83PCI Journal | Summer 2012
Corrosion of steel in reinforced concrete structures is one of the main factors limiting the service life of bridge decks and parking structures. Chlorides
from deicing salts or a marine environment act as catalysts
for the corrosion of steel in concrete. Corrosion mitigation
requires expensive maintenance, repair, or replacement.
The use of glass-fiber-reinforced polymer (GFRP) bars as
internal reinforcement is a possible solution to corrosion
of steel bars. In addition to their noncorrosive properties,
GFRP bars have higher strength than steel bars and are
light and easy to handle, which makes them attractive as
reinforcement for certain concrete elements, such as slabs.
However, GFRP bars have different mechanical properties
from steel; GFRP bars behave in a linear elastic manner
until rupture, which makes concrete members reinforced
with GFRP bars vulnerable to brittle failure.
Considerable research has been undertaken to investigate
both flexural and shear performance of GFRP-reinforced
concrete structures. Despite the differences in material
properties compared with steel bars, the prediction of
flexural capacity using the strain compatibility approach is
still effective. The behavior of lightweight concrete slabs
reinforced with GFRP bars without shear reinforcement
is a topic of active research. Prediction of shear capac-
■ This paper reports an experimental investigation of the flexural
and shear performance of concrete structures reinforced with
glass-fiber-reinforced polymer (GFRP).
■ Simply supported slabs of both normalweight and lightweight
concretes with compressive strengths in excess of 8000 psi
(55 MPa) were tested.
■ Modified compression field theory first- and second-order
equations can provide accurate yet conservative predictions of
the behavior of GFRP-reinforced concrete despite the differ-
ences in mechanical properties between GFRP and steel.
■ The predictions are less conservative for lightweight than for
Shear capacity of concrete
slabs reinforced with
bars using the modified
compression field theory
Ruifen Liu and Chris P. Pantelides
Summer 2012 | PCI Journal84
ity is essential in the design of GFRP reinforced concrete
members, as Guide for the Design and Construction of
Structural Concrete Reinforced with FRP Bars (ACI
440.1R-06)1 recommends that such members be designed
as overreinforced, making them vulnerable to shear failure.
There is little research available on GFRP-reinforced slabs
constructed with high-strength normalweight or light-
The modified compression field theory (MCFT) is an
analytical model with 15 equations that produce accurate
estimates of shear strength for steel-reinforced concrete
members.2 Bentz and Collins3 reduced the MCFT equations
to two, which still accurately estimate the shear strength of
steel-reinforced concrete members.4 Hoult et al.5 found that
crack widths are affected by both a size effect and a strain
effect regardless of the type of reinforcement used; they
also showed that the two MCFT equations proposed by
Bentz and Collins work equally well in predicting the shear
capacity of normalweight concrete slabs reinforced with
steel or FRP reinforcement.
Sherwood et al.6 demonstrated that the width of a member
does not affect the shear stress at failure for steel-rein-
forced concrete members, which indicates that the MCFT
could be used for both beams and slabs. Bentz et al.7 found
that despite the brittle nature of the reinforcement, FRP-
reinforced large concrete beams behave similarly in shear
to steel-reinforced concrete beams. In this paper, a series
of 20 tests is presented to investigate the influence of slab
width and depth, slab span, concrete compressive strength,
and type of concrete (lightweight versus normalweight) on
the shear strength of GFRP-reinforced slabs. The maxi-
mum deflection of the slabs under service loads satisfied
the American Association of State Highway and Transpor-
tation Officials’ AASHTO LRFD Bridge Design Specifica-
tions8 in the tests for the slabs designed for flexure accord-
ing to ACI 440.1R-06 guidelines (Pantelides et al.9).
