Engineering Structures 24 (2002) 1523–1534 www.elsevier.com/locate/engstruct
Bolted large seismic steel beam-to-column connections Part 1: experimental study Egor P. Popov, Shakhzod M. Takhirov ∗ Pacific Earthquake Engineering Research Center, University of California at Berkeley, Richmond, CA 94804-4698, USA Received 17 April 2001; received in revised form 15 November 2001; accepted 29 May 2002
Abstract Two large bolted steel moment-resisting connections were studied by experiments. These connections were single-sided beamto-column assemblies that are representative of exterior beam-to-column connections, and they were composed of W36×150 Grade 50 beams and W14×283 Grade 50 columns. T-stubs were cut from W40×264 sections of Grade 50 steel. The T-stub stems were welded to the beams and prestressed by bolts to the beam flanges in the shop. Final beam-to-column assembly required no additional welding: the T-stub flanges were bolted to the column and the column shear tab was bolted to the beam web. The specimens had two symmetrically located T-stubs with different stem geometry: Specimen 1 had rectangular-shaped stems, whereas Specimen 2 had U-shaped stems. During the cyclic testing the beam deformation was minimal controlled by active participation of the T-stub flanges: a separation between T-stub flanges and the column flanges was observed. This separation was caused by bending plastic deformation in the T-stub flanges and plastic deformation in the high-strength bolts. This phenomenon allowed energy dissipation and prevented severe buckling of the beam flanges and beam web. 2002 Elsevier Science Ltd. All rights reserved. Keywords: Steel; Bolted moment-resistant connection; T-stub; High-strength bolt; Full-scale experimental study; Yielding; Fillet shop welded; Field bolted; Seismic
1. Introduction The typical welded steel moment frame connection used in seismically active zones in the United States failed to provide the expected ductile behavior in the 1994 Northridge earthquake in Los Angeles, California. Damage occurred at the beam-to-column joints and included fractures of full penetration welds, cracks in beam flanges, and cracks through the column sections. The generally accepted detail of attaching steel beams to columns in seismic applications consists of attaching shear tabs to the column and direct welding beam flanges with or without cover plates, to column flanges. Numerous tests on this type of connection were supported by the National Science Foundation (NSF) with many specimens donated by the fabricators. The testing of the specimens was organized by the SAC Joint Venture. One
Corresponding author. Tel.: +1 510 231 9563; fax: +1 510 231 9572. E-mail address:
[email protected] (S.M. Takhirov). ∗
of the most recent comprehensive studies on welded steel beam-to-column connections was conducted by the research group at the Pacific Earthquake Engineering Research Center (PEER), University of California, Berkeley [1]. The research program focused on the response of steel moment-resisting connections reinforced with flat steel plates that served to relocate the plastic hinge away from the face of the column. The moment capacity of such connections depends on the cyclic endurance of the flange and the cover plate welds in both tension and compression. In the report [1], two types of welded plate-reinforced connections were tested and analyzed namely, the cover-plate and flange-plate connections. The objective of this research is to provide an alternative bolted connection having low installation cost and high reliability. Bolted and riveted connections have been used for decades and have performed well in past earthquakes, particularly when encased in concrete, as was traditionally done for fireproofing until the late 1950s [2–4]. The bolted connection has the advantage of eliminating the difficulties of field welding. Also pro-
0141-0296/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 1 4 1 - 0 2 9 6 ( 0 2 ) 0 0 0 8 6 - X
1524
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
posed is an approach to the connection design that avoids weld failures and facilitates shop welding with field bolting. An attempt at the above approach on several end-plate connections was made by K. C. Tsai and E. P. Popov at the University of California, Berkeley [5]. Three endplate specimens (Specimen 10, 10R, and 12) were tested under cyclic loading. In all cases, the beam section was fully welded to the end plate, using fillet welds without any web cope for backup bars. Specimen 10 with no ribs over the beam flanges did not give satisfactory results; after premature bolt failure it was modified and tested again as Specimen 10R. This modified specimen with ribs over the beam flanges and larger bolts behaved very well under cyclic loading. Note that the required large thickness