Archive of SID

Transaction A: Civil Engineering

Vol. 17, No. 6, pp. 471{481
c Sharif University of Technology, December 2010
Research Note

A Procedure for Predicting the Behavior of

FRP Con ned Concrete Using the FE Method
D. Mosto nejad1; and H. Saadatmand2
Abstract. The plastic behavior of concrete and enhancement of the ductility and compressive strength
of reinforced concrete members are indispensable consequences of concrete con nement. Fiber Reinforced
Polymers (FRP) are well known as con ning materials for circular columns. This study describes how
to predict the behavior of concrete con ned with Carbon Fiber Reinforced Polymers (CFRP) using a
non-linear analysis. The results of 29 experimental studies are used, where con ning composites have
been unidirectional CFRP. The stress-strain behavior of the concrete members from these experiments is
estimated as bilinear curves and, by extracting the necessary data, speci c expressions for modeling of
the nonlinear behavior of con ned concrete are presented. The presented relationships are veri ed using
the results of 16 distinct experiments. The relationships are applicable in the con nement modeling by
considering the hoop rupture strain of the CFRP attached to the concrete, and by using the Tsai-Wu
failure criterion. As part of the present study, concrete specimens con ned with a CFRP composite were
modeled with ANSYS software using the presented relationships. The results show the suitability of the
model selected, such that the stress-strain curves obtained from the software are properly applicable in the
parametric studies conducted on the con ned concrete subjected to axial load and exural moment.
Keywords: Con ned concrete; Column; Fiber Reinforced Polymer (FRP); Finite Element Analysis
(FEA).

INTRODUCTION now, Lam and Teng proposed a mathematical model

by taking into consideration the type of ber used
Over the past two decades, there has been increased in composite, but without considering the premature
usage of FRP composites to strengthen various struc- composite layer rupture attached to the concrete [4].
tural elements. The e ect of using composite sheets Furthermore, Pessiki et al. (2001) conducted a valuable
to con ne concrete columns is so valuable that it has parametric study on columns of reinforced concrete
attracted special e orts and research. Many studies con ned with composites [5]. In 2000, Mirmiran et
have been conducted to predict the stress-strain curve al. proposed a concept for the nite element modeling
and to develop a mathematical model to explain the of FRP tubes using ANSYS software, without consid-
con ned concrete behavior. Researchers, including ering the possibility of premature hoop rupture in the
Saa et al. [1], as well as Samaan et al. [2], con- composite [6].
ducted experiments on FRP tubes lled with concrete. More recent work on the subject includes the
Further, Xiao and Wu (2000) proposed models based study of Rousakis et al., in 2008, who presented
on the results of experiments conducted on cylinders an analytical modeling of plastic behavior for FRP
of concrete con ned in FRP attached to concrete [3]. con ned concrete members [7], as well as the study of
Examining the models and studies undertaken until Eid and Paultre, in 2007, on a plasticity-based model
for circular columns con ned with FRP sheets [8],
1. Department of Civil Engineering, Isfahan University of Tech- and an elastoplastic con nement model for circular
nology (IUT), Isfahan, P.O. Box 84156, Iran.
2. Department of Civil Engineering, Khoraskan Azad Univer- columns presented by Eid et al. in 2007 [9]. Fur-
sity, Isfahan, Iran. thermore, Cervenka and Papanikolaou, in 2007, pre-
*. Corresponding author. E-mail: dmosto @cc.iut.ac.ir sented a 3-D combined fracture-plastic material model
Received 30 January 2010; received in revised form 24 July 2010; for concrete, including an orthotropic smeared crack
accepted 2 October 2010 model for concrete cracking based on the Rankine

