模糊聚类的最大树法在机械运行状态中的应用

The Method of Maximum Tree Based on Fuzzy Classification and Application in Monitoring and Diagnosis of Machinery

采用模糊数学中的基于模糊相似关系的聚类方法——最大树法,以相似矩阵代替传递闭包对实际机组运行状态进行聚类分析,并利用模糊识别技术进行故障分析。通过一段时间的实践,取得了较好的效果。该方法简单、实用,它一方面减少了求传递闭包的工作量,另一方面对计算机在线监测的机组束说,使用该方法分类、识别后,可只保留同类型的一组数据,数据有效压缩,大大节省了计算机的内存空间。

This paper introduced fault pattern methods of maximum tree in Fuzzy sets and recognition.By using Fuzzy similarity matrix instead of trasitive closure,the method of maximum tree can not only reduce calculative quality than using trasitive closure but also save computer memory in monitoring of on —time machine —set.The methods used in real machine sets have obtained better results than ever.