Vous êtes sur la page 1sur 11

Flexural strengthening of concrete beams

using externally bonded composite


materials
Michael J. Chajes*, Theodore A. Thomson Jr, Ted F. Januszka and William
W. Finch Jr
Department of Civil Engineering, University of Delaware, 137 Dupont Hall, Newark, DE
19716, USA
Received 22 March 1994; revised and accepted 3 May 1994
In this study, a series of reinforced concrete beams were tested in four-point bending to determine the
ability of externally bonded composite fabrics to improve the beams' flexural capacity. The fabrics used
were made of aramid, E-glass and graphite fibres, and were bonded to the beams using a two-part
epoxy. The different fabrics were chosen to allow a variety of fabric stiffnesses and strengths to be
studied. The external composite fabric reinforcement led to a 36 to 57% increase in flexural capacity and
a 45 to 53% increase in flexural stiffness. For the beams reinforced with E-glass and graphite fibre
fabrics, failures were a result of fabric tensile failure in the maximum moment region. The beams
reinforced with aramid fabric failed due to the crushing of the compression concrete. In addition to the
test results, an analytical model based on the stress-strain relationships of the concrete, steel and
composite fabrics is presented. Using the model, beam response is computed and compared with the
experimental results. The comparisons indicate that the flexural behaviour of composite-fabric-
reinforced concrete beams can be accurately predicted using the described method.
Keywords: composite materials; flexural strength; reinforced concrete beams
Many modern cities are faced with a rapidly deterio-
rating infrastructure. Because of the prohibitive cost of
replacing large numbers of deteriorated structures,
research efforts have focused on methods of strengthen-
ing existing structures. The rehabilitation of concrete
structures represents one of the more challenging
problems faced by engineers today. The research pre-
sented here focuses on flexural strengthening of rein-
forced concrete beams through the use of externally
applied composite materials.
Research and field application of steel plates bonded
to concrete has been the subject of ongoing research for
many years. Steel plates bonded to the tension face of
concrete beams have been shown to be effective in (i)
increasing flexural capacity and (ii) enhancing flexural
stiffness thereby reducing deflections and controlling
crackingt". Included in these works are studies regarding
the anchorage of steel plates bonded to concrete beams.
Stress concentrations at the ends of the plates can lead to
premature debonding of the plates, and several different
end-anchorage schemes were employed to help alleviate
this problem. This work provides useful insight into the
methods of bonding plates to concrete, the types of adhe-
sives that can be used, and the various modes of failure.
While steel plates have the advantage of possessing a
high strength-to-weight ratio (i.e. fairly thin plates can be
*Correspondence to Dr Michael J. Chajes
used), they have the disadvantages of being susceptible to
corrosion and debonding.
As an alternative to steel, the use of advanced com-
posite materials for structural rehabilitation shows great
promise. Along with having a high strength-to-weight
ratio, composite materials have the beneficial character-
istics of being non-corrosive and generally resistant to
chemicals. Several survey papers have discussed promis-
ing applications of composite materials for a variety of
civil structures'"!", In terms of strengthening concrete
structures, research efforts have included (i) the bonding
of fibre-reinforced-plastic (FRP) composite plates to re-
inforced concrete and prestressed concrete beams to
improve flexural stiffness and strength's-"; (ii) the wrap-
ping of concrete columns with fibreglass/epoxy jackets to
provide the additional flexural and shear strength needed
in seismic regions>; (iii) the confinement of concrete
columns using composite materials, thereby increasing
the columns' axial capacityv"; and (iv) the wrapping of
concrete beams with composite fabric/epoxy jackets to
provide additional shear strength>?'.
Work involving the bonding of composite plates to the
tension face of concrete beams can be divided into two
categories: (i) the bonding of non-prestressed composite
plates, and (ii) the bonding of prestressed composite
plates. In the studies involving non-prestressed com-
posite plates bonded to reinforced concrete beams, vari-
ous types of glass, carbon and aramid fibre composite
0950-0618/94/03/0191-11
1994 Butterworth-Heinemann Ltd Construction and Building Materials 1994 Volume 8 Number 3 191
Flexural strengthening of concrete beams: M. J. Chajes et al.
Figure 1 Test beam geometry, strain gauge locations and applied
loading
loads do not have a detrimental effect on the fabric-
reinforced beams. Both of these tests should help to
identify the most appropriate fabrics for this type of
application.
I 127mm I
Q
2 # 3 Bars
Test Set-Up
I 127mm I

BeamCross-Sections
1#3Bar
I 127 mm I
D
I 406.4mm
31.75 mm
Experimental study
Test programme
This programme consisted of testing 14under-reinforced
rectangular concrete beams. The tests were designed to
evaluate the effect of externally bonded, composite fabric
reinforcement on the beams' flexural behaviour. The
beams were loaded monotonically to failure in four-
point bending using an 890 kN capacity Tinius Olsen
universal testing machine (see Figure 1).
Beam design
Fourteen rectangular beams with span lengths of 1.12 m
and cross-sectional dimensions of 76.2 mm x 127 mm
were used (see Figure 1). The set of beams consisted of
three control beams having only traditional steel reinfor-
cement (beams CI, C2, C3); three sets of three beams,
each having the same steel reinforcement as the control
beams but with additional, externally applied aramid, E-
glass and graphite fibre fabric reinforcement (beams AI,
A2, A3, EI, E2, E3, GI, G2, G3); and two beams having
twice the steel reinforcement of the control beams but no
external composite fabric reinforcement (beams Sl, S2).
