Seismic assessment of frame structures subjected to ground motions selected based on the conditional mean spectrum

Abstract

The nonlinear response history analysis has become an essential tool for assessing the performance of a structure subjected to seismic shaking. To perform this type of analysis, a set of recorded ground motions representative of the seismic conditions at the site are usually selected, scaled and/or modified to later be used as uniform base excitation in a structural model. The uniform hazard spectrum, and more recently, the uniform risk spectrum has been specified as the target spectrum for ground motion selection and scaling in design codes. These spectra are considered a conservative and unrealistic target because their spectral shape is not representative of any individual ground motion, as they represent the envelope of all possible earthquake scenarios that went into a seismic hazard evaluation. The Conditional Mean Spectrum (CMS) appears in 2006 as an alternative to the probabilistic spectra, with a spectral shape that resembles those of realistic ground motions. For the analysis of a structure, the number of conditional periods, T*, used for defining the CMS has been a topic of continuous research because it is desired to run as few as possible while still capturing the widest range of structural response. To understand the impact of T* on the seismic structural behavior, this investigation studies the distribution of several engineering demand parameters (EDP) of ductile RC-frame structures, analyzed under three seismic intensity levels, using 3168 ground motions which resulted from evaluating 15 conditioning periods. The results confirm that the conditional period value used for the ground motion selection has a significant impact on the seismic response of the structure.