2/hr of NEFMAC can be produced.

3. SURVEY OF DEVELOPMENT^]
Table 2 is a survey of NEFMAC research and development (R&D) at the
present. R&D was begun in 1984. In the beginning, fundamental R&D was
done independently and then joint R&D was done with universities to
clarify NEFMAC's fundamental properties. Later NEFMAC was applied to
356

an actual concrete structure in both independent and joint R&D.

As for fundamental R&D, we clarified NEFMAC's tensile property and
conducted studies that related to the spliced lap joint mechanism[3]. These
results proved that NEFMAC is effective as reinforcement for concrete
structures.
In the joint R&D done with universities, shear and compressive proper-
ties i.e. NEFMAC's fundamental mechanical characteristics were clarified.
Then the mechanism of anchorage to concrete was determined. It became
clear that the shear strength is about 50% of the tensile strength in case of
TFGF highbrid type[3]. Anchorage capacity larger t h a n the tensile
strength can be secured by developing NEFMAC as much as 2 intervals of
transverse reinforcement into concrete[3], and so on. Additionally, tests on
chemical resistance under constant tensile deformation[l], tests conducted
in acid hot spring atmosphere[6], bending test after sustained loading and
so on confirmed NEFMAC's durability. Moreover to define the behavior of
concrete reinforced with NEFMAC in fire and at high temperature,
NEFMAC's heat resistance was also studied.
We then verified, in cooperation with universities and for our techno-
logical development project, to verify the applicability of NEFMAC to con-
crete structures. For wall panel development, in-plane bending-shear tests
under reversed cyclic lateral loading and bending tests under out-of-plane
loading were done, and structural characteristics clarified. For slab panel
development, beam-anchoring methods, flexural behavior, and long-term
deflection were confirmed. Besides, to improve flexural behavior we studied
the possibility of using the pretension prestress method[4]. Also, wall and
slab fire resistance were studied and the applicability of NEFMAC was veri-
fied. Furthermore, for the applications to columns and beams, flexural,
shear, and fatigue behavior and especially flexural behavior after long-term
exposure were confirmed.

4. MATERIAL PROPERTIES OF NEFMAC[5]

4.1 Tensile and Compressive P r o p e r t i e s
Standard specification of NEFMAC is shown in Table 3. In this table,
"Max. Load" is guaranteed. Examples of tensile test results of NEFMAC are
shown in Fig. 3.
The tensile testing methods and the compressive testing method of
NEFMAC are shown in Fig. 4 and Fig. 5 respectively. For tensile testing,
three grip types are adopted. In any grip type, the tensile test goes well and
tensile strength of NEFMAC is evaluated correctly. For compressive test-
ing, the method of confining the buckling of NEFMAC by concrete blocks is
adopted.
Relation between the data of tensile and compressive max. load is
shown in Fig. 6. Also the relation between the data of tensile and compres-
sive rigidity is shown in Fig. 7. The symbols represent the average value
and the arrows represent the ranges of the data of the five specimens.
From these figures, it is clear that the following consideration will be led.
In case of type G NEFMAC, the data of compressive max. load are almost
equal to the data of tensile max. load. In case of type C NEFMAC, however,
the compressive data and values are about 40% of the tensile ones. As for
the rigidity, the compressive and tensile data are almost equal regardless of
the types of NEFMAC.
 357

4.2Durability of NEFMAC
4.2.1Durability Test C o n d u c t e d i n Acid Hot Spring Atmo-
sphere[6]
(1) Purpose of t h e Test
The durability of NEFMAC h a s been confirmed in t e s t s based on the
J I S K 7107 "Test Method of Chemical Resistance of Plastics Under Constant
Tensile Deformation," for which various chemicals were used. Its actual in-
use durability, however, had not been proved.
Through t h e cooperation of t h e Ohita Construction office of the J a p a n
H i g h w a y Public Corporation, t h e d u r a b i l i t y t e s t on NEFMAC was per-
formed at t h e Beppu Alum Spa. This c h a p t e r is based on t h e data obtained
from t h e first to t h e twelfth month of t h e t e s t s .
(2) Test Method
Types and q u a n t i t i e s of t h e t e s t pieces are given in Table 4. Each t e s t
piece was installed so as 300mm r e m a i n e d u n d e r g r o u n d and 600mm were
left exposed to t h e air (surface environment). N a m e p l a t e s were attached to
each group of t h e t e s t pieces for t h e purpose of classification. Installed t e s t
pieces are shown in Photo 3. Test types a n d procedures are given in Table 5.
(3) Test Results
a) Underground t e m p e r a t u r e distribution and pH a t place where test pieces
were buried
U n d e r g r o u n d t e m p e r a t u r e d i s t r i b u t i o n a n d r e s u l t s of p H m e a s u r e -
m e n t s at t h e place where t e s t pieces were buried a r e shown in Fig. 8. Tem-
p e r a t u r e w a s m a i n t a i n e d at a b o u t 80C a t t h e lowest h o r i z o n t a l b a r ,
roughly 60C at t h e middle horizontal bar, a n d about 50C at t h e horizontal
bar nearly t h e surface. Max. tensile load of b u r i e d t e s t pieces were t a k e n as
s t r e n g t h of a section h e a t e d to approximately 60C.
b) Results of Visual E x a m i n a t i o n
As seen in Table 6, t h e r e were almost no changes in t h e surface sec-
tions of t h e NEFMAC samples. However, changes were seen in t h e under-
ground sections; among t h e m , a progressive i n c r e a s e in coloring t h a t re-
sulted from deeper embedding was included. This led to blackening of t h e
lowest h o r i z o n t a l b a r . Color c h a n g e could h a r d l y be noticed after two
m o n t h s ; t h u s , degree of coloring was considered to be governed by under-
ground t e m p e r a t u r e r a t h e r t h a n t i m e - d e p e n d e n t change.
The surface sections of t h e welded wire m e s h and steel reinforcement
samples had r u s t on t h e m ; after two m o n t h s , t h e s t a t e of t h e r u s t did not
change. However, t e m p e r a t u r e and pH greatly influenced t h e u n d e r g r o u n d
sections. Test piece melting could be noticed four m o n t h s later for welded
wire mesh and twelve m o n t h s l a t e r for steel reinforcement. T h u s , tension
t e s t s could not be r u n on those samples.
c) Results of Tension Tests
From t h e t i m e - d e p e n d e n t change in t h e t e n s i l e load r e t e n t i o n
ratio(Fig. 9), it is clear t h a t t h e surface sections of NEFMAC G4 and G10
and t h a t of steel reinforcement D10 do not lose s t r e n g t h even after twelve
m o n t h s in an a t m o s p h e r e of t h e exposure site. In contrast, t h e tensile load
retention ratio of welded wire mesh 4 0 was reduced to roughly 50% because
of r u s t .
358

When underground sections came into contact with soil of about 60C
and a pH of 1.7(Fig. 8), corrosion and deterioration was more striking than
in the above-ground sections. For welded wire mesh and steel reinforce-
ment, corrosion progressed in a linear way. 4 0 welded wire mesh melted
away in about 100 days and D10 steel reinforcement in about 220 days.
With NEFMAC, strength reduction could be seen after twelve months while
G10 held onto 70% for the same period. Since G10 has a large area, more
time is required to reach the 60% plateau of the retention ratio of G4.

4.2.2 Behaviors of Beams Subjected to Sustained Load[7]

(1) Test Methods
a) Beam Specimens and Loading Method
The dimensions of the beams and the loading method are given in Fig.
10. Reinforcement was arranged in the axial direction only. The clear cover
was 25mm.
b) Types of NEFMAC and mechanical properties
Three types of NEFMAC, type C, typeG and type H, were used. They
were formed in grid shape of 10cm pitch. The criteria for strength of
NEFMAC was 1.2 times the tensile strength of reinforcing steel (D10,
SD345).
(2) Test Results
a) Deformation behaviors under sustained load
Strains of reinforcements and average curvatures under sustained load
decreased with increasing tensile rigidity of the reinforcement. A difference
was prominently seen between rates of change in curvatures in the defor-
mation behaviors of RC and NEFMAC beams subjected to sustained loads.
In Fig. 11, the amount of increase in curvature is indicated as the rate of the
change in curvature immediately after loading. The absolute values of cur-
vature, as mentioned previously, were higher with the beams u s i n g
NEFMAC of low tensile rigidity than with the RC beams.

b) Mechanical Behavior of Beams Subjected to Different Sustained Loads

Flexural strength tests were performed on beams after load had been
sustained for one year. Strains of reinforcements, flexural crack width, de-
flections and failure loads were determined. The relationship between av-
erage flexural crack width and load for type G beams subjected to different
sustained loads is shown in Fig. 12. The crack width at a load of 0 tf corre-
sponds to the residual crack width immediately before the flexural strength
test. The arrows in the figure indicate values of sustained load before the
test.
The increases in flexural crack width for beams with large sustained
loads (since flexural cracking during sustained loading had developed con-
siderably) showed roughly the same trend, and the effect of sustain loads
being different could hardly be seen. Beams subjected to sustained loads of
100MPa(as converted to stress intensity of reinforcing steel in RC beams)
had large flexural crack width at low load as compared to virgin beams. But
with increasing test load, the difference between the three beams became
small. And, they became close to beams with heavy sustained loads.
As a result of the above, it was recognized that with NEFMAC rein-
forcement, the flexural behavior after sustained loading was influenced by
 359

whether flexural cracks had developed considerably or not. There were

cases when flexural behavior was hardly affected by the magnitude of the
sustained load.

5. STRUCTURAL CHARACTERISTICS OF NEFMAC REIN-

FORCED CONCRETE
5.1 Flexural Behavior[2]
To apply NEFMAC reinforced concrete panels to curtain walls and
fender plates for small ships, we conducted bending tests to clarify flexural
behavior. Table 7 shows the physical properties of NEFMAC and the con-
crete used in testing.
To develop curtain walls , we confirmed the wind resistance of the pre-
cast concrete panels by bending tests. Shapes of specimens are shown in
Fig. 13 and load-deflection relationships of specimens in Fig. 14. Bending
tests were also done on specimens undergoing the fire resistance test ac-
cording to Japanese Industrial Standard; Fig. 14 shows test results.
The aforementioned tests clarified that the flexural cracking load ex-
ceeds the design load and that max. strength of the precast concrete panels
declines little, even after the fire resistance test. We also learned that the
existing bending theory for reinforced concrete can be applied to calculated
initial stiffness (Ig), post-cracking stiffness (Icr) and max. strength (Pu).
Additionally, we clarified through both experiments and analysis that crack
width under the design load is 0.3mm or less.
Fender plates for small ships were originally designed with epoxy
coated steel reinforcing bars. However, the comparison of flexural behavior
(exhibited in bending tests) of two precast concrete panel specimensone
with epoxy coated steel reinforcing bars and the other with NEFMAC rein-
forcementled to the use of NEFMAC in some of the panels. The dimen-
sions of the specimens are shown in Fig. 15, physical properties of the mate-
rials used in Table 7 and the load-deflection relationship of the specimen in
Fig. 16. As NEFMAC has a small Young's modulus, the bending stiffness of
the specimen dropped after flexural cracking, causing the deflection to be-
come approximately 5 times larger than that at yielding (i.e., within the
elastic range) of the epoxy-coated steel reinforcing bars. This result showed
that a NEFMAC reinforced specimen has a large capacity of absorbing en-
ergyan advantage for a fender plate function.

5.2 S h e a r Behavior[8]
Concrete shear walls reinforced by NEFMAC were compared to those
reinforced by conventional deformed steel bars.
Three specimens were tested. The dimensions of the specimens and the
bar arrangement of columns and beams are shown_in Fig. 17. Columns and
beams are reinforced by conventional steel bars. The wall of specimen No. 1
is reinforced by deformed bars as used in columns and beams(steel shear
wall). Fig. 18 shows the detail of the bar arrangement in the wall of speci-
men No. 1. The walls of the specimen Nos. 2 and 3 are reinforced with
NEFMAC(NEFMAC shear wall). The reinforcement details are shown in
Figs. 19 and 20, respectively. NEFMAC is formed ina grid shape so it is
devided into two parts: wall reinforcement(refer to the right side of Figs. 19
360

and 20)and the anchorage to columns and beams (refer to the left side of
Figs. 19 and 20). Table 8 shows the physical properties of NEFMAC and the
concrete used inthe test.
The loading method is shown in Fig. 21. The specimens were loaded
with a horizontal force by two jacks. In order to supply the uniform shear
stress to the wall, a jack pushed one side of the wall beam and the other jack
pulled the other side. The axial force of 16.5 tf(N), which devided by the
section total area of columns was about 2MPa, was applied vertically at the
top of the columns.
The relations between load and deformation are shown in Figs. 22, 23
and 24. The ultimate failure aspect are shown in Photos 4, 5 and 6. When
the deformation of the specimen No. 1 reached about 5mm, the lateral load
reached its peek. At large deformation level, the compressive failure oc-
curred at the base of column and at the corner of wall and the strength was
declined gradually without recovering the loss due to cyclic loading. The
behavior of the specimen No. 2 was similar to that of the specimen N o . l
until the deformation of 12mm, but it was observed that the number of
cracks was less and the width was wider than in the case of the specimen
No. 1. At the deformation of 12mm, noise was heard. It was regarded as the
break of the fiber. At the deformation of 13mm, the failure happened sud-
denly. On the specimen No. 3, cracks occurred over the wall and the
strength increased even after the repetition of the deformation of about
5mm. After the strength reached max. at the deformation of 8mm, the noise
regarded as the break od the fiber was heard, and the failure occurred sud-
denly. The deformation of the two NEFMAC shear walls was over 1/100 in
joint translation angle when the strength reached max. Until the max.
strength, the specimen No. 2 showed almost the same behavior between the
strength and the deformation as that of the steel shear wall No.l.
In the experiment, the shear wall reinforced by NEFMAC showed al-
most same reaction as that reinforced by conventional deformed bars until
the joint translation angle became 1/100. The shear walls reinforced by
NEFMAC had enough strength and deformability and behavior similar to
that of the shear wall reinforced by steel bars. We found that NEFMAC is
applicable to concrete shear walls in place of steel bars.

6. APPLICATIONS^]
As of the end of March 1992, l,150,000m2 of NEFMAC had been used
for concrete reinforcement. What follows is a discussion of its major uses.
Welded mesh is sometimes used to reinforce shotcrete of the New Aus-
trian Tunneling Method (NATM). However, due to its high rigidity, large
gaps are caused between the welded mesh and primary shotcrete. There-
fore, more secondary shotcrete is needed to fill in the large gaps. This in-
creases costs. Additionally, it requires much more labor for mesh setting.
In contrast, due to NEFMAC's low rigidity, only small gaps are caused; less
secondary shotcrete helps cut cost. The simplified method that has been
developed to set NEFMAC by air gun stapling has significantly reduced set-
ting time and overall construction period[10]. Because of high durability,
NEFMAC has been used for reinforcing shotcrete of oil storage rock tanks
(Photo 7).
Also NEFMAC has been used to reinforce shotcrete in the repair a rail
way tunnel (Photo 8). Its non-conductivity was recognized and proved espe-
cially useful in the prevention of electric shock.
 361

In general, shotcrete in tunnels is lined with concrete such as road tun-

nels. This lining is often reinforced with steel bars; however the tunnel s
narrowness demands a significant amount of time and labor for arranging
reinforcement bars. Whereas NEFMAC, lightweight and formed into either
curved of flat planes in factories, drastically reduces work inside tunnels.
Photos 9 and 10 are examples of NEFMAC application; arrangement time
was reduced by roughly 70%.
Because NEFMAC does not rust, it is a promising concrete reinforce-
ment material for ocean coastal structures. Photo 11 shows the concrete
pontoon. A block of styrene foam is contained in the concrete skin. Even if
the cracks are occur in the concrete skin, the pontoon does not sink. Then,
providing for the cracking of concrete, the rust free reinforcement is
adopted. Photo 12 depicts application of NEFMAC to the concrete cushion
panels for small vessels, which were attached to a landing bridge.
NEFMAC has also been applied to the stop log for water intake of a steam
power plant. Two years have passed since NEFMAC has been used for rein-
forcing the pontoon, three years for reinforcing panels and six years for
reinforcing the stop log; no problem has occurred thus far.
Photo 13 shows the construction site of the control building in the ant-
arctic base of Japan. This building had to be constructed by the members of
the expedition not trained for construction, in a short period of time.
NEFMAC was adopted because of its lightweight properties and ease of
prefabucation.
NEFMAC has often been adopted because it is not magnetic and does
not conduct electricity. One example of such usage is in the reinforcement
of a GRC OA floor. Approximately 160,000m2 of NEFMAC has been used
for this purpose (Photo 14). The major reason for its use is that NEFMAC
completely inhibits the flow of stray currents and can greatly raise the
amount of energy absorbed during bending of GRC. NEFMAC has also been
used to reinforce the concrete used in the foundation of an earth magnetism
observatory, for which no metals could be used (Photo 15). This kind of
performance would also prove beneficial in concrete girder reinforcement in
linear motor cars.
NEFMAC was also adopted for use in reinforcing a lightweght concrete
curtain building wall, which was approved by Japan's Minister of Construc-
tion (Photo 16). Although there are many legal barriers that make it diffi-
cult to adopt advanced composite materials in buildings in Japan, it is ex-
pected that the need for lightweight, highly durable building materials will
increase.

