A Modeling Approach for Power Device Junction Temperature Estimation in Vehicle Mild Hybrid Power System Leveraging Layer-Anchored Cauer-to-Foster With Control Algorithm-Based Power Losses Calculation

The junction temperature estimation of power semiconductor devices is crucial for reliability and performance in a wide range of power electronics applications, including vehicle power systems. This paper proposes a compromise solution between estimation accuracy and computational cost in specific vehicle mild hybrid power systems, leveraging Layer-anchored Cauer-to-Foster (LACF) with control algorithm-based power losses (CAPL) calculation. The proposed model includes the application scenario, electrical model and thermal network model. First, the power losses are calculated according to the electrical stress of power devices that are mapped from system level operating conditions with motor control algorithms by building an application scenario electrical model. Second, the Cauer thermal network model is constructed based on the 3D geometric model of power module and power circuit. Third, the thermal parameters of the Cauer network are extracted using a FEA-based, data-driven procedure and then converted into Foster-network parameters. Finally, the calculated power losses and the Foster-network parameters are used to compute the junction temperature of power devices. The proposed approach is validated by both simulations and experiments. The results from the validation show that the maximum estimation error is within 2.4 °C and the proposed method is effective for power device junction temperature estimation. The proposed modeling approach is in a model-based form and can be converted to program code for on-line estimation on embedded computing hardware. It can be extended to other power device-based applications.