Stochastic Modeling and Synthesis of Near-Fault Forward-Directivity Ground Motions

Abstract

The stochastic modeling and synthesis of near-fault forward-directivity ground motions for the prescribed seismic scenario is presented in this study. This stochastic model, which is established in the orientation of the strongest pulse, combines the directivity pulse component represented by the Gabor wavelet pulse model and the high-frequency component characterized by the modulated filtered white noise model. In this regard, the pulse feature, evolutionary intensity and frequency contents are reasonably considered. Then, the empirical predictive equations that relate the parameters of the proposed stochastic model to the variables that describe earthquake source, path, site conditions, and near-fault rupture-to-source geometry are developed based on both regression and correlation analysis. The basic procedure for the synthesis of possible forward-directivity ground motions under the prescribed seismic scenario is provided as well. The efficacy of the stochastic model and the simulation procedure is confirmed by comparing the simulated motions with the recorded ones in terms of time-series and response spectra values, which indicate that the synthetic ground motions are capable of reproducing the salient characteristics, as well as the natural variability, of the recorded ones. Therefore, the proposed stochastic modeling and simulation method is expected to be a desirable alternative and complement for recorded ground motions in the near-fault seismic risk assessment.