Optimized model predictive control with reduced switching states for neutral point clamped converters

The implementation of multilevel converters, particularly the Neutral-Point Clamped (NPC) converter, has risen in modern power electronics owing to its ability to enhance power quality and voltage levels. Model Predictive Control (MPC) has emerged as an important strategy in NPC converter control, offering empowerment in managing multiple objectives with rapid dynamic responses. However, conventional MPC suffers from computational burdens, delaying its practical implementation. This paper proposes a Reduced Switching States MPC (RSS-MPC) approach for NPC converters, significantly reducing computational time by segmenting the prediction horizon into discrete zones with fewer potential switching states. Simulation results demonstrate the effectiveness of the proposed approach in achieving control objectives, maintaining robustness against disturbances, and variations in grid voltage, thereby presenting a promising solution for practical implementation in industrial applications.