Authors: Afrah Chopdar, Anubrata Das, Ms Khushi, Associate Professor Dr. Dhanamjayulu C
Abstract: Electric vehicles (EVs) are transforming transportation by offering high efficiency and sus-tainability. However, their powertrains—particularly motors, inverters, and control units—require reliable fault detection mechanisms to prevent sudden failures and ensure safety. This project proposes a simulation-driven Machine Learning (ML) approach for Fault Detection and Diag-nosis (FDD) in EV powertrains. Drawing on prior work by Khan et al. (2024) and Zhang et al. (2022), this study develops a simplified, reproducible framework for academic use. The methodology involves synthetic signal generation of a Permanent Magnet Synchronous Motor (PMSM), fault injection, statistical feature extraction, and classification using Support Vector Machine (SVM) and Random Forest (RF). Results show over 90% classification accuracy with low latency, validating the feasibility of lightweight ML in real-time embedded EV applications.
International Journal of Science, Engineering and Technology