Seismic anisotropy gradient inversion method and system based on sparse representation

Abstract

The invention discloses a seismic anisotropy gradient inversion method and system based on sparse representation. The method comprises the steps of obtaining an intermediate variable and an anisotropy gradient according to logging data; establishing a training set according to the intermediate variable, and learning a sparse dictionary corresponding to the intermediate variable from the training set through a K-SVD method; performing interpolation processing on the intermediate variable, and establishing an initial model of anisotropic gradient inversion in an anisotropic medium; establishing an objective function according to the anisotropy gradient theory and the sparse representation theory, wherein it is assumed that the objective function is in the same work area and the stratum has continuity in the transverse direction, and sparse representation on the sparse dictionary corresponding to the intermediate variable serves as a constraint condition; and taking the initial model as an initial value, conducting inversion calculation through the objective function, obtaining the inversion result of the intermediate variable, obtaining an inversion result of the anisotropy gradient according to the relation between the anisotropy gradient and the intermediate variable, wherein the inversion result tends to logging data with higher resolution and is higher in accuracy.

Description

technical field [0001] The invention belongs to the field of seismic data interpretation, in particular to a seismic anisotropy gradient inversion method and system based on sparse representation. Background technique [0002] Seismic anisotropy gradient inversion is to use the pre-stack azimuth gather seismic data of surface observations, and constrain the known geological laws and drilling and logging data to image the spatial structure and anisotropy characteristics of underground rock formation fractures. process. Anisotropic gradient inversion provides an important basis for reservoir prediction in oil and gas exploration, and the use of wide-azimuth seismic data from surface observations can better invert anisotropic gradients. [0003] The anisotropic gradient inversion framework is consistent with the pre-stack isotropic inversion. It is a model-based pre-stack inversion method and an important means of predicting fractured reservoirs. This inversion method starts ...

AI Technical Summary

Problems solved by technology

Similarly, in 2018, She et al. applied the dictionary learning method to the pre-stack AVO elastic parameter inversion for the first time, and proposed a new method using sparse representation as a regularization term, which is effective both in numerical simulation and in actual data. Very good results, this is a breakthrough in inversion history; however, this method can only solve the inversion of isotropic elastic parameters, and cannot invert seismic data in anisotropic media

Images