A Model-Based Multi-Objective Optimized Active Gate Driving Method

The rapid switching of SiC MOSFETs causes severe overshoots, while parasitic uncertainties lead to model mismatch. This paper proposes a hybrid active gate driving strategy integrating in-situ parameter identification with multi-objective optimization. First, a dual-variable controller achieves orthogonal decoupling of gate current amplitude and duration. Second, a systematic initialization strategy is developed: particle swarm optimization (PSO) calibrates parasitic parameters in-situ to ensure model accuracy, followed by NSGA-II for global optimization. Third, a runtime mechanism using bilinear interpolation maps optimized solutions to real-time control, resolving the conflict between computational burden and response speed. Experimental results demonstrate that the proposed method achieves a superior overshoot-loss trade-off and exhibits robust engineering tolerance against parameter deviations.