Experimental study on seismic behavior of high strength steel frames: Global response

Engineering Structures xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Engineering Structures journal homepage: www.elsevier.com/locate/engstruct

Experimental study on seismic behavior of high strength steel frames: Global response Fangxin Hu a,b, Gang Shi c,d,⇑, Yongjiu Shi c,d a

School of Civil Engineering and Transportation, South China University of Technology, Guangzhou 510640, China State Key Laboratory of Subtropical Building Science, South China University of Technology, Guangzhou 510640, China c Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry, Department of Civil Engineering, Tsinghua University, Beijing 100084, China d Beijing Engineering Research Center of Steel and Concrete Composite Structures, Tsinghua University, Beijing 100084, China b

a r t i c l e

i n f o

Article history: Received 3 January 2016 Revised 2 November 2016 Accepted 4 November 2016 Available online xxxx Keywords: High strength steel Steel frame Experiment Seismic behavior Global response

a b s t r a c t Cyclic tests were performed on six full-scale single-bay two-story frames, including one frame using Q345 (fy = 345 MPa) ordinary strength steels, and another five frames using Q460 (fy = 460 MPa) high strength steels or Q890 (fy = 890 MPa) ultra-high strength steels in only columns or both beams and columns. The frames were designed to provide strength and ductility for earthquake resistance with energy dissipation located at member ends by flexural yielding, and panel zones by shear yielding. Cover-plate reinforced connections were specified to relocate the plastic hinge beyond the nose of cover plates and away from the face of column to improve connection performance. As the first of two companion papers on this experimental study, this paper presents detailed procedures to design the specimens, and then provides the outline of test program including test setup, loading protocol and instrumentation. After that, global responses in experimental results, including test observations of each specimen throughout the test and their hysteresis and backbone curves, deformation and energy dissipation in each half-cycle and in total, were described. The results evidenced that satisfactory seismic behavior was identified by using high strength steel columns with compact or noncompact sections in frames, and the maximum overall drift ratio reached 4.0%. Even the frame with ultra-high strength steel columns with slender sections accommodated an overall drift ratio of 3.0%. No soft-story mechanism occurred in all the frames. Local responses in beams, columns and connections, especially in panel zones, will be examined in the second companion paper. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction 1.1. Background and motivation With the rapid development of new production and welding techniques, high strength steels with nominal yield strength fy P 460 MPa (ultra-high strength steels with fy P 690 MPa) have been used in engineering structures, due to their substantial advantages in architectural style, structural safety and economic benefit [1–4]. A comprehensive literature review in recent research advances of high strength steel structures has been conducted by the authors [5]. However, the majority of research advances in that paper focused attention on static behavior, such as cross-sectional strength, stability of beams and columns, shear behavior of panel ⇑ Corresponding author at: Key Laboratory of Civil Engineering Safety and Durability of China Education Ministry, Department of Civil Engineering, Tsinghua University, Beijing 100084, China. E-mail address: [email protected] (G. Shi).

zones and direct analysis for frames made of high strength steels under static load [6–23]. Quite few studies have been found with respect to seismic behavior, since multi-story frame structures using high strength steel members represent an innovation in current seismic design. As the first step to investigate seismic behavior of high strength steel structures, Huang et al. [24,25], Shi et al. [26,27], Wang et al. [28], Dusicka et al. [29] and Miyazaki et al. [30] carried out cyclic loading experiments of high strength steel materials and weld connections, so that constitutive models could be calibrated to describe accurately their nonlinear cyclic plasticity. Then, with respect to strength, ductility and energy dissipation capacity of structural members involving high strength steels, Ricles et al. [8], Green [31] and Suzuki et al. [32] conducted a series of cyclic loading experiments on welded I-shaped beams made of HSLA-80 steels (fy = 552 MPa) and high strength steels with tensile strength equal to 590 MPa, respectively, and carried out finiteelement analyses to evaluate the influence of material properties, cross-section geometry and applied loading; Dusicka et al. [33]

http://dx.doi.org/10.1016/j.engstruct.2016.11.013 0141-0296/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

2

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

experimentally examined cyclic response of HPS485 W steel (fy = 485 MPa) shear links and satisfactory results were obtained showing a potential advantage of low overstrength; Fukumoto and Kusama [34], Kuwamura and Kato [35], Shi et al. [36,37],Wang et al. [38] and Chen et al. [39], respectively, investigated cyclic behavior of columns under compression or combined compression and bending, which were made of various high strength steels including HT80 steels (fy = 700 MPa) made in Japan, Q460 (fy = 460 MPa) and Q690 (fy = 690 MPa) steels produced in China. Lin et al. [40,41] proposed a new kind of built-up columns made of H-SA700 steel plates (fy = 700 MPa) by high strength bolts and investigated their flexural performance by bending test, and then developed corresponding bolted beam-to-column connections for such built-up columns. Sato et al. [42] developed weld-free beam-to-column connections for H-SA700 steel members using knee brace damper and friction damper to dissipate energy. Sun et al. [43] also examined cyclic behavior of beam-to-column bolted connections using Q690 steel in columns and end plates. Furthermore, Matsui and Mitani [44] conducted an early experimental study on inelastic cyclic behavior of SM58 steel (fy = 470 MPa) frames subjected to constant vertical and alternating horizontal forces. Recently Tenchini et al. [45,46] and Dubina et al. [47,48] combined the use of high strength steels in non-dissipative members and mild carbon steel in dissipative zones, and lots of work were contributed to assess seismic performance of those dualsteel structures by experiments on structural components and analyses on frame buildings. In Japan, damage-free design method [49] has been proposed to greatly facilitate utilization of high strength steels and experimental studies by Shinsai et al. [50], Nakai et al. [51], Takeuchi et al. [52,53] were carried out on frames using H-SA700 steels in beams and columns and mild steels like SN400, SS400 in knee braces, concentric braces, or bucklingrestrained braces, so that those high strength steel members remained elastic even under the extremely rare earthquakes. The previous research clearly evidenced the effectiveness and advantages by using high strength steels in structural members desired to remain elastic under earthquakes, such as in columns in moment frames or both beams and columns in braced frames [54]; however, there is still a gap before high strength steels can be widely used in practice, since no systematic seismic design methods associated with such steels have been proposed in current codes. Even additional rules [55] to extend Eurocode 3 [56] up to steel grades S700 (fy = 700 MPa) are questionable due to a lack of substantial research background. Therefore, comprehensive investigation on high strength steel structures, especially full-scale experimental studies, are in urgent need in order to evaluate their system performance under strong earthquakes, and to develop safe and reliable seismic design guidelines for such steel structures. 1.2. Objective This paper presents a full-scale cyclic test program to study seismic behavior of single-bay two-story high strength steel frames, including two homogenous steel frames using Q460 (fy = 460 MPa) steels in both beams and columns, two hybrid steel frames using Q460 (fy = 460 MPa) steels in columns and Q345 (fy = 345 MPa) steels in beams, and one hybrid steel frame using Q890 (fy = 890 MPa) steels in columns and Q345 (fy = 345 MPa) steels in beams. Another homogenous ordinary strength steel frame using Q345 (fy = 345 MPa) steels has also been designed as the dummy specimen. Firstly, detailed design procedures and test program are illustrated in Sections 2 and 3, respectively. Then, observations throughout the test are described in Section 4. System global responses including strength, deformation and energy dissipation capacity are demonstrated in Section 5. Finally, several practical implications associated with the safety of current seismic

design are discussed in Section 6 by exploring test results, and conclusions are summarized in Section 7. Detailed local responses in beams, columns and connections, in particular, panel zones, are introduced in the companion paper. This study will provide significant research basis for the new Design Specification of High Strength Steel Structures [5] being codified in China. 2. Design of frame specimen 2.1. Prototype building Based on the constraints imposed by the laboratory, six specimens were designed, constructed, and tested, each of which consists of a single-bay two-story moment frame. In order to relate these specimens to an actual structure, the specimens were designed to be extracted from a six-story prototype building. Fig. 1(a) and (b) shows the plan and elevation of the prototype building, which was assumed to be located on stiff soil in Beijing, China. The typical bay span and story height were 6 m and 2.7 m in both directions. The characteristic dead (D) and live (L) loads were 6 kN/m2 and 2 kN/m2, respectively, for the floors and the roof. Effective seismic weights were taken as a combination of all dead loads and half of live loads (D + 0.5L), and were 4536 kN for the floors and roof, resulting in a total seismic weight of the building equal to 27,216 kN. The design followed Chinese Code for design of steel structures [57] with the seismic force stipulated in Chinese Code for seismic design of buildings [58]. The response spectrum adopted for seismic design is shown in Fig. 1(c), where the maximum response acceleration (amax) depends on the earthquake level (i.e., frequently occurred earthquake FOE, design basis earthquake DBE, or maximum considered earthquake MCE), and the characteristic period (Tg) is 0.35 s depending on the location. 2.2. Member design Beams and columns were made of welded wide-flanges in the shop, as shown in Fig. 2(a). Taking into account the lateral restraint provided by the slab in the actual building, the beams were verified as having sufficient flexural strength under three load combinations considered for ultimate limit state, i.e., 1.2D + 1.4L, 1.35D + 0.7
1.4L, and D + 0.5L + XEhk2, where Ehk2 was the seismic load corresponding to DBE defined in Fig. 1(c) and X was structural characteristic factor (i.e., Ds in Japanese code [59], the reciprocal of behavior factor q in Eurocode 8 [60] or response modification factor R in US code [61]) depending on the ductile behavior of structures. A recommended value of X = 0.4 in [57] was used. The flexural strength of beams was checked as [57],

