Simulating Conditional Ground Motions for Time-History Analyses

Session: Earthquake Ground Motion and Impacts [Poster]
Type: Poster
Date: 4/30/2020
Time: 08:00 AM
Room: Ballroom
Description: 

In time-history analyses of structures, input ground motions not only should meet specific level of a selected intensity measure, but also must be consistent with scenarios (magnitude and distance) compatible with the site’s hazard. In this work, a conditional ground motion simulation approach is proposed to produce input ground motions for time history analyses. The essence of this method is based on sampling from the joint distributions of Fourier amplitude and duration conditioned on a scenario and an intensity measure (e.g., spectral acceleration) level. We use Bayes’ theorem to produce the updated distribution starting from a prior obtained from ground motion models (i.e., ground motion prediction equations) for Fourier amplitude and duration. The prior is updated based on the desired level of the selected Intensity measure. The sampling is done through the Metropolis Hasting algorithm. The necessary input for simulating conditional ground motions are Fourier amplitude Ground Motion Prediction Equations, the correlation between Fourier amplitude of different frequencies and the correlation between each frequency Fourier amplitude and ground motion duration. Such models have been developed by the authors using the Japanese Kik-Net ground motion database.

Presenting Author: Mahdi Bahrampouri

Authors