Research On Adaptive Subtraction Method Of Multiple Waves

At present,the methods to suppress multiple waves can be summarized into two categories: one is the direct filtering method based on the difference separability of primary waves and multiple waves;the other is the prediction subtraction method based on wave equation theory.For the complex seabed seismic data in the deep sea,the multiple wave has the characteristics of significant,strong energy and wide distribution,so the prediction subtraction method is the most common and effective method to suppress the multiple waves of the marine seismic data.Among these predictive subtraction methods,the conventional least squares adaptive subtraction algorithm has become the most widely used algorithm in industry due to its simple principle,computation efficient and small memory requirement.However,the traditional least squares adaptive subtraction algorithm not only has multiple wave residuals,but also damages the primary wave energy at the intersection part of primary and multiple waves,affecting the imaging quality of seismic section.To solve these problems,we first introduce a multi-channel two-norm adaptive subtraction algorithm,considering the lateral distribution characteristics of seismic records,in order to restrain its damage to primary wave energy.At the same time,we compare and analyze the performance of the adaptive subtraction algorithm based on one-norm energy minimization by using the feature that one-norm is robust to large anomalies.In addition,considering that the dip angle of the same events at the intersection of multiple and primary waves is different,we combine the Tau-p transform algorithm to suppress multiple waves,and improve the conventional least squares adaptive subtraction algorithm.When the assumption of seismic wavelet consistency is not valid,we introduce the adaptive subtraction technique based on pattern recognition to suppress multiple waves.Finally,combined with complex seismic data,the advantages and disadvantages of each adaptive subtraction algorithm are summarized and suggestions for improvement are put forward.