中文摘要：

针对目前大庆油田一类油层聚驱区块分析中，由于油层物性、原油性质、水驱开发阶段、聚合物溶液性质等指标相近，造成应用常规聚类分析进行区块归类评价困难，并且收敛速度慢的问题，利用墨西哥草帽函数对常规FCM算法进行改进．该（IFKN）方法在保证较高寻优概率基础上，对邻域调整进行侧抑制作用，加快了收敛速度，节省了计算时间12％以上．目前，通过大庆油田20个已完成注聚的工业化区块计算显示，结果可靠、快速，为下步二类、三类聚驱油层进行效果评价奠定了基础．

英文摘要：

for now, the reservoir＆oil characteristics, waterflooding status, polymer solution properties of the first class polymerflooding oil layers in Daqing oilfield are very similar, which causing the general clustering course difficult, and the rate of convergence slow. FCM algorithm was improved by importing Mexico strawhat function. This new method（IFKN） can assure a high optimization probability, and the convergence rate is faster by side suppression when adjusting neighborhood, consuming time was shorten over 12%. The results using this method in 20 polymerflooding blocks showed that, the results was reliable and fast. This method could build a foundation for effect evaluation of second or third class polymerflooding blocks.