Probability-deterministic broadband ground motion simulation method considering seismic hazard and physical processes

Abstract

Near-fault seismic data are rare yet vital for site-specific analyses, but traditional methods struggle to capture their unique features. This study proposes a novel seismic ground motion prediction method that integrates the advantages of both probabilistic and deterministic approaches. This method incorporates the target structure's natural vibration period and evaluates all possible seismic risks within the potential seismic source zone rather than limiting the analysis to seismic response patterns under assumed scenarios. This approach establishes a series of probabilistically defined seismic events. A broadband hybrid seismic source model, which accounts for high-frequency random scattering, is introduced. The deterministic physics-based broadband hybrid method for seismic motion simulation simulates the low-frequency component (≤1 Hz) using the Spectral Element Method (SEM) and models the high-frequency component (>1 Hz) using a modified three-dimensional Stochastic Finite Fault Method (3D-EXSIM). The resulting output comprises a series of probabilistically significant broadband ground motion simulations designed for the actual site with varying return periods. The method is employed as a case study to analyze a rare earthquake scenario for a subway station (with a natural vibration period T = 1.0 s) in the Binhai New Area, Tianjin, to clarify and validate the rationality and advancement of this method for the seismic fortification of critical building structures in near-fault regions. The results demonstrated that this method is more reliable than the artificial ground motion synthesis model using the Uniform Hazard Spectrum (UHS) and effectively captures significant near-fault effects. This method is valuable for developing seismic codes and achieving urban planning and infrastructure protection in active fault areas.