See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/280206837

Reinforced HSC Beams Strengthened with CFRP Plates under Bending

Article  in  Kuwait Journal of Science & Engineering · January 2009

CITATIONS READS

0 15

1 author:

Seyed Hamid Hashemi

Arak University
32 PUBLICATIONS   305 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Seyed Hamid Hashemi on 21 July 2015.

The user has requested enhancement of the downloaded file.

Kuwait J. Sci. Eng. 36 (1B) pp. 1-31, 2009

Reinforced HSC beams strengthened with CFRP plates

under bending

SEYED HAMID HASHEMI*, ALI AKBAR MAGHSOUDI** AND

REZA RAHGOZAR***
*
Faculty of Engineering, Arak University, Arak, Iran
E-mail: h-hashemi@araku.ac.ir
**
E-mail: rahgozar@mail.uk.ac.ir ,
***
E-mail: maghsoudi.a.a@mail.uk.ac.ir

ABSTRACT
Poststrengthening and retro®tting is a growing reality, as existing structures are required
to meet the demands of modern society. Apart from the need to increase load capacity,
upgrading of a structure may be necessary due the deterioration of the structure on
account of corrosion or accidental damage, a change in the structural system, or to
rectify initial design and construction faults. A commonly observed mode of failure for
beams strengthened using carbon-®ber-reinforced polymer composite material plates is
due to the plate peeling o€ prematurely and unpredictably at relatively low magnitudes
of applied loading. End plate anchorages and long unanchored plate lengths, which can
add signi®cantly to the overall cost of a strengthening solution, overcome this problem.
The objective of this investigation is to study the e€ectiveness of the length of carbon
FRP laminate on performance of repaired reinforced high strength concrete (HSC)
beams. This objective is achieved by conducting the ¯exural four-point testing of
reinforced HSC beams with a di€erent amount of tensile reinforcement that is
strengthened with di€erent plate con®guration up to failure and calculating the ¯exural
response. The result of tests show that, in contrast with a control beam, initial cracking
loads of strengthened beams increase slightly, while ductility decreases and ultimate
loads increase considerably. Examination of the crack distribution indicated that the size
and density of the cracks were signi®cantly less in the strengthened beams than in the
control specimens, thereby producing a more durable system, which is less susceptible to
the ingress of water and other potentially corrosive solutions. After the experimental
study, three-dimensional nonlinear ®nite element models (FEMs) are adopted by
ANSYS to examine the structural behavior of the reinforced high strength concrete
(HSC) beams strengthened with FRP sheets. Lastly, a comparison between the ®nite
element analysis results and the experimental data available on the specimens is made.
2 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

The load de¯ection plots obtained from the numerical study show good agreement with
the experimental results.
Keywords: CFRP Plates; Finite Element Model; serviceability; HSC; ¯exural;
strengthening

INTRODUCTION
Strengthening, upgrading and retro®tting of existing structures are some major
challenges that modern civil engineering is currently facing. One of the most
promising answers to these needs is the use of strips made of ®ber reinforced
polymers (FRP) bonded to the tensile face of the member. Comprehensive
experimental investigations conducted in the past have shown that this
strengthening method has several advantages over traditional ones, especially
due to the high strength, low weight and improved durability of the composite
material.
In FRP-strengthened beams, failure may occur due to beam shear, ¯exural
compression, FRP rupture, FRP debonding or concrete cover ripping as
presented by Ascione & Feo (2000), Bonacci & Maalej (2000, 2001). Based on
experimental results conducted by Teng et al. (2003), the most common failure
mode is due to debonding of FRP plate or spalling of the concrete cover. These
failure modes are undesirable because the FRP plate cannot be fully utilized. In
addition, such premature failures are generally associated with a reduction in
deformability of the strengthened members. Premature failure modes are caused
by interfacial shear and normal stress concentration at FRP cut-o€ points and
at ¯exural cracks along the beam.
Extensive testing of such strengthened members has been carried out over the
last two decades. A number of failure modes for RC beams bonded with FRP
sot plates have been observed in numerous experimental studies to date (e.g.
Alagusundaramoorthy et al. 2003, Arduini & Nanni 1997, Chahrour & Soudki
2005, GangaRao & Vijay 1998, Hashemi et al. 2007, Maalej & Bian 2001,
Nguyen et al. 2001, Oh et al. 2003, Rahimi & Hutchinson 2001, Saadatmanesh
& Ehsani 1991 and Toutanji et al. 2006).
Based on existing studies, schematic representation of typical failure modes
observed in tests is shown in Fig. 1. These failure modes are termed: Type (1)
¯exural failure by crushing of compressive concrete which could happen before
or after yielding of tensile steel reinforcement; Type (2) rupture of the FRP
laminate after yielding of the steel in tension; Type (3) cover delamination at the
end of FRP (shear delamination of the concrete cover); Type (4) debonding of
the FRP from the concrete substrate: Type (4a) plate end interfacial debonding,
 Reinforced HSC beams strengthened with CFRP plates under bending 3

Type (4b) inter-facial debonding induced by ¯exural crack, Type (4c) interfacial
debonding induced by ¯exural shear crack, and Type (5) shear failure.

(a) Failure type (1) (b) Failure type (2)

(c) Failure type (3) (d) Failure type (4a)

(e) Failure type (4b) (f) Failure type (4c)

(g) Failure type (5)

Figure 1. Failure modes of RC beams ¯exurally strengthened with a FRP plate

Failure types (3) and (4a) have been studied experimentally and analytically
by Ascione & Fao (2000), Maalej & Bian (2001) and Malek et al. (1998). These
types of failure are common in cases where the ends of the FRP sheets are not
properly anchored. Failure Types (4b) and (4c) depend on the bond-slip
4 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

behaviour between FRP sheets and concrete. According to Sebastian (2001) and
Teng et al. (2003), the corrosion of longitudinal steel bars and the change of the
reinforcing bar ratio in the vicinity of large bending moments and shear forces
increase the probability of these types of failures. When FRP reinforcement is
being used to increase the ¯exural strength of a member, it is important to verify
that the member will be capable of resisting the shear forces associated with the
increased ¯exural strength. To avoid failure Type 5, the potential for shear
failure of the section should be considered by comparing the design shear
strength of the section to the required shear strength. If additional shear
strength is required, FRP laminates oriented transversely to the section can be
used to resist the applied shear forces.
In spite of many recent studies on the behaviour of reinforced concrete beams
strengthened using FRP composites, the e€ect of the reinforcing bar ratio on the
behaviour and the strength of these beams has not yet been explored. The
reinforcing bar ratio of beams a€ects the pattern and the width of cracks due to
the e€ect of bending and shear. The in¯uence of the FRP composites on ¯exural
strengthening of reinforced concrete beams should depend on the width and
spacing of these cracks. The ductility of beams, also, depends on the reinforcing
bar ratio. This paper presents the test results of six HSC beam specimens
strengthened by carbon FRP plates. The main variable parameter in these tests
are the reinforcing bar ratio and CFRP length. The types of failure of these
beams are also investigated.
In the case of a beam in four-point bending, the ratio of the length of the
bonded plate within the shear span of the beam (p), to the shear span of the
beam (a), generally referred to as the p/a ratio, is an important parameter in this
respect. In this study, the e€ect of di€erent p/a ratios on the ultimate response of
six reinforced HSC beams is studied in an attempt to develop a relationship
between the bonded length of a plate within the shear span and the load at
which failure by plate peel-o€ occurs. Two di€erent plate lengths were
considered, and although each plate length resulted in di€erent ultimate loads,
the strain gradient in the bonded plate length in the shear spans was the same
for all beams when failure by plate peel-o€ occurred.