7. REFERENCES
1. Hayashi, K., Sekine, K., Sekijima, K. and Nakatsuji, T., Application of
FRP Grid Reinforcement for Concrete and Soil, 46th Annual Conference,
Composite Institute, The Society of Plastics Industry, Inc., February 1991,
ss. 12-D pp. 1-7
2. Sugita, M., Nakatsuji, T., Sekijima, K. and Fujisaki, T., Application of
FRP Grid Reinforcement to Precast Concrete Panel, Advanced Composite
Materials in Bridges and Structures, ACMBS-1, October 1992, pp. 331-340
3. Fujisaki, T., Sekijima, K., Matsuzaki, Y. and Okumura, H., New Material
for Reinforced Concrete in place of Reinforced Steel Bar, IABSE Symposium
in Paris-Versailles, September 1987, pp. 413-418
4. Sekijima, K. and Hiraga, H., Fiber Reinforced Plastics Grid Reinforce-
362

ment for Concrete Structures, IABSE Symposium in Brussels 1990, pp. 593-
598
5. Nakatsuji, T. Mechanical Properties of NEFMAC, NEFCOM Corporation
Technical Leaflet, June 1990
6. NEFCOM Corporation, "NEFMAC Durability Test Conducted in Spa At-
mosphere," June 1989
7. Tsuji, Y., Sekijima, K., Nakajima, N. and Saito, H., "Mechanical Behav-
iors of Concrete Beams Reinforced with Grid shaped FRP and Effects of
Chemical Prestress," Concrete Library of JSCE No.18, December 1991, pp.
211-221
8. Fujisaki, T., Kokusho, S., Kobayashi, K., Hayashi, S., Ito, C. and Arai, M.,
Application of New Fiber Reinforced Composite Material (NFM) to Concrete
Shear Wall, Report of the Research Laboratery of Engineering Materials,
Tokyo Institute of Technology, No.15, 1990, pp. 529-534
9. Nakatsuji, T., Sugita, M., Fujimori, T. 'FRP Grid Reinforcement for Con-
crete and Soil', Transportation Research Board 69th Annual Meeting, Paper
No.89 CP057, January 1990
10. Ikeda, K., Sekijima, K. and Okamura H., New Materials for Tunnel Sup-
ports, IABSE 13th Congress in Helsinki, June 1988, pp. 27-32

Table 1 Characteristics of NEFMAC

Non-corrosive Improvement of the durability of con-

crete structures used under severe
Excellent resistance to alkalis, acids, conditions where damage by salt and
and chemicals chemicals is to be expected

Usage of continuous fibers Effective usage of fibers and appear-

ance of a hybrid effect(*) by mixtures
of different kinds of fibers

Assurance of the strength of the inter- Assurance of the enough anchorage to

sections of the grid the concrete

Availability of lapped splice joint

Light-weight (specific gravity *? 2) Improving of the productivity in the

field

Possibility of forming of intricate

shaped objects in a single body piece

Non-magnetic Applicability to a structure requiring

non - magnetic properties

* Hybrid effect: In a composite material which is reinforced with more than two
kinds of fibers, break starts from the fiber with the least elongation and other
breaks take place in order. The final break in the material occurs when the fiber
with the largest elongation breaks. In this process, the stress-strain relationship
becomes non-linear and shows the same phenomenon as the yielding of a
reinforcing steel bar.
 363

Al y y y y v y y y
-Pin
/
t>J y

Resin \ 1

\ 1
3
.Mold
7ied

=
1i & = = = ^T! A A ff
Fig. 1 Batch Process for Production of NEFMAC

UV Cure Resin
Trim
Cut
Package.

Fig. 2 Continuous Process for Production of NEFMAC

 Table 2 Research and Development Items of NEFMAC

Purpose Examination
Clarification of Fundamental Properties of NEFMAC
-| Mechanical Properties Tensile Strength of Fibers
Tensile Strength and Hybrid Effect
Compressive Strength
Shear Strength
Reversed Cyclic Loading (Tension/Compression)
Strength of Cross Point
Bond Mechanism Anchorage to Concrete
Lap Splice Joint
Durability Chemical Resistance under Constant Tensile Deformation
H Tensile Strength under Spa Atmosphere
H Creep Fracture
Heat Resistance Tensile Strength under/after Heating/Cooling
Proof of the Applicability to Concrete Structures
Earthquake Resisting Wall Behavior under Reversed Cyclic Lateral Loading
Precast Wall Structure Shear Resistance of Vertical Joint
Wall Panel Flexural Behavior ,
T Fire Resistance
| Slab Panel h NEFMAC s Anchorage to Beam
Time-Dependent Deflection
Flexural Behavior
Fire Resistance
Effectiveness of Prestressing
- Beam and Column Effectiveness of Main and Shear Reinforcement
h Confined Effect by Lateral Reinforcement
Fatigue Behavior
Flexural Behavior after Sustained Loading
Flexural Behavior with Exposed NEFMAC in Air
Effectiveness of Chemical Prestressing
- Shotcrete and Concrete Lining | Work Productivity
 Table 3 Standard Specification of NEFMAC

Sectional Max. Load Tensile Young' s Weight

Type Bar Not Area Strength ModuI us
(mm1) (tonf) (kgf/mm') (kgf/mm') (g/m)

G2 4.4 0.26 7.5

G G3 8.7 0.52 15
G4 13.1 0.78 22
Glass Fiber G6 35.0 2.1 60 3000 60
+ G10 78.7 4.7 130
Resin G13 131 7.8 220
G16 201 12.0 342
G19 297 17.7 510

H H6 39.5 2.1 65
H10 88.8 4.7 147
Glass Fiber/ H13 148.0 7.8 53 3700 244
Carbon Fiber H16 223 12.0 368
+ H19 335 17.7 553
Resin H22 444 23.4 733

C C6 17.5 2.1 25
C10 39.2 4.7 56
Carbon Fiber C13 65.0 7.8 92
+ C16 100 12.0 120 10000 142
Resin C19 148 17.7 210 1.0 2.0 3.0
C22 195 23.4 277
Elongation(%)
A A6 16.2 2.1 21
A10 36.2 4.7 46 Fig. 3 Load vs. Elongation of NEFMAC
Aramid Fiber A13 60.0 7.8 130 5700 77
+ A16 92.3 12.0 118
Resin A19 136 17.7 174

Note : ltonf = 9.8kN

2
lkgf/mm =9.8MPa
 Steel-sleeve i s . 6. NEFMAC ON
ON
/
/ ^-Jb-4
u u
190 300 190
680

(a) Steel-sleeve Grip Type

Glass-cloth
with resir /
/
u u-
*---- -;
J,-- --
NEFMAC

200 300 200

700 j

(b) Glass-cloth Grip Type Concrete

NEFMAC

Fig. 5 Compressive Testing Method of NEFMAC

(c) Resin-mortar Grip Type.

Fig. 4 Tensile Testing Method of NEFMAC

 u en
M
7 h o 700
Average Range of Data Type G
X
O Type G
9 Type H 600
Type C
-f-
TO a 500

5 4 400

1 1 300

200
Average Range of Data
O Type G
<J TypeH
100
Type C
4-
Note : lkgf = 9.8N
J I I L
600 700
xlO'(kgf)
Tensile Rigidity

Fig. 6 Tensile Max. Load vs. Compressive Max. Load Fig. 7 Tensile Rigidity vs. Compressive Rigidity

as
368

Table 4 Types and Quantities of Exposed Test Pieces

\ Material
Dia-^s. NEFMAC Welded wire Steel
mesh reinforcement
meter ^v

i 4mm l l p c s . (G4) l l p c s . (40)

10mm llpcs. (G10) llpcs. (D10)

(Each material was tested to confirm initial strength.)

Table 5 Test Types and Procedures

Type of test Procedure

Visual Corroded state was observed and recorded.

examination Photographs were taken.

Three vertical bars were collected from each sample

by cutting horizontal bars away. These vertical bars
were cut and separated into underground and surface
sections. Samples were subjected to a tension test at
Tension test Matsuzaki Laboratory, Architectural Division,
Faculty of Engineering, Science University of Tokyo.
Maximum load and fracture conditions were
recorded. NEFMAC test pieces were reinforced at
chucked sections and then tested
 369

*-o

r
2.0
<&-- <t

H i.o

-50 -40 -30 -20 -10 0 +10

Surface

Measuring position(cm)

Fig. 8 Underground Temperature Distribution and pH of the Place Where Test

Pieces were Buried
370

Table 6 Results of Visual Examination

Initial state 2 months later 4 months later 6 months later 12 months later
Test piece
Mar. 3,'88 May 18, '88 Jul. 6, '88 Sep. 2 1 / 8 8 Mar. 7, '89

Surface Translucent No change Same as left Same as left Same as left

camel color
NEFMAC Under- Same as above Lowest No change in Same as left Same as left
G4 ground horizontal bar coloring
was blackened.
Surface Same as above No change Same as left Same as left Same as left

NEFMAC Under- Same as above Lowest No change in Same as left Same as left
G10 ground horizontal bar coloring
was blackened.
Surface Rust-free Rust (red) No change in Same as left Same as left
Welded wire generated rusted state
mesh Under- Same as above Partially melted Melted portion Mostly melted
40 ground away increased away
Surface Same as above Rust (red) No change in Same as left Same as left
Steel- generated rusted state
reinforce- Under- Same as above Section reduced Section further Partially melted Mostly melted
ment ground by rusting reduced by away away
D10 rusting

- /-D10( surface)
L ^

'
/ GlO(surface)
1.0
L ~r G4 (surface)
.2 0.8 L "" ' ~._^^ G10( underground)

\ \ G4 (underground)
\ \
0.6h \\ V
l
\
\
\\
X

\N
4 { surface)
0.4 r
\
V

4
0.2 L 4 <# (underground)
\
s
D10(underground)

h
1 1 1 1 1 1 1 L I I I L_ t L _, L
100 200 300 400

Duration time(days)

Fig. 9 Tensile Load Retention Ratio and Duration

 371

PC tendon
Spring

Cross point of grids

-+-LL+. --f-- + -Q- + S- + -Q-+-G +-0- + -

o o o o o

( ) ( )
II ti
II O O O O O
II
II ^ > Contact t i p
II
-It- 4--0- + - 0 - + - 0 - + -G- +-G- + - ._JL + .
^ W^
100 300 4@100=400 300 100

Fig. 10 Test Method of Sustained Loading

Stress of reinforcing
200 r RC
bar=150MPa

150 \

100 \

Type G

50

200

Time after loading(day) Unloading

L
-50

Fig. 11 Rate of Change in Curvature

372

stress ( M P a ) time

n o n 28 days

o 0

1 0 0
One year
D 1 5 0

v 2 0 0

Note : ltonf = 9.8kN

2
initial loads
during sustained
load tests

200 400 600 800 1000 1200 1400

Flexural crack width(10" , mm)

Fig. 12 Relationship between Flexural Crack Width and Load

Table 7 Material Properties of Bending Tests

Applications Materials Type Max.(Yield)Load(kN) Stiffness (kN)

Curtain Wall NEFiMAC H8 37.14 2,195

G8 30.38 1,519

Light Weight Fc= 30.9MPa

Concrete

Fender Plate NEFMAC G19 172.5 8,683

Reinforcing Bar D19 Yield = 100.9 55,929

(SD295A)

Concrete Fc= 16.4MPa

 373

150 3.600

I I I
L-l-L 1 I |i I I I h i I I l i l I I I I | | I

1
-41I T
[ M i ! i II I11' I I' [YlD J111 1
II1 II ill II111II11II1
II 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 G8 7 ch-^'MccuAr/iV

1 1 11 1
1

ng I

1 1111 11 11 1111 111111111

* I11111'1'1 M1 II11 I I'
111111 I 11
h iil ll
1 I I' | |i
m
rL1J.1
t 1 t I 111 i I I I> ' I I >111 I
i I I I 111 I Ii I Ii III 1 1
1 i

30
H r*\ I
30
Fig. 13 Details of Test Curtain Wall
90

Fig. 14 Load-Deflection Relationships of Test Curtain Walls

374

4,280

50 190 19@200=3.800

IT)
f- fo" o o

o
in
r-
3 II
in P J
r
\
si
> in
o
in

in
P
5 CO 1

m >

>,
>
lo _ _J L
n:=i

- 1
Ui

Fig. 15 Reinforcement of Fender Plate

Epoxy Coated Steel Bar

s NEFMAC/
/
L / C

/ ^ 1
>

L // ^ ^^ [| 1
140
1/ ^^ 1.600
4,280
400

\L^S^ Note : ltonf = 9.8kN r

1 1 J 1 L i 1 L_
10 12 14

Deflection(cm)

Fig. 16 Load-Deflection Relationships of Specimens

 375

Table 8 Material Properties of Shear Walls

Materials Type Max. (Yield) Load(kN) Stiffness(kN)

NEFMAC H6 24.70 1,490

H3 12.05 764

Reinforcing Bar D16 Yield=70.56 35,389

D6(Column) Yield= 12.74 5,821

D6(Wall) Yield=11.96 5,870

Concrete Fc=21.6~
23.8MPa

STDD6@100

Fig. 17 The Dimensions of the Specimens and the Bar Arrangement of the
Columns and Beams
376

06@100-

length of anchorage 35D

I I I I I I I I I I

Fig. 18 The Bar Arrangement in the Wall (No. 1)

length of outstanding 25mm

H6@100

(in nail)
(for anchorage)

Fig. 19 The Bar Arrangement in the Wall (No. 2)

length of outstanding 20mm

(for anchorage)
Fig. 20 The Bar Arrangement in the Wall (No. 3)

mm 1
-T- r
reaction frame
ii) i i h i iztcfc
rolIer
m r r t
200tf i
.
load cell. . IL -JToil jack

n
in
Pin -Pin

^
Sffflr
-oil jack c e| |

m
LU LJJ LU u uuuu u u u
LU LiJ J
Fig. 21 Loading Method
 377

Fig. 22 The Relations between Load and Deformation (No. 1)

Shear Force(tf)
80

60

40
\
20
^/ Deforme tion(mm)

f/ 10 2o :0 40 50

^-20

--40
Note : 1tf = 9.8kN
--60

--80

Fig. 23 The Relations between Load and Deformation (No. 2)

Shear Force(tf)
80

60

40
)
20 t
y Deformati on(mm)
V i0 20 30 40 50

--20

--40
Note : 1tf = 9.8kN
--60

--80

Fig. 24 The Relations between Load and Deformation (No. 3)

378

Photo 1 An Example of the Form of NEFMAC

Photo 2 Intersections of a Grid

 379

Photo 3 Installed Test Pieces

Photo 4 The Ultimate Failure Aspect (No. 1)

380

Photo 5 The Ultimate Failure Aspect (No. 2)

Photo 6 The Ultimate Failure Aspect (No. 3)

 381

Photo 7 Oil Storage Rock Tanks

Photo 8 Railway Tunnel

382

Photo 9 Invert of Water Conveyance Tunnel

Photo 10 Arch of Water Conveyance Tunnel

 383

Photo 11 Pontoon

Photo 12 Cushion Board

384

Photo 13 Control Building in the Antarctic Base of Japan

Photo 14 OA Floor
 385

Photo 15 Foundation of Earth Magnetism Observatory

Photo 16 Curtain Wall

Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 387
1993 Elsevier Science Publishers B.V. All rights reserved.

