View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by Universidade do Minho: RepositoriUM

FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

DESIGN MODELS FOR SHEAR STRENGTHENING OF REINFORCED

CONCRETE BEAMS WITH EXTERNALLY BONDED FRP
COMPOSITES: A STATISTICAL VS RELIABILITY APPROACH

J.L.T. LIMA 1 J.A.O BARROS 2

1 2
MSc Student; Associate Prof., Department of Civil Engineering, University of Minho, Portugal

Keywords: analytical, bonded, externally, FRP, models, shear, strengthening

1 INTRODUCTION

Experimental studies conducted worldwide on reinforced concrete (RC) beams strengthened in

shear with externally bonded (EBR) fiber-reinforced polymers (FRP) over the last years evince the
reliability and efficacy of such technique for structural retrofitting. For elements with shear resistance
deficiencies, an higher load carrying capacity may be achieved by bonding FRP reinforcement
systems with the fibers as orthogonal as practically possible to the critical shear crack plane for an
optimal configuration, or with the fibers normal to the beam axis for a more practical setting. Common
configurations of strengthening (Fig.1) include the full wrapping of the cross section (O), U jacketing
along 3 sides (U) and side bonding in the beam web (S). Additional mechanical anchorage systems
can be provided to enhance the behavior of U or S configurations where the available bond length is
short (U+ and S+).

Fig. 1: Common externally bonded FRP strengthening configurations

Each of the aforementioned strengthening configurations may be set in several possible

arrangements (Fig. 2) regarding the fiber orientation, the use of discrete strips or continuous sheets,
and the overlay of sheets with different fiber orientations, among others.

Fig. 2: Possible arrangements for externally bonded FRP strengthening

2 ANALYTICAL FORMULATIONS FOR FRP SHEAR REINFORCEMENT DESIGN

As an outcome of the increasing demand stimulated by a continuous growth in field applications,

several proposed analytical formulations [1-3] have been implemented into reference design
guidelines, providing the guidance for design, detailing, and installation of FRP based strengthening
systems. The present study addresses the shear provisions included in fib [4], ACI [5], CNR [6] and
the Australian Standard [7] design guidelines. The later follows an analytical model previously
introduced by Chen and Teng (CT) [8, 9]. All of the aforementioned design models rely on the
approach where shear strength of a strengthened member is attained by the sum of the contributions
from the reinforcing steel, Vs, and concrete, Vc, with the one deriving from the FRP ,Vf , as follows:

Vr = Vc + Vs + Vf (1)

1
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

where Vc and Vs, may be calculated according to provisions existing on current design codes,
independently of the FRP strengthening system adopted. The methodology to estimate the design
value of the FRP contribution in shear, Vfd, according to each of the aforementioned design proposals
is briefly described in Table 1. Figure 1 summarizes the adopted notation defining the geometric
properties of a generic beam reinforced in shear with externally bonded FRP.

Fig. 3: Adopted notation to define the main geometric properties of an FRP shear reinforcement

Table 1: Vfd calculation methodology (cont.)

fib design proposal: ACI design proposal:

⎛ w ⎞
Vfd = 0.9 ⋅ ε fed ⋅ Ef ⋅ ρf ⋅ bw ⋅ d ⋅ ( cot θ + cot β ) ⋅ sin β Vfd = φ ⋅ψ f ⋅ ⎜⎜ 2 ⋅ tf ⋅ f ⋅ ffe ⋅ (sin β + cos β ) ⋅ d f ⎟⎟
⎝ sf ⎠
2 ⋅ tf ⋅ w f 2 ⋅ tf ⋅ sin β
ρf = (strips) ; ρf = (cont.) φ =0.85 ; ψ f =0.95 (O) ; ψ f = 0.85 (U,S)
bw ⋅ sf bw

ffe = E f ⋅ε fe
0.8 ⋅ ε fe
ε fed = ; γ f =1.2 / 1.3 / 1.35
γf i) Full wrapping configuration (O):