Mb 6 fb Wb

ð1Þ

where Mb is the maximum bending moment under any load combination, Wb is the moduli of the cross-section of beams, and fb is the design strength of beam flange materials accounting for the partial factor cR, i.e. fb = fyb/cR where fyb is the yield strength of beam flanges. The columns in each floor were verified in combined compression and bending under the same two gravity load combinations (1.2D + 1.4L and 1.35D + 0.7
1.4L), and another amplified seismic load combination, i.e., D + 0.5L + 1.1gyXai,minEhk2, where gy is material overstrength factor (taken as 1.1) taking into account the possibility that the actual yield strength of steel is higher than nominal yield strength, and Xai,min is the actual minimum characteristic factor of the i-th floor which is determined by,

Xai;min ¼

min


 ðW b f y  MGE Þ=MEhk2

all beams in i-th floor

ð2Þ

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

3

h

h

l

h

h

l

h

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

l

h

Test frame (b)

l

(g) max max=

l

=(Tg/T)0.9 0.45

0.16 (FOE) 0.45 (DBE) 0.90 (MCE)

max

max max

l

=[0.20.9-0.02(T-5Tg)]

l

l

l

0.1 Tg

5Tg

(a)

6.0

T (s)

(c)

Cap plate

tfc tfc

c

2700

db

tfb

6

6

high strength bolt

tw

dc

6

tw b

35o

CJP weld

Cover plate

tfb

6

390

Fig. 1. Prototype building: (a) plan, (b) elevation and (c) response spectrum for seismic design.

Panel zone

Shear tab

bc

bb Beam cross-section

Column cross-section

Cover plate

Column

CJP weld

2700

30 230

Base plate

35o

Side view of the connection

180-dc/2 132 Stiffener

CJP weld

10

624 Column base

Cover plate

Beam

Continuity plate

6000

Top view of the connection

(a)

(b)

Fig. 2. Test frame specimen: (a) elevation drawing, and (b) connection details.

where MGE is the design bending moment due to effective seismic weights (i.e., D + 0.5L), MEhk2 is the design bending moment due to seismic load under DBE (i.e., Ehk2). In-plane and out-of-plane stability of columns were checked as [57],

Nc

u x Ac

þ

bm Mc 6 fc W c ð1  0:8Nc =N0Ex Þ

N0Ex ¼ p2 EAc =ð1:1k2x Þ

ð3Þ ð4Þ

Nc

uy Ac

þ

Mc

ub W c

6 fc

ð5Þ

where Nc and Mc are the most unfavorable combination of axial force and bending moment under any load combination; Ac and Wc are area and moduli of the cross-section of columns, respectively; ux and uy are in-plane and out-of-plane flexural buckling coefficients, respectively; ub is lateral-torsional buckling coefficient; kx is in-plane slenderness of columns; bm is equivalent

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

4

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

moment factor; E is elastic modulus of the material, and fc is the design strength of column flange materials accounting for the partial factor cR, i.e. fc = fyc/cR where fyc is the yield strength of column flanges. In addition, all members were designed to satisfy the requirements associated with serviceability limit state, i.e., deflection of beams was limited to l/250 (l is beam span) under the load combination (1.0D + 1.0L), and story drift ratio should not exceed 1/250 under the seismic load (1.0Ehk1) in FOE as shown in Fig. 1(c). The final cross-sections for beams and columns of six specimens are summarized in Table 1, where the cross-section classification is also indicated. Specimens are labeled with the steel grade used in beams and columns. The first dummy specimen is an ordinary strength steel frame used for both testing experimental facilities and comparison with other high strength steel frame specimens. 2.3. Connection design To avoid premature fracture in beam-to-column welds, coverplate reinforced moment connections were designed, as shown in Fig. 2(b). These connections utilized two cover plates which were wider than beam flanges and located outside the top and bottom beam flanges. The design guidelines proposed by Kim et al. [62] were referred to. Cover plates were shop fillet welded to beam flanges, and both of them were 35° grooved and CJP (complete joint penetration) welded to the column flange with prequalified welding wires. Ceramic backing bars were used, so that they could be removed conveniently after welding. Cover plates used the same material and thickness with beam flanges. The length of cover plates in all specimens was set equal to 340 mm, and their width was 30 mm wider (15 mm at each side) than beam flanges to facilitate comparison of responses. Such results would guarantee the composite cover plate-beam section at the face of column remained elastic assuming that a plastic hinge formed in the beam section at a distance db/3 (db is beam depth) from the end of cover plates. Three-sided fillet welds of 10 mm in size (2 longitudinal and 1 transverse) as shown in Fig. 2(b) were adopted that could develop the yield strength of cover plate. The improved access hole detail specified in [57] was followed as shown in Fig. 2(b), which is the same with that in AISC 358 [63]. The beam web was connected to a shear tab which was shop CJP welded to the column flange, using three class 10.9 s high strength bolts of M24 in size to develop design shear force in the beam (see Eq. (7)). In order to ensure an effective hierarchy of yielding under strong earthquakes, the strong-column weak-beam requirement specified in [57] writes,

P

W c ðf yc  Np =Ac Þ P P1 1:1gy ðW b f yb þ V b sÞ

ð6Þ

where Np is axial force in the column under the amplified seismic load combination (D + 0.5L + 1.1gyXai,minEhk2), s is the distance from the assumed plastic hinge to the column face (i.e., s = lp + db/3 where lp is the length of cover plate), and Vb is design shear force in the beam given by,

V b ¼ V GE þ

2W b f yb

ð7Þ

L  dc  2lp  23 db

where VGE is design shear force due to effective seismic weights (D + 0.5L), L is beam span, dc is column depth. The specimens were designed with different strength ratios but all less than 1.0 according to Eq. (6), as shown in the SCWB column in Table 1, to investigate the safety of current strong column-weak beam design formula and the effect of column yielding. Another critical issue in beam-to-column connections is the design of panel zones. It is generally recognized that inelastic panel zone deformation can substantially increase the deformation capacity of beam-column connections. As such, FEMA [64] proposed a balanced design to share the inelastic deformation between the beam under flexural yielding and the panel zone under shear yielding. In spite of this, the requirement in [57] for the strength of panel zone is given as, 4 h h t f 3 ob oc p yv

0:85

P P1 ðW b f yb þ V b sÞ

ð8Þ

where hob and hoc are the distance between flange centroids of the beam and column, respectively; tp is panel zone thickness, including doubler plates, if any; fyv is the yield strength of panel zone in shear pffiffiffi (i.e., f yv ¼ f y = 3). In order to investigate the effect of yielding in panel zones, the specimens were designed with different strength ratios according to Eq. (8), as shown in the SPWB column in Table 1. In particular, in specimen B460-C460-1 a doubler plate of the same material and thickness to the column web was added for each beam-to-column connection to examine such a case with very strong panel zones. Instead of embedding anchor bolts in foundation RC beams, 30 mm thick rectangular base plates (624 mm
340 mm) were shop CJP welded to the bottom of columns, and were connected using 24 high strength bolts to very strong and rigid pedestals, which in turn were securely tied down to the strong floor. Such column base connections could be treated as fully fixed. To avoid potential failure in CJP welds, two 20 mm thick trapezoidal stiffener plates were fillet welded to column flanges and the base plate in each column base connection, as shown in Fig. 2(a).

2.4. Material properties All the specimens were fabricated with welded wide-flange sections. Tensile coupon tests were carried out using five identical specimens for each kind of steel plate [65], and average measured results were obtained and shown in Table 2, where E is elastic modulus, fy is yield strength, est is the strain at the end of yield plateau, and fu and eu are ultimate strength and strain, respectively. Note that tensile stress-strain curves for Q890 ultra-high strength steels did not exhibit an obvious yield plateau. Thus, the proof stress corresponding to 0.2% residual plastic strain was treated as yield strength.

Table 1 Design of frame specimen. Specimen no.

