Experimental investigation of extended end plate joints to concrete-filled steel tubular columns

Journal of Constructional Steel Research 79 (2012) 56–70

Contents lists available at SciVerse ScienceDirect

Journal of Constructional Steel Research

Experimental investigation of extended end plate joints to concrete-filled steel tubular columns Jingfeng Wang a, b,⁎, Liping Chen a a b

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

a r t i c l e

i n f o

Article history: Received 10 January 2012 Accepted 25 July 2012 Available online 19 August 2012 Keywords: Concrete filled steel tubular (CFST) Semi-rigid Extended end plates Blind bolts Anchorage extension

a b s t r a c t An experimental programme to obtain the behaviour of blind bolted extended end plate joints to circular or square concrete-filled steel tubular (CFST) columns under monotonic loading has been conducted. In order to enhance the strength and stiffness of the connections, the anchorage extensions are provided to the blind bolts to link the connection back into the concrete with the tubular. This paper investigated the effect of the end plate thickness and the column section type on the static behaviour and failure modes of the tested connections. The structural performance of the blind bolted extended end plate connections was evaluated in terms of the moment–rotation relationship, connection rigidity, the deformation pattern and the strain response. The test results showed that the blind bolted extended end plate connection to CFST columns exhibits high strength and stiffness, while its connection rotation capacity satisfies the ductility requirement for earthquake resistance in aseismic region. The experimental studies also demonstrated that the strength and stiffness of the connections can be improved by providing anchorage extensions to the blind bolts, and utilising moderately thick end plates leads to joints approaching full strength for the extended end plate connections. Crown Copyright © 2012 Published by Elsevier Ltd. All rights reserved.

Notation B D H L Lb L0 hb bfb tfb twb tp E Ec EIb fy fu fcu n N Nu P Pu

Width of square steel tube Exterior diameter of circular steel tube Column length of test joint Beam length of test joint Beam span Distance between the load application point and the beam end welded to the end plate Beam section height Beam flange width Beam flange thickness Beam web thickness End plate thickness Young's modulus for steel Young's modulus for concrete Flexural rigidity for the beam Steel yield strength Steel ultimate strength Concrete compressive cube strength Axial load level, n = N / Nu Axial load applied to CFST column Ultimate axial load applied to CFST column Test load on the beam tip Maximum test load on the beam tip

⁎ Corresponding author. Tel./fax: +86 551 2901434. E-mail address: [email protected] (J. Wang).

M My Me Mm Mf Mbp θr θb θc θr,y θr,e θr,m θr,f Kie Kse Cm Cθ Ck εyr,b εyr,w εyr,e εyr,c

0143-974X/$ – see front matter. Crown Copyright © 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jcsr.2012.07.016

Connection moment Yield moment defined by the test joint Design moment capacity defined by EC3 specification, Me = 0.67Mu Ultimate moment of the test joint Moment of the test joint at failure state, Mf = 0.85Mu Design plastic moment resistance of the beam Connection rotation Beam rotation Column rotation Connection rotation corresponding to the yield moment of the connection Connection rotation corresponding to the design moment capacity of the connection Connection rotation corresponding to the ultimate moment of the connection Connection rotation corresponding to the moment of the connection at failure state Initial stiffness of connection Service-level stiffness of connection Connection moment coefficient Connection rotation coefficient Connection stiffness coefficient Yield strain of steel beam flange Yield strain of steel beam web Yield strain of end plate Yield strain of steel tube

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

1. Introduction The use of structural hollow sections (SHS) in practice engineering is attractive due to aesthetics, reliability of manufacture and a high strength to weight ratio. Concrete-filled steel tubular (CFST) columns exhibit excellent structural and constructional benefits, which include high strength and fire resistance, large stiffness and ductility [1]. Their use in this capacity is inhibited by problems in making connections to other members. Early developments in overcoming the connection problem included additional fittings, through-bolt connections, internal or external diaphragm plates, and passing the beam continuously through the column [2–10]. However, the use of these connection methods has not always been convenient in construction practice. The use of standard dowel bolts is often impossible as there is rarely access to the inside of the tube to allow for tightening. Recently the development of the blind fasteners allows the bolt installation from one side of the connection only without the need for access within the SHS column section [11]. In the context of structural engineering, the commercially available blind bolts include the Lindapter Hollo-bolt, the Ajax ONESIDE bolt, the Huck high strength blind bolt, and Flowdrill, as shown in Fig. 1. Each type of fastener differs in the bolt components, resistance mechanism and method of installation. In order to explore the possibility of using the blind bolt techniques, several investigations have been conducted on different blind bolted structural connections, such as T-stub connection, flush or extended end plate connection, and angle connection, which are connected to the face of the column. This highlights the need for an extensive study on the structural behaviour of the different blind bolted connections. Barnett et al. [12] reported an experimental study on T-stub connections to square CFST columns using different type Hollo-bolts under tensile loading. Goldsworthy and Gardner [13,14], and Yao et al. [15] carried out a series of cyclic tension experiments on blind bolted T-stub connections to the circular CFST columns, to investigate the use of extensions to blind bolts and stiffness of a T-type element. Yao et al. [16] completed a joint test to investigate double built-up tee connections to square CFST columns using Ajax ONESIDE bolts with anchorage extensions under cyclic loading. Lee et al. [17–19] reported a testing research project in collaboration with Ajax Engineered Fasteners and Australian Tube Mills to develop various type blind bolted connections to unfilled hollow section columns using the