i) Full wrapping configuration (O): ε fe = 0.004 ≤ 0.75 ⋅ ε fu

0.30
⎛
2
fcm 3 ⎞ ii) Side bonding or U jacketing configuration (U, S):
ε fe = 0.17 ⋅ ⎜⎜ ⎟⎟ ⋅ ε fu
⎝ Ef ⋅ ρf 1000 ⎠ ε fe = kv ⋅ ε fu ≤ 0.004
ii) Side bonding or U jacketing configuration (U, S): k1 ⋅ k 2 ⋅ Le
kv = ≤ 0.75
⎧ 0.56 11900 ⋅ ε fu
⎛ ⎞
2
fcm 3
⎪0.65 ⋅ ⎜ ⎟⎟ ⋅ 10
−3
23300
⎪⎪ ⎜ E ⋅ ρ 1000 ⎛f ⎞
23
Le =
⎝ f f ⎠ k1 = ⎜ ck ⎟ ;
ε fe = min ⎨ 0.30 ⎝ 27 ⎠ ( t f ⋅ Ef )
0.58

⎪ ⎛ ⎞
2
fcm 3
⎪ 0.17 ⋅ ⎜⎜ ⎟⎟ ⋅ ε fu df − Le d − 2 ⋅ Le
⎩⎪ ⎝ Ef ⋅ ρf 1000 ⎠ k2 = (U) ; k 2 = f (S)
df df

Notation:

ε fed - design value of effective FRP strain; φ - shear strength reduction factor [10];
ε fe - mean value of effective FRP strain; ψ f - additional reduction factor for FRP;
ε fu - FRP ultimate tensile strain; kv - bond reduction coefficient;
γ f - partial factor for FRP reinforcement; k1 - modif. factor regarding the concrete strength;
ρf - FRP reinforcement ratio; k 2 - modif. factor regarding the FRP configuration;
Ef - elasticity modulus of FRP reinforcement; Le - effective bond length of FRP reinforcement;
fcm - concrete average compressive strength; fck - concrete characteristic compressive strength;

2
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

Table 1 (cont.): Vfd calculation methodology

CNR design proposal: CIDAR (CT) design proposal:

i) Full Wrapping configuration (O) wf

Vfd = 2 ⋅ ffed ⋅ tf ⋅ ⋅ hfe ⋅ (cot θ + cot β ) ⋅ sin β
sf '
1 wf
Vfd = ⋅ 0.9 ⋅ d ⋅ ffed ⋅ 2 ⋅ tf ⋅ (cot θ + cot β ) ⋅ hfe = zb − zt ; zb = 0.9 ⋅ d − dfb ; zt = dft
γ Rd sf '
⎡ 1 Le ⋅ sin β ⎤ ffed = Df ⋅ ffd ,max
ffed = ffdd ⋅ ⎢1 − ⋅ ⎥+
⎢⎣ 6 min {0.9 ⋅ d , hw } ⎥⎦
i) Failure by FRP rupture (O)
1 ⎡ Le ⋅ sin β ⎤
+ (φR ⋅ ffd − ffdd ) ⋅ ⎢1 − ⎥
2 ⎢ min {0.9 ⋅ d , hw } ⎦⎥
⎣ ⎛ z ⎞
 Df = 0.5 ⋅ ⎜ 1 + t ⎟
≥0 ⎝ zb ⎠
rc rc ⎧ 1
φR = 0.2 + 1.6 ⋅ ; 0≤ ≤ 0.5 ⎪⎪ γ ⋅ φR ⋅ ffu , ε f ≤ 1.5%
bw bw ffd ,max =⎨ f
⎪ 1 ⋅ φR ⋅ Ef ⋅ ε f , ε f > 1.5%
⎪⎩ γ f
Ef ⋅ t f 0.80 2 ⋅ Ef ⋅ Gfk
Le = ; ffdd = ⋅ φR = 0.80 ; γ f = 1.25
2 ⋅ fctm γ fd tf

2 − w f sf ' ii) Failure by FRP debonding (U , S)

