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

net/publication/245308172

Stress–Strain Model for Laterally Confined Concrete

Article  in  Journal of Materials in Civil Engineering · December 2005

DOI: 10.1061/(ASCE)0899-1561(2005)17:6(607)

CITATIONS READS
53 463

3 authors:

Weena P. Lokuge Jay G. Sanjayan

University of Southern Queensland  Swinburne University of Technology
88 PUBLICATIONS   294 CITATIONS    290 PUBLICATIONS   7,543 CITATIONS   

SEE PROFILE SEE PROFILE

Sujeeva Setunge
RMIT University
206 PUBLICATIONS   1,558 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Geopolymers View project

Retrofitting Public Buildings for Energy and Water Efficiency View project

All content following this page was uploaded by Weena P. Lokuge on 19 February 2017.

The user has requested enhancement of the downloaded file.

 Stress–Strain Model for Laterally Confined Concrete
Weena P. Lokuge1; J. G. Sanjayan2; and Sujeeva Setunge3

Abstract: High strength concrete 共HSC兲 with highly desirable structural properties can lead to significant cost savings in heavily loaded
lower story columns of concrete structures. Its use has however, been limited by a concern regarding an increased brittleness compared
to normal strength concrete. It is well established that the ductility of HSC columns can be increased by confinement of the core of
concrete columns by lateral steel reinforcement. Confining pressure applied by the reinforcement is a function of the lateral strain of
concrete. Therefore establishing axial stress, axial strain, and lateral strain relationships is a timely concern. Based on shear failure of
concrete, a simple strain-based model is proposed in this paper. It is developed using prevailing test results for HSC with active
confinement.
DOI: 10.1061/共ASCE兲0899-1561共2005兲17:6共607兲
CE Database subject headings: High strength concretes; Stress strain relations; Concrete structures; Brittleness; Confinement;
Failures.

Introduction taken by a research team at Monash Univ. 共Attard and Setunge

The use of high strength concrete 共HSC兲 in building systems is
increasing around the world due to many advantages it offers.
However, its use is constrained by the well-known fact that HSC
is more brittle than normal strength concrete 共NSC兲. The issue of
ductility is extremely important in structures subjected to cyclic
loading conditions.
The above concerns have ld to a number of research projects
around the world 共Martinez et al. 1984; Pessiki and Pieroni 1997兲.
In all cases the experimental studies covered the behavior of
small-scale circular or square columns with a limited configura-
tion of steel reinforcements, tested under laboratory conditions.
One disadvantage of these types of experiments is that the value
of the minor principal stresses cannot be controlled, since the
confinement provided by lateral steel is passive. Most of the ex-
isting models for the stress–strain behavior of confined concrete
are based on column tests which incorporate the proposed effec-
tive area of confinement. There is no agreement among the re-
searchers about the confinement effectiveness of the column core
共Attard and Foster 1995; Cusson and Paultre 1995; Foster
et al. 1998兲. In most studies, the confining pressure is assumed to
be that corresponding to the yield stress of the confining steel.
A different approach to address the same problem is being
1
PhD Student, Dept. of Civil Engineering, Monash Univ., Clayton,
VIC 3800, Australia 共corresponding author兲. E-mail: weena.lokuge@
eng.monash.edu.au
2
Associate Professor, Dept. of Civil Engineering, Monash Univ.,
Clayton VIC 3800, Australia.
3
Sr. Lecturer, School of Civil and Chemical Engineering, RMIT
Univ., Melbourne 3001, Australia.
Note. Associate Editor: Christopher K. Y. Leung. Discussion open
until May 1, 2006. Separate discussions must be submitted for individual
papers. To extend the closing date by one month, a written request must
be filed with the ASCE Managing Editor. The manuscript for this paper
was submitted for review and possible publication on October 1, 2002;
approved on February 3, 2004. This paper is part of the Journal of
Materials in Civil Engineering, Vol. 17, No. 6, December 1, 2005.
©ASCE, ISSN 0899-1561/2005/6-607–616/$25.00.
1996; Candappa et al. 1999; Candappa 2000; Lokuge et al. 2000兲.
The behavior of plain concrete is studied under multiaxial stress
conditions with the aim of establishing a constitutive material
model for short-term loadings. Candappa 共2000兲 has proposed an
experimental program to investigate the complete axial stress,
axial strain, and lateral strain behavior of HSC using active con-
finement. Passive confinement 共using lateral reinforcement such
as spirals, ties, steel tubes, or other form of material like carbon
fiber兲 depends on lateral dilation of the concrete under axial load
and the stress–strain relationship of the confining material 共Can-
dappa et al. 1999兲 which necessitates the establishing of axial
stress–lateral strain behavior of confined concrete.
This paper describes a novel approach in predicting the stress–
strain relationship of HSC subjected to active lateral confinement.
The proposed model formulation is based on the experimental
results reported by Candappa 共2000兲. Although the proposed
strain-based model was developed for concrete with active con-
finement, it is extended for the case of passive confinement using
an iterative procedure.
Existing Models
Based on the test results, various stress–strain models for con-
fined concrete have been proposed, such as Sheikh and Uzumeri
共1982兲, Mander et al. 共1988兲, and Cusson and Paultre 共1995兲
models. Existing stress–strain models for confined, unconfined
normal, as well as high strength concrete can be divided into three
broad categories. One group of researchers used a form of equa-
tion proposed by Sargin et al. 共1971兲 共Table 1兲. The second group
of researchers proposed second order parabola for the ascending
branch and a straight line for the descending branch and their
studies were based on equations proposed by Kent and Park
共1971兲 共Table 2兲. The third group developed stress–strain relation-
ships based on equations suggested by Popovics 共1973兲 共Table 3兲.
In these models, selected parameters were included in the stress–
strain curves and then they were calibrated using the test results.
In these stress–strain models, 共␴1 , ␧1兲 are the coordinates of
any point in the stress–strain curve, ␧co is the peak axial strain of