TEST BEAM DETAILS

Test Specimens
The length, width, and depth (l2b2h) of all beams were kept as 300021502250
mm. Each concrete beam was reinforced with two 16-mm diameter (A series)
and two 22-mm diameter (B series steel bars) for tension and two 10-mm-
diameter steel bars for compression along with 10-mm-diameter bars at a
 Reinforced HSC beams strengthened with CFRP plates under bending 5

spacing of 90 mm center-to-center for shear reinforcement. The spacing of

stirrups and minimum and maximum reinforcement ratios are in accordance
with the provision of the American Concrete Institute (ACI318-05).
Di€erent types of electrical resistance disposable strain gauges, manufactured
by TML Measurements Group (Japan), were pasted on the CFRP sheets and on
internal reinforcing bars at di€erent locations. The demec and electrical gauges
were also attached along the height of beams to measure the concrete strains.
These values can be used to ®nd out the strain distribution and the moving
neutral axis depth of the beams tested. All beams were loaded in four-point
bending to failure with a clear span of 2.7 m, and loading points were located at
450 mm on either side of the mid-span location. The load was applied step-by-
step up to failure in a load control manner of test beams. During the test, crack
width, the strains on steel, CFRP plates and concrete, and vertical de¯ections
were measured using LVDTs to an accuracy of 0.01 mm (Fig. 2). The strain
gauges, LVDTs, and the load cell were connected through a data acquisition
system to a computer and the data was recorded and stored in the computer.

Figure 2. Beam details and measurement schemes for half of the test specimen (unit: millimeter)

The remaining four beams were strengthened with di€erent CFRP plate
lengths bonded symmetrically about the midspan of the beam with various
tensile bar reinforcement ratios. For the ®rst pair of strengthened beams, the
plates were bonded along almost the full length of the beam (p/a ratio of 0.94).
The reactions developed during the tests thus anchored the plates to the beams
at the support locations. In subsequent pairs of beams, the e€ect of reduced
plate length p/a ratio of 0.67 were investigated.

Material properties
For each beam three 10021002100mm concrete cube specimens were made at
the time of casting and were kept with the beams during curing. The average 28-
day concrete cube strength (fcu) was 96.2 MPa. The relationship of cylinder
strength (f'c) and cube strength assumed as f'c=0.8 fcu, thus the average
6 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

compressive strength (f'c) was 77 MPa. The measured yield and maximum
tensile strength of the 10 and 16 mm rebar was 420.6, 634.1 and 412.5, 626.4
MPa respectively.
The Sika CarboDur S-512 system, comprising CFRP composite material
plates and a two-part epoxy resin adhesive (Sikadur 30), was used for the
external plating on the strengthened beams. The Sika CarboDur S laminate, 1.2
mm thick and 50 mm wide, is a pultruded composite consisting of unidirectional
carbon ®bers embedded within an epoxy resin. This material has a Young's
modulus of 165 GPa with a lower bound ultimate tensile strength of 2,800 MPa
parallel to the ®bers. Prior to bonding of the plates, the beams were ground
using a mechanical grinder to expose the aggregate and to obtain a clean sound
surface, free of all contaminants such as cement laitance, grease, and dirt, and
the plate surface was wiped clean with an acetone solution. The adhesive was
applied to both the plate and the beam in a pro®led shape such that the spread
of the adhesive from one surface to the other prevented the formation of air
bubbles within the bonded system. Granular beads, 3 mm in diameter, were
embedded in the adhesive to maintain a uniform thickness, and the adhesive was
given a minimum of three days to cure prior to testing.

Major test variables

The main test variables considered in the present study include the FRP plate
lengths and tensile bars. The FRP plate lengths varied from 210 to 260 cm and
the bar reinforcement ratio varies from 1.2 to 2.4%. The test program is
summarized in Table 1. Of the six beams tested, two were set aside as control
beams and were not strengthened (AH0, BH0), two beam were strengthened
with a 260 cm plate length of CFRP as the p/a ratio of 0.94 (AHF, BHF) and
two beams were strengthened with a 210 cm plate length of CFRP (AHD,
BHD) at the p/a ratio of 0.67. The ®rst letter of beam name refers to bar
reinforcement ratio as A=1.2% and B=2.4%, the letter H refers to high
strength concrete and the last letter (F & D) refer to p/a ratio.

Table 1. Test parameters and specimen identi®cations

LFRP
Test beam As (mm) p/a ratio
AH0 2816 0 0
AHF 2816 2600 0.94
AHD 2816 2100 0.67
BH0 2822 0 0
BHF 2822 2600 0.94
BHD 2822 2100 0.67
 Reinforced HSC beams strengthened with CFRP plates under bending 7

Test Results and Discussions

The cracking patterns and failure of tested beams are shown in Fig. 3. The control
beams (AH0, BH0) without strengthening plates were designed to fail in ¯exure, and
failure was by crushing of the concrete in the compression zone after tension steel
yield. All of the strengthened test beams exhibited the debonding of the FRP plate
from the substrate of the concrete and failed in the same manner.

Figure 3. Failure con®guration of control and FRP beams at ultimate state

8 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Tables 2 and 3 summarize the test results in terms of ¯exural loading capacity
and de¯ection for yield and ultimate stage at mid-span section. Table 2 and 3
also show the increase of yield and peak load according to the various
strengthening lengths of CFRP. The rates of increase of peak loads varied from
19.3 to 26.8%, depending upon the strengthening method. The results clearly
demonstrate the accepted bene®cial e€ects of CFRP plates with regard to
sti€ening and strengthening of the beams.

