人工神经网络在测井解释中的应用  

APPLICATION OF ARTIFICIAL NEURAL NETWORKS IN LOG ANALYSIS

作　　者：刘争平[1] 何永富[1] 

机构地区：[1]成都理工学院应用地球物理系,成都理工学院计算中心

出　　处：《地球物理学报》1995年第A01期323-330,共8页Chinese Journal of Geophysics

基　　金：国家自然科学基金

摘　　要：定量储层参数计算和定性岩性类型识别是测井解释的主要地质目的．应用人工神经网络计算方法，解释了多个油区的６０多口井的测井资料．结果表明，该计算方法不仅可以克服以上两类问题在传统测井解释方法中所遇到的高度复杂非线性建模困难，而且还可以高度统一地处理这两类问题，极大地简化了解释中的数学手段．The quantitation calcuation of reservoir parameters and qualitative recognition of lithology are two major goals of geological interpretation from well logs. We process and analyze the logging data from over 60 wells in several oil fields based on BP mdel. The interpretation results have Proved that the new algorithm can not only overcome the difficulty of highly complex non-linear model of traditional well logging interpretation, but also highly unify the two methods which compute the reservoir parameters and recognize the lithology. The new algorithm has greatly simplified the mathematical means in the traditional interpretation.

关 键 词：测井 神经网络 非线性 BP模型 

分 类 号：P631.81[天文地球—地质矿产勘探]