An Automatic Optimization Method for Electrical and Thermal Performance of stacked DBC Power Module

To meet stringent performance of electric vehicle controllers, it is highly needed to optimize the electrical and thermal performance for power semiconductor modules. As stacked DBC (Direct Bonded Copper) power modules have higher power density and enable the parallel of more chips, the demand for optimization is increasingly imperative. In this paper, an automated optimization method based on GA (genetic algorithm) and LBM (lattice Boltzmann method) was proposed for stacked DBC power modules with water-cooled PinFin heatsink. With proposed automatic generation method for electrical layout and heatsink structures, GA is used to select candidates. Because of the expanded optimization space, there is a growing requirement for faster and more precise evaluation methods. The proposed evaluation method based on LBM can assess layout parasitic inductance and the flow velocity and temperature distribution of power module with PinFin heatsink, which verifies an advantage in balancing speed and accuracy compared to traditional evaluation methods. According to optimization results, a 1200 V/ 1400 A, 16 chips in parallel, SiC MOSFET power module was fabricated in a standard EconoDUAL packaging size. Dynamic and static parameter testing as well as thermal testing were conducted, and experiment results demonstrate the effectiveness of proposed automatic optimization method.