Journal of Constructional Steel Research 118 (2016) 76–90

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Shear capacity of concrete-filled steel plate composite coupling beams

Hong-Song Hu a,⁎, Jian-Guo Nie b, Yu-Hang Wang c
a
Disaster Prevention Research Institute (DPRI), Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan
b
Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry, Dept. of Civil Engineering, Tsinghua University, Beijing 100084, China
c
School of Civil Engineering, Chongqing University, Chongqing 400045, China

a r t i c l e i n f o a b s t r a c t

Article history: The concrete-filled steel plate (CFSP) composite coupling beam is a newly developed form of coupling beam that
Received 12 April 2015 possesses high deformation and energy dissipation capacities. In this paper, the shear capacity of concrete-filled
Received in revised form 24 October 2015 steel plate composite coupling beams was investigated. An analytical model was proposed based on the force
Accepted 28 October 2015
mechanism of the composite coupling beam and was proved to have adequate accuracy compared with the avail-
Available online xxxx
able test results. The shear capacity of the composite coupling beam was found to be governed by flexural
Keywords:
strength, shear strength, or flexural-shear strength, depending on the material and geometrical properties of
Concrete-filled steel plate composite coupling the coupling beam. By incorporating the effect of flexural-shear interaction, a unified equation was finally
beams developed.
Shear capacity © 2015 Elsevier Ltd. All rights reserved.
Analytical model
Design equations
Concrete filled steel tube

1. Introduction significantly alleviated due to the presence of the embedded steel

Coupled shear wall systems that consist of two or more shear walls
connected by coupling beams are efficient lateral force resisting sys-
tems; such systems have been widely used in mid-rise and high-rise
buildings. As the building height increases, the internal forces in the
structural members increase significantly, which makes the conven-
tional reinforced concrete (RC) coupled shear wall system uneconomi-
cal for use in super high-rise buildings. To overcome this problem, a
new coupled shear wall system, referred to as the concrete-filled steel
plate (CFSP) composite coupled shear wall system, was proposed by
the authors [1], as shown in Fig. 1. In the proposed coupled shear wall
system, CFSP composite wall piers are coupled by CFSP composite cou-
pling beams. Because the two structural members have similar configu-
rations (both consist of surface steel plates and concrete infill), it is easy
to design them with matched stiffness, load-carrying, and deformation
capacities.
Over the past three decades, great efforts have been made to develop
new forms of coupling beams to improve the seismic performance of
coupled shear wall system. The steel plate-reinforced concrete compos-
ite coupling beam is an effective alternative for the conventional RC
coupling beam. Compared with the conventional RC coupling beam,
the steel plate-reinforced composite coupling beam can resist larger
shear forces and withstand substantial inelastic deformations [2].
The pinching problem of the conventional RC coupling beam was
⁎ Corresponding author.
E-mail address: hhsong05@gmail.com (H.-S. Hu).
plate. In addition to the experimental studies, finite element models
[3] and design procedures [4] were developed for this composite
coupling beam.
The composite coupling beam proposed by Gong and Shahrooz [5]
was similar to the steel plate-reinforced composite coupling beam,
though the embedded steel plate was replaced by an I-shaped steel
beam. Component and subassembly testing [6] were both conducted
to study the seismic behavior of the composite coupling beams, the
beam-wall connections and the overall behavior of the composite
coupled wall system. Based on the experimental studies, guidelines for
the proper design and detailing of the composite coupling beams and
beam-wall connections were proposed [7].
In the U.S., recent research efforts on coupled shear wall systems
were focused on the hybrid coupled shear wall system, in which the re-
inforced concrete shear walls were connected by steel coupling beams.
Pushover analysis [8–9] and nonlinear dynamic analysis [10] were con-
ducted, and design recommendations were developed (e.g., [11]).
Several research studies have been conducted on the CFSP compos-
ite shear walls and coupling beams. For the CFSP composite shear walls,
experimental and theoretical studies have been performed and design
recommendations have been made [12–13]. Experiments have been
conducted to study the seismic behavior and mechanism of CFSP com-
posite coupling beams. Six coupling beam specimens with varying
values of the span-to-height ratio, steel plate thickness, and bending-
to-shear capacity ratio were tested under reversed cyclic loading [1].
The progressions of the limit states, hysteretic behavior, deformation ca-
pacity, and energy dissipation of the coupling beams have been studied.

http://dx.doi.org/10.1016/j.jcsr.2015.10.023
0143-974X/© 2015 Elsevier Ltd. All rights reserved.
 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90
Fig. 1. Details of the CFSP composite coupled shear wall system.
Based on the test results, the internal force and deformation responses
of the composite coupling beams were studied in a companion paper
[14]. This study focuses on the shear capacity of CFSP composite cou-
pling beams.
As shown in Fig. 2(b), when the coupled shear wall is subjected to
lateral forces, coupling beams experience constant shear force, and
the points of inflection are located at approximately the mid-span
because the lateral stiffness of the wall piers is usually substantially
larger than that of the coupling beams. Therefore, the shear capacity
Fig. 2. Internal force distributions of coupling beam.
77
of the coupling beam, Vu, which is the maximum shear force attained
in the coupling beam, can be determined as the smaller of the shear
forces corresponding to the flexural strength at the coupling beam
ends, Mu,fle, and shear strength, Vu,sh, if the flexural-shear interaction
is ignored:


V u ¼ min V u;sh ; 2M u;fle =lb ð1Þ
where lb. is the length of the coupling beam span.
Eq. (1) has been adopted in current design specifications (e.g., ACI
318-11 [15]; ANSI/AISC 341-10 [16]) to calculate the shear capacities
of various forms of coupling beams, including conventional reinforced
concrete coupling beams, diagonally reinforced concrete coupling
beams [17], steel coupling beams and steel reinforced concrete compos-
ite coupling beams. The design equations of the shear strength, Vu,sh, are
different for different forms of coupling beams because their shear
mechanisms are different. The force mechanism of CFSP composite cou-
pling beams is distinct from that of reinforced concrete or steel rein-
forced concrete composite coupling beams, so the current design
equations for these coupling beams are not applicable to CFSP compos-
ite coupling beams. Thus, both analytical models and simplified design
equations for the shear capacity of CFSP composite coupling beams
must be established.
2. Force mechanism
Six concrete-filled steel plate composite coupling beams were tested
by the authors [1]. The coupling beam width and height were 150 mm
and 300 mm, respectively, and the key parameters are shown in
Table 1. A schematic diagram of the test setup is shown in Fig. 3. The
specimens were rotated 90° from their actual orientation in a building
to be attached into the test setup through high-strength bolts. The cen-
ter of action of the actuator passed through the center of the beam span,
so the point of inflection was located at the mid-span of the coupling
beam, and equal rotations at the two beam ends were maintained,
which simulated the typical boundary conditions expected in real build-
ings. The specimens were subjected to reversed cyclic loading with
gradually increased lateral displacement.
The distributions of the longitudinal stress, σsx, and shear stress, τsxy,
of the steel plates along the height of beam end at the maximum shear
78 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

Table 1
Calculated and test results of the composite coupling beams.

