用ISODATA统计法识别碳酸盐岩气水层

PICKING OUT THE GAS/WATER BEARING FORMATION FROM CARBONATE ROCK USING THE ISODATA STATISTICS SCHEME

成都地质学院

用常规测井资料和传统的解释方法来识别气、水层,在储层性质复杂、地层流体性质多变等情况下,其难度很大,精度不高。本文中,我们采用了动态聚类算法与模糊理论相结合的ISODATA统计法,根据六条常规测井曲线,综合从不同侧面反映储层性质的多种参数,达到自动识别碳酸盐岩地层的储集层及气、水性质的目的。用这种方法,我们对数口井作了验证和预测,得到了满意的结果。

It is worldwide acknowledged as a pretty tough problem and produces poor dep endability while recognizing the gas / water bearing zone under the circumstances of complex reservoir and varied fluid. What’s presented in the paper is, based on 6 conventional well logs and using the ISODATA statistics, which is a dynamic classification algorithm integrated with fuzzy theory, to synthesize the multiple parameters reflecting the characteristics of reservoir from different indirect sources with the aim of being capable of identifying spontaneously the reservoir of carbonate formation as well as the properties of gas/ water. By utilizing such an approach, a number of wells have been verified and predicted. And thus a satisfactory result is achieved.