Seismic Characterization of Steel Buildings Subjected to Megathrust Earthquakes

While most of the current research on the seismic performance of Steel Special Resisting Frames (SSRFs) focuses on seismic hazard due to shallow crustal earthquakes, research on the response of SSRFs subjected to megathrust earthquakes is generally neglected. This study describes a research effort aimed at shedding needed light on the behavior of Steel Special Moment Frames (SSMFs) at locations where the seismic hazard is due mainly to subduction-type earthquakes, such as Chile, Japan or the Pacific Northwest of the U.S. Recent studies have shown that accurate estimations of seismic demands require a vector-valued ground motion intensity measure (IM) to characterize the seismic hazard. Hence, the vector-valued IM considered in this study is comprised of the spectral acceleration at the fundamental period of the structure (Sa(T1)), the shape of the ground motion spectrum in the vicinity of Sa(T1) (SaRatio), and the ground motion significant duration (DS5-95). A set of 40 SSMFs designed per the latest US seismic design regulations is considered. The number of stories ranges from 2 to 20. Further, two sets of records (two horizontal components per record) are considered: 22 records due to shallow crustal earthquakes (based on the FEMA P695 Far-Field record set) and 22 records due to megathrust subduction earthquakes. The latter were selected in accordance with the FEMA P695 selection criteria. 2D models (as suggested by ATC-72) of the archetypes are subjected to Incremental Dynamics Analysis. Using a hazard-consistent framework for seismic demand assessment, the collapse capacities of the archetypes under both shallow crustal and megathrust subduction earthquakes are characterized, and significant differences are found. The results of this study are relevant because the calibration of current seismic design codes aims at providing system- and component-level force and displacement limit states consistent with a uniform probability of collapse equal to 1% in 50 years. Such uniformity is not possible unless the calibration process accounts for the
location-dependent correct type of hazard source.