Gfk = 0.03 ⋅ k b ⋅ fck ⋅ fctm ; kb = ≥ 1
1 + bf 400
⎧ 2 1- cos ( π2 ⋅ λ )
ii) U jacket configuration (U) ⎪⎪ ⋅ , λ ≤1
Df = ⎨ π ⋅ λ sin ( 2 ⋅ λ )
π
; λ = Lmax Le
⎡ 1 Le ⋅ sin β ⎤ ⎪1- ⋅ π - 2 , λ >1
ffed = ffdd ⋅ ⎢1 − ⋅ ⎥ ⎩⎪ π ⋅ λ
⎣⎢ 3 min {0.9 ⋅ d , hw } ⎦⎥
⎧ hfe
⎪ sin β , (U) Ef ⋅ t f
iii) Side bonding configuration (S) ⎪
Lmax = ⎨ ; Le =
1 sin β w f ⎪ hfe , (S)
fck
Vfd = ⋅ min {0.9 ⋅ d , hw } ⋅ ffed ⋅ 2 ⋅ tf ⋅ ⋅ ⎪⎩ 2 ⋅ sin β
γ Rd sin θ sf '

zred ,eq ⎛ Leq ⎞

2
⎧ 1
⎪ ⋅φ ⋅ f
ffed = ffdd ⋅ ⋅ ⎜ 1 − 0.6 ⋅ ⎟ γ f R fu
min {0.9 ⋅ d , hw } ⎜⎝ zred ,eq ⎟ ⎪
⎠ ffd ,max = min ⎨
⎪ 1 ⋅ 0.35 ⋅ β ⋅ β ⋅ Ef ⋅ fck
zred ,eq = zred + Leq ⎪γ f L w
tf
⎩

zred = min {0.9 ⋅ d , hw } − Le ⋅ sin β

⎧λ , λ ≤1 2 − w f ( sf, ⋅ sin β )
suf βL = ⎨ ; βw =
Leq = ⋅ sin β λ >1 1 + w f sf, ⋅ sin β
ffdd / Ef ⎩1 ,
Notation:

γ Rd - partial factor for the resistance model (1.2); φR - reduction factor due to local stress in corners;
f - average concrete tensile strength; λ - normalized maximum bond length;
ctm

ffed - design value for the FRP effective stress; Df - stress distribution factor;

ffd - design value for the ultimate FRP stress; ffd ,max - maximum design stress in FRP;
ffdd - design value for the FRP debonding stress; ffu - ultimate FRP tensile stress;
Gfk - bonded joint specific fracture energy; hfe - effective height of the bonded reinforcement;

kb - covering / scale coefficient; β L - bond length coefficient;

suf - FRP slip at debonding (0.20mm);
βw - strip width coefficient;

3
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

3 PERFORMANCE APPRAISAL OF THE ANALYTICAL FORMULATIONS

3.1 Database assembly

To assess the accuracy of the theoretical predictions obtained with the aforementioned analytical
formulations, a database (DB) containing the experimental results from 212 beams strengthened with
EBR was collected from published literature, upgrading previous compiled databases [11-12]. The
criteria adopted in this task was to collect the largest amount of data with a wide spectrum of test
results regarding the beams geometry, concrete properties, longitudinal steel reinforcement ratios,
shear steel reinforcement ratios, FRP properties and strengthening configurations. Aiming to reduce
the influence of erroneous and inconsistent data present in the DB, the analysis was performed not
only in the integral database (IDB), but also in partial subsets of the data – reduced databases (RDB).
Detailed information on the databases characteristics and the considered beams may be found in [13].

3.2 Results obtained using the integral database (IDB)

For each described design model, the obtained values of Vfd are compared with Vf,exp and a χ
factor corresponding to the Vf,exp/Vfd ratio is evaluated. Figure 4 plots the predicted against
experimental values, where a 45º solid line establishes the division between the safe previsions from
the unconservative ones and a dashed line traces an “ideal safety trend” corresponding to χ=1.5.
300 300

χ=1 χ=1
250 250
ACI design model − Vfd (kN)
fib design model − Vfd (kN)

200 200
χ = 1.5 χ = 1.5

150 150

100 100

50 50

0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Vf,exp (kN) Vf,exp (kN)
300 300

