Thin-Walled Structures 104 (2016) 211–224

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Full length article

Partial interaction theory to analyze composite (steel–concrete) shear

wall systems under pure out-of-plane loadings
Saeid Sabouri-Ghomi a, Younes Jahani b, Anjan K. Bhowmick c,n
a
Civil Engineering Department, K.N. Toosi University of Technology, Tehran, Iran
b
Civil Engineering, K.N. Toosi University of Technology, Tehran, Iran
c
Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, Quebec, Canada

art ic l e i nf o a b s t r a c t

Article history: This paper presents a closed form solution for the analysis and design of composite steel concrete (SC)
Received 9 November 2015 shear wall systems subjected to pure out-of-plane loads with partial interaction theory. This method
Received in revised form takes into account the flexibility of connection between plate and concrete. SC walls under out-of-plane
21 February 2016
loads can be considered a slab under distributed loads; therefore, for obtaining the formulation of these
Accepted 13 March 2016
systems, a strip beam-slab is considered. These walls are subjected to soil pressure when used in deep
excavations water hydraulic pressure when used as marine structures, and ice moving pressure when
Keywords: used in offshore structures etc. For providing the interaction between steel and concrete, shear con-
Partial interaction nectors and, for calculating the out-of-plane loadings, classic methods are utilized. The existence of
Composite shear wall
concrete in composite shear walls not only prevents the steel plate from buckling, but also plays an
Thin steel plate shear wall
important role in out-of-plane resistance. To investigate the effect of concrete on the behavior of these SC
Out-of-plane resistance
Buckling shear walls, two cases are considered in this study: shear walls with and without concrete. To validate
Finite element model the accuracy of the proposed method, a number of shear walls were modeled and analyzed by ABAQUS
software and compared with the results from theoretical formulations. Results indicate that the proposed
interaction theory is well capable of predicting the deformation and stress distribution of the composite
shear walls.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Steel plate shear wall (SPSWs) is an effective lateral resistance
system against earthquake and wind loads. When lateral loads are
applied to this wall, the thin infill plate buckles in shear while
developing the diagonal tension field action that resists the lateral
forces. Low weight, low construction costs, high construction
speed, high stiffness, and high capacity to dissipate the energy are
the advantages of these systems. For in-plane loading, many stu-
dies have been done in the past decades: the earliest study was
done by Thorburn et al. [1] that introduced replacing the plate
with the simple strip model to represent the behavior of the un-
stiffened SPSW system under in-plane loads. Roberts and Sabouri-
Ghomi [2] applied quasi-static loads to unstiffened steel plate
shear walls with an opening at the center of the plate. Driver and
Grondin [3], Lubell et al. [4], Romero [5], Seilie and Hooper [6],
n
Corresponding author.
E-mail addresses: sabouri@kntu.ac.ir (S. Sabouri-Ghomi),
younes.jahani@yahoo.com (Y. Jahani),
anjan.bhowmick@concordia.ca (A.K. Bhowmick).
Sabelli and Bruneau [7], and Bhowmick [8] have studied the
seismic behavior of steel plate shear walls. In another study, Sa-
bouri-Ghomi and Sajjadi [9] studied the behavior of SPSW systems
with and without stiffeners.
In the past decades, composite structural wall systems are used
in seismic zones as the primary lateral resistance system for
building structures. Composite constructions utilize exterior steel
plate, infill concrete, shear connector, and network of steel re-
inforcement. Optimum performance of the composite system de-
pends on transfer of stresses between concrete and steel. This
transfer is provided by shear connectors. SC composite walls
combine benefits of both reinforced concrete and steel plate shear
walls: reinforced concrete is inexpensive, massive, and stiff, while
steel walls are strong, lightweight, and easy to assemble. The op-
timal combination of the properties of the two materials, steel and
concrete, makes the structures safe and economical and the steel
can improve the concrete behavior in tension. Saari et al. [10]
studied the behavior of the shear stud in composite shear walls
under seismic loads. Link and Elwi [11] presented the capacity of
composite walls under the transverse and longitudinal loads due
to moving ices in the offshore structures. Hajjar [12] summarized

