1

Strength AnaIysis of SteeI-Concrete Composite Beams in

Combined Bending and Shear
Qing Quan Liang, M.ASCE

; Brian Uy, M.ASCE

; Mark A. Bradford, M.ASCE

; and Hamid R.
Ronagh

Abstract: Despite experimental evidences, the contributions oI the concrete slab and
composite action to the vertical shear strength oI simply supported steel-concrete composite
beams are not considered in current design codes, which lead to conservative designs. In this
paper, the Iinite element method is used to investigate the Ilexural and shear strengths oI
simply supported composite beams under combined bending and shear. A three-dimensional
Iinite element model has been developed to account Ior geometric and material nonlinear
behavior oI composite beams, and veriIied by experimental results. The veriIied Iinite
element model is than employed to quantiIy the contributions oI the concrete slab and
composite action to the moment and shear capacities oI composite beams. The eIIect oI the
degree oI shear connection on the vertical shear strength oI deep composite beams loaded in
shear is studied. Design models Ior vertical shear strength including contributions Irom the
concrete slab and composite action and Ior the ultimate moment-shear interaction are
proposed Ior the design oI simply supported composite beams in combined bending and
shear. The proposed design models provide a consistent and economical design procedure Ior
simply supported composite beams.
CE Database subject headings: Bending; Composite structures; Finite element method;
Shear; Strength.
1
Australian Postdoctoral Fellow, School oI Civil and Environmental Engineering, The
University oI New South Wales, Sydney, NSW 2052, Australia; ProIessor, Central South
University, Changsha, Hunan, China. E-mail: q.liangunsw.edu.au
2
Associate ProIessor, School oI Civil and Environmental Engineering, The University oI
New South Wales, Sydney, NSW 2052, Australia.
2
3
ProIessor oI Civil Engineering, School oI Civil and Environmental Engineering, The
University oI New South Wales, Sydney, NSW 2052, Australia.
4
Lecturer, Department oI Civil Engineering, University oI Queensland, Brisbane, QLD 4072,
Australia.
Introduction
Steel-concrete composite beams have been extensively used in building and bridge
construction. Composite action in a composite beam is achieved by means oI mechanical
shear connectors. Headed stud shear connectors are usually welded to the top Ilange oI a steel
beam to resist longitudinal slip and vertical separation between the concrete slab and the steel
beam. Concrete slabs can be either solid slabs or composite slabs incorporating proIiled steel
sheeting. Composite beams under applied loads are oIten subjected to combined actions oI
bending and vertical shear. Despite experimental evidences, the contributions Irom the
concrete slab and composite action to the vertical shear strength oI a simply supported
composite beam is not considered in current design codes, such as AS 2327.1 (1996),
Eurocode 4 (1994) and LRFD (1999), which result in conservative designs (Johnson and
Anderson 1993). In order to design composite beams consistently and economically, it is
necessary to develop new design models Ior shear strength including contributions Irom the
concrete slab and composite action and Ior moment-shear interactions.
Experimental studies on the ultimate strength oI steel-concrete composite beams in combined
bending and shear have been oI interest to researchers. Johnson and Willmington (1972)
conducted experiments on continuous composite beams in combined negative bending and
vertical shear. Their test results indicated that longitudinal steel reinIorcement in the concrete
slab increases the strength and stiIIness in vertical shear oI a composite beam. Allison et al.
(1982) tested Iive composite plate girders and one steel plate girder under negative bending
3
and shear to Iailure. Porter and CheriI (1987) studied experimental behavior oI simply
supported composite plate girders loaded primarily in shear. They proposed a shear strength
model that incorporates contributions Irom both the concrete slab and the steel plate girder Ior
the design oI composite beams.
Research on the behavior oI composite beams with web openings indicated that the concrete
slab contributes signiIicantly to the vertical shear strength oI a composite section at web
openings. Tests on short-span composite plate girders with web openings have been carried
out by Narayanan et al. (1989) and Roberts and Al-Amery (1991). These tests showed that the
shear strength oI a composite plate girder is signiIicantly higher than that oI a steel plate
girder alone iI adequate shear connectors are provided in the composite girder. In addition, the
composite action under predominantly shear loading depends on the tensile or pullout strength
oI the shear connectors. Analytical models including a contribution Irom the concrete slab
were proposed Ior determining the shear strength oI composite plate girders. Experiments
conducted by Clawson and Darwin (1982) and Donahey and Darwin (1988) indicated that the
behavior oI composite beams with web openings is largely controlled by the moment-shear
ratio at the opening. Darwin and Donahey (1988) proposed an equation to express the ultimate
moment-shear relationship Ior composite beams with web openings.
Numerical analysis methods have been used to analyze the inelastic behavior oI composite
beams. Yam and Chapman (1968) presented an iterative numerical method Ior the inelastic
analysis oI simply supported composite beams. The inelasticity oI steel, concrete and shear
connections was taken into account in the analysis. Hirst and Yeo (1980) used a standard
Iinite element program to analyze composite beams with partial and Iull shear connection.
Quadrilateral elements were employed to simulate discrete stud shear connectors. The
