Experimental and analytical investigation of panel zone behavior in steel moment frames

Abstract

Steel moment resisting frames (MRFs) are one of the most commonly used lateral force resisting system for steel building construction located in regions of high seismic risk. Although steel moment frames were studied extensively following the 1994 Northridge Earthquake, one critical design issue is not yet completely understood: the role of the panel zone. Recent U.S. building codes have significantly increased the required shear strength of the panel zone for seismic applications. This leads to increased column sizes or the need for costly doubler plates. A number of past research studies showed that shear yielding of panel zones results in highly ductile behavior and provides a major contribution to frame deformation as an excellent source of energy dissipation. However, these same studies suggested that large panel zone shear deformation can cause high strain concentrations near the beam flange groove welds and can cause premature fracture at the joint. The overall goal of this research is to provide additional data to help answer the question: how much panel zone participation should be permitted in the inelastic seismic response of a steel moment frame? Despite a number of past studies on this issue, there are sharply conflicting views of how panel zones should be treated in design, both within the research community as well as within the building regulatory community. At the crux of the disagreements are concerns regarding fracture induced by panel zone yielding. There appears to be broad agreement that panel zone yielding is a highly ductile process. However, there is broad disagreement on the role that panel zone yielding plays in joint fracture. While there have been a number of past experimental studies on moment frame subassemblies with weak panel zones, the availability of data on large-scale specimens constructed using current US connection detailing and welding practices is very scarce. Further, the experimental database lacks critical data on panel zone behavior in deep columns and for columns subjected to significant axial forces. The primary focus of this dissertation is to provide much needed large-scale experimental data on the cyclic loading performance of steel moment frame joints with weak panel zones. The experimental study was supplemented by finite element analysis of the test specimens, intended to provide additional insights into the response of the test specimens. Cyclic loading tests were conducted on ten large-scale interior steel moment connections to study the seismic performance of the connections. The key variables for the tests were: panel zone strength; beam and column size; beam-to-column connection detail; and column axial stress. Nine of the ten test specimens performed well and met the acceptance criteria of 0.04 radian story drift angle for special moment frames in the current AISC Seismic Provisions. One specimen failed by fracture of the column flange just prior to achieving 0.04 radian story drift angle. In the experimental program, the weak panel zone specimens showed excellent performance, developing drift capacities as large as or larger than the strong panel zone specimens. The weak panel zone specimens exhibited less beam buckling and therefore less strength degradation than the strong panel zone specimens. Shear yielding of the panel zones in these specimens showed highly ductile behavior with stable hysteretic loops. The panel zone was an excellent energy dissipater. The results of this test program showed that acceptable performance of moment frame joints can be achieved over a range of panel zone strengths, varying from very weak to very strong. The experimental results also show that joints with weak panel zones can achieve very large inter-story drift angles prior to fracture. These experimental results suggest that current U.S. code requirements for panel zone strength merit reevaluation. Finite element models were developed of the test specimens, including a global model using shell elements and a local joint model using solid elements. The overall load deflection response of the test specimens predicted by the global models showed good agreement with the experimental results. Local joint models were developed to examine the potential for ductile fracture initiation using the Rupture Index. These studies suggested that the weak panel zone specimens should be far more prone to fracture than the strong panel zone specimens. This prediction, however, was inconsistent with the experimental observations. This suggest that using the Rupture Index as an indicator of the potential for ductile fracture initiation when comparing differing moment frame joint design options, as has been done by previous researchers, should be approached with caution.