Behavior of Reinforced Concrete T-Beams Strengthened
Behavior of Reinforced Concrete T-Beams Strengthened
Behavior of Reinforced Concrete T-Beams Strengthened
This paper presents results of a wide and extensive experimental beams strengthened with externally bonded carbon fiber-
investigation on reinforced concrete (RC) T-beams retrofitted in reinforced polymer (CFRP) fabric. The parameters of the
shear with externally bonded carbon fiber-reinforced polymer study were set as follows: 1) the CFRP ratio (that is, the number
(CFRP). In total, 22 tests were performed on 4520 mm-long T-beams. of CFRP layers); 2) the internal shear steel reinforcement
The parameters investigated were as follows: 1) the CFRP ratio
ratio (that is, spacing); and 3) the shear length to the beam
(that is, the number of CFRP layers); 2) the internal shear steel
reinforcement ratio (that is, spacing); and 3) the shear length to depth ratio, a/d (that is, deep beam effect).
the beams depth ratio, a/d (that is, deep beam effect). The main The objectives of this paper are as follows:
objective of the study was to analyze the behavior of RC T-beams To investigate the shear performance, including the
strengthened in shear with externally applied CFRP by varying the mode of failure, of RC beams strengthened with CFRP
aforementioned parameters. The results showed that the contribution in terms of the CFRP; the internal transverse steel
of the CFRP to the shear resistance is not in proportion to the reinforcement, hereafter called the transverse steel; and
CFRP thickness (that is, the stiffness) provided, and depends on the shear span to depth ratios;
whether the strengthened beam is reinforced in shear with internal
transverse steel reinforcement. Results also confirmed the influence of To analyze the behaviour of the CFRP, the internal
the ratio a/d on the behavior of RC beams retrofitted in shear with transverse and longitudinal steel reinforcement, and the
external fiber-reinforced polymer (FRP). Finally, comparison of concrete struts, while the above parameters are varied; and
the shear resistance values predicted by ACI 440.2R-02, CSA To verify the reliability of ACI 440.2R-02 (ACI
S806-02, and fib TG9.3 guidelines, with the test results clearly Committee 440 2002), CSA S806-02 (Canadian Standards
indicated that the guidelines fail to capture important aspects, Association 2002), and fib TG9.3 (2001), hereafter called
such as the presence of the transverse steel and the ratio a/d on the the guidelines.
one hand, and overestimates the shear resistance for high FRP
thickness (and hence high FRP stiffness), on the other.
RESEARCH SIGNIFICANCE
Keywords: polymer; reinforced concrete; shear; strain; strengthening.
Most research studies on shear strengthening with FRP
composites are mainly focused on the properties and the
performance of the FRP and often involve rectangular beam
INTRODUCTION
test specimens of reduced sizes. Also, the lack of data on the
One of the techniques used to strengthen existing reinforced
strains experienced by the different components (FRP,
concrete (RC) members involves externally bonding fiber
concrete, and steel) makes it difficult, if not impossible, to
reinforced polymer (FRP) composite materials by means of
fully grasp the prevailing shear resistance mechanisms. The
epoxy adhesives. This technique improves the structural
proposed research was targeted to address these and other
performance of a member (Neale 2000; Meier 1995). The wide
important aspects. It is believed that the findings of this study
use of this strengthening method for various structures,
contribute to the understanding of the resistance mechanisms
including buildings and bridges, has demonstrated its efficiency
involved for RC beams strengthened in shear with externally
and its convenience (Bakis et al. 2002; Clarke 2000).
bonded FRP. This understanding is of paramount importance
Strengthening of beams and slabs in flexure and confinement
because it leads to a more rigorous approach toward safer
of circular columns have been well documented. A review
and rational design guidelines.
of research studies on shear strengthening, however,
revealed that experimental investigations are still needed
(Bousselham and Chaallal 2004; Matthys and Triantafillou EXPERIMENTAL PROGRAM
2001). Research studies carried out in recent years have The experimental program (Table 1) involves 22 tests
provided valuable findings, particularly with regard to the effect performed on 11 full-scale T-beams. The control specimens,
of the stiffness of the composite on the shear strength not strengthened with CFRP, are labelled 0L, whereas the
enhancement (Triantafillou and Antonopoulos 2000; Khalifa specimens retrofitted with CFRP are labelled 0.5L, 1L, or
and Nanni 2000). Other parameters that also influence the 2L, corresponding to 0.5, 1, and 2 bonded layers of CFRP,
shear resistance mechanism, however, were not sufficiently respectively. The letters DB (deep beam) and SB (slender
studied (Bousselham and Chaallal 2004). Shear steel
reinforcement and shear span to depth ratio (a/d) are examples ACI Structural Journal, V. 103, No. 3, May-June 2006.
