Construction and Building Materials 87 (2015) 6777

Contents lists available at ScienceDirect

Construction and Building Materials

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

Arching behaviour of precast concrete slabs in a deconstructable

composite bridge deck
H. Valipour , A. Rajabi, S.J. Foster, M.A. Bradford
Centre for Infrastructure Engineering and Safety (CIES), School of Civil and Environmental Engineering, UNSW Australia, UNSW Sydney, NSW 2052, Australia

h i g h l i g h t s
 A deconstructable steelconcrete composite bridge deck is proposed.
 Experimental data on arch behaviour of precast RC deck slabs are provided.
 Efciency of cross-bracing and transverse ties for inducing arch action is studied.
 Application of bolted shear connectors in a deconstructable deck is studied.

a r t i c l e

i n f o

Article history:
Received 4 December 2014
Received in revised form 3 April 2015
Accepted 9 April 2015

Keywords:
Arching action
Bolted shear connector
Composite deck
Deconstruction
Rehabilitation

a b s t r a c t
This paper describes the results of the testing of precast concrete slabs in a deconstructable composite
steelconcrete system for the construction of bridge decks. Benign arching action is utilised to carry
the point (wheel) loads to the supports and to develop the required slab capacity; the failure mode
and loaddeection response of the precast concrete slabs being investigated in the study. Twelve
half-scale precast reinforced concrete slab strips were tested, with the slabs being attached to steel girders using friction grip bolts to provide shear connection between the deck and the supporting steel girders. The systems were tested under a monotonically increasing point load, which simulates vehicle
wheel loading. The conguration and proportion of the reinforcing steel bars and the types of transverse
cross-bracing and transverse straps were the main test variables. It is concluded that friction grip bolted
shear connectors can prevent relative slip between the steel girders and concrete deck slabs, so that the
equilibrating tension force in the cross-bracing/transverse straps, required to develop compressive arching in the slabs, can be developed. The arching effect in the slabs is very benecial, and cannot be ignored
in rational structural design processes.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
In reinforced concrete (RC) exural members, cracking of the
section in the tensile zone is associated with a change in the neutral axis (NA) position that, in turn, causes an axial extension of the
member as the neutral axis moves away from the centroidal axis
and towards the farthest compressive bre. In steelconcrete composite bridge decks, this axial extension of the RC deck slab can be
prevented by adjacent spans and cross-bracing/transverse diaphragms, generating a compressive thrust in the restrained RC
deck slab [1]. This phenomenon, known as compressive membrane
(or arching) action, can signicantly increase the post-cracking
stiffness as well as the exural and punching shear capacities of
laterally restrained RC deck slabs [210].
Corresponding author. Tel.: +61 2 9385 6191.
E-mail address: H.Valipour@unsw.edu.au (H. Valipour).
http://dx.doi.org/10.1016/j.conbuildmat.2015.04.006
0950-0618/ 2015 Elsevier Ltd. All rights reserved.

The enhancing effect of arching action on the ultimate load

capacity of restrained RC deck slabs has been recognised and
implemented in some design standards [11,12]. Moreover, to
resolve the issues associated with corrosion of the internal steel
reinforcing bars and to extend the service life of concrete deck
slabs, the concept of mobilising this arching action to develop
steel-free deck slabs has been proposed by some researchers
[4,5,13]. In existing steel-free deck slabs that rely on the development of arching action, the internal steel bars are typically
replaced by external transverse straps with or without cross-bars
[14]. The straps rest (or are welded) on the top ange of the steel
girders and close to the soft of the deck slab and the concrete slab
is separated from the top ange of the steel girders by a haunch
that encases the cross-bars and a small portion of the transverse
straps. Alternatively, the lateral restraint/connement required
for the development of arching action in the concrete slab can be
provided by threaded bars (i.e. steel or CFRP/GFRP) and/or

68

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

Fig. 1. Geometry, cross-section, conguration of restraining/conning system in transverse direction and test set up for precast RC slabs connected to steel girders using
PFBSCs.

Table 1
Designation of specimens, details of exural reinforcement and bar conguration and transverse conning/restraining system.

Designation of specimens#

Reinforcement-1

Reinforcement-2

q = Ast/bd (%)

c1 (mm)

c2 (mm)

Transverse restraining system 45  45  5EA

M6B
M4B
B4B
B6B
M6
M4
B4
M6S
M6BS
M4BS
B4BS
B6BS

6N10
4N10
4N10
6N10
6N10
4N10
4N10
6N10
6N10
4N10
4N10
6N10

4N10
6N10

4N10
6N10

1.5
1.0
0.7
1.0
1.5
1.0
0.7
1.5
1.5
1.0
0.7
1.0

45
45
25
25
45
45
25
45
45
45
25
25

25
25

25
25

Bracing
Bracing
Bracing
Bracing

Strap
Bracing + strap
Bracing + strap
Bracing + strap
Bracing + strap

Mx and Bx designations are used for slabs with main exural reinforcement at middle and bottom layer, respectively and x denotes the number of reinforcing steel bars.