Table 2. Mid-span test results of the control and CFRP strengthened

beams at yield stage
Yield Increase Decrease Pyield
load over over Strain (micron)
Test Pyield control control Tensile Compressive
beam (kN) (%) y …mm† (%) CFRP rebar rebar Concrete
AH0 63.93 --- 11.89 --- --- 1972 -3261 -2007
AHF 76.7 19.9 11.13 -6.4 2872 1980 -550 -1190
AHD 71.3 11.5 10.15 -14.6 2779 1943 -376 -1317
BH0 122.2 --- 13.325 --- --- 2063 -703 -1450
BHF 122 0 13.04 -2.1 2446 1959 -889 -1794
BHD 124 1.4 12.18 -8.6 3061 1940 -953 -1452

Table 3. Mid-span test results of the control and CFRP strengthened

beams at ultimate stage
Peak Increase Decrease Pult
load over over Strain (micron)
Test Failure Pult control control Tensile Compressive
beam mode (kN) (%) y …mm† (%) CFRP rebar rebar Concrete
AH0 Concrete 81.25 --- 72 --- - 2316 2245 3600
Crushing
AHF CFRP 103 26.8 23.23 67.7 5308 4996 -525 1503
Debonding
AHD CFRP 101 24.3 19.96 72.3 5521 12680 -469 1911
Debonding
BH0 Concrete 134.9 --- 57 --- - 14224 45 3480
Crushing
BHF CFRP 161 19.3 24.38 57.2 6184 7508 -1122 2861
Debonding
BHD CFRP 163.3 21 25.05 56 6858 14872 -813 2710
Debonding
 Reinforced HSC beams strengthened with CFRP plates under bending 9

The strain response of FRPs is di€erent from that of conventional steel,

which yields after elastically, deforming to relatively small values of strain
0.2% for Grade 60 (410 MPa) and 0.14% for Grade 40 (280 MPa). FRP
materials exhibit elastic deformation to relatively large strain values before
rupture. As a result, when FRPs are used for ¯exural strengthening of
concrete beams reinforced with conventional steel, the steel reinforcement
may yield before the FRP contributes any additional capacity to the beam.
Therefore, it can be dicult to obtain a signi®cant increase in yield load or
sti€ness for a beam. When an increase in beam yield load or sti€ness is
required, larger cross sections of FRPs must be used (before the steel
yields), which generally increases the cost of strengthening. Although using
some special, low-strain ®bers, such as ultra-high-modulus carbon ®bers may
appear to be a solution. They can result in brittle failures due to ®ber
failure. Taking advantage of the high strength of FRPs during ¯exural
strengthening of RC beams is limited by the bond capacity between them
and the concrete surface. In many cases, debonding occurs (Nguyen et al.
2001 and Rahimi & Hutchinson 2001) at stress levels that are a small
fraction of the FRPs' strength.
As the amount of steel reinforcement increases, the additional strength
provided by the carbon FRP external reinforcement decreases. The same
amount of CFRP reinforcement increased more than 27% the ¯exural strength
of a lightly reinforced beam (20% of balanced ratio), but only increased by
19.3% the strength of a moderately reinforced beam (40% of balanced ratio).
The tension steel in control beams AH0 and BH0 reached its yield
strength before the compressive strain in concrete reached 0.003 and the
beams failed by crushing of concrete. Even though the control beams failed
by crushing of concrete, since the failure was initiated by yielding of tension
steel, the mode of failure was mentioned to be under reinforced tension
failure, thus the behavior of the two control beams, AH0 and BH0, was a
ductile ¯exural response.
For control beams after the ®rst visible cracks observed, the cracking
became extensive and crack widths increased steadily. The shape of the load
de¯ection curves indicates a loss of sti€ness at a load of approximately 64
kN for AH0 and 122 kN for BH0. This was due to the yielding of the
tensile reinforcement and occurred at a midspan de¯ection of 21 mm for
AH0 and 13.3 mm for BH0. After this point, large ¯exural cracks opened
during the test and eventual ultimate collapse was by concrete crushing
within the compression zone, a photograph of which is presented in Fig. 4.
The ultimate loads recorded were 81.25 and 134.9 kN for AH0 and BH0,
respectively.
10 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Figure 4. Flexural failure picture of control beam AH0

From the experimental observation, the debonding failure can be

explained as follows: due to the ¯exural cracks formed in the constant
moment region as the load increased, the bond between FRP and concrete
started to fracture at a certain load level, and the failure propagated
towards the shear span until most parts of the FRP composites detached
from the concrete substrate. It can be seen that the bond between the FRP
and concrete surface is not strong enough to ensure the rupture of the
composites with one layer of carbon ®ber plate, thus the FRP-concrete bond
strength controls the failure mode in this study. Fig. 5 shows a photograph
of this failure mode.

Figure 5. Debonding failure of FRP from concrete substrate

The load de¯ection response for each of the test beams is plotted in Fig. 6 for
 Reinforced HSC beams strengthened with CFRP plates under bending 11

mid-span and in Fig. 7 for load point section. In general, the strengthened
beams were sti€er and less ductile than the control specimens, and they had
higher ultimate loads.

Figure 6. Load-de¯ection response of test beams at midspan

Figure 7: Load-De¯ection Response of Test Beams under Point Load

12 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Ductility
Ductility is an important factor for any structural element or structure especially
in the seismic regions. A ductile material is one that can undergo large strains
while resisting loads. When applied to RC members, the term ductility implies
the ability to sustain signi®cant inelastic deformation prior to collapse. Since
CFRP repair is a fairly new innovation and the HSC behavior is also like a
brittle material. Therefore, understanding the e€ect of such materials on the
ductility of a RC beam is critical.
Ductility has generally been measured by a ratio called a ductility index or
factor (). The ductility index is usually expressed as a ratio of rotation (),
curvature (), or de¯ection (displacement) () at failure to the corresponding
property at the yielding of steel (Maghsoudi & Akbarzadeh 2006). In the case of
beams strengthened with FRP laminates, there is usually no clear yield point.
However, it was shown that de¯ection based on tension steel yielding can be
used as a criterion of ductility to evaluate comparative structural performance
of FRP bonded RC beams.
The ductility index in this study is obtained based on de¯ection (d )
computation, and is de®ned as the mid-span de¯ection, at peak load,
divided by the mid-span de¯ection at the point where the steel starts
yielding. Table 2 and 3 show the test results of the beams for yield and
ultimate stage and Table 4 shows the experimental de¯ection ductility ratio
and percent decrease of ductility with respect to the control beam for each
of the specimens.

