Unstiffened Steel Plate Shear Wall Performance under Cyclic Loading

453 I o r l o o t im a d r s e s li and by in-plane shear resistance of the infill panels, anchored betwe ened load. postb with from the re can d exhib Re 1Str Broad 2As Main 3As Main 4Re Main No 1, 200 be file paper 14, 19 Vol. 1 0460/$ pacity of the steel infill panels in designing the primary lateral eth- nfill sign im- pro- gle- ens rpo- han cu- em- cese ized ing 97) ame th a tory any nel- teel om- JOURNAL OF STRUCTURAL ENGINEERING / APRIL 2000 / en the frame members. If desired the panels can be stiff- or chosen thick enough to preclude shear buckling under It is much more efficient, however, to utilize the full uckling shear strength of the infill panel and dispense additional intermediate stiffeners. One problem arising the buckling and tension field action, however, would be duced capacity upon load reversal until a tension field evelop in the opposite direction. This phenomenon is ited through a pinched shape of the hysteresis curves. search programs worldwide have investigated various pa- uct. Designer, Read Jones Christoffersen Ltd., 3rd Floor, 1285 West way, Vancouver, BC, Canada V6H 3X8. soc. Prof., Dept. of Civ. Engrg., Univ. of British Columbia, 2324 Mall, Vancouver, BC, Canada V6T 1Z4. soc. Prof., Dept. of Civ. Engrg., Univ. of British Columbia, 2324 Mall, Vancouver, BC, Canada V6T 1Z4; corresponding author. s. Assoc., Dept. of Civ. Engrg., Univ. of British Columbia, 2324 Mall, Vancouver, BC, Canada V6T 1Z4. te. Associate Editor: Brad Cross. Discussion open until September 0. To extend the closing date one month, a written request must d with the ASCE Manager of Journals. The manuscript for this was submitted for review and possible publication on December 98. This paper is part of the Journal of Structural Engineering, 26, No. 4, April, 2000. qASCE, ISSN 0733-9445/00/0004-0453– 8.00 1 $.50 per page. Paper No. 19862. load resisting system for buildings. As a result, a design m odology for steel plate shear walls using thin unstiffened i panels was incorporated as an appendix to the Canadian de code for steel construction (‘‘Limit’’ 1994). It proposes a s plified strip model for use in conventional frame analysis grams. Previously published research typically focused on sin panel specimens, specimens of very small scale, or specim incorporating stiffened infill panels. Most specimens inco rated boundary frames with stiffness significantly higher t that of the infill panels. Limited experience has been do mented with medium- to large-scale steel shear wall ass blies incorporating narrow infill panels. Research by Cac et al. (1993) on quarter-scale, three-story specimens util panel width-to-height aspect ratios of 1.5 and lateral load at the roof level only. Driver et al.’s experimental work (19 with a 50% scaled four-story specimen, representing a fr of a typical commercial building, incorporated panels wi slightly wider aspect ratio of 1.7. In all these tests, the s shear mode dominated the behavior of the system. Little if experience has been documented to validate current pa shear-oriented analysis techniques against tall unstiffened s shear walls in which global flexural deformation modes d inate. UNSTIFFENED STEEL PLATE S UNDER CYCL By Adam S. Lubell,1 Helmut G. L. Prion and Mahm ABSTRACT: In the last few decades, steel plate shea resisting elements in several buildings around the wor performance of unstiffened thin steel plate shear walls f postbuckling strength of the panels is relied upon for m web of a plate girder. Experimental testing was conduc specimens, under cyclic quasi-static loading. Each spec and floor-to-floor dimensions of 900 mm, representing Identification of load-deformation characteristics and the primary objectives of the testing program. Good energy achieved. Primary inelastic damage modes included yiel in the single-panel tests and yielding of the columns fo compared with simplified tension field analytical model limit-states design of steel structures. The models, in gen of the specimens, with less satisfactory results for ela analytical results, the adequacy of existing design guide INTRODUCTION For the past few decades, experimental and analytical stud- ies have been conducted on the use of steel plate shear walls as primary lateral load resisting elements in buildings. These studies have identified unique performance characteristics, in- cluding high elastic stiffness properties, large displacement ductility capacities, and stable hysteresis behavior. Steel shear walls have therefore been proposed as viable structural alter- natives to resist lateral loads in medium- and high-rise steel construction, particularly in areas of high seismic risk. A steel shear wall frame consists of column and beam ele- ments augmented by steel infill shear panels, provided over the height of a framing bay. Its form is analogous to that of a plate girder, vertically cantilevered from its base, with the col- umns