intermediate diaphragms [4,15]. Amongst different alternatives,

transverse straps permanently connected to the concrete slab
and top ange of the steel girder appears to be the most efcient
conning system, whilst threaded steel bars can be replaced easily
[4].
In current engineering practice, the composite interaction
between the concrete slab and steel girders is typically achieved
by welded headed shear studs buried permanently in the cast
in situ concrete or in pockets lled with grout for precast slabs.
However, this form of construction is not conducive to deconstruction and it also hinders the speedy and cost-effective replacement
of defective slabs and/or the transverse conning system.
The ability of bolted shear connectors in developing efcient
composite action between the steel girders and precast concrete
slabs has been demonstrated through several studies [1623]. In
particular, recently conducted three-point bending tests on

composite beams and push-out tests on composite connections

with post-installed friction-grip bolted shear connectors (PFBSCs)
have shown that the composite efciency and fatigue strength of
PFBSCs are signicantly higher than those for beams having stud
shear connectors [19,20,24]. Furthermore, the steelconcrete composite decks with PFBSCs can be deconstructed easily, so that the
possibility for future reuse and the recycling of the structural components are maximised [16,17]. Thus, composite bridge decks with
PFBSCs can allow for speedy and cost-effective rehabilitation,
replacement and repair of deteriorated deck slabs with minimal
trafc disruption.
In this paper, the ability of PFBSCs in preventing the relative slip
between the precast concrete slabs and steel girders in a transversely-conned steelconcrete composite bridge deck is explored
experimentally. The anchoring provided by PFBSCs allows for the
development of compressive membrane (or arching) action that

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

69

Fig. 2. Locations of (a) LVDTs and inclinometers and (b) steel and concrete strain gauges on precast slab.

can signicantly enhance the load carrying capacity and stiffness of

RC exural members in this application. A total of twelve half-scale
single-span precast RC slab strips connected compositely to steel
girders using PFBSCs are tested under a monotonically increasing
displacement-controlled point load applied at mid-span. The
arrangement of the exural steel reinforcement (the number and
location of the bars) as well as the type and stiffness of the transverse connement/restraint (i.e. cross-bracing and straps) provided
for the precast slabs are the main variables in the experimental
program. In addition to the vertical displacement and applied load,
the rotation and horizontal translation (i.e. transverse extension of
the slab) as well as strain in the steel bars, cross-bracing, straps and
the concrete were measured during the tests. The experimental
data is used to evaluate the structural performance (viz. the exural behaviour and load carrying capacity) and the ductility of transversely restrained precast bridge deck slabs with PFBSCs.
2. Experimental program
2.1. General
To study the arching behaviour of transversely-restrained steelconcrete composite bridge decks with PFBSCs, a 15 m long simply supported steelconcrete composite bridge deck was designed according to the Australian bridge standard
AS5100.6-2004 [25]. The bridge deck comprised of 200 mm thick RC slabs connected compositely to 1200WB249 welded steel beams spaced 2.0 metres apart.
The loads considered for the strength limit state design of the concrete deck slabs
include gravity loads (i.e. the self-weight of the structure plus 50 mm of surfacing
asphalt), the stationary distributed trafc load denoted S1600 and a wheel point
load of 80 kN, as specied in the Australian bridge standard AS5100.2 for loading
[26]. In the experimental program, the response of a 1200 mm wide middle slab
strip was investigated. Twelve half-scale model precast slabs were constructed with
different quantities and congurations of the reinforcement (i.e. middle, bottom
and top layers) and four different types of transverse connement/restraint (i.e.
no straps or cross-bracing, cross-bracing only, straps only and a combination of
straps and bracing) were provided for the slab. The cross-bracing and horizontal
straps were made of 45  45  5EA equal-leg angles of Grade 300PLUS. At each
end, the cross-bracing and horizontal straps were bolted to the stiffeners and the

top ange of the steel girders respectively using high-strength 8.8 grade bolts of
16 mm diameter (Fig. 1). The precast slabs were tested under a monotonically
increasing displacement-controlled point load applied at mid-span, to replace
wheel loading on the bridge.

2.2. Geometry and test setup

The precast slabs that were tested complied with minimum design requirements of the Australian standard AS5100.5-2004 [27] for concrete bridges. Fig. 1
shows the test setup for the half-scale single (non-continuous) precast reinforced
concrete slab strip connected compositely to the steel girders using PFBSCs. The
load was applied on the precast RC slabs through a 300 kN hydraulic actuator
operated under displacement control at a rate of 0.1 mm/s (Fig. 1). The overall
dimensions of the tested slabs, size of the sections, conguration of the restraining/conning system in the transverse direction and the testing setup are shown
in Fig. 1. Details of the reinforcing bar sizes and congurations (i.e. bottom and middle) and the concrete cover to the steel bars (c1 and c2 in Fig. 1) are given in Table 1.
The M or B designation in Table 1 was used for slabs with main exural reinforcing
bars at middle or bottom layer, respectively. With regard to the reinforcing ratios
(q), the specimens can be classied as being over-reinforced (series M6 with
q = 1.5%), those with high reinforcement (series M4 and B6 with q = 1.0%) and those
which are moderately-reinforced (series B6 with q = 0.7%).
At each end, two M16 8.8/F post-installed bolts were used to connect the precast RC slabs to the top ange of a 310UB32 steel girder of Grade 300PLUS (Fig. 1).
These post-installed friction grip bolted shear connectors were tightened to a shank
tension of 0.4fuf = 330 MPa, where fuf = 830 MPa is the tensile strength of class 8.8
bolts. The cross-bracing and straps for the transverse connement of the slabs were
connected to the web-stiffeners and top ange of the 310UB32 steel girders respectively by the M16 bolts (Fig. 1). The post-tensioning stress induced in all of the friction grip bolts was checked using a calibrated torque wrench. The pre-drilled bolt
holes in the top ange and web stiffeners of the steel girder were 18 mm in diameter and the holes in the precast slab were 20 mm in diameter to allow for easy
installation of the PFBSCs, and to facilitate potential deconstruction.

