Adaptive Time-frequency Analysis Method And Its Application In Seismic Signal Analysis

At present,most signals in nature are non-stationary signals,and time-frequency analysis is one of the powerful means to deal with non-stationary signals.The relationship between time and frequency of signals can be seen from the results of time-frequency analysis,and the strength of signal energy can also be reflected.In this paper,the analysis and processing of non-stationary signal is taken as the research topic,and the adaptive time-frequency analysis method and its application in seismic signal analysis are studied emphatically.The main research content is divided into the following aspects:1.The method of signal noise reduction is studied.When the non-stationary signal is affected by noise,the result of time-frequency analysis is often not ideal,which is not convenient to analyze the internal characteristics of the signal.According to the traditional least mean square algorithm with fixed step size,it is difficult to take into account the small steady-state error and high convergence rate,a variable step size least mean square algorithm is proposed to denoise the signal.2.A novel adaptive generalized S-transform algorithm is proposed to solve the problems of fixed time window shape,single time-frequency resolution and poor energy aggregation of standard S-transform.A novel generalized Gaussian window function is designed,which takes the window shape as the constraint condition to carry out nonlinear constraints on the parameters,and then takes the time-frequency aggregation degree as the objective function of parameter optimization to obtain the optimal result,and further obtain the time-frequency distribution with the best effect.Finally,the time-frequency ridge extraction method and intrinsic chirp component decomposition method are combined to further improve the performance of timefrequency resolution.3.Aiming at the problem of cross interference items in non-stationary signal analysis by WVD method with fixed kernel function,a new local adaptive directional kernel reiteration method based on WVD is proposed to weaken the influence of cross interference items.The improved adaptive transform kernel method is used to analyze the time and frequency of the signal after noise reduction.Meanwhile,according to the measured value of concentration as the stopping condition,the energy aggregation of the signal is improved through iteration.4.Based on the above research results,analyze the Rayleigh wavelet signal model and apply it to the identification of thin interbeds and seismic spectral decomposition.In the analysis of the Reyker wavelet,the two improved methods can obtain better Time-frequency analysis results,and can identify the signal characteristics well.Two improved Time-frequency analysis methods are used to identify wedge-shaped thin layers.The simulation results show that the improved method can still clearly distinguish the stratum interface even when the stratum is relatively thin.Natural Seismic wave data are selected for analysis.Compared with traditional methods,the resolution of the improved method is higher.Simulation results show that the improved method can effectively improve the time-frequency resolution and energy aggregation in Time-frequency analysis.A seismic signal processing software was designed using the App Designer development platform of MATLAB software.Time-frequency analysis,signal pre-processing,time-domain analysis and other functions can be used by importing the rake wavelet signal.By comparing the Time-frequency analysis results of signals with different methods,the characteristics of signals can be observed intuitively.