基于遗传算法的SiC MOSFET导通电阻模型

提出了一种简洁、新颖的SiC MOSFET器件导通电阻模型,该模型采用遗传算法对其不同温度下的导通电阻进行准确描述。相比于传统的导通电阻建模方案,采用进化算法可以准确地得到MOSFET导通电阻与结温之间的关系。并探究了不同种群规模、交叉率和变异率对算法的影响。为了验证模型的准确性,采用一款自主封装的1 200 V/90 A SiC MOSFET模块去验证模型的静态特性,并与传统导通电阻建模方案进行对比,其最大误差为4.1%。

Genetic Algorithm Based SiC MOSFET On-state Resistance Model

In this paper, a compact and novel SiC MOSFET on-state resistance model is proposed. The presented model is evaluated by fitting the on-state resistance of SiC MOSFET under different junction temperature using Genetic Algorithm (GA). Compared to conventional on-state resistance fitting methods, the relationship between on-state resistance and junction temperature can be accurately expressed by using evolutionary algorithm. The influences of different population size, crossover rate and mutation rate on the algorithm is investigated. To verify the proposed model, the static characteristics of a self-packing 1200V/90A SiC MOSFET are tested under different junction temperature, and compared with conventional methods, the maximum error is 4.1%.