Study And Application Of High Resolution Processing New Method For Shale Gas Reservoir Seismic Data

In order to effectively improve the signal-to-noise ratio and resolution of shale gas seismic data,high-resolution and high-precision multi-wave joint high-quality shale prediction is achieved.This dissertation has studied the new method of the high resolution processing for shale gas reservoir seismic data.The main research work includes the following contents:1.Firstly,because the wavelet transform has the characteristics of infinite classification,and the fractional transform has the characteristics of arbitrary angle rotation,in this paper,a method of signal-noise separation based on coherent noise model in the wavelet fractional domain is proposed.This method can solve the difficulty of the signal-noise coupling separation in a single shot.Secondly,according to the principle that the effective signal of post-stack data has good coherence,a coherent signal enhancement method based on adaptive gradual mixing is proposed.This method fits the pure data using a small time window,and use adaptive gradual mixed method in a large time window to improve the quality of the coherent enhanced profile.2.According to the principle of high order cumulant and the characteristics of signal purity spectrum,a mixed-phase wavelet deconvolution method with signal purity spectrum constraint is proposed.This method uses multi-channel superposition iterative algorithm to obtain high-precision mixed-phase wavelet,which can solve the problem of multi-solution of phase unwrapping.At the same time,the signal purity spectrum is calculated by using the adjacent channel cross-correlation power spectrum algorithm,which can solve the difficulties in estimating the accuracy of non-horizontal medium data.The mixed-phase wavelet deconvolution with purity spectrum constraint can improve the resolution of single shot data and effectively control the pollution of high frequency noise.3.Firstly,a new method of time-variant Colored spectrum extension driven by well-logging signals is proposed based on the principle of conventional spectral whitening and blue-filter deconvolution.This method uses interactive,time-varying and real-time processing method to extend middle and high frequencies,and uses the well-logging informations to calculate the blue compensation factor automatically so as to improve real-time quality control and fidelity of spectral extension processing data.Secondly,according to the characteristic that the amplitude spectrum of the real data is more consistent with the Gaussian distribution,a Gaussian spectral simulation deconvolution method is proposed to preserve the low frequency information.4.Firstly,in order to improve the anti-noise of sparse blind convolution,a high-order statistical constraint sparse deconvolution method is proposed.This method uses the high-precision mixed-phase wavelet as a priori constraint to iteratively obtain the reflection coefficient with higher accuracy,which is beneficial to improve the signal-to-noise ratio of high resolution wideband recovery profile.Secondly,in order to overcome the bottleneck constrained by the frequency range of the data itself in the conventional high-resolution deconvolution processing,a method of frequency spectrum inversion based on model impedance and sparse double constraints is proposed in this paper.This method uses logging impedance and sparsity constraints to build a new objective function,which is beneficial to the iterative precision of alternating direction algorithm,and the high-resolution broadband restoration profile enhances the resolution of thin layers.5.Using the method of previous chapters to improve the signal-to-noise ratio and resolution of multi-wave data,then this paper takes multi-wave data in real area as an example,through the analysis of rock physics and seismic characteristics,with the use of PP-wave and PS-wave velocity ratio,the high-precision prediction of high quality shale thickness and prediction of reservoir characteristics are carried out to achieve the goal of high-resolution and high-quality shale prediction.