| During the 10 th Five-Year Plan period,the National Seismological Network was fully digitalized,so far,data storage is abundant and convenient.The JCZ-1 ultra-wideband seismograph was developed by Cai Yaxian’s team.The observation bandwidth is wider than that of the CTS-1 Very Wideband Seismometer,which is 50Hz-120 s.It reaches 50Hz-DC and can record signals with periods greater than 120 s.The number of JCZ-1 seismographs working at stations across the country has reached 16 and continuously records all day ground vibration signals.We can use abundant signal data to explore the connection between daily ground vibration changes and earthquake precursors.A large number of algorithms for processing and analyzing data have been developed on the basis of the Fourier transform,and one of them is the time-frequency analysis method.The Fourier transform is a single frequency domain analysis.When the data is analyzed by the time-frequency method,this one-dimensional space is expanded.It is two-dimensional space,time-frequency domain.Earthquakes are caused by physical changes under the earth’s crust,so earthquake signals excited by this change must be a typical non-stationary signal.In this paper,Fourier transform and Morlet wavelet transform are introduced at first,four groups of experiments are designed,and the advantages and disadvantages of the two methods in non-stationary signal processing are compared.It can be concluded that the Morlet wavelet transform is obviously more suitable for seismic signals.in conclusion.This thesis designs a seismic signal processing algorithm based on Morlet wavelet transform which was used to analyze the Wenchuan earthquake in 2008 and earthquakes with a magnitude greater than 7.0 in the Middle East from 2010 to 2016.That is,the Yushu M7.1 earthquake in 2010,the Lushan M7.0 earthquake in 2013,and the Yutian M7.3 earthquake in 2014.Before the Wenchuan earthquake,the energy amplitudes of the two frequency bands 0.1-0.18 Hz and 0.2-0.25 Hz increased.Before the Yushu earthquake in Qinghai Province,the energy amplitudes in the 0.2-0.3Hz frequency band at most stations have increased,and all began at the day before the earthquake,there is also a slight increase in energy amplitude in the 0.1-0.2Hz band.Prior to the Lushan earthquake in Sichuan,the energy amplitude in the 0.2-0.3Hz frequency band increased at some stations.Before the Yutian earthquake in Xinjiang,six stations located at the nearest epicentral epicenter experienced an increase in energy amplitude in the 0.1-0.2 Hz frequency band,and the 0.2-0.3 Hz frequency band increase appeared at almost all stations.The time-frequency results of the Wenchuan earthquake are in good agreement with Hu Xiaogang’s research findings,the time-frequency characteristics of the Yushu,Lushan and Yutian earthquakes are similar.However,there is no uniform feature in the Fourier transform spectrogram.For example,the Yushu earthquake only fluctuates in the 0.1 to 0.2 Hz band at the Chengdu station in Sichuan,and the Yutian earthquake only exhibits extreme values in the 0.1 to 0.2 Hz range at the Golmud station in Qinghai.In summary,this paper uses Morlet wavelet transform to process the JCZ-1 ultra-wideband seismometer raw data,counts the number of earthquakes above Ms7.0 magnitude occurred in mainland China since 2008,analyzes the pre-earthquake data of these large earthquakes recorded by 16 domestic JCZ-1 seismographs separately.The trigger mechanism of different earthquakes,the fault zone where the epicenter is located,and the form of internal movement are all different,each earthquake has its own and different pre-earthquake time-frequency characteristics.This paper transforms the original data,analyzes the time spectrum,and draws useful conclusions.It shows that the transform method used is a feasible method for statistical analysis of time-frequency features before the earthquake. |
PDF Full Text Request