You are on page 1of 11

Bond Strength of Fiber Reinforced Polymer Rebars

in Normal Strength Concrete


Roman Okelo, A.M.ASCE,1 and Robert L. Yuan, P.E., M.ASCE2
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

Abstract: The bond behavior of reinforcing bars in concrete is a critical issue in the design of reinforced concrete structures. This study
focuses on the bond strength of fiber reinforced polymer 共FRP兲 rebars in normal strength concrete. Four different types of rebars were
tested using the pullout method: aramid FRP 共AFRP兲; carbon FRP 共CFRP兲; glass FRP 共GFRP兲, and steel. This involved a total of 151
specimens containing 6, 8, 10, 16, and 19 mm rebars embedded in a 203 mm concrete cube. The test embedment lengths were five, seven,
and nine times the rebar diameter 共db兲. For each rebar, the test results include the bond stress–slip response and the mode of failure. The
test results showed that the bond strength of an FRP rebar is, on average, 40–100% the bond strength on a steel rebar for pullout failure
mode. Based on this research, a proposal for the average bond strength of straight FRP rebars in normal strength concrete is made, which
verifies an existing bond strength relationship 共GFRP兲 and extends its application to AFRP and CFRP. It is an expression that is a function
of the rebar diameter, and the concrete compressive strength.
DOI: 10.1061/共ASCE兲1090-0268共2005兲9:3共203兲
CE Database subject headings: Corrosion resistance; Concrete structures; Pull-out resistance; Bonding strength; Structure
reinforcement.

Introduction high specific strength, high fatigue resistance, lightweight 共25%


of steel reinforcement兲, magnetic transparency, and non-
Corrosion of steel reinforcement in structural concrete members conductivity, and ease of handling at job site and cutting.
such as bridge decks, pavements, parking garages, wastewater Despite these advantages of FRP materials over steel rein-
treatment plants, and port and marine structures, leads to a pre- forcement, a direct substitution between FRP and steel rebar is
mature concrete deterioration and loss of serviceability or capac- not possible due to various differences in the mechanical and
ity. This is due to the exposure to aggressive environments, deic- physical properties of the two materials. The main differences that
ing salts, industrial chemicals, combinations of moisture, or affect design are: the tensile strength of FRP is significantly
temperature. Professionals have always faced the challenge of greater than steel; the modulus of elasticity for FRP 共AFRP and
GFRP兲, is much lower than steel, hence deflection and crack
corrosion decay in different ways. Methods such as galvanization,
widths may control design; the bond strength to concrete may
the use of stainless steel rebars, cathodic protection systems,
differ; and FRPs are linear elastic up to failure and exhibit no
epoxy coatings, concrete additives, etc., have been used. Although
ductility or yielding. The adaptation of FRP as a concrete rein-
effective in some situations, these methods may still be unable to
forcement has been slow for a number of reasons, including high
completely eliminate the problems of steel corrosion 共Kessler and
cost and lack of familiarity with the new technology 共e.g., lack of
Powers 1998兲.
standards for the rebar deformation geometry and composition as
Composite materials are made of fibers embedded in poly- given for steel in ASTM A 616/A兲.
meric resin, also known as fiber-reinforced polymers 共FRPs兲. The It is worth noting that a number of bond tests have already
most common fibers are made of aramid FRP 共AFRP兲, carbon been performed by several authors on FRP rebars 共see e.g., ACI
FRP 共CFRP兲, and glass FRP 共GFRP兲. The resin matrix that binds 440.1R-01; Brown and Bartholomew 1993; Larralde and Silva-
the fibers together to form composites are those of the polyester, Rodriguez 1993; Ehsani et al. 1995, 1996; Malvar 1995;
vinyl ester, and epoxy groups. The outstanding characteristics of Benmokrane et al. 1996; Focacci et al. 2000; Katz 2000; De
these materials, make them an ideal material of choice to address Lorenzis and Nanni 2001, 2002; and Pecce et al. 2001兲. However,
the infrastructure problem. These include: corrosion resistance, due to the complexity of the parameters influencing the bond
1
Visiting Assistant Professor, Dept. of Civil Engineering, Lamar
Univ., P.O. Box 10024, Beaumont, TX 77710. E-mail: okelorx@ Table 1. Composition and Characteristics of Concrete
hal.lamar.edu Concrete mix M0 M1 M2
2
Professor, Chair, Dept. of Civil Engineering, Lamar Univ., Box
10024, Beaumont, TX 77710. E-mail: robert.yuan@lamar.edu Water 共kg/ m 兲3
182 185 189
Note. Discussion open until November 1, 2005. Separate discussions Cement 共kg/ m3兲 349 408 503
must be submitted for individual papers. To extend the closing date by Coarse aggregate 共kg/ m3兲a 905 905 905
one month, a written request must be filed with the ASCE Managing Fine aggregate 共kg/ m3兲b 881 798 715
Editor. The manuscript for this paper was submitted for review and pos-
Air content 共%兲 2.5 2.5 2.5
sible publication on August 26, 2002; approved on June 5, 2003. This a
paper is part of the Journal of Composites for Construction, Vol. 9, No. Maximum aggregate size was 12.7 mm.
b
3, June 1, 2005. ©ASCE, ISSN 1090-0268/2005/3-203–213/$25.00. Size interval was 0 – 4.75 mm.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 203

