Classifying Transformer Winding Fault Type, Location and Extent using FRA based on Support Vector Machine

In this paper, four common winding faults in power transformers (axial displacement (AD), serial capacitance variation (VSC), ground capacitance variation (VGC), open circuit (OC)) are simulated on a transformer winding model to classify the fault type, location and extent, by applying an intelligent methodology for diagnosing transformer faults, depends on building a comprehensive database by collecting Frequency Responses Analysis (FRA) related to health and faulty conditions and analyzing them using statistical and mathematical indicators, this base that can inventory all possible faults in terms of location and extent, which is used to train a support vector machine (SVM) classifier on the faults included in it, which is then able to classify any new data. The results of the tests showed that the proposed method is characterized by high accuracy in detecting the type of defect, determining its location and the extent of its occurrence, It also contributes to the development of the application of machine learning on transformers.