The beams with additional steel reinforcement were used
to allow a comparison of beams having additional exter-
nal fabric reinforcement with beams having additional
internal steel reinforcement. Cross-sections of the vari-
ous types of beams are shown in Figure 1.
All beams were designed to fail in flexure. The control
beams and the beams with additional steel reinforcement
were designed according to the specifications of the ACI
Building Code". In satisfying the ductility requirement
a single Grade 60#3 bar (9.525
mm diameter with a yield strength of 413 MPa) was used.
Two # 3 bars were used in beams having extra steel
plates have been tested
1s
-
25
These studies have shown
that the externally reinforced concrete beams exhibited
substantial increases in flexural capacity and stiffness. In
the earliest work, many of the beams failed due to the
debonding of the composite plate. In later studies, adhe-
sives with sufficient strength to transfer the shear forces
between the composite and concrete were found. When
these adhesives were used, the failure mode with regard
to bond was a shear failure of the concrete. Finally,
analytical models were developed for predicting the
flexural behaviour of the beams.
The technique of prestressing composite plates and
then bonding them to reinforced concrete beams has also
been studied>:". These studies found that prestressed,
bonded reinforcement produced a moderate gain over
non-prestressed reinforcement. The most significant
improvement found was the control of cracking. In one
study>, a theory for the optimal design of externally
bonded composite plates was presented.
While FRP composite plates can be used as an effective
means of providing additional reinforcement, they do
possess some drawbacks including (i) the need for a flat
surface for bonding, (ii) the cost associated with manu-
facturing large plates, and (iii) the difficulty in achieving
a bond between the concrete and the composite plate
sufficient enough to prevent debonding from governing
the failure mode.
As an alternative to the use of composite plates, this
paper investigates the strengthening of reinforced con-
crete beams using epoxy-bonded composite fabrics made
of aramid (Kevlar*), E-glass, and graphite fibres. Like
the composite plates, these fabrics are non-corrosive and
possess high strength-to-weight ratios. They also have
the beneficial qualities of (i) being able to conform to
irregular surface geometries, (ii) being manufactured in
long lengths, and (iii) being able to be bonded to beams
in such a way as to develop full tensile capacity prior to
debonding. Jackets made from glass-fibre fabrics expox-
ied to concrete columns have been studied as a possible
seismic retrofitting procedure and have been found to
perform quite well>.
This research can be directly applied to the upgrading
and rehabilitation of concrete beams. In situations where
bridges will be required to handle increased traffic, com-
posite fabrics can be adhered to the beams, thereby
increasing allowable loads. Likewise, damaged beams
can be reinforced and restored to their original capacity.
Because of their ability to conform to the shape of the
beam, composite fabrics can be used in a wide variety of
situations, and by wrapping the exposed cross-section,
strong bonds can be achieved and a simultaneous
increase in shear capacity can be gainedv-". It should be
noted that additional environmental and fatigue testing
must be run before actual field application is attempted.
The environmental testing should investigate the ability
of the various fabrics to withstand aggressive environ-
ments, while the fatigue testing should verify that cyclic
*Kev!ar is the registered tradename of a family of aramid fibres pro-
duced by the E.I. duPont de Nemours and Co. Inc., Wilmington, DE,
USA
192 Construction and Building Materials 1994 Volume 8 Number 3
Flexural strengthening of concrete beams: M. J. Chajes et al.
0.025
-Aramid (1 layer)
......... E-glass (3 layers)
- - - - -Graphite (2 layers)
0.010 0.Q15 0.020
Strain (mm1mm)
....
.'
,
,
,
..'
0.005
,
,
,
,
,
,
,
,
, .'
, .'
, .:
r .'
, .'
, .'
0...jL..-'-............+ - - ' - ~ ' - ' - f - - ' - - - " - - - ' - ~ + - ' - ~ - ' - 1 - - ' - - - ' - ' - - - ' - - +
0.000
Ultimate strength,};' (MPa)
Table I Beams tested
250
Beams Internal steel External fabric
reinforcement reinforcement
200
CI, C2, C3 I #3 bar
I
SI, S2 2#3 bars
150
AI, A2, A3 I #3 bar Aramid (one layer)
El, E2, E3 I #3 bar E-glass (three layers)
'"
Gl, G2, G3 I #3 bar Graphite (two layers) '"
4)
100
J:l
tf.l
50
Concrete batch
Table 2 Concrete strengths
1
2
3
4
5
34.8
33.6
36.0
42.5
42.1
Figure 2 Stress-strain behaviour of impregnated composite fabrics
reinforcement (beams Sl and S2). The beams with exter-
nal composite fabric reinforcement were designed so that
the tensile capacity of the fabric was close to the yield
load of the additional # 3 bar (29.4 kN). To accomplish
this, different numbers of fabric layers were used accord-
ing to the ultimate strength of the various types of fabric.
According to the tensile capacity of the aramid, E-glass
and graphite fabrics (see section on composite fabrics),
one layer of aramid fabric (29.5 kN), three layers of E-
glass (24.9 kN) and two layers of graphite fabric (29.4
kN) were used. Shear reinforcement was not needed in
any of the beams. Table 1 provides a summary of the
various beams tested.
Beam fabrication
A concrete mix having a water-to-cement ratio of 0.50 by
weight was used to cast the 14 beams. Type I Portland
cement was used, and the maximum aggregate size was
9.5 mm. Five batches of concrete were needed to fabri-
cate the 14 beams and associated standard test cylinders
(152.4 mm diameter by 304.8 mm high). The beams were
cast in the University of Delaware Structures Laboratory
and were allowed to cure in a water bath for 28 days.
