Construction and Building Materials 25 (2011) 3956–3965

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Testing of full-scale concrete bridge deck slabs reinforced with fiber-reinforced

polymer (FRP) bars
K. Bouguerra, E.A. Ahmed, S. El-Gamal, B. Benmokrane ⇑
Dept. of Civil Engineering, University of Sherbrooke, Quebec, Canada

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents an experimental study investigating the behavior of FRP-reinforced concrete bridge
Received 16 November 2010 deck slabs under concentrated loads. A total of eight full-scale deck slabs measuring 3000-mm long by
Received in revised form 23 March 2011 2500-mm wide were constructed. The test parameters were: (i) slab thickness (200, 175 and
Accepted 14 April 2011
150 mm); (ii) concrete compressive strength (35–65 MPa); (iii) bottom transverse reinforcement ratio
Available online 10 May 2011
(1.2–0.35%); and (iv) type of reinforcement (GFRP, CFRP, and steel). The slabs were supported on two par-
allel steel girders and were tested up to failure under monotonic single concentrated load acting on the
Keywords:
center of each slab over a contact area of 600  250 mm to simulate the footprint of sustained truck
Fiber-reinforced polymer (FRP)
Concrete
wheel load (87.5 kN CL-625 truck). All deck slabs failed in punching shear. The punching capacity of
Bridge deck slab the tested deck slabs ranged from 1.74 to 3.52 times the factored load (Pf) specified by the Canadian High-
Design way Bridge Design Code (CHBDC) CAN/CSA S6-06. Besides, the ACI 440.1R-06 punching strength equation
Punching strength greatly underestimated the capacity of the tested slabs with an average experimental-to-predicted
punching capacity ratio (Vexp/Vpred) of 3.17.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction cutting-edge technology has emerged as one of the most cost-

effective alternatives to traditional solutions.
Corrosion of steel reinforcement in concrete structures is a ma- Recent advances in polymer technology have led to the devel-
jor durability problem that leads to structural degradation and opment of the latest generation of FRP reinforcing bars (in partic-
consequent costly repairs. In a bridge system, concrete bridge ular glass FRP (GFRP) bars). These corrosion-resistant bars have
decks deteriorate faster than any other bridge component due to shown promise as a way to further protect bridges and public
direct exposure to environment, de-icing salts and ever-increasing infrastructure from the devastating effects of corrosion. With FRP
traffic loads. A considerable amount of North America’s transporta- being standardized through the recently published CSA specifica-
tion infrastructure is in need of repair or replacement and consti- tions (CAN/CSA S807-10 [3]) and bars of the highest quality being
tutes a major problem when measured in terms of rehabilitation produced, FRP bars are emerging as a realistic and cost-effective
costs and traffic disruption [1]. The American Society of Civil Engi- alternative to traditional steel reinforcement for concrete struc-
neers (ASCE) estimates that over $1.6 trillion would be needed in tures under severe environmental conditions.
the next five years in the U.S. to alleviate potential problems with At the Department of Civil Engineering, University of Sher-
civil infrastructure. Canada’s deficit for its municipal infrastruc- brooke, a multi-phase experimental program investigating the
ture, which represents 70 percent of the country’s total infrastruc- structural performance of restrained FRP-reinforced concrete
ture was estimated to be $60 billion in 2004, and is expected to bridge deck slabs under different loading conditions has been com-
grow by $2 billion dollars per year (ISIS Canada [2]). Worldwide, pleted. The first phase [4] investigated the restrained FRP-rein-
governments and industrial firms are looking for infrastructure forced concrete bridge deck slabs under static loading condition
systems that are stronger, longer lasting, more resistant to corro- acting at the center of the slab on an area equivalent to a wheel
sion and less costly to build and maintain. Engineers all over the load from a CL-625 truck specified by the CHBDC CAN/CSA S6-06
world are searching for new and affordable construction materials [5]. The second phase [6] investigated the behavior of the re-
as well as innovative approaches and systems to solve these prob- strained FRP-reinforced concrete bridge deck slabs under constant
lems. An effective solution to this problem is corrosion-resistant and variable amplitude fatigue loading. The load was also applied
materials, such as fiber-reinforced polymer (FRP) composites. This at the center of the slab on an area equivalent to a wheel load from
a CL-625 truck specified by the CHBDC CAN/CSA S6-06 [5]. The
aforementioned two phases [4,6] considered different reinforce-
⇑ Corresponding author. ment types and ratios maintaining a 200-mm slab thickness for
E-mail address: [email protected] (B. Benmokrane). all test specimens and using normal strength concrete.