Specimen designation Specimen parameters Calculated results Vu,tes (kN) Vu,cal/Vu,tes

lb/hb tsf (tsw) (mm) fyf (fyw)a (N/mm2) fc (N/mm2) σw1/fyw τw1/fyw σw2/fyw τw2/fyw xs/hb xc/hb Vu,cal (kN)

CFSCB-1 2.5 5 (5) 336 (336) 38.7 0.77 0.43 1.09 0.19 0.31 0.38 475 550 0.86
CFSCB-2 2.5 3 (3) 354 (377) 39.6 0.65 0.48 1.11 0.16 0.25 0.31 345 378 0.91
CFSCB-3 2.5 8 (3) 338 (384) 43.2 −0.41 0.54 0.99 0.30 0.35 0.34 482 514 0.94
CFSCB-4 1.33 5 (5) 288 (296) 40.1 0.37 0.55 1.05 0.24 0.30 0.50 708 741 0.96
CFSCB-5 1.33 3 (8) 354 (294) 40.9 0.80 0.42 1.06 0.23 0.36 0.50 838 793 1.06
CFSCB-6 1.33 5 (5) 278 (330) 39.4 0.48 0.53 1.05 0.24 0.32 0.50 754 772 0.98
Average 0.95
Standard deviation 0.07
a
fyf and fyw are the steel plate strengths at the maximum shear capacity.

capacity are shown in Fig. 4 for Specimen CFSCB-1 (with span-to-
height ratio of 2.5) and CFSCB-4 (with span-to-height ratio of
1.33). Because the concrete infill could supply compression resis-
tance, most parts of the steel web plates resisted tension forces,
and the steel web plates at the compression region tended to sustain
more shear forces than those at the tension region. It was expected
that this stress distribution pattern created the largest shear capacity
in the composite coupling beam. In Fig. 4, Vu and Mu are the shear
force and bending moment at the beam end at the maximum shear
capacity, respectively.
The internal forces sustained by the steel plates can be obtained
from integrating the steel plate stresses along the beam height, and
the internal forces of the concrete infill can then be deduced from
global equilibrium equations. At the maximum shear capacities, the
ratios of the shear force to the axial compression force in the con-
crete infill were 0.48, 0.44, 0.64, 0.77, 0.64 and 0.88 for specimens
CFSCB-1 through CFSCB-6, respectively, which were close to the
values of the corresponding beam height-to-span ratios (0.4 for
specimens CFSCB-1 through CFSCB-3 and 0.75 for specimens
CFSCB-4 through CFSCB-6). This result implied that a diagonal com-
pression strut was formed in the concrete infill. A similar diagonal
strut mechanism has also been observed in beam-column moment
connections in composite frames [18]. The compression force associ-
ated with the diagonal concrete strut also resisted a portion of the
bending moment at the beam end. At the maximum shear capacities,
the ratios of the eccentricity of the concrete compression force, dc, to
the coupling beam height, hb, were equal to 0.38, 0.42, 0.17, 0.27,
0.26 and 0.28 for specimens CFSCB-1 through CFSCB-6, respectively.
Except for Specimen CFSCB-3, in which measurement errors might
have occurred, the values of dc/hb all exceeded 0.25, i.e., the depth
of equivalent rectangular stress block was equal to 0.5hb. Therefore,
an upper bound value of 0.5hb may be assumed for the depth of the
equivalent rectangular stress block, which corresponds to the largest
bending moment that the concrete infill can supply.
The concrete stresses at the beam end are approximately distributed
as shown in Fig. 5(a). The longitudinal concrete stress, σcx, decreased
from the maximum at the compression edge to 0 at the point where
the concrete infill started to separate from the steel plates, since no
shear connectors were included between the steel plates and the concrete
infill. The distribution of concrete shear stress, τcxy, was consistent with
that of the longitudinal stress. The shear stress was approximately equal
to 0 where the longitudinal stress was 0. Compressive stresses perpendic-
ular to the longitudinal axis of the coupling beam, σcy, also existed in the
concrete infill due to Poisson's effect. The concrete stresses at some typical
locations at the coupling beam end are analyzed in Fig. 5(b). The Mohr's
circle of stress narrowed gradually from the compression edge to the po-
sition where the longitudinal stress, σcx was equal to 0, and the principal
compressive stress, σc2, rotated from along the longitudinal axis to per-
pendicular to the longitudinal axis. The value of the major principal stress,
σc1, basically remained 0 along the beam height.

Fig. 3. Schematic diagram of the test setup for CFSP composite coupling beam testing.
 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 79

Fig. 4. Stress distributions of steel plates at the beam end.

3. Analytical model

3.1. Basic assumptions

The steel plate slenderness and welds at the beam end connections should satisfy the requirements specified in design codes or speci-
fications (e.g., ANSI/AISC 341-10 [16]), so the local buckling of the steel plates and fracture of the welds will not control the shear capacity
of the CFSP composite coupling beam. Another five assumptions are made based on the stress analysis results presented in the previous
section:

(1) The tensile strength of concrete is ignored.

(2) The steel web stresses are evenly distributed at the compression and tension regions, as shown in Fig. 6(b), and the steel web stresses satisfy
the von Mises yield criteria.
(3) The axial compressive force, Nuc, and shear force, Vuc, sustained by the concrete infill at the beam ends satisfy the following equation:

V uc hb
¼ ¼ tanθ ð2Þ
Nuc lb

where hb is the coupling beam height, lb. is the coupling beam span, and θ is the angle between the diagonal and longitudinal axis of the coupling
beam. The average longitudinal compressive stress, σ c , and average shear stress, τc , in the concrete infill at the beam ends satisfy the Mohr's circle
relationship, as shown in Fig. 6(d), where the major principal stress is equal to 0, and the minor principal stress is equal to −fc. The internal forces and
average stresses satisfy

Nuc ¼ σ c bc xc ð3Þ

V uc ¼ τc bc xc ð4Þ

where bc is the width of concrete infill and xc is the compression depth corresponding to the average stresses. Substituting Eqs. (3) and (4) into Eq. (2)
gives

τc
¼ tanθ: ð5Þ
σc
80 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

Fig. 5. Stress analysis of the concrete infill.

From the Mohr's circle relationship shown in Fig. 6(d), we have

σ c ¼ f c cos2 θ ð6Þ

τc ¼ f c sinθ cosθ ð7Þ

where fc is the axial compressive strength of concrete. Substituting Eqs. (6) and (7) into Eqs. (3) and (4) leads to

Nuc ¼ f c bc xc cos2 θ ð8Þ

 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 81

Fig. 6. Assumed stress distributions at the coupling beam ends.

V uc ¼ f c bc xc sinθ cosθ: ð9Þ

(4) Limited by the flexural and shear deformations of the coupling beam, there is an upper bound value for the compression depth of concrete, xc.
For the composite coupling beams with common proportions, this upper bound value, xc,lim, can be taken as 0.5hb.

The analytical model is also intended to be available for concrete filled steel tubular beam-columns subject to axial force, N. For this case, the upper
bound value is taken as
8  
< N
0:5hb 1− ; Nb0
xc; lim ¼ Nu ð10Þ
:
0:5hb ; N≥0

where N is taken as positive for tension, and the axial compression strength, Nu, is given by

Nu ¼ 2 f yf bc t sf þ 2f yw hb t sw þ f c bc hc ð11Þ

where fyf and fyw are the yield strengths of the steel flange and web plates, respectively, tsf and tsw are the thicknesses of the steel flange and web
plates, respectively, and hc is the height of concrete infill, which can be considered equal to hb. Eq. (10) satisfies the boundary conditions that xc,lim
is equal to 0.5hb for N = 0 and is equal to hb for N = −Nu.

(5) From the lower bound theorem, it can be assumed that, on the premise of satisfying the structural equilibrium equations, the largest shear
force that the stress distributions in the steel plates and concrete infill can create is the shear capacity of the composite coupling beam or
other composite elements. Combined with assumption (2), we have

3 2 2
σ þ 3τ2w1 ¼ f yw ð12Þ
4 w1

3 2 2
σ þ 3τ2w2 ¼ f yw ð13Þ
4 w2
where σw1 and τw1 are the longitudinal stress and shear stress at the compression region of steel webs, respectively (σw1 is taken as positive for
compression), and σw2 and τw2 are the longitudinal stress and shear stress at the tension region of steel webs, respectively (σw2 is taken as positive
for tension).

