Journal of Building Engineering 31 (2020) 101359

Contents lists available at ScienceDirect

Journal of Building Engineering

journal homepage: http://www.elsevier.com/locate/jobe

Numerical and theoretical analysis of seismic behaviour of CFDSP

composite shear walls
Jian Li *, Feng Li , Caiwei Liu , Jijun Miao
School of Civil Engineering, Qingdao University of Technology, Qingdao, 266033, China

A R T I C L E I N F O A B S T R A C T

Keywords: In the present study, the seismic behaviour of concrete-filled double-steel-plate (CFDSP) composite shear walls
CFDSP composite shear wall was numerically and theoretically investigated. Finite element (FE) models of CFDSP composite shear walls were
FE model designed, validated, and analyzed. Based on the FE models, the influence of the width-to-thickness ratio, height-
Restoring force model
to-width ratio, axial compression ratio, thickness of steel plates, number of stiffeners and spacing of shear studs
Seismic performance
High-rise buildings
on the seismic behaviour of CFDSP composite walls was also investigated. The failure mode, hysteretic curve,
skeleton curve, stiffness, shear-bearing capacity, and ductility of CFDSP composite shear walls are discussed. It
was revealed that CFDSP composite shear walls make full use of the steel and concrete materials and exhibit
excellent seismic performance. Additionally, a restoring force model for describing the hysteretic behaviour of
CFDSP composite shear walls has been proposed and discussed. The predicted hysteretic curves obtained from
the restoring force model are in good agreement with the test results. This research confirmed that CFDSP
composite shear walls are suitable lateral resisting members of high-rise buildings.

1. Introduction Several scholars have devoted themselves to the study of the me­
chanical behaviour of composite shear walls. Hossain et al. [16] con­
Due to the development of innovative technologies of building ducted experimental and theoretical investigations on the axial
structures, more and more large-span structures [1,2], high-rise build­ compression behaviour of composite shear walls with and without an
ings [3], and large bridges [4] have been built. In these buildings, opening. Based on the test results, the axial load-deformation response,
several composite members have been gradually developed to make full axial strength, steel-concrete interaction, failure mode, and stress–strain
use of the property of different materials. These members include the development were discussed. Finally, the authors proposed theoretical
column [5–9], beam [4,10], wall [3], and joint [11]. Among these formulae to predict the axial load-bearing capacity of the composite
composite members, the composite shear wall, which has excellent shear wall. Ji et al. [17] carried out quasi-static tests on two single-plate
lateral stiffness, is one of the main components of high-rise buildings reinforced composite shear walls and three double skin composite shear
[12,13]. Recently, increasingly innovative high-rise buildings have been walls under an axial force and lateral cyclic loading. It was found that
constructed worldwide. As a result, a new concrete-filled double-­ the formulae in the current codes underestimated the shear-bearing
steel-plate (CFDSP) composite shear wall, which generally comprises capacity of the composite shear wall. Zhang et al. [18] presented the
two layers of steel plates, internal concrete, through-wall bolts, and effect of shear studs on the composite action of the steel-composite wall.
stiffeners, has been developed. The through-wall bolts and stiffeners are It was found that the shear studs could prevent local buckling of the steel
placed between two layers of steel plates to prevent the local buckling of face plates and provide composite action between the steel plates and
the steel plates, and they can help improve the shear-bearing capacity, concrete infill of SC walls. To overcome the quality problems caused by
ductility, and global behaviour of the CFDSP composite shear wall. welding, Yousefi, et al. [19] proposed a new construction method in
Compared with the traditional shear wall, the CFDSP composite shear which only one end of the connector was welded to the steel plates while
wall has the advantages of higher load-bearing capacity and better the other was embedded in the concrete core. Next, the authors carried
seismic behaviour in construction [14,15]. Thus, the CFDSP composite out a push-out test to study the performance of the composite shear wall
shear wall has been widely used in high-rise buildings. manufactured using the new construction method. Finally, equations for

* Corresponding author.
E-mail address: [email protected] (J. Li).

https://doi.org/10.1016/j.jobe.2020.101359
Received 1 December 2019; Received in revised form 10 March 2020; Accepted 17 March 2020
Available online 23 March 2020
2352-7102/© 2020 Elsevier Ltd. All rights reserved.
J. Li et al. Journal of Building Engineering 31 (2020) 101359

