基于动态模糊聚类分析的模糊集划分新方法

摘要

由于模拟电路具有容差特征,用故障字典法进行故障诊断时,传统的模糊集划分方法就存在弊端.针对这种弊端;采用模糊理论中的动态聚类分析法来改善传统模糊集的划分;首先用蒙特卡罗分析法进行电路容差分析,然后用动态模糊聚类分析方法,对电路故障状态进行分类,形成新的模糊集划分方法;实验结果证明了新划分方法的可行性和有效性,并显著提高电路故障的检测率和隔离率.%As the characteristics of analog circuits with tolerance, using fault dictionary for fault diagnosis, in the traditional division method on fuzzy sets existing drawbacks is unreasonable. In response to this unreasonable, use dynamic cluster analysis method to improve the traditional division of fuzzy set. The first, Monte Carlo analysis to the circuit tolerance analysis, then the dynamic fuzzy clustering analysis method to classify the circuit fault condition, the last, the formation of a new fuzzy partition method. Experimental results show that the new division of the feasibility and effectiveness.