Steel moment-resisting connections can be reinforced with cover or flange plates to protect the potentially vulnerable beam-to-column groove welded joint by relocating the plastic hinge away from the face of the column. A number of reinforced connection tests have been conducted at Berkeley since the 1994 Northridge, California, earthquake with mixed results. Specimens of varying sizes, reinforcing plate geometry, weld geometry, and panel zone strengths have been tested. Specimen plastic rotations have ranged from 0.00 to 0.05+ radian.
In 1998 the SAC Joint Venture funded a coordinated analytical and experimental study to resolve questions related to the design and detailing of steel moment-resisting connections reinforced with cover or flange plates. (For cover plated connections, the beam flange and cover plate assemblies are groove welded to the column flange. For flange plate connections, only the flange plates are welded to the column.) As part of this study, five full-scale single-sided joints have already been tested at Berkeley and five additional specimens will be tested in the next three months. Various joint details have been developed for the ten specimens to investigate the effects on specimen response of reinforcing plate geometry, plate-to-flange fillet weld, restraint to lateral torsional bucking, panel zone strength, and the utility of yielding reinforcing plates. All ten specimens are fabricated using W14x176 Grade 50 columns and W30x99 Grade 50 beams. The beam web is compact per the AISC Seismic Provisions but the beam flange is not.
The nonlinear finite element analysis program ABAQUS is being used to investigate the behavior of cover-plated connections and welded flange-plated connections of varying geometry to predict the local and global response of the test specimens, and to identify locations for instrumentation of the test specimens. Two types of models are under development. One type uses solid elements to calculate linear and nonlinear (plastic) strain profiles and specimen susceptibility to brittle fracture. Shell elements are used for the second type of model to characterize large-displacement responses and flange and web local buckling. Data from the elastic analysis of one cover-plated connection are presented in figure 1. (Only one-quarter of the beam-column connection is shown in the figure.) In this specimen, the cover plate is fillet welded to the beam flange on three sides, including across the nose of the cover plate. Critical sections can be seen at the nose of the cover plate and at the face of the column. Welding across the nose of the cover plate appears to improve the response of the connection.
Five SAC specimens (four-cover plate and one-flange plate)
have been tested up to the time of this writing.
Fig. 1. ABAQUS data for specimen 3
The specimens were instrumented to capture responses at critical cross sections including the ends of the cover plates, beam flanges, beam web, column panel zone, continuity plates, and panel-zone doubler plate. All specimens tested to date have responded in a ductile manner, although web and flange local buckling of the beam immediately beyond the nose of the reinforcing plate has led to a rapid loss of strength and stiffness. Brittle fractures at the beam-column intersection have been avoided. A photograph of specimen 3 at a beam plastic rotation of 3.82-percent radian is shown in figure 2. Spalling of the whitewash paint indicates the extent of flange and web local buckling.
Fig. 2. Photograph of specimen 3 during testing
The analysis of experimental data is currently under way. Beam
plastic rotations greater than 3-percent radian have been recorded
in all five specimens although the loss in strength and stiffness
has been significant, as can be seen in figure 3.
Fig. 3. Hysteretic response for specimen 3
The remaining five specimens consist of four-flange plate and one-cover plate. Data will be published in a PEER report early in 2000. The authors, Shakzod Takhirov, Vitelmo Bertero, and Andrew Whittaker comprise the Berkeley team working on the SAC project.
University of California, Berkeley
University of California, Berkeley