Transformers Faults Prediction Using Machine Learning Approach

The maintenance of transformers is crucial for ensuring their proper functioning. Due to the high expenses associated with maintenance, finding alternative methods to maintain these expensive electrical components has become a priority, as opposed to relying solely on traditional methods. In this paper, Machine Learning algorithms were used for fault prediction in transformers. These algorithms were trained using measurements data of the three-phase currents and voltages. Several algorithms were employed and evaluated to determine the performing ones. Results show that K-Nearest Neighbor algorithm and Decision Trees gave the best accuracy.