深度域地震子波提取方法综述

Review on depth-domain seismic wavelet estimation

由于深度域地震子波具有对介质速度依赖的特性,褶积模型无法直接用于提取深度域地震子波。准确提取深度域地震子波对于深度域地震资料的反演和解释具有重要意义。首先,分析了深度域地震子波提取中存在的2个主要问题,即深度域地震子波对介质速度的依赖特性和深度域测井数据与地震数据的采样间隔非一致性,并提出了相应对策。然后,对目前已有的深度域地震子波提取方法进行了梳理,根据是否在方法中用到测井信息,将方法分为确定性地震子波提取方法和统计性地震子波提取方法2大类,并对每一种方法的优缺点及适用条件进行了总结。最后,对深度域地震子波提取方法的未来发展提出以下建议:①深度域统计性地震子波提取方法有待进一步发展;②将地震反演结果与测井资料之间的相关性纳入到子波提取结果优劣的评价标准中;③使用地震子波模型减少子波提取中需要求解的未知量个数,以提升方法的计算效率;④将一维深度域地震子波提取方法扩展至提取二维和三维深度域地震子波;⑤利用机器学习方法的优势,建立不依赖褶积模型的深度域地震子波提取方法,或建立可以省去子波提取的深度域地震反演方法。

The seismic wavelet in the depth domain is a function of the subsurface-medium velocity.Its wave number changes with the medium velocity.The“linear time-invariant”condition does not hold in the depth domain,thus the convolution model cannot be directly used to extract seismic wavelet information in such domain.The accurate estimation of wavelet information is essential for depth-domain seismic inversion and interpretation.In this review,two main problems were analyzed and discussed,namely the variation of the seismic wavelet waveform with the medium velocity in the depth domain and the inconsistency of the sampling interval between depth-domain well-log and seismic data.Countermeasures to overcome these issues were suggested.Subsequently,existing depth-domain seismic wavelet estimation methods were described and classified.According to whether well logging information is used,these methods were divided into deterministic and statistical seismic wavelet estimation methods.The advantages,disadvantages,and applicability conditions of each method were summarized.Finally,the following suggestions were made for improving these methods:①further developing statistical seismic wavelet estimation methods in the depth domain;②utilizing the correlation between seismic inversion results and well-log information as one of the criteria to evaluate seismic wavelet estimation results;③utilizing seismic wavelet models to reduce the number of unknowns in the wavelet estimation and improve computational efficiency;④establishing two-dimensional and three-dimensional depth-domain seismic wavelet estimation methods;⑤establishing depth-domain seismic wavelet estimation methods that do not depend on the convolution model or that omit seismic wavelet estimation using machine learning.