3.2. Equilibrium equations

The axial tensile force, Nus, and the shear force, Vus, sustained by the steel plates are

Nus ¼ 2ð−σ w1 xs þ σ w2 ðhb −xs ÞÞt sw þ ðσ f2 −σ f1 Þbc t sf ð14Þ

V us ¼ 2ðτ w1 xs þ τ w2 ðhb −xs ÞÞt sw ð15Þ

where xs is the compression depth of the steel webs, and σf1 and σf2 are the longitudinal stresses of the compression and tension flanges, respectively
(σf1 is taken as positive for compression, and σf2 is taken as positive for tension). From the equilibrium of axial forces, we have

Nus −Nuc −N ¼ 0: ð16Þ

82 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

Substituting Eqs. (8) and (14) into Eq. (16) gives

2ð−σ w1 xs þ σ w2 ðhb −xs ÞÞt sw þ ðσ f2 −σ f1 Þbc t sf −f c bc xc cos2 θ−N ¼ 0: ð17Þ

The total shear force at the beam end section is

V u ¼ V us þ V uc ¼ 2ðτw1 xs þ τw2 ðhb −xs ÞÞt sw þ f c bc xc sinθ cosθ: ð18Þ

The bending moment at the beam end section is

M u ¼ σ f1 bc t sf ð0:5hb −0:5t sf Þ þ σ f2 bc t sf ð0:5hb −0:5t sf Þ þ σ w1 xs  2t sw ð0:5hb −0:5xs Þ þ σ w2 ðhb −xs Þ  2t sw  0:5xs þ Nuc ð0:5hc −0:5xc Þ
¼ 0:5ðσ f1 þ σ f2 Þbc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw þ 0:5 f c bc xc cos2 θðhc −xc Þ:
ð19Þ

The total shear force and bending moment at the beam end section satisfy

Mu ¼ V u  0:5lb : ð20Þ

Substituting Eqs. (18) and (19) into Eq. (20) gives

0:5ðσ f1 þ σ f2 Þbc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw þ 0:5f c bc xc cos2 θðhc −xc Þ ¼ ðτw1 xs  2t sw þ τw2 ðhb −xs Þ  2t sw þ f c bc xc sinθ cosθÞ  0:5lb : ð21Þ

Because hc is approximately equal to hb and hb = lbtanθ,

0:5f c bc xc hc cos2 θ ¼ 0:5f c bc xc lb sinθ cosθ: ð22Þ

Substituting Eq. (22) into Eq. (21) leads to

0:5ðσ f1 þ σ f2 Þbc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw −ðτw1 xs þ τw2 ðhb −xs ÞÞt sw lb −0:5 f c bc x2c cos2 θ ¼ 0: ð23Þ

3.3. Solving procedure

For CFSP composite coupling beams with different geometrical and material properties, one of the following cases may occur for the steel flanges:
both of the compression and tension flanges yield, only the tension flange yields, or neither of the flanges yields.
For the case in which the compression and tension flanges both yield,

σ f1 ¼ σ f2 ¼ f yf : ð24Þ

Thus, the equilibrium equations, Eqs. (17) and (23) can be simplified to

2ð−σ w1 xs þ σ w2 ðhb −xs ÞÞt sw −f c bc xc cos2 θ−N ¼ 0 ð25Þ

f yf bc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw −ðτ w1 xs þ τ w2 ðhb −xs ÞÞt sw lb −0:5f c bc x2c cos2 θ ¼ 0: ð26Þ

The assumption (5) is then equivalent to that the shear capacity is the maximum value of Vu in Eq. (18) on the premise of satisfying Eqs. (25) and
(26), where σw1, σw2, xs, and xc are the independent variables. Define the following functions:

f ðσ w1 ; σ w2 ; xs ; xc Þ ¼ 2ð−σ w1 xs þ σ w2 ðhb −xs ÞÞt sw − f c bc xc cos2 θ−N ð27Þ

g ðσ w1 ; σ w2 ; xs ; xc Þ ¼ f yf bc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw −ðτw1 xs þ τw2 ðhb −xs ÞÞt sw lb

ð28Þ
−0:5 f c bc x2c cos2 θ

Lðσ w1 ; σ w2 ; xs ; xc Þ ¼ V u þ λ1 f þ λ2 g ð29Þ

where λ1 and λ2 are the Lagrange multipliers. To determine the maximum value of Vu, we have
8   
>
> ∂L σ w1 σ w1
>
> ¼ − −2λ1 þ λ2 hb −xs þ lb xs t sw ¼ 0
>
> ∂σ w1  2τ w1 4τ w1
>
>  
>
> ∂L σ σ
>
< ¼ −
w2
þ 2λ1 þ λ2 xs þ
w2
l ðhb −xs Þt sw ¼ 0
∂σ w2 2τ w2 4τ w2 b ð30Þ
>
> ∂L
>
> ¼ ð2ðτ w1 −τw2 Þ−2λ1 ðσ w1 þ σ w2 Þ þ λ2 ððσ w1 þ σ w2 Þðhb −2xs Þ−ðτw1 −τw2 Þlb ÞÞt sw ¼ 0
>
>
> ∂xs
>
>
> ∂L
>
: ¼ f c bc cosθð sinθ−λ1 cosθ−λ2 cosθxc Þ ¼ 0:
∂xc
 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 83

From the first two equations of Eq. (30), we have

8 σ w2 σ w1
>
> ðhb −xs Þ −xs
>
> 4τ 4τ
>
> λ1 ¼  w2 w1
>
> σ w1 σ w2
>
> hb þ þ
< l
4τw1 4τw2 b
σ σ ð31Þ
>
> w1
þ
w2
>
>
>
> λ2 ¼ 2τ
 w1 2τ w2 
>
>
>
> σ w1 σ w2
: hb þ þ l
4τw1 4τ w2 b

Substituting Eq. (31) into the third equation of Eq. (30) gives

σ w1
4ðτw1 −τ w2 Þ þ ðσ w1 þ σ w2 Þ
 τ w1

xs ¼ hb : ð32Þ
σ w1 σ w2
ðσ w1 þ σ w2 Þ þ
τw1 τw2

Substituting Eqs. (31) and (32) into the fourth equation of Eq. (30) gives

σ w1 τ w1 −τw2
tanθ þ þ
 2τw1 σ w1  þ σ w2
xc ¼ hb : ð33Þ
σ w1 σ w2
þ
2τ w1 2τ w2

Therefore, the values of σw1, τw1, σw2, τw2, xs and xc can be computed by solving the system of equations in Eq. (34), and the shear capacity can be
determined by substituting these values into Eq. (18). Eq. (34) is a system of nonlinear equations, and it is unable to obtain the explicit solutions; thus,
a numerical method is required to solve these equations.

8
> 3 2
> 2
> σ w1 þ 3τ w1 −f yw ¼ 0
2
>
> 4
>
>
>3 2
> 2
>
> σ w2 þ 3τ 2w2 −f yw ¼ 0
>
>4
>
< 2ð−σ w1 xs þ σ w2 ðhb −xs ÞÞt sw − f bc xc cos2 θ−N ¼ 0
c
f yf bc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw −ðτw1 xs þ τw2 ðhb −xs ÞÞt sw lb −0:5 f c bc x2c cos2 θ ¼ 0 ð34Þ
>
>    
>
>
> x ðσ þ σ Þ σ w1 þ σ w2 − 4ðτ −τ Þ þ ðσ þ σ Þ σ w1 h ¼ 0
>
>
> s w1 w2 w1 w2 w1 w2
>
>  τw1 τw2
  τ w1 b
>
>

>
> xc σ w1 σ w2 σ w1 τ w1 −τ w2
: þ − tanθ þ þ h ¼ 0 If xc Nxc; lim ; take xc ¼ xc; lim :
2τ w1 2τ w2 2τw1 σ w1 þ σ w2 b

There may be no solutions for Eq. (34), which implies that the compression flange does not yield. Thus, the longitudinal stress of the
compression flange, σf1 , should also be treated as an independent variable. Using the similar derivations as shown in Eq. (30) through
Eq. (33) leads to
8
>
> 3 2 2 2
> 4 σ w1 þ 3τ w1 −f yw ¼ 0
>
>
>
>
>
> 3 σ 2 þ 3τ 2 −f 2 ¼ 0
>
>
>
> 4 w2 w2 yw
 
>
>
>
> 2 ð−σ x þ σ 2
w2 ðhb −xs ÞÞt sw þ f yf −σ f1 bc t sf − f c bc xc cos θ−N ¼ 0
>
> 
w1 s

>
<
0:5 f yf þ σ f1 bc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw −ðτ w1 xs þ τ w2 ðhb −xs ÞÞt sw lb −0:5f c bc x2c cos2 θ ¼ 0 ð35Þ
>    
>
> σ w1 σ w2 σ w1
>
> xs ðσ w1 þ σ w2 Þ þ − 4ðτw1 −τ w2 Þ þ ðσ w1 þ σ w2 Þ h ¼0
>
> τw1 τw2 τ w1 b
>
>   
>
> σ w1 σ w2 σ w1 τw1 −τ w2

>
> xc þ − tanθ þ þ hb ¼ 0 If xc Nxc; lim ; take xc ¼ xc; lim
>
> σ þ σ
>
> 2τ w1 2τ w2 2τ w1 w1 w2
>
> σ σ
>
: ðxs −t sf Þ w2 þ ðxs −t sf þ hb Þ w1 ¼ 0:
τ w2 τw1

The values of σf1, σw1, τw1, σw2, τw2, xs and xc can be computed by solving the system of equations in Eq. (35), and the shear capacity can then be
determined. There may also be no solutions for Eq. (35). This lack of solutions may occur when the thickness of steel flanges is much larger than that
of steel webs, i.e., both the compression and tension flanges will not yield. For this case, the solution is

8
>
> σ w1 ¼ σ w2 ¼ 0 pffiffiffi
<
τ w1 ¼ τw2 ¼ f yw = 3
ð36Þ
>
> xs ¼ 0
:
xc ¼ xc; lim
84 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

Because the coupling beams are primarily subject to shear force and bending moment, the case that the tension flange does not yield due to
high axial compressive force is not included in this model. However, the equations can be similarly derived by taking σf2 as an independent
variable.