www.SID.ir
Archive of SID
472 D. Mosto nejad and H. Saadatmand

failure criterion and a plasticity model for concrete Regarding Tsai-Wu's rupture criterion, the in-
crushing [10]. Moreover, Karabinis et al., in 2008, teraction between biaxial stresses is studied when it
used a 3-D FE analysis to study the e ect of external applies to anisotropic materials. For this model, using
con nement by a FRP sheet jacket on concrete columns the mechanical properties of two major directions of
using an advanced Drucker{Prager type loading-failure the composite, a closed curve is de ned as the rupture
function [11]. A plastic-damage model was also used in criterion on the plane of biaxial stresses. The point of
the FE modeling of con ned concrete by Yu et al. in intersection of this curve with the compressive-tensile
2010 [12]. stress curve on biaxial stress coordinates indicates a
Accessibility to a reliable and rather simple - stress pair that results in FRP rupture. Tsai-Wu's
nite element modeling procedure on the behavior of rupture criterion comply with Equation 1 [14], where
con ned concrete is of great importance, in order 1 and 2 is the rupture compressive-tensile stress
to conduct the necessary studies for predicting the pair.
behavior of reinforced concrete members con ned in
FRP composite and subject to a variety of loading f1 1 + f2 2 + f11 12 + f22 22 + f12 1 2 = 1 k2 : (1)
combinations. On this basis, the present work has
drawn on the results of 45 experiments conducted Parameters fi , fj and fij (i; j = 1; 2) in Equa-
on columns wrapped in CFRP composite, so that tion 1 are obtained, taking into consideration the
by using the results of 29 studies, the relations for mechanical properties of the composite and using
prediction of the behavior of con ned concrete are Equations 2 to 6, where F1t and F2t are the ten-
obtained. These relationships are examined using the sile strengths, and F1c and F2c are the compressive
results of 16 other studies. These relationships are so strengths along and perpendicular to the composite
proposed as to be applicable in modeling by appropri- ber. Parameter k, varying from zero to one, as
ate software like ANSYS, taking into consideration the per Equation 7, is the ratio of the shear stress to
premature rupture of the bers wrapped around the the shear strength of the composite. In Figure 1,
column. the rupture curves for k = 0 and k = 1 for
a kind of composite from carbon ber, whose me-
chanical properties are determined by the producer,
RUPTURE CRITERIA FOR FRP have been plotted. In the present study, the curve
ATTACHED TO COLUMN relating to k = 0 has been used, as recommended
by [13].
In order to decide on the mechanical properties of
composite sheets, a at coupon tensile test, based f1 = 1=F1t 1=F1c ; (2)
on ASTM D 3039, is used. Based on laboratory
observations, the ultimate destruction of the con ned f11 = 1=F1t F1c ; (3)
column accompanies the FRP rupture. From review
of the experimental results provided in Table 1, it
can be concluded that the hoop strain rupture, "h , in
the CFRP composite attached to concrete, on average,
and with a standard deviation of 0.022, is 58% of the
tensile strain rupture resulting from the at coupon
composite test, "t . One reason for this premature
rupture is transfer of the axial compressive stress of the
concrete column to the composite, while the composite
is under hoop strain, resulting in the FRP rupture
under biaxial stresses [3]. On this basis, Fam and
Rizkalla proposed the Tsai-Wu rupture criteria for a
FRP composite as an anisotropic material subjected
to biaxial stresses [13]. Figure 1. Failure curves of Tsai-Wu's biaxial stresses.

Table 1. Comparison of the results obtained from the nite elements method with the observed results, and the
parameters used in the Drucker-Prager yield criterion for the test samples I and II.
Sample c   fcc fcc c Error "c "c c Error
(MPa) (MPa) (MPa) (%) (mm/mm) (mm/mm) (%)
I 9.09 33.317 13.327 49.4 49.5 +0.2 0.0124 0.0114 -8.06
II 10.64 38.17 0 96.5 90.53 -6.2 0.0174 0.0175 +0.57

www.SID.ir
Archive of SID
Predicting the Behavior of FRP Con ned Concrete Columns 473

