Mobile QR Code
Title Multi-fault Mode Diagnosis System for EV PTC Heaters using VAE-based Generative Algorithms
Authors 조도현(Do-Hyeoun Cho)
DOI https://doi.org/10.5573/ieie.2025.62.12.143
Page pp.143-149
ISSN 2287-5026
Keywords Electric vehicles; PTC heater; VAE; Generative models; Fault diagnosis; Multiple failure modes
Abstract This study proposes a VAE-based generative algorithm for multi-fault mode diagnosis of electric vehicle PTC heaters. Four major fault modes - PTC element degradation, cooling system clogging, heater coil open circuit, and power supply anomaly - were systematically analyzed, and specialized VAE models were designed for each mode. Real-time diagnosis is performed utilizing multi-sensor data from temperature-humidity, MEMS vibration, and acoustic sensors. Experimental results demonstrate that the proposed system achieved 94.5% detection accuracy, an average improvement of 11.6% over conventional methods. Through VAE data augmentation, 89.7% accuracy was secured with only 50 rare fault data samples, and real-time processing capability was confirmed with an inference time of 2.3ms.