Modeling and Validation of Junction Temperature Estimation in High-Power SiC MOSFET Inverters for Electric Vehicle Applications

The growing concern with reducing CO2 emissions and the demand for sustainable mobility have promoted the adoption of electric vehicles and the use of wide bandgap semiconductors. Thermal management for semiconductor devices is crucial to ensure reliability and efficiency, with junction temperature being a key variable that cannot be directly measured during operation. In this context, this paper proposes a method for junction temperature estimation of SiC MOSFET devices applied to electric traction inverters. A Luenberger state observer is applied to estimate the junction temperature from accessible variables such as ambient temperature, heat sink temperature, and calculated power losses. The method is formulated in a state-space representation, verified through simulations, and validated using an experimental setup, consisting of a single-phase full-bridge inverter. Infrared thermal imaging and physical sensors were employed to compare estimated and measured values. The results demonstrated high accuracy and fast dynamic response of the proposed observer, confirming its robustness and applicability for embedded thermal monitoring in power converters.