57

Ajax ONESIDE bolts. The research blind bolted connections to SHS columns included T-stub connection, channel side plate connection and channel T-stub connection. France et al. [20–22] conducted a series of joint tests under monotonic loading to explore the moment capacity and rotational stiffness of the end plate connections to square SHS or CFST columns with flowdrill connectors. Loh et al. [23] reported a static experimental study on flush end plate composite connections to square CFST columns with Hollo-bolts, in order to investigate the effects of shear connection and reinforcement ratio in composite joints. Wang et al. [24,25] studied the static and hysteretic behaviour of flush end plate joints to circular or square CFST columns using Hollo-bolts, and concluded that the blind bolted flush end plate connection behaves in a semi-rigid and partial strength manner according to EC3 specification [26]. Mirza and Uy [27] described the test results of composite flush end plate connections to square CFST columns with Ajax fasteners subjected to low-probability, high-consequence loading. Elghazouli et al. [28,29] reported a series of test programme to examine the behaviour of angle connections between open beams and tubular columns by means of Lindapter Hollo-bolts. Experimental results involving extended end plate connections comprised I-beam sections connected to H-section columns have previously been described [30–33]. However, little effort was devoted to studying steel beams connected to CFST columns by blind fasteners and extended end plates. Korol et al. [34] conducted a test programme on five extended end plate connections between W-shape beams and square SHS columns using blind bolts under monotonic loading. Mourad et al. [35] reported two cyclic load tests on blind bolted extended end plate connections of wide flange beams to square SHS columns, and investigated the effect of joint flexibility on the response of the frames under dynamic loading. France et al. [20] carried out testing on four extended end plate connections to square unfilled hollow section columns using flowdrill connectors, in order to investigate their strength and rotational stiffness characteristics. France et al. [20] observed that utilising moderately thick end plate leads to joint approaching full strength for extended end plate connections. However, little is known about the effect of concrete filling of tubes on the performance of blind bolted extended end plate connections. France et al. [22] performed tests on two extended end plate connections to concrete filled square steel tubular columns using flowdrill connectors, and also concluded that the strength and

a) Lindapter Hollo-bolt

b) Ajax oneside bolt

c) Huck Ultra-Twist

d) Flowdrill

Fig. 1. Blind bolts.

58

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

stiffness of concrete-filled flowdrill joints increased obviously when compared to their unfilled equivalent. In this paper, the test programme included four blind bolted extended end plate connections to square or circular CFST columns. The parameters investigated were the column section type and the end plate thickness. It is feasible and convenient for steel tubes with square sections connected to steel beams using plane end plates, but little attention has been paid to the blind bolted extended end plate connections to circular CFST columns. To resolve this problem, Wang et al. [24,25] proposed an innovative approach of blind bolting a flush end plate to concrete-filled circular hollow columns. Hence, this paper also studies circular CFST column connections using curved end plates and blind bolts. Furthermore, this paper also investigates that an extension was provided to the blind bolt head by welding, in order to improve the tension behaviour of the blind bolted connections to the thin tube wall. Recently Gardner and Goldsworthy [13,14] and Yao et al. [15] verified that the extensions to the blind bolts were very effective in relieving the stress concentration on the thin tube wall, but there is no analysis or experimental study of extended end plate connections using blind bolts with anchorage extensions. In this paper the structural behaviour of the blind bolted extended end plate connections is evaluated in terms of the moment–rotation relationship curves, connection rigidity, failure modes, and the strain distribution within critical components.

2. Experimental programme

a) Circular CFST connection (MEC1 and MEC2)

b) Square CFST connection (MES1 and MES2)

2.1. Test specimens The test specimens (shown in Fig. 2) are representative of exterior beam-to-column joints in a moment-resisting frame. To investigate the effect of the column section type and the end plate thickness, four extended end plate joints to concrete-filled steel tubular columns with blind bolts were tested under monotonic loading. Table 1 summarises the detail of the test specimens. Fig. 1 provides the design details of the beam-to-column connections. The columns for specimens MEC1 and MEC2 are concrete filled circular steel tubes with a cross-section 200 × 10 mm; the columns for specimens MES1 and MES2 are concrete filled square steel tubes with a cross-section 200 × 200 × 10 mm. Meanwhile, the beams were commercial H-shape steel sections of a cross-section HN300× 150 ×6 × 10 mm for all test specimens. The test joints were fabricated and erected in the workshop at the laboratory. The bare steel joint assembly was constructed firstly. The steel beams and columns are assembled by means of extended end plate connections with blind bolts. All the bolts for the connections are tightened by a torque wrench with a torque value of 442 N.m, so as to ensure consistency. Self-consolidating concrete (SCC) mix was filled in the circular or square steel tubular columns after the erection of the steel framework. In the experimental programme, the column section type and the thickness of the end plate were the parameters that varied. The extended end plate was fastened to the circular or square tube by blind bolts with extensions into the concrete core, as illustrated in Fig. 3. The extensions to the bolts were 20 mm diameter 50 mm length high strength reinforcing bars of grade 335 N/mm2, as shown in Fig. 4. The reinforcing bars were welded to the head of the bolt to form a complete unit. Gardner and Goldsworthy [13,14] studied the behaviour of the blind bolts with extensions in T-type element of a moment-resisting connection to concrete-filled circular steel tube column by cycle loading tests. These tests showed that provision of extensions onto blind bolts into the concrete-filled circular steel tube column has obviously improved the strength and initial stiffness of the connections. The blind bolt used in the tests is Grade 10.9 M20, namely that exterior diameter of the bolts is 20 mm and the ultimate strength of the bolts is 1000 N/mm 2. The ratio of the yielding strength and the ultimate strength of the bolts is 0.9. These were tightened

c) Curved endplate

d) Rectangular endplate

Fig. 2. Details of extended end plate connection specimens (in mm).

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

Rebar extension

Table 1 Information of the test specimens. Specimen number

Column section B or D mm

Column length H mm

Beam section hb × bfb × twb × tfb mm

Beam length L mm

Endplate thickness tp mm

MEC1

200 × 10 (circular) 200 × 10 (circular) 200 × 10 (square) 200 × 10 (square)

1625

300 × 150 × 6 × 10

1700

12

1625

300 × 150 × 6 × 10

1700

18

1625

300 × 150 × 6 × 10

1700

12

1625

300 × 150 × 6 × 10

1700

18

MEC2 MES1 MES2

59

Blind bolt

50 Fig. 4. Blind bolts with extension.

2.3. Experimental setup and loading programme

to 442 N.m torque following the specifications by GB50017 [36]. The experimental setup photo is shown in Fig. 5.

