Method for picking arrival time of seismic phase based on LSTM (Long Short Term Memory) recurrent neural network

Abstract

The invention discloses a method for picking an arrival time of a seismic phase based on an LSTM (Long Short Term Memory) recurrent neural network. The method comprises the following steps: (1) acquiring original seismic waveform data, performing cutoff processing on a waveform, and outputting equilong waveform data comprising a P (Primary) wave and an S (Secondary) wave; (2) preprocessing waveform data in a data set, and then dividing the data set into a training data set and a testing data set; (3) constructing a structure of the LSTM recurrent neural network; (4) training an LSTM recurrentneural network model and testing a trained model by using the testing data set, wherein when a testing result meets an accuracy requirement, the training is completed; and (5) deploying a trained LSTMrecurrent neural network model in a waveform analysis system, analyzing seismic waveform data, and picking an arrival time of the P wave and an arrival time of the S wave. By adopting the technical solutions provided by the invention, the anti-noise performance is good, the picking for the arrival time of the P wave and the arrival time of the S wave is excellent, and the method has very good technological value and application prospect.

Description

technical field [0001] The invention relates to the technical field of information processing, in particular to a method for picking up the arrival time of seismic phases based on an LSTM cyclic neural network. Background technique [0002] When an earthquake occurs, various seismic phases are produced, the most critical phases are longitudinal wave (P wave) and shear wave (S wave). In seismic monitoring, picking up the arrival time of P-wave and S-wave is the key link of seismic source location and seismic phase identification technology. Through real-time analysis of the waveform data collected by the station's seismometer, seismic events are detected, and the arrival time of P-wave and S-wave is picked up, combined with other technical means, the staff can determine the source location and magnitude, and report it to the superior department in time. and notify other departments. The traditional seismic phase arrival time picking method mostly extracts features from the ...

AI Technical Summary

Problems solved by technology

In addition, the stations are distributed in different areas, and the types of instruments are various. The collected waveform data is mixed with noise and redundant information, which brings great challenges to the work of seismic phase arrival. time solution

Images