See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/300377206

Behaviour of composite steel-concrete beams under elevated temperatures

Chapter · November 2012

DOI: 10.1201/b15320-27

CITATIONS READS

0 314

3 authors, including:

Olivia Mirza Brian Uy

Western Sydney University UNSW Sydney
64 PUBLICATIONS   427 CITATIONS    302 PUBLICATIONS   4,907 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Demountable connections for steel and composite structures View project

Demountable connections for steel-concrete composite structures View project

All content following this page was uploaded by Olivia Mirza on 25 April 2016.

The user has requested enhancement of the downloaded file.

 Journal of Constructional Steel Research 65 (2009) 662–674

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

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

Behaviour of headed stud shear connectors for composite steel–concrete beams

at elevated temperatures
O. Mirza 1 , B. Uy ∗
School of Engineering, University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW 1797, Australia

article info a b s t r a c t

Article history: The behaviour of composite steel–concrete beams at elevated temperatures is an important problem. A
Received 13 November 2007 three-dimensional push test model is developed herein with a two-dimensional temperature distribution
Accepted 10 March 2008 field based on the finite element method (FEM) and which may be applied to steel–concrete composite
beams. The motivation for this paper is to increase the awareness of the structural engineering
Keywords: community to the concepts behind composite steel–concrete structural design for fire exposure. The
Composite steel–concrete beams
behaviour of reinforced concrete slabs under fire conditions strongly depends on the interaction of the
Elevated temperature
Finite element method
slabs with the surrounding elements which include the structural steel beam, steel reinforcing and shear
connectors. This study was carried out to consider the effects of elevated temperatures on the behaviour
of composite steel–concrete beams for both solid and profiled steel sheeting slabs. This investigation
considers the load–slip relationship and ultimate load behaviour for push tests with a three-dimensional
non-linear finite element program ABAQUS. As a result of elevated temperatures, the material properties
change with temperature. The studies were compared with experimental tests under both ambient
and elevated temperatures. Furthermore, for the elevated temperature study, the models were loaded
progressively up to the ultimate load to illustrate the capability of the structure to withstand load
during a fire. It is concluded that finite element analysis showed that the shear connector strength under
fire exposure was very sensitive. It is also shown that profiled steel sheeting slabs exhibit greater fire
resistance when compared with that of a solid slab as a function of their ambient temperature strength.
© 2008 Elsevier Ltd. All rights reserved.

1. Introduction because the mechanical properties of the material decrease as

the temperature rises and likewise, the structural resistance of
Composite steel–concrete beams are formed by connecting a member decreases with temperature rise. Fire safety design
concrete slabs to a supporting structural steel beam. In the is an important aspect of building design because a properly
early 1900s, composite beams were considered favourable for designed building system greatly reduces the hazards to life and
bridge design but in recent decades, composite steel and concrete limits property loss. Research on fire safety design started almost
structures have been employed extensively in modern high-rise 80 years ago [1].
buildings. The flexural strength of composite beams is greatly When a composite steel–concrete beam is subjected to fire,
influenced by the strength and ductility of the shear connectors both the structural steel beam and concrete slab are exposed
between the structural steel beam and the concrete slab. The directly to fire. On the other hand, shear connectors are indirectly
behaviour of the shear connectors is mainly relevant to the design heated by heat transfer from the structural steel as illustrated
for the relationship between shear force transmitted and the in Fig. 1. Fire will cause these elements to lose their mechanical
degree of slip at the interface of the steel and concrete. strength with respect to the temperatures reached. However, the
Fire is a very complex phenomenon which can cause structural mechanical behaviour of composite beams exposed to fire is much
damage. Fires can occur at any time in buildings, and the safety more complicated because of the different materials present.
The behaviour and strength of the shear connector are two
of occupants and maintaining the integrity of the structure are of
of the main factors affecting the behaviour of a composite
major importance. The response of a structural member exposed
steel–concrete beam. Therefore, a three-dimensional finite ele-
to fire is governed by the rate that it is heated, and this is
ment model of a push test using ABAQUS [2–4] was developed to
simulate the behaviour of shear connectors under elevated tem-
peratures. The results of numerical push tests under fire are com-
∗ Corresponding author. Tel.: +61 2 4736 0228; fax: +61 2 4736 0137. pared with the selected push tests under ambient temperatures.
E-mail addresses: o.mirza@uws.edu.au (O. Mirza), b.uy@uws.edu.au (B. Uy). The main objectives of this paper are to develop a three-
1 Tel.: +61 2 4736 0402; fax: +61 2 4736 0137. dimensional finite element model with a two-dimensional
0143-974X/$ – see front matter © 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2008.03.008
 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 663