http://dx.doi.org/10.1016/j.tws.2016.03.013
0263-8231/& 2016 Elsevier Ltd. All rights reserved.
212 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224
Nomenclature
γ slip between the layers
Q load on one shear connector
q shear flow
s space between shear connectors
n number of shear connectors
k stiffness of shear connectors
K the factor to control the slope of the concrete stress–
strain curve
F force between layers
εct concrete strain in tension region
εpc steel plate strain in compression region
M bending moment; subscripts c for concrete and p for
steel plate
E modules of elasticity; subscripts c for concrete and p
for steel
I second moment of area, subscripts c for concrete and
the recent research on a composite lateral resistance system.
Shanmugam et al. [13] investigated the ultimate load behavior of
double skin composite (DSC) slabs. Some experimental tests were
done and compared with finite element results. In the other study,
Tong et al. [14] presented an experimental study of the cyclic be-
havior of a composite structural system; the one-bay, two-storey
test specimen was built in the one-third scale. Zhao and Astaneh-
Asl [15] used a concrete cover on the steel plate to increase the
stiffness and enhance the buckling behavior of the plate. Vas-
dravellis et al. [16] investigated the behavior of steel–concrete
composite beams subjected to the combined effects of negative
bending and axial compression. Rafiei et al. [17] presented the
development of experimental and finite element models to si-
mulate the behavior of a novel composite shear wall system con-
sisting of two skins of steel and an infill of concrete under in-plane
loadings. Varma et al. [18] presented the finite element model for
predicting the behavior and failure of composite walls subjected to
a combination of in-plane forces and moments. Sener and Varma
[19] found the out-of-plane shear strength of composite walls.
They indicated that the steel faceplates had a minor influence on
the out-of-plane resistance, but they had a major influence on the
out-of-plane failure mode. Zhang et al. [20] found the effects of
shear connectors on local buckling and level of composite action in
steel concrete composite walls. The results based on experimental
and numerical database indicated that when the normalized shear
connector spacing-to-plate thickness ratio, (s/tp × Fy/E ) is less
than 1.0, yielding in steel plate occurs before local buckling.
It has been observed that most of the past studies on SPSWs
and composite shear walls focused mainly on in-plane loads. The
objective of the present work is to present the out-of-plane re-
sistance of SC composite shear wall systems. A closed form solu-
tion to analyze and design simply supported composite steel-
concrete (SC) shear wall systems subjected to out-of-plane loads is
presented by considering the partial interaction theory. The out-
of-plane loads considered in this study are water hydraulic pres-
sure, soil pressure, and ice pressure on offshore structures that
could act on a structure as a static load. Under these loads, the
walls are bent and deformation occurred in plates. For limiting this
deformation, composite shear walls are used. Wright et al. [21–23]
conducted experimental studies of simply supported double skin
composite beams and also presented a closed form solution for
analysis of these composite beams. Dogan and Roberts [24] com-
pared experimental deformations of steel-concrete-steel sandwich
beams with full and partial interaction theories. Xie et al. [25]
p for steel plate
A area; subscripts c for concrete and p for steel plate
t thickness; subscripts c for concrete and p for steel
plate
ht total depth of beam section
W distributed load on beam per unit length
y vertical deflection of the beam
σ′cu ultimate strength of the concrete in compression
σ′tu ultimate strength of the concrete in tension
σ′el elastic limit of the concrete
σc strength of the concrete in compression
σcc existing stresses in the concrete in compression region
σct existing stresses in the concrete in tension region
σp existing stresses in the steel plate
b length of the composite wall
h width of the composite wall (span of beam)
dss shear connector diameter
presented experimental and numerical studies on the static be-
havior of the friction-welded connections with the bar loaded in
shear. In the other study, Xie et al. [26] presented an experimental
investigation on the static behavior of steel-concrete-steel beams.
The objective of this study is to develop theoretical formula-
tions for analysis and design of SC composite shear walls under
out-of-plane loading. This formulation obtained by the partial in-
teraction theory considered flexibility between plate and concrete.
To validate the proposed theoretical formulation, finite element
models were developed and results from FE analyses were com-
pared with the theoretical relations.
2. General theory
In this section a closed form solution for analysis and design of
steel-concrete (SC) composite shear wall systems subjected to out-
of-plane loads is presented. In the first step, theoretical formula-
tions are obtained from simple bending and partial interaction
theory and, then in the second step, the results are compared with
those from finite element analyses. Fig. 1 presents the general
form of the problem, to which out-of-plane loads are applied. This
type of loads is applied from soil pressure in deep excavations in
high-rise buildings, water hydraulic pressure on marine structures,
ice moving pressure in offshore structures, etc. In the deep ex-
cavations, structural walls are subjected to the soil pressure;
therefore, these systems must be designed for these kinds of loads.
For calculating the static and dynamic pressures of the soil, Cou-
lomb [27] and Mononobe-Okabe [28,29] methods are utilized re-
spectively. Furthermore, ice actions on offshore structures are
among the main concerns for engineering activities in cold areas
with ice-infested waters and can be categorized as out-of-plane
loads. Løset et al. [30–32] studied the effect of these kinds of load
on the offshore structures.
In the present study, because of rigid diaphragms, the beams
are not considered and simple supports are assumed for con-
sidering the critical situations; then, the walls act as a one-way
slab with cylindrical deformation.
2.1. Partial interaction theory to design the composite shear walls
For this study, for modeling of the interactions in the interface
of the layers of SC composite shear walls, partial interaction theory
is utilized. In this method, theoretical relations are obtained by
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224 213

Fig. 1. Details of composite shear wall.

Fig. 2. (a) Slab with simple supports, (b) details of beam-slab strip model, and (c) section along the beam.
214 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224