0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.04.028
 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965 3957

Recently, there is an increase in the use of high-strength con- in the reference slab for comparison. The tensile properties of the reinforcing FRP
bars were determined by testing of representative specimens in accordance with
crete (HSC) in bridges due to its superior strength and stiffness.
ACI 440 3R-04 B.2 [9] Test Method ‘‘Test Method for Longitudinal Tensile Properties
Carpenter [7] reported that the use of high-strength concrete of FRP Bars.’’ Table 1 summarizes the tensile properties of the reinforcing bars used
(HSC) in bridge applications can result in: (1) greater compressive in this study.
strength per unit cost, per unit weight, and per unit volume; (2) The test slabs were divided into four groups. The first and second groups were
higher modulus of elasticity resulting in reduced deflection; and designed to investigate the effect of slab thickness and concrete compressive
strength, respectively. The third and fourth group, however, were designed to inves-
(3) increased tensile strength. The enhanced properties which in-
tigate the effect of reinforcement ratio and type. Slab designation is as follows: The
cludes compressive and tensile strengths, stiffness, and modulus first letter indicates reinforcement type (S for steel, G for Glass FRP, and C for Car-
of the HSC would contribute to achieving the design capacity with bon FRP), the middle number indicates the thickness of the slab followed by a letter
smaller dimensions which reduces the load that has to be resisted representing the strength of the concrete (N for normal strength and H for high
strength). Numbers at the end (only in Group 3) represent the reinforcement ratio
by the other supporting elements. Thus, the third phase of this
when not equal to 1.2%.
experimental program was planned to investigate the effect of slab The first group, Group 1, comprised three normal strength concrete deck slabs
thickness, reinforcement type and ratio, and the concrete strength (G-200-N, G-175-N and G-150-N) with three different thicknesses (200, 175 and
(normal and high-strength concretes) on the structural perfor- 150 mm), respectively. Group 2 comprised a deck slab; G-175-H with identical rein-
mance and capacity of restrained FRP-reinforced bridge deck slabs. forcement to the G-175-N in Group 1; however 64.8 MPa HSC was used for the G-
175-H slab. For the slabs of these two groups (Group 1 and Group 2), the bottom
The test results of this phase are presented and discussed herein.
transverse GFRP reinforcement ratio was 1.2% while the GFRP reinforcement ratio
in all other directions was set to 50% of the bottom transverse reinforcement which
yield a reinforcement ratio of 0.6%. Group 3 comprised two slabs reinforced with
2. Experimental program
GFRP bars: G-175-N-0.7 and G-175-N-0.35 with bottom transverse reinforcement
ratios equal to 0.7 and 0.35%, respectively. The fourth group comprised two slabs,
2.1. Test specimens
(S-175-N and C-175-N), reinforced with different types of reinforcing bars; steel
and CFRP, respectively. The reinforcement ratio in the other directions for the
A total of eight full-scale concrete deck slabs of 3000-mm long  2500-mm
FRP-reinforced slabs of Group 3 and 4 was maintained to 0.35%. The dimensions
wide  (150, 175, and 200-mm) thick, as shown in Fig. 1, were constructed and
and reinforcement details of the deck slabs are shown in Fig. 1 while the complete
tested. All slabs were reinforced with two mats of orthogonal reinforcing bars.
characteristics of the deck slab are listed in Table 2.
Three different diameter of glass FRP (GFRP) bars were used and designated as
No. 13 (12.7 mm), 15 (15.9 mm), and 20 (19.1 mm) according to the CAN/CSA
S807-10 [3] and one diameter of carbon FRP (CFRP) bars designated as No. 10 2.2. Test setup and procedure
(9.5 mm) [8]. The bars were fabricated using a pultrusion process with a fiber con-
tent of 73% and 78% by weight in a vinyl ester resin for the GFRP and CFRP bars, Similar to the first two phases of this research program [4,6] and to simulate
respectively. The FRP bars had a sand-coated surface to enhance bond performance bridge deck slabs with restrained edges, the slabs were supported on two steel gird-
between the bars and surrounding concrete. In addition, 10 M steel bars were used ers spaced at 2000 mm center-to-center. The girders were connected together by

2500 mm GFRP No. 15 @ 200 mm GFRP No. 13 @ 255 mm

25 25
150 200 125 175
25 25
GFRP No. 20 @ 150 mm GFRP No. 15 @ 200 mm
G-200-N G-175-N-0.7
GFRP No. 15 @ 230 mm GFRP No. 13 @ 255 mm

25 25
125 175 125 175
25 25
600
A
GFRP No.15 @ 115 mm GFRP No. 13 @ 255 mm
A G-175-N G-175-N-0.35
250

GFRP No. 15 @ 270 mm Steel 10M @ 230 mm

25 25
100 150 125 175
25 25
GFRP No. 15 @ 140 mm Steel 10M @ 230 mm
G-150-N S-175-N
GFRP No. 15 @ 230 mm GFRP No. 15 @ 230 mm

25 25
125 175 125 175
25 25
GFRP No. 15 @ 115 mm CFRP No. 10 @ 125 mm
G-175-H C-175-N
(a) Typical dimensions of the tested slabs (b) Section A-A
Fig. 1. Dimensions and reinforcement details of the tested slabs.

Table 1
Mechanical properties of the reinforcing bars.

Bar typea Bar designation Diameter (mm) Areaa (mm2) Modulus of elasticity (GPa) Tensile strength (MPa) Ultimate strain (%)
b
Steel 10 M 11.3 100 200 fy = 453 eyb = 0.23
GFRP No. 13 12.7 129 41.0 ± 2.1 769 ± 33 1.87 ± 0.02
No. 15 15.9 199 41.6 ± 1.5 778 ± 16 1.87 ± 0.01
No. 20 19.1 284 44.5 ± 1.3 637 ± 15 1.37 ± 0.03
CFRP No. 10 9.5 71 122 ± 2.4 1444 ± 18 1.23 ± 0.07
a
FRP bars are designated according to the CSA S807-10 [3].
b
fy and ey are the yield strength and yield strain of the steel bars, respectively.
3958 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965

Table 2
Concrete strength and reinforcement details of the tested deck slabs.

