Journal of Constructional Steel Research 128 (2017) 289–304

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Experimental and analytical investigation of semi-rigid CFST frames withexternal SCWPs

Jingfeng Wang a,b,⁎, Beibei Li a, Jinchao Li a

a School of Civil Engineering, Hefei University of Technology, Anhui Province 230009, Chinab Anhui Civil Engineering Structures and Materials Laboratory, Anhui Province 230009, China

⁎ Corresponding author at: School of Civil Engineering,Anhui Province 230009, China.

E-mail address: jfwang008@163.com (J. Wang).

http://dx.doi.org/10.1016/j.jcsr.2016.08.0190143-974X/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 7 June 2016Received in revised form 8 August 2016Accepted 23 August 2016Available online 7 September 2016

An experimental and numerical research on the seismic performance of semi-rigid concrete-filled steel tubular(CFST) frames with external sandwich composite wall panels (SCWPs) was reported. Four specimens of semi-rigid CFST frames with external SCWPs and one specimen of pure semi-rigid CFST frame subjected to low-cyclicloading were conducted. Failure modes, horizontal load versus displacement relation curves were analyzed. Thetest specimens exhibited good hysteretic behavior, energy dissipation and ductility. Finite element (FE) analysismodeling was developed and the results obtained from the FE model matched well with the experimental re-sults. Extensive parametric studies have been carried out to investigate the effect of steel strength, column slen-derness ratio and steelwire diameter ofwall, etc. on the strength and stiffness of the typed composite frames. Theopening ratio and location of the SCWPswere also discussed. The experimental study and numerical analysiswillprovide the scientific basis for design theory and application of the SCWPs in fabricated steel structure building.

© 2016 Elsevier Ltd. All rights reserved.

Keywords:Sandwich composite wall panels (SCWPs)Concrete-filled steel tubular (CFST)Blind boltsSeismic behaviorFinite element (FE) analysis

1. Introduction

Widespread and unanticipated brittle fractures occurred in weldedsteel beam-column connections in the 1994 Northridge and the 1995Kobe earthquakes. To avoid extensive welding and the required hightolerance, the static and seismic behavior of blind bolted joints to CFSTcolumns has been studied by experiments and FE analysis, such asMirza and Uy [1], Lee et al. [2], Wang et al. [3–6], Ataei et al. [7], andWang et al. [8]. However, little studies focused on the semi-rigid con-crete-filled steel tubular (CFST) frameswith external sandwich compos-ite wall panels (SCWPs).

The development and construction of lightweight pre-fabricatedsandwich panels in building construction are a growing trend in Chinadue to its high strength, reduced weight, good thermal insulation,money-saving and better fire resistance, etc. The SCWPs studied inthis paper consisted of two outside layers separated by an insulationlayer. The outside layers were constructed of precast concrete and thecore layer was polystyrene foam. Diagonal steel wire with alternatingdirection was welded to steel wire mesh which embedded into eachconcrete layer to form space truss connectors. The SCWPs could beprefabricated in factory and assembled in spot and speed constructionschedule up to maximum extent. However, previous earthquake dam-age surveys indicated that the wall destruction and collapse caused

Hefei University of Technology,

large casualties and property losses and scare studies have been doneto investigate failure mechanism of composite frame with SCWPs.Thus, the cooperativework and failuremodes of semi-rigid CFST frameswith external SCWPs under the earthquake action are a research topicwith a high priority.

Up to now, a great deal of study on the seismic behavior and interac-tion between the panels or blocks and H-shaped steel frame have beenconducted. Flanagan et al. [9] tested nine steel frames with hollow claytile under in-plane loading; the experimental results showed that allinfills failed by corner crushing. Markulak [10] reported the hystereticbehavior of steel frames infilled with three different masonry infilltypes: perforated clay blocks, lightweight AAC blocks and newly pro-posed combination of these materials. Tasnimi et al. [11] conductedsix brick-infilled steel frames with openings. Moghaddam [12] carriedout eleven experimental tests on masonry and concrete infilled steelframes. Fang et al. [13] completed a shaking table test of a full-scalesteel frame with ALC external panels. Tong et al. [14] and Sun et al.[15] investigated seismic behavior of the semi-rigid steel frame withRC infill walls. Matteis and Landolfo [16] investigated the behavior ofsandwich panels inserted into a pin-jointed frame systems. Hou et al.[17] described a cyclic loading test results of H-shaped steel frameswith sandwich composite panels.

Apart from experimental research, many accurate theoreticalmodels and FE analysis on the behavior of steel frames with variousinfill walls have been proposed. Saneinejad and Hobbs [18] proposed amethod to predict the strength and stiffness of concrete or masonryinfilled steel frames. Dawe et al. [19] set up a series of complex

Steel beam

SCWP

1600

Square CFST column

185

15001350

1100125 125

125

125

1225

150×150×6

HN150×75×5×7

Connecting plate

(a) Specimen WSF1 and WSF2

Steel beam

15001350

HN150×75×5×7

185

1600

150×150×6

1100125 125

1225

125

125

Square CFST column SCWP

Connecting plate

(b) Specimen WSF3

1600

185

15001350

1100125 125

125

125

1225

150×150×6

HN150×75×5×7Steel beam

Square CFST column

SCWP

Steel brace

Connecting plate

-150×8

(c) Specimen WSF4

Fig. 1. Dimension of specimens (unit: mm).

290 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

calculation model for steel frames with masonry infilling walls.Doudoumis [20] simulated single-bay, single-story infilled framesthrough a precise FE micromodel to investigate the elastic behavior ofinfilled frames. Moghaddam [12] proposed an approximate methodbased on FE analysis to evaluate cracking strength and shear resistanceof steel frameswithmasonry or concrete infills. Chen and Liu [21] devel-oped a FEmodel to study the effect of opening size and opening locationon the seismic behavior ofmasonry infilled steel frames.Matteis [22] re-ported a profitable methodology to predict seismic response of light-weight sandwich shear walls infilled pin-jointed steel frames andaccounted for the contributing effect to the structural behavior byusing ABAQUS.

These above-mentioned experimental studies and numerical analy-sis mainly focused on the behavior of rigid, semi-rigid or hinged H-shaped steel frame with infilled various type of masonry, brick, RCwalls and ALC panels, etc. In addition, few scholars studied the compli-cated interaction of composite structurewith ALCwalls.Wang et al. [23]investigated the rigid circular CFST frames with ALC panels under in-plane cyclically increasing lateral loads by five specimens, the failuremodes, hysteretic curves, energy dissipation and ductility were ana-lyzed. However, little literature on the experiment and numerical anal-ysis of semi-rigid CFST frames with SCWPs has been reported.

This paper is to investigate the seismic performance of semi-rigidCFST frames with external SCWPs. Four specimens of semi-rigid CFSTframes with external SCWPs and one specimen of pure semi-rigidCFST framewere tested under cyclic loading. Moreover, the FE programABAQUS was applied in the analysis. Comparisons between the FE ana-lytical results and the experimental results indicated that the FE modelcould well predict the horizontal load versus corresponding displace-ment relations of the semi-rigid CFST frames with SCWPs. In addition,twelve parameters were also completed to investigate the effect of var-iation of parameters on the structure performance, such as steelstrength, column slenderness ratio, steel wire diameter on walls andcolumn axial load level. The opening ratio and location of the SCWPswere also discussed.

