Research On The Relationship Between Peak Seismic Characteristic Parameter And The Seismic Early P Wave

Because of the great significance of earthquake prediction in disaster prevention, it is valued by the majority of researchers in the current seismic studies. After station detected the relevant parameters of seismic waves, the value of them can be predicted by computing when they reach the warning area and this useful to make earthquake prediction and reduce casualties. However, the relationship between the parameters of early seismic waves and follow-up peak is no clear, resulting in difficulty to establish reasonable and effective prediction model.Artificial neural network has obvious advantages to solve complex network and it is widely used in various fields. So in the face of this ambiguity of the relational model established, artificial neural network technology can provide a better solution. This paper introduced the approach of artificial neural network to predict characteristic parameters of the earthquake and applied MATLAB to simulation study in order to get the relationship between the seismic waves early data and seismic wave peak.The core content of this article is to use the parameters of the early characteristics of the seismic waves to predict the peak ground acceleration of seismic waves. Summarized previous methods in study of the magnitude forecast, this article used the early peak acceleration, peak velocity, peak displacement, and excellent cycle peak of P wave as the basis of the prediction and by neural network to establish relationship between them and the peak acceleration. And it also used principal component analysis to filter parameters, in order to achieve the objective of dimensionality reduction and made a comparative analysis between the predicted results of dimensionality reduction parameters and early results.Seismic data simulation and analysis is achieved by BP neural network model which is established by MATLAB. The analysis shows that the result of earthquake prediction is more accurate when combined amplitude parameters and cycle parameters. After principal component analysis, it determined the most three informative impact parameter:early peak acceleration, peak velocity, peak displacement. The result showed that it was worse than forecast with five parameters.