2.2. Material properties Table 2 summarises the results of the material tests of the steel coupons used in the specimens. The nominal yield strength and tensile strength of the Grade 10.9 M20 blind bolts were determined as 900 N/mm 2 and 1000 N/mm 2, respectively. The self-consolidating concrete (SCC) mix proportions of the concrete were as follows: cement 450 kg/m3; blast furnace slag 170 kg/m3; water 181 kg/m3; sand 815 kg/m3; coarse aggregate 815 kg/m3; and additional high-range water reducer (HRWR) 4.3 kg/m3. To simulate the casting process, an SCC sample was forced to flow through obstacles under static pressure with a typical test set-up (L-box). The flow time from the sliding door to the front door of the L-box and the flow speed, as well as the flow distance of the SCC, were recorded. The fresh properties of the SCC mixture were as follows: slump flow 259 mm; concrete temperature 27 °C; unit weight 2385 kg/m 3; flow time 15 s; flow speed 52 mm/s; and flow distance 800 mm. In all the concrete mixes, the fine aggregate used was silica-based sand; and the coarse aggregate was carbonate stone. The size of the concrete cubes was 150 × 150× 150 mm in the test of cube compressive strength and 100× 100 × 300 mm for the modulus of elasticity. The compressive strength of the concrete was determined by standard cylinder compression tests. Material properties of concrete cube are shown in Table 3. The compressive cube strength (fcu) of the self-consolidating concrete was found to be 44.34 N/mm 2 and the modulus of elasticity (Ec) was 32,604 N/mm2 at 28 days. On the day of testing, the compressive strength from cube samples was 48.27 N/mm2 and the modulus of elasticity was 33,521 N/mm2.

a) Circular HSS

The general arrangement of the test setup is illustrated in Fig. 6. The axial loading was applied by a hydraulic jack reacting against a steel frame onto an upper support at the top of the CFST column. The level of axial load (n) in the specimens was selected as 0.6, which generally reflected the real axial level of the CFST columns in a practical building structural system, n = N/Nu, where N is the axial load applied to the column. The nominal strength Nu was calculated by using the specification for CFST structures DB34/T 1262‐2010 [37] and using the recorded steel and concrete mechanical properties. One hydraulic actuator of 500 kN capacity was used to apply the load to the beam end in the vertical direction. The load direction of the hydraulic actuator on the beam end is upward to refrain from lateral movement. In order to restrict the plane beam-to-column joint from lateral movement, the upper end of the CFST column was supported by a horizontal articulated equipment, which was connected with a RC reaction wall. The lower end of the CFST column rested on a bed of solid reinforced concrete by using vertical articulated equipment. In order to investigate the behaviour of the blind bolted extended end plate connections under monotonic loading, the test procedure was divided into two different phases: the preloading phase and the formal loading phase. In the preloading phase, all the measurement channels were scanned to record the initial readings before any loads were applied. Then, all the loading jacks were increased by 10 kN and after a while reduced to zero. In the formal loading phase, the vertical actuator was operated in the displacement control mode after the axial load of column was preloaded. This was found to be safe to control the loading, especially during the larger deflection range of the beam. 2.4. Instrumentation The beam tip displacement was automatically recorded by the hydraulic actuator acting on the beam tip. In addition, four Linear Variable Displacement Transducers (LVDTs) were mounted to measure

b) Square HSS Inner steel wall

Inner steel wall Blind bolts

Fig. 3. Plan view of the blind bolts secured to the HSS.

Blind bolts

60

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

Fig. 5. Experimental setup photograph.

the connection rotation and side-sway of the specimens. A total of forty seven strain gauges were employed in each specimen. The layout of the LVDTs is illustrated in Fig. 7. Strain gauges were used to monitor the strains in the beam flanges and webs, end plate and the steel tube. Placement of the beam strain gauges was mainly concentrated in the beam closer to the connection. All readings were recorded using a microcomputer. The layout of the strain gauges is illustrated in Fig. 8. 3. Test results 3.1. Specimen MEC1 Specimen MEC1 had 12 mm thick curved end plate connected to the circular CFST columns. Before the actuator displacement reached a level of 19.58 mm corresponding to a load of 84.6 kN the load– displacement curve of the beam (shown in Fig. 9) is essentially elastic. The end plate welded with the beam tension flange primarily appeared with a bending deformation at an actuator displacement of 39 mm. With the load increment to 70 mm, there was deformation in the beam compressive flange. The load attained was a maximum of 137.72 kN at an actuator displacement of 114 mm, and failure of the connection was evident. The specimen MEC1 test was terminated when signs of a large beam deformation were observed at an actuator displacement of 180 mm. The maximum deformation of the end plate

Table 2 Material properties of steel. Specimen number

Steel wall thickness mm

Yield stress N/mm2

Ultimate stress N/mm2

Young's modulus N/mm2

Elongation at fracture (%)

Steel beam flange Steel beam web Circular steel tube Square steel tube Endplate-1 Endplate-2

10 6 10 10 12 18

349.3 312.5 331.8 328.1 323.3 274.4

492.0 508.3 484.5 490.6 436.7 414.4

1.87 × 105 2.16 × 105 1.94 × 105 2.01 × 105 1.98 × 105 1.93 × 105

16.5 17.4 18.2 21.7 31.0 24.8

in beam compressive flange was about 12 mm from the column wall. The beam compressive flange was in flow plastic state; the left side of the flange with the distance of 52 mm from the column wall plumped up 12 mm; and the right side of the flange with the distance of 160 mm from the column wall caved in 50 mm. 3.2. Specimen MEC2 Specimen MEC2 had 18 mm thick curved end plate connected to the circular CFST columns. The trend of load–displacement curve of the beam (shown in Fig. 9) was primarily the same to specimen MEC1. However, the deformations of the end plate and the column flange in specimen MEC1 were markedly more than those in specimen MEC2. The load attained was a maximum of 144.02 kN at an actuator displacement of 84 mm, and the test strains in the end plate were still in elastic stage. The specimen MEC2 test stopped when signs of remarkable bucking deformation in compressive beam flange were found at an actuator displacement of 180 mm. Moreover, the left side of beam compressive flange with the distance of 90 mm from the column wall plumped up 50 mm; and the right side of the flange with the distance of 138 mm from the column wall caved in 60 mm. 3.3. Specimen MES1 Specimen MES1 had 12 mm thick rectangular end plate connected to the square CFST columns. A slight sound was emitted from the connection zone at 46 mm actuator displacement. This phenomenon maybe explained: with the tension bolts slowly moving out from the core concrete, it led to the concrete cracking and spalling in the tension. After 64 mm actuator displacement, a distinctive rattle sound was observed and the load fell 4 kN. However, the load still increased for a little while. It was deduced that steel rebars welding the bolts exhibited anchorage failure. The above reason is validated by observing the specimen with cut away columns exposing the steel tube. The load attained was a maximum of 118.51 kN at an actuator displacement of 205 mm, and failure of the connection was evident. The specimen MES1 test was terminated when signs of welding crack between the beam tension flange and the end plate were observed at an actuator displacement of 275 mm. The end plate was in flow plastic state and exhibited obvious