The concrete strengths for the five batches are shown
in Table 2. The average strength for the five batches was
37.8 MPa, and none differed by more than 12.5% from
this average. The Grade 60 steel reinforcing bars used
were tested and found to have an actual yield strength of
493.0 MPa.
Composite fabrics
The three types of composite fabrics used in this study
were (i) plain-weave aramid (Kevlar") fabric, (ii) crow-
foot satin-weave E-glass fabric, and (iii) plain-weave
graphite fabric. All of these woven composite fabrics are
made up of fibres oriented at 0 and 90 with an equal
distribution of fibres in each direction. The use of the
three different fabrics allowed the effect of varying the
stiffness of the external reinforcement to be studied. It
should be noted, however, that the use of some of these
fabrics may prove to be undesirable if they are found to
lack acceptable environmental durability.
For each type of fabric, three tensile test specimens
impregnated with the same adhesive used to bond the
fabrics to the concrete (see next section) were prepared.
The number of layers of fabric used for the specimens
was the same as used for the beams (one layer of aramid
fabric, three layers of E-glass fabric and two layers of
graphite fabric). All test specimens were cured under a
vacuum. After curing, the tensile specimens were loaded
to failure using an Instron universal testing machine.
From the test results, their respective elastic moduli,
failure strains and ultimate strengths were computed.
Stress-strain plots for the three impregnated composite
fabrics are shown in Figure 2, and a summary of the
average test values is presented in Table 3.
Adhesive selection
Based on ongoing research by the authors dealing with
the bonding of composite plates to concrete surfaces,
Sikadur 32 was selected as the adhesive for this appli-
cation. Sikadur 32 is a two-component, high-modulus,
high-strength, construction epoxy. The two parts are
mixed in a ratio of 1:1and can be cured at room tempera-
ture.
Since the previous bond tests involved composite
plates bonded to concrete, additional bond tests using
Sikadur 32 and the three types of composite fabrics were
conducted. A series of pull-off tests were run using the
set-up shown in Figure 3. The tests involved the bonding
of a 1 inch wide piece of epoxy-impregnated composite
fabric to a concrete block. Nine specimens having bond
lengths of 25.4, 50.8 and 76.2 mm were tested (three at
each length). The tests indicate that a single layer of
aramid, a triple layer of E-glass, and a double layer of
graphite fabric can be expected to develop full tensile
capacity in approximately 50.8 mm for both the E-glass
and graphite fabric, and in approximately 76.2 mm for
the aramid fabric. Based on these results, Sikadur 32 was
deemed to be a suitable adhesive choice. A summary of
the bond test results is given in Table 4.
Construction and Building Materials 1994 Volume 8 Number 3 193
Flexural strengthening of concrete beams: M. J. Chajes et al.
Table 3 Properties of resin-impregnated composite fabrics
Composite fabric Specimen thickness Modulus of elasticity Failure strain, 4u Failure stress, a,u
(mm) E,(MPa) (mm mm-
I
) (MPa)
Aramid (one layer) 1.04 11020 0.0225 223
E-glass (three layers) 1.42 13090 0.0122 138
Graphite (two layers) 1.22 22050 0.00748 190
p p
t t
.-Composite Fabric---. Composite Fabric (no end tabs)
(b)
Front View SideView
Figure 4 Externally bonded fabric: (a) beams AI, EI, E2, E3, GI, G2,
G3; (b) beams A2, A3
Figure 3 Experimental set-up for bond tests
Table 4 Bond test results
Bond length (mm) Aramid fabric E-glass fabric Graphite fabric
25.4 B B B
B B F
B F F
50.8 B F F
B F F
B F F
76.2 F F F
F F F
F F F
B - Bond failure
F - Fabric tensile failure
Bonding of composite fabric
Prior to the bonding of the fabric to the beams, the
concrete surface was mechanically abraded using a
grinding wheel, creating a somewhat porous surface. The
fabric was coated with adhesive on both sides and placed
onto the tension face of the beam, which itself had been
coated with adhesive. The fabric was then smoothed to
ensure a uniform distribution of adhesive and placed in a
vacuum bag. The beam was allowed to cure under
vacuum for I day and for an additional 2 days once it
was removed from the vacuum.
The two bonding schemes used are shown in Figure 4.
Originally, all beams were bonded without end tabs as
shown in Figure 4(a). During the initial round of testing,
however, the first aramid-reinforced beam (beam AI)
experienced some fabric debonding. As a result, end tabs
were added (see Figure 4b) as suggested in a study by
Ritchie et a/.
21

Instrumentation and test procedure


The test procedure consisted of loading all 14 beams
monotonically to failure according to the configuration
Figure 5 Typical beam during testing
shown in Figure I. Failure was defined as (i) the crushing
of the concrete in the compression region, (ii) tensile
failure of the composite fabric, or (iii) the debonding of
the composite fabric. During the testing, the deflections
of all beams were measured at the end points, at the
location of the applied loads, and at midspan. A photo-
graph of a typical beam taken during testing is shown in
Figure 5.
To provide additional information about the behav-
iour of the externally reinforced beams, electrical resis-
tance strain gauges were used to measure strains in both
the concrete and composite on one of each type of fabric-
reinforced beam (beams A2, E2 and G2). Four gauges
were mounted through the depth at midspan with an
additional two gauges mounted on the fabric on either
side of the midspan gauge. The location of all strain
gauges is shown in Figure I.
194 Construction and Building Materials 1994 Volume 8 Number 3
Flexural strengthening of concrete beams: M. J. Chajes et al.