Three-dimensional fabric reinforcement

H. Nakagawa, M. Kobayashi, T. Suenaga, T. Ouchi, S. Watanabe and K. Satoyama

Kajima Technical Research Institute, 19-1, Tobitakyu 2-Chome, Chofu-shi, Tokyo 182,
Japan

Abstract
The three-dimensional fabric studied as a reinforcement for concrete is a stereo-fabric
made of rovings of fibers, woven into three directions and impregnated with epoxy resin.
Fiber material, number of filaments, and distance between rovings can be varied easily.
Efficient production is also possible since three-dimensional weaving, resin impregnation,
and hardening can all be done by an automatic weaving machine.
The authors investigated the flexural and fire resistance behaviors of three-dimensional
fabric reinforced concrete (3D-FRC) toward applying the material to building panels. The
results demonstrate that 3D-FRC panels have sufficient flexural strength and rigidity to
withstand design wind loads, and fire resistance of 60 minutes was achieved.
3D-FRC panels have been used for curtain walls, parapets, partition walls, louvers, etc.,
and installations amount to 7,000 m2 in Japan.

1. INTRODUCTION

In a worldwide high-tech boom, construction industries are now adopting the use of new
materials. One reason for this trend is the deterioration of reinforced concrete as a conse-
quence of the corrosion of steel bars. Another is labor savings in construction work, necessi-
tated in Japan by a labor shortage and desires to reduce working hours and construction costs.
Pressure arising from these factors has accelerated the increase in size, decrease in weight,
and prefabrication of construction members. Fiber reinforced composite materials have
been attracting more and more attention in recent years. With the potential advantage of
their light weight, high strength, and long durability, they are expected to be applied to high-
rise buildings and sea front structures [1].
Three-dimensional fabric recently developed in Japan is a stereo-fabric made of rovings
of fibers, woven into three directions and impregnated with epoxy resin [2]. Flexural and
fire resistance behaviors of three-dimensional fabric reinforced concrete (3D-FRC) were
investigated toward applying the material to building panels.

2. THREE-DIMENSIONAL FABRIC

The intensive development of advanced composite materials reinforced with fibers such as
carbon fiber and aramid fiber has been closely linked to the various ways these fibers can be
configured. Figure 1 shows various fiber configurations for composite materials, classifying
geometrical shapes as dimensions and fiber directions as axes [3]. Following this figure, a
388

Axis 1 2
Dimens ioi Mono-ax i a I Biaxial Triaxial Multi-axial
1D Pultrusion rod

2D
Pre-impreg- Plane weave Triaxial Multi-axial
nation sheet weave weave
4- 5- 6-
axial axial axial
c Weft

CO

c
Multi-ply Triaxial 3D Multi-axial weave
3D braid weave weave
3D
c

_
CD

CD
^
CD
Laminate type H or I beam Honey-comb
c type
CO

Reprinted from : Composite Materials Technology Handbook for

Future Industries, Japanese Standards Association, p. 135

Figure 1. Fiber configurations for composite materials

pultrusion rod is classified as mono-axial-ID, a unidirectional pre-impregnation sheet (UD

pre-preg sheet) as mono-axial-2D, and braided fiber rovings and laminated UD pre-preg
sheets as mono-axial-3D.
Three-dimensional fabric (3D fabric) is made by weaving fiber rovings in three dimen-
sions. Complex configurations for triaxial and multi-axial-3D fabrics can be created, as
indicated in Figure 1. An important feature of 3D fabrics is that they can readily be manufac-
tured in configurations which suit the required performance, as opposed to the conventional
mono-axial fiber reinforced composite materials which are extremely anisotropic.
The 3D fabric investigated in this study is woven as triaxial-3D and impregnated with
epoxy resin into rovings (see Figure 2). It was devised for reinforcing concrete, and varia-
tions with different distances between rovings can easily be produced. The fabric maintains
its configuration after being placed in concrete because, as shown in Figure 3, there are two
types of Z axis rovingsmain rovings (Z;) and sub-rovings (Z0)which are intertwined with
X and Y axes rovings between them. The triaxial intersections of this 3D fabric are, there-
fore, stronger than those of conventional 3D fabrics.
 389

Figure 2. 3D fabric Figure 3. Fiber configuration for

3D fabric

3. PROPERTIES OF 3D-FRC

3D-FRC is concrete reinforced with 3D fabric in place of steel bars. By changing the
fiber material, the number of filaments, and the distance between rovings, fiber reinforcement
appropriate to various stress conditions can be produced. Moreover, the mechanical bond
strength between fibers and matrix is high because of the checkered pattern structure of the
rovings.
The authors worked with carbon fiber made from polyacrylonitrile (PAN-type carbon
fiber), aramid fiber, vinylon fiber, or AR-glass fiber for 3D fabrics. The physical properties
of the fibers are shown in Table 1.
Figure 4 shows the tensile stress-strain curves of FRP rods which compose 3D fabric.
For this figure, tensile stress was calculated on the basis of net fiber cross-sectional area
(epoxy resin area was excluded). The tensile stress-strain relationships are approximately
linear up to the ultimate tensile strength. The tensile strength and modulus of elasticity of
FRP rods are 70% to 80% of those of filaments. It is considered that 20% to 30% of fila-
ments are not aligned under equal tension.

3.1 Flexural behavior

Four-point flexural tests were conducted on planks made with concrete containing short
vinylon fibers mixed in the matrix (VFRC), with and without 3D fabric reinforcement to
investigate the effects of the presence of 3D fabrics and short vinylon fibers on flexural
characteristics. The specifications of the 3D fabrics and the mix proportions of VFRC are
listed in Table 2 and 3. PAN-type carbon fiber and aramid fiber were used for the X and Y
axes of 3D fabrics. The Z axis fibers of all specimens were PAN-type carbon fibers, a
specification derived out of consideration for ease of manufacturing. Low shrinkage cement
390

Table 1
Physical properties of fibers
Diameter Specific Tensile Modulus of
Fiber type Elongation
( u m) gravity strength elasticity
(GPa) (GPa) (%)
PAN-type
carbon fiber 7.0 1.8 3.63 235 1.5

Aramid fiber 12.4 1.4 3.04 73.5 4.4

Vinylon fiber 14.2 1.3 1.47 36.3 6.6

AR-glass fiber 13.5 2.7 1.47 73.5 2.0

0 1.0 2.0 3.0 4.0 5.0

Strain {%)

Figure 4. Tensile stress-strain curves of FRP rods

was used to prevent cracking and deformation of the planks due to drying shrinkage. Micro
balloons, which are minute hollow glass balls made from pozzolan (particle size less than 100
), were used to reduce weight. Loading tests were performed using four-point loading
with equal spacing (see Figure 5).
Load-deflection curves for CF48 and AF16 specimens are shown in Figure 6. They
reveal the reinforcing effects of short vinylon fibers mixed in the matrix. The higher is the
vinylon fiber content, the larger is the increase of modulus of rupture (MOR) and flexural
rigidity in the second region. In the same way, the higher the vinylon fiber content, the
smaller is the drop in strength after crack occurrence.
The curves for specimens without short vinylon fibers have an apparent yielding range
 391

Table 2
Specifications of 3D fabrics (flexural behavior test)
Number of Pitch of rovings Tension
Specimen filaments (mm)
No. Fiber type reinforcement
X, Y axes Z axis X, Y axes Z axis ratio Pt (%)
CF48 PAN-type 48KF 0.54
carbon fiber 12KF 30 20
CF72 72KF 0.79
AF16 16KF PAN-CF 0.56
Aramid fiber 12KF 30 20
AF24 24KF 0.82

Table 3
Mix proportions of VFRC

w/c S/C
Fiber content Cement Aggregate Admixture
(%) V, (vol.%)
0.0
47.3 0.13 1.0 Low shrinkage Micro balloon
cement Methylcellulose
1.5

LO
cvi

3
I I I I 1 I I I 1 1 I 1 I I I I I I I I I I I I I I I I I I I I I I I I I 1I I 1 I I I I I I | LO

hool 400 1 400 1 400 hool -

1o
1 CM

I 1 I 1 11 1 I 1 1 1 1 M 3-D fabric (360 x 20 x 1.380) o <->>

U i 11111111111111111111111111111IIII11 ^ o
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 1 /tt\
111II111 1II1II11M111IIIIM1IIII1II11IIII1111
L> 1 11

|10 @30 10|
1.400 I

Figure 5. Geometry of flexural test specimen

392

0 100 100100 0 0 100 100 100200 200 200

Deflection Deflection

Figure 6. Load-deflection curves (CF48, AF16)

,-N ID
z CF72>
CO

\ 20 |
/CF48
o /^~AF24
CD

_3io CO
CO
CD
1
+-
CO

^AF16 10 75

5 3

LL

J ^ r ^ N o n - i IDF

50 100 150 200

Deflection (mm)

Figure 7. Load-deflection curves (Vf=1.0%)

 393

after the first crack point. This is because the main stress of matrix on extreme tension fiber
is released, and stress shifts to the 3D fabric tension reinforcements which causes the fibers to
elongate. In the case of specimens with short vinylon fibers, on the other hand, the apparent
yielding range did not appear because the matrix could withstand the stress even after crack
occurrence since the short fibers served as a bridge.
Figure 7 shows the load-deflection curves for specimens reinforced with different 3D
fabrics and a 1.0% short vinylon fiber content. The flexural characteristics of 3D-FRC
depend on the properties of the matrix up to the initial crack occurrence, and on the material
and fiber content of the 3D fabric in the range beyond. Rupture elongation of fiber influ-
ences maximum deflection, and tensile strength, modulus of elasticity and fiber content influ-
ence MOR and flexural rigidity.

3.2 Thermal durability

The thermal durability of flexural strength was also examined by testing specimens soaked
in 80C water or autoclaved at 180C. The specifications of 3D fabrics and the mix propor-
tions of the mortar matrix for specimens tested for thermal durability are listed in Table 4 and
5. PAN-type carbon fiber, aramid fiber, vinylon fiber, and AR-glass fiber were used for
the X axis of 3D fabrics. The matrix was river sand mortar made with high-early-strength
Portland cement. Some specimens were soaked in 80C water for a designated period (up
to 30 days) after four weeks standard curing (20C, 60% RH), while others were cured by
autoclaving at 180C from one to five times (five hours per session). Loading tests were
performed under four-point loading with equal spacing (see Figure 8).
Figure 9 shows the relation between MOR and days soaked in 80C water, and Figure 10
the relation between MOR and number of autoclave sessions. While the MOR of specimens
with AR-glass fiber decreased with longer periods of soaking in hot water, the MORs of
other specimens changed little after soaking for up to 30 days. Also, the rate at which the
MOR of specimens with AR-glass fiber changed was extremely low compared with that of
conventional GRC made with short AR-glass fibers [4].
The MORs of specimens autoclaved at 180C all declined after the first session (see Figure
10). This is assumed to be a result of the deterioration of epoxy/matrix bond strength.
After subsequent sessions MORs of carbon fiber and aramid fiber specimens were relatively
unchanged and that of AR-glass fiber specimens fell gradually. The MOR of vinylon fiber
specimens declined radically after autoclaving as thermal deterioration destroyed the reinforc-
ing capacity of these fibers.

4. EXAMPLES OF 3D-FRC BUILDING PANEL APPLICATIONS

4.1 Partition panelschlorine gas storeroom, Higashi-Murayama purification plant

This was the first application of 3D-FRC planks to an actual structure. The original
design called for the wall between a chlorine gas storeroom and an adjacent room to be made
of pre-cast concrete panels reinforced with stainless steel bars. However, due to concerns
over possible leakage of chlorine gas and the need for an extremely durable material, and
because site conditions allow only limited space for installation, light and durable 3D-FRC
was used. Figure 11 shows the geometry of a standard panel and the photograph in Figure
12 shows installation in progress. The 3D fabric was made of PAN-type carbon fiber
(48KF) and the matrix was VFRC with 1.0% content of vinylon short fibers. The weight of
a standard panel is 250kg, and the panels were installed over an area of 80 m2.
394

Table 4
Specifications of 3D fabric (thermal durability test)
Number of Pitch of rovings Tension
Specimen Fiber type filaments (mm)
No. reinforcement
X axis Y axis Z axis X,Y axes Z axis ratio Pt (%)

CF72 PAN-type
carbon fiber 72KF 24KF 12KF 20 30 0.45

AF24 Aramid
fiber 24KF PAN-CF
24KF
PAN-CF
12KF 20 30 0.47

VF18 Vi nylon 18KF PAN-CF PAN-CF 20 30 0.45

fiber 24KF 12KF

GF19 AR-glass 19KF PAN-CF PAN-CF 20 0.45

fiber 24KF 12KF 30
KF : 1000 filaments

Table 5
Mix proportion of mortar matrix

w/c S/C Cement Aggregate Admixture

High-early-strength River sand Water reducing
45 2.2 Portland cement admixture

r~^ o
QJ] >*
~7X 5~ I 70
50 | 200 I 200 | 200 150
700

Figure 8. Geometry of durability test specimen

 395

PAN-type
carbon fiber

Aramid fiber

AR-glass fiber

Vinylon fiber
Limit of
proportionality
(mean of all)
10 30
Days soaked in 80 C water

Figure 9. Relation between MOR and days soaked in 80C water

PAN-type
carbon fiber

Aramid fiber

AR-glass fiber
Limit of
proportionality
(mean of all)
Vinylon fiber
2 3 4 5
Number of autoclaving (180Cx 5hours)

Figure 10. Relation between MOR and number of autoclave sessions

396

3DF (PAN-CF)
@30 (X, Y) 2 0 (Z)

o
o In
^CO In
IS
for handling

[-60X50X2.3

150

Figure 11. Geometry of standard panel Figure 12. Installation in progress

(Higashi-Murayama purification plant) (Higashi-Murayama purification plant)

4.2 Parapet panelsSuidobashi Building, Tokyo Dental College

The exterior of this building (see Figure 13) was composed of steel-stud-framed [4]
granite-faced curtain walls. The curtain wall panels are made of concrete reinforced with
short carbon fibers. However, steel-stud-framed panels were not suitable for the parapet
because corrosion of the steel stud frames due to weathering was likely. This is the reason
durable 3D-FRC panels were used.
Fig. 14 shows the geometry of a standard panel. The rovings used for 3D fabrics were
aramid fibers (24KF) for the X and Y axes and carbon fibers (12KF) for the Z axis, in consid-
eration of possible trouble concerning radio wave interference. VFRC (Vf=1.25%) was
used for the matrix. Panel-type pieces of 3D fabric were arranged in rows a maximum 600
mm wide along the X axis, and were bound to the pieces of beam-type 3D fabric and pieces
of 2D joint sheet fabric. The weight of a standard panel is about one ton, and the panels
 397

cover 600 m2.

Wind resistance tests were conducted on a specimen identical to the largest panels used on
the Suidobashi Building (3573 mm wide and 2133 mm high). The design wind loads for this
case were large at 5.0 kN/m2 on the negative pressure side and 3.3 kN/m2 on the positive
pressure side. The wind load-displacement and wind load-strain relationships, both as
measured at the center of upper rib where maximum displacement occurred, are shown in
Figure 15. Displacement and strain of the specimen display elastic relationships with varia-
tion of wind load, indicating that 3D-FRC panels have stable deformation behavior and suffi-
cient flexural strength and rigidity to withstand the design wind loads.

