Seismic Signal Time-frequency Analysis And Its Application To Blast Distinguishing

In practical earthquake observation, it is very important for earthquake analysis and prediction to distinguish earthquake with blast. There are a lot of methods to distinguish blast and earthquake, such as wave comparison, relation between amplitude and period, spectrum, P wave first motion, amplitude ratio, coda reduction, etc. In these methods, the seismic wave was regarded as one type of stationary signal and was analyzed with Fourier transform. With the development of knowledge, seismic signal was regarded as one type of nonstationary signal and the Fourier transform could not describe the nonstationary characteristic of seismic signal. It is very difficult to analyze the nonstationary characteristic of seismic signal. Modernized digital seismograph provides a lot of digital seismic wave with high sampling rate, high resolving power, and broadband. That makes it convenient to analyze seismic signal with time-frequency analysis method.In this article, the theory of Fourier analysis, wavelet analysis was introduced, and the merits of time-frequency analysis were elaborated. The characteristic of weakening the cross distribution of three kind of time-frequency presentation, which are Born-Jordan, Choi-Williams, Zhao-Atlas-Marks time-frequency presentation, is discussed in this article. In conclusion, Zhao-Atlas-Marks time-frequency presentation has weakest cross distribution, and is most suitable for analyzing earthquake waveform signal. 11 blasts and 40 earthquakes are single out to time-frequency analyze, and Zhao-Atlas-Marks time-frequency presentation of these events are obtained. At last, the difference of Zhao-Atlas-Marks time-frequency presentation between blast and earthquake was compared. There are four differences as follows:1. Average frequency of earthquake's P wave is larger than that of blast;2. Average frequency of earthquake's S wave is larger than that of blast,...