ctbuh.org/papers Title:

Transcription

1 ctbuh.org/aers Title: Authors: Subject: Keywords: Influence of Panel Zone Strength and Beam Web Connection Method on Seismic Performance of Reduced Beam Section Steel Moment Connections Cheol H. Lee, Associate Professor, Seoul National University Jin-Ho Kim, Graduate Student, Seoul National University Jae-Hoon Kim, Research Engineer, Seoul National University Sang-Woo Jeon, Research Engineer, Seoul National University Structural Engineering Seismic Steel Structure Publication Date: 2004 Original Publication: Paer Tye: CTBUH 2004 Seoul Conference 1. Book chater/part chater 2. Journal aer 3. Conference roceeding 4. Unublished conference aer 5. Magazine article 6. Unublished Council on Tall Buildings and Urban Habitat / Cheol H. Lee; Jin-Ho Kim; Jae-Hoon Kim; Sang-Woo Jeon

2 Influence of Panel Zone Strength and Beam Web Connection Method on Seismic Performance of Reduced Beam Section Steel Moment Connections Cheol-Ho Lee 1, Jae-Hoon Kim 2, Sang-Woo Jeon 3, Jin-Ho Kim 4 1 Associate Professor, Deartment of Architecture, Seoul National University 2 Graduate Student, Deartment of Architecture, Seoul National University 3 Research Engineer, Research Institute of Industrial Science and Technology 4 Research Engineer, Research Institute of Industrial Science and Technology Abstract This aer resents test results on eight reduced beam section (RBS) steel moment connections. The testing rogram addressed bolted versus welded web connection and anel zone (PZ) strength as key variables. Secimens with medium PZ strength were designed to romote energy dissiation from both PZ and RBS regions such that the requirement for exensive doubler lates could be reduced. Both strong and medium PZ secimens with a welded web connection were able to rovide satisfactory connection rotation caacity for secial moment-resisting frames. Secimens with a bolted web connection erformed oorly due to remature brittle fracture of the beam flange at the weld access hole. If fracture within the beam flange groove weld was avoided using quality welding, the fracture tended to move into the beam flange base metal at the weld access hole. The exerimental and analytical results of this study aeared to indicate that conventional sli-critical design method can not revent sliage of the web bolts. A criterion for a balanced PZ strength that imroves the lastic rotation caacity while reduces the amount of beam buckling are also roosed. Keywords: 1. Introduction As a resonse to the widesread damage in connections of steel moment-resisting frames that occurred during the 1994 Northridge, California and the 1995 Kobe, Jaan earthquakes, a number of imroved beam-to-column connection design strategies have been roosed. Of a variety of new designs, the reduced beam section (RBS) connection has been shown to exhibit satisfactory levels of ductility in numerous tests and has found broad accetance in a relatively short time (Chen 1996; Plumier 1997; Zekioglu et al. 1997; Engelhardt et al. 1998). In the RBS connection a ortion of the beam flanges at some distance from the column face is strategically removed to romote stable yielding at the reduced section and to effectively rotect the more vulnerable welded joints. This weakening strategy also reduces the seismic force demand on the column and the anel zone. Although this tye of moment connection has been widely used in the ast few years, there remain several design issues that should be further examined (for examle, Jones et al. 2002; Gilton and Contact Author: Cheol-Ho Lee, Assoc. Professor, Det. of Architecture, Seoul National Univ. San 56-1 shinlim-dong, Kwanak-gu, Seoul Korea. Tel: Fax: ceholee@snu.ac.kr Uang 2002; Chi and Uang 2002). The rimary objective of this exerimental study was to investigate the effects of beam web connection tye and anel zone strength on the seismic erformance of RBS connections 2. Testing Program 2-1. Design of Test Secimens A total of eight full-scale test secimens were designed and groued as Set No. 1 and Set No. 2 (Table 1). Tyical geometry and seismic moment rofile for the design of the radius-cut RBS are shown in Figs. 1 and 2. The grade of steel for the beams was SS400 with a secified minimum yield strength of 235 Ma (34 ksi); SM490 steel was used for the columns and the secified minimum yield strength was 324 Ma (47 ksi). The tensile couon test results are summarized in Table 2. The RBS design followed the recommendations by Iwankiw (1997) and Engelhardt et al. (1998). The beam end length (a) and the total length of the RBS zone (b) were chosen as 25% and 75% of the beam deth, resectively. These dimensions were selected to minimize the reduction in flange area. The resulting distance from the centerline of the RBS to the column face was 62.5% of the beam deth. The strain hardened lastic moment at the RBS hinge was calculated using the exected yield strength of the beam (F ye = 313 Ma) and a strain hardening 556 CTBUH 2004 October 10~13, Seoul, Korea

