1、2023 年第 38 卷 第1期2023,38(1):0271-0284地球物理学进展Progress in Geophysicshttp:/wwwprogeophyscnISSN 1004-2903CN 11-2982/P唐晓敏,殷雪松,吕亚娟,等2023 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 地球物理学进展,38(1):0271-0284,doi:106038/pg2023GG0082TANG XiaoMin,YIN XueSong,L YaJuan,et al2023 Determination of permeability of medium-
2、low porosity and extra-low permeability reservoirs basedon pore structure reservoir classification:a case study of S reservoir in block B Progress in Geophysics(in Chinese),38(1):0271-0284,doi:10 6038/pg2023GG0082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 poros
3、ity and extra-lowpermeability reservoirs based on pore structure reservoir classification:a case study of S reservoir in block B唐晓敏1,2,殷雪松1,2,吕亚娟3,宋延杰1,2,陈学洋3,易俊3TANG XiaoMin1,2,YIN XueSong1,2,L YaJuan3,SONG YanJie1,2,CHEN XueYang3,YI Jun3收稿日期2022-05-22;修回日期2022-10-10投稿网址http:/www progeophys cn基金项目黑
4、龙江省省属本科高校基本科研业务费项目(HBHZX202003)资助第一作者简介唐晓敏,女,1981 年生,博士,副教授,从事测井资料解释方法研究 E-mail:txmdqpi163 com1 东北石油大学地球科学学院,大庆1633182 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1633183 大庆油田有限责任公司第九采油厂地质研究所,大庆1638531 School of Earth Sciences,Northeast Petroleum University,Daqing 163318,China2 Accumulation and Development of Uncon
5、ventional Oil and Gas,State Key Laboratory Cultivation Base Jointly-Constructed byHeilongjiang Province and the Ministry of Science and Technology,Daqing 163318,China3 Institute of Geology,No 9 Oil Production Company,Daqing Oilfield Limited Company,Daqing 163853,China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
6、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
7、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AbstractInordertosolvetheproblemofpoorcalculationaccuracyofpermeability”ofshalyandcalcareous formation with medium-low porosity and extra-low permeability in S reservoir in block B,the influencesof lithology,physical property and pore
8、structure onpermeability of the reservoirs are analyzedThe resultsshow that porosity,shale content,calcium content andpore structure are the main factors,and the pore structureis the key factor that affect the permeability of extra-lowpermeability reservoirs in S reservoir in block B Based oncombina
9、tionofmercuryinjectioncurve,distributioncharacteristics of pore-throat radius and flow unit indexrepresenting the change of pore structure,the extra-lowpermeability reservoir can be divided into three typesaccording to different pore structure,and the standard fordetermining the type of extra-low pe
10、rmeability reservoir isestablished After neutron log,density log,acoustic log,microspherically focusedlog,andabsolutevalueofdifference between deep and shallow laterolog are selectedas sensitive logging response and parameter,identificationmethods of reservoir type are proposed based on four kindsof
11、 machine learning algorithms including decision treemethod,K-Nearest Neighbor method,BP neural networkmethod and Support Vector Machine(SVM)method The地球物理学进展www progeophys cn2023,38(1)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关键词中低孔特低渗储层;含泥含钙;
12、渗透率;孔隙结构;储层类型;机器学习中图分类号P631文献标识码Adoi:10 6038/pg2023GG0082discriminant accuracy of reservoir type is all improved andincreases in the order with four kinds of machine learningalgorithms Also,the accuracy of discriminant method ofreservoirtypeisthehighestat92.2%,andthediscriminant result for three typ
13、es of reservoirs is all bestbased on SVM The permeability calculation formula isestablished respectively for three types of reservoirs Theinterpretation results for wells show that the accuracy ofpermeability calculated with machine learning reservoirclassificationissignificantlyhigherthanthatofperm
14、eability calculated on unclassified reservoir,and theaccuracy of permeability calculated with SVM reservoirclassification is the highest for S reservoir in block BKeywordsMedium-low porosity and extra-low permeabilityformation;Shaly and calcerous;Permeability;Pore structure;eservoir types;Machine le
15、arning0引言渗透率是衡量储层渗流能力好坏和预测储层产能大小的重要参数,其求取精度直接关系到储层产能计算的准确性(王清辉等,2021;赵军等,2021)B区块 S 油层储层岩性以粉砂岩、泥质粉、钙质粉砂岩为主,泥质含量主要分布区间为 8%18%,钙质含量主要分布区间为 0.22%20%,孔隙度介于 5%19%之间,渗透率介于(0.03 30)1015m2之间,属于含泥含钙中低孔特低渗储层 因此,建立一种适用于 B 区块 S 油层特低渗储层渗透率的计算方法,以提高渗透率计算精度,是十分有必要的渗透率的求取主要有两大类方法,第一类是基于岩石的微观孔隙结构特征,将孔隙空间等效为不同的物理模型,从
16、而推导出岩石渗透率计算公式(荆万学,2002;景成等,2013;谢伟彪等 2014),该方法具有较好的数学物理意义(于华等,2020),但岩石内部结构复杂多变,等效的物理模型并不能反应真实的岩石内部结构,因此,建立的渗透率预测模型精度有限(孙建孟和闫国亮,2012)第二类是利用岩心分析岩性物性参数或测井参数,采用统计分析方法,建立岩石渗透率计算公式(楚泽涵和谢京,1994;杜超等,2013),这种方法简便、实用,对于中高孔渗储层,渗透率计算精度较高,但对于低孔渗储层,由于复杂孔隙结构往往成为影响储层渗透率主要因素(何胜林等,2017),故导致渗透率计算误差较大 为了提高低孔渗储层渗透率计算精度,可根据孔隙结构好坏对储层进行分类,按不同储层类型建立渗透率计算公式(程梦薇,2016)压汞法是研究岩石孔隙结构最常用和直接方法,根据毛管压力曲线形态和孔喉半径分布特征以及排驱压力、中值压力等参数可实现储层分类(李彦山等,2009;王剑峰等,2010;李郑辰等,2013;杨玲等,2014;韩博华等,2021;崔改霞等,2021),但压汞法建立的储层类型判别标准多是定性的,即使有定量标准,其参数也需