A New Transformer Faults Diagnosis Method based on an Improved Fuzzy C-Means Clustering Algorithm

The correct functioning of transformers guarantees the proper operation of the power sector. The accurate diagnosis of the problems associated with the transformers consider the basis for the maintenance of the transformers and the possibility of continuing their proper operation and functioning. There are many methods used for this purpose; one of them is the Three-Ratio Method. This manner is commonly used for fault diagnosis by emitting the gas Analysis of the Oil-Immersed transformer. Occasionally, these values don't align with the specific type of factual fault. The Three Ratio Method, which utilizes the Fuzzy C-Means Clustering (FCM) algorithm, is a suitable approach for addressing the issue in the classical Three Ratio Method. However, this algorithm suffers from the random initial selection problem. This paper deals with the defect of random initial selection of the FCM algorithm to improve its performance for accurate transformer fault diagnosis. For the evaluation, we compare the proposed method with the conventional FCM and Improved FCM (I-FCM), where several experiments are executed on some sets of the common IRIS dataset. The results show that the proposed method is overall better than the I-FCM and the conventional FCM.