基于孔隙结构储层分类的中低孔特低渗储层渗透率确定——以B区块S油层为例  

Determination of permeability of medium-low porosity and extra-low permeability reservoirs based on pore structure reservoir classification:a case study of S reservoir in block B

作　　者：唐晓敏[1,2] 殷雪松 吕亚娟 宋延杰 陈学洋[3] 易俊 TANG XiaoMin;YIN XueSong;LYaJuan;SONG YanJie;CHEN XueYang;YI Jun(School of Earth Sciences,Northeast Petroleum University,Daqing 163318,China;Accumulation and Development of Unconventional Oil and Gas,State Key Laboratory Cultivation Base Jointly-Constructed by Heilongjiang Province and the Ministry of Science and Technology,Daqing 163318,China;Institute of Geology,No.9 Oil Production Company,Daqing Oilfield Limited Company,Daqing 163853,China)

机构地区：[1]东北石油大学地球科学学院,大庆163318 [2]非常规油气成藏与开发省部共建国家重点实验室培育基地,大庆163318 [3]大庆油田有限责任公司第九采油厂地质研究所,大庆163853

出　　处：《地球物理学进展》2023年第1期271-284,共14页Progress in Geophysics

基　　金：黑龙江省省属本科高校基本科研业务费项目(HBHZX202003)资助。

摘　　要：针对B区块S油层含泥含钙中低孔特低渗储层渗透率计算精度低的难题,分析岩性、物性、孔隙结构对储层渗透率的影响,明确了孔隙度、泥质含量、钙质含量、孔隙结构是影响B区块S油层特低渗储层渗透率的主要因素,其中,孔隙结构是影响特低渗储层渗透率的关键因素.综合运用压汞曲线、孔喉半径分布特征以及流动单元指数反映特低渗储层孔隙结构变化,将特低渗储层按不同孔隙结构划分成3种类型,建立了特低渗储层类型的判别标准.利用中子测井、密度测井、声波测井、微球形聚焦测井、深浅侧向电阻率测井差值的绝对值等5个储层类型识别的敏感测井响应及参数,使用决策树法、最邻近结点法、BP神经网络法和支持向量机法建立了4种基于机器学习的储层判别方法,储层类型判别准确率依次提高,其中,基于支持向量机的储层类型判别方法判别准确率最高92.2%,且对3种类储层判别效果均很好.针对3类储层分别建立了渗透率计算公式.实际井解释结果表明,基于机器学习储层分类的渗透率模型计算B区块S油层特低渗储层渗透率精度明显高于储层分类前渗透率计算精度,其中,基于支持向量机储层分类计算的渗透率精度最高.In order to solve the problem of poor calculation accuracy of permeability”of shaly and calcareous formation with medium-low porosity and extra-low permeability in S reservoir in block B,the influences of lithology,physical property and pore structure on permeability of the reservoirs are analyzed.The results show that porosity,shale content,calcium content and pore structure are the main factors,and the pore structure is the key factor that affect the permeability of extra-low permeability reservoirs in S reservoir in block B.Based on combination of mercury injection curve,distribution characteristics of pore-throat radius and flow unit index representing the change of pore structure,the extra-low permeability reservoir can be divided into three types according to different pore structure,and the standard for determining the type of extra-low permeability reservoir is established.After neutron log,density log,acoustic log,microspherically focused log,and absolute value of difference between deep and shallow laterolog are selected as sensitive logging response and parameter,identification methods of reservoir type are proposed based on four kinds of machine learning algorithms including decision tree method,K-Nearest Neighbor method,BP neural network method and Support Vector Machine(SVM)method.The discriminant accuracy of reservoir type is all improved and increases in the order with four kinds of machine learning algorithms.Also,the accuracy of discriminant method of reservoir type is the highest at 92.2%,and the discriminant result for three types of reservoirs is all best based on SVM.The permeability calculation formula is established respectively for three types of reservoirs.The interpretation results for wells show that the accuracy of permeability calculated with machine learning reservoir classification is significantly higher than that of permeability calculated on unclassified reservoir,and the accuracy of permeability calculated with SVM reservoir classification is the highest for S reservoir in bl

关 键 词：中低孔特低渗储层 含泥含钙 渗透率 孔隙结构 储层类型 机器学习 

分 类 号：P631[天文地球—地质矿产勘探]