Denoising And Hydrocarbon Detection Of Seismic Data Based On Modal Decomposition Method

With the development of seismic exploration,the requirements for the quality of the seismic data are getting higher.Since seismic data is often accompanied by noise during the acquisition process,it is necessary to denoise the original seismic data to obtain seismic data with high SNR,and perform hydrocarbon detection on the basis of high signal-to-noise ratio data for the further reservoir prediction.Therefore,based on the decomposition method of VMD modality,this paper decomposes the signal into a series of modal components,eliminates the noise component in the time domain,and achieves the goal of removing noise.At the same time,the VMD decomposition method is combined with HT and generalized S transform respectively.Based on the joint timefrequency analysis method of VMD,spectral decomposition and extraction of attenuation gradient properties are performed in the time-frequency domain to achieve the purpose/goal of hydrocarbon detection.In the direction of denoising processing: this paper starts from the modal decomposition method and compares the modal decomposition methods such as Empirical Mode Decomposition(EMD),Ensemble Empirical Mode Decomposition(EEMD),Complete Ensemble Empirical Mode Decomposition(CEEMD)and VMD.EMD,EEMD and CEEMD are all based on the development of EMD decomposition.The average of the minimum envelopes is obtained for each eigenmode function.This method of modal decomposition is not derived from strict mathematical formulas.but an empirical decomposition form.Meanwhile,due to the action of the interpolation function,there is an end effect at both ends of the signal,which affects the accuracy of modal decomposition.The modal decomposition based on VMD has undergone rigorous mathematical derivation,and its decomposition process is of adaptability and flexibility since the signal can be decomposed into IMF components of a certain frequency band.Compared with the EMD modal decomposition method,the decomposition process is well overcome the difficulties existed before.Modal aliasing problem in.In order to highlight the advantages of VMD decomposition,the CEEMD and VMD methods are applied to denoise the noisy theoretical model and the actual seismic data.Comparing the characteristics of the two decomposition methods,the VMD-based decomposition method can better preserve the useful signal and eliminate the noise interference,thus data processing results with high signal to noise ratio is obtained.In the direction of hydrocarbon detection,the following researches are carried out:(1)The time-frequency analysis methods of short-time Fourier transform,wavelet transform,S transform and generalized S transform are compared respectively.The generalized S transform is compared with the first three time-frequency analysis.The method has higher time-frequency focusing,and a generalized S-transform is better for spectral decomposition.(2)Combining the modal decomposition methods of EMD,EEMD,CEEMD and VMD with Hilbert transform,the time-frequency analysis method based on modal decomposition is obtained.The time-frequency analysis method based on VMD and CEEMD is applied to the theoretical model and the actual seismic data processing respectively.The theoretical model analysis shows that the time-frequency analysis method based on VMD has higher resolution and can accurately describe the time of the signal-frequency change trend;in the application of actual seismic data processing results,the result based on the VMD spectral decomposition processing is much better than that of the CEEMD,since the horizon is more continuous and the resolution of the profile is higher.(3)Considering that the VMD decomposition method can overcome the shortcomings of modal aliasing between multiple frequency components,the VMD and CEEMD decomposition methods are combined with the generalized S transform respectively to extract high frequency and low frequency seismic profiles,according to "low frequency accompanying shadow" and the attenuation gradient property to indicate the hydrocarbon-bearing layer of the study area.By comparing the low frequency accompaniment detection method based on VMD decomposition,the resolution of spectral decomposition is higher,the characteristics of low frequency accompaniment are more distinct,and the VMD decomposition as well as generalized S transformation can obtain continuous attenuation gradient profile to accurately detect oil and gas.