4
material properties oI stud elements were modiIied to make them equivalent in strength and
stiIIness to the actual shear connectors in composite beams. A three-dimensional bar element
has been developed by Razaqpur and NoIal (1989) Ior modeling the nonlinear behavior oI
shear connectors in composite beams. An empirical shear-slip relationship was used to
express the stiIIness properties oI the bar element.
Al-Amery and Roberts (1990) presented a nonlinear analysis oI composite beams with partial
shear connection by using a Iinite diIIerence method. Salari et al. (1998) Iormulated a
composite beam element based on the Iorce analysis method Ior the nonlinear analysis oI
composite beams with deIormable shear connectors. A distributed spring model was used to
simulate shear connectors. Thevendran et al. (1999) utilized the Iinite element soItware
ABAQUS to study the ultimate load behavior oI composite beams curved in plan. Shell
elements were used to model the concrete slab and the steel beam whilst a rigid beam element
was employed to simulate stud shear connectors. Sebastian and McConnel (2000) described a
nonlinear Iinite element program Ior modeling composite beams. Axial springs with empirical
shear-slip relations were used to model discrete shear connectors. A kinematic model was
proposed by Fabbrocino et al. (2000) Ior analyzing continuous composite beams with partial
interaction and bond. Baskar et al. (2002) investigated the ultimate strength oI composite plate
girders under negative bending by using the Iinite element soItware ABAQUS. Furthermore,
Liang et al. (2004a) has undertaken nonlinear Iinite element analyses on continuous
composite beams in combined bending and shear. In their study, design Iormulas
incorporating contributions Irom the concrete slab and composite action were proposed Ior
the vertical shear strength and the ultimate strength interaction oI continuous composite
beams.
5
In this paper, the ultimate Ilexural and shear strengths oI simply supported composite beams
in combined bending and shear are investigated by using the Iinite element analysis method.
A three-dimensional Iinite element model, which accounts Ior geometric and material
nonlinear behavior oI composite beams, is described in detail. The Iinite element model is
veriIied by corresponding experimental results. The veriIied Iinite element model is then used
to study the interaction behavior oI composite beams subjected to combined actions oI
bending and shear. The eIIects oI shear connection on the vertical shear strength oI composite
beams are investigated. Based on the numerical results, design models Ior vertical shear
strength and Ior moment-shear interactions are developed Ior the design oI simply supported
composite beams.
Finite EIement AnaIysis
General
The general-purpose Iinite element program ABAQUS version 6.3 (2002) was used in the
present study to investigate the ultimate Ilexural and shear strengths oI composite beams
subjected to combined bending and shear. A three-dimensional Iinite element model has been
developed to account Ior geometric and material nonlinear behavior oI composite beams. The
concrete slab, steel Ilanges and web were modeled by Iour-node doubly curved thick/thin
shell elements with reduced integration. A 3D-beam element was employed to simulate
discrete stud shear connectors. The von Mises yield criterion was used in the nonlinear
analysis to treat the plasticity oI steel material with Iive integration points through the
thickness. A typical Iinite element discretization oI a composite beam used in the present
study is shown in Fig. 1.
6
Steel Modeling
Steel Section
Tests indicate that structural steels in uniaxial tension exhibit strain hardening behavior that is
diIIerent Irom the elastic-perIectly plastic assumption (Kemp et al 2002). The stress-strain
curve with strain hardening used in the nonlinear analysis has shown to predict well the
behavior oI structural steel (Liang and Uy 2000; Liang et al. 2004b). In the present study,
structural steel sections were modeled as an elastic-plastic material with strain hardening. A
bilinear stress-strain relationship shown in Fig. 2 was used Ior steel sections in both
compression and tension. Material properties, such as the Young's modulus, Poisson's ratio,
the yield stress, the ultimate strength and the ultimate strain, need to be input to deIine the
stress-strain curve. Experimental values oI the yield and ultimate strengths were used in the
analysis Ior steel sections. An ultimate strain oI 0.25 was assumed Ior mild structural steel.
Steel Reinforcement
Steel reinIorcing bars in concrete slabs were modeled in the present study as smeared layers
with a constant thickness in shell elements. The thickness oI a steel layer was calculated as the
area oI a reinIorcing bar divided by the spacing oI reinIorcing bars. In the input data Iile,
reinIorcement in a concrete slab was deIined by the Rebar Layer option within the shell
section that deIined the concrete slab. Four layers were used to represent the top and bottom
longitudinal and transverse reinIorcing bars in the concrete slab in a composite beam. The
cross-sectional area oI the reinIorcing bar, spacing, distance Irom the mid surIace oI the
concrete slab, material property name, angle to the reIerence axis and the reIerence axis were
input to deIine each rebar layer. The material property oI reinIorcing bars was deIined in the
7
material section. The bilinear stress-strain relationship shown in Fig. 2 was also used in the
present study Ior reinIorcing bars.
Concrete Modeling
Concrete in Compression
Concrete in compression was modeled as an elastic-plastic material with strain soItening. The
stress-strain relationship Ior concrete in uniaxial compression proposed by Carreira and Chu
(1985) was adopted in the present study as
( )