acting as flanges, the beams as stiffeners, and the infill panel as the plate girder web. When subjected to lateral load- ing in the plane of the wall, forces are resisted through the flexural and coupled axial response of the columns and beams HEAR WALL PERFORMANCE C LOADING ,2 Carlos E. Ventura,3 Member, ASCE, ud Rezai4 walls have been introduced as primary lateral load d. This paper presents research from a study on the r medium- and high-rise buildings. In this concept, the st of the frame shear resistance, similar to the slender ed on two single- and one four-story steel shear wall en consisted of a single bay with column-to-column quarter-scale model of a typical office building core. stresses induced in the structural components were the dissipation and displacement ductility capacities were ing of the infill plates combined with column yielding the multistory frame. The experimental results were , based on recommendations in the Canadian code for ral, provided good predictions of the postyield strength tic stiffness calculations. From the experimental and nes were assessed. rameters and construction details associated with steel shear wall construction. A limited amount of information is currently available regarding the postbuckling behavior of steel plate shear walls for the purpose of developing simple expressions for the analysis and design of this structural system. A number of static and quasi-static cyclic tests performed on large- and small-scale specimens, with accompanying analytical models, have been reported in Canada (Timler and Kulak 1983; Trom- posch and Kulak 1987; Driver et al. 1997), the United States (Caccese et al. 1993; Elgaaly and Liu 1997), Japan (Yamada 1992), and England (Roberts and Sabouri-Ghomi 1991). Xue and Lu (1994) reported on an extensive numerical study of a hypothetical 12-story steel building frame incorporating infill panels with various attachment details. All of these studies examined the behavior of steel plates throughout the entire range of loading, from elastic to plastic and from prebuckling to postbuckling stages. The results obtained from the studies unanimously support the rationale of using the postbuckling strength, tension field action, and stable energy dissipation ca- 454 / JOURNAL OF STRUCTURAL ENGINEERING / APRIL 2000 This paper presents the results of a study of the performance of single- and multistory unstiffened steel shear walls sub- jected to cyclic quasi-static loading. It is a subset of a larger collaborative study established in 1994, between researchers at the University of British Columbia, the University of Al- berta, and practicing engineers in industry from Canada and the United States. Studies included quasi-static testing, dy- namic shake table experimentation, numerical and analytical investigations, and a comparative design study to assess the economic feasibility of the system. The objective of the re- search program presented herein was to verify the structural response of the steel plate shear wall specimens under an in- dustry standard cyclic seismic loading regime and provide a benchmark for a subsequent shake table testing program (Re- zai 1999). The suitability of this structural system for areas of high seismic risk was to be verified, with current design guide- lines assessed on the basis of the obtained performance char- acteristics. In particular, the validity of applying the current simplified design methodology to multistory steel shear wall frames was to be assessed. A detailed overview of this collab- orative project and findings to date have been presented by Timler (1998). EXPERIMENTAL TEST PROGRAM Laboratory testing was conducted using two single-panel specimens (SPSW1 and SPSW2), and one four-story specimen (SPSW4). Each test was conducted under fully reversed cyclic quasi-static loading in both the elastic and inelastic response regions of the specimens. The overall strength, elastic post- buckling stiffness, formation of diagonal tension field action combined with diagonal compression buckling of the infill plates, stability of the panel hysteresis curves, effects of beam and column rigidities, and the interaction between the frame action and the shear panel behavior were the main issues in- vestigated during the quasi-static test program. Standardized test protocols (‘‘Guidelines’’ 1992) and per- formance evaluation methods were employed, to permit direct comparisons between the specimen configurations without al- tering the basic parameters of frame stiffness, infill panel prop- erties, and story aspect ratio. Specimens The test specimens represented 25% scale models of one bay of a steel-framed office building core. A column-to-col- umn centerline spacing of 900 mm was used for each speci- men, with unstiffened infill panels having width to height as- pect ratios of 1:1. This panel aspect ratio would be at the narrow end of typical building bay proportions, although it is within the practical range