Analysis Of Multiple Elimination Methods And Discussion On Its Application

Multiple attenuation has always been a difficult problem in seismic data processing.The existence of multiple waves results in lower SNR (signal to noise ratio) of theseismic data and interferes with the identification of effective information, which,thereby, brings great difficulties to the velocity analysis and data migration, influencesthe reliability of the seismic tomography, and furthermore, the false structural profileformed will eventually influence the correct interpretation of the regional structure.At present, there are mainly two kinds of methods of multiple attenuation, includingde-noising method based on the spatial characteristics differences between first-break andmultiple waves and predict-subtract method based on the multiple waves’internalperiodicity and predictability. The former could be considered when there is no obviouslateral variation in subsurface structure and less demanding on the accuracy of the results,and the latter is mainly divided into two kinds of methods, including both predictive de-convolution method which is good at dealing with the data consisting of strong multipleperiodicity in shallow water area seismic data processing and wave field extrapolationmethod, by which better result could be obtained in the complex structural area.However, it still remains unavoidable on how to subtract the multiple data predicted fromthe input data when applying predict-subtract based method. Due to the reflectioninterface features and model error, it’s often unable to predict both amplitude and travel-time perfectly. However, the adaptive filter with dynamic adaptive modification ofcontrolling parameters could be realized to filter the signal real-time-feedbackcorrectively, which in turn helped to achieve the goal of adaptive subtraction of multiplewaves predicted.Furthermore, the SRME (Surface Related Multiple Elimination) method, one of themost important breakthroughs in seismic data processing over the past thirty years, is adata driven method, which means no consideration of the regional structuralcharacteristics. For seismic data with complex underwater conditions, using wave fieldextrapolation method or SMRE method might be a good choice, but it’s better to useSRME method if the water is deep and there exist strong reflection interface under theseabed. To sum up, during the actual seismic data processing, based on the knowledge levelon the transcendental information on the specific area; the method of applying variouseffective analysis and identification methods comprehensively before designingprocessing flow of multiple attenuation is generally widely adopted by now. In order toobtain overall best results, the realization of optimized combination of better multipleattenuation method depend on the optimal adoption during the different processing stepsand data domain.In this article, we first did a comparative study on the current multiple attenuationtechnique based on Vershcuur’s classification suggestion, by which the multipleattenuation methods are mainly divided into three, including travel-time characteristicdifference methods, predict-subtract methods, and data driven surface multipleelimination method.Then, we realized surface multiple elimination algorithm in two dimensions and itsvalidation was confirmed by the test on both simulated and actual data. Furthermore wedid comparative analysis on the adaptive subtractive method and realized the adaptivefilter algorithm based on LMS (Least Mean Square) algorithm and did valid test usingrelative data.Finally, we discussed the influence of multiple on tomography, indicating that on onehand, based on the current theory and technology, we must consider multiple waves asnegative factors and the elimination is unavoidable, on the other hand, the multiple wavescontains more subsurface structure information and the further study is necessary,including transforming multiple waves to first-break and using both multiple and first-break directly to do structure inversion imaging, in order to obtain more accuratestructural profile and provide more detailed and precise materials for studying on near-surface geological structure.