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005 / 607

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Table 1. Stress–Strain Models for Confined Concrete Based on Sargin et al. 共1971兲
Y = 关AX + 共D − 1兲X2兴 / 关1 + 共A − 2兲X + DX2兴

Researcher A D
Sargin et al. 共1971兲 Ec␧co / kf c 0.65− 7.25f c ⫻ 10−3
Wang et al. 共1978兲 Different parameters for ascending and descending branches
Ahmad and Shah 共1982兲 E / Ep 1.111+ 0.876A − 4.0883共␶oct / f c兲
El-Dash and Ahmad 共1995兲 Ec / E p 共16.5/ 冑 f c兲关f l / 共s / dsp兲兴0.033
Attard and Setunge 共1996兲 Eti␧cc / f cc 共A − 1兲 / ␣关1 − 共f pl / f cc兲兴 + A2共1 − ␣兲 / ␣2共f pl / f cc兲关1 − 共f pl / f cc兲兴
2

Assa et al. 共2001兲 Ec␧cc / f cc 关共␧80 / ␧cc兲2 − 共0.2A + 1.6兲共␧80 / ␧cc兲 + 0.8兴 / 0.2共␧80 / ␧cc兲2

unconfined concrete strength f c, ␧cc is the peak axial strain of
confined concrete strength f cc, f l is the confining pressure, Ec is
the elastic modulus of concrete, Ei is the initial tangent modulus,
and E f is the secant modulus of concrete measured at peak stress.
Sargin et al. 共1971兲 proposed a form of equation 共Table 1兲 and
various versions of this were proposed by many researchers in
predicting the stress–strain curves for unconfined and confined
concrete 共Ottosen 1979; Martinez et al. 1984兲. Wang et al. 共1978兲
used a similar form of an equation but the parameters were dif-
ferent for the ascending and descending branches. Ahmad and
Shah 共1982兲 proposed another version of the same form of equa-
tion for concrete confined by steel reinforcement. Later El-Dash
and Ahmad 共1995兲 and Attard and Setunge 共1996兲 modified this
form of equation and calibrated the model parameters by using
confined and unconfined HSC. The Attard and Setunge 共1996兲
model was shown to be applicable for a wide range of concrete
strengths 共20– 130 MPa兲. Assa et al. 共2001兲 proposed stress–
strain relationships, using two parameters and validated it using
test results for confined concrete with a strength range of
20– 90 MPa.
Kent and Park 共1971兲 suggested a second-degree parabola for
the ascending branch 共independent of the confinement兲 and a lin-
ear descending branch 共depending on the confinement兲 for con-
crete confined by rectangular hoops. Park et al. 共1982兲 modified
this model by including strength and strain increase due to the
confinement. Scott et al. 共1982兲 and Sheikh and Uzumeri 共1982兲
also used a similar kind of equation for parabolic ascending
branch and linear descending branch and it was validated for
NSC. Samra 共1990兲 used the model proposed by Kent and Park
共1971兲 and showed a weakness in the ACI Building Code for
detailing columns for ductility.
Saatcioglu and Razvi 共1992兲 suggested an analytical model for
confined NSC and proved it to be valid for circular, square, and
rectilinear columns. This model consists of a parabolic ascending
branch, linear descending branch, and a constant residual
strength.
Application of this model was confirmed for NSC and it over-
estimated the initial modulus of elasticity when applied to HSC
共Razvi and Saatcioglu 1999兲. Therefore a relationship first pro-
posed by Popovics 共1973兲 and later used by Mander et al. 共1988兲
for NSC was used by Cusson and Paultre 共1995兲 and Razvi and
Saatcioglu 共1999兲 for HSC. Mendis et al. 共2000兲 later modified
the Scott et al. 共1982兲 model to predict the behavior of normal and
high strength confined concrete.
Shah et al. 共1983兲 developed a stress–strain equation for con-
fined concrete in an attempt to predict the envelope curve of
cyclically loaded columns. They pointed out that the Kent and
Park 共1971兲 model has a discontinuity between the ascending and
the descending branches and using a linear relationship for the
descending branch is not accurate. By noting that the curve pro-
posed by Sargin et al. 共1971兲 does not give good descending
branches for unconfined concrete, Shah et al. 共1983兲 proposed a
different form of equation for the ascending branch and an expo-
nential equation for the descending branch. This model is valid
for normal weight confined concrete.
Based on an equation suggested by Popovics 共1973兲, Carreira
and Chu 共1985兲 proposed a model for plain concrete in compres-
sion. Mander et al. 共1988兲 developed a similar relationship which
was extended to predict the stress–strain curves for confined con-
crete subjected to cyclic loading. The major difference between
this model and the other reported models is that it has one con-
tinuous function for both ascending and descending branches. The
ascending branch of the Cusson and Paultre 共1995兲 model is a
relationship originally proposed by Popovics 共1973兲 and the de-
scending branch proposed was a modified version of Shah et al.
共1983兲. The model was proven to be valid from the experimental
studies carried out for a wide range of concrete 共60– 120 MPa兲.
Wee et al. 共1996兲 proposed a modified version of the Carreira and