Figure 6 (a) Stress-strain curve for concrete, (b) internal strain, stress
and force distributions for beam
Theoretical analysis
General methodology
Beams properly designed according to the ACI Code>'
will fail in flexure due to the crushing of the compression
concrete. The ultimate strength analysis of such under-
reinforced beams is relatively straightforward. This is
partially due to the fact that the design procedure forces
the longitudinal reinforcing steel to be at yield when the
beam fails. Because composite materials loaded in ten-
sion behave linear elastically to failure (see Figure 2), the
analysis of concrete beams reinforced with composite
materials must be handled differently than the ACI
approach. An iterative analytical method for predicting
the flexural behaviour of reinforced concrete beams with
externally bonded composite plates has been presented
by An et al.. This method, which involves assuming a
strain distribution for the cross-section and then utilizing
the material constitutive relationships to check equili-
brium, will form the basis for the analysis used herein.
Cross-Section Strain
(a)
(b)
Stress Force
Assumptions
Before proceeding to the details of the analytical pro-
cedure, the following assumptions are noted: (i) plane
sections remain plane; (ii) no slip betwen the longitudinal
reinforcing steel and the concrete; (iii) no slip between
the impregnated composite fabric and the concrete; (iv)
no shear failure of the concrete adjacent to the bonded
fabric reinforcement; (v) concrete carries no tension; and
(vi) maximum strain in the concrete in compression is
0.003.
Material stress-strain relationships
Use of the analytical model assumes that the dimensions
of the beam are known and that complete stress-strain
relationships are known for the steel rebar and impreg-
nated composite fabric in tension, and the concrete in
compression. In our case, the steel rebar is assumed to
behave elastic-perfectly plastic, and the steel tensile
stressfs is given by
Values for the modulus" ultimate strain and ultimate
strength for the three composite fabrics are given in
Table 3.
For the concrete, Hognestad's idealized stress-strain
curve for concrete in compression, described by Park and
Paulay>, is used. This relationship, shown in Figure
6(a), is given by
(4)
and
where
So = 2fc"/s; (6)
fs = E,6, for 0 ~ 6, < 6y (1)
fc"';:::,0.92fc' for 32.8 MPa ~ fc' ~ 4 4 . 8 MPa (7)
where E, is the modulus of the epoxy-impregnated fabric,
Gr is the fabric strain, and Gru is the ultimate fabric strain.
where E, is the elastic modulus of the steel, 6, is the strain
in the steel, and hand 6y are the yield stress and asso-
ciated yield strain. From tension tests, it was found that
the # 3 bars used have a yield strength of 493 MPa and a
modulus of 200000 MPa. It is assumed that all beams
fail before the steel reaches its ultimate strain.
The impregnated composite fabrics are modelled as
behaving linear elastically to failure, and the fabric
tensile stressfr is given by
and
I, = h for 6y~ 6,
fr = E
rGr
for 0 ~ 6r < Gru
(2)
(3)
E; = 4730.jfc'MPa for normal weight concrete (8)
in which fc is the stress in the concrete, fc' is the cylinder
strength, fc" is the maximum compressive stress reached
in the concrete, e; is the strain in the concrete, So is the
strain in the concrete at maximum stress, Seu is the maxi-
mum strain in the concrete, and E, is the initial tangent
modulus of the concrete in compression.
Analytical method
The method presented is iterative in nature and is best
suited for computer solution. It begins by selecting a
value of concrete strain at the outer compression fibre
and guessing the associated location for the neutral axis.
Based on the values chosen, the assumptions discussed
above, and the stress-strain relationships for the con-
crete, steel and composite fabric, a state of strain, stress
Construction and Building Materials 1994 Volume 8 Number 3 195
Flexural strengthening of concrete beams: M. J. Chajes et al.
Table 5 Test results for each beam
Beam Concrete strength j, Failure mode Experimental beam strength
(MPa) (N)
CI 34.8 Concrete crushing 10 182
C2 34.8 Concrete crushing II 236
C3 36.0 Concrete crushing II 125
SI 33.6 Concrete crushing 18330
S2 33.6 Concrete crushing 16986
AI " 42.1 Fabric debonding 16269
A2 42.5 Concrete crushing 14756
A3 34.8 Concrete crushing 16888
EI 33.6 Fabric tensile failure 15 280
E2 42.1 Fabri c tensile failure 15282
E3 42.5 Fabric tensile failure 153 79
GI 42.5 Fab ric tensile failure 15059
G2 42.1 Fabri c tensile failur e 17035
G3 36.0 Fabric tensile failure 14489
Cont rol beam strength
(N)
10849
b
to 849
b
to 849'
10849'
10849'
10849
b
10 849
b
10849'
10849'
10849
h
to 849
h
Percentage increase
+ 69.0
+56.6
+ 50.0
+ 36.0
+55.7
+ 40.8
+ 40.9
+ 41.8
+ 38.8
+57.0
+ 33.6
"This was the only aramid-reinforced beam without end tabs
'Average strength of control beams CI, C2 and C3
Figure 8 Load versus midspan deflection behaviour of fabric-
reinforced beams after fabric failure
- ControlBeam(C2)
......... E-glassBeam(E3)
--ControlBeam (C2)
-Steel Beam(SI)
- - - - -Aramid Beam (A3)
.........E-glassBeam(E3)
- -Graphite Beam (G3)
10.0 20.0 30.0 40.0 50.0 60.0
MidspanDisplacement(nun)
10.0 20.0 30.0 40.0 50.0 60.0
MidspanDisplacement(nun)
Load versus midspan deflection plot for typical beams Figure 7
20000
15000
~
-g 10000
.3
5000
16000
14000
.'
.'