predicting the ultimate shear-bearing capacity of the structure were 2. Experimental program
proposed. Seo et al. [20] used a mechanics-based model (MBM) to
present the fundamental in-plane shear force–shear strain response of 2.1. Test
the steel plate composite wall and introduced an estimated value for the
strength reduction factor (ϕ) by means of reliability analysis. In previous studies, a series of tests pertaining to the seismic
The seismic performance of members has an important influence on behaviour of eight CFDSP composite shear wall specimens was per­
the safety of structures [21]. Until now, the composite shear wall has formed, and it has been reported in a companion paper [3]. For the
been used as a lateral loading-resisting system for a long time, and it CFDSP composite shear wall specimens, steel plates were constructed
plays a key role in the seismic behaviour of high-rise buildings. Hossain using the Q235 steel, and an infill concrete core was poured with the
et al. [22] conducted experimental and numerical studies on the seismic material being C30 or C50 concrete. The thickness of flange plates and
behaviour of composite shear walls. It was indicated that the composite stiffeners was equal to that of steel plates, and they were welded
shear wall exhibited excellent capacity for energy absorption and together. The dimensions of these CFDSP composite shear wall speci­
strength retention under post impact load. Subsequently, Rafiei, et al. mens are presented in Fig. 1 and Table 1. The CFDSP composite shear
[23] performed further research on the seismic performance of such wall specimens were labelled from CFDSP1 to CFDSP6. The wall thick­
composite wall systems. They theoretically predicted the maximum top ness T was 120 mm. The wall width B was 1200 mm. In the test pro­
displacement of the composite shear wall under in-plane impact loading. gramme, eight parameters including the calculated height of wall H,
Ghomi et al. [24] tested the shear-bearing capacity of composite shear thickness of steel plates t, number of stiffeners n, spacing of shear studs s,
walls with and without stiffeners. According to the test results, the au­ through-wall bolts, concrete grade, and axial compression ratio β were
thors proposed equal energy curves and the plate–frame interaction selected. The axial compression ratio β was defined as the ratio of axial
theory (PFI), which could accurately describe the hysteretic behaviour force to product of nominal strength and section area. During the
of composite shear walls. Nguyen et al. [25] performed a numerical loading process, the CFDSP composite shear wall specimens were sub­
study to investigate the effects of key design variables on the in-plane jected to constant axial force and reversed cyclic lateral loading. The
shear capacity of steel plate composite shear walls. The predicted re­ loading scheme is shown in Fig. 2.
sponses including the global force displacement relationship, damage of
infill concrete and steel plates, shearing force contribution, and
connector behaviour, were in good agreement with the measured data. 2.2. Failure modes and test results
Elmatzoglou et al. [26] developed a three-dimensional finite element
(FE) model, which could efficiently predict the overall performance of a Basically, all specimens had similar failure modes. The main failure
double-steel plate concrete wall in terms of accuracy and time reduction. mode of the CFDSP composite shear wall specimens is presented in Fig. 3
The developed model was validated in terms of the axial compression (a). As the load increased, first, the steel plates yielded at the edge of the
and monotonic, cyclic in-plane shear loading using results from tests flange. When the CFDSP composite shear wall specimens entered the
reported in the literature. Zhao et al. [27] proposed a four-linear back­ plastic phase, the paint layer of the steel plates began to detach. Local
bone curve and a hysteretic rule based on the results of thirty-two buckling appeared at the bottom of steel plates, signifying that the ul­
in-plane cyclic loading tests. The authors also suggested reduction fac­ timate load-bearing capacity of the CFDSP composite shear wall speci­
tors for the elastic modulus of steel and concrete, which reflect the mens had been reached.
plasticity extension and damage accumulation. Zhu et al. [28] proposed The load–displacement hysteretic curve, which reflects the rela­
an innovative composite wall with the end steel shapes setting consid­ tionship between the horizontal load V and deformation δ, is an
ered as a constraint. The authors conducted quasi-static tests on ten important index for the evaluation of the seismic performance. The
composite shear walls under vertical loading. It was indicated that the load–displacement hysteretic curves of the CFDSP composite shear wall
seismic performance of these composite shear walls was considerably specimens are shown in Fig. 4. At the beginning of the cyclic loading
enhanced with the setting of end steel shapes. To calculate the mono­ process, the load–displacement curve exhibited a linear characteristic,
tonic response of the steel plate composite shear wall, Epackachi, et al. implying the elastic phase of the CFDSP composite shear wall specimens.
[29] presented a simplified procedure to predict the flexure and flex­ As the load increased, steel plates began to buckle at the bottom, and the
ure–shear responses of rectangular steel plate composite shear walls concrete was crushed at the corresponding position. Residual deforma­
using moment-curvature and shearing force–shearing strain relation­ tion appeared and became increasingly larger, resulting in an increase in
ships. The simplified model was validated using finite element numer­ the area of the hysteretic loop. Before the ultimate load-bearing capacity
ical results. was reached, two hysteretic curves at the same loading level overlapped,
The aforementioned references indicate that the specified design signifying that the strength and stiffness degradation were inconspic­
method for accurately estimating the seismic behaviour of the CFDSP uous. After the ultimate load was applied, the strength and stiffness of
composite shear wall is limited. Therefore, the main objective of this the CFDSP composite shear wall specimens weakened rapidly. Overall,
study was to realize complete understanding of the CFDSP composite the CFDSP composite shear wall exhibited a full hysteretic curve and
shear wall by means of numerical and theoretical studies. A brief good seismic behaviour.
description of the experimental programme which has been carried out Skeleton curves can reflect the load-bearing capacity, stiffness, and
by the author is reported to present the seismic behaviour of such walls. ductility of the CFDSP composite shear wall. The skeleton curves of
Subsequently, the FE models of these composite shear wall specimens partial CFDSP composite shear wall specimens are shown in Fig. 5. The
are described. To achieve reliable FE results, the initial geometric im­ height-to-width ratio considerably influenced the seismic behaviour of
perfections, material, geometrical non-linearity, interface friction were the CFDSP composite shear wall specimens. As the height-to-width ratio
taken into consideration. To validate the reliability of the FE models, the increased, the shear-bearing capacity notably decreased. Owing to the
FE and test results were compared. Furthermore, parametric studies
were performed by using the proposed FE models and varying the width-
to-thickness ratio, height-to-width ratio, axial compression ratio,
thickness of steel plates, number of stiffeners, and spacing of shear studs.
Finally, a restoring force model that could describe the hysteretic
behaviour of the CFDSP composite shear wall was proposed.

Fig. 1. Cross-section of the CFDSP composite shear wall specimens.

2
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Table 1
Dimensions of the CFDSP composite shear wall specimens.
Specimens H/mm T/mm B/mm t/mm n s Through- wall bolts Concrete β

CFDSP1 2160 120 1200 8 3 Φ8@160 M8@160 C30 0.4

CFDSP2 2160 120 1200 8 3 None None C30 0.4
CFDSP3 2160 120 1200 8 1 None None C30 0.4
CFDSP4 3000 120 1200 8 3 Φ8@160 M8@160 C30 0.4
CFDSP5 2160 120 1200 6 3 Φ8@160 M8@160 C30 0.25
CFDSP6 2160 120 1200 8 3 Φ8@160 M8@160 C50 0.25

Fig. 2. Loading scheme of tests.

