Classification of transformer faults using frequency response analysis based on cross-correlation technique and support vector machine

One of the most important methods for transformers fault diagnosis (especially mechanical defects) is the frequency response analysis (FRA) method. The most important step in the FRA diagnostic process is to differentiate the faults and classify them in different classes. This paper uses the intelligent support vector machine (SVM) method to classify transformer faults. For this purpose, two groups of transformers have been tested. First, the necessary measurements are performed on the model transformers under healthy conditions and under various fault conditions (axial displacement, radial deformation, disc space variation, short-circuits, and core deformation). Then, by dividing the frequency ranges of the measured transfer functions of the transformer, a new feature based on the cross-correlation technique is proposed for SVM training and validation. After the training process, by applying the data obtained from real transformers, the performance of SVM in different modes is evaluated and compared. Finally, the most appropriate feature has been provided.