Table 4. Experimental ductility ratio of the test beams

Decrease
De¯ection over control
ductility ratio beam
Series Test beam … ˆ u †
 (%)
y
AH0 6.05 ----

A AHF 2.1 65.2

AHD 1.97 67.4

BH0 4.25 ----

B BHF 1.87 56

BHD 2.05 51.8

 Reinforced HSC beams strengthened with CFRP plates under bending 13

Load-Strain Responses:
The relationship between concrete strains measured experimentally on the
compression face at mid-span for each of the test beams is plotted and shown in Fig.
8. There is a similar increase in strain for all the beams at low loads. However,
cracking of the concrete in the tension zone results in larger increments of strain in
the control specimens. For these beams, concrete strain varies almost linearly with
load, after initial cracking, until yielding of the tension steel. Following yield, strain
increases rapidly with each increment of load, and ultimately the concrete crushes as
the beam collapses. On the other hand, concrete strain in the strengthened beams
remains more or less linear up to failure of the beam and is not signi®cantly a€ected
by concrete cracking or yielding of the tension steel. These results demonstrate that
the e€ect of the plate is to reduce strain in the compression ®bers of the beam. The
presence of the plate draws the neutral axis lower in the section and hence, places a
greater volume of concrete in compression, resulting in lower strain and enabling a
more ecient use of the existing material. Thus, externally bonded CFRP plates
may also be bene®cially used to reduce concrete compressive stresses in addition to
acting as additional tensile reinforcement.

Figure 8. Mid-span load- compression top ®ber of concrete strain

The relationship between concrete, CFRP and tensile bar steel strains measured
experimentally at mid-span section (Sec. 1) and load point section (Sec. 2) for all
strengthened beams are plotted in Figs. 9-12. Figs. 9-12 indicate that each curve
consists of almost three straight lines with di€erent slopes. The ®rst turning point
(A) indicates the cracking of concrete in the tension zone. The second turning
point (B) refers to the yielding tension steel. The yielding and maximum load
(ultimate load) can be found for each beam from its load-strain curve.
14 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

For all strengthened beams, the tensile steel and CFRP strains are essentially
the same at loads below the cracking of the concrete. After cracking, the strains
in steel exceeded those of the CFRP laminate. As the load approached the
yielding load for the strengthened beam, the strains in steel increased more
rapidly than those in the CFRP. This is because the CFRP had begun to debond
from the concrete surface near the cracks. It was noted that the tensile steel
strains were always higher than the CFRP strains.

Figure 9. Mid-span and load point load-strain curves of CFRP, tensile steel
and extreme top concrete ®ber for beam AHD

Figure 10. Mid-span and load point load-strain curves of CFRP, tensile steel
and extreme top concrete Fiber for beam AHF
 Reinforced HSC beams strengthened with CFRP plates under bending 15

Figure 11. Mid-span and load point load-strain curves of CFRP, tensile steel
and extreme top concrete ®ber for beam BHD

Figure 12. Mid-span and load point load-strain curves of CFRP, tensile steel
and extreme top concrete ®ber for beam BHF
16 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

FINITE ELEMENT ANALYSIS APPROACH

Nonlinear FE analysis is performed using ANSYS, a general purpose ®nite
element program. This section introduces the elements chosen from the software
library and the analytical approach and assumptions used in the analysis. In this
study due to symmetry, only one loading plate and one support plate are
needed. The support is a 5x15x2 cm steel plate, while the plate at the load point
is 10x15x2 cm. The combined volumes of the plate, support, and beam with the
FE mesh for half length of the beam model is shown in Fig. 13.

Figure 13. FE mesh of concrete, steel load plate and steel support plate

Material Modeling

Concrete:
The SOLID65, three-dimensional (3D) reinforced concrete solid element, is used
to represent the concrete in the models. The element, using a 22222 set of
Gaussian integration points, is de®ned by eight nodes having three translational
degrees of freedom at each node. This element is capable of cracking in tension
and crushing in compression, although during this study, it was found that when
the crushing capability of the concrete is turned on, the ®nite element beam
models fail prematurely. Basically, a pure compression failure of concrete is
unlikely. In a compression test, the specimen is subjected to a uniaxial
compressive load and secondary tensile strains induced by Poisson's e€ect occur
perpendicular to the load. Because concrete is relatively weak in tension, these
actually cause cracking and the eventual failure. Therefore, the crushing
 Reinforced HSC beams strengthened with CFRP plates under bending 17

capability was turned o€ and the cracking of the concrete controlled the failure
of the ®nite element models.
The most important implementation of the SOLID65 element is the proper
de®nition of nonlinear material properties. The responses of concrete under
loading are characterized by distinct nonlinear behavior. Complete stress-strain
curves for concrete are needed to accurately predict a whole range of beam
behavior, from service loading up to failure and post-failure responses.
Additionally, the descending branch is needed since a portion of the concrete
compression zone is usually in this range of strains at the ultimate limit state.
The response of a RC structure under load depends largely on the stress-
strain relation of the materials and the stress magnitude. In compression, the
stress-strain relation of concrete displays nearly linear elastic behavior up to
about 30 % of the compressive strength. After reaching the ultimate strength
(f 0c ) the diagram descends into a softening area, and follows until failure takes
place by crushing at an ultimate strain ("cu ). The concrete behavior under
tension, on the other hand, is di€erent to compression. The stress-strain curve
here is nearly linearly elastic up to the maximum tensile strength. After this
point, the concrete cracks and the strength decrease gradually to zero.
The uniaxial stress-strain behavior of concrete has been proposed by many
empirical formulas. The simplest of the linearly elastic-perfectly plastic model,
which was used by Uy (1998), is shown in Fig. 14. To de®ne a failure surface for
concrete, two strength parameters, i.e., ultimate tensile and compressive
strengths are needed. A three-dimensional failure surface for concrete is shown
in Fig. 15 that represents the 3-D failure surface for states of stress, which are
biaxial or nearly biaxial.

Figure 14. Concrete stress-strain curve (Uy 1998)

18 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

In Fig. 15, f 0c and fr are the ultimate uniaxial compressive and tensile strength
of the concrete, respectively. If the most signi®cant nonzero principal stresses are
in the xp and yp directions, the three surfaces presented are for zp greater than
zero, equal to zero, and less than zero. Although the three surfaces, shown as
projections on the xp - yp plane, are nearly equivalent, and the 3-D failure
surface is continuous, the mode of material failure is a function of the sign of zp .
For example, if xp and yp are both negative and zp is slightly positive, cracking
would be predicted in a direction perpendicular to the zp direction. However, if
zp is zero or negative, the material is assumed to crush (ANSYS 2003).