f2 = 1=F2t 1=F2c ; (4) modulus of elasticity, Efrp , and composite thickness,

tfrp .
f22 = 1=F2t F2c ; (5)
k = (fcc fc0 )=fl ; (10)
0: 5
f12 = 1=2 (f11 f22 ) ; (6)
fl = 2tfrp Efrp "h =D: (11)
k = 6 =F6 : (7) The dilation angle must be in such a way that the
axial strain corresponding to the rupture hoop strain
FINITE ELEMENTS MODELING OF of FRP is obtained equal to the ultimate axial strain
CONCRETE CONFINEMENT corresponding to the ultimate compressive strength of
the concrete. When the dilation angle increases up to
To model the concrete and FRP, solid 65 and shell the internal friction angle, the ultimate axial strain is
41 elements, available in ANSYS software, are used. reduced.
Solid 65 is an eight-node element with the capability
of considering the nonlinear behavior of concrete, and
shell 41 is a membranous four-node linear element VERIFICATION OF THE NON-LINEAR
consistent with the linear behavior of FRP. Each MODELING OF CONFINED CONCRETE
node of both elements has three degrees of freedom. IN THE SOFTWARE
The modulus of elasticity of concrete, p taken as an To study the ANSYS performance using the above
isotropic material, is estimated at 3950 fc0 (MPa), relationships, the behaviors of two concrete specimens
and Poisson's coecient at 0.18, based on Ahmad and con ned in CFRP are modeled. The modeling is
Shah's model (1982) [2]. The mechanical properties conducted based on tests done by Xiao and Wu in 2000.
of FRP are de ned as an anisotropic material. Based The specimens marked as I and II are concrete cylin-
on the software recommendations for consideration ders with a uniaxial compressive strength of 33.7 and
of increased compressive strength and the ultimate 43.8 MPa, respectively, con ned in CFRP composite
axial strain of con ned concrete, Drucker-Prager's with a modulus of elasticity of 105 GPa and thicknesses
yield criterion is used as the solid 65 behavioral of 0.38 and 1.14 mm [3].
model [15]. Since the behavior of con ned concrete is modeled
by software using the Drucker-Prager yield criterion,
Drucker-Prager's Behavioral Model hence based on the above relations, in order to deter-
mine and set the parameters of the model, it is essential
This model is suitable and executable for granular to determine the hoop rupture strain of the composite,
materials such as soil, stone and concrete [15]. Drucker- "h , and the compressive stress corresponding to the
Prager's model is expressed by de ning three parame- strain (i.e. the compressive stress of the con ned
ters: cohesion, internal friction angle and the dilation concrete, fcc ) from the results of the experiment.
angle, referred to as input to the con ned concrete. Having these parameters, the coecient of cohesion
The rst parameter, i.e. the cohesion, c, is greater and the internal friction angle are determined using the
than zero, measured in units of force applied to the Drucker-Prager model. It is also possible to compute
surface area. The second parameter, i.e. the internal the dilation angle using the determined ultimate axial
friction angle, , is expressed in degrees, and the third is strain, "c . Having determined the parameters required
dilation angle, , which varies from zero to the value of for nite elements analysis, axial loading in the non-
the internal friction angle. To set the rst and second linear analysis is continued up to the point where
parameters, Mirmiran et al. proposed the following the strain in the respective element of the composite
equations for composite tubes lled with concrete [6]: reaches the laboratory rupture strain, "h , in the model.
In Table 1, fcc c , axial stress and strain corresponding
c = fc0 (1 sin )=(2 cos ); (8) to this stage of loading, fcc c and "c c , compared
p to the laboratory values, as well as the values for
 = 2(tan 1 k 45 ): (9) the parameters of the Drucker-Prager yield criterion,
are speci ed. The negligible error values of the non-
In Equations 8 and 9, if the ultimate compressive stress linear modeling indicate the credibility of the non-
corresponding to the con ning stress fl , equals fcc , the linear modeling of the concrete cylinders con ned in
con ning coecient, k, is obtained using Equation 10. composite sheets.
The value of con ning stress corresponding to the In Figures 2a and 2b, the axial stress curves versus
rupture hoop strain of the composite ("h ) is obtained axial strain and lateral strain are plotted for speci-
from Equation 11; given the diameter of the sample, D, mens I and II, respectively. The curves indicate the

www.SID.ir
Archive of SID
474 D. Mosto nejad and H. Saadatmand

RELATIONSHIPS PROPOSED TO
COMPUTE THE COMPRESSIVE
STRENGTH AND THE ULTIMATE AXIAL
STRAIN OF CONFINED CONCRETE
In order to work out the relationships required, 29
test results conducted by Watanable, Harris and Xiao
on carbon ber samples were used [3,4]. In these
tests, the mechanical properties of the composite were
determined using the at coupon tensile test. The
test samples include cylinders with a diameter of D =
100 and 152 mm, the compressive strength of the
uncon ned concrete varied from fc0 = 26.2 to 55.2 MPa,
the FRP thickness varied from tfrp = 0.14 to 02 mm,
the composite sti ness of Efrp = 38.1 to 630 GPa, and
the ultimate tensile strength of the at sample varied
from ffrp = 580 to 2873 MPa. The speci cations of
the test samples are listed in Table 2 where fcc is the
compressive strength of the con ned concrete and L is
is the length of the concrete cylinders. Furthermore,
fo is the intersection point of the second linear part of
the stress-strain graph with the stress axis based on the
values given in [4].
The stress-strain curve for the concrete con ned
with FRP composites is estimated in accordance with
the overall model in Figure 3 in a bilinear curve. Using
this estimation, the intercept of the second linear part
of the graph with the stress axis is de ned as fo , as
taken in [2]. In Figure 3, E1 and E2 are the slopes
of the second linear portions of the stress-strain graph
and, respectively, correspond to the curves for lateral
and axial strain. In this study, the curve for the bilinear
stress-strain is estimated so that an equal value is
obtained from an intercept of the second linear portions
of both graphs of axial stress/lateral strain and axial
Figure 2. Axial stress versus axial strain and lateral stress/axial strain with the stress axis (point fo ). This
strain in terms of changes in dilation angle for a) specimen estimation leads to the following equation:
I, and b) specimen II.
E2 =E1 = "h ="c : (12)
comparison of the observed results with the results of
the non-linear analysis of the specimens. In Figures 2a Based on the aforementioned discussions for con ne-
and 2b, the curves are plotted for each sample, along ment modeling, it is necessary to determine the ulti-
with the values of the internal friction angle and the mate compressive strength, fcc , corresponding to the
concrete cohesion coecient given in Table 1. The hoop rupture strain of FRP. Here, based on Figure 3,
curves show that an increase in the dilation angle, from fcc can be de ned in accordance with the following
zero to a coecient of , decreases the corresponding
axial strain, "h .
Based on Equations 8 and 11, it is evident that
in order to set the parameters of the Drucker-Prager
model, parameters fcc and "c corresponding to the "h
of the FRP composite must be given. Where the test
results of the con ned concrete are not available, such
values must be predicted using certain relationships.
For this purpose, suitable equations are expanded and
presented in the following section using the results Figure 3. Bi-linear stress-strain curve of FRP con ned
listed in Table 1. concrete.