List of notation
A1p A1s cross-sectional area at the bottom of concrete
A2p A2s cross-sectional area at the top of concrete
a, b, c parameters for elevated temperature behaviour for
composite steel–concrete beams
Cp , C a specific heat for concrete or steel structures
E Young’s modulus
Ea,θ slope of linear elastic range
En design action effect for ambient condition
Ef design action effect for fire condition
fc, fc0 characteristic compressive strength of concrete
fc,θ ultimate stress of concrete
fy,θ effective yield strength of structural steel
fp,θ proportional limit of structural steel
G dead load
H enthalpy
kc,t reduction coefficient for tensile strength of concrete
P pressure Fig. 1. Cross-sectional diagram of composite steel and concrete beam subjected to
fire.
Pu ultimate load
Q live load
T, θ temperature reinforcing steel, profiled steel sheeting and shear connectors, the
t time constitutive laws are represented by the stress–strain relationships
εc concrete compressive strain of the materials. In this paper, the mechanical behaviour at
ε0c strain corresponding to fc0 ambient and elevated temperatures is considered. When elevated
εc 1 θ strain at ultimate stress of the concrete temperature is involved, the main properties required to carry
εcu1θ ultimate strain of the concrete out an accurate calculation of the temperature distribution in a
εps strain value before strain hardening begins composite cross-section are the specific heat, thermal expansion
εp,θ strain at the proportional limit and thermal conductivity.
εt,θ limiting strain for yield strain according to temper-
ature 2.2. Concrete properties
εu,θ ultimate strain according to temperature
εy,θ yield strain according to temperature
εus ultimate yield strain of the steel structure 2.2.1. Concrete at ambient temperature
εys yield strain of the steel structure Plain concrete was recommended by Carreira and Chu [5],
ξ load ratio between design effects in fire condition where the stress in compression is assumed to be linear up to a
with respect to design effect in ambient condition stress of 0.4fc0 . Beyond this point, stress is represented as a function
κs , κp the ratio of the proportional cross-sectional area of of strain according to Eq. (1).
the concrete at high temperature
fc0 γ(εc /ε0c )
γ parameter used to define stress–strain curve for σc = (1)
concrete γ − 1 + (εc /ε0c )
λ thermal conductivity
where
σc concrete compressive stress
σt
 0 3
tensile strength of concrete  f 
σus γ =  c  + 1.55 and ε0c = 0.002.
ultimate stress of the steel material 32.4
σys yield stress of the steel material
∆l/l thermal expansion For concrete in tension, the tensile stress is assumed to increase
linearly relative to strain until the concrete cracks. After the
concrete cracks, the tensile stresses decrease linearly to zero. The
temperature field with the field being constant along the section value of strain at zero stress is usually taken to be 10 times the
in the vicinity of the stud in order to simulate the shear connector strain at failure, which is shown in Fig. 2.
behaviour for both a solid slab and profiled slab with elevated tem-
perature for the push test and to analyse how the temperature af- 2.2.2. Thermal properties of concrete
fects the behaviour of the push tests. Subsequently, the behaviour
of shear connectors in composite beams at elevated temperatures 2.2.2.1. General. An important design consideration for concrete
will be discussed herein. includes the effects of fire. The behaviour of concrete slabs
subjected to fire conditions is complex. In a fully developed fire,
2. Mechanical behaviour of the constituent materials to prevent fire spread to the upper floors, the slab has to carry and
withstand the applied loads and prevent collapse during and after
2.1. General the fire.
The effect of fire, which is not generally considered in typical
In general, constitutive laws are used to define the stress–strain structural design practice, involves the thermal conductivity,
characteristics of a material. The accuracy of the analysis is specific heat and high thermal expansion of the concrete. This will
dependent on the constitutive laws used to define the mechanical cause the surrounding structure to respond against these effects
behaviour. In materials such as concrete, structural steel and and generate compressive forces in the heated concrete slab.
664 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

Fig. 2. Stress–strain relationship for concrete at ambient temperature, Carreira and

Chu [5]. Fig. 4. Specific heat of concrete, EC2 British Standards Institution [7].

Fig. 3. Thermal conductivity of concrete, EC2 British Standards Institution [7]. Fig. 5. Concrete thermal expansion, EC2 British Standards Institution [7].

2.2.2.2. Thermal conductivity. Thermal conductivity is the capabil-

ity of a material to conduct heat, and is defined as the ratio of heat
flux to the temperature gradient. It represents the uniform flow
of heat through concrete of unit thickness over a unit area sub-
jected to a unit temperature difference between the two opposite
faces [6]. The thermal conductivity of siliceous aggregate concrete
as represented in Eurocode 2, British Standards Institution [7] in
section 3.3.3 is shown in Fig. 3.

2.2.2.3. Specific heat. The specific heat of a material, as defined

by Harmathy [8], is the amount of heat per unit mass which is
required to change the temperature of the material by a degree.
It is represented by Eq. (2).

δH
 
Fig. 6. Compressive stress–strain relationship at elevated temperature for
Cp = (2)
δT p
concrete, EC2 British Standards Institution [7].

where H is enthalpy, T is temperature, and P is pressure. according to temperature. The stress–strain relationship of con-
The specific heat of concrete with siliceous aggregates as a crete with siliceous aggregates expressed as a function of tem-
function of temperature according to Eurocode 2, British Standards perature according to Eurocode 2, British Standards Institution [7]
Institution [7] in section 3.3.2 is shown in Fig. 4. follows Eqs. (3)–(5), and the distributions given in Figs. 6 and 7 rep-
resent the compressive and tensile stress–strain behaviour of the
2.2.2.4. Thermal expansion. Due to its isotropic nature, concrete
concrete, respectively.
exhibits thermal expansion when it is subjected to a temperature
change. According to Bazant and Kaplan [6], cracking occurs Compressive stress–strain relationship:
when stresses develop in concrete structures due to non-uniform
3εfc,θ
thermal expansion. The thermal expansion of concrete with σc (θ) =  3  for ε ≤ εc1,θ (3)
2+ εε

siliceous aggregates expressed as a function of temperature εc1,θ c1,θ
according to Eurocode 2, British Standards Institution [7] in section
3.3.1 is shown in Fig. 5. σc = 0, ε = εcu1,θ (linear behaviour is adopted)
for εc1(θ) < ε ≤ εcu1,θ (4)
2.2.2.5. Stress–strain relationship of concrete at elevated tempera-
tures. The most substantial consequence of fire on a concrete slab where fc,θ = Ultimate stress of concrete
is the stiffness and strength degradation which may lead to even- εc1,θ = Strain at ultimate stress of the concrete
tual collapse. It is important to study the concrete property changes εcu1,θ = Ultimate strain of the concrete.
 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 665

Table 1
Stress–strain value for structural steel beam, shear connectors, profiled steel
sheeting and steel reinforcing
Element σus εps εus
Steel beam 1.28σys 10εys 30εys
Steel reinforcing 1.28σys 9εys 40εys
Profiled sheeting – 20εys –
Shear connectors – 25εys –

Fig. 7. Tensile stress–strain relationship at elevated temperature for concrete, EC2

British Standards Institution [7].

