Structural Evaluation of Full-Scale FRP-Confined Reinforced

Concrete Columns
Antonio De Luca, M.ASCE1; Fabio Nardone2; Fabio Matta, A.M.ASCE3; Antonio Nanni, F.ASCE4;
Gian Piero Lignola5; and Andrea Prota6

Abstract: The external confinement of RC columns by means of externally bonded fiber-reinforced polymer 共FRP兲 laminates is a well
established technique for strengthening and retrofitting purposes. This paper presents a pilot research that includes laboratory testing of
full-scale square and rectangular RC columns externally confined with glass and basalt-glass FRP laminates and subjected to pure axial
load. Specimens that are representative of full-scale building columns were designed according to a dated American Concrete Institute
共ACI兲 318 code 共i.e., prior to 1970兲 for gravity loads only. The study was conducted to investigate how the external confinement affects
peak axial strength and deformation of a prismatic RC column. The results showed that the FRP confinement increases concrete axial
strength, but it is more effective in enhancing concrete strain capacity. The discussion of the results includes a comparison with the values
obtained using existing constitutive models.
DOI: 10.1061/共ASCE兲CC.1943-5614.0000152
CE Database subject headings: Confinement; Deformation; Fiber reinforced polymer; Laminated materials; Concrete columns;
Strain; Structural analysis.
Author keywords: Confinement; Deformability; FRP laminates; Full-scale RC columns; Strengthening; Volumetric strain.

Introduction become a competitive alternative to conventional rehabilitation

A growing number of RC building and bridge structures is in need
of retrofitting and strengthening. There is a number of causes to
this need: deterioration caused by environmental effects; damage;
change in use of the structures; higher load demand as a conse-
quence of more severe code requirements; and, higher strength
and ductility demand to correct design or construction errors. The
use of fiber-reinforced polymer 共FRP兲 composite materials has
1
Postdoc, Dept. of Civil, Architectural, and Environmental Engineer-
ing, Univ. of Miami, 105 McArthur Engineering Bldg., Coral Gables, FL
33146-0630 共corresponding author兲. E-mail: adeluca@miami.edu
2 behavior of confined plane concrete cylinders subjected to pure
Postdoc, Dept. of Structural Engineering, Univ. of Naples, Federico
II, Via Claudio 21, Naples, P.O. Box I-80125, Italy. E-mail:
fabio.nardone@unina.it
3
Assistant Professor, Dept. of Civil and Environmental Engineering,
Univ. of South Carolina, 300 Main St., Room C210, Columbia, SC
29208. E-mail: mattaf@engr.sc.edu
4
Lester and Gwen Fisher Endowed Scholar, Professor and Chair of
Dept. of Civil, Architectural, and Environmental Engineering, Univ. of
Miami, 323 McArthur Engineering Bldg., Coral Gables, FL 33146-0630.
E-mail: nanni@miami.edu; and Professor, Dept. of Structural Engineer-
ing, Univ. of Naples, Federico II, Via Claudio 21, Naples, P.O. Box
I-80125, Italy. E-mail: antonio.nanni@unina.it
5
Assistant Professor, Dept. of Structural Engineering, Univ. of Naples,
Federico II, Via Claudio 21, Naples, P.O. Box I-80125, Italy. E-mail:
glignola@unina.it
6
Assistant Professor, Dept. of Structural Engineering, Univ. of Naples,
Federico II, Via Claudio 21, Naples, P.O. Box I-80125, Italy. E-mail:
aprota@unina.it
Note. This manuscript was submitted on January 4, 2010; approved
on June 29, 2010; published online on June 30, 2010. Discussion period
open until July 1, 2011; separate discussions must be submitted for indi-
vidual papers. This paper is part of the Journal of Composites for Con-
struction, Vol. 15, No. 1, February 1, 2011. ©ASCE, ISSN 1090-0268/
2011/1-112–123/$25.00.
techniques, whereby confinement of RC columns is one of the
most attractive applications. In FRP-confined concrete, the inter-
action between the two materials allows for the enhancement of
concrete strength and ultimate strain. In the case of small plain
concrete cylinders, the properties of the two materials are used in
the most desirable and successful way: 共1兲 the transverse FRP is
loaded in tension due to concrete dilation, thus containing con-
crete after its internal cracking and providing lateral confining
pressure; and 共2兲 the concrete is loaded in triaxial compression
due to the restraining action of the FRP laminate, thus leading to
a substantial improvement in strength and ultimate strain. The
axial loads has been extensively studied, and confinement effec-
tiveness has been experimentally proven since the late 1970s
共e.g., Kurt 1978; Fardis and Khalili 1982; Nanni and Bradford
1995; Mirmiran and Shahawy 1997; Karbhari and Gao 1997;
Spoelstra and Monti 1999; Fam and Rizkalla 2001; Shehata et al.
2002; Campione and Miraglia 2003; Lam and Teng 2003a,b; Mat-
thys et al. 2005; Matthys et al. 2006; Harajli 2006; Saenz and
Pantelides 2007; Wu et al. 2009; Toutanji et al. 2010兲. In concrete
columns with circular cross section, the confining effectiveness of
the FRP jacket is optimal since the geometrical configuration al-
lows the fibers to be effective on the entire cross section 共Lam and
Teng 2003a,b兲. Prismatic cross sections behave differently: as it is
well recognized, the confining pressure is high at the corners and
low along the flat sides, and the cross section is only partially
confined 共Mander et al. 1988; Lam and Teng 2003b兲. Confine-
ment of a rectangular cross section still enhances concrete
strength and ultimate strain, but its effectiveness is not as tangible
as that on a circular cross section 共Rocca et al. 2006; Rocca et al.
2008兲. A number of studies have been conducted on FRP-
confined RC prismatic columns, and several analytical models
have been proposed 共e.g., Mirmiran et al. 1998; Wang and Re-