www.SID.ir
Archive of SID
Predicting the Behavior of FRP Con ned Concrete Columns 475

Table 2. Characteristics of the samples used in presenting the proposed relationships [3,4] (continued).
Source of Data Fiber D L fc0 tfrp ffrp Efrp "c "h fo fcc
Type (mm) (mm) (MPa) (mm) (MPa) ( MPa) (MPa) (MPa)
Watanable et al. Carbon 100 200 30.2 0.17 2716 224600 0.0151 0.01 32 46.6
(1997) [16]
Watanable et al. Carbon 100 200 30.2 0.5 2873 224600 0.0311 0.0088 35 87.2
(1997) [16]
Watanable et al. Carbon 100 200 30.2 0.67 2658 224600 0.0415 0.0066 35 104.6
(1997) [16]
Watanable et al. HM 100 200 30.2 0.14 1579 628600 0.0058 0.002 30 41.7
(1997) [16] Carbon
Watanable et al. HM 100 200 30.2 0.28 1824 629600 0.0088 0.002 36 56
(1997) [16] Carbon
Watanable et al. HM 100 200 30.2 0.42 1285 576600 0.013 0.002 40 63.3
(1997) [16] Carbon
Harries et al. Carbon 152 610 26.2 1 580 38100 0.0144 0.0089 33.9 50.6
(1998) [17]
Harries et al. Carbon 152 610 26.2 2 580 38100 0.0175 0.0106 33.9 64
(1998) [17]
Xiao & Wu Carbon 152 305 33.7 0.38 1577 105000 0.012 0.0084 31.2 47.9
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 0.38 1577 105000 0.014 0.0115 31.2 49.7
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 0.38 1577 105000 0.0124 0.0087 31.2 49.4
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 0.76 1577 105000 0.0165 0.0091 36 64.6
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 0.76 1577 105000 0.0225 0.01 36 75.2
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 0.76 1577 105000 0.0216 0.009 36 71.8
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 1.14 1577 105000 0.0245 0.0086 38.4 82.9
(2000) [3]
Xiao & Wu Carbon 152 305 33.7 1.14 1577 105000 0.0303 0.009 38.4 95.4
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 0.38 1577 105000 0.0039 0.0035 50.4 52.1
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 0.76 1577 105000 0.0157 0.0092 50.4 84
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 0.76 1577 105000 0.0137 0.01 50.4 79.2
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 0.76 1577 105000 0.0166 0.0101 50.4 85
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 1.14 1577 105000 0.0174 0.0079 50.4 96.5
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 1.14 1577 105000 0.0168 0.0071 50.4 92.6
(2000) [3]
Xiao & Wu Carbon 152 305 43.8 1.14 1577 105000 0.0175 0.0084 50.4 94
(2000) [3]

