Ain Shams Engineering Journal (2016) 7, 73–88

Ain Shams University

Ain Shams Engineering Journal

www.elsevier.com/locate/asej
www.sciencedirect.com

CIVIL ENGINEERING

Finite element modeling of steel concrete

beam considering double composite action
Ashraf Mohamed Mahmoud *

Department of Civil Engineering, Faculty of Engineering, Modern University for Technology and Information, Cairo, Egypt1

Received 4 January 2015; revised 23 February 2015; accepted 10 March 2015

Available online 4 May 2015

KEYWORDS Abstract Steel concrete composite construction has gained wide acceptance as an alternative to
Finite element modeling; pure steel or concrete construction. Ansys 11 computer program has been used to develop a
Ansys 11; three-dimensional nonlinear finite element model in order to investigate the fracture behaviors of
Experimental; continuous double steel-concrete composite beams, with emphasis on the beam slab interface.
Double steel-concrete Three beam models with varying number of the head studs have been addressed. The associated
composite beams; constitutive results such as the ultimate loads, the maximum deflections, the interface slip and slip
Slip; strain values are presented. A parametric study has been carried out in order to investigate the effect
Head studs of some parameters on their fracture capabilities, such as steel beam height, lower slab thickness
and length, studs diameter and arrangement method. By comparing these results with the available
experimental data, the proposed model is found to be capable of analyzing steel-concrete composite
beams to an acceptable accuracy.
 2015 Faculty of Engineering, Ain Shams University. Production and hosting by Elsevier B.V. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction connected by metal devices known as shear connectors. One

type of these connectors is called head studs as shown in
The use of composite structures is increasingly present in civil Fig. 1. The main functions of these studs are to allow for the
construction works. Steel-concrete composite beams, particu- joint behavior of the beam-slab, to restrict longitudinal slip-
larly, are structures consisting of two materials, a steel section ping and uplifting at the elements interface and to take shear
located mainly in the tension region and a concrete section, forces. Double steel-concrete composite continuous beam is a
located in the compression cross-sectional area, both new structural system developed on the basis of single steel-
concrete one, in which there is also a bottom reinforced con-
* Address: 5-Elkaem Makam khedr khedr Street, from Gesr El-Suez crete slab connected to a steel profile in the negative moment
Street, Nozha, Flat No. (4), Floor number (2), Heliopolis, Cairo, regions through the head studs, therefore with two interfaces.
Egypt. Tel.: +20 20100 5466 324. Comparing with the traditional single steel-concrete composite
E-mail address: ashraf_amin78@yahoo.com continuous beam, its advantage is that effectively limits the
1
University Web-site: http://www.mti.edu.eg
crack width of the negative moment area, and also improves
Peer review under responsibility of Ain Shams University.
the stress state of section, so that it is suitable to the composite
continuous beam with a larger span. The mechanical proper-
ties of the double composite beam obviously depend on their
Production and hosting by Elsevier respective properties and interactions. In the negative applied

http://dx.doi.org/10.1016/j.asej.2015.03.012
2090-4479  2015 Faculty of Engineering, Ain Shams University. Production and hosting by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
74 A.M. Mahmoud

Nomenclature

Ec young modulus of elasticity of concrete N2 studs number in each line of lower slab
Et young modulus of elasticity of steel reinforcement Pult. ultimate load capacity
bars Dmax maximum deflection
fc average cubic compressive strength of concrete. Ustud stud diameter
Ft yield stress of steel reinforcement bars rxp principle stress in X - principle direction
hstud stud height ryp principle stress in Y - principle direction
N1 studs number in each line of upper slab rzp principle stress in Z - principle direction

girder bridges due to the synergetic combination of the con-

cast-in-place concrete slab crete and steel. The plastic design in the framework of the
Eurocode through an existing elastically designed bridge is also
head studs welded wire reinforcing mesh introduced. Xu et al. [2] discussed the improvement of the local
shear
connectors buckling strength of continuous double composite box girders
by adding a concrete slab to the steel bottom flange. The
mechanical properties in concrete crack, formation of sectional
plastic hinge are also investigated. Tan et al. [3] utilized exper-
imental tests to provide further information and conclusions
regarding composite steel-concrete beam specimens by examin-
ing the behavior of multi-span composite steel-concrete beams.
main steel beam
These beams are subjected to combined actions of torsion and
flexure for both full and partial shear connection and compar-
cold formed
steel deck ing the disparity in the varying degrees of shear connection.
secondary steel beam Lin and Yoda [4] studied the mechanical performance of the
horizontally curved continuous composite steel-concrete
(a) Illustrative sketch of roof slab with composite action.
beams subjected to combined hogging (negative) bending
and torsion, in order to investigate the effect of curvature on
both elastic and inelastic behaviors of these beams in the inte-
rior support regions. Henriques et al. [5] presented a general-
ized beam theory (GBT) formulations specially designed for
performing efficient linear analysis of steel-concrete composite
bridges and elastoplastic collapse analysis of thin-walled steel
members and extended for including the non-linear reinforced
concrete material behavior of steel-concrete composite beams.
Liang et al. [6] have undertaken nonlinear finite element anal-
ysis on continuous composite beams in combined bending and
shear. In their study, design formulas incorporating contribu-
tions from the concrete slab and composite action were pro-
posed for vertical shear strength and the ultimate shear
interaction of continuous composite beams. A finite element
(b) Composite beam system with head studs shear connectors. model is presented by Liang et al. [7] to investigate the flexural
and shear strengths of simply supported composite beams
Figure 1 Steel-concrete composite section with studs shear under combined bending and shear. In this research, the
connectors. numerical results are verified and compared with the available
experimental results. Sebastian and McConnel [8] described a
nonlinear finite element program for modeling composite
bending moment area, the concrete slab cracks under tension beams. Axial springs with empirical shear slip relations were
and then the interface slip occurs between steel profile and con- used to model discrete shear connectors. Hirst and Yeo [9]
crete slab, with non-linear features, it makes great impact on used a standard finite element program to analyze composite
the structure of the internal forces and deformation. beams with partial and full shear connection. Quadrilateral
Therefore, it is necessary to present a finite element model to elements were employed to simulate discrete and stud shear
study the mechanical properties of the double steel-concrete connectors. The material properties of stud elements were
composite beam in negative moment regions. modified to make them equivalent in strength and stiffness
Although many experimental and theoretical studies for the to the actual shear connectors in composite beams. Al-
traditional single steel-concrete composite beam have been Amery and Roberts [10] presented a nonlinear analysis of
done, few research studies have been found in references to composite beams with partial shear connection by using a
the double steel-concrete composite continuous beam. finite difference method. Salari et al. [11] formulated a compos-
Rozsas [1] investigated the plastic reserve of composite plate ite beam element based on the force analysis method for the
Finite element modeling of steel concrete beam 75