2.3. Materials
The steel reinforcing bars were 10 mm diameter ribbed bars with a nominal
yield strength of 500 MPa. Two sets of tensile tests (three specimens in each set)
on the bars were conducted; the mean yield strength being fy = 575 MPa. The average ultimate strain of the bars was eu 0:08 and the ultimate strength was
f u 680 MPa.

70

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

Crack began on the top surface

of slab & ran between PFBSCs

Crushing of concrete

Specimen M6B
Bolt hole

Crack began on the top surface

of slab & ran between PFBSCs

Crushing of concrete
Specimen B6BS

Bolt hole

(a)

and this mode of failure was associated with yielding of reinforcing

steel bars at mid-span. The second mode of failure was triggered by
cracks that developed on the top surface of the slabs between the
bolted connectors and was then followed by crushing of the concrete in the compressive zone at the mid-span of the slabs
(Fig. 3a). For the third mode of failure, only compressive crushing
of the concrete on top surface of the slab at mid-span occurred
(Fig. 3b). It is noteworthy that except for the rst mode of failure,
the other modes of failure were fairly brittle owing to crushing of
the compressive concrete at mid-span.
The modes of failure for all specimens tested are reported in
Table 2, and it can be seen that the transverse connement/restraint provided by straps and cross-bracings has altered the mode
of failure. It is seen further that only specimens M4 and B4 (with no
transverse connement/restraint) had a ductile mode, with the
failure of other specimens partly associated with crushing of the
concrete at the slab mid-span. In particular, the somewhat brittle
failure of specimen M6 with no transverse connement/restraint
can be attributed to over-reinforcement (approximately 1.5%)
and possibly the higher than normal yield strength of the steel
bars.
At the conclusion of each test, the bolted shear connectors were
untightened and inspected visually; despite extensive cracking,
crushing and severe damage in the precast slabs, no evidence of
damage or deformation was observed in the bolted shear connectors and they could be taken out of the sleeve easily and so were
conducive to the deconstruction, repair or rehabilitation of the
bridge decks.
3.2. Global response

Specimen M6S
Bolt hole

Crushing of concrete

(b)
Fig. 3. Brittle modes of failure associated with (a) development of cracks on top
surface of slabs between bolted connectors followed by crushing of concrete in
compressive zone at mid-span and (b) crushing of compressive concrete at midspan.

The average compressive strength of the concrete used for the precast slabs at
the time of testing of each specimen was fcm = 38 MPa, having been determined
from the average of three 300 mm by 150 mm diameter cylinders strengths in
accordance with AS1012.9.
2.4. Instrumentation
In addition to the applied load and vertical displacement at the mid-span of the
precast RC slab, the horizontal transverse deection and rotation at each end of the
slab were measured using LVDTs and inclinometers respectively (Fig. 2a). The
strains in the concrete and longitudinal reinforcing bars were measured at various
sections along the transverse length of the precast slab. In total, ve strain gauges
(three steel strain gauges on the reinforcement and two concrete strain gauges)
were mounted along each precast slab strip. The locations of the concrete and steel
strain gauges along the precast slab are shown in Fig. 2b. In addition, two strain
gauges, viz. St-B-SG(1) and St-B-SG(2), were mounted on the cross-bracing and a
strain gauge (St-S-SG) was attached to the strap.

3. Test results

Plots of the applied load versus vertical displacement at the

mid-span of the precast slabs are shown in Fig. 4; the small dip following the rst peak load in these loaddeection diagrams was
caused by the development of cracks on the top surface of the precast slab. These cracks ran between the PFBSCs as shown in Fig. 3a.
The second peak in the loaddeection response of the slabs was
associated with compressive crushing of the concrete at mid-span
(Fig. 3).
The load versus rotation response of the left and right ends of
the precast slabs is shown in Fig. 5, exhibiting a non-linear
relationship between the load and average rotation of the slab
ends. The rotation data for some of the specimens are not reported
in Fig. 5, because they were spoiled by the noise in the data acquisition system.
The end rotations for different specimens at a load of 54.2 kN,
which is the peak load capacity of specimen M4 with no bracing
or straps, are shown in Fig. 6. It is seen that the straps and, particularly, the cross-bracing in the transversely conned bridge deck
provide the slabs with partial rotational xity. The rotational stiffness of the end supports for the single-span precast slabs is much
greater than the idealised pinned boundary conditions of simplysupported members. Accordingly, accurate analysis of transversely
conned/restrained precast slabs necessitates modelling of the
rotational stiffness as well as translational stiffness provided by
the transverse conning system (e.g. cross-bracing and/or straps).
The two LVDTs attached to the sides of slabs (Fig. 2a) were used
to determine the horizontal displacement in the transverse direction and the algebraic sum of these displacements is the total elongation of the slabs. The load versus elongation response of the
precast slabs is shown in Fig. 7.