Figure 13. Suidobashi Building, Tokyo Dental College

398

3-D fabric (aramid, beam type)

6 0 (X,Z), 30 (Y)
/3-D fabric (aramid, panel type)
6 0 (X,Y), 20 (Z)

Figure 14. Geometry of standard panel (Suidobashi Building)

-10i 2
Max load (8.8kN/m ) load (8.8.kN/m2)
CM
E
\
z
Q-
Design iuau
(3.3kN/m 2 )

4 2 400
i
t
> l\ 400
" ( Tim)

Desi gn lo ad
(-5.C ) k N / -R

I Max loa<i 4
(-8.8kN/r MaxJoad (-8.8kN/m 2 )
r '"""""' 1-10- 10 1
(Wind load vs. displacement) (Wind load vs. strain)

Figure 15. Relation between wind load and displacement and strain
 399

43 Curtain wallsSea Fort Square, Shinagawa, Tokyo

3D-FRC was first applied on a large scale (1500 m2) to the 23-story Sea Fort Square
building in the sea front area of Shinagawa, Tokyo (see Figure 16). A single unit of this tile-
finished curtain wall consists of a 3D-FRC panel and a steel stud frame [5], the two being
connected by bolts at intervals of 600 mm. The original design specified aluminum panels
for this curtain wall, but due to concerns over possible salt-induced aluminum corrosion and
because a significantly heavier wall material such as conventional precast concrete would not
suit the load-bearing capacity of the steel skeleton, light and durable 3D-FRC was adopted.
Figure 17 shows the geometry of the panels. The rovings used for the 3-D fabrics were
PAN-type carbon fibers (24KF in X and Y axes and 12KF in Z axis). Joint rods made of
120KF carbon fibers were used to bind two pieces of 3D fabrics in beam panels.
60-minute fire resistance tests were conducted according to the JIS A 1304 standard fire
test procedure [6] to obtain the approval of the relevant authorities. The specimens were
modified from column panels to suit the size of the furnace and were not finished with tile;
all other specifications were the same as for the panels to be used on the building. The
temperature in the furnace was held at the standard level, and temperature and deformation at
several points on the specimens were measured.
Maximum temperature and deformation at each point measured are listed in Table 6, and
temperature histories are shown in Figure 18. Maximum temperature on the unexposed
surface of the specimens was 103C when heated from the exterior face, and 110C when
heated from the interior face. Both are well below the 260C maximum allowable tempera-
ture specified by JIS A 1304 regulations. Maximum temperature of the 3D fabric on the
exposed side was 516C when heated from the exterior face and 349C when heated from the
interior face, while at the 3D fabric on the unexposed side maximum temperatures when
heated from the exterior and interior faces were 152C and 110C, respectively. Consider-
ing the results of tensile strength tests of 3D fabrics after heating to 400C, shown in Figure
19, the unexposed side rovings of 3D fabrics were not damaged by these temperatures though
the exposed side rovings apparently deteriorated.
Maximum deformation occurred 10 minutes after the tests were started, and was 28 mm
under heating from the interior face. Most of the deformation was probably caused by
deflection of the steel frame due to the temperature differential between the outside and the
inside of the steel frame. There were no deformations, fractures, or cracks which would
weaken fire resistance, and fire resistance of 60 minutes was achieved.

5. CONCLUSION

3D-FRC panels have been used for curtain walls, permanent forms (see Figure 20), lou-
vers (Figure 21 and 22), and for other applications in addition to the examples described
above. Installations now amount to 7,000 m2. All members to which the panels have been
applied are substructural, since there are fewer restrictions on such applications. There are
still problems to be solved before FRP reinforced materials, including 3D-FRC, can be used
for primary structural members such as columns, beams, slabs, and bearing walls. A few of
these problems are: those related to various laws and regulations, including the Japan Build-
ing Standards Law; the cost of FRP bars; various mechanical and physical properties; and
durability.
FRP reinforced materials, including 3D-FRC, will first replace steel reinforced concrete in
structures designed for extremely adverse conditions, in small-scale structures, and in sub-
structural members before they come into wider use, as accumulated technologies lead to the
gradual improvement of these materials.
400

Figure 16. Sea Fort Square

Figure 17. Geometry of curtain wall panel (Sea Fort Square)

 401

Table 6
Fire resistance test results
Heated from Heated from
the exterior face the interior face
Maximum Time Time
Maximum (min)
(min)
Exposed 937.5 60 971.7 60
3D-FRC surface
panel
Unexposed
surface 103.0 69.5 110.2 103
Exposed 83.1 84 952.7 60
Temperature side
Steel frame
(C) Unexposed 56.5 101.5 864.6 60.5
side
Exposed 515.5 61 348.6 62.5
side
3-D fabric
Unexposed 151.7 73 109.9 77
side
Deformation Center of 3D-FRC panel 3.6 12 28.0 10
(mm)

1000

Time (min)

Figure 18. Temperature histories (heated from interior face)

402

-5 4.0
_
CD

2 3.0 .^^..

2.0

1.0
100 200 300 400
Temperature (C)

Figure 19. Relation between tensile strength (after heating)

and maximum temperature for 3D fabrics

Figure 20. 3D-FRC permanent forms (mock up test)

 403

Figure 21. 3D-FRC louvers (Shinjuku district heating center)

Figure 22. 3D-FRC louvers (curing in progress)

6. ACKNOWLEDGMENTS

The authors are grateful to Dr. Akihama of FRC Corporation for extending his valuable
advice in this research effort. The authors would also like to thank Arisawa Mfg. Co., Ltd.
for supplying the 3D fabric.
404

7. REFERENCES

1 T. Fujisaki, K. Sekijima, Y. Matsuzaki and H. Okamura, New Material for Reinforced

Concrete in Place of Reinforcing Steel Bars, IABSE Symposium in Paris-Versailles
(1987) pp. 413-418.
2 K. Fukuta, T. Kitano, Y. Nagatsuka, E. Aoki, M. Funabashi, S. Shirai, M. Anahara and K.
Murayama, Composite Materials Reinforced with Three Dimensional Fabric, Progress of
Japan-Sweden Research Cooperation in Composite Materials (1988) pp. 14-17.
3 K. Fukuta, Composite Materials Technology Handbook for Future Industries, Japanese
Standards Association (1990), edited by H. Maki, pl35.
4 S. Akihama, T. Suenaga, M. Tanaka and M. Hayashi, Properties of GFRC with Low
Alkaline Cement, Fiber Reinforced Concrete Symposium ACI, Fall Convention in Balti-
more (1986) SP105-11 pp. 189-209.
5 S. Akihama, K. Ogawa, T. Suenaga, I. Uchida, H. Fujii and M. Hayashi, Development of
New GFRC Cladding, 6th Biennial Congress of the GRCA in Edinburgh (1987) pp. 4 1 -
52.
6 Method of Fire Resistance Test for Structural Parts of Buildings, Japanese Industrial
Standard, A 1304 (1975).
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 405
1993 Elsevier Science Publishers B.V. All rights reserved.

A New Three-Dimensional FRP Reinforcement

T. Yonezawa3, S. Ohnoa, T. Kakizawaa, K. Inoueb, T. Fukatab, and R. Okamotob
a
Takenaka Corporation
2-5-14 Minamisuna, Koto-ku, Tokyo, JAPAN
b
Shikibo Ltd.
1500-5 Sibahara Minami, Youkaichi-shi, Shiga, JAPAN

Abstract

This paper reports a three-dimensional FRP (fibre reinforced plastics) reinforcement

which has recently been developed for concrete structural members. This reinforcement called
BE3D, is produced in a streamlined manufacturing system which is introduced in this paper.
The results of experiments on the tensile properties of the BE3D reinforcement are described.
Related to the improvement of the BE3D reinforcement, the effects of deformations of FRP
bars on bonding properties are investigated and discussed both in the experiment and the
numerical analysis. Flexural loading tests on slab specimens reinforced with BE3D were
carried out and the results are discussed.

1. INTRODUCTION

Research and development of FRP materials as alternatives to steel bars or prestress

cables for concrete reinforcement has recently been conducted. These FRP materials have
many advantages, i.e. high strength, rust free, and non-magnetic PROPERTIES. They have
great potential for future applications. Also, FRP reinforcements can be made in different
shapes. At present one-dimensional reinforcements such as rods or cables are used mainly for
prestressed concrete considering the reinforcing effect. However, the utilization of FRP
reinforcements as alternatives to ordinary steel bars should not be neglected because there are
many reinforced concrete structures where the excellent properties of FRP could be utilized. In
general, ordinary steel reinforcements for concrete structural members are three-dimensionally
assembled at the construction site. However, in the case of FRP materials, it may not be suitable
to assemble the various FRP components at the construction site. Since work such as bending
or connecting requires special skills, if bending and connecting are improperly done, the
properties of FRP components may be harmed. At this point, three-dimensional FRP
reinforcements assembled at factories are more appropriate than one-dimensional
reinforcement for construction uses. Since FRP reinforcements are also very light, even large
three-dimensional FRP reinforcements can be easily used. This will make construction more
efficient, thus shortening the period of work. Three-dimensional FRP reinforcements have
been developed for this reason.
406

2. THREE-DIMENSIONAL FABRIC FRP

REINFORCEMENT

The use of manufacturing system for

woven three-dimensional fabrics is a typical
method of manufacturing three-dimensional
reinforcement. As shown in Figure 1, three-
dimensional fabrics are made by weaving many
filaments into a solid body. Several
manufacturing method for three-dimensional
fabrics have been developed; and the basic idea is
shown in Figure 2. As it can easily be seen from
the densely woven fabrics in Figure 1, the
original three-dimensional fabrics were for
reinforcing plastics. Therefore, in order to use
them to reinforce concrete structural members, Figure 1. Various three-dimensional
the fabrics should be woven with a much wider fabrics
fiber spacing to ensure the concrete or mortar
penetrates into the fabric. Also, since the original
three-dimensional fabrics are not impregnated by
resin in the manufacturing process, a polymer
should be impregnated in order to utilize the
strength of the filaments. These problems reduce
the efficiency at production of reinforcements.

3. A NEWLY DEVELOPED THREE-

DIMENSIONAL REINFORCEMENT

Reinforcement for concrete beam Figure 2. Weaving concept for three-

members consists of the longitudinal main dimensional fabrics
reinforcement and the shear reinforcement. In the
case of slabs or walls, two directional main reinforcements are normally incorporated in both
the compressive and tensile zones. The spacing of ordinary reinforcement is quite wide
compared with the three-dimensional fabrics produced by the method described above. In
practice, the spacing of the fiber strands should be several times the aggregate size to smoothly
cast the concrete. Three-dimensional reinforcements with wide spacing can be technically
produced using a three-dimensional weaving machine if some improvements of the machine
are made. However, improvements may decrease productivity. On the other hand, when the
spacing of the fiber strands is extended, thicker fibre strands are required to obtain the same
reinforcing effects. This means that the manufacturing system should be able to produce thicker
fibre strand. The method using the three-dimensional weaving machine, however, limits the
thickness of the strand and the width or depth of the reinforcement depending on the machine
size. To improve this, a new method to rationally produce the three-dimensional reinforcement
 407

has been investigated, and the following

method, called the base element three-
dimensional reinforcement method (BE3D
method,) is proposed.

3.1 THE CONCEPT OF THE BASE

ELEMENT (BE3D) METHOD

In general, concrete structural Figure 3. The basic component

members are reinforced by compressive, (called the Base Element)
tensile, and shear reinforcement. Given this,
in the proposed method, the basic
component, which consists of two main bars
and crossing bars trussing the two main bars,
is produced first.
The basic component shown in Figure 3 is
called the BE (base element). This base
element can be used as a component of
three-dimensional reinforcements with
different shapes. The simple three-
dimensional reinforcement is formed by
crossing the reinforcement called the
combined bar around the BE placed in
parallel; this is called the combined element
(CE).
Figure 4. Variation of the combined element
The simplest application is when the reinforcement
two main bars of the BE act as compressive
and tensile reinforcements, and the cross
bars of the BE and the combined bars of the
CE act as shear reinforcements. Further
complex shapes can be assembled by
combining these combined elements. For
example, as shown in Figure 4, a rectangular
section for a slab and beam, a T shape, a
circle section, and a hollow section are
possible. Figure 5 shows the combined
element for a precast hollow slab.

Figure 5. Hollow slab model reinforced with

3.2 MANUFACTURING METHOD OF
the Combined Element
THE BASE ELEMENT

As described in the above section, the base element consists of two main bars and the
crossing bars trussing the two main bars. The main bars of the base element are considered to be
408

the principal reinforcement, thus their quality should be as high as possible. To achieve high
productivity and high quality, the pultrusion method is adopted for the manufacturing the main
bars.

The two main bars are simultaneously extracted in parallel with the required spacing by
the pultrusion method. Next, in the intermediate process of the hardening of the main bars, after
which the fibre strand (called the cross bar) is impregnated with resin, it is wound around the
two main bars. Then the base element is cured and hardened again in a heater.

The cross bar is a reinforcement that unifies the main bars by winding around them,
eventually, its curvature at the position of the main bars becomes quite large. It is known that
the strength of FRP tends to decrease at the corner when it has a large curvature. In this method,
the section of the cross bar is made thin like a tape in order to avoid such effects. The cross bar
acts as a unifying component and provides shear reinforcement. Normally there are not many
cross bars. Figure 6 shows the basic construction process of the manufacturing machine of the
base element.

Winding of the cross bar No.l chuck Cutter

No.2 chuck

Main bar Cross bar BE

Quill & fiber roving
Ja
- " 1 jOQOQQOO

\
Resin impregnation bath / Automatic conveyor
No. 1 curing oven
Pultrusion die No.2 curing oven

Figure 6. Machinery manufacturing process for the Base Elements

3.3 MANUFACTURING METHOD OF THE COMBINED ELEMENT

The techniques for making the combined element is the same as those for making the
base element. The combined elements are made by winding the combined bar around the
several base elements placed in parallel. In the case of mass production, the combined bars,
after being impregnated with resin, are wound around the base elements continuously produced
by the multi BE production machine, then they are hardened in a hardening oven. Therefore,
the CE are continuously produced.
 409

A batch manufacturing system is adopted in the pilot plant or when complex combined
elements are manufactured. In the batch system, the combined elements are produced in the
following order. Several base elements cut to the required size are first set in the frame. In the
next stage, the combined bars impregnated with resin are wound around the base elements by
rotating the frame. They are then hardened in a heater.

3.4 ADVANTAGES OF THE BE3D METHOD

In the BE3D method, the manufacturing method can be easily be adjusted to provide
the main reinforcement and the shear reinforcement required by the design. Therefore,
reinforcements can be efficiently obtained. Furthermore, this manufacturing method is very
flexible in terms of the size or the shape of the manufactured reinforcement. As a result, the
BE3D method is a streamlined production system

4. MECHANICAL PROPERTIES OF A
BE3D REINFORCEMENT Table 1. Properties of carbon fibers
and epoxy resin
4.1 TENSILE PROPERTIES
Tensile
3.6 GPa
strength
It may be difficult to directly assess the
Carbon Elastic
mechanical and the physical properties of a fiber modulus
220 GPa
BE3D reinforcement itself, because the main Ultimate
bars, the cross bars and the combined bars have 1.3%
strain
different properties. Density 1.8
Diameter 7 I
In terms of design, it is much more
important to know the properties of the Tensile 54 MPa
Epoxy strength
individual reinforcing component. This section
resin Flexural
describes the basic properties of the components. strength
54 MPa
The most important properties are those of the
Elastic
main bars in the BE3D reinforcement. In the modulus
4.0 GPa
BE3D reinforcement developed at present,
carbon fibers and epoxy resin are used. The
mechanical properties of the carbon fibers and Table 2. Mechanical properties of
epoxy resin are shown in Table 1. To assess the FRP main bars (04mm)
properties of the main bars, the main bars were
Average Standard
produced by the pultrusion method in the same
values deviation
machine without the cross bars. Also, the cross
bars for the tensile tests were made in the same Tensile
1826 MPa
102 MPa
manner as the main bars. strength (n=15)

Elastic 3.3 GPa

131 GPa
A standard tensile test for FRP modulus (n=15)
reinforcements has not yet been established, but
410

some recommendations have recently

been proposed0}. For this test, the test
method proposed by the Japan Society of
Civil Engineers (JSCE) research
committee was referred to. In the case of
FRP materials, it is difficult to directly
grip the FRP bars with a chuck and load it
as with steel bars, because the
concentration of stress at the grips causes
earlier failure at the grips. Therefore,
anchoring parts are normally provided at
both ends of the FRP bar. In this test, the
end of specimen is put into an aluminum
pipe and fixed in the center of the pipe by Figure 7. Tensile specimen after failure
epoxy resin. The length of the test section ( 4mm plain FRP bar)
is 40 cm; the anchoring section is 20cm.
Loading tests were conducted using an Amsler type testing machine at a loading speed of 500
MPa/min (stress on the fiber), and at a room temperature of 20C. Both the extensometer (gage
length: 100mm) and the electric wire strain gages (gage length: 5 mm) were used to measure
the strain of the FRP bars. The 4mm diameter main bars and the cross bars were tested.

The results of the tensile test on the main bars are given in Table 2. Figure 7 shows the
specimen after tensile failure, breakage of the FRP in the middle can be seen. The average
tensile strength of the <j> 4mm main bars is 1866 MPa; this value is obtained by dividing the
maximum load by the nominal cross-sectional area of the bars. The fiber volume fraction of
these FRP bars is 55%, so the strength of the fiber is calculated to be 3395 MPa. Since the
original strength of the carbon fibers is 3670 MPa, the obtained value corresponds to 92% of the
original value.