3.4. Verification

3.4.1. CFSP composite coupling beam tests

For the CFSP composite coupling beams tested by the authors, the sectional dimensions, bb × hb, were all 150 mm × 300 mm, and the values
of other parameters are shown in Table 1. The shear capacities of the coupling beam specimens were calculated using the above analytical
model, and the calculated results and the ratios of the calculated shear capacities, Vu,cal, to the measured shear capacities, Vu,tes, are shown
in Table 1. In the calculation, the steel plate strengths were calculated using the constitutive model proposed by Shi et al. [19] to account for
the strain hardening. The friction force of the test setup was deducted from the measured shear capacities [14]. As shown in Table 1, the cal-
culated shear capacities agree well with the measured shear capacities, which indicates that the analytical model can predict the shear capacity
with adequate accuracy. As the span-to-height ratio, lb./hb, increases, the ratio of the bending moment to the shear force increases, which re-
sults in the increase of σw1/fyw and σw2/fyw and the decrease of τw1/fyw and τw2/fyw. As the ratio of tsf/tsw increases, greater bending moment
resistance can be provided by the steel flanges, so the values of τw1/fyw and τw2/fyw increase to provide higher shear force resistance. For spec-
imen CFSCB-3, the calculated results show that the steel webs are all in tension, thus enabling a larger axial compressive force to be sustained
by the concrete infill, which results in a greater shear force resistance provided by the concrete infill. The concrete compression depth, xc, is
larger for the specimens with smaller span-to-height ratio, in which the shear force dominates.

3.4.2. Concrete filled square steel tubular beam-column tests

Several concrete filled square steel tubular beam-columns subjected to different levels of axial compressive force were tested by Tomii and
Sakino [20], and Sakino and Ishibashi [21]. The test setup was similar to that for testing CFSP composite coupling beams as shown in Fig. 3. The

Table 2
Calculated and test results of the concrete filled square steel tubular beam-columns.

Reference Specimen Specimen parameters Calculated results Vu,tes Vu,cal/Vu,tes

designation (kN)
lb/hb tsa fyb fc N (kN) σw1/fyw τw1/fyw σw2/fyw τw2/fyw xs/hb xc/hb Vu,cal
(mm) (N/mm2) (N/mm2) (kN)

Tomii and Sakino [20] I0-20 4.0 2.29 194 38.5 0 0.25 0.56 1.11 0.16 0.22 0.21 40.1 40.8 0.98
I0-30 6.0 2.29 194 36.4 0 0.73 0.45 1.13 0.11 0.20 0.20 27.2 27.6 0.99
I-23 4.0 2.29 194 45.3 −184 1.13 0.13 1.14 0.10 0.45 0.48 58.5 68.2 0.86
II-12 2.0 2.24 310 30.9 −109 0.75 0.44 0.96 0.32 0.42 0.60 122 128 0.95
II-13 2.0 2.24 310 30.9 −164 0.96 0.32 0.90 0.36 0.53 0.65 124 128 0.97
II-15 2.0 2.24 310 30.9 −273 1.14 0.10 0.89 0.37 0.78 0.75 111 119 0.94
II-23 4.0 2.24 310 23.4 −142 1.08 0.21 1.04 0.25 0.54 0.60 63.3 70.0 0.90
II-25 4.0 2.24 310 23.4 −254 1.15 0.07 1.03 0.26 0.79 0.75 54.7 60.3 0.91
II-32 6.0 2.20 303 24.2 −97 1.09 0.20 1.11 0.17 0.46 0.49 42.3 45.0 0.94
II-33 6.0 2.20 303 24.2 −146 1.12 0.14 1.10 0.17 0.55 0.58 42.2 45.3 0.93
II-35 6.0 2.20 303 24.2 −243 1.15 0.05 1.12 0.14 0.74 0.75 37.6 41.1 0.91
III-12 2.0 2.98 298 23.9 −108 0.72 0.45 0.72 0.45 0.50 0.60 134 149 0.90
III-13 2.0 2.98 298 23.9 −162 0.89 0.37 0.59 0.50 0.57 0.65 134 149 0.90
III-15 2.0 2.98 298 23.9 −270 1.09 0.20 0.18 0.57 0.74 0.75 119 145 0.82
III-22 4.0 2.97 297 23.5 −107 1.02 0.28 1.03 0.26 0.48 0.55 76.1 86.6 0.88
III-23 4.0 2.97 297 23.5 −161 1.08 0.21 1.02 0.28 0.57 0.63 75.4 87.2 0.86
III-25 4.0 2.97 297 23.5 −268 1.14 0.11 0.95 0.33 0.76 0.75 67.3 81.3 0.83
III-32 6.0 2.97 286 23.5 −107 1.09 0.19 1.10 0.18 0.49 0.52 50.4 55.9 0.90
III-33 6.0 2.97 286 23.5 −161 1.12 0.14 1.09 0.19 0.58 0.60 49.8 54.8 0.91
III-35 6.0 2.97 286 23.5 −268 1.15 0.07 1.07 0.21 0.76 0.75 44.0 52.0 0.85
IV-12 2.0 4.22 292 22.9 −130 0.72 0.45 0.56 0.51 0.53 0.60 168 196 0.86
IV-13 2.0 4.22 292 22.9 −196 0.87 0.38 0.39 0.54 0.59 0.65 165 196 0.84
IV-23 4.0 4.13 299 22.7 −195 1.07 0.21 0.95 0.33 0.61 0.65 95.3 98.0 0.97
IV-25 4.0 4.13 299 22.7 −325 1.13 0.12 0.76 0.44 0.78 0.75 83.2 88.0 0.95
IV-32 6.0 4.25 294 23.2 −131 1.09 0.18 1.08 0.20 0.53 0.56 67.6 69.8 0.97
IV-33 6.0 4.25 294 23.2 −198 1.12 0.14 1.06 0.22 0.61 0.64 66.1 71.0 0.93
IV-35 6.0 4.25 294 23.2 −327 1.14 0.08 0.99 0.30 0.78 0.75 58.0 61.0 0.95
Sakino and Ishibashi [21] IIS1.5-2 3.0 2.20 291 20.5 −87.2 0.92 0.35 0.99 0.30 0.46 0.57 74.6 81.8 0.91
IIS1.5-3 3.0 2.20 291 20.5 −130 1.03 0.26 0.96 0.32 0.55 0.64 74.5 81.4 0.91
IIS1.5-5 3.0 2.20 291 20.5 −218 1.13 0.12 0.82 0.40 0.78 0.73 65.9 78.3 0.84
IIIS1.5-2 3.0 2.98 290 22.0 −105 0.93 0.34 0.95 0.33 0.49 0.60 94.6 108 0.88
IIIS1.5-3 3.0 2.98 290 22.0 −158 1.02 0.27 0.88 0.37 0.58 0.64 93.6 112 0.84
IIIS1.5-3 3.0 2.98 290 22.0 −263 1.12 0.14 0.65 0.48 0.78 0.74 82.2 99.3 0.83
IVS1.5-2 3.0 4.25 316 22.0 −137 0.94 0.34 0.85 0.39 0.54 0.60 133 158 0.84
IVS1.5-3 3.0 4.25 316 22.0 −207 1.02 0.27 0.75 0.44 0.61 0.65 130 157 0.83
IVS1.5-5 3.0 4.25 316 22.0 −344 1.11 0.16 0.31 0.56 0.77 0.74 117 145 0.81
Average 0.90
Standard deviation 0.05
a
ts is the thickness of steel tube plate.
b
fy is the yield strength of steel tube plate.
 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 85

ends of the beam columns were fixed and the point of inflection was located in the middle of the span, as in the case of coupling beams, so the
analytical model can be directly used to calculate the shear capacities of the beam-column specimens. The steel tubes were 100 mm wide for all
of the specimens. The specimen parameters and calculated results are listed in Table 2. The measured shear capacity, Vu,tes, is taken as the max-
imum shear force for chord rotations smaller than 0.02 rad before the strain hardening was expected to be significant. The ratios of the calcu-
lated shear capacities to the measured shear capacities range from 0.81 to 0.99, with an average value of 0.90.