3.1. Modes of failure

3.3. Local response
Three different modes of failure were observed in the precast
slabs tested. The rst mode of failure was only observed in specimens M4 and B4 (with no transverse connement or restraint),

The load versus strain response in the reinforcing bars at the

mid-span of the slab (the results from strain gauge St-SG(1)) are

##

69.8
107.9
141.1
68.0
54.2
72.8
88.7
102.7

84.1
114.0
136.6

M4B
B4B
B6B
M6
M4
B4
M6S
M6BS

M4BS
B4BS
B6BS

75.7

70.1
103.6
135.9

86.8

87.5

2nd

45.2
63.6
91.5

45.2
63.6
91.5
64.0
45.3
63.7
64.0
64.0

64.0

Plastic
analysis

##

(Exp.)/Pu

1.67
1.79
1.49

1.55
1.63
1.48
1.06
1.20
1.14
1.39
1.36

1.37

(Plastic)

Pu

23.9
24.2
24.8

14.9
16.9
22.5
24.2
26.6
25.7
23.2
25.6

18.8

1st

62.1
79.9
114.1
54.5
37.9
40.2
66.8
83.0
70.9
87.6
101.9

27.1

70.4

Py
(kN)

Yield

20.9
19.5
24.9

27.2

24.6

2nd

du (mm)

15.1
11.1
11.9

12.2
8.3
12.0
14.2
9.2
8.3
11.7
14.4

11.7

dy
(mm)

45.5
48.6
65.8

38.7
46.7
52.8
23.8
16.1
33.2
52.8
47.2

35.6

P0.001
(kN)

ec = 0.001#

6.72
6.19
7.43

5.09
4.21
4.89
5.35
2.61
2.43
8.31
6.43

4.05

d0.001
(mm)

1.79
2.18
2.08

1.72
2.35
2.08
1.70
2.89
3.10
1.98
1.89

2.10

(2nd)/dy

du

3.28
6.11
4.71

2.99
6.29
4.92
2.98
10.9
11.9
4.04
3.21

3.64

Energybased

Load (kN)

Crushing of concrete
Crushing of concrete

Load (kN)

Crushing of concrete
Development of cracks between bolt holes followed by concrete
crushing at mid-span

Crushing of concrete
Yielding of steel bars

Development of cracks between bolt holes followed by concrete

crushing at mid-span

Mode of failure 45  45  5EA

Load (kN)

6.57
9.17
6.93

5.28
9.27
12.30
12.92
34.31
23.19
4.69
8.66

11.64

J
factor

Ductility index l or J factor

Compressive strain of concrete at mid-span on the top surface of the slab.

Plastic analysis was conducted assuming that the precast slab strips are simply supported, ultimate strength fu = 680 MPa in the analysis.

89.3

1st

Exp.

Pu (kN) peak load

M6B

Designation of
specimens

Table 2
Peak load capacity Pu, load Py corresponding to onset of steel yielding, mode of failure, ductility index l, mid-span deection dy at onset of steel yielding and mid-span deection du corresponding to failure load.

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

20
0
0

100

140

20

0
0
10

71

140

120

Concrete crushing

100
80

60

40
Cracks develop between
bolted connectors
10
20

120

20

B6B
B4B
M6B
M4B

Deflection (mm)

30

40

20
0

10
20
30
Deflection (mm)

40

Cracks develop between

bolted connectors

30

40

(a)

140
Plastic hinge (yielding
of steel bars)

Concrete crushing

80

60
M6S
B4
M6
M4

40

(b)

120
Concrete crushing

100

80

60

B6BS
B4BS
M6BS
M4BS

Deflection (mm)
40

(c)

Fig. 4. Load versus vertical displacement at mid-span of slabs (a) transversely

conned/restrained by cross bracing (b) without bracing and straps or with only
straps (c) transversely conned/restrained by cross bracings plus straps.

shown in Fig. 8. It is seen that in all cases, reinforcing bars yielded

well before the load capacity of the slabs was attained. The load
versus compressive strain for the concrete on the top surface and
at the mid-span of the slabs (the results from strain gauge CSG(1)) are shown in Fig. 9, from which it can be seen that the compressive strain in the concrete has reached values greater than
0.003 (adopted typically by different standards as being the maximum compressive strain in concrete). Accordingly, it is concluded
that the failure of the specimens is not fully brittle, and some level
of ductility is available in the transversely conned or restrained
precast bridge deck slabs.
Curves of the load versus tensile strain ebu in the straps and
bracing (being from strain gauges St-B-SG(1) and St-S-SG) are
shown in Fig. 10. It is seen that at all stages of the loading, the tensile strains and stresses in the cross-bracing and straps remain well

72

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

100

Rotation (Degree)

120

3.5

Load (kN)

140

80
60
B6B-Left 40
B4B-Left
M6B-Left 20
M4B-Left

B6B-Right
B4B-Right
M6B-Right
M4B-Right

-4

-3 -2 -1 0
1
2
3
Rotation at end of slab (degree)

100

1.5

N
on

Load (kN)

120

0.5

(a)
140

2.5

0
-5

Only
strap

Only
bracing

Transverse confinement

Strap
+
bracing

Fig. 6. Average rotation at end of precast slabs at load of 54.2 kN (peak load
capacity of specimen M4 with no strap and/or bracing).

