摘要
基于BP神经网络算法,选择山东地区天然与非天然地震事件为研究对象,选取P波初动方向、P波初动振幅与S波最大振幅比以及P波最大振幅与S波最大振幅比作为神经网络输入元,构建合理的网络模型,搜集大量的地震样本数据进行训练,实现对山东地区天然与非天然地震事件的识别。震例预测结果显示,P波初动方向和振幅比可以作为识别天然地震和非天然地震主要依据,BP神经网络方法具有对地震事件类型识别的可行性。
Based on the BP neural network algorithm,natural and non-natural seismic events in Shandong region are selected as the research objects,and the initial motion direction of P wave,the maximum amplitude ratio of P wave initial motion amplitude to S wave,and the maximum amplitude ratio of P wave to S wave are selected as the input elements of the neural network.We constructed reasonable network model and collected a large number of training sample data to identify the natural earthquakes and non-natural earthquakes events in Shandong.Prediction results of earthquake examples show that the initial direction and amplitude ratio can be used as the main basis for identifying natural and non-natural earthquakes.The BP neural network method is a feasible method for identifying earthquake events types.
作者
刘方斌
曲均浩
张志慧
Liu Fangbin;Qu Junhao;Zhang Zhihui(Earthquake Administration of Shandong Province,Jinan 250102,Shandong,China)
出处
《计算机应用与软件》
北大核心
2020年第1期106-109,185,共5页
Computer Applications and Software
基金
山东省地震局科研基金项目(JJ1802Y)
山东省地震局科技创新团队基金项目(SDST-07-2018)
关键词
BP神经网络
天然地震
非天然地震
识别
BP neural network
Natural earthquakes
Non-natural earthquakes
Identification