Ground-motion simulation using stochastic finite-fault method combined with a parameter calibration process based on historical seismic data

Accurate simulation of ground motion is an important basis for the seismic design of engineering structures. The stochastic finite-fault method, which takes into account the source, path, and site effects, is comprehensively applied in simulating ground motion. However, the uncertainty in path and site parameters can affect the reliability of the simulation results. Therefore, it is of great significance to accurately determine these parameters. In this study, a parameter calibration process based on historical seismic data was proposed, where the genetic algorithm was adopted and the optimal combination of parameters was obtained through the best fit between the observed and simulated 5%-damped pseudo-spectral acceleration. Based on the 2019 Changning Ms 5.6 earthquake records, the calibrated parameters were obtained. In addition, a model bias analysis was performed and the simulation results were compared with those predicted by the ground motion prediction equations, which verified the effectiveness of the parameter calibration process. Furthermore, the ground motion of Changning Ms 6.0 earthquake was synthesized using the calibrated parameters, and the blind simulation was carried out at Gongxian middle school where ground motion was not recorded. The results show that in the area with complex terrain, the parameters reflecting the geological conditions are obtained through calibration, which forms effective input conditions in the stochastic finite-fault method, so that the ground motion can be well reproduced. Additionally, it also provides a theoretical basis for disaster prevention planning and implementation.