AI Based Incipient Fault Detection and Monitoring for Prevention of Powertrain Failures in EVs

Electric vehicles are gaining popularity but face many challenges, A critical concern for potential buyers is the risk of powertrain failure, which can result from faults in key components like the motor, battery, and inverter. These failures sometime require expensive repairs or replacements, making them economically unfeasible. Traditional safeguards such as pyro fuses only protect against catastrophic failures but do not address incipient stage of fault that gradually degrade component health over time. This research proposes an AI-based system for online fault identification and condition monitoring of EV powertrains during startup. The system employs artificial neural networks to detect faults at their early stages, providing detailed information about the faulty components via a warning notification. This approach aims to prevent further damage by disconnecting power supply immediately when fault occurs, thereby extending the lifespan of EVs and reducing maintenance costs. The study evaluates the performance characteristics ANN in classifying different types of faults, highlighting their effectiveness in early detection and condition monitoring.