Wavelet Transform and Least Square SVM Technique for Zone and Fault Classification in Electrical Transformer

This article offers an innovative method for power transformer fault zone identification and fault type categorization. A hybrid technique considering wavelet transform (WT) and least square support vector machine (LSSVM) has been applied for a successful outcome. The suggested integrated WT-LSSVM classifier runs using the current signal acquired from the transformer's main and secondary. PSCAD software is used to model the power system simulation, while MATLAB is used to develop the method. Numerous in-zone and out-of-zone faults were created for the validation of the algorithm. Around 12000 fault cases have been generated by altering the system and fault parameters. It has been found that the proposed fault classifier scheme is precise and faithful in the presence of varying system and fault scenarios. The suggested scheme provides classification accuracy of more than 99% in terms of zone identification and fault categorization. Thus, the outcome validates the efficacy of the suggested scheme for accurately classifying power transformer failures.