of the end plate (a connection based on the design of specimens 10 and 10R) may require shims during assembling. A W18×40 beam section for Specimen 10 and a W21×44 beam section for Specimen 12 were used. The similar research on end-plate connections was also recently pursued by T. M. Murray and his associates in 2000 at Virginia Polytechnic Institute (VPI) [6] with good results. They achieved a number of successful tests with W36×150 beams. Eleven beam-to-column extended moment end-plate connection tests were performed. Four of these tests were conducted with a W36×150 beam and a W14×257 column. It appears that for larger or heavier beams the use of ribs over beam flanges at columns would be required. A comprehensive study of full-size beam-to-column connections employing new types of bolted connections was conducted at Lehigh University as a joint industry study with ICF Kaiser Engineers [7]. A total of eight connection tests was performed. The research showed that bolted connections are capable of providing rigid moment connections with cyclic plastic rotational capabilities in excess of equivalent welded joints, but with the same rigidity as welded connections. The beam depth ranged from 16 in. (406 mm) to 36 in. (914 mm). The bolted connection was achieved by using all-welded haunch brackets and pipe brackets. An extensive excellent study of bolted connections was been done in 2000 at Georgia Institute of Technology by R. T. Leon and his associates. [8]. Six beamto-column full-scale connection tests were performed. The idea of the connection design was similar to the one used in this paper: the beam was bolted to the column by using T-stubs cut from available beam sections. In two specimens the beam was made from a W21×44 beam section, whereas in the remaining four specimens a W21×55 beam section was used. In all specimens a W14×145 column section was used for the columns. The study on bolt size and T-stub size was conducted. The work was very comprehensive, but was limited to small and medium-size members.
The results of tests on 48 T-stub specimens are presented and discussed in [9]. These tests were carried out in order to study the behavior, failure modes, and ductility of this detail of bolted connection. The main variables tested include the size of the T-stub, the gauges of the bolts, and the type and diameter of the bolts. The tests were conducted for different size T-stubs: the largest size of the tested specimens was W33×169. The simplified theoretical model of the T-stub based on these results is presented in [10]. The newly developed and tested connection at University of California, Berkeley is somewhat related to the end-plate connection but is more versatile because it is more readily adaptable to a larger range of heavier beams. The new connection depends on the use of A490 1-1/4 in. (31.75 mm) bolts in tension in oversized round holes (as in the column flange and in the T-stub flange as well) and shop fillet welds. Two specimens based on this connection design were fabricated and tested; the cyclic test procedure of the specimens was based on [11]. In the aftermath of the Northridge earthquake, a connection plastic rotation of 0.03 radian under cyclic loading, with a loss of flexural strength less than 20%, has been suggested as an acceptable rotation limit to differentiate between ductile and brittle connections for special moment-resisting frames (see, for example [9]). The primary goal of the project was to develop a large bolted connection with small amount of shop welding that would have a ductile behavior under cyclic loading. The testing procedure consisted of two stages: (1) an experimental study on the high-strength bolts used in the connection and (2) testing of Specimen 1 and Specimen 2. 2. Tension tests on A490 1-1/4 in. (31.75-mm) bolts In order to conduct a ductility study of the A490 11/4 in. (31.75 mm) bolts, two tests were performed. For the first test, a special device was built simultaneously to test the shank of the bolt and the threaded part of the bolt below the nut. The actual failure occurred in the threaded region. The remarkable ductility of the A490 bolt was clearly demonstrated as shown in the load versus elongation diagram of Fig. 1. In another experiment the excellent ductility of the A490 bolt material was shown in a specimen of constant diameter machined from this bolt. Fig. 2 shows the stress-strain diagram for this test. The ultimate stress for A490 bolts is specified as 150 ksi (1033 MPa). The measured ultimate stress was very close to this specification value. 3. Test specimen design and detailing The details of the two newly developed connections using the A490 1-1/4 in. (31.75-mm) bolts are shown in
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 1.
Load versus elongation for A490 1 1/4 in (31.75 mm) bolt.
