Smart Universal Parameter Fitting Method for Modeling Static SiC Power MOSFET Behavior

For efficient converter design, modeling of switching components must be both accurate and fast. A variety of simulation models for SiC Power MOSFETs has been developed. To achieve adequate accuracy, models are adjusted to experimental data by applying fitting methods. However, proposed universal fitting methods result in high-dimensional optimization problems. Their computational inefficiency leads to either long run times or low accuracy. This paper proposes a new universal parameter fitting method that is at least twice as fast and yields a more accurate outcome than presented methods under comparable conditions. It effectively reduces the dimensionality of the parameter fitting optimization problem by sensitivity-based reduction of optimization parameters. Thereby, it optimizes both model adjustment speed and accuracy, which are critical factors for an efficient converter design process. The method does not compromise flexibility as it is universally applicable to any SPICE model and can meet any model architecture preference.