The maximum load of the cross bars, which consists of 48000 filaments of <j> 7 m
carbon fibers, was 6227N. Since the cross bars have thin irregular sections, it is difficult to
determine their cross sectional area. Accordingly, when expressed as fiber strength, a reading
of 3550 MPa was obtained, and this corresponds to about 95 % of the original fiber strength.

The values of the elastic modulus shown in Table 2 are measured by the extensometer.
The average value is about 130 GPa; this was almost the same as the value measured by the
wire strain gages. Although the experimental results of 4mm diameter main bars are mainly
shown in this paper, diameters from 4 mm to 10 mm are available for the main bars of the BE3D
reinforcement. Also, main bar the fiber volumes of 50% to 65 % can be produced.

4.2 BONDING PROPERTIES OF THE BE3D REINFORCEMENT

In general, reinforcements must bond well with the concrete matrix if the reinforced
concrete is to perform well. In the case of the BE3D reinforcement, there are two bonding
mechanisms with concrete, i.e. mechanical bonding at the intersection points of the main bars
 411

and cross bars, and the physico-chemical bonding on the surface of the main bars. Although the
bonding properties required of a reinforcement usually depend on the proportions or type of the
structural members, there are cases in which higher bonding properties to the BE3D
reinforcement are required, because the BE3D reinforcement has quite a high strength.
Therefore, a method for increasing the mechanical bonding of the main bars was investigated.
Two methods for achieving this were devised. One method is to wind a thread with resin around
the FRP bars; and the other is to directly produce deformations (lugs) on the surface of the FRP
bars with epoxy resin. With regard to the deformed main bars, an analytical investigation using
the finite element method (FEM) was also conducted before the experiments. The results were
taken into account in manufacturing the deformed main bars. In this section, the analytical
results of the deformed main bars and the experimental results of BE3D reinforcement are
described.

4.2.1 FEM ANALYSIS ON THE

BONDING PROPERTIES OF
THE DEFORMED FRP BARS
Table 3. Analyzed cases
When the deformed main bars are Case Analized Load
Conditions
produced, many factors such as rigidity, No. effects (N)
spacing, the shapes of the deformations, Lug Angle
1 45, 60, 90 4904
and the adhesion area are thought to affect (deg)
the bond properties. In order to chose the Lug spacing
2 20,30,50, 110 2942
optimum conditions, an FEM analysis was (mm)
carried out. The analytical conditions and Lug height 9807
the mode analyzed are given in Table 3 and 3 (mm)
0.6, 1.0

Figure 8, respectively. Using NASTRAN Lug Rigidty

4 (elastic modulus of 1.0, 3.7, 10.0 9807
two dimensional plane analysis, the effects resin) (GPa)
of rigidity, spacing, height, and the shapes
of the deformations (lugs) on the FRP bars
were investigated with the bonding
properties. The following conclusions were
obtained. I I I I I I I 1 1 1 1 1 1 1 | 1 I 1 1 1 I 4

I IIII I
CONCRE K1
a) Effect of lug angie II II I I
To evaluate the bonding properties of the I I I I I I I I I I I I I I I I
1 lfi
(low Young's modulus)-L IIII' i II11 M Lii
deformed FRP bars, both stress in the | I I I I I \\\ I I I I I I I l l
11 i \\ \\r
concrete and at the adhesion interface l fTTf^W- spacir
MM ^ M:

between the lugs and the FRP bars are

M i l -I l!'
1'<1 1
i 1 l 1 ' 1 M ta
1 ' ' ' 1 i
f
1M
important. As shown in Figure 9, the shapes ! angle 1 j j \ \-
^ ^11
^ >|
1
' ^11
"*'r
of the lugs have very little effect on the
I { 1 i l l
r n r UU
n
shear stress in the concrete area above the 1
I
1
1
1 !
1 i . M "-UAD 1 **

lugs and the tensile stress in the lugs. This A A A---- ' ' '--,

indicates that the shape is not the principal

factor on the bond behavior. Figure 8. Typical modeling for FEM analysis
412

b) Effect of lug spacing

It is thought that the stress conditions in 100
the concrete will be altered by reducing the TENSILE STRESS OF LUG
number of the FRP bars. However, as indicated 80
in Figure 10, the shear stress on the concrete CO
CL
60
IT
over the lugs changes little as the ratio of lug
spacing to FRP bar diameter varies from 2 to 14. 40
The compressive stress on concrete in front of

n
SHEAR STRESS OF CONC.
the lugs changes in proportion to this ratio. The 20
same tendency is also observed as regards shear
stress at the interface of the lugs and the FRP 0
50 60 70 80 90 100
bars. Even though the applied load increases, the Angle of Lug
shear stress and compressive stress on the
concrete are thought to be far bellow the Figure 9. Effect of lug shape
assumed strength of the concrete (30 MPa
compressive, 7.5MPa : shear) for every ratio.
While the shear stress at the interface of lugs and
the bars has not yet been obtained by
experiment, the increasing shear stress at the
interface is so great that it may reach the limit
earlier than the stresses on the concrete. This
implies that the bonding failure at the interface
between the lugs and bars may occur.

c) Effect of lug height

Considering the actual manufacturing
4 8 12
variations of the lugs of the FRP bars, the effect
Lug Spacing/Bar Diameter
of the lug height is also analyzed in relation to
lug spacing. In the analysis, the stress Figure 10. Effect of lug spacing
distribution and displacement of the </> 10mm
FRP bars with the lug heights of 0.6 mm and 0.8
mm were compared. Differences in lug height
did not influence the stress distribution in the
concrete and displacement at the end of the bar.

d) Effect of the elastic modulus of the lugs

The elastic modulus of the epoxy resin
lugs is approximately 3.7 GPa.This value can be
controlled somewhat by changing the
composition of the resin. The effect of different
elastic modulus (1.0, 3.7,10.0 GPa) of the lugs 0 4 8 12
on bonding properties was checked in the Elastic Modulus (GPa)
analysis. As shown in Figure 11, even if the
values of the elastic modulus change from 1.0 to Figure 11. Effect of the elastic modulus
10.0 GPa, the shear stress in the concrete just of the lugs
 413

above the lugs does not change at all. The shear stress in the interface region of the lugs and the
FRP bar changes little with an increase in the value of the elastic modulus. The same tendency
can be seen in the compressive stress. This means that the availability of the elastic modulus in
the usual manufacturing method does not affect the bonding properties very much.

4.2.2 BONDING BEHAVIOR OF PULL-OUT TESTS

The bonding strength of the main bars of a BE3D reinforcement was evaluated by pullout
tests. Table 4 shows the test program for pullout bonding tests. In this experiment, three types
of FRP bars were tested: plain FRP bars, plain FRP bars intersected with cross bars, and
deformed bars. As explained earlier, the deformed bars were manufactured by two method, i.e.
winding the thread with resin around the FRP
bars, and directly producing lugs on the Table 4. Spacimens of pullout bond tests
surface of the FRP bars with resin. As
variables in the experiment on the deformed Name
(mm
Type of bars Shapes

FRP bars, the pattern of the lugs, the spacing 4N 4 Plain bar I I
of the lugs, the adhesion area between the lugs 10N 10
4D1
and the FRP bars were selected. Although a 4 10 mm
4D2 Deformed by
standard test method for the bonding of FRP
10 fiber winding l \ \ \ \ \ \ \ l
10D1
bars has not yet been established, the JSCE 10D2
Committee on Continuous Fiber Reinforcing

plain bar
10N-BE3 10
with cross bar
Materials has tentatively proposed a test
Deform spacing w-pattem
method for bonding based on the pullout test W20 10 -ed bar 20mm y<J\y\y\
(1)
of steel bars . In this experiment, the tests (lugs
Z1-20 20mm z-pattem
were carried out using the proposed test formed
15mm IIMMNMI
Z2-15 10 with
method.
Z3-15 resin) 15mm
wide area H
a) Bonding behavior of plain bars 15 or 20 mm

The bonding resistance of plain FRP

bars is considered to be a function of the

"
20
adhesion and friction between the bar's
surface and concrete. Figures 12 and 13 show I
the bond stress-slip relation -ship curves of the 15
d e f o r m e d 4 (4 D_1)
Q_
pullout tests for the <f> 4mm (4N) and <f> 10mm 2 1-A
plain FRP bars (ION). As illustrated in the 10
figures, the bars started to slip when stress CO deforied
reached a certain value; the slip displacement T3 0 4(4D-2)

then increased drastically without a noticeable < p l a i n 4(4N)-|-

increase of stress. The maximum bonding

stress of the plain FRP bars was about 4 MPa. | j |
0.05 0.10 0.15 0.20 0.25 0.30
Although this value was higher than the value Slip Displacement (mm)
of other plain FRP bars reported earlier^, it
was considered to be the result of the asperity Figure 12. Bond stress - slip displacement
of the FRP bars. curves (4 FRP bars)
414

b) Bonding of the plain bars with cross bar 14

! '. ! I 1
intersection
1(2 deformed --,
In order to check the effect of mechanical deformed
10(1 OD-2) lO(lOD-l) ;
bonding due to the intersection of the cross 10
bars, the plain bars intersecting cross bars were \ plain(cross bar)
8
tested. As shown in Figure 13, the bond stress - ; ; 10 (10N-BE3)
slip relationship curve is similar to those of the ; f T !

plain bars, but with a slight increase in bond

stress after slip started. The maximum bond y^- I plain 10 (ION)
stress was 6 - 7 MPa. This result suggest that the
i i !! i
intersection of the cross bars has an effect in 0.05 O.10 0.15 0.20 0.25 0.30
increasing the maximum bond stress of 1 - 2 Slip Displacement (mm)
MPa per intersection, but no effect on the initial
Figujre 13. Bond stress - slip displacement
bond stress once slip started. Once the adhesion curves ( FRP bars )
at the intersection of the cross bar fails, slip
behavior becomes similar to that of plain FRP 20,
bars.
Z2-15mm Z3-15mm-WIDE
c) Bond behavior of fibre wound deformed bars
Both thread wound and resin formed
a.
2
CO 15
\XBB.
deformed FRP bars have mechanical bonding
from the deformations (lugs) in addition to
c [/ Z1-20mm
1
surface adhesion and friction. As shown in -20mm

Figure 12, the shapes of the bond stress-slip

curves of the fiber wound deformed FRP bars
are different from those of the plain FRP bars. 0 0.05 0.10 0.15 0.20 0.25 0.30
Slip started at higher stress than on the plain Slip Displacement (mm)
bars. Even after slip started, the bond stress Figure 14. Bond stress -slip displacement
increased with increasing slip displacement. curves (the deformed FRP bars
Maximum stresses are achieved at the slip formed with resin)
displacement of 0.05 - 0.5 mm, and the values
are about 13 MPa for 4mm bars, and 9 -10 MPa for </> 10mm bars. Although two types of
deformed bars which were wound by thread with different thickness were tested, the difference
of the effect was not obvious. This may be supported by the conclusion on the height of lugs
height in the FEM analysis. In the examination of the internal failure condition of the pullout
specimens, complete exfoliation of the wound fibre strand from the main FRP bars was
observed. This phenomenon is considered to be completely different from that of deformed
steel bars.

d) Bond behavior of deformed bars formed by resin

Figure 14 shows the test results of the deformed bars formed by resin. In this test, two
deformation patterns ( w -pattern and z -pattern with the same lug spacing of: 20 mm) were
tested, but the difference in bond behavior was not obvious. The maximum bond stress of the
deformed bars is approximately 12-14 MPa for a lug spacing of 15mm, and 10 MPa for a lug
spacing of 20mm. These values correspond to the maximum bond strength of deformed steel
 415

bars. On the other hand, it was also concluded that the maximum bond stress becomes high if
the adhesion area between the lugs and the FRP bars increases.

5. FLEXURAL BEHAVIOR OF A SLAB REINFORCED BY A BE3D

REINFORCEMENT

The three-dimensional FRP reinforcement (BE3D) described in this paper can be used in
various structural members such as slabs and walls. However, since the mechanical properties
of FRP reinforcements differ from ordinary steel reinforcement as described in the above
sections, further research is required on the mechanical behavior, durability and design method
of FRP reinforced concrete. With regard to the BE3D reinforcement, flexural loading tests were
carried out using concrete slabs reinforced with BE3D reinforcement in order to obtain basic
data for the design. In this section, test
results are reported. ,300 (unit; mm)

* I *
5.1 Test program
-*"
The slab specimens used were 350 x
150 x2000 mm (b x h x 1); flexural loading
tests were carried out in the manner shown 010
"h Hi j H| CM

in Figure 15. The slabs were reinforced by

1
in * 10 -

30 30 30
BE3D as shown in Figure 15. The test 30
-+4
program is given in Table 5. Slab No. 1 has No.1(^10) No.2(^4)
a BE3D reinforcement consisting of 52
main bars of </> 4 mm; 26 of these were
placed in the tensile zone and the others in
the compressive zone. The BE3D
reinforcement of No.2 consists of 4 pieces
30 30
of the base element, including 8 main bars H-
of 10 mm. Of these main bars 4 are in No.3(01O)
the tensile zone and 4 are in the
compressive zone. In this test, the slab Figure 15. Test set-up and cross section of
the slab specimens

Table. 5 Flexural test program and results of the BE3D reinforced slabs
Cracking load Flexural strength Shear strength
Type of Reinfor- Predicted
(MN) (MN) (MN)
No. reinfor- cing ratio failure mode * * 'Hem ** ** Pum *** * * * j Psm
(mm) cement (%) PcrrvPcc '<5* Pum Puc PSC
1 Puc Psm J ^PTm
rc<
compressive
1 10 CE 0.6 16.0:13.2:1.21 143.6 130.4 1.10 154.8:
failure
compressive
2 4 CE 0.6 12.3JH.8h.04 149.5 133.1 1.12 154.8;
failure
only
3 10 0.6 shear failure 14.8J13.2J1.12 130.4 137 121.3J 1.13
BE
* Pcm: Measured cracking load ** Pern: Measured flexural strength *** Pcm: Measured shear strength
Pec : Calculated cracking load Pec : Calculatedflexuralstrength Pec : Calculated shear strength
416

specimens of No.l and No.2 were

designed to fail in No. 1
compression.Specimen No.3 was I
reinforced with only pieces of the
base element consisting of 2 main
bars of </> 10 mm. The reinforcement
A
ii/im\\\]\ i *
i
has no combined bars. Slab specimen
No.3 was prepared to check the effect No .2
of shear reinforcement of the
combined bars.
A
//r/iMTf\ i
i
5.2 Results of the flexural loading
tests
No .3
The test results are also given
in Table 5. Figure 16 illustrates the
cracking patterns of the loaded
i /7^/i/tm i
slabs.Slab specimens with BE3D
reinforcement show good crack Figure 16. Cracking pattern of the slabs
distributions. The crack spacing is
less than 50 - 100 mm; it is almost
the same as those of the cross bars or the
combined bars. The observed crack
spacing is considered smaller than for the
case of conventional steel reinforced
concrete members. Figure 17 shows the
load-deflection curves measured at the o

center of the slabs. The values calculated _j

for cracking load using the concrete

flexural strength had a good agreement
with the measured values. As illustrated in
Figure 17, after cracking, the load
increases linearly, with increasing
deflection until failure. In the case of FRP (mm)
reinforced concrete, since the reduction of Deflection
rigidity after cracking is smaller than with
Figure 17. Load-deflection curves of the
steel reinforced concrete, the limitation of
slabs
deflection under service conditions may
become a problem. The limit depends on the types of structural members or the required service
conditions. If the limitation of deflection is assumed to be less than 1/250 of the span length, the
service load level will be less than 1/3 to 1/4 of the ultimate load; and this is considered to be a
reasonable strength level. From this viewpoint, the FRP reinforcement can be directly applied
to structural members with relatively short span such as slab or wall for building construction.
When the span is much longer and the limit for deflection is more severe, an FRP
reinforcement with much higher elastic modulus may be used, or a partially prestressed FRP
 417

reinforcement may be introduced to the structural members to improve the performance in

service conditions.

