Model predictive control for single phase T-type neutral point clamping ZCZ-filtered inverters

The model predictive control (MPC) predicts all the possible behaviors and selects an optimal switching state which minimizes the predefined cost function at the next step, and is emerging as a powerful control method in power electronics. However, for a single phase T-NPC inverter with the conventional MPC method, only three different voltage vectors can be generated, which significantly impairs the system performance. Therefore, aiming at improving the current quality, a real-time duty ratio calculation scheme based on the MPC control is proposed. Moreover, the error of the two capacitor voltages is embedded into the current reference for balancing the capacitor voltage. The improved MPC control strategy achieves the balanced capacitor voltage and the minor steady-state error by providing another control degree of freedom. The effectiveness of the proposed MPC strategy is