112 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Table 1. Test Matrix
Note: 25.4 mm= 1 in.; 쏗25.4-mm bar= # 8 bar; and 쏗12.7-mm bar= # 4 bar.
strepo 2001; Campione and Miraglia 2003; Lam and Teng 2003b;
Kumutha et al. 2007; Wu and Wang 2009兲. These models, how-
ever, do not converge to similar predicted values.
Two databases reported by Hassan and Chaallal 共2007兲 and
Rocca et al. 共2008兲 assemble relevant experimental data reported
by several writers on RC prismatic columns externally confined
with FRP laminates and tested under compressive axial load from
1994 to 2007. The following can be noted: only 16 of the 113
column specimens included in the two databases 共14% of the
entire population兲 have short section sides larger than 254 mm 共10
in.兲; only 1 column specimen is higher than 2.44 m 共8 ft兲; the ratio
between specimen height and short section side is always smaller
than 5; 82 out of 113 column specimens 共73%兲 are square; 16
共14%兲 are rectangular with a side-aspect ratio 共ratio between large
and short section sides兲 of about 1.5; 15 共13%兲 are rectangular
with a side-aspect ratio of about 2; 86 column specimens 共76%兲
were confined with carbon FRP 共CFRP兲, 9 共8%兲 with glass FRP
共GFRP兲, and 16 共18%兲 with aramid FRP 共AFRP兲.
Full-scale experiments are generally limited by high cost and
availability of high-capacity testing equipment. Full-scale experi-
ments are therefore critical not only to validate a new technology,
but also to produce compelling evidence to justify rational design
methodologies.
When considering column confinement, glass fibers are par-
ticularly attractive. First, they have the highest ultimate strain of
any “high-modulus” fiber; second, their low fatigue and creep-
rupture resistance are not a detrimental factor in this type of ap-
plication. In addition, the shortage of carbon fiber supplies that
recently affected the market, as well as the development of high-
performance glass fibers with lower manufacturing costs made
GFRP cost competitive with CFRP laminates, thereby inducing an
important increase in the demand of glass fibers over the last
years. At the same time, continuous basalt fibers started to be-
come commercially available. Basalt fibers offer an alternative to
glass fibers due to their desirable characteristics, including for
example thermal stability 共Sim et al. 2005兲.
The research program presented herein aims at providing ex-
perimental evidence to remedy the lack of data needed to charac-
terize the behavior of full-scale RC columns strengthened with
FRP laminates subjected to pure compressive load. The condition
of pure axial load is atypical for RC columns given that they
always transmit axial compressive loads together with bending
JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 113
Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
moments, but this condition represents the first step to understand
the mechanics of FRP confinement. An important novelty of this
experimental program is the size of the column specimens, which
provides the opportunity to investigate and validate the use of
glass and glass-basalt hybrid laminates as confining systems on
the basis of experimental evidence representative of real cases. In
particular, this study aims at:
• Investigating the effectiveness of the FRP confinement in re-
lation to different cross-sectional geometries and sizes;
• Studying the deformability enhancement due to FRP confine-
ment;
• Investigating the contribution of GFRP and hybrid glass-basalt
FRP 共HFRP兲 laminates to concrete confinement;
• Assessing the equivalence of confined column performance
when different glass fibers of comparable quality are used;
• Assessing the contribution of the glass-basalt hybrid system on
concrete confinement; and
• Comparing the experimental values of the normalized axial
strength with those obtained using selected analytical models.
Experimental Program
The test matrix, summarized in Table 1, was designed considering
different factors, namely: shape factor 共side-aspect ratio兲, volume
factor 关volume-aspect ratio based on a benchmark volume of
610⫻ 610⫻ 3050 mm3 共24⫻ 24⫻ 120 in.3兲兴, FRP volumetric
ratio 共ratio between the total volume of confining FRP and vol-
ume of confined concrete兲, type and amount of FRP plies. The
specimens were intended to represent real size building columns
designed according to the American Concrete Institute 共ACI兲 318
code in use prior to 1970 共ACI 318-63兲 for gravity loads only.
Three series of column specimens were considered: Series S-1
cross section corresponds to a shape factor of 1.0 and a volume
factor of 1.0; Series R-1 cross section to a shape factor of 1.45
and a volume factor of 1.0; Series R-0.5 cross section to a shape
factor of 1.43 and a volume factor of 0.5. Three different types of
fiber fabrics were used: two types of conventional glass fiber
sheets from two different manufacturers 共which are denoted
herein as “Type A” and “Type B”兲; and a hybrid glass-basalt fiber
sheet 共glass-to-basalt fiber ratio 2:1兲. For each series one speci-
 Cross-section layout Cross-section layout
610 by 610 mm 508 by 737 mm Cross-section layout
B B 356 by 508 mm
B
2 12.7 cross-ties 2 12.7 cross-ties
12.7 tie 4 25.4 bars 12.7 tie
12.7 tie
8 25.4 bars 8 25.4 bars
A A
A A
A A B
B B

Section A-A and B-B Section A-A Section B-B Section A-A Section B-B

12.7 steel ties 12.7 steel ties 508 mm 12.7 steel ties
508 mm 508 mm
@ 50.8 mm on center @ 50.8 mm on center @ 50.8 mm on center

12.7 steel ties 12.7 steel ties 3048 mm 12.7 steel ties
3048 mm 10 ft mm
3048
@ 406 mm on center @ 406 mm on center @ 406 mm on center

12.7 steel ties

12.7 steel ties 508 mm 12.7 steel ties
508 mm @ 50.8 mm on center 508 mm
@ 50.8 mm on center @ 50.8 mm on center

737 mm 508 mm
610 mm 508 mm 356 mm
14in.

a) b) c)

Fig. 1. Reinforcement layout for series 共a兲 S-1; 共b兲 R-1; and 共c兲 R-0.5

men was kept as-built and used as benchmark. For the first series
共S-1兲, three specimens were confined with both the glass FRP
systems 共Types A and B兲 and with the HFRP system; for the
second series 共R-1兲, one specimen was confined with type A glass
FRP and the other with the HFRP; for the third series 共R-0.5兲, one
specimen was confined with Type A glass FRP, two specimens
with Type B glass FRP 共with different number of plies兲, and one
specimen with the HFRP system.
A three-part denomination is used to identify each specimen.
The first part identifies the cross-sectional geometry: “S” stands
for square 共shape factor of 1.0兲 and “R” for rectangular 共shape
factor of either 1.45 or 1.43兲. The digits of the second part indi-
cate the volume factor 共1 or 0.5兲. The third part identifies type and
number of plies: GA for Type-A glass, GB for Type-B glass and H
for hybrid, with 2, 5, or 8 plies.
Specimen Design
The column specimens were designed using the ACI 318-63
code-mandated minimum amount of longitudinal reinforcement
and minimum tie area at maximum spacing. ACI 318-63 requires
that the total area of longitudinal bars be larger than 1.0% of the
gross section area, Ag; and that the vertical spacing of the ties be
the smallest of 16 longitudinal bar diameters 共to prevent bar buck-
ling兲, 48 tie diameters 共to ensure sufficient tie area to restrain the
lateral displacement of the longitudinal bars兲, and the least lateral
dimension of the column 共to develop the maximum strength of
the concrete core兲. Figs. 1共a–c兲 shows the reinforcement layout
and the cross section layout for each series. The total cross-
sectional area of the longitudinal bars was kept at 1.0% using
eight 25.4-mm diameter 共No. 8兲 bars for Series S-1 and R-1, and
four 25.4-mm diameter 共No. 8兲 bars for Series R-0.5. For Series
S-1 and R-1, 12.7-mm diameter 共No. 4兲 ties were used, spaced at
406 mm 共16 in.兲 on-center, which corresponds to the requirement
to prevent bar buckling. For Series R-0.5, 12.7-mm diameter 共No.
4兲 ties were spaced at 356 mm 共14 in.兲 on-center, which corre-
sponds to the requirement to develop the maximum strength of
the concrete core. The specimens’ dimensions were also selected
to ignore slenderness effects. Steel ties with 12.7-mm diameter
共No. 4兲 spaced at 50.8 mm on-center 共2 in.兲 were used at the two
ends of the specimens to prevent failure in these zones.
The FRP plies were applied by manual layup in the transverse
direction. Prior to the application of the FRP, all corners were
rounded with a radius of about 25.4 mm 共1 in.兲. The number of
FRP plies was five in the case of the Type A glass fiber sheets,
two in the case of the Type B glass fiber sheets, and eight in the
case of the hybrid glass-basalt fiber sheets. Given that the fiber
types were all of comparable quality, the number of plies was
designed in order to have the same FRP volume ratio for all
column specimens, with the exception of Specimen R-0.5-5GB.
In typical field applications, the number of plies ranges between 3
and 6 in the case of 600 g per square meter 共1.1 lb/sq yd兲 yield,
and between 2 and 3 in the case of 900 g per square meter 共1.7
lb/sq yd兲 yield. Specimen R-0.5-5GB was designed to have a data
point at a high FRP confinement ratio.
Materials
The specimens were fabricated at a precast plant one at a time
using the same concrete mix design. The average concrete com-
pressive strength, f ⬘c , is reported in Table 2. Concrete strength is
based on the results of compression tests on 150-mm diameter by
304 mm 共6 by 12 in.兲 cylinder samples, 100-mm diameter by 202
mm 共4 by 8 in.兲 cylinder samples, or 92.7 mm diameter by 114
mm 共3.71 by 7.50 in.兲 core samples, per ASTM C39. For Speci-