3 factor of 1.1. Engelhardt et al. (1998) recommended that the moment at the face of the column be limited to aroximately 85 to 100 ercent of M, where M = exected lastic moment of the beam. In this study the trimmed flanges were sized to limit the moment at the Table 1 Test Secimens Secimen Beam and column Panel zone strength DB700-SW DB700-MW DB700-SB DB700-MB DB600-MW1 DB600-MW2 DB600-SW1 DB600-SW2 H700X300X13X24 H428X407X20X35 H700X300X13X24 H428X407X20X35 H700X300X13X24 H428X407X20X35 H700X300X13X24 H428X407X20X35 H600X200X11X17 H400X400X13X21 H600X200X11X17 H400X400X13X21 H600X200X11X17 H588X300X12X20 H606X201X12X20 H588X300X12X20 Strong (10 mm doubler late, SM490) Beam web connection method Set No. 1 a (mm) b (mm) c (mm) reduction (%) Welded Medium Welded Strong (10 mm doubler late, SM490) Bolted Medium Bolted Set No. 2 Medium Welded Medium Welded Strong Welded Strong Welded Table 2 Tensile Couon Test Results Member Couon Yield strength (MPa) Beam 304 H700X300X13X24 Web 364 Column 343 H428X407X20X35 Web 358 Beam 326 H600X200X11X17 Web 343 Column 358 H400X400X13X21 Web 374 Beam 295 H606X201X12X20 Web 333 Column 374 H588X300X12X20 Web 405 Tensile strength (MPa) Yield ratio (%) CTBUH 2004 October 10~13, Seoul, Korea 557

4 column face to about 90 ercent of M as follows. Fig. 1. Tyical geometry of the radius-cut RBS Fig. 2. Seismic moment rofile for RBS design m = α Z F = (1.1) Z F (1) act RBS ye RBS act Lb 0.90 M M f = m (2) ' L The reduction in flange area at the RBS center was 37% and 40% for Set No. 1 and Set No. 2, resectively (see Table 1). The flange reduction in Set No. 1 was slightly less than the 40% minimum reduction limit of the SAC recommendation (SAC 2000). The anel zones were then designed by using either of the following two equations for the anel zone design strength: V V 2 3b cf tcf ( 0.75)(0.6Fycd ct ) 1 + (3) dbd ct = 2 3b cf tcf ( 0.6Fycd ct ) 1 + (4) dbd ct = where F yc = yield strength of the column web, d b = beam deth, d c = the column deth, t = thickness of the anel zone, b cf = the column flange width, and t cf = the column flange thickness. Eq. (3) was imlemented in the AISC Seismic Provision (AISC 1997). Secimens with ye V d anel zone designed by Eq. (3) are defined as strong anel zone secimens in this study. The anel zone satisfying Eq. (3) is exected to remain essentially elastic during the entire test and almost all the lastic rotation of the connection will be develoed by the beam. In Set No. 1, nominally identical steel shaes were used for the beams and columns, resectively. When Eq. (3) was used for the anel zone strength, doubler lates of 10 mm thickness were rovided to secimens DB700-SB and DB700-SW. The doubler lates were lug-welded to the column web to revent remature local buckling under large cyclic inelastic shear deformations (AISC 1997, AWS 2000). Eq. (4), which is adoted in the 2002 AISC Seismic Provisions, was used to design medium anel zone secimens. This equation, which does not include the resistance factor (0.75), reresents the anel zone shear strength at 4 times the shear strain at yield (Krawinkler 1978). To estimate the required shear strength of the anel zone, it is worthy of note that a reduction factor of 0.8 on beam yielding was included in the 1997 AISC Seismic Provisions to account for the effect that gravity loads might inhibit the develoment of lastic hinges on both sides of a column. However, based on the reasoning that there is no assurance that this will be the case, esecially for one-sided connections and for erimeter frames where gravity loads may be relatively small, the 2002 AISC Seismic Provisions require that the exected shear demand in the anel zone be determined from the summation of the moments at the column faces as determined by rojecting the exected moments at the lastic hinge oints to the column faces without considering the resence of gravity moments. Four medium anel zone secimens were included in this testing rogram (DB700-MW and DB700-MB in Set No. 1, DB600-MW1 and DB600-MW2 in Set No. 2). Secimens DB600-MW1 and DB600-MW2 in Set No. 2 were identical excet for a slight difference in the RBS length: that is, the RBS length was taken as 85% (DB600-MW1) and 65% (DB600-MW2) of the beam deth. Most of the ast tests have been conducted on secimens with a fully welded beam web. Recently, Jones et al. (2002) indicated that the use of a welded web connection does rovide some benefit to the connection erformance and it tends to reduce the vulnerability of the connection to weld fracture. To further investigate the influence of the beam web connection, two bolted web secimens, DB700-SB and DB700-MB, were included in Set No. 1. With a sli coefficient of 0.33, the sli-critical bolted web connection consisted of eight fully tensioned- M22-F10T high strength bolts. The bolts were tightened with the calibrated wrench method with a secified tension level of 201 kn. The ultimate strength of the bolted web connection was about two times the exected maximum beam shear. In Set No. 2, all the beam webs were groove-welded to the column flange. Continuity lates equal in thickness to the beam flange were rovided in all secimens. Electrodes with a secified minimum Chary V-Notch (CVN) toughness of 26.7 Joule at 28.9 ο C (20 ft-lb at 20 ο F ) was secified for 558 CTBUH 2004 October 10~13, Seoul, Korea

