基于对应分析理论的油气储层流体识别新方法

NEW METHOD OF RESERVOIR FLUID IDENTIFICATION BASED ON CORRESPONDENCE ANALYSIS THEORY

储层流体性质的识别,在统计学方面出现过模糊聚类分析、判别分析、正态概率分布法以及因子分析等。使用这些方法进行判别,其结果大部分都是定性或者半定量的,而对应分析方法是在因子分析基础上发展起来的一种多元统计分析方法,该方法充分利用测井、录井、试油等方面的信息,把R型与Q型因子分析结合起来,对变量与样品统一进行分析研究。根据对应分析原理,利用VB6.0语言编程对吐哈油田塔河地区17口井25层进行了分析研究,其判别结果经试油资料证实,符合率可达88.0%,实现了对储层流体性质的有效识别。

There are some methods including fuzzy clustering analysis,discriminant analysis,normal probability distribution and factor analysis in statistics for discriminating reservoir liquid property. The results of using the methods to discriminate reservoir liquids are qualitative or semi-quantitative. However,correspondence analysis is a method of multivariable statistical analysis,based on factor analysis. The variables and samples are unified in this model in which information of logging and testing for oil is adequately used and R factor analysis and Q factor analysis are united. According to correspondence analysis principle,the thesis analyzed and studied the 25 layers of 17 wells of TUHA oilfield by using VB6.0. 88% of the results of Correspondence Analysis are right,which can be proved by testing for oil information. It makes efficiently judgments in discriminating reservoir fluid property.