Fig. 9. Stress–strain relationship for structural steel at ambient temperature, Loh

et al. [9].

Fig. 8. Elastic modulus of concrete with temperature, EC2 British Standards

Institution [7].

Tensile stress–strain relationship:

σt (θ) = kc,t (θ)σt (5)
where σt = tensile strength of concrete
kc,t = reduction coefficient for tensile strength of concrete
kc,t (θ) = 1.0 for 20 ◦ C ≤ θ ≤ 100 ◦ C
kc.t (θ) = 1.0 − 1.0(θ − 100)/500 for 100 ◦ C ≤ θ ≤ 600 ◦ C.
Fig. 10. Thermal conductivity for structural steel, EC3 British Standards
Fig. 6 illustrates that the compressive strength of the concrete Institution [10].
decreases when temperature increases but the ultimate strain
of the concrete increases with temperature. The tensile strength According to Loh et al. [9], the stress–strain relationship
of the concrete also decreases with an increase in temperature, for structural steel is represented as a simple elastic–plastic
as depicted in Fig. 7. A tensile stress can also be obtained for model with strain hardening. The mechanical behaviour for both
temperatures up to 500 ◦ C. compression and tension is assumed to be similar. Fig. 9 represents
The modulus of elasticity of the concrete in Fig. 8 decreases the stress–strain relationship for steel and Table 1 indicates the
with an increment in temperature. The reduction of the modulus different values of stress–strain for each material.
of elasticity is due to the rupture of bonds in the microstructure of
the cement paste when the temperature increases and is the result
2.3.2. Thermal properties of structural steel, reinforcing steel, shear
of the onset of rapid short-term creep.
connectors and profiled steel sheeting

2.3. Structural steel, reinforcing steel, shear connectors and profiled 2.3.2.1. General. For thin-walled steel sections, the thickness is
steel sheeting properties such that their temperature across the section is considered
uniform. Similarly for concrete, the effects of thermal conductivity,
2.3.1. Structural steel, reinforcing steel, shear connectors and profiled specific heat and high thermal expansion of the structural steel,
steel sheeting at ambient temperatures steel reinforcing, profiled steel sheeting and shear connectors are
The stress–strain characteristics of reinforcing steel, shear considered when the temperature changes.
connectors and profiled steel sheeting are essentially similar to
structural steel. Their behaviour is initially elastic after which 2.3.2.2. Thermal conductivity. The thermal conductivity of steel
yielding and strain hardening develops. A piecewise linear depends mainly on the amount of alloying elements and on the
approach was found to be sufficiently accurate to represent the heat treatment. The thermal conductivity of steel, λ according to
stress–strain relationship. Moreover, these curves are utilised in Eurocode 3, British Standards Institution [10] in section 3.4.1.3 is
the model when the stress–strain data is not available. presented in Fig. 10.
666 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

Fig. 11. Specific heat of structural steel, EC3 British Standards Institution [10].

Fig. 13. Stress–strain relationship at elevated temperature for structural steel, EC3
British Standards Institution [10].

Fig. 12. Thermal expansion of structural steel, EC3 British Standards Institu-
tion [10]. Fig. 14. Modulus of elasticity of structural steel at elevated temperatures EC3
British Standards Institution [10].

2.3.2.3. Specific heat. The specific heat of the steel Ca is expressed

in Eurocode 3, British Standards Institution [10] in section 3.4.1.2
σ(θ) = 0 for ε = εu,θ (10)
and is shown in Fig. 11. where εp.θ = fp,θ /Ea,θ

2.3.2.4. Thermal expansion. Harmathy [11] stated that the thermal

εy,θ = 0.02
expansion of steels depends mainly on the heat treatment used. εt,θ = 0.15
The coefficient of thermal expansion of steel at room temperatures εu,θ = 0.20
is expected to be 11.4 × 10−6 m−1 C−1 . Furthermore, the
a2 = (εy,θ − εp,θ )(εy,θ − εp,θ + c/Ea,θ )
thermal elongation of structural and reinforcing steel according
to Eurocode 3, British Standards Institution [10] is evaluated in b2 = c(εy,θ − εp,θ )Ea,θ + c2
section 3.4.1.1 and is illustrated in Fig. 12. (fy,θ − fp,θ )2
c=
(εy,θ − εp.θ )Ea,θ − 2(fy,θ − fp,θ )
2.3.2.5. Stress–strain relationship of structural steel, reinforcing
steel, shear connectors and profiled steel sheeting at elevated tem- fy,θ = effective yield strength
peratures. Most normal constructional steels have well-defined fp,θ = proportional limit
yield strengths at normal temperatures. Harmathy [8] and Ea,θ = slope of the linear elastic range
Buchanan [12] stated that the yield plateau becomes less no-
ticeable with a temperature rise. Upon further temperature in- εp,θ = strain at the proportional limit
crease, the ultimate strength of the steel declines steadily. The εy,θ = yield strain
stress–strain relationships may be applied to steel in both tension εt,θ = limiting strain for yield strength
and compression. The effects of high temperature on creep have
εu,θ = ultimate strain.
also been taken into account.
The stress–strain relationship of structural steel as a function The ultimate strength of the structural steel decreases when the
of temperature according to Eurocode 3, British Standards temperature increases, as illustrated in Fig. 13. Furthermore, the
Institution [10] follows Eqs. (6)–(10) and the stress–strain modulus of elasticity decreases with an increase in temperature.
relationships are illustrated in Fig. 13. The relationship of the modulus of elasticity of the structural steel
according to temperatures is illustrated in Fig. 14.
σ(θ) = εEa,θ for ε ≤ εp,θ (6)
2 0.5
σ(θ) = fp,θ − c + (b/a)[a − (εy,θ − ε) ]
2
3. Push test finite element analysis at elevated temperatures
for εp,θ ≤ ε ≤ εy,θ (7)
3.1. General
σ(θ) = fy,θ for εy,θ ≤ ε ≤ εt,θ (8)
σ(θ) = fy,θ [1 − (ε − εt,θ )/(εu,θ − εt,θ )] According to Cooke et al. [13], the fire resistance of composite
for εt,θ < ε < εu,θ (9) steel–concrete beams is obtainable in terms of their load carrying
 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 667

Fig. 16. Finite element mesh and boundary condition of the solid slab model.