Figure 15. 3-D failure surface for concrete

Reinforcing Steel:
In addition to the rebar capability of the element, the reinforcing steel can be
modelled in ANSYS using a series of two node link (truss) elements, called
LINK8, which has three degrees of freedom, translations in the nodal x, y, and z
directions, and which is also capable of plastic deformation. Unlike concrete,
steel is very uniform and as such, generally the speci®cation of a single stress-
strain relation is adequate to de®ne it numerically. Typical stress-strain curves
for reinforcing steel bars used in concrete construction are obtained from
standard tensile tests. Practically speaking, steel exhibits the same stress-strain
curve in compression as in tension. The stress-strain relation at ®rst exhibits an
initial linear elastic portion, followed by a yield plateau until a strain hardening
region is reached in which stress again increases with strain before eventually
dropping o€ as microscopic fracture occurs.
 Reinforced HSC beams strengthened with CFRP plates under bending 19

As for the program, it is mentionable that ANSYS does not introduce a speci®c
stress-strain curve for steel reinforcement. In order to model the behavior of the
deformed steel reinforcement, several standard tests have been carried out on the
actual samples of the reinforcements that were used for fabrication of the
specimens, and an average of the test results was used to model the stress-strain
curve of the steel reinforcements using multilinear modelling. For this, the
nonlinear stress-strain relation was approximated by a series of straight line
segments in such a way that the critical part of the curve, such as onset of steel
yielding and strain hardening, were simulated by piecewise linear models. The
multilinear models of the stress-strain curves which were used by Schneider (1998)
show a good agreement with experimental results. This shown in Fig. 16.

Figure 16. Idealized Stress-Strain curve for steel (Schneider 1998)

Fiber-Reinforced Polymer:
The SOLID46, 3D-layered structural solid element, is used to represent the FRP
materials. The element has eight nodes with three translational DOFs at each
node. Assuming a perfect interlaminate bond, no slippage is allowed between
the element layers. The FRP laminates are considered brittle materials, and the
stress-strain relationship is roughly linear up to failure. Consequently, in this
study it is assumed that the stress-strain relationships for the FRP laminates are
linearly elastic.

Nonlinear Analysis Approach

The nonlinear behavior of the beams under loading is caused not only by
the transition of concrete from an uncracked to cracked state, but also by
20 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

the nonlinear material properties of concrete in compression and in the steel

as it yields. To obtain accurate predictions, Newton-Raphson equilibrium
iteration is used to resolve for the nonlinear behavior of the models using
an incremental process controlled by force and displacement convergence
tolerances. To avoid non-convergence, the load must be increased gradually,
and tolerances in both force and displacement criteria may have to be
gradually increased along the loading history. In the sensitivity study, the
maximum force and displacement convergence tolerances required were 0.5
and 1%, respectively. The maximum number of iterations per loading
increment is 50.
In ANSYS, there are four convergence criteria: F (force), M (moment), U
(displacement), and ROT (rotation). In this study, the convergence was
sought by initially trying the default convergence settings. If convergence did
not occur, the number of substeps was increased. In a case in which the
force convergence was still unachievable, the convergence tolerance was
relaxed. For this, the convergence tolerance limits were increased to a
maximum of 5 times the default tolerance limits (0.5% for force checking
and 5% for displacement checking). In order to help with the convergence,
a strategy was to apply displacements to the system rather than the loads.
The forces required to produce these pre-chosen displacements were then
considered as the beam-tip loads. The entire load-displacement curves of the
specimens were captured at the end of the analysis using this strategy. For
this, the response of the structure was recorded at every 50 substep
intervals. The failure was de®ned as the situation at which the solution for
the last load increment did not converge. The program then noted that the
models exceeded the displacement limitation of the ANSYS program,
indicating a signi®cantly large de¯ection.

Analysis Assumptions:
The following are the general analysis assumptions made for the test beam
models in this study in order to provide reasonable simulations for the complex
behavior:
1 - The bond between each element/material type is assumed perfect; that is,
there is no slippage between concrete and reinforcing steel bars, concrete
and FRP laminates, and di€erent FRP layers. Unless the failure mode of a
structure involves a bond failure, the perfect bond assumption used in the
structural modeling will not cause a signi®cant error in the predicted load-
de¯ection response.
 Reinforced HSC beams strengthened with CFRP plates under bending 21

2 - A constant Poisson's ratio of 0.2 is assumed for concrete throughout the

loading history.
3 - For the shear transfer coecients () for closed and open cracks in the
SOLID65 element, values can range from 0.0 to 1.0. A value of 0.0 refers to
a smooth crack, while 1.0 refers to a rough crack. These factors are used to
determine how much shear force can be transferred across open or closed
cracks. For closed cracks,  is always assumed to equal 1.0. However, for
an open crack,  varies from 0.05 to 0.5 and in this study, a  value of 0.2
for open cracks is used. This value is selected based on results in
Chansawat et al. (2006).
4 - The concrete is assumed to be isotropic prior to cracking and orthotropic
after cracking. The steel is assumed to be isotropic. The FRP material is
assumed to be specially orthotropic and transversely isotropic. That is, the
material properties in the two directions that are both perpendicular to the
®ber direction are identical.
5 - Time-dependent nonlinearities such as creep, shrinkage, and temperature
change are not included in this study.

COMPARISON BETWEEN EXPERIMENTAL DATA

AND NUMERICAL RESULTS
Load- Displacement Curves:
In this section, the numerical results for all specimens are presented and
compared with experimental values. Figures 17 and 18 contain a comparison
between the load-displacement curves predicted by ANSYS and the test results
for all specimens. As is seen, the agreement is reasonable. The results of ANSYS
match the plain specimen better than the strengthened specimens. This may be a
result of bond slip between FRP and concrete that is ignored in the current
analysis.
In general, the strengthened beams were sti€er and less ductile than the
control specimens with a higher ultimate load. As a result, compared to a
beam reinforced heavily with steel only, beams reinforced with both steel
and CFRP have adequate deformation capacity, in spite of their brittle
mode of failure.
22 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Figure 17. Comparison between load-displacement curves predicted by ANSYS

and the test result for A group specimens

Figure 18. Comparison between load-displacement curves predicted by ANSYS

and the test result for B group specimens

Table 5 shows a summary of the experimental and numerical ¯exural

behavior of all test beams in terms of ¯exural loading capacity and de¯ection.
The results clearly demonstrate the accepted bene®cial e€ects of CFRP plates
with regard to sti€ening and strengthening of the beams.
 Reinforced HSC beams strengthened with CFRP plates under bending 23

Table 5 shows the increase of yield and peak load according to the various
strengthening CFRP lengths. The rates of increase of peak loads varied from
19.3 to 26.8% for experimental data and 17.3 to 43% for numerical results,
depending upon the strengthening method.

