The Improvement Of C3Coherence Algoirthm And The Application In Fault Identification

Seismic coherence analysis technology was officially put forward by Bahorichand Farmer who are from Amoco oil company as an independent seismic attributes inthe65th SEG annual meeting in1995, then, the technology has its unique applicationand works well in faults interpretation.In seismic coherence cube, the points of low coherent value are the formationdiscontinuous points, so we can come to know the distribution of faults in the workarea. Three dimensional coherent datas is more intuitive and precise than conventionalfault interpretation, reducing human intervention and experience factors andimproving the efficiency and accuracy of the interpretation, shortening the cycle ofgeophysical exploration and interpretation, so the study of three-dimensional seismiccoherence technology is very meaningful.There are three generations of coherence cube algorithm: the first generation ofcoherent algorithm is Cl coherent algorithm which is based on the cross correlation;the second generation coherent algorithm is C2coherent algorithm which is based onmulti-channel similarity; the third generation algorithm is C3coherent algorithmwhich is based on the characteristic structure. Compared to C1and C2coherentalgorithm, C3coherent algorithm has a higher lateral resolution and antinoiseperformance, but because of failing to consider the problem of the formation dip, theillusion of the low coherence value appeared in formation dip area, makes C3coherent algorithm’ coherent image quality descending,leading to the faultinterpretation errors. In addition, the calculation process is based on the analysis ofthe characteristic structure of the covariance matrix, so the amount of calculation islarge and the calculation time is long. When the seismic data is large, it will reduce itsoperating efficiency. For this, this paper improved C3coherent algorithm in thefollowing:Firstly, we need to consider formation dip in C3coherence cube algorithm, thispaper use plane-wave destruction filters to get formation dip, then it can get thecoherent value from seismic datas based on control of formation dip.Studies haveshown that C3coherent algorithm improved in formation dip has a higher resolution,eliminating the low coherence area’ interference led by formation dip effectively andhighlighting effective fault information, but this method’ improvement is obvious information that the dip is relatively large in faults interpretation.Secondly, in order to improve the efficiency of operation of C3coherentalgorithm, In this paper, Aitken acceleration method and origin translational acceleration method reference based on power method are introduced to C3coherentalgorithm. C3coherent algorithm is based on the characteristic structure, powermethod is a iterative methods of calculating the matrix eigenvalues and correspondingeigenvectors, Aitken acceleration method,which is used in power method to acceleratthe convergence of power method,is a mathod that accelerat the convergence ofconvergent sequence. Studies have shown that, compared to the previousalgorithm,C3coherent algorithm based on the Aitken acceleration improved has theSame result in the same error precision range, but it greatly reduce the number ofiterations and the running time, while the accelerating effect of origin translationalacceleration method is not good.