Fig. 15. Position of shear connectors in push test specimens.

capacity, reliability and thermal insulation. Therefore the main

components affecting these are the concrete slab, structural steel,
steel reinforcing, profiled steel sheeting and shear connectors.
In this study, the finite element package ABAQUS was used
to investigate the behaviour of shear connectors in composite
steel–concrete beams with both solid and profiled steel sheeting
slabs under elevated temperatures.

3.2. Experimental investigation for push tests

Fig. 17. Finite element mesh and boundary condition of the profiled steel sheeting
There are three different experimental series results which slab model.
were compared with the finite element analysis mentioned. The
push tests performed by Lam and El-Lobody [14] and El-Lobody A four-node doubly curved thin shell element (S4R) was used
and Young [15] considered shear connection of a solid and for the profiled steel sheeting because it is the most appropriate
profiled slab, respectively, at ambient temperatures and these type of element to model thin walled steel structures. The S4R
were compared with the finite element analysis method results. element has six degrees of freedom per node and proved to
In this paper, the authors only considered the position of the give accurate solutions and also permits quadratic deformation
shear connectors in the middle of the profiled steel sheeting. over four nodal coordinates, membrane action and plain strain
Shear connectors are usually placed either in the strong side or behaviour.
weak side of the profiled slab. Fig. 15 shows the position of the Finally a two-node linear three-dimensional truss element
shear connectors in the push test. In order to avoid the adverse (T3D2) for steel reinforcing was adopted where the axial direction
effect of rib punch-through failure and other failure modes such is released using an option of an equation provided in ABAQUS.
as longitudinal splitting and concrete pull-out, shear connectors Furthermore, the authors were not too concerned about the slip
are preferred to be placed on the strong side. The strong side between reinforcing bar and concrete in this analysis.
position for shear connectors can lead to significant improvements Figs. 16 and 17 illustrate the finite element mesh used to
in strength and ductility of the connectors. Mirza and Uy [16] represent half of a stud of the push test. The aim of this was to
showed that the strength of the shear connectors increased by reduce the simulation cost and computational time. The generated
about 11% when they were placed on the strong side. Conversely, mesh was designed to give optimal accuracy where the fine mesh
when the shear connectors were positioned on the weak side, the surrounds the shear connectors and a coarse mesh was used
strength of the shear connectors showed a reduction of about 13%. elsewhere. A sensitivity analysis was conducted by the authors to
Other authors used similar push tests but at elevated temperatures ensure the accuracy of the developed model and its convergence
and these were analysed against ambient temperatures. The with mesh refinement in Mirza and Uy [16].
experimental investigation undertaken by Zhao and Aribert [17] at
elevated temperatures with different loading conditions was also 3.4. Boundary conditions and load application
compared with the finite element analysis.
In Figs. 16 and 17, the nodes that lie on the other symmetrical
3.3. Finite element type and mesh surface (Surface 1) for concrete, shear connectors, structural steel
beam, steel reinforcing and profiled steel sheeting are restricted
Three-dimensional solid elements were used to model the push from moving in the x-direction. All the nodes in the middle of the
off test specimens in order to achieve an accurate result from structural steel beam web, which are designated as Surface 2, are
the finite element analysis. For both the concrete slab and the restricted to move in the y-direction. All the nodes of the concrete
structural steel beam, a three-dimensional eight-node element and the profiled steel sheeting, which are designated Surface 3, are
(C3D8R) was used because it is a solid element and it was also restricted to move in the z-direction.
used to improve the rate of convergence. A three-dimensional For the application of load, a static concentrated load was
thirty-node quadratic brick element (C3D20R) for shear connectors applied to the centre of the web employing the modified RIKS
was used because it is a second-order element consisting of 20 method which can be obtained through a series of iterations for
nodes and these provide higher accuracy when compared with each increment for a non-linear structure. The RIKS method was
the first-order elements. They capture stress concentrations more used for nonlinear analysis to ensure that any unloading was
effectively and are better for modelling geometric features. captured. Furthermore, this method was used to predict unstable
668 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

and nonlinear collapse of a structure. The load magnitude was

used as an additional unknown and was solved simultaneously
for the loads and displacements. To achieve accurate results in
ABAQUS, the RIKS method has the ability to use the arc length
along the static equilibrium in load–displacement space. The
initial increments will be adjusted if the finite element model
fails to converge. Finally, the value of load after each increment
is computed automatically. The final result will either be the
maximum value of the load or the maximum displacement value.

3.5. Analysis procedure

3.5.1. General
Fig. 18. ISO834 Fire standard curve.
The response of a composite steel–concrete structural member
exposed to fire is governed by the rate of heating. This is
because the mechanical properties of materials decrease as the
temperature rises and, likewise, the structural resistance of a
member decreases with a temperature rise.
In this study, two components were evaluated. They comprise
structural and temperature analyses. The structural analysis is
where the structure is simulated at ambient temperature, and the
result was compared with experimental studies. The temperature
analysis simulated the behaviour of the structure as a function of
time using measured temperature distributions of the structural
elements from independent tests.