considering the flexibility of the connection between plate and ⎛ 1 M ⎛t ⎞

s d2F 1 ⎞ M ⎛t ⎞
concrete. When the loads are applied to these walls, the slip takes = F ⎜⎜ + ⎟⎟ − c ⎜ c ⎟ − p ⎜⎜ p ⎟⎟
nk dx 2
⎝ Ep A p ⎝
EcAc ⎠ EcIc 2 ⎠ EpIp ⎝ 2 ⎠
place between the layers. The amount of this slip depends on the (7)
shear stiffness of the connections. Because of the direction of loads
Taking moments around the centroid of the concrete leads to:
in Fig. 1, the shear walls act as a slab with distributed load. For
designing the slabs, a beam-slab strip is considered and, then, the M = Mc + Mp + Fdm (8)
results are extended to the other parts. A strip with 1000 mm
width is assumed. Therefore, a flexural beam with 1000 mm width Using the assumption that all the layers are bent to the same
and simple supports was designed. The height of this beam was curvature, it is also possible to write:
equal to the wall thickness. The details are shown in Fig. 2.
d2y M Mp M Mp
In this section, the aim is to obtain the formulation for calcu- − 2
= c = = c +
dx E I
cc E I
pp E I
cc E pIp (9)
lating the slip between layers, stresses, and deformation along the
composite beams, as shown in Fig. 2. Following assumptions are Substituting Eq. (8) in Eq. (9) gives:
considered:
d2y M Mp M − Fdm
− = c = =
(1) The materials are elastic. dx2 EcIc EpIp ∑ EI (10)
(2) Deflections are small.
(3) The shear connection between the layers is assumed to act as where
a continuous connection.
(4) The distribution of strains through the depth of layers is linear.
∑ EI = EcIc + EpIp (11)
(5) In every section of the beam, each layer is bent to the same and
radius of curvature and no buckling or separation of the layers
occurs. tc + tp
dm =
(6) No friction between the layers is considered. 2 (12)
(7) The concrete stays uncracked in both compression and tension
For distributed load on simple supported beam, the moment
regions (this assumption is considered because of the execu-
equation is obtained by:
tive problems and the impossibility of replacing cracked con-
crete; therefore, the thickness of the concrete must be con- wx
M= (h − x)
sidered large enough in order to stay uncracked). 2 (13)
(8) For simplicity, the steel reinforcements are not considered in
where x is the distance measured from the left support to the
the design process and they are only assumed as thermal steel
differential element in Fig. 2(b).
reinforcements.
Substituting (Eqs. (10) and 12) in Eq. (7) gives:
The theoretical relations are obtained through the following
s d2F ⎛ 1 1 d2 ⎞ ⎛ d ⎞
steps: 2
= F ⎜⎜ + + m ⎟⎟ − M ⎜ m ⎟
nk dx E A
⎝ p p E A
c c ∑ EI ⎠ ⎝ ∑ EI ⎠ (14)
The slip between the layers (concrete and steel plate) is ob-
tained as After simplification, Eq. (14) can be written as follows:
Q qs 2
γ= = dF
k nk (1) − A1F = − A2 M
dx2 (15)
The shear flow is equal to the rate of force change between
where
layers:
dF nk ⎛ 1 1 d2 ⎞
q= − A1 = ⎜⎜ + + m ⎟⎟
dx (2) s ⎝ Ep A p EcAc ∑ EI ⎠ (16)
Eq. (2) can be substituted in Eq. (1) and provide: and
s dF
γ=−
(3) nk ⎛ dm ⎞
nk dx A2 = ⎜ ⎟
s ⎝ ∑ EI ⎠ (17)
The rate of change of slip equals the strain difference at the
interface of concrete and steel plate: Using of Eq. (13) and considering the distributed load on the
beam and simple supported condition, Eq. (15) is solved as:
dγ s d2F
=− = εct −εpc ⎛ ⎞⎡
dx nk dx2 (4) ⎤
⎜ A2 ⎟⎢ e A1 x + e A1 (h − x) ⎛ x2 hx 1 ⎞⎥
These strains at the interface of the layers can be obtained by a F = W ⎜ ⎟⎢ −⎜ − − ⎟⎥
force and a moment that act on the centroid of each of the layers, ⎝ ⎠⎣ (
⎜ A1 ⎟⎢ A1 1 + e h A1 ) ⎝ 2 2 A1 ⎠⎥
⎦ (18)
as shown in Fig. 2. The strains are derived from simple bending
theory as the following equation: with differentiation from Eq. (18) and substitution in Eq. (3),
the slip between the layers is obtained as follows:
F M ⎛t ⎞
εct = − + c ⎜ c⎟ ⎛ ⎞
Ec Ac EcIc ⎝ 2 ⎠ (5) ⎛ sA w ⎞⎜ e A1 x − e A1 (h − x) h⎟
γ = ⎜ 2 ⎟⎜ x − − ⎟
⎝ nkA1 ⎠⎜
εpc =
F
−
Mp ⎛ tp ⎞
⎜ ⎟
⎝ (
A1 1 + e h A1 )
2⎟
⎠ (19)
Ep A p E p Ip ⎝ 2 ⎠ (6)
For calculating the deformed shape of the beam, simple
(Eqs. (5) and 6) are substituted in Eq. (4) as follows: bending theory is utilized as follows:
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224 215

d2y M − Fdm Step 1: for given values of parameters such as concrete cover
− = thickness, steel plate thickness, shear connector diameter, shear
dx2 ∑ EI (20)
connector spacing, stiffness of shear connectors, length of span,
Substituting (Eqs. (13) and 18) in Eq. (20) gives: material property and distributed loads, strains in both concrete
and steel plate are calculated using (Eqs. (5) and 6). These
⎛ ⎞
⎜ A2 dm ⎟ parameters can be selected based on recommendations pro-
d2y Fdm − M
= = w⎜ ⎟ vided in steel standard, AISC 341-10 [33] or other published
dx2 ∑ EI ⎜ A1 ∑ EI ⎟
⎝ ⎠ literature [19,20,34]. For calculating the strains, moments in
⎡ ⎤ concrete ( Mc ) and in steel plate ( Mp ) and section force ( F ) are
⎢ e A1 x + e A1 (h − x) ⎛ x2 hx 1 ⎞⎥ wx required.
⎢ − ⎜ − − ⎟⎥− (h − x) Step 2: the moments in steel plate, Mp , and concrete cover, Mc ,
⎣⎢ A1 1 + e (
h A1 ⎝ 2 2
) A1 ⎠
⎦⎥
2 ∑ EI
(21) are obtained from Eq. (10).
Step 3: the value of section force, F , in the steel plate and
With double integration from Eq. (21), the deformed shape of concrete cover is obtained from Eq. (18).
the beam is obtained as follows: Step 4: once the strain values at the interface of the layers are
⎛ ⎞⎡ ⎤ obtained from Eqs. (5) and (6), the stress distribution in height
⎜ A d ⎟⎢ e A1 x + e A1 (h − x) ⎛ x4 hx 3 1 2⎞⎥ of the section can be calculated by multiplying the strain values
y = w ⎜ 2 m ⎟⎢ −⎜ − − x ⎟⎥ with modulus of elasticity of steel and concrete. The difference
⎜ A1 ∑ EI ⎟⎢ A12 1 + e h A1
⎝ ⎠⎣ (
⎝ 24 12 2A1 ⎠⎥
)
⎦ of the strain at the interface of the layer should be distributed
⎛ ⎞ between the layers, and then the strain in the extreme fiber in
w ⎜ hx 3 x4 ⎟ the compression region of concrete and in the tension region of
− ⎜ − ⎟ + c1x + c2 the steel plate should be corrected accordingly.
2 ∑ EI ⎜ 6 12 ⎟
⎝ ⎠ (22) Step 5: deformation of specimen along the length of the span is
then calculated from Eq. (25).
where c1 and c2 are the integration constants and obtained by
Step 6: steps 1–5 are repeated until all the assumptions are
boundary conditions. For simply supported beam and distributed
satisfied.
load, the constants are derived as follows:

wdmA2
if x= 0 → y = 0 → c2 = −
A13 ∑ EI (23) 3. Finite element analysis of SC composite walls

