应用人工神经网络识别水淹层  

Identification of flooded Formation withArtificial Neural Network

作　　者：张玎[1] 梅红 冉文琼 

机构地区：[1]江汉石油学院物探系,长庆石油勘探局测井工程处

出　　处：《测井技术》1996年第3期210-214,共5页Well Logging Technology

摘　　要：张玎冉文琼等：应用人工神经网络识别水淹层，测井技术，１９９６（３）２０，２１０～２１４。油层水淹后，水淹层的内部物性明显区别于油层的原始状态。由于水淹状况复杂多变，使用一般测井解释方法识别水淹层具有很大困难。本文研究了应用人工神经网络ＢＰ算法处理测井资料来识别水淹层的基本原理和具体实现，给出了ＢＰ算法的改进方法及在两个油田的实际应用效果。Because of water encroachment,the petrophysical properties of watered--outzone are quite different from that of the original oil bearing zone.It is difficult toidentify the flooded formation with the conventional interpretation methods sincethe conditions of the flooded formation are complicated and changeable. The basicprinciple and the specific procedures to identify flooded formation with the improvedBP algorithm of artificial neural network are detailed in the paper.How to improveBP algorithm and its application results in two oilfields are also given, which aredesirable.

关 键 词：神经网络 水淹层 收敛(数学) 模式识别 误差 测井资料 

分 类 号：P631.84[天文地球—地质矿产勘探]