3.5.2. Structural analysis

The structural analysis is performed for structures in the
ambient condition. This push test specimen is similar to the
standard push test according to Eurocode 4, British Standards
Institution [18] but only one shear connector is connected to
each flange. This is due to the assumption that the load from the
structural steel beam is transferred equally to the shear connectors.
In order to reduce the simulation time and utilise structural
symmetry, only a quarter of the push test specimen was modelled. Fig. 19. Temperature distribution diagram for solid slab.
The material properties are defined in Section 2.2.1 for concrete,
and Section 2.3.1 for the structural steel, reinforcing steel, headed
shear stud connectors and profiled steel sheeting. In more complex analyses, different heating models are
considered to represent the temperature development in different
3.5.3. Temperature analysis zones of the fire compartment. This is the case, for instance, when
the fire is transferred from the structural steel beam to the concrete
The temperature analysis is performed independently of
the structural analysis. To perform the temperature analysis, slab, shear connector and profiled steel sheeting. Steel elements
the geometry of the cross-section is similar to the structural have an adverse behaviour in this respect due to the very high
analysis specimen. Conversely, its material properties are defined thermal conductivity of the steel. Rapid heating of the whole
in Sections 2.2.2 and 2.3.2 for concrete and structural steel, profile takes place as a result. In comparison, composite elements
respectively. The materials in the section can vary from element exhibit a favourable behaviour due to the great thermal inertia of
to element, and their properties are temperature dependent. the elements and the low thermal conductivity of the concrete.
Zhao and Kruppa [19] stated that the mechanical behaviour is An accurate computer simulation of heat flow through a
much more complicated when the temperature changes because composite deck slab has been completed by Cooke et al. [13]
there are two materials involved, which are mainly concrete and using a program called FIRES-T to determine the distribution
steel. At elevated temperatures, the different heating conditions of temperatures throughout the slab according to time. Lamont
directly influence their own mechanical behaviour which will et al. [21] also studied composite steel–concrete structure
consequently modify their initial interaction effect. The push test temperature distributions using HADAPT, and compared these
specimen model is similar to the structural analysis model in order with the Cooke et al. [13] approach. Their results were in good
to compare the results.
agreement, therefore, the resulting temperature distributions have
Fire is usually represented by a temperature–time curve which
been adopted here by the authors for the application of the finite
gives the average temperature reached during a fire in a small
element ABAQUS analysis. The idealised temperature distribution
sized compartment or in the furnaces used for fire resistance tests.
diagrams are presented in Figs. 19 and 20 for solid and profiled
International standards are based on the standard fire defined by
the heat exposure given by ISO834 [20], and is represented by the slabs, respectively. Figs. 19 and 20 show that the concrete,
curve shown in Fig. 18. structural steel beam, profiled steel sheeting and shear connectors
are divided into layers which is necessary to differentiate the
T = 345 log10 (8t + 1) + 20 (11) temperature distributions according to time. Table 2 provides the
◦
where T is the average furnace temperature, in C; and t is the time, temperature distribution of the layers with respect to time. The
in minutes. layers refer to Figs. 19 and 20, respectively.
 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 669

Table 2
Temperature changes according to time
Time (t) (min) Layer (T ) (◦ C)
A B C D E F G H J K L

10 775 675 575 575 475 375 275 175 75 38 19

20 939 839 739 739 639 539 439 339 239 139 39
30 1042 942 842 842 742 642 542 442 342 242 142
60 1145 1045 945 945 845 745 645 545 445 345 245
90 1200 1105 1005 1005 905 805 705 605 505 405 305
120 1200 1149 1049 1049 949 849 749 649 549 449 349
180 1200 1200 1100 1100 1000 900 800 700 600 500 400

Table 3
Dimensions and concrete strength for finite element models

Series Specimen Dimension Exposure to fire (min) Concrete strength (N/mm2 ) Comments
Sheeting B (mm) h D (mm)
type (mm)

1 A1 Solid 600 150 600 Ambient 34.5 Load until structure failure
* A2a-A2g Solid 600 150 600 10–180 34.5 Load at 17.5 kN per stud
* A3a-A3g Solid 600 150 600 10–180 34.5 Load at 25 kN per stud
* A4a-A4g Solid 600 150 600 10–180 34.5 Load at 50 kN per stud
2 LTA1 Solid 619 150 619 Ambient 35.0 Load until structure failure
* LTA2a-LTA2g Solid 619 150 619 10–180 35.0 Load until structure failure
LPTA1 W-Dek 900 115 600 Ambient 35.5 Load until structure failure
* LPTA2a-LPTA2g W-Dek 900 115 600 10–180 35.5 Load until structure failure
3* LTA3a-LTA3g Solid 619 150 619 10–180 35.0 20% of ultimate load
* LPTA3a-LPTA3g W-Dek 900 115 600 10–180 35.5 20% of ultimate load
* LTA4a-LTA4g Solid 619 150 619 10–180 35.0 40% of ultimate load
* LPTA4a-LPTA4g W-Dek 900 115 600 10–180 35.5 40% of ultimate load
* LTA5a-LTA5g Solid 619 150 619 10–180 35.0 60% of ultimate load
* LPTA5a-LPTA5g W-Dek 900 115 600 10–180 35.5 60% of ultimate load
* a denotes 10 min fire exposure, b denotes 20 min fire exposure, c denotes 30 min fire exposure, d denotes 60 min fire exposure, e denotes 90 min fire exposure, f denotes
120 min fire exposure and g denotes 180 min fire exposure.

temperatures with different load levels. The load levels considered

were 20, 40 and 60% of the ultimate load. In this paper, two
different kinds of slab were studied, namely, a push test for a
solid slab and a push test for a profiled trapezoidal type slab. The
dimensions and concrete strength of the finite element models are
shown in Table 3.