J. Compos. Constr., 2005, 9(3): 203-213


mate states. The most commonly used test procedures to address
these issues are: pullout tests 共with centric or eccentric placement
of the reinforcing rebar in the concrete specimen兲, beam tests
共e.g., RILEM type兲, and splice tests. Depending on cover thick-
ness and confinement, the failure mode can be pullout or splitting.
With these in mind, this study focuses on the bond of straight
FRP reinforcing bars in normal strength concrete using the pull-
out test method. A total of 151 specimens were tested 共Yuan and
Okelo 2001; Okelo 2002兲. Based on these results, an empirical
model for the average bond strength of straight FRP rebars in
normal strength concrete is presented.
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

Experimental Program

Materials
Fig. 1. Surface deformations, characteristics of rebars 共R1–R14兲 and
Concrete
specimen wedges
Normal strength concrete was prepared in the laboratory accord-
ing to ASTM Standard C 192–95, with the composition as given
behavior 共the most significant being: diameter of rebar, spacing, in Table 1. Type I Portland cement was used. The nominal maxi-
embedment length, and concrete compressive strength兲, and the mum coarse aggregate size was 12.7 mm, and the fine aggregate
different types of currently commercially available FRP rebars, a size interval was 0 – 4.75 mm. Forms constructed out of plywood
comprehensive relationship for the bond strength of straight FRP were used to cast the concrete around the rebar. The concrete mix
rebars in concrete has not yet been established. This is the goal of was designed based on the absolute volume method and it was
this study. nonair entrained with an air content of 2.5%. The concrete was
Bond of reinforcing bars to concrete influences the behavior of mixed in a ⬃0.2 m3 effective capacity rotating mixer. The fine
structural concrete in many respects. It affects the anchorage of and coarse aggregates were first mixed together, then the cement
rebars, the strength of lap splices, and the serviceability and ulti- was added and allowed to mix thoroughly. Finally the water was

Table 2. Geometrical Properties of Rebars


Means to improve bondd,e

Deformations by resin,
surface undulations/
indentations or ribs
Fiber/ Bar
rebar diameterc Sand Surface Helical Deep dents Spacing Height
Firm typea,b 共mm兲 coating texture wrapping 共groves兲 共mm兲 共mm兲
A A/R11 10 ⫹ — — — — —
C/R12 10 ⫹ — — — — —
B G/R2 6 ⫹ — ⫹ — ⫹ ⫹
G / R2* 10 ⫹ — ⫹ — ⫹ ⫹
G/R1 16 ⫹ — ⫹ — ⫹ ⫹
G / R1* 19 ⫹ — ⫹ — ⫹ ⫹
C/R4 10 — ⫹ — — — —
C/R5 10 — ⫹ — ⫹ — —
C/R3 10 — — — ⫹ — —
C G/R7 10 — — — — 5.72 0.76
G/R6 19 — — — — 11.43 1.52
C/R8 10 — — — — 5.72 0.76
D G/R13 8s — ⫹ — — — —
E G/R9 10 ⫹ — — — — —
F G/R10 10 — — — — 4.29 0.71
G S/R14 10 — — — — 6.65 0.38
a
A = aramid; C = carbon; G = glass; and S = steel.
b*
= R1 and R2 have similar surface configuration.
c
s = square cross section.
d
+ = means to improve bond available.
e
— = not available.

204 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 203-213


Table 3. Mechanical Properties of Rebars
Fiber/ Bar Fiber Tensile Elastic
rebar diameterc Density content strength modulus
Firm typea,b 共mm兲 Resind 共g / cm3兲 共%兲 共MPa兲 共GPa兲
A A/R11 10 Vinyl ester 1.3 ⬃49 1,350 60
C/R12 10 1.5 ⬃58 2,800 130
B G/R2 6 Vinyl ester 2.0 艌70 825 40.8
G / R2* 10 760
G/R1 16 655
G / R1* 19 620
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

C/R4 10 — — — ⬃1,648 ⬃138


C/R5 10 — — — ⬃1,648
C/R3 10 — — — ⬃517
C G/R7 10 Modified 1.9 60 772 38.7
G/R6 19 Vinyl ester 772 38.7
C/R8 10 — — — — —
D G/R13 8s — — — 900 45
E G/R9 10 — — — — —
F G/R10 10 Vinyl ester 1.5–2.0 — 517–1,207 41–55
G S/R14 10 NA 7.9 NA 550 200
a
A = aramid; C = carbon; G = glass, and S = steel.
b*
= R1 and R2 have similar surface configuration.
c
s = square cross section.
d
— = no information available.

added and allowed to mix until the concrete became uniform in bond forces. The bond length should be such that the bond
appearance. The ratio of cement to water 共w/c兲 was 0.52 for mix stresses are almost evenly distributed during the loading process.
M0, 0.45 for mix M1, and 0.38 for mix M2. Immediately after Fig. 2 shows the test setup, the loading frame, dimensions of the
mixing, the slump of each batch of concrete was measured. test specimen, and bonded length of the rebars for the test. The
rebars were pulled from 203 mm concrete cube. The test embed-
Fibers and Resin ment lengths were five, seven, and nine times the rebar diameter
The FRP reinforcing bars 共AFRP, CFRP, and GFRP兲 used in this 共db兲. All the rebars were 762 mm long. Contact between the rebar
study were supplied by international manufacturers. The surface and concrete along the debonded length was broken using a soft
deformations, characteristics of the rebars and the specimen plastic tubing to reduce the influence of the area close to the
wedges used are shown in Fig. 1. These rebars had nominal di-
ameter of 6, 8, 10, 16, and 19 mm. They are generally manufac-
tured using the so-called pultrusion process, and are made of con-
tinuous longitudinal fibers bound together with a thermosetting
vinyl ester resin 共see Table 3兲. The bond improvement is achieved
through the following means: sand coating, surface texture, heli-
cal wrapping with sand coating, deep dents 共grooves兲, and defor-
mations by resin 共surface undulations or indentations or ribs兲 as
given in Table 2, showing their geometrical properties. In Table 3,
their mechanical properties are presented. The fiber volume frac-
tion ranges from about 50 to more than 70%. The tensile strengths
were: 550 MPa 共steel兲; 1,350 MPa 共AFRP兲; 517– 1,207 MPa
共GFRP兲; and 517– 2,800 MPa 共CFRP兲. The Young’s moduli
共axial direction兲 provided by each manufacturer were: 200 GPa
共steel兲; 60 GPa 共AFRP兲; 38.7– 55 GPa 共GFRP兲; and 130– 138
GPa 共CFRP兲. All the FRP rebars follow a linear stress–strain be-
havior up to failure.

