Study On Seismic Identification Method Of Fault In Complex Fault Block Oil Reservoir

Underground faults have a dual role for reservoir development,which is an important channel for oil and gas migration and accumulation,and also leads to the escape of oil and gas.For complex fault block reservoirs,accurate identification and detailed description of faults play an important role in determining reservoir range,calculating reserves,deploying well locations and formulating production plans.It can be said that the identification and description of faults is always the top priority in the exploration and development of fault block reservoirs.The identification and description methods of faults include geological,logging and seismic methods,among which seismic methods are the most important means.Because of the complex spatial pattern of fault,the imaging accuracy of seismic data at fault development is poor,and the signal-to-noise ratio is low,more effective fault identification methods and techniques are needed.From the preprocessing of seismic data(" entropy " preserving edge filtering),to the method improvement(spectral variance coherence and magic matrix),to the new method and intelligent optimization algorithm proposed(rotating rhombus interpretation method and optimal selection ant colony algorithm),this paper carried out systematic theoretical research and application.The theoretical research has been tested by geotechnical 3D model,Marmousi2 model and Qdome 3D model,and actual applications have been tested in several blocks such as Yong3,Xin25,Gao89 and HZ,and the application effect is good.The pre-processing of seismic data can provide optimized data with high signal-to-noise ratio and edge information maintained for subsequent seismic methods to accurately identify faults.Combined with the concept and function of "entropy" in the imageology,this paper introduces it into the estimation of the structural complexity of seismic data,and realizes a new method of edge preservation and denoising based on "entropy".This method can objectively evaluate the structural complexity of a block of data to add the second derivative information related to the structure in anisotropic diffusion filtering,thereby better protect the edge information of special geological bodies such as faults.The eigenvalue coherence has different characterization,this paper first deeply analyzes and interprets the physical meaning.Then the coherence technique based on spectral variance is realized by combining the principle of variance attribute with spectral analysis and eigenvalue coherence to solve the problem of conventional coherence that is easily disturbed by inclined strata.Compared with conventional coherence,this method can better suppress the disturbance of inclined strata and has better ability to recognize small faults.Finally in order to better strengthen the recognition effect of coherence algorithm,this paper combines the operator constructed by the magic matrix with the coherence algorithm to realize the coherence enhancement algorithm based on the magic matrix.It can identify faults in different directions,thereby improving the spatial recognition accuracy of complex faults.Rhombus(body)compared with other mathematical patterns,its long diagonal has obvious pointing in the direction of the action,when the rhombus long diagonal and fault in the same direction,the right length of the diagonal size can make account for larger proportion,along the fault data to participate in the calculation of the other regions involved in calculating ratio,virtually to calculate with the directional,benefit to the precise identification of fault,etc.Therefore,this paper proposes a new attribute extraction method based on rotating rhombus,its rationality and effectiveness have verified by applications of the model and actual fault shattered zone data.This paper develops an ant colony algorithm based on optimal selection for fault automatic and fast tracking by combining the gradient direction calculation,threshold selection and ant colony algorithm.The method is based on the previously extracted high SNR coherence data volume,combined with the image threshold segmentation algorithm to select the optimal threshold and the principal component analysis(PCA)gradient direction estimation to calculate the optimal direction of ant tracking.The automatic control and tracking of the algorithm enables the ants tracking to always be in the optimal choice,which improves the efficiency of the algorithm and the accuracy of fault recognition,makes this algorithm have a good application prospect.