B345-C345 B345-C460-1 B460-C460-1 B345-C460-2 B460-C460-2 B345-C890

Beam

Column

Steel

Section

Steel

Section

Q345 Q345 Q460 Q345 Q460 Q345

H340
170
8
10 (Compact) H340
170
8
10 (Compact) H340
110
10
10 (Compact) H340
110
8
16 (Compact) H340
110
10
10 (Compact) H340
110
8
16 (Compact)

Q345 Q460 Q460 Q460 Q460 Q890

H300
250
10
10 (Noncompact) H240
220
10
10 (Noncompact) H240
220
10
10 (Noncompact) H170
170
12
12 (Compact) H170
170
12
12 (Compact) H240
150
6
6 (Slender)

SCWB

SPWB

0.80 0.78 0.72 0.52 0.48 0.67

0.80 0.84 1.59 0.68 0.65 0.98

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

5

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx Table 2 Material properties of steel plates used in the test. Sample plates

E (MPa)

fy (MPa)

est (%)

fu (MPa)

eu (%)

Steel Gr.

Thickness (mm)

Q345

8 10 16

199,400 195,700 196,900

452 408 391

2.64 2.79 2.23

564 493 531

15.76 17.69 18.03

Q460

10 12

196,000 197,500

548 490

2.11 2.04

659 567

11.92 8.69

Q890

6

192,800

908

–

967

5.30

3. Test program

3.2. Construction sequence

3.1. Test setup

All the specimens were erected in the laboratory. Firstly, two columns were lifted and moved into position on the pedestals tied to the strong floor and temporarily secured by several 10.9 s M24 high strength bolts at base. Secondly, the lower and top beams were lifted up and inserted between two columns, and then temporarily bolted to the shear tabs on the interior side of column flange at both ends. After the whole specimen was aligned to insure that it was plumb and fit up, the brackets and wide flange sections providing lateral restraint for columns and beams were installed. Finally, the bolts at column bases were tightened to minimum required pretension forces using torque wrench. Welding of beam-to-column connections were continued using CO2 gas shielded arc welding. All welding consumables were low hydrogen wires and under-matched welding was used between steels of different grades. After cover plates and beam flanges were fully welded, the bolts connecting the beam web to the shear tab were tightened with required pretension forces. The final configuration of a specimen after construction is shown in Fig. 3(b).

The overview of final configuration of the test setup is shown in Fig. 3. The specimens were designed to fit in a 3-D loading frame system. Two 500 kN actuators with ± 250 mm stroke length were used, with one installed at each level. With these actuators, the target load corresponding to the maximum base shear that could be developed by the specimens was within the capacity and an upper level displacement equal to about 4% of the specimen height can be imposed. Strong brackets were installed on the reaction wall to attach the actuators. Another two 2500 kN jacks which were connected to the heavy built-up beam of the loading frame but could slide horizontally with the specimens’ sway were used to apply axial load on the top of both columns to represent the seismic weight in the prototype building. Column bases were fixed to the pedestals on the strong floor. To prevent lateral instability of the specimens, eight brackets were installed on the loading frame to provide lateral restraint for columns at each level, and several wide flange sections which were installed on two heavy columns tied down to the strong floor were used as out-of-plane bracing system for the top and lower beams at quarter positions, as shown in Fig. 3(a). It should be noted that for the first two specimens (B345-C345 and B345-C460-1) only brackets were used. Since obvious lateraltorsional buckling of beams occurred in the latter specimen, the wide flange sections were added to continue the test for the other specimens.

West-side (negative)

3.3. Loading protocol After applying a constant axial load of 756 kN (i.e. the effective seismic weight supported by each side column in Fig. 1) on the cap plate which was shop CJP welded to the top of each column, all the specimens were tested quasi-statically, with a prescribed history of the top displacement (i.e. the displacement at the upper beam level) imposed. The displacement at the upper beam level was monitored and controlled during the entire test process; however,

Jack

East-side (positive)

Jack

Actuator Jack

Lateral restraint

Jack

Actuator

Actuator

Specimen

Wide flange sections Brackets

Brackets Actuator

Specimen

Pedestal

Pedestal Pedestal

View from south side

View from west side (a)

View from north-east side (b)

Fig. 3. Test setup: (a) scheme, and (b) real.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

6

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

the lower level actuator was force controlled. Throughout the test, the force applied at the lower level was one twentieth of the instantaneous force measured in the load cell of the upper level actuator using the force-displacement mixed control system developed by Pan et al. [66]. This makes the lateral force pattern imposed on the two-story specimen equivalent to an inverted triangular distribution on the six-story prototype building. The displacement protocol recommended in AISC Seismic Provisions [67] was imposed at the upper level, as shown in Fig. 4. Due to the limit of actuator stroke, however, after two cycles at 3% overall drift ratio (i.e. the top displacement divided by the specimen height of 5400 mm) the specimen were tested at a constant amplitude of 4% overall drift ratio until whenever major events such as unanticipated fracture or weld cracking were observed, or the lateral strength of the specimen decreased from its peak value by 15%. 3.4. Instrumentation

Overall drift ratio (%)

5 4 3 2 1

0.375%

0.5%

4.0% 3.0% 2.0% 1.5% 0.75% 1.0%

0 -1 -2 -3 -4 -5 Fig. 4. Loading protocol for the top displacement.

250 200 150 100 50 0 -50 -100 -150 -200 -250

Top displacement (mm)

Each specimen was extensively instrumented with displacement transducers, strain gauges and rosettes, and load cells, as shown in Fig. 5. A load cell attached to the head of each actuator (LC-1 and LC-2 in Fig. 5(a)) and jack (LC-3 and LC-4 in Fig. 5(a)) measured the horizontal and axial load applied. Displacement transducers having a variety of gauge lengths consisted of those used to measure actuator displacements, floor level displacements (D-1 and D-2 in Fig. 5(a)) and shear deformation of panel zones (D9 to D-12 in Fig. 5(a)) at the east end of the specimen (i.e. away from the actuator side). The movement of both column base plates was also monitored horizontally and vertically (D-3 to D-8 in Fig. 5 (a)), as well as out-of-plane displacements in the middle of both beams (D-13 and D-14 in Fig. 5(a)). Nearly 134 strain gauges and rosette strain gauges were utilized for each specimen, most of which attached to portions of beams and columns where elastic behavior was expected. Four strain gauges were glued on each column surface at two cross-sections (Section S2 in Fig. 5(a)), each located at a distance of 750 mm measured toward the column mid-height from the center line of beams. Similarly, two cross-sections (Section S1 in Fig. 5(a)) on each beam were also measured, but at a distance of 1500 mm measured from the center line of columns. Those strain gauges were aligned with the flange quarter positions as shown in Fig. 5(b), and engineering principles were then used to estimate curvatures, average axial strains, bending moments, shears, and axial loads in these members. Strain gauges and rosettes were also attached to cover-plate connections, panel zone regions and continuity plates to measure strains for evaluation of load transfer mechanism and subsequent comparison with numerical predictions, as shown in Fig. 5(b). Furthermore, six strain gauges were glued to each column surface (Section S3 in Fig. 5(a)) which was just beyond the nose of stiffener plates at the column base, and were aligned with the

flange quarter positions and centerline (see Fig. 5(b)) to examine local plastic behavior. A total of 156 instrumentation channels were connected to the data acquisition system to record the data for processing.

4. Test observations Before conducting the formal test, two small complete cycles of 0.1% overall drift ratio were performed to check the operation of data acquisition and control systems. According to the loading protocol, for each specimen the test process was divided into several stages of loading (amplitudes or phases). Every loading stage contained several complete cycles, i.e., 6 cycles for loading stages with 0.375%, 0.5% and 0.75% overall drift ratios, respectively, 4 cycles for the 1.0% loading stage, and 2 cycles for subsequent loading stages. Major observations in each of these cycles for each specimen are described sequentially below. 4.1. Specimen B345-C345 The entire frame essentially remained elastic and no special findings were found during 0.375%, 0.5%, 0.75% and 1.0% loading stages. At the first half-cycle of the 1.5% loading stage when the top displacement approached the positive peak, notable panel zone deformation in the first story was observed and the story deformation had a tendency to concentrate at the first story. At the second cycle of the same loading stage, local buckling of column flanges was found at both east-side and west-side column bases, when the peak first story drift ratio was about 1.4%. The test was continued to the 2.0% loading stage where additional significant local buckling of column webs and thus plastic hinges at column bases developed at the second cycle. After passing the second positive peak top displacement at the 3.0% loading stage, some cracks initiated between column flanges and the web at the east-side column base due to rather severe local buckling, and instantaneous axial squash at both column bases was noticed which resulted in significant vertical shortening. The test was stopped after the first complete cycle of the 4.0% loading stage. An evolution of buckling deformation in the east-side column base from the 1.0% loading stage till the end of test (story drift ratio at the first story is also included as the average value from both positive and negative peaks at the corresponding loading stage) and the final deformation of the west-side column base are shown in Fig. 6. Despite significant shear deformation in panel zones, no crack or fracture occurred in all the beam-to-column connections. 4.2. Specimen B345-C460-1 No special observations were made anywhere through loading stages with overall drift ratios less than or equal to 2.0%. Local flange buckling at both column bases was observed at the positive peak of the first half-cycle of the 3.0% loading stage. When the top displacement continued to move toward the negative peak to the west, twist of the lower beam was found due to the absence of out-of-plane deformation restraint system (see Fig. 7(a)). At the 4.0% loading stage, even the upper beam began to exhibit slight torsion. Some minor local buckling of the web in column bases was noted, and the column flange buckled obviously in the firststory east-side beam-to-column connection at this stage (see Fig. 7(b)). The test was stopped after the second complete cycle since the measured base shear reduced from its peak value obviously due to twist of beams. Evolution of buckling deformation in column bases are shown in Fig. 8. All the beam-to-column connections were nearly undamaged.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