Table 5. Numerical and experimental mid-span de¯ection and load in yield

and ultimate stage of RC beams strengthened with CFRP plates
Yield Stage Ultimate Stage
Increase Decrease Increase Decrease
over over over over
ecruoS

Test Load control control Load control control

beam Py(kN) (%) y …mm† (%) Pu(kN) (%) y …mm† (%)
AH0 63.93 --- 11.89 --- 81.25 --- 72 ---
latnemirepxE

AHF 76.7 20 11.13 6.3 103 26.8 23.23 67.7

AHD 71.3 11.5 10.15 14.6 101 24.3 19.96 72.3

AH0 69.2 --- 13.5 --- 75.4 --- 94.3 ---
laciremuN

AHF 75.7 9.3 12.1 10.4 100 32.6 21.3 77.4

AHD 72.6 4.9 12.8 5.2 107.8 43 23.2 75.4

BH0 122.2 --- 13.325 --- 134.9 --- 57 ---
latnemirepxE

BHF 122 0 13.04 2.1 161 19.3 24.38 57.2

BHD 124 1.5 12.18 8.6 163.3 21 25.05 56

BH0 132.4 --- 17.3 --- 140.2 --- 68.9 ---
laciremuN

BHF 134.2 1.3 9.8 43.3 172.3 22.9 28.2 59.1

BHD 141.5 6.9 11.3 34.7 164.4 17.3 29.4 57.3

SERVICEABILITY PERFORMANCE BASED

ON EXPERIMENTAL RESULTS
The widths of cracks in ¯exural members depend upon crack spacing, the
quality of the bond between concrete, reinforcing bars and FRPs, and above all,
strain in reinforcement. For steel-reinforced sections, ACI 318-05 requires that
the nominal moment strength, multiplied by the strength reduction factor
(=0.9), be equal to or greater than the moment Ms at service multiplied by an
appropriate load factor. The same code requires a load factor of 1.2 and 1.6 for
24 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

dead load and live loads, respectively. Considering an average load factor of 1.4,
the service stress in steel is (0.9/1.4) times the speci®ed yield stress. This is
approximately equal to 64% of the speci®ed yield stress.
For the steel used in this paper, the speci®ed yield stress is about 400 MPa (60
ksi); the resulting maximum strain in the reinforcement at service is
approximately 1286210-6. When the service strain in steel is permitted to reach
1286210-6, a width of cracks of 0.4 mm is allowed in codes and standards.
Fig. 19 presents the measured maximum observed ¯exural crack width versus
the maximum applied loads for tested beams. The crack width measurements
were made using a hand-held microscope of 40X magni®cation with a least
count of 0.02mm.
For serviceability limit state during the beams test, the experimental values of
steel tensile strain ("s ), the extreme layer of concrete compressive strain ("c ),
strain of FRP, crack width (for four di€erent environmental conditions, i.e., 0.1-
0.4mm), and de¯ection are measured at mid-span section (Sec. 1) and load point
section (Sec. 2) and given in Table 6. However, their normalized experimental
values are given in Table 7.

Figure 19. Experimental applied load versus crack width for tested beams
 Reinforced HSC beams strengthened with CFRP plates under bending 25

Table 6. Experimental measured values for serviceability limit state at mid-span

and load point section
STRAIN210-6 (mm/mm)
CRACK TENSILE DEFLECTION
WIDTH TEST LOAD BAR CONCRETE FRP (mm)
(mm) BEAM (kN) Sec. 1 Sec. 2 Sec. 1 Sec. 2 Sec. 1 Sec. 2 Sec. 1 Sec. 2
AH0 2.83 83 30 55 57 - - 0.52 0.4

0.1 AHF 16.7 536 406 166 757 686 663 1.7 1.56

AHD 15.33 333 311 410 320 707 604 1.09 1.12

AH0 5.67 167 65 110 115 - - 1.1 0.85

0.2 AHF 48 1147 1114 775 1225 1603 1749 6.33 5.59

AHD 37.33 960 935 854 570 1576 1318 4.72 4.32

AH0 8.5 250 100 165 179 - - 1.61 1.31

0.3 AHF 70 1699 1678 1010 1390 2569 2200 9.82 8.62

AHD 60 1595 1570 1072 1311 2377 2061 8.23 7.45

AH0 11.5 340 132 220 230 - - 2.2 1.72

0.4 AHF 84 2955 5309 1285 1932 3869 3299 13.9 12.12

AHD 84 9125 5570 1509 1578 4230 3625 14.83 13.13

BH0 7.66 101 119 75 1098 - - 0.55 0.48

0.1 BHF 31.3 542 469 509 432 421 984 3.12 2.84

BHD 37.33 592 497 585 263 1185 893 3.02 2.85

BH0 15.3 202 238 149 2202 - - 1.09 1.08

0.2 BHF 48.7 812 725 659 728 826 1416 4.92 4.44

BHD 57.33 919 773 808 527 1666 1289 5.03 4.62

BH0 28.3 481 480 392 2405 - - 2.32 2.13

0.3 BHF 109.33 1756 1616 1570 1347 2144 2721 11.7 10.32

BHD 105.3 1636 1406 1011 1309 2668 2180 10.07 9.03

BH0 42.29 728 721 517 2576 - - 3.91 3.59

0.4 BHF 122 1959 1791 1794 1432 2446 2964 13.04 11.51

BHD 134 2315 1872 1741 1540 3432 2817 13.5 12.11
26 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Table 7. Normalized test results for serviceability limit state of mid-span

and load point section

CRACK STRAIN210-6 (mm/mm) DEFLECTION

WIDTH TEST LOAD TENSILE BAR CONCRETE (mm)
(mm) BEAM (kN) Sec. 1 Sec. 2 Sec. 1 Sec. 2 Sec. 1 Sec. 2
AH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.1 AHF 5.90 6.46 13.53 3.02 13.28 3.27 3.90

AHD 5.42 4.01 10.37 7.45 5.61 2.10 2.80
AH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.2 AHF 8.47 6.87 17.14 7.05 10.65 5.75 6.58

AHD 6.58 5.75 14.38 7.76 4.96 4.29 5.08
AH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.3 AHF 8.24 6.80 16.78 6.12 7.77 6.10 6.58

AHD 7.06 6.38 15.70 6.50 7.32 5.11 5.69
AH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.4 AHF 7.30 8.69 40.22 5.84 8.40 6.32 7.05

AHD 7.30 26.84 42.20 6.86 6.86 6.74 7.63
BH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.1 BHF 4.09 5.37 3.94 6.79 0.39 5.67 5.92

BHD 4.87 5.86 4.18 7.80 0.24 5.49 5.94
BH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.2 BHF 3.18 4.02 3.05 4.42 0.33 4.51 4.11

BHD 3.75 4.55 3.25 5.42 0.24 4.61 4.28
BH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.3 BHF 3.86 3.65 3.37 4.01 0.56 5.04 4.85

BHD 3.72 3.40 2.93 2.58 0.54 4.34 4.24
BH0 1.00 1.00 1.00 1.00 1.00 1.00 1.00

0.4 BHF 2.88 2.69 2.48 3.47 0.56 3.34 3.21

BHD 3.17 3.18 2.60 3.37 0.60 3.45 3.37
 Reinforced HSC beams strengthened with CFRP plates under bending 27

The tensile bar and concrete extreme compressive strain values are converted
in terms of stress for di€erent environmental ¯exural crack widths (i.e., 0.1-
0.4mm) permitted in almost all codes, and the obtained values are shown in
Table 7 for two sections. The ratios of fs/fy and fc/f'c are presented in Table 8 for
convenience.