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

61

Table 3 Material properties of concrete cube. Specimen label

Size of cube (mm)

Age (days)

Cube compressive strength (N/mm2)

C11 C12 C13 C21 C22 C23 Average C31 C32 C33 Average

150 × 150 × 150 150 × 150 × 150 150 × 150 × 150 150 × 150 × 150 150 × 150 × 150 150 × 150 × 150

28 28 28 28 28 28

40.18 38.31 44.71 50.04 46.67 46.13 44.34

100 × 100 × 300 100 × 100 × 300 100 × 100 × 300

28 28 28

Young's modulus (N/mm2)

33,460 32,554 31,798 32,604

deformation; the welding crack was about 25 mm; the left side of the beam compressive flange with the distance of 54 mm from the column wall plumped up 4 mm; and the right side of the flange with the distance of 80 mm from the column wall plumped up 5 mm.

Age at testing (days)

Cube compressive strength (N/mm2)

52 52 52 52 52 52

48.19 47.56 50.02 46.57 49.14 48.16 48.27

52 52 52

34,537 32,146 33,880 33,521

3.4. Specimen MES2 Specimen MES2 had 18 mm thick rectangular end plate connected to the square CFST columns. Due to the thicker end

Reaction frame Reaction wall 500kN MTS actuator Reaction frame Restraining beam

5000kN jack Restraining device

Specimen Anchor device RC beam Anchor bolts

RC bed device Thick bar

Fig. 6. Experimental setup.

a) Circular CFST connection (MEC1 and MEC2) b) Square CFST connection (MES1 and MES2) LVDT (horizontal deformation)

300

LVDT (vertical deformation)

LVDT (horizontal deformation) LVDT (vertical deformation)

L1

LVDT (sway deformation)

300

L1

L4

LVDT (sway deformation)

L4

L3

300

300

300

300

Young's modulus (N/mm2)

L3

L2

L2

Fig. 7. Layout of LVDTs.

62

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

a) Circular CFST connection (MEC1 and MEC2) Top flange 1

7

80

8 300

250

9 300

30

30 1

26

30 29 28

320

27

23 24 25

50 50

30

35 38 34 36 37 41 39 40 44 47 42 45 43 46

33

2 3 4 5

320

30

50 50

40 40

6 80

32

6

10 250

80

Bottom flange

13 14

18 19

11

15

20

16 17

21 22

300

12 300

31

b) Square CFST connection (MES1 and MES2) Top flange 8 300

300

30

1

320

30

35 38 36 34 37 41 39 4047 44 42 45 43 46

33

2 3 4 5

320

30

4040 23 24 25 27 26

30

6 80

32

Bottom flange 10

6 80

250

11 300

12 300

50 50

250

80

9

13 14

18 19

15

20

16 17

21 22

31

30 29 28

50 50

7

1

Fig. 8. Layout of strain gauges.

plate, the weld cracking in the steel beam end has not been also observed during the test, and differs from specimen MES1. The maximum load of specimen MES2 was larger than that of specimen MES1, as shown in Fig. 9. The load attained was a maximum of

125 kN at an actuator displacement of 50.13 mm. The specimen MES2 test was stopped at an actuator displacement of 98.84 mm when remark signs of buckling deformation of beam tension flange were observed.

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

160

63

4. Experimental analysis and discussion

MEC2 MEC1

4.1. Deformation patterns

Load (kN)

120

80

MES2

MES1

40

0

0

50

100

150

200

250

300

Displacement (mm) Fig. 9. Load–displacement relationship of the beam ends.

The maximum deformation of the end plate was about 18 mm. Meanwhile, the left side of the beam compressive flange with the distance of 120 mm from the column wall plumped up 40 mm; and the right side of the flange with the distance of 140 mm from the column wall caved in 14 mm.

In the case of the blind bolted extended end plate connections to CFST columns, failure (seen in Figs. 10–13) occurred in the following modes: (1) deformation of the end plate; (2) bucking deformation of the beam compressive flange; (3) bucking deformation of the beam web; (4) outward deformation of the column flange; (5) welding crack between the beam compressive flange and the thin end plate for square CFST column joints; (6) anchorage fracture of the tensile bolts with extensions in the square columns; and (7) crushing of the core concrete due to the larger connection rotation. The failure of the joints was related with the end plate thickness and the column section type. Examination of the joints after testing revealed that specimens MEC1 and MEC2 were nominally identical, except that the end plate thickness was 12 mm and 18 mm respectively. However, the end plate deformation of specimen MEC1 was more significant than that of specimen MEC2. Due to the thinner end plate, the end plate strains in specimen MEC1 were beyond the material yielding strain and the end plate deformation appeared in the tension beam flange, but the end plate strains in specimen MEC2 were still elastic without any bending deformation. Compared with the end plate, CFST column and bolts, the steel beam became the weakest member. The maximum deformation of the end plate

Fig. 10. Connection deformation in test MEC1.

Fig. 11. Connection deformation in test MEC2.

64

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

Fig. 12. Connection deformation in test MES1.

Fig. 13. Connection deformation in test MES2.

for specimen MEC1 was 12 mm, while specimen MEC2 had obvious deformation of the end plate, illustrated in Figs. 10 and 11. A similar effect of the end plate thickness also occurred for specimens MES1 and MES2, but the end plate defamation in the circular CFST column connections was obviously lower than that of the square CFST column connections. For specimen MES1 with 12 mm thick end plate, the end plate was the weakest member and completely came into plastic stage after testing. The welding between the beam compressive flange and the end plate cracked and teared. The 18 mm thick end plate strains in specimen MES2 beyond the yielding strain,

a) MEC1

but the bending deformation of the end plate was less than that of specimen MES1. The maximum deformation of the end plate for specimens MES1 and MES2 was 24 mm and 18 mm, shown in Figs. 12 and 13. In order to determine the failure mode of the blind bolts inside the composite columns, the steel tubes were exscinded using gasoline weld-cutting equipments without pressure. The cracks and failure of the exposed concrete infill were observed and recorded. Then, the core concrete was removed to obverse the deformation of the bolts. Fig. 14 illustrates the inner failure of specimens MEC1

b) MES1

Fig. 14. Blind bolts in specimens with concrete removed.