2. Experimental program

2.1. Test specimens

In order to explore the effect of the wall concrete type, the wall con-nection type, the brace setting and the wall setting on the seismic be-havior and failure modes of the typed composite structure, fourspecimens of semi-rigid CFST frames with external SCWPs and onespecimen of semi-rigid CFST frame were tested and analyzed in thispaper. Details of the specimen are illustrated in Fig. 1. The columns forall specimens are concrete-filled square steel tubes with a cross sectionof 150 × 150 × 6 mm and its length is 1785 mm. The self-compactingconcrete was filled in the square steel columns. The beams were de-signed as hot-rolled H-shaped steel section with a cross-section of150 × 75 × 5 × 7 mm and the length of the beams is 1350 mm. Thesteel beams and columns were fabricated through extended end plateconnections with blind bolts which were two rows of M16 Grade 10.9high strength bolts. The ratio of the yielding strength to the ultimatestrength of the bolt is 0.9. All bolts for the beam-to-column connectionswere finally tightened to a torque of 228 Nm according to specificationGB50017 [24].

The extended end plate was fastened to square steel tubes by blindbolts with hooked extensions into the concrete core for the purpose ofreducing the deformations of the tube wall and the end plate, as seenin Fig. 2. The hook-typed extension to the bolt is high strength reinforc-ing rebar with 16 mm in diameter, 70 mm in horizontal length and35 mm in hooked length and the yield strength of the extensions is ofgrade 335 N/mm2. The high strength bolt welded with a hooked rein-forcing rebar to form a complete unit. Gardner and Goldsworthy [25,26] concluded that the hooked anchorage welded to the blind bolts

into CFST column could obviously improve the strength and the initialstiffness of the joints.

The sandwich composite wall panel is a three-layer element com-prising of two outside layers and one core insulation layer. The outsidelayers were constructed of precast concrete and the core layer was

Blind boltsExtended end plate

Square CFST column

150×150×6Steel beam

HN150×75×5×7

35

70 16

M16 Grade10.9

270

3075

7560

38 3875

30

Diameter 18

(a) Extended end plate connection (b)Extended end plate

Fig. 2. Detail of blind bolted end plate joint (unit: mm).

1100 105105

1355

8030

850B

B

230230

AA

B

B

A A

230 230850

105 1051100

8013

5530

Embedded part

Steel brace

Embedded part

(a) Wall set with embedded part (b) Wall set with embedded part and steel braces

Diagonal steel wire3@100

Steel wire mesh 3@50

100 50

3535

30100

Concrete layer

Polystyrene foam layer

Concrete layer

Steel wire mesh 3@50

Reversed diagonal steel wire3@100

(c) Wall section A-A

Diagonal steel wire3@100

100 50

3535

30100

Reversed diagonal steel wire3@100

Polystyrene foam layer

Concrete layer

Concrete layer

Steel wire mesh 3@50

Steel wire mesh 3@50

(d) Wall section B-B

Fig. 3. Detail of sandwich composite wall panels (unit: mm).

291J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

292 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

polystyrene foam. A square welded steel wire mesh of 3 mm diameterbars with 50 × 50 mm openings was established as the longitudinaland transverse reinforcement and embedded in the two outside layers.The connectors were made of 3 mm diameter steel bar and welded tosteel wire mesh with an inclined angle of 45° to a horizontal line. Be-sides, connectors were welded along with alternating direction andover a distance of 100 mm. These connectors running the full heightof the panels were used to tie the two outside concrete layers to formspace truss connectors and to provide shear transfer between the layers.Fig. 3 illustrates the details of SCWPs. The embedded parts were pre-buried at the four corners of SCWPs. The connecting plates wereweldedon the columnflange or beamflange. Then, the SCWPswere in assemblyto semi-rigid CFST frames through bolts drawn into the embedded partand connecting plate to form complete structures, as shown in Fig. 1.

A summary of the test specimens is represented in Table 1. The wallconnection type ofWSF1,WSF2 andWSF4was that the SCWPswas con-nected with both beams and columns, however, specimen WSF3 wasthat the SCWPs was connected with columns. Meanwhile, a pair ofsteel braces with a width of 150 mm and a thickness of 15 mm wereset on specimen WSF4. Specimen WSF5 was a pure blind-bolted CFSTframe without SCWPs. In addition, ceramsite concrete was casted inthe specimen WSF2, ordinary concrete was used in specimens WSF1,WSF3 and WSF4.

2.2. Experimental setup and loading protocol

The arrangement of the experimental setup is illustrated in Fig. 4 andpictures of test site for some specimens were shown in Fig. 5. One hy-draulic actuator of 1000 kN in capacity was employed to apply theload to the column end and was controlled to simulate seismic loading.The actuator mounted between the specimen and RC reaction wall.Eight anchor bolts and four hydraulic jacks were set on the steel groundbeam to transfer the shear forces from the specimen to the ground floorand prevented slip displacement. Due to the limitation of loading condi-tions and consideration of safety, no vertical load was applied on theCFST column during the test.

For investigating the behavior of the semi-rigid CFST frameswith ex-ternal SCWPs and make experiments going smoothly, the preloadingstage and the formal loading stage were applied during the loading pro-cess. In the preloading stage, the hydraulic actuator pushed +5 mm tothe structures and after that dropped to zero, and then pulled −5 mmon the frames and after a while reduced to zero again. Repeating thisprocess twice ensures that the apparatus was working well in subse-quent experimental procedures. In the formal loading stage, the dis-placement control mode was applied in accordance with ATC-24 [27]guidelines until test specimens damaging or having larger deforma-tions. The displacement of specimen corresponding to the yield load(0.7Pmax, and Pmax is assumed as theoretical value of maximum load ofspecimens ) is defined as the theoretical yield displacement (Δy). Thetheoretical yielding displacement Δy and ultimate loading capacityPmax were determined based on finite element analysis. And the Δy

was determined as 12 mm during the loading process. Two cycleswere employed to specimen at each displacement level of 0.25Δy,0.5Δy and 0.7Δy; three cycles were employed to specimen at each dis-placement level of 1Δy, 1.5Δy and 2Δy; two cycles were imposed ateach displacement level of 3Δy, 5Δy, 6Δy, 7Δy.

Table 1Information of the tested specimens.

Specimen Column sectionB × D × t (mm)

Beam sectionhb × tw × tf × bf (mm)

Wall

WSF1 □150 × 150 × 6 H150 × 75 × 5 × 7 OrdinWSF2 □150 × 150 × 6 H150 × 75 × 5 × 7 CeramWSF3 □150 × 150 × 6 H150 × 75 × 5 × 7 OrdinWSF4 □150 × 150 × 6 H150 × 75 × 5 × 7 OrdinWSF5 □150 × 150 × 6 H150 × 75 × 5 × 7 –

2.3. Material properties

Themeasured steel coupons applied in the specimens were listed inTable 2. Tensile coupons cut from steel tubes and sheets ( used in beams,columns, end plates, braces, connecting plates, embedded parts andsteel wire) were conducted to obtain their yield stress (fy), ultimatestress (fu), modulus of elasticity (Es) and elongation at fracture (δ).The testing yield stress and ultimate strength of the Grade 10.9 M16blind bolts were measured as 928 N/mm2 and 1015 N/mm2,respectively.