Group Slab Slab thickness (mm) fc0 ðMPaÞ Reinforcement ratio (%) Reinforcement configuration
Bottom transverse direction (axial stiffness, MN/m) Other directions
1 G-200-N 200 49.1 1.20 No. 20 @ 150 mm (84.2) No. 15 @ 200 mm
G-175-Na 175 35.2 1.20 No. 15 @ 115 mm (72.0) No. 15 @ 230 mm
G-150-N 150 35.2 1.20 No. 15 @ 140 mm (59.1) No. 15 @ 270 mm
2 G-175-H 175 64.8 1.20 No. 15 @ 115 mm (72.0) No. 15 @ 230 mm
3 G-175-N-0.7 175 53.1 0.70 No. 15 @ 200 mm (41.4) No. 13 @ 255 mm
G-175-N-0.35 175 53.1 0.35 No. 13 @ 255 mm (20.7) No. 13 @ 255 mm
4 S-175-N 175 42.3 0.30 10 M @ 230 mm (86.9) 10M@230 mm
C-175-N 175 40.3 0.40 No. 10 @ 125 mm (69.3) No. 15 @ 230 mm
a
Used for comparison with Groups 2, 3, and 4.

two cross frames spaced at 3000 mm. Each cross frame consisted of an X-shaped spaced 180 mm apart with a pitch of 250 mm. To ensure equal and uniform edge
bracing with 55  55  6 mm angles, as shown in Fig. 2a. Each slab was bolted to clamping force, all bolts were hand tightened to an equal torque of 160 N m using
the top flange of the steel girders through two rows of holes at each edge using a torque wrench. In addition, a 3-mm thick neoprene pad was used between the
25-mm diameter steel bolts and two steel channels. The two rows of the holes were steel sections (channel and top flange of girder) and the concrete slab. The steel

Fig. 2. Test setup: (a) Supporting system. (b) Restrained edge details.
 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965 3959

diameter equal to the spacing between the two girders as shown

in Fig. 4.
Generally, the carrying capacities of the tested deck slabs ran-
ged from 362 to 732 kN representing 1.74–3.52 times the factored
design load (Pf = 208.25 kN) specified by the CHBDC CSA S6-06 [5].
The high level of conservativeness is due to the general behavior of
the restrained bridge deck slabs and mode of failure. Hewitt [10],
Hewitt and Batchelor [11]; and Perdikaris and Beim [12] reported
that reinforced concrete restrained bridge deck slabs subjected to
concentrated wheel loads failed in punching shear rather than flex-
ure as assumed by conventional design. Furthermore, Hewitt and
Batchelor [11] reported that when a deck slab is restrained, there
is no need for reinforcement to resist the wheel loads due to the
compressive membrane action which is similar to the arching ac-
tion in the reinforced concrete beams. A minimum amount of rein-
forcement was proposed for serviceability reasons. This concept
was adopted by the Canadian Highway Bridge Design Code (CAN/
CSA S6-06 [5]) through its empirical method. However, this meth-
od does not account for any design parameters other than the
effective depth of the slab and the modulus of elasticity of the rein-
Fig. 3. Testing of a slab specimens. forcing bars such as the concrete strength or the effective slab
span. Ahmed and Benmokrane [13] concluded that the empirical
bolts replace shear studs, which provide edge restraint to the slab in both longitu- method in most cases overestimates the required bottom trans-
dinal and transverse directions [4]. Fig. 2b shows the details of the edge-restraint verse reinforcement.
configuration that were utilized herein. The deck slabs were tested up to failure To evaluate the design methods of the Canadian Highway
using one concentrated load over a contact area of 600  250 mm to simulate the
Bridge Design Code CAN/CSA S6-06 [5], the bridge deck slabs are
foot print of the truck wheel load; CL-625 Truck (CAN/CSA S6-06) acting on the cen-
ter of each slab. In all tests, the load was applied through a 60-mm thick steel plate designed using the flexural design method and the empirical de-
over a 10-mm thick neoprene pad. The load was monotonically applied at a load sign method. As the deck slab specimens were edge restrained,
controlled rate of 5 kN/min. During loading, the formation of cracks on the bottom they were designed as a continuous slab over girders. The required
surface of the deck slabs were marked and recorded. Fig. 3 shows a photo of the test amount of reinforcing bars determined by those methods is listed
setup during the testing of a slab specimen.
in Table 4. From this table it can be noticed that, the empirical de-
sign method in most cases yields higher reinforcement amount
3. Test results and discussion than the flexural design method which was also reported by
Ahmed and Benmokrane [13]. This is referred to the fact that the
The test results are summarized in Table 3 and the final crack empirical design method is affected only by the thickness of the
patterns of the slabs at failure are shown in Fig. 4. However, the concrete section and the tensile modulus of the reinforcing bars.
following section will discuss the results and the effect of the dif- Increasing the depth to enhance the performance or decreasing
ferent parameters on the behavior of the tested slabs. The design the effective spacing between the girders will not be reflected on
service load, Pser, of the deck slabs was taken as 1.4  0.9  the required amount of reinforcement. On the other hand, chang-
87.5 = 110.25 kN, where 87.5 is the maximum wheel load of the ing the any of those parameters will affect the reinforcement
design truck (CL-625 Truck – CAN/CSA S6-06 [5]), 1.4 is the impact amount resulted from the flexural design method. However, the
coefficient, and 0.9 is the live load combination factor. The design design of the bridge deck slab is often controlled by the crack
factored load was taken as Pf = 1.4  1.7  87.5 = 208.25 kN, where width criterion when the flexure design method is employed [13].
1.7 is the live load combination factor (CAN/CSA S6-06 [5]). From Table 4 it can be noticed that, even though the failure load
of the tested slabs was higher than the factored load of the CAN/
3.1. Ultimate capacity and mode of failure CSA S6-06 [5], the design of the bridge deck slabs using flexural de-
sign method gave more reinforcement than what was provided in
For all tested deck slabs, the failure mode was punching shear some specimens. Of interest, the flexural design of slabs G-175-N
around the loaded area. The top surface of the failure zone had and G-175-H yielded the same amount of reinforcement as the
an elliptical shape passing through the corners of the loaded area. high-strength concrete did not affect the calculated amount of
The bottom surface had an approximately circular shape with a reinforcement. This is due to the fact that the design of the section

