基于模糊C均值聚类与Bayes判别的致密油储层分类评价  

Classification and Evaluation of Tight Oil Reservoirs Based on Fuzzy C-Means Clustering and Bayes Discrimination

作　　者：王伟[1] 康胜松[1] 高峰[1] 郭粉转[1] 张亮[1] Wang Wei;Kang Shengsong;Gao Feng;Guo Fenzhuan;Zhang Liang(Shaanxi Yanchang Petroleum(Group)Co.,Ltd.,Xi'an Shaanxi 710075,China)

机构地区：[1]陕西延长石油(集团)有限责任公司,陕西西安710075

出　　处：《特种油气藏》2020年第5期118-124,共7页Special Oil & Gas Reservoirs

基　　金：陕西省科技统筹创新项目“延长难采储量有效动用开发技术研究”(2016KTCL01-02)。

摘　　要：为解决常规致密油储层因缺乏岩心分析资料而难以进行储层分类评价的难题,通过优选属性参数,应用模糊C均值算法划分致密油储层最优类别,再运用Bayes判别分析法建立储层类别与常规测井属性的关系式,利用常规采油井的测井属性判别储层的类别。现场实践应用表明:利用该方法划分的鄂尔多斯盆地志丹地区长7Ⅱ^6、长7Ⅰ^2射孔砂体储层类别与试油结果一致,研究区203口井储层划分吻合率为89.7%。该研究对高效开发致密油藏具有一定的指导意义。In order to solve the problem that the conventional tight oil reservoirs are difficult to classify and evaluate due to the lack of core analysis data,the optimal type of tight oil reservoirs was classified by fuzzy C-means algorithm through optimizing attribute parameters;then Bayes discriminant analysis was applied to establish the relationship between reservoir types and conventional well logging attributes,and the well logging attributes of conventional oil production wells were utilized to distinguish the reservoir types.Field applications show that the reservoir types of Chang 7Ⅱ^6 and Chang 7Ⅰ^2 perforated sand bodies in Zhidan area of the Ordos Basin are consistent with the oil test results.The coincidence rate of reservoir classification of 203 wells in the study area is 89.7%.This study has a certain guiding significance for the efficient development of tight oil reservoirs.

关 键 词：致密油 储层分类 模糊C均值聚类 Bayes判别 储层评价 鄂尔多斯盆地 

分 类 号：TE348[石油与天然气工程—油气田开发工程]