Table 8. Experimental values of concrete and steel strain and stress for di€erent
permissible ¯exural crack widths of the control and FRP strengthened beams
TENSILE CONCRETE
CRACK BAR STRESS STRESS
WIDTH TEST fs(Mpa) fs/fy fc(Mpa) fc/f'c
(mm) BEAM Sec. 1 Sec. 2 Sec. 1 Sec. 2 Sec. 1 Sec. 2 Sec. 1 Sec. 2
AH0 16.6 6 0.04 0.01 2.06 2.14 0.03 0.03
0.1 AHF 107.2 81.2 0.26 0.20 6.15 26.26 0.08 0.34
AHD 66.6 62.2 0.16 0.15 14.79 11.66 0.19 0.15
AH0 33.4 13 0.08 0.03 4.10 4.29 0.05 0.06
0.2 AHF 229.4 222.8 0.56 0.54 26.83 40.18 0.35 0.52
AHD 192 187 0.47 0.46 29.29 20.20 0.38 0.26
AH0 50 20 0.12 0.05 6.12 6.62 0.08 0.09
0.3 AHF 339.8 335.6 0.83 0.82 34.01 44.68 0.44 0.58
AHD 319 314 0.78 0.77 35.83 42.55 0.47 0.55
AH0 68 26.4 0.17 0.06 8.11 8.47 0.11 0.11
0.4 AHF 410 410 1.00 1.00 41.84 57.88 0.54 0.75
AHD 410 410 1.00 1.00 47.78 49.52 0.62 0.64
BH0 20.2 23.8 0.05 0.06 2.81 36.58 0.04 0.48
0.1 BHF 108.4 93.8 0.26 0.23 18.16 15.55 0.24 0.20
BHD 118.4 99.4 0.29 0.24 20.70 9.64 0.27 0.13
BH0 40.4 47.6 0.10 0.12 5.53 63.58 0.07 0.83
0.2 BHF 162.4 145 0.40 0.35 23.12 25.34 0.30 0.33
BHD 183.8 154.6 0.45 0.38 27.86 18.77 0.36 0.24
BH0 96.2 96 0.23 0.23 14.17 67.48 0.18 0.88
0.3 BHF 351.2 323.2 0.86 0.79 49.32 43.53 0.64 0.57
BHD 327.2 281.2 0.80 0.69 34.03 42.50 0.44 0.55
BH0 145.6 144.2 0.36 0.35 18.43 70.51 0.24 0.92
0.4 BHF 391.8 358.2 0.96 0.87 54.75 45.78 0.71 0.59
BHD 410 374.4 1.00 0.91 53.50 48.57 0.69 0.63
28 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

In general for normal strength concrete (NSC), the analysis of a section may
be considered linear, when the coecient of fc/f'c and fs/fy do not exceed the
values of 0.5 and 0.62 respectively. Although in literature, no suggestion was
founded for HSC (Maghsoudi & Akbarzadeh 2006) ¯exural beams. However,
here the same coecients are assumed for concrete and steel in HSC beams
strengthened with FRP. Based on this assumption, it is obvious that,
irrespective of the amount of ,  0 , fy and f 0c for all tested beams at mid-span
section, the coecient fc/f'c is equal to or less than 0.5 for permissible crack
widths of up to 0.2mm.

CONCLUSIONS
The external bonding of a CFRP plate o€ers an extremely e€ective means of
strengthening reinforced HSC beams in ¯exure. Flexural tests carried out in
this test program demonstrated an increase in sti€ness and ultimate strength
for reaction anchored plates. Examination of the crack distribution indicated
that the size and density of the cracks were signi®cantly less in the
strengthened beams than in the control specimens, thereby producing a more
durable system, which is less susceptible to the ingress of water and other
potentially corrosive solutions.
The e€ectiveness of the external plates in strengthening was not reduced as
the plate lengths were shortened. Each of the strengthened beams with
unanchored plates failed by plate peel-o€ when the strain gradient in the plate
within the shear span reached sensibly the same value. This is an important
result that requires further research to examine whether this same limit can be
applied universally.
The results of tests performed in this study indicate that signi®cant increase in
the ¯exural strength can be achieved by bonding CFRP plates to the tension
face of high strength reinforced concrete beams. The gain in the ultimate
¯exural strength was more signi®cant in beams with lower steel reinforcement
ratios than compared to a beam reinforced heavily with steel only. Beams
reinforced with both steel and CFRP have adequate deformation capacity, in
spite of their brittle mode of failure.
The extreme compressive strain of concrete ®bre in the strengthened
beams remains more or less linear up to the failure of the beam and is not
signi®cantly a€ected by concrete cracking or yielding of the tension steel.
These results demonstrate that the e€ect of the strengthening plate is to
reduce strain in the compression ®bers of the concrete. It was also found
that for all strengthened test beams, the tensile steels strains were always
higher than the CFRP strains.
 Reinforced HSC beams strengthened with CFRP plates under bending 29

The ®nite element model results show good agreement with observations and
data from the experimental full-scale beam tests. This numerical study can be
used to predict the behavior of reinforced concrete beam strengthened with FRP
more precisely by assigning appropriate material properties to develop design
rules for strengthening a RC member using FRP.