114 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Table 2. Concrete Strength applied. ASTM Grade 60 steel bars and ties were used for all
Concrete sample Average specimens. Unidirectional continuous fiber sheets were used for
Specimen 共diameter and height Number strength, the FRP systems, where the properties of the fiber sheets as pro-
code are in millimeters兲 of tests f c⬘ 共MPa兲 vided by the manufacturers are summarized in Table 3.
S-1-control 152 by 304 6 37.3
S-1-5GA 152 by 304 3 48.6 Test Setup and Procedure
S-1-2GB 94.2 by 190 共core兲 3 37.1
The instrumentation in all the specimens consists of electrical
S-1-8H 152 by 304 3 44.4
strain gauges located on the longitudinal and transverse steel re-
R-1-control 152 by 304 3 48.0 inforcement at the level of the midheight cross section, and on
R-1-5GA 94.2 by 190 共core兲 3 56.4 the external surface of the specimen: onto the concrete surface at
R-1-8H 94.2 by 190 共core兲 3 47.6 the midheight section for the control specimens, and on the FRP
R-0.5-control 102 by 203 3 34.7 jacket at critical locations 共corner areas and midsection on
R-0.5-5GA 102 by 203 3 53.8 each face of the prismatic specimens兲 along the perimeter of the
R-0.5-2GB 102 by 203 6 46.4 cross section at midheight of the strengthened specimens. Addi-
R-0.5-5GB 102 by 203 6 49.7 tionally, LVDT sensors were used to measure the vertical
R-0.5-8H 102 by 203 3 46.8 displacement of the specimen, and to evaluate the horizontal 共in-
Note: 6.895 MPa= 1,000 psi; 25.4 mm= 1 in. plane兲 dilation at the midheight cross section, along the two
sides and the two diagonal directions. The stroke of the LVDT
ranged between ⫾12.25 mm 共⫾0.5 in.兲 and ⫾490 mm 共⫾20 in.兲.
mens S-1-2GB, R-1-5GA, and R-1-8H, concrete cylinder samples The control specimens were tested using a 22.2 MN
were delivered to the laboratory at different times. This could 共5 million lbforce兲 testing machine. The tests of the strengthened
have caused inconsistent results; therefore, concrete cores were specimens were conducted using a 53.4 MN 共12 million lbforce兲
taken from the cast specimens and used to define concrete testing machine. The control specimens and the strengthened ones
strength. According to ASTM C39, if the concrete core length to were tested, respectively, 5 and 18 months after casting. Special
diameter ratio is higher than 1.75, the concrete compressive care was taken such that each specimen was plumb and centered
strength can be taken as is and no correction factor has to be with respect to the cross head of the machine. A thin layer of
high-strength grouting paste was cast onto the base platen 共below
the specimen兲 and another one was cast on the top surface of the
Table 3. FRP System Properties specimen. The load was applied concentrically with a displace-
Filament yarn Glass A Glass B Hybrid ment control rate of 0.5 mm/min 共0.02 in./min兲. The same rate
properties fabric fabric fabric was maintained when the sample concrete cylinders and cores
were tested. The loading sequence included five load cycles, each
Type of fibers Glass Glass Basalt-glass
of which was repeated once, with increments of one fifth of the
Ratio in volume 100% 100% 33.3–66.6%
expected capacity.
Tensile modulus 共MPa兲 76,948 72,397 88,945–76,948
Tensile strength 共MPa兲 3,399 3,241 4,840–3,399
Tensile strain 共%兲 4.7 4.5 3.15–4.7
Experimental Results and Discussion
Sheet Glass A Glass B Hybrid
properties fabric fabric fabric
The test results are summarized in Table 4. For each specimen,
Ply thickness 共mm兲 0.246 共0.480兲 0.589 共1.27兲 0.120 共0.284兲 the following is reported: average concrete compressive strength,
Weight 共g / m2兲 600 900 340 f ⬘c 共as defined in the section “Materials”兲; maximum load applied,
Note: 6.895 MPa= 1 ksi; 25.4 mm= 1 in.; 0.542 g / m2 = 1 lb/ yd2; gross Ppeak; load at failure, Pu, 共defined as the load corresponding to
laminate properties in square brackets. 75% of the peak load兲; ratio between load at failure and maxi-

Table 4. Test Results

f c⬘ Ppeak Pu Pu ⌬peak ⌬u ⌬u ␴c,peak ␴c,peak / f c⬘

Specimen code 共MPa兲 共MN兲 共MN兲 Ppeak 共mm兲 共mm兲 ⌬peak f c⬘ 共␴c,peak / f c⬘兲control
S-1-control 37.3 12.5 9.87 0.789 6.53 8.94 1.37 0.779 1.00
S-1-5GA 48.6 17.8 N/A N/A 7.01 N/A N/A 0.893 1.15
S-1-2GB 37.1 13.0 9.79 0.753 9.07 15.93 1.76 0.821 1.05
S-1-8H 44.4 15.6 12.1 0.775 6.10 28.1 4.61 0.843 1.09
R-1-control 47.9 16.5 16.3 0.987 7.09 7.72 1.09 0.826 1.00
R-1-8H 56.4 17.0 12.8 0.750 8.64 14.63 1.69 0.870 1.05
R-0.5-control 34.7 5.52 5.05 0.915 8.00 8.38 1.06 0.746 1.00
R-0.5-5GA 53.8 8.66 6.49 0.750 7.70 15.32 1.99 0.847 1.13
R-0.5-2GB 46.4 7.11 5.36 0.754 6.83 18.21 2.66 0.748 1.00
R-0.5-5GB 49.7 8.73 6.86 0.785 9.14 30.73 3.36 0.879 1.18
R-0.5-8H 46.8 7.97 6.04 0.758 8.00 27.69 3.46 0.844 1.13
Note: N/A= not applicable; 6.895 MPa= 1,000 psi; 4.448 MN= 1,000 kip; 25.4 mm= 1 in.; and ␴c,peak = 共Ppeak − Pbar兲 / Ac.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 115

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 Axial strain (%) Axial strain (%)
0 0.167 0.333 0.500 0.667 0.833 1.00 1.17 1.33 0 0.167 0.333 0.500 0.667 0.833 1.00 1.17 1.33
1 1

Normalized concrete axial stress

Normalized concrete axial stress

S-1-5GA 610 mm 508 mm

S-1-2GB
0.8 610 mm 0.8 736 mm

0.6 0.6
S-1-control S-1-8H R-1-control R-1-8H

0.4 0.4

0.2 0.2

0 0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Axial deformation (mm) Axial deformation (mm)

Fig. 2. Normalized axial stress versus axial deformation 共Series S-1兲. Fig. 3. Normalized axial stress versus axial deformation 共Series R-1兲.
Failed Specimens S-1-control and S-1-8H are also shown. Failed Specimens R-1-control and R-1-8H are also shown.