Fig. 3. Comparison of failure mode.

constraint of steel plates, the ductility of the inner concrete of the CFDSP in Fig. 6, the steel plates and stiffeners were meshed using S4R elements,
composite shear wall specimens improved significantly, and this effect and the inner concrete was meshed using C3D8 elements. In order to
could be enhanced further with increase in the thickness of the steel simplify the simulation, for nonlinear connector elements, the ratio of
plates. The use of stiffeners could increase the ductility and shear- length to diameter was 30:1, and its main function was to constrain the
bearing capacity of the CFDSP composite shear wall specimens, and out of plane deformation of steel plates on both sides. For the shear
the shear studs could delay the local buckling of steel plates. behaviour, due to the symmetrical loading in the thickness direction, the
deformation of the two sides of the steel plate will not be inconsistent.
3. Finite element simulation Therefore, the shear force on the nonlinear connector elements was zero.
Even if there is a big inconsistency between the two steel plates at the
In the companion paper [3], the mechanical behaviour of the CFDSP later stage of loading, this effect can be ignored in the experiment. So the
composite shear wall was discussed primarily on the basis of experi­ nonlinear connector elements adopted the axial tension compression
mental investigation. To better understand the outcome of the experi­ element. The shear studs were meshed using T3D2 elements. Generally,
mental programme, analysis was performed by means of numerical S4R is a shell element with 4-node doubly curved thin or thick shell,
simulation using the non-linear code ABAQUS. reduced integration, hourglass control, finite membrane strains; C3D8 is
a solid element with 8-node linear brick; and T3D2 is a bar element with
2-node linear 3-dimension truss.
3.1. Geometric dimensions and element type

The geometric dimensions of FE models were the same as those of the 3.2. Material properties
CFDSP composite shear wall specimens described in the companion
paper. The stiffness of the loading beam and the deviation of the as­ The material properties of the components were obtained from the
sembly were ignored. The vertical and horizontal forces were assumed tests introduced in the companion paper. In the simulation, a linear
to be applied at the centre of the CFDSP composite shear wall. As shown kinematic hardening model was adopted to simulate the constitutive

3
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Fig. 4. Load-displacement hysteretic curves obtained from the experimental and FE results.

relationship of the steel plates. The points of yield strength and ultimate assumed that the concrete fails via the mechanisms of crack and crush.
strength of the linear kinematic hardening model could be determined According to reference [30] and the strength obtained from reference
via the test results, as shown in Fig. 7. The yield strength and ultimate [3], the stress–inelastic strain (σ –εin) curve and the damage coefficient of
tensile strength of steel was 280 MPa and 470 MPa, respectively. The inner concrete could be derived, as shown in Fig. 8. The compressive
kinematic hardening factor C could be calculated using Eq. (1), where σo strength of concrete was 32 MPa. The tensile strength of concrete was
represents the yield stress, σ lim represents the ultimate strength, and εpl 2.01 MPa.
is the ultimate plastic strain.
σlim σ0 3.3. Interaction and initial geometric imperfections
C¼ (1)
εpl
In the FE analysis, the Mohr–Coulomb frictional model (Fig. 9) and
A damage plasticity model was employed to simulate the constitutive
hard contact were employed to consider the interaction between the
relationship of inner concrete. For the damage plasticity model, it was
steel plates and infill concrete. Based on the experimental conditions,

4
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Fig. 5. Skeleton curves.

Fig. 6. FE model.

the internal faces of steel plates were chosen as the master surfaces and
the concrete surface was the slave surface. Tie constraints were applied
to the connection of stiffeners and concrete.
Because the wall specimen is not an ideal vertical member, it will
have initial bending. Therefore, it was necessary to consider the initial
geometric imperfections in FE models. The most unfavourable initial
imperfection distribution form was applied to the wall model before
loading. According to the experimental phenomena and numerical
analysis, it is observed that before the overall instability of the wall, the
bottom steel plate had entered the plasticity, and the steel plate can not
continue to bear the axial force, leading to steel plate buckling. Due to
the supporting effect of concrete, the form of drumbeat can only be
outward. This was a very complex elastic-plastic buckling process. In the
elastic buckling mode, it is difficult to find the corresponding buckling
form. Adding initial imperfections according to the elastic buckling form
was an approximate method. In this test, due to the bearing of certain
axial force, the mechanical property of wall specimens was similar to
that of columns. When it was damaged, the material first entered into
Fig. 7. Stress-strain curve of steel.
yield, which is originally an elastic-plastic buckling process. Macro­
scopically, adding the first-order buckling mode can better reflect the

5
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Fig. 8. Stress-strain curve of concrete.

was similar to that observed in the test results. Considering the hyster­
etic response, the hysteretic curves obtained from the FE simulations
and tests were similar. Considering the skeleton curves, the initial
stiffness and shear-bearing capacity of the FE skeleton curves were in
good agreement with the experimental skeleton curves.
Consequently, good agreements existed in the verification of the FE
modelling against the representative results of tests conducted consid­
ering the failure mode, load–displacement hysteretic curves and skel­
eton curves of the CFDSP composite shear wall.