Nonlinear finite element analysis of steel-concrete composite

wh3 ⎛ A1 − dmA2 ⎞ whdmA2
walls subjected to out-of-plane loads were carried out to verify the
if x= h → y = 0 → c1 = ⎜ ⎟−
24 ∑ EI ⎝ A1 ⎠ 2A12 ∑ EI (24) results obtained from the theoretical relations. In total, fourteen
Substituting (Eqs. (23) and 24) in Eq. (22) gives the final (14) specimens were considered, twelve of which had a concrete
cover attached with steel plate and, for the other two, concrete
equation of beam deformation shape:
cover was removed to show its effect on the behavior of the walls
wA2 dm under out-of-plane loads. In all the models, loading and boundary
y= (e A1 x
+e A1 (h − x)
)
A13 ( 1 + e ) ∑ EI
h A1 conditions were constant and other parameters such as concrete
cover thickness, steel plate thickness and space between shear
A1 − A2 dm connectors were changed. Details of the specimens are indicated
+
24A1 ∑ EI
(
wh3x − 2whx 3 + wx 4 ) in Table 1 and also Fig. 3 presents the column details.
A2 dm wA2 dm In the present study, the magnitude of out-of-plane load, which
+
2A12 ∑ EI
( wx 2
− whx − ) A13 ∑ EI (25) was calculated from soil pressure and was assumed about
0.27 (N/mm2) for this study, was applied as a distributed load on
In working with the proposed theoretical formulation for de- the walls. For the first eight specimens, S1, S2, S3, S4, S5, S6, S7 and
sign and analysis of SC composite walls following steps shall be S8, listed in Table 1, only the concrete thickness varied and all
considered: other parameters were assumed to remain same. For the next

Table 1
Summary of details of specimens.

Specimen Column b (mm) h (mm) tp (mm) σ′cu (Mpa) tc (mm) s (mm) dss (mm) tp Optimal mesh size (mm)
ρ=
tp + tc

S1 H Shape 5700 3000 8 280 180 20 0.028 50

S2 290 0.027
S3 300 0.026
S4 310 0.025
S5 320 0.024
S6 330 0.0236
S7 340 0.0229
S8 30 350 0.022
S9 10 300 0.032
S10 12 300 0.038
S11 14 300 0.045
S12 8 330 225 0.0236
S13 8 – – – – –
S14 15 – – – – –
216 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224
Fig. 3. Details of column sections.
three specimens, S9, S10 and S11, only the steel plate thickness
was changed and in specimen S12, the space between shear con-
nector was varied. Sener and Varma [19] and Kai et al. [20] pre-
tp
viously used steel plate reinforcement ratios ( ρ =
t p + tc
1.5% and 5%. As observed from Table 1, steel plate reinforcement
ratios calculated for all the selected specimens were between 2.2%
and 4.5%. AISC 341-10 [33] recommends use of a minimum of
200 mm of concrete cover when concrete is provided on one side
of the steel plate of composite plate wall. Thus, when selecting
concrete cover thickness this minimum requirement was main-
tained. The concrete thickness was calculated by the theoretical
relation presented in section two, and it was observed when the
concrete thickness was equal or greater than 280 mm, all the as-
sumptions were satisfied. In calculating the thickness of concrete,
it was assumed that the concrete had a linear behavior, and only
elastic limit of the concrete in compression and ultimate strength
of concrete in tension were checked. Thus, in each step, the stress
in the top face and bottom face of the concrete was calculated with
the theoretical relations obtained in Section 2, and when these
stresses were less than the limit values, the selection thickness
was acceptable. The limit values were elastic limit of the concrete
in compression, σ′el , and ultimate strength of concrete in tension,
σ′tu , as shown in Table 2. Therefore, for comparison between the-
oretical relation and FE models, tension behavior of the concrete
was not modeled in finite element simulations. Johnson [34] re-
commends use of shear connector diameter between 13 mm and
25 mm for composite structures. In this study, a shear connector
diameter of 20 mm is used for all the selected specimens. In ad-
dition, Johnson [34] recommends the following limit for shear
connector diameter ( dss ) to steel plate thickness ( tp ) ratio:
Table 2
Details of concrete material in the present study.
σ′cu (MPa) σ ′el = 0.4σ ′cu (MPa) m Ec (MPa)
30 12 2.56
dss
< ( 2.5~2.7)
tp (26)
For this research, steel plate thicknesses were selected so that
shear connector diameter to steel plate thickness ratio for all the
specimens remained lower or equal to 2.5.
3.1. Development of finite element model
Finite element software, ABAQUS [35], was used for the simu-
lation of the selected specimens. Finite element models included
two columns, a steel plate sheet, concrete, and shear connectors.
Columns and steel plate were modeled using linear quadrilateral
shell element (S4R), concrete was modeled by linear hexahedral
solid element (C3D8R), and shear connectors were modeled by
beam elements (B31). When selecting FE mesh, it is important to
choose appropriate mesh size that will yield accurate FE analysis
results while reduce computational time as much as possible. A
mesh convergence analysis was conducted to select optimum
mesh sizes for columns, steel plate and concrete cover. In this
study, it was observed that the meshing refinement was not sen-
sitive after having the maximum mesh size of less than 50 mm for
shell and solid elements and 10 mm for beam elements. Fig. 4
shows the 3D-view of SC composite wall simulation and also, Fig. 5
indicates the finite element mesh of composite walls. As men-
tioned previously, for investigating the effects of concrete on the
out-of-plane behavior of the walls, two specimens were modeled
without concrete and shear connectors. Fig. 6 illustrates the finite
element model of steel plate shear wall.
The connections between the column and plate and also shear
) between connectors and plate were provided by tie constraint in ABAQUS.
When the weld metal was considered the same as the steel ma-
terial, one could use the tie constraint. The interaction between
plate and columns with concrete was considered and tangential
behavior (frictionless) and hard contact was used for interaction.
Shear connectors were embedded in concrete to improve the
concrete behavior and reduce the concrete cracks.
In this study, rigid diaphragm was assumed and for this reason,
the beams were not modeled. This assumption was based on the
fact that the floor was long and thus the longitudinal displacement
of the floor could be omitted (Fig. 1). Also, to obtain maximum
deformations at the mid span, simple supports were used for
boundary conditions (Figs. 1 and 4).
3.2. Materials behavior
3.2.1. Concrete
In the past decades, many researchers have made valuable
contributions to understanding the behavior of concrete and de-
veloped methods for the simulation of concrete to use in finite
element software. Due to non-linear stress–strain relation under
multi-axial stress conditions, strain softening, and anisotropic
stiffness reduction as well as time dependent behavior and pro-
gressive cracking caused by tensile stresses, concrete behavior is
more complicated than steel materials [36]. The behavior of con-
crete in the compression is linear up to 30 percent of the ultimate
compressive strength of concrete [36] and, generally, the behavior
ε0 K for (εc /ε0) > 1 σ′tu (MPa)
27386 0.001796 1.15 3.83
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224 217