Twenty slabs were tested to investigate the behavior of
GFRP-reinforced concrete slabs constructed with high-
strength normalweight or lightweight concrete. The
construction variables included unit weight and compres-
sive strength of concrete, slab span and depth, slab width,
and reinforcement ratio. Four series of slabs were built
with different dimensions or reinforcement ratios. Figure 1
shows the top and bottom reinforcement for series A and
B slabs. Series A and B slabs have the same width (2 ft
[0.6 m]) but different spans and depths. Series C slabs have
the same reinforcement, thickness, and span as series A
slabs, but their widths are 6 ft (1.8 m). Series D slabs have
the same dimensions as series C slabs, but series D slabs
have a GFRP reinforcement ratio approximately half that
of series C slabs.
The normalweight concrete used in this study was ready-
mixed concrete incorporating a 3/4 in. (19 mm) crushed
limestone. The specified compressive strength of both
normalweight and lightweight concretes was 6000 psi
(41 MPa); however, several batches were cast at different
times and consequently the concrete compressive strength
for the normalweight concrete at the time of testing ranged
from 8500 psi (59 MPa) to 12,600 psi (87 MPa) and for
lightweight concrete from 8100 psi (56 MPa) to 10,900 psi
The lightweight concrete used was sand-lightweight con-
crete, which had a coarse aggregate (expanded shale) size
of 1/2 in. (13 mm). The unit weight of the sand-lightweight
concrete used was 123 lb/ft3 (1970 kg/m3).
The GFRP bars used for construction were no. 5 (16M)
bars. The tensile strength of the specific lot of GFRP
bars used in these tests was 103,700 psi (715 MPa),
and the modulus of elasticity was 6280 ksi (43 GPa), as
determined from tensile tests of the bars according to
Table 1 shows the concrete compressive strength at the
time of testing, the actual reinforcement ratio, and the bal-
anced reinforcement ratio.
Test setup and procedure
All slabs were tested as simply supported members on two
reinforced concrete beams (Fig. 2). Elastomeric pads 6 in.
(150 mm) wide and 2 in. (50 mm) thick were placed on the
supporting beams so that the slabs could rotate freely near
the support without coming into contract with the beams.
The load was applied using a hydraulic actuator through a
10 in. × 20 in. × 1 in. (250 mm × 500 mm × 25 mm) steel
bearing plate for all slabs, which simulates the area of a
double-tire truck load on a bridge deck.8 The steel bearing
plate was placed directly on the concrete surface of the
panels. The wider panels are subjected to a combination of
one-way shear and punching shear. The load was applied
as a series of half-sine downward cycles of increasing am-
plitude without stress reversals. The load application was
displacement controlled at a constant rate of 0.2 in./min
(5 mm/min). The loading scheme was intended to simu-
late repeated truck loading applied to the slab of a precast
concrete bridge deck.
During testing, all slabs developed flexural cracks at low
loads and additional diagonal cracks as the loads increased.
Ultimately, the slabs failed in diagonal tension (Fig. 3).
After formation of the critical diagonal crack near one of
85PCI Journal | Summer 2012
the two supports, the concrete crushed on the compression
face of the slabs. All slabs failed the same way regard-
less of concrete type (normalweight or lightweight), slab
dimensions, or amount of reinforcement. In a few tests, a
few GFRP bars in the top mat near the outer edges of the
slab snapped and sheared off after the ultimate load was
reached, shortly before the ultimate deflection (Fig. 3).
This occurred after the concrete cover had spalled off and
the bars were exposed, and was the result of the GFRP
bars trying to carry the compression forces arising from
the applied load. The GFRP bars in the bottom mat did not
fracture in any of the tests even though they experienced
significant tensile strain and deformation. Table 1 shows
the concrete compressive strength at the time of testing,
the actual reinforcement ratio, the balanced reinforcement
ratio, and the experimental shear capacity.
Figure 1. Dimensions for top and bottom glass-fiber-reinforced polymer reinforcement mat for slabs. Note: no. 5 = 16M; 1 in. = 25.4 mm; 1 ft = 0.305 m.
2 ft 8 ft 2 ft
24 No. 5 at 6 in.
.3 in. 3 in.
Series C slabs
2 ft 8 ft 2 ft
18 No. 5 at 8 in.
. 4 in. 4 in.
5 in. 9.
Series D slabs
2 ft 9 ft - 6 in. 2 ft