Engineering Structures 77 (2014) 6583

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Finite element analysis of new composite oors having cold-formed steel

and concrete slab
Yazdan Majdi a,, Cheng-Tzu Thomas Hsu b, Mehdi Zarei b
a
b

Arup, 77 Water Street, New York, NY 10005, United States

Department of Civil and Environmental Engineering, New Jersey Institute of Technology, University Heights, Newark, NJ 07102, United States

a r t i c l e

i n f o

Article history:
Received 17 November 2013
Revised 22 July 2014
Accepted 23 July 2014
Available online 15 August 2014
Keywords:
Composite oor system
Cold-formed steel
Light-gauge steel
Concrete slab
Continuous shear connector
Finite element modeling

a b s t r a c t
In this research, the structural behavior of a new type of composite oor system is explored through nite
element modeling. The new composite oor incorporates cold-formed (light-gauge) steel proles as the
joist on bottom, a corrugated steel deck as the formwork for concrete, a continuous hat channel (furring
channel) as the shear connector and nally a concrete slab on top. All steel parts in the system are coldformed and connected together by self-drilling fasteners.
In the present study, a comprehensive three-dimensional nite element modeling is performed for this
composite oor system. A local bond-slip model is applied to simulate the slip of the shear connector
inside the concrete slab. A nonlinear analysis is performed on the composite oor considering all different
types of structural nonlinearities and the behavior of the system is monitored from beginning of loading
all the way to a dened point of failure. Results of nite element analyses are compared with experimental data.
Further, parametric studies are conducted to determine the effect of shear connectors slip on reducing
ultimate strength and initial stiffness of such a oor system.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
The basic idea in composite construction is to use the advantages of both steel and concrete materials while avoiding their inherent disadvantages. In order for this idea to work, steel and concrete
parts should be fully connected so that no delamination and/or slip
can occur between the two parts. This research focuses on numerically studying a new type of composite oor system with coldformed steel having an innovative type of shear connector. The
oor system is illustrated in Fig. 1.
In this composite oor system [1,2], a continuous furring
channel has been used instead of shear studs, as the shear connector (Fig. 2). The structural system consists of: steel joists
comprising two back-to-back C-sections, a corrugated steel deck
(which acts as formwork for the concrete slab), furring channel
as the shear connector and a cast-in-place concrete slab with
transverse reinforcement. Self-drilling fasteners are used for connecting all the steel parts. The furring channel is expected to
transfer shear ow from the steel joist to the slab through its
Corresponding author. Tel.: +1 (212) 896 3000.
E-mail addresses: ym34@njit.edu (Y. Majdi), hsuc@adm.njit.edu (C.-T.T. Hsu),
mz42@njit.edu (M. Zarei).
http://dx.doi.org/10.1016/j.engstruct.2014.07.030
0141-0296/ 2014 Elsevier Ltd. All rights reserved.

bond-slip behavior with concrete [31]. Thus, the bond-slip mechanism is the key parameter in studying the effectiveness of the
furring channel.
In numerical simulation of structures having steel embedded in
concrete, it is customary to assume that there is a perfect connectivity between the two materials (i.e., no slip condition). Such an
assumption is accepted to be appropriate where the shear transfer
is provided by shear studs or other similar interlocking mechanisms [3]. However, in the composite system discussed above,
the shear transfer relies on the bond-slip behavior of the smooth
steel surface with concrete which is a new idea never used before.
Thus, it is important to investigate the performance of such shear
transfer mechanism.
This research uses computer modeling (simulation) to study the
structural behavior of such a composite oor system up to a
dened point of failure. Finite element method has been used for
structural modeling and analysis. Results of the analyses are compared with experimental data for the purpose of verication.
Further, capability of the newly introduced continuous shear
connector in maintaining composite action between steel joist
and concrete slab has been studied and the effect of shear connectors slip on behavior of the system has been numerically
evaluated.

66

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 1. The cold-formed composite oor system, subject of the current research.

2.1. Past experimental studies

Fig. 2. Proposed continuous furring channel as the shear transfer member [1].

2. Literature review
A composite oor system conventionally consists of a reinforced concrete slab that is supported on a set of steel joists. Steel
and concrete parts should be fully connected such that shear ow
can be transferred between them. The composite integrity is provided by shear connectors. Types of shear connectors include:
studs, channels, stiffened angles and at bars as illustrated in
Fig. 3. The most-often used connectors are shear studs [4].
Cold-formed steel sections are made by directly shaping thin
steel sheets, strips, plates or at bars with no heating involved.
The thickness of these sections usually ranges from 0.014900 to
0.2500 [4]. Cold-formed steel sections generally have the advantage
of being lighter, cheaper and environmentally greener, while equal
in stiffness and strength compared to regular hot-rolled sections. It
should be noted that cold-formed steel shows a gradual yielding in
stressstrain curve as compared to hot-rolled steel with sharp
yielding.
Due to the mentioned advantages, use of cold-formed steel
beams as joists in composite oor systems instead of their hotrolled counterparts has come into interest in recent years. However, as the cold-formed steel sections are light gauge (small in
thickness), it is impractical to weld shear studs to them as it was
traditionally done to hot-rolled joists [5]. This especially raises
the problem of shear transfer between steel joists and the concrete
slab. Various types of shear connectors suggested up to this date
are shown to be costly and there is little information in open technical literature about the performance and capacities of them [6,5].