Fig. 4. Finite element model of composite shear wall.

Fig. 5. Finite element mesh of composite shear wall.

Popovics [37] and Yip [38] to simulate the concrete behavior in

compression. In this method, the value of compressive concrete
stress ( σcu ) is a function of concrete strain (εc ) that is obtained from
the following equation:
σc m(εc /ε0)
=
σ ′cu m − 1 + (εc /ε0)mk (27)

where σ′cu is the ultimate compressive strength of concrete, m is a

curve-fitting factor (equal to m = 0.8 + σ ′cu /17), Ec is the initial
tangent modulus (can be taken as Ec = 5000 σ ′cu ), ε0 is strain
when σc reaches σ′cu ; ε0 = (σ ′cu /Ec )(m /(m − 1)), and K is the factor to
control the slope of the stress–strain curve in ascending and
descending branches [38]:
K = 1.0 if (εc /ε0) ≤ 1

K = 0.67 + σ ′cu /62≥1.0 if (εc /ε0) > 1 (28)

Fig. 7 presents the behavior of the concrete in compression.

Table 2 shows the details of concrete material used in this study.
Fig. 6. Finite element model of steel plate shear wall.

3.2.2. Steel
of concrete is considered linear up to 30–40 percent of the ulti- Behavior of steel materials, unlike concrete, is known and be-
mate compressive strength. In this study, concrete behavior in cause of convenience, many tests have been done on it. For si-
compression was based on the numerical method developed by mulating the steel behavior in ABAQUS software, true stress and
218 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224
Fig. 7. Compression behavior of concrete.
Fig. 8. Stress–strain curve for mild steel.
logarithmic plastic strain are used. The relationships between true
stress and engineering stress as well as logarithmic plastic strain
and engineering plastic strain are obtained from the following
equations [17]:
σtrue = σnom(1 + εnom)
and
ε lnpl = ln(1 + εnom) − σtrue/Ep
In this study, two types of steel materials are used. The mild
strength steel is used for column and steel plates. The high
strength steel is used for shear connectors. Fig. 8 and Fig. 9 present
the steel stress-strain relation for two types of materials. Table 3
presents the steel material details used in this study.
3.3. Validation of FE model
The finite element model developed was validated by com-
paring published test results with the corresponding FE analysis
results. A single storey SPSW specimen tested by Lubell et al.
(2000) and a steel–concrete–steel sandwich beam tested by Xie
et al. (2007) were selected for the validation. Pushover analyses
were carried out and the resulting curves are compared with re-
sults from the tests.
3.3.1. Lubell et al. (2000) SPSW specimen
One single storey SPSW specimen (designated as SPSW2 in the
literature [4]) tested by Lubell et al. (2000) is modeled in ABAQUS
following the procedure outlined above. Details of the test speci-
men are available in the literature [4]. As published in the litera-
ture [4], a yield strength of 380 MPa was selected for the beams
and columns and for the infill plate, yield strength was taken as
Fig. 9. Stress–strain curve for high strength steel.
Table 3
Steel material details in present study.
Steel type Nominal yield True yield Nominal ulti- True ultimate
stress stress mate stress stress
(MPa) (MPa) (MPa) (MPa)
High strength 400 400.76 600 690
steel
Mild strength 235 235.45 370 425.5
steel
320 MPa. A monotonic pushover analysis was conducted with a
lateral load applied along the center-line of the top beam. FE mesh
of the single storey SPSW is shown in Fig. 10(a). Fig. 10
(b) compares the pushover analysis results with the results ob-
tained from the test. It is observed that the FE model agrees very
well with the test results. The initial stiffness was predicted very
well and the ultimate strength was underestimated by only 3%.
3.3.2. Xie et al. (2007) steel–concrete–steel sandwich beam
specimen
(29)
Since the purpose of the present work is to present the ana-
lytical model to analyze composite SC walls under pure out-of-
plane loads, the accuracy of the finite element model is further
(30) validated for a composite steel-concrete-steel sandwich beam
tested by Xie et al. [26]. In total 18 specimens were tested by Xie
et al. [26]. Different parameters such as number of shear con-
nectors, steel and concrete thickness, material strengths, and also
beam span were varied among the specimens. Details of the spe-
cimens are summarized in the literature [26]. One of the speci-
mens (designated as BS7 in the literature [26]) is selected and si-
mulated in ABAQUS software. The load is applied with two 100 t
static jacks in mid-span of the specimen. The concrete is attached
to steel plate with shear connectors. Experimental load versus
mid-span deflection relation of BS7 specimen is compared with
the load-deflection relation obtained from the FE model. The FE
mesh of BS7 specimen is illustrated in Fig. 11(a). Fig. 11(b) shows
that the FE model is in good agreement with the experimental
force deflection curve of specimen BS7.
4. Analysis of results
The selected twelve SC walls (S1–S12 as presented in Table 1)
were analyzed using the validated FE model. Results from the FE
analyses were compared with the results obtained from theore-
tical relation. In this study, the deformation along the length of the
span of the specimens was plotted. As mentioned in Section two,
concrete must stay in the linear region. Therefore, the thickness of
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224 219

Fig. 10. Validation of Lubell et al. (2000) specimen: (a) FE mesh; (b) Pushover curves.

