地震波データの自動検側方式とオンライン処理システムにおける稼動実験

概要

論文の詳細を見る
In a number of cases, the time series of seismic wave data under study may be considered to be stationary in the wide sense, at least within certain time intervals. A computer based system has been developed for the automatic measurement of the arrival times of seismic signals. The parametric models of an autoregressive type are applied to the small sections of ground noise and seismic wave signals containing the P and S and unidentified X phases of microearthquakes. In measuring the arrival time of the P waves, autoregressive (AR) models are given for two separated sections, one for ground noise section before the P waves and the other for the section containing the P wave signals. According to the definition of the AR model, the prediction error calculated for the given model is represented by the normal distribution. Therefore, the maximum likelihood estimate associated with each of the two models provides an effective method for the measurement of the P arrival time. As a practical application of the method, the Akaike's Information Criterion (AIC) is introduced. An extensive application of the method for the measurement of the arrival times of S and unidentified phases has proved to give satisfactory results. Experiments using the on-line system linked with the telemetering notwork of microearthquake observation in the Kanto district show arrival time measurements almost exactly the same as those measured by experienced interpreters, giving the P arrival times within the error of ±0.01 sec for about 95% of the detected events for P and within the error of ±0.1 sec for 90% for S. The time required for the determination of arrival time for each chanel is about one second.
1982-01-08

地震波データの自動検側方式とオンライン処理システムにおける稼動実験

スポンサーリンク

概要

著者

関連論文

スポンサーリンク