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

a) MEC1 and MEC2

65

b) MES1 and MES2

Fig. 15. Moment–rotation relationship of test specimens.

4.2. Moment–rotation relationship The behaviour of beam-to-column connections is typically represented by a M–θr curve that describes the relationship between the applied bending moment, M, and the corresponding connection rotational rotation, θr. The bending moment, M, acting on the connection corresponds to the applied load, P, multiplied by the distance between the load application point and the beam end welded to the end plate, L0. M ¼PL

ð1Þ

The connection rotation, θr, is defined as the change in angle between the centre lines of the beam and the column, θb and θc. The connection rotations and corresponding connection moments were used to construct the moment–rotation curves for the connections shown in Fig. 15. Some key characteristic points (illustrated in Fig. 15) are marked in the moment–rotation curves of the test connections: ① means the end plate yielding deformation when the end plate near the beam tension flange deviated from the column wall; ② means the bucking deformation in the beam compressive flange; ③ means the bolt extension breakage in tension. The typical moment–rotation relationship of connection and key parameters are shown in Fig. 16. In Fig. 16, My and θr,y are respectively the yield moment defined by the test joint and corresponding connection rotation; Me and θr,e are respectively the design moment capacity defined by EC3 specification and corresponding connection rotation, hereinto Me = 0.67Mm; Mm and θr,m are respectively the ultimate moment and corresponding connection rotation; Mf and θr,f are respectively the moment at failure state and corresponding connection rotation, hereinto Mf = 0.85Mm. The ultimate moment capacity, Mm, initial stiffness, Kie, service-level stiffness, Kse, and rotation capacity, θr,f, of each joint are summarised

in Table 4. The initial stiffness and service-level stiffness of the connections are defined as the scant flexural stiffness corresponding to the stage when the moment reaches 20% and 60% of the flexural capacity, respectively. In order to evaluate the connection behaviour, connection moment coefficient Cm, connection rotation coefficient Cθ and connection stiffness coefficient Ck are used in Table 5, where, Cm1 = My / Mm, Cθ1 = θr,m / θr,y, Cθ2 = θr,m / θr,e, Cθ3 = θr,f / θr,y, and Ck1 = Kse / Kie. It is concluded that the yield moments of the test connections are about 73%– 91% of the ultimate moment capacities, while the connection rotation coefficient Cθ1 = 2.25–3.33, Cθ2 = 4.02–4.84, and Cθ3 = 3.87–4.51. The service-level stiffness of the test connections is about 63%–87% of the initial stiffness. The test results showed that the typed connection exhibits high strength and ductility. The response of all specimens was ductile as exhibited by significant rotation capacity. It was demonstrated that strength and stiffness of the blind bolted extended end plate joints may be affected by the end plate thickness and the column section type. An increased the end plate thickness enhanced stiffness and strength of nominally identical connections. For circular CFST column connections, the maximum strength capacity and initial stiffness of MEC2 were enhanced by 3.3% and 8.2% respectively to that of MEC1. For square CFST column connections, the maximum

M (kN.m)

and MES1 with cut away column. The observed results showed that except for the concrete near the bolts in tension which was cracked, no obvious deformation appeared in the core of the joint and there was also no sign of bending or shear deformation of the bolts in the tests. Meanwhile, the anchorage rebar fracture of the tensile bolts in the specimen MES1 with square CFST columns was also found. All the blind bolt connections performed satisfactorily in an acceptable manner. The test results show that the blind bolted extended end plate connections exhibited favourable strength, stiffness and deformation performance for use in a moment resisting frame.

Mm Mf My Me

S j,ini θ r,e θ r,y

θ r,m

θ r,f

θ r (mrad) Fig. 16. Key parameters in moment–rotation relationship of test joints.

66

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

Table 4 Measured moments, stiffness and rotations of connections. Specimen

θr,y (mrad)

My (kN.m)

θr,e (mrad)

Me (kN.m)

θr,u (mrad)

Mu (kN.m)

θr,f (mrad)

Mf (kN.m)

Kie (kN.m/mrad)

Kse (kN.m/mrad)

MEC1 MEC2 MES1 MES2

27.29 22.59 38.07 23.58

184.64 204.12 137.95 170.33

14.9 12.64 26.64 12.98

145.47 150.26 125.83 133.6

69.82 50.77 126.96 62.76

217.12 224.26 187.81 193.09

114.23 88.83 147.24 106.26

185.18 191.78 159.64 164.13

13.38 14.48 10.26 13.76

10.90 12.63 6.42 11.20

strength capacity and initial stiffness of MES2 were respectively reduced by 2.8% and 34.1% to that of MES1. However, the rotation capacities of connections with a 12 mm thick end plate are obviously larger than those of connections with an 18 mm thick end plate. For example, the rotation capacity of MEC1 increased by 28.6% to that of MEC2, while the rotation capacity of MEC3 increased by 38.6% to that of MEC4. Moreover, the maximum strength and initial stiffness of the joint type for circular CFST columns are more than that for square CFST columns, in the same of the connected endplates of blind bolted connection. For the 12 mm thick endplate, the maximum strength capacity and initial stiffness of the joint type for circular CFST columns respectively were enhanced by 15.6% and 30.4% to that for square CFST columns. For 18 mm thick end plate, the maximum strength capacity and initial stiffness of the type joint for circular CFST columns were respectively enhanced by 16.1% and 5.2% to that for square CFST columns. 4.3. Evaluation of connection rigidity There is still no particular specification for the blind bolted end plate connections to CFST columns. In order to assert the connection rigidity in the moment–rotation curves for the extended end plate connections to CFST columns, the typed connection may be classified by EC3 Part 1-8 [26], specially for steel connections is used to evaluate the connections with composite columns tentatively in this paper. The classification of connections in EC4 Part 1‐1 [38] has been made consistent with the approach in EC3 Part 1‐8 [26]. EC4 Part 1‐1 [38] is restricted to the composite connections in braced frames only. A connection can be classified in terms of its rigidity and strength as follows. 4.3.1. Classification by rigidity A connection can be classified as rigid, nominally pinned, or semi-rigid according to its rotational stiffness by comparing its initial stiffness with classification boundaries suggested in EC3 Part 1‐8 [26] as:

