Does the number of applied ground motions matter on the structural responses?

This study discusses one of the sources of bias in estimating the accurate structural seismic demand, i.e., the number of required response history analyses. Specifically, this study investigates how many ground motions (GMs) are required for seismic demand analysis to obtain a stable and accurate structural response. Unlike the previous studies, a special attention is devoted to using hazard consistent GMs for performing response history analysis and estimating the structural responses. To this end, the structural seismic demand of an 8-story building is evaluated through code- and intensity-based seismic assessment frameworks. Then, the error induced on the seismic demand due to using a finite number of response history analyses is evaluated by comparing the responses of considered buildings subjected to different sets of GMs with a reference structural response. All GMs in this study are selected based on the generalized conditional intensity measure framework, which allows to select hazard consistent GMs. In addition, the reference structural response is obtained using a large number of GMs that appropriately characterize the seismic hazard of the site considered. The results demonstrated that at least 25 hazard consistent GMs are required to be used to accurately and reliably estimate the collapse fragility curves. In addition, unlike the 2016 edition of ASCE/SEI 7, which required to use at least 11 GMs, the results explained that seven well-chosen GMs are sufficient to accurately and stably estimate the structural responses for deign purposes if the average structural response is used for evaluation