mum load, Pu / Ppeak; axial deformation when the maximum load

共peak兲 was reached, ⌬peak; ultimate axial deformation, ⌬u 共defined
as the axial deformation corresponding to Pu兲; ratio between ul- mented in Fig. 2. The confined specimens behaved similarly with
timate axial deformation and axial deformation at peak, ⌬u / ⌬peak; respect to each other: upon attaining the peak load, the load
concrete axial stress at peak, ␴c,peak, normalized with respect to steadily decreased while the axial deformation continued increas-
the average concrete compressive strength, f ⬘c , 共␴c,peak / f ⬘c 兲; ratio ing due to the confining action of the FRP wrap. Failure occurred
between normalized concrete axial stress at peak and correspond- by rupture of the FRP laminates. In particular, fiber rupture al-
ing value of the control specimen for the reference series, ways initiated in the proximity of a corner and then propagated
共␴peak / f ⬘c 兲 / 共␴peak / f ⬘c 兲control. The axial deformation is rendered as toward the sides. The average concrete axial peak stress, ␴c,peak
the average of the measurements from the LVDTs. The concrete ranged between 82 and 89% of the average concrete compressive
axial stress at peak, ␴c,peak, is computed as the difference between strength, f c⬘, when the peak capacity was reached. The increment
the peak load 共Ppeak兲 and the load carried by the reinforcement in concrete strength due to confinement was 15, 9, and 5% for
共Pbar兲, divided by the net area of concrete 共Ac兲, where Pbar is S-1-5GA, S-1-8H, and S-1-2GB, respectively. An important im-
given as the total area of reinforcing steel 共As兲 times the nominal provement in deformability in the postpeak behavior was experi-
yield stress 共f y兲. enced by the confined specimens. The ultimate axial deformation
recorded, ⌬u, was about 461 and 176% of ⌬peak for S-1-8H and
S-1-2GB, respectively. While testing Specimen S-1-5GA, a prob-
Strength and Failure Modes
lem on the data acquisition system occurred that caused the loss
Failure of the control specimens initiated with vertical cracks fol- of the data postpeak. Fig. 2 also shows the failure of the FRP
lowed, first, by lateral displacement of the longitudinal bars that wrap in Specimen S-1-8H.
contributed to the splitting of the concrete cover and, finally, by
crushing of the concrete core and buckling of the longitudinal Series R-1
bars. All FRP-confined columns failed due to rupture of the FRP Fig. 3 shows the normalized concrete axial stress versus axial
jacket; cracking of the concrete core developed after the maxi- deformation plot for the specimens of Series R-1. Failure of the
mum load was attained, and longitudinal bar buckling was visible benchmark specimen, R-1-control, was sudden and accompanied
after the postmortem removal of the ruptured FRP jacket and by an explosive noise. The peak capacity was attained at an av-
concrete cover. erage concrete axial stress equal to about 83% of f ⬘c . Once the
peak load was attained, the load dropped almost instantly without
Series S-1 early warning, as no cracking of the concrete was observed until
Fig. 2 plots the normalized concrete axial stress 共defined as the the final crushing. Failure occurred when the load was about 99%
ratio between ␴c,peak and f ⬘c 兲 with respect to the axial deformation of Ppeak. The ultimate axial deformation recorded, ⌬u, was only
of the specimens of Series S-1. The failure of the benchmark about 109% of ⌬peak. Failure occurred at the lower-half of the
Specimen S-1-control, was brittle and occurred at the center of specimen, as shown in Fig. 3. Specimen R-1-5GA experienced a
the upper-half of the specimen. The peak capacity was attained premature failure localized at the top of the specimen due to stress
when the average concrete axial stress was equal to about 78% of concentration and has not been taken into account in this study. A
the average concrete compressive strength, f ⬘c . The load stabilized 102 mm 共4 in.兲 wide strip at the top end of Specimen R-1-5GA
at the level of the peak load before it suddenly dropped. Cracking was left unconfined; the splitting of the concrete cover in this
of the concrete was observed before splitting of the concrete region caused the premature rupture of the FRP jacket. Failure of
cover and buckling of the longitudinal bars. The ultimate axial Specimen R-1-8H occurred due to rupture of the FRP laminate.
deformation recorded, ⌬u, was about 137% the value at peak load, Failure started at one of the corners at the higher-half of the
⌬peak, while the load dropped to 78% of Ppeak. Crushing of the specimen and then expanded to the adjacent sides 共Fig. 3兲. The
concrete core and buckling of the longitudinal bars are docu- peak load was reached when the average concrete axial stress was

116 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 Axial strain (%)
0 0.167 0.333 0.500 0.667
1 Strain gauges were attached on the longitudinal steel bars at the
Normalized concrete axial stress
0.8
R-0.5-5GA
0.6
R-0.5-2GB
R-0.5-control
0.4 R-0.5-8H
0.2
0 the confining action provided by the FRP jacket delayed the col-
0 5 10 15 20
Axial deformation (mm)
Fig. 4. Normalized axial stress versus axial deformation 共Series
R-0.5兲. Failed Specimens R-0.5-control, R-0.5-5GB, and R-0.5-8H
are also shown.
about 86% of f ⬘c . The increment of concrete strength due to con-
finement was about 5%, while the ultimate axial strain was about
169% of ⌬peak.
Series R-0.5
The normalized concrete axial stress versus axial deformation
plot for the specimens of Series R-0.5 is shown in Fig. 4. Speci-
men R-0.5-control failed similarly to Specimen S-1-control. The
failure affected the entire upper-half of the column. Concrete
cracking was heard before the concrete cover spalled and the steel
bars buckled. The peak load was reached when the average con-
crete axial stress was about 75% of f ⬘c . The failure was brittle
共failure occurred when the load was about 92% of Ppeak兲, with a
measured ultimate axial strain of 106% of ⌬peak. Specimens
R-0.5-5GA, R-0.5-8H and R-0.5-5GB behaved similarly to each
other. The average concrete axial stress ranged between 84 and
88% of f ⬘c when the peak load was reached. The increment in
concrete strength was 13% for both R-0.5-5GA and R-0.5-8H,
and 18% for R-0.5-5GB, as result of a higher FRP confinement
ratio. After the peak load was attained, for both R-0.5-5GA and
R-0.5-8H, the load gradually decreased with increasing axial de-
formations. R-0.5-5GA failed when the axial deformation almost
doubled the value at peak. In the case of specimen R-0.5-8H,
instead, the load stabilized at about 80% of the peak load, and the
specimen failed when the axial deformation was about 346% of
⌬peak. For Specimen R-0.5-5GB, the load remained nearly con-
stant after reaching the peak load, with increasing axial deforma-
tions; then, the load decreased quite suddenly and stabilized at
about 90% of the peak load. The specimen failed when the axial
load was about 78% of Ppeak and the axial deformation was about
336% of ⌬peak. The failure mode was similar to that of the speci-
mens of Series S-1 and R-1, as documented in Fig. 4 for Speci-
mens R-0.5-5GB and R-0.5-8H. Specimen R-0.5-2GB did not
experience any gain in concrete strength. The peak load was
reached when the average concrete axial stress was about 75% of
f ⬘c and the ultimate axial deformation was about 266% of ⌬peak.
Factors contributing to this strength result may include those af-
fected by preparation, setup, and execution of the test itself.
Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
Vertical Strain in Reinforcing Bars
0.833 1.00 1.17 1.33
508 mm level of the midheight cross section at the middistance between
356 mm
two adjacent ties. The axial strain in the bars recorded for the
unstrengthened specimens is either close to or exceeds the yield
strain and is slightly higher than that measured in the confined
specimens. This can be explained as follows. In the case of un-
confined specimens, the strain gauges stopped reading immedi-
R-0.5-5GB ately after the concrete cover split since they were located at the
interface between bar and concrete cover. The compressive strain
recorded on the steel bars by the strain gauges was not affected by
the lateral displacement of the bars since concrete failure and bar
buckling happened concurrently. For confined specimens, instead,
25 30 35 40
umn failure and the strain gauges continued reading after concrete
cover splitting. Strain readings were affected by the lateral deflec-
tion of the steel bars. The compressive strain was therefore re-
duced by the tensile strains induced by the lateral bending of the
bars since the strain gauges were located on the exterior side of
the rebars. Experimental measurements were not used for the
analysis, and, given also the small percentage of steel reinforce-
ment, the assumption of yielded steel is considered reasonable for
analysis purposes.
Volumetric Response
The impact of the external confinement provided by the FRP
jacket on column strength enhancement and postpeak deformabil-
ity is analyzed with respect to the volumetric response. It has
been widely recognized 共Park and Paulay 1975; Pantazopoulou
and Mills, 1995; Neville 1996; Spoelstra and Monti 1999兲 that
axially loaded unconfined concrete contracts in volume up to
about 90% of its peak strength; then, the direction of the volume
change reverses and results in dilation when the peak strength is
reached. Beyond this point, the volume expansion grows at a
higher rate as the softening branch develops until failure occurs.
When concrete is confined with FRP jackets, the volume expan-
sion may be effectively constrained, and the unstable crack
growth controlled. Provided that the jacket is sufficiently thick,
concrete expansion can be curtailed 共Mirmiran et al. 1998; Spoel-
stra and Monti 1999; Pessiki et al. 2001; Harries and Kharel
2002; Carey and Harries 2005; Lam and Teng 2003a,b兲. Plain
concrete dilation ratio 共defined as the ratio of transverse to axial
strain兲 has an initial value 共Poisson’s ratio兲 generally found to be
about 0.20, begins to increase nonlinearly when concrete starts
cracking, and grows indefinitely until failure. The presence of the
FRP jacket affects the concrete dilation ratio before and after the
concrete reaches its peak capacity by significantly restraining the
growth in volume and allowing large axial deformations 共Mirmi-
ran et al. 1998; Pessiki et al. 2001兲.
Herein, the axial strain was measured as the LVDT-measured
axial deformation averaged along the entire height of the column.
The volumetric strain 共change in volume per unit volume of con-
crete兲 is calculated as the sum of the axial strain and the two
transverse strains at the mid-height cross section 共measured along
the two orthogonal directions in the plane of the cross section兲. A
positive volumetric strain indicates volume reduction, whereas a
negative value indicates expansion. The volumetric strain repre-
sents a parameter indicative of the response of the cross section in
its entirety, unlike the dilation ratio which indicates how the cross
section tends to deform along different directions. The dilation
ratio is defined as the ratio between the average transverse strain
JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 117
 0.005
S-1-control R-0.5-control