3.4.3. Simply supported concrete filled steel tubular beam-column tests

The analytical model can also be used to calculate the load-carrying capacities of simply supported rectangular concrete filled steel tubular (CFST)
beam-columns subjected to concentrated lateral force, as shown in Fig. 7. For this case, the beam span in the analytical model, lb., should be taken as 2
times the length of the test span, 2l1, as shown in Fig. 7. The specimen parameters [22–29] and the calculated results of the existing tests are shown in
Table 3. For the test of Lu and Kennedy [22], the steel strengths, fy, were taken as the stresses at the strain value of 0.018, which approximately
corresponded to the maximum shear forces. In the tests of Han [24] and Han et al. [26], the compressive strengths of concrete were tested using
cubic coupons. The axial compressive strength, fc, used in the calculation was taken as 0.76fcu according to the Chinese building code [30], where
fcu is the cubic compressive strength. The ratios of the calculated shear capacities to the measured shear capacities have an average value of 0.94
and a standard deviation of 0.09. Therefore, the analytical model also has a high accuracy for predicting the load-carrying capacities of the simply
supported CFST beam-columns.

4. Computational formulas

4.1. Critical variable determination

Four coefficients, k11, k12, k21 and k22 are introduced to express the relationships between the steel web stresses and the yield strength of
steel web plates:

σ w1 ¼ k11 f yw ; τw1 ¼ k12 f yw
ð37Þ
σ w2 ¼ k21 f yw ; τw2 ¼ k22 f yw

and denote

F 1 ¼ f c bc hb cos2 θ; F 2 ¼ f yw hb t sw ; F 3 ¼ f yf bc t sf : ð38Þ

Fig. 7. Simply supported concrete filled rectangular steel tubular beam-columns subjected to concentrated lateral force.
86 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

Table 3
Calculated and test results of the simply supported CFST beam-columns.

Reference Specimen Specimen parameters Calculated results Vu,tes Vu,cal/Vu,tes

designation (kN)
bb × hb l1/hb tsa fyb fc N (kN) σw1/fyw τw1/fyw σw2/fyw τw2/fyw xs/hb xc/hb Vu,cal
(mm × mm) (mm) (N/mm2) (N/mm2) (kN)