4.2. Damaged mechanism

At the beginning of the loading process, the steel plates and inner
Fig. 9. Coulomb’s friction model. concrete remained in the elastic phase. Therefore, the deformation of
these two components was coordinated. When the load increased to a
specific value, the concrete cracked under the tensile stress and under­
stress characteristics. Therefore, the first-order buckling mode with an
went crushing under the compressive stress at the edge of the CFDSP
amplitude of H/1000 was adopted conservatively. The FE model with
composite shear wall. At this time, the steel plate did not yield, and its
the initial geometric imperfections is illustrated in Fig. 10.
stress was about in the range of 180 MPa–230 MPa. During this time, the
stiffness of the CFDSP composite shear wall decreased. Due to the
4. Finite element simulation
confinement effect of the steel plates, the CFDSP composite shear wall
could continue to resist the load. As the horizontal force increased
4.1. Verification of the FE models
continually, the steel plates entered into the plastic phase, and the
stiffness declined rapidly. As the plastic region became increasingly
Verification of the FE models was performed against the test results
larger, local buckling occurred at the bottom of the steel plates. At this
in terms of the failure mode, load–displacement hysteretic curves, and
time, the CFDSP composite shear wall reached its ultimate bearing ca­
skeleton curves. The comparisons in terms of the main failure mode,
pacity. Fig. 11 shows the stress and strain values of the steel plates at the
hysteretic response, and skeleton curves of the CFDSP composite shear
damage state. The steel plates at the bottom of the CFDSP composite
wall between the FE and test results are presented in Figs. 3–5, respec­
shear wall yielded, and their equivalent plastic strain was 0.1. Fig. 12
tively. Considering the main failure mode, it was observed that local
shows a plot of the damage state of the inner concrete. Figs. 11 and 12
buckling occurred at the bottom of the steel plates. This phenomenon
shows the strain diagram of wall specimens at a horizontal displacement
of about 15 mm. Compressive damage of the inner concrete occurred at
the corners of both sides of the CFDSP composite shear wall, while its
tensile damage was concentrated mainly on the corners and on the
location of the shear studs.

5. Parametric studies

To facilitate further understanding of the seismic behaviour of the

CFDSP composite shear wall, parametric studies were carried out. All
the FE models had the same width B, whose value was 1200 mm. Six
parameters including the width-to-thickness ratio, height-to-width ratio,
axial compression ratio, thickness of steel plates, number of stiffeners,
and spacing of shear studs were considered. The values of these pa­
rameters were varied. The axial compression ratio β took values of 0,
0.2, 0.4, 0.6, and 0.8; the height-to-width ratio H/B values were 1.2, 1.4,
1.8, 2.2, and 2.6; the width-to-thickness ratio B/T values were 4, 6, 8,
10, and 12; the thickness of the steel plates, t, took values of 6 mm, 8
mm, and 12 mm; the values of the number of stiffeners, n, were 1, 3, and
Fig. 10. Initial geometric imperfections. 5; and the spacing of shear studs, s, took values of 80 mm, 160 mm, and

6
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Fig. 11. Stress and strain of steel plates at the damage state.

Fig. 12. Damaged state.

240 mm. The dimensions of the FE models adopted in the parametric the “FE-E00 group contains the CFDSP composite shear wall models with
studies are listed in Table 2. FE-Basic represents the basic FE model with different amounts of stiffeners, and the “FE-F00 group contains the CFDSP
β ¼ 0, H/B ¼ 1.8, B/T ¼ 10, t ¼ 8 mm, n ¼ 3, and s ¼ 160 mm. The “FE- composite shear wall models with different spacings of shear studs. The
A00 group contains the CFDSP composite shear wall models with different FE results are presented in Table 2. K1 is the initial stiffness; Vy, Vmax,
width-to-thickness ratios, the “FE-B00 group contains the CFDSP com­ and Vu represent the loads at the yield state, ultimate state, and
posite shear wall models with different axial compression ratios, the collapsed state, respectively; δy, δmax, and δu represent the displacements
“FE-C00 group contains the CFDSP composite shear wall models with at the yield state, ultimate state, and collapsed state, respectively; and μ
different height-to-width ratios, the “FE-D00 group contains the CFDSP is the displacement ductility coefficient, which can be calculated using
composite shear wall models with different thicknesses of steel plates, Eq. (2). These loads and displacements are plotted in Fig. 13. For the

7
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Table 2
Characteristics of nineteen FE models.
No. β H/B B/T t/mm s/mm n K1/kN/mm Yield status Ultimate state Collapsed state μ
Vy/kN δy/mm Vmax/kN δmax/mm Vu/kN δu/mm

FE-Basic 0.2 1.8 10 8 160 3 227 1119.09 6.97 1256.60 13.23 1068.11 29.90 4.29
FE-A1 0.2 1.8 4 8 160 3 429 1820.46 6.91 2110.40 12.22 2088.90 16.55 2.39
FE-A2 0.2 1.8 6 8 160 3 317 1415.67 6.71 1616.50 10.82 1412.70 30.00 4.47
FE-A3 0.2 1.8 8 8 160 3 261 1224.86 6.77 1383.90 11.61 1176.32 28.64 4.23
FE-A4 0.2 1.8 12 8 160 3 205 1045.12 7.08 1169.70 12.32 994.25 21.45 3.03
FE-B1 0 1.8 10 8 160 3 225 1125.88 7.73 1242.80 13.18 1232.60 16.27 2.11
FE-B2 0.4 1.8 10 8 160 3 223 1001.33 6.04 1132.90 10.80 962.97 14.35 2.38
FE-B3 0.6 1.8 10 8 160 3 196 781.16 4.72 872.93 8.84 741.99 15.45 3.27
FE-B4 0.8 1.8 10 8 160 3 153 487.16 3.41 522.20 4.58 443.87 8.33 2.44
FE-C1 0.2 1.2 10 8 160 3 603 1551.27 3.72 1770.30 7.05 1504.76 14.32 3.85
FE-C2 0.2 1.4 10 8 160 3 423 1426.51 4.95 1618.00 8.59 1375.30 26.54 5.36
FE-C3 0.2 2.2 10 8 160 3 133 910.39 9.31 1015.10 16.54 1015.70 16.77 1.80
FE-C4 0.2 2.6 10 8 160 3 84 755.00 11.80 838.02 16.09 838.03 16.09 1.36
FE-D1 0.2 1.8 10 6 160 3 197 889.78 6.58 1005.90 10.80 855.02 22.20 3.37
FE-D2 0.2 1.8 10 12 160 3 288 1566.81 7.38 1747.10 13.22 1661.40 30.00 4.06
FE-E1 0.2 1.8 10 8 160 1 221 1026.44 6.39 1150.90 9.53 978.27 13.47 2.11
FE-E2 0.2 1.8 10 8 160 5 234 1192.96 7.23 1337.90 12.80 1235.70 30.00 4.15
FE-F1 0.2 1.8 10 8 80 3 228 1129.52 7.04 1268.10 13.48 1150.30 30.00 4.26
FE-F2 0.2 1.8 10 8 240 3 227 1113.62 6.91 1250.10 12.09 1062.59 23.06 3.34