7

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

S2

750

LC-3 West 1500 LC-1 S1

S1

LC-4 East 1500 D-12 D-11D-2 Load cell

D-14

S2

S2

S1

750

LC-2

Section S2

Section S3 Displacement transducer

S2

750

S2

Section S1

S1

D-10

D-13

D-9 D-1

Top view of the connection

Strain gauge

S2 Strain rosette S2

750

S2 S3

D-7 D-8

S3 D-6

Column

Beam

D-3

D-4 D-5

Side view of the connection

(a)

(b)

Fig. 5. Instrumentation: (a) load cells and displacement transducers, and (b) strain gauges and rosettes.

crack initiation

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 6. Buckling deformation in the east-side column base at the end of the last cycle of (a) 1.0% loading stage (0.8% at the first story), (b) 1.5% loading stage (1.4% at the first story), (c) 2.0% loading stage (2.0% at the first story), (d) 3.0% loading stage (3.3% at the first story), (e) 4.0% loading stage (4.6% at the first story), and (f) in the west-side column base after test.

4.3. Specimen B460-C460-1

Fig. 7. (a) Twist of the lower beam and (b) local buckling of the column flange in the first-story east-side beam-to-column connection.

Since out-of-plane instability of beams occurred in the previous specimen, several wide flange sections were added to provide lateral restraint for beams before continuing the test. Similar to specimen B345-C460-1, no special observations were made anywhere until at the first half-cycle of the 3.0% loading stage, when local buckling deformation was observed in column flanges at bases. At the 4.0% loading stage, some torsion occurred at the top of the second-story west-side column, thus resulting in out-of-plane bending deformation of the associated second-story beam in the connection region. Additional local web buckling in column bases also developed at this stage. The measured base shear of this specimen after two complete cycles of the 4.0% loading stage was still above 85% of the peak maximum value experienced in cycles of the 3.0% loading stage; therefore, an additional third cycle at the 4.0% loading stage was imposed. When the specimen moved to the west in this cycle, load popping noises were heard when a crack in the first-story east-side beam-to-column connection initiated and quickly propagated through the entire continuity plates and

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

8

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

Doubler plate

fracture

Beam

Beam

Column flange

Colu mn flan ge

Fig. 8. Buckling deformation in the east-side column base at the end of (a) the second cycle of 2.0% loading stage (1.8% at the first story), (b) the first and (c) second cycle of 3.0% loading stage (2.7% at the first story), (d) the first and (e) second cycle of 4.0% loading stage (3.5% at the first story), and (f) in the west-side column base after test.

Column web fracture Continuity plate

Continuity plate

(a)

(b)

Fig. 9. Fracture between the column flange and continuity plates and the column web: view from (a) north side and (b) south side.

column web along the edge of the doubler plate to form a partial fracture between the column flange and other plate components (see Fig. 9). Significant load drop was observed compared with previous cycles and the test was stopped after the third cycle was completed. Evolution of buckling deformation in column bases are shown in Fig. 10. The side view of the first-story beam-tocolumn connection at the east side after test is shown in Fig. 11. The inside column flange experienced a notable bending deformation due to the failure of continuity plates to transfer the load from cover plates and beam flanges.

fracture

Fig. 11. View of the first-story east-side beam-to-column connection in specimen B460-C460-1 after test.

4.4. Specimen B345-C460-2 local buckling

No special findings were observed before the 4.0% loading stage. Due to the compact section used for columns, notable local buckling of flanges just began to develop in column bases at the first half-cycle of the 4.0% loading stage when the top displacement approached the positive peak. As shown in Fig. 12, column flanges

(a)

local buckling

(b)

Fig. 12. Local buckling of column flanges at the top of second-story (a) east-side and (b) west-side columns.

Fig. 10. Buckling deformation in the east-side column base at the end of (a) the first and (b) second cycle of 3.0% loading stage (2.8% at the first story), (c) the first, (d) second and (e) third cycle of 4.0% loading stage (4.1% at the first story), and (f) in the west-side column base after test.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

at the top of both second-story columns buckled at this loading stage, which indicated the formation of plastic hinges. The test was stopped after three complete cycles were finished, although almost no load drop was observed. Evolution of buckling deformation in column bases are shown in Fig. 13. All the beam-to-column connections were nearly undamaged.

9

5.2. Strength and stiffness

5. Frame response

Backbone curves were constructed from base shear-overall drift ratio hysteresis curves by connecting end points in each loading stage for both the first and last cycle, as shown in Fig. 22. For specimen B345-C345, an obvious degradation of peak base shear between the first and second cycle at the 2% loading stage was mainly due to the quick deterioration of the strength of column bases as a result of local buckling (see Fig. 6(c)), and compared with the maximum base shear developed at the 1.5% loading stage, the lateral strength experienced about 10%, 20% and 30% deterioration at the 2%, 3% and 4% loading stages, respectively; while for specimens B345-C460-1 and B460-C460-1, such degradation in lateral strength at the 4% loading stage was about 15% and 7% respectively compared with the maximum based shear developed at the 3% loading stage. This was attributed to the twist of the lower beam (see Fig. 7(a)) and the partial fracture of connection (see Fig. 9), respectively. Furthermore, specimen B345-C890 was substantially damaged in column bases (see Fig. 15(d)) after the first cycle at the 3% loading stage, and it could not even afford a second complete cycle at this loading stage before a sudden squash in severely buckled column bases was triggered by the constant axial load (see Fig. 15(e) and (f)). This damage resulted in nearly 50% drop in the positive base shear between the first and second cycle. No special degradation of base shear was found for specimens B345-C460-2 and B460-C460-2 from the very beginning of test to the maximum 4% overall drift ratio. This could be explained by the rather compact cross-sections used for columns, which provided a stable strength although some local buckling occurred due to large cumulated plastic deformation (see Figs. 13 and 14). As shown in Fig. 22, lateral stiffness (Ke) for each specimen was evaluated by the initial elastic slope based on a best linear fit to peak points within the 0.5% loading stage in backbone curves. It is clear that by using high strength steels, smaller cross-sections could be used in individual members and this greatly reduced global stiffness of the frame specimen.

5.1. Lateral force-drift ratio relationship

5.3. Deformation and energy dissipation capacity

Hysteresis plots of base shear versus controlled overall drift ratio and corresponding story shear versus story drift ratio with respect to the first and second stories for each specimen are shown in Figs. 16–21. Key observations in the previous section can be summarized here as local buckling of column bases (LB), partial fracture of the connection (CF) and beam twist (BT), the onset of which is marked in the corresponding hysteresis curves. Worth noting for specimen B345-C460-1 (see Fig. 17), although the overall drift ratio was controlled in symmetric predefined protocol, there is an obvious asymmetry in the hysteresis curve of each story, with the first story taking a higher percentage of the overall drift ratio during cycles in the negative direction (toward the west side) than when loaded in the positive direction, and vice versa for the second story. This asymmetry occurred at the 4.0% loading stage, when the lower level beam was found to exhibit significant twist associated with much out-of-plane deformation at positive loading (i.e., actuators pushed the specimen to the east). In this case horizontal displacements were different at the west and east edges of the specimen, which resulted in the difference of the measured story drift. For the other specimens lateral story displacements evolved more symmetrically. Based on the test observations and hysteresis curves, peak base shear generally occurred right after the onset of visible local buckling in column bases of all specimens. It was also recognized that the frame was able to deform plastically with substantial lateral stiffness and strength beyond the partial fracture of the connection, as shown in specimen B460-C460-1 (see Fig. 18).