Maximiliano et al. [7] were the rst researchers who studied

composite beams consisting of cold-formed double channel shape
connected to a concrete slab by shear connectors made of coldformed steel. Push-out tests were performed to nd the load-slip
behavior of them. According to the results from composite beam
tests, ultimate moments were in average 30% higher than the values obtained from elastic section analysis and 9% lower than the
values obtained from plastic section analysis.
Hanaor [5] conducted tests on composite beams with coldformed double channels and tried different types of connecting
technologies between the steel beam and concrete slab which consisted of cold-formed channel sections in both screwed and welded
conditions, powder actuated nails, expansion anchors, through
bolts and concrete screws. These connecting technologies were primarily studied in push-out tests and then were used in testing of
composite beams with cold-formed double channels. The results
showed that the response of such composite beams, if designed
and executed properly, is highly ductile. The researcher also concluded that the capacity of cold-formed shear connectors can be
conservatively determined according to cold-formed section design
codes. Fig. 4 shows the shear connectors developed by Hanaor.
Lakkavalli and Liu [6] conducted experiments on composite
beams with cold-formed double channel joists through many
large-scale tests. In their conguration, joist anges were cast into
the concrete slab. Four types of shear transfer mechanisms were
tried in their work: (1) natural surface bond between steel and
concrete, (2) shear transfer enhancements in forms of pre-fabricated bent-up tabs, (3) pre-drilled holes and (4) self-drilling
screws. Test results by these researchers indicated that specimens
employed with shear transfer enhancements showed a marked
increase in strength and reduced deection as compared with
those relying only on a natural bond between steel and concrete
to resist shear. Among the shear-transfer enhancements investigated (Fig. 5), bent-up tabs provided the best performance at both
the strength and serviceability limit states, followed by drilled
holes in the embedded anges. The use of self-drilling screws
resulted in the lowest strength.
Punurai [8] conducted a series of bending tests on the composite beam conguration discussed in this research and obtained
experimental loaddisplacement graphs for them.

2.2. Past nite element work

Barbosa and Ribeiro [9] investigated modeling of simplysupported reinforced concrete beams using ANSYS concrete

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

67

Fig. 3. Different types of shear connectors for composite beams [4].

Fig. 4. Hanaors shear connectors [5].

elements. They tried different types of plasticity models combined

with concrete cracking/crushing features. Their results suggest that
using Von Mises plasticity and isotropic work hardening rule
together with no crushing feature in concrete elements can lead
to a much better solution convergence and longer loaddeformation path in nite element modeling of structures involving
concrete in ANSYS.

Salari et al. [10] used the basic governing equations with the
force method to develop a new composite beam element for the
purpose of modeling composite beams with deformable shear connectors under small displacement.
Sebastian and McConnel [11] developed a Finite Elements (FE)
program for the analysis of general composite structures with
the ability to model ribbed composite slabs on proled steel sheeting. They used a specialized stub element with empirical nonlinear
shear force-slip relationships at the concretesteel beam interface
to permit the modeling of either full or partial shear connector
action.
Ayoub and Filippou [12] presented an inelastic beam element
for the analysis of steelconcrete girders with partial composite
action under monotonic and cyclic loads. Their element was
derived from a two-eld mixed formulation with independent
approximation of internal forces and transverse displacements.
Using the displacement formulation and the nite elements
technique, DallAsta and Zona [13] performed non-linear analysis
of composite steelconcrete beams and illustrated some aspects
related to the convergence issues by comparing solutions
deriving from nite elements with 8, 10 and 16 degrees of
freedom.
Amadio and Fragiacomo [14] focused on the effective width
evaluation for analysis of steelconcrete composite beams. Using
a parametric study carried out through the ABAQUS code, the
researchers analyzed the most important parameters that inuence the effective width and presented some preliminary criteria
for an adequate design.
Baskar et al. [15] used ABAQUS software for nonlinear nite element analysis on steelconcrete plate girders subjected to negative
bending. The researchers used thin shell elements for the girder
and solid elements for the concrete slab while shear studs were
used as shear connectors. In order to model the shear connection
between steel and concrete, two attempts were made: In the rst
attempt, bond strength at the interface of steel and concrete and
the shear strength of studs were combined and modeled as a horizontal shear friction between the two materials using ABAQUS
surface interaction technique. According to the researchers, this
method had resulted in larger deections in comparison to the
experimental data and also had resulted in severe discontinuities
between the steel and concrete surfaces. In a second attempt, the
researchers used beam elements to model the studs into the

68

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 5. Shear transfer mechanisms used by Lakkavalli and Liu [6].

concrete slab. The results from the second attempt were closer to
the experimental values and the model behaved more consistently.
Faella et al. [16] used a nite element procedure to study the
effect of nonlinearity of the shear connection in composite beams.
The results showed that a signicant increase in deformation of
steelconcrete composite beams occurs due to the nonlinearity
of shear connection.
Fragiacomo et al. [17] developed of a numerical procedure for
studying steelconcrete composite beams with regard to both
the collapse analysis and long-term behavior at the serviceability
limit state. The researchers specially considered rheological phenomena of concrete (creep and shrinkage). Later, in a related work,
Macorini et al. [18] presented a nite element model suitable for
long-term analysis of steelconcrete composite beams. The
researchers used one-dimensional beam elements for the steel
prole and two-dimensional shell elements for the concrete slab
where the two types of elements were interconnected by means
of link elements to account for deformability of the connection.
Cas et al. [19] presented a new nite element formulation for
non-linear analysis of two-layer composite beams and columns
with an interlayer slip. They introduced a layered frame element
which could account for geometrical and material nonlinearities
and also time-dependent shrinkage and creep in concrete. However the element could only take small amount of slip between
the layers.
Chung and Sotelino [20] studied different nite element modeling techniques for composite beams consisting of a steel girder and
a steel deck. In order to evaluate the accuracy of each technique,
results from each model were compared to the experimental data
from full-scale laboratory and eld tests. Fig. 6 shows all of those
modeling techniques. Having assumed a full composite action,
rigid link elements were used to connect the two steel parts in
all models. The researchers concluded that if shell elements are
used for modeling of the girder, a high level of mesh renement
is required to achieve convergence due to the displacement incompatibility between the drilling Degree of Freedom (DOF) of the web
element and the rotational DOF of the ange element. They also
concluded that the economical eccentric beam model (model no.
4 in Fig. 6) is accurate enough for prediction of exural behavior
of the composite beam.
Barth and Wu [21] conducted nonlinear nite element analyses
to study the behavior of concrete slab on steel stringers. They used
ABAQUS software and the results of simulation were veried by