Specimen
For the experimental determination of the local bond slip relation
between rebar and concrete, pullout tests with centric rebar place-
ment in concrete can be used. This method is popular because it
provides a simple means of comparing the relative bond devel-
oped by different rebars and materials. The confining action pro-
vided by the surrounding concrete mass or reinforcement should Fig. 2. Test setup, loading frame, dimensions of test specimen, and
be adequate so as to minimize the risk of splitting the concrete by bonded length of rebars for test

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 205

J. Compos. Constr., 2005, 9(3): 203-213


Table 4. Bond Test Results for Specimens 共db = 6, 8, and 10 mm兲
Concrete strength Ultimate stated

Compressive Splitting Load Average bond Free end Loaded end Modified bond
Fiber/ rebar f c⬘ T Pmax stress u slip ␦fe slip ␦le stress u*e Failure
Specimena typeb,c 共MPa兲 共MPa兲 共kN兲 共MPa兲 共mm兲 共mm兲 共MPa兲 modef
6-2-5#1 G/R2 44.3 3.38 12.6 19.9 — 0.03 18.9 P
6-2-5#2 9.6 15.1 1.22 1.4 14.3 P
6-2-5#3 9.3 14.7 0.84 1.3 14 P
8-2-5#1 G/R13 44.9 2.89 17.6 13.7 0.84 0.69 12.9 P
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

8-2-5#2 23.9 18.7 0.48 0.36 17.6 P


8-2-5#3 11.3 8.8 0.66 1.24 8.3 P
10-2-5#1 S/R14 42.8 3.58 33.9 23.8 0.36 8.15 22.9 P
10-2-5#2 35.6 24.9 2.87 22.66 24.2 P
10-2-5#3 34.9 24.5 1.75 20.98 23.7 P
10-2-5#1 G / R2* 38.6 3.21 28.6 20.1 0.61 0.71 20.5 A
10-2-5#2 30.8 21.6 0.03 0.89 22.0 A
10-2-5#3 29.8 20.9 0.03 0.74 21.2 P
10-2-5#1 G/R7 41.3 3.11 31.3 21.9 0.69 1.55 21.6 P
10-2-5#2 32.8 22.9 0.23 0.43 22.6 P
10-2-5#3 33.9 23.8 0.23 1.14 23.4 P
10-2-5#1 C/R8 42.7 3.28 28.4 19.9 0.63 0.61 19.3 P
10-2-5#2 37.6 26.4 0.30 0.66 25.5 P
10-2-5#3 31.5 22.1 0.56 0.74 21.4 P
10-2-5#1 A/R11 42.5 3.37 18.6 13.1 — 0.30 12.7 P
10-2-5#2 19.3 13.6 — 0.48 13.2 P
10-2-5#3 11.7 8.2 — 0.23 8.0 P
10-2-5#1 C/R12 41.3 3.08 32.6 22.9 — 0.23 22.5 P
10-2-5#2 30.4 21.3 — 0.15 21.0 P
10-2-5#3 28.9 20.3 — 0.18 19.9 P
10-2-5#1 G/R9 44.8 2.88 22.2 15.5 0.18 0.28 14.7 A
10-2-5#2 28.6 20.1 0.46 0.28 19.0 A
10-2-5#3 30.2 21.2 0.23 0.10 20.0 A
10-2-5#1 G/R10 46.4 3.49 15.8 11.1 0.56 0.81 10.3 P
10-2-5#2 16.1 11.3 1.55 2.44 10.5 P
10-2-5#3 14.5 10.2 0.61 1.09 9.5 P
10-2-5#1 C-SG/R3 44.6 3.18 18.9 13.3 11.23 10.69 12.6 P
10-2-5#2 17.8 12.5 12.24 13.33 11.8 P
10-2-5#3 12.8 8.9 9.12 9.17 8.5 P
10-2-5#1 C-T/R4 20.9 14.6 0.10 0.33 13.9 P
10-2-5#2 16.7 11.7 — 0.05 11.1 P
10-2-5#3 17.4 12.2 0.08 0.58 11.5 P
10-2-5#1 C-TG/R5 26.7 18.7 1.75 3.07 17.8 P
10-2-5#2 24.3 17.1 1.14 3.15 16.2 P
10-2-5#3 25.2 17.7 0.74 0.99 16.7 P
10-2-5#1 G / R2* 60.4 3.79 19.8 13.9 0.56 0.30 11.3 A
10-2-5#2 23.7 16.6 0.71 0.20 13.5 A
10-2-5#3 29.9 20.9 — 0.46 17.1 A
10-2-9#1 33.7 13.1 0.10 0.89 10.7 A
10-2-9#2 26.7 10.4 0.30 0.33 8.5 A
10-2-9#3 25.4 9.9 0.56 0.10 8.1 A
10-0-5#1 41.9 3.11 28.2 19.8 0.48 1.22 19.4 A
10-0-5#2 23.2 16.3 7.06 8.59 15.9 P
10-0-5#3 31.9 22.4 0.84 1.24 21.9 R

