Evaluation of simulated ground motions for seismic assessment of a steel frame structure using multi-criteria selection and scaling approach based on evolutionary algorithms

Abstract

Recently, simulated ground motion records have supplanted real records as a key alternative. Unlike real motions, simulations encompass various intensities, source-to-site distances, and site characteristics. However, determining whether they are appropriate for engineering applications takes time and effort. Another challenging topic is the proper selection and scaling of ground motion records to assess the seismic performance of structures. This study aims to investigate the difference between the real and simulated motions selected according to the code design spectra. Simulations are accomplished through the stochastic finite-fault method, considering the uncertainty of the rupture of the North Tabriz Fault Plane in northwestern Iran, one of the hazardous regions with a seismic gap. A real ground motion dataset with consistent seismological characteristics is compiled. The records are selected and scaled through a multicriteria approach using the evolutionary algorithm. In this approach, the mean spectrum and the (period-depended) dispersion fit best with the target spectrum and its dispersion. Variations in real and simulated record sets are assessed using several ground motion intensity metrics. This study also investigates the effectiveness of the simulated motions for seismic demand evaluation of a three-story steel moment frame.