横向非均匀介质中的曲率属性稳健提取方法研究

项目名称： 横向非均匀介质中的曲率属性稳健提取方法研究

项目编号： No.41504092

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 天文学、地球科学

项目作者： 王晓凯

作者单位： 西安交通大学

项目金额： 22万元

中文摘要： 曲率属性（包含结构曲率和振幅曲率）不但能够反映地层的几何形态，还能够刻画地层的横向非均质性，因此在复杂油气藏勘探开发中能够发挥重要的作用。叠后地震数据不但包含地层几何信息，而且还含有地层横向非均质性信息。常规方法直接利用地震数据估计倾角，地层横向非均质性将严重影响倾角估计的准确性，进而大幅降低两种曲率属性提取的准确性和稳定性。本项目针对此问题展开研究：拟利用相空间变换从三维叠后地震数据中剥离出仅含有地层几何形态的数据，以此为基础，结合主分量分析技术和多窗估计技术提出准确估计倾角的方法，进而提出结构曲率属性的稳健提取方法；在准确倾角的基础上，拟以主分量分析为工具，提出振幅曲率的稳健快速提取方法并对其解译。本项目的研究成果能够准确、精细地刻画地层几何形态及非均质性，为复杂油气藏精细勘探开发提供有力支撑。

中文关键词： 储层预测；横向非均匀介质；曲率；主分量分析；小波变换

英文摘要： The curvature attributes (include structural curvature and amplitude curvature) can not only reflect the geometry of layer but also characterize the lateral nonhomogeneity. Therefore, curvature attributes play an important role in the exploration and development of complex oil-gas reservoir. In application, the stacked seismic data contains not only the geometric information of layer but also the lateral nonhomogeneity of layer. If the stacked seismic data is directly used to estimate dip/azimuth, the effect of lateral nonhomogeneity of layer would be involved, which will reduce the precision and robustness of structural curvature and amplitude curvature. This project focuses on this issue. We would use one-dimensional/high-dimensional continuous wavelet transform to extract the instantaneous phase, which only contains the geometric information of layer. Will the help of principal component analysis and multiwindow estimate, we will accurately estimate the dip based on instantaneous phase. The robust structural curvature can be extracted according to the dip estimation result. Finally, with the assistance of principal component analysis and accurate dip, we would propose a robust amplitude curvature extraction method and interpret this amplitude curvature. If this project runs smoothly, the result would support the exploration and development of complex oil-gas reservoir strongly, especially in characterizing the geometry and lateral nonhomogeneity of layer.

英文关键词： reservoir prediction;lateral inhomogeneous media;curvature attribute;principle component analysis;wavelet transform