206 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 203-213


Table 4. 共Continued.兲
Concrete strength Ultimate stated

Compressive Splitting Load Average bond Free end Loaded end Modified bond
Fiber/ rebar f c⬘ T Pmax stress u slip ␦fe slip ␦le stress u*e Failure
Specimena typeb,c 共MPa兲 共MPa兲 共kN兲 共MPa兲 共mm兲 共mm兲 共MPa兲 modef
10-0-7#1 33.7 16.9 0.43 1.32 16.5 A
10-0-7#2 29.1 14.6 0.79 0.69 14.2 A
10-0-7#3 31.5 15.8 0.41 1.04 15.4 A
10-0-9#1 33.5 13.0 0.36 1.19 12.7 A
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

10-0-9#2 29.8 11.6 0.46 0.97 11.3 A


10-0-9#3 35.8 13.9 0.10 1.24 13.6 R
a
BD-M-K: BD= bar diameter; M = concrete mix, and K = factor for embedment length.
b
A = aramid; C = carbon 共surface treatment: S = smooth; G = grove, and T = textured兲; and G = glass.
c*
= R1 and R2 have similar surface configuration.
d
— = not measured.
u = u ⫻ 共40/ f c⬘兲0.5.
e *

f
A = anchorage; P = pullout, and R = rebar fracture.

bearing plate. Three nominally identical specimens were tested were carried out in displacement control mode so as to obtain the
for each specimen type to obtain some measure of uncertainty. postpeak behavior. The load was measured with the electronic
The concrete was placed in three layers of approximately load cell of the machine. The slips of the rebar relative to concrete
equal thickness and each layer was rodded 25 times with 16 mm at the loaded end and at the free end were measured with four
diameter tamping rod. After molding, the specimens were initially LVDTs. Output from the MTS machine and the LVDTs were re-
cured by covering them with plastic sheet to prevent moisture loss corded using an automatic data acquisition system.
for 24 h. Thereafter, the molds were removed, and for final curing
the specimens were stored for 14 days in a moist room. During
this period, they were sprayed daily with water so as to maintain Results
moisture on the surfaces at all the times.
Research indicates that there can be a difference in bond
Experimental results for the bond testing are given in Table 4
strength when reinforcing bars are situated vertical or horizontal
共db = 6, 8, and 10 mm兲; in Table 5 共db = 10 and 16 mm兲, and in
at casting. In this study the concrete was cast with the rebar in the
Table 6 共db = 19 mm兲. The specimen identification is in the form:
vertical position. For each batch of concrete mixed, nine 共or six兲
BD-M-K; where BD is the rebar diameter in millimeters, M is the
152.4 by 304.8 mm standard cylinders were cast and cured under
concrete mix, and K is the multiplication factor for embedment
the same conditions as the specimens. Six 共or three兲 of the control
length 共K ⫻ rebar diameter兲. Failure was defined as the point of
cylinders were for determining strength in compression and the
maximum pullout load during the test. Corresponding maximum
other three were for the tensile strength according to the Brazilian
nominal bond stress and slip values were then defined as those
splitting test. At some time prior to testing 共after 14 days兲, the test
values occurring at the point of failure. Since the concrete com-
control cylinders for compressive strength were capped with stan-
pressive strength of the batches varied, a modified bond strength
dard sulfur capping compound to provide plane surfaces on the
is also given in the table for comparison purposes. The modified
ends of the cylinders.
bond strength is obtained by normalizing the test results with
respect to a nominal concrete strength of 40 MPa, using the as-
Testing Equipment sumption that, within the concrete strength range used, bond
A critical issue in testing FRP reinforcing bar is its low transverse strength is proportional to the square root of the compressive
strength. Hence, the rebars cannot be tested in tension using the strength. Generally, the test results indicate that the longer the
same gripping systems that are used for steel rebars. If an FRP embedment length, the smaller the value of the average bond
rebar is loaded using traditional wedge-shaped frictional grips strength. The decrease in average bond strength with increasing
e.g., vee specimen wedges 共see Fig. 1兲, the combination of high embedment length can be explained by considering the actual
compressive stresses and mechanical damage caused by the ser- bond stress distribution 共instead of assuming it constant兲 which
ration on the wedge surfaces will lead to premature failure at the shows a peak near the loaded end of the specimen.
grip zone. For this study, special, round specimen wedges with
alloy on the surface as given in Fig. 1, were used for some of the Mode of Failure
tests. Otherwise, the vee specimen wedges were used in conjunc-
tion with copper sleeves at the gripping end of the rebar. The type of failure observed for each test is given in the corre-
The tests were performed using a MTS servo hydraulic testing sponding Tables 4–6. In most cases the specimens failed by pull-
machine with a capacity of 445 kN. A loading frame 共see Fig. 2兲, out of the rebar. But, failure at the anchorage; rupture of the rebar;
mounted in the lower grip of the machine by means of a high or splitting of the enclosing concrete were also observed. In two
strength bolt db = 19 mm, was used to transfer the reaction from cases, the rebar slipped out of the wedge. Pullout failure occurred
the specimen to the machine. The load was applied to the rein- once the shear strength of the bond between the rebar and the
forcement bar at a rate of 0.274 mm/ min. Hence, all the tests concrete was exceeded. The ultimate bond strength of the speci-