www.SID.ir
Archive of SID
476 D. Mosto nejad and H. Saadatmand

Table 2. Continued.
Source of Data Fiber D L fc0 tfrp ffrp Efrp "c "h fo fcc
Type (mm) (mm) (MPa) (mm) (MPa) (MPa) (MPa) (MPa)
Xiao & Wu Carbon 152 305 55.2 0.76 1577 105000 0.0057 0.0039 68.6 74.6
(2000) [3]
Xiao & Wu Carbon 152 305 55.2 0.76 1577 105000 0.0081 0.0047 68.6 77.6
(2000) [3]
Xiao & Wu Carbon 152 305 55.2 0.76 1577 105000 0.0138 0.0083 68.6 77
(2000) [3]
Xiao & Wu Carbon 152 305 55.2 1.14 1577 105000 0.0143 0.0076 61.2 106.5
(2000) [3]
Xiao & Wu Carbon 152 305 55.2 1.14 1577 105000 0.0145 0.0085 61.2 108
(2000) [3]
Xiao & Wu Carbon 152 305 55.2 1.14 1577 105000 0.0118 0.007 61.2 103.3
(2000) [3]

equation:
fcc = E1  "h + fo : (13)
To provide relationships to determine E1 and fo in
order to compute fcc from Equation 13, the tests results
presented in Table 2 are used. Equations 14 and 15 are
obtained from regression on the test results and where
R2 is the coecient of correlation, the regression curve
for E1 is shown in Figure 4a.
fo = 5:1 + 1:239f 0 + 0:0708(Efrp tfrp =Df 0 );
c c

R2 = 0:93; (14)
E1 = 4189:7 ln(Efrp tfrp =Dfc0 ) 7228:4;
R2 = 0:85: (15)
Based on the test results given in Table 2, the ratio
de ned in Equation 12 is directly proportional to
the uniaxial compressive strength of the concrete and
diameter of the samples and inversely proportional to
modulus of elasticity, FRP thickness and the ultimate
tensile strain of the at coupon of FRP, "t . By applying
regression to the test results, using a linear function, as
shown in Figure 4b, Equation 16 is proposed for com-
putation of the ultimate axial strain, "c , corresponding
to the rupture hoop strain, "h . Note that parameter x
in Figure 4b is the fraction used in the right hand side
of Equation 16.
0
"h fc0:7 D0:5
= 7:1925 + 0:0288;
"c (Efrp tfrp )0:7 "0t :04
R2 = 0:87: (16)
Figure 4. Regression curve on the test data to determine;
To evaluate the validity of the proposed relationships, a) slope of E1 ; b) ratio of the ultimate lateral strain to the
16 test results, where cylindrical concrete samples were ultimate axial strain.

www.SID.ir
Archive of SID
Predicting the Behavior of FRP Con ned Concrete Columns 477

con ned in composite sheets of carbon bers, were PREDICTION OF THE BEHAVIOR OF
used [4,5,18]. The mechanical properties of FRP and CONCRETE CYLINDERS CONFINED IN
the concrete used in the samples, along with their test CFRP USING THE FINITE ELEMENT
results, are listed in Table 3. METHOD
In Figure 5a, having computed the values of fo
and E1 from Equations 14 and 15, and then substituted Based on discussions of previous sections, the following
into Equation 13, the values of the predicted ultimate steps are taken for modeling of the con nement of
strength from this equation versus the observed values concrete cylinders with FRP composites in the case of
are shown with an error range of 20%. Furthermore, a lack of test results:
the ultimate axial strains corresponding to the ultimate 1. As an initial estimation of "h for FRP wrapping,
hoop strains computed according to Equation 16, the hoop rupture strain is taken equal to 58% of
versus the observed values, are shown with an error the rupture strain obtained from the at coupon
range of 20% in Figure 5b. tensile test.

Table 3. Characteristics of the samples used to evaluate the proposed relationships [4,5,18].
Source of Data Fiber D L fc0 tfrp ffrp Efrp "c "h fcc
Type (mm) (mm) (MPa) (mm) (MPa) (MPa) (mm/mm) (mm/mm) (MPa)
Pessiki et al. Carbon 152 610 26.22 1 580 38700 0.0144 0.009 50.6
(2001) [5]
Pessiki et al. Carbon 152 610 26.23 2 580 38700 0.0165 0.0072 64
(2001) [5]
Harmon & Slattery Carbon 51 102 41 0.09 3500 235000 0.0117 0.0116 86
(1992) [19]
Harmon & Slattery Carbon 51 102 41 0.18 3500 235000 0.0165 0.0106 117
(1992) [19]
Harmon & Slattery Carbon 51 102 41 0.34 3500 235000 0.0257 0.0081 158
(1992) [19]
Harmon & Slattery Carbon 51 102 41 0.69 3500 235000 0.0365 0.0032 241
(1992) [19]
Toutanji Carbon 76 305 31.8 0.22 3485 228000 0.0179 0.0075 98.7
(1999) [20]
Toutanji Carbon 76 305 31.8 0.33 2940 373000 0.016 0.0043 96
(1999) [20]
Shahawy et al. Carbon 152 305 19.4 0.5 2275 82700 0.0159 0.0075 33.8
(2000) [18]
Shahawy et al. Carbon 152 305 19.4 1 2275 82700 0.0221 0.0061 46.4
(2000) [18]
Shahawy et al. Carbon 152 305 19.4 1.5 2275 82700 0.0258 0.006 62.6
(2000) [18]
Shahawy et al. Carbon 152 305 19.4 2 2275 82700 0.0356 0.0062 75.7
(2000) [18]
Shahawy et al. Carbon 152 305 19.4 2.5 2275 82700 0.0342 0.0062 80.2
(2000) [18]
Shahawy et al. Carbon 152 305 49 0.5 2275 82700 0.0062 0.0063 59.1
(2000) [18]
Shahawy et al. Carbon 152 305 49 1 2275 82700 0.0097 0.0063 76.5
(2000) [18]
Shahawy et al. Carbon 152 305 49 1.5 2275 82700 0.0126 0.0064 98.8
(2000) [18]
Shahawy et al. Carbon 152 305 49 2 2275 82700 0.019 0.0064 112.7
(2000) [18]