4.3.2. Classification by strength A connection can be classified as full-strength, nominally pinned, or of partial strength in terms of strength by comparing its moment resistance with the plastic moment resistance of the beam. • full strength, if Mu ≥ Mbp; • nominally pinned, if M u ≤ 0:25Mbp ; otherwise, it is partial strength. Mbp is the design plastic moment resistance of the beam. According to the above classification method of connections, the analysis results for non-sway frame and sway frames are shown in Fig. 17. In Fig. 17, horizontal ordinate, θ, and vertical ordinate, m, respectively indicate dimensionless of connection rotation and moment. m ¼ M=M bp

ð2Þ

θr EIb ⋅ Mbp Lb

ð3Þ

θ¼

Fig. 17 shows that the blind bolted extended end plate connections in the specimens presented herein may be classified as semi-rigid and full strength, except that specimen MES1 behaves semi-rigid and partial strength. It is also found that the rotation capacities of the specimens satisfy the ductility requirement of no less than 30 mrad for earthquake resistance, suggested by FEMA-350 [39]. 4.4. Strains of steel beam

• rigid, if Sj;ini ≥ kb EIb =Ib ; where kb = 8 for non-sway frames and kb = 25 for sway frames. • nominally pinned, if Sj;ini ≤ 0:5EIb =Ib ; otherwise, it is semi-rigid. EIb is the flexural rigidity for the beam and Lb is the beam span. Table 5 Moment coefficient, rotation coefficient, and stiffness coefficient of connection. Specimen

Cm1

Cθ1

Cθ2

Cθ3

Ck1

MEC1 MEC2 MES1 MES2

0.85 0.91 0.73 0.88

2.56 2.25 3.33 2.66

4.69 4.02 4.77 4.84

4.19 3.93 3.87 4.51

0.81 0.87 0.63 0.81

The strain gauges on the steel beam were arranged a distance of 80 mm from the column face, as shown in Fig. 18(a). The measured cross-section strains of the steel beam end in the specimens are illustrated in Fig. 19, where, the positive values and negative values denote tensile strains and compressive strains, respectively. It demonstrates that except for the strains of the beam flange, the strains were linearly distributed over the section of the beam when the ratio of the test load to the test ultimate load (P / Pu) of the beam was less than 0.69. However, the distribution of strains over the section of the beams became non-linear when P / Pu is greater than 0.69, as shown in Fig. 19. Meanwhile, the compressive strain in the bottom flange of the beam also increased significantly. The strain gauges on the top and bottom flanges in the steel beam were arranged in Fig. 18(b). Fig. 20 shows the response of the top and bottom flange strains in the steel beam adjacent to the connection region for all specimens tested. The yield strain of the steel beam flange was 1868 με. It should also be noted that strain levels of the top and bottom flanges for the connections with an 18 mm thick end plate are obviously less than those for the connections with a 12 mm thick end plate; the beam flange strains for the connection to circular CFST column are less than those for the connection to square CFST columns.

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

a) Non-sway frame

b) Sway frame 1.2

1.2 Rigid/full strength connection

1

MEC1

MES2

0.8

0.6 0.4 0.2 0

0.4

0.4

Pinned connection

0.2

0.8

1.2

1.6

2

MES2

MEC1

0.6

MES1

MEC2

Rigid/full strength connection

1

m

m

0.8

0

67

0

MES1

MEC2

Pinned connection 0

0.4

0.8

θ

1.2

1.6

2

θ Fig. 17. Classification of tested joints.

4.5. Strains of end plate Fig. 21 illustrates the strains of the end plate of all tested specimens. The yield strains of the end plates with 12 mm and 18 mm thickness were 1633 με and 1422 με, respectively. The maximum tensile and compressive strains of the end plate with 12 mm thickness basically yield, while the maximum strains of the end plates with 18 mm thickness are below their yield strain. Especially in specimen MEC2 with circular CFST columns, the maximum strain of the end plate is only 51% of the yield strain. The strain distribution for the strain gauge numbers 14 and 16 in all specimens is shown in Fig. 21. The test results demonstrated that the strains of the end plate for concrete-filled circular steel tube column connections were less than those for concrete-filled square steel tube column connections. The strains of the end plate with 18 mm thickness for circular column connection are maximum in all specimens, but the strains of the end plate with 12 mm thickness for square column connection are minimum. This phenomenon illustrates that the end plate thickness and the column section type can affect the strain distribution of the end plates.

The test results show that the steel tube strains of all specimens at the maximum moment during the tests are not beyond the yield strain of the circular and square steel tubes. With the thickness of the end plate increasing, the strain level of the steel tube reduces 15–38% compared with specimens MEC1 and MEC2, while it reduces 7–20% compared with specimens MES1 and MES2. Moreover, the strains in the circular steel tube of the column wall are less than those in the square steel tube, assumed with the same thickness end plate. It is evident that the thickness of the end plate and the column section type may affect the strain distribution of CFST columns. The variation of steel tubular strains of the column in the panel zone against the connection moment is shown in Fig. 23. It is shown that these strains do not exceed the yield strain of the steel tube. Moreover, the strains in the circular steel tube of the column in the panel zone are less than those in the square steel tubular of the column; the strains in the steel tubular of the column in the panel zone reduce when the thickness of the end plate is increased. Therefore, the effects of the end plate thickness and the column section type on the strain distribution of the concrete-filled steel tubular columns were validated again.

