Parameters Optimization Of A Behavioural Spice Model Of An Automotive Grade Sic Mosfet Using Particle Swarm Optimization Algorithm

Power device modeling plays an important role at the design stage of power conversions systems, especially in the automotive and aircraft sectors. In this context, it is important to obtain in short time an accurate model that account for the mutual interaction of thermal and electrical quantities. The behavioral model is the best approach because it consists of a set of equations that emulates the device behavior at terminals level and can be easily implemented in a SPICE simulation tool, largely adopted both in industrial and academic contexts. The major issue of the behavioral model is to choose the equations parameters such that the simulations results fit the measurement data. In this paper, this problem is tackled by means of a particle swarm optimization algorithm. It is applied to automatically find the optimal parameters for an accurate behavioral model of an automotive grade SiC MOSFET that properly fits the datasheet curves. The comparison between the waveforms obtained by experimental tests and the SPICE simulations in a double test has confirmed the accuracy of the model.