4.2. Solid slab for Zhao and Aribert [17]

In order to verify that the finite element analysis was

accurately undertaken, the experimental investigation of Zhao
and Aribert [17] was used for calibration purposes. Zhao and
Aribert [17] completed push test experiments at ambient and
at elevated temperatures. The specimens had a mean concrete
compressive strength of 34.5 N/mm2 . The push test results for
Fig. 20. Temperature distribution diagram for profiled slab.
ambient temperatures of the finite element analysis are provided
in Fig. 21. Experimental results showed that the shear connectors
had a shear force carrying capacity of ultimate load of 125 kN and
4. Results and discussion
the finite element model produced an ultimate load of 131 kN
which resulted in a 5% discrepancy. The push tests at elevated
4.1. General temperatures with different loading and comparison with the
finite element models are shown in Fig. 22. The results are in good
For the purposes of this paper, finite element push tests agreement with the experimental investigation. The temperature
were modelled to determine the load–slip behaviour of the shear and slip are measured at the bottom shear stud where it is welded
connectors. The first analyses considered were compared with the to the steel beam. According to Fig. 23, stress distribution contours
experimental investigations undertaken by Zhao and Aribert [17] proved that the higher the load applied, the lower the temperature
for solid slabs. For the parametric studies, both solid and profiled the specimens could withstand.
slabs were considered to study the effects of temperature on
the behaviour of the shear connectors. The second series of 4.3. Parametric studies
finite element models considered ambient temperatures in order
to compare the experimental tests. The finite element models 4.3.1. Solid slab
were modified to study the temperature changes with time. The For solid slabs, the finite element model is compared with the
temperature changes with time are illustrated in Table 2. The third experimental study undertaken by Lam and El-Lobody [22]. The
series of analyses involved finite element models comprised of specimens had a concrete compressive strength of 35 N/mm2 .
670 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

Fig. 21. Shear capacity of shear connectors.

Fig. 24. Comparison between experimental push test and finite element model for
solid slab.

load was 102 kN with 6 mm slip and 113 kN with 9 mm slip

for the experimental result and finite element model results,
respectively. The failure mode of the shear connectors herein is
similar to that described by Yam [23] and Lam and El-Lobody [22].
Fig. 25 illustrates that the shear connectors experienced significant
deformation around their base. In the experiments, yielding of
the stud element was discerned near the shear connector collar
followed by maximum compressive stress being reached by the
concrete elements around the shear connector.
Another finite element model analysis carried out was to study
the effect of temperature changes on the shear connector. This
series of analyses was able to obtain the load versus slip curve
Fig. 22. Slip according to temperature. as shown in Fig. 26 where the strength of the shear connector
decreased with the increment of the temperature.
The second series of finite element analyses are highlighted in Fig. 26 illustrates that the strength of the shear connector
Fig. 24. It was demonstrated that the experimental result and the decreases when the temperature increases. Moreover, the ultimate
finite element model results are in good agreement. The maximum load is reduced when the temperature increases. Initially, for the

Fig. 23. Stress distribution for different loading according to time.

 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 671

Fig. 25. Stress contours and deformed shape for solid slab.

first 10 min of the fire, the ultimate load reduces by 35% compared
with the ultimate load at ambient temperatures. After 180 min of Fig. 26. Comparison of push tests with temperature change versus time.

fire, the ultimate load decreased to a maximum of 57%. The failure

is governed by the ultimate stress of the concrete. The ultimate
stresses of concrete decreased with respect to temperature. When
the structures reach the ultimate stress, the concrete started to
crack which allowed failure to occur. Based on the results of the
push test finite element analyses, a strength reduction with respect
to time has been carried out.
The ultimate load ratio [Pu (θ)/Pu (20 ◦ C)] which relates to
different time periods of the push tests can be determined from
Eqs. (12) to (15) and is illustrated in Fig. 27.
Pu (θ)
= −3.45 × 10−2 t + 1
Pu (20 ◦ C)
for 0 min ≤ t ≤ 10 min (12)
Pu (θ) Fig. 27. Shear resistance of shear connectors at elevated temperature according to
= −9.4 × 10−3 t + 0.7483
Pu (20 ◦ C) time.

for 10 min ≤ t ≤ 20 min (13) The maximum load was 83 kN with 1.9 mm slip and 84 kN
Pu (θ) with 1.3 mm slip for the experimental result and finite element
= −1.10 × 10 t + 0.5837
−3
Pu (20 ◦ C) model results, respectively. The profiled slab revealed that initial
cracking occurred in the middle of the slab along the trough of
for 20 min ≤ t ≤ 120 min (14)
the profiled slab, which is caused by concrete failure. The concrete
Pu (θ)
= −6 × 10 t + 0.4527
−5 element in the trough of the profiled slab reached a maximum
Pu (20 ◦ C) stress before the shear connector element. Both the experimental
for 120 min ≤ t ≤ 180 min (15) tests and finite element model demonstrated that the failure mode
was dominated by the concrete deforming where the concrete
where Pu (θ) is the ultimate load in degrees Celsius, Pu (20 C) is the ◦
crushed and cracked before the shear connectors fractured near
ultimate load at 20 ◦ C, and t is the time in minutes. the weld collar. The failure mode observed is similar to that
Fig. 27 illustrates that the ultimate load decreases significantly mentioned by El-Lobody and Young [15]. Fig. 31 shows that the
from 0 up to 20 min, and after that the reduction is not as large. profiled slab experienced less deformation and lower stresses
Therefore, in reality, the first 20 min of the fire is crucial because in the concrete and shear connector which is attributed to the
the structure loses about 35% of its strength. Different stress strength contribution of the profiled steel sheeting.
contours and deformed shapes of the concrete slab, structural steel Another analysis was undertaken to study the effect of
beam and shear connectors of different fire ratings can be observed temperature changes on the shear connector. This series of
in Fig. 28. analyses were undertaken to attain the load versus slip curves as
In order to study the behaviour and strength of the shear shown in Fig. 32, and the result is similar to that mentioned in
connectors when exposed to temperature changes, several load Section 4.3.1, i.e. the strength of the shear connector degrades with
levels were applied to the push test until it failed. Fig. 29 shows an increment in temperature.
that, during fire exposure, the maximum load that the push test Even though the ultimate load is reduced when the temperature
can withstand is up to 40% of the ultimate load for solid slabs up to is increased, when compared with the solid slab, the reduction of
180 min. If the load is 60% of the ultimate load, the structure can ultimate load is much lower. Initially, for the first 10 min of fire, the
resist only 10 min of fire exposure before the structure fails. ultimate load decreased by 25% compared with the ultimate load
at ambient temperature. After 180 min of fire, the ultimate load
4.3.2. Profiled slab decreased by a maximum of 31%.
For the profiled steel sheeting slab, the finite element model The ultimate load ratio [Pu (θ)/Pu (20 ◦ C)] related to different
is compared with the experimental study undertaken by El- time periods of the push tests is determined from Eqs. (16) to (18)
Lobody and Young [15]. The specimen had a concrete strength of and is plotted in Fig. 33.
35.5 N/mm2 . The second series of finite element model analyses Pu (θ)
are shown in Fig. 30. = −2.51 × 10−2 t + 1
Pu (20 ◦ C)
The results shown in Fig. 30 confirm that the experimental
result and the finite element model results are in close agreement. for 0 min ≤ t ≤ 10 min (16)
672 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