www.SID.ir
Archive of SID
478 D. Mosto nejad and H. Saadatmand

in order to estimate the dilation angle of the

con ned concrete, taken as the third parameter of
the Drucker-Prager yield criterion.
7. Having gone through the above steps by modeling
with appropriate software like ANSYS, the non-
linear analysis of con ned concrete is performed
with monotonic loading. The loading is continued,
at least up to the point where the hoop strain
considered in the rst step is reached, and the
compressive strength and the ultimate axial strain
of the con ned concrete specimen is determined.
Accordingly, by computing the increased trend of
compressive stress versus axial and hoop strains,
the curves representing the axial and lateral stress-
strain relationships are plotted as the results of the
modeling.
8. Since the CFRP composite attached to the column
is exposed to biaxial stresses, axial compression
and hoop tension, a more precise value of the
hoop rupture strain, is obtained based on Tsai-Wu's
failure criterion.
9. The hoop rupture strain of FRP composite, "h ,
obtained from step 8, is compared with the es-
timated value in the rst step, and in the event
of an unacceptable di erence the above steps are
repeated based on the new hoop strain, to achieve
an acceptable convergence.
Note that upon the aforementioned convergence,
a model of the con ned concrete is obtained using
the nite element method. This curve can be used
when the test results for the con ned concrete column
are not available. This curve enables us to conduct
preliminary studies on con ned concrete, in order to
Figure 5. a) Computed compressive strengths using study the behavior of the con ned concrete column
Equations 13 through 15 versus the real values. b) under loading combinations. By computing the error
Computed ultimate axial strains using Equation 16 versus of the compressive strength and hoop, as well as axial
the real values. strains, the validity of the nite element modeling can
be determined.
2. Based on Equation 11, the lateral con ning pressure
(fl ) corresponding to "h is computed.
MODELING TWO SAMPLE CONFINED
3. The values fo and E1 are determined using Equa- CONCRETE CYLINDERS
tions 14 and 15, then, the ultimate compressive
stress of the con ned concrete is obtained from In order to state the results of modeling based on the
Equation 13. above steps, two assumed concrete cylinders A and B
4. The con nement coecient, k, is calculated from con ned in CFRP, 0.8 and 0.6 mm in thickness, are
Equation 10. modeled. The diameter and height of specimens A and
B are considered as 152 and 305 mm, respectively, and
5. Using Equations 8 and 9, and considering the the compressive strength of the concrete is assumed
con nement coecient, k, the cohesion coecient, to be 32 MPa. The ultimate tensile and compressive
c, and the internal friction angle of the con ned strengths and modulus of elasticity of CFRP, based on
concrete, , are determined as the rst and second the data provided in [14], are considered, respectively,
Drucker-Prager yield criterion. as 2,860, 1,875 and 177,000 MPa along the bers,
6. Using Equation 16, the ultimate axial strain cor- and 49, 246 and 10,800 MPa perpendicular to the
responding to the hoop strain ("h ) is obtained, bers.

www.SID.ir
Archive of SID
Predicting the Behavior of FRP Con ned Concrete Columns 479

