Fatigue lifetime prediction of power modules under power cycling using a volume-weighted averaging technique

With the continuous increase in power density of electric vehicle (EV) power modules, double-sided cooling (DSC) technology has become a focal point of research. However, the commercialization of DSC power modules is significantly hindered by their complex manufacturing processes and high development costs. Traditional power cycling reliability assessment methods, which are costly and time-consuming, are not conducive to rapid technological iteration and advancement. This study proposes a lifetime prediction model based on volume-weighted average strain energy density, utilizing a global volume-weighted averaging technique rather than focusing on the maximum stress and strain locations at solder joints. Through least-squares curve fitting analysis, numerous volume-weighted average strain energy density simulation results were correlated with corresponding power cycling test results, allowing the determination of material-specific fatigue coefficients in the lifetime prediction model. The feasibility of the proposed model was validated using commercially available insulated-gate bipolar transistor (IGBT) power modules. Furthermore, the influence of different buffer materials on the lifetime of DSC IGBT power modules was thoroughly evaluated. This study provides valuable insights for accelerating the development of DSC power modules in the future.