Columns of all specimens were filled with self-compactingconcrete and ordinary concrete or ceramsite concrete were uti-lized on the SCWPs. For obtaining the material properties of con-crete used in the test, each group of tests had three concretecubes with size of 150 × 150 × 150 mm for compressive strengthand 100 × 100 × 300 mm for elastic modulus. The compressivestrength of core concrete in steel tubes was 43.31 N/mm2 at28 days and the modulus of elasticity was 33,015 N/mm2. Thecompressive strength of ordinary concrete used in the SCWPswas found to be 32.53 N/mm2 at 28 days and the modulus of elas-ticity was 30,611 N/mm2. The compressive strength of ceramsiteconcrete was 51.50 N/mm2 at 28 days and the modulus of elastic-ity was 34,797 N/mm2, as shown in Table. 3.

3. Finite element modeling

3.1. General descriptions

In order to deeply study the mechanical behavior and failure modesof semi-rigid CFST frames with external SCWPs, the key factors formodeling the typed structure include thematerial constitutive law, con-nection between elements, element type and mesh, boundary condi-tions and solution method.

3.2. Material modeling of steel

The constitutive relation of steel was developed by elastic-plasticmodel in software ABAQUS. A simplified bilinear stress-strain curvewas introduced to the high strength blind bolt with hooked extensionand the steel wire and the hardening stage modulus was determinedas 0.01Es, where Es is themeasured elastic modulus of steel. Amulti-lin-ear stress-strain relationship was adopted to steel beam, steel tube, endplate, connecting plate and embedded part. The first part of the multi-linear stress-strain relationship represents the elastic part with a mea-sured elastic modulus (Es) and the rest of model for steel assumes asso-ciated plastic flow. Abdel-Rahman and Sivakumaran [28] proposed amulti-linear stress-strain relation model to simulate accurately me-chanical behavior of cold-formed steel tubes, where a cold-formedcross section was divided into two zones: a corner zone and a flatzone. Sully and Hancock [29] concluded that a residual stress of 0.4 fycould be applied in the corner zone and a residual stress of (0.24–0.0006B)fy be used for the flat zone (where fy is the yield strength inunit MPa and B is the thickness of the steel tube in unit mm,respectively).

concrete type Connection type Brace setting

ary concrete Connect with beams and columns –site concrete Connect with beams and columns –ary concrete Connect with columns –ary concrete Connect with beams and columns Steel braces

– –

RC reaction wall

1000kN jack

SCWP

Pull rod

1000kN MTS actuator

Steel ground beam

Anchor bolts

Loading device

1000kN jack

Fig. 4. Experimental setup.

293J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

3.3. Material modeling of concrete

The concrete damaged plastic model could be used to simulate theconcrete mechanical properties under cyclic or monotonic loadsthrough the ABAQUS library. The model enables to input a multi-linearuniaxial compressive stress-strain curve. Due to the confinement of thesteel tube wall, the transverse deformation of concrete was limited andthe initial defect of concrete was slowed down. The constitutive law ofthe concrete filled in the steel tube and the concrete used in theSCWPs were different. For the concrete infilled steel tube, a practicaland accurate model is proposed by Han et al. [30]. The compressivestress-strain curve of core concrete is expressed as follows:

y ¼2x−x2 x≤1ð Þ

xβ x−1ð Þη þ x

xN1ð Þ

8<: ð1Þ

where χ = ε/ε0, y = σ/σ0, σ0 = fc' (N/mm2), ε0 = εc + 800ξ0.2 × 10−6,

εc = (1300 + 12.5 fc' ) × 10−6; for square CFST, η = 1.6 + 1.5χ, β

¼ ð f 0cÞ

0:1=ð1:2

ffiffiffiffiffiffiffiffiffiffiffi1þ ξ

pÞ . Here, ξ is the confinement factor and ξ =

(Asfy)/(Acfck); As and Ac are respectively the cross-sectional areas of thesteel tube and core concrete; fy and fck are respectively the yield strengthof the steel tube and the characteristic strength of the concrete. The elas-

tic modulus and Poisson's ratio were equal to 4730ffiffiffiffiffif0c

qand 0.2,

(a) Outside external wall

Fig. 5. Photos o

respectively, which was determined by ACI 318 [31]. Where fc' denotes

the cylinder compressive strength of concrete. The relationship be-tween cube compression strength of concrete and cylinder compressionstrength of concrete is given by CEB-FIP [32]. For the core concrete intensile condition, the tension stress is supposed to increase linearlywith respect to strain until the concrete cracking, and then the tensionstress reduces proportionally to zero at a strain which is ten times ofthe strain at concrete cracking. In addition, the damaged plastic modelproposed by specification GB50010-2010 [33] was applied to simulatethe compressive and tensile behavior of ordinary concrete or ceramsiteconcrete casted on the SCWPs. It could better reflect a certain degree ofconcrete damage by introducing damage factor in the process ofcompressing or tensioning. The uniaxial compressive stress-straincurve of concrete used in SCWPs was defined as follows:

σ ¼ 1−dcð ÞEc � ε ð2Þ

dc ¼1−

ρcnn−1þ xn

x≤1

1−ρc

αc x−1ð Þ2 þ xxN1

8><>:

ð3Þ

where ρc = fc,r/(Ec·εc,r), n= (Ec·εc,r)/(Ec·εc,r- fc,r), x= ε/εc,r. The fc,r canbe defined as fck and equaled to 20.1MPa and 32.4MPa, respectively, forordinary concrete and ceramsite concrete according to GB50010-2010[33]. The αc and εc,r equaled to 0.74 and 1470 × 10−6 for ordinary

(b) Inside external wall

f test site.

Table 2Material properties of steel.

Specimen Thickness(mm)

Yield stress(N/mm2)

Ultimate stress(N/mm2)

Young's modulus(N/mm2)

Elongation at fracture(%)

Steel tube 6 361.3 492.3 1.97 × 105 17.5Steel beam flange 7 371.5 508.3 2.01 × 105 17.4Steel beam web 5 308.5 498.2 2.14 × 105 16.9End plate 16 364.4 474.4 2.08 × 105 22.8Steel brace 8 377.2 482.9 1.86 × 105 18.6Connecting plate 12 341.2 452.6 2.15 × 105 19.4Embedded part 12 354.2 487.6 2.08 × 105 20.1Steel wire 3 461.6 592.5 1.88 × 105 18.2

Rigid block

Square CFST columnSteel beam

Extended end plate

294 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

concrete; the αc and εc,r equaled to 1.50 and 1680 × 10−6 for ceramsiteconcrete, according to GB50010-2010 [33]. The uniaxial tensile stress-strain curve of concrete used in SCWPs was defined as follows:

σ ¼ 1−dtð ÞEc � ε ð4Þ

dt ¼1−ρt 1:2−0:2x5

� �x≤1

1−ρt

αt x−1ð Þ1:7 þ xxN1

8<: ð5Þ

where ρt = ft.,r/(Ec ⋅εt,r), x = ε /εc,r. The ft.,r can be defined as ftk andequaled to 2.01 MPa and 2.64 MPa, respectively, for ordinary concreteand ceramsite concrete according to GB50010-2010 [33]. The αc andεc,r equaled to 1.25 and 95 × 10−6 for ordinary concrete; the αc andεc,r equaled to 2.19 and 110 × 10−6 for ceramsite concrete, accordingto GB50010-2010 [33].

3.4. Numerical model description

Based on the experimental specimens,finite element (FE) analysis ofthe semi-rigid CFST frameswith external SCWPswas developed and an-alyzed. A finemesh of C3D8R is used for all steel parts and concrete partdescribed above and the three-dimensional two-node linear element(T3D2) is used to simulate steel wire truss applied in SCWPs. A reason-ablemesh density was ensured bymeshing test, typical meshes of com-posite frames were illustrated in Figs. 6 and 7. At least two elements aremeshed on the thickness of steel beam, steel tube, extended end plate,connecting plate and embedded part to reach accurate mechanical per-formance. Partition command for geometric elements was used to di-vide the structural components to achieve mesh convergence,especially for the extended end plate, blind bolts, embedded part,connecting plate, steel tube and core concrete at the joints and SCWPat four corners. The global mesh size of 70 mm was applied for the

Table 3Material properties of concrete.

