模糊聚类分析及其在测井识别油气层中的应用

Fuzzy cluster analysis and its application in identification of reservoir bed with logging data.

介绍了基于模糊等价关系的“传递包”聚类和基于软划分的模糊ISODATA聚类2种模糊聚类分析方法的 基本原理,充分利用2种方法的优点,提出了一种将K-L变换与上述2种模糊聚类方法相结合的分类识别方 法。应用该方法对东濮凹陷深部的气层和干层进行了分类识别,结果表明,该方法能够较好地提高分类的准确 性。

This paper briefs the basic principles of ' transitive closure ' cluster analysis based on fuzzy equivalent relation and fuzzy ISODATA cluster analysis based on soft division technique, and advances a classification and recognition method by combining K-L transformation and the two fuzzy cluster methods above, which makes full use of the advantages of both cluster methods. The method was applied to classify hydrocarbon-bearing beds and hydrocarbon-free ones from deep Dongpu depression. The resulted classifica- tion was verified by the well testing data of this area.