Optimal selection and scaling of ground motion records compatible with input energy and acceleration spectra

Abstract

Nonlinear response history analysis is the primary tool for risk-targeted design and seismic performance evaluation of structures. These analyses require the selection of a set of ground motions that satisfy predetermined conditions such as spectral acceleration. Numerous efforts have been made so far to obtain ground motion records which are expected to represent possible earthquakes. Even though spectral acceleration-based ground motion scaling is a common procedure, recent studies showed that structural response can be better represented through the energy content of the records. To this end, this study aims to develop an energy and acceleration spectra-compatible record selection and scaling methodology to achieve higher efficiency and lower bias in the predicted structural response. The efficiency of the proposed method is evaluated through the standard deviations of the computed story drifts of benchmark structures resulting from the records processed by either the proposed or commonly used methods. The results demonstrated that considering input energy together with spectral acceleration for the selection and scaling of the records can considerably reduce the bias in structural response, especially for structures located on stiff soils.