用神经网络方法识别碳酸盐岩裂缝系统中的气与水  

GAS WATER RECOGNITION IN CARBONATE ROCK FRACTURE SYSTEMS BY USE OF NEURAL NETWORK

作　　者：李亚林[1] 雷晓[2] 林万昌 易宗富 周阿波[3] 蒋骥[3] 李兴甫 

机构地区：[1]四川石油管理局地调处成都物探研究中心 [2]西南石油学院 [3]四川石油管理局地调处

出　　处：《天然气工业》2000年第1期32-36,共5页Natural Gas Industry

摘　　要：碳酸盐岩裂缝系统中所含流体性质的确定（ 即究竟是含气还是含水或二者均有） 是四川盆地天然气勘探中所面临的关键问题和难点之一。在用川南沈公山构造的三维地震资料提取出多种地震特征参数和综合利用已有的钻井、测井资料的基础上，通过使用具有无教师监督和自聚类特点的自组织映射神经网，提出了一套地震特征参数筛选和气、水识别分类器设计的方法。由于应用该方法能够将沈公山构造上已知井所钻遇裂缝系统中的气、水进行正确区分，因此有可能利用该方法对未钻遇裂缝系统中的流体性质（气或水）作出判断。It is one of the key and difficult problems in natural gas exploration in Sichuan Basin to identify the property of the fluid (water or gas or both of them) in carbonate rock fracture systems.On the basis of extracting various seismic feature parameters from the 3 D seismic data of Shengongshan structure in South Sichuan and in combination with the data on drilling and logging,a set of methods selecting seismic feature parameters and designing the classifier of gas water recognition are proposed by use of an unsupervised,self clustering and self organizing mapping neural network.Because the gas and water in the frature systems encountered by the wells drilled in Shengongshan structure can be correctly divided by these methods,it is possible to identify the property of the fluid (gas or water) in the undrilled fracture systems by the same methods.

关 键 词：神经网络 碳酸盐岩 裂缝性油气藏 储集层流体 

分 类 号：P618.13[天文地球—矿床学]