nonlinear analysis of composite beams with deformable shear Based on this investigation, a simplified analytical model
connectors. Thevendran et al. [12] utilized the finite element through Ansys 11 software is developed in order to enables
software ABAQUS to study the ultimate load behavior of the prediction of the fracture behavior. Its results are com-
composite beams curved in plan. Shell elements were used to pared with the previously available experimental investigated
model the concrete slab and the steel beam while a rigid beam models introduced by Duan et al. [24]. The results demonstrate
element was employed to simulate stud shear connectors. a better approximation for the failure criteria in both cases.
Reiner [13] and Stroh and Sen [14] presented a double steel-
concrete composite continuous beam as a new structural 3. Methodology and the analytical model
system developed on the basis of single steel-concrete compos-
ite beam, in which there is also a bottom reinforced concrete The objective of this section is to describe the finite element
slab connected to a steel profile in the negative moment regions model features common to double steel-concrete composite
through the shear connectors, therefore with two interfaces. beams being considered. The Ansys 11 finite element package
This research was accompanied by the determination of the was used to carry out the modeling. The applied load was iter-
crack width limits of the negative moment area, and the ated step by step using the Newton-Raphson method.
improvement of the stress state of section, and later applied Solid65 element was used to model the concrete. This ele-
for the composite continuous beam with a larger span. ment has eight nodes with three degrees of freedom at each
Newmark et al. [15] introduced the partial collaboration node translations in the nodal x, y, and z directions. The ele-
theory which is used later for deriving the elastic stiffness ment is capable of plastic deformation, cracking in three
matrix in the negative moment region for a double composite orthogonal directions, and crushing. A schematic of the ele-
beam element and for studying and verifying the double ment was shown in Fig. 2a. A Link8 element was used to
composite continuous beam models, and consequently the model steel reinforcement. This element is a 3D spar element
composite action effect as illustrated by Duan et al. [16–18]. and it has two nodes with three degrees of freedom translations
Nagai et al. [19] tested a double composite girder under pure in the nodal x, y, and z directions. This element is capable of
hogging moment and measured its ultimate bending moment plastic deformation and element was shown in Fig. 2b. The
strength. Duan et al. [20] and Yang and Duan [21] focused modeling of the head studs shear connectors was done by
on the problems of interface slip, deformation, ultimate bearing the BEAM 188 elements, which allow for the configuration
capacity, and the effective flange width of concrete slab for of the cross section, enable consideration of the nonlinearity
the double steel-concrete composite beams. Wang et al. [22] of the material and include bending stresses. This element
presented the elastic analysis of double composite beam defor- was indicated in Fig. 3a. SOLID185 is used for the modeling
mations using the Goodman elastic sandwich method. Yen of the steel beam. It is defined by eight nodes having three
et al. [23] discussed the ultimate load behavior and elastic degrees of freedom at each node, translations in the nodal x,
deformations of steel box girders containing composite bottom y, and z directions. The element has plasticity, hyperelasticity,
flanges. Duan et al. [24] performed beam collapse tests for stress stiffening, creep, large deflection, and large strain
three models of double steel-concrete composite continuous
beam. These tests aimed to report the load–deflection curve, the
ultimate flexural capacity, and the interface slip and slip strain
values between steel and concrete along the span direction.
The objective of the current paper was to demonstrate a 4
P
proposed analytical finite element model of continuous double 5
O
steel-concrete composite beams to estimate the fracture behav-
ior and interface slip values of tested specimens produced by 6
Duan et al. [24], through Ansys 11. The analytical model N Rebar
and the results of system level study can be of interest in assess- M
z
2 3
ing progressive collapse resistance of existing structures con-
tain double steel-concrete composite beams and in the design L y
of new structures. x K
Z

2. Research significance I
Y 1
X
J
The target of this research is to demonstrate a better analytical
understanding of double steel-concrete composite beams. Figure 2a Solid65 – 3D solids modeling.
Thereby, the focus should be set on the analysis of the maxi-
mum increase in strength and deflection capacity due to the
existing of double composite action. Therefore, the principal J
purpose is the nonlinear finite element analysis of continuous
steel-concrete composite beams containing double composite Z
x
action and head studs shear connectors. Within this frame-
work, several aspects should be investigated such as the Y
I
load–deflection response of the composite beam, and the grad- x
ual evolution of slip and slip-strain values at the beam-slap
interface up to failure considering double composite action. Figure 2b Link8 –3D spar modeling.
76 A.M. Mahmoud