Table 2. Stress–Strain Models for Confined Concrete Based on Kent and Park 共1971兲
Researcher Ascending branch 共␴1兲 Descending branch 共␴1兲
Kent and Park 共1971兲 f cc关2共␧1 / 0.002兲 − 共␧1 / 0.002兲 兴
2
f cc关1 − Zm共␧1 − 0.002兲兴
Sheikh and Uzumeri 共1982兲 Kf c关2共␧1 / ␧cc兲 − 共␧1 / ␧cc兲2兴 f cc关1 − Zm共␧1 − ␧cc兲兴
Park et al. 共1982兲 Kf c关2共␧1 / 0.002K兲 − 共␧1 / 0.002K兲2兴 Kf c关1 − Zm共␧1 − 0.002K兲兴
Scott et al. 共1982兲 Same as Park et al. 共1982兲
Samra 共1990兲 Same as Kent and Park 共1971兲
Saatcioglu and Razvi 共1992兲 f cc关2共␧1 / ␧cc兲 − 共␧1 / ␧cc兲2兴1/共1+2K兲 f cc关1 − Zm共␧1 − ␧cc兲兴
Saatcioglu et al. 共1995兲 Same as Saatcioglu and Razvi 共1992兲
Razvi and Saatcioglu 共1999兲 f ccxr / 共r − 1 − xr兲 f cc关1 − Zm共␧1 − ␧cc兲兴
Mendis et al. 共2000兲 Kf c关2共␧1 / ␧cc兲 − 共␧1 / ␧cc兲2兴 Kf c关1 − Zm共␧1 − ␧cc兲兴
Shah et al. 共1983兲 f cc兵1 − 关1 − 共␧1 / ␧cc兲兴A其 f cce−k共␧1 − ␧cc兲
1.15