The overall plastic displacement at the upper beam level in each half-cycle and the total cumulative one are shown in Fig. 23 for each specimen, where the total cumulative plastic displacement for all six specimens increased substantially after cycles at the 2% loading stage. As expected, those specimens using high strength steels (especially specimen B345-C890) have obviously smaller plastic displacement than the one using ordinary strength steels (specimen B345-C345) in the same half-cycle before column bases buckled severely. For example, at the first positive half-cycle of the 2% loading stage, specimens B345-C345, B345-C460-2 and B345C890 have 52.0 mm, 33.9 mm and 20.2 mm, respectively. Similarly, specimen B345-C345 has the largest cumulative plastic displacement (1102.5 mm), while specimens B345-C460-2 and B460C460-2 have 840.8 mm and 801.5 mm respectively, at the end of first complete cycle of the 4% loading stage. However, as for the deformation capacity at the end of test, specimen B345-C890 has the smallest cumulative plastic displacement (543.7 mm), while specimens B345-C460-2 and B460-C460-2 have the largest (more than 1500 mm) since even no critical failure or significant strength degradation was found until the test was stopped due to the limit of stroke of the actuators. Examining the energy dissipation in each half-cycle and the total cumulative one as shown in Fig. 24, note that the energy dissipation typically increases as the half-cycle number increases, but the energy dissipation per half-cycle drops after substantial strength degradation (see the last half-cycle for specimens B345-C460-1, B460-C460-1 and B345-C890). Similar results to

4.5. Specimen B460-C460-2 Similar to specimen B345-C460-2, no notable phenomenon was found until the 4.0% loading stage. The flanges in column bases exhibited local buckling at the third half-cycle of the 4.0% loading stage when the specimen moved toward the east. Although additional six cycles were imposed after the second complete cycle at the 4.0% loading stage, only slight deterioration in the lateral strength was observed and the test was stopped manually. Evolution of buckling deformation in column bases are shown in Fig. 14. All the beam-to-column connections were nearly undamaged. 4.6. Specimen B345-C890 No special observations were found anywhere until the 1.5% loading stage, when local buckling in flanges and webs of both column bases began to develop at the first half-cycle toward the positive peak. When the second positive peak top displacement at the 3.0% loading stage was reached, cracks initiated between column flanges and the web at the east-side column base due to rather severe local buckling, and similar to specimen B345-C345, instantaneous axial squash at both column bases was noticed which resulted in significant vertical shortening. The test was stopped when unloaded with the top displacement toward zero. Evolution of buckling deformation in column bases are shown in Fig. 15. All the beam-to-column connections remained undamaged.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

10

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

Fig. 13. Buckling deformation in the east-side column base at the end of the second cycle of (a) 2.0% (1.8% at the first story) and (b) 3.0% loading stages (2.8% at the first story), (c) the first, (d) second and (e) third cycle of 4.0% loading stage (3.7% at the first story), and (f) in the west-side column base after test.

Fig. 14. Buckling deformation in the east-side column base at the end of (a) the second cycle of 3.0% loading stage (2.8% at the first story), (b) the second, (c) fourth, (d) sixth and (e) eighth cycle of 4.0% loading stage (3.8% at the first story), and (f) in the west-side column base after test.

crack initiation

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 15. Buckling deformation in the east-side column base at the end of (a) the second cycle of 1.5% loading stage (1.3% at the first story), (b) the first and (c) second cycle of 2.0% loading stage (1.9% at the first story), (d) the first and (e) second cycle of 3.0% loading stage (3.6% at the first story), and (f) in the west-side column base after test.

0 -250 -500 -6

-4

-2

0

2

4

6

Story shear (kN)

250

2nd story

1st story

LB Test stop

Story shear (kN)

Base shear (kN)

500

-6

-4

-2

0

2

4

6

-6

-4

-2

0

2

Overall drift ratio (%)

Story drift ratio (%)

Story drift ratio (%)

(a)

(b)

(c)

4

6

Fig. 16. Hysteresis curves of B345-C345: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

11

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

0 -250 -500 -6

-4

-2

0

2

4

6

Story shear (kN)

250

2nd story

1st story

LB BT Test stop

Story shear (kN)

Base shear (kN)

500

-6

-4

-2

0

2

4

6

-6

-4

-2

0

2

Overall drift ratio (%)

Story drift ratio (%)

Story drift ratio (%)

(a)

(b)

(c)

4

6

Fig. 17. Hysteresis curves of B345-C460-1: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

0 -250 -500 -6

-4

Story shear (kN)

250

2nd story

1st story

LB CF Test stop

Story shear (kN)

Base shear (kN)

500

-2

0

2

4

6

-6

Overall drift ratio (%)

-4

-2

0

2

4

6

-6

Story drift ratio (%)

(a)

-4

-2

0

2

4

6

Story drift ratio (%)

(b)

(c)

Fig. 18. Hysteresis curves of B460-C460-1: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

0 -250 -500 -6

-4

-2

2nd story

0

2

4

6

Story shear (kN)

250

1st story

LB Test stop

Story shear (kN)

Base shear (kN)

500

-6

-4

-2

Overall drift ratio (%)

0

2

4

6

-6

-4

Story drift ratio (%)

-2

(b)

(a)

0

2

4

6

Story drift ratio (%)

(c)

Fig. 19. Hysteresis curves of B345-C460-2: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

0 -250 -500 -6

-4

-2

0

2

4

6

Story shear (kN)

250

2nd story

1st story

LB Test stop

Story shear (kN)

Base shear (kN)

500

-6

-4

-2

0

2

4

6

-6

-4

-2

0

2

Overall drift ratio (%)

Story drift ratio (%)

Story drift ratio (%)

(a)

(b)

(c)

4

6

Fig. 20. Hysteresis curves of B460-C460-2: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

the deformation capacity are found by comparing the total cumulative energy dissipation for those specimens, except that specimen B460-C460-1 dissipated 34% more cumulative energy than specimen B345-C460-1 as shown in Fig. 24(b) and (c), although a similar cumulative plastic displacement capacity (difference is within 15%) was observed between them in Fig. 23 (b) and (c). This can be explained by the much less dissipative

failure mode (i.e. twist of the lower beam) in specimen B345-C460-1. Again, specimen B460-C460-2 has the largest capacity in cumulative energy dissipation at the end of test (more than 600 kJ), but followed by specimen B460-C460-1. A key aspect of behavior of interest to engineers and researchers is whether soft-story mechanisms form in the specimens as inelastic displacement amplitudes increase. The ratios of story

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

12

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

0 -250 -500 -6

-4

2nd story

-2

0

2

4

6

-6

Story shear (kN)

250

1st story

LB Test stop

Story shear (kN)

Base shear (kN)

500

-4

-2

0

2

4

6

-6

-4

-2

0

2

Overall drift ratio (%)

Story drift ratio (%)

Story drift ratio (%)

(a)

(b)

(c)

4

6

Fig. 21. Hysteresis curves of B345-C890: (a) base shear versus overall drift ratio, (b) the first and (c) second story shear versus story drift ratio.

Base shear (kN)

500 250

500 First cycle Last cycle

250

0

500 First cycle Last cycle

0

0

Ke=4.92kN/mm

Ke=4.07kN/mm

-250 -500 -6

-2

0

2

4

6

-500 -6

Overall drift ratio (%)

-250 -4

-2

(a) Base shear (kN)

2

4

6

250

-250 2

4

6

-500 -6

0

2

4

6

(c)

First cycle Last cycle

First cycle Last cycle

250 0

Ke=2.51kN/mm

-250 0

-2

500

0

Ke=2.67kN/mm

-2

-4

(b)

First cycle Last cycle

-4

-500 -6

Overall drift ratio (%)

500

0

-500 -6

0

Overall drift ratio (%)

500 250

Ke=3.74kN/mm

-250 -4

First cycle Last cycle

250

Ke=2.74kN/mm

-250 -4

-2

0

2

4

6

-500 -6

-4

-2

0

2

4

Overall drift ratio (%)

Overall drift ratio (%)

Overall drift ratio (%)

(d)

(e)

(f)

6

200

Overall plastic disp. Cum. plastic disp.

Overall plastic disp. Cum. plastic disp.

Overall plastic disp. Cum. plastic disp.

1500 1200

150

900

100

600

50

300

0

250 200

Half-cycle number

Half-cycle number

Half-cycle number

(a)

(b)

(c)

Overall plastic disp. Cum. plastic disp.

Overall plastic disp. Cum. plastic disp.

Overall plastic disp. Cum. plastic disp.

0

1500 1200

150

900

100

600

50

300

0

Half-cycle number

Half-cycle number

Half-cycle number

(d)

(e)

(f)

Cum. plastic disp. (mm)

250

Cum. plastic disp. (mm)

Overall plastic disp. (mm)

Overall plastic disp. (mm)

Fig. 22. Backbone curves of specimens: (a) B345-C345; (b) B345-C460-1; (c) B460-C460-1; (d) B345-C460-2; (e) B460-C460-2 and (f) B345-C890.