Z the research concerns solely symmetrically continuous double

1 K
steel-concrete composite beams. The cross sections for all the
4
2 models, namely SCB1, SCB2, and SCB3, are constructed by
I
3 a top concrete slab along the whole beam length with tension
reinforcement 7U8/m’ in each direction, and by a 1000 mm
Y
Y 188 length bottom concrete slab over interior support, whereas
the upper and lower slab thickness was 80 mm.
x
x
Z J 5
4. Material properties of the proposed model
Figure 3a Beam188 – 3D quadratic beam modeling.
Table 2 summarizes the values of the material properties for all
composite beam model components, i.e., reinforced concrete
slab, steel beam, and head studs. For the steel beam and head
4 M O,P
P studs, the maximum tensile strength obtained from the exper-
5 O N
imental test as ft = 235 MPa and the young modulus of elas-
I K,L
6 ticity as 2.06 · 105 Mpa. As mentioned above, Solid65
J element is used to simulate the concrete. According to
Prism option
M
N 3 M,N,O,P Fanning [25], this element requires linear and multilinear iso-
2
L I tropic material properties to properly model concrete. For
K,L
K the linear isotropic material, the concrete cube compressive
J strength obtained from the experimental test as 47 MPa, and
Z Tetrahedral option (not
I
1 recommended) the young modulus of elasticity as 4.62 · 104 Mpa. The multi-
Y J linear isotropic material uses the Von Mises failure criterion
X
along with the William and Warnke [26] model to define the
Figure 3b Solid 185 – 3D solids modeling. failure of the concrete. A three-dimensional failure surface
for concrete is shown in Fig. 8. The most significant non-zero
principal stresses are in the x and y directions respectively.
capabilities. It also has mixed formulation capability for Three failure surfaces are shown as the projections on the
simulating deformations of nearly incompressible elastoplastic rxp - ryp plane. The mode of failure is the function of the sign
materials, and fully incompressible hyperelastic materials as of rzp (principal stress in Z direction). For example, if rxp and
shown in Fig. 3b. TARGE170 and CONTA173 elements were ryp, both are negative (compressive) and rzp is slightly positive
used to represent the contact slab-steel beam interface. These (tensile), cracking would be predicted in a direction perpendic-
elements are able to simulate the existence of pressure between ular to rzp. However, if rzp is zero or slightly negative, the
them when there is contact, and separation between them material is considered as crushed. Implementation of the
when there is not. The two material contacts also take into William and Warnke [26] material model in Ansys 11 requires
account friction and cohesion between the parties. The stud different constants that must be defined. Shear behavior of
shear connector was considered as a clamped metal pin in SOLID65 element in Ansys 11 is controlled by two-shear
the steel section, with rotations and translations made compat- transfer coefficient for open and closed cracks. These coeffi-
ible. On the slab connector interface, translational referring to cients represent conditions at the crack allowing for the possi-
the Y and Z axes was also made compatible and, at the Node bility of shear sliding across the crack face. A number of
below the pin head, there was a consideration of coupling in preliminary analysis were attempted in this study with various
the X direction to represent the mechanical anchoring between values for the shear transfer coefficients (for open and closed
the head of the connector and the concrete slab. The geometry cracks) within the below indicated ranges, but Ansys conver-
of these elements is as shown in Fig. 4a–c. An eight-node solid gence problems were encountered at the following entering val-
element, Solid 45, was used to model the steel plates under the ues of the William and Warnke [26] constants:
load. The element is defined with eight nodes having three
degrees of freedom at each node in the nodal x, y, and z direc- 1. Shear transfer coefficient for open crack was entered as 0.5.
tions. The geometry and node locations for this element type Its recommended range is from 0.2 to 0.5 as presented by
are as shown in Fig. 5. Three double steel-concrete composite Razaghi et al. [27].
beam models with the same material properties and cross sec- 2. Shear transfer coefficient for closed crack was entered as 1.
tion shape were analyzed. The only difference between them is Its recommended range is from 0.0 (for representing a
that the arrangement of the head studs. Two lines with differ- smooth crack, i.e., complete loss of shear transfer), to 1
ent number of head studs for the top and the bottom slabs (for representing a rough crack, i.e., no loss of shear trans-
were proposed as reported in Table 1. The geometry of the fer), as suggested by Razaghi et al. [27].
proposed model components is as shown in Fig. 6a–g. In 3. Uniaxial tensile cracking stress which was based upon the
order to saving Ansys 11 – computational time significantly, modulus of rupture; and was entered as 4.70 Mpa.
a quarter of full composite beams have been modeled as shown 4. Uniaxial crushing stress was based on the uniaxial uncon-
in Fig. 7a and b. All the investigated models are constrained fined compressive strength, and was entered as 47.0 Mpa,
at edge ab in the directions y and z, while edge cd is con- to turn on the crushing capability of the concrete element
strained in the directions x and z. In addition, other directions as discussed by Kachlakev and Miller [28].
were free of constraints as indicated in Fig. 7a and b. Thus, 5. Biaxial crushing stress.
Finite element modeling of steel concrete beam 77

Steel
reinforcement of
upper slab

Contact Head studs

elements of
upper slab Steel
beam
Contact
Y elements of Steel
lower slab reinforcement of
lower slab
X Z
(a) Finite element modeling of steel and contact elements

Target segment
element

Node-to-surface
Surface to surface contact element
contact element CONTA175
(CONTA173)

3D line-to-line
contact element
CONTA176
(b) TARGE170 Geometry

Associated Target
Surface

Contact
elements

Surface of
solid/shell element
(c) CONTA173 Geometry

Figure 4 Geometry of TARGE170 and CONTA173 elements.