Table 3
Summary of the test results.

Group Slab Cracking load (kN) Ultimate load (kN) Ultimate load/Pf Maximum deflection (mm) Maximum strain at service load (le)
Service Failure Bars Concrete
1 G-200-N 115 732 3.52 0.45 22.9 624 180
G-175-N* 113 484 2.32 0.72 17.9 214 200
G-150-N 107 362 1.74 1.49 17.1 552 200
2 G-175-H 130 704 3.38 0.45 21.7 214 140
3 G-175-N-0.7 118 549 2.64 0.48 20.4 223 120
G-175-N-0.35 98 506 2.43 1.04 26.4 1293 298
4 S-175-N 121 550 2.64 0.52 23.8 78 146
C-175-N 103 530 2.55 0.78 16.7 905 276
*
Used for comparison with Groups 2, 3, and 4.
3960 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965

(a) G-150-N (b) G-175-N

(c) G-150-N (d) G-175-N

(e) G-175-H (f) G-175-N-0.35

(g) C-175-N (h) S-175-N

Fig. 4. Crack pattern of the slab specimens at failure.

was based on failure in flexure when in fact the failure mode was though the sections are designed to resist the same straining ac-
punching shear. Considering the actual mode of failure in design tions. Besides, for slabs with thickness of 150 mm there was a
may lead to reducing the FRP reinforcement amount which will re- big difference between the reinforcement amount calculated from
sult in reducing the initial costs of the FRP-reinforced concrete the flexural and empirical methods.
bridge deck slabs. Furthermore, the slab G-15-N with a 150 mm The test results revealed that the ultimate carrying capacity of
thickness, which is less than the CHBDC (CAN/CSA S6-06 [5]) min- the slab specimens was affected by the slab thickness, reinforce-
imum allowable thickness of 175 mm, gave very high reinforce- ment ratio, and concrete strength. Table 3 shows that the punching
ment ratio (2%) when designed using the flexural method. capacities of the tested slabs were proportional to the slab thick-
Also, the empirical method of the CAN/CSA S6-06 [5] does not ness. In Group 1, when the slab thickness was decreased by
account for the concrete strength. Of interest, increasing the slab 12.5% (from 200 to 175 mm) for slabs G-200-N and G-175-N and
thickness resulted in increasing the reinforcement amount even the ultimate capacity decreased by about 34%. Similarly when

Table 4
Design of the bridge deck slabs according to the CHBDC (CSA S6-06) [5].

Slab Thickness (mm) fc0 ðMPaÞ Bottom transverse direction Bottom transverse direction
Flexural design method Empirical design method
G-200-N 200 49.1 GFRP No. 15 @ 135 mm GFRP No. 15 @ 100 mm
G-175-N 175 35.2 GFRP No. 15 @ 120 mm GFRP No. 15 @ 115 mm
G-175-H 175 64.8 GFRP No. 15 @ 120 mm GFRP No. 15 @ 115 mm
G-150-N 150 35.2 GFRP No. 15 @ 85 mm GFRP No. 15 @ 140 mm
 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965 3961

800 120

700

Expiremental punching
100

Normalized Punching
capacity, Vexp (kN)

√ f'c
600

Capacity, Vexp /√
80
500

400 60

300
40
200 Vexp
Vexp /√f'c 20
100

0 0
125 150 175 200 225
Slab thickness (mm)

(a)
800

700
Expiremental punching

600
capacity (kN)

500

400

300

200

100

0
0 20 40 60 80 100
Concrete compressive strength (MPa)

(b)
560 84

550 82
Expiremental punching

Normalized Punching
capacity, Vexp (kN)

Capacity, Vexp /√f'c

540 80

530 78

520 76

510 74

500 72
Vexp
490 70
Vexp /√f'c
480 68
0 0.25 0.5 0.75 1 1.25 1.5
Reinforcement ratio %
(c)
Fig. 5. Experimental punching capacity versus: (a) Slab thickness (Group 1). (b) Concrete compressive strength (Group 2). (c) Bottom transverse reinforcement ratio (Group
3).