comparison with experimental data. Having assumed a full composite action between the steel girder and concrete slab, multipoint contacts were used between the nodes of two materials to
model the shear transfer mechanism.
Queiroz et al. [22] conducted nite element modeling of composite beams with shear studs using ANSYS software. The researchers
performed modeling for both exible and rigid shear connections
between steel and concrete using links and springs as shown in
Fig. 7 (bond link elements). The reliability of their models was demonstrated by comparisons with experimental data. They concluded
that their simulation was able to successfully predict load
deection response of composite beams subjected to concentrated
or uniformly distributed loading. The researchers also concluded
that the continuation of the shear connection beyond the beam supports of simply-supported beams can favorably affect not only the
overall system response, but also the slip and the stud force distributions along the beam. Furthermore, they showed that partial interaction in shear connection between steel and concrete can signicantly
affect both strength and deection of the composite beam.
Zhao and Li [23] studied the nonlinear mechanical behavior and
failure process of a bonded steelconcrete composite beam by
using nite element method. The researchers used a threedimensional FE model and discussed major reasons for failure of
composite beams.
Nguyen et al. [24] investigated the behavior of continuous composite beams with discrete shear connection using a nonlinear
mixed nite element model. The researchers paid special attention
to the hogging moment regions where cracking occurs.
Queiroz et al. [25] used a simplied two-dimensional nite element modeling for evaluation of composite beams by ANSYS software. They used plastic beam elements to model steel beam and
concrete slab and a collection of links to model the shear studs
(see Fig. 8). Their model was capable of simulating both full and
partial composite action between steel and concrete. The results
were validated by comparison with experimental data and more
sophisticated three-dimensional models. The conclusion was that
this type of modeling, with fewer numerical convergence problems, faster time of analysis and far more simplicity, is accurate
enough for prediction of the behavior of a composite beam system
Tsavdaridis et al. [26] performed nite element modeling of
composite ultra-shallow oor beams (USFB) using ANSYS. In their
system, concrete slab lies within the steel anges and is connected
through the web opening providing enhanced longitudinal and

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

69

Fig. 6. Different techniques in nite element modeling of girder-steel deck composite [20].

vertical shear resistance. The researchers compared their results

from ANSYS nite element modeling to experimental data and a
very well agreement was observed. In that research, the bond
between steel and concrete was accounted for in computer models
using various MohrCoulomb friction coefcients.
3. Modeling assumptions
3.1. General considerations
In nite element modeling of the oor system, a single steel
joist plus an effective width of the concrete slab and steel deck
are considered. The effective width is chosen based on AISC

recommendations [3]. In geometry of the models, concrete ribs,

steel deck ribs and joist openings are included. Concrete slab is
modeled using 3D solid elements while all steel parts, due to their
small thicknesses, are modeled using 2D shell elements. All steel
members are represented with their mid-planes and corner curves
of the steel joist are approximated with sharp edges. Due to the
symmetry, only one quarter of the system is required to be modeled. In all models, steel joist is made of two AISI standard C sections (introduced in Table 1 for each model) and steel deck is an
AISI standard 9/1600 cold-formed deck (gauge 20) with proof stress
of 33 ksi (proof stress is dened in section 3.4). The continuous
shear connector is a cold-formed hat channel developed by Hsu
et al. [1] as shown in Fig. 2 and has a proof stress of 33 ksi.

70

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 7. Finite element modeling of a composite beam with shear studs as shear connectors [22].

Fig. 8. Two-dimensional nite element modeling of a composite beam using beam elements [25].

Fig. 9. Support conditions and loading pattern in the models.

2. All models are loaded with two vertical concentrated loads

(equal to P/2) in one third and two third of the span. This type
of loading will cause a pure bending condition between the
applied forces. Total loading on the model will be equal to the
summation of the two applied forces (=P).
3. It is assumed that self-weight of the system also appears as two
equivalent concentrated loads in one third and two third of the
span. This idealization causes minor error in the results as the
effect of self-weight is insignicant. Thus, in ANSYS models,
the total load, P, comprises two portions: the self-weight (dead
load) and the piston load (live load).

3.2. Support conditions and loading pattern

3.3. Elements meshing
Support conditions and loading pattern are chosen as follows
(see Fig. 9):
1. All models are assumed to be simply-supported in both ends. In
order to create such condition, the nodes at the bottom of steel
joist are prevented from vertical translation, yet they are free to
rotate and are allowed to have horizontal translations.

Finer mesh is used around the furring channel where high stress
gradient is expected. The meshing becomes coarser in horizontal
direction at far concrete edges where stress gradient is expected
to be very low. Since high stress gradient is expected in vertical
direction all long the beam section, a ne meshing is considered
vertically for the concrete slab and the steel joist.

71

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 10. Uniaxial stressstrain model for cold-formed steel (Fy = proof stress).

Fig. 11. Uniaxial compressive stressstrain model for concrete with f c 3200 psi.

On the faces where different parts are in contact or attached by

screws, a perfect node to node matching is performed in meshing.
General meshing of the models are rened in multiple steps until
further renement does not signicantly affect the analysis results.
Refer to Figs. 17 and 18 for general meshing of the system.
3.4. Modeling cold-formed steel material

Fig. 12. Local bond-slip model for cold-formed steel embedded in concrete.

In the present modeling, the cold-formed steel material is

assumed to follow a multi-linear stressstrain pattern in uniaxial
tension. The pattern is chosen to agree with actual stressstrain

72

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 16. Parameters of Cohesive Zone Material as used by ANSYS [27].

Fig. 13. Visualization of slip between steel and concrete.

curves obtained from tension tests on material coupons. Fig. 10

depicts the assumed multi-linear patterns for the cold-formed
steel. Note since cold-formed steel typically does not show a clear
yield point in its stressstrain graph, an offset yield stress (also
known as proof stress) is dened instead of yield stress for that
by intersecting an offset line with the stressstrain curve. The offset line is drawn from the 0.2% strain value with a slope parallel to
the initial slope of the curve [4] (see Fig. 10).
For all cold-formed steel materials, von Mises yield criterion is
used with associative ow rule and isotropic (work) hardening rule
[27]. Poissons ratio for cold-formed steel is assumed to be equal to
0.3.
3.5. Modeling concrete material

Fig. 14. Distribution of bond stress over the embedded steel in general case.

A multi-linear uniaxial stressstrain relationship in compression is assumed for concrete based on the following numerical
expression as suggested by Desayi and Krishnan [28]:

Fig. 15. Test set-up introduced by Majdi et al. [31] to study bond-slip behavior between the proposed furring channel and concrete.

