Identifying Explosion Event By Seismic Wave Energy Attenuation Rate

With the development of society and economy, the level of economic activity continuously improved,the annual occurrences of explosion are more and more frequent every year. If the seismic signals data being observed in seismological observatories are not appropriately processed, explosion events are very easily as regarded as earthquake events and vice versa. The correct classification of the earthquake and explosion is very important not only to the researches of seismology but also to the identifying of suspect dangerous explosion. First of all, this article describes the earthquake data formats which often to be used and the acquisition method of the public earthquake waveform data. There are several acquisition methods to the public earthquake waveform data:Obtaining earthquake data by China Earthquake Data Center; By the Incorporated Research Institutions for Seismology(IRIS) Data Management Center (DMC) obtaining seismic data; And other ways to get data through the International Seismological Centre (ISC). Introduction of the seismic data formats which often used, mainly on the SAC format, SEED format and SEGY format. Then we explained the relationship between the wave count value and the seismic displacement. In the third chapter, we explained in detail the theory and method of the starting point of the waveform signal, and use the EMD (Empirical Mode Decomposition)and improved STA/LTA(short-term to long-term average) method to determined and give experiments results icon. The experimental results achieve the desired effect.This article is based on extracting the characteristics of natural earthquakes and artificial blasting waveforms to be studied. First, we extracted valid waveform for the data which come from35natural earthquakes of Beijing surrounding areas and27times artificial blasting event, then propose two new features:the largest S-wave attenuation rate and the average attenuation rate, that we use the two characteristics to identify the natural earthquake and artificial blasting events. We used the tool of identification is the least squares support vector machine (LS-SVM), the result is: using the characteristics of the largest S-wave attenuation rate and Corner frequency to recognize was89.33%, the characteristics of the average attenuation rate and Corner frequency to recognize was90.33%, the characteristics of the largest S-wave attenuation rate the average attenuation rate to recognize was84%, all of the three characteristics to recognize was92.33%.Then we achieved the recognition rate of91.5%through use of multi-core learning methods and use the characteristics of the average attenuation rate and Comer frequency. Therefore, these two characteristics can be well applied to recognize the natural earthquakes and artificial blasting events.