Lu and CB12 152 × 152 1.5 4.43 450 47 0 0.47 0.53 1.05 0.24 0.32 0.37 313 313 1.00
Kennedy CB13 152 × 152 3.0 4.43 450 42.8 0 0.96 0.32 1.12 0.14 0.31 0.32 167 162 1.03
[22] CB15 152 × 152 5.1 4.43 450 41.2 0 1.08 0.20 1.14 0.09 0.31 0.31 102 92.8 1.10
CB22 152 × 152 1.6 8.95 492 46.9 0 0.73 0.44 0.99 0.29 0.39 0.48 603 621 0.97
CB31 152 × 254 1.0 6.17 462 46.7 0 0.63 0.48 1.03 0.26 0.35 0.48 807 807 1.00
CB33 152 × 254 3.0 6.17 462 45.2 0 1.07 0.22 1.13 0.11 0.34 0.36 307 275 1.12
CB35 152 × 254 5.0 6.17 462 44.3 0 1.12 0.13 1.15 0.07 0.34 0.35 186 163 1.14
CB41 152 × 254 1.0 9.04 433 46.2 0 0.69 0.46 0.98 0.30 0.40 0.50 1035 1087 0.95
CB45 152 × 254 5.0 9.04 433 43.8 0 1.13 0.12 1.14 0.07 0.38 0.38 241 221 1.09
CB52 254 × 152 1.5 6.17 462 47.1 0 −0.42 0.54 0.87 0.38 0.41 0.38 577 616 0.94
CB53 254 × 152 3.0 6.17 462 42.1 0 0.68 0.46 1.08 0.20 0.31 0.31 323 317 1.02
CB55 254 × 152 5.1 6.17 462 40.5 0 0.98 0.30 1.13 0.13 0.29 0.30 199 186 1.07
Uy [23] HS6 126 × 126 4.8 3 300 50 0 1.00 0.29 1.14 0.08 0.21 0.21 44 46.5 0.94
HS12 156 × 156 4.8 3 300 50 0 0.97 0.32 1.15 0.07 0.18 0.18 55 56.5 0.97
NS6 186 × 186 4.8 3 300 32 0 1.01 0.28 1.14 0.08 0.21 0.22 64 69.6 0.92
NS12 246 × 246 4.9 3 300 38 0 0.92 0.35 1.15 0.07 0.16 0.16 87 86.3 1.01
NS18 306 × 306 4.9 3 300 38 0 0.85 0.39 1.15 0.07 0.14 0.14 110 102 1.08
Han [24] RB1-1 120 × 120 2.1 3.84 330 20.7 0 0.85 0.39 1.06 0.23 0.37 0.41 114 117 0.97
RB2-1 120 × 120 2.1 3.84 330 26.8 0 0.81 0.41 1.07 0.21 0.34 0.38 116 121 0.96
RB2-2 120 × 120 2.1 3.84 330 26.8 0 0.81 0.41 1.07 0.21 0.34 0.38 116 129 0.90
RB2-3 120 × 120 2.1 3.84 330 26.8 0 0.81 0.41 1.07 0.21 0.34 0.38 116 127 0.92
RB3-1 120 × 120 2.1 5.86 321 23.8 0 0.89 0.36 1.05 0.24 0.39 0.44 161 164 0.98
RB3-2 120 × 120 2.1 5.86 321 23.8 0 0.89 0.36 1.05 0.24 0.39 0.44 161 166 0.97
RB4-1 120 × 120 2.1 5.86 321 30.4 0 0.87 0.38 1.07 0.22 0.37 0.42 164 166 0.99
RB4-2 120 × 120 2.1 5.86 321 30.4 0 0.87 0.38 1.07 0.22 0.37 0.42 164 170 0.96
RB5-1 120 × 150 1.7 2.93 294 26.2 0 0.69 0.46 1.08 0.20 0.30 0.34 115 126 0.91
RB5-2 120 × 150 1.7 2.93 294 26.2 0 0.69 0.46 1.08 0.20 0.30 0.34 115 126 0.91
RB6-1 90 × 120 2.1 2.93 294 26.2 0 0.91 0.35 1.10 0.17 0.33 0.36 69.9 84.4 0.83
RB6-2 90 × 120 2.1 2.93 294 26.2 0 0.91 0.35 1.10 0.17 0.33 0.36 69.9 80.8 0.86
RB7-1 90 × 150 1.7 2.93 294 26.2 0 0.88 0.37 1.10 0.18 0.33 0.38 97.9 114 0.86
RB7-2 90 × 150 1.7 2.93 294 26.2 0 0.88 0.37 1.10 0.18 0.33 0.38 97.9 118 0.83
RB8-1 60 × 120 2.1 2.93 294 26.2 0 1.03 0.26 1.11 0.16 0.37 0.41 56.0 73.6 0.76
RB8-2 60 × 120 2.1 2.93 294 26.2 0 1.03 0.26 1.11 0.16 0.37 0.41 56.0 71.2 0.79
Gho and B01 150 × 150 2.4 4.88 438 56.3 0 0.83 0.40 1.11 0.16 0.29 0.31 221 224 0.99
Liu [25] B02 150 × 150 2.4 4.87 438 56.3 0 0.83 0.40 1.11 0.16 0.29 0.31 221 241 0.92
B03 150 × 150 2.4 4.92 438 87.5 0 0.71 0.46 1.12 0.14 0.24 0.25 232 270 0.86
B04 150 × 150 2.4 4.84 438 87.5 0 0.70 0.46 1.12 0.14 0.24 0.25 229 278 0.82
B05 150 × 200 1.8 5.94 495 56.3 0 0.85 0.39 1.09 0.19 0.33 0.37 446 468 0.95
B06 150 × 200 1.8 5.92 495 56.3 0 0.85 0.39 1.09 0.19 0.33 0.37 445 482 0.92
B07 150 × 200 1.8 5.93 495 87.5 0 0.77 0.43 1.11 0.16 0.28 0.31 468 498 0.94
B08 150 × 200 1.8 5.82 495 87.5 0 0.77 0.43 1.11 0.16 0.27 0.30 460 480 0.96
B09 149 × 250 1.5 5.91 409 70.6 0 0.75 0.44 1.10 0.17 0.28 0.33 539 590 0.91
B10 149 × 250 1.5 5.89 409 70.6 0 0.75 0.44 1.10 0.17 0.28 0.33 538 582 0.92
B11 149 × 250 1.5 5.88 409 90.9 0 0.68 0.47 1.11 0.15 0.25 0.29 554 602 0.92
B12 149 × 250 1.5 5.85 409 90.9 0 0.68 0.47 1.11 0.15 0.25 0.29 552 596 0.93
Han et al. SVA-1 100 × 100 3.5 1.9 282 66.7 0 0.64 0.48 1.14 0.09 0.16 0.15 29.3 30.9 0.95
[26] SVA-2 100 × 100 3.5 1.9 282 66.7 0 0.64 0.48 1.14 0.09 0.16 0.15 29.3 28.5 1.03
SSCA-1 100 × 100 3.5 1.9 282 66.7 0 0.64 0.48 1.14 0.09 0.16 0.15 29.3 29.5 0.99
SVB-1 200 × 200 1.75 1.9 282 66.7 0 −0.74 0.44 1.08 0.20 0.31 0.09 119 120 0.99
SVB-2 200 × 200 1.75 1.9 282 66.7 0 −0.74 0.44 1.08 0.20 0.31 0.09 119 157 0.76
SSCB-1 200 × 200 1.75 1.9 282 66.7 0 −0.74 0.44 1.08 0.20 0.31 0.09 119 162 0.73
SB1-1 140 × 140 1.5 3 235 50.1 0 −0.37 0.55 1.08 0.20 0.27 0.21 117 152 0.77
SB1-2 140 × 140 1.5 3 235 50.1 0 −0.37 0.55 1.08 0.20 0.27 0.21 117 131 0.89
SB2-1 140 × 140 3 3 235 50.1 0 0.63 0.48 1.13 0.11 0.18 0.18 61.5 70 0.88
SB2-2 140 × 140 3 3 235 50.1 0 0.63 0.48 1.13 0.11 0.18 0.18 61.5 61.7 1.00
SB3-1 140 × 140 6 3 235 50.1 0 1.02 0.28 1.15 0.06 0.17 0.17 31.2 36 0.87
SB3-2 140 × 140 6 3 235 50.1 0 1.02 0.28 1.15 0.06 0.17 0.17 31.2 35 0.89
SB4-1 180 × 180 1.25 3 235 50.1 0 −0.75 0.44 1.01 0.28 0.39 0.17 179 167 1.07
SB4-2 180 × 180 1.25 3 235 50.1 0 −0.75 0.44 1.01 0.28 0.39 0.17 179 192 0.93
SB5-1 180 × 180 2.5 3 235 50.1 0 0.20 0.57 1.13 0.12 0.18 0.15 96.6 84.2 1.15
SB5-2 180 × 180 2.5 3 235 50.1 0 0.20 0.57 1.13 0.12 0.18 0.15 96.6 92.7 1.04
SB6-1 180 × 180 5 3 235 50.1 0 0.88 0.37 1.15 0.06 0.14 0.14 49.2 55.3 0.89
SB6-2 180 × 180 5 3 235 50.1 0 0.88 0.37 1.15 0.06 0.14 0.14 49.2 51.7 0.95
Shawkat CFT-S1 102 × 150 1 3.4 380 31 0 0.52 0.52 1.01 0.28 0.36 0.50 232 226 1.03
et al. CFT-S2 102 × 150 2 3.4 380 31 0 0.93 0.34 1.10 0.18 0.34 0.38 129 141 0.91
[27] CFT-S3 102 × 150 3 3.4 380 31 0 1.05 0.24 1.13 0.12 0.34 0.35 88 91 0.97
Guler et al. BF2.5–1 80 × 80 5.6 2.51 288 150 0 0.88 0.37 1.15 0.05 0.12 0.12 19.8 22.0 0.90
[28] BF2.5–2 80 × 80 5.6 2.51 288 150 0 0.88 0.37 1.15 0.05 0.12 0.12 19.8 23.2 0.85
BF2.5–3 80 × 80 5.6 2.52 288 150 0 0.89 0.37 1.15 0.05 0.12 0.12 19.8 20.6 0.96
BF3-1 80 × 80 5.6 3.02 277 152 0 0.94 0.34 1.15 0.05 0.13 0.13 22.5 23.7 0.95
BF3-2 80 × 80 5.6 3.04 277 152 0 0.94 0.34 1.15 0.05 0.13 0.13 22.6 24.2 0.93
BF3-3 80 × 80 5.6 3.01 277 152 0 0.94 0.34 1.15 0.05 0.13 0.13 22.4 24.9 0.90
Cai et al. C1 600 × 600 0.9 20 460 56.2 0 −0.16 0.57 0.94 0.33 0.37 0.5 8659 9057 0.96
 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 87

Table 3 (continued)

Reference Specimen Specimen parameters Calculated results Vu,tes Vu,cal/Vu,tes

designation (kN)
bb × hb l1/hb tsa fyb fc N (kN) σw1/fyw τw1/fyw σw2/fyw τw2/fyw xs/hb xc/hb Vu,cal
(mm × mm) (mm) (N/mm2) (N/mm2) (kN)

[29] C2 600 × 600 0.9 20 460 56.2 −4433 0.42 0.54 0.85 0.39 0.42 0.55 9410 10,154 0.93
C3 600 × 600 0.65 20 460 56.2 0 −0.48 0.53 0.59 0.50 0.48 0.5 10,193 11,192 0.91
C4 600 × 600 0.65 20 460 56.2 −4433 0.10 0.58 0.46 0.53 0.48 0.55 11,109 11,968 0.93
Average 0.94
Standard 0.09
deviation
a
ts is the thickness of steel tube plate.
b
fy is the yield strength of steel tube plate.

Since hb − tsf ≈ hb and hb/lb. = tanθ, Eq. (34) can be rewritten as

8
>
> 3 2 2
> k11 þ 3k12 −1 ¼ 0
>
>
> 4
>
>3 2
>
> 2
k þ 3k22 −1 ¼ 0
>
>
>
> 421  
>
> xs F 1 xc N
>
>
> 2 −k11 xs þ k21 1− h
<
− −
F 2 hb F 2
¼0
b
       ð39Þ
> F x x x x 1 F 1 xc 2
> 3 þ ðk11 þ k21 Þ s 1− s − k12 s þ k22 1− s
> −0:5 ¼0
>
> F
>
> x2  hb hb  hb hb tanθ
 F 2 hb
>
> k11 k21 k11
>
>
s
ðk11 þ k21 Þ þ − 4ðk12 −k22 Þ þ ðk11 þ k21 Þ ¼0
>
> hb  k
>
> 12 k22   k12 
>x
> k11 k21 k11 k12 −k22 xc xc; lim xc xc; lim
>
> c
þ − tanθ þ þ ¼ 0 If N ; take ¼ :
:
hb 2k12 2k22 2k12 k11 þ k21 hb hb hb hb

For a given composite coupling beam, k11, k12, k21, k22, xs/hb and xc/hb are unknowns in Eq. (39). The solution of Eq. (39) is related only to the
values of F1/F2, F3/F2, N/F2 and tanθ. Similarly, the equations for the case in which the compression flange does not yield can be rewritten as