4.6. Strains of column The distribution of the steel tube strains along the longitudinal and transverse direction of the column wall close to the steel beam is illustrated in Fig. 22. Fig. 22 also shows the steel tube strains of the column wall close to the steel beam for test specimens at maximum moment. In Fig. 22, the positive values denote tensile strain, while negative values denote compressive strains. The yield strains of the circular and square steel tubes with 10 mm thickness are 1710 με and 1632 με, respectively.

a) Cross section of beam end

Strain gauges of beam top flange

End plate

1 2

Strain gauges of beam web

The following observations and conclusion can be drawn with the limitation of the experimental research reported in this paper. This paper has described and reported on static tests of joints between open section beams and CFST columns using blind bolts with extension. The test demonstrated that utilising moderately thick end

b) Top and bottom flanges

3 4 5

60 60 60 60 60

Strain gauges of beam top flange

5. Conclusions

9

8

7

1

12

11

10

6

6 Strain gauges of beam bottom flange

80

300 Fig. 18. Layout of strain gauges in steel beam.

Strain gauges of beam bottom flange 250 80 300

End plate

68

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

2000

2000

1000

1000

Beam section strain (µε)

b) MEC2

Beam section strain (µε)

a) MEC1

0 -1000

α=0.22 α=0.47 α=0.58 α=0.69 α=0.79 α=0.90

-2000 -3000 -4000

0

50

100

150

200

250

0 -1000

-3000 -4000

300

α=0.22 α=0.47 α=0.58 α=0.69 α=0.79 α=0.90

-2000

0

50

Distance from base of beam (mm)

2000

2000

1000

1000

Beam section strain (µε)

d) MES2

Beam section strain (µε)

c) MES1

0 -1000

α=0.22 α=0.47 α=0.58 α=0.69 α=0.79 α=0.90

-2000 -3000 -4000

0

50

100

150

200

250

100

150

200

250

300

Distance from base of beam (mm)

300

0 -1000

α=0.22 α=0.47 α=0.58 α=0.69 α=0.79 α=0.90

-2000 -3000 -4000

0

50

Distance from base of beam (mm)

100

150

200

250

300

Distance from base of beam (mm)

Fig. 19. Strain distribution on a beam cross-section from specimens.

plates leads to joints approaching full strength for extended end plate connections. The strength and stiffness of the extended end plate joints to CFST columns with high strength blind bolts may be influenced by the end plate thickness and the column section type, which is the same with the test results of flush end plate joints. The strength and stiffness of nominally identical connections can be remarkably enhanced by

250

MES1

End plate (18mm) yield strain ε yr,e =1422με MEC2_14#

200 MEC1

150

Moment (kN.m)

Moment (kN.m)

250

Yield Strain of beam flange ε MES1 yr,f =1868με MES2

200

an increased end plate thickness. The test results also demonstrated that the strength and stiffness of the connections can be improved by providing anchorage extensions to the blind bolts, according to failure modes and strain response. The blind bolted extended end plate connection to CFST columns is classified as semi-rigid and full strength by EC3 specification, except that the square CFST column connection with thin end plates behaves in a semi-rigid and partial strength manner.

MES2

MEC2

MEC1

100 MEC2

50

MEC2_16# MEC1_16#

150

MES2_14#

MEC1_14#

MES2_16#

100 MES1_14#

50

0 -3000

-2000

-1000

0

1000

Beam flange strain (µε) Fig. 20. Beam flange strain response.

2000

MES1_16#

End plate (12mm) yield strain ε yr,e =1633με

Bottom flange strain gauge No.6

Top flange strain gauge No.1

3000

0 -3000

-2000

-1000

0

1000

End plate strain (µε) Fig. 21. End plate strain response.

2000

3000

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

a) Strain gauge No. 23-25

b) Strain gauge No. 28-30 2000

Steel tube strain (µε)

Steel tube strain (µε)

0

-500

-1000

MEC1 MEC2 MES1

steel tube yield strain -1500 Square ε yr,c

MES2

=1632 με

Circular steel tube yield strain ε yr,c =1710 με

-2000 320

340

360

380

400

MEC2 MES1

1000

MES2

500

-400

-380

-1000

Steel tube strain (µε)

-500

MEC1 MEC2 MES1

steel tube yield strain -1500 εSquare =1632 με yr,c

MES2

Circular steel tube yield strain εyr,c =1710 με

20

-360

-340

-320

d) Strain gauge No. 30-32 2000

0

Square steel tube yield strain ε yr,c =1632 με

Vertical distance from centre of column (mm)

0

Steel tube strain (µε)

MEC1

0 -420

420

c) Strain gauge No. 25-27

Circular steel tube yield strain ε yr,c =1710 με

1500

Vertical distance from centre of column (mm)

-2000

69

40

60

80

Vertical distance from centre of column (mm)

Circular steel tube yield strain εyr,c =1710 με MEC1

1500

MEC2

Square steel tube yield strain εyr,c =1632 με

MES1 MES2

1000

500

0 -420

-400

-380

-360

-340

-320

Vertical distance from centre of column (mm)

Fig. 22. Steel tube strains at maximum moment.

The blind bolted extended end plate connection to CFST columns exhibits high strength and stiffness, while its rotation capacity satisfies the ductility requirement for earthquake resistance in seismic region. The work in this paper provides a basis for further theoretical study of mechanical behaviour of the typed blind bolted connections. Acknowledgements The research reported in this paper is part of Project 51178156 and Project 50808062 supported by the National Natural Science Foundation of China. This financial support is highly appreciated, and the authors also wish to thank Prof. Liu B.K. and Zhou A. for their assistance in the tests. The authors would also like to acknowledge the Project 200803591022 supported by Doctoral Fund of Ministry of Education of China. References [1] Han LH, Yao GH, Zhao XL. Tests and calculations of hollow structural steel (HSS) stub columns filled with self-consolidating concrete (SCC). J Construct Steel Res 2005;61(9):1241-69. [2] Schneider SP, Alostaz YM. Experimental behaviour of connections to concretefilled steel tubes. J Constr Steel Res 1998;45(3):321-52. [3] Elremaily A, Azizinamini A. Experimental behavior of steel beam to CFT column connections. J Constr Steel Res 2001;57(10):1099-119. [4] Kang CH, Shin KJ, Oh YS, et al. Hysteresis behavior of CFT column to H-beam connections with external T-stiffeners and penetrated elements. Eng Struct 2001;23(9): 1194-201. [5] Beutel J, Thambiratnam D, Perera N. Cyclic behaviour of concrete filled steel tubular column to steel beam connections. Eng Struct 2002;24(1):29-38. [6] Cheng CT, Chung LL. Seismic performance of steel beams to concrete-filled steel tubular column connections. J Constr Steel Res 2003;59(3):405-26.