608 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Table 3. Stress–Strain Models for Confined Concrete Based on Popovics 共1973兲
Researcher
Popovics 共1973兲
Carreira and Chu 共1985兲
Mander et al. 共1988兲
Hsu and Hsu 共1994兲
Cusson and Paultre 共1995兲 f cc兵k共␧1 / ␧cc兲 / 关k − 1 − 共␧1 / ␧cc兲 兴其
Wee et al. 共1996兲
Hoshikuma et al. 共1997兲
Chu 共1985兲 model noting that the Wang et al. 共1978兲 model is
complicated for routine use and that the Carreira and Chu 共1985兲
model does not represent the descending branch of the curve for a
wider range of concrete strength. For the ascending branch of the
stress–strain curves, Wee et al. 共1996兲 used the equation proposed
by Carreira and Chu 共1985兲 but for modeling the descending
branch, two correction factors were introduced. Hsu and Hsu
共1994兲 used a modified version of the Carreira and Chu 共1985兲
model for the ascending branch and an exponential function for
the descending branch. Hoshikuma et al. 共1997兲 developed a new
model after observing that a second order parabola for the ascend-
ing branch can reflect only three boundary conditions of the four
boundary conditions which should be reflected by the ascending
branch of the stress–strain curves. None of the above models were
capable of predicting the volumetric strain of concrete.
Ottosen 共1979兲 proposed a theoretical model based on nonlin-
ear elasticity where the secant values of Young’s modulus 共Es兲
and Poisson’s ratio 共␯a兲 are changed accordingly. However Ot-
tosen 共1979兲 suggested parameter values only for the ascending
branch of the stress–strain curves. The Ottosen model has the
capability of predicting axial stress, axial strain, and lateral strain
behavior. Candappa 共2000兲 used the Ottosen 共1979兲 model and
developed a stress-based model which was proven to be valid for
HSC. However further studies carried out by the writers found
that it did not predict the results from other sources. As a result
the writers have approached the problem from a fresh perspective
and developed the model presented in this paper.
Model Formulation
The model formulation is based on the experimental results re-
ported by Candappa 共2000兲, which were obtained under active
lateral confinement.
Experimental Results
Candappa et al. 共2001兲 used four grades of concrete 共40, 60, 75,
and 100 MPa兲 and three confining pressures 共4, 8, and 12 MPa兲
as test variables in the experimental program. Since the tests were
performed in duplicate, a total of 24 tests have been reported.
Further details of the experimental program can be found in Can-
dappa et al. 共1999, 2001兲 and Candappa 共2000兲. However, for the
convenience of the reader, a brief outline of the experimental
setup and procedures are presented here.
Cylindrical specimens 共98 mm in diameter and 200 mm high兲
were prepared according to AS1012, Part 2 and cured according to
AS1012, Part 8. Having both ends properly ground, pores filled
with plaster, two strain gages were placed longitudinally on two
diametrically opposite sides. Similarly, another two strain gages
were placed laterally at the middle third on two diametrically
JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005 / 609
Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Ascending branch 共␴1兲 Descending branch 共␴1兲
f c共␧1 / ␧co兲关n / n − 1 + 共␧1 / ␧co兲 兴
n
f cc ⫻ ␤ / 共␤ − 1 − x␤兲
f cc ⫻ r / 共r − 1 − xr兲
f ccn ⫻ ␤ / 共n␤ − 1 − xn␤兲 0.3f cce−0.8共x − xd兲
0.5
k k2
f cce−k1共␧1 − ␧cc兲
f cc ⫻ ␤ / 共␤ − 1 − x␤兲 k1 f cc ⫻ ␤ / 共k1␤ − 1 − xk2␤兲
Ec␧1关1 − 共1 / n兲共␧1 / ␧cc兲n−1兴 f cc − Edes共␧u − ␧cc兲
opposite sides. The specimen thus prepared was then accommo-
dated in a triaxial cell. The required pressure was applied to the
specimen using oil through a flexible polyurethane membrane. A
bleed valve was utilized to maintain the required confining
pressure.
The two strain gages placed longitudinally were used to mea-
sure the axial deformations. Since the strain gage readings are
unreliable after reaching the peak strength, linear variable dis-
placement transducers 共LVDTs兲 were also used to measure the
axial deformation of concrete. In all the experimental curves,
axial strains were verified using these two methods of measure-
ments. Similarly, lateral strains were also measured by strain
gages as well as a clip gage.
Basis of Model Formulation
The proposed model formulation has several novel key aspects,
based on the following observations:
1. Lateral strain, axial strain relationship—Normalized lateral
strains were plotted against normalized axial strains for all
four concrete strengths and for all confining pressures. It is
clearly seen that for a particular concrete strength, these
curves have a similarity in the shape irrespective of the con-
fining pressure applied. The gradient of these curves is di-
rectly related to an important parameter, Poisson’s ratio.
2. Shear stress, shear strain relationship—in an attempt to de-
velop a stress–strain model, normalized shear stress versus
normalized shear strain was plotted. Further analysis indi-
cated that these curves coincide for a particular concrete
compressive strength with different confining pressures. This
observation is consistent with the theory proposed by Nielsen
共1999兲.
3. At the peak axial stress volumetric strain becomes zero—
Volumetric strain is an important factor especially when lat-
eral confinement is involved. Normalized volumetric strain
versus normalized axial strain curves have a similar shape.
The writers found that the volumetric strain is close to zero at
peak axial stress.
Initially the uniaxial compressive strength of concrete f c has to
be given as an input variable. Corresponding axial strain ␧co is
assumed to be 0.002. The strain-based model is formulated based
on theory of shear failure with three relationships as follows:
1. Relationship between peak axial stress, corresponding axial
strain and lateral strain;
2. Relationship between axial strain and lateral strain at any
stress; and
3. Relationship between any shear stress and shear strain which
is used to find a relationship between axial stress, axial
strain, and lateral strain.
In the reported studies so far, a second order parabolic curve,
or a form of an equation, was proposed for the ascending branch
 perimental data reported by Candappa 共2000兲. Each line in these
graphs represents the average of two test results. The curves are
for generalized lateral strain 共with respect to lateral strain at peak
stress兲, ¯␧2 versus generalized axial strain 共with respect to axial
strain at peak stress兲 ¯␧1

␧2
¯␧2 =
␧⬘cc

and

Fig. 1. Typical stress–strain relationship for confined and unconfined
concrete
of the stress–strain curves of confined concrete. For the descend-
ing branch, past researchers proposed a linear relationship or an
exponential function. However some models could only predict
the behavior of NSC. In this paper the writers propose complete
stress–strain curves based on a basic theory of shear failure of
concrete. At the peak stress, the ascending and descending
␧1
¯␧1 = 共1兲
␧⬘cc
where ␧1 and ␧2 = axial and lateral strains, respectively;
␧cc = axial strain corresponding to peak axial stress, and
␧⬘cc = corresponding lateral strain. The gradient of these curves will
be related to Poisson’s ratio. It will be a constant at the beginning
and increases gradually. Therefore the observed behavior can be
approximated by a pairwise equation

冉 冊
冦 冧
branches are compatible. The proposed stress–strain relationship ␧1
consists of different exponential functions for ascending and de- ␯ai if ␧1 艋 ␧⬘
␧2 ␧cc

冉 冊
scending branches for unconfined and confined concrete as illus- = 共2兲
trated in Fig. 1. ␧⬘cc ␧1 a
if ␧1 ⬎ ␧⬘
␧cc
Predicting Relationship Between Axial Strain
a = material parameter which depends on the uniaxial concrete
and Lateral Strain
strength. The values of a for each concrete strength were found
In order to find a relationship between axial strain and lateral using the best fit curves and summarized in Table 3. a can be
strain, the curves shown in Fig. 2 were developed based on ex- approximated by