5 flux-cored arc welding. Weld access hole configurations followed the SAC recommendations (SAC 2000). Fig. 3 shows the connection details for secimen DB700-SB. In Table 1, the following abbreviations were used for the secimen designation: S= strong anel zone, M= medium anel zone, W= welded web, and B= bolted web. 3. Test Result and Discussion 3-1. Effects of Beam Web Connection Method The cyclic resonses of the secimens in Set No. 1 are resented in Fig. 5. The ordinate is exressed in terms of the normalized moment at the column face; the normalization was based on the nominal lastic moment of the original (unreduced) beam section. Both strong and medium anel zone secimens with a welded web connection develoed satisfactory levels of ductility required for secial moment frames. Fig. 3. Secimen DB700-SB moment connection details 2-2. Test setu and Loading The secimens were mounted to a strong floor and a strong wall. An overall view of the test setu is shown in Fig. 4. Fig. 5. Normalized moment versus story drift ratio relationshi (Set No.1) But secimens with a bolted web connection erformed oorly due to remature brittle fracture of the beam flange at the weld access hole (see Fig. 6). A comlete fracture across the beam flange width was develoed in both cases. Fig. 6. Beam bottom flange fracture of secimen DB700-SB at 2% story drift Fig. 4. Test Setu Lateral restraint was rovided at a distance of 2500 mm from the column face. The secimens were tested statically according to the SAC standard loading rotocol (Krawinkler et al. 2000). The beam ti dislacement corresonding to 1% story drift ratio was 38 mm. The test secimens were instrumented with a combination of dislacement transducers and strain gages to measure global and local resonses. Whitewash was ainted in the connection region to monitor yielding. Fig. 7 shows the lastic hinge formation in the welded web secimens. Significant yielding of the anel zone in secimen DB700-MW was evident from the flaking of the whitewash. Secimen DB700-SW exhibited excellent connection rotation caacity u to 6% story drift without fracture. Fig. 7. Connection region of secimens DB700-MW and DB700-SW CTBUH 2004 October 10~13, Seoul, Korea 559

