A Quantitative Seismic Topographic Effect Prediction Method Based upon BP Neural Network Algorithm and FEM Simulation

作     者：Qifeng Jiang Mianshui Rong Wei Wei Tingting Chen Qifeng Jiang;Mianshui Rong;Wei Wei;Tingting Chen

作者机构：Shandong Seismic Hazard Prevention CenterShandong Earthquake AgencyJinan250014China Shandong Institute of Earthquake EngineeringJinan250021China Key Laboratory of Urban Security and Disaster Engineering of China Ministry of EducationBeijing University of TechnologyBeijing100124China Publicity and Education CenterShandong Earthquake AgencyJinan250014China Shandong Earthquake StationShandong Earthquake AgencyJinan250014China 

出 版 物：《Journal of Earth Science》 (地球科学学刊（英文版）)

年 卷 期：2024年第35卷第4期

页      面：1355-1366页

核心收录：

学科分类：070801[理学-固体地球物理学] 07[理学] 0708[理学-地球物理学] 

基　　金：supported by the National Natural Science Foundation of China(No.51878625) the Collaboratory for the Study of Earthquake Predictability in China Seismic Experimental Site(No.2018YFE0109700) the General Scientific Research Foundation of Shandong Earthquake Agency(No.YB2208) 

主　　题：seismic topographic effect finite element method BP neural network algorithm earthquake disaster prevention 

摘      要：Topography can strongly affect ground motion,and studies of the quantification of hill surfaces’topographic effect are relatively *** this paper,a new quantitative seismic topographic effect prediction method based upon the BP neural network algorithm and three-dimensional finite element method(FEM)was *** FEM simulation results were compared with seismic records and the results show that the PGA and response spectra have a tendency to increase with increasing elevation,but the correlation between PGA amplification factors and slope is not obvious for low *** BP neural network models were established for the prediction of amplification factors of PGA and response *** kinds of input variables’combinations which are convenient to achieve are proposed in this paper for the prediction of amplification factors of PGA and response spectra,*** absolute values of prediction errors can be mostly within 0.1 for PGA amplification factors,and they can be mostly within 0.2 for response spectra’s amplification *** input variables’combination can achieve better prediction performance while the other one has better expandability of the predictive ***,the BP models only employ one hidden layer with about a hundred nodes,which makes it efficient for training.