12000
~
10000 :
-g 8000
0
...J
6000
4000
2000
and 52) also failed in flexure owing to concrete crushing.
Like the control beams, these beams also developed
initial flexural tension cracks at loads near 2000 N. How"
ever, because of the additional tensile reinforcement, the
steel in these beams yielded at loads ranging from 15500
to 16500 N and finally failed at loads between 17000 and
and internal resultant forces can be established (see
Figure 6b). By summing the resultant forces, the equili-
brium of the section can be checked. If the resultant
forces are not in equilibrium, a different location for the
neutral axis is guessed while the selected value for the
strain at the outer fibre of the concrete is maintained.
Once a neutral axis location that satisfies equilibrium is
found, the internal moment, curvature and strain any-
where in the cross-section can be computed. By following
this procedure, and by allowing the value of the strain at
the outer fibre of the concrete to begin at a small value
and incrementally increa se, the flexural behaviour of the
section can be establi shed. Failure of the section is said to
occur either when the strain in the concrete reaches its
ultimate compression strain E:cu (taken to be 0.003 in this
study), or when the composite fabric reaches its ultimate
tensile strain Cfu'
All of the beams tested were analysed using a
computer program that followed the procedure des-
cribed above . The analyses provide predictions regarding
the ultimate capacity of the beams, their mode of failure
and the load-deformation behaviour.
Experimental and analytical results
The results of the 14 beam tests are summarized in Table
5 and Figures 7-9. These results, along with the corres-
ponding analytical predictions, are discussed in the
sections that follow.
Beam behaviour and mode of failure
Upon being loaded , each of the three control beams (CI ,
C2 and C3) began developing flexural tensile cracks in
the const ant moment region at loads around 2000 N. At
loads around 9000 N, the reinforcing steel in these beams
yielded. Finally, each of the three control beams failed in
flexure at loads of around II 000 N as a result of the
compression concrete crushing. The load-deflection
behaviour of a typical control beam is shown in Figures 7
and 8, while a picture showing beam C3 after failure can
be seen in Figure 9.
The two beams with additional steel reinforcement (51
196 Construction and Buil ding Mater ials 1994 Volume 8 Number 3
Flexural strengthening of concrete beams: M. J. Chajes et al.
18500 N. Figure 7includes a typical load-deflection plot
for these beams.
The general flexural behaviour to failure of all the
composite-fabric-rein forced beams was similar, although
the flexural stiffness and final mode of failure var ied
depending upon the fabric used. Like the standard steel-
reinforced beams, the fabri c-reinforced beams also de-
veloped flexural cracks at loads around 2000 N. After
cracking, the beams continued to deform at stiffnesses
between that of the control beams and the beams having
additional steel reinforcement . At loads around 12 000
N, the steel in these beams began to yield as evidenced by
the noticeable change in flexural stiffness seen in Figur es
7 and 8. Because the impregnated fabric remains elastic,
Figure 9 Photographs of typical beams after failure: (a) control beam
with I # 3 bar (C3); (b) aramid fabric reinforced without end tabs (AI);
(c) aramid fabric reinforced with end tabs (A2); (d) E-glass fabric
reinforced (E3); (e) graphite fabric reinforced (G3)
these beams exhibit an appreciable amount of flexural
stiffness until failure occurs at loads ranging from
approximately 15000 to 17000 N. Figur e 7shows typical
load-deflection behaviour for the beams reinforced with
each of the three types of fabric.
For all of the beams reinforced with E-glass and gra-
phite fabrics, beam failure resulted following the tensile
failure of the fabric. For the beams instrumented with
strain gauges, the strains recorded just prior to failure are
similar to the values obtained from tensile tests. Follow-
ing the tensile failure of the fabric, the beams' resistance
dropped to that of the control beams, and the beams
continued to resist load until the compression concrete
crushed. This behaviour is shown for one of the E-glass
Construction and Building Materials 1994 Volume 8 Number 3 197
Flexural strengthening of concrete beams: M. J. Chajes et al.
Table 6 Average ultimate strength increase for beams Table 7 Average stiffness increase for beams in service load region
Beams Experimental Control beam Percentage Beams Experimental Control beam Percentage
beam strength strength increase beam stiffness stiffness increase
(N) (N) (N mm-
I
) (N mm")
CI, C2, C3 10 848
a
CI, C2, C3 1203
a
81,82 17658
b
10 848
a
+62.8 81,82 1805
b
1203
a
+50.0
A2,A3 15822
b
10 848
a
+45.9 AI, A2,A3 I 843
b
1203
a
+53.2
EI, E2, E3 153J4b 10 848
a
+41.2 EI, E2, E3 1752
b
1203
a
+45.6
GI, G2, G3 15528
b
10 848
a
+43.1 GI,G2, G3 1744
b
1203
a
+45.0
-Average value for control beams
hAveragevalues
-Average value for control beams
bAverage values
Table 8 Measured versus theoretical ultimate strength of beams
beams, as compared with the control beams, are evident
in the load-deflection plots of Figures 7 and 8. To quan-
tify and compare the flexural stiffness of the beams in the
service load region, stiffnesses of the various beams were
computed by finding the slope of the line connecting a
point at the origin of the load-deflection plot and a point
on the displacement curve corresponding to a deflection
of L/360 (where L is the beam's span length). The aver-
age of these flexural stiffnesses for each of the different
types of beams is given in Table 7. From the results in
Table 7, we can see that both the fabric-reinforced beams
and the beams with additional steel reinforcement exhi-
bited similar increases in stiffness in the working load
region ranging from 45.0 to 53.2%.
