A Self-Healing Prognostic Framework for Fault-Resilient SVPWM-Based Electric Vehicle Traction Inverters

Abstract

This paper presents a Self-Healing Prognostic-Enabled Power Module Reconfiguration (SH-PEPMR) framework integrated with a three-phase Space Vector Pulse Width Modulation (SVPWM) traction inverter for electric vehicle applications. The proposed method is numerically evaluated on a 100-kW inverter with a 600-V DC link supplying a permanent magnet synchronous motor under healthy, degraded, and fault-injected operating conditions. Device degradation is emulated by a 30% increase in ON-state resistance and a 25% increase in switching delay, while an open-switch fault is introduced at 0.5 s. Under degraded operation without self-healing, output voltage total harmonic distortion increases from 2.1% to 6.8%, and peak junction temperature rises from 74 °C to 102 °C, with the prognostic health index (PHI) dropping from 1.0 to 0.4 within 2,150 operating hours. When SH-PEPMR is activated, adaptive SVPWM micro-reconfiguration reduces effective switching stress by 25%, restores voltage and current waveforms, and limits THD to 2.5% after fault occurrence. Junction temperature rise is reduced by approximately 20–25 °C, while inverter efficiency remains above 95%. Prognostic analysis indicates a 35–45% extension in remaining useful life, with PHI maintained above 0.6 beyond 3,000 operating hours. The numerical results demonstrate that prognostics-driven self-healing control significantly improves reliability and lifetime of EV traction inverters without hardware redundancy.