Automatic Fault Detection and Analysis Of Electric Vehicle Charging Station by Machine Learning

As the transition to Electric Vehicles (EVs) accelerates, enhancing the resilience of electrical systems becomes increasingly crucial. A key component in EVs is the Power Electronic Converter (PEC), which not only powers the electric motors but also reconfigures the grid charging of the battery bank. This paper introduces a fault detection system specifically designed for a prevalent type of charger used in EVs. The system aims to identify faults that may lead to malfunctions during the operation of the onboard electric charger, thereby enabling a supervisory system to execute corrective measures and maintain uninterrupted charging operations. The proposed fault detection method leverages a Machine Learning (ML) technique known as a classification algorithm. The system is trained using grid-side Total Harmonic Distortion (THD) and load resistance values to classify potential faults accurately. Extensive testing demonstrates that the proposed system achieves a fault prediction accuracy of around 94.59 %, showcasing superior performance in terms of training and execution times as well as overall accuracy. This study underscores the effectiveness of integrating advanced ML techniques in enhancing the reliability and efficiency of EV charging systems, thus contributing to a smoother and faster transition to electric mobility.