Comparison of analytical and experimental results
Using the analytical method discussed earlier, the flex-
ural behaviour and ultimate flexural capacity of the three
types offabric-reinforced beams were computed. Table 8
contains a comparison of the average measured beam
strengths and the values obtained using the analytical
model. From these results, we see that the model is quite
accurate in predicting the ultimate capacity of the vari-
ous beams. The computed values differed by 17.0, 2.6
and 4.5% respectively for the aramid-, E-glass- and gra-
phite-reinforced beams. These differences are quite small
when one considers the expected variation associated
with the behaviour of reinforced concrete structures. In
fact, computed values of ultimate capacity for the beams
having only steel reinforcement differed by approxima-
tely 20% from the experimental results.
Another useful comparison involves looking at the
strain behaviour of the composite fabric. In Figure 10,
plots of the applied load versus fabric strain for the three
beams with strain gauges (A2, E2 and G2) are compared
with values computed analytically. Like the computation
"Based on an average concrete strength
hAverage values
+17.0
-2.6
-4.5
Percentage
difference
I3 523
a
15728"
16266"
Theoretical
beam strength
(N)
15822
b
153J4b
15528
b
Experimental
beam strength
(N)
Beams
A2,A3
EI, E2, E3
GI,G2,G3
beams in Figure 8. Figure 9 shows typical E-glass- and
graphite-reinforced beams after failure.
The first aramid-reinforced beam tested (A1) failed
when a sudden debonding of the aramid fabric occurred.
The other two beams with aramid fabric (A2 and A3)
utilized end tabs and did not experience debonding.
These beams, unlike the beams with E-glass and graphite
reinforcement, reached the ultimate compression strain
of the concrete before reaching the fabric's tensile capa-
city. As a result, these two beams failed owing to crush-
ing of the compression concrete (see Figure 9). This
difference in failure mode from the other fabric-rein-
forced beams is understandable since the ultimate strain
of the impregnated aramid fabric is twice that of the
impregnated E-glass fabric and three times that of the
impregnated graphite fabric.
It is important to note that while the fabric-reinforced
beams do not display as much ductility as the beams
reinforced with steel alone, they do exhibit some measure
of ductility prior to failure. If ductility is taken to be the
ultimate displacement divided by the displacement
occurring when the steel yields (Llu/Ll
y
) , the steel-rein-
forced beams have ductilities on the order of four to five
while the fabric-reinforced beams have ductilities on the
order of two or three. It should be noted that the ductili-
ties of the E-glass and graphite fabric-reinforced beams
are similar to steel-reinforced beams if the behaviour
after fabric failure is considered (see Figure 8).
Changes in strength and stiffness
All of the externally reinforced beams showed significant
increases in ultimate flexural capacity as compared with
the control beams. As shown in Table 5, the increases
ranged from 33.6 to 57%. When the experimental values
are averaged, we find that the E-glass-, graphite- and
aramid-reinforced beams showed similar increases in
strength over the control beams of 41.2,43.1 and 45.9%
respectively (see Table 6). By comparison, the average
increase in strength resulting from the additional steel
reinforcement was 62.8%. The effect of the externally
applied reinforcement on ultimate flexural capacity
becomes very similar to that of beams having additional
internal steel reinforcement when one recalls that the
actual (tested) strength of the additional steel was 15.7%
higher than the strength of the aramid fabric, 16.0%
higher than the strength of the E-glass fabric, and 28.9%
higher than the strength of the graphite fabric.
Differences in flexural stiffness of the fabric-reinforced
198 Construction and Building Materials 1994 Volume 8 Number 3
Flexural strengthening of concrete beams: M. J. Chajes et al.
Finally, based on the analytical computations, the
percentage of the fabric-reinforced beams' flexural re-
sistance contributed by the fabric reinforcement at both
initial yield of the steel reinforcement and at the ultimate
load is shown in Table 9. From these numbers we see that
the composite reinforcement contributes between 23 and
42% of the moment resistance at initial yielding of the
steel reinforcement, and between 43 and 54% of the
moment resistance at failure. ......... E-glass Beam(Theoretical)
-E-glass Beam(Measured)
4000
2000
6000
sooo
16000
14000
12000
gl0000
]
Conclusion
The results of this study indicate that externally applied
composite fabrics can be effectively used to rehabilitate
or strengthen concrete beams, and analytical methods
needed to describe their behaviour are available. For
each of the three types of fabrics tested, increases in
flexural strength similar to those found in beams having
additional steel reinforcement were achieved. Specifi-
cally, the beams reinforced with epoxy-impregnated ara-
mid, E-glass and graphite fabrics displayed increases in
flexural capacities of 53.2, 45.6 and 45.0%, respectively,
over control beams having only internal steel reinforce-
ment. The fabric-reinforced beams also displayed an
approximately 40% increase in flexural stiffness. Finally,
the bond between the fabric and the concrete, combined
with the additional anchorage provided by easily applied
end tabs, ensured failure of either the concrete in com-
pression or the fabric in tension. Both of these failure
modes displayed a reasonable amount of ductility prior
to failure.
In this paper, an analytical method was also presented.
By comparing the measured and computed behaviour of
the three types of fabric-reinforced beams, the method
was found to predict the beams' ultimate flexural capaci-
ties accurately. The method was also accurate in predict-
ing the mode of failure. For the beams reinforced with
fabrics having smaller ultimate strain capacities (E-glass
and graphite), predicted and observed failure resulted
from the tensile fracture of the fabric. For the beams
reinforced with aramid fabric, which has a higher ulti-
mate strain capacity, the predicted and observed mode of
failure was crushing of the compression concrete.