the slab thickness was decreased by 25% (from 200 to 150 mm) for sive strength by 83% (from 35.4 to 64.8 MPa) increased the punch-
slabs G-200-N and G-150-N the ultimate capacity decreased by ing capacity by 45% (from 484 to 704 kN) as shown in Table 3. This
about 50%. The punching capacity versus the slab thickness rela- is due to the dependency of the punching shear strength of rein-
tionship in Fig. 5a also shows that the punching capacity is directly forced concrete slabs on the concrete strength. However, this is
proportional to the slab thickness. The higher the slab thickness not reflected in the flexural design method incorporated in the
the higher the punching capacity. This is referred to that increasing Canadian Highway Bridge Design Code (CSA S6-06 [5]).
the slabs thickness results in increasing the surface area that re- For the slabs
pffiffiffiffi of Group 3, the normalized punching capacity
sists the punching shear stresses which yields higher punching curve (V exp = fc0 ) in Fig. 5c shows that punching capacity is propor-
shear capacity. tional to reinforcement ratio. It can be noticed that increasing the
For slabs of Group 2, the punching capacity increased with the reinforcement ratio increased the punching capacities of the tested
increase in concrete compressive strength as shown in Fig. 5b. slabs. For Group 4, it can be noticed that the type of reinforcement
For slabs G-175-N and G-175-H, increasing the concrete compres-
3962 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965

800 800
700 700

Load (kN)
600 600
Load (kN)

500 500

400 400

300 G-200-N 300

G-200-N
G-175-N 200 G-175-N
200
G-150-N
G-150-N 100
100
0
0 -3000 -1500 0 1500 3000 4500 6000 7500 9000 10500 12000
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5
Strains (micro-strain)
Crack Width (mm)
(a) (a)
800 800

700 700

600 600

Load (kN)
Load (kN)

500
500
400
400
300
300
200 G-175-N
200 G-175-N
G-175-H 100 G-175-H
100
0
0 -3000 -1500 0 1500 3000 4500 6000 7500 9000 10500 12000
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 Strains (micro-strain)
Crack Width (mm)
(b)
(b)
800
800
700
700
600
Load (kN)

600
500
Load (kN)

500
400
400
300
G-175-N
300
G-175-N 200 G-175-N-0.7
200 100 G-175-N-0.35
G-175-N-0.7
100 G-175-N-0.35 0
0 -3000 -1500 0 1500 3000 4500 6000 7500 9000 10500 12000
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 Strains (micro-strain)
Crack Width (mm)
(c)
(c)
800
800
700
700 600
Load (kN)

600 500
Load (kN)

500 400
400 300 S-175-N
300 200 G-175-N
S-175-N
C-175-N
200 G-175-N 100
100 C-175-N 0
-3000 -1500 0 1500 3000 4500 6000 7500 9000 10500 12000
0
0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 Strains (micro-strain)
Crack Width (mm) (d)
(d) Fig. 7. Load–maximum strains relationships for tested slabs: (a) Group 1. (b) Group
2. (c) Group 3. (d) Group 4.
Fig. 6. Load–crack width relationships for test deck slabs: (a) Group 1. (b) Group 2.
(c) Group 3. (d) Group (4).

3.2. Cracking behavior

did not show a significant effect on the punching capacities of the All tested slabs had almost similar cracking patterns. The first
tested slabs. cracks appeared directly under the loaded area and were oriented
 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965 3963

800.0 with a lower cracking load was observed in slab G-175-N-0.35,

700.0 the slab with the lowest reinforcement ratio. The slabs of Group
4, with different types of reinforcement, also showed similar crack
600.0
patterns. In this group, slab C-175-N had a lower cracking load
Load (kN)

500.0 than the two other slabs. The lower cracking loads observed in
400.0 slabs G-175-N-0.35 (Group 3) and C-175-N (Group 4) could be ex-
300.0 plained by the presence of hairline shrinkage cracks that were ob-
served in the slabs before testing.
200.0 G-200-N Fig. 6 shows the variation of maximum measured crack widths
100.0 G-175-H against applied loads for all tested slabs. The load-crack width rela-
0.0 tionships were almost linear for all FRP-reinforced concrete slabs.
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 Slab S-175-N reinforced with steel showed a bilinear curve due
Crack Width (mm) to the yielding of the steel bars. Fig. 6a shows the load-crack width
curves of the slabs of Group 1. It can be noticed that slabs G-200-N
(a)
and G-175-N had almost similar crack width values. Higher crack
width values were recorded in slab G-150-N. This indicates that
800.0
decreasing the thickness from 200 to 175 mm did not affect crack
700.0
widths. However, decreasing the thickness to 150 mm, which is
600.0 less than the CHBDC (CAN/CSA S6-06 [5]) minimum allowable
Load (kN)

500.0 thickness of 175 mm, resulted in wider cracks. For the two slabs
400.0
of Group 2, Fig. 6b shows that the load-crack width relationships
of slabs G-175-N and G-175-H were similar up to failure. This indi-
300.0
cates that the concrete strength did not have significant influence
200.0 G-200-N on the crack widths. Fig. 5c shows the load-crack width curves of
100.0 G-175-H Group 3 slabs. It can be clearly seen that decreasing the bottom
0.0
transverse reinforcement ratio by increasing the spacing between
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25 bars significantly increased crack width values. This is in good
Deflection (mm) agreement with several investigations which directly specify the
maximum bar spacing to control cracking [14,15].
(b) The load-crack width curves of the slabs of Group 4 are shown
in Fig. 6d. The three slabs of this group had similar axial stiffness of
800
the bottom transverse reinforcement. Slabs of the same axial stiff-
700 ness may demonstrate the same behavior, however, slab S-175-N
reinforced with steel, showed higher crack widths after steel
Load (kN)

600
500 yielding.