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

73

Plasticity is considered in the concrete model using von Mises

yield criterion, associative ow rule and isotropic (work) hardening
[27]. No crushing is assumed for the concrete.
When tensile principal stress in a point attains the value of fct
(splitting tensile strength), concrete is assumed to crack perpendicular to the direction of stress in that point. The tensile stress in
point of crack will suddenly be dropped to zero. Tensile behavior
of concrete is assumed to be linear before cracking. The value of
fct is assumed equal to the average splitting tensile strength suggested in the commentary of ACI code [29] section R8.6.1:

q
0
fct 6:7 fc
In which

4
0
fc

is in psi, or:

q
0
fct 0:56 fc
0

Fig. 17. ANSYS nite element modeling of the proposed composite oor system
3D view.

In which fc is in MPa.
Poissons ratio for concrete is assumed equal to 0.2.
3.6. Modeling structural components

Fig. 18. ANSYS nite element modeling of the proposed composite oor system
3D view of steel parts only.

Concrete slab is modeled using SOLID65 element in ANSYS [30].

The element is dened by eight nodes having three degrees of freedom at each node: translations in the nodal x, y, and z directions.
The element is capable of considering reinforcement and cracking
in the concrete. The reinforcement and cracks are assumed
smeared in the element.
Cold-formed steel parts are the furring channel, steel deck and
the steel joist. Since these parts all consist of thin steel sheets,
SHELL181 element is found to be suitable to model them [30].
SHELL181 is a four node element with six degrees of freedom at
each node: translations in the x, y, and z directions, and rotations
about the x, y, and z-axes. The element is well-suited for linear,
large rotation, and/or large strain nonlinear applications. Change
in shell thickness as a result of large-deformation is accounted
for in nonlinear analyses.
Refer to Figs. 17 and 18 for elements meshing and types of elements used for steel and concrete parts.
3.7. Bond-slip behavior between furring channel and concrete

Ec e
 2
1 ee0

e0

2fc
Ec

In the proposed composite oor system, shear transfer between

the steel joist and concrete slab relies on the bond-slip behavior
between the faces of the continuous furring channel and concrete.
A local bond-slip model is required to simulate such behavior in
Finite Element Analysis. Local bond-slip model is referred to a

where f is the stress at strain e, e the strain at stress f, e0 the strain at

0
the ultimate compressive strength, fc and Ec is the concrete modulus
of elasticity (equal to initial slope of the curve), determined from
the equation suggested in ACI code [29]:

q
0
Ec 57; 000 fc
In which

0
fc

is in psi, or:

q
0
Ec 47; 000 fc
0

In which fc is in MPa.
The assumed multi-linear model is depicted in Fig. 11 for
0
fc 3200 psi (22.06 MPa). As seen in this gure, the model consists
of 6 linear portions. The softening branch of stressstrain curve is
approximated by a perfectly plastic behavior beyond the point of
ultimate stress to overcome convergence problems. Similar
0
multi-linear models with 6 lines are used for other fc values.

Fig. 19. ANSYS nite element modeling of contacts between concrete slab and steel
parts.

74

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

d1 0:001100 0:279 mm
q
sf 0:054fc0 8:433 fc0  173
df

0
0:369 lnfc

5
6

 0:595

7
0
fc and

where d1 and df are in inches,

sf are in psi and
0
2000 psi 6 fc 6 6000 psi.
In SI units, the above equations will take the form:

sf 0:054fc0 0:7 fc0  1:193

df 9:373 ln

q
0
fc 8:213

Fig. 20. ANSYS nite element modeling of self-drilling fasteners.

Fig. 21. ANSYS nite element modeling of the furring channel with bent ribs on the
lips.

graph or function which at any arbitrary point of contact, expresses

bond stress (shear stress transferred over the steel and concrete
faces, s) in terms of the slip between the contact faces (d).
Majdi et al. [31] developed and veried such a local bond-slip
model. The researchers began their work with setting up and solving the bond-slip governing differential equation based on a bilinear behavior (Fig. 12) to obtain the necessary mathematical
equations. Then a new test set-up was introduced for the experimental studies followed by a new innovative procedure to calibrate the mathematical equations based on selected points from
the test data. Subsequently, a local bond-slip model was developed
by curve tting through four sets of results. Finally, the researchers
veried validity of their proposed model.
Fig. 13 depicts how the slip (d) between two materials is dened
as the difference between deformation of steel (us) and deformation of concrete (uc). In Fig. 14, bond stress distribution is shown
over a portion of embedded steel in a general case as a result of
the assumed bi-linear model. Further, the test set-up to obtain
experimental bond-slip data is shown in Fig. 15. Note the furring
channel shown in the set-up is identical to the furring channel
used as shear transfer member in the current study.
Eventually, the following equations were proposed to determine the parameters of the bi-linear bond-slip model based on
0
the concrete strength, fc [31,32]:

where fc and sf are in MPa and df is in mm.

In the Finite Element Models, the concrete deck is connected to
the furring channel through surface-to-surface contact and target
elements, CONTA173 and TARGET170 [30], to simulate the bondslip condition between the two materials (see Fig. 19).
A series of contact elements are created over the surface of
embedded steel shell elements sharing common nodes with them.
Similarly, a series of target elements are considered over the
surface of concrete solid elements which are in contact with steel
elements. Thus the actual behavior between steel and concrete
surfaces can be represented through the interaction of contact
and target elements [33]. This interaction can be programmed to
follow a special constitutive law equivalent to the developed local
bond-slip models.
ANSYS Bilinear Cohesive Zone Material Model [27] can be
employed to simulate the constitutive law between steel and concrete surfaces. In this model it can be assumed that contact elements can have tangential slip over the target elements
according to the rule shown in Fig. 16. ANSYS Bilinear Cohesive
Zone Material Model is based on the model proposed by Alfano
and Criseld [34]. According to that model (referring to Fig. 16),
debonding begins at point A and is completed at point C. After debonding has been initiated in point A, it is assumed to be cumulative
and any unloading and subsequent reloading occurs in a linear
elastic manner along line OB at a more gradual slope. Parameter
dt is obtained from:

dt

t
ut  u
ut



uct
t
uct  u


8

And is used to determine the Bond Stress at any point from:

st K t ut 1  dt :

3.8. Contact between concrete slab and steel deck

The contact behavior between concrete slab and steel deck is
also considered and is modeled with the pair of CONTA173/
TARGE170 elements as described before. In this case, the two faces
are constrained such that they are not able to penetrate into each
other; however they are free to slip and/or become separated. The
contact surfaces are assumed friction-free and no-penetration is
the only constrain (see Fig. 19).
3.9. Modeling connectivity with screws
All short screws are assumed to create a no-slip condition
between the two connecting steel faces. The effect of bending in
long screws is accounted for by modeling beam elements between
the furring channel and the steel joist. Additionally, screws are
assumed to have adequate strength to maintain connectivity
between steel parts until the end of analysis.
CONTA178 element is used to model the connectivity between
any two faces attached with a short screw [30]. This connectivity is

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

75

Fig. 22. Loaddeformation graphs for models A1p and A1b compared with experimental data.