80
where Dy and Du are the displacements corresponding to the onset
of steel yielding and to the peak load capacity respectively.
The energy-based ductility in this study is dened as [29]:

60
40
M6S-Left
M6-Left 20
M4-Left

M6S-Right
M6-Right
M4-Right

-5

-4

-3 -2 -1 0
1
2
3
Rotation at end of slab (degree)

where Wy is the area under the loaddeection response curve from

rst loading to the deection corresponding to onset of steel bar
yielding and W0.75u is the area under the loaddeection curve from
rst loading to the deection at which the load has reduced to 75%
of the peak load. Similar energy-based measures have been used by
other researchers to evaluate the structural ductility of concrete
members reinforced with low-ductility steel bars or strengthened
with FRP sheets [28,30].
The J factor was introduced by Mufti et al. [31] and adopted by
Canadian Highway Bridge Design Code [12]. It is dened as:

100

Load (kN)

120

80
60
B6BS-Left 40
B4BS-Left
M6BS-Left 20
M4BS-Left

-5

B6BS-Right
B4BS-Right
M6BS-Right
M4BS-Right

0
-3
-1
1
3
Rotation at end of slab (degree)

J Cc  Cs
5

(c)
Fig. 5. Load versus average rotation at end of precast slab (a) transversely conned/
restrained by cross bracing (b) without bracing and straps or with only straps (c)
transversely conned/restrained by combination of cross bracing and straps.

below the yield strain and stress of the Grade 300PLUS steel. The
friction grip bolts employed in the transverse conning/restraining
system have effectively prevented slip in the transverse direction
and only a small slip was observed for the bolts connecting the
straps to the top ange of the steel girder in specimen M6S (see
Fig. 10a).
3.4. Ductility of specimens
Different denitions have been used by researchers to evaluate
the structural ductility of reinforced concrete members. For the
purpose of this study, three different measures, viz. the displacement-based ductility lD, the energy-based ductility lE and J factor
are employed to evaluate the structural ductility of the system. The
displacement-based ductility (deformability) is [28]:

lD

Du
;
Dy

(b)
140

W 0:75u
;
Wy

lE

wu
w0:001

Mu

;
M 0:001

where C c and C s are curvature and strength factors, respectively, and

are dened as the ratio of curvature w, or bending moment M, values at ultimate load to the corresponding values at concrete compressive strain of 0.001. The strain of 0.001 is assumed to be the
beginning of inelastic deformation in concrete or the strain in concrete under service load condition [32].
A slightly modied J factor is used in this study,

 

Du
Pu

;
D0:001
P0:001

where Du/D0.001 is the mid-span displacement at ultimate relative

to the value at concrete compressive strain of 0.001 and Pu/P0.001
is the load at ultimate relative to the value at a concrete compressive strain of 0.001. The J factor adopted in study is similar to
robustness index proposed by Van Erp [32].
The displacement-based and energy-based ductility indices and
the J factor of the transversely restrained precast slabs are given in
Table 2 and the variation of the ductility index lE and J factor with
respect to the reinforcing ratio q and the ratio of conning system
stiffness KConning system (in the transverse direction) divided by the
axial stiffness of slab K Slab are shown in Fig. 11a and b. It is observed
that all the measures used in this study consistently represent the
relative ductility of transversely restrained slabs. Specimen M4
without transverse restraint (lE = 10.9 and lD = 2.89) and specimen
B4 without transverse restraint (lE = 11.9 and lD = 3.10) have the

73

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

140

Load (kN)

120
100
80
60
B6B
B4B
M6B
M4B

40
20
0
0

2
3
4
5
Elongation of slab (mm)

(a)

(a)

140

Load (kN)

120
100
80
60
40

M6S
B4
M6
M4

20
0
0

1
2
3
4
5
Elongation of slab (mm)

(b)
(b)

140

Load (kN)

120
100
80
60
40

B6BS
B4BS
M6BS
M4BS

20

0
0

1
2
3
4
5
Elongation of slab (mm)

(c)
Fig. 7. Load versus elongation of precast slab measured by horizontal LVDTs at end
of precast slab (a) transversely conned/restrained by cross bracing (b) without
bracing and straps or with only straps (c) transversely conned/restrained by
combination of cross bracing and straps.

highest ductility indices, whereas the lowest ductility index

(lE = 2.98 and lD = 1.70) belongs to the series M6, which are over
reinforced. Transversely-conned precast slab decks with medium
to low reinforcing proportions (i.e. B4B, B6B, B4BS and B6BS) have
medium ductility indices (lE = 4.71 to 6.29). Moreover, it is seen
that the transversely restrained series B slabs with a low reinforcing
ratio (q = 0.7%) has ductility index lE, exceeding 6.0. The J factor for
all transversely restrained slabs was above the minimum requirement of 4.0 specied in CHBDC code [12] for rectangular sections
with FRP reinforcements.
In this study, the stiffness of conning system KConning system
and the slab K Slab are obtained from

K Confining system


n 
X
Es Ai
cos2 h;
li
i1

n total number of bracings=straps

(c)
Fig. 8. Load versus tensile strain at mid-span in reinforcing bars (St-SG1) for precast
slabs (a) transversely conned/restrained by cross bracing (b) without bracing and
straps or with only straps (c) transversely conned/restrained by combination of
cross bracing and straps.

K Slab



Ec Ac
;
lc

where Es 200 GPa is the elastic modulus of steel, Ai 394 mm2

is the cross-sectional area of a single cross-bracing/strap, li is the
length of cross-bracing/strap and h is the angle between the
cross-bracing/strap and the horizontal plane. For cross-bracings
li 1045 mm and h 12 and for straps li 1020 mm and
h 0 . Ec is the concrete modulus of elasticity and it is taken as
Ec 29 GPa, Ac 600  100 6  104 mm2 is the cross-sectional
area of concrete slab and lc 1100 mm is the span length of slab.
It is noteworthy that the effect of reinforcing bars on the axial stiffness of slab has been ignored in Eq. (6).