Specimen Specimen dimension(mm)

Age(day)

fcu(N/mm2)

Ec(N/mm2)

CS1 150 × 150 × 150 28 45.26CS2 150 × 150 × 150 28 41.96CS3 150 × 150 × 150 28 42.71Average 43.31 33,015CO1 150 × 150 × 150 28 32.78CO2 150 × 150 × 150 28 33.31CO3 150 × 150 × 150 28 31.51Average 32.53 30,611CC1 150 × 150 × 150 28 50.22CC2 150 × 150 × 150 28 52.98CC3 150 × 150 × 150 28 51.31Average 51.50 34,797

Note: CS, CO and CC represent material test of self-consolidating core concrete, ordinaryconcrete and ceramsite concrete, respectively.

whole model by structural mesh technology. The local mesh size ofthe structural components depended on the level of partition on eachcomponent and was all b70 mm. Both material and geometric nonline-arities are considered in the analytical process and theNewton-Raphsonequilibrium iterationmethodwas used to solve the nonlinear problems.

Hard contact was adopted to simulate the behavior of normal con-tact and Coulomb friction model was applied to calculate the tangentialforce of the interface to the surface among steel components and coreconcrete. The surface-to-surface contact including the contact betweensteel tube and core concrete, the contact between steel tube and extend-ed end plate, the contact between bolt head and extended end plate, thecontact between bolt shank and bolt holes on the steel tube and extend-ed end plate, the contact between bolt nut and steel tube, the contactbetween embedded part and connecting plate. The continuity and con-vergence were taken into account on the surface contact. The frictioncoefficient between steel components is taken as 0.45 according toGB50017-2003 [24]. Many scholars studied the interaction betweenthe steel and concrete [34–38] on single column. The friction coefficientbetween the two components was respectively taken as 0.2, 0.25, 0.3,0.47 and 0.6 to coincide better with each test results [34–38]. The fric-tion coefficient between steel tube and core concrete was respectively0.25, 0.47, 0.6 and 0.8 for the finite element model of specimen WSF1to investigate the effect of friction coefficient between steel and coreconcrete on the overall behavior of the typed composite frame. The FEcalculation results demonstrated that FE model with different frictioncoefficient marginally affected the horizontal load versus displacementof the typed composite frame. The friction coefficient between steel

Horizontal load

SCWP

Connecting plate

Embedded part

Fig. 6. FE analysis model.

(a) Steel beam (b) Extended end plate (c) Steel tube (d) Core concrete

(g) Embedded part

tlobdnilB)f(PWCS)e( s (h) Connecting plate

Fig. 7. 3-D idealization of the structural components.

295J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

tube and core concrete was taken as 0.6 according to Han et al. [39] andcoincided better with the test results. The friction coefficient betweensteel and core concrete on other models of tested specimens and para-metric study models were all 0.6. The surface of bolt extension andcore concrete was tied to approximately simulate the provision ofhooked anchorage on the blind bolts into core concrete. The contact be-tween embedded part andwall concretewas tied to couple fully. For thewelding between steel elements and the contact between the outsideconcrete layers and foam-core layer, all degree of freedom for thesame nodes at the contact surface were tied to couple fully. The steelwire truss was embedded in the SCWPs based on the imbeddingprinciple.

The bottom surface of the CFST columns was fixed against all de-grees-of-freedom to simulate the spot condition of experimental speci-mens. The horizontal load is transmitted to the typed frame by a rigidblock whose elastic modulus is 1 × 102 N/mm2 and Poisson's ratio is0.0001. Two steps were set to complete the FE analysis, pre-tighteningforce of 155 kN was imposed on the high strength blind bolt at thebeam-to-column joints at the first step, then, a horizontal displacementwas loaded on the column end at the second step.

4. Experimental and numerical results

4.1. Failure modes

The observed and numerical failure modes of the specimen werecompared, as illustrated in Fig. 8. The failure modes that occurred inthe test process are matchedwell with the numerical results. Cracks oc-curred on the wall surface around the embedded parts and then pene-trated to the opposite side of wall to form slitted cracks with theincreasing of horizontal load. The crushing of concrete around the em-bedded parts was also observed due to larger horizontal displacementand connection rotation. Local buckling occurred on the beam flangesand webs at the beam end. For the specimen WSF2 namely that thesemi-rigid CFST frame with external precast ceramsite concreteSCWPs and the SCWPs is connectedwith beams and columns, the num-ber and distribution of cracks on the precast ceramsite concrete SCWPsare less than that of the precast ordinary concrete SCWPs of specimenWSF1. For the specimen WSF3 namely that the semi-rigid CFST framewith external precast ordinary concrete SCWPs and the SCWPs is con-nected with columns, the number and distribution of cracks on the

(a) Specimen WSF1

(b) Specimen WSF2

(c) Specimen WSF3

(d) Specimen WSF4

(e) Specimen WSF5

Fig. 8. Observed and predicted failure modes.

296 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

297J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

SCWPs aremore than that of the SCWPs connectedwith beams and col-umns (WSF1). For theWSF4 that SCWPs is set with diagonal braces, thecracks propagated to the middle part of the SCWPs and stress concen-tration around the connections can be eased.

From Fig. 8, it can be found that the damage Mises stress of WSF2was higher than that of WSF1, WSF3, WSF4 because of the usage ofceramsite concrete in SCWPs for WSF2. While comparing Fig. 8(a)with Fig. 8(d), the damage Mises stress WSF4 was lower than that ofWSF1 and WSF3, though the steel braces were set on specimen WSF4and the ultimate bearing capacity was higher than that of WSF1 andWSF3, deformation of the SCWPswas restricted and the SCWPs showedas more brittle and these phenomena can also be found in the test. Forspecimen WSF5, buckling deformation at the beam top flange and bot-tom flange were shown in Fig. 8(e). The FE analysis agreed reasonablywith the test results in term of the location of the local buckling andthe magnitude of the buckling deformation. All specimens behaved inan acceptable manner and the test of specimens WSF1, WSF2, WSF3,WSF4 were stopped when the signs of severe crush and spalling of

-150 -100 -50 0 50 100 150-300

-200

-100

0

100

200

300

P/ k

N

/ mm

TestedPredicted

WSF1

(a) Specimen WSF1

-150 -100 -50 0 50 100 150-300

-200

-100

0

100

200

300

WSF3

TestedPredicted

/ mm

P/ k

N

(c) Specimen WSF3

-150 -100 -50 0-300

-200

-100

0

100

200

300

WSF5

P/ k

N

/ m(e) Specimen

Fig. 9. Load (P)-displacement (Δ) r

concrete occurred around the embedded parts and local buckling oc-curred on the beam end, while the specimen WSF5 was stopped be-cause that beam flange fractured partly and too large overalldeformation occurring on the frame to resist horizontal load. No unex-pected failures occurred in the test and FE process. It could be suggestedthat a generally good consistency is reached between the tested andpredicted results.

