Electro-thermal Models of Power Modules for Stochastic Optimization of Inverters

Stochastic optimization methods are commonly used in the design process of power electronic systems. These require models of crucial parts of the system which are both scalable and computationally efficient. This paper presents thermal models for traction inverter power modules based on scalable chip area that provide high speed of execution and accuracy. The thermal model includes chip-chip interaction, heat spreading and the effects of chip density inside the package. For single-sided modules the temperature error is within 5 % for single chip modules and 17% for multi chip modules compared to 3D FEM results