Fig. 23. Loaddeformation graphs for models A2p and A2b compared with experimental data.

assumed to be over multiple pairs of nodes rather than one single

pair of node on the contacting faces. CONTA178 is a node-to-node
contact element. The element has two nodes with three degrees of
freedom at each node: translations in the X, Y, and Z directions.
Each node of the element is attached to one of the two connecting
faces; such that the two nodes will have coincident coordinates in
the space. The contact behavior between each node pair is
assumed perfectly bonded in all directions. The average of forces

transmitted between the two sides of the contact can be interpreted as forces in the screw connecting the two faces.
BEAM188 element is used to model the connectivity of long
screws between furring channel and the steel joist [30]. Once again
the connectivity is assumed to be over multiple pairs of nodes
rather than one single pair of node. BEAM188 is a two node line
element with 6 (or 7) degrees of freedom at each node. These
include translations in the x, y, and z directions and rotations about

76

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Fig. 24. Loaddeformation graphs for models A3p and A3b compared with experimental data.

Fig. 25. Model A1b: distribution of shear force in short fasteners (connecting steel
deck to the joists) over half-span length.

Fig. 27. Model A1b: deection of the system at ultimate load.

Fig. 26. Model A1b: distribution of shear force in long fasteners (connecting furring
channel to the joists) over half-span length.

the x, y, and z directions (a seventh degree of freedom which is

warping magnitude is optional for the element and is not used in
the present modeling). One node of the element is attached to
the bottom face of furring channel while the other node is attached
to the top face of steel joist. The average of shear forces and bending moments in all beam elements can be interpreted as the shear
force and bending moment in the attaching long screw. Values of
axial and exural stiffness for the beam elements group is adjusted
to match the actual stiffness of the connecting screw.
Refer to Fig. 20 for nite element modeling of fasteners.

Fig. 28. Model A1b: distribution of exural stress in the concrete slab at ultimate
load (steel parts not shown for clarity).

77

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Table 1
List of nite element models with their specications.
Model name

Steel joist sectiona

Concrete slabs heightb,

in (cm)

Concrete slabs
width, ft (m)

Span length,
ft (m)

f c , ksi (MPa)

Fyc, ksi (MPa)

Shear connector
type

Bond slip
behavior applied?

A1p
A1b
A2p
A2b
A3p
A3b
B1p
B1b
B2p
B2b
B3p
B3b
B4p
B4b
C1p
C1b
C2p
C2b
C3p
C3b
C4p
C4b

600S200-68
600S200-68
600S200-68
600S200-68
600S200-54
600S200-54
600S200-68
600S200-68
600S200-68
600S200-68
600S200-68
600S200-68
600S200-68
600S200-68
600S200-97
600S200-97
600S200-68
600S200-68
600S200-68
600S200-68
600S200-54
600S200-54

3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3 (7.62)
3 (7.62)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3.5 (8.89)
3 (7.62)
3 (7.62)
3 (7.62)
3 (7.62)

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3

12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
8 (2.44)
8 (2.44)
10 (3.05)
10 (3.05)
12 (3.66)
12 (3.66)
16 (4.88)
16
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)
12 (3.66)

3.2
3.2
3.2
3.2
2.3
2.3
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2
3.2

41
41
41
41
37
37
41
41
41
41
41
41
41
41
41
41
41
41
41
41
41
41

Regular
Regular
With cut
With cut
With cut
With cut
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular
Regular

NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES
NO
YES

(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)
(0.914)

(22.06)
(22.06)
(22.06)
(22.06)
(15.86)
(15.86)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)
(22.06)

(282.68)
(282.68)
(282.68)
(282.68)
(255.11)
(255.11)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)
(282.68)

&
&
&
&

bent
bent
bent
bent

a
Designation as adopted by AISI standard [36], e.g., 600S200-68 means cold-formed C section with total height of 6 in., ange width of 2 in. and nominal thickness of
0.068 in. (for SI, 1 in. = 2.54 cm).
b
Measure from top of lower steel decks rib.
c
Proof stress of cold-formed steel, obtained from 0.2% offset method.

3.10. Furring channel with cut and bent ribs

As a modication to the regular set-up, some cuts are made
on the shear connectors lips at longitudinal distances of every
1 foot (30.48 cm) to study their effect on the systems structural
behavior. The cut edges are then bent up to form some ribs
which will interlock into the concrete and provide better
integrity [2]. Finite element modeling of these bent ribs is shown
in Fig. 21. Surfaces of the ribs are in contact with the concrete.
Additionally, a hinge connection has been considered at the
bottom of bent ribs.
4. Analysis assumptions
Nonlinear analysis is performed on each model using the
NewtonRaphson method. All models are loaded gradually in
multiple steps and in a displacement based control procedure.
The amount of loading is increased until the maximum principal compressive strain in concrete elements on the top attains
the value of 0.003 which is the amount of ultimate strain for concrete proposed by the ACI code [29]. The analysis will stop then
and the total amount of the applied load in the last step will be
recorded as the ultimate strength (ultimate load) of the system.
Post crushing and post peak behaviors of the composite oor system are not taken into account.
It is observed that solution convergence is extremely difcult to
achieve as a result of a high level of nonlinearity in the system
(material nonlinearity, large deformation and contact behavior).
In order to overcome the convergence issues, the following techniques are introduced and used:
1. All models are loaded in a displacement based control procedure rather than a load based control procedure.
2. If rebar are not designed in the concrete slab in certain directions, a very small value of reinforcement (arbitrarily chosen
as 0.1% of the concrete volume) smeared in concrete elements
is used as ctitious reinforcement to prevent numerical