Fig. 2. Normalized lateral strain versus normalized axial strain

610 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 Fig. 3. Normalized volumetric strain factor versus normalized axial strain factor

a = 0.0177f c + 1.2818 共3兲 ⬘…

␧⬘ can be obtained by equating the right hand side of Eq. 共2兲.
Initial Poisson’s ratio 共␯ai 兲 is defined in Candappa 共2000兲
␯ai = 8 ⫻ 10−6共f c兲2 + 0.0002f c + 0.138
Therefore if axial strain ␧cc and lateral strain 共␧⬘cc兲 corresponding
to peak axial stress can be expressed, the relationship between
axial strain and lateral strain at any time is fully defined by
Eq. 共2兲.
Attard and Setunge 共1996兲 suggested equations for axial strain
corresponding to peak axial stress ␧cc
␧cc
␧co
= 1 + 共17 − 0.06f c兲
fl
fc
冉冊
where f l = confining pressure and ␧co = axial strain corresponding
to peak uniaxial compressive strength 共generally assumed to be
0.002兲. Peak axial stress for confined concrete f cc is defined as
Predicting Lateral Strain at Peak Axial Stress „␧cc
In order to find a relationship between axial and lateral strain at
peak axial stress, normalized volumetric strain factor 共␧¯ v兲 versus
normalized axial strain factor 共␧ ¯ 1兲 graphs were plotted for all
共4兲 concrete batches and all confining pressures. Normalized volu-
metric strain factor 共␧
¯ v兲 is defined as
␧v ␧1 + 2␧2
¯␧v = = 共9兲
␧v,max ␧v,max
共5兲
where ␧v,max = maximum volumetric strain.
Normalized axial strain factor 共␧ ¯ 1兲 is the same as defined in
Eq. 共1兲. The curves are shown in Fig. 3. It is clearly seen that
there is a similarity in all the curves irrespective of concrete
strength or confining pressure. Initially volumetric strain in-
creases gradually and then starts decreasing. That is, with the

冉 冊 k
increasing axial stress, the concrete first contracts and then gradu-
f cc fl ally expands. At a particular point the contraction changes to ex-
= +1 共6兲
fc ft pansion. At this point the volumetric strain is the maximum. This
point corresponds to an axial strain of 0.55␧cc for 40 MPa con-
where k = constant given by
crete and 0.7␧cc for 100 MPa concrete, where ␧cc is the axial

冉
k = 1.25 1 + 0.062
fl
fc
冊共f c兲−0.21 共7兲
strain at peak axial stress.
It is interesting to note that there is a point where volumetric
strain changes its sign. That is, at this point of zero volume, the
and f t = tensile strength. As silica fume was not used in this concrete comes back to its original volume. When the experimen-
project tensile strength is given by tal data are carefully analyzed for the peak axial stress it is ob-
served that volumetric strain becomes zero when peak axial stress
f t = 0.9 ⫻ 0.32共f c兲0.67 共8兲
is reached. Normalized volumetric strain factor and axial strain
In this study Eq. 共5兲 is used to predict the axial strain correspond- factor corresponding to the peak stress is marked as “⫻” in each
ing to peak axial stress, which is determined by Eq. 共6兲. curve in Fig. 3. Therefore at peak stress

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005 / 611

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 Fig. 4. Normalized shear stress factor versus normalized shear strain factor

␧1 + 2␧2 ␧1 + ␧2
¯␧v = =0 共10兲 ␥max = 共13兲
␧v,max 2
and where ␴1 = axial stress; f l = lateral stress 共confining pressure兲;
␧1 = axial strain; and ␧2 = lateral strain. Shear stress factor is de-
␧cc = 2␧⬘cc 共11兲
fined as
By introducing Poisson’s ratio, Eq. 共11兲 can be written as
maximum shear stress at any stress level ␶max
␯af = 0.5 共12兲 ¯␶max = =
maximum shear stress at peak ␶mp
where ␯af = secant value of Poisson’s ratio at peak stress. This has 共14兲
been addressed in the literature by a few researchers. Ottosen
共1979兲 suggested a value of 0.36 for ␯af and his study is said to be Similarly shear strain factor is defined as
valid up until failure. He further stated that secant value of Pois- maximum shear strain at any stress level ␥max
son’s ratio must always be less than 0.5. Dahl 共1992兲 proposed ¯␥max = =
maximum shear strain at peak ␥mp
0.5 for ␯af for triaxial tests and later Cusson and Paultre 共1995兲
and Candappa et al. 共2001兲 found that it is approximately 0.5. 共15兲
Based on the experimental results by Candappa 共2000兲 and ob-
Shear stress factor 共14兲 against shear strain factor 共15兲 graphs
serving the volumetric behavior for both normal and high strength
were plotted for all the concrete batches and for all confining
concrete 共Fig. 3兲, it can be confirmed that the secant value of
pressures. They are shown in Fig. 4. The relationship between
Poisson’s ratio at peak stress is 0.5.
these factors can be approximated by the following exponential
functions:
Relationship Between Axial Stress, Axial Strain,
and Lateral Strain
Maximum shear stress 共␶max兲 and maximum shear strain 共␥max兲
¯␶max = 再 1 − e−c␥¯ max before peak
2
ed␥¯ max − d after peak
冎 共16兲