5.2. Influence of the axial compression ratio

The skeleton curves of the FE models with different axial compres­

sion ratios are shown in Fig. 14(b). The effect of the axial compression
ratio on the seismic performance of the CFDSP composite shear wall was
investigated according to Fig. 14(b). It was found that when the axial
compression ratio was larger than 0.2, the axial compression ratio
exerted considerable influence on the shear-bearing capacity of the
CFDSP composite shear wall. However, the effect of the axial compres­
sion ratio on the initial stiffness could be ignored. According to Table 2,
when the axial compression ratio increased from 0.2 to 0.8, the yield
load and ultimate load decreased from 1119.09 kN to 487.16 kN
(decrease of 43.5%) and 1256.6 kN–522.2 kN (decrease of 41.6%),
respectively.

Fig. 13. Geometric drawing method. 5.3. Influence of the height-to-width ratio

The skeleton curves of the FE models with different height-to-width

displacement δy at the yield state, it can be determined by the geometric
ratios are shown in Fig. 14(c). As the height-to-width ratio increased, the
drawing method since there was no obvious yield point in the skeleton
shear-bearing capacity and initial stiffness decreased. When the height-
curve. The displacement δu at the collapsed state corresponded to the
to-width ratio increased from 1.2 to 2.6, the initial stiffness, yield load,
load Vu at the collapsed state, and the load Vu is equal to 0.85Vmax. The
and ultimate load decreased from 603 kN/mm to 84 kN/mm (decrease
skeleton curves of these FE models were plotted, as shown in Fig. 14.
of 13.9%), 1551.27 kN–755 kN (decrease of 48.7%), and 1770.3
δu kN–838.02 kN (decrease of 47.3%), respectively. It was observed that
μ¼ (2)
δy when the values of the height-to-width ratio were 2.2 and 2.6, the
displacement ductility coefficient was considerably small, because as the
5.1. Influence of the width-to-thickness ratio height-to-width ratio increased, the failure mode of the CFDSP com­
posite shear wall changed from shear collapse to bending collapse.
The skeleton curves of the FE models with different width-to-
thickness ratios are shown in Fig. 14(a). Fig. 14(a) clearly illustrates 5.4. Influence of the thickness of steel plates
the effect of the width-to-thickness ratio on the seismic performance of
the CFDSP composite shear wall. The width-to-thickness ratio influ­ The skeleton curves of the FE models with different thicknesses of
enced the shear-bearing capacity and initial stiffness. As the width-to- steel plates are shown in Fig. 14(d). As the thickness of the steel plates
thickness ratio increased from 4 to 12, the initial stiffness, yield load, increased from 6 mm to 8 mm, the initial stiffness, yield load, and ul­
and ultimate load decreased from 429 kN/mm to 205 kN/mm (decrease timate load increased from 197 kN/mm to 288 kN/mm (increase of
of 47.8%), 1820.46 kN–1045.12 kN (decrease of 57.4%), and 2110.4 146.2%), 889.78 kN–1566.81 kN (decrease of 176.1%), and 1005.9
kN–1169.7 kN (decrease of 55.4%), respectively. However, the dis­ kN–1747.1 kN (decrease of 173.7%), respectively. Therefore, the
placements corresponding to the yield and ultimate states were nearly thickness of steel plates played a key role in the seismic behaviour of the
equivalent. When the width-to-thickness ratio was in the range of [6, CFDSP composite shear wall.
12], the displacement ductility coefficient of the CFDSP composite shear
wall decreased with increase in the width-to-thickness ratio. 5.5. Influence of the amount of stiffeners

The skeleton curves of the FE models with different numbers of

stiffeners are shown in Fig. 14(e). According to Fig. 14(e), the number of

8
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Fig. 14. Skeleton curves in parametric studies.

stiffeners primarily affected the shear-bearing capacity and ductility. 5.6. Influence of the spacing of shear studs
When the number of stiffeners increased from 1 to 5, the yield load and
ultimate load increased from 1026.44 kN to 1192.96 kN (increase of The skeleton curves of the FE models with different spacing of shear
116.2%) and from 1150.9 kN to 1337.9 kN (increase of 116.2%), studs are shown in Fig. 14(f). It was found that the influence of the
respectively. In addition, as the number of stiffeners increased, the ul­ spacing of shear studs on the seismic behaviour of the CFDSP composite
timate displacement increased, leading to improved ductility. However, shear wall was relatively small. Moreover, the reduction in the spacing
the effect of the number of stiffeners on the initial stiffness could be of shear studs could improve the ductility.
ignored.

