Predictive Power Control for T-Type Nested Neutral Point Clamped Inverter Without Weighting Factor and Reduced Computational Burden

This paper offers a simplified method for direct power control employing model predictive control (MPC) for a four-level T-type Nested Neutral Point Clamped inverter-connected grid. A discrete-time model is used to forecast the grid current and power for evaluating the cost function. Unlike the traditional MPC, which utilizes exhaustive searches for all viable switching combinations, the proposed method only considers the possible candidates based on the needed inverter voltage sector for the optimization loop during the initial stage. During the second phase, the redundant switching status, which is considered the best output vector, will be employed to balance the capacitor voltages. Therefore, the computation load is not only dramatically reduced but also eliminated the weighting factors tuning by the suggested approach. The proposed method’s efficacy is substantiated through comparative simulation studies conducted under both steady-state and dynamic-state operations.