and is similar to ratios reported by other researchers. Each specimen consisted of S7538 col- umns, kept continuous through the height of the frame, and S7538 beams. To better anchor the internal panel forces, the second single-story specimen (SPSW2) incorporated an addi- tional S7538 top beam welded along adjoining flange times, whereas the four-story SPSW4 specimen utilized a deep stiff S200334 beam at its roof level. Full moment connections were provided at all beam-column joints by continuous fillet welds of the entire beam section to the column flanges. Full flange continuity stiffeners were used at all joints. The infill panels were constructed from 1.5 mm (16 gauge) hot-rolled steel plate. A schematic of the SPSW4 specimen is shown in Fig. 1. SPSW1 and SPSW2 are equivalent to the first story of the SPSW4 specimen, with the top beam variation for SPSW2 as noted above. Typical material yield strengths for the boundary frame (380 MPa) and infill panels (320 MPa) were determined through tension coupon testing of representative samples. These tests FIG. 1. Specimen SPSW4 Shown in Test Frame FIG. 2. Typical Stress-Strain Characteristics of: (a) Infill Panel; (b) Boundary Frame also verified that ‘‘hot-rolled’’ material was used, with the de- sired yield plateau and strain hardening properties represen- tative of full-scale construction (Fig. 2). No significant varia- tions in the plate material properties in the two orthogonal directions were observed. The specimen design was based on the multistory frame, where the geometry was largely governed by size limitations and height restrictions of the test facilities and by the maxi- mum base shear capacity of the shaking table actuator to be used for subsequent tests on a similar specimen. Loading Horizontal loading was applied by hydraulic actuators through welded tabs aligned with the plane of the column web, over the height of the beam column joints. In the case of the four-story SPSW4 specimen, equal horizontal loads were ap- plied at each floor. All testing was conducted in a self-reacting frame, significantly stiffer than the specimens. Gravity loading of the SPSW4 specimen was simulated by means of steel masses attached at each story level. The vertical load at each story was 13.5 kN, providing a total of 54.0 kN of added gravity load. The weights on the upper three stories FIG. 3. Load-Deformation Curve for SPSW1 were bolted to the specimen through the center of each beam- column joint. The gravity load on the first story was trans- ferred by channel sections welded to the outside of the column flanges at the joint locations. This was done to keep the shear panel of the first story free from view obstructions. No external gravity loading was applied to the SPSW1 and SPSW2 spec- imens. SPSW2 and SPSW4 were tested according to procedures as recommended in ATC-24 (‘‘Guidelines’’ 1992). This guideline provides for a standardized measure of the seismic perfor- mance of steel structure components when tested under cyclic quasi-static conditions. Testing was conducted in a force-con- trolled manner, until a displacement corresponding to a point of ‘‘significant’’ yielding dy was achieved. Loading was then conducted cyclically in a displacement-controlled manner, at multiples of the yield displacement.The longitudinal displace- ment of the first-floor beam was used as the control parameter for all tests. The SPSW1 specimen was cycled with one or two cycles at each load level, to a global yield corresponding to 180-kN base shear and about 9-mm story displacement. Cyclic loading in the postyield region was applied using one cycle per dis- placement increment, until 4 3 dy, followed by pushover load to test termination at 7 3 dy. Testing was terminated when a lateral bracing member exhibited distress due to out-of-plane restraint forces resulting from the large longitudinal motion. The SPSW2 specimen was subjected to three cycles at each load level, with a global yield corresponding to 190-kN base shear and about 6-mm story displacement. Failure occurred due to a column fracture at a displacement of 6 3 dy. SPSW4, under three cycles per load level, achieved a global yield at about 150-kN base shear and 9-mm deformation at the first floor. A maximum displacement of 1.5 3 dy was achieved prior to a global instability failure, propagated by yielding of the columns. Full details of the loading histories for each speci- men were reported by Lubell (1997). Discussion Well-defined elastoplastic load deformation envelopes were achieved for each test specimen. Hysteresis loops were S- shaped, stable, and full (Figs. 3–5). In the elastic region, a high initial stiffness with little energy dissipation was evident. In the postyield region, several well-defined segments of the load deformation curves represented the various stages of loading, unloading, and the reversal of the infill plate buckles. Within these phases, regions could be isolated corresponding to predominantly