PCIM INDIA - BEST POSTER AWARD
Real-Time EV Powertrain Fault Diagnostics Using IoT and FMEA- Guided Machine Learning Strategy
This paper presents a retrofittable, IoT-based onboard diagnostics device for electric vehicles, utilizing a decision tree algorithm based on FMEA principles as well as real-time multi-source data from motor, controller, and battery.
In combination with the monitoring of critical switching parameters such as VGS, VDS, and bootstrap voltage, Risk Priority Numbers (RPN) are computed for eight critical failure modes, from MOSFET failures to DC-link capacitor degradation. Insights in the system architecture as well as live measurement results from real driving operation are presented. The approach enables further innovation due to early fault detection, risk prioritization, and proactive maintenance in applications such as electric two-wheelers, e-mobility fleets, and predictive maintenance of EV powertrains. This paper showcases its capabilities in an actual electric vehicle with a cloud-connected IoT dashboard and automatic fault alerts via a mobile/web app.
The provider of this whitepaper
Download free whitepaper