Volumetric strain
(1-2v) (1-2v)
0.000
S-1-2GB R-0.5-GB

-0.005
S-1-8H
R-0.5-8H
-0.010

-0.015 a) b)
2.5
2.500
R-0.5-8H
2.0002
Dilation ratio

R-0.5-5GB
1.500
1.5
S-1-8H
1.0001
S-1-2GB
0.5
0.500
R-0.5-control
S-1-control
0.0000 c) d)
2.5
2.500
0.5 0.5
S-1-control R-0.5-control
0.4 0.4
2.0002
Dilation ratio

0.3 0.3
0.2 0.2
1.500
1.5
0.1 0.1
0.0 0.0
1.0001
0 0.001 0.002 0.003 0 0.001 0.002 0.003

0.5 1 Transverse 1 Transverse 1

0.500
2 1
2 2
0.0000 2
e) f)
2.5
2.500
1
S-1-8H 1 R-0.5-8H 1
2.0002 2
Dilation ratio

3 4 Transverse 3
1.500
1.5
Transverse 3
1.0001
Diagonal
1
0.5
0.500 3
Diagonal 4 Transverse
2
0.0000
0 0.002 0.004 0.006 0.008 0.01 0.0 0 0.002 0.004 0.006 0.008 0.01 0.01

Axial strain Axial strain

g) h)

Fig. 5. Volumetric strain⫺axial strain and dilation ratio⫺axial strain relationships: 关共a兲, 共c兲, 共e兲, and 共g兲兴 Series S-1; 关共b兲, 共d兲, 共f兲, and 共h兲兴 R-0.5

along each of the two orthogonal directions in the plane of the
cross section and the axial strain. The midheight cross section is
less affected by the boundary conditions and, even though failure
may occur elsewhere, it is taken as representative of the behavior
of the column.
Series S-1
Figs. 5共a and c兲 show the volumetric strain⫺axial strain response
and the dilation ratio⫺axial strain response, respectively, of the
specimens of Series S-1 with the exception of Specimen S-1-
5GA, for which the data for the postpeak branch were not avail-
able. The initial slope of all the curves in Fig. 5共a兲 is close to
共1 – 2␯兲 共where ␯ = Poisson’s ratio of the concrete assumed equal
to 0.20兲, which corresponds to the elastic condition. The curves
deviate from this line and reach their maximum 共point of reversal
in volumetric strain兲 as the load approaches its peak value. This
point corresponds to the onset of uncontrolled crack growth lead-
ing to failure in the case of the control specimen 共S-1-control兲,
whose postpeak branch has a limited extent and rapidly develops
into failure. In the case of the externally confined specimens 共S-
1-2GB and S-1-8H兲, the larger development of the postpeak
branch clearly shows that the external FRP jacket provides a lat-
eral constraint for the cracked concrete and reverses the dilation
process of the concrete. In Fig. 5共c兲, the dilation ratio for the
control specimen has an average value of about 0.2 through axial
strains up to 0.0028, past which it rapidly increases. No readings
are available in the postpeak zone because the loss of confine-
ment, crushing of concrete and buckling of longitudinal bars oc-
curred almost instantaneously. The dilation ratio for Specimens
S-1-2GB and S-1-8H 关Fig. 5共c兲兴 ranges between 0.15 and 0.25 up
to axial strains of about 0.002. Past this level, the dilation ratio for

118 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
both specimens increases rapidly and reaches a limit value of 305 mm 305 mm
about 1.5 when the axial strain is about 0.0035. Beyond this point, 305 mm 305 mm
both curves begin to decrease almost with the same slope. Speci-
men S-1-2GB failed when the axial strain was about 0.005 and
the dilation ratio dropped to about 0.8. Specimen S-1-8H reached
an axial strain at failure of about 0.009 with a dilation ratio drop-
ping to about 0.5. Fig. 5共e兲 shows the plot of the dilation ratio of
Specimen S-1-control measured along the two transverse direc-
tions, whereas Fig. 5共g兲 shows the dilation ratio of Specimen
a) b)
S-1-8H measured along one of the two transverse directions and
along the two diagonals. In case of the control specimen, the
dilation ratio along both transverse directions is similar, that is, 254 mm 254 mm
the cross section deforms symmetrically along these axes. For 178 mm 178 mm
Specimen S-1-8H, after reaching its peak capacity, the concrete
core tends to expand much more along the transverse directions
rather than along the two diagonals. As expected for a square
column, the dilation ratio is symmetrical along the two diagonals.