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

120
100
80
60

B6B
B4B
M6B
M4B
-5000

-4000

-3000

150
120

Load (kN)

140

strain at ultimate
in design

Load (kN)

74

90

Slip
60

B6B-bracing
B4B-bracing
M6B-bracing
M4B-bracing
M6S-strap

40
30

20
0
-2000

-1000

0
0

100

Strain (mm/mm)

60
40

M6S
M6
M4

60
B6BS-bracing
B4BS -bracing
M6BS-bracing
M4BS-bracing

20
0

-2000

-1000

120
100
80

Load (kN)

140

strain at ultimate
in design

60
B6BS
B4BS
M6BS
M4BS
-3000

40
20
0
-2000

-1000

100

200
300
400
Strain ( mm/mm)

500

(b)

(b)

-4000

90

30

Strain (mm/mm)

-5000

120

Load (kN)

80

Load (kN)

120
100

-3000

500

150

140

strain at ultimate
in design

-4000

400

(a)

(a)

-5000

200
300
Strain (mm/mm)

Strain (mm/mm)

(c)
Fig. 9. Load versus concrete compressive strain at mid-span on top surface of
precast slabs (a) transversely conned/restrained by cross bracing (b) without
bracing and straps or with only straps (c) transversely conned/restrained by
combination of cross bracing and straps.

The stiffness of a single cross-bracing and a strap obtained from

Eq. (5) is 72.5 and 77.3 kN/mm, respectively, and stiffness of the
precast slab obtained from Eq. (6) is 1580 kN/mm.
3.5. Strength enhancement provided by arching action
The maximum loads carried by the specimens are given in
Table 2. The theoretical peak load capacities of the specimens
can be calculated using an elementary plastic analysis assuming
pinned ends; the theoretical over the experimental load capacities
of the specimens are given in Table 2. The results show that ignoring the arching effect and the rotational restraint lead to overly
conservative design. For specimens M4BS and B4BS, for example,
the failure load is underestimated by 67% and 79%, respectively.

Fig. 10. Load versus tensile strain in cross bracing/straps for the slabs transversely
conned/restrained by (a) only cross bracing or straps and (b) cross bracing plus
straps.

The load capacities of the precast concrete deck slabs with

respect to the reinforcement ratio q and the relative stiffness
(KConning system/KSlab) for the transverse conning system (i.e.
cross-bracing and/or straps) are shown in Fig. 11c. It should be
noted that the load capacity of the tested precast slabs depends
on the reinforcement ratio and conguration as well as the stiffness of the transverse conning system, which is characteristic of
strength enhancement provided by compressive membrane action
[2].
With regard to Fig. 11a and b, it can be concluded that mobilising the arching action and the subsequent strength enhancement
in the transversely conned precast slab decks (TCPSD) come at a
cost of lower ductility. It is observable that for slabs with less than
1% tensile reinforcement, the measures of ductility (i.e. lE and J)
consistently decrease with an increase in the stiffness of conning
system in the transverse direction. However, it should be noted
that the variation of ductility indices for the TCPSD depends not
only on the transverse stiffness provided for the precast slabs but
also on the conguration of the conning system (i.e. only bracing,
only strap, strap + bracing) as well as reinforcing ratio in the slab.
For example, in series M with q = 1.5%, it is seen that the J factor
increases from 4.69 to 11.64 (Fig. 11b), when the KConning system/
KSlab ratio increases from 0.1 (with only strap) to 0.185 (with only
bracing). This increase in the J factor (despite increase in the transverse stiffness of conning system) can be attributed to the larger
rotational stiffness provided for the slabs by the crass bracings
compared to the system with only straps.
The normalised experimental peak load capacities of the transversely conned slabs with respect to identical slabs without any
transverse restraint (i.e. ratio of the peak load capacity of the slab
with transverse connement over the peak load capacity of the
identical slab without transverse connement) for slabs with

75

Normalised peak load capacity

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

1.6

1.4

Only
bracing

No transverse
confining system

Strap +
bracing

1.2
M6 series (p=1.5%)
M4 series (p=1.0%)
B4 series (p=0.7%)

1
0

(a)

0.1
0.2
KConfining system / KSlab

0.3

Fig. 12. Normalised peak load capacity (ratio of the peak load capacity of the slab
with transverse connement over the peak load capacity of the identical slab
without transverse connement) for slabs with different transverse conning
system.

(b)

(c)
Fig. 11. Variation of (a) ductility index lE and (b) J factor and (c) peak load capacity
of slabs, with respect to reinforcing steel proportion and relative stiffness of
transverse conning system.

derive simple empirical formulas that can capture the enhancing

effect of arching action with sufcient accuracy.
Valipour et al. [36] used an enhanced bre element model to
conduct a parametric study on a generic model (see Fig. 13a) and
determine the inuence of different parameters including span/
depth ratio, concrete compressive strength, reinforcing ratio, stiffness of transverse conning system KConning system and rotational
stiffness Kh of end supports (see Fig. 13a) on the enhancing effect
of arching action in RC beams. Fig. 13b shows the strength
enhancement factor versus relative stiffness of transverse conning system (KConning system/Kaxial) obtained from the parametric
study conducted by Valipour et al. [36]. It is noteworthy that the
strength enhancement factor in Fig. 13b is dened as the ratio of
peak load capacity of the transversely conned/restrained RC
member over peak load capacity of the identical RC member without transverse conning system (KConning system = 0). Furthermore,
the xed and pinned conditions in Fig. 13b refer to the end support
conditions adopted for analyses of the generic model in Fig. 13a.
The plots shown in Fig. 13b are used to estimate an upper and
lower bound value for strength enhancement factor in the tested
slabs and the results are compared with the experimental results
in Fig. 14. It is seen that the bound of strength enhancement predicted by the results of the parametric study correlates reasonably
well with the test data.