9
J. Li et al. Journal of Building Engineering 31 (2020) 101359

6. Theoretical derivation
1 0:0002
Vmax ¼ ½0:0346ft ðT 2tÞðL 2tÞ 0:0018N� þ fy Aw
H=B þ 18:671 H=B
6.1. Skeleton curves
0:0003
þ fy Ap
6.1.1. Assumption H=B þ 0:078
In this study, a tri-linear model was developed to simulate the skel­ (5)
eton curve of a CFDSP composite shear wall subjected to reversed cyclic
lateral loading. The tri-linear skeleton curve model could be defined as δmax ¼
0:421H 0:454
ft ðT 2tÞðL 2tÞ
shown in Fig. 15. The yield point, ultimate bearing capacity point, and G1 A1 þ 0:141G2 A2 ð2:489 H=BÞ2
the collapse load point were selected as the key points for the skeleton
! (6)
113:354
curve. þ fy A1 2:197N
21:662 H=B

6.1.2. Determination of the key points

According to the aforementioned discussion, the parameters that 6.1.2.3. Collapsed status
� �
affect the key points of the tri-linear skeleton curve model include 1 0:184 0:01
Vu ¼ fy Aw þ f y Ap (7)
mainly the axial force N, the height-to-width ratio H/B, and the thick­ H=B þ 14:964 H=B H=B 0:497
ness of steel plates t. Based on the formulae proposed in JGJ 3–2010 [31]
�
and the statistical regression method, relevant values of the key points 1:239H 2:552
δu ¼ ft ðT 2tÞðL 2tÞ
could be calculated using Eqs. (3) – (8). G1 A1 0:354G2 A2 H=B 2:368
� (8)
0:998
þ fy A1 0:27N
6.1.2.1. Yield status H=B 1:296
1 0:0001
Vy ¼ ½0:0446ft ðT 2tÞðL 2tÞ 0:00026N� þ f y Aw
H=B þ 32:478 H=B 6.1.3. Comparison of skeleton curves
0:0003 The key points of the tri-linear skeleton curve model were calculated
þ fy Ap by means of the proposed formulae. The relevant skeleton curves are
H=B þ 0:064
(3) shown in Fig. 5. It can be noted that the proposed formulae agree well
with the experimental curves in terms of the initial stiffness, yield
35:765H 0:0002 strength, ultimate strength, and displacement corresponding to the ul­
δy ¼ ft ðT 2tÞðL 2tÞ timate strength.
G1 A1 þ 0:128G2 A2 ð2:522 H=BÞ2
! (4)
0:404 6.2. Hysteretic rules
þ fy A1 0:0136N
12:594 H=B
In most studies, the basic hysteretic rules usually incorporate the
Here, fy is the characteristic value of yield strength of steel; ft is the stiffness deterioration and strength deterioration. These rules are used to
characteristic value of concrete tensile strength; G1 is the shear modulus investigate the degradation law and cumulative damage failure rule of
of steel; G2 is the shear modulus of concrete; H, B, and T represent the the strength and stiffness of the CFDSP composite shear wall.
total height, width, and thickness of the CFDSP composite shear wall,
respectively; t is the thickness of steel plates; N is the axial force applied 6.2.1. Stiffness deterioration rule
at the specimens; L is the effective height of the CFDSP composite shear The experimental hysteretic curves indicate that the change in the
wall; A1 and A2 are the sectional areas of the steel plate and concrete unloading stiffness was inconspicuous before the yield point. However,
respectively; and Aw and Ap are the sectional areas of the stiffeners and with increase in the displacement, the unloading stiffness decreased step
the webs, respectively. by step. To determine the unloading stiffness Kj1 of the CFDSP composite
shear wall, regression analysis based on the test results and numerical
6.1.2.2. Ultimate status simulation was carried out. In Eq. (9), Kj1 and K1 respectively represent
the unloading stiffness of the 1st cycle at each loading level and the
initial stiffness. Further, δy and δ respectively represent the yield
displacement and horizontal displacement at the top of the CFDSP
composite shear wall.
Kj1 0:64
¼� � �1:54 �2 (9)
K1
1 þ 0:32 δδy

6.2.2. Strength deterioration rule

The test results indicate that when the amplitude of the horizontal
displacement was constant, the shear-bearing capacity of the CFDSP
composite shear wall decreased with increase in the number of cyclic
loading loops. After the peak point, the strength deterioration of the
CFDSP composite shear wall became more distinct. An independent
variable δj/δu was considered to investigate the strength deterioration
law, as expressed in Eq. (10).

Fig. 15. Tri-linear skeleton curve model.

10
J. Li et al. Journal of Building Engineering 31 (2020) 101359

� �10:92
δ
6.3. Restoring force model
2:421 þ 0:755 δyj
Vjm
¼ � �10:92 (10) Based on the tri-linear skeleton curve model and hysteretic rules, a
Vj1
restoring force model was developed to investigate the seismic behav­
δ
2:483 þ δyj
iour of the CFDSP composite shear wall under reversed cyclic lateral
where Vj1 and Vjm represent the peak load of the 1st cycle and the mth loading. The softening during the reloading was considered in the
cycle at the j loading level, respectively; m represents the number of the model.
cycle; and δj and δy represent the horizontal displacement at the j loading To verify the accuracy of the restoring force model, the horizontal
level and the yield displacement, respectively. load–displacement hysteretic curves pertaining to the tests and the
restoring force model were compared. The key points in terms of yield
status, ultimate status, and collapsed status can be calculated firstly by
Eqs. (3)–(8). Therefore, the skeleton curve can be obtained. According to

Fig. 16. Comparison of load-displacement hysteretic curve of tests and restoring force model.