Fig. 29. Slip at different load levels according to time.

Fig. 30. Comparison between experimental push test and finite element model for
profiled sheeting slab.

Fig. 31. Stress contours and deformation shape for profiled steel sheeting slab.

Fig. 28. Stress contours and deformed shapes at various times.

Pu (θ)
= −6 × 10−4 t + 0.7549
Pu (20 ◦ C)
for 10 min ≤ t ≤ 60 min (17)
Pu (θ)
= −2 × 10−4 t + 0.7279
Pu (20 ◦ C)
Fig. 32. Comparison of push test with temperature changes according to time.
for 60 min ≤ t ≤ 180 min (18)
where Pu (θ) is the ultimate load at temperature θ degrees Celsius, The ultimate load decreases significantly up to 20 min, and
Pu (20 ◦ C) is the ultimate load at 20 ◦ C (ambient temperature), and after that, the incremental reduction value is not as significant.
t is the time in minutes. Different stress contours and deformed shapes of the concrete slab,
 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674 673

Fig. 33. Shear resistance of shear connectors at elevated temperature according to

time.

structural steel beam, profiled steel sheeting and shear connector

of varying fire ratings can be observed in Fig. 34.
The third series of finite element analyses are similar to the
series described in Section 4.3.1. Fig. 35 showed that, during fire
exposure, the maximum load that the push test can withstand is
up to 60% of the ultimate load for profiled slabs up to 180 min.
This result illustrates the potential role of the steel sheeting as
a protective layer to the concrete slab. If the load is 80% of the
ultimate load, the structure fails when exposed to fire.
Figs. 27 and 33 reveal that the ratio of the shear resistance at
180 min of fire exposure compared with ambient temperature is
0.45 and 0.7 for the solid and profiled slabs, respectively. Fig. 36
shows the cross-sectional area for both the solid and profiled slab.
From Fig. 36, the ratio of proportional area in a high temperature
region, κ is determined using Eq. (19):
A1s A 1p
κs = or κp = (19)
A1s + A2s A1p + A2p

where κs and κp are the ratio of proportional area of concrete at

high temperature; A1s and A1p are the cross-sectional area at the
bottom of the concrete; A2s and A2p are the cross-sectional area at
the top of the concrete.
Fig. 36 shows that the proportional area in a high temperature
region for a solid slab and a profiled slab is 0.45 and 0.27,
respectively. It appears that the failure mode of shear connectors
was greatly influenced by the proportional area of concrete in
a high temperature region, which was higher for the solid slab.
Subsequently on this basis, profiled slabs perform better at high
temperatures in comparison to ambient temperature.

4.4. Different load distributions

For design purposes, Australian Standard AS/NZ 1170.0 [24] is

used in order to determine the basic combinations for the strength
limit state for the strength of the structures. The loads when it is in
the ambient condition and when exposed to fire are illustrated in
Eqs. (20) and (21), respectively. The load ratio factor ξ is the ratio
Fig. 34. Stress contours and deformed shapes at various times.
between distributed loads when exposed to fire over distributed
load under normal conditions as shown in Eq. (22).
0.55 shows that the solid slab can resist the fire for 30 min before
En = 1.2G + 1.5Q (20) failure occurs. For the profiled slab, the structures can withstand
Ef = 1.1G + 0.4Q (21) the fire for more than 180 min as shown in Fig. 33. This is because
the profiled steel sheeting acts as a protective layer for composite
ξ = Ef /En (22) steel–concrete structures during a fire. Moreover, according to
where G is the dead load of the structures and Q is the live load of Yu et al. [25], due to the inherent efficiency of the profiled slab,
the structures. the displacement of the concrete is lower when compared to a
When G = Q , the distributed load for the ambient condition is conventional solid slab. Therefore Figs. 27 and 33 are very useful for
En = 2.7G, the distributed load when exposed to fire is Ef = 1.5G designers to estimate the appropriate fire exposure for both solid
and the load ratio factor is ξ = 0.55. From Fig. 27, a load ratio of and profiled slabs in different combination factors.
674 O. Mirza, B. Uy / Journal of Constructional Steel Research 65 (2009) 662–674

basis for temperature analysis. This may not be the case for real
buildings. However, the trends in this paper may be of use for
further studies.
When designers use the ultimate limit state to check the
strength, different load combination factors are used. Figs. 27 and
33 are very useful to estimate the fire exposure time before the
structures fail.
Further experimental research is considered necessary in order
to validate the results above to enhance the understanding of fire
resistance of composite steel–concrete structures.

