Predicting the Time Lag between Primary and Secondary Waves for Earthquakes using Artificial Neural Network

Abstract

Seismic waves experienced prior to earthquake are the primary and the secondary waves. This paper investigates the time lag after the primary wave before the occurrence of the secondary (destructive) wave. The aim is to allow for necessary warning signals and safety steps to be taken prior to the impending disaster. Seismometer records from previous earthquakes were used in this investigation, putting into consideration the time lag between the primary and secondary waves. Other parameters considered include the magnitude, the epicenter distance, the seismic station‘s distance and the direction (in azimuths). Consequently, a prediction model was developed from the derived data using Artificial Neural Network (ANN). Data obtained from earthquakes of magnitude 6.0 to 7.0, based on Richter‘s scale, was used to train the ANN. The results therein showed high performance, with regression values greater than 0.9 and root mean squared errors of 0.1003-0.1148 for the most satisfactory architecture. The final results showed that the developed ANN model achieved a high performance, hence, adequate for this type of application.