8
>
> 3 2 2
>
> 4 k11 þ 3k12 −1 ¼ 0
>
>
>
>3 2
>
> 2
k þ 3k22 −1 ¼ 0
>
>
> 421
>   !
>
>
>
> xs σ f1 F 3 F 1 xc N
>
> 2 −k x þ k 1− þ 1− − − ¼0
>
>
11 s 21
h f F F h F
>
> ! b yf 2 2 b 2
>
<       
σ F3 xs xs xs xs 1 F 1 xc 2
0:5 1 þ f1 þ ðk11 þ k21 Þ 1− − k12 þ k22 1− −0:5 ¼0 ð40Þ
>
> f yf F 2 hb hb hb hb tanθ F 2 hb
>
>    
>
> xs k11 k21 k11
>
> ðk11 þ k21 Þ þ − 4ðk12 −k22 Þ þ ðk11 þ k21 Þ ¼0
>
>
> hb 
> k12 k22   k12 
>
> k12 −k22
>
> xc k11 k21 k11 xc xc; lim xc xc; lim
>
> þ − tanθ þ þ ¼ 0 If N ; take ¼
>
> hb 2k12  2k22 
 2k12 k11 þ k21 hb hb hb hb
>
>
>
> xs t sf σ w2 xs t sf σ w1
>
: − þ − þ1 ¼ 0:
hb hb τw2 hb hb τ w1

Because tsf/hb ≈ 0, the solution of Eq. (40) is also related only to the values of F1/F2, F3/F2, N/F2 and tanθ. The solution for the case in which neither
of the flanges yield, Eq. (36), can be rewritten as

8
> k11 ¼ k21 ¼ 0 pffiffiffi
>
>
>
< kx12 ¼ k22 ¼ 1= 3
>
s
¼0 ð41Þ
> hb
>
>
> x x
>
: c ¼ c; lim
hb hb

The shear capacity of the composite coupling beam can be expressed as

  
xs xs xc
V u ¼ 2 k12 þ k22 1− F 2 þ F 1 tanθ : ð42Þ
hb hb hb

Therefore, the value of Vu/F2 is only related to the values of F1/F2, F3/F2, N/F2 and tanθ.
88 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90

4.2. Design equations

For the coupling beam with large span-to-height ratio, the shear capacity is primarily dominated by the flexural strength at the beam end. The
flexural dominated capacity, Vu,fle, can be approximately calculated by (see derivation in Appendix A)

 
2F 2 þ F 1 þ N
V u;fle ¼ 2 tanθ F 3 þ ð F 2 −0:5N Þ : ð43Þ
4 F2 þ F1

For the case in which neither of the steel flanges yields, the shear capacity is dominated by the shear strength of the coupling beam. The shear
dominated capacity, Vu,sh, can be obtained by substituting Eq. (41) into Eq. (42)

2 pffiffiffi xc; lim

V u;sh ¼ 3F2 þ F 1 tanθ: ð44Þ
3 hb

For composite coupling beams with bb = 200 mm, hb = 400 mm, tsf = 16 mm, tsw = 4 mm, fy = 235 N/mm2, fc = 30 N/mm2 and N = 0, the
variation of the shear capacity with the beam span-to-height ratio, lb./hb, is shown in Fig. 8. When the span-to-height ratio is close to 1 or larger
than 6, the shear capacities of these composite coupling beams can be accurately predicted by


V u ¼ min V u;sh ; V u;fle : ð45Þ

When the span-to-height ratio is between 1.5 and 6, Eq. (45) will overestimate the shear capacities of these composite coupling beams, and the
overestimate reaches the maximum value when Vu,sh and Vu,fle are equal. This phenomenon is caused by the flexural-shear interaction.
To investigate the influence of flexural-shear interactions to the shear capacity of coupling beam, a series of composite coupling beams covering
the common parameter values was designed. The axial force, N, was taken as 0. The beam width was fixed as 200 mm, and the values of hb/bb, lb./hb,
tsw/bb, tsf/hb, fy and fc were varied parametrically according to Table 4 while satisfying the limitations of the width-to-thickness ratios: hb/tsw ≤ 100
and bb/tsf ≤ 50. In total, 8400 composite coupling beams were obtained. Fig. 9 shows the relationship between the ratios of the shear capacities
calculated from Eq. (45), Vu2, to the shear capacities calculated by the analytical model, Vu1 (referred to as “exact solution” hereafter), and the
corresponding ratios of Vu,fle to Vu,sh for the designed composite coupling beams. Based on Fig. 9, a reduction coefficient, αv, which captures the
relationship between Vu1/Vu3 and Vu,fle/Vu,sh, can be introduced to consider the effect of flexural-shear interaction:

8



< 1
V u;fle =V u;sh −1
≤0:5

αV ¼ −0:3
V u;fle =V u;sh −1
þ 1:15


 ð46Þ
:
V u;fle =V u;sh −1
N0:5
1

Therefore, the design equations for the shear capacity become


V u ¼ α V  min V u;sh ; V u;fle : ð47Þ

The shear capacities calculated from Eq. (47), Vu3, and the exact shear capacity, Vu1, are compared in Fig. 10 for the 8400 composite coupling
beams. As shown in Fig. 10, Vu3 and Vu1 are in good agreement, and the relative errors are mostly below 10%. Although Eq. (46) is determined
from composite coupling beams not subjected to axial force, it is also applicable to the case that the axial force exists. As shown in Fig. 11, Vu3 and
Vu1 are also in good agreement for composite coupling beams subjected to different levels of axial force.

Fig. 8. Variation of the shear capacity with the span-to-height ratio.

 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90 89

Table 4
Values for each parameter.

Parameter Value

bb (mm) 200
hb/bb 2, 2.5, 3, 3.5, 4
lb/hb 1, 1.25, 1.5, 2, 2.5, 3, 3.5, 4, 5, 6
tsw/bb 0.01, 0.015, 0.02, 0.025, 0.03, 0.035, 0.04
tsf/hb 0.01, 0.015, 0.02, 0.025, 0.03, 0.035, 0.04
fy (N/mm2) 235, 345
fc (N/mm2) 20, 30, 40, 50

Fig. 9. Relationship between Vu2/Vu1 and Vu,fle/Vu,sh.

5. Conclusions where the values of k11 and k21 are adjustable, and the values of k21 and
k22 are dependent on k11 and k21 through the von Mises yield criteria, as
An analytical model for the shear capacity of concrete-filled steel shown in the first two equations in Eq. (39). From the equilibrium of
plate composite coupling beams was proposed based on the force mech- axial forces, we have
anism of composite coupling beams. It was found that the shear capacity
was primarily related to the values of F2, F1/F2 and F3/F2 (F1, F2 and F3 are  
xs xs
defined in Eq. (38)) and the beam height-to-span ratio, tanθ. The shear 2 1−2 k11 F 2 − F 1 −N ¼ 0: ð49Þ
hb hb
capacity of the composite coupling beam may be governed by flexural
strength, shear strength, or flexural-shear strength, depending on the
material and geometrical properties of the coupling beam. A unified Eq. (49) can be rewritten as
equation, Eq. (47), incorporating the effect of flexural-shear interaction
was developed based on the analytical model.
xs 2k11 F 2 −N
¼ : ð50Þ
hb 4k11 F 2 þ F 1
Acknowledgments

The work reported in this paper was supported by the National Key Therefore, the flexural strength at the beam end is
Technology R&D Program of China (2011BAJ09B01), National Science
Fund of China (51178246) and Tsinghua University Initiative Scientific M u ¼ f yf bc t sf ðhb −t sf Þ þ ðσ w1 þ σ w2 Þxs ðhb −xs Þt sw þ 0:5 f c bc xc cos2 θðhc −xc Þ
Research Program (2010Z03078). Hong-Song Hu was supported by an  
2k11 F 2 −N 2k11 F 2 −N
overseas fellowship from the Japan Society for the Promotion of Science ¼ F 3 hb þ 2k11 F 2 hb 1−
4k11 F 2 þ F 1 4k11 F 2 þ F 1
(JSPS, Award Number P 14374). The support is gratefully acknowledged.   ð51Þ
2k11 F 2 −N 2k11 F 2 −N
The writers are also grateful to Prof. M. R. Eatherton of Virginia Tech for þ 0:5 F 1 1−
4k11 F 2 þ F 1 4k11 F 2 þ F 1
reviewing the manuscript and providing the writers with important 2k11 F 2 þ F 1 þ N
comments on the content of this paper. ¼ F 3 hb þ ðk11 F 2 −0:5N Þhb :
4k11 F 2 þ F 1