400
300 3.3. Strains in reinforcement and concrete
200
G-200-N
Fig. 7 shows the load-maximum concrete and reinforcement
100 G-175-H
strains for the tested slabs. Before cracking, all the tested slabs
0 showed almost similar concrete and reinforcement strains. After
-3000 -2000 -1000 0 1000 2000 3000 4000 5000 6000 7000 8000
cracking, different strains were recorded. At failure, the maximum
Strains (micro-strain)
strains in the bottom transverse reinforcement were measured in
(c) slab G-175-N-0.35 (Group 3) which had the lowest reinforcement
ratio and slab S-175-N (Group 4) due to the yielding of the steel
Fig. 8. Comparison between G-200-N and G-175-H: (a) Crack width. (b) Deflection. reinforcement as listed in Table 3. For all FRP-reinforced slabs,
(c) Strains.
the maximum measured strains in the bars at service and at fac-
tored load levels were about 9% and 20% of their ultimate strains,
respectively. At failure, the maximum measured strains in those
in the longitudinal direction parallel to the supporting beams. As bars were about 66 % of their ultimate strains. These maximum
the load increased, subsequent cracks propagated in the radial strains were measured in slab G-175-N-0.35.
direction away from the loaded area. For all test slabs, the cracking Fig. 7a shows that increasing slab thickness decreased the mea-
loads ranged between 98 to 130 kN as listed in Table 3. For the sured strains in the GFRP bars after cracking. At the factored design
three slabs of Group 1, the cracking loads were greater than the load level (Pf = 208.25 kN), the maximum measured strains in the
service load (Pser), except for slab G-150-N with a 150-mm thick- reinforcement were 1990, 2390 and 2495 micro-strain for slabs
ness. It should be mentioned that 150 mm is below the CAN/CSA G-200-N, G-175-N, and G-150-N, respectively. These values were
S6-06 [5] minimum slab thickness of 175 mm. For the two slabs 30%, 38%, and 51% of the measured strains at failure (6690, 6224,
of Group 2, it was noticed that the slab G-175-H had more cracks and 4872 micro-strain). This provides an ample warning and safety
on its tension face. This could be related to its HSC which increased factors of about 3.33, 2.63, and 1.96 before failure, respectively.
the bond strength between the GFRP bars and the concrete compo- Fig. 7b shows that lower strains were recorded in slab G-175-H
nents, consequently, increasing the number of cracks. It was also compared to slab G-175-N. This indicates that as the concrete
observed that increasing the concrete strength from 35.2 MPa (G- strength was increased, the measured reinforcement and concrete
175-N) to 64.8 MPa (G-175-H) increased the cracking load from strains decreased. At the factored design load level, the maximum
114 to 130 kN. For the slabs of Group 3, with different bottom measured strains in the reinforcement were 2390 and 1575 micro-
transverse GFRP reinforcement ratios, less cracks accompanied strains for slabs G-175-N and G-175-H, respectively. These values
3964 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965

Table 5
Punching strength capacity models of FRP-RC elements.

Reference Equation
JSCE [16] Vc = bdbpbrfpcdbo;0.5dd
pffiffiffiffi
bd = (1000/d)1/4 6 1.5, bp = (100qfEf/Es)1/3 6 1.5, br = 1 + 1/(1 + 0.25u/d), fpcd = 0.2 fc0 6 1.2 MPa
El-Ghandour et al. [17,18] pffiffiffiffiEf 1=3
V c ¼ 0:33 fc0 Es bo;0:5d d
Matthys and Taerwe [19] ð100qEf =Es fc0 Þ1=3
V c ¼ 1:36 1=4 bo;1:5d d
d
qffiffiffiffi
Ospina et al. [20] E
V c ¼ 2:77ðqf fc0 Þ1=3 Efs bo;1:5d d
pffiffiffiffi
El-Gamal et al. [21] V c ¼ 0:33 fc0 bo0:5d da
a = 0.5(qfEf)1/3(1 + 8d/bo;0.5d)
pffiffiffiffi
ACI 440.1R-06 [22] V c ¼ 45 fc0 bo;0:5d kd
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
k ¼ 2qf nf þ ðqf nf Þ2  qf nf

Note: Vc is the punching shear capacity (N); Ef is the modulus of elasticity of the FRP reinforcement (MPa); Es is the modulus of
elasticity of steel (MPa); d is the effective slab depth; qf is the FRP reinforcement ratio; u is the perimeter of the loaded area (mm); fc0
is the compressive strength of the concrete (MPa); bo;0.5d is the critical perimeter at a distance of 0.5d from the column face (mm);
bo;1.5d is the critical perimeter at a distance of 1.5d from the column face (mm); nf is the ratio of modulus of elasticity of FRP bars to
modulus of elasticity of concrete.

Table 6
Comparison between predicted and experimental punching shear strengths.

