Research And Application Of Iterative Inverse Spectral Decomposition Based On Shape Regularization

In the field of seismic data analysis and interpretation,spectrum decomposition technology has always played a very important role.Spectral decomposition can transform the signal from a single time domain to a time-frequency domain,which provides the possibility to dig deep into the seismic information hidden in the frequency domain with the stratigraphic,structural and lithologic information.With the increasing difficulty of seismic exploration,the time-frequency resolution and analysis accuracy of traditional time-frequency analysis methods have been difficult to meet the increasing demand for seismic exploration.Therefore,exploring spectral decomposition methods with better spectral decomposition performance and adapting to the requirements of high-resolution and high-precision in seismic exploration in the new era not only gives full play to the advantages of spectral decomposition methods in analyzing and processing seismic data,but also acquires new seismic exploration requirements.The inevitable requirement for breakthrough.This paper reviews the method principles of several conventional time-frequency analysis methods,including Short-Time Fourier Transform(STFT),Continuous Wavelet Transform(CWT),S and Generalized S-Transform,Wigner-Ville Distribution and Matching Pursuit(MP).The principle of spectrum decomposition of the above methods,physical meaning of the parameters,advantages and disadvantages of time-frequency decomposition are reviewed in this part.Based on the former introduction,a new inversion spectral decomposition method is introduced--Iterative inversion spectral decomposition method based on shaping regularization.First at all,starting from the principle of geophysical inversion and shaping regularization,the algorithm flow of the method is deduced.Based on the spectral decomposition characteristics,an adaptive shaping operator and a step factor that can control the convergence speed of the iteration are constructed.Then,the influence of the relevant parameters of the algorithm on the time-frequency analysis results is explored,such as the number of iterations,the value of the step size,the type and length of the time window,and other factors.It is found that: the number of iterations mainly controls the frequency resolution,the value of the step size can control the velocity of convergence and the length ofthe time window has a greater influence on the time resolution;a good time-frequency decomposition result is a comprehensive reflection of various parameters.Next,this paper construct analog signals and synthetic seismograms with different variations to test the time-frequency decomposition performance of this method compared to traditional spectral decomposition methods.Time-frequency decomposition performances mainly include time-frequency resolution,amplitude aggregation,and noise immunity ability in amplitude spectrum,and application of phase spectrum in signal interpretation.The test results show that the iterative inversion spectral decomposition method based on shaping regularization has better time-frequency resolving power and anti-noise performance than continuous wavelet transform and generalized S-transform,and this method can improve frequency resolution while maintaining time resolution.What's more,the abnormal point in the phase spectrum is also closely related to the position of the demarcation point of the wavelet in the signal.Finally,the iterative inversion spectral decomposition method based on shaping regularization is applied to the spectrum analysis of gas-bearing reservoirs.The results of the well-by-channel spectral decomposition and the results of the frequency-divided sections show that the method accurately reveals the information related to the location and boundary of gas-bearing reservoirs and has good practical application effects.