Thermal Impedance Parameters Identification for Power Modules Based on Measured $T_{j}$

Accurate thermal analysis of power semiconductor devices is critical for evaluating reliability and operational lifetime. However, achieving high-fidelity thermal analysis remains challenging due to two key obstacles. First, the thermal resistance between the semiconductor junction and the environment cannot be directly obtained by the user. Second, although thermal cross-coupling within the power module significantly affects the temperature distribution, this phenomenon is often neglected in manufacturer datasheets. To overcome these limitations, this study proposes a thermal network-based parameter identification method that explicitly incorporates thermal coupling effects. The approach enables accurate extraction of coupled thermal parameters through a physically constrained identification framework combined with a collaborative multi-condition identification strategy. The validity of the identified parameters is verified by comparing the predicted junction temperatures with experimental measurements under different operating scenarios.