Slab Vexp (kN) JSCE [16] El-Ghandour et al. [17,18] Matthys and Taerwe [19] Ospina et al. [20] El-Gamal et al. [21] ACI 440.1R-06 [22]
Vexp/Vpred Vexp/Vpred Vexp/Vpred Vexp/Vpred Vexp/Vpred Vexp/Vpred
G-200-N 732 1.51 1.44 1.59 1.23 1.13 2.57
G-175-N 484 1.32 1.28 1.53 1.06 1.03 2.68
G-150-N 362 1.29 1.21 1.45 1.06 1.02 2.55
G-175-H 704 1.92 1.37 1.82 1.26 1.10 3.28
G-175-N-0.75 549 1.79 1.18 1.81 1.26 1.13 3.43
G-175-N-0.35 506 1.99 1.04 2.02 1.39 1.26 4.18
C-175-N 530 1.49 0.93 1.65 0.97 1.08 3.47
Mean 1.62 1.21 1.70 1.18 1.11 3.17
S.D. 0.28 0.18 0.20 0.15 0.08 0.60
COV (%) 18 15 12 13 7 19

Vexp: Experimental punching strength.

Vpred: Predicted punching strength.

were about 38% and 21% of the measured strains at failure (6224 3.5. Comparison between experimental and predicted punching
and 7566 micro-strain). strengths
Fig. 7c shows that decreasing the bottom transverse reinforce-
ment ratio significantly increased the measured reinforcement The punching shear strengths of the tested bridge deck slabs
strains after cracking. At service load level, much greater strains were predicted using the available models in the literature that
were recorded in slab G-175-N-0.35 because it was cracked at a predict the punching strength of FRP-reinforced slabs. This in-
load less than the service load. At factored design load level, the cludes the models of the Japan Society of Civil Engineers [16], El-
maximum measured reinforcement strains were 2390, 2969, and Ghandour et al. [17,18], Matthys and Taerwe [19], Ospina et al.
3831 micro-strains for slabs G-175-N, G-175-N-0.7 and G-175-N- [20], El-Gamal et al. [21], and ACI 440.IR-06 [22] Table 5 lists these
0.35, respectively. For the slabs of Group 4, Fig. 7d shows that al- models that were used to predict the punching capacity of FRP-
most similar strains were recorded in the FRP-reinforced concrete reinforced concrete slabs tested herein. The predicted punching
slabs. For slab S-175-N, slightly lower strains were recorded before shear strengths were compared to the experimental values as gi-
yielding of the steel bars. After yielding, greater strain values were ven in Table 6. It can be noticed that the model proposed by El-
recorded. Gamal et al. [21] yielded good yet conservative prediction with
an average ratio of Vexp/Vpred os 1.11 and corresponding coefficient
of variation of 7%. On the other hand, the ACI 440.1R-06 [22]
3.4. Comparison between G-200-N and G-175-H slabs
underestimates the punching strength of bridge deck slabs rein-
forced with FRP bars. The average ratio of Vexp/Vpred is 3.17 with a
The behaviors of slabs G-200-N and G-175-H were compared in
coefficient of variation of 19%.
deflection, cracking, and strain as shown in Fig. 8. From this com-
parison it can be seen that the two slabs with the same quantity of
reinforcement but with different thicknesses and fabricated using 4. Conclusions
normal and high-strength concrete respectively, showed the same
behavior. It may be concluded that a reduction in the deck slab This paper presented the experimental test results of eight full-
thickness may be allowed by increasing the concrete strength. scale bridge deck slabs reinforced with FRP bars. The slabs were di-
However, this is not reflected in the design procedures of the vided into four groups to investigate the following parameters:
CSA S6-06 as previously mentioned. slab thickness, concrete compressive strength, reinforcement ratio,
 K. Bouguerra et al. / Construction and Building Materials 25 (2011) 3956–3965 3965

and reinforcement type. The test results were presented in terms of Also, many thanks for Pultrall Inc. (Thetford Mines, Quebec) for
cracking, deflection, strains in concrete and reinforcement, mode of generously providing the FRP materials. Special thanks to François
failure and ultimate capacity. The main findings of this investiga- Ntacorigira, technician at the Department of Civil Engineering,
tion can be summarized as follows: University of Sherbrooke, for his help in the fabrication and testing.