) / ( 1
/
'
c c
'
c c
'
c
c
+
=

(1)
where
c
compressive stress in concrete;
c
strain in concrete;
'
c
f the cylinder
compressive strength oI concrete;
'
c
the strain corresponding to
'
c
f ; and is deIined by
55 . 1
4 . 32
3
'
c
+ =
f
(2)
in which
'
c
f is in MPa. The strain
'
c
is usually taken as 0.002. A stress-strain curve Ior
concrete with a compressive strength oI 42.5 MPa is shown in Fig. 3. In the present study, the
stress-strain behavior oI concrete in compression was assumed to be linear elastic up to
0.4
'
c
f . Beyond this point, it was in the plastic regions in which plastic strain was input to
deIine the stress-strain relationship in the Iinite element model. The Iailure ratio option was
8
used to deIine the Iailure surIace oI concrete. The ratio oI the ultimate biaxial compressive
stress to the ultimate uniaxial compressive stress was taken as 1.16. The ratio oI the uniaxial
tensile stress to the uniaxial compressive stress at Iailure was taken as 0.0836.
Concrete in Tension
The behavior oI concrete and reinIorcement in a concrete slab was modeled independently.
The interaction between the concrete and reinIorcing bars was simulated approximately by the
tension stiIIening model. The model assumes that the direct stress across a crack gradually
reduces to zero as the crack opens. Tension stiIIening was deIined in the present study using
stress-strain data. The stress-strain relationship as shown in Fig. 4 assumes that the tensile
stress increases linearly with an increase in tensile strain up to concrete cracking. AIter
concrete cracking, the tensile stress decreases linearly to zero as the concrete soItens. The
value oI tension stiIIening is an important parameter that aIIects the solution oI a nonlinear
analysis oI reinIorced concrete. Tension stiIIening is inIluenced by the density oI reinIorcing
bars, the bond, the relative size oI the aggregate compared to the rebar diameter and the Iinite
element mesh. For heavily reinIorced concrete slabs, the total strain at which the tensile stress
is zero is usually taken as ten times the strain at Iailure in the tension stiIIening model.
However, it has been Iound that this value was not adequate Ior concrete slabs in composite
beams (Basker et al. 2002; Liang et al. 2004a). In the present study, a total strain oI 0.1 was
used Ior reinIorced concrete slabs in composite beams.
Shear Retention
9
The reduction in shear modulus due to concrete cracking was deIined as a Iunction oI direct
strain across the crack in the shear retention model. The shear modulus oI cracked concrete is
deIined as
c
G G = , where
c
G is the elastic shear modulus oI uncracked concrete and is a
reduction Iactor, which is given by
) / 1 (
max c
= Ior
max c
< (3)
0 = Ior
max c
(4)
in which
c
direct strain across the crack. The shear retention model states that the shear
stiIIness oI open cracks reduces linearly to zero as the crack opening increases. Parameters
max
0.005 and 95 . 0 = were used in the present study to deIine the shear retention oI
concrete, as suggested by Thevendran et al. (1999) and Liang et al. (2004a).
Shear Connector Modeling
Wright (1990) suggested that the stud shear connection should be modeled as a discrete
connection to accurately predict the nonlinear behavior oI composite beams with partial
interaction. A three-dimensional beam element was employed in the present study to model
discrete stud shear connectors. Shear connectors were assumed to connect the middle plane oI
the concrete slab and the top Ilange oI the steel beam. The cross-sectional area oI the beam
element was modiIied to make it equivalent in both strength and stiIIness to the actual stud
shear connector in a composite beam. The bilinear stress-strain relationship illustrated in Fig.
2 was used Ior the shear connector material. Pin jointed truss elements with an eIIective
stiIIness were used in place oI shear connectors to transIer direct stress Irom the concrete slab
10
to the top Ilange oI the steel beam. This model can be used to simulate any degree oI shear
connection in composite beams.
Solution Method
The nonlinear response oI structural concrete is highly discontinuous due to cracking. To
prevent the discontinuity, the Controls, AnalysisDiscontinuous option was speciIied in the
nonlinear analysis oI composite beams. Local instabilities oIten occur in the nonlinear
analysis oI reinIorced concrete elements because oI the large amounts oI cracking. The
modiIied Riks method was thereIore used in the present study to prevent the local
instabilities. The automatic load control scheme was employed. The deIlection at midspan oI
the composite beam was monitored in the analysis.
Validation of Finite Element Models
The Iinite element model developed herein has been used to analyze a simply supported
composite beam (E1) tested by Chapman and Balakrishnan (1964) and the results are
compared with corresponding experimental data in this section. The span oI the composite
beam under a point load was 5.5 m. The cross section oI the composite beam is shown in Fig.
5. Material properties oI the composite beam are given in Table 1. The Iinite element
discretization oI the composite beam is shown in Fig. 1. The concrete slab was modeled with
13 61 elements. The Ilange oI the steel beam was modeled using 2 60 elements while
the steel web was modeled with 3 60 elements. The load-deIlection curve oI the composite