4. Concluding remarks
different transverse conning system (i.e. cross-bracing, strap,
cross-bracing + strap) are shown in Fig. 12. It can be seen that
the strength enhancement due to arching action varies inversely
with the reinforcing bar ratio, i.e. those precast slabs with a lower
reinforcing ratio (series B4) have a higher strength enhancement.
3.6. Experimental versus analytical strength enhancement
As demonstrated in Table 2, simple plastic analyses ignore the
effect of arching action and consistently underestimate the load
capacity of the restrained precast slabs. Accordingly, several
attempts have been made to develop analytical methods that can
capture the enhancing effect of arching action [3335]. The existing analytical models are, however, cumbersome and require several iterations to predict the failure load of restrained RC members
with reasonable accuracy [33]. Accordingly, there is still a need to

A set of twelve half-scale precast reinforced concrete slabs were

connected to steel girders using post-installed friction grip bolted
shear connectors (PFBSCs). The slabs were tested under a monotonically increasing point load applied at the mid-span to simulate
wheel loading on a deconstructable composite deck. In the experimental program, the application of three types of transverse conning system (i.e. straps, cross-bracing and a combination of
straps and cross-bracing) in conjunction with friction grip bolted
connections for mobilising arching action and enhancing the peak
load capacity of the precast concrete slabs was studied. Although
the experimental work was focused mainly on steelconcrete composite decks with PFBSCs, the results of the tests can be also
extended to the rehabilitation and structural strength assessment
of conventional steelconcrete composite decks, where the application of cross-bracing (or considering the effect of existing
cross-bracing or cross-beams) can signicantly enhance the load

76

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

(a)

(b)
Fig. 13. (a) Outline of the generic model and denition of parameters (b) strength enhancement factor (i.e. peak load capacity of the transversely conned RC member over
peak load capacity of the RC member without transverse conning system KConning system = 0) versus relative stiffness of conning system, obtained from a parametric study
using generic bre-element models (span/depth = 11) [36].

Normalised peak load

2.5

Analytic-Fixed (p=0.8%)
Analytic-Fixed (p=1.2%)
Exp. (p=1.5%)
Exp. (p=1.0%)
Exp. (p=0.7%)
Analytic-Pinned (p=0.8%)
Analytic-Pinned (p=1.2%)

1.5

1
0.1

0.2
KConfining system / KSlab

0.3

Fig. 14. Comparison between experimental and analytical strength enhancement

predicted by the results of parametric study [36].

carrying capacity of concrete slab decks and, consequently, prevent

unnecessary repair and/or replacement of these decks.
The experimental program in this study provided benchmark
data for evaluating the structural performance (i.e. the mode of
failure, loaddeection response and ductility) and for validating
numerical models of the transversely-restrained precast RC slabs
of composite decks with PFBSCs. In specic regard to the arching
behaviour of precast slabs, the following conclusions are drawn:
 PFBSCs can effectively prevent relative slip between precast
slabs and steel girders in the transverse direction. This is of particular importance for the development of arching action in
deconstructable composite bridge decks with transversely conned precast slab decks (TCPSD) where the restraining system

(cross-bracing and straps) is connected mechanically (using

friction bolts) to the steel girders instead of slab itself, in order
to facilitate dismantling of the structure.
 Despite extensive damage (cracking and crushing) and fairly
large displacements and rotations that the precast slabs experienced at their ultimate loading state, no apparent damage or
deformation was observed in the bolted shear connectors and
the PFBSCs could be removed from the sleeves easily. This
demonstrates the ability of proposed system to accommodate
future repair, rehabilitation and/or dismantling.
 Two distinctive modes of failure were identied for transversely-conned precast slab decks with PFBSCs; the rst mode
of failure was triggered by the development of cracks on the top
surface of the slabs between the PFBSCs and then followed by
crushing of the concrete at mid-span, whereas the second mode
of failure was only associated with crushing of the concrete at
mid-span.
 The development of arching action in the single-span TCPSDs
enhanced the peak load capacity, by up to 51%, but the higher
load capacity of TCPSDs came at a cost of lower ductility, particularly for slabs with a reinforcing ratio in excess of 1%. For
slabs reinforcing with less than 1% tensile reinforcement, which
is typical of slabs in practice, a good compromise between ductility and strength enhancement can be achieved by adjusting
the stiffness of the transverse conning system.
References
[1] Valipour H, Farhangvesali N, Foster S. A generic model for investigation of
arching action in reinforced concrete members. Constr Build Mater
2013;38:74250.
[2] Farhangvesali N, Valipour H, Samali B, Foster S. Development of arching action
in longitudinally-restrained reinforced concrete beams. Constr Build Mater
2013;47:719.