χ=1 χ=1
250 250
CNR design model − Vfd (kN)

CT design model − Vfd (kN)

200 200
χ = 1.5 χ = 1.5

150 150

100 100

50 50

0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Vf,exp (kN) Vf,exp (kN)

Fig. 4 - Vf,exp vs Vfd scatterplots regarding fib, ACI, CNR and CT design models

A large scatter is observed in the experimental vs predicted design values for all of the considered
analytical formulations. Table 2 summarizes the main descriptive statistical measures regarding the χ
factor, namely minimum (MIN) and maximum (MAX) values, the average (AVG) that represents a
global safety factor associated with the design procedure, the standard deviation (STD) and the
coefficient of variation (COV) that are indicators of accuracy. The first quartile (Q1) that cuts off the
lowest 25% of data, the median (MED) corresponding to the 50th percentile and the third quartile (Q3)
that cuts off the highest 25% of data are also included.

4
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

Table 2: Statistical values of the χ factor computed from the IDB

χ Min Q1 MED AVG Q3 MAX STD COV

FIB 0.00 0.730 1.198 1.22 1.718 3.278 0.666 0.546
ACI 0.00 0.980 1.903 2.017 2.831 5.961 1.255 0.622
CNR 0.00 1.126 2.108 2.528 3.541 9.261 1.846 0.730
CT 0.00 0.875 1.370 1.431 1.962 5.454 0.826 0.577

The obtained results show that the fib design model presents, in average, the lowest safety factor
while the safest predictions are attained with CNR. The largest scattering is attained by the CNR
model (COV=0.73) while the least scattered model is fib (COV=0.55). The CT model globally presents
a good performance with an average value of χ =1.43 and COV = 0.58.
However, from the structural safety point of view, a classification system based only on the main
descriptive statistics measures regarding the behavior of the χ factor may not provide enough
information to assess the reliability of a design proposal, considering that for structural purposes
having χ=0.5 is worst than χ=2.0, which is not taken into account on the statistical analysis.
To overcome this limitation a weighed penalty classification system was applied, based on the
“Demerit Points Classification” (DPC) model proposed by [14], where a penalty (PEN) is assigned to
each range of χ ratios according to Table 3, and the total of penalties determines the performance of
each design model.

Table 3: Reliability analysis based on structural safety – IBD

χ Classification PEN FIB ACI CNR CT

< 0.75 Extr. Dangerous 10 55 32 28 45
0.75 - 1.00 Dangerous 5 30 22 15 23
1.00 - 1.25 Reduced Safety 2 26 18 16 30
1.25 - 1.75 Appropriate Safety 0 53 26 27 43
1.75 - 3.00 Conservative 1 47 65 58 65
> 3.00 Extr. Conservative 2 1 42 68 6
∑ PEN 801 615 581 702

From Table 3 it can be noticed that the fib design model presents the weakest performance, with
the highest number of penalty points corresponding to 40% of Predictions Against Safety (PAS, χ<1),
while the best results are attained by the CNR design proposal with the lowest of number of PAS
(20%). The CNR model also provides the highest number of extremely conservative values (32%).

3.3 Results obtained using the RDB

The high scattering found in the previous analysis performed over a DB with 212 beams with
highly differentiated characteristics, proves that none of the studied design models simulates with
enough accuracy the generic behavior of RC beams strengthened in shear with externally bonded
CFRP. It was also found that all the aforementioned design proposals provided a large amount of
unsafe values for Vfd, especially in the range 0 < Vf,exp< 100 kN. Such can be related with a significant
number of experimental results where, without a clear understanding, the load carrying increase due
to the FRP reinforcement is either null or extremely small, disturbing the global performance of the
considered analytical models. From the above considerations, the consistence of results obtained with
the IDB was appraised by means of removing from the analysis those observations, which in the
authors’ belief, lead into incoherent results. A reduced database (RDB) containing 130 beams
extracted from the IDB was assembled. A beam was removed from DB when fulfils one of the
following conditions: i) statistical outliers; ii) beams reinforced with bidirectional fibers; iii)
reinforcement systems with special anchorage mechanisms; iv) beams that show poor performance in
all of the aforementioned design models (χ < 0.25).
Figure 5 presents the obtained results with the RDB, providing for each design model a scatterplot
of the Vfd vs Vf,exp relationship, an histogram of the χ ratio distribution and a “box and whiskers” plot of
the χ ratio variation related with the reinforcement configuration. The box plot diagram graphically
depicts the statistical five-number summary, which consists of the smallest non-outlier observation,
lower quartile (Q1), median, upper quartile (Q3), and largest non-outlier observation.