are defined as where c and d = material parameters. They depend on the uniaxial
␴1 − f l concrete strength. Their values for each concrete strength were
␶max = found using the best fit curves and summarized in Table 3. Sub-
2
stituting shear stress 共␶
¯ max兲 and shear strain factors 共␥
¯ max兲 in terms
and of axial stress 共␴1兲, axial strain 共␧1兲, lateral strain 共␧2兲, maximum

612 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 Fig. 5. Procedure in drawing stress–strain curves for confined concrete 共given ␧1 and f c兲
shear stress at peak 共␶mp兲, and maximum shear strain at peak
共␥mp兲, complete deformational behavior of concrete can be ex-
pressed as
再 冎
2␶mp共1 − e−c共␧1+␧2/2␥mp兲兲 + f l before peak
␴1 = 2 共17兲
2␶mp共ed共␧1 + ␧2/2␥mp兲 − d兲 + f l after peak
Thus, having established the axial strain and lateral strain rela-
tionship and shear strain factor and shear stress factor relation-
ship, complete stress–strain curves can now be generated.
Model Application in Columns
with Passive Confinement
In the literature, many experimental results on cylindrical con-
crete columns subjected to passive confinement are available. In
the last 2 decades, the use of fiber reinforced polymer composites
as the method of confinement has been gaining increasing popu-
larity. Experimental results on cylindrical columns confined by
fiber reinforced plastic sheets are ideal for verification of the pro-
posed model in this paper. From the many available experimental
results, it was decided to use the experimental results reported by
Candappa 共2000兲, who used carbon fiber reinforced plastic sheets
as the method of confinement. The availability of the experimen-
tal data for axial stress, axial strain, and lateral strain in electronic
form for the writers was the main reason for this selection.
JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005 / 613
Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Experimental Program
There were two test variables, namely the compressive strength of
concrete and the level of confinement provided. Three compres-
sive strengths 共30, 60, and 100 MPa兲 and three levels of confine-
ment 共1, 2, and 3 wraps兲 were investigated. Tests were performed
in duplicate for each wrapping configuration and each compres-
sive strength.
The average tensile strength of carbon fiber composites was
741.3 MPa and the average Young’s modulus was 101,920 MPa.
Specimen preparation method used by Candappa 共2000兲 is de-
scribed here in detail.
The cylindrical specimens 共150 mm diameter and 300 mm
high兲 were allowed to air dry for about 1 week after being taking
out from the curing bath. They were then sand blasted to remove
any loose particles and the dust was wiped off thoroughly. Any
surface pores were filled with a quick setting filler. Having ap-
plied an epoxy-based primer onto the concrete surface, it was
allowed to cure overnight. Before fiber wrapping, the concrete
specimen surface was slightly sanded.
The required length of fiber wrap was cut with an allowance of
30 mm for overlapping. Laminating resin of equal mass as the
weight of the fiber wrap was applied on the surface of the fiber
wrap using a soft brush. Distributing laminating resin and squeez-
ing any excess resin out was performed using a bubble breaker.
The composite fiber was then wrapped tightly around the concrete
specimen and allowed to cure for 2 weeks.
 Fig. 6. Stress–strain curves for carbon fiber wrapped concrete
The LVDTs were used to measure the axial deformations. The
clip gage used to measure the lateral deformations in actively
confined concrete was modified to measure the lateral strains in
passively confined concrete columns.
The confining pressure exerted by the fiber reinforced plastic
sheets on the concrete core is of passive type. As the axial stress
increases, the corresponding lateral strain in concrete increases,
and a tensile hoop stress is developed in the confining sheets
which is balanced by the uniform radial pressure due to the lateral
expansion of concrete. The hoop strength of the carbon wrap was
assumed to be the tensile strength of the carbon fiber composite
共741.3 MPa兲. The diameter of the specimen was 150 mm and the
thickness of one layer of carbon fiber was 0.24 mm.
Results
The constitutive model for HSC presented in this paper is applied
to obtain the behavior of carbon fiber wrapped concrete columns.
614 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005
Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
The analytical findings are compared with the experimental re-
sults outlined above. Comparisons are shown in Fig. 5.
When higher confining pressure is applied to HSC 共100 MPa
concrete with 3 carbon fiber wrap兲, there may be bond failure
共bond between concrete and carbon fiber兲. However, in the model
prediction a perfect bond between concrete and carbon fiber is
assumed. This may be the reason for the difference in model
prediction and the experimental curve for 100 MPa concrete with
3 carbon fiber wraps. Apart from this, all the other specimens are
in good agreement with the model predictions.
Many researchers have noted that the strain measured in the
confining FRP at rupture is in many cases lower than the ultimate
strain of FRP tested for tensile strength 共Lorenzis 2001兲. The
recorded hoop strains corresponding to rupture had a range of
50–80% of the failure strain obtained in the tensile tests 共Xiao
and Wu 2000兲. This phenomenon considerably affects the accu-
racy of the model.
Conclusions ␴1 ⫽ axial stress of confined concrete;
␶max ⫽ maximum shear stress;
1. The proposed stress–strain model for confined concrete can ¯␶max ⫽ normalized maximum shear stress factor; and
be summarized as shown in Fig. 6. ␶m,p ⫽ maximum shear stress corresponding to peak axial