Appendix A. Approximate solution for the flexural dominated

The shear capacity corresponding to the above flexural strength is
capacity
 
To calculate the flexural dominated capacity of composite coupling Mu 2k11 F 2 þ F 1 þ N
V u;fle ¼ ¼ 2 tanθ F 3 þ ðk11 F 2 −0:5NÞ : ð52Þ
beams, the following can be assumed: 0:5lb 4k11 F 2 þ F 1

8 Vu, fle should also satisfy

< k11 ¼ k21
k ¼ k22 ð48Þ
: 12 V u;fle ≤V u : ð53Þ
xs ¼ xc
90 H.-S. Hu et al. / Journal of Constructional Steel Research 118 (2016) 76–90
Fig. 10. Comparison of Vu3 and Vu1.
Fig. 11. Comparison of Vu3 and Vu1 for composite coupling beams subjected to axial force.
Substituting Eqs. (43) and (42) into Eq. (53) and using the relation-
ships in Eqs. (48) and (50) lead to
2 sections, Can. J. Civ. Eng. 21 (1) (1994) 111–130.
lb 1 2F 1 F 3 þ 8k11 F 2 F 3 þ 4k11 F 22 −N 2
¼ ≥
hb tanθ 2k12 F 2 ð4k11 F 2 þ F 1 Þ
2 [24] L.H. Han, Flexural behaviour of concrete-filled steel tubes, J. Constr. Steel Res. 60 (2)
2ð F 1 =F 2 Þð F 3 =F 2 Þ þ 8k11 F 3 =F 2 þ 4k11 −ðN=F 2 Þ2
¼ :
8k11 k12 þ 2k12 F 1 =F 2
Therefore, for the composite coupling beams that satisfy Eq. (54),
the shear capacity of the coupling beam can be approximately calcu-
lated by Eq. (52). Eq. (52) will underestimate the shear capacity of
the coupling beam. For a larger value of k11, the shear capacity calcu-
lated by Eq. (52) is closer to the exact solution, whereas the prereq-
uisite given by Eq. (54) is stricter. When taking k11 and k21 as 1.0, the
shear capacities calculated using Eq. (52) were all not smaller than
91.4% of the corresponding exact solutions for the 8400 investigated
composite coupling beams. Therefore, the value of k 11 in Eq. (48)
through Eq. (54) can be taken as 1.0 to obtain solutions with ade-
quate accuracy.
References
[1] J.G. Nie, H.S. Hu, M.R. Eatherton, Concrete filled steel plate composite coupling
beams: experimental study, J. Constr. Steel Res. 94 (2014) 49–63.
[2] W.Y. Lam, R. Su, H.J. Pam, Experimental study on embedded steel plate composite
coupling beams, J. Struct. Eng. ASCE 131 (8) (2005) 1294–1302.
[3] R.K.L. Su, W.Y. Lam, H.J. Pam, Behaviour of embedded steel plate in composite
coupling beams, J. Constr. Steel Res. 64 (10) (2008) 1112–1128.
[4] R.K.L. Su, W.Y. Lam, A unified design approach for plate-reinforced composite
coupling beams, J. Constr. Steel Res. 65 (3) (2009) 675–686.
[5] B. Gong, B.M. Shahrooz, Concrete–steel composite coupling beams. I: component
testing, J. Struct. Eng. ASCE 127 (6) (2001) 625–631.
[6] B. Gong, B.M. Shahrooz, Concrete–steel composite coupling beams. II: subassembly
testing and design verification, J. Struct. Eng. ASCE 127 (6) (2001) 632–638.
[7] B. Gong, B.M. Shahrooz, Steel–concrete composite coupling beams—behavior and
design, Eng. Struct. 23 (11) (2001) 1480–1490.
[8] S. El-Tawil, C.M. Kuenzli, M. Hassan, Pushover of hybrid coupled walls. I: design and
modeling, J. Struct. Eng. ASCE 128 (10) (2002) 1272–1281.
[9] S. El-Tawil, C.M. Kuenzli, Pushover of hybrid coupled walls. II: analysis and behavior,
J. Struct. Eng. ASCE 128 (10) (2002) 1282–1289.
[10] M. Hassan, S. El-Tawil, Inelastic dynamic behavior of hybrid coupled walls, J. Struct.
Eng. ASCE 130 (2) (2004) 285–296.
[11] S. El-Tawil, K.A. Harries, P.J. Fortney, B.M. Shahrooz, Y. Kurama, Seismic design of
hybrid coupled wall systems: state of the art, J. Struct. Eng. ASCE 136 (7) (2010)
755–769.
[12] J.G. Nie, H.S. Hu, J.S. Fan, M.X. Tao, S.Y. Li, F.J. Liu, Experimental study on seismic
behavior of high-strength concrete filled double-steel-plate composite walls, J.
Constr. Steel Res. 88 (2013) 206–219.
[13] H.S. Hu, J.G. Nie, M.R. Eatherton, Deformation capacity of concrete-filled steel plate
composite shear walls, J. Constr. Steel Res. 103 (2014) 148–158.
[14] H.S. Hu, J. Nie, M.R. Eatherton, Internal force and deformation of concrete-filled steel
plate composite coupling beams, J. Constr. Steel Res. 92 (2014) 150–163.
[15] ACI Committee, Building Code Requirements for Structural Concrete (ACI 318-11)
and Commentary (ACI 318R-11), American Concrete Institute (ACI), Farmington
Hills, MI, 2011.
[16] American Institute of Steel Construction (AISC), ANSI/AISC 341-10 Seismic
Provisions for Structural Steel Buildings, American Institute of Steel Construction,
Chicago, 2010.
[17] T. Paulay, J.R. Binney, Diagonally reinforced coupling beams of shear walls, ACI Spec.
Publ. 42 (1974) 579–598.
[18] G.G. Deierlein, T.M. Sheikh, J.A. Yura, J.O. Jirsa, Beam-column moment connections
for composite frames: part 2, J. Struct. Eng. ASCE 115 (11) (1989) 2877–2896.
[19] Y. Shi, M. Wang, Y. Wang, Experimental and constitutive model study of structural
steel under cyclic loading, J. Constr. Steel Res. 67 (8) (2011) 1185–1197.
[20] M. Tomii, K. Sakino, Experimental studies on concrete filled square steel tubular
beam-columns subjected to monotonic shearing force and constant axial force,
Trans. AIJ 281 (1979) 81–90.
[21] K. Sakino, H. Ishibashi, Experimental studies on concrete filled square steel tubular
short columns subjected to cyclic shearing force and constant axial force, Trans.
AIJ 353 (1985) 81–89.
[22] Y.Q. Lu, D.L. Kennedy, The flexural behaviour of concrete-filled hollow structural
[23] B. Uy, Strength of concrete filled steel box columns incorporating local buckling, J.
Struct. Eng. ASCE 126 (3) (2000) 341–352.
ð54Þ (2004) 313–337.
[25] W.M. Gho, D. Liu, Flexural behaviour of high-strength rectangular concrete-filled
steel hollow sections, J. Constr. Steel Res. 60 (11) (2004) 1681–1696.
[26] L.H. Han, H. Lu, G.H. Yao, F.Y. Liao, Further study on the flexural behaviour of
concrete-filled steel tubes, J. Constr. Steel Res. 62 (6) (2006) 554–565.
[27] W. Shawkat, W. Fahmy, A. Fam, Cracking patterns and strength of CFT beams under
different moment gradients, Compos. Struct. 84 (2) (2008) 159–166.
[28] S. Guler, A. Copur, M. Aydogan, Flexural behaviour of square UHPC-filled hollow
steel section beams, Struct. Eng. Mech. 43 (2) (2012) 225–237.
[29] J. Cai, W. Liang, H. Lin, Experimental study on shear resistance performance of con-
crete filled square steel tubular columns, J. Shenzhen Univ. Sci. Eng. 29 (3) (2012)
4–9 (in Chinese).
[30] GB 50010-2010 Code for Design of Concrete Structures, China Architecture & Build-
ing Press, Beijing, 2010 (in Chinese).