Fig. 11. Validation of finite element model for Xie et al. (2007): (a) FE mesh; (b) Load versus deflection curves.

Fig. 12. Comparing the finite element model and theoretical relations for calculating the deformation of S1.

Fig. 13. Comparing the finite element model and theoretical relation for calculating the deformation of S2.
220 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224

Fig. 14. Comparing the finite element model and theoretical relation for calculating the deformation of S3.

Fig. 15. Comparing the finite element model and theoretical relation for calculating the deformation of S4.

Fig. 16. Comparing the finite element model and theoretical relation for calculating the deformation of S5.

Fig. 17. Comparing the finite element model and theoretical relation for calculating the deformation of S6.

Fig. 18. Comparing the finite element model and theoretical relation for calculating the deformation of S7.

Fig. 19. Comparing the finite element model and theoretical relation for calculating the deformation of S8.

the concrete, which stays linear in compression and is not cracked relation in Section two and it was equal to 280 mm which satisfied
in the tension, is acceptable. For out-of-plane load of 0.27 (N/mm2), all the assumptions. Figs 12–23 indicate the deformations of the
the concrete cover thickness was obtained from theoretical SC composite walls along the length of the span. For better
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224 221

Fig. 20. Comparing the finite element model and theoretical relation for calculating the deformation of S9.

Fig. 21. Comparing the finite element model and theoretical relation for calculating the deformation of S10.

Fig. 22. Comparing the finite element model and theoretical relation for calculating the deformation of S11.

Fig. 23. Comparing the finite element model and theoretical relation for calculating the deformation of S12.

Table 4
2.35% for specimen S12. For all the specimens (S1–S12), the aver-
Mid-span displacement of the specimens. age difference in mid span deformation between FE analysis and
partial interaction theory is about 3.5 percent. The small difference
Specimen Displacement (mm) between the results of finite element and theoretical relations was
Finite element Partial interaction theory Difference (percent)
because of consideration of the beam-slab strip instead of the full
SC wall.
S1 2.34 2.41 2.99 Fig. 24 presents the comparison of the results of finite element
S2 2.14 2.21 3.27 and theoretical relations for first eight specimens (S1–S8). It is
S3 1.97 2.03 3.05
observed that with increasing the thickness of the concrete, mid-
S4 1.82 1.87 2.74
S5 1.68 1.73 2.97 span deformation is decreased. Thus, for specimen S8, which had a
S6 1.55 1.6 3.23 concrete cover of 350 mm, the maximum out-of-plane deforma-
S7 1.44 1.48 2.77 tion (from FE analysis) at mid-span was reduced by about 43% than
S8 1.34 1.38 2.98
the maximum out-of-plane deformation for specimen S1, which
S9 1.89 1.83 3.17
S10 1.8 1.72 4.44 had concrete thickness of 280 mm.
S11 1.76 1.62 7.95 For specimens S3, S9, S10 and S11, only the steel plate thickness
S12 1.7 1.66 2.35 was varied and other parameters such as shear connector spacing
and concrete cover remained same. Maximum out-of-plane de-
formations for these four specimens were compared in Fig. 25. It is
representing the difference between theoretical and finite element observed that with an increase in steel plate thickness, mid-span
results, the vertical axes were plotted on a larger scale. It is ob- deformation of the SC composite wall decreases. Thus, for speci-
served that the proposed theoretical formulation is well capable of men S11, which has a steel plate thickness of 14 mm, the max-
predicting the deformation of the walls. Table 4 presents the va- imum out-of-plane deformation at mid-span is reduced by about
lues of the deformations of the selected walls. As shown in Table 4, 11% in compare to specimen S3, which has a steel plate thickness
the maximum difference in mid span deformation between the FE of 8 mm. Fig. 25 also shows that the results from FE analyses are in
analysis and the proposed partial interaction theory is equal to excellent agreement with the results from theoretical relations.
7.95% for specimen S11, and the minimum difference is equal to In addition to concrete thickness and steel plate thickness
222 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224

Fig. 24. Comparing the finite element model and theoretical relation for calculating the deformation of specimens S1–S8 with different concrete cover.

Table 5
Stress distribution in mid-span of the specimens.

Specimen Stress (MPa)

Finite element Partial interaction theory

σp σcc σct σp σcc σct

S1 32.13 11.21 3.72 33.85 11.97 3.71

S2 30.74 10.49 3.57 32 11.3 3.49
S3 29.43 9.85 3.43 30.36 10.69 3.3
S4 28.11 9.26 3.29 28.81 10.13 3.12
S5 27.18 8.73 3.17 27.38 9.62 2.96
S6 25.96 8.37 3.2 26.06 9.14 2.82
S7 24.97 7.91 3.05 24.83 8.69 2.68
S8 24 7.49 2.97 23.68 8.28 2.55
S9 24.73 9.67 2.97 27.6 9.85 3.07
S10 21.16 9.53 2.63 25.8 9.12 2.82
Fig. 25. Comparing the finite element model and theoretical relation for calculating S11 18.32 9.42 2.37 24.1 8.45 2.6
the deformation of specimens S3, S9, S10 and S11 with different steel plate S12 26.5 8.6 3.81 28 9.15 2.7
thickness.

