摘要
本文在总结四川气田当前判别气水层测井解释方法的基础上,提出了利用ISODATA模糊聚类分析判别气水层的方法.该方法利用双侧向、声波时差、中子、密度和自然伽玛等常规测井资料,提取出7个对判别地层储集流体性质有利的聚类参数,对单井样本集或地区性标准样本集进行模糊聚类并求出聚类中心,同时得到各样本隶属于气、水、干、泥4种类别的模糊隶属度.这种方法的判别结果,受测井响应方程及m、R_w等解释参数的选取影响非常小。由于在计算程序中加入了根据地区实践总结的若干专家经验和限制条件,因此在四川地区裂缝—孔隙型碳酸盐岩储集层中判别含流体性质取得了明显的效果,在18口样本井中符合率达98%,在10口预测未知井中判别气、水层,经测试验证符合率达90%,并能解释一些常规解释所错判的疑难井问题。
This paper presents a method for distinguishing gas-and water bearing beds withISODATA clustering analysis derived from the log interpretation technology currently usedin Sichuan. This method extracts 7 clustering parameters which are useful to fluidcharacterization from the conventional log information such as Dual Laterolog, Sonicelapsed time, Density, Neutron and Gamma Ray. It clusters fuzzily sample set acquiredfrom single well or standard regional sample set and finds out the clustering center. It alsopoints out to which extent those clustering points represent gas zone, water zone, dry zoneand shales. With this method the distinguished result is less affected by the selection of log re-sponse equations and such parameters as m, R_w etc. Since the local geological informationand limitation are incorporated into the program, good results have been obtained with thismethod in discriminating the fluid property of fractured carbonate reservoir in Sichuan.