2. When confined concrete is stressed, initially it contracts 共re- stress.
ducing in volume兲. After a certain stress, it starts to expand
共increasing in volume兲. The maximum volumetric strain oc-
curs at this point where the contraction turns into expansion. References
This point occurs at an axial strain of 0.5␧cc for 40 MPa
concrete and 0.7␧cc for 100 MPa concrete; where ␧cc is the Ahmad, S. H., and Shah, S. P. 共1982兲. “Stress-strain curves of concrete
axial strain at peak axial stress. confined by spiral reinforcement.” ACI J., 79共6兲, 484–490.
3. At the time of peak axial stress, the volume of concrete is Assa, B., Nishiyama, M., and Watanabe, F. 共2001兲. “New approach for
back to its original unloaded volume. That is, the volumetric modeling confined concrete. I: Circular columns.” J. Struct. Eng.,
strain is observed to be back to zero at peak axial stress. This 127共7兲, 743–750.
means that the magnitude of lateral strain at peak axial stress Attard, M. M., and Foster, S. J. 共1995兲. “Ductility of high strength con-
is half of corresponding axial strain. This is a very useful crete columns.” Technical Rep. No. UNICIV R-344, School of Civil
Engineering, The Univ. of New South Wales, New South Wales,
observation for constitutive models which has not been re-
Australia.
ported in the literature. Attard, M. M., and Setunge, S. 共1996兲. “Stress-strain relationship of con-
4. The proposed strain-based stress–strain model is a new ap- fined and unconfined concrete.” ACI Mater. J., 93共5兲, 432–442.
proach in predicting the behavior of HSC subjected to active Candappa, D. 共2000兲. “The constitutive behavior of high strength con-
lateral confinement. crete under lateral confinement.” PhD thesis, Monash Univ., Clayton,
5. Proposed model can be applied to concrete with active as VIC, Australia.
well as passive confinement. It is proven to be generally in Candappa, D. P., Sanjayan, J. G., and Setunge, S. 共2001兲. “Complete
close agreement with the experimental test results for con- triaxial stress-strain curves of high-strength concrete.” J. Mater. Civ.
crete confined by carbon fiber wraps. Eng., 13共3兲, 209–215.
Candappa, D. P., Setunge, S., and Sanjayan, J. G. 共1999兲. “Stress versus
strain relationship of high strength concrete under high lateral con-
finement.” Cem. Concr. Res., 29共12兲, 1977–1982.
Carreira, D. J., and Chu, K. H. 共1985兲. “Stress–strain relationship for
Notation plain concrete in compression.” ACI J., 83共6兲, 797–804.
Cusson, D., and Paultre, P. 共1995兲. “Stress–strain model for confined high
The following symbols are used in this paper: strength concrete.” J. Struct. Eng., 121共3兲, 468–477.
A ; D ⫽ constants in form of equation proposed by Sargin Dahl, K. K. B. 共1992兲. “A failure criterion for normal and high strength
et al. 共1971兲; concrete.” Project 5 Rep. No. 5.6, American Concrete Institute,
a ⫽ material constant in Eq. 共2兲; Detroit.
c , d ⫽ material constants in Eq. 共17兲; El-Dash, K. M., and Ahmad, S. H. 共1995兲. “A model for stress-strain
Es ⫽ secant value of Young’s modulus; relationship of spirally confined normal and high-strength concrete
columns.” Mag. Concrete Res., 47共171兲, 177–184.
f c ⫽ uniaxial compressive strength of concrete;
Foster, S. J., Liu, J., and Sheikh, S. A. 共1998兲. “Cover spalling in HSC
f cc ⫽ peak axial stress of confined concrete; columns loaded in concentric compression.” J. Struct. Eng., 124共12兲,
f l ⫽ confining pressure; 1431–1437.
f t ⫽ tensile strength of concrete; Hoshikuma, J., Kawashima, K., Nagaya, K., and Taylor, A. W. 共1997兲.
X ⫽ ␧1 / ␧cc; “Stress-strain model for confined reinforced concrete in bridge piers.”
Y ⫽ ␴1 / f cc; J. Struct. Eng., 123共5兲, 624–633.
␥max ⫽ maximum shear strain; Hsu, L. S., and Hsu, C. T. T. 共1994兲. “Complete stress-strain behavior of
¯␥max ⫽ normalized maximum shear strain factor; high-strength concrete under compression.” Mag. Concrete Res.,
␥m,p ⫽ maximum shear strain corresponding to peak axial 46共169兲, 301–312.
stress; Kent, D. C., and Park, R. 共1971兲. “Flexural members with confined con-
crete.” J. Struct. Div. ASCE, 97共7兲, 1969–1990.
␧⬘ ⫽ axial strain at point where shape of axial strain and
Lokuge, W. P., Sanjayan, J. G., and Setunge, S. 共2000兲. “Design of high
lateral strain curves deviate; strength concrete columns for ductility. Concrete in Australia.” J.
␧co ⫽ axial strain corresponding to peak uniaxial Concrete Institute of Australia, 26共3兲, 17–20.
compressive strength; Lorenzis, L. D. 共2001兲. “A comparative study of models on confinement
␧cc ⫽ axial strain corresponding to peak axial stress of of concrete cylinders with FRP composites.” Research Rep. Prepared
confined concrete; for Chalmers Univ. of Technology, Goteborg, Sweden.
␧⬘cc ⫽ lateral strain corresponding to peak axial stress of Mander, J. B., Priestley, M. J. N., and Park, R. 共1988兲. “Theoretical
confined concrete; stress-strain model for confined concrete.” J. Struct. Eng., 114共8兲,
␧v ⫽ volumetric strain; 1804–1826.
¯␧v ⫽ normalized volumetric strain factor; Martinez, S., Nilson, A. H., and Slate, F. O. 共1984兲. “Spirally reinforced
high-strength concrete columns.” ACI J., 81共5兲, 431–442.
␧v,max ⫽ maximum volumetric strain;
Mendis, P., Pendyala, R., and Setunge, S. 共2000兲. “Stress-strain model to
␧1 ⫽ axial strain; predict the full-range moment curvature behavior of high-strength
¯␧1 ⫽ normalized axial strain factor; concrete sections.” Mag. Concrete Res., 52共4兲, 227–234.
␧2 ⫽ lateral strain; Nielsen, M. P. 共1999兲. Limit analysis and concrete plasticity, CRC, Boca
¯␧2 ⫽ normalized lateral strain factor; Raton, Fla.
␯af ⫽ Poisson’s ratio at peak axial stress; Ottosen, N. S. 共1979兲. “Constitutive model for short-time loading of con-
␯ai ⫽ initial Poisson’s ratio; crete.” J. Eng. Mech. Div., 105共1兲, 127–141.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005 / 615

Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Park, R., Priestley, M. J. N., and Gill, W. D. 共1982兲. “Ductility of square-
confined concrete columns.” J. Struct. Div. ASCE, 108共4兲, 929–950.
Pessiki, S., and Pieroni, A. 共1997兲. “Axial load behavior of large-scale
spirally-reinforced high-strength concrete columns.” ACI Struct. J.,
94共3兲, 304–314.
Popovics, S. 共1973兲. “A numerical approach to the complete stress–strain
curve of concrete.” Cem. Concr. Res., 3共5兲, 583–599.
Razvi, S., and Saatcioglu, M. 共1999兲. “Confinement model for high-
strength concrete.” J. Struct. Eng., 125共3兲, 281–289.
Saatcioglu, M., and Razvi, S. R. 共1992兲. “Strength and ductility of con-
fined concrete.” J. Struct. Eng., 118共6兲, 1590–1607.
Saatcioglu, M., Salamat, A. H., and Razvi, S. R. 共1995兲. “Confined col-
umns under eccentric loading.” J. Struct. Eng., 121共11兲, 1547–1556.
Samra, R. M., 共1990兲. “Ductility analysis of confined columns.” J. Struct.
Eng., 116共11兲, 3148–3161.
Sargin, M., Ghosh, S. K., and Handa, V. K. 共1971兲. “Effects of lateral
reinforcement upon the strength and deformation properties of con-
crete.” Mag. Concrete Res., 23共75–76兲, 99–110.
Scott, B. D., Park, R., and Priestley, N. 共1982兲. “Stress-strain behavior of
616 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / NOVEMBER/DECEMBER 2005
View publication stats Downloaded 15 Jul 2010 to 139.86.13.152. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
concrete confined by overlapping hoops at low and high strain rates.”
ACI J., 79共1兲, 13–27.
Shah, S. P., Faគtis, A., and Arnold, R. 共1983兲. “Cyclic loading of spirally
reinforced concrete.” J. Struct. Eng., 109共7兲, 1695–1710.
Sheikh, S. A., and Uzumeri, S. M. 共1982兲. “Analytical model for concrete
confinement in tied columns.” J. Struct. Div. ASCE, 108共12兲,
2703–2722.
Standards Association of Australia. 共2000兲. “Methods of testing con-
crete.” AS1012, Australia.
Wang, P. T., Shah, S. P., and Naaman, A. E. 共1978兲. “Stress-strain curves
of normal and lightweight concrete in compression.” ACI J., 75共11兲,
603–611.
Wee, T. H., Chin, M. S., and Mansur, M. A. 共1996兲. “Stress-strain rela-
tionship of high strength concrete in compression.” J. Mater. Civ.
Eng., 8共2兲, 70–76.
Xiao, Y., and Wu, H. 共2000兲. “Compressive behavior of concrete confined
by carbon fiber composite jackets.” J. Mater. Civ. Eng., 12共2兲,
139–146.