instabilities caused by an extensive concrete cracking at large

deformation. Note this ctitious reinforcement does not result
in a signicant error as the stress observed in cracked concrete
is very close to zero (see Fig. 28).1
3. In each analysis, initially a large value of convergence tolerance
was allowed for the solution. This value is then reduced step by
step as the analysis results are monitored. The reduction in the
convergence tolerance is continued until further decrease
would have no considerable effect on the analysis results. In
all the analyses, convergence tolerance is based on both displacement and force criteria.
4. In places of concentrated loads, local concrete stiffener blocks
are modeled inside the C-channels, similar to the actual conditions in experimental tests. This prevents extensive local yielding and cracking which could prematurely terminate the
analysis.
5. In all models, loading is applied in multiple steps such that the
maximum increase in deections of the system is limited to
0.01 inch in a loading step.
5. Results of nite element analyses
5.1. Designation of models
Each model is designated by a capital letter A, B, C
followed by a number and ending with either letter p or
b.
The rst capital letter species the category of the model.
Models of category A have regular or with cut & bent ribs furring
channel and their results will be compared with the existing experimental data by Punurai [8]. Category B models are created to
study the effect of shear connectors slip in concrete on the systems behavior for different span lengths. Category C models
are created to study the effect of shear connectors slip in concrete
1
For interpretation of color in Figs. 27 through 31, the reader is referred to the web
version of this article.

78

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

on the systems behavior for different values of sections moment

of inertia.
Letter p in the end of a models name shows that the shear
connector in that model is perfectly bonded with the surrounding
concrete while letter b shows that the shear connector in that
model has a bond-slip behavior inside the concrete following the
bi-linear model proposed by Majdi et al. [31].
A list of the models together with their specications is given in
Table 1. In all models, height of concrete is measured from the
lower steel decks rib to the top of concrete and Fy represents proof
stress of cold-formed steel.
5.2. Results for category A models
In Figs. 2224, loaddeformation graphs for category A
models are illustrated and compared with experimental data.
Contour result at ultimate load for deection of the composite
beam A1b is shown in Fig. 27.1 Also, exural stress in concrete,
exural stress in steel, shear connectors slip distribution and shear
connectors bond stress distribution are depicted in Figs. 2831,
respectively at mid-span of model A1b. Note Fig. 28 implies
that Inelastic Neutral Axis at ultimate load lies at the boarder
of yellow and brown. Also, since exural stress is close to zero
in the brown zone, that zone can be assumed as the cracked
concrete.
Further, Figs. 25 and 26 depict distribution of shear force in
each fastener for model A1b at ultimate load.
Based on studying the results of category A models, the
following conclusions can be drawn:
Results of nite element modeling for the proposed composite
oor system are in acceptable agreement with the experimental
data.
The nite element analyses presented here lead to more conservative results as compared to the experiments. This can be
attributed to several reasons such as: 1 Self-weight is modeled
in two concentrated points rather than uniformly along the
beam. 2 Post cracking tensile strength of concrete is neglected.
3 Materials under bending can be stronger than what they are
in pure tension, especially in large deformations. 4 The
smooth stress/strain curve of materials is approximated with
a multi-linear pattern. 5 Theoretical assumptions regarding
material plasticity and large deformation are conservative. 6
Systematic and random errors always exist in experiments.
Cut and bent ribs of the shear connector do not signicantly
improve the ultimate strength and the general structural behavior of the system based on the present nite element modeling
and the test results by Punurai [8].
Referring to Figs. 2831, it is observed that the new proposed
shear connector (the furring channel) is successfully capable
of developing a composite action in the system. As seen in
Fig. 30, slip of the furring channel inside concrete slab is within
reasonable range. Also, Fig. 31 shows distribution of the bond
stress on surface of the furring channel agrees with shear diagram of the composite beam which proves the shear transfer
has been successfully provided by the furring channel up to
the ultimate load.

Fig. 29. Model A1b: distribution of exural stress in steel parts at ultimate load
(concrete slab not shown for clarity).

Fig. 30. Model A1b: distribution of furring channels slip from the end support to
the mid-span at ultimate load.

5.3. Results for category B and C models

Loaddeformation graphs for category B models are shown in
Fig. 32. It is observed that as the span length increases, the effect of
bond-slip on structural response of the system reduces.
All category C models have the same span lengths of 12 feet
(3.66 m) and the same material properties and concrete slab
width. However, different values of sections moments of inertia

Fig. 31. Model A1b: distribution of furring channels bond stress from the end
support to the mid-span at ultimate load (bond stress shown is the summation of
front and back faces).

are used in the nite element modeling by changing the joist

thickness and the height of concrete slab. The loaddeformation
graphs for these models are shown in Fig. 33. It is observed that

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

79

Fig. 32. Loaddeformation graphs for category B models.

Fig. 33. Loaddeformation graphs for category C models.

as the sections moment of inertia decreases, the effect of bondslip on structural response of the system reduces. The moment
of inertia calculated for category C models are based on nonslip condition and no crack in concrete, to provide the easiest calculations for users.
Eventually, it is concluded that for each one of models in category B and C, the value of moment of inertia over span length
can be chosen as a parameter to represent the effectiveness of
bond-slip behavior in structural response of the system.

Values of ultimate loads and initial slopes of loaddeformation

graphs (initial stiffness) are compared versus moment of inertia
over span length as depicted in Tables 2 and 3, respectively. The
amounts of percentage reduction in ultimate loads and initial
slopes are plotted versus the parameter of moment of inertia over
span length as shown in Figs. 34 and 35, respectively.
In order to quantify the effect of bond-slip behavior on ultimate
load and stiffness of the proposed composite oor system, the
following equations are derived by curve tting:

80

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Table 2
Comparison of ultimate loads for different values of moments of inertia over span lengths.
Model

M.I./span, in3 (cm3)

B1p & B1b

0.5450
(8.9309)
0.4360
(7.1448)
0.3634
(5.9551)
0.2725
(4.4655)
0.4672
(7.6560)
0.3634
(5.9551)
0.3073
(5.0357)
0.2587
(4.2393)

B2p & B2b

B3p & B3b
B4p & B4b
C1p & C1b
C2p & C2b
C3p & C3b
C4p & C4b

Ultimate total load, kips (kN)

% Reduction due to bond-slip

Perfect bond condition

Bond-slip applied

26.455 (117.68)

22.961 (102.14)

13.207

19.643 (87.38)

18.836 (83.79)

4.108

17.964 (79.91)

17.644 (78.48)