5
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

300
0.8 6
χ=1
LRM - Linear Regression Model Normal Density
NLR - Non Linear Regression Model Est. Desnsity
250
5

0.6
fib design model − Vfd (kN)

200 4

Relative Frequency
χ = 1.5

χ fib
150 0.4 3

100 2

0.2

1
50

0.0 0
0
0 50 100 150 200 250 300 0 1 2 3 4 5 6 O S U
Vf,exp (kN) χ fib Reinf. Configuration
300
0.8 6
χ=1
LRM Normal Density
NLR Est. Desnsity
250
5

0.6
ACI design model − Vfd (kN)

200 4
Relative Frequency

χ = 1.5

χ ACI
150 0.4 3

100 2

0.2

1
50

0.0 0
0
0 50 100 150 200 250 300 0 1 2 3 4 5 6 O S U
Vf,exp (kN) χ ACI Reinf. Configuration
300
0.8 6
χ=1
LRM Normal Density
NLR Est. Desnsity
250
5

0.6
CNR design model − Vfd (kN)

200 4
Relative Frequency

χ = 1.5
χ CNR

150 0.4 3

100 2

0.2

1
50

0.0 0
0
0 50 100 150 200 250 300 0 1 2 3 4 5 6 O S U
Vf,exp (kN) χ CNR Reinf. Configuration
300
0.8 6
χ=1
LRM Normal Density
NLR Est. Desnsity
250
5

0.6
CT design model − Vfd (kN)

200 4
Relative Frequency

χ = 1.5
χ CT

150 0.4 3

100 2

0.2

1
50

0.0 0
0
0 50 100 150 200 250 300 0 1 2 3 4 5 6 O S U
Vf,exp (kN) χ CT Reinf. Configuration

Fig. 5 – Analysis results with the RDB (from top to bottom: fib, ACI, CNR and CT design models)

6
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

Table 4: Statistical values of the χ factor computed from the RDB

χ MIN Q1 MED AVG Q3 MAX STD COV ∑ PEN

FIB 0.119 0.889 1.389 1.387 1.76 3.278 0.592 0.427 354
ACI 0.596 1.310 2.141 2.221 2.88 5.463 1.066 0.48 233
CNR 0.411 1.734 2.503 2.886 3.69 8.931 1.653 0.573 219
CT 0.39 1.051 1.522 1.655 2.03 5.454 0.784 0.474 294

The values in Table 4 show that, despite the global improvement in the design models
performance with the RDB, the results follow the same trend as for the IDB analysis, thus ratifying the
consistency of the collected data.
Figure 6 plots the safe (PSS) vs unsafe (PAS) predictions diagrams for both the IDB (a) and the
RDB (b). From their analysis, it is mandatory to emphasize that all the studied design models show a
poor performance taking into account the large amount of unsafe predictions for the design value of
the FRP contribution in shear.

PAS PSS PAS PSS

1 1

0.75 0.75
0.60
0.74 0.68 0.70
0.80 0.78
0.87 0.94
0.5 0.5

0.25 0.25
0.40
0.26 0.32 0.30
0.20 0.22
0.13 0.06
0 0
FIB ACI CNR CT FIB ACI CNR CT
a) b)