6 Fig. 8 shows a comarison of the normalized maximum moment at the centerline of the RBS (i.e., assumed lastic hinge location). The normalization was based on the actual lastic moment of the narrowest reduced beam section. At a given story drift ratio, the figure shows that the bolted web secimens were less efficient in develoing moment caacity. Tsai and Poov (1988) indicated that web bolts tyically sli during testing, leaving the welded flanges alone to resist the total moment. Fig. 9 comares the cyclic flexural strain resonses of secimens DB700-SB and DB700-SW near the groove weld of the beam bottom flange u to the drift level secimen DB700-SB fractured. Much higher strain demand on the bolted web secimen is evident. These measured results aear to be consistent with the observation by Tsai and Poov. Normalized maximum moment at RBS brittle fracture DB700-SB DB700-MB DB700-MW DB700-SW Story drift ratio (%) Simlified finite element simulation of sli behavior was conducted using the general finite element analysis rogram ABAQUS. The three-dimensional finite element model for secimen DB700-MB was reared using eight-node continuum element (C3D8I in ABAQUS). The surface interaction between the shear ta and the beam web was formulated by the Coulomb friction model with an elastic sli with % of the characteristic element length. An elastic sli of this minimal amount was required to circumvent numerical convergency roblem. Pretension force was simulated by alying a air of concentrated forces to the front side of the shear ta and the back side of the beam web (see Fig. 10). A beam ti force of 499 kn was alied to simulate the beam shear corresonding to the strain hardened lastic moment at the RBS center. Fig. 11 shows the sli resonse based on the analytical model. It is evident that, contrary to the design assumtion, conventional sli-critical design can not revent sliage of the web bolts. Goel et al. (1997) also ointed out that the area in the middle of the beam web near the shear tab is virtually devoid of stresses and much of the shear force is transferred through the beam flanges, thus leading to overstressing of the beam flanges. The measured cyclic shear strain resonses are resented in Fig. 12. These measured results suort the foregoing observations by Goel et al. The shear transfer mechanism in the RBS connection is still not consistent with that redicted by the classical beam theory. Fig. 8. Comarison of normalized maximum moment at RBS (Set No. 1) Fig. 10. Finite element model for sliage simulation 0.18 Weld access hole Horizontal sli(mm) Beam ti force(kn) Shear ta A Vertical sli(mm) Beam ti force(kn) Fig. 9. Comarison of measured flexural strain resonses near the groove weld Fig. 11. Sli resonse at the uer right corner of the shear ta (oint A) 560 CTBUH 2004 October 10~13, Seoul, Korea

7 Medium PZ Strong P Z Normalized LTB amlitude (% ) (normalized by beam flange width) DB700-MW DB700-SW DB600- MW1 DB600- MW2 DB600- SW1 DB600- SW2 Fig. 14. Comarison of LTB amlitudes at 4% story drift cycle Fig. 12. Measured cyclic shear strain resonses (secimen DB700-SB) 3-2. Effects of Panel Zone Strength The lots shown in Fig. 13 indicate that all secimens in Set No. 2, with both strong and medium anel strengths, exhibited satisfactory connection ductility. Fig. 14 resents a comarison of the lateral-torsional buckling (LTB) amlitudes measured u to the 4% story drift cycles. Because both the beam and the anel zone contributed to lastic rotation in the medium anel zone secimens, LTB amlitudes were reduced. This is a sure advantage to reducing the tendency for global instability of the RBS beam. The cyclic strain hardening factor comuted at the RBS center based on the measured yield strength of the beam was of similar magnitude between the medium and strong anel zone secimens, and reached an average value of 1.27 at 4% story drift. This value is higher than that assumed 1.1 in AISC 2002 and 1.15 in FEMA 350. Fig. 13. Normalized moment versus story drift ratio relationshi (Set No. 2) Effects of anel zone strength on some connection resonses are summarized in Table 3. For the urose of analyzing the effects of anel zone strength, the Krawinkler s recommendation (Eq. 4), which includes the contribution of the column flange to the ost-yield strength, was used as a measure of the anel zone strength. The measured yield strength in Table 2 was used to calculate the anel zone strength. As a measure of the beam strength, the anel zone shear force V RBS, P corresonding to the actual lastic moment of the RBS was used; a similar strength measure was used by Roeder (2002). For a one-sided moment connection, V RBS, P can be comuted as follows: V RBS, P M L /2 + d /2 d = 1 d L /2 e H RBS, P b c b b b c (6) where M RBS, P = actual lastic moment at the RBS center based on the measured yield stress, H c = column height, and refer to Fig. 2 for some remaining symbols. In Table 3, secimen DB700-SW was excluded because the tensile couon test results for the doubler lates were not available. To augment the database, test results from Yu et al. (2000), Chi and Uang (2002) was included. Since available test results show that the anel zone can easily develo a lastic rotation of 0.01 rad. without causing distress to the beam flange groove welds, and Table 3 shows that the anel zone at this deformation level would dissiate about 30% to 40% of the total energy, for a balanced design it is suggested that either of the following criterion be satisfied in design: Table 3 Effects of anel zone strength on lastic rotation and energy dissiation Secimen PZ strength relative to beam V / V RBS, y Panel zone lastic rotation at 4% story drift ratio (rad) Energy dissiation by anel zone u to 4% story drift cycle (%) DB700-MW DB600-MW DB600-MW DC2* DB600-SW DB600-SW Negligible Negligible * From Chi and Uang (2002) CTBUH 2004 October 10~13, Seoul, Korea 561