Having gone through the aforementioned 10 steps, Table 4. Comparison of the results obtained from the
the cohesion, internal friction angle and dilation angle nite elements model with the values computed from the
of the con ned concrete in the Drucker-Prager model proposed relationships.
for sample A are determined as 8.012 MPa, 36.8 degrees Sample A B
and zero degrees, respectively; and, for sample B as
7.63 MPa, 38.97 degrees and zero degrees, respec- Parameter Value Error Value Error
tively. Based on the selected mechanical properties for (%) (%)
the FRP composite, the failure surface in the Tsai- fcc p (MPa) 95.25 -6.2 87.2 6.85
Wu rupture criterion is as shown in Figure 6a. In fcc (MPa) 89.32 80.35
c
this gure, negative and positive signs are used for
compressive and tensile stresses, respectively. Further- "h p (mm/mm) 0.0085 +1.2 0.009 -1.1
more, the axial compressive stresses versus hoop tensile
stresses during monotonic loading for the composite "h c (mm/mm) 0.0086 0.0089
element are plotted in the gure. The tensile stress "c p (mm/mm) 0.026 -26.0 0.023 -17.4
corresponding to the rupture hoop strain of FRP is
"c c (mm/mm) 0.0192 0.019

determined by measuring and plotting the compressive

and tensile stresses of the composite element. Based
on Tsai-Wu's rupture criterion, the lateral rupture
hoop strain for samples A and B are calculated as
0.85% and 0.90%. The stress-strain curves obtained
from the modeling of each sample are plotted in
Figure 6b.
In Table 4, the predicted values of fcc , "h and "c
for samples A and B, obtained through the aforemen-
tioned steps and proposed expressions, are indicated
using subscript \p". Furthermore, the calculated
values of the same variables by setting the parameters
needed in the nite element modeling in ANSYS are
represented by subscript \c" in Table 4. The table
shows good agreement between the results, except for
"c . The error obtained for the ultimate axial strain
is due to the fact that in certain cases the nite
element model obtained by the above method predicts
smaller values for axial strain, compared to actual
values as observed in the modeling of certain test
results.

CONCLUSIONS
In this study, certain relationships for predicting the
ultimate compressive strength and the ultimate axial
strain corresponding to a given ultimate hoop strain
in concrete columns con ned with FRP sheets were
proposed for use in the nite elements modeling. The
relationships were evaluated using certain available test
results. Furthermore, the required steps for modeling a
non-linear analysis of con nement, using the proposed
relationships, were referred to and, based on these
Figure 6. a) Intercept of the axial compressive stress and proposed relationships, two concrete cylinder samples
the hoop tensile stress curves for the critical element in were modeled. The results of the study are as follows:
CFRP composite with Tsai-Wu's failure curve. b)
Stress-strain curves for samples A and B obtained from 1. The relationships proposed to determine the be-
nite elements modeling. havior of con ned concrete using nite elements,