Fig. 6 –PAS and PSS ratios with the: a) IDB, b) RDB

3.4 Influence of other parameters not explicitly considered in the analytical formulations
Such poor performance showed by the aforementioned analytical formulations indicates that the
relative influence of the considered parameters is simulated deficiently and the effect of others, not
explicitly taken into account, should not be neglected.
Figure 7 presents some relationships that are supposed to affect the performance of the analytical
models namely, the shear force gain ratio, Vf,exp ⁄ (Vr,exp−Vf,exp), the influence of the longitudinal
reinforcement percentage, ρsl, and the influence of shear steel reinforcement presence are
investigated.
All the studied analytical formulations seem to show an increase of the χ factor with beams where
the global shear force gain is higher. Such trend is observed for both discrete (DISC) and continuous
(CONT) reinforcement arrangements, being more obvious for the fib and CT models while a more
diffuse pattern is observed within the ACI and CNR models.
From the interaction between the χ factor and the longitudinal steel reinforcement it is found that χ
tends increase with the increment of the Esρs ⁄ Efρf ratio suggesting a major interaction between the
FRP and longitudinal reinforcement. On the plotted diagrams, beams with conventional shear
reinforcement are set aside from those without stirrups decoupling the interactions between these two
phenomena. It proved that both kinds of beams (with and without stirrups) follow the same trend
regarding the χ vs Esρs ⁄ Efρf relationship.
The plots of the χ vs Eswρsw ⁄ Efρf relation suggest that beams without stirrups may have higher χ
factors with the studied analytical models than those with higher shear reinforcement ratio.
Nevertheless it is the authors’ belief that the pattern found in the χ vs Eswρsw ⁄ Efρf scatter when Asw>0
may also be influenced by the longitudinal reinforcement ratio. In a future approach these phenomena
should be investigated, making several analysis of χ vs Eswρsw ⁄ Efρf for different clusters of ρsl.

7
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

6 6 6
CONT Asw=0
DISC Asw>0 LRM
LRM LRM
5 5 5

4 4 4
χ fib

χ fib

χ fib
3 3 3

2 2 2

1 1 1

0 0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100 120 140 0 2 4 6 8 10 12
Shear Force Gain Esl ρsl / Ef ρf Esw ρsw / Ef ρf
6 6 6
CONT Asw=0
DISC Asw>0 LRM
LRM LRM
5 5 5

4 4 4
χ ACI

χ ACI

χ ACI
3 3 3

2 2 2

1 1 1

0 0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100 120 140 0 2 4 6 8 10 12
Shear Force Gain Esl ρsl / Ef ρf Esw ρsw / Ef ρf
6 6 6
CONT Asw=0
DISC Asw>0 LRM
LRM LRM
5 5 5

4 4 4
χ CNR

χ CNR

χ CNR

3 3 3

2 2 2

1 1 1

0 0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100 120 140 0 2 4 6 8 10 12
Shear Force Gain Esl ρsl / Ef ρf Esw ρsw / Ef ρf
6 6 6
CONT Asw=0
DISC Asw>0 LRM
LRM LRM
5 5 5

4 4 4
χ CT

χ CT

χ CT

3 3 3

2 2 2

1 1 1

0 0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 20 40 60 80 100 120 140 0 2 4 6 8 10 12
Shear Force Gain Esl ρsl / Ef ρf Esw ρsw / Ef ρf

Fig. 7 – Influence of shear gain, longitudinal and shear reinforcement on the models performance