1. The entire group of tested slabs failed in punching shear failure References
with a very similar cracking pattern except slab G-175-H which
showed more cracks on its tension face. This could be related to [1] Yunovich M, Thompson N. Corrosion of highway bridges: economic impact and
control methodologies. Am Concr Inst, ACI Concr Int 2003;25(1):52–7.
the use of HSC which increased the bond strength between the [2] Intelligent Sensing for Innovative Structures (ISIS). Durability of fibre
GFRP bars and the concrete components. reinforced polymers in civil infrastructure. Durability monograph. Canadian
2. The bottom transverse reinforcement ratio was the main network of centers of excellence on intelligent sensing for innovative
structures. MB (Canada): University of Manitoba; 2006.
parameter affecting crack widths. Decreasing the bottom trans- [3] Canadian Standard Association (CSA). Specification for fibre-reinforced
verse reinforcement ratio significantly increased crack widths. polymers. CAN/CSA S807-10. Rexdale, Ontario, Canada; 2010. 27 p.
3. The measured strains were affected by slab thickness, concrete [4] El-Gamal S. Behavior of restrained concrete bridge deck slabs reinforced
with reinforcing bars under concentrated load. Ph.D. Thesis, Sherbrooke
strength, and reinforcement ratio. For all FRP-reinforced slabs, (Quebec): Department of Civil Engineering, Université de Sherbrooke; 2005.
the maximum measured strains in the bars at service load level, 227 p.
factored load level, and at failure were about 9%, 20%, and 66% [5] Canadian Standard Association (CSA). Canadian highway bridge design code.
CAN/CSA-S6-06. Rexdale, Ontario, Canada; 2006.
of their ultimate strains, respectively. These maximum strains
[6] El-Ragaby A. Fatigue behavior of concrete bridge deck slabs reinforced with
were measured in slab G-175-N-0.35 with the smallest bottom glass FRP reinforcing bars. Ph.D. Thesis, Department of Civil Engineering,
transverse reinforcement ratio. Université de Sherbrooke, Sherbrooke, Quebec; 2007. 171 p.
[7] Carpenter JE. Applications of high-strength concrete for highway bridges.
4. The punching capacities of the tested slabs were significantly
Public Roads 1980;44(2):76–83.
affected by the slab thickness and the concrete compressive [8] Pultrall. V-ROD FRP reinforcing bars data sheet; 2008. <www.pultrall.com>.
strength. Decreasing the deck slab thickness by 12.5% and 25% [9] ACI Committee 440. Guide test methods for fiber reinforced polymer (FRP) for
decreased the punching capacities by 34% and 51%, respectively. reinforcing or strengthening concrete structures. ACI 440.3R-04. Farmington
Hills (MI, USA): American Concrete Institute; 2004. 41 p.
For the slabs of Group 2, increasing the concrete compressive [10] Hewitt BE. An investigation of the punching strength of restrained slabs with
strength by 83% increased the punching capacity by 45%. particular reference to the deck slabs of composite I – beam bridges. Ph.D.
5. The type of bottom transverse reinforcement, when similar Thesis, Kingston (ON, Canada): Queen’s University; 1972.
[11] Hewitt BE, Batchelor de VB. Punching shear strength of restrained slabs. J
axial stiffness was used, did not affect cracking, deflections, Struct Div, ASCE 1975;10(9):1837–53.
strains, or punching capacities of the tested deck slabs. [12] Perdikaris PC, Beim S. RC bridge decks under pulsating and moving loads. J
6. The behavior of the slab G-200-N was very similar to that of the Struct Eng, ASCE 1988;114(3):591–607.
[13] Ahmed E, Benmokrane B. Design of concrete bridge deck slabs reinforced with
slab G-175-H. This means that the reduction of the deck slab GFRP bars according to the CHBDC CSA-S6. In: Proceedings of the SMSB-10
thickness can be recovered by increasing the concrete strength. conference, Niagara Falls, ON, Canada; 2010.
7. Considering the actual mode of failure which is the punching [14] Beeby AW. The prediction of crack widths in hardened concrete. Struct Eng
1979;57A(1):9–17.
shear failure in the design of restrained bridge deck slabs may
[15] Frosch RJ. Another look at cracking and crack control in reinforced concrete.
lead to a reduction in the amount of reinforcement which will ACI Struct J 1999;96(3):437–42.
result in a reduction in initial costs of the GFRP-reinforced con- [16] Japan Society of Civil Engineers (JSCE). In: Machida A, editor. Recommendation
for design and construction of concrete structures using continuous fibre
crete bridge deck slabs.
reinforcing materials. Concrete engineering series 23; 1997. 325 p.
8. The current ACI 440.1R-06 [22] punching shear model underes- [17] El-Ghandour AW, Pilakoutas K, Waldron P. New approach for punching shear
timates the punching shear strength of the tested slabs. The capacity prediction of fiber reinforced polymer reinforced concrete flat slabs.
average ratio of Vexp/Vpred was 3.1 and the corresponding COV ACI SP 188. American Concrete Institute; 1999. p. 135–44.
[18] El-Ghandour AW, Pilakoutas K, Waldron P. Punching shear behavior of fiber
was 19%. On the other hand, El-Gamal et al. [21] model yielded reinforced polymers reinforced concrete flat slabs: experimental study. J
good yet conservative prediction with an average Vexp/Vpred of Compos Constr, ASCE 2003;7(3):258–65.
1.11 and a corresponding COV of 7%. [19] Matthys S, Taerwe L. Concrete slabs reinforced with frp grids. II: Punching
resistance. ASCE, J Compos Constr 2000;4(3):154–61.
[20] Ospina CE, Alexander SDB, Roger Cheng JJ. Punching of two-way concrete slabs
with fiber-reinforced polymer reinforcing bars or grids. ACI Struct J
Acknowledgements 2003;100(5):589–98.
[21] El-Gamal SE, El-Salakawy EF, Benmokrane B. A new punching shear equation
for two-way concrete slabs reinforced with FRP bars. ACI SP-230. American
The authors acknowledge the financial support received from Concrete Institute; 2005. p. 877–94.
the Natural Science and Engineering Research Council of Canada [22] ACI Committee 440. Guide for the design and construction of concrete
reinforced with FRP bars. ACI 440.1R-06. Farmington Hills (MI, USA): American
(NSERC), the Fond Quebecois pour la Recherche en Nature et
Concrete Institute; 2006. 41 p.
Technologie (FQRNT), and Ministry of Transportation of Quebec.