用自组织神经网络方法实现测井相定量识别  

WELL LOGGING FACIES IDETIFICATION USING SELF-ORGANIZATION NEURAL NETWORK

作　　者：魏莲[1] 肖慈珣[1] 

机构地区：[1]成都理工大学,四川成都610059

出　　处：《物探化探计算技术》2001年第4期324-327,352,共5页Computing Techniques For Geophysical and Geochemical Exploration

摘　　要：作者在本文中介绍了一种利用自组织神经网络进行测井相识别的方法。自组织神经网络是一种无导师学习算法 ,它通过网络自身的调节对输入进行聚类 ,近年来广泛应用于各类模式识别问题中。这里我们采用从测井曲线中提取出反映沉积环境的信息参数 ,再利用自组织神经网络进行测井相判别。此法从沉积成因角度出发 ,判别沉积相模式 ,消除了测井曲线中的不确定因素 。The paper introduces a method to identify well logging facies using self organization neural network, which is a non-supervisor learning algorithm. It extracts information from input data adaptively and now is largely used in various kinds of pattern identifications. In the method, we first extract the depositional environment information parameters from well logging data. Then depositional facies patterns are identified using the self organization neural network. As the method to identify depositional facies is based on depositional genetic factors, it can remove some uncertain factors from well data and is more representative.

关 键 词：自组织神经网络 测井相 曲线形态 模式识别 沉积环境 测井资料 

分 类 号：P631.8[天文地球—地质矿产勘探] TP183[天文地球—地质学]