基于不同邻域系统的马尔可夫链模型的储层岩相随机模拟  

Stochastic Simulation of Reservoir Lithofacies Based on Markov Chain Models with Different Neighborhood Systems

作　　者：李军[1] 郝天珧[1] 刘建[2] 周翠英[3] 徐亚[1] 赵百民[1] 

机构地区：[1]中国科学院地质与地球物理研究所,北京100029 [2]胜利油田滨南采油厂地质所,山东滨州256606 [3]中国矿业大学资源与安全工程学院,北京100083

出　　处：《现代地质》2006年第4期621-627,共7页Geoscience

基　　金：中国科学院知识创新重大项目(KZCX1-SW-18-01);国家自然科学基金项目(40620140435;40674046)

摘　　要：针对在油气储层随机模拟中马尔可夫链模型的不同方向的转移概率矩阵求取困难的问题,提出一种二维剖面中不同方向的转移概率矩阵求取方法,这种方法的提出使得不同阶次的各向同性和各向异性的邻域系统的转移概率矩阵的求取变得容易可行。随后,对不同邻域系统的马尔可夫链模型采用蒙特卡罗抽样方法进行了储层岩相随机模拟试验。最后比较了不同邻域系统岩相模拟的结果并探讨了在储层研究中的适用性。For the difficulty of getting transition probability matrixes in various directions in Markov chain models, the paper presents a method to figure out it, which makes getting transition probability matrixes of different neighborhood systems of Markov chain models easier and more feasible. Furthermore, Monte Carlo sampling method is used to simulate reservoir lithofacies, based on different neighborhood systems of Markov chain models. In the end, the different simulated results are compared and the characteristics and applications of different neighborhood systems in the study of reservoir are discussed.

关 键 词：马尔可夫链 邻域系统 转移概率矩阵 岩相随机模拟 

分 类 号：TE319[石油与天然气工程—油气田开发工程]