frame action and to combined frame and infill plate action. Increased energy dissipation was achieved with each dis- placement level increment in the postyield region. Some de- crease in energy dissipation between subsequent cycles at the same load level was noted, due to local damage. The principal sequence of significant inelastic action in the single-panel specimens consisted of yielding of the infill panel followed by yielding of the boundary frame. The columns of JOURNAL OF STRUCTURAL ENGINEERING / APRIL 2000 / 455 FIG. 6. Deformation and Yield Pattern of SPSW2 after 6 3 dy FIG. 5. Load-Deformation Curves for SPSW4: (a) First Floor; (b) Fourth Floor FIG. 4. Load-Deformation Curve for SPSW2 the multistory SPSW4 specimen yielded before significant in- elastic action occurred in the infill panel, resulting in a state of global instability. Inelastic response in the specimens re- sulted from infill panel yielding (SPSW1, SPSW2), plate tear- ing (SPSW2), localized weld fractures (SPSW1, SPSW2), and the formation of plastic hinges in the boundary frame (top beam—SPSW1; bottom of columns—SPSW1, SPSW2, SPSW4; top of columns—SPSW2). Significant inelastic shear deformations were also observed in the columns of SPSW2 near the top beam-column joints. Deformation of the boundary frame elements in each spec- imen resulted from local anchorage of internal panel forces and the global deflection characteristics of the specimen. Sig- nificant ‘‘pull-in’’ of the columns, caused by the tension field a as it was specimen formatio and botto ‘‘hourgla trates th SPSW2 minimize tension fi hed to amage ly alter ss effi- verify defor- d over equent bled a d flex- asured uators, ieving at the strated 1. In- hinge of the sulted, m and FIG. 8. Angle of Inclination—B FIG. 7. Column Profile for SPSW4 456 / JOURNAL OF STRUCTURAL ENGINEERING / APRIL 2000 se Shear Relationship for SPSW4 prompted termination of the test at 7 3 dy. Boundary frame strain recordings from the multistory spec- imen verified that little flexural action occurred in the first- floor interior beam of the steel shear wall, due to opposing panel forces generated on adjacent stories. Little flexural ac- tion was also recorded near a first-floor beam-column joint, in both the column and beam elements. This suggested that a localized area around each joint remained stiff, due in part to the full perimeter infill panel attachment, regardless of the de- gree of beam-column joint fixity. Uniaxial strain gauges were mounted in horizontal and ver- tical directions near some of the fifth-points on the infill panel of SPSW1. Strain recordings suggested that the angle of in- clination of the infill panel principal strains varied in some relationship with the applied base shear levels. Even though the flexible top beam, rotating under load, would contribute to this relationship, this relationship occurred at low load levels when beam rotation would be minimal, suggesting that the basic mechanisms involved in establishing load paths in steel shear walls are at least partially responsible for the varied an- gle. Strain rosettes were affixed to each side of the first- and second-story infill panels of the SPSW4 specimen, to deter- mine the in-plane principal strain characteristics during various transferred from the infill plates, was observed in all s. In the SPSW2 specimen, the inward column de- n resulted in the formation of plastic hinges at the top m of each column, with the specimen taking on an ss’’ shape at the conclusion of the test. Fig. 6 illus- e deformation and yielding patterns of specimen after the completion of three cycles at 4 3 dy. To the secondary deformation that would result from eld anchorage with the columns, Xue and Lu (1994) have recommended that the infill panels should be attac the girders only (and not the columns). Although the d to the columns would be reduced, this would significant the load path characteristics and possibly result in a le cient system. No experimental results are available to Xue and Lu’s numerical model of the resulting system To better quantify the intra- and interfloor column mations, the SPSW4 specimen was heavily instrumente the first story, with moderate instrumentation at subs floor levels. The global deflected column shape resem shear mode over the first story and a combined shear an ure mode for the upper stories. Fig. 7 shows the me deformations of the column opposite the hydraulic act at various base shear levels for the load level prior to ach significant yielding. The importance of having a flexurally stiff member top and bottom of the steel shear wall stack was demon by the relatively poor performance of specimen SPSW sufficient stiffness permitted the formation of a plastic in the horizontal member, preventing proper anchorage tension field. A decrease in out-of-plane stability re which increased the demand on the lateral support syste phases of the loading history. This instrumentation showed that signific