4.2. Loading-displacement relationship

The horizontal loads and corresponding displacements were used toestablish the horizontal load-displacement curves for the typed struc-tures, as shown in Fig. 9. Fig. 9 shows the effects of the wall concretetype, the wall connection type, the steel brace setting and the wall set-ting on the load-displacement curve for the typed structure. The ulti-mate strength and elastic stiffness of the semi-rigid CFST frames withexternal SCWPs significantly improved when compared with puresemi-rigid CFST frame. When the SCWPs connected with beam and

-150 -100 -50 0 50 100 150-300

-200

-100

0

100

200

300

/ mm

P/ k

N

TestedPredicted

WSF2

(b) Specimen WSF2

-150 -100 -50 0 50 100 150-300

-200

-100

0

100

200

300

TestedPredicted

/ mm

P/ k

N

WSF4

(d) Specimen WSF4

50 100 150

TestedPredicted

mWSF5

elationships of test specimens.

298 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

column, compared with specimen WSF1 that the SCWPs used ordinaryconcrete, the ultimate strength of specimen WSF2 was improved obvi-ously owing to its SCWPs using higher strength ceramsite concrete.When the SCWPs adopted the same ordinary concrete, specimenWSF1 which adopted SCWP-beam and column connection type pos-sessed better integrality than that of specimen WSF3 which usedSCWP-column connection type. Thus, the seismic behavior of specimenWSF1 was superior to specimenWSF3. The hysteresis loop of specimenWSF4 was fuller and without significant pinch, it showed that the ulti-mate strength of specimen WSF4 was improved when the wall setwith diagonal steel braces. The seismic behavior of specimen WSF4was better than that of specimen WSF1 for the reason that the steelbraces shared certain earthquake action for the whole structure. Speci-menWSF5without SCWPs had good deformation capacity, but the ulti-mate strength and elastic stiffness were the smallest in all testspecimens. The test results confirmed that the external SCWPs coulddramatically improve the strength and stiffness of the CFST frames.

Fig. 9 also gives the comparisons between numerical and experi-mental load-displacement curves. The failure load (Pf,t) and the failuredisplacement (Δf,t) of test are respectively the horizontal load and cor-responding displacement when the load falls to 85% of maximumload; the failure displacement of verified numerical simulation equaledto the failure displacement (Δf,t) of test, the failure displacement corre-sponding to the load is defined as the failure load of the verified numer-ical simulation. On the whole, the numerical analysis results werematched reasonably with the test results. The damage of corner panelsand the local buckling of beams can be simulated. It is also noted thatthe elastic stiffness, ultimate strength and correspond displacement ofthe finite element models were relatively larger than that of the testedresults for the reason that the manufacturing and processing defects ofSCWPs and semi-rigid CFST frames were not considered on the FEmodel. The three layers of SCWP on FE model were tied together andnot considered the real interaction between them. The FE theoreticalanalysismodel developed in this paper could be used in subsequent nu-merical analysis of semi-rigid CFST frames with external SCWPs.

5. Parametric analysis

The FE modeling tools were used to present and discuss the resultsof parametric analysis on the effect of structural mechanical behaviorand working mechanism in this section. The main factors that affectthe performance of the typed structure and influence regularity weredetermined.

Based on the above-mentioned model, a parametric investigation ofthe main variables in a structure design of semi-rigid CFST frames withSCWPswas conducted to ascertain the discipline on the elastic stiffness,ultimate strength and corresponding displacement. Twelve parameterswhich may influence the seismic behavior of the typed structure weredivided into three aspects: material parameters, geometric parametersand load parameters. Specific parameters were set as follows:

Material parameters: steel strength (fy) and concrete strength (fcu)of SCWPs; and

Geometric parameters: concrete layer thickness of walls (t), steelwire diameter (d), slenderness ratio of CFST column (λ), beam to CFSTcolumn linear stiffness ratio (ki), steel ratio (a), beam to CFST columnyield strength ratio (km), end plate thickness (tep) and blind bolt diam-eter (db); and

Load parameters: axial compression ratio of the CFST column (n)and blind bolt pretension force (P).

The characteristic calculating example which is commonly used inpractical engineering is described as follows:

Steel beam: H350 × 175 × 7 × 11 mm; span length: L = 6000 mm;steel strength: fy = 345 N/mm2.

CFST column: steel tube □300 × 300 × 10mm; column height: H=3000 mm; steel strength: fy = 345 N/mm2; concrete strength: fcu =60 N/mm2; axial compression ratio: n = 0.6.

SCWPs: thickness of the core polystyrene foam layer: t = 30 mm,thickness of the outside concrete layer: t = 40 mm.

Connection details: extended end plate: 300 × 600 × 16 mm; highstrength blind bolt: 10.9 Grade M20; bolt extension: 75 mm lengthand 20 mm diameter.

In this paper, the slenderness ratio (λ) is defined as 4H/B, where B isthe width of steel tube. The beam to CFST column linear stiffness ratio(ki) is defined as ib/ic, where ib and ic are the linear stiffness of steelbeam and CFST column. The beam to CFST column yield strength ratio(km) is defined as Mb/Mc , where Mb and Mc are the sectional yieldingmoment capacities of steel beam and CFST column by specificationDB34/T 1262-2010 [40], respectively. The steel ratio (a) is defined asAs/Ac, where Ac and As are the net cross-sectional areas of core concreteand steel tube, respectively. The axial compression ratio (n) is definedby N/Nu, where N and Nu are the compressive load exerted on theCFST column and the design value of axial compression bearing capacityof the CFST column by DB34/T 1262-2010 [40], respectively. λ, a, ki andkm for this typical model are respectively 35, 0.15, 0.5, 0.6. For the FEmodelswith various parameters, the failure load and correspondingdis-placement is the maximum point on each horizontal load versus dis-placement curve.

5.1. Effect of steel strength (fy)

In order to investigate the effect of steel strength on the behavior ofthe typed structure, four kinds of strength grade of steel commonlyemployed in the engineering which are 235, 345, 420, 550 MPa in thisstudy. The recorded horizontal load versus horizontal displacement(P-Δ) relationship was compared with change of the steel yieldstrength, as seen in Fig. 10(a). It could be observed that the elastic stiff-ness and ultimate strength were increased dramatically with the in-crease of steel yield strength. When the steel yield strength of thestructure was set to 345, 420, 550 MPa, the elastic stiffness of theframe was raised by 19.5%, 24.7%, 28.6%, respectively, compared withthe frame using the yield strength 235 MPa and the correspondingvalues of the ultimate strengthwas increased by 23.5%, 57.1%, 77.3%, re-spectively. When steel strength fy ≤ 420 MPa, the ultimate strength ofthe structure increases greatly when increasing the steel strength,whereas the incremental amplitude of the ultimate strength is slowdown when fy N 420 MPa.

5.2. Effect of concrete strength of SCWPs (fcu)

Fig. 10(b) shows the effect of concrete strength (fcu) of SCWPs on theP-Δ curves for the structure. The cube compressive strengths of concreteare fit to 30, 60, 80, 100 MPa in the FE analysis to explore this changes.The results of the comparisons showed that both the stiffness andstrength of the structure are dramatically affected by concrete strength(fcu) casted on SCWPs. The elastic stiffness and ultimate strength wereincreased dramatically with the increase of concrete strength (fcu).When the concrete strength (fcu) of the structure was set to 60, 80,100 MPa , the elastic stiffness of the frame was raised by 4.5%, 25.2%,52.3%, respectively, compared with the frame using the concretestrength fcu = 30 MPa and the corresponding values of the ultimatebearing capacity was also increased by15.5%, 18.6%, 26.5%, respectively.