beam obtained by the present study is compared with that obtained by experiments in Fig. 6.
It can be observed Irom Fig. 6 that the initial stiIIness oI the composite beam predicted by the
11
Iinite element model is the same as that oI experimental one. The ultimate load obtained by
the present study was 494 kN, which is 95.3 oI the experimental value. The nonlinear Iinite
element analysis conIormed the experimental observation that the composite beam Iailed by
crushing oI the top concrete slab at midspan. It can be concluded that the Iinite element model
developed herein is reliable and conservative in predicting the ultimate strength oI composite
beams.
Load-DefIection Behavior
The Iinite element model developed has been used to investigate the ultimate load behavior oI
simply supported composite beams with various moment/shear ratios ( ) under combined
actions oI bending and shear. A point load was applied to the midspan oI all composite beams
in the analysis. The span oI the composite beam (E1) tested by Chapman and Balakrishnan
(1964) was varied to give diIIerent combinations oI moment and shear while other conditions
oI the composite beam were unchanged. The moment/shear ratios used in the analysis were
0.4, 0.5, 0.75, 1.0, 1.25, 1.75 and 2.75 m, which correspond to the spans oI 0.8, 1.0, 1.5, 2.0,
2.25, 3.5 and 5.5 m, respectively. Since these composite beams with various spans were
actually cut short Irom the one tested by Chapman and Balakrishnan (1964), the degree oI
shear connection was approximately same Ior all cases. Material properties given in Table 1
were used Ior all cases.
The load-deIlection curves obtained Irom the results oI the nonlinear Iinite element analysis
on composite beams with various moment/shear ratios are shown in Fig. 7. It can be seen
Irom Fig. 7 that the response oI composite beams to applied loads is initially linear. AIter
concrete cracking, steel yielding and large deIormations, the nonlinear load-deIlection
12
behavior is observed. It is seen that the strength and stiIIness oI composite beams decrease
with an increase in the moment/shear ratio. The ultimate load oI composite beams decreases
with an increase in the moment/shear ratio. This is justiIied by the Iact that Ior the same
composite section, increasing the span oI the composite beam will reduce the load carrying
capacity oI the composite beam. When the moment/shear ratio was high, the composite beam
Iailed by Ilexure. In contrast, the composite beam Iailed by shear when the moment/shear
ratio was low such as the beam with a 0.4 or 0.5, as indicated in Fig. 7. The ultimate
loads oI these two beams are almost the same as they reach the same ultimate shear strength
oI the same composite section.
Moment-Shear Interaction
The behavior oI composite beams depends on the moment/shear ratios. The moment-shear
interaction strength oI composite beams has been investigated by undertaking nonlinear
analyses on composite beams with various spans and with the same cross section and material
properties, as discussed in the previous section. The ultimate moment and shear capacities oI
these composite beams can be calculated Irom the ultimate loads predicted by the nonlinear
Iinite element analysis, and are shown in Fig. 8. The ultimate moment capacity is almost not
aIIected by the vertical shear when the moment/shear ratio is high such as the point with the
maximum moment capacity shown in Fig. 8. When the vertical shear exceeds halI oI the
ultimate shear strength oI the composite section, the ultimate Ilexural strength oI the
composite beam is reduced with an increase in the vertical shear. A weak interaction between
moment and shear strength is observed when the moment/shear ratio is low. This means that
the composite beam can withstand similar maximum shear Iorces while the applied moments
are quite diIIerent. The maximum ultimate shear strength oI the composite beam obtained
Irom the results oI the Iinite element analysis is 804 kN, while it is only 439 kN according to
13
AS 2327.1 (1996). Through composite action, the concrete slab increases the maximum shear
strength oI the composite beam by 85.
Effect of Shear Connection on VerticaI Shear Strength
The eIIect oI the degree oI shear connection on the ultimate moment capacities oI simply
supported composite beams is reIlected in design codes, such as AS 2327.1 (1996), Eurocode
4 (1994) and LRFD (1999). The codes assume that the web oI the steel beam resists the entire
vertical shear, and do not consider the eIIect oI shear connection on the vertical shear strength
oI composite beams. This assumption allows Ior a simple model to be given but results in
conservative designs. In real composite construction, the vertical shear strength oI a
composite beam is in Iact a Iunction oI the degree oI shear connection (Donahey and Darwin
1988). To quantiIy this eIIect, a simply supported composite beam with a span oI 0.8 m and
with various degrees oI shear connection has been analyzed. This deep composite beam is a
non-Ilexural member where the shear load is transIerred to the supports by a strut-and-tie
model, as reported by Liang et al. (2000, 2002). The composite beam was a shortened version
oI the one tested by Chapman and Balakrishnan (1964). The cross section oI the composite
beam is shown in Fig. 5. Only the cross-sectional area oI stud shear connectors was modiIied