11
J. Li et al. Journal of Building Engineering 31 (2020) 101359

Eqs. (9) and (10), the hysteretic rule can be determined. Combined the Appendix A. Supplementary data
skeleton curve with the hysteretic rule, the restoring force curve of the
wall specimens can be plotted by programming based on the corre­ Supplementary data to this article can be found online at https://doi.
sponding loading rule. As shown in Fig. 16, the restoring force model org/10.1016/j.jobe.2020.101359.
agreed well with the experimental curves, especially in the stage of
strength deteriorationn. References
In summary, the proposed restoring force model could effectively
reflect the mechanical characteristics of the CFDSP composite shear [1] Z. Xiong, X.N. Guo, Y.F. Luo, et al., Elasto-plastic stability of single-layer
reticulated shells with aluminium alloy gusset joints, Thin-Walled Struct. 115
wall, and it could satisfy the general requirements of structural (2017) 163–175, https://doi.org/10.1016/j.tws.2017.02.008.
engineering. [2] Z. Xiong, X.N. Guo, Y.F. Luo, et al., Numerical analyses of aluminium alloy gusset
joints subjected to bending moment and axial force, Eng. Struct. 152 (2017) 1–13,
https://doi.org/10.1016/j.engstruct.2017.09.005.
7. Conclusions and expectations [3] Y.F. Luo, X.N. Guo, J. Li, et al., Experimental research on seismic behavior of the
concrete-filled double-steel-plate composite wall, Adv. Struct. Eng. 11 (2015)
In this paper, the seismic behaviour of the CFDSP composite shear 1845–1858, https://doi.org/10.1260/1369-4332.18.11.1845.
[4] J.L. Wang, L. Tian, L.S. Luo, Theoretical and experimental study on deflection of
wall was discussed via FE simulations and theoretical analysis. The steel-concrete composite truss beams, Steel Compos. Struct. 29 (1) (2018) 91–106,
following conclusions were drawn: https://doi.org/10.12989/scs.2018.29.1.091.
Good agreements existed in the validation of the FE modelling [5] M.X. Xiong, D.X. Xiong, J.Y.R. Liew, Behaviour of steel tubular members infilled
with ultra high strength concrete, J. Constr. Steel Res. 138 (2017) 168–183,
against the representative test results obtained considering the failure
https://doi.org/10.1016/j.jcsr.2017.07.001.
mode, load–displacement hysteretic curve, and load–displacement [6] Y.C. Tang, L.J. Li, W.X. Feng, et al., Study of seismic behavior of recycled aggregate
skeleton curve of the CFDSP composite shear wall. Therefore, the pro­ concrete-filled steel tubular columns, J. Constr. Steel Res. 148 (2018) 1–15,
posed FE simulations could be used with confidence to predict the https://doi.org/10.1016/j.jcsr.2018.04.031.
[7] Y.L. Li, X.L. Zhao, P.K. Singh Raman, Theoretical model for concrete-filled stainless
seismic behaviour of CFDSP composite shear walls. steel circular stub columns under axial compression, J. Constr. Steel Res. 157
To facilitate further understanding of the seismic behaviour of the (2019) 426–439, https://doi.org/10.1016/j.jcsr.2019.03.010.
CFDSP composite shear wall, the effects of the width-to-thickness ratio, [8] Z. Xiong, J.P. Deng, F. Liu, et al., Experimental investigation on the behavior of
GFRP-RAC-steel double-skin tubular columns under axial compression, Thin-
axial compression ratio, height-to-width ratio, thickness of steel plates, Walled Struct. 132 (2018) 350–361, https://doi.org/10.1016/j.tws.2018.08.026.
number of stiffeners, and spacing of shear studs on the seismic behaviour [9] Y.C. Guo, W.Y. Gao, J.J. Zeng, et al., Compressive behavior of FRP ring-confined
of the CFDSP composite shear wall were investigated. (1) A larger width- concrete in circular columns: effects of specimen size and a new design-oriented
stress-strain model, Construct. Build. Mater. 201 (2019) 350–368, https://doi.org/
to-thickness ratio corresponded to better shear-bearing capacity of the 10.1016/j.conbuildmat.2018.12.183.
composite shear wall. (2) A large axial compression ratio could reduce [10] M.X. Xiong, D.X. Xiong, J.Y.R. Liew, Flexural performance of concrete filled tubes
the shear-bearing capacity of the CFDSP composite shear wall. (3) The with high tensile steel and ultra-high strength concrete, J. Constr. Steel Res. 132
(2017) 191–202, https://doi.org/10.1016/j.jcsr.2017.01.017.
shear-bearing capacity and initial stiffness reduced with increase in the [11] S.C. Jiang, X.N. Guo, Z. Xiong, et al., Experimental studies on behaviour of tubular
height-to-width ratio. (4) Increasing the thickness of the steel plates was T-joints reinforced with grouted sleeve, Steel and Composite Structures, Int. J. 23
an effective method to improve the shear-bearing capacity of the com­ (5) (2017) 585–596, https://doi.org/10.12989/scs.2017.23.5.585.
[12] Y. Wu, D.Y. Kang, Y.B. Yang, Seismic performance of steel and concrete composite
posite shear wall. (5) The ductility of the CFDSP composite shear wall
shear walls with embedded steel truss for use in high-rise buildings, Eng. Struct.
increased with an increase in the number of stiffeners. (6) The ductility 125 (2016) 39–53, https://doi.org/10.1016/j.engstruct.2016.06.050.
of the CFDSP composite shear wall improved with a decrease in the [13] B. Rassouli, S. Shafaei, A. Ayazi, Experimental and numerical study on steel-
spacing of shear studs. concrete composite shear wall using light-weight concrete, J. Constr. Steel Res. 126
(2016) 117–128, https://doi.org/10.1016/j.jcsr.2016.07.016.
A tri-linear restoring force model and hysteretic rules were proposed [14] J.Y.R. Liew, B.Y. Jia, Y.H. Zhen, Steel-concrete-steel sandwich composite
to model the complete range of the seismic behaviour of CFDSP com­ structures-recent innovations, J. Constr. Steel Res. 130 (2017) 202–221, https://
posite shear walls. Good agreement existed in the validation of the doi.org/10.1016/j.jcsr.2016.12.007.
[15] X. Hao, Y.D. Liu, Analysis of factors influence bearing capacity of double steel
theoretical model and the test results. concrete shear Wall, Constr. Technol. 43 (2016) 90–91.
[16] K.M.A. Hossain, L.K. Mol, M.S. Anwar, Axial load behavior of pierced profiled
Declaration of competing interest composite walls with strength enhancement devices, J. Constr. Steel Res. 110
(2015) 48–64, https://doi.org/10.1016/j.jcsr.2015.03.009.
[17] X.D. Ji, X.F. Jia, J.R. Qian, Experimental study on shear behavior of steel-plate
We declare that we have no financial and personal relationships with composite shear walls, J. Build. Struct. 36 (2015) 46–55.
other people or organizations that can inappropriately influence our [18] K. Zhang, A.H. Varma, S.R. Malushte, et al., Effect of shear connectors on local
buckling and composite action in steel concrete composite walls, Nucl. Eng. Des.
work, there is no professional or other personal interest of any nature or 269 (2014) 231–239, https://doi.org/10.1016/j.nucengdes.2013.08.035.
kind in any product, service and/or company that could be construed as [19] M. Yousefi, M. Ghalehnovi, Push-out test on the one end welded corrugated-strip
influencing the position presented in, or the review of, the manuscript connectors in steel-concrete-steel sandwich structure, Steel Compos. Struct. 24 (1)
(2017) 23–35, https://doi.org/10.12989/scs.2017.24.1.023.
entitled.
[20] J. Seo, A.H. Varma, K. Sener, et al., Steel-plate composite (SC) walls: in-plane shear
behavior, database, and design, J. Constr. Steel Res. 119 (2016) 202–215, https://
CRediT authorship contribution statement doi.org/10.1016/j.jcsr.2015.12.013.
[21] H.F. Li, J. Luo, F. Han, J.X. Luo, Experimental study on seismic behavior of new
steel box bridge piers with embedded energy dissipation shells, Int. J. Steel Struct.
Jian Li: Conceptualization, Project administration, Methodology, 19 (3) (2019) 952–969, https://doi.org/10.1007/s13296-018-0181-0.
Resources, Writing - original draft. Feng Li: Formal analysis, Software, [22] K.M.A. Hossain, S. Rafiei, M. Lachemi, et al., Finite element modeling of impact
Writing - review & editing. Caiwei Liu: Data curation, Investigation. shear resistance of double skin composite wall, Thin-Walled Struct. 117 (2016)
101–111, https://doi.org/10.1016/j.tws.2016.06.002.
Jijun Miao: Funding acquisition, Supervision, Validation. [23] S. Rafiei, K.M.A. Hossain, M. Lachemi, et al., Impact shear resistance of double skin
profiled composite wall, Eng. Struct. 140 (2017) 267–285, https://doi.org/
Acknowledgements 10.1016/j.engstruct.2017.02.062.
[24] S.S. Ghomi, S.R.A. Sajjadi, Experimental and theoretical studies of steel shear-walls
with and without stiffeners, J. Constr. Steel Res. 75 (2012) 152–159, https://doi.
The authors gratefully acknowledge the financial support provided org/10.1016/j.jcsr.2012.03.018.
by the National Natural Science Foundation of China (Project Nos. [25] N.H. Nguyen, A.S. Whittaker, Numerical modeling of steel-plate concrete
composite shear walls, Eng. Struct. 150 (2017) 1–11, https://doi.org/10.1016/j.
5160081693 and 51608285). engstruct.2017.06.030.
[26] M. Elmatzoglou, A. Avdelas, Numerical modelling of double-steel plate composite
shear walls, Computation 5 (1) (2017) 12, https://doi.org/10.3390/
computation5010012.