1525
were bolted to the columns. The design calculation was based on widely used references [12–16] and because of space limitations is not presented here. The calculation was conducted prior to manufacturing of the specimens and was published earlier [17]. The connection design uses the tee stem-to-beam flange bolts to enforce the tee stem and the beam flange joint performance. One of the goals of the experimental study (presented here) and the nonlinear numerical analysis (presented in companion paper) was to evaluate the right location of these bolts and the possibility of omitting them. The geometric location and the number of pre-stressing 1-in. (25.4-mm) bolts in Specimen 2 were different from that of Specimen 1. In Specimen 1 these bolts were located along two rows (per each beam flange): one close to the face of the column, the other close to the end of the T-stub stem. Specimen 2 had only one row of these bolts with the location slightly closer to the end of the fillet weld: the difference in distance was about 0.8 in. (20 mm) between Specimens 1 and 2.
4. Test setup
Fig. 2. Stress versus strain for coupon test of A490 1 1/4 in (31.75 mm) bolt material [coupon was 3/8 in (9.53 mm) in diameter].
Figs. 3 and 4. In both cases the attachment of beams to a column is made using structural tees cut from W shapes (T-stubs). A large choice of such sections is available. By rotating the beam all fillet welds can be done in the shop in a down-hand position. Shop experience in fabricating these two specimens was very encouraging. These steel moment-resisting connections were singlesided beam-to-column assemblies that are representative of exterior beam-to-column connections. The beams were fabricated from a W36×150 section of A572-Gr.50 steel, the columns were fabricated from a W14×283 section of A572-Gr.50 steel. The T-stubs were made from a W40×264 section of A572-Gr.50 steel. The global dimensions and geometry of the specimens are shown in Fig. 5. The material properties of the connection members from mill certificate data are presented in Table 1. The specimens had two symmetrically located T-stubs with different stem geometry: Specimen 1 had rectangular-shaped stems, Specimen 2 U-shaped stems. The T-stub stems were welded and prestressed to the beam flanges in the shop, and later the T-stub flanges
The specimens were tested in the Structural Research Laboratory of PEER, at the Richmond Field Station, UC Berkeley. The test setup was designed to accommodate specimens with columns in a vertical position (Fig. 5). The specimens were attached to horizontal and vertical frames. The horizontal steel frame was prestressed to the strong floor. The columns in the test specimens were attached to the horizontal frame and the vertical reaction frame using short segments of W14×311 to achieve nearpinned boundary conditions. The load was applied to the cantilever beam end by a 400 kips (1780 kN) hydraulic actuator, through a clevis bolted to the beam end plate. The testing setup had a displacement capacity of ±7.75 in. (197 mm) and a load capacity of ±350 kips (1558 kN). No axial load was applied to the column. The test was conducted using the beam end displacement control. The beam end was at a distance of 131 in. (3330 mm) from the column face. To prevent out-of-plane movement of the beam, a vertical bracing system was provided near the beam end. The photograph in Fig. 6 shows a view of a test in progress.
5. Test results Table 2 presents the loading protocol for both specimens. The testing program consisted of symmetric, stepwise-increasing displacements imposed by the actuator at the tip of the beam. To study the small deformation response of the connection in post-yielding stage, two cycles with a small amplitude of 0.42 in. (11mm) were imposed after cycles with 1.83 in. (47 mm) beam tip
1526
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 3. Design details of Specimen 1.
Fig. 4. Design details of Specimen 2.
displacement. For the test specimens, 0.01 radian story drift corresponds to 1.31 in. (33 mm) at the beam tip. 5.1. Specimen 1 The specimen sustained all loading steps up to and including the 5.20-in. (132-mm) beam tip displacement cycles without significant damage. Testing was stopped because the maximum load for the test setup was reached. A close-up side view of the specimen after the test is presented in Fig. 7. During the last set of the load
reversals a slight buckling in the beam web and flanges was observed. The cyclic yielding of the T-stub flanges and high-strength bolts was observed, with the gap between the T-stub and column flanges opening and closing periodically. The residual gap in the upper Tstub is shown in Fig. 8. The plot of applied force versus beam tip displacement is presented in Fig. 9. The load-carrying capacity of the connection was very stable up to the last stage of cyclic testing, without any capacity degradation, but not in other tests [1,5]. The maximum beam end displace-
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 5.
Test setup for both specimens.
ment was 5.2 in. (132 mm) and the maximum imposed load was 345 kips (1535 kN). The connection maximum story drift was very close to 0.04 radian and the maximum moment at the center line of the column was 48,645 kips × in. (5497 kN × m). Fig. 10 shows the applied moment versus the connection plastic rotation.