4
P
5
O
6
Element coordinate M y
system (shown for N
KEYOPT(4) = 1) 2 y 3
L x
Table 1 Studs arrangement for the upper and lower slabs.
Z
y K
x
Model Number of studs in each line
I Upper slab (N1) Lower slab (N2)
Z x 1
J
SCB1 94 28
Y Surface Coordinate System
SCB2 82 28
X SCB3 82 24
Figure 5 Solid45 – 3D solids modeling.
78 A.M. Mahmoud

Solid 65 elements

(b) Slab modeling (a) quarter beam modeling

Beam 188 elements Link 8 elements

(d) Studs modeling (c) Reinforced steel modeling

Conta173 and Targe170 Solid 185 elements

elements

(e) Steel beam modeling

(f) Contact elements modeling

Solid 45 elements

(g) loading plate modeling

Figure 6 Geometry components of the all beam models.

6. Ambient hydrostatic stress state for use with constant 7 performance of the proposed model. The comparison consists
and 8. of the tests performed by Duan et al. [24] and the results
7. Biaxial crushing stress under the ambient hydrostatic stress obtained by the proposed finite element model. The proposed
state (constant 6). model delivered valuable outputs concerning the behavior of
8. Uniaxial crushing stress under the ambient hydrostatic the continuous double steel-concrete composite beams such
stress state (constant 6). as the strength capacity, the maximum deflection, the interface
9. Stiffness multiplier for cracked tensile condition. slip and slip strain of the upper and lower slab of the double
composite beam models.
Coefficients from 5 to 9 were implemented as zero value, as
discussed by Wolanski and B. [29], in order to encounter the 5.1. load–deflection relationship
Ansys convergence problem.
The load–deflection curves analyze the different performance
5. Validation of the analytical model of the double steel-concrete composite model with respect to
the strength and deflection capacities. Figs. 9–11 illustrate
The comparison of the results from the analytical model to the the load–deflection curves obtained by both the proposed
experimentally obtained results enables the validation of the and experimental approaches for the models SCB1, SCB2,
Finite element modeling of steel concrete beam 79

P 7 φ 8/m' P
1 2 1

1 7 φ 8/m' 1
2
1.45 m 1.45 m 1.45 m 1.45 m
5.80 m
0.60 m 0.60 m
0.075 m
0.08 m
Y 0.08 m 7 φ 8/m' 0.15 m
7 φ 8/m' 0.15 m
Z 0.08 m
0.35 m
X
Section 1 - 1 Section 2 - 2
(a) Geometry of proposed full-scale model

P P 7 φ 8/m'
2 1 a 2 1
a

b 1 b 1
2 2
Y 7 φ 8/m'

0.30 m
Z 0.30 m
c
X c 0.08 m
0.08 m 7 φ 8/m' 0.15 m
60 mm 7 φ 8/m' 0.15 m
d 0.08 m
d 0.175 m
0.30 m 0.30 m
13 mm c c
Stud dimensions 0.08 m 0.08 m

Y 0.15 m
0.15 m
d
X 0.08 m
d
0.175 m
Z Section 1 - 1 Section 2 - 2

(b) Geometry of proposed quarter model

Figure 7 Geometry and cross sections dimensions of all beam models.

maximum load capacity and the maximum deflection values

Table 2 Material properties of the proposed model.
for the three proposed models. Good agreement is noticed
(1) Concrete between the values of the two approaches.
Concrete strength (fc) 47 Mpa Also, it has to be noticed that the developed models exhib-
Young modulus of elasticity (Ec) 4.62 · 104 Mpa ited a softer performance than that of the experimental results.
Poison’s ratio (c) 0.3
This is due to the following reasons:
(2) Steel
Maximum tensile strength (ft) 235 Mpa 1. The William-Warnke failure criteria in Ansys cannot suit-
Young modulus of elasticity (Et) 2.06 · 105 Mpa ably predict the behavior of reinforced concrete structures,
Poison’s ratio (c) 0.2 as it does not consider the material softening properly due
(3) Studs to the varying range of its constants values, such as the
Maximum tensile strength (ft) 235 Mpa shear transfer coefficient for open and closed crack. In
Young modulus of elasticity (Et) 2.06 · 105 Mpa addition, for this kind of failure criteria, the crushed ele-
Diameter (Ustud) 13 mm ments are removed from the model and that could lead to
Height (hstud) 60 mm premature failure, which is not consistent with the real
behavior of reinforced concrete structures.
and SCB3 respectively. An increase in the proposed strength 2. Due to the possibility of the inaccuracy in modeling the
capacity values of approximately 32%, 27%, and 29% postyield behavior of steel rebar material, there is some-
compared to the experimentally obtained one is observed. what none agreeable between the finite element results
Tables 3 and 4 show the significant comparison of the and those of experimental results for postyield behavior.
80 A.M. Mahmoud

σyp

Cracking Cracking

σxp

Cracking
σzp > 0 (cracking)
σzp = 0 (crushing)
Figure 11 Load verses deflection curve for beam model SCB3.
σzp < 0 (crushing)

Table 3 Comparison of the load capacity results at collapse.

Beam model Load capacity, Pult. (kN) % Difference
Figure 8 Failure surface for concrete, William and Warnke
material model [26]. Proposed Experimental
SCB1 266.00 234.00 13.67
SCB2 265.50 233.00 13.73
SCB3 264.30 232.00 13.79

Table 4 Comparison of the maximum deflection results at

collapse.
Beam model Maximum deflection, Dmax. (mm) % Difference
Proposed Experimental
SCB1 14.49 14.61 0.80
SCB2 14.54 15.37 5.40
SCB3 14.43 14.61 1.20
Figure 9 Load verses deflection curve for beam model SCB1.