Series R-0.5
To describe the volumetric response of specimens of Series R-0.5,
Specimens R-0.5-control, R-0.5-5GB, and R-0.5-8H were se- c) d)
lected. Figs. 5共b and d兲 show the volumetric strain-axial strain
and the dilation ratio-axial strain relations, respectively. The volu-
metric response of Specimen R-0.5-control is similar to that ex- Fig. 6. Change in volume of representative one-quarter unit element
perienced by Specimen S-1-control. In the case of the confined for Specimen S-1-8H at 共a兲 peak load; 共b兲 failure; and Specimen
column specimens, the FRP jacket constrains the volume dilation R-0.5-8H at 共c兲 peak load; and 共d兲 failure
of the concrete core, but its effectiveness is not sufficient to re-
verse the volumetric expansion as for the square columns. For
corner is smaller, but comparable to the transverse ones. For the
Specimen R-0.5-5GB, given the higher FRP amount compared to
square column, the transverse expansion at failure is significantly
Specimen R-0.5-8H, a reversal point is reached but not main-
higher than the diagonal one 关Fig. 6共b兲兴. This experimental obser-
tained. The dilation ratio for the control specimen has a constant
vation confirms the generally accepted assumption that in pris-
average value of about 0.30 until failure occurs 关Fig. 5共d兲兴. The
matic cross sections the confining pressure is higher at the corners
dilation ratio for Specimens R-0.5-5GB and R-0.5-8H 关Fig. 5共d兲兴
than along the flat sides. For the rectangular column under failure
ranges between 0.30 and 0.45 up to axial strains of about 0.0015.
load 关Fig. 6共d兲兴, instead, the displacements of the midpoint on the
Past this level, the dilation ratio for both specimens increases at a
long side and of the corner are of about the same magnitude and
high rate until a value of about 2 at axial strains of about 0.003.
much larger than the displacement of the midpoint on the short
Beyond this point, the dilation ratio continues to increase at a low
side, respectively. The commonly accepted assumption is that the
rate until failure. Specimen R-0.5-5GB failed when the axial
area of effective confinement is defined by four parabolas within
strain was about 0.009 and the dilation ratio was about 2.1. Speci-
which the concrete is fully confined 共and outside of which negli-
men R-0.5-8H reached an axial strain at failure of about 0.008
gible confinement occurs兲. Based on this experimental evidence,
with a dilation ratio close to 2.3, whereas its square counterpart
it appears that the two parabolas along the short sides may be
共S-1-8H兲 reached an axial strain at failure of about 0.009 with a
disregarded, since the dilation in the direction orthogonal to the
dilation ratio close to 0.5. Fig. 5共f兲 plots the dilation ratio of
short side is small. As the rectangular column has more FRP than
Specimen R-0.5-control measured along the two transverse direc-
the square one, and yet the jacket is less effective, the lower level
tions, while Fig. 5共h兲 illustrates the dilation ratio for Specimen
of performance has to be attributed to the cross-sectional shape.
R-0.5-8H measured along the two transverse directions 共short and
Moreover, it appears that for a prismatic cross section, the incre-
long side兲 and along one of the two diagonals. The dilation ratio
ment in concrete strength is only possible when concrete contracts
along the long side is much smaller than the one along the short
in volume. Fig. 7 shows a typical axial stress⫺volumetric strain
side for both the control and confined specimens. In the case of
relationship for FRP-confined prismatic columns 共Specimen S-1-
the control specimen, the development of unstable cracks makes
8H兲. The FRP confinement allows the reversal in volume change,
the long side dilation ratio to rapidly increase until failure.
and, during the process of volume dilation, concrete experiences
Figs. 6共a–d兲 shows the change in volume of a representative
large axial deformations without significantly increasing its axial
one-quarter unit element of Specimens S-1-8H and R-0.5-8H, re-
strength.
spectively, when the peak load is reached 共a and c兲 and at failure
共b and d兲 by plotting the displacements of the midpoints of the
two sides and of the corner and the axial shortening. To make the Review of Existing Constitutive Models
change in volume visible, all changes in length are amplified for FRP-Confined Prismatic Columns
using the same magnification factor. Given the symmetry, only
one-quarter of the cross section is drawn. For both square and
Theoretical Prediction of Increment in Concrete
rectangular shapes, when the peak load of the column is reached
Strength
no significant cross-sectional shape change is noted 关Figs. 6共a and
c兲兴, namely: the displacements of the midpoints of the two sides The increment of concrete strength achieved by each specimen is
are of the same order of magnitude and the displacement of the compared with the theoretical prediction given by the analytical

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 119

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 1.25 2
Normalized axial stress Peak load
1.75
1
1.5
0.75
1.25
  cc    cc 
0.5     1
 c THEOR  c  EXP

Campione and Miraglia

Campione and Miraglia

Campione and Miraglia

Volume Volume
0.75

Wang and Restrepo

Wang and Restrepo

Wang and Restrepo

0.25 dilation contraction
0.5

Kumutha et al.

Kumutha et al.

Kumutha et al.
Mirmiran et al.

Mirmiran et al.

Mirmiran et al.
Lam and Teng

Lam and Teng

Lam and Teng

Wu and Wang

Wu and Wang
Wu and Wang
0
-0.01 -0.008 -0.005 -0.003 0 0.0025 0.005 0.25

Volumetric strain
Fig. 7. Typical volumetric strain⫺normalized axial stress relation for
FRP-confined prismatic concrete column 共Specimen S-1-8H兲
models proposed by Mirmiran et al. 共1998兲, Wang and Restrepo
共2001兲, Campione and Miraglia 共2003兲, Lam and Teng 共2003b兲,
Kumutha et al. 共2007兲, and Wu and Wang 共2009兲. The experimen-
tal and theoretical values of the concrete axial stress at peak are
shown in Tables 5 and 6. The ratios between the theoretical pre-
dictions and the experimental values are plotted in Figs. 8 and 9.
The ordinates of the histograms represent the following:
0
S-1-5GA S-1-2GB S-1-8H
Fig. 8. Predicted concrete axial strength enhancement for square col-
umns based on f ⬘c 共top arrows兲 and strength of control column 共bot-
tom arrows兲
study as f ⬘c was found to be about 20% higher than the axial
concrete strength of an as-built column. Based on research carried
out at Lehigh University and the University of Illinois in the early
1930s 共Slater and Lyse 1931a,b,c; Lyse and Kreidler 1932; Lyse
1933; Richart 1933; Richart and Brown 1934兲, the maximum con-

冉 冊 冒冉 冊
␴cc
␴c THEOR
␴cc
␴c EXP
共1兲
crete stress in columns was found to be about 85% of the com-
pressive strength of a 152.4 diameter by 308.4 mm 共6 by 12 in.兲
concrete cylinder, at a strain where the yield stress of the rein-
where 共␴cc / ␴c兲THEOR is the theoretical ratio between the predicted forcing steel, f y, was attained. This result is in agreement with the
confined concrete strength and the relative concrete cylinder experimental values found for the control specimens and shown
strength, f ⬘c , and 共␴cc / ␴c兲EXP is the experimental ratio between the in Table 2 共second column from the right兲. This issue had never
confined concrete strength and the unconfined concrete strength. been addressed before because the models base their predictions
When the ratio 共␴cc / ␴c兲EXP is computed as ␴peak / f c⬘ 共second col- on scaled cylinders with height-to-diameter ratio equal to 2,
umn from the right in Table 4兲, the experimental results are over- whereas the work presented herein refers to full-scale elements
estimated by the models. Conversely, the predictions tend to with a height-side ratio equal or larger than 5, representative of
slightly underestimate the experimental results if the ratio building columns 共Blanks and MacNamara 1935兲.
共␴cc / ␴c兲EXP is computed as 共␴peak / f ⬘c 兲 / 共␴peak / f c⬘兲control 共first col- The following discussion is limited to the comparison of the
umn from the right in Table 4兲. In Figs. 8 and 9 the arrows theoretical models with the predictions based on the strength of
highlight the difference between the comparison of the theoretical the control column. As can be noted from the histograms reported
predictions with the experimental results based on the cylinder in Fig. 8, Mirmiran et al.’s, Wang and Restrepo’s, Campione and
strength, f ⬘c , and the same comparison based on the strength of Miraglia’s, and Wu and Wang’s models tend to underestimate the
control column. All models rely on the assumption that the con- increment in concrete strength for the square column specimens.
crete strength of an as-built unconfined column is equal to that of Kumutha et al.’s model overestimates the experimental results.
a control cylinder, f c⬘. This assumption was not verified in this Lam and Teng’s and Wang and Restrepo’s models are also accu-
rate in predicting the concrete strength enhancement for the rect-
angular column specimens 共Fig. 9兲.
Table 5. Experimental and Theoretical Concrete Axial Stress at Peak
共Square Columns兲
Lateral Pressure and Strain in FRP
Models S-1-5GA S-1-2GB S-1-8H
The definition of the lateral confining pressure exerted by the FRP
Experimental 43.4 30.5 37.4 jacket on the concrete core is key to the prediction of the incre-
Mirmiran et al. 共1998兲 52.0 39.0 48.0 ment in concrete strength 共Jiang and Teng 2007, Teng and Lam
Wang and Restrepo 共2001兲 50.7 40.1 47.1 2004, Teng et al. 2007, and Teng et al. 2009兲. The lateral pressure
Campione and Miraglia 共2003兲 50.7 39.0 46.1 depends on the strain in the FRP and varies depending on the
Lam and Teng 共2003b兲 56.3 38.7 45.6 cross-sectional shape 共circular or prismatic兲. For circular cross
Kumutha et al. 共2007兲 62.2 43.7 48.5 sections, the lateral pressure is “ideally” uniformly distributed
Wu and Wang 共2009兲 52.0 48.8 51.8 around the perimeter and, therefore, it can be derived based on

Table 6. Experimental and Theoretical Concrete Axial Stress at Peak 共Rectangular Columns兲
Models R-1-8H R-0.5-5GA R-0.5-2GB R-0.5-5GB R-0.5-8H
Experimental 41.1 45.6 34.7 43.7 39.5
Wang and Restrepo 共2001兲 49.3 57.0 49.0 56.5 49.2
Lam and Teng 共2003b兲 49.5 59.7 51.1 61.6 49.8
Kumutha et al. 共2007兲 54.7 75.0 63.5 91.9 57.6

120 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 2.5 tions the dilation ratio is smaller along the diagonals than
2.25
along the two transverse directions, whereas for rectangular
2
1.75 cross sections it is high if measured along the short trans-
  cc    cc 
1.5 verse direction and along the diagonals, and low along the
    1.25 long transverse direction.
 c THEOR  c  EXP
1 Wang&Restrepo 3. Existing semiempirical prediction models do not converge to

Wang&Restrepo

Wang&Restrepo

Wang&Restrepo

Wang&Restrepo
0.75
Kumutha et al.