Acknowledgements

The authors would like to thank the Australian Research

Fig. 35. Slip at different load levels according to time.
Council Linkage Grants Scheme and BlueScope Lysaght, Sydney for
providing funding for this project and the University of Western
Sydney for providing a conducive environment for the authors to
prepare this paper.

References

[1] Inberg SH. Test of severity of building fire. Quarterly National Fire Association;
1928.
[2] Pawtucket RI. ABAQUS, Theory manual version 6.5. Hibbitt, Karlsson and
Sonrensen Inc.; 2006.
[3] Pawtucket RI. ABAQUS User’s manual version 6.5. Hibbitt, Karlsson and
Sonrensen Inc.; 2006.
Fig. 36. Sectional area for solid slab and profiled slab. [4] Pawtucket RI. ABAQUS Analysis user’s manual version 6.5. Hibbitt, Karlsson
and Sonrensen Inc.; 2006.
[5] Carreira D, Chu K. Stress–strain relationship for plain concrete in compression.
5. Conclusions Journal of ACI Structural 1985;82(11):797–804.
[6] Bazant PZ, Kaplan MF. Concrete at high temperature: Material properties
This paper has considered four key issues for push tests and mathematical models. In: Concrete design and construction series. UK:
Longman Group Limited; 1996.
of composite steel–concrete beams. An accurate finite element [7] British Standards Institution. Design of concrete structures, Part 1.1 General
model has been developed to investigate the behaviour of shear rules and rules for buildings. London: British Standard Institute; 2004. ENV
connection in composite steel–concrete beams for both solid and 1992-1-1.
[8] Harmathy TZ. Thermal properties of concrete at elevated temperatures.
profiled slabs at elevated temperatures. Based on the comparisons Journal of Materials 1970;5(1):47–74.
between the results obtained from the finite element models and [9] Loh HY, Uy B, Bradford MA. The effects of partial shear connection in the
available experimental results, it was observed that they are in hogging moment region of composite beams Part II — analytical study. Journal
of Constructional Steel Research 2003;60:921–62.
good agreement. All the failure modes were accurately predicted [10] British Standards Institution. Design of the steel structures, Part 1.1 General
by the finite element model and a maximum discrepancy of 10% rules and rules for buildings. London: British Standard Institute; 2005. ENV
was observed when comparing the finite element model with 1993-1-1.
[11] Harmathy TZ. Fire safety design and concrete, concrete design and construc-
experimental studies.
tion series. UK: Longman Group Limited; 1993.
Three differences were observed between the profiled and [12] Buchanan A. Structural design for fire safety. UK: John Wiley & Sons Ltd.; 2001.
solid slabs. One of them is the failure mode. For the solid slabs [13] Cooke GME, Lawson RM, Newman GM. Fire resistance of composite deck slabs.
The Structural Engineer 1988;66(16):253–67. 267.
considered, the failure mode was dominated by shear connection [14] Lam D, El-Lobody E. Behaviour of headed stud shear connectors in composite
fracture whereas for the profiled slab, failure was dominated by beam. Journal of Structural Engineering-ASCE 2005;131(1):96–107.
concrete crushing and cracking, as discussed in Section 4. Stresses [15] El-Lobody E, Young B. Performance of shear connection in composite beams
with profiled steel sheeting. Journal of Constructional Steel Research 2006;
in the shear connector and concrete are lower compared with the
62(7):682–94.
solid slab due to the addition of the steel profiled sheeting. It can [16] Mirza O, Uy B. Effect of steel fibre reinforcement on the shear connection
also be observed that the solid slab generally has a higher ultimate of composite steel–concrete beams. Journal of Advanced Steel Construction.
2007 [in press].
load compared with that of the profiled steel sheeting.
[17] Zhao B, Aribert JM. Essais et caracterisation du comportement de goujons
Composite steel–concrete beams subjected to fire are also en cisaillement a des temperatures elevees. Construction Metallique 1992;3:
discussed in this paper. Numerical analyses of fire resistance 27–39.
of shear connectors were also investigated using the finite [18] British Standards Institution. Design of composite steel and concrete
structures, Part 1.1 General rules and rules for buildings. London: British
element analysis package, ABAQUS. These studies have allowed Standard Institute; 2004. ENV 1994-1-1.
the investigation of the shear resistance of the connectors as a [19] Zhao B, Kruppa J. Experimental and numerical investigation of fire behaviour
function of time at elevated temperatures. The finite element of steel and concrete composite beams. In: Proceedings of the engineering
foundation conference 1997. New York (NY, USA): ASCE; 1996. p. 129–42.
analysis allowed the elucidation of the failure of shear connectors [20] ISO834. Fire resistance tests. Element of building construction. 1999. 1999-09-
at elevated temperatures. 15.
As established the ultimate load of a solid slab is higher than [21] Lamont S, Usmani AS, Gillie M. Behaviour of a small composite steel frame
structure in a “long-cool” and a “short-hot” fire. Fire Safety Journal 2004;39(5):
that of a profiled slab, but Figs. 29 and 35 have shown that 327–57.
the profiled steel sheeting acted as a protective layer for the [22] Lam D, El-Lobody E. Behavior of headed stud shear connectors in composite
concrete slab. The profiled steel sheet slabs can withstand 60% of beam. Journal of Structural Engineering-ASCE 2005;131(1):96–107.
[23] Yam LCP. Design of composite steel–concrete structures. London: Surrey
their ultimate load at elevated temperature compared to ambient University Press; 1981.
temperature. Solid slabs can only attain 40% of their ultimate load [24] AS/NZ 1170.0. Structural design action, Part 0: General principles. Australian/
at elevated temperature compared to ambient temperature. This New Zealand Standards; 2002.
[25] Yu XM, Huang ZH, Burgess IW, Plank RJ. Thermal and structural behaviour of
is very useful for structural efficiency purposes. Furthermore, this orthotropic slabs in fire. In: Fourth international workshop on structures in
study was based on results where the ISO 834 fire was used as a fire. 2006.