5.3. Effect of the concrete layer thickness of SCWPs (t)

The effect of the thickness of concrete panel where t = 30, 40,50 mm was carried out. Fig. 10(c) displays a comparison of the P-Δcurves of the structure, indicating that the ultimate strength increasesslightly and the elastic stiffness remains basically constantwith increas-ing the thickness of concrete panel. This result demonstrates that boththe ultimate strength and elastic stiffness of the structure are not nota-bly affected by the outside concrete layer thickness.

299J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

5.4. Effect of steel wire diameter of SCWPs (d)

The change of steel wire diameter of SCWPs could affect the ultimatestrength of the structure. Three different diameters are used to plot thehorizontal load (P) versus horizontal displacement (Δ) relationship, asdisplayed in Fig. 10(d). It demonstrates that the ultimate strength in-creases greatly with an increase in the steel wire diameter of SCWPs.While the steel wire diameter had almost no influence on the stiffnessof the structure. When the diameters are 6, 14 mm, the ultimatestrength is raised by 12.2%, 34.9%, respectively, compared to theSCWPs using 3 mm diameter steel wire.

5.5. Effect of slenderness ratio of CFST column (λ)

The study of the impact of slenderness ratio of CFST column whereλ=20, 40, 60 was implemented. By changing the height of the column

0 50 100 150 200 2500

300

600

900

1200

1500

1800

fy =235MPa

fy =345MPa

fy =420MPa

fy =550MPa

P/ k

N

/ mm

(a) Steel yield strength (

t=30t=40t=50

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/

kN

(c) Concrete layer thickness of SCWPs (

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

=20 =40 =60

(e) Column slenderness ratio (

Fig. 10. Effect of various parameters o

to achieve the goal of the variation of slenderness ratio in the finite ele-ment models. Fig. 10(e) provides a comparison of the horizontal loadversus horizontal displacement (P-Δ) relationship with the change ofslenderness ratio. It is found that the slenderness ratio had significantinfluence on ultimate strength and elastic stiffness, both ultimatestrength and elastic stiffness of the structure dramatically decrease asthe λ increase.

5.6. Effect of beam to CFST column linear stiffness ratio (ki)

Three different beams to CFST column linear stiffness ratios wereadopted in the process of simulation which included 0.25, 0.5, 1 andthen the horizontal load-displacement curves are depicted, as shownin Fig. 10(f). It is shown that the ultimate carrying capacity is affectedgreatly by the parameter ki, but only a minor influence exhibited onthe elastic stiffness of the structure. When the beam to CFST column

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

fcu

=30MPa

fcu

=60MPa

fcu

=80MPa

fcu

=100MPa

b) Concrete strength of SCWPs

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

d=3d=6d=14

d) Steel wire diameter

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

ki=0.25

ki=0.5

ki=1

f) Beam to column linear stiffness ratio

n P-Δ relations for the structures.

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

a =0.05a =0.10a =0.15

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

km=0.4

km=0.6

km=0.8

(g) Steel ratio (h) Beam to column yield strength ratio

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

tep=10mm

tep=16mm

tep=28mm

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

db=16mm

db=20mm

db=28mm

(i) Extended end plate thickness (j) Blind bolt diameter

0 50 100 150 200 2500

300

600

900

1200

1500

1800

/ mm

P/ k

N

n=0.2n=0.4n=0.6n=0.8

0 50 100 150 200 2500

300

600

900

1200

1500

1800

P/ k

N

/ mm

P/P0=0.25P/P0=0.75P/P0=1.0P/P0=1.5

(k) Column axial load level (l) Blind bolt pretension force

Fig. 10 (continued).

300 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

linear stiffness ratio of the structurewas arranged for 0.5, 1, the ultimatestrength of the frame was enhanced by 10.3%, 14.5%, respectively, com-pared with the ki equaling to 0.25. The reason for the result of the sim-ulations was that the constraint from beam to column would beenhanced with the increasing of ki.

5.7. Effect of steel ratio (a)

The steel ratio of the CFST column cross-section is a significant de-sign parameter in practical engineering. By changing the steel tubethickness to achieve different steel ratio and the FE modeling withsteel ratio of 0.05, 0.1, 0.15. Fig. 10(g) illustrates the P-Δ curves gainedfrom the FE results. The ultimate strength increases with increasingsteel ratio, while the elastic stiffness of the structure hardly changed at

all. Comparingwith the steel ratio of 0.05, the ultimate strength is raisedby 16.9%, 27.6%, respectively, when the steel ratios of the FE modelingare 0.1 and 0.15.

5.8. Effect of beam to CFST column yield strength ratio (km)

The horizontal load versus displacement under monotonic loadingand corresponding beam to CFST column yield strength ratio with 0.4,0.6, 0.8 were plotted in Fig. 10(h). It is suggested that the ultimatestrength of the frame increase with the incremental ratio km while theelastic stiffness of the structure is rarely changed. When the ratiokm ≤ 0.6, the ultimate strength observably increases as the ratio km is in-creasing. However, the pace of increase of the ultimate strength isslowed down when km N 0.6.

0 50 100 150 200 2500

300

600

900

1200

/ mm

P/ k

N

=0

=10%

=25%

= 40%

=60%

=80%

=100%

Fig. 11. Effect of door opening ratio on P-Δ relation.

301J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

5.9. Effect of extended end plate thickness (tep)

Changing the thickness of extended end platewould also impact theultimate strength of the structure. Fig. 10(i) shows the P-Δ curves of thestructure with different tep. It is suggested that the ultimate strength ofthe structure increases as the extended end plate thickness increases,while the elastic stiffness is not sensitive to different thickness of the ex-tended end plate. Comparing to the ultimate strength of the framewhich is tep = 28mm, the ultimate strength of frames in which the ex-tended end plate thickness was set to 10, 16 mm decreased by 34.2%,6.1%, respectively.

5.10. Effect of blind bolt diameter (db)

Fig. 10(j) exhibits the load-displacement relationship with differentblind bolt diameters. It can be seen that the ultimate strength increaseswith the increase of the blind bolt diameter, but small changes occurredin the elastic stiffness of the frame. When the blind bolt diameter of theframe was designed as 20, 28 mm, the ultimate load was improved by8.9%, 12.9%, respectively, compared with the frame using 16 mm diam-eter blind bolts.

5.11. Effect of axial compression ratio of the CFST column (n)

The axial compression ratio of the CFST column has an importantinfluence on the behavior of the frame. n ranges from 0.2 to 0.8 de-veloped in the FE models to illustrate the horizontal load versus dis-placement curves, as shown in Fig. 10(k). It seems that both theultimate strength and elastic stiffness of the frames are decreasedas the axial compression ratio n increases. When the column axialload level of the structure was set to 0.4, 0.6, 0.8, the ultimatestrength of the frame was reduced by 6.8%, 17.6%, 27.9%, respective-ly, when comparing with n = 0.2 and the elastic stiffness wasdropped by 5.6%, 12.8%, 21.3%, respectively.

5.12. Effect of blind bolt pretension force (P)

The research of the effect of pre-tightening force of the blind boltwhere P=0.25P0, 0.75P0, 1.0P0, 1.5P0was carried out. Hereinto, thenor-mal pretension forces P0 = 155 kN for 10.9 Grade M20 high strengthblind bolts according to GB50017-2003 [24]. Fig. 10(l) provides a com-parison of the horizontal load versus horizontal displacement curvesvariation in the blind bolt pretension force. It can be shown that the ul-timate strength of the frame increases slightly when increasing preten-sion force of blind bolts especially for P N 1.0P0.

