基于多尺度数据融合Markov链模型的岩性随机模拟  

Lithology stochastic simulation based on Markov chain models integrated with multi-scale data

作　　者：李军[1] 熊利平[1] 方石[2] 唐林[3] 霍红[1] 

机构地区：[1]中国石化石油勘探开发研究院,北京100083 [2]吉林大学地球科学学院,吉林长春130061 [3]中国石化胜利油田分公司河口采油厂,山东东营257200

出　　处：《石油学报》2010年第1期73-77,共5页Acta Petrolei Sinica

基　　金：中国石油化工集团公司科技攻关项目(JP06001)资助

摘　　要：Markov链模型在储层随机建模中发挥着越来越重要的作用,但难以融合岩心、测井、地震等多尺度数据限制了它在实际中的应用。依据前人研究的结果,提出了将多尺度数据融入到Markov链模型中的相关方法和公式,即将大尺度数据作为条件数据以贝叶斯公式表达,同时利用公式将小尺度数据转换为井点硬数据。应用此方法对SL盆地Y地区过井剖面进行的岩性模拟表明,相对于无数据融合的方法,此方法能更加直观、准确地揭示薄岩性层的分布。The Markov chain models have played a more important role in the reservoir stochastic modeling,but the models are difficult to be integrated with the multi-scale data such as logging,core data and seismic data,which limits the application of the models.A new method and some formula were proposed for integrating the multi-scale data with the Markov chain models.The large-scale data were added into the models and taken as the conditional data,and the small-scale data were used to get exact data of well points by formula.The application of the method to simulate lithology of a section across wells in Y region of SL Basin shows that the fine lithology distribution obtained from the new method is more accurate and distinctive than that of the previous method.

关 键 词：MARKOV链模型 多尺度数据 融合方法 岩性随机模拟 薄岩层分布 

分 类 号：TE19[石油与天然气工程—油气勘探]