Kumutha et al.

Kumutha et al.

Kumutha et al.

Kumutha et al.
the same predictions for the ultimate axial capacity of full-
Lam & Teng

Lam & Teng

Lam & Teng

Lam & Teng

Lam & Teng

0.5
0.25 scale FRP-confined prismatic concrete columns. The incre-
0 ment in strength due to the FRP confinement is based on the
strength of a control concrete cylinder, f c⬘, rather than the
concrete strength of the as-built unconfined column. The
maximum stress reached by concrete in columns is only
Fig. 9. Predicted concrete axial strength enhancement for rectangular about 85% of the cylinder strength, therefore, theoretical
columns based on f ⬘c 共top arrows兲 and strength of control column models should use 0.85f c⬘, not just f ⬘c .
共bottom arrows兲 4. Difference in the FRP material manufacturers does not affect
performance when confining materials are of comparable
quality. The contribution to column confinement of the hy-
equilibrium considerations. For prismatic cross sections, the lat- brid glass-basalt FRP laminates was similar to that of the
eral pressure is not uniform along the perimeter: it is higher in GFRP laminates.
proximity of the corners and lower along the sides. However, in The condition of pure axial loads is uncommon for RC col-
the most commonly adopted confinement models for prismatic umns, which are typically subjected to compressive loads com-
columns, the lateral pressure is derived by using the same expres- bined with bending moments that may also be produced by lateral
sion used for circular cross sections, which is adjusted to account loads. More evidence is needed to evaluate the effects of realistic
for the different shape. It is widely acknowledged that for FRP- loading scenarios.
confined circular columns 共for commonly used numbers of FRP
plies兲, the final point in the stress-strain relationship defines both
the peak load and the ultimate axial strain. The peak stress is Acknowledgments
reached when the FRP ruptures, and the lateral confining pressure
is the maximum attained. Therefore, it is reasonable to express the The writers gratefully acknowledge the support of the NSF
lateral confining pressure in terms of the ultimate strain in the Industry/University Cooperative Research Center for “Repair of
FRP when failure occurs. On the contrary, as discussed in this Buildings and Bridges with Composites” 共RB2C兲 at the Univer-
paper and in agreement with earlier research 共Pessiki et al. 2001; sity of Miami, of the “REte dei Laboratori Universitari di Ingeg-
Shehata et al. 2002; Rocca et al. 2008兲, in the case of FRP- neria Sismica” 共RELUIS兲 at the University of Naples “Federico
confined prismatic concrete columns, while failure coincides with II,” of Mapei S.p.A., and of Fyfe Co. LLC. Special thanks are
rupture of the FRP, the peak capacity of the column occurs at an extended to Derek Schesser, undergraduate researcher at the Uni-
FRP strain that is significantly lower than its ultimate value. Con- versity of Miami; to the Fritz Engineering Laboratory at Lehigh
sequently, it does not appear logical to relate the lateral pressure University, in particular to Mr. Frank Stokes and Mr. Gene Mat-
to the ultimate strain in the FRP. lock; and to the Building and Fire Research Laboratory at the
National Institute for Standards and Technology 共NIST兲, in par-
ticular to Mr. Steven Cauffman and Mr. Frank Davis, for the
Conclusions assistance in planning and conducting the tests.

Based upon the experimental evidence gained through the full-

scale experiments presented in this paper, the following conclu- Notation
sions are drawn:
1. The axial load-axial deformation behavior of a prismatic The following symbols are used in this paper:
concrete column laterally confined by means of an FRP Ac ⫽ net area of concrete;
jacket of thickness representative of field applications is Ag ⫽ gross section area;
characterized by a linear elastic branch almost up to the peak As ⫽ total area of reinforcing steel;
load, and a descending postpeak branch until failure. In pris- f c⬘ ⫽ average cylinder compressive strength;
matic columns, the FRP confinement effectiveness is more f y ⫽ nominal steel yield stress;
significant in terms of enhancement of concrete axial defor- Pbar ⫽ load carried by the reinforcement;
mation rather than increment in axial strength. The presence Ppeak ⫽ maximum load applied;
of the FRP jacket allows a “growth” in volume of the con- Pu ⫽ load at failure;
crete core by offsetting buckling of the longitudinal bars and Pu / Ppeak ⫽ ratio between load at failure and maximum
by delaying unstable crack propagation. load;
2. The shape of the cross section influences the effectiveness of ⌬peak ⫽ axial deformation when the maximum load
the confinement. Effectiveness is higher for square shapes 共peak兲 was reached;
than for rectangular ones, and decreases as the side-aspect ⌬u ⫽ ultimate axial deformation;
ratio of a rectangular cross section increases. The transverse ⌬u / ⌬peak ⫽ ratio between ultimate axial deformation and
expansion of the concrete core in the plane of the cross sec- axial deformation at peak;
tion, defined by means of the dilation ratio, changes with the ␯ ⫽ Poisson’s ratio of the concrete 共assumed equal
direction along which it is evaluated. For square cross sec- to 0.20兲;

