Beruflich Dokumente
Kultur Dokumente
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
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.
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.
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.
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
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
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
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.
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
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
Vacuum = 6 0 9 mm Hg
Pressure = 621 KPa
Resin Cure
177
Q:
Resin h \
< Flow / ! !\
Q: 121
LU
a.
UJ
21
TIME
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
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
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
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.
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.
a
Measured with impregnated strand.
^Measured with unimpregnated strand.
28
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
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.
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.
/computer console
to tensioner cylinder
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
Mat
Forming Guide
Roving Heated Die
To Puller
Curing
Agent Resin
Finished Part
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
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
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
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
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.
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 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.
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.
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.
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
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.
* > ! (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
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 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
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.
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
CHEMICAL PROPERTIES
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.
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
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
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.
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
Isotropic Material
Youngs Modulus E
Shear Modulus G
Poisson Ratio v
Bulk Modulus K
Plane Strain Bulk Modulus KP
74
G = ^ (17)
2(1+v)
3K-2G (18)
V =
2(3K + G)
K = ^ (19)
3(1-2v)
Kp = -2 (20)
p
l-2v
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
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.
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 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 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.
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.
FAILURE PROPERTIES
Isotropie Material
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.
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.
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.
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)
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.
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.
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.
[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
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.
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
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.
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.
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
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
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
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
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.
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
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
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.
22.45 m
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 %
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
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].
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
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
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.
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
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
40-
50 100 YEARS
TIME
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
'
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
~"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
"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].
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
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
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
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.
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.
4. INSTITUTIONAL EFFORTS
Table 3 Trends of committees on the application of the continuous fiber bars for
structures in Japan
Promoters Committees
Table 3 Trends of committees on the application of the continuous fiber bars for
structures in Japan (continued)
Promoters Committees
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)
5.1. 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.
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)
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.
Investigation of Function of
FRP Cable to be Required
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
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
Appendix
Examination of Testing Method of
1. Standard Test Method- Tensile Strength
2. Example of Structural Design- Structural Behavior of Frame
Heat Cycle at Dry and Wet State by Bond
Test
3. Evaluation Standard of Durability- Chemical Resistance
Behavior of RC and PC Beams Subjected
to Long-Term Loading
Tensile Strength at High Temperature
4. Evaluation Standard of Fire Resistance- Flexural Behavior of Beams under Heating
Flexural Behavior of Beams after Heating
Table 6 Schedule of research items (building engineering)
Chemical Resistance
Tensile Strength at
Low Temperature
Evaluation of 2 and
3 Dimensional FRPR
Structural Behavior
of Frame
Table 6 Schedule of research items (building engineering) (continued)
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
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.
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.
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.
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
SOURCES:
1. Watson, 1989 2. Pigliacampi, 1989 3. Tokyo Rope, 1990
133
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
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.
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.
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
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
3.4 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
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)
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
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.
5.1 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
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
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]
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.
kd
nAr
SECTION STRESSES
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.
->fr
Ar I I reinforcement strain
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
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.
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.
6.3.2 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
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.
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.
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.
8. RESEARCH NEEDS
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.
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
REFERENCES
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
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.
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.
Ehsani, M. and Saadatmanesh, H. (1990). "Fiber Composite Plates for Strengthening Bridge
Beams". COMPOSITE STRUCTURES, Vol. 15, No. 4 p. 343.
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.
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.
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.
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.
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 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., 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.
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.
Matzeg, R. and Santrach, D. (1991). "Effects of the Glass Fiber Type on Corrosion of FRP
Laminates". SAMPE QUARTERLY, October p. 12.
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.
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 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., 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.
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.
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.
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.
Tokyo Rope Manufacturing Co., Ltd. (1990). "Carbon Fiber Composite Cable". Corporate
Report
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.
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.
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.
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
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.
3. MECHANICAL PROPERTIES
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
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
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.
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
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.
^-<ar 2 +b)E u
= e u 0<r<R (la)
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
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.
2R2 - (R-c) 2
a- " cR2 (2R-c) (5a)
c 2 (2R-c) 2 -R 4
b - 2cR2 (2R-c) (5b)
c c
= <5c>
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.
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
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.
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:
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.
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.
w
max = 0 7 6 E s 8 s vdcA x 10"3 ( 12 )
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
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)
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
- = 0 , 4 ^ ^ x 1 0 - 3 (.9,
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.
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.
d
Id = ^ b (23)
= / (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)
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.
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.
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
8. REFERENCES
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
(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
CH - C H - Type 7901
I
OH High crystallinity
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
(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
n
2 4 6 8 16 18
Elongation (%)
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.
Condition Fiber*
Chemicals
Cone Temp .Time Vinylon Rayon Nylon Polyester
(%> ( e c) (hrs) -6
Sodium hydrosulfite 1 70 10 96 87 96 97
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
80
Strength
retention ^0 A
of fiber
Polyethylenp
(%) fiber
40
20
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)
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
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
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
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
(xl60)
x^.