0

Fig. 23. Overall plastic displacement in each half-cycle and total cumulative overall plastic displacement of specimens: (a) B345-C345; (b) B345-C460-1; (c) B460-C460-1; (d) B345-C460-2; (e) B460-C460-2 and (f) B345-C890.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

13

60

Energy dissipation Cum. energy dissipation

Energy dissipation Cum. energy dissipation

600 450

40

300

20

150

0

Energy dissipation (kJ)

Energy dissipation Cum. energy dissipation

Half-cycle number

Half-cycle number

Half-cycle number

(a)

(b)

(c) 600

80 60

0

Energy dissipation Cum. energy dissipation

Energy dissipation Cum. energy dissipation

Energy dissipation Cum. energy dissipation

450

40

300

20

150

0

Half-cycle number

Half-cycle number

Half-cycle number

(d)

(e)

(f)

0

Cum. energy dissipation (kJ)

80

Cum. energy dissipation (kJ)

Energy dissipation (kJ)

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

Fig. 24. Energy dissipation in each half-cycle and total cumulative energy dissipation of specimens: (a) B345-C345; (b) B345-C460-1; (c) B460-C460-1; (d) B345-C460-2; (e) B460-C460-2 and (f) B345-C890.

Half-cycle number

(a) 100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

(c) 100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

(e)

Story deformation ratio (%)

2nd story 1st story

100 90 80 70 60 50 40 30 20 10 0

Story deformation ratio (%)

100 90 80 70 60 50 40 30 20 10 0

according to the loading protocol, but this difference was insignificant). In last few half-cycles, however, an increase to 50–60% was observed. This change in behavior occurred when column bases developed significant plastic buckling. Note that there is an obvious asymmetry for specimen B345-C460-1 (see Fig. 25(b)), with the first story contributing less than 40% in half-cycles with positive displacement and more than 40% for

100 90 80 70 60 50 40 30 20 10 0

Story deformation ratio (%)

Story deformation ratio (%)

Story deformation ratio (%)

Story deformation ratio (%)

deformation at each story to the top displacement are shown in Fig. 25 for each half-cycle. For all six specimens, story deformation ratios for the first story are essentially close to 40% throughout the test. Such a result can be explained by the larger lateral stiffness exhibited by the first story where column bases were fixed, although its story shear was slightly greater than that in the second story (i.e., the ratio of shear forces was always kept as 1.05

100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

(b)

2nd story 1st story Half-cycle number

(d)

2nd story 1st story Half-cycle number

(f)

Fig. 25. Story deformation ratio of specimens: (a) B345-C345; (b) B345-C460-1; (c) B460-C460-1; (d) B345-C460-2; (e) B460-C460-2 and (f) B345-C890.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

Energy dissipation ratio (%)

In global aspect, a direct comparison of test results can be made with the seismic design procedure employed initially to determine the reduced seismic force for finalizing cross-sections of individual structural members. As introduced in Section 2.2, a tentative value of 0.4 for structural characteristic factor (or a reduction factor of 2.5) has been assumed for design of all the specimens in consideration of nonlinear structural response and energy dissipation capacity. After the test the obtained backbone curves in Fig. 22 were used to evaluate the actual reduction factor and other structural properties. In order to facilitate an easy calibration, general

100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

Energy dissipation ratio (%)

100 90 80 70 60 50 40 30 20 10 0

Energy dissipation ratio (%)

(a)

100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

(c)

2nd story 1st story Half-cycle number

(e)

Energy dissipation ratio (%)

6. Practical implications

nonlinear backbone curves were idealized as elastic-perfectly plastic, as shown in Fig. 27, where the ordinate represents base shear, while the abscissa is the top displacement at the upper beam level. The idealized elastic stiffness (Ke) is based on the initial elastic slope shown in Fig. 22 and the individual data point at each loading stage corresponds to the average between the first and last cycle in both positive and negative peaks in Fig. 22. The idealized yield base shear was determined based on an equal energy absorption, while the ultimate limit state was characterized by 15% drop in base shear from its peak. The displacement ductility is then defined as l = Du/Dy, where Du and Dy represent the ultimate (or collapse) and yield displacement, respectively, for the idealized behavior. The overstrength factor is s = Vy/Vd, where Vy and Vd represent the idealized yield and design base shear, respectively. Using the equal displacement rule proposed by Newmark and Hall [68], the actual reduction factor is calculated as R = l
s. The results are reported in Table 3 for all six specimens. For specimens B460-C460-1, B345-C460-2 and B460-C460-2, because after several cycles at the 4% loading stage where the actuator at the upper beam level reached almost its maximum stroke, the measured base shear hasn’t decreased from its peak value by 15%, Table 3 can only provides conservative results based on an ultimate overall drift ratio of 4.0%. That’s why ‘‘>” is used to indicate corresponding results. Note that the reduction factor used in design is based on design earthquake, while it is clear that the definition in Fig. 27 corresponds to the ultimate or collapse limit state under maximum considered earthquake, which is about two times as large as design earthquake. Thus, R-factors obtained in Table 3 are compared with a design value of 2.5
2 = 5. It is evident that all the specimens have much larger reduction factors than expected in design, which indicates a safe initial design. This can be explained by the significant overstrength as shown in Table 3, which in turn may result from the substantially higher measured

100 90 80 70 60 50 40 30 20 10 0

Energy dissipation ratio (%)

negative displacement. It is believed that in that test unsymmetrical displacement was caused by out-of-plane deformation and change of length of the lower twisted beam under compression in positive half-cycles. Ratios of energy dissipation at each story to the overall energy dissipation are shown in Fig. 26 for each half-cycle, where the dissipated energy was calculated as the area enclosed by the hysteresis curve shown in Figs. 16–21. It indicates that for all six specimens, around 50% of total energy was dissipated in each story; however, in last several half-cycles a tendency to dissipate more in the first story (up to 70%) of specimen B345-C890 (see Fig. 26(f)) was resulted from local plastic buckling of column bases while the second story remained essentially elastic by using very high strength steels in columns. Obvious fluctuations can be observed for specimen B345-C345 (see Fig. 26(a)) in first few half-cycles, and the contribution of the first story to the overall energy dissipation can be as much as 90%. This result deserves no special attention since the whole specimen remained basically elastic and the friction between lateral restraints and the specimen led to additional energy dissipation.

100 90 80 70 60 50 40 30 20 10 0

Energy dissipation ratio (%)

14

100 90 80 70 60 50 40 30 20 10 0

2nd story 1st story Half-cycle number

(b)

2nd story 1st story Half-cycle number

(d)

2nd story 1st story Half-cycle number

(f)

Fig. 26. Energy dissipation ratio of specimens: (a) B345-C345; (b) B345-C460-1; (c) B460-C460-1; (d) B345-C460-2; (e) B460-C460-2 and (f) B345-C890.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

15

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

Vu

Base shear (kN)

Vy

Note that because only one specimen using Q890 high strength steels whose ultimate strain was as low as 5.3% (see Table 2) was tested in this paper, it is of course not confident to conclude the definite safety to apply such ultra-high strength steels in various engineering structures. The pioneering experimental study here, however, still reveals some potential advantages and seismic capacity of those ultra-high strength steel frames. It should also be kept in mind that the above discussion is only based on the frame behavior under quasi-static loading conditions, but there is always a gap between the cyclic test behavior and the real one subject to earthquake ground motions. As such, further work, especially nonlinear time history analyses, are required to comprehensively evaluate the effect of using high strength steels, not only on structural behavior, but also on economic benefit.

15%Vu

A2 A1

A3 Equal areas A1+A3=A2

Vd Ke y

u

Top displacement (mm) Fig. 27. Idealization of the backbone curve.

7. Conclusions steel strength than the code-specified nominal one. The redistribution of internal forces in inelastic range, minimum serviceability requirements (deflection or drift) and multiple load combinations should also contribute to such overstrength. Another interesting result is, although high strength steels are believed to have poorer ductility than ordinary strength steels (see Table 2), the global ductility exhibited by frame systems using high strength steels is not necessarily poorer (see ductility factors in Table 3). The member ductility which is associated with crosssection slenderness is a more rigorous representation to affect global behavior [69]. The deformation capacity is also correlated with the global stability related to second-order effect under seismic weight, and the local stability in column bases subjected to high axial forces, both of which can lead to a substantial loss of strength. Overall drift ratio corresponding to the ultimate limit state is calculated as hu = Du/H (H is 5400 mm) and summarized in Table 3. Since the discussion in the previous section has revealed that the lateral top displacement was well distributed among each story, this ultimate overall drift ratio is believed to conservatively represent the capacity of story drift ratio that could be developed by the test specimens. Note that in Chinese seismic code [58] a maximum demand of story drift ratio being 2.0% is allowed under MCE, while in US seismic provisions [67] the capacity of story drift ratio is required to be at least 2.0% or 4.0% for intermediate or special moment frames (IMFs or SMFs). Therefore, the ordinary strength steel specimen B345-C345 satisfies the Chinese code and US seismic provisions for IMFs, while specimens B345-C460-1, B460C460-1, B345-C460-2 and B460-C460-2 succeed to be well qualified for SMFs in the US. Even the specimen B345-C890 whose columns were made of ultra-high strength steel slender sections, conform to drift requirements in Chinese seismic code or IMFs in the US. In addition, the elastic story drift ratio before the first significant yield is believed to be around 1.0% in practice [67]. The equivalent yield overall drift ratio calculated as hy = Dy/H and shown in Table 3 for each specimen demonstrates that such empirical assumption is well representative of ordinary strength steel frames, but for those using high strength steels, this value is about 1.4% and can even reach 2.0%.