Figure 12a Interface slip values of the upper slab for beam
Figure 10 Load verses deflection curve for beam model SCB2. model SCB1.

As a result of these two statements, there is disparity length. Figs. 12a–14b illustrate the slip-beam length curves
between the proposed model results and those of Duan et al. obtained by both the proposed and experimental approaches
[24] for the pre- and postyield behavior. for the models SCB1, SCB2, and SCB3 respectively. A reduc-
tion in the proposed slip values of approximately 37%, 31%,
5.2. Interface slip values along the beam length and 47% compared to the experimentally obtained one is
observed for the upper slabs. In contrast, an increase of
The slip-beam length curves analyze the different performance approximately 21%, 30%, and 28% for the lower slabs is
of the double steel-concrete composite model with respect to noticed. Good agreement is noticed between the values of
the slip values at collapse along the composite beam model the two approaches for the cases of the upper and lower slabs.
Finite element modeling of steel concrete beam 81

Figure 12b Interface slip values of the lower slab for beam
model SCB1. Figure 14a Interface slip values of the upper slab for beam
model SCB3.

Figure 13a Interface slip values of the upper slab for beam
Figure 14b Interface slip values of the lower slab for beam
model SCB2.
model SCB3.

values of approximately 34%, 52%, and 63% compared to the

experimentally obtained one is observed for the upper slabs. In
addition, an increase of approximately 35%, 74%, and 62%
for the lower slabs is noticed. Somewhat notable non-agreeing
values are observed between the values of the two approaches
for the cases of the upper and lower slabs. Fig. 19 shows the
steps of the interface slip strain calculation for the upper slab
of proposed model SCB1 as an example of the others.

6. Parametric studies

To further improve the understanding of the strength capacity

and the fracture behavior of the continuous double steel-con-
Figure 13b Interface slip values of the lower slab for beam crete composite beams having head studs shear connectors,
model SCB2. parametric studies were performed to investigate the impact
of the presence or absence of lower slab at the interior support,
and the variation of the steel beam height. In addition, the
Fig. 15 shows the steps of the beam-slab interface slip calcula- variation of the lower slab length and thickness, and the vari-
tion for the upper slab of proposed model SCB1 as an example ation of the studs arrangement and diameter are also studied.
of the others.
6.1. The influence of removing the lower slab
5.3. Interface slip strain values along the beam length
The case study under consideration involves the influence of
The slip strain-beam length curves analyze the different perfor- removing the lower slab on the mechanical and geometrical
mance of the double steel-concrete composite model with characteristics of the beam models at failure, such as the
respect to the slip strain values at collapse along the composite strength and the deflection capacity values. The study was con-
beam model length. Figs. 16a–18b illustrate the interface slip ducted on three proposed models SB1, SCB2, and SCB3
strain-beam length curves obtained by both the proposed respectively. Fig. 20 illustrates the effect of varying composite
and experimental approaches for the models SCB1, SCB2, action on the fracture characteristics (strength and maximum
and SCB3 respectively. An increase in the proposed slip strain deflection) of the proposed model.
82 A.M. Mahmoud

Interface between concrete and steel for the upper slab

Concrete values

Steel values

Difference values

All values are multiplied by 10-3

Figure 15 Difference between the interface longitudinal displacements of concrete and steel along the beam length direction for the
upper slab of the beam model SCB1 (slip values).

Figure 16a Interface slip strain values of the upper slab for beam Figure 16b Interface slip strain values of the lower slab for beam
model SCB1. model SCB1.
Finite element modeling of steel concrete beam 83

Figure 17a Interface slip strain values of the upper slab for beam Figure 18b Interface slip strain values of the lower slab for beam
model SCB2. model SCB3.

the lower slab. It has to be noted that in case of existing the

lower slab, the proposed values of the maximum deflection
decrease by a significant average amount of 20% for all beam
models.

6.2. The influence of varying the steel beam height

In this part, the effect of changing the height of steel beam on

the characteristics of the collapse stage for the continuous dou-
ble steel-concrete composite beam is investigated. Five steel
beam heights of values 110 mm, 130 mm, 150 mm, 170 mm,
and 190 mm were proposed and applied to the model SCB1,
Figure 17b Interface slip strain values of the lower slab for beam as a case study. Fig. 21 demonstrates the effect of varying steel
model SCB2. beam height on the fracture characteristics of the proposed
model.
Fig. 21.a presents a comparison of the results of the ulti-
mate load values. This study was applied to the model SCB1
with the same height values as indicated previously. It has to
be noted that the case of the beam model with steel beam
height of 110 mm had the minimum ultimate load value,
whereas the case of the steel beam height of 190 mm had the
maximum one. The increase in the ultimate load for two con-
secutive heights (e.g. 130 mm and 150 mm) reached a signifi-
cant value of approximately 30% for all models.
Fig. 21.b compares the results obtained for the maximum
deflection values, taking into account the same model and
the proposed steel beam heights as indicated above. It has to
be observed that the case of the 110 mm steel beam height
Figure 18a Interface slip strain values of the upper slab for beam had the maximum value of the maximum deflection, whereas
model SCB3. the case of the steel beam with height of 190 mm had the min-
imum one. The decrease in the maximum deflection values for
two consecutive heights (e.g. 130 mm and 150 mm) reached a
Fig. 20.a compares the results obtained for the ultimate significant value of approximately 12% for all models.
load values, taking into account the presence or absence of
the lower slab. It has to be noted that in case of existing the 6.3. The influence of varying lower slab length
lower slab, the proposed ultimate load values increase by an
amount of 0.075% in the case of model SCB1, and by an This part contains study of the impact of changing the lower
amount of 0.11% in the case of model SCB2. This increasing slab length on the collapse stage characteristics for the
value is become 0.068% in the case model SCB3. One can continuous double steel-concrete composite beam. Four lower
observe that the presence of the lower slab increases the slab lengths of values 1000 mm, 1200 mm, 1400 mm, and
strength capacity by an average amount 0.08% for all experi- 1600 mm were proposed and applied to the model SCB1, as
enced composite models. This means that the existence of the a case study. Fig. 22 exhibits the effect of varying the lower
lower slab has a minor effect on the strength capacity values. slab length on the fracture characteristics of the proposed
Fig. 20.b presents a comparison of the results of the maxi- model.
mum deflection values. Again, the above three models were Fig. 22.a presents a comparison of the results of the ulti-
investigated twice in order to experience the effect of removing mate load values. This study was applied to the model SCB1
84 A.M. Mahmoud