of such parameters. MS No. 04-129 received April 12, 2005, and reviewed under Institute publication
policies. Copyright 2006, American Concrete Institute. All rights reserved, including
To address these areas, the authors conducted a large the making of copies unless permission is obtained from the copyright proprietors. Pertinent
discussion including authors closure, if any, will be published in the March-April
experimental investigation on the shear performance of RC 2007 ACI Structural Journal if the discussion is received by November 1, 2006.
specimen in the nomenclature (Fig. 2); and b) the other Fig. 3Instrumentation: (a) strain gauges on transverse
beam end zone is tested, but this time it is the end zone and longitudinal steel and embedded in concrete; and (b)
already tested that is overhung and unstressed. In this case, crack gauges on CFRP.
the load is applied at a distance a = 3d, from the nearest
support, which corresponds to a SB specimen in the nomen-
clature (Fig. 2). The sequence of loadingSpecimen DB control conditions at 2 mm/minute The signals from the
then SBwas enforced because the specimens and the setup gauges and the displacement sensors were captured and
were designed for that order. monitored using an automatic data acquisition system.
Instrumentation
ANALYSIS OF RESULTS
To meet the objective and the scope of the study, a very
comprehensive and carefully engineered measuring scheme The results related to the global behavior will be presented
was adopted for the project. in terms of: a) the load at rupture and the gain in capacity due
The vertical displacement was measured at the position to the CFRP; b) the load versus deflection relationship and
under the applied load and at the midspan using linear the gain in stiffness due to the CFRP; and c) the cracking
displacement sensors. The latter were also installed at each pattern and the failure modes observed. The strain data gathered
side of the supports perpendicular to the flange plan to will be used to study the transverse steel and the CFRP
control any undesired sway or tilt effects. Strain gauges were responses as the parameters (that is, the CFRP ratio, the
glued on transverse steel to measure stirrup deformations shear steel ratio, and a/d) are varied.
during the different loading stages and to monitor any
yielding (Fig. 3). The deformations experienced by the Overall response
CFRP wrap were measured using displacement sensors Table 5 presents the loads attained at failure; the experi-
known as crack gauges. These gauges were fixed vertically mental shear resistance due to concrete, due to the transverse
on the lateral faces of the specimens at the same positions steel, and due to the CFRP; as well as the shear capacity gain
along the longitudinal axis as the strain gauges on the due to the CFRP. Note that the values provided in Table 5
stirrups. Thus, the CFRP and the transverse steel responses were derived on the basis of the following assumptions
can be conveniently compared at the same positions during implicitly admitted in the guidelines: a) the shear resistance
the different stages of loading. due to concrete is the same whether the beam is retrofitted in
Likewise, strain gauges were also installed in parallel on shear with FRP or not and whether the retrofitted beam is
the longitudinal steel bars, on and in the concrete and on the reinforced with transverse steel or not; and b) the contribution
CFRP wrap at the tension zone of the specimens (Fig. 3). The of the transverse steel is the same for both retrofitted and
data monitored by these gauges will be a valuable tool and nonretrofitted beams.
will help explain the observed phenomena during the course The results show that the contribution of the CFRP to the
of the testing and hopefully give a better understanding of shear resistance is greater for the DB specimens with no
the resistance mechanisms. transverse steel (62% gain) than for the corresponding SB
specimens (50% gain). With the transverse steel, these gains
Testing and recording drastically decrease to reach 15% on average for the DB
The load was applied using a 500 kN capacity MTS specimens, whereas no gain is observed for the SB specimens.