H. Valipour et al. / Construction and Building Materials 87 (2015) 6777

[3] Ruddle ME, Rankin GIB, Long AE. Arching actionexural and shear strength
enhancements in rectangular and Tee beams. Proc Inst Civ Eng Struct Build
2002;156:6374.
[4] Bakht B, Lam C. Behavior of transverse conning systems for steel-free deck
slabs. ASCE J Bridge Eng 2000:13947.
[5] Mufti AA, Bakht B, Newhook JP. Precast concrete decks for slab-on-girder
systems: a new approach. ACI Struct J 2004;101(3):395402.
[6] Hon A, Taplin G, Al-Mahaidi RS. Strength of reinforced concrete bridge decks
under compressive membrane action. ACI Struct J 2005;102(3):393403.
[7] Klowak CS, Mufti AA. Behaviour of bridge deck cantilever overhangs subjected
to a static and fatigue concentrated load. Constr Build Mater 2009;
23(4):165364.
[8] Mufti AA, Newhook JP. Punching shear strength of restrained concrete bridge
deck slabs. ACI Struct J 1998;95(4):37581.
[9] Taylor SE, Rankin B, Cleland DJ, Kirkpatrick J. Serviceability of bridge deck slabs
with arching action. ACI Struct J 2007;104(1):3948.
[10] Taylor SE, Mullin B. Arching action in FRP reinforced concrete slabs. Constr
Build Mater 2006;20(12):7180.
[11] ACI Innovation Task Group, Report on Bridge Decks Free of Steel
Reinforcement. Report No. ACI ITG-3-04, American Concrete Institute,
Farmington Hills, MI, 2004.
[12] CAN, CSA-S6-06. Canadian Highway Bridge Design Code (CHBDC). Ontario,
Canada: Canadian Standard Association; 2006.
[13] Klowak C, Memon A, Mufti A. Static and fatigue investigation of second
generation steel-free bridge decks. Cement Concr Compos 2006;28(10):8907.
[14] Mufti AA, Jaeger LG, Bakht B, Wegner LD. Experimental investigation of FRC
slabs without internal steel reinforcement. Can J Civ Eng 1993;20(3):398406.
[15] Bakht B, Mufti AA, Tadros G. Providing transverse connement in deck
slabs. Montreal, Canada: Annu. Conf. of Can. Soc. for Civ. Eng.; 1996. p. 44354.
[16] Ataei A, Bradford MA. FE Modelling of sustainable semi-rigid ush end plate
composite joints with deconstructable bolted shear connectors. Palm Cove,
Queensland, Australia: International Conference on Composite Construction
CCVII; 2013.
[17] Bradford MA, Pi YL. Numerical modelling of composite steel-concrete beams
for life-cycle deconstructability. Hong Kong: 1st International Conference on
Performance-Based and Life-Cycle Structural Engineering; 2012.
[18] Rowe M, Bradford MA. Partial shear interaction in deconstructable composite
steel-concrete composite beams with bolted shear connectors. Miskolc,
Hungary: International Conference on Design, Fabrication and Economy of
Welded Structures; 2013.
[19] Kwon G, Engelhardt MD, Klinger RE. Experimental behavior of bridge beams
retrotted with post-installed shear connectors. ASCE J Bridge Eng 2011;
16(4):53645.

77

[20] Kwon G, Engelhardt MD, Klinger RE. Behavior of post-installed shear

connectors under static and fatigue loading. Constr Steel Res 2010;
66(4):53241.
[21] Marshall WT, Nelson HM, Banerjee HK. An experimental study of the use of
high-strength friction-grip bolts as shear connectors in composite beams.
Struct Eng 1971;49(4):1718.
[22] Dallam LN, Harpster JL. Composite beam tests with high-strength bolt shear
connectors. Report No. 683, University of Missouri-Columbia, MO, USA; 1968.
[23] Dallam LN, Push out tests with high strength bolt shear connectors. Report No.
687, University of Missouri-Columbia, MO, USA; 1968.
[24] Lee SSM, Bradford MA. Sustainable composite beams with deconstructable
shear connectors. 5th International Conference on Structural Engineering,
Mechanics and Computation, Cape Town, South Africa. 5th International
Conference on Structural Engineering, Mechanics and Computation. Cape
Town, South Africa; 2013.
[25] AS5100.6. Bridge design steel and composite construction. Australian
Standard; 2004.
[26] AS5100.2. Bridge design Part 2: design load. Australian standard; 2004.
[27] AS5100.5. Bridge Design, Part 5: concrete. Australian Standard; 2004.
[28] Spadea G, Swamy RN, Bencardino F. Strength and ductility of RC beams
repaired with bonded CFRP laminates. J Bridge Eng 2001;6(5):34955.
[29] Grace NF, Soliman AK, Abdel-Sayed G, Saleh KR. Behavior and ductility of
simple and continuous FRP reinforced beams. J Compos Constr
1998;2(4):18694.
[30] Sakka ZI, Gilbert I. Effect of reinforcement ductility on strength and failure
modes of one-way reinforced concrete slabs. Report No. UNICIV R-450, The
University of New South Wales, NSW, Australia; 2008.
[31] Mufti AA, Newhook JP, Tadros G. Deformability versus ductility in concrete
beams with FRP reinforcement. Montreal, Quebec: Proceeding of the second
international conference on advanced composite materials in bridges and
structures; 1996. p. 18999.
[32] Erp GM. Robustness of bre composite structures loaded in exure. Hong
Kong, China: Proceeding of international conference on FRP composites in civil
engineering (CICE); 2001. p. 14216.
[33] Yu J, Tan KH. Analytical model for the capacity of compressive arch action of
reinforced concrete sub-assemblages. Mag. Concr. Res. 2014;66(3):10926.
[34] Taylor SE, Rankin GIB, Cleland DJ. Arching action in high-strength concrete
slabs. Proc Inst Civ Eng Struct Build 2001;146(4):35362.
[35] Rankin GIB, Long AE. Arching action strength enhancement in laterallyrestrained slab strips. Proc Inst Civ Eng Struct Build 1997;122(4):4617.
[36] Valipour HR, Vesali N, Samali B, Foster S. Reserve of strength in reinforced
concrete frames: analysis of arching action. Aust J Struct Eng 2014;
15(2):16175.