부산대학교 도서관

Active gate drive (AGD) is one of the key techniques to utilize SiC power devices' fast-switching capability without compromising the drawbacks. However, the increasing complexity in the circuit design makes it hard to have an optimized operation without expert know-hows or considerable design efforts. This article demonstrates a digital-twin-compatible metaheuristic optimization system for a digital AGD. It offers a totally-digital control of switching characteristics of power devices through genetic-algorithm-based optimization. An individually developed digital AGD is adopted to generate a genetically expressed gate-voltage waveform by a multibit gate signal sequence. The optimization system is verified in simulation and experiment by successfully obtaining the optimum Pareto-front solutions, which clarify the relationship between the selected gate voltage levels and the device characteristics. The optimization is also performed for variable operating conditions and for different SiC mosfets. The results of this article offer feasibility for a software-based optimization of gate drive design.