Interface between concrete and steel for the upper slab

Concrete values

Steel values

Difference values

All values are multiplied by 10-3

Figure 19 Difference between the interface longitudinal strains of concrete and steel along the beam length direction for the upper slab
of the beam model SCB1 (slip strain values).

with the same lower slab length values as mentioned above. It 6.4. The influence of varying lower slab thickness
has to be observed that the case of the beam model with lower
slab length of 1600 mm had the maximum ultimate load value, The influence of changing the lower slab thickness on the char-
whereas the case of the lower slab length of 1000 mm had the acteristics of the collapse stage for the continuous double steel-
minimum one. The increase in the ultimate load for two con- concrete composite beam is studied herein. Four lower slab
secutive slab lengths (e.g. 1200 mm and 1400 mm) reached thicknesses of values 80 mm, 100 mm, 120 mm, and 140 mm
non-notable value of approximately 0.6% for all models. were proposed and executed to the model SCB1, as a case
Fig. 22.b compares the results obtained for the maximum study. Fig. 23 explicates the effect of varying the lower slab
deflection values, taking into account the same model and thickness on the fracture characteristics of the proposed
the proposed lower slab lengths as indicated above. It has to model.
be noted that the case of the beam model involving lower slab Fig. 23.a presents a comparison of the results of the ulti-
length of 1600 mm had the minimum value of the maximum mate load values. This study was applied to the model SCB1
deflection, whereas the case of the lower slab length of with the same lower slab thickness values as mentioned above.
1000 mm had the maximum one. The decrease in the maxi- It has to be noted that the case of the beam model with lower
mum deflection values for two consecutive slab lengths (e.g. slab thickness of 80 mm had the minimum ultimate load value,
1200 mm and 1400 mm) reached a remarkable value of whereas the case of the lower slab thickness of 140 mm had the
approximately 5% for all models. maximum one. The increase in the ultimate load for two
Finite element modeling of steel concrete beam 85

1000 1200 1400 1600

SCB1 SCB2 SCB3 (a) Ultimate load as a function of the lower

a. Ultimate load as a function of the composite action slab length of model SCB1

1000 1200 1400 1600

SCB1 SCB2 SCB3 (b) Maximum deflection as a function of the lower

slab length of model SCB1
b. Maximum deflection as a function of the composite action
Figure 22 Fracture characteristics as a function varying lower
Figure 20 Fracture characteristics as a function varying com-
slab length of model SCB1.
posite action.

consecutive slab thicknesses (e.g. 100 mm and 120 mm)

reached non-remarkable value of approximately 0.85% for
all models.
Fig. 23.b compares the results obtained for the maximum
deflection values, taking into account the same model and
the proposed lower slab thicknesses as mentioned above. It
has to be observed that the case of the beam model including
lower slab thickness of 80 mm had the maximum value of the
110 130 150 170 190 maximum deflection, whereas the case of the lower slab thick-
ness of 140 mm had the minimum one. The decrease in the
(a) Ultimate load as a function of the steel maximum deflection values for two consecutive slab thick-
beam height of model SCB1 nesses (e.g. 100 mm and 120 mm) reached a slightly remark-
able average value of approximately 3.5% for all models.

6.5. The influence of varying the head studs arrangement

In order to complete the parametric study, the effect of

changing the arrangement of the head studs on the character-
istics of the collapse stage for the continuous double steel-
concrete composite beam is discussed. Three cases of head
studs arrangement were proposed. The first case is when
the studs were fully arranged along the whole length of the
110 130 150 170 190
upper and the lower interface slab-steel beam surfaces. The
second case is when the studs were arranged with staggered
(b) Maximum deflection as a function of the steel shape, while the third case is when the studs were completely
beam height of model SCB1 removed. This study was applied to the model SCB1, as a
case study. Fig. 24 indicates the effect of varying the studs
Figure 21 Fracture characteristics as a function varying steel arrangement on the fracture characteristics of the proposed
beam height of model SCB1. model.
86 A.M. Mahmoud

80 100 120 140

Fully Staggared Completely
arranged arranged removed
(a) Ultimate load as a function of the lower Studs arrangement
slab thickness of model SCB1
(a) Ultimate load as a function of the studs
arrangement of model SCB1

(b) Maximum deflections as a function of the lower Fully Staggared Completely

slab thickness of model SCB1 arranged arranged removed
Studs arrangement
Figure 23 Fracture characteristics as a function varying lower
(b) Maximum deflection as a function of the studs
slab thickness of model SCB1.
arrangement of model SCB1

