Enhancing Electrical Transformer Fault Prediction with Deep Learning: A Focus on ANN-Based Classification

Electrical transformers play an essential role in the power distribution network, acting as key connectors that facilitate the flow of electricity over large distances to meet diverse power needs. Despite their reliability, the rare occurrence of faults emphasizes the necessity for diligent maintenance to maintain their peak performance. The significant expenses linked to both the maintenance and repair of these components highlight the imperative for innovative maintenance strategies that extend beyond traditional approaches. This situation underscores the urgent need for maintenance strategies that surpass conventional methods. Our study leverages two unique datasets: one featuring data on three-phase currents and voltages and the other including seven essential indicators of transformer health. To predict faults within these datasets, we employ advanced deep learning techniques, specifically using Feedforward Back-propagation neural networks, demonstrating the potential of modern methodologies in enhancing the reliability and efficiency of electrical transformers.