Research On High Resolution Time-frequency Analysis Method For Seismic Signal

In the analysis of non-stationary signals,as a popular signal processing method,time-frequency analysis is no longer limited to a single time-domain or frequency-domain analysis.Its task is to describe how the spectral component of the signal changes in time,so as to establish a time-frequency joint distribution that shows the signal energy in time and frequency at the same time.For the analysis and research of seismic signals,highresolution time-frequency analysis method is more conducive to the exploration of geophysical fields such as formation identification and oil and gas exploration.Aiming at the problems of low time-frequency resolution and poor energy aggregation when processing seismic signals,three modified high-resolution time-frequency analysis methods suitable for seismic signal processing are proposed in this paper.The main research contents are as follows:1.Aiming at the defects of fixed window function and low resolution of traditional short-time Fourier transform,a high-order adaptive synchrosqueezing short-time Fourier transform is proposed.By introducing window parameters,the time-frequency aggregation degree corresponding to different parameters is calculated,and the timefrequency window parameters with the highest aggregation are selected according to the signal characteristics.On this basis,high-order synchrosqueezing transformation is carried out to further improve the time-frequency resolution of seismic signal analysis.2.Aiming at the failure of the instantaneous frequency estimation in the traditional time-frequency synchronous compression transform to the seismic signal,a new secondorder synchrosqueezing wavelet transform algorithm is proposed based on the rake wavelet as the mathematical model.Firstly,a wavelet base suitable for rake wavelet is proposed,and then the improved wavelet base is used to match the seismic signal,and then the spectral peaks are aligned to complete the correction of the reference frequency.Combined with the second-order synchrosqueezing transform,the time-frequency resolution and time-frequency energy aggregation are improved.3.Aiming at the problems of fixed time window shape,single time-frequency resolution and poor energy aggregation of standard S-transform,a new adaptive generalized S-transform algorithm is proposed.A new type of generalized Gaussian window function is designed.The new parameters are nonlinearly constrained with the window shape as the constraint condition,and then the time-frequency aggregation degree is taken as the objective function of parameter optimization to obtain the optimal result,so as to obtain the best time-frequency distribution.4.Combined with the above research results,it is applied to thin interbed identification and seismic spectrum decomposition.Firstly,two thin layer models,wedgeshaped thin layer and horizontal thin layer,are established,and three improved timefrequency analysis methods are used for identification.The analysis results show that this method can clearly distinguish the formation interface even when the formation is relatively thin.In the field of seismic spectrum decomposition,natural seismic wave data are selected and the method in this paper is applied to time-frequency analysis.Compared with traditional methods,the resolution of this method is higher.The experimental results prove the effectiveness and superiority of this method in improving resolution and aggregation in time-frequency analysis.