to give diIIerent degrees oI shear connection while other conditions oI the composite beam
were unchanged. Material properties given in Table 1 were used in the analysis.
Fig. 9 shows the ultimate shear strength oI the composite beam with various degrees oI shear
connection obtained Irom the Iinite element analysis. It can be observed Irom Fig. 9 that the
vertical shear strength oI the composite beam increases with an increase in the degree oI shear
connections ( ). This conIirms experimental Iindings presented by Donahey and Darwin
14
(1988). When 1 > , the vertical shear strength is not aIIected by the degree oI shear
connection. This indicates that the composite beam exhibits Iull shear connection. It is also
observed Irom Fig. 9 that the vertical shear strength oI a composite beam with Iull shear
connection is 29.5 higher than that oI the one without composite action.
Proposed Design ModeIs
Design Models for Vertical Shear Strength
Experiments and nonlinear Iinite element analyses indicated that the concrete slab and
composite action make signiIicant contributions to the vertical shear strength oI a composite
beam. To take advantage oI composite actions, a design model Ior the vertical shear strength
oI simply supported composite beams with any degree oI shear connection is proposed as
( ) 295 . 0 1
o uo
+ =J J ) 1 0 ( (5)
where
uo
J the ultimate shear strength oI the composite beam in pure shear;
o
J the ultimate
shear strength oI the non-composite beam in pure shear (with zero degree oI shear
connection); and the degree oI shear connection. It should be noted that the pullout
Iailure oI stud shear connectors results in the damage oI composite action. II this occurs, the
ultimate shear strength oI the damaged composite beam (
uo
J ) should be taken as
o
J Ior
saIety. The proposed design model Ior vertical shear strength is compared with the results
obtained Irom the nonlinear Iinite element analysis in Fig. 10. It is shown that the design
model agrees very well with numerical predictions.
15
II no shear connection is provided between the concrete slab and the steel beam, the two
components will work independently to resist vertical shear. The superposition rule can be
applied to the vertical shear strength oI the non-composite section. The vertical shear strength
oI a non-composite beam can be expressed by
s c o
J J J + = (6)
where
c
J the contribution oI the concrete slab; and
s
J the shear capacity oI the web oI the
steel beam. Tests indicated that the pullout Iailure oI stud shear connectors in composite
beams might occur (Narayanan et al. 1989). This Iailure mode may reduce the shear
resistance oI the concrete slab. ThereIore, the contribution oI the concrete slab (
c
J ) should be
taken as the lesser oI the shear strength oI the concrete slab
slab
J and the pullout capacity oI
stud shear connectors
p
T . The shear strength oI the concrete slab is proposed as
( )
ec
3 / 1
'
c slab
16 . 1 A f J = (7)
where
'
c
f the compressive strength oI the concrete (MPa); and
ec
A the eIIective shear area
oI concrete. The eIIective shear area oI concrete in a solid slab can be evaluated as
c c I ec
) ( D D b A + = , in which
I
b the width oI the top Ilange oI the steel beam and
c
D the
total depth oI the concrete slab. For a composite slab with proIiled steel sheeting orientated
perpendicular to the steel beam,
ec
A can be taken as ) )( (
r c c r I
h D D h b + + , in which
r
h the
rib height oI the proIiled steel sheeting. The eIIect oI longitudinal steel reinIorcement in the
concrete slab is not considered in Eq. (7). The model gives a good estimate to the shear
16
strength oI the concrete slab in which there is little longitudinal steel reinIorcement passing
through the eIIective shear area in a composite section in the positive moment region.
The pullout capacity oI stud shear connectors in composite beams with solid slabs can be
expressed as
ct c c s p
| 2 ) ( | f h s h d T + + = (pair studs) (8)
ct c c s p
) ( f h h d T + = (single stud) (9)
where
s
d the head diameter oI the stud;
c
h the total height oI the stud; s the transverse
spacing oI studs; and
ct
f the tensile strength oI concrete (MPa). The pullout capacity oI stud
shear connectors in composite slabs incorporating proIiled steel sheeting should be calculated
using the eIIective pullout Iailure surIace in Eqs. (8) and (9).
The shear capacity oI the web oI the steel beam can be determined by (Trahair and BradIord
1991)
w w yw w s
6 . 0 t d f J = (10)
where
yw
f the yield strength oI the steel web (MPa);
w
d the depth oI the steel web;
w
t the
thickness oI the steel web; and
w
the reduction Iactor Ior slender webs in shear buckling.
The reduction Iactor
w
is equal to 1.0 Ior stocky steel webs without shear buckling.
Design Model for Strength Interaction
17
Both the ultimate moment and shear capacities oI a composite beam under combined actions
oI bending and shear are a Iunction oI the degree oI shear connection. The eIIect oI the
vertical shear on the ultimate moment capacity oI composite beams is considered in AS
2327.1(1996) and Eurocode 4 (1994) by using interaction equations. However, design codes
allow only the shear strength oI the steel web to be considered in the interaction equations. To
determine the Ilexural and shear strengths oI simply supported composite beams, design
model Ior strength interactions is proposed as
1
6
uo
u
6
uo
u
=