hydraulic jack. All the tests were performed under displacement This clearly confirms the observations made in recent studies
Deflection response
Figure 4 and 5 show the curves representing the shear
force versus the midspan deflection for Series S0 and S1,
respectively. The figures feature two distinct sets of curves
Fig. 5Shear force versus midspan deflectionSeries S1. corresponding to deep beams (upper curves) and slender
beams (lower curves). The quasi-linear behavior of the
curves is characteristic of a shear failure. Compared to
(Chaallal et al. 2002; Pellegrino and Modena 2002; Li et al. slender specimens, the deep specimens featured a higher
2002) that the contribution of FRP to the shear resistance of a overall stiffness, but were more brittle (Fig. 5). Figure 5
beam with transverse steel differs from that of the same reveals no overall gain in stiffness due to the CFRP on specimens
beam but with no transverse steel. It is also observed that with transverse steel. The specimens with no transverse steel
doubling the thickness (for example, from 0.5L to 1L or from showed a very minor change in overall stiffness due to the
Failure mode
All the tested specimens failed in shear, except those of
Series SB-S2, which failed in flexure (refer to Table 5). No
specimen failed by debonding, delamination, or fracture of
the CFRP. The shear failure occurred by crushing of the
concrete struts. In the retrofitted specimens, it was evident
from the sudden appearance of a crack on the compression
table (flange). This crack progressed rapidly and announced
an imminent failure (Fig. 6). Note that in the specimens with
transverse steel, crushing of concrete occurred after the
transverse steel had yielded. Therefore, these specimens did
not fail by premature crushing of concrete. Failure by flexure
occurred by yielding of the longitudinal steel in the maximum
moment zone, followed by crushing of the concrete in the
compression zone at very large deformations (Fig. 6).
Cracking
In the deep specimens with no CFRP (DB-S0-0L, DB-S1-0L,
and DB-S2-0L), cracking occurred at a shear force of
approximately 80 kN. Specimen DB-S0-0L featured one
principal crack propagating at an average angle of 36 degrees Fig. 6Typical view of specimen at failure: (a) shear
from the support to the load point, which is typical of deep failure; and (b) flexural failure.
beam behavior. In specimens with internal steel stirrups
(DB-S1-0L and DB-S2-0L), in addition to the principal crack
and parallel to it, other finer cracks developed (Fig. 7(a)). The
ultimate load was attained as the principal crack extended
deeper into the compression zone.
In the slender specimens with no CFRP (SB-S0-0L,
SB-S1-0L, and SB-S2-0L), the cracking pattern depends on
the transverse steel spacing. Specimen SB-S0-0L featured a
principal crack that initiated at the support and progressed
rapidly towards the compression zone at an angle of
approximately 24 degrees (Fig. 7(d)). In Specimen SB-S1-0L
(refer to Fig. 7(e)), cracking was rather widespread and
propagated at a greater angle (with respect to longitudinal
axis) compared to Specimen SB-S0-0L, because the crack
angle increased from 24 to 38 degrees (Fig. 7(d)). In
Specimen SB-S2-0L, which failed in flexure, the first flexural
cracks appeared in the maximum moment zone at an applied
shear force of 65 kN. The first diagonal cracks appeared in
prolongation of flexural cracks at an applied shear force of
approximately 100 kN, then stabilized at approximately
275 kN. The flexural cracks continued to progress within the
maximum moment zone, until they reached the compression
zone. This was then followed by crushing of the concrete at
an applied shear force of approximately 295 kN.
In the retrofitted specimens with the U-shape continuous
wrap adopted, the crack propagation could not be monitored Fig. 7Crack patterns at failure.
during the course of the testing, except during the final phase
of loading where cracks suddenly appeared on the compression conditions, as well as the cracking pattern and extent, the
table. At this stage, the applied load had already attained CFRP wrap was carefully peeled off with some difficulty.
95% of its ultimate value. Examination of the specimens The following observations were made: 1) The concrete was
after tests revealed an expansion of the concrete evidenced completely pulverized. Confined by the CFRP wrap, the
by a bulge within the cracking zone. To examine the concrete concrete struts were subjected to stresses well beyond their
Strains analysis
This part of the study investigates the behavior of the
CFRP, the transverse steel, the longitudinal steel, and the
concrete struts, as the thickness of the CFRP varies. As
mentioned earlier, extensive instrumentation for strain
monitoring was carefully engineered to provide the information
and data much needed for the understanding of the shear
resistance mechanisms involved in beams retrofitted with
FRP. It must be realized that all the recorded data was
Fig. 8Shear force versus vertical CFRP strains in terms of subjected to careful examination, analysis, and comparisons.
number of layersdeep beams, Series S0 (crack gauge 1). For obvious reasons, however, it is not possible to report all
the findings in this paper. For more details, the reader is
referred to Bousselham (2005).