1.781

13.264 (59.00)

13.223 (58.82)

0.309

23.043 (102.50)

21.412 (95.25)

7.078

17.964 (79.91)

17.644 (78.48)

1.781

15.429 (68.63)

15.31 (68.10)

0.771

12.911 (57.43)

12.88 (57.29)

0.240

Table 3
Comparison of initial slopes of loaddeformation graphs (stiffness) for different values of moments of inertia over span lengths.
Model

B1p & B1b

B2p & B2b
B3p & B3b
B4p & B4b
C1p & C1b
C2p & C2b
C3p & C3b
C4p & C4b

M.I./span, in3 (cm3)

0.5450
(8.9309)
0.4360
(7.1448)
0.3634
(5.9551)
0.2725
(4.4655)
0.4672
(7.6560)
0.3634
(5.9551)
0.3073
(5.0357)
0.2587
(4.2393)

Initial slope of the graph, kips/in (kN/cm)

% Reduction due to bond-slip

Perfect bond condition

Bond-slip applied

74.282 (130.088)

62.791 (109.964)

15.469

38.5 (67.424)

33.875 (59.324)

12.013

23.041 (40.351)

20.774 (36.381)

9.839

10.358 (18.140)

9.555 (16.733)

7.752

30.518 (53.445)

26.667 (46.701)

12.619

23.041 (40.351)

20.774 (36.381)

9.839

19.445 (34.053)

17.776 (31.131)

8.583

16.142 (28.269)

14.943 (26.169)

7.428

Fig. 34. Percentage reduction in ultimate total load due to bond-slip behavior plotted versus moment of inertia over span length.

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

81

Fig. 35. Percentage reduction in initial slope of loaddeformation graph (stiffness) due to bond-slip behavior plotted versus moment of inertia over span length.

 5:42
I
PRU 394:16
L
 
I
PRS 27:78
L

10
11

where PRU is the Percentage reduction in ultimate load due to

bond-slip behavior (shear connectors slip) [%]; PRS the Percentage
reduction in stiffness (initial slope of loaddeformation graph)
due to bond-slip behavior (shear connectors slip) [%]; I the Moment
of inertia of the composite section [in4] and L the Span length of the
composite beam [in]. For calculation of moment of inertia (I), concrete should be converted to an equivalent steel and no cracking
should be assumed in concrete. Further, effects of steel deck and
concrete slab ribs should be neglected [35].
In SI units, the equations will take the following form:

 5:42
I
PRU 0:0004
L
 
I
PRS 1:691
L
where LI has unit of cm3.

6. Simplied section analysis

Basic principles of mechanics of materials can be employed to
perform section analysis on the proposed composite oor system
to determine ultimate moment capacity, elastic deformation, and
average shear ow between steel joists and concrete slab. Majdi
and Hsu [35] have published a technical report which completely
covers the procedure and details for analysis and design of the proposed composite oor system based on section analysis and code
requirements.
The key assumptions in section analysis of this oor system are:
1 There is perfect bond condition between steel joist and concrete slab; However, corrections on systems ultimate strength
and deection can be applied using Eqs. (10) and (11) to account
for the effect of shear connectors slip. 2 Proof stress of coldformed steel is used as the yield stress. 3 Participation of steel
deck in strength and stiffness of the system is negligible. This is
proven by both nite element analysis [32] and analysis of the
cross section [35]. 4 Effect of concrete slab ribs should be
neglected [3].
Fig. 36 illustrate inelastic section analysis to determine ultimate
moment capacity of the system.

Fig. 36. Section analysis of the proposed composite oor system assuming inelastic behavior of materials.

82

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

Table 4
Comparison of ultimate loads obtained from F.E. modeling, section analysis and
experiments.
Model

A1p
A1b
A2p
A2b
A3p
A3b

Ultimate total load, kips (kN)

From F.E. modeling

From section analysis

From experiment

17.964 (79.908)
17.644 (78.484)
17.966 (79.917)
17.874 (79.507)
11.05 (49.153)
10.987 (48.873)

17.289 (76.905)

19.733 (87.777)

17.289 (76.905)

19.981 (88.880)

11.459 (50.972)

12.041 (53.561)

Table 5
Comparison of average shear ow at ultimate load obtained from F.E. modeling and
section analysis.
Model

A1b
A2b
A3b

Average shear ow on top of joist at ultimate load, kips/in (kN/cm)

From F.E. modeling

From section analysis

0.857 (1.501)
0.881 (1.543)
0.605 (1.060)

0.914 (1.601)
0.914 (1.601)
0.654 (1.145)

Results of systems ultimate load for category A models from

nite element analyses, experiment data and section analyses are
compared in Table 4. Further, average shear ow for category A
models obtained from nite element analyses (shown in Figs. 25
and 26) are compared in Table 5 with the results from section analyses. As concluded from those tables, there is quite acceptable
agreement between results from the three methods. Note that
the present nite element modeling and section analysis have
achieved more conservative results in comparison to experiments.

7. Conclusions
The following conclusions can be drawn from this study:
Use of multi-linear material nonlinearity with von Mises yield
criteria, associative ow rule and isotropic (work) hardening
combined with large-deformation geometrical nonlinearity
and nonlinear contact behavior between steel and concrete produced reasonable results in nite element analysis of the discussed oor system and results are in good agreement with
the experimental data. The analysis by ANSYS terminates in
computation as soon as the compressive principal strain on
the top of the concrete reaches the value of 0.003 as suggested
in the ACI code and the corresponding load is considered as the
ultimate total load on the system. Post-crushing and post-peak
behaviors of the composite oor system are not taken into
account in this study.
To overcome the convergence issues of the highly nonlinear
system of this study, several techniques are used. For instance,
if rebar are not designed in the concrete slab in certain directions, a very small value of reinforcement smeared in concrete
elements is used as ctitious stiffness to prevent numerical
instabilities.
The proposed furring channel is capable of effectively providing
the necessary shear transfer between steel and concrete to
achieve a composite action. Bond-slip behavior of the furring
channel inside concrete has been accounted for in the models.
Applying bond-slip behavior between steel and concrete in the
models will decrease the ultimate strength and the initial stiffness of the system as compared to the perfect bond condition.
This decrease is more signicant for systems with higher value

of moment of inertia over span length. Empirical equations are

derived and presented to determine the percentage decrease in
the ultimate strength and stiffness.
The ultimate loads obtained from nite element analyses,
experimental data and section analyses have been found to be
in quite acceptable agreements. Results from the nite element
modeling and section analysis are generally more conservative
compared to experiments. Further, results of the average shear
ow between steel joist and concrete from the nite element
analysis and section analysis have been found to be acceptably
close.