On the whole, the parameters described above were investigated indetail to make it clear that the seismic performance of the semi-rigidCFST frames with external SCWPs. The numerical analysis results dem-onstrated that the ultimate strength of the structure is obviously affect-ed by steel strength, concrete strength of SCWPs, steel wire diameter ofSCWPs, slenderness ratio, steel ratio, beam to CFST column yieldstrength ratio, extended end plate thickness, blind bolt diameter andcolumn axial compression ratio. The elastic stiffness of the structure isobviously affected by steel strength, concrete strength of SCWPs, col-umn slenderness ratio and column axial compression ratio.

6. Discussion of SCWPs with openings

The walls with openings is very common in the practical engineer-ing. The openings are located in different position of walls to meet therequirements of flexible architectural function. Several experimentsand numerical analysis have been conducted and presented to investi-gate the in-plane behavior and reductions with appearance of openingsin infilled walls [11,21]. However, infills in those literatures were ma-sonry with openings. For deeply investigating the behavior of semi-

rigid CFST frames infill SCWPs with openings, the opening size andopening location of SCWPs were considered.

6.1. Effect of door opening ratio (ω)

The door opening was set on the center of SCWPs and the openingratio has a prominent influence on the ultimate strength and elasticstiffness of the typed structure. In this paper, the opening ratio (ω) is de-fined as Ao/Ap, where Ao and Ap are the area of door opening and wholewall, respectively. Six different opening ratio (ω) were used in the FEmodel which were 10%, 25%, 40%, 60%, 80% and 100%. Fig. 11 showsthat the strength and stiffness of SCWPs decreased markedly owing tothe set of an opening. The drop velocity slows down with the increaseof opening size, especially when 40% ⩽ ω ⩽ 80%. Compared with thesolid frame, when the opening ratio is 10%, 25%, 40%, 60%, 80% and100%, the ultimate strength of the structure respectively reduced by4.6%, 19.7%, 36.2%, 37.1%, 40.3% and 65.1%; the elastic stiffness of thestructure respectively decreased by 2.3%, 4.3%, 5.5%, 8.7%, 11.4% and43.2%.

6.2. Effect of door opening location

At the same door opening ratioω=40%, threemodelswith differentdoor opening locations including central, left-side and right-side andone model together with door and window opening were considered,as shown in Fig. 12. Fig. 13 provides a comparison of the horizontalload versus displacement curve for the door opening location. It canbe found that as the opening transfers from the right-side to center toleft-side, the ultimate strength and stiffness increase. That is to say,when the opening location is toward the loaded side, the stiffness andstrength are the largest. Compared with the left-side door opening,the ultimate strength of walls with right-side, central, together withdoor and window opening decreased by 14.7%, 7.8%, 4.5%; the elasticstiffness of the structure respectively decreased by 42.7%, 29.1%, 6.4%.

6.3. Effect of window opening location

At the same window opening ratio of 40%, three models with differ-entwindow opening locations including central, left-side and right-sideand one model together with door and window opening were consid-ered, as shown in Fig. 14. The horizontal load versus displacementcurve for the window opening location is shown in Fig. 15. It showsthat as the opening transfers from the right-side to center to left-side,the ultimate strength and stiffness decrease. Compared with the right-leanedwindow opening, the ultimate strength ofwalls with left-leaned,central, together with door and window opening decreased by 14.2%,10.5%, 5.3%; the elastic stiffness of the structure respectively decreasedby 46.2%, 31.8%, 16.9%.

(a) Model-CDO (Central door opening) (b) Model-LDO (Left-side door opening)

(c) Model-RDO (Right-side door opening) (d) Model-TDWO (Together with door and window openings)

Fig. 12. Configuration of door openings.

302 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

The numerical analysis of walls with openings shows that both stiff-ness and strength of the typed structure decreasewith an increase in theopening area. Meanwhile, the reduction rate is also linked with theopening location. The ultimate strength and stiffness decrease whenthe door opening moves away from the loaded side. However, the ulti-mate strength and stiffness decrease when the window opening closesto the loaded side.

7. Conclusions

Based on the experimental and numerical results, the findings ob-tained may be summarized within the limitation of the study:

(1) Themain failuremodes of the semi-rigid CFST frameswith exter-nal SCWPs include the cracks emerged and propagated around

0 50 100 150 200 2500

300

600

900

P/ k

N

/ mm

Model-CDO

Model-LDO

Model-RDO

Model-TDWO

Fig. 13. Effect of door opening locations on P-Δ relation.

the embedded parts, concrete spalling around the embeddedparts and the local buckling of beam flange and web. Comparedwith traditional wall panels, the SCWPs exhibit better integra-tion.

(2) The ultimate strength and initial stiffness of the semi-rigid CFSTframes with external SCWPs significantly improved when com-pared with pure semi-rigid CFST frame. As for the SCWPs usingordinary concrete, the ultimate strength of the specimen thatthe SCWPs connected with columns and beams is larger thanthat of specimen that the SCWPs connected with columns.When the SCWPs connected with beams and columns, the ulti-mate strength of the specimen that the SCWPs using highstrength ceramsite concretewas larger than that of the specimenthat the SCWPs using ordinary concrete. In addition, the set of di-agonal steel braces could improve obviously structural ultimatestrength but become more brittle.

(3) The FE analysis modeling on semi-rigid CFST frames with exter-nal SCWPs was established in this study. The FE modeling wasvalidated in terms of the failure modes and horizontal load-dis-placement relationship curves of the typed structure. It isshown that the FE modeling can be used to predict the behaviorof the typed structure with an acceptance precision.

(4) The extensive numerical parameter analysis results showedthat the ultimate strength of the structure is obviously affect-ed by steel strength, concrete strength of SCWPs, steel wirediameter of SCWPs, column slenderness ratio, steel ratio,beam to column yield strength ratio, end plate thickness,blind bolt diameter and column axial load level. The elasticstiffness of the structure is obviously affected by steel strength,concrete strength of SCWPs, column slenderness ratio andcolumn axial load level.

(a) Model-CWO (Central window opening) (b) Model-LWO (Left-side window opening)

(c) Model-RWO (Right-side window opening) (d) Model-TDWO (Together with door and window openings)

Fig. 14. Configuration of window openings.

303J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

(5) A reduction in both stiffness and strength of SCWPs with theincrease of opening ratio and the reduction rate is also linkedwith the opening location at the same opening ratio. The ulti-mate strength and stiffness decrease when the door openingoffset moves away from the loaded side while the ultimatestrength and stiffness decrease when the window openingcloses to the loaded side.

Acknowledgments

This work described in paper is supported by the National Natu-ral Science Foundation of China (project 51478158 and project51178156). New Century Excellent Talents in University (projectNCET-12-0838), is greatly appreciated.

0 50 100 150 200 2500

300

600

900

/ mm

P/ k

N

Model-CWO

Model-LWO

Model-RWO

Model-TDWO

Fig. 15. Effect of window opening locations on P-Δ relation.

References

[1] O. Mirza, B. Uy, Behavior of composite beam-column flush end-plate connectionssubjected to low-probability, high-consequence loading, Eng. Struct. 33 (2) (2011)647–662.

[2] J. Lee, H.M. Goldsworthy, E.F. Gad, Blind bolted moment connection to unfilled hol-low section columns using extended T-stub with back face support, Eng. Struct. 33(5) (2011) 1710–1722.

[3] J.F. Wang, L.H. Han, B. Uy, Behaviour of flush end plate joints to concrete-filled steeltubular columns, J. Constr. Steel Res. 65 (4) (2009) 925–939.

