A Fast Dual Vector Model Predictive Voltage Control Strategy for the Permanent Magnet Synchronous Generator System Based on T-Type Three-Level Converter

In this paper, a fast dual-vector model predictive voltage control (FDV-MPVC) strategy is proposed for the permanent magnet synchronous generator (PMSG) system based on a T-type three-level converter, aiming to significantly reduce computational burden, eliminate the need for weighting factors, and enhance steady-state and dynamic performance while directly addressing the key challenges in conventional finite control set model predictive control (FCS-MPC) methods. First, inherent characteristics of redundant small vectors are utilized to achieve neutral point (NP) voltage balance, eliminating the requirement for the associated weighting factor. Second, based on the sectors of spatial voltage vectors and NP balance strategy, only two voltage vectors are selected from five candidate vectors, thereby avoiding the exhaustive evaluation of all feasible voltage vectors while improving computational efficiency and dynamic response. Finally, a modulation model is employed to calculate the duration for each selected vector, and an optimized vector selection strategy is introduced to minimize the switching frequency of the converter. A comprehensive analysis of the converter topology, the PMSG system model, and the proposed control system is presented. To validate the effectiveness of the FDV-MPVC strategy, a 10 kW T-type three-level converter prototype using Silicon Carbide (SiC) devices was developed. Both simulation and experimental results demonstrate that the FDV-MPVC strategy can efficiently achieve stable DC voltage output with enhanced steady-state accuracy and dynamic response.