8
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

4 CONCLUSIONS

Based on the information available in the literature regarding the shear strengthening of RC beams
with externally bonded CFRP, a comprehensive database was assembled containing experimental
results of 212 beams. The results obtained from a statistical analysis carried out on such database
demonstrate that none of the analytical formulations predicts with enough accuracy the contribution of
the EBR CFRP systems for the shear strengthening of RC beams. A large scatter of the χ=Vf,exp/Vfd
was found within all the design models, even when a reduced database (RDB) was used in the
analysis.
Using the RDB the average of the χ factor varies between 1.4 (fib) and 2.9 (CNR) and the
coefficient of variation is comprehended between 43% (fib) and 57% (CNR). From a statistical point of
view the CT model can be pointed as the one with the best performance since it always combines an
appropriate global safety factor (AVG χ = 1.67) with the one of most least scattered behaviors (COV χ
= 47%). Although the large amount of calculated Vfd values that are against safety suggest that all of
the aforementioned models are still not robust enough for generalized practical design purposes.
A reliability analysis and classification based on structural safety was also implemented. Among
the studied formulations, the fib design model presented the most unsafe results from all, while the
safest results were attained with the CNR design code provisions. CNR also provided the largest
amount of extremely conservative predictions especially for the S type strengthening configuration.
The influence of some parameters not explicitly considered on the analytical models was
assessed, proving that the performance of the aforementioned design models is subordinated to the
global attained shear force gain. Furthermore, the influence of conventional steel reinforcement
(longitudinal and transversal) proved to be significant, and none of the studied analytical models
explicitly considers these parameters to determine the FRP contribution to shear. Other parameters
not taken into account in the analytical formulations may also influence the behavior of strengthened
beam. The collected database provides a significant source for data mining techniques in order to
decouple the interactions between all the phenomena involved. Thus, more investigation in this field is
needed in order to improve the existing design models.

ACKNOWLEDGEMENTS

The study reported in this paper forms a part of the research program “CUTINSHEAR -
Performance assessment of an innovative structural FRP strengthening technique using an integrated
system based on optical fiber sensors” supported by FCT, POCTI/ECM/59033/2004. The first author
acknowledges the support provided by the grant in the ambit of this research project.

REFERENCES

[1] Triantafillou, T. “Shear strengthening of reinforced concrete beams using epoxy bonded FRP
composites” ACI Structural Journal 11, 9 (March-April 1998), 107-115.
[2] Khalifa et al “Contribution of externally bonded FRP to shear capacity of RC flexural member”
Journal of Composites for Construction ASCE 2, 4 (1998), 195-202.
[3] Monti, G., and Liotta, M. “FRP-strengthening in shear: tests and design equations”. em 7th
International Symposium on Fiber Reinforced Polymer (FRP) Reinforcement for Concrete
Structures (FRP7RCS), 2005, ACI Symposium Publication 230.
[4] fib “Bulletin 14 - Externally Bonded FRP reinforcement for RC structures” Technical report,
2001, Task Group 9.3 FRP (fibre reinforced polymer) reinforcement for concrete structures.
[5] ACI “440.2R-02: Guide for the design and construction of externally bonded FRP systems for
strengthening of concrete structures”, 2002, Reported by ACI Committee 440.
[6] CNR-DT200 “Guidelines for design, execution and control of strengthening interventions by
means of fibre reinforced composites”, 2004, National Research Council.
[7] CIDAR - “Design guideline for RC structures retrofitted with FRP and metal plates: beams and
slabs” Draft 3 - submitted to Standards Australia, 2006, The University of Adelaide.
[8] Chen, J.F. e Teng, J.G. - “Shear Capacity of FRP Strengthened RC Beams: FRP Rupture”
Journal of Structural Engineering, ASCE 2003: 129(5): 615–625.

9
FRPRCS-8 University of Patras, Patras, Greece, July 16-18, 2007

[9] Chen, J.F. e Teng, J.G. - “Shear Capacity of FRP Strengthened RC Beams: FRP Debonding”
Construction and Building Materials 2003: 17(1): 27– 41.
[10] ACI 318 “318-02/318R-02: Building code requirements for structural concrete and
commentary”, 2002, Reported by ACI Committee 318.
[11] Bousselham, A., and Chaallal, O. “Shear strengthening reinforced concrete beams with fiber
reinforced polymer: assessment of influencing parameters and required research''. ACI
Structural Journal 110, 2 (March-April 2004), 219-227.
[12] Aprile, A., and Benedetti, A. “Coupled flexural-shear design of R/C beams strengthened with
FRP” Composites Part B: Engineering 35, 1 (January 2004), 1-25.
[13] Lima, J.L.T. “Assessment of the effective strain in FRP laminates used in shear strengthening
of reinforced concrete beams”, MSc thesis (in preparation) 2006, Univ. Minho, Portugal.
[14] Collins, M.P. - “Evaluation of shear design procedures for concrete structures”, 2001, A Report
prepared for the CSA technical committee on reinforced concrete design.

10