J
J
M
M
(11)
where
u
M the ultimate moment capacity oI the composite beam in combined bending and
shear;
uo
M the ultimate moment capacity oI the composite beam in pure bending;
u
J the
ultimate shear strength oI the composite beam in combined bending and shear; and
uo
J the
ultimate shear strength oI the composite beam in pure shear. The proposed design model Ior
strength interaction is compared with the results Irom the Iinite element analysis in Fig. 8. It is
seen Irom Fig. 8 that the proposed design model agrees very well with the numerical results.
The ultimate moment capacity (
uo
M ) oI a composite section with any degree oI shear
connections in pure bending can be determined by the rigid plastic analysis method (Oehlers
and BradIord 1999; AS 2327.1, 1996; Eurocode 4, 1994; LRFD 1999). The ultimate shear
strength (
uo
J ) oI a composite beam under pure shear loading can be evaluated by Eq. (5). Any
point (
u
M ,
u
J ) on the moment-shear interaction curve shown in Fig. 8 corresponds to the
applied moment/shear ratio that deIines the load path. This means that the ultimate
18
moment/shear ratio (
u u
/J M ) is also equal to . II the applied moment/shear ratio is known,
the ultimate moment and shear capacities oI a composite beam in combined bending and
shear can be determined by solving the only unknown in Eq. (11).
ConcIusions
The ultimate Ilexural and shear strengths oI simply supported composite beams under
combined bending and shear have been investigated by using the Iinite element method in this
paper. A three-dimensional Iinite element model, which incorporates geometric and material
nonlinear behavior oI the reinIorced concrete slab, stud shear connectors and the steel beam in
a composite beam, has been presented Ior the nonlinear analysis oI composite beams with any
degrees oI shear connection. The eIIects oI the concrete slab on the Ilexural and shear
strengths were taken into account in the analysis. The load-deIlection behavior oI composite
beams with various moment/shear ratios has been demonstrated. The eIIects oI the degree oI
shear connection on vertical shear strength oI composite beams have also been studied.
Design models Ior the vertical shear strength and Ior ultimate moment-shear interactions have
been developed Ior the design oI simply supported composite beams under combined actions.
The Iinite element models developed in this study predict well the ultimate strength oI
composite beams in combined bending and shear. Numerical results indicate that the vertical
shear strength oI composite beams increases with an increase in the degree oI shear
connection. The design model Ior vertical shear strength is proposed as a Iunction oI the shear
capacity oI the non-composite section and the degree oI shear connection. The proposed shear
strength equation Ior the non-composite section comprises contributions Irom the concrete
slab and the steel beam. The behavior oI composite beams depends on the moment/shear
19
ratio. II the applied moment and shear Iorce at the cross section oI a composite beam is
known, the moment-shear interaction equation developed can be used to determine the
ultimate moment and shear capacities oI the composite beam. Although the proposed design
models have been based on the nonlinear analysis oI the tested composite beam, similar
design models Ior continuous composite beams have been veriIied by experimental results
(Liang et al. 2004a). The design models presented in this paper are applicable to simply
supported composite beams with any section. The proposed design models take account oI the
eIIects oI the concrete slab and composite action on both the ultimate moment and shear
capacities oI composite beams, and thus provide a consistent and economical design
procedure Ior simply supported composite beams.
AcknowIedgments
This work has been supported by the Discovery-Projects Grants provided by the Australian
Research Council. The Iinancial support is grateIully acknowledged.
Notation
The following svmbols are used in this paper.
ec
A eIIective shear area oI concrete;
I
b width oI the top Ilange oI steel beam;
w
d depth oI the web oI steel beam;
c
D total depth oI the concrete slab;
s
d diameter oI the head oI headed stud shear connector
20
c
E Young`s modulus oI concrete;
s
E Young`s modulus oI steel;
'
c
f
cylinder compressive strength oI concrete;
ct
f concrete tensile strength
su
f ultimate strength oI steel;
sy
f yield strength oI steel;
yw
f yield stress oI the web oI steel beam;
G shear modulus oI cracked concrete;
c
G elastic shear modulus oI uncracked concrete;
c
h height oI shear connector
r
h rib height oI proIiled steel sheeting;
u
M ultimate moment capacity oI composite beam;
uo
M ultimate moment capacity oI composite beam in pure bending;
p
T pullout capacity oI stud shear connectors
w
t thickness oI steel web;
c
J shear contribution oI the concrete slab
o
J ultimate shear strength oI non-composite beam;
s
J ultimate shear strength oI the steel web
slab
J shear strength oI the concrete slab
u
J ultimate shear strength oI composite beam in combined bending and shear
uo
J ultimate shear strength oI non-composite beam in pure shear;
moment/shear ratio, J M / = ;
21
w
reduction Iactor Ior slender web;
degree oI shear connection;