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 207

J. Compos. Constr., 2005, 9(3): 203-213


Table 5. Bond Test Results for Specimens 共db = 10 and 16 mm兲
Concrete strength Ultimate statec

Compressive Splitting Load Average bond Free end Loaded end Modified bond
Fiber/rebar f c⬘ T Pmax stress u slip ␦fe slip ␦le stress u* d Failure
Specimena typeb 共MPa兲 共MPa兲 共kN兲 共MPa兲 共mm兲 共mm兲 共MPa兲 modee
10-0-5#1 G/R7 39.4 2.74 23.9 16.8 1.75 2.64 16.9 P
10-0-5#2 25.4 17.8 — 1.35 18.0 P
10-0-5#3 28.2 19.8 0.64 0.71 20.0 P
10-0-7#1 31.0 15.6 1.22 2.92 15.7 P
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

10-0-7#2 31.7 15.9 1.19 1.42 16.0 P


10-0-7#3 28.9 14.5 0.38 1.12 14.6 P
10-0-9#1 32.6 12.7 0.56 0.48 12.8 A
10-0-9#2 36.4 14.2 0.38 0.89 14.3 A
10-0-9#3 34.3 13.4 0.23 1.50 13.5 A
10-1-5#1 30.8 2.32 25.0 17.5 1.42 0.79 20.0 P
10-1-5#2 21.5 15.1 2.84 0.94 17.2 P
10-1-5#3 21.1 14.8 1.60 0.74 16.8 P
10-1-7#1 33.9 17.0 0.05 0.86 19.3 A
10-1-7#2 34.3 17.2 0.89 0.91 19.6 R
10-1-7#3 33.4 16.8 1.47 0.84 19.1 R
10-1-9#1 34.3 13.4 0.41 1.09 15.2 R
10-1-9#2 36.3 14.1 0.99 2.44 16.1 R
10-1-9#3 36.4 14.2 0.03 0.61 16.2 R
10-2-5#1 46.4 3.49 25.4 17.8 0.20 1.19 16.5 P
10-2-5#2 22.4 15.7 — 0.10 14.6 P
10-2-5#3 32.8 23.0 0.48 0.15 21.3 P
16-0-5#1 G/R1 29.7 2.53 75.1 19.0 0.99 4.04 22.0 P
16-0-5#2 74.2 18.8 2.92 4.88 21.8 P
16-0-5#3 48.4 12.2 3.94 5.18 14.2 P
16-0-5a#1 39.0 2.65 84.7 21.4 3.89 4.85 21.7 P
16-0-5a#2 63.4 16.0 5.44 6.58 16.2 P
16-0-5a#3 79.0 20.0 3.99 4.50 20.2 P
16-0-7#1 107.2 19.4 1.42 2.72 19.6 R
16-0-7#2 106.8 19.3 0.20 3.07 19.5 R
16-0-7#3 100.9 18.2 4.88 5.69 18.4 P
16-0-9#1 102.5 14.4 0.38 1.47 14.6 R
16-0-9#2 108.4 15.2 0.48 1.65 15.4 R
16-0-9#3 103.8 14.6 0.20 1.91 14.7 R
16-1-5#1 41.8 3.46 51.2 12.9 0.33 0.74 12.7 SL
16-1-5#2 75.3 19.0 3.58 3.94 18.6 P
16-1-5#3 71.7 18.1 4.70 5.44 17.7 P
16-1-7#1 103.4 18.7 4.47 4.93 18.3 P
16-1-7#2 99.5 17.9 4.17 5.21 17.6 P
16-1-7#3 87.7 15.8 3.15 4.34 15.5 P
16-1-9#1 97.1 13.6 2.01 2.97 13.3 R
16-1-9#2 110.3 15.5 0.46 1.35 15.1 R
16-1-9#3 111.8 15.7 0.38 2.34 15.4 R
16-2-5#1 45.5 3.25 62.5 15.8 0.81 0.58 14.8 SL
16-2-5#2 70.1 17.7 4.04 5.56 16.6 P
16-2-5#3 75.1 19.0 3.25 4.14 17.8 P
a
BD-M-K: BD= bar diameter; M = concrete mix, and K = factor for embedment length.
b
A = aramid; C = carbon 共surface treatment: S = smooth; G = grove; and T = textured兲, and G = glass.
c
— = not measured.
u = u ⫻ 共40/ f c⬘兲0.5.
d *

e
A = anchorage; P = pullout; R = rebar fracture, and SL= bar slipped out of wedge.

208 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 203-213


Table 6. Bond Test Results for Specimens 共db = 19 mm兲
Concrete strength Ultimate stated

Compressive Splitting Load Average bond Free end Loaded end Modified bond
Fiber/rebar f c⬘ T Pmax stress u slip ␦fe slip ␦le stress u*e Failure
Specimena typeb,c 共MPa兲 共MPa兲 共kN兲 共MPa兲 共mm兲 共mm兲 共MPa兲 modef
19-2-5#1 G / R1* 53.3 2.88 100.9 17.7 — 1.85 15.3 S
19-2-5#2 114.0 20.0 — 3.78 17.3 S
19-2-5#3 95.3 16.7 — 2.39 14.5 S
19-2-7#1 124.0 15.5 1.12 0.91 13.5 S
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

