神经网络在石油测井解释中的应用综述  

The application of network techonology to petroleum logging interpretation

作　　者：赵军龙[1] 李纲[1] 麻平社 巩泽文[1] 蒙灵飞[1] 李甘[1] 

机构地区：[1]西安石油大学,油气资源学院,西安710065 [2]中国石油测井有限公司长庆事业部,西安710201

出　　处：《地球物理学进展》2010年第5期1744-1751,共8页Progress in Geophysics

基　　金：陕西省工业攻关计划基金项目(2009K10-15)资助

摘　　要：常规测井解释往往通过"四性关系"研究建立储层参数解释模型,将测井信息转化成地质参数开展储层评价,但对于复杂岩性储层、低渗低阻等储层而言,将测井信息转化为储层参数及利用测井曲线开展储层评价往往存在多解性.作为一种非线性数学方法,神经网络具有多信息融合、综合预测的功能,在解决一些复杂、非线性问题领域展示了强大的生命力.将神经网络引入石油测井解释中,综合利用多种测井信息开展岩性、物性及含油性分析,可有效提高石油测井解释的精度和效率;利用神经网络技术开展流体性质识别和储层裂缝研究可有效破解此类世界性难题.为此,本文对神经网络技术在石油测井解释中的应用进行了综述,并对应用中应注意的事项进行分析,最后进行了应用展望.By means of four-property relation research, logging data is transformed into reservoir geology parameters in the conventional logging interpretation. However, there is ambiguity in the evaluateration of complex lithology reservoirs and bad permeability and low resistivity reservoirs. The network has the function of much information amalgamation and synthetically prediction and displays the active vitality of resolving complex and nonlinear problems. The accuracy and efficiency of logging interpretation would be enhanced when the network is used in the analyzing lithology and physical properties and oil-bearing. The world problem would be solved when the network is applied to fluid property identification and the reservoir fracture study. This paper presents a review and thoughts on this aspect, and discuss its application.

关 键 词：神经网络 测井解释 综述 注意事项 展望 

分 类 号：P631[天文地球—地质矿产勘探]