REFERENCES
ACI Committee 318-2005. Building code requirements for structural concrete (ACI 318-05) and
Commentary (ACI 318R-05). American Concrete Institute, Farmington Hills, MI.
Alagusundaramoorthy, P. & Harik I. & Choo, C. 2003. Flexural behavior of R/C beams
strengthened with carbon ®ber reinforced polymer sheets or fabric. Journal of Composites for
Construction 7(4): 292-301.
ANSYS, I., 2003. ANSYS Manual Set., ANSYS, Inc., Canonsburg, PA 15317, USA.
Arduini, M & Nanni, A. 1997. Behavior of pre-cracked RC beams strengthened with carbon FRP
sheets, Journal of Composites for Construction. ASCE 1(2): 63-70.
Ascione, L & Feo, L. 2000. Modeling of composite concrete interface of RC beams strengthened
with composite laminates. Composites Part B: Engineering 31: 535-40.
Bonacci, J. & Maalej, M. 2000. Externally bonded ®ber-reinforced polymer for rehabilitation of
corrosion damaged concrete beams. ACI Structural Journal 97(5): 703-711.
Bonacci, J. & Maalej, M. 2001. Behavioral trends of RC beams strengthened with externally
bonded FRP. Journal of Composite for Construction 5(2): 102-113.
Chahrour, A. & Soudki, K. 2005. Flexural response of reinforced concrete beams strengthened with
end-anchored partially bonded carbon ®ber-reinforced polymer strips. Journal of Composites
for Construction 9(2): 170-177.
Chansawat, K. & Yim, S. & Miller, T. 2006. Nonlinear ®nite element analysis of a FRP-
strengthened reinforced concrete bridge. Journal of Bridge Engineering 11(1): 21-32.
Fanning, P. & Kelly, O. 2001. Ultimate response of RC beams strengthened with CFRP plates.
Journal of Composites for Construction 5(2): 122-127.
Gangarao, H. & Vijay, P. 1998. Bending behavior of concrete beams wrapped with carbon fabric.
Journal of Structural Engineering, ASCE 124(1): 3-10.
Hashemi, S.H. & Rahgozar, R. & Maghsoudi, A. 2007. Finite element and experimental
serviceability analysis of HSC beams strengthened with FRP sheets. American Journal of
Applied Sciences, AJAS 4 (9): 725-735.
Maalej, M. & Bian, Y. 2001. Interfacial shear stress concentration in FRPstrengthened beams.
Composite Structures 54: 417-426.
Malek, A. & Saadatmanesh, H. & Ehsani, M. 1998. Prediction of failure load of R/C beams
strengthened with FRP plate due to stress concentration at the plate end. ACI Structural
Journal 95(2): 142-152.
Maghsoudi, A. & Akbarzadeh, H. 2006. Flexural ductility of HSC members. Structural Engineering
and Mechanics an International Journal 24(2): 195-213.
Nguyen, D. & Chan, T. & Cheong, H. 2001. Brittle failure and bond development length of CFRP-
concrete beams. Journal of Composites for Construction, ASCE 5(l): 12-17.
Oh, B. & Cho, J. & Park, D. 2003. Static and fatigue behavior of reinforced concrete beams
strengthened with steel plates for ¯exure. Journal of Structural Engineering 129(4): 527-535.
Rahimi, H. & Hutchinson, A. 2001. Concrete beams strengthened with externally bonded FRP
plates. Journal of Composites for Construction, ASCE 5(1): 44-56.
30 Seyed Hamid Hashemi, Ali Akbar Maghsoudi and Reza Rahgozar

Saadatmanesh, H. & Ehsani, M. 1991. RC beams strengthened with GFRP plates. I: Experimental
study. Journal of Structural Engineering 117(11): 3417-3433.
Sebastian, W. 2001. Signi®cance of mid-span debonding failure in FRP-plated concrete beams.
Journal of Structural Engineering 127(7): 792-8.
Schneider, S. 1998. Axially loaded concrete-®lled steel tubes. Journal of Structural Engineering
124(10): 1125-1138.
Teng, J. & Smith, S. & Yao, J. & Chen, J. 2003. Intermediate crack-induced debonding in RC
beams and slabs. Construction and Building Materials 17: 447-62.
Toutanji, H. & Zhao, L. & Anselm, E. 2006. Veri®cations of design equations of beams externally
strengthened with FRP composites. Journal of Composites for Construction 10(3): 254-264.
Uy, B. 1998. Local and post-local buckling of concrete ®lled steel welded box columns. Journal of
Constructional Steel Research 47: 47-72.

Submitted : 25/6/2007
Revised : 25/5/2008
Accepted : 3/6/2008
 Reinforced HSC beams strengthened with CFRP plates under bending 31

#FGOzyG MG(yz< ?"ZJA| I(tyG ?+y9f B"}S ?\ 9f N+SQ@ G Q

,"DyG I(s BJ@ f(<QwyG a9+yz< ?|(fO}y ?=vQ}y G G

QGR(sGQ 9\ Q h jO(Zt| Q=v C ,zf ,,}V9$ O+}I O+S

cGQB - cGQB ?g|9F - ?SO"%yG ?+zv

?+!GQ* G ?+|;S G ?*Q(%}GyG

?Y;L

h| upG(A* fC ) E y ?F9J< IQp(A| >+@GQAv , f9G@Q:G B+=DAyGh u=T}yG #+ ZJAyG

k
>+vQAzy {zJAyG d;L 9*QhQ\ f( w* 9}<Q >+vQAyG Q*(a@ ,h}AG}zy ?D*OJyG J 9=zaA}yG
L ;gy hC >+vQAyG , p Q++jAyG >=T< hC 9| KO 9I >=T< ?G@9"yG H(+ g yG hC {v5AyG >=T<
9 qL
b v v G fG (v$ ?vdvI;v}vyGh ?vgvF9vWvyG QG(v _}G #v| 9 =y
.A v"v v Gh y
,v h}G ~v+v}vZvAvyG H(v+v f
> = T< ($ ) FRP( f( <QwyG a + z 9 y < ?|(fO}yGh ?=vQ}yG #FG OzyG MG( z y < ?}fO}yG VQG( gzy
k
?*9%! fE . + 9 =T! {+zs {+}J@ O"f K OJ * j PyGh &!Gh} u<9TyGh hs(A}yG Q+i QWtAyG
d( zJy ?+y9f ?+y9}FE ?qzw@ r+]@ fC # w}* , AyG ?z*(ayG KF9qZyGh , SGQ}yG ?J+qY
. ?zwW}yG gP$ {J@ Os #+ ZJAyG

a + z9 y < ?|(fO}yG ?=vQ}yG #FG OzyG a + C9 y Q+Cz@ ?SGQO ($ EJ=yG GP$ # | a O % yG

a O%yG GP$ .9%J+zZ@ ~@ Ay
, Gh ?"ZJ}yG ?+A"}S G VQG( g yG AGOCh c( zS ) zf f( <QwyG
9%Z+ZM@ ~@ , AyG ?+"}S G VQG( gzy W 9t! h<Q}G [Jp AG QFE u*Q_ # f utJA*
~C # h| {ZqyG ?zIQ| ) AI NSQ}yG OWyG # | ?qzAM| J + 9 }v eGOMAS9< ?qzAM| MG( z y<
.A " 9 J!:G ?<9GASG H 9TI

?"ZJ}yG MG( ~ y y ,FG OA<:G YOZAyG ,jO 9f M( y h| ?!Q9t}y9< &!C #+ =@ U( JqyG ?G+A!
k
.I Q+ = v IQ( Z< {]q}yG {+}JAyG OGO R*h ?!hQ}yG {t@ 9}"+< ;+zs OGO *R

?y9I , p Q+Dw< {sC B!9v Yh OZyG gP$ ?p9Dvh ~GI fC Y OZAyG h*R(@ [Jp #+
q
< Oty
?+!9w|E #+ =@ 9"ASGQO fC d( tyG 9""w}* zfh
&+ . *O? 9gyG y
MG( }G , p 9%"f ?"ZJ}yG MG( }Gy
. {v5Azy ?==T}yG i QL } G OG( }yGh A9}yG d( LO< {sC ?FQO< QCzA@ ?"+A| MG ( C Ly 9A!E
View publication stats