19-2-7#2 130.3 16.3 1.55 3.15 14.1 S


19-2-7#3 135.5 17.0 — 1.30 14.7 A
19-2-9#1 48.3 3.23 104.7 10.2 0.38 1.40 9.3 S
19-2-9#2 121.2 11.8 2.72 3.84 10.7 P
19-2-9#3 114.8 11.2 — 0.81 10.2 S
19-1-5#1 45.6 3.48 119.9 21.0 2.59 4.72 19.7 S
19-1-5#2 86.0 15.1 2.11 2.06 14.1 P
19-1-5#3 97.1 17.0 4.60 5.46 15.9 P
19-1-7#1 44.1 3.39 115.1 14.4 2.77 4.14 13.7 S
19-1-7#2 105.3 13.2 5.69 5.54 12.6 P
19-1-7#3 132.9 16.7 1.32 2.72 15.9 S
19-1-9#1 74.1 7.2 — 0.61 6.9 S
19-1-9#2 74.9 7.3 — 0.30 7.0 S
19-1-9#3 106.8 10.4 0.23 0.79 9.9 S
19-0-5#1 35.0 2.91 86.9 15.2 4.50 3.28 16.3 P
19-0-5#2 76.0 13.3 3.78 3.61 14.3 P
19-0-5#3 84.0 14.7 5.31 5.69 15.8 P
19-0-5#4 68.6 12.0 0.08 0.38 12.9 P
19-0-7#1 33.5 3.30 92.7 11.6 3.68 3.94 12.7 P
19-0-7#2 114.4 14.3 2.92 3.73 15.7 S
19-0-7#3 91.6 11.5 0.99 0.64 12.5 P
19-0-9#1 95.8 9.3 4.39 6.30 10.2 P
19-0-9#2 92.0 9.0 4.01 4.78 9.8 P
19-0-9#3 97.3 9.5 2.51 3.35 10.4 P
19-0-5#1 G/R6 38.2 3.18 91.1 16.0 — 0.61 16.4 A
19-0-5#2 106.4 18.7 0.23 1.42 19.1 R
19-0-5#3 84.9 14.9 0.94 0.43 15.2 S
19-0-7#1 92.5 11.6 0.66 1.07 11.9 S
19-0-7#2 110.7 13.9 0.28 1.14 14.2 R
19-0-7#3 85.5 10.7 0.10 0.74 11.0 S
19-0-9#1 76.4 7.4 — 0.79 7.6 S
19-0-9#2 62.8 6.1 1.22 0.58 6.3 S
19-0-9#3 56.7 5.5 1.47 0.46 5.6 S
19-1-5#1 40.0 3.51 96.4 16.9 0.71 1.30 16.9 R
19-1-5#2 99.2 17.4 0.99 1.63 17.4 R
19-1-5#3 103.1 18.1 — 1.12 18.1 R
19-1-7#1 98.3 12.3 0.48 0.15 12.3 R
19-1-7#2 95.5 12.0 0.81 0.74 12.0 R
19-1-7#3 103.8 13.0 0.36 0.97 13.0 R
19-1-9#1 65.1 6.4 0.10 0.61 6.4 S
19-1-9#2 60.6 5.9 0.99 0.36 5.9 S
19-1-9#3 64.7 6.3 0.79 0.38 6.3 S
19-2-5#1 49.0 2.95 94.3 16.5 0.20 0.97 14.9 R
19-2-5#2 102.3 17.9 0.61 0.36 16.2 R

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 209

J. Compos. Constr., 2005, 9(3): 203-213


Table 6. 共Continued.兲
Concrete strength Ultimate stated

Compressive Splitting Load Average bond Free end Loaded end Modified bond
Fiber/rebar f c⬘ T Pmax stress u slip ␦fe slip ␦le stress u*e Failure
Specimena typeb,c 共MPa兲 共MPa兲 共kN兲 共MPa兲 共mm兲 共mm兲 共MPa兲 modef
19-2-5#3 89.4 15.7 0.10 1.17 14.2 R
19-2-7#1 94.4 11.8 0.28 0.91 10.7 R
19-2-7#2 105.7 13.3 0.58 1.50 12.0 R
19-2-7#3 93.6 11.7 0.74 0.86 10.6 R
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

a
BD-M-K: BD= bar diameter; M = concrete mix, and K = factor for embedment length.
b
A = aramid; C = carbon 共surface treatment: S = smooth, G = grove, T = textured兲; and G = glass.
c*
= R1 and R2 have similar surface configuration.
d
— = not measured.
u = u ⫻ 共40/ f c⬘兲0.5.
e *

f
A = anchorage; P = pullout; R = rebar fracture, and S = splitting.

mens was dictated by the shear strength of the concrete surround- ␦le = ␦t − ␦c 共2兲
ing the rebar and by the geometrical properties of the rebar,
mostly the surface configuration 共deformation and shape兲.
PL
␦c = 共3兲
EA
Average Bond Stress–Slip Relationships
where ␦le = loaded end slip 共mm兲; ␦t = average of the measured
The average bond stress is calculated as slips 共mm兲; ␦c = slip correction due to rebar elongation 共mm兲;
P P = tensile load 共N兲; L = length between the top surface of bonded
u= 共1兲 length and the average point of attachment of the LVDTs on the
␲dblbf rebar 共mm兲; E = modulus of elasticity 共MPa兲; and A = nominal
where u = average bond stress 共MPa兲; P = tensile load 共N兲; cross sectional area 共mm2兲. The free end slip is the difference
db = effective rebar diameter 共mm兲; and lbf = embedment length between the LVDT readings on the rebar and on the concrete.
共mm兲. The loaded end slip is calculated, taking into account the Shown in Figs. 3–8, are the results in terms of bond stress–slip
adjustment for the elastic elongation of the rebar between the response. The identification is in the form: A, B-C-BD-M-K;
actual loaded end of the embedment length and the attachment where A is the rebar type, B is the firm, C is the fiber type, BD is
point of the LVDTs as given below the rebar diameter in millimeters, M is the concrete mix, and K is