1527
The maximum plastic rotation of the connection was about 0.026 radian that is about 65% of the total maximum rotation of the connection (0.04 radian). During the test the visible opening between the Tstub flanges and the column flange was observed. The amplitude of the T-stub flange and column flange relative displacement was monitored by a displacement transducer. During mutual compression of these two flanges the displacement is negative, whereas the tension in the T-stub stem increases this distance and it becomes positive. For a relatively high positive value of this relative displacement, a visible gap opening was observed. The relative displacement increases when the tension force in the T-stub stem became larger than the prestress force (the T-stub flanges were prestressed to the column flange). Further increase in the tension force caused the yielding of the high-strength bolts. Another reason for this relative displacement increase was that yielding occurred in the T-stub flange. This relative displacement between the flanges is referred in this paper as “gap opening”. The gap opening between the top T-stub flange and the column flange versus the imposed load at the beam tip is presented in Fig. 11. The maximum measured gap opening at the top T-stub during the test was 0.31 in. (8 mm). The connection rotation caused by these
Table 1 Material properties No
Part of connection
Yield stress, ksi (MPa)
Ultimate stress, ksi (MPa)
Section sizea
Gradea
1 2 3
Beam Column T-section
56.6 (390.0) 52.0 (358.3) 64.0 (441.0)
74.4 (512.6) 66.0 (454.7) 79.0 (544.3)
W36×150 W14×283 WT40×264
Gr50 Gr50 Gr50
a
According to US notations.
Fig. 6.
View of the test in progress.
1528
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Table 2 Testing program for both specimens Step No
1
2
3
4
5
6
7
8
9
10
Beam end displacement, in. (mm) No of cycles
0.28
0.42
0.57
0.88
1.20
1.83
0.42
2.46
3.85
5.20
(7.1) 6
(10.7) 6
(14.5) 6
(22.4) 6
(30.5) 4
(46.5) 2
(10.7) 2
(62.5) 2
(97.8) 3
(132.1) 6a
a
Only 2 cycles at this level were performed for Specimen 2.
Fig. 7.
Fig. 8.
Specimen 1 after the test (side view).
Residual gap opening in the top T-section (after the test).
openings in both T-stubs is presented in Fig. 12. The plots show that the maximum connection rotation caused by the plastic deformations in the T-stub and highstrength bolts was about 0.01 radian or 25% of the total beam rotation (0.04 radian). The relatively high value of the connection rotation caused by the gap opening in the
T-stubs shows the active participation of the T-stubs in the connection global deformation. The relative beam rotation was calculated by subtracting (from the connection total rotation) the rigid body rotation of the column, the rotation caused by the gap opening in the T-stubs, and the panel zone shear
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 9. Imposed load versus beam end displacement for Specimen 1.
1529
Fig. 12. Imposed load versus connection rotation caused by gap opening (Specimen 1).
deformation. The maximum rotation of the beam itself was about 0.006 radian or about 15% of the connection total rotation (0.04 radian). This small value explains the minimal deformations in the beam with slight buckling in the beam flanges and the beam web. 5.2. Specimen 2
Fig. 10.
Moment versus connection plastic rotation for Specimen 1.
Fig. 11. Relative displacement between column and top T-section flanges (Specimen 1).
The specimen sustained all loading steps up to the 5.20 in. (132 mm) beam tip displacement cycles and failed at the first ramp of the last cycle. The fracture was caused by a crack in the stem of the lower T-stub. The crack line started at the end of the fillet weld and went through the nearest hole for a 1-in. (25.4 mm) bolt. Testing was stopped at the end of this cycle. At the end of the test a slight buckle in the beam web and flanges was observed. A photo of the side view after the testing is presented in Fig. 13. The possibility of the T-stub stem fracture in Specimen 2 was expected. As previously mentioned, the holes for 1-in. (25.4-mm) bolts were located (comparable to in Specimen 1) closer to the end of the fillet weld, which created a weak cross section very close to the stress concentration zone. The test of Specimen 2 demonstrated that the use of 1-in. (25.4-mm) bolts requires a greater distance between the bolt and the end of the fillet weld. The nonlinear numerical analysis presented in the second part of the paper reveals that the bolts can be omitted altogether. Similar as in the Specimen 1 test, cyclic yielding of the T-stubs and high-strength bolts was observed, with the gap opening and closing between the T-stub and the column flanges. The residual gap in the top T-stub is shown in Fig. 14. The crack in the stem of the bottom T-stub is shown in Fig. 15. The location was close to the k-line of the
1530
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 13.
