Fault Diagnosis Of Electric Apparatus Based On Fuzzy Clustering Analysis

With the development of the electric apparatus, the traditional way cannot fill the urgent need of rapidness and validity in modern fault diagnosis technique field. Now this diagnosis technique is developing to a new phase, the intelligent diagnosis phase. How to finding a new diagnosis method to make full use of the expert's experience and those diagnosis theories is very important. Based on this idea, in this paper fuzzy clustering analysis is used to diagnose electric apparatus fault.The theory of the fuzzy clustering analysis based electric apparatus fault diagnosis technique is to classify the object by faults according to the relationship or the comparability between the objects. There are two kind of fuzzy clustering analysis used in this paper: the method transitive closure and ISODATA. The former one can cluster fast but the result is not good enough;the latter can get a good result, but the original matrix has a great effect on its result. So in this paper these two methods are used together to diagnose the fault, which means the method transitive closure is used to diagnose and the ISODATA to validate the result. The concept of weight also is used to improve these two methods. In this way the result of the diagnosis can have a better accuracy. Two cases are used to be diagnosed by this method, the result it get is the same to the traditional ways, and this method is proved to be the effective.