Dual-Vector Model Predictive Control for NPC Three-Level Inverters

To address the challenges of high current harmonics, complex multi-objective optimization, and high computational demands in three-level inverters using finite control set model predictive control (FCS-MPC), this study proposes a dual-vector control strategy. First, a finite state machine (FSM) is utilized to define the restricted control set and vector switching transitions, ensuring stable operation of the inverter by preventing excessive phase and line voltage jumps. Next, a cost function is optimized to identify the optimal and suboptimal voltage vectors, effectively resolving multi-objective optimization conflicts. The dwell time of these vectors is calculated based on their cost function values, enabling the simultaneous application of two voltage vectors in the next control cycle to reduce output current harmonics. A simulation platform (MATLAB/Simulink) and an experimental prototype (DSP28335-based) were developed to validate the proposed strategy. Both simulation and experimental results confirm the superior performance of the proposed method.