Before this type of rehabilitation procedure can be
safely applied, further studies involving the durability of
these externally reinforced beams are needed. These tests
should provide information regarding the ability of the
fabrics to withstand aggressive environments and cyclic
loads.
0.010 O.OOS 0.004 0.006
Strain (mm1mm)
......... Aramid Beam(Theoretical)
- Aramid Beam(Measured)
........ Graphite Beam(Theoretical)
- Graphite Beam(Measured)
0.002

0.000
5000
g
'i ioooo
.3
5000
g
'0 10000
.3
(b)
15000

0.000 0.002 0.004 0.006 0.008 0.010 0.012
Strain (mmlmm)
15000
(c)
(a)

0.000 0.002 0.004 0.006 O.OOS 0.010 0.012
Strain (mm1mm)
Figure 10 Load versus fabric strain diagrams: (a) aramid fabric;
(b) E-glass fabric; (c) graphite fabric
of ultimate strengths, the computed load versus fabric
strain behaviour agrees reasonably well with the meas-
ured behaviour. Even for the results of the aramid-rein-
forced beam, in which the computed and measured
values differ most notably, the general stiffness exhibited
before and after the steel yielded are quite similar. In
fact, the assumption that the concrete carries no tension
is responsible for a lot of the discrepancy between the
two curves in this case.
Acknowledgements
The authors of this study would like to thank Mark
Courtney and Hercules Incorporated for the graphite
fabric used in these tests. We would also like to extend
thanks to the Center for Composite Materials, with spe-
cial thanks to Anthony Thiravong, for giving us access to
their equipment and supplies. Finally, we would like to
thank Doug Baker and Michael Davidson for their use-
ful contributions in preparing for the tests.
Construction and Building Materials 1994 Volume 8 Number 3 199
Flexural strengthening of concrete beams: M. J. Chajes at al.
Table 9 Contribution of composite fabric to flexural resistance
Fabric-reinforced beams Moment resisted by
steel-concrete couple
(N m)
Moment resisted by
fa bric-concrete couple
(N m)
Percentage of moment
resisted by fabric-concrete couple
1560
1535
1523
Nomenclature
nal reinforcement of concrete beams using fiber reinforced plas-
tics. ACI Struct. J. 1991,88,490-500
22 Rostasy, F.S., Hankers, e. and Ranisch, E.B. Strengthening of
R/C- and PIC-structures with bonded FRP plates. Advanced Com-
posite Materials in Bridges and Structures, CSCE, Sherbrooke,
Canada, 1992,pp. 253-263
23 Saadatmanesh, H. and Ehsani , M.R. Flexural strength of exter-
nally reinforced concrete beams. Proc. First Materials Engineering
Congr., ASCE. 1990, pp. 1152-1161
24 Saadatmanesh, H. and Ehsani, M.R. Fiber composite plates can
strengthen beams. Concr. Int., ACI 1990, 12(3),65-71
25 Saadatrnanesh, H. and Ehsani , M.R. RC beams strengthened with
GFRP plates. I: Experimental stud y. J. Struct . Eng. 1991, 117(11),
3417-3433
26 Karam, G.N. Optimal design for prestressing with FRP sheets in
structural members . Advanced Composite Materials in Bridges and
Structures, CSCE, Sherbrooke, Canada, 1992, pp. 277-285
27 Triantafillou, T.e. and Deskovic, N. Innovative prestressing with
FRP sheets: Mechanics of short-term behaviour. J. Eng. Mech.
' 1991, 117(7), 1652- 1672
28 Triantafillou, T.e., Deskovic, N. and Deuring, M. Strengthening
of concrete structures with prestre ssed fiber reinforced plastic
sheets . ACI Struct. J . 1992,89(3),235-244
29 Priestley, M.J ., Fyfe, E. and Scible, F. Column retrofit using
fiberglass /epoxy jackets. First Annual Seismic Research Workshop.
CalTrans , Sacramento. California. 1991, pp. 217-224
30 Katsumata, H. , Yoshirou, K. and Toshikaza, T. A study with
carbon fiber for earthquake-resistant capacity of existing rein-
forced concrete columns. Proc. Ninth World Conf on Earthquake
Engineering, Tokyo, Japan, 1988, Vol. 7, pp. 517-522
31 Saadatmanesh, B., Ehsani , M,R. and Li, M.W. Behavior of'exter-
nally confined columns. Fiber-Reinforced-Plastic Reinforcement
for Concrete Structures, SP 138, American Concrete Institute,
Detroit, 1993, pp. 249-265
32 Chajes, M.J., Januszka, T.F., Thompson, TA., Finch, W.W. and
Mertz, D.R. Shear strengthening of reinforced concrete beams
using externally applied composite fabrics . ACI Struct. J. In press
33 Dolan, c:, Rider , W., Chajes, M.J. and DeAscanis, M. Pre-
stressed concrete beams using non-metallic tendons and external
shear reinforcement. Fiber-Reinforced-Plastic Reinforcement for
Concrete Structures, SP 138, American Concrete Institute, Detroit,
1993, pp. 475--495
34 ACI Committee 318. Building code requirements for reinforced
concrete (ACI 318-89) Revised 1992. American Concrete Insti-
tute, Detroit, 1992
35 Park, R. and Paulay, T. Reinforced Concrete Structures, John
Wiley, New York , 1975
(a) At load corresponding to init ial yielding of steel reinforcement
Aramid 1541
E-glass 1529
Graphite 1519
(b) At ultimate load
Aramid
E-glass
Graphite
References
Hamoush, S.A. and Ahmad, S.H. Debonding of steel plate-
strengthened concrete beams. J. Struct. Eng. 1990,116(2),356--371
2 Jones, R., Swamy, R.N. and Ang, T.H. Under- and over-rein -
forced concrete beams with glued steel plat es. Int. J. Cern. Compos.