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 121

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
 ␴c ⫽ concrete axial stress;
␴c,peak ⫽ concrete axial stress at peak;
␴c / f ⬘c ⫽ normalized concrete axial stress;
␴c,peak / f ⬘c ⫽ normalized concrete axial stress at peak;
共␴cc / ␴c兲EXP ⫽ experimental ratio between the increment in
concrete strength and the unconfined concrete
strength;
共␴cc / ␴c兲THEOR ⫽ theoretical ratio between the confined
concrete strength and the unconfined
concrete strength; and
共␴peak / f ⬘c 兲 /
共␴peak / f ⬘c 兲control ⫽ ratio between normalized concrete axial
stress at peak and corresponding value of
the control specimen for the reference
series.
References
American Concrete Institute 共ACI兲. 共1963兲. “Building code requirements
for reinforced concrete.” ACI 318-63, Detroit.
Blanks, R. F., and McNamara, C. C. 共1935兲. “Mass concrete tests in large
cylinders.” Proc., ACI J., Paper JL31-12, American Concrete Institute,
New York, 280–303.
Campione, G., and Miraglia, N. 共2003兲. “Strength and strain capacities of
concrete compression members reinforced with FRP.” Cem. Concr.
Compos., 25, 31–41.
Carey, S. A., and Harries, K. A. 共2005兲. “Axial behavior and modeling of
confined small-, medium-, and large-scale circular sections with car-
bon fiber-reinforced polymer jackets.” ACI Struct. J., 102共4兲, 596–
604.
Fam, A. Z., and Rizkalla, S. H. 共2001兲. “Confinement model for axially
loaded concrete confined by circular fiber-reinforced polymer tubes.”
ACI Struct. J., 98共4兲, 451–461.
Fardis, M. N., and Khalili, H. 共1982兲. “FRP-encased concrete as a struc-
tural material.” Mag. Concrete Res., 34共121兲, 191–202.
Harajli, M. H. 共2006兲. “Axial stress-strain relationship for FRP confined
circular and rectangular concrete columns.” Cem. Concr. Compos.,
28共10兲, 938–948.
Harries, K. A., and Kharel, G. 共2002兲. “Behavior and modeling of con-
crete subject to variable confining pressure.” ACI Mater. J., 99共2兲,
180–189.
Hassan, M., and Chaallal, O. 共2007兲. “Fiber-reinforced polymer confined
rectangular columns: Assessment of models and design guidelines.”
J. Compos. Constr., 104共6兲, 693–702.
Jiang, T., and Teng, J. G. 共2007兲. “Analysis-oriented stress-strain models
for FRP-confined concrete.” Eng. Struct., 29, 2968–2986.
Karbhari, V. M., and Gao, Y. 共1997兲. “Composite jacketed concrete under
uniaxial compression—Verification of simple design equations.” J.
Mater. Civ. Eng., 9共4兲, 185–193.
Kumutha, R., Vaidyanathan, R., and Palanichamy, M. S. 共2007兲. “Behav-
iour of reinforced concrete rectangular columns strengthened using
GFRP.” Cement Concr. Compos., 29共8兲, 609–615.
Kurt, C. E. 共1978兲. “Concrete filled structural plastic columns.” J. Struct.
Div., 104共1兲, 55–63.
Lam, L., and Teng, J. G. 共2003a兲. “Design-oriented stress-strain model
for FRP-confined concrete.” Constr. Build. Mater., 17, 471–489.
Lam, L., and Teng, J. G. 共2003b兲. “Design-oriented stress-strain model
for FRP-confined concrete in rectangular columns.” J. Reinf. Plast.
Compos., 22共13兲, 1149–1186.
Lyse, I. 共1933兲. “Fifth report on column tests at Lehigh University.”
J. American Concrete Institute Proceedings, 29, 433–442.
Lyse, I., and Kreidler, L. 共1932兲. “Fourth progress report on column tests
at Lehigh University.” J. American Concrete Institute Proceedings,
28共1兲, 317–346.
Mander, J. B., Priestley, M. J. N., and Park, R. 共1988兲. “Theoretical
122 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011
Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org
stress-strain model for confined concrete.” J. Struct. Eng., 114共8兲,
1804–1826.
Matthys, S., Toutanji, H., Audenaert, K., and Taerwe, L. 共2005兲. “Axial
load behavior of large-scale columns confined with fiber-reinforced
polymer composites.” ACI Struct. J., 102共2兲, 258–267.
Matthys, S., Toutanji, H., and Taerwe, L. 共2006兲. “Stress-strain behavior
of large-scale circular columns confined with FRP composites.”
J. Struct. Eng., 132共1兲, 123–133.
Mirmiran, A., and Shahawy, M. 共1997兲. “Behavior of concrete columns
confined by fiber composite.” J. Struct. Eng., 123共5兲, 583–590.
Mirmiran, A., Shahawy, M., Samaan, M., Echary, H. E., Mastrapa, J. C.,
and Pico, O. 共1998兲. “Effect of column parameters on FRP-confined
concrete.” J. Compos. Constr., 2共4兲, 175–185.
Nanni, A., and Bradford, N. M. 共1995兲. “FRP jacketed concrete under
uniaxial compression.” Constr. Build. Mater., 9共2兲, 115–124.
Neville, A. 共1996兲. Properties of concrete, Wiley, New York.
Pantazopoulou, S. J., and Mills, R. H. 共1995兲. “Microstructural aspects of
the mechanical response of plain concrete.” ACI Mater. J., 92共6兲,
605–616.
Park, R., and Paulay, T. 共1975兲. Reinforced concrete structures, Wiley,
New York.
Pessiki, S., Harries, K. A., Kestner, J., Sause, R., and Ricles, J. M.
共2001兲. “The Axial behavior of concrete confined with fiber reinforced
composite jackets.” J. Compos. Constr., 5共4兲, 237–245.
Richart, F. E., 共1933兲. “Reinforced concrete column investigation.”
J. American Concrete Institute Proceedings, 29共2兲, 275–284.
Richart, F. E., and Brown, R. L. 共1934兲. “An investigation of reinforced
concrete columns.” Bulletin no. 267, Univ. of Illinois Engineering
Experiment Station.
Rocca, S., Galati, N., and Nanni, A. 共2006兲. “Experimental evaluation of
FRP strengthening of large size reinforced concrete columns.” Center
for Infrastructure Engineering Studies (CIES), Rep. No. 06-63, Univ.
of Missouri-Rolla, Rolla, Mo.
Rocca, S., Galati, N., and Nanni, A. 共2008兲. “Review of design guidelines
for FRP confinement of reinforced concrete columns of noncircular
cross sections.” J. Compos. Constr., 12共1兲, 80–92.
Saenz, N., and Pantelides, C. P. 共2007兲. “Strain-based confinement model
for FRP-confined concrete.” J. Struct. Eng., 133共6兲, 825–833.
Shehata, L. A. E. M., Carneiro, L. A. V., and Shehata, L. C. D. 共2002兲.
“Strength of short concrete columns confined with CFRP sheets.”
Mater. Struct., 35, 50–58.
Sim, J., Park, C., and Moon, D. Y. 共2005兲. “Characteristics of basalt fiber
as a strengthening material for concrete structures.” Composites, Part
B, 36, 504–512.
Slater, W. A., and Lyse, I. 共1931a兲. “First progress report on column tests
at Lehigh University.” J. American Concrete Institute Proceedings,
27共2兲, 677–730.
Slater, W. A., and Lyse, I. 共1931b兲. “Second progress report on column
tests at Lehigh University.” J. American Concrete Institute Proceed-
ings, 27共3兲, 791–835.
Slater, W. A., and Lyse, I. 共1931c兲. “Third progress report on column tests
at Lehigh University.” J. American Concrete Institute Proceedings,
28共11兲, 159–166.
Spoelstra, M. R., and Monti, G. 共1999兲. “FRP-confined concrete model.”
J. Compos. Constr., 3共3兲, 143–150.
Teng, J. G., Jiang, T., Lam, L., and Luo, Y. Z. 共2009兲. “Refinement of
a design-oriented stress-strain model for FRP-confined concrete.”
J. Compos. Constr., 13共4兲, 269–278.
Teng, J. G., and Lam, L. 共2004兲. “Behavior and modeling of fiber rein-
forced polymer-confined concrete.” J. Struct. Eng., 130共11兲, 1713–
1723.
Teng, J. G., Yu, T., Wong, Y. L., and Dong, S. L. 共2007兲. “Hybrid FRP-
concrete-steel tubular columns: Concept and behavior.” Constr. Build.
Mater., 21共4兲, 846–854.
Toutanji, H., Han, M., Gilbert, J., and Matthys, S. 共2010兲. “Behavior of
large-scale rectangular columns confined with FRP composites.”
J. Compos. Constr., 14共1兲, 62–71.
Wang, Y. C., and Restrepo, J. I. 共2001兲. “Investigation of concentrically
loaded reinforced concrete columns confined with glass fiber rein-
forced polymer jackets.” ACI Struct. J., 98共3兲, 377–385.
Wu, H. L., Wang, Y. F., Yu, L., and Li, X. R. 共2009兲. “Experimental and
computational studies on high-strength concrete circular columns con-
fined by aramid fiber-reinforced polymer sheets.” J. Compos. Constr.,
13共2兲, 125–134.
Wu, Y. F., and Wang, L. M. 共2009兲. “Unified strength model for square
and circular concrete columns confined by external jacket.” J. Com-
pos. Constr., 135共3兲, 253–261.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / JANUARY/FEBRUARY 2011 / 123

Downloaded 25 Jan 2011 to 143.225.98.141. Redistribution subject to ASCE license or copyright. Visithttp://www.ascelibrary.org