Fig. 14.
Specimen 2 after the test (side view).
Specimen 2: residual gap opening in top T-section.
T-stub and was parallel to it. The crack started from the end of the fillet weld, continued through the nearest hole for the 1-in. (25.4-mm) bolt and ended at the next bolt hole. The arrows on the photo trace the crack line. As previously mentioned the 1-in. (25.4 mm) bolts were located closer to the weld end point in Specimen 2, and the chosen location of the bolts reduced the load capacity of the T-stub stem and the beam flange joint section. This was the main reason for failure in Specimen 2. The plot of applied force versus beam tip displacement is presented in Fig. 16. The load-carrying capacity of the connection was very stable up to the moment of failure at the bottom of the T-stub. The maximum beam end displacement was 5.2 in. (132 mm) and the maximum imposed load was 327 kips (1455 kN). The connection maximum drift was very close to 0.04 radian,
and the maximum moment at the center line of the column was 486,107 kips × in. (5210 kN × m). Fig. 17 shows the applied moment versus the connection plastic rotation. The maximum connection plastic rotation was about 0.033 radian, about 80% of the total maximum rotation of the connection (0.04 radian). The value for the connection plastic rotation is relatively high because it was calculated including the failure stage of the specimen. Similar to the previous specimen test, the visible gap opening was observed between the T-stub and column flanges. The gap opening between the top T-stub flange and column flange versus imposed load at the beam tip is presented in Fig. 18. The maximum measured gap opening at the top T-stub during the test was 0.28 in. (7 mm). The beam rotation caused by these openings in
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
1531
Fig. 15. Specimen 2: close view of the crack line.
Fig. 16. Imposed load versus beam end displacement (Specimen 2).
Fig. 18. Relative displacement between column and top T-section flanges (Specimen 2).
both T-stubs is presented in Fig. 19. The plots show that the maximum beam rotation connected with the plastic deformations in the T-stub and high-strength bolts was about 0.007 radian, or 18% of the total beam rotation (0.04 radian). The relatively high value of the connection rotation caused by the gap opening in the T-stubs shows the active participation of the T-stubs in the connection global deformation. The maximum relative rotation of the beam was about 0.015 radian, or about 38% of the connection total rotation (0.04 radian). The calculated rotation includes the stages of the specimen failure; that is why it was much higher then in the case of Specimen 1. 6. Experimental results and conclusions Fig. 17.
Moment versus connection plastic rotation (Specimen 2).
A brief summary of the experimental results and the key parameters characterizing the performance of the
1532
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
Fig. 19. Imposed load versus connection rotation due gap opening (Specimen 2).