Figure 24 Fracture characteristics as a function varying studs

Fig. 24.a presents a comparison of the results of the ulti- arrangement of model SCB1.
mate load values. This study was applied to the model SCB1
with the same cases of the studs arrangement as mentioned
above. It has to be noted that the change of the shape of the with the same values of the head studs diameters as stated
studs arrangement has no influence on values of the ultimate above. It has to be noted that the change of the head studs
load. diameter has no influence on values of the ultimate load.
Fig. 24.b compares the results obtained for the maximum Fig. 25.b compares the results obtained for the maximum
deflection values. It has to be observed that the beam model deflection values. It has to be concluded that the beam model
including fully studs arrangement had the minimum value of containing 13 mm head studs diameter had the minimum value
the maximum deflection, whereas the case of the completely of the maximum deflection, whereas the case of the studs diam-
removed head studs had the maximum one. The increase in eter of 22 mm had the maximum one. The increase in the max-
the maximum deflection values for two consecutive studs imum deflection values for two consecutive studs diameter (e.g.
arrangement (e.g. fully and staggered arrangement) reached a 16 mm and 19 mm) attained a very slightly notable average
very slightly remarkable value of approximately 0.08% for value of approximately 0.05% for all models.
all models.
7. Conclusions
6.6. The influence of varying the head studs diameter
This paper investigates the behavior of the continuous steel-
The effect of changing the value of the diameter of the head concrete composite beam taking into account the existence
studs on the characteristics of the collapse stage for the contin- of the double composite action and the head stud shear
uous double steel-concrete composite beam is examined as a connectors.
part of this study. Four cases of head studs diameter of values Based on the finite element numerical study and the exper-
13 mm, 16 mm, 19 mm, and 22 mm were suggested and imple- imentally available results, the following main conclusions can
mented to the model SCB1, as a case study. Fig. 25 clarifies the be extrapolated:
effect of varying the studs diameter on the fracture character-
istics of the proposed model. 1. A numerical proposed model based on the finite element
Fig. 25.a presents a comparison of the results of the ulti- theory can be used to examine the geometrical and
mate load values. This study was applied to the model SCB1 mechanical characteristics in steel-concrete composite
Finite element modeling of steel concrete beam 87

This means that the shear slip has a significant contribu-

tion to composite beam deformation, which cannot be
negligible.
5. Parametric studies were carried out to look at the impact
of removing the lower slab, the effect of varying steel
beam height and the lower slab length and thickness,
and the effect of changing the head studs arrangement
and diameter. These studies were performed to investi-
gate the effect of these parameters on the strength and
13 16 19 22 the deflection capacity of the steel-concrete composite
beams having double composite action.
(a) Ultimate load as a function of the 6. The presence of the lower slab increases the proposed
studs diameter of model SCB1 strength capacity values by an average amount 0.08%
for all experienced composite models, leading to a minor
effect on the strength capacity. Moreover, the proposed
values of the maximum deflection decrease by a signifi-
cant average amount of 20% for all beam models when
removing the lower slab.
7. In comparison with the five suggested cases of steel
beam height involved in the parametric study, it can
be observed that the more increase the steel beam height
is the bigger the ultimate load values are.
8. Moreover, this study showed that the smaller the lower
13 16 19 22 slab length or thickness is the smaller the ultimate load
values and the bigger the maximum deflection values
(b) Maximum deflection as a function of the are.
studs diameter of model SCB1 9. It can be noted that the change of the shape of the studs
arrangement has no influence on the values of the ulti-
Figure 25 Fracture characteristics as a function varying studs mate load. In addition, the beam model including fully
diameter of model SCB1. studs arrangement had a minimum value of the maxi-
mum deflection, whereas the case of the completely
removed head studs had the maximum one.
beam with double composite action, resulting in a good
10. In comparison with the five head stud diameters sug-
agreement when comparing to available full-scale test
gested in this study, one can concluded that the change
data.
of this parameter has no effect on the values of the ulti-
2. The comparison of the strength capacity values obtained
mate load. In addition, it has to be noted that the smal-
by the proposed and experimental models leads to a
ler the head studs diameter is the smaller the maximum
good agreeable between them. An average increase in
deflection values are.
the proposed strength capacity values of approximately
29% compared to the experimentally available data was
concluded for all proposed models. However, a softer
performance of the validation figures (load – deflection References
curves) is observed for the developed models than that
[1] Rozsas A. Plastic design of steel–concrete composite girder
of the experimental results. This is mainly due to the
bridges. M.sc thesis, department of structural engineering, faculty
varying range of the William-Warnke constants values, of civil engineering: Budapest (Hungary); 2011.
which must be chosen carefully by a sensitivity analysis [2] Xu C, Su Q, Wu C. Experimental study on double composite
in order to encounter the Ansys convergence problems action in the negative flexural region of two-span continuous
as mentioned above. composite box girder. J Constr Steel Res 2011;67(10):1636–48.
3. An increase in the proposed interface steel-concrete slip [3] Tan El, Uy B, Hummam G. Behavior of multi-span composite
values of approximately 38% compared to the experi- steel–concrete beams subjected to combined flexure and torsion.
mentally available data was observed, leading to slightly Research and applications in structural engineering, mechanics
non-agreeable results. and computation, London, UK; 2013. p. 1397–402.
4. An increase in the proposed interface steel-concrete slip [4] Lin W, Yoda T. Numerical study on horizontally curved steel–
concrete composite beams subjected to hogging moment. Int J
strain values of approximately 49% and 55% compared
Steel Struct, USA 2014;14(3):557–69.
to the experimentally available data was observed for
[5] Henriques D, Goncalves R, Gamotim D. Nonlinear analysis of
both the upper and the lower slabs respectively, leading steel–concrete beams using generalized beam theory. In: 11th
to somewhat non-agreeing values between them. This is World Congress on Computational Mechanic, Barcelona, Spain;
due to the difference in values of friction (shear slip) at 20–25 July 2014.
the slab-steel beam interface between the analytical [6] Liang Q, Uy B, Bradford M, Ronapf H. Ultimate strength of
and experimental approaches, because of the presence continuous composite beams in combined bending and shear. J
of the contact elements for simulating this friction. Constr Steel Res 2004;60(8):1109–28.
88 A.M. Mahmoud