Fig. 3. Typical average bond stress–slip relationship for aramid Fig. 4. Typical average bond stress–slip relationship for carbon
rebars 共aramid fiber reinforced polymer兲, firm A, R11, db = 10 mm rebars 共carbon fiber reinforced polymer兲, firm A, R12, db = 10 mm

210 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 203-213


Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Typical average bond stress–slip relationship for carbon Fig. 7. Typical average bond stress–slip relationship for glass rebars
rebars 共carbon fiber reinforced polymer兲, firm B, R5, db = 10 mm 共glass fiber reinforced polymer兲, firm B, R1, db = 16 mm

the multiplication factor for embedment length 共K ⫻ rebar diam- The oscillating behavior in the postpeak zone is due to the
eter兲. The behavior is characterized by an initial steep increase in shearing off of the sand coating on the AFRP bar 共see Fig. 3兲. A
the bond stress with little slippage, followed by softening once the similar behavior has also been reported by Zhang and
ultimate bond stress is exceeded. Up to failure bond can be attrib- Benmokrane 共2000兲 for a grouted Arapree rod. Analogously, in
uted to bearing 共deformed rebars only兲, adhesion, and friction Fig. 6 the early oscillations is due to the re-engaged mechanical
between the rebar and concrete. Once the adhesive bond is bro- interlocking after shearing off of the rib on the CFRP bar. Later
ken, at failure friction reduces as the rebar is pulled out further on, the period of oscillations correspond approximately to the
and the contact surface is worn. deformations spacing on the rebar. As mentioned above, the

Fig. 6. Typical average bond stress–slip relationship for carbon Fig. 8. Typical average bond stress–slip relationship for steel rebars
rebars 共carbon fiber reinforced polymer兲, firm C, R8, db = 10 mm 共steel兲, firm G, R14, db = 10 mm

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 211

J. Compos. Constr., 2005, 9(3): 203-213


This relationship indicates that the data further verifies the
relationship proposed for GFRP bars by Ehsani et al. 共1996兲, dif-
fering only in the constant—14.7 versus their value of 14.25.
However, the current study extends the applicability of the rela-
tionship to all FRP rebars 共AFRP, CFRP, and GFRP兲.

Conclusions

A total of 151 pullout tests were carried out on FRP 共AFRP,


CFRP, and GFRP兲 and steel reinforcing bars to study their differ-
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

ent properties on the bond behavior. For each of the tested rebars,
average bond stress–slip relationships and failure mode were ex-
perimentally determined. The results of this study can be summa-
rized as follows: 共1兲 surface deformations similar to the ones on
steel are very effective means of bond enhancement; 共2兲 other
surface deformations or indentations obtained by stressing an ex-
ternal helicoidal strand and deep dents 共groves兲 are acceptable
Fig. 9. Comparison of proposed equation and test results for all means of bond improvement; 共3兲 sand coating and surface texture
rebars are other possibilities of obtaining better bond performance com-
pared to a smooth surface; 共4兲 the mode of failure of bond is
mainly dependent on the concrete compressive strength, the shape
loaded end slip is computed based on the reading from two 共and composition兲 of the bar’s surface, the cover thickness, and
LVDTs. In this case, if one of the readings was not representative embedment length. For shorter embedment lengths with low com-
of the behavior it was ignored 共see Fig. 8, specimen 1兲. The pressive strengths, actual pullout of the rebar occurs. For short
behavior is realistic, since during the whole loading process, the embedment lengths with higher compressive strengths and for
loaded end slip is larger than the free end slip. Damage of the longer embedment lengths with low compressive strengths, split-
interface initiates at the loaded end and propagates towards the ting failures may occur; 共5兲 pullout load increases with increasing
free end. embedment length, but the average bond strength decreases, due
to the fact that the pullout load increase is not proportional to the
increase in embedment length; 共6兲 as for steel, the average bond
Empirical Model strength of FRP rebars decreases as the reinforcing bar diameter
increases; 共7兲 the bond strength of an FRP rebar is typically 40–
The applied load and the strength of a reinforced concrete mem- 100% that of a steel rebar for pullout failure mode; 共8兲 the
ber are variable quantities. Hence, the conditions of overload and Young’s modulus of the rebar namely the type of fiber 共AFRP,
understrength contribute to the cause of failure. To reduce the CFRP, and GFRP兲 appears to have some influence on the bond
probability of failure to an acceptable level, most present codes strength, and 共9兲 a proposal for the average bond strength of
for reinforced concrete specify the use of the overload factors straight FRP rebars in normal strength concrete is made which is
applied to the loads and the understrength factors applied to the a function of 共f ⬘c 兲0.5.
strength. Analogously, this has to be accounted for in obtaining
the average bond strength of FRP rebars in concrete.
For the 151 bond pullout specimens, the following assump- Acknowledgments
tions are made 共Okelo 2002兲: 共1兲 the average bond strength, for a
given concrete range is normally distributed; 共2兲 the characteris- The writers wish to acknowledge the following for their corpora-
tic, or specified, average bond strength is that expected to be tion and support in form of material donations, and financial re-
exceeded by a large proportion of the concrete in a structure. For source: Doug Gremel of Hughes Brothers; Sireg Spa; Marshall
this study, we use 75%, indicating that 25% of the predicted av- Industries Composites; Eurocrete; Pultrall; Re-bars Chili, and Tim
erage bond strength in the structure would be lower than the Bradberry of The Texas Department of Transportation 共TxDOT兲,
specified strength; 共3兲 the average bond strength is proportional to Austin.
共f ⬘c 兲0.5, and 共4兲 as for steel, the average bond strength of FRP
rebars decreases as the reinforcing bar diameter increases. Based
on these considerations, a relationship as given in Fig. 9 关see Eq. Notation
共4兲兴 is used to represent the average bond force as a function of
the concrete compressive strength for all the tested rebars. The The following symbols were used in this paper:
figure shows that most of the test results are enclosed by the
A ⫽ nominal cross sectional area of fiber reinforced
relationship. Hence the average bond strength for a straight FRP
polymer reinforcement 共mm2兲;
rebar in normal strength concrete can be expressed as follows:
db ⫽ effective rebar diameter 共mm兲;
冑 f c⬘ E ⫽ guaranteed modulus of elasticity of fiber reinforced
u f = 14.7 共MPa兲 共4兲 polymer defined as mean modulus of sample of test
db
specimens minus three times standard deviation 共MPa兲;
in which u f = average bond strength in MPa; f ⬘c = specified con- f ⬘c ⫽ specified compressive strength of concrete 共MPa兲;
crete compressive strength in MPa; and db = effective rebar diam- 冑f ⬘c ⫽ square root of specified compressive strength of
eter in millimeters. concrete 共MPa兲;