The ultimate strength can be calculated from the equilibrium of internal forces using the
ordinary theory for steel reinforced concrete, whereby the distribution of linear strain in the
cross section and the equivalent stress block is assumed. The calculated values and the
measured value have reasonable agreement. The ultimate deflection of the specimens is about
1/40 for the span; this deflection is considered sufficient from the viewpoint of energy
absorption.

In the case of FRP reinforced concrete, the failure mode tends to be brittle, but if the
structural member greatly deforms at the ultimate state, the deflection may be useful for
noticing the failure. Although the adequate failure mode is a debatable point for FRP reinforced
concrete, there are no problems for practical use in terms of strength.

On the other hand, shear failure does not normally occur in the adopted specimen size
and the loading condition. However, since the high strength of FRP reinforcement increase the
ultimate strength of the member, shear failure may occur if the amount of the shear
reinforcement is less than required. In the test, the slabs of No.l and No.2 were designed to
avoid shear failure due to the shear resistance of both the cross bars and the combined bars.
Since slab No.3 had no combined bars, the amount of shear reinforcement was not sufficient to
prevent shear failure. As a result, slab No.3 failed in shear mode, although its ultimate strength
was almost the same as those of slabs No.l and No.2. In the flexural loading test, the cross bars
were broken before the concrete in the compressive zone reached the ultimate state, and the
slabs suddenly lost strength. This could be seen from the obvious shear cracking, shown in the
cracking pattern in Figure 16. In the case of the BE3D reinforcement, the combined bars and
the cross bars could be controlled to resist the shear force.

6. APPLICATIONS

At present, the authors are working to apply the BE3D reinforcement to the following
types of structures, where its properties such as high-strength, zero-magnetism and corrosion
resistance are of great utility.

(1) Application to structure with controlled

magnetic field
Recently-developed medical equipment for
Magnetic Resonance Imaging requires an
environment in which there are no magnetic
field or the magnetic field are highly stable.
Equipment which operates in the actively
controlled magnetic field also needs to be
placed in a room containing no materials which
can affect the magnetic filed. To meet these Figure 18. Two directional BE3D
requirements, the authors are developing a reinforcement
418

steel-free concrete building whose slab

members and walls are reinforced with the BE3D reinforcement
BE3D reinforcement in two directions, as
shown in Figure 18.
The non-magnetic features of the BE3D
mean it can also be used in guide ways for
magnetic-levitation ground transportation
systems ( linear motor train), because steel
reinforcement may lead to magnetic
resistance. At this stage, the pre-cast concrete 425 350 425
guide ways reinforced with the BE3D and 1700 (unit: mm)
FRP prestressing cables are under study.
Figure 19. PC bridge section reinforced
(2) Application to concrete structures in a with ^
corrosive environment
Deterioration of concrete structures in a corrosive environment is caused mainly by corrosion
of steel reinforcing bars and pre-stressed tendons. The BE3D reinforcement is suitable for use
in bridge decks and offshore structures. The authors are planning of the use of the BE3D as the
reinforcement of deck slabs for simply supported pre-stressed concrete bridge (see Figure 19).
In another area, there has been much demand for greenery on building walls and structures
adjacent to roads, therefore, concrete that promotes plants growth has been developed. Since
such concrete suitable for plants shows low-acidity, however, it is necessary to use non-
corrosive reinforcing materials such as the BE3D. This is another example of the possible uses
of the BE3D reinforcement.

7. CONCLUSION

There has been much interest in FRP reinforcement as an alternative material to steel
bars because it has high strength and non-magnetic and rust free properties. This paper has
discussed the development of a three-dimensional FRP reinforcement called the BE3D
reinforcement. The BE3D reinforcement, produced in a new streamlined manufacturing
system, shows good mechanical properties and excellent reinforcing effects. This three-
dimensional reinforcement is considered to be applicable to various structural members such as
slabs or walls, although the establishment of and the consensus on technical foundations such
as standard test methods or design criteria will be needed. Further research and development of
the BE3D reinforcement is continuing so its advantages can be fully derived and applied.

REFERENCES
(1) JSCE Research Subcommittee on Continuous Fiber Reinforcing Materials, Application of
Continuous Fiber Reinforcing Materials To Concrete Structures, Concrete Library of
JSCE, No.19, Tokyo, JAPAN, 1992, pp.15-28
(2) Itoh, H., Maruyama, M. and Okamura H., Study on Bond Characteristics of Deformed
Fiber Reinforced Plastic Rods, Proc. of JCI. Vol.11, No.l, 1989, pp. 777-782
 419

(3) Mizutani, T., Noritake, K., Honma, H., Asai, H., Anchoring Mechanism of Deformed
AFRP Rods, Proc. of JSCE Anual Congress, Vol.44, V, 1989, pp.552-553
(4) Akihama, S., Suenaga, T. and Nakagawa, H., Mechanical Properties of three-Dimensional
Fabric Reinforced Concrete Members, Proc. of JCI, Vol.10, N0.2, 1988, pp.677-682
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 423
1993 Elsevier Science Publishers B.V.

CFRP BONDED SHEETS

U. Meier, M. Deuring, H. Meier and G. Schwegler

Swiss Federal Laboratories for Materials Testing and Research (EMPA),

CH-8600 Duebendorf, Switzerland

Abstract
Since 1984 in the static and fatigue loading tests at the EMPA laboratories, carbon-
fiber reinforced plastic (CFRP) sheets have successfully been employed for the post-
strengthening of flexural beams with a span up to 7 m. The results of this com-
prehensive research program show that the calculation of flexure in reinforced
concrete elements post-strengthened with untensioned or with tensioned carbon-
fiber reinforced epoxy-resin sheets can be performed analogous to conventional
reinforced or prestressed concrete. In both cases long time fatigue tests on 7 m girders
displayed an outstanding fatigue performance of this strengthening technique. A
new approach also allows an increase of shear strength of the post-strengthened
systems. In 1991, for the first time real structures were strengthened with CFRP in
Switzerland e.g. the multispan box beam Ibach bridge near Lucerne with a total
length of 228 meters and the City Hall of Gossau St. Gall. 1992 followed the historic
wooden bridge near Sins which was strengthened for heavy trucks.

1. INTRODUCTION

In previous papers [2-4,6] the advantages and disadvantages of post-strengthen-

ing by means of steel plates have been discussed at length as well as the reasons for
replacing steel plates with carbon-fiber reinforced plastics (CFRP), above all in
bridge construction. The most important reason is clear from the following observa-
tions: within the framework of long-term creep tests on beams strengthened with
steel plates at the EMPA, the residual strength was determined after 15 years of
exposure to weathering. In several cases, traces of corrosion were discovered at the
joint steel/adhesive.

2. STRENGTHENING WITH NON-PRETENSIONED CFRP SHEETS

In the years 1984 to 1989, carbon-fiber reinforced epoxy-resin was successfully

employed for the first time at the EMPA for post-strengthening purposes. Loading
tests were performed on 26 flexure beams having a span of 2 m and 1 beam having
a span of 7 m [2]. The research work shows the validity of the strain-compatibility
424

method in the analysis of cross-sections. This implies that the calculation of flexure
in reinforced concrete elements post-strengthened with carbon-fiber reinforced
epoxy-resin sheets can be performed analogous to conventional reinforced concrete
elements. The work also shows that possible occurrence of shear cracks may lead to
a peeling-off of the strengthening sheet. Thus, the shear crack development repre-
sents a design criterion. Flexural cracks are spanned by the CFRP sheet and do not
influence the loading capacity. In comparison to the unstrengthened beams, the
strengthening sheets lead to a much finer cracking distribution [4]. A calculation
model [2], developed for the CFRP sheet anchoring agrees well with experimental
results.
Differences in the temperature expansion coefficients of concrete and the carbon-
fiber reinforced epoxy resin result in stresses at the joints with changes of tempera-
ture. After one hundred frost cycles ranging from +20 C to -25 C, no negative
influence of the loading capacity of the post-strengthened 3 beams was found [2].

The following failure modes were observed in the load tests (see also Figure 1):
- Tensile failure of the CFRP sheet (1). The sheets failed more or less suddenly, with
a sharp explosive snap. The impending failure was always announced far in
advance by cracking sounds.
- Classical concrete failure in the compressive zone (2) of the beam.
- Continuous peeling-off of the CFRP sheets due to an uneven concrete surface (3).
For thin sheets (less than 1 mm) applied with a vacuum bag, an extremely even
bonding surface is required. If the surface is too uneven, the sheet will slowly peel
off during loading.
- Shearing of the concrete in the tensile zone (3) (also observed as secondary failure).
- Interlaminar shear (4) within the CFRP sheet (observed as secondary failure).
- Failure of the reinforcing steel in the tensile zone (5). This failure mode was only
observed during fatigue tests.

The following failure modes were not yet observed but are theoretically possible:
- Cohesive failure within the adhesive (6)
- Adhesive failure at the interface CFRP sheet/adhesive (7)
- Adhesive failure at the interface CFRP concrete/adhesive (8).

For post-strengthening with CFRP sheets we recommend the design rule that the
CFRP sheets should fail during yielding of the steel reinforcing bars before crushing

rebar

concrete

rebar
adhesive
CFRP sheet
5 1 6 7 8
Figure 1: Failure modes
 425

Table 1
Properties of Hybrid and CFRP sheets
_> A Sheet type Nr.
Property ^ 2 3 4 5
2/3 T 300
Fiber type 1/3 E-glass T300 - T700 M46J
Fiber volume fraction [%] - 70 51 66 -
Longitudinal strength [MPa] 960 2,000 1,900 2,300 2,600
Longitudinal Young's modulus [GPa] 80 147.5 129 152 305
Strain at failure [%] 1.15 1.36 1.47 1.51 0.85
Density [g/cm3] - 1.58 1.46 1.45 __

of the concrete in the compressive zone. Yielding of the steel bars should not occur
before reaching the permitted loads.
Kaiser [2] investigated a beam with a 2.0 m span under fatigue loading. The cross-
section was 300 mm wide and 250 mm deep. The existing steel reinforcement
consisted of 2 rebars of 8 mm diameter in the tension as well as the compression
zones. This beam was post-strengthened with a hybrid sheet with the dimensions 0.3
x 200 mm (see Table 1, sheet type Nr. 1). The fatigue loading was sinusoidal at a
frequency of 4 Hz. The test set-up corresponded to a four-point flexure test with
loading at the one-third points. The calculated stresses in the hybrid sheet and the
steel reinforcement are listed in Table 2. After 480,000 cycles the first fatigue failure
occurred in one of the two rods in the tension zone. After 560,000 cycles the second
rod broke at another cross-section. After 610,000 cycles a further break was observed
in the first rod and after 720,000 cycles, in the second rod. The first damage appeared
after 750,000 cycles. It was in the form of fractures of individual rovings of the sheet.
The beam exhibited gaping cracks, which however were bridged by the hybrid sheet.
The relatively sharp concrete edges rubbed against the hybrid sheet at every cycle.
After 805,000 cycles, the sheet finally failed. This test was executed with unrealis-
tically high steel stresses. The goal of the test however was to gain insight into the
failure mechanism after a complete breaking of the steel reinforcement. This goal
was achieved. It was remarkable to observe how much the hybrid sheet could
withstand after failure of the steel reinforcement.
In 1991/92 [1] the EMPA performed a further fatigue test on a beam with a span
of 6.0 m under more realistic conditions. The dimensions and reinforcement of the

Table 2: Fatigue loading Table 3: Fatigue loading

and stresses of the 2 m beam and stresses of the 6 m beam
, ! r, XT, Stresses [MPa] T j M Stresses [MPa]
Loads [kN] rebarg hybridsheet Loads [kN] rebarg CFRP sheet

minimum 1 21 11 minimum 125.8 131 102

maximum 19 407 205 maximum 283.4 262 210
426

View
F/41
41 |F/4
IF/4 F/41 IF/4
IF

mH3I

3501
F/2
1360
I I
520
CFRP sheet 1=5750

2240 J520 1360

F/2
|35d

Cross section
TJ V75
^ 42la=6
d=6
- 42 d=10
3
- 4 d=26
- CFRP sheet 200 x 1

Figure 2: View and cross section of the beams with 6 m span

(all dimensions are in mm)

beam can be seen in Figure 2. The total carrying capacity F of this beam amounted
to 610 kN without the CFRP. Through the bonding of a CFRP sheet having the
dimensions 200 x 1 mm (sheet type Nr. 2, Table 1), the carrying capacity was
increased by 32% to 815 kN. In Figure 3, typical force/deflection diagrams are
presented for this type of beam. The calculated stresses in the CFRP sheet and the
steel reinforcement are given in Table 3. The beam was subjected to this loading for
10.7 million cycles. The crack development was observed after every 2 million cycles
[1]. After 10.7 million cycles the tests were continued in a climatic room. The
temperature was raised from room temperature to 40C and the relative humidity
to at least 95%. The goal of this test phase was to verify that the bonded CFRP sheet

O without post-strengthening
post-strengthened with non pretensioned CFRP sheets
post-strengthened with pretensioned CFRP sheets

30 40 50 60 70
Deflection [mm]

Figure 3: Typical load/deflection curves for beams shown in Figure 2

 427

(1) anchorage zone

(2) shear strengthening element of
advanced composites
(3) adhesive
(4) CFRP sheet for flexural
strengthening
(5) CFRP plate for load distribution

Figure 4a: Cross section of beam with shear strengthening arrangement

can withstand very high humidity with simultaneous fatigue loading. Already at the
beginning of this test phase the CFRP sheet was nearly completely saturated with
water. After a total of 12.0 million cycles the first reinforcement steel failed due to
fretting fatigue. The joint CFRP sheet/concrete did not present even the slightest of
problems. In the continuation of the test, the external loads were held constant
(Table 3) whereas the stresses in the reinforcement steel and the
CFRP sheet correspondingly increased. After 14.09 million cycles the second
reinforcement steel failed, likewise due to fretting fatigue. The cracks bridged by the
CFRP sheet rapidly grew. After failure of the third reinforcement rod due to yielding
of the remaining overloaded steel the CFRP sheet was sheared off.

3. STRENGTHENING WITH PRETENSIONED CFRP SHEETS

AND SHEAR STRENGTHENING
A further fatigue test was carried out at the EMPA, analogous to that described in
the previous section. The only difference is that the CFRP sheet was prestressed
(50% of strength of sheet type Nr. 2). 30 million cycles were performed without any
evidence of damage whatever.
In the last months a new method was developed and tried out in several tests. This
allows an effective strengthening of the shearing force areas without the use of steel.
One possible implementation is illustrated in the beam cross-section of Figures 4a
and 4b. The inner stirrup reinforcement is supplemented by a prestressed or non-
prestressed applied external strengthening, made of advanced composite materials.
These are braided or unidirectional in form. Depending on the application, carbon
fiber may be employed as well as aromatic polyamid and glass fibers. The pretensioning
material is wrapped around the cross-section on one side and anchored on the
opposite side in the compression zone. The modulus of elasticity and geometry should
be chosen so as to minimize the loss of tensioning force due to creep of the element
to be strengthened and relaxation of the prestressing material.
Many bridges in need of rehabilitation are deficient not only with respect to
flexural resistance. Often it is necessary to strengthen mainly the shear resistance.
In order to contribute at once to the strengthening and to relieve the inner stirrup
reinforcement, the reinforcement located externally to the element should be
additionally prestressed. In this way, crack formation in the shear force region can
be precluded or the cracks more finely distributed, in case shear cracks develop.
428

Figure 4b: View of T-beam with shear strengthening arrangement

In some cases it can be advantageous to provide additional prestressing to the

flexure-strengthening sheets. In this way, the serviceability of the structure can be
improved and the shearing off of the sheets due to shear failure of the concrete in the
tension zone can be avoided. An initial publication of this EMPA research effort
appeared recently [7]. Detailed results will soon follow [1]. The procedure for
applying a prestressed sheet is shown schematically in Figure 5. When the
pretensioning force is too high, failure of the beam due to pretension release will
occur at the two ends, because of the development of high shear stresses in the
concrete layer just above the CFRP sheet. Therefore the design and construction of
the end regions requires careful attention. Tests and calculations have shown that

Figure 5: Procedure for applying a prestressed sheet

(1) adhesive
(2) prestressed CFRP sheet
(3) shear strengthening arrangement according to Figure 4a
 429

without special end anchoring, CFRP sheets shear off from the end zones immedi-
ately with a prestress of over 5% of their failure strength. In order to achieve a
technically and economically rational prestress considerably higher degrees of
prestressing in the range of 50% are necessary.
At the EMPA end anchorings for flexure beams in accordance with Figures 4 and
5 were developed and successfully tested. In contrast to pure shear strengthening,
the advanced composites which wrap around the sheet must most definitely be
prestressed. This will built up a multi-axial stress condition in the concrete as much
as possible and also interlock cracks. This way failure at the two ends of the CFRP
sheets can be avoided.