[7] Liang Q, Uy B, Bradford M, Ronapf H. Strength analysis of [22] Wang G, Wang FJ, et al. Theoretical analysis of double
steel–concrete composite beams in combined bending and shear. J composite beam deformation in elastic state by Goodman elastic
Struct Eng, ASCE, USA 2005;131(10):1593–600. sandwich method. Chin Railway Sci 2006;27(5):66–70.
[8] Sebastian W, McConnel RE. Nonlinear finite element analysis of [23] Yen BT, Huang T, et al. Steel box girders with composite bottom
steel–concrete composite structures. J Struct Eng, ASCE, USA flanges. In: Official proceedings, 3rd annual international bridge
2000;126(6):662–74. conference. Pittsburgh, PA (US); 1986. p. 79–86.
[9] Hirst MJS, Yeo MF. The analysis of composite beams using [24] Duan SJ, Wang JW, Zhou QD, Wang HL. An experimental study
standard finite element programs. Comput Struct on double steel–concrete composite beam specimens. Challenges,
1980;11(3):233–7. opportunities and solutions in structural engineering and con-
[10] Al-Amery RIM, Roberts TM. Nonlinear finite difference analysis struction, London; 2010. p. 209–14.
of composite beams with partial interaction. Comput Struct [25] Fanning P. Nonlinear models of reinforced and post tensioned
1990;35(1):81–7. concrete beams. Lecture, Department of Civil Engineering
[11] Salari MR, Spacone E, Shing B, Frangopol DM. Nonlinear University College Dublin Earls fort Terrace. Dublin, Ireland; 2001.
analysis of composite beams with deformable shear connectors. J [26] William KJ, Warnke ED. Constitutive model for the triaxial
Struct Eng, USA 1998;124(10):1148–58. behavior of concrete. Proc of the Int Assoc Bridge Structural
[12] Thevendran V, Chen S, Shanmungam NE, Liew JWR. Nonlinear Engineering, ISMES, Bergamo 1975;19:174.
analysis of steel–concrete composite beams curved in plan. Finite [27] Razaghi J, Hosseini A, Hatami F. Finite element method
Elem Anal Des 1999;32(3):125–39. application in nonlinear analysis of reinforced concrete structures.
[13] Reiner S. Bridges with double composite action. Struct Eng Int, Second Nat Congr Civil Eng 2005.
UK 1999;1:32–6. [28] Kachlakev D, Miller T. Finite element modeling of reinforced
[14] Stroh SL, Sen R.** Steel bridge with double composite action: concrete structures strengthened with FRP laminates. Oregon
innovative design. In: 5th International bridge engineering con- state University; May 2001.
ference, tampa, FL (US). Transportation Research Record, April [29] Wolanski J, B.S. Flexural behavior of reinforced and prestressed
3–5, 2000. vol. 1, 1696. p. 299–309. concrete beams using FRP element Analysis, A thesis submitted
[15] Newmark NM, Siess CP, Viest IM. Test and analysis of composite to the faculty of the graduated school, Marquetee university, in
beams with incomplete interaction. Proc, Soc Exp Stress Anal partial fulfillment of the requirement for the degree of master of
1951;9:75–92. science; May 2004.
[16] Duan S, Niu R, Xu J, Zheng H. A finite element model for double
composite beam. Challenges, opportunities and solutions in
structural engineering and construction, London; 2010. p. 197– Ashraf Mohamed Mahmoud graduated from
202. Ain Shams University in 1987. He obtained
[17] Duan SJ, Huo JH, Zhou QD. The research on calculation his M.Sc. degree in 1997 and Ph.D. degree in
method of the ultimate bearing capacity of double steel– 2004 from the same university. His fields of
concrete composite beam. J Shijiazhuang Railway Inst interest include the modelling of the structural
2007;20(4):1–4. systems using finite element programs for
[18] Duan SJ, Zhou QD, et al. Experimental study on bearing checking their mechanical behaviours. Also,
capacity of double steel and concrete composite continuous interested in the field of the effect of
beams. J Railway Sci Eng 2008;5(5):12–7. strengthening the steel and concrete elements
[19] Nagai M, Inaba N, et al. Experimental study on ultimate strength using the CFRP and GFRP laminates.
of composite and double composite girders. In: Proceedings of 8th The study of the stability of the cable struc-
Pacific structural steel conference steel structures in natural tures, including the geometric and material
hazards, 2007. p. 329–34. nonlinearities of the different elements for this type of structures is also
[20] Duan SJ, Duan YJ, Zhang ZG. The interface slip expression of one of his interest.
double steel–concrete composite beam under concentrated load. J
Shijiazhuang Railway Inst 2007;20(2):1–4.
[21] Yang XW, Duan SJ. The effective width of reinforcement bars for
double steel–concrete composite beam. Eng Mech 2008;25(A1):
184–8.