基于低秩与稀疏优化分解的三维角度域绕射波分离与成像方法

Fractured-cavernous carbonate reservoir has become the main petroleum target in China. How to predict the relevant caverns and fractures is a challenging problem. Conventional seismic exploration method mainly employs reflection and its limited resolution makes the prediction work hard to perform. Seismic diffraction carries the information of small-scale geologies. We intend to develop a diffraction separation and imaging method by low-rank and sparse decomposition in 3D angle domain. First, construct an imaging matrix in angle domain by 3D shot gather, and then utilize low-rank and sparse decomposition method to separate diffraction, where the separated diffraction keeps a good waveform because of its generation by the same shot and thus has advantage for high-resolution imaging. Using diffraction amplitude coefficient in uniform theory of diffraction, we design a diffraction compensation imaging kernel function in 3D angle domain for solving the multi-path diffraction imaging problem and improve imaging quality of diffractions. This method has great potential for clarifying the interior stories of carbonate reservoir, and has significance in favorable zone evaluation and exploration risk reduction.

碳酸盐岩缝洞型油气藏是我国油气资源主要勘探目标，其中洞穴、孔洞和裂缝预测难题亟待解决。常规地震勘探方法以反射波为主，分辨率有限，无法对之有效识别。绕射波是小尺度地质体信息载体，本项目拟研究基于低秩与稀疏优化分解的三维角度域绕射波分离与成像方法，首先针对三维炮集数据构建角度域成像矩阵，其次研究最优化理论中的低秩与稀疏优化分解方法分离绕射波，分离出的绕射波由同一震源激发，波形一致性强，有利于高分辨率成像；在此基础上，利用一致性绕射理论中的绕射波振幅系数，设计三维角度域补偿式绕射波成像核函数，解决多路径绕射波成像问题，提高绕射波成像质量。本项目有望解决缝洞型储层内幕成像难题，在寻找碳酸盐岩有利发育带，减少勘探风险上具有重要意义。

本项目针对当前我国碳酸盐岩缝洞型油气藏预测难题，通过理论研究、数值模型测试和实际资料应用，系统研究了基于低秩与稀疏优化分解的三维角度域绕射波分离与成像方法。首先针对三维炮集数据构建角度域成像矩阵，其次研究最优化理论中的低秩与稀疏优化分解方法分离绕射波，分离出的绕射波由同一震源激发，波形一致性强，有利于高分辨率成像；在此基础上，利用一致性绕射理论中的绕射波振幅系数，设计三维角度域补偿式绕射波成像核函数，解决多路径绕射波成像问题，提高绕射波成像质量。本项目研究的方法提高了缝洞型储层内幕成像精度，在寻找碳酸盐岩有利发育带，减少勘探风险上具有重要意义。