8 V 0.70 V RBS, 0.90 (7) Industrial Science and Technology (RIST) in Korea is gratefully acknowledged. References This recommended range attemts to achieve the following: (1) to minimize the use of exensive doubler lates, which often require welding near the k area of the column, (2) to reduce the amount of beam buckling amlitude (i.e., beam torsion), and (3) to encourage the anel zone to rovide about 0.01 rad. lastic rotation, which corresonds to about 30% to 40% of the total energy dissiation in the connection region. 4. Conclusions The results of this study are summarized as follows. (1) Both strong and medium anel zone secimens with welded web connection exhibited satisfactory levels of connection ductility required of secial moment-resisting frames. Secimens with a bolted web connection erformed oorly due to remature brittle fracture of the beam flange at the weld access hole. If fracture within the beam flange groove weld in a bolted web connection was avoided by using quality welding, fracture tended to move into the beam flange base metal at the weld access hole. The exerimental and analytical results of this study aear to imly that the high incidence of base metal fracture in secimens with a bolted web attachment is related to, at least in art, the increased demand on the beam flanges due to the web bolt sliage and the actual load transfer mechanism which is comletely different from that usually assumed in connection design. (2) For welded-web RBS moment connections, test results showed that the anel zone could easily develoed a lastic rotation of 0.01 rad. without distressing the beam flange groove welds. With this level of inelastic deformation, the anel zone would dissiate about 30% to 40% of the energy. Allowing the anel zone to deform inelastically at this level also reduces the magnitude of beam distortion (e.g., lateral torsional buckling) by about a half. A criterion for a balanced PZ strength that imroves the lastic rotation caacity while reduces the amount of beam buckling is resented. Acknowledgments Funding for this research rovided by the Korea Earthquake Research Center (KEERC Project No. R ) and the Research Institute of 1) Chen, S. J., Yeh, C. H., and Chu, J. M. Ductile steel beam-to-column connections for seismic resistance. J. Struct. Engrg., ASCE, 122(11), 1996: ) Chi, B. and Uang, C.-M. Cyclic resonse and design recommendations of reduced beam section moment connections with dee column. J. Struct. Engrg., ASCE, 128(4), 2002: ) Engelhardt, M. D., Winneberger, T., Zekany, A. J., Potyraj, T. J. Exerimental investigations of dogbone moment connections. Engrg. J., 35(4), AISC, Fourth Quarter, 1998: ) Gilton, C. S. and Uang, C.-M. Cyclic resonse and design recommendations of weak-axis reduced beam section moment connections. J. Struct. Engrg., ASCE, 128(4), 2002: ) Goel, S. C., Stojadinovic, B., and Lee, H.-K. Truss analogy for steel moment connections, Eng. J. 34(2), 1997: ) HKS(1998). ABAQUS User s Manual, Version 5.8 Hibbit, Karlson & Sorenson, Inc 7) Iwankiw, N. Ultimate strength consideration for seismic design of the reduced beam section (internal lastic hinge). Engrg. J. 34(1), 1997: ) Jones, S. L., Fry, G. T., and Engelhardt, M. D. Exerimental evaluation of cyclically loaded reduced beam section moment connections. J. Struct. Engrg., ASCE, 128(4), 2002: ) Krawinkler, H. Shear in beam-column joints in seismic design of steel frames. Engrg. J. 15(3), 1978: ) Krawinkler, H., Guta, A., Medina, R., and Ruco, N. Loading histories for seismic erformance testing of SMRF comonents and assemblies. Reort No. SAC/BD-00/10. SAC Joint Venture, Sacramento, Calif., ) Plumier, A The dogbone: back to the future. Engrg. J. 34(2), 1997: ) Roeder, C. W. General issues influencing connection erformance. J. Struct. Engrg., ASCE, 128(4), 2002: ) SAC. Seismic design criteria for new moment-resisting steel frame construction. Reort No. FEMA 350, SAC Joint Venture, Sacramento, Calif., ) American Institute of Steel Construction (AISC). Seismic rovisions for structural steel buildings., American Institute of Steel Construction, Chicago, 1997 and ) American Welding Society (AWS). Structural welding code-steel, AWS D1.1: Section 3.10, American Welding Society, FL., ) Tsai, K. C. and Poov, E. P. Steel beam-column joints in seismic moment resisting frames. EERC Reort, UCB/EERC-88/19, Univ. of California, Berkeley, ) Zekioglu, A., Mozaffarian, H., Chang, K. L., and Uang, C.-M. Designing after Northridge. Modern Steel Constr., 37(3), 1997: CTBUH 2004 October 10~13, Seoul, Korea