4. APPLICATIONS

4.1 The I b a e h bridge

The bridge to be repaired (Figures 6a and 6b), located in the County of Lucerne, was
completed in 1969. It is designed as a continuous, multispan box beam with a total

Figure 6a: Ibach bridge near Lucerne/Switzerland. The holes in the accidentally
damaged web are arrowed.
430

Figure 6b: Ibach bridge near Lucerne / Switzerland. Strengthening work: the first
CFRP sheet is applied after completion of the preparation of the concrete surface
for the three laminates to be bonded to the bridge. The holes in the accidentally
damaged web are arrowed.

length of 228 meters. The damaged span of the bridge has a length of 39 meters. The
box section is 16 meters wide, with a central, longitudinal web.
Core borings were performed to mount new traffic signals. In the process, a
prestressing tendon in the outer web was accidentally damaged, with several of its
wires completely severed by means of an oxygen lance. As a result, the granting of
authorizations for special, heavy convoys across the bridge was suspended until after
completion of the repair work. Since the damaged span crosses Highway N2, the
traffic lanes in direction Lucerne on this highway had to be closed during the repair
work, which could therefore only be conducted at night.
Carbon fiber-reinforced plastics (CFRPs) are fifty times more expensive, per
kilogram, than the steel used to this date (Fe 360) for the reinforcement of existing
structures. Do the unquestionably superior properties of CFRPs justify their high
price? When one considers that, for the repair of the Ibach bridge, 175 kg of steel could
be replaced by a mere 6.2 kg of CFRP, the high prices suddenly no longer seem so
outrageous. Furthermore, all the work could be carried out from a mobile platform,
thus eliminating the need for expensive scaffolding.
The bridge was repaired with 3 CFRP sheets of the dimensions 150 x 5000 x 1.75
mm (2 sheets) and 150 x 5000 x 2.00 mm (1 sheet) according to sheet type Nr. 3 in
Table 1. A loading test with 840 kN vehicle demonstrated that rehabilitation work
with the CFRP sheets was very satisfactory. The experts participating in the repair
of the Ibach bridge were pleasantly surprised about the simplicity of applying the 2
mm thick and 150 mm wide CFRP sheets.
 431

4.2 Historie Wooden Bridge i n Sins

The covered wooden bridge in Sins/Switzerland (Figure 7) was built in 1807 in
accordance with the design of Josef Ritter of Lucerne. On the Sins side the original
supporting structure is almost completely preserved, even today. The Chams side
was blown up for strategic purposes on November 10, 1847 during the Civil war. In
1852, the destroyed half of the bridge was rebuilt with a modified supporting
structure. On the Sins side, the supporting structure consists of arches. These are
strengthened with suspended and trussed members. On the Chams side, the
supporting structure is made up of a combination of suspended and trussed members
with interlocking tensioning transoms. Originally the bridge was dimensioned for
horse-drawn vehicles. Today, vehicles with a load of 200 kN are permitted.

Figure 7: River Reuss with the historic wooden bridge in Sins/Switzerland with
two spans of 30.8 m each.

In the 185 year history of the bridge a great variety of rehabilitation efforts were
undertaken. Loading tests performed by the EMPA and the ETH Zrich indicated
that the pavement and several crossbeams no longer met the requirements of heavy
traffic. A project involving the construction of a pretensioned concrete bridge several
hundred meters upstream was opposed by the residents. Thus, in 1992 the wooden
bridge urgently had to be rehabilitated. It was decided to replace the old wooden
pavement with 20 cm thick bonded wooden planks, transversely pretensioned. This
technique described in the Ontario Bridge Design Code in 1983 was further
developed at the ETH Zrich. Two of the most highly loaded cross-beams were
strengthened by the EMPA using carbon fiber reinforced epoxy resin sheets. Each of
these cross-beams were constructed of two solid oak beams placed one upon the other.
432

A cross section of the bridge with the strengthening is shown in Figure 8. In order to
increase the thickness, wooden blocks were inserted between the beams. The lower
beams were 37 cm high and 30 cm wide, the upper beams 30 cm thick and 30 cm wide.
Cross-beam Nr. 14 was strengthened with 1.0 mm thick CFRP sheets made of
high-modulus fibers (sheet type Nr. 5, Table 1); on the upper side the width amounted
to 250 mm and and on the lower side 200 mm. Cross-beam Nr. 15 was strengthened
with 1.0 mm thick CFRP sheets made of high-strength fibers (sheet type Nr. 4, Table
1); on the upper side the width amounted to 300 mm and and on the lower side 200
mm.
Before bonding the sheets, the bonding surface was planed with a portable system.
The bonding and installation were accomplished with the same material and the
analogous clamping system as for the Ibach bridge (Figure 6b).
The strengthened cross-beams of the Sins bridge, subjected to extremely high
loading and reinforced with CFRP sheets, help to gain practical experience and
confidence in this method for preservation of historic bridges. Thus, in the future,
similar structures may be rehabilitated in this manner.
The historic wooden bridge in Sins is a valuable structure, both from esthetic and
technical
viewpoints. It is also of value historically and under protection as a national
monument. For the post-strengthening of such structures the technique with CRP
sheets is especially suited since the thin but extremely stiff and strong sheets are
hardy noticed and therefore do not detract from the original design of the structure.

CFRP

Figure 8: Cross section of historic wooden bridge in Sins.

Selected cross beams were strengthened with CFRP sheets
 433

Figure 9: Concrete slab of the City Hall of Gossau strengthened with black CFRP
sheets. The part of the slab, which will be cut in, is marked with a dashed line.
This work was executed by Stahlton AG, 8034 Zrich/Switzerland

4.3 The City Hall of Gossau

The City Hall of Gossau St. Gall was the first building in Switzerland where the
CFRP sheet strengthening technique was used. Within a renovation program it was
decided that an elevator should be added. As a result a rectangular hole had to be cut
in a concrete slab. Before cutting, the future edges of the hole were strengthened with
CFRP sheets as shown in Figures 9. Due to aesthetic reasons the architects were
persuaded to use thin high strength CFRP sheets instead of thick steel plates. After
painting the CFRP sheets "disappear".

5. OUTLOOK

In the expanding range of non-metallic tensile elements based on high strength

fibers, CFRP sheets, the non-corrosive tensile elements composed of endless parallel
carbon filaments in a epoxy resin, may be considered to be a very promising
alternative to steel plates for strengthening applications where long-term durability
is required. Carbon sheets combine the qualities of very high strength with an
outstanding fatigue performance and light weight for easy handling. The sheets are
durable under practically every type of environmental attack which may occur in or
around concrete structures.
In Switzerland the application of non-tensioned CFRP sheets to strengthen
existing structures successfully crossed the threshold to application in 1991. For
434

wider acceptance of this efficient strengthening method European design recom-

mendations and codes are required.
Concerning the bending and shear strengthening method with pretensioned
CFRP sheets we are not yet as far. It will take at least another two years of R&D work
to cross the threshold to applications since the handling of the above described
pretensioning method is not yet practical.
Today the use of CFRP sheets in civil engineering is in its infancy, but there are
clear indications that it will be an excellent choice for a multitude of rehabilitation
projects on bridges, dams, tubes, high pressure pipes, buildings or even historic
monuments.

6. REFERENCES

1 Deuring, M. (1993). Post-strengthening of Concrete Structures with Pretensioned

Advanced Composites. Will be published in German in summer 1993 by the EMPA
as Research Report No. 224, EMPA Dbendorf, CH-8600 Dubendorf/Switzerland.
2 Kaiser, H.P. (1989). Strengthening of Reinforced Concrete with Epoxy-Bonded
Carbon-Fiber Plastics. Doctoral Thesis, Diss. ETH Nr. 8918, ETH Zrich, CH-
8092 Zurich/Switzerland (in German).
3 Meier, U. (1987). Bridge Repair with High Performance Composite Materials.
Material und Technik, 15, 125-128 (in German and in French).
4 Meier, U. and Kaiser H.P. (1991). Strengthening of Structures with CFRP
Laminates. Proceedings Advanced Composite Materials in Civil Engineering
Structures. MT Div/ASCE/Las Vegas, J a n 31, 1991.
5 Meier, U. and Deuring, M. (1991). The application of fiber composites in bridge
repair. Strasse und Verkehr, 77, page 775.
6 Meier, U. (1992). Carbon Fiber-Reinforced Polymers: Modern Materials in Bridge
Engineering. Structural Engineering International, 2, 7-12.
7 TriantafiUou, T.C., Deskovic, N. and Deuring, M. (1992). Strengthening of
Concrete Structures with Prestressed Fiber reinforced Plastic Sheets. ACI Struc-
tural Journal, 89, 235-244.
Fiber-Reinforced-Plastic (FRP) Reinforcement for
Concrete Structures: Properties and Applications
A. Nanni (Editor) 435
1993 Elsevier Science Publishers B.V. All rights reserved.

A retrofitting method for reinforced concrete structures using carbon fiber

Y. Kobatake, K. Kimura and H. Katsumata

Structural Engineering Department, Technical Research Institute of Obayashi Co.,

4-640 Shimokiyoto, Kiyose, Tokyo, Japan

1. INTRODUCTION

Some of existing reinforced concrete chimneys in Japan lack sufficient earth-

quake resistance and have been often damaged or partially destroyed to as much
as 2/3 or more of their total height during severe earthquakes. The design of
these chimneys was going by the design regulations of these days which did not
demand enough flexural strength in the top part of chimneys, thereby the retro-
fit of longitudinal reinforcement should be performd for such chimneys in the
dengerous regions occuring a severe earthquake. As the retrofitting work is gen-
erally done at high and narrow places, the retrofitting material is required
to possess (l)high strength, (2)light weight and (3)durability.
In 1987, a new retrofitting method using carbon fiber, featuring high strength
and light weight, was developed and introduced into the construction practice by
Obayashi Co. and Mitsubishi Kasei Co. of Japan. The method is applied to retro-
fit chimneys using an automatic carbon fiber winding machine.
So far, the method has been applied to 14 cases in which retrofitting of chim-
neys was needed and, in 1991, has been evaluated as one of the most effective
retrofitting methods by Building Disaster Prevention Association of Japan. Among
the retrofitting methods using advanced materials (carbon, aramid, etc.) for
existing reinforced concrete chimneys or columns of buildings, the method devel-
oped by authors is only one evaluated publicly in Japan at present.
Furthermore, reinforced concrete highway bridge columns in Japan have a por-
tion in which main reinforcing bars are terminated at mid-height for economic
reasons.
In some cases, these parts do not possess sufficient lateral resistance and duc-
tility, thus, such bridge columns would have a heavier damage during a severe
earthquake. Expanding the application field of this method to ductility and str-
ength retrofitting for highway bridge columns is under cooperative study by Oba-
yashi Co. and Highway Public Co. of Japan, since 1990.
In this paper, the preceding applications of this retrofitting method are pre-
sented.
436

2. RETROFIT OF CHIMNEYS

2.1. RETROFITTING METHOD FOR EXISTING CHIMNEYS

Ordinary retrofitting methods for existing chimneys are:

(1) After cutting the upper part of a chimney, the place is exchanged for a new
stainless steel tube. If employing this method, the use of chimney should be
stopped.
(2) The chimney is encased with a steel tube and mortar is grouted into the gap
between the concrete surface of chimney and the tube. If employing this method,
the increase of retrofitted chimney's weight should be considered, and the re-
trofitting zone enlarged.
The new method developed by the authors, as illustrated in Figure 1, is carri-
ed out by gluing carbon fiber UD tape (Unidirectionally arranged carbon fibers
impregnated with epoxy resin) onto the existing concrete surface. This method
can overcome above problems and have following various merits:
(1) Use of the chimney does not need to be stopped since only the outside of
chimney is retrofitted.
(2) The retrofitting zone is not enlarged since increased weight accompanied
with retrofitting is negligible because of the light weight material.
(3) Durability of concrete is improved because the concrete surface is covered
with carbon fiber UD tapes.
(4) Superior durability is expected even under severe environments, like on-
shore or industrial regions.
A detailed procedure of retrofit for chimneys is presented as follows:
(1) Removing of the lightning conductor and ladder
(2) Substrate treatment and/or arrangement of concrete surface
The substrate treatment is to paint primer onto the concrete surface.

Figure 1. The concept of retroffiting Figure 2. Application of UD tape

method using carbon fiber
 437

Figure 3. Winding of carbon fiber strands Figure 4. The view of execution

This primer which is one of epoxy resin penetrates into concrete and helps adhe-
sion between concrete and carbon fiber UD tape.
(3) Gluing carbon fiber UD tape in the longitudinal direction
After adhesive epoxy resin is painted on the concrete surface, carbon fiber
UD tapes are glued along the axial direction on the whole surface of the retro-
fitting zone. As the need arises, the procedure of gluing carbon fiber UD tape
is repeated, until the retrofitted structure has sufficient strength against the
demanded force.
(4) Winding carbon fiber strand impregnated with epoxy resin along the hoop
direction.
(5) Restoration of the lightning conductor and ladder.
(6) Painting according to provisions for safety of airplanes (paint airbeacon).
Different phases in the execution of this method are shown in Figure 2 through
4.

2.2. THE EFFECT OF RETROFIT ON FLEXURAL STRENGTH

The structural efficiencies on flexural strength have been confirmed through

experiments. The specimens, modeled on an existing reinforced concrete chimney,
are three circular hollow reinforced concrete beams, as shown in Figure 5.
Within the concrete, 10-D3 bars are arranged as longitudinal reinforcement and
3.2mm0 spiral hoop bar is arranged as transverse reinforcement. The test para-
meter of specimens is the amount of carbon fiber glued onto the outside surface.
Material properties are shown in Table 1.
The experimental apparatus is shown in Figure 6. Monotonous load is applied at
two points of the specimen: bending moment is constant between loading points.
The load-displacement relationship is shown in Figure 7. From this figure, it is
found that the maximum strength is much improved and the displacement at maximum
strength becomes large as the amount of longitudinal carbon fiber is increased.
438

Reinforced concrete Thickness = 28 mm

X

200 | 500 | 400 |500 [200

1,800

4-"
3
2 @7
unit; mm

m
Reinforcement

Figure 5. Dimensions of specimen Figure 6. Experimental apparatus

Table 1.
Material properties

Yield strength Tensile strength Young's modulus

Steel bar 283 MPa 437 MPa 214X10 3 MPa

Carbon fiber 2971 240 X 1 0 3
(Pitched fiber)

Compressive Strength Young's modulus

Concrete 16. 7 MPa 16.6X10 3 MPa

Table 2. Carrying load

Specimen Longitudinal Flexural cracking Maximum load

fiber ratio(%) (a) P E (b) PA (a)/(b) (c) P E (d) pA (c)/(d) a

Un-retrofitted 0 22.6 kN 18.1 kN 1.24 21.3 k 17.8 kN 1.20

Retrofitted 1 0.4 32.4 18.5 1.75 106.1 118.1 0.90 0.84
Retrofitted 2 0.8 38.4 18.9 1.50 176.5 207.1 0.85 0.82

P E and PA : experimental and analytical value, respectively

a ' effective ratio of carbon fiber

Flexural cracking and maximum load are shown in Table 2, compared with the
analytical values. The analysis is carried out as the same way as the ordinary
flexural analysis of a reinforced concrete section, that is, the strength of
 439

retrofitted specimen is obtained by a superposition of the strength of longitu-

dinal reinforcing bars and carbon fiber.
The maximum carrying load of the un-retrofitted specimen is much smaller than
the others; and the load of the most retrofitted specimen is the highest. The
experimental maximum load, including another three specimens' data besides those
in Table 2, is 0.85 to 0.93 times of the analytical value. In order to express
effectiveness of retrofit with carbon fiber, the effective ratio a of carbon
fiber is defined from experimental and analytical values as the next equation:

a = {P(e)-Pc(e)} / {P(a)-Pc(a)}

where, P(e)= experimental maximum carrying load of the retrofitted specimen

P(a)= analytical maximum carrying load of the retrofitted specimen
Pc(e)= experimental ultimate carrying load of the un-retrofitted specimen
Pc(a)= analytical ultimate carrying load of the un-retrofitted specimen

Load (kN) <

Retrofitted 2
200

150
3 / Retrofitted 1
/ -^ /
O
100
^ / ^- ' V
/ .* X ((.
50 Un-retrofitted

Displacement (mm)
-H 1 (-
10 15 20

Figure 7. Load-displacement relationship

The value a, including another specimens' data besides in above table, is

0.82 to 0.88 and the average value is 0.85. We can estimate the flexural
strength of specimen gluing carbon fiber UD tape as follows:

P = Pc + a P f

where, Pc = flexural carrying capacity of reinforced concrete beams

Pf = contribution to flexural capacity of carbon fiber, employing the
full strength of carbon fiber
440

2.3. THE FUNDAMENTAL PERFORMANCE OF RETROFITTING METHOD

It is important for efficient retrofit that the carbon fiber adhere reliably
the concrete surface. The following fundamental parameters should be determinded
in order to establish the retrofitting method:
(1) The length of the carbon fiber fixed to concrete
(2) The length of a lapped joint to carbon fiber UD tapes
(3) The durability of the adhesive strength between concrete surface and carbon
fiber.