www.SID.ir
Archive of SID
480 D. Mosto nejad and H. Saadatmand

in order to model that concretes con ned in FRP 9. Eid, R., Dancygier, A.N. and Paultre, P. \Elastoplastic
are suitable and executable. con nement model for circular concrete columns", J.
of Struc. Eng., 133(12), pp. 1821-1831 (2007).
2. The stress-strain curve obtained using a non-linear 10. Cervenka, J., Papanikolaou, V.K. \Three dimensional
analysis explicitly illustrates the behavior of the combined fracture-plastic material model for con-
con ned concrete using the software. By using this crete", International Journal of Plasticity, 24(12), pp.
model, a variety of studies can be done on the 2192-2220 (2008).
behavior of con ned reinforced concrete by means 11. Karabinis, A.I., Rousakis, T.C. and Manolitsi, G.E.
of the available software. \3D nite-element analysis of substandard RC columns
3. Drucker-Prager's material behavior model can be strengthened by ber-reinforced polymer sheets", J. of
used in the modeling of concrete columns con ned Comp. for Constr., 12(5), pp. 531-540 (2008).
in FRP composites, taking into consideration the 12. Yu, T., Teng, J.G., Wong, Y.L. and Dong, S.L. \Fi-
premature rupture of the composite. For this nite element modeling of con ned concrete-II: Plastic-
purpose, the ultimate compressive strength of the damage model", Eng. Struc., 32(3), pp. 680-691
concrete corresponding to the rupture hoop strain (2010).
of the FRP composite must be computed. 13. Fam, A.Z. and Rizkalla, S.H. \Con nement model
for axially loaded concrete con ned by circular ber-
4. Using Tsai-Wu's failure criterion, the biaxial tensile reinforced polymer tubes", ACI Struc. J., 98(4), pp.
and compressive stresses causing premature hoop 451-461 (2001).
rupture in a FRP composite can well be determined. 14. Daniel, I.M. and Ishai, O., Engineering Mechanics of
Composite Materials, Oxford University Press, Inc.,
New York (1994).
REFERENCES 15. ANSYS User's Manual 5.4, ANSYS, Inc., Canonsburg,
1. Saa , M., Toutanji, H.A. and Li, Z.J. \Behavior Pennsylvania (1997).
of concrete columns con ned with ber reinforced 16. Watanable, K., Nakamura, H., Honda, T., Toyoshima,
polymer tubes", ACI Mater. J., 96(4), pp. 500-509 M., Iso, M., Fujimaki, T., Kaneto, M., and Shirai,
(1999). N. \Con nement e ect of FRP sheet on strength
and ductility of concrete cylinders under uniaxial
2. Samaan, M., Mirmiran, A. and Shahawy, M. \Model compression", Non-Metallic (FRP) Reinforcement for
of concrete con ned by ber composites", J. of Struct. Concrete Structures, Proc., 3rd Int. Symposium, Japan
Eng., 124(9), pp. 1025-1031 (1998). Concrete Institute, Sapporo, Japan, 1, pp. 233-240
3. Xiao, Y. and Wu, H. \Compressive behavior of con- (1997).
crete con ned by carbon ber composites jackets", J. 17. Harries, K.A., Kestner, J., Pessiki, S., Sause, R.
of Mater. in Civil Eng., 12(2), pp. 139-146 (2000). and Ricles, J. \Axial behavior of reinforced concrete
columns retro t with FRPC jackets", Proc., 2nd Int.
4. Lam, L. and Teng, J.G., Stress-Strain Models for FRP Conf. on Composites in Infrastructures (ICCI), H.
Con ned Concrete, Research Center for Advanced Saadatmanes and M.R. Ehsani, Eds., Tucson, Arizona,
Technology in Structural Engineering, Hong Kong pp. 411-425 (1998).
Polytechnic University (2002).
18. Shahawy, M., Mirmiran, A. and Beitelman, T. \Test
5. Pessiki, S., Harries, K.A., Kestner, J.T., Suse, R. and and modeling of carbon-wrapped concrete columns",
Ricles, J.M. \Axial behavior of reinforced concrete Composites; Part B: Eng., 31(6-7), pp. 471-480 (2000).
columns con ned with FRP jackets", J. of Compos. 19. Harmon, T.G. and Slattery, K.T. \Advanced compos-
for Const., 5(4), pp. 237-245 (2001). ite con nement of concrete", Proc., 1st Int. Conf. on
6. Mirmiran, A., Zagers, K. and Yuan, W. \Nonlinear Advanced Composite Materials in Bridges and Struc-
nite element modeling of concrete con ned by ber tures, K.W. Neale and P. Labossiere, Eds., Canadian
composites", Finite Element in Analysis and Design, Society for Civil Engineering, Sherbrooke, Canada, pp.
35(1), pp. 79-96 (2000). 299-306 (1992).
20. Toutanji, H.A. \Stress-strain characteristics of con-
7. Rousakis, T.C., Karabinis, A.I., Kiousis, P.D. and crete columns externally con ned with advanced ber
Tepfers, R. \Analytical modelling of plastic behaviour composite sheets", ACI Mater. J., 96(3), pp. 397-404
of uniformly FRP con ned concrete members", Com- (1999).
posites; Part B: Eng., 39(7-8), pp. 1104-1113 (2008).
8. Eid, R. and Paultre, P. \Plasticity-based model for cir- BIOGRAPHIES
cular concrete columns con ned with bre-composite
sheets", Engineering Structures, 29(12), pp. 3301-3311 Davood Mosto nejad was born in 1960, and entered
(2007). Tehran University to study Civil Engineering in 1978.

www.SID.ir
Archive of SID
Predicting the Behavior of FRP Con ned Concrete Columns 481

He obtained his B.S. in 1985, his M.S. in Structural En- Isfahan University of Technology, as well as at Carleton
gineering in 1987 from Isfahan University of Technology University in Canada and the University of Arizona in
(IUT) and, nally, his Ph.D. in Structural Engineering- USA.
Concrete Structures in 1996 from Carleton University
in Canada. Dr. Mosto nejad has been working as a Hassan Saadatmand was born in 1978. He obtained
faculty member in the Civil Engineering Department his B.S. in Civil Engineering in 2001, and then his M.S.
of Isfahan University of Technology (IUT) since 1987, in Structural Engineering in 2004 from Yazd University.
and is now working as a university professor. Dur- Since that time, he has been working as a lecturer in the
ing the previous two decades, he has been teaching Department of Civil Engineering of Azad University of
di erent courses on reinforced concrete structures at Khoraskan.

www.SID.ir