Switching Loss Model for SiC MOSFETs Based on Datasheet Parameters Enabling Virtual Junction Temperature Estimation

SiC MOSFETs are widely employed in power converters due to their superior efficiency and reliability at high temperatures. For this reason, it is crucial to implement accurate thermal models capable of indirectly estimating the junction temperature and its fluctuations: both are caused by power losses in the device. In this framework, the evaluation of switching losses remains the most challenging task. To enable real-time monitoring of the junction temperature, this work presents the development of a virtual sensor specifically designed for SiC MOSFETs. The sensor relies on a num-analytical model (NAM), which employs only datasheet parameters and leverages electrical quantities-namely, bus voltage and current-available from sensors integrated into power converter systems. The proposed NAM is implemented in MATLAB using an iterative algorithm that accounts for the main physical phenomena involved in switching transitions. The computed energy losses are then used to thermally model the SiC MOSFETs within the PLECS environment, where a digital twin of an all-SiC board is created. Finally, the accuracy of the model is validated by comparing simulation results with experimental efficiency data obtained from a real half-bridge converter, with explicit consideration of measurement uncertainty.