The test on carbon fiber development length

The outline of the specimen is shown in Figure 8. The specimen is elongated by

application of tension at each end. The variable parameter is the carbon fiber
length (L) adhered to concrete. The state of damage is shown in Figure 9.

x
Load
L.UU y\

Re-bar

40
Development
length UD tape
Notch
!-3 15
m Concrete block
/
UD tape
| 100 1
Unit: mm

Figure 8. The outline of specimen Figure 9. The state of damage

Table 3.
Test results

The carbon fiber Maximum load(kN) The state of damage

development length(L)(cm)

4 11.0 carbon fiber peels off

10 13.6 carbon fiber peels off
20 14.9 carbon fiber breaks
 441

The experimental results are shown in Table 3. From these results, we find
that the carbon fiber development length should be more than 20 cm.

The test on development length of lapped carbon fiber joint

The outline of the specimen is shown in Figure 10. The test is performed the
same way as before. The state of damage is shown in Figure 11.
The experimental results are shown in Table 4. From these results, we find
that the length of a lapped joint to carbon fiber UD tapes is necessarily more
than 10 cm.

100
Load

Lapped joint
length UD tape

cocrete block

UD tape
If
Unit: mm
Re-bar

Figure 10. The outline of specimen Figure 11. The state of damage

Table 4.
Test results

Development length of Maximum load The state of damage

lapped joint (cm) (kN)

4 11.8 carbon fiber peels off

10 16.5 carbon fiber breaks
16 16.3 carbon fiber breaks

The test on the durability of the adhesive strength

The outline of specimen and a way of test is shown in Figure 12. The specimen
442

consists of two concrete blocks, 4 0 X 4 0 X 100mm, which are jointed with a sheet
of carbon fiber UD tape. The variable is the time of accelerated artificial
exposure.
The test was performed to expose the specimen to ultraviolet rays using expo-
sure apparatus (Carbon Arc type). After exposure, the specimen was elongated by
the application of tension at each end.
The experimental results are shown in Figure 13. From these results, we find
that the adhesive strength between concrete surface and carbon fiber decreases
with the time of accelerated artificial exposure, a decrease of about 10 % at
2,000 hours and about 15 % at 4,000 hours compared with the non-exposure speci-
men.

Load y\
Maximum load (kN)
25
5
If Concrete block
Mean value
UD_ 1 4
HE &
tape CD
CD

f LO UD t a p e specimen 3
CSi

LO
CSJ
CD
2
CD

4 1
( X 1 0 3 hr)

hi Specimen Unit: mm 0 -JJ*-

0 1 2 3 4
Time of accelerated
Loading test
artificial exposure

Figure 12. The outline of specimen Figure 13. Maximum load time of
accelerated exposure relationship

2.4. DESIGN CONCEPT FOR RETROFIT USING CARBON FIBER

The basic concept of the retrofitting method (Figure 1) consists of increasing

the longitudinal reinforcement of the chimney by gluing carbon fiber UD tapes
to its outside surface along the axial direction. Whenever shear and/or thermal
stress along the hoop direction cannot be ignored, transverse reinforcing is
performed by winding the carbon fiber strands around the chimney's body.
And then, the carrying load of retrofitted chimney is to be above the carrying
load demanded by current design regulations.

2.5. RESULTS ON THE RETROFIT OF CHIMNEYS

The actual results on the retrofit of chimneys executed from 1983 to 1992 are
 443

shown in Table 5. This method has been used often lately, not only for retrofit
of chimneys but also repair of concrete structures, for example a slab or ex-
ternal wall of a building, or sacred arch at a Shinto shrine.
The amount of carbon fiber used for the retrofit of chimneys from 1987 to 1992
is to be about 1, 650 kg.

Table 5 : Field results

Use of chimney Site Date of Height(m) Diameter(m)

execution (at the base)

Equipment for heating Tokyo 1987.12 15.0 1.4

(Technical research institute)
Industrial plant Aichi Pref. 1988.4 33.4 3.0
Chemical plant Chiba Pref. 1988.12 45.0 2.4
Equipment for heating Chiba Pref. 1989.3 35.0 3.0
(Airport terminal)
Chemical plant Osaka 1989.9 37.0 3.7
Chemical plant Shizuoka Pref. 1989.11 90.0 6.2
Pharmaceutical plant Mie Pref. 1990.7 55.0 3.9
Oil manufacture plant Okinawa Pref. 1991.3 40.0 4.2
Textile plant Shizuoka Pref. 1991.4 60.0 3.9
Iron plant Fukuoka Pref. 1991.7 70.0 7.4
Industrial plant Kanagawa Pref. 1991.12 50.0 3.8
Industrial plant Yamaguchi Pref. 1991.12 50.0 4.2
Chemical plant Niigata Pref. 1992.8 30.0 2.7
Thermal power plant Okinawa Pref. 1992.9 42.0 3.7

3. RETROFIT OF BRIDGE COLUMNS

3.1. RETROFITTING METHOD FOR EXISTING BRIDGE COLUMNS

In generally, the next two retrofitting techniques for reinforced concrete

bridge columns are used:
(1) Encasing with reinforced concrete (reinforced concrete jacketing)
Around the existing column, longitudinal and transverse reinforcement bars are
arranged and concrete is placed. The thickness of the additional part is 100 to
150 mm increase weight should be less.
(2) Encasing with steel tube (steel plate jacketing)
Around the existing column, steel tube is constructed by welding steel plates,
and epoxy resin is pumped into the gap of roughly 3 mm thickness between the
steel and the existing concrete.
444

The former is inexpensive method, but has the following problems:

Increase in dimensions and weight of the retrofitting column is not negli-
gible. It influences the natural frequency of the bridge, that is, input level
of earthquake ground motions. So, it is necessary to re-evaluate the safety mar-
gin of the substructures and foundations.
The work for integration of existing and new concrete, which are to chip
the joint surface of existing concrete and to fix anchor bolts into existing
concrete, is difficult to carry out. And also the process is noisy and may cause
environmental problem in some cases.
On the other hand, when the latter is used, the increase in dimensions and
weight is small. However, the following problems arise:
This method is expensive.
The integration, fixing anchor bolts into existing concrete and injecting
adhesives or grouting mortar to the gap between existing concrete and new steel
plates, is difficult to carry out well.
In order to solve these problems and reduce construction labor, the authors
have studied a new retrofitting technique using carbon fiber, which is one of
brand-new and high-tech materials. This method (the outline is shown in Figure
14) utilizes excellent advantage of carbon fibers such as high strength(2942 MPa),
high modulus of elasticity (245X10 3 MPa), lightness and excellent durability,
providing the following merits:
(1) Increase in dimensions and weight of retrofitted columns is negligible.
(2) Construction work is fairly easy.
(3) Superior durability is obtained.
In order to verify the retrofitting performance of this new technique with
carbon fiber, static and cyclic loading tests, employing 1/3 scaled specimens,
were carried out. The outline of static and cyclic loading tests and the concept
of design are described here. Other investigations on materials and durability
of retrofitted members and further details on retrofitting bridge columns have
been carried out. Some of them are referred to in Ref.5.

Iixistin concrete pier

Figure 14. The outline of new retrofitting method for bridge column
 445

3.2. THE OUTLINE OF TESTS

Specimen

The scale of the specimens tested is one-third of the actual bridge columns.
The dimensions and the bar arrangement of the specimens are shown in Figure 15.
In specimens No.1, and No.3, there was a termination of main reinforcement at
the height of 90 cm from the foundation, at which the amount of longitudinal
reinforcement was reduced to 1/2 of the lower part. In specimen No.2, there was
no termination of main bar to investigate performance of retrofit at the bottom
part of the column. The longitudinal reinforcement were D16 deformed bars (grade
SD30). The transverse reinforcement were 6 mm round bars (grade SR24) with 20
cm spacing; percentages of longitudinal and transverse reinforcement were 1. 2 %
and 0.047 %, respectively.
The parameters of the specimen, range of reinforcement and the aim of retrofit,
are summarized in Table 6. Specimen No. 1 is an unstrengthened specimen, and spe-
cimen No.2 is only transversely strengthened by carbon fiber strands in the 60cm

Table 6.
Specimens

No. of Termination Retrofit Range of reinforcement

specimen portion Termination portion Bottom
Upper (mm) Lonwer(mm)

1 Included Nothing
2 Not included S 600
3 Included F+S 300 250 650

Note: F; Flexural reinforcement S; Shear reinforcement

Unit: mm
Specimen No. 3 Specimen No. 2

Figure 15. Dimension of specimen

446

Figure 16. Loading apparatus

region from the base. The specimen No.3 is longitudinally strengthened by doubly
layered UD tapes in the upper 30cm and lower 25cm regions from the termination
section and is transversely strengthened by carbon fiber strands from the upper
30cm of the termination section to the base as shown in Figure 15.

Loading

Loading apparatus is shown in Figure 16. For specimens, lateral and cyclic
load were applied. After several loadings with the allowable tensile stress
level of the reinforcing bar, a repeated loading regime was performed with dis-
placement on the basis of the yield displacement yt2yt . The constant
axial compressive force of 14.4tf (averaged axial stress is 0.59 MPa), which
corresponds to reaction force against the dead load of the superstructures of
the prototypes, was applied in all cases.

Test results

Test results are summarized in Table 7, and load displacement relationships

are shown in Figures 17 through 19.

Failure pattern

Specimen No. 1 : At the loading stage of 2<5y, the longitudinal reinforcement

above the termination section yielded, and the cracks propagated around that
section. At 3(5y, concrete crushed due to buckling of the compressive reinforce-
 447

ment above the termination section during the first cycle loading. Then the sus-
tained load decreased. At the termination section, the spalling of the concrete
coverage and buckling of compressive longitudinal reinforcement was observed.
Specimen No. 2 : In the negative loading cycle of 5<5y, the concrete in the
bottom part of the column expanded because of buckling of longitudinal reinfor-
cement. On the same time, fracture of carbon fibers and heavy crush of concrete
took place and lateral resistance decreased.

J| J I I I I I I I ! 1 I L

Figure 17. Load-displacement Figure 18. Load-displacement

relationship(No. 1) relationship (No.2)

Figure 19. Load-displacement relationship (No.3)

448

Specimen No. 3 : The longitudinal reinforcement yielded just above the streng-
thened section and at the base almost the same time. When the load increased up
to 4(5 y , the bond between UD tape and concrete depressed causing the decrease of
load. With further load increase, large flexural cracks occured at the lower end
of the UD tape-strengthened section followed by the buckling of reinforcement
during the last cycle of 6<5 y . However, even at the displacement cycle of 7<5 y ,
the reduction of the sustained load was fairly small. The carbon fibers locally
ruptured.

Table 7.
Test results

No,. of Crack Yieild Maximum Ultimate Failure

specimen Load Load Disp. Load Disp. Load Disp. position
(kN) (kN) (mm) (kN) (mm) (kN) (mm)

1 24.5 110.8 22.8 114.7 68.1 114.7 68.3 termination

2 19.6 127.5 26.2 142.2 130.0 142.2 131.0 bottom
3 24.5 127.5 27.0 134.4 53.7 101.0 ^201 bottom

Crack load is the load when the first crack appears.

Yield load is the load when the longitudinal reinforcement of the tensile side
begins to yield.
Ultimate load is the load when the concrete is heavily damaged, for example, it
crushes or a crack opens widely; or carbon fibers begin to fracture critically.

Retrofitting effect

Comparing with specimen No. 1, improvement of flexural strength (maximum load)

by applying carbon fiber sheets in the longitudinal direction is verified, and
damaged part of specimen No.3 moved from the main bar termination to the bottom
part of the column. Ductility was also improved. From these results, it is con-
cluded that retrofitting method using carbon fiber is effective for seismic re-
trofit.

3.3. THE CONCEPT OF DESIGN

The basic concept of design for retrofitting method using carbon fiber is
ductility-oriented. In order to avoid brittle failure, bridge columns are retro-
fitted as follows:
(1) The shear strength of the columns should be greater than the flexural
strength.
(2) The lateral resistance of the columns should be maintained when the columns
are subjected to large deformation.
 449

(3) The lateral resistance should be determined by the flexural strength of the
bottom part of the columns. Because shear span is long and superior ductility
can be easily obtained.
So, if bridge columns do not have enough flexural resistance at the main bar
termination, flexural strengthening is needed, and usually, shear strengthening
is also conducted.
For the bottom part of the columns, the transverse strengthening option is
employed in order to improve ductility. This option has two meaningful effects;
one is shear strengthening, and the other is confinement of concrete.
The amount of shear strengthening was calculated, based on the following
concept:
(1) bridge columns should not fail in shear
(2) shear force acting in the columns should be only resisted by the retrofit-
ting carbon fibers
(3) the amount of carbon fibers can be converted to the amount of reinforcing
steel bars by Eq.1, just as the case for flexural strengthening

ACF = A 8 ( o r B y / aCF) (1)

where, A C F required amount of carbon fibers

A8 required amount of reinforcing steel bars
Cf ay yield strength of reinforcing steel bars
ffCF design strength of carbon fibers during an earthquake.

(4) the contribution of shear reinforcement to the shear capacity is the same
as defined in the current code:

V s = P w aWy-b-d/1.15 (2)

where, Pw contribution of shear reinforcement to shear capacity

Pw percentage of shear reinforcement
Cf wy yield strength of shear reinforcement
b column width
d effective depth of column

4. CONCLUSIONS

The outline of the retrofitting method using carbon fiber and its applications
are described. Lately the retrofit of concrete structures, such as, reinforced
concrete chimneys, bridge columns, bridge girders and slabs, has become the cen-
tre of wide interest in Japan. Therefore, we think the retrofitting method using
advanced materials (carbon, aramid, etc.) will be adopted widely in the future.
As problems yet to be solved in the future, the following points are left:
(1) The decline of the materials' price
450

(2) The disposal of disused materials. From the view of environmental protec-
tion for the earth, the disposal of Fiber Reinforced Plastics (FRP) is be-
coming a world-wide problem.

ACKNOWLEDGEMENTS

We wish to thank Mr. Kensuke Yagi and his fellows of Mitsubishi Kasei Co. for
their considerable assistances with respect to the development of the retrofit-
ting method for chimneys. We also would like to express 'profound appreciation to
Tetsuo Matsuda and his fellows of Japan Highway Public Co. with respect to the
investigations of the retrofit for bridge columns.

References

1 K. Kimura, Y. Kobatake, H. katsumata, T. Takeda, K. Yagi, T. Tanaka, and

T. Sawanobori, Summaries of technical papers of annual meeting, Architectural
Institute of Japan, (1988)
2 K. Kimura, Y. Kobatake, M. Yosizaki, and J. Nomura, Report of Obayashi Corpo-
ration Technical Research Institute, No.37(1988)
3 N. Higashida, H. Ishida, Y. Kobatake, Y. Okajima, H. Katsumata, and K. Yagi,
Proceedings of the 45th annual meeting, the Japan civil engineering
association,(1990)
4 Japan Road Association, Design Specifications for Highway Bridges (1990)
5 H. Katsumata, S. Ohno, Y. Kobatake, T. Matsuda, H. Fujiwara, and N. Higashida,
Proceedings of International Symposium on Earthquake Disaster Prevention at
CENAPRED,(1992)