0 7 14 28
pH=12.7
CLATEC ROD has high adhesive strength with concrete (bonding force
required to pull out from concrete) approximate as same as deformed
steel bar.
Adhesive Maximum
strength adhesive
Strength
CLATEC ROD 6D 14.8 rod broke
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.
Steel 3
-#ti =4
f*l
-i 1 1 r
I Stirrup
lf 50
-1000- 4IOO V
CLATEC ROD 6 D
50]Bending test
Deflection (mm)
Figure11 Bending LoadDeflection curve
208
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
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)
(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.
5. APPLICATIONS
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- 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
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)
(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.
(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.
Shotcrete
Anchor bolt or
CLATEC ROD Castin place
> \ concrete layer
u-^->.. W k s , ,
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
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.
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.
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.
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.
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.
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.
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
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 |
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.
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.
Room temperature
Frequency: 9.1 kHz
Coil Heating tube
(Length=l. Om Inner dia. =16mm0)
mmimmi \
Test piece
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.
(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.
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.
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 {%)
(1) Creep
Result of creep tests with CFCC is shown in Fig. 3-3. Creep tests were conducted
under the following conditions:
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)
(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
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)
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)
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.
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
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
OB
|j 20
0
Q_
3E
LU
- -20
5 10 15 20 25
TIME (Hrs)
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.
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.
4. PERFORMANCE AS REINFORCEMENT
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
4.2 Flexibility
Compared with other materials, shearing characteristics of CFCC are shown in the following.
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
(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
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
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.
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.
CFCC
Resin
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
(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
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
Fixing box
Fixed saddle for bending
2.2 10
<)
? 2 ()
< '
==1.8 1i -10
)
li.e ii {i -20
I
CO [ \\ !
1
S1.4 1
-30 g
CO
. 1
. -50
10 5 10 15 20 25 30 35
BENDING ANGLE (DEGREE)
5. APPLICATIONS
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:
8 200
The Hakui Bicycle Road #15 Bridge is a bicycle bridge. Construction was completed in
Ishikawa Prefecture in March of 1991.
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
The Hakui Bicycle Road #16 Bridge is a bicycle bridge. Construction was completed in
Ishikawa Prefecture in March of 1992.
700
640
H
CFCC 6=1. 5mm
lea @ @ ro ta fa a &
4 300
3500 4Qd
CO
J 2 3 L K 22L_8f
U j U
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.
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
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
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
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
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
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
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.
fe
RF-Girder irf*
rr
RAH
i^
,
1 1L
77777777777777
10,625
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>
ends center
p?i **!
Table 8
T h e outline of CFRP tendons
Type of CFRP tendons three-strand multicable
( 3-CFRP 12.5 0 )
Length of CFRP tendons 20, 950 mm
per one cable
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
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
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
RP-Girder
G
-69
-54
(unit : x 10 ")
L/-36 I 2F-Girder
(-):compression strain
2,5001 5,000
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. 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)
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.
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
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
o
o
VO H
%
00
V>
-I
3. 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.
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
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
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
Rod strength
(GPa) 2.40
2.20 L
2.00 |_
1.80
1 2 3
Tensioning Speed (%/min)
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
2.20 \-
2.00 U
1.80
1 2 3
Rod strength
(GPa) 2.40
2.20 r-
-20 0 20
Sample temperature (C)
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
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
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
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.
2000
Stirrup 6
100
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
0.3
i
0.2
i
0.1
1
0 4.0
4 X10
6
1.0 2.0 3.0
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
&
5250 (Beam Length = 10500) 250
50 ,250 2 X 220 15X300 = 4500 50 5|
440 D13 150
D13>1
| DIP AFRP
4 D10
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).
500
400
Bending
Moment
300
(kN.m)
200
CalculativeValue
I
Measured Value
100
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.)
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)
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
J
AFRP girder section
500 14 000 14 000 9 000 14 000 9 000 500
Berth length L = 6I,OOO
3 ^
AFRP tendon en
Yi
bo n n o o o no'
Cross-section of girder
CZ 1 1 1 1 1
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.
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,
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.
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>
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
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)
150
n ^ H L - DiSDlacement meter
4 \Reinforced bar
I I
(mm) Pull out Displacement (mm)
I Pull
4. DURABILITY
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
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)
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.
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\
Figure 12. FiBRA tendons for building Figure 13. Installation of a beam is
floor. proceeding.
Figure 14. FiBRA-bolt with an anchor. Figure 15. Floor panel for raised
floor.
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.
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
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
- 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
Modulus of elasticity
(N/mm 2 ) 210.000 205.000 51.000 64.000 146.000
structures
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
120
F
E
\ 100
-z.
_c
b"
<( 80
OJ
cn
c
60
\
40
^~~~~
3 70
<
-
J4 10 5 10| 10 7 1C
| |
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)
#:
i - POLYSTAL - bar prestressed to 50%
1
Aqueous cement solution pH - 13,0
Outer steel tube as abutment
Prestressing anchorage
Outer steel tube as support
POLYSTAL- bar
prestressed to 50%
#=8-
L
~ PE perforated tube
{
Cement mortar
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.