Since relatively few studies of complete high strength steel frame systems have been conducted, this study focuses on determining through experimental means a better understanding of seismic behavior of high strength steel frame systems, and using that understanding to improve current seismic design based on ordinary strength steels. A total of five full-scale, single-bay, twostory high strength steel frame specimens and one ordinary strength steel frame specimen for comparison have been tested quasi-statically to obtain experimental data on their seismic behavior. Several conclusions with respect to their system global response and some recommendations for future research are summarized as follows: (1) The frame specimens using ordinary strength steels (specimen B345-C345) and Q460 high strength steels in either columns only or beams and columns (specimens B345-C460-1, B460-C460-1, B345-C460-2 and B460-C460-2) were capable of accommodating an overall drift ratio of 4.0% (the maximum story drift ratio of about 4.6%, 4.4%, 4.1%, 4.3%, 4.2%, respectively) without collapse due to global lateral instability or local complete fracture under cyclic loading conditions. Both scenarios with compact or noncompact columns were examined to clearly find that more significant local buckling and thus, more severe damage occurred at column bases of the latter. The two specimens using compact high strength steel columns (specimens B345-C460-2 and B460-C460-2) had the largest deformation capacity and the smallest peak strength degradation among all the specimens; while the two specimens using high strength steels for both beams and columns (specimens B460-C4601 and B460-C460-2) exhibited the largest energy dissipation capacity. (2) Twist and notable out-of-plane deformation of beams were observed in specimen B345-C460-1 due to the absence of lateral restraint. This indicated that the lateral-torsional buckling issue was more critical by using high strength steels to reduce cross-sectional size. The test results evidenced that the deformation capacity might not be affected

Table 3 Structural properties of tested specimens. Specimen no.

Vd (kN)

Vy (kN)

Dy (mm)

Du (mm)

l

s

R

hy (rad)

hu (rad)

B345-C345 B345-C460-1 B460-C460-1 B345-C460-2 B460-C460-2 B345-C890

60.5 55.4 53.9 53.6 53.7 53.4

282.3 324.5 389.7 203.3 194.6 192.9

57.4 79.7 104.2 76.2 77.5 70.4

132.5 213.1 >212.4 >215.4 >215.7 141.2

2.31 2.67 >2.04 >2.83 >2.78 2.01

4.67 5.86 7.23 3.79 3.62 3.61

10.8 15.7 >14.7 >10.7 >10.1 7.2

1.1% 1.5% 1.9% 1.4% 1.4% 1.4%

2.5% 4.0% >4.0% >4.0% >4.0% 2.6%

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

16

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx

(3)

(4)

(5)

(6)

so much by lateral-torsional buckling; however, this led to obvious lower energy dissipation capacity as seen from less plump hysteresis cycles. The CJP welds in beam-to-column connections exhibited satisfactory behavior in all the specimens throughout the test, which benefited from reinforcing cover plates to reduce local stress level. Thus, good performance was developed by those cover-plate reinforced connections, as expected. Although fracture occurred thoroughly in continuity plates and part of the column web in a first-story beam-tocolumn connection in specimen B460-C460-1, it did not necessarily result in complete failure of the whole frame since a considerable residual strength was developed. The slender column base cross-section in specimen B345C890 buckled severely at an overall drift ratio up to 3.0% (the first story drift ratio is about 3.6%), which caused about 50% drop in the lateral strength. In all the specimens, the first story generally contributed about 40% of top displacement in elastic stage which increased to 50–60% after plastic hinges and local buckling developed at column bases, and about 50% of total energy dissipation throughout the test. No concentration of deformation or soft-story mechanism was found. Seismic force reduction factors evaluated for all the specimens were much larger than the design value, and this can be explained by significant overstrength including substantially higher measured steel strength than code-specified. The ultimate capacity of overall drift ratio assessed based on an equal energy idealization demonstrated satisfied highly ductile behavior (not smaller than 4.0%) for specimens B345-C460-1, B460-C460-1, B345-C460-2 and B460C460-2, and moderately ductile behavior (larger than 2.0%) for specimens B345-C345 and B345-C890. Further numerical analyses are needed to fully characterize reliable performance criteria and safe reduction factors, and identify the effects of high strength steels and cross-section slenderness.

Acknowledgements The authors acknowledge the financial support for this work from the Natural Science Foundation of China (Grant No. 51478244), and the Excellent Young Scientist Programme of the National Natural Science Foundation of China (No. 51522806). References [1] Raoul J, Günther HP. Use and application of high-performance steels for steel structures. Zürich, Switzerland: IABSE-AIPC-IVBH; 2005. [2] Bjorhovde R. Development and use of high performance steel. J Constr Steel Res 2004;60(3):393–400. [3] Bjorhovde R. Performance and design issues for high strength steel in structures. Adv Struct Eng 2010;13(3):403–11. [4] Haaijer G. Economy of high strength steel structural members. J Struct Div (ASCE) 1961;87(8):1–23. [5] Shi G, Hu FX, Shi YJ. Recent research advances of high strength steel structures and codification of design specification in China. Int J Steel Struct 2014;14 (4):873–87. [6] Shokouhian M, Shi YJ. Flexural strength of hybrid steel I-beams based on slenderness. Eng Struct 2015;93:114–28. [7] Ma JL, Chan TM, Young Ben. Experimental investigation of cold-formed high strength steel tubular beams. Eng Struct 2016;126:200–9. [8] Ricles JM, Sause R, Green PS. High-strength steel: implications of material and geometric characteristics on inelastic flexural behavior. Eng Struct 1998;20 (4):323–35. [9] McDermott JF. Plastic bending of A514 steel beams. J Struct Div (ASCE) 1969;95(9):1851–71. [10] Shokouhian M, Shi YJ, Head M. Interactive buckling failure modes of hybrid steel flexural members. Eng Struct 2016;125:153–66. [11] Usami T, Fukumoto Y. Local and overall buckling of welded box columns. J Struct Div (ASCE) 1982;108(3):525–42.