Lightweight Concr. 1982, 1(1), 19-32
3 Macdonald, M.D. and Calder, A.J.J. Bonded steel plating for
strengthening concrete structures. Int . J. Adhes. Adhes. 1982, 4,
119-127
4 Ong, K.C.G. , Mays, G.e. and Cusens, A.R. Flexural tests of steel-
concrete open sandwiches, Mag. Concr. Res. 1982, 34(120),
130-138
5 Roberts, T.M. and Haj i-Kazemi , H. Theoretical study of the
behaviour of reinforced concrete beams strengthened byextemal1y
bonded steel plates. Proc. Inst. Civ. Eng. 1989,87(2),39-55
6 Swamy, R.N. , Jones, R. and Bloxham, J.W. Crack control of
reinforced concrete beams through epoxy bonded steel plates.
Adhesion Between Polymers and Concrete, RILEM 86, 1986, pp.
542-555
7 Swarny, R.N., Jones , R. and Bloxham, J.W. Structural beha viour
of reinforced concrete beams strengthened by epoxy-bonded steel
plates. Struct . Eng. Part A 1987,65(2),59-68
8 Theillout, J.N. Repair and st rengthening of bridges by means of
bonded plates . Adhesion Between Polymers and Concrete, RILEM
86, 1986,pp. 601-621
9 Ballinger, CA. Development of composites for civil engineering,
Advanced Composite Mat erials in Civil Engineering Structur es,
ASCE, Las Vegas, Ne-ada, 1991, pp. 228-30i
10 Ballinger, e.A. Development of fiber-reinforced plastic products
for the construction market - how has and can it be done?
Advanced Composite Materials in Bridges and Structures, CSCE,
Sherbrooke, Canada, 1992, pp. 3-13
II Bank, t .c. Questioning composites. Civ. Eng. ASCE 1992, 63(1),
64-65
12 Head, P.R. Design methods and bridge forms for the cost effective
use of ad vanced composites in bridges . Advanced Composite
Materials in Bridges and Structures, CSCE, Sherbrooke, Canada,
1992, pp. 15-30
13 McCormick, F.e. Advancing structural plastics into the future .
J . Struct , ot ASCE 1988,114(3),235-243
14 Measures, R.M. Smart structure s - a revolution in civil engineer-
ing, Advanced Composites Mat erials in Bridges and Structure,
CSCE, Sherbrooke, Canada, 1992, pp, 31-59
15 Meier, U. Carbon fiber-reinforced polymers: modern materials in
bridge engineering. Struct, Eng. Int . 1992,2, 7-12
16 Saad atmanesh, H. and Ehsani, M.R. Application of'fiber-compo-
sites in civil engineering. Proc. 7th St ructures Congr., ASCE. 1989,
pp. 526--535
17 Sotiropoulos, S.N. and GangaRoa, B. Bridges systems made of
FRP components. 2nd Workshop on Bridge Engineering Research
in Progress, NSF, Reno, Nevada, 1990, pp. 295-298
18 An, W., Saadatmanesh, H. and Ehsani , M.R. RC beams streng-
thened with FRP plates. 2: Analysi s and parametric study.
J . Struct . Eng. 1991, 117(11),3434-3455
19 Meier, U. and Kaiser, H. Strengthening of structures with CFRP
laminates. Advanced Composite Mat erials in Civil Engineering
Structures, ASCE, Las Vegas, Nevada , 1991, pp. 224-232
20 Meier, U. , Deuring, M., Meier , H. and Schwegler, G. Strengthen-
ing of structures with CFRP laminates: Research and applications
in Switzerland. Advanced Composite Mat erials in Bridges and
Structures, CSCE, Sherbrooke, Canada, 1992, pp. 243-251
21 Ritchie, P.A., Thomas, D.A., Lu, L.W. and Conelly, G.M. Exter-
A
b
C
de
d
s
458.9
756.1
1095
1188
1661
1782
23.0
33.0
41.9
43.3
52.0
53.9
aramid-reinforced beam
width of beam
control beam
distance from extreme compression fibre to
centroid of tension fabric
distance from extreme compression fibre to
centroid of tension reinforcement
E-glass-reinforced beam
modulus of elasticity of concrete
modulus of elasticity of resin-impregnated
fabric
200 Construction and Building Materials 1994 Volume 8 Number 3
Flexural strengthening of concrete beams: M. J. Chajes et al.
!e"
if
is
h
G
h
p
resultant force in compression concrete
force in fabric reinforcement
force in tensile steel reinforcement
concrete compression stress
compressive strength of concrete from
cylinder test
maximum compressive stress of concrete
fabric tensile stress
steel tensile stress
specified yield strength of non-prestressed
steel reinforcement
graphite-reinforced beam
depth of beam
applied load
CTr
CTru
CT
s
beam reinforced with additional steel
strain in the concrete
ultimate compressive strain of the concrete
strain in fabric reinforcement
ultimate tensile strain of the fabric
strain in concrete at maximum stress
strain in tensile steel reinforcement
ratio of non-prestressed steel tension reinfor-
cement to cross-sectional area of beam
reinforcement ratio producing balanced
strain condition
stress in fabric reinforcement
ultimate tensile stress of the fabric
stress in steel tensile reinforcement
Construction and Building Materials 1994 Volume 8 Number 3 201

Vous aimerez peut-être aussi