[7] Ricles JM, Peng SW, Lu LW. Seismic behavior of composite concrete filled steel tube column-wide flange beam moment connections. J Struct Eng ASCE 2004;130(2): 223-32. [8] Wang WD, Han LH, Uy B. Experimental behaviour of steel reduced beam section to concrete-filled circular hollow section column connections. J Constr Steel Res 2008;64(5):493-504. [9] Han LH, Li W. Seismic performance of CFST column to steel beam joint with RC slabs: experiments. J Constr Steel Res 2010;66(11):1374-86. [10] Li W, Han LH. Seismic performance of CFST column-to-steel beam joints with RC slab: analysis. J Constr Steel Res 2011;67(1):127-39. [11] Klippel S. Recent design developments with blind mechanically operated bolt systems for use with hollow section steelwork. J Constr Steel Res 1998;46(1–3):267-8. [12] Barnett T, Tizani W, Nethercot A. The practice of blind bolting connections to structural hollow sections: a review. Steel Compos Struct 2001;1:1–16. [13] Gardner AP, Goldsworthy HM. Experimental investigation of the stiffness of critical components in a moment-resisting composite connection. J Constr Steel Res 2005;61(5):709-26. [14] Goldsworthy HM, Gardner AP. Feasibility study for blind-bolted connections to concrete-filled circular steel tubular columns. Struct Eng Mech 2006;24(4):463-78. [15] Yao H, Goldsworthy H, Gad E. Experimental and numerical investigation of the tensile behavior of blind-bolted T-stub connections to concrete-filled circular columns. J Struct Eng 2008;134(2):198-208. [16] Yao H, Goldsworthy H, Gad E. Structural behaviour of beam-to-RHS column connection with blind bolts. In: Fragomeni, et al, editor. The 21st Australasian conference on the mechanics of structures and materials; 2010. p. 53-8. Melbourne. [17] Lee J, Goldsworthy HM, Gad EF. Blind bolted T-stub connections to unfilled hollow section columns in low rise structures. J Constr Steel Res 2010;66(8–9):981-92. [18] Lee J, Goldsworthy HM, Gad EF. Blind bolted moment connection to sides of hollow section columns. J Constr Steel Res 2011;67(12):1900-11. [19] Lee J, Goldsworthy HM, Gad EF. Blind bolted moment connection to unfilled hollow section columns using extended T-stub with back face support. Eng Struct 2011;33(5):1710-22. [20] France JE, Davison JB, Kirby PA. Strength and rotational response of moment connections to tubular columns using flowdrill connectors. J Constr Steel Res 1999;50(1):1–14. [21] France JE, Davison JB, Kirby PA. Strength and rotational stiffness of simple connections to tubular columns using flowdrill connectors. J Constr Steel Res 1999;50(1): 15-34.

70

J. Wang, L. Chen / Journal of Constructional Steel Research 79 (2012) 56–70

a) MEC1

b) MEC2

250

250 200

Moment (kN.m)

Moment (kN.m)

200 41# 150

40# 39#

100 50 0 -1000

40# 39#

150 100

41# 50

-500

0

500

0 -1000

1000

-500

Tube wall strain (µε)

c) MES1

d) MES2

250

250

41#

150 100 50 0 -1000

500

1000

39#

200

Moment (kN.m)

Moment (kN.m)

40#

39#

40#

200

0

Tube wall strain (µε)

150

41#

100 50

-500

0

500

1000

0 -1000

-500

0

500

1000

Tube wall strain (µε)

Tube wall strain (µε) Fig. 23. Steel tube strain response in the panel zone.

[22] France JE, Davison JB, Kirby PA. Moment-capacity and rotational stiffness of endplate connections to concrete-filled tubular columns with flowdrilled connectors. J Constr Steel Res 1999;50(1):35-48. [23] Loh HY, Uy B, Bradford. The effects of partial shear connection in composite flush end plate joints: part I—experimental study. J Constr Steel Res 2006;62(4): 378-90. [24] Wang JF, Han LH, Uy B. Behaviour of flush end plate joints to concrete-filled steel tubular columns. J Constr Steel Res 2009;65(4):925-39. [25] Wang JF, Han LH, Uy B. Hysteretic behaviour of flush end plate joints to concrete-filled steel tubular columns. J Constr Steel Res 2009;65(8–9):1644-63. [26] CEN. Eurocode 3: design of steel structures—part 1‐8: design of joints. ENV 1993-1-8. Brussels: CEN; 2005. [27] Mirza O, Uy B. Behaviour of composite beam–column flush end-plate connections subjected to low-probability, high-consequence loading. Eng Struct 2011;33(2): 647-62. [28] Elghazouli AY, Malaga-Chuquitaype C, Castro JM, Orton AH. Experimental monotonic and cyclic behaviour of blind-bolted angle connections. Eng Struct 2009;31(11): 2540-53. [29] Malaga-Chuquitaype C, Elghazouli AY. Component-based mechanical of blindbolted angle connections. Eng Struct 2010;32(10):3048-67. [30] Ghobarah A, Osman A, Korol RM. Behaviour of extended end-plate connections under cyclic loading. Eng Struct 1990;12(1):15-27.

[31] Adey BT, Grondin GY, Cheng JJR. Extended end plate moment connections under cyclic loading. J Constr Steel Res 1998;46(1–3):435-6. [32] Girao Coelho AM, Bijlaard FSK, Silva. Experimental assessment of the ductility of extended end plate connections. Eng Struct 2004;26(9):1185-206. [33] Shi YJ, Shi G, Wang YQ. Experimental and theoretical analysis of the moment– rotation behaviour of stiffened extended end-plate connections. J Constr Steel Res 2007;63(9):1279-93. [34] Korol RM, Ghobarah A, Mourad S. Blind bolting W-shape beams to HSS columns. J Struct Eng 1993;119(12):3463-81. [35] Mourad S, Ghobarah A, Korol RM. Dynamic response of hollow section frames with bolted moment connections. Eng Struct 1995;17(10):737-48. [36] GB50017-2003. Code for design of steel structures. Beijing: China Planning Press; 2003 [in Chinese]. [37] DB34/T 1262‐2010. Technical specification for concrete-filled steel tubular structures. Hefei: Anhui Engineering Construction Standard Design Press; 2010 [in Chinese]. [38] CEN. Eurocode 4: design of composite steel and concrete structures—part 1‐1: general rules and rules for buildings. ENV 1994-1-1. Brussels: CEN; 2004. [39] FEMA-350. Recommended seismic design moment-frame buildings. Federal Emergency Management Agency; 2000.