Research on transformer fault diagnosis based on fuzzy c-means and improved normalization method
Cui Qing1, Fang Xin2, Zhang Zhilei1, Wang Tao1, Zhang Tianwei3
1. Hebei Electric Power Company Shijiazhuang Power Supply Company, Shijiazhuang 050051; 2. College of Electrical Engineering, Northeast Electric Power University, Jilin, Jilin 132012; 3. Beijing Runwei Tianhua Power Technology Co., Ltd, Beijing 102211
Abstract：Dissolved gas analysis (DGA) is an important method for transformer fault diagnosis. A transformer fault diagnosis model based on fuzzy c-means algorithm (FCM) is established in this paper. In order to study the influence of the different normalization methods of the sample (considering the sensitivity of different gas reaction faults) on the clustering results in the FCM algorithm model, firstly, three methods are used to normalize the dissolved gas component samples, including deviation standardization, general concentration normalization method and characteristic concentration normalization method. Then the normalized sample is used as the input of the FCM algorithm, and the fault type of the sample is determined by the obtained membership matrix. The example calculation results show that the normalization of characteristic concentration can improve the accuracy of fault diagnosis.
2019, Vol. 20 