CFRP strainFigure 8 and 9 present the curves of the
shear force versus the strains in the CFRP wrap for deep
specimens with no transverse steel and for slender specimens
with transverse steel spaced at s = d/2, respectively. For
convenience, the locations of the strain gauges are provided
along the curves in the figures. It is observed that the curves
have the same tendency and feature three phases. In the
initial stage of loading, the CFRP does not contribute to the
load-carrying capacity. In the second stage, the CFRP begins
to strain at an applied shear force of approximately 105 kN
for deep specimens and 85 kN for slender specimens. The
CFRP strain continued to increase under increasing applied
Fig. 9Shear force versus vertical CFRP strains in terms of shear force up to a certain threshold, the level of which
number of layersslender beams, Series S1 (crack gauge 2). differs from one specimen to another depending on the
CFRP thickness (the thicker the CFRP, the lower this
compressive unconfined strength; 2) in deep specimens, one threshold level). In deep specimens for instance (Fig. 8), this
principal diagonal crack generally extended from the support level was 4720 microstrains in Specimen DB-S0-0.5L,
zone to the point load zone (Fig. 7(b) and (c)). In some cases, 2580 microstrains in Specimen DB-S0-1L, and 1900 micros-
however, in addition to this principal crack, a few fine diagonal trains in Specimen DB-S0-2L. In the third stage, the CFRP
cracks also developed. The crack angle was found to be strain started to decrease, drastically at times, as the shear
approximately 36 degrees, that is, unchanged by the addition force increased. This is shown by the reversing of the curves
of the CFRP; and 3) in slender specimens pertaining to in Fig. 8 and 9 and can be explained as follows. Although no
Series S0, only one principal crack was observed. As for sign of debonding was observed during the course of the test,
Specimen SB-S0-0L, it was relatively less inclined, with a the few popping noises heard here and there lead to believe
crack angle of approximately 22 degrees (Fig. 7(f)). In that local debonding could have occurred and may explain
contrast, the specimens of Series S1, such as SB-S1-0L, the CFRP strain decrease, which incidentally had no impact
showed rather widespread cracking, with an average crack on the applied loading, which in fact continued to increase.
angle of approximately 38 degrees. Again, it is seen that It may be argued that the increase of the applied load is due
neither the cracking pattern nor the crack angle were to the so-called redistribution of loadings between the
modified by the CFRP retrofit. transverse steel and the concrete struts. This was not the
The fact that in slender beams the cracking pattern is case, however, because no related change has been observed
influenced by the presence of the transverse steel and not by in the behavior of these two components.
that of the CFRP can be explained by the level of loading to Transverse steel reinforcement strainFigure 10 and 11
failure attained in the case where transverse steel is present present for the deep and slender specimens, respectively, the
(refer to Table 5). A higher level of loading translates into curves for the applied shear force versus the strains in the
more diagonal cracks prior to failure. In contrast, the transverse steel in terms of the CFRP thickness. These
specimens with no transverse steel experienced a lower curves indicate that the behavior of the transverse steel went
level of loading to failure, which occurred well before through three phases during loading. In the first initial phase,
diagonal cracks could proliferate within the test zone. In no noticeable contribution of the transverse steel to the
deep specimens, the presence of transverse steel and/or FRP resistance was observed. In the second phase, the first diagonal
did not alter the cracking pattern. This is attributed to the cracks initiated and the transverse steel started to strain. In
behavior of such deep specimens, where the transverse steel the deep specimens, for instance, this phase started at an
and/or the externally applied FRP contribute less to the shear average applied shear force of approximately 75 kN for the
resistance compared to slender specimens and may therefore control specimens, and 100 kN for the retrofitted specimens.
affect the cracking pattern. The transverse steel strain continued to increase with
Finally, it may of interest to note that no crack was increasing load until either the transverse steel yielded or
observed outside the tested zone of a = 1.5d during the first rupture of the specimen occurred. In the third stage, the
Fig. 11Shear force versus transverse steel strain in terms Fig. 13Shear force versus longitudinal steel strain in terms of
of number of layersslender beams (strain gauge 3). number of layersslender beams (under load point).