tested specimens is presented in Table 3. Both specimens had very close values for the elastic stiffness of the connection and the maximum imposed load at the tip of the beam. The global behavior of Specimen 1 demonstrated the ductility of the specimen: up to 0.04 radian story drift with minimal beam deformation and with no capacity degradation. The maximum value of the connection plastic rotation was 0.026 radian for Specimen 1 without any loss of the flexural strength. The further testing in order to achieve 0.03 radian connection plastic rotation was impossible because of the test setup limitations. During Specimen 2 testing the connection plastic rotation achieved the maximum value of 0.033 including the rotation at the failure stage. Relative rotation of the beam was very small, 0.006 radian (for Specimen 1), compared to the specimen total rotation of 0.04 radian. The observed behavior was caused by the plastic bending deformation of the T-stub flanges and the A490 1-1/4 in. (31.75-mm) bolts yielding. The main lesson from the failure of Speci-
men 2 is that the A490 1-in. (25.4-mm) bolts have to be moved further from the face of the column and located far enough from the end of the fillet weld. The nonlinear numerical analysis presented in the second part of the paper provides a proof that the bolts can be omitted altogether without significant change in the connection global behavior. Remarkable local behavior was demonstrated by the specimens. The most important parts of the connection are the T-stub stem connection to the beam flanges and the connection between the column and T-stub flanges. The flakes of the brittle paint applied to the specimen and the strain gauge reading showed plastic deformation in the T-stub stem close to the k-line. The strain gauges were glued to the top surface of the T-stub stem and the beam flange, as shown in Fig. 20. The maximum values of the strain at a certain location during a chosen cycle of the test are presented in Fig. 21 (the cycle with low amplitude numbered step 7 in Table 3 is omitted here). The horizontal axis represents the strain value in percent and the vertical axis represents the maximum imposed load during a chosen cycle. The curve denoted by filled squares corresponds to the strain in the middle of the Tstub stem (close to k-line) measured by gauge 1 and the curve denoted by filled circles corresponds to the strain in the middle of the beam flange (close to the end of Tstub stem) measured by gauge 2. The plots presented in this figure show joint behavior of the T-stub stem and the beam flange. The yielding in the T-stub stem starts at about 247 kips (1100 kN). The imposed load causing the yielding in the beam flanges is higher and equals to 290 kips (1300 kN) as can be seen from Fig. 21. Yielding in the middle part of the beam flange occurred after the yielding in the T-stub stem and flanges. 6.1. Advantages The design and performance of the proposed beamto-column connections shows the following advantages:
Table 3 Short summary of test results Key parameters
Specimen 1
Specimen 2
Yield load, kips (kN) Beam end displacement at yield point, in. (mm) Elastic stiffness of connection, kips/in. (kN/m) Maximum beam end displacement, in. (mm) Beam end displacement at failure, in. (mm) Maximum imposed load, kips (kN) Maximum imposed moment, kips × in. (kN × m) Maximum connection rotation, radian Maximum connection plastic rotation, radian Maximum rotation due gap opening, radian Maximum relative beam rotation itself, radian
230 (1025) 1.2 (30.5) 180 (31536) 5.2 (132.1) N/A 345 (1535) 48645 (5496.9) 0.040 0.026 0.010 0.006
230 (1025) 1.2 (30.5) 178 (31186) 5.2 (132.1) 3.5 (89) 327 (1455) 46107 (5210.1) 0.040 0.033 0.007 0.015a
a
This value is high because it includes beam rotation after the bottom beam flange failure.
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
앫
앫
앫
Fig. 20.
Strain gauge locations in Specimen 1.
앫
1533
clamping force between the column and the T-stub flanges, the widely available torque multiplier WRIGHTTOOL Model 9S393A was used. The device does not produce any noise and has an accuracy of ±5%. When the testing program was finished, Specimen 2 was disassembled. The procedure was conducted at the PEER Structural Research Laboratory and did not require any torching; the beam was simply unbolted from the column. This procedure demonstrated that repairing and replacing the beam with a new T-stub is possible. Actual beam deformation is minimal, because it is controlled by active participation of the T-stub flanges, T-stub stems, the high-strength 1-1/4 in. (31.75-mm) bolts, and the beam flanges during cyclic input. The construction based on this design requires shims during erection, but with shims properly installed, the connection develops less residual strain than the welded one. The design of the connection allows eliminating large quantities of field welding and greatly helps the connection work to keep up with the steel erection.
6.2. Disadvantages and suggested improvements The chosen design and the failure of Specimen 2 show the disadvantages and suggested improvements:
Fig. 21.
Strain gauge data for Specimen 1.
앫 Specimen 1 has achieved 0.026 radian plastic rotation under cyclic loading with no degradation in the loadcarrying capacity and without any failure. This value is very close to the plastic rotation limit of 0.03 radian used to differentiate between ductile and brittle connections for special moment-resisting frames. 앫 The chosen connection design needs less welding work and this work can be done in a welding shop in convenient welding positions. This eliminates the expensive inspection of the welding work. 앫 Final assembling with bolts is a relatively easy procedure and does not require a rigorous quality-assurance inspection. In order to achieve the required
앫 The test of Specimen 2 demonstrated that the use of 1-in. (25.4-mm) bolts (as used in Specimen 2 to prestress the T-stub stem to beam flange) requires a greater distance between the bolt and the end of the fillet weld. Alternatively, it appears that the bolts can be omitted altogether. 앫 As noted before, as with any other bolted connection, the connection based on the proposed design requires shims for field assembly. 앫 One more disadvantage of the connection is that beams with welded top and bottom T-stubs require more shipping space during transportation. 6.3. Demand on steel material high quality The chosen design assumes that the specification satisfied material properties of the joint parts in the connection. The material property of the used steel can be obtained from mill certificate data or from specially conducted coupon tests on samples. 앫 The material properties of any part of the connection are important. The most important is that the steel material of the T-stub must be carefully selected. 앫 The steel material of 1-1/4 in. (31.75-mm) bolts has to be high quality as used in the tested connections,
1534
E.P. Popov, S.M. Takhirov / Engineering Structures 24 (2002) 1523–1534
and has to be very close to the specification for the high strength bolts.