SC specimen increases. From the results of FE analysis, it was ob-

Fig. 26. Comparing the finite element model and theoretical relation for calculating
the deformation of specimens S6 and S12 with different shear connector spacing.
another parameter, shear connector spacing was studied. In speci-
men S12, the space between shear connector was changed to
225 mm. Results from both FE analysis and theoretical relations
were compared with results for specimen S6, which had shear
connector spacing of 180 mm. It is observed from Fig. 26 that for
both finite element and theoretical relations, with an increase in the
space between the shear connectors, mid-span deformation of the
served that by increasing the shear connector spacing from 180 mm
to 225 mm, the mid span deformation was increased about 9.7%.
As mentioned earlier, the assumption of the theoretical rela-
tions was based on the linear behavior of the materials. Also, it
was assumed that the concrete had to be in the linear region in
both compression and tension areas. For investigating the accuracy
of these assumptions, the maximum stress in the mid-span of the
selected specimens (S1–S12) was obtained and compared with
finite element models in Table 5. It is observed from Table 5 that
the maximum stresses in the steel plate, σp , in the concrete in
compression region, σcc , and in the concrete in tension region, σct ,
obtained from FE models are in excellent agreement with the
stresses obtained from Partial interaction theory. As shown in
Table 5, for stress distribution in steel plate, the maximum dif-
ference between FE analysis and theoretical relation is related to
S11 and is equal to 5.78 MPa, and the minimum difference is re-
lated to S6, which is equal 0.1 MPa. Also, the maximum and
minimum difference in compressive stress in concrete, are related
to S11 and S9, and equal to 0.97 and 0.18 MPa, respectively. Fur-
thermore, the S12 specimen has a maximum difference in stress
for concrete in tension (equal to 1.11 MPa) and the specimen, S1,
has a minimum difference (equal to 0.01 MPa). This differences
between FE analysis and theoretical relations are due to some
assumption that were considered in obtaining the theoretical
 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224
Fig. 27. Comparing the finite element results of specimens; S1, S3 and S6 with concrete cover, S13 and S14 without concrete cover.
relations, such as considering the full SC wall instead of 1000 mm
beam-slab strip in FE analysis.
The limit value for the steel plate is the nominal yield stress in
steel material, which is equal to 235 MPa and the limit values for
concrete cover in compression and tension are the ultimate
strength in compression and ultimate strength in tension, which
are equals to σ ′el = 0.4σ ′cu = 12MPa and σ ′tu = 0.7 σcu = 3.83MPa ,
respectively. The maximum stresses of the specimens at mid span
were obtained from both FE analysis and theoretical relations and
compared with limit values. The specimens which had stresses
less than these limit values were accepted. It is observed from
Table 5 that all the stresses are less than the limit values.
For investigating the effect of concrete in out-of-plane de-
formation, two specimens (S13 and S14) were simulated without
concrete. Boundary condition, load magnitude, and other details
were same as the previous specimens and only the concrete and
shear connectors were removed. The deformations of specimens
S1, S3 and S6 (with concrete cover) and S13 and S14 (without
concrete cover) along the span of the SC composite walls are
presented in Fig. 27. It is observed from Fig. 27 that for specimen
13, which does not have any concrete cover and has a steel plate of
thickness of 8 mm, out-of-plane deformation is the highest in
compare to the other specimens. Thus, concrete cover can sig-
nificantly decrease the out-of-plane deformation magnitude of SC
composite walls. Also, FE analysis of S13 and S14 specimens shows
that by increasing the steel plate thickness, mid-span deformation
can be decreased. Based on the observations from results of ana-
lysis of all the selected specimens, it would be practical to increase
the concrete cover or to increase the steel plate thickness to limit
the out-of-plane deformation of SC composite walls when sub-
jected to out-of-plane loadings.
5. Conclusion
This paper presents a closed form solution for the analysis and
design of composite (steel-concrete) shear wall systems subjected to
pure out-of-plane loads with partial interaction theory. The theore-
tical formulations were obtained by considering the flexibility of
connection between plate and concrete. To validate the accuracy of
the proposed theoretical formulation, 12 SC walls were modeled and
analyzed using ABAQUS software. The results from the FE models
were compared with the theoretical relations. The results indicated
that the proposed closed form solution had good capability for pre-
dicting the deformation and stress distribution of the composite
shear walls. The small difference between theoretical relations and
finite element results was due to the consideration of the beam-slab
strip, instead of the complete shape of the specimens. Also, the re-
sults demonstrated that both steel plate and concrete cover are ef-
fective in limiting the out-of-plane deformation and with removal of
the concrete cover, out-of-plane deformation increased significantly.
In addition, shear connector spacing was found to be an important
223
parameter that affects the behavior of SC composite walls and the
out-of-plane deformation of SC composite walls increased with an
increase in shear connector spacing.
References
[1] L.J. Thorburn, G.L. Kulak, C.J. Montgomery, Analysis of steel plate shear walls,
Structural Engineering Report no. 107, Edmonton, AB: Department of Civil En-
gineering, University of Alberta, 1983.