parameter used to deIine stress-strain curve Ior concrete;

c
strain in concrete;
'
c

strain in concrete corresponding to
'
c
f ;
max
maximum direct strain;
s
strain in steel;
su
ultimate strain in steel;
sy
yield strain in steel;
Poisson`s ratio;
c
compressive stress in concrete;
s
stress in steel; and

reduction Iactor.
References
ABAQUS Users Manual, Jersion 6.3 (2002). Hibbitt, Karlsson and Sorensen, Inc.,
Pawtucket, R.I.
Al-Amery, R. I. M., and Roberts, T. M. (1990). 'Nonlinear Iinite diIIerence analysis oI
composite beams with partial interaction. Comput. & Struct. 35(1), 81-87.
Allison, R. W., Johnson, R. P., and May, I. M. (1982). 'Tension Iield action in composite
plate girders. Proc. Inst. Civ. Eng. Part 2, 73, 255-276.
22
AS 2327.1 (1996). Composite Structures, Part I. Simplv Supported Beams. Standards
Australia, Sydney.
Baskar, K., Shanmugam, N. E., and Thevendran, V. (2002). 'Finite-element analysis oI steel-
concrete composite plate girder. J. Struct. Eng. 128(9), 1158-1168.
Carrreira, D. J., and Chu, K. H. (1985). 'Stress-strain relationship Ior plain concrete in
compression. ACI Struct. J. 82(11), 797-804.
Chapman, J. C., and Balakrishnan, S. (1964). 'Experiments on composite beams. The Struct.
Eng. 42(11), 369-383.
Clawson, W. C., and Darwin, D. (1982). 'Tests oI composite beams with web openings. J.
Struct. Eng. 108(ST1), 145-162.
Darwin, D., and Donahey, R. C. (1988). 'LRFD Ior composite beams with unreinIorced web
openings. J. Struct. Eng. 114(3), 535-552.
Donahey, R. C., and Darwin, D. (1988). 'Web openings in composite beams with ribbed
slabs. J. Struct. Eng. 114(3), 518-534.
Eurocode 4 (1994). Design of Composite Steel and Concrete Structures, Part 1.1, General
Rules and Rules for Buildings.
Fabbrocino, G., ManIredi, G., and Cosenza, E. (2000). 'Analysis oI continuous composite
beams including partial interaction and bond. J. Struct. Eng. 126(11), 1288-1294.
Hirst, M. J. S., and Yeo, M. F. (1980). 'The analysis oI composite beams using standard Iinite
element programs. Comput. & Struct. 11, 233-237.
Johnson, R. P., and Anderson, D. (1993). Designers Handbook to Eurocode 4. Thomas
TelIord, London.
Johnson, R. P., and Willmington, R. T. (1972). 'Vertical shear in continuous composite
beams. Proc. Inst. Civ. Eng. Part 2, 53, 189-205.
23
Kemp, A. R., ByIield, M. P., and Nethercot, D. A. (2002). 'EIIect oI strain hardening on
Ilexural properties oI steel beams. The Struct. Eng. 29-35.
Liang, Q. Q., and Uy, B. (2000). 'Theoretical study on the post-local buckling oI steel plates
in concrete-Iilled box columns. Comput. Struct. 75(5), 479-490.
Liang, Q. Q., Uy, B., BradIord, M. A., and Ronagh, H. R. (2004a). 'Ultimate strength oI
continuous composite beams in combined bending and shear. J. Construct. Steel Res.
60(8), 1109-1128.
Liang, Q. Q., Uy, B., and Steven, G. P. (2002). 'PerIormance-based optimization Ior strut-tie
modeling oI structural concrete. J. Struct. Eng. 128(6), 815-823.
Liang, Q. Q., Uy, B., Wright, H. D., and BradIord, M. A. (2004b). 'Local buckling oI steel
plates in double skin composite panels under biaxial compression and shear. J. Struct.
Eng. 130(3), 443-451.
Liang, Q. Q., Xie, Y. M., and Steven, G. P. (2000). 'Topology optimization oI strut-and-tie
models in reinIorced concrete structures using an evolutionary procedure. ACI Struct. J.
97(2), 322-330.
LRFD (1999). Load and Resistance Factor Design Specification for Steel Buildings.
American Institution oI Steel Construction.
Narayanan, R., Al-Amery, R. I. M., and Roberts, T. M. (1989). 'Shear strength oI composite
plate girders with rectangular web cut-outs. J. Construct. Steel Res. 12, 151-166.
Oehlers, D. J., and BradIord, M. A. (1999). Elementarv Behaviour of Composite Steel and
Concrete Structural Members. Butterworth-Heinemann, OxIord.
Porter, D. M., and CheriI, Z. A. E. (1987). 'Ultimate shear strength oI thin webbed steel and
concrete composite girders. Proc. Int. Conf. on Steel and Aluminium Structures, CardiII,
UK, R. Narayanan et al. ed., Elsevier London, 55-64.
24
Razaqpur, A. G., and NoIal, M. (1989). 'A Iinite element Ior modeling the nonlinear
behaviour oI shear connectors in composite structures. Comput. & Struct. 32(1), 169-174.
Roberts, T. M., and Al-Amery, R. I. M. (1991). 'Shear strength oI composite plate girders
with web cutouts. J. Struct. Eng. 117(7), 1897-1910.
Salari, M. R., Spacone, E., Shing, B., and Frangopol, D. M. (1998). 'Nonlinear analysis oI
composite beams with deIormable shear connectors. J. Struct. Eng. 124(10), 1148-1158.
Sebastian, W., and McConnel, R. E. (2000). 'Nonlinear FE analysis oI steel-concrete
composite structures. J. Struct. Eng. 126(6), 662-674.
Thevendran, V., Chen, S., Shanmugam, N. E., and Liew, J. Y. R. (1999). 'Nonlinear analysis
oI steel-concrete composite beams curved in plan. Finite Elements Ana. Des. 32, 125-139.
Trahair, N. S., and BradIord, M. A. (1991). The Behaviour and Design of Steel Structures.
Chapman & Hall, London.
Wright, H. D. (1990). 'The deIormation oI composite beams with discrete Ilexible
connection. J. Construct. Steel Res. 15, 49-64.
Yam, L. C. P., and Chapman, J. C. (1968). 'The inelastic behaviour oI simply supported
composite beams oI steel and concrete." Proc. Inst. Civ. Eng. 41, 651-683.
25
Fig. 1. Typical Iinite element mesh Ior the composite beam
f
su
0
f
sy

