基于BP神经网络识别的曲堤油田低阻油层研究  

Study on Low Resistivity Oil Layer of Qudi Oilfield Based on BP Neural Network Recognition

作　　者：陈铭 彭作磊 Chen Ming;Peng Zuolei(School of Geosciences, China University of Petroleum(East China), Qingdao 266580, China)

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛266580

出　　处：《宁夏大学学报（自然科学版）》2020年第1期98-103,共6页Journal of Ningxia University(Natural Science Edition)

基　　金：中国石油华北油田校企合作基金资助项目(HBYT-CY3-2011-JS-345)。

摘　　要：作为一种非常规储层,低阻油层已成为现阶段油田增储增产的主要来源之一.曲堤油田主力油层馆陶组和沙河街组具有明显低阻特征,常规测井解释方法难以识别.为更大限度利用测井资料,提高低阻油层测井解释精准度,在研究区内低阻油层特征的基础上,从岩石物理成因等方面对曲堤油田馆三段、沙三段和沙四段低阻油层的成因进行分析;用BP神经网络对低阻油层测井曲线数学特征进行学习和训练,并对其进行识别;建立孔隙度、渗透率、束缚水饱和度等参数模型对低阻油层进行精细定量评价.结果表明,岩石粒度细、地层水矿化度高、黏土矿物含量高等是油层低阻成因的主要因素;用BP神经网络可有效划分油层、水层、干层等,识别准确率在88%以上;孔隙度、渗透率、含水饱和度等模型的精度分别为78.33%,79.62%,64%.As a kind of unconventional reservoir,low resistivity reservoir has become one of the major sources of production increment for oil field nowadays.The major reservoirs:Shahejie and Guantao Formations of Qudi oil field have distinct low-resistivity characteristic that it cannot be recognized by conventional logging interpretation method.The article analyzes the origination of low-resistivity reservoirs formation of Guan 3 segment,Sha 3 segment,and Sha 4 segment in Qudi oil field from the aspect of rock physical genesis on the basis of analysis and summary of low-resistivity reservoirs characteristic in the area to make more use of logging information and enhance the well logging interpretation precision of low-resistivity reservoirs;it studies and trains the mathematical feature of low-resistivity reservoirs logging curve in this area with BP neural network,and identifies it;it establishes the parameter models such as porosity,penetrance,irreducible water saturation,etc.for fine quantitative evaluation on low-resistivity reservior.The result shows that,fine rock grain size,high formation water salinity,and the influence of clay minerals are the principal factors for low resistance causes of oil reservoir in this area;BP neural network can effectively partition oil layer,water layer,and dry layer with the identification accuracy over 88%;the model precision of porosity,penetrance,and water saturation are 78.33%,79.62%,64%respectively.

关 键 词：曲堤油田 低阻油层 BP神经网络 油层评价 

分 类 号：P62[天文地球—地质矿产勘探] TP183[天文地球—地质学]