[12] Rasmussen KJR, Hancock GJ. Tests of high strength steel columns. J Constr Steel Res 1995;34(1):27–52. [13] Shi G, Ban HY, Bijlaard FSK. Tests and numerical study of ultra-high strength steel columns with end restraints. J Constr Steel Res 2012;70:236–47. [14] Ban HY, Shi G, Shi YJ, Wang YQ. Overall buckling behavior of 460MPa high strength steel columns: experimental investigation and design method. J Constr Steel Res 2012;74:140–50. [15] Ban HY, Shi G, Shi YJ, Bradford MA. Experimental investigation of the overall buckling behavior of 960 MPa high strength steel columns. J Constr Steel Res 2013;88:256–66. [16] Wang YB, Li GQ, Chen SW, Sun FF. Experimental and numerical study on the behavior of axially compressed high strength steel columns with H-section. Eng Struct 2012;43:149–59. [17] Wang YB, Li GQ, Chen SW, Sun FF. Experimental and numerical study on the behavior of axially compressed high strength steel box-columns. Eng Struct 2014;58:79–91. [18] Shi G, Zhou WJ, Lin CC. Experimental investigation on the local buckling behavior of 960 MPa high strength steel welded section stub columns. Adv Struct Eng 2015;18(3):423–37. [19] Shi G, Zhou WJ, Bai Y, Lin CC. Local buckling of 460 MPa high strength steel welded section stub columns under axial compression. J Constr Steel Res 2014;100:60–70. [20] Shi G, Jiang X, Zhou WJ, Chan TM, Zhang Y. Experimental study on column buckling of 420 MPa high strength steel welded circular tubes. J Constr Steel Res 2014;100:71–81. [21] Girão Coelho AM, Bijlaard FSK, Kolstein H. Experimental behaviour of highstrength steel web shear panels. Eng Struct 2009;31:1543–55. [22] Jordão S, Simões da Silva L, Simões R. Design formulation analysis for high strength steel welded beam-to-column joints. Eng Struct 2014;70: 63–81. [23] Li TJ, Liu SW, Chan SL. Direct analysis for high-strength steel frames with explicit-model of residual stresses. Eng Struct 2015;100:342–55. [24] Huang YH, Wada A, Iwata M, Watanabe A. Inelastic behavior of high strength steels under cyclic gradient stress. In: Proceedings of the 4th pacific structural steel conference. p. 367–74. Singapore. [25] Huang YH, Onishi Y, Hayashi K. Inelastic behavior of high strength steels with weld connections under cyclic gradient stress. In Proceedings of the 11th world conference on earthquake engineering, Acapulco, Mexico; 1996; Paper No. 1745. [26] Shi G, Wang M, Wang YQ, Wang F. Cyclic behavior of 460 MPa high strength structural steel and welded connection under earthquake loading. Adv Struct Eng 2013;16(3):451–66. [27] Shi G, Wang M, Bai Y, Wang F, Shi YJ, Wang YQ. Experimental and modeling study of high-strength structural steel under cyclic loading. Eng Struct 2012;37(4):1–13. [28] Wang YB, Li GQ, Cui W, Chen SW, Sun FF. Experimental investigation and modeling of cyclic behavior of high strength steel. J Constr Steel Res 2015;104:37–48. [29] Dusicka P, Itani AM, Buckle IG. Cyclic response of plate steels under large inelastic strains. J Constr Steel Res 2007;63(2):156–64. [30] Miyazaki Y, Nara S, Ono K, Miyoshi T, Mori S. Characteristics of plastic hysteresis of high performance steels and the constitutive equations. J Struct Eng (AIJ) 2007;53A:25–32 [in Japanese]. [31] Green PS. The inelastic behavior of flexural members fabricated from high performance steel Doctoral Thesis. Bethlehem, Pennsylvania, US: Lehigh University; 2000. [32] Suzuki T, Ogawa T, Ikarashi K, Obata M. Effects of the yield ratio on the behavior and energy absorbing capacity of high strength steel beams under the repeated load. J Struct Constr Eng (AIJ) 1996;485:135–44 [in Japanese]. [33] Dusicka P, Itani AM, Buckle IG. Cyclic behavior of shear links of various grades of plate steel. J Struct Eng (ASCE) 2010;136(4):370–8. [34] Fukumoto Y, Kusama H. Local instability tests of plate elements under cyclic uniaxial loading. J Struct Eng (ASCE) 1985;111(5):1051–67. [35] Kuwamura H, Kato B. Inelastic behavior of high strength steel members with low yield ratio. In: Proceedings of 2nd Pacific structural steel conference, Victoria, Australia. p. 429–37. [36] Wang JJ, Shi G, Shi YJ. Experimental research on behavior of 460 MPa high strength steel I-section columns under cyclic loading. Earthq Eng Eng Vibr 2014;13(4):611–22. [37] Shi G, Wang JJ, Bai Y, Shi YJ. Experimental study on seismic behavior of 460 MPa high strength steel box-section columns. Adv Struct Eng 2014;17 (7):1045–60. [38] Wang YB, Li GQ, Cui W, Chen SW. Seismic behavior of high strength steel welded beam-column members. J Constr Steel Res 2014;102:245–55. [39] Chen SW, Lu ZL, Li GQ, Wei M, Wang YB, Chen X. Cyclic loading tests of Q690D high strength steel welded columns. J Build Struct 2014;35(12):97–103 [in Chinese]. [40] Lin XC, Okazaki T, Chung YL, Nakashima M. Flexural performance of bolted built-up columns constructed of H-SA700 steel. J Constr Steel Res 2013;82:48–58. [41] Lin XC, Okazaki T, Nakashima M. Bolted beam-to-column connections for built-up columns constructed of H-SA700 steel. J Constr Steel Res 2014;101:469–81. [42] Sato A, Kimura K, Suita K, Inoue K. Development of weld-free beam-to-column connection of H-SA700A high strength steel for building structures. J Struct Constr Eng (AIJ) 2009;74(646):2355–63 [in Japanese].

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013

F. Hu et al. / Engineering Structures xxx (2016) xxx–xxx [43] Sun FF, Sun M, Li GQ, Xiao Y, Wei M, Liu LX. Experimental study on seismic behavior of high-strength steel beam-to-column end-plate connections. J Build Struct 2014;35(4):116–24 [in Chinese]. [44] Matsui C, Mitani I. Inelastic behavior of high strength steel frames subjected to constant vertical and alternating horizontal loads. Proceedings of 6th world conference on earthquake engineering, New Delhi, India, vol. 3. p. 3169–74. [45] Tenchini A, D’Aniello M, Rebelo C, Landolfo R, Simões da Silva L, Lima L. Seismic performance of dual-steel moment resisting frames. J Constr Steel Res 2014;101:437–54. [46] Tenchini A, D’Aniello M, Rebelo C, Landolfo R, Simões da Silva L, Lima L. High strength steel in chevron concentrically braced frames designed according to Eurocode 8. Eng Struct 2016;124:167–85. [47] Dubina D, Stratan A, Dinu F. Dual high-strength steel eccentrically braced frames with removable links. Earthquake Eng Struct Dynam 2008;37 (15):1703–20. [48] Dubina D, Vulcu C, Stratan A, Ciutina A, Grecea D, Ioan A, et al. High strength steel in seismic resistant building frames. Final Report HSS-SERF, DirectorateGeneral for Research and Innovation, European Commission; 2015. [49] Nakai M, Nakamura Y, Maeda S, Tanaka T, Asai H, Suzuki Y. Proposal for damage-free design method of steel structure utilizing high strength steel under great earthquake. J Struct Constr Eng (AIJ) 2011;76(666):1443–51 [in Japanese]. [50] Shinsai N, Suita K, Koetaka Y. Mechanical performance and seismic response of knee brace damper structure of H-SA700 high strength steel members made by undermatched welds. J Struct Constr Eng (AIJ) 2012;77(682):1959–68 (in Japanese). [51] Nakai M, Tsuda K, Mase S, Narihara H, Okayasu T, Asai H, et al. Performance verification through full-scale static loading tests for a structural system using high strength steel. J Struct Constr Eng (AIJ) 2013;78(687):1007–16 [in Japanese]. [52] Takeuchi T, Ohyama T, Ishihara T. Seismic performance of high-strength steel frames with energy dissipation braces-Part 1: Cumulative cyclic deformation capacity of high-strength steel frames with energy dissipation braces. J Struct Constr Eng (AIJ) 2010;75(655):1671–9 [in Japanese]. [53] Takeuchi T, Ohyama T, Matsui R. Seismic performance of high-strength steel frames with energy dissipation braces-Part 2: Cumulative deformation capacity evaluation of high-strength steel beam ends subjected to cyclic bending moment. J Struct Constr Eng (AIJ) 2011;76(661):695–702 [in Japanese].

17

[54] Longo A, Montuori R, Nastri E, Piluso V. On the use of HSS in seismic-resistant structures. J Constr Steel Res 2014;103:1–12. [55] Eurocode 3. Design of steel structures-part 1–12: Additional rules for the extension of EN 1993 up to steel grades S700. EN 1993-1-122007. Brussels: European Committee for Standardization; 2007. [56] Eurocode 3. Design of steel structures-part 1–1: General rules and rules for buildings. EN 1993-1-1-2005. Brussels: European Committee for Standardization; 2005. [57] GB 50017-201X. Code for design of steel structures. Beijing, China; 2013 [draft for approval in Chinese]. [58] GB 50011-2010. Code for seismic design of buildings. Beijing, China: China Architecture & Building Press; 2010 [in Chinese]. [59] AIJ. Guidelines for limit state design of steel structures. Tokyo, Japan: The Architectural Institute of Japan; 1998 [in Japanese]. [60] Eurocode 8. Design of structures for earthquake resistance-part 1: general rules, seismic actions and rules for buildings. EN1998-12004. Brussels: European Committee for Standardization; 2004. [61] ASCE/SEI 7-10. Minimum design loads for buildings and other structures. Reston, Virginia, US: American Society of Civil Engineers; 2010. [62] Kim T, Whittaker AS, Gilani ASJ, Bertero VV, Takhirov SM. Cover-plate and flange-plate reinforced steel moment-resisting connections PEER Report 2000/ 07. Berkeley, California, US: Pacific Earthquake Engineering Research Center, University of California; 2000. [63] ANSI/AISC 358-10. Prequalified connections for special and intermediate steel moment frames for seismic applications. Chicago, Illinois, US: American Institute of Steel Construction; 2014. [64] FEMA-355D. State of the art report on connection performance. SAC Joint Venture; 2000. [65] GB/T 228.1-2010. Metallic materials-Tensile testing-Part 1: Method of test at room temperature. Beijing, China: China Standards Press; 2011 [in Chinese]. [66] Pan P, Zhao G, Lu XZ, Deng KL. Force-displacement mixed control for collapse tests of multistory buildings using quasi-static loading systems. Earthquake Eng Struct Dynam 2014;43(2):287–300. [67] ANSI/AISC 341-10. Seismic provisions for structural steel buildings. Chicago, Illinois, US: American Institute of Steel Construction; 2010. [68] Newmark NM, Hall WJ. Earthquake spectra and design. Berkeley, California, US: Earthquake Engineering Research Institute; 1982. [69] Gioncu V, Mazzolani FM. Ductility of seismic resistant steel structures. London, UK: Spon Press, Taylor & Francis Group; 2002.

Please cite this article in press as: Hu F et al. Experimental study on seismic behavior of high strength steel frames: Global response. Eng Struct (2016), http://dx.doi.org/10.1016/j.engstruct.2016.11.013