【摘要】 地质统计学反演结合了随机建模和地震反演的优势,并充分地利用了测井数据纵向分辨率高以及三维地震数据横向分辨率高的特点,从而可以获得高分辨率的反演结果.传统的地质统计学反演通过两点之间的空间关系来描述地质体结构信息,其不足之处在于无法进行具有复杂结构地质信息的模拟,为了克服这些不足,多点地质统计学应运而生,多点地质统计学利用空间多个点的组合模式来描述地质结构信息,因此更适合于进行具有复杂结构地质体的模拟.在给出地质统计学反演理论框架的基础上,介绍了传统的两点地质统计学反演的理论基础和研究思路,阐述了发展中的多点地质统计学基本概念和几个算法,最后,介绍了一种基于SIMPAT算法的多点地质统计学反演以及应用效果.更多还原

【Abstract】 Combining the advantage of stochastic modeling and seismic inversion,geostatistics inversion can acquire high resolution inversion results.Geostatistics inversion provides a quantitative way to integrate the high vertical resolution of well logs with the dense aerial coverage of post-stack threedimensional seismic data.Therefore,it can potentially infer vertical variations of resolution similar to that of well logs,and at worst of vertical resolution equal to that of the seismic data at locations distant from wells. Traditional geostatistics attempts to quantify geologic information in the form of twopoint spatial correlations.However,two-point statistics cannot capture complex curvilinear structures in geology. The emerging alternative to these traditional techniques is the use of multiple-point geostatistics(MPG).The basic idea behind MPG is to go beyond the two-point,variogram-based modeling and to model the reservoir using multiple-point relations.Based on the formulation of the inverse problem as an inference problem,this article introduces the theoretical basis for traditional geostatistics inversion,describes the basic concepts and several algorithms of the developing MPG.Finally,a novel MPG inversion method and its application effect are introduced.更多还原