12
J. Li et al. Journal of Building Engineering 31 (2020) 101359

[27] W.Y. Zhao, Q.Q. Guo, Z.Y. Huang, et al., Hysteretic model for steel-concrete [30] GB 50010–2010, Code for Design of Concrete Structures, Ministry of Housing and
composite shear walls subjected to in-plane cyclic loading, Eng. Struct. 106 (2016) Urban Rural Development of the People’s republic of China, General
461–470, https://doi.org/10.1016/j.engstruct.2015.10.031. administration of quality supervision, inspection and quarantine of the people’s
[28] L.M. Zhu, D.Y. Zhou, M.Y. He, Experimental research on seismic behavior of republic of China, Beijing, 2010.
composite concrete and steel plate shear walls with binding bars, J. Build. Struct. 6 [31] JGJ 3-2010, Technical Specification for Concrete Structures of Tall Building,
(2013) 93–102. Ministry of Housing and Urban-Rural Development of the People’s republic of
[29] S. Epackachi, A.S. Whittaker, Y.N. Huang, Analytical modeling of rectangular SC China, General administration of quality supervision, inspection and quarantine of
wall panels, J. Constr. Steel Res. 105 (2015) 49–59, https://doi.org/10.1016/j. the people’s republic of China, Beijing, 2010.
jcsr.2014.10.016.

13