Research on mechanical fault diagnosis of Ultra high voltage GIS based on the combination of neighbor algorithm and FCM

Ultra high voltage Gas Insulted Switchgear(GIS) plays an important role in the switch, control, protection. Through studying on Ultra high voltage GIS, we can realize the online real time state of the ultra high voltage GIS. This paper proposes an approach to distinguish the mechanical faults combined with the neighbor algorithm and the fuzzy C means algorithm. The new algorithm takes the advantage of both methods. It calculates the initial cluster centers by means of fuzzy C-means algorithm and uses neighbor algorithm to get the final classification results. It can not only solve the problem of storage requirements, large amount of calculation and long computing time in neighbor algorithm, but also overcome the inaccuracy of cluster centers obtained from FCM which results in low accuracy of GIS mechanical fault diagnosis. We use a dataset of a vacuum circuit breaker and GIS in the experiments to evaluate the three methods. The experimental results show that using the new mechanical fault diagnosis method obtains the highest accuracy. Therefore, the new algorithm proposed in this paper is more effective than the traditional algorithm and it is more suitable for the mechanical fault diagnosis of GIS.