212 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005

J. Compos. Constr., 2005, 9(3): 203-213


L ⫽ length between top surface of bonded length and Ehsani, M. R., Saadatmanesh, H., and Tao, S. 共1995兲. “Bond of hooked
average point of attachment of linear variable differential glass fiber reinforced plastic 共GFRP兲 reinforcing bars to concrete.”
transducers on rebar 共mm兲; ACI Mater. J., 92共4兲, 391–400.
lbf ⫽ basic development length of fiber reinforced polymer Ehsani, M. R., Saadatmanesh, H., and Tao, S. 共1996兲. “Design recom-
mendations for bond of GFRP rebars to concrete.” J. Struct. Eng.,
rebar 共mm兲 共embedment length兲;
122共3兲, 247–254.
P ⫽ tensile load 共N兲; Focacci, F., Nanni, A., and Bakis, C. E. 共2000兲. “Local bond–slip rela-
Pmax ⫽ maximum tensile load 共N兲; tionship for FRP reinforcement in concrete.” J. Compos. Constr.,
T ⫽ splitting concrete strength 共MPa兲; 4共1兲, 24–31.
u ⫽ average bond stress 共MPa兲; Katz, A. 共2000兲. “Bond to concrete of FRP rebars after cyclic loading.” J.
u f ⫽ average bond strength 共MPa兲; Compos. Constr., 4共3兲, 137–144.
␦c ⫽ slip correction due to rebar elongation 共mm兲; Kessler, R. J., and Powers, R. G. 共1998兲. “Corrosion of epoxy-coated
Downloaded from ascelibrary.org by National Institute of Technology Calicut on 09/28/19. Copyright ASCE. For personal use only; all rights reserved.

␦fe ⫽ free end slip 共mm兲; rebars—Keys segmental bridge—Monroe County.” Rep. No. 88-8A,
␦le ⫽ loaded end slip 共mm兲; and Florida Department of Transportation, Materials Office, Corrosion
␦t ⫽ average of the measured slips 共mm兲. Research Laboratory, Gainesville, Fla.
Larralde, J., and Silva-Rodriguez, R. 共1993兲. “Bond and slip of FRP
rebars in concrete.” J. Mater. Civ. Eng., 5共1兲, 30–40.
Malvar, L. J. 共1995兲. “Tensile and bond properties of GFRP reinforcing
References bars.” ACI Mater. J., 92共3兲, 276–285.
Okelo, R. 共2002兲. “Development and splice lengths criteria for straight
Benmokrane, B., Tighiouart, B., and Chaallal, O. 共1996兲. “Bond strength fiber reinforced polymer composite bars in tension.” PhD dissertation,
and load distribution of composite GFRP reinforcing bars in con- The Univ. of Texas at Arlington, Arlington, Tex.
crete.” ACI Mater. J., 93共3兲, 246–253. Pecce, M., Manfredi, G., Realfonzo, R., and Cosenza, E. 共2001兲. “Experi-
Brown, V. L., and Bartholomew, C. L. 共1993兲. “FRP reinforcing bars in mental and analytical evaluation of bond properties of GFRP bars.” J.
reinforced concrete members.” ACI Mater. J., 90共1兲, 34–39. Mater. Civ. Eng., 13共4兲, 282–290.
De Lorenzis, L., and Nanni, A. 共2001兲. “Characterization of FRP rods Yuan, R. L., and Okelo, R. 共2001兲. “Round robin tests on the bond
as near-surface mounted reinforcement.” J. Compos. Constr., 5共2兲, strength of FRP reinforcement.” Technical Rep. for the Texas Dept. of
14–21. Transportation (TxDOT), Austin, Civil Engineering Dept., The Univ.
De Lorenzis, L., and Nanni, A. 共2002兲. “Bond between near-surface of Texas at Arlington, Arlington, Tex.
mounted fiber-reinforced polymer rods and concrete in structural Zhang, B., and Benmokrane, B. 共2000兲. “Prediction of tensile capacity of
strengthening.” ACI Struct. J., 99共2兲, 123–132. bond anchorages for FRP tendons.” J. Compos. Constr., 4共2兲, 39–47.

JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / MAY/JUNE 2005 / 213

J. Compos. Constr., 2005, 9(3): 203-213

You might also like