[4] J.F. Wang, L.H. Han, B. Uy, Hysteretic behaviour of flush end plate joints to concrete-filled steel tubular columns, J. Constr. Steel Res. 65 (8–9) (2009) 1644–1663.

[5] J.F. Wang, L. Zhang, B.F. Spencer, Seismic response of extended end plate joints toconcrete-filled steel tubular columns, Eng. Struct. 49 (2) (2013) 876–892.

[6] J.F. Wang, B.F. Spencer, Experimental and analytical behavior of blind bolted mo-ment connections, J. Constr. Steel Res. 82 (2) (2013) 33–47.

[7] A. Ataei, M.A. Bradford, H.R. Valipour, Experimental study of flush end plate beam-to-CFST column composite joints with deconstructable bolted shear connectors,Eng. Struct. 99 (2015) 616–630.

[8] Z.B. Wang, Z. Tao, D.S. Li, L.H. Han, Cyclic behaviour of novel blind bolted joints withdifferent stiffening elements, Thin-Walled Struct. 101 (2016) 157–168.

[9] R.D. Flanagan, R.M. Bennett, In-plane behavior of structural clay tile infilled frames, J.Constr. Steel Res. 125 (6) (1999) 590–599.

[10] D. Markulak, I. Radić, V. Sigmund, Cyclic testing of single bay steel frames with var-ious types of masonry infill, Eng. Struct. 51 (2) (2013) 267–277.

[11] A.A. Tasnimi, A. Mohebkhah, Investigation on the behavior of brick-infilled steelframes with openings: experimental and analytical approaches, Eng. Struct. 33 (3)(2011) 968–980.

[12] H.A. Moghadam, M.G. Mohammadi, M. Ghaemian, Experimental and analytical in-vestigation into crack strength determination of infilled steel frames, J. Constr.Steel Res. 62 (12) (2006) 1341–1352.

[13] M.J. Fang, J.F. Wang, G.Q. Li, Shaking table test of steel frame with ALC external wallpanels, J. Constr. Steel Res. 80 (1) (2013) 278–286.

[14] X.D. Tong, J.F. Hajjar, A.E. Schultz, C.K. Shield, Cyclic behavior of steel frame struc-tures with composite reinforced concrete infill walls and partially-restrained con-nections, J. Constr. Steel Res. 61 (1) (2005) 531–552.

[15] G.H. Sun, R.Q. He, Q. Gu, Y.Z. Fang, Cyclic behavior of partially-restrained steel framewith RC infill walls, J. Constr. Steel Res. 67 (12) (2011) 1821–1834.

[16] G.D. Matteis, R. Landolfo, Structural behaviour of sandwich panel shear walls: an ex-perimental analysis, Mater. Struct. 32 (5) (1999) 331–341.

304 J. Wang et al. / Journal of Constructional Steel Research 128 (2017) 289–304

[17] H.T. Hou, C.X. Qiu, J.F. Wang, G.Q. Li, An experimental study on sandwich compositepanel infilled steel frames, Adv. Steel Constr. 8 (3) (2012) 226–241.

[18] A. Saneinejad, B. Hobbs, Inelastic design of infilled frames, J. Constr. Steel Res. 121(4) (1995) 634–650.

[19] J.L. Dawe, Y. Liu, C.K. Seah, A parametric study of masonry infilled steel frames, Can.J. Civ. Eng. 28 (1) (2001) 149–157.

[20] I.N. Doudoumis, Finite element modeling and investigation of the behavior of elasticinfilled frames under monotonic loading, Eng. Struct. 29 (6) (2007) 1004–1024.

[21] X. Chen, Y. Liu, Numerical study of in-plane behaviour and strength of concrete ma-sonry infills with openings, Eng. Struct. 82 (2014) 226–235.

[22] G.D. Matteis, R. Landolfo, Modelling of lightweight sandwich shear diaphragms fordynamic analyses, J. Constr. Steel Res. 53 (1) (2000) 33–61.

[23] B.Wang, J.F. Wang, X.D. Gong, B.K. Liu, Experimental studies on circular CFST frameswith ALC walls under cyclic loadings, Int. J. Steel Struct. 14 (4) (2015) 755–768.

[24] GB50017-2003, Code for Design of Steel Structures, China Plan Press, Beijing, 2010[in Chinese].

[25] A.P. Gardner, H.M. Goldsworthy, Experimental investigation of the stiffness of criti-cal components in a moment-resisting composite connection, J. Constr. Steel Res. 61(5) (2005) 709–726.

[26] H.M. Goldsworthy, A.P. Gardner, Feasibility study for blind-bolted connections toconcrete-filled circular steel tubular columns, Struct. Eng. Mech. 24 (4) (2006)463–478.

[27] ATC-24 Guidelines for Cyclic Seismic Testing of Components of Steel Structures[S],Applied Technology Council, Redwood City (CA), 1992.

[28] N. Abdel-Rahman, K.S. Sivakumaran, Material properties models for analysis of cold-formed steel members, J. Struct. Eng. 123 (9) (1997) 1135–1143.

[29] R.M. Sully, G.J. Hancock, Behavior of cold-formed SHS beam-columns, J. Struct. Eng.122 (3) (1996) 326–336.

[30] L.H. Han, G.H. Yao, X.L. Zhao, Tests and calculations of hollow structural steel (HSS)stub columns filled with self-consolidating concrete (SCC), J. Constr. Steel Res. 61(9) (2005) 1241–1269.

[31] ACI 318-02, Building Code Requirements for Reinforced Concrete and Commentary.Farmington Hills (MI), American Concrete Institute, Detroit, USA, 2002.

[32] Comite Euro-International du Beton, Bulletin D'information No. 213/214, CEB-FIPModel Code 1990, Concrete Structures, 1993 Lausanne.

[33] GB50011-2010, Specification of Structural Seismic Design, China Building IndustryPress, Beijing, 2010 [in Chinese].

[34] J. Garzón-Roca, J.M. Adam, P.A. Calderón, I.B. Valente, Finite element modelling ofsteel-caged RC columns subjected to axial force and bending moment, Eng. Struct.40 (7) (2012) 168–186.

[35] H.T. Thai, B. Uy, M. Khan, Z. Tao, F. Mashiri, Numerical modelling of concrete-filledsteel box columns incorporating high strength materials, J. Constr. Steel Res. 102(11) (2014) 256–265.

[36] M. Pagoulatou, T. Sheehan, X.H. Dai, D. Lam, Finite element analysis on the capacityof circular concrete-filled double-skin steel tubular (CFDST) stub columns, Eng.Struct. 72 (2014) 102–112.

[37] J. Moon, C.W. Roeder, D.E. Lehman, H.E. Lee, Analytical modeling of bending of circu-lar concrete-filled steel tubes, Eng. Struct. 42 (12) (2012) 349–361.

[38] L.H. Han, S.H. He, F.Y. Liao, Performance and calculations of concrete filled steeltubes (CFST) under axial tension, J. Constr. Steel Res. 67 (11) (2011) 1699–1709.

[39] L.H. Han, W.D. Wang, X.L. Zhao, Behaviour of steel beam to concrete-filled SHS col-umn frames: finite element model and verifications, Eng. Struct. 30 (6) (2008)1647–1658.

[40] DB34/T 1262-2010, Technical Specification for Concrete-Filled Steel Tubular Struc-tures, Anhui Engineering Construction Standard Design Press, Hefei, 2010 [inChinese].

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具