sy

su
Strain
s
S
t
r
e
s
s


s
Fig. 2. Stress-strain curve Ior steel with strain hardening
26
0
10
20
30
40
50
0 0.001 0.002 0.003 0.004 0.005 0.006
Strain
c
S
t
r
e
s
s


c

(
M
P
a
)
Fig. 3. Stress-strain curve Ior concrete in compression
f
ct
0

tu
to
Strain
c
S
t
r
e
s
s


c
Fig. 4. Stress-strain curve Ior concrete in tension
27
Fig. 5. Cross-section oI the composite beam (Chapman and Balakirshnan 1964)
0
100
200
300
400
500
600
0 10 20 30 40 50 60 70 80 90
Deflection at midspan (mm)
L
o
a
d

(
k
N
)
Experiment (Chapman and
Balakirshnan 1964)
Finit e element
Fig. 6. Comparison oI results by Iinite element modeling with experimental data
28
1.75
1.25
1.0
0.75
0.5
0.4
0
200
400
600
800
1000
1200
1400
1600
1800
0 10 20 30
Deflection at midspan (mm)
L
o
a
d

(
k
N
)
Fig. 7. Load-deIlection curves oI composite beams with various moment/shear ratios
0
0.2
0.4
0.6
0.8
1
1.2
0 0.2 0.4 0.6 0.8 1 1.2

u
/
uo

u
/

u
o
Finite element
Proposed design model
Fig. 8. Moment-shear interaction oI composite beams
29
0
200
400
600
800
1000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Degree of shear connection

u
o

(
k
N
)
Fig. 9. EIIects oI shear connection on vertical shear strength oI composite beams
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Degree of shear connection

u
o
/

o
Finite element
Proposed design model
Fig. 10. Proposed design model Ior vertical shear strength oI composite beams
30
TabIe 1. Material properties used in the analysis oI composite beams
Material Property Value
Yield stress,
sy
f (MPa)
265
Ultimate strength,
su
f (MPa)
410
Young`s modulus,
s
E (MPa)
3
10 205
Poisson`s ratio, 0.3
Structural steel
Ultimate strain,
su

0.25
Yield stress,
sy
f (MPa)
250
Ultimate strength,
su
f (MPa)
350
Young`s modulus,
s
E (MPa)
3
10 200
Poisson`s ratio, 0.3
ReinIorcing bar
Ultimate strain,
su

0.25
Compressive strength,
'
c
f (MPa)
42.5
Tensile strength,
ct
f (MPa)
3.553
Young`s modulus,
c
E (MPa)
32920
Poisson`s ratio, 0.15
Concrete
Ultimate compressive strain,
cu

0.0045
Spacing, (mm) 110
Number oI rows 2
Yield stress,
sy
f (MPa)
435
Ultimate strength,
su
f (MPa)
565
Young`s modulus,
s
E (MPa)
3
10 200
Poisson`s ratio, 0.3
Stud shear connector
Ultimate strain,
su

0.25
Captions for Figures and Tables
Fig. 1. Typical Iinite element mesh Ior the composite beam
Fig. 2. Stress-strain curve Ior steel with strain hardening
Fig. 3. Stress-strain curve Ior concrete in compression
Fig. 4. Stress-strain curve Ior concrete in tension
31
Fig. 5. Cross-section oI the composite beam
Fig. 6. Comparison oI results by Iinite element modeling with experimental data
Fig. 7. Load-deIlection curves oI composite beams with various moment/shear ratios
Fig. 8. Moment-shear interaction oI composite beams
Fig.9. EIIects oI shear connection on vertical shear strength oI composite beams
Fig.10. Proposed design model Ior vertical shear strength oI composite beams
TabIe 1. Material properties used in the analysis oI composite beams