References
[1] Hsu CTT, Punurai S, Munoz P. United States of America Patent 7,779,590; 2010.
[2] Hsu CTT, Punurai S, Punurai W, Majdi Y. New composite beams having coldformed steel joists and concrete slab. J Eng Struct 2014;71C:187200.
[3] AISC. Specication for structural steel buildings. Chicago, Illinois: American
Institute of Steel Construction, 2005; 2011.
[4] Yu WW, LaBoube RA. Cold-formed steel design. Hoboken, NJ: John Wiley &
Sons; 2010.
[5] Hanaor A. Tests of composite beams with cold-formed sections. J Constr Steel
Res 2000:24564.
[6] Lakkavalli BS, Liu Y. Experimental study of composite cold-formed steel Csection oor joists. J Constr Steel Res 2006:9951006.
[7] Maximiliano M, Walter AN, Jose JS. Cold-formed shear connectors for
composite constructions. In: 14th International specialty conference on coldformed steel structures, October 1516, St. Louis, Missouri; 1998. p. 40921.
[8] Punurai S. Behavior of composite beams with cold-formed steel joists and
concrete. Newark, NJ: New Jersey Institute of Technology, PhD Dissertation;
2007.
[9] Barbosa AF, Ribeiro GO. Analysis of reinforced concrete structures using ANSYS
nonlinear concrete model. Comput Mech, New Trends Appl 1998:17.
[10] Salari M, Spacone E, Shing PB, Frangopol DM. Nonlinear analysis of composite
beams with deformable shear connectors. J Struct Eng 1998;124(10):114858.
[11] Sebastian WM, McConnel RE. Nonlinear FE analysis of steelconcrete
composite structures. J Struct Eng 2000;126(6):66274.
[12] Ayoub A, Filippou FC. Mixed formulation of nonlinear steelconcrete
composite beam element. J Struct Eng 2000;126(3):37181.
[13] DallAsta A, Zona A. Nonlinear analysis of composite beams by a displacement
approach. Comput Struct 2002;80(2730):221728.
[14] Amadio C, Fragiacomo M. Effective width evaluation for steelconcrete
composite beams. J Constr Steel Res 2002;58(3):37388.
[15] Baskar K, Shanmugam NE, Thevendran V. Finite-element analysis of
steelconcrete composite plate girder. J Struct Eng, ASCE 2002;128(9):
115868.
[16] Faella C, Martinelli E, Nigro E. Shear connection nonlinearity and deection of
steelconcrete composite beams: a simplied method. J Struct Eng, ASCE
2003;129(1):1220.
[17] Fragiacomo M, Amadio C, Macorini L. Finite element model for collapse and
long-term analysis of steelconcrete composite beams. J Struct Eng
2004;130(3):48997.
[18] Macorini L, Fragiacomo M, Amadio C, Izzuddin BA. Long-term analysis of steel
concrete composite beams: FE modeling for effective width evaluation. Eng
Struct 2006;28(8):111021.
[19] Cas B, Saje M, Planinc I. Non-linear nite element analysis of composite planar
frames with interlayer slip. Comput Struct 2004;82:190112.
[20] Chung W, Sotelino ED. Three-dimensional nite element modeling of
composite girder bridges. J Eng Struct 2006;28:6371.
[21] Barth K, Wu H. Efcient nonlinear nite element modeling of slab on steel
stringer bridges. J Finite Elem Anal Des 2006;42:130413.
[22] Queiroz FD, Vellasco PCGS, Nethercot DA. Finite element modeling of
composite beams with full and partial shear connection. J Constr Steel Res
2007;63:50521.
[23] Zhao G, Li A. Numerical study of a bonded steel and concrete composite beam.
Comput Struct 2008;86(1920):18308.
[24] Nguyen Q, Hijaj M, Uy B, Guezouli. Analysis of composite beams in the hogging
moment regions using a mixed nite element formulation. J Constr Steel Res
2009;65(3):73748.
[25] Queiroz FD, Queiroz G, Nethercot DA. Two-dimensional FE model for
evaluation of composite beams I: Formulation and validation. J Constr Steel
Res 2009;65:105562.
[26] Tsavdaridis KD, DMello C, Huo BY. Computational study modelling the
experimental work conducted on the shear capacity of perforated concrete
steel ultra shallow oor beams (USFB). In: 16th Conference on concrete,
Paphos, Cyprus; 2009.
[27] Ansys Inc. Theory reference for the mechanical APDL and mechanical
applications. Canonsburg, PA: Ansys Incorporation; 2009a.
[28] Desayi P, Krishnan S. Equation for the stressstrain curve of concrete. ACI J
1964;61:34550.

Y. Majdi et al. / Engineering Structures 77 (2014) 6583

[29] ACI Committee 318. Building code requirements for structural concrete (ACI
318-05) and commentary (ACI 318R-05). Farmington Hills, MI: American
Concrete Institute; 2011.
[30] Ansys Inc. Element reference. Canonsburg, PA: Ansys Incorporation;
2009b.
[31] Majdi Y, Hsu CT, Punurai S. Local bond-slip behavior between cold-formed
metal and concrete. J Eng Struct 2014;69C:27184.
[32] Majdi Y. Finite element analysis of composite oors with cold-formed steel
and concrete slab. Newark, NJ: New Jersey Institute of Technology, PhD
Dissertation; 2013.

83

[33] Ansys Inc. Contact technology guide. Canonsburg, PA: Ansys Incorporation;
2009c.
[34] Alfano G, Criseld M. Finite element interface models for the delamination
analysis of laminated composites: mechanical and computational issues. Int J
Numer Meth Eng 2001;50:170136.
[35] Majdi Y, Hsu CTT. Technical manual for analysis and design of composite oors
with cold-formed steel and concrete slab having continuous furring
channel. Newark, NJ: New Jersey Institute of Technology; 2011.
[36] AISI. Cold-formed steel design. Washington, DC: American Iron and Steel
Institute; 2002.