[2] T.M. Roberts, S. Sabouri-Ghomi, Hysteretic characteristics of unstiffened perforated
steel plate shear panels, Thin Walled Struct. 14 (1992) 139–151.
[3] R.G. Driver, G.Y. Grondin, Steel plate shear walls: now performing on the main stage,
Mod. steel Constr. 41 (9) (2001) 58–59.
[4] A.S. Lubell, H.G.L. Prion, C.E. Ventura, M. Rezai, Unstiffened steel plate shear wall
performance under cyclic load, J. Struct. Eng. 126 (4) (2000) 453–460.
[5] E.M. Romero, Construcción compuesta en edificios altos, in: Proceedings of the VII
Simposio Internacional de Estructuras de Acero, Veracruz, Mexico, 2003.
[6] Ignasius F. Seilie, D. Hooper John, Steel plate shear walls: practical design and
construction, Mod. steel Constr. 45 (4) (2005) 37–43.
[7] R. Sabelli, M. Bruneau, Steel plate shear walls. Steel design guide, Chicago, Am. Inst.
Steel Constr. (2006).
[8] A.K. Bhowmick, Seismic behavior of steel plate shear walls with centrally placed
circular perforations, Thin Walled Struct. 75 (2014) 30–42.
[9] Saeid Sabouri-Ghomi, Seyed Ramin Asad Sajjadi, Experimental and theoretical
studies of steel shear walls with and without stiffeners, J. Constr. Steel Res. 75
(2012) 152–159.
[10] William K. Saari, Jerome F. Hajjar, Arturo E. Schultz, Carol K. Shield, Behavior of
shear studs in steel frames with reinforced concrete infill walls, J. Constr. Steel Res.
60 (10) (2004) 1453–1480.
[11] R.A. Link, A.E. Elwi, Composite concrete-steel plate walls: analysis and behavior,
J. Struct. Eng. 121 (2) (1995) 260–271.
[12] Jerome F. Hajjar, Composite steel and concrete structural systems for seismic en-
gineering, J. Constr. Steel Res. 58 (5) (2002) 703–723.
[13] N.E. Shanmugam, Ghanshyam Kumar, V. Thevendran, Finite element modelling of
double skin composite slabs, Finite Elem. Anal. Des. 38 (7) (2002) 579–599.
[14] Xiangdong Tong, Jerome F. Hajjar, Arturo E. Schultz, Carol K. Shield, Cyclic behavior
of steel frame structures with composite reinforced concrete infill walls and par-
tially-restrained connections, J. Constr. Steel Res. 61 (4) (2005) 531–552.
[15] Qiuhong Zhao, Abolhassan Astaneh-Asl, Cyclic behavior of traditional and in-
novative composite shear walls, J. Struct. Eng. 130 (2) (2004) 271–284.
[16] G. Vasdravellis, B. Uy, E.L. Tan, B. Kirkland, Behavior and design of composite beams
subjected to negative bending and compression, J. Constr. Steel Res. 79 (2012)
34–47.
[17] S. Rafiei, K.M.A. Hossain, M. Lachemi, K. Behdinan, M.S. Anwar, Finite element
modeling of double skin profiled composite shear wall system under in-plane
loadings, Eng. Struct. 56 (2013) 46–57.
[18] A.H. Varma, S.R. Malushte, K.C. Sener, Z. Lai, Steel-plate composite (SC) walls for
safety related nuclear facilities: design for in-plane forces and out-of-plane mo-
ments, Nucl. Eng. Des. 269 (2014) 240–249.
[19] K.C. Sener, A.H. Varma, Steel-plate composite walls: experimental database and
design for out-of-plane shear, J Constr. Steel Res. 100 (2014) 197–210.
[20] Z. Kai, A.H. Varma, S.R. Malushte, S. Gallocher, Effect of shear connectors on local
buckling and composite action in steel concrete composite walls, Nucl. Eng. Des.
269 (2014) 231–239.
[21] H.D. Wright, T.O.S. Oduyemi, H.R. Evans, The design of double skin composite ele-
ments, J. Constr. Steel Res. 19 (2) (1991) 111–132.
[22] H.D. Wright, T.O.S. Oduyemi, Partial interaction analysis of double skin composite
beams, J. Constr. Steel Res. 19 (4) (1991) 253–283.
[23] T.O.S. Oduyemi, H.D. Wright, An experimental investigation into the behavior of
double-skin sandwich beams, J. Constr. Steel Res. 14 (3) (1989) 197–220.
[24] O. Dogan, T.M. Roberts, Comparing experimental deformations of steel-concrete-
steel sandwich beams with full and partial interaction theories, Int. J. Phys. Sci. 5
(10) (2010) 1544–1557.
[25] M. Xie, N. Foundoukos, J.C. Chapman, Experimental and numerical investigation on
the shear behavior of friction-welded bar-plate connections embedded in concrete,
J. Constr. Steel Res. 61 (5) (2005) 625–649.
[26] M. Xie, N. Foundoukos, J.C. Chapman, Static tests on steel–concrete–steel sandwich
beams, J. Constr. Steel Res. 63 (6) (2007) 735–750.
224 S. Sabouri-Ghomi et al. / Thin-Walled Structures 104 (2016) 211–224
[27] C.A. Coulomb, Essai sur une application des regles des maximis et minimis a
quelque problems de statique relatifs 1’architecture, Mem. d’Acad. Roy. Pres. Divers.
7 (1773).
[28] Saburo Okabe, General theory of earth pressure, J. Jpn. Soc. Civil Eng. 12 (1) (1926)
311.
[29] Nagaho Mononobe, Haruo Matsuo, On the determination of earth pressures during
earthquakes, in: Proceedings of the World Engineering Congress, 9, 1929.
[30] S. Løset, K.N. Shkhinek, E. Uvarova, Evaluation of Existing Ice Force Prediction
Methods, LOLEIF Report 3, 8, 1999.
[31] S. Løset, K.N. Shkhinek, E. Uvarova, An overview of the influence of structure width
and ice thickness on the global ice load, Proc. Port. Ocean. Eng. Arct. Cond. (1999)
425–434.
[32] S. Løset, K.N. Shkhinek, T. Kärnä, Global ice load dependency on structure width and
ice thickness, Proc. Port. Ocean. Eng. Arct. Cond. (2003) 857–868.
[33] AISC, Seismic provisions for structural steel buildings, ANSI/AISC 341 10 Chicago,
Illinois: American Institute of Steel Construction AISC, 2010.
[34] R.P. Johnson, Composite Structures of Steel and Concrete: Beams, Slabs, Columns,
and Frames for Buildings, John Wiley and Sons, United States, 2008.
[35] Sorensen, Hibbitt Karlsson, Sorensen, ABAQUS /standard user’s manual, Version
6.13. Pawtucket, RI: HKS Inc.; 2013.
[36] S.V. Chaudhari, M.A. Chakrabarti, Modeling of concrete for nonlinear analysis using
finite element code ABAQUS, Int. J Comput. Appl. 44 (7) (2012) 14–18.
[37] Sandor Popovics, A numerical approach to the complete stress–strain curve of
concrete, Cem. Concr. Res. 3 (5) (1973) 583–599.
[38] W.K. Yip, Generic form of stress–strain equations for concrete, Cem. Concr. Res. 28
(4) (1998) 499–508.