Acknowledgements Special thanks are due to Professor Vitelmo V. Bertero of the University of California, for detailed review of the experimental results, useful comments and recommendations to improve the presentation of the results. Special thanks are also due to Ms. Janine Hannel of the PEER for her help with the editing of the paper.
References [1] Kim T, Whittaker AS, Gilani ASJ, Bertero VV, Takhirov SM. Cover-plate and flange-plate reinforced steel moment-resisting connections. Report No. PEER 2000/07. Berkeley, CA: Pacific Earthquake Engineering Research Center, University of California. 2000. [2] Munse WH. Fifty years of riveted, bolted, and welded steel construction. J Constr Div ASCE 1976;102(3):437–47. [3] Roeder CW, Leon RY, Preece FR. Strength, stiffness, and ductility of older steel structures under seismic loading. Re. SGEM 94-4. Seattle, WA: Dept. of Civil Engineering, University of Washington. 1994. [4] Leon RT. Seismic performance of bolted and riveted connections. background reports on metallurgy, fracture mechanics, welding, moment connections and frame system behavior. (SAC report 9509). Sacramento, CA: FEMA 288 SAC Joint Venture 1997;5:1–5:83.
[5] Tsai KC, Popov EP. Cyclic behavior of end-plate moment connections. J of Struct Engineering ASCE 1990;116(11):2917–30. [6] Murray TM, et al. Cyclic testing of bolted moment end plate connections. Blacksburg, VA: Struct. and Materials Lab., Virginia Polytechnic Institute and State University. 2000. [7] Kasai K, Hodgson I, Bleiman D. Rigid-bolted repair method for damaged moment connections. Engineering Structures 1998;20(4-6):521–32. [8] Leon R, et al. Tests on bolted connections. Report No. SEMM 00-02. Atlanta, GA: School of Civil and Environmental Engineering, Georgia Institute of Technology. 2000. [9] Swanson JA, Leon RT. Bolted steel connections: tests on T-stub components. J Structural Engineering ASCE 2000;126(1):50–6. [10] Swanson JA, Leon RT. Stiffness modeling of bolted T-stub connection components. J of Structural Engineering ASCE 2001;127(5):498–505. [11] SAC. Protocol for fabrication, inspection, testing, and documentation of beam-column tests and other experiments. Rep. No. SAC/BD-97/02, Washington, D.C.: Joint Venture, Strategic Air Command. 1997. [12] AISC. Manual of steel construction. Load & resistance factor design. Vol. 1: Structural members, Specifications & Codes., 2nd Chicago: American Institute of Steel Construction, Inc, 1995a. [13] FEMA-267. Interim guidelines: evaluation, repair, modification and design of welded steel moment frames. FEMA Report No. 267. Washington, D.C.: Federal Emergency Management Agency. 1995a. [14] FEMA-267. Interim guidelines: Advisory No. 1. supplement to FEMA 267. FEMA Report No. 267. Washington, D.C.: Federal Emergency Management Agency. 1995b. [15] AISC. Manual of steel construction. Load & resistance factor design. Vol. 2: Connections., 2nd Chicago: American Institute of Steel Construction, Inc, 1995b. [16] Kulak GL, Fisher JW, Struik JHA. Guide to design criteria for bolted and riveted joints. New York, NY: Wiley, 1987. [17] Popov EP, Takhirov SM. Large seismic steel beam-to-column connections. Moraga, CA; Steel Tips: Structural Steel Educational Council. 2001.