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
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
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
|- 22.45
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
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
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.
5. SENSOR TECHNOLOGY
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
6. TRIAL LOADINGS
Control of prestressing
Control device
H *&
1HLV prestressing tendon withl
integrated optical fiber sensorj
)ata
processing
temperature
tension
attenuation
cross section
Fig. 28: Measurement layout
SLAB 20 SLAB 21 SLAB 22
LAYER 4 LAYER 4 LAYER 4
*
i L/v
f
I
-*
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
sensor section 2
/ V
sensor pier 1 /
10
o.o \/y \v.
-10 measuring -
-20 time / sec.
-30
7. CONCLUSION
I N S T I T U T FR BAUTECHNIK
Anstalt des ffentlichen Rechts
ZULASSUNGSBESCHEID
Der
Zulassungsnummer:
Der zugelassene Gegenstand darf nur verwendet werden, wenn seine Herstellung
berwacht ist und dies am Verwendungsort geprft werden kann.
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.
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.
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.
~) 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
2000 ./+
1800
_1600
/E=121kN/mm2
1400
\
-1200
2 1000
To
800
600 /
E=111kN/mm2
400
200
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.
^ /
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
days years
100 100
10
I 1
10'
i2
10 5
101
IT
:
10 10 3 4
10 ~ 10 e
10 e 2
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).
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.
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
200
160
140
120
Load cycle 1
fw
100 Load cycle 2
I
..I
o
0 /
60
40
20 - Prestressing _
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)
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
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.
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.
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.
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
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.
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
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
* 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.
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
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
Glass-cloth
with resir /
/
u u-
*---- -;
J,-- --
NEFMAC
NEFMAC
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
\ Material
Dia-^s. NEFMAC Welded wire Steel
mesh reinforcement
meter ^v
*-o
r
2.0
<&-- <t
H i.o
Measuring position(cm)
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
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)
PC tendon
Spring
( ) ( )
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
Stress of reinforcing
200 r RC
bar=150MPa
150 \
100 \
Type G
50
200
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
G8 30.38 1,519
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
30
H r*\ I
30
Fig. 13 Details of Test Curtain Wall
90
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
s NEFMAC/
/
L / C
/ ^ 1
>
L // ^ ^^ [| 1
140
1/ ^^ 1.600
4,280
400
Deflection(cm)
H3 12.05 764
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-
H6@100
(in nail)
(for anchorage)
(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
Shear Force(tf)
80
60
40
\
20
^/ Deforme tion(mm)
f/ 10 2o :0 40 50
^-20
--40
Note : 1tf = 9.8kN
--60
--80
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
Photo 11 Pontoon
Photo 14 OA Floor
385
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
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.
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
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
,-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
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.
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
Table 5
Mix proportion of mortar matrix
r~^ o
QJ] >*
~7X 5~ I 70
50 | 200 I 200 | 200 150
700
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
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)
3DF (PAN-CF)
@30 (X, Y) 2 0 (Z)
o
o In
^CO In
IS
for handling
[-60X50X2.3
150
-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
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
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)
-5 4.0
_
CD
2 3.0 .^^..
2.0
1.0
100 200 300 400
Temperature (C)
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
Abstract
1. INTRODUCTION
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.
\
Resin impregnation bath / Automatic conveyor
No. 1 curing oven
Pultrusion die No.2 curing oven
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.
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)
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.
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.
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:
lugs and the tensile stress in the lugs. This A A A---- ' ' '--,
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.
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.
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
"
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)
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.
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
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
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.
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.
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
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
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
Cross section
TJ V75
^ 42la=6
d=6
- 42 d=10
3
- 4 d=26
- CFRP sheet 200 x 1
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]
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.
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
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
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 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
5. OUTLOOK
6. REFERENCES
1. INTRODUCTION
2. RETROFIT OF CHIMNEYS
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.
1,800
4-"
3
2 @7
unit; mm
m
Reinforcement
Table 1.
Material properties
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
a = {P(e)-Pc(e)} / {P(a)-Pc(a)}
Retrofitted 2
200
150
3 / Retrofitted 1
/ -^ /
O
100
^ / ^- ' V
/ .* X ((.
50 Un-retrofitted
Displacement (mm)
-H 1 (-
10 15 20
P = Pc + a P f
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.
x
Load
L.UU y\
Re-bar
40
Development
length UD tape
Notch
!-3 15
m Concrete block
/
UD tape
| 100 1
Unit: mm
Table 3.
Test results
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 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
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)
Figure 12. The outline of specimen Figure 13. Maximum load time of
accelerated exposure relationship
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.
Figure 14. The outline of new retrofitting method for bridge column
445
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
1 Included Nothing
2 Not included S 600
3 Included F+S 300 250 650
Unit: mm
Specimen No. 3 Specimen No. 2
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
Failure pattern
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
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
Retrofitting effect
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
(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)
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