Dynamic Mode Decomposition Based Reduced-Order Temperature Field Prediction Method for IGBT Module

Detailed temperature field distribution of high-power insulated gate bipolar transistor (IGBT) modules is essential information for reliability analysis and thermal design of power electronic systems. This paper proposes a dynamic mode decomposition (DMD) based reduced-order temperature field prediction (RTFP) method for IGBT module. The IGBT temperature field (ITF) snapshot data obtained from finite element method (FEM) simulation is decomposed to obtain the dynamic modes. And then, the dominant modes among dynamic modes are chosen to establish a reduced-order model, which can quickly predict the ITF. The proposed DMD based RTFP method only requires a small number of sample snapshots to capture the dynamic modes of ITF and selects a small number of dominant modes to predict the ITF, and accordingly it improves the computational efficiency and reduces the disk space occupation. In addition, it does not require complex mathematical models and can obtain detailed ITF distribution. Finally, the accuracy of the proposed method is verified by simulation and experiment.