Multiple Elimination Research Based On Feedback Method

Multiple elimination is always an important work in seismic data processing. Multiple reduces the resolution of seismic data, disturbs the recognition of primary, and damages the quality of seismic data processing seriously. So it is significant that how to elimination all kinds of multiples effectively and reserve the primary possibly in seismic exploration.The elimination of multiple based on wave equation consider not only the kinematics characteristic of wave propagation but dynamic characteristic, and also avoid some shortages when we use some filtering methods. We believe that the elimination of multiple based on wave equation will have a bright future, because it doesn't need any priori information and have a better result.First, this thesis systemically discusses all kinds of methods about multiple elimination, including filtering methods and wave equation-based methods. Then I study the basic theory and procedures of prediction subtraction method based on wave equation. I discuss the prediction and subtraction detailedly. Based on the analysis of traditional SRME method, I achieve the single ensemble SRME method. This method reduce the request of computer's hardware, and increase the computation speed. I study some detail issues about the extrapolation of near offset wavefield, the affection of inverse source wavelet and multiple encoding scheme. As a part of prediction subtraction method, this paper discusses adapt subtraction algorithm deeply, make detailed theoretical derivation of single channel least-square matching filter method and constrained cross-equalization expanded multichannel matching filter method, and compare these two methods with models. Establish the relation between the effectiveness of SRME and the iterative times and length of time window, which is used for quality control of multiple elimination. Then, I summarize and compare the several commercial seismic processing software, and develop a software focused on multiple elimination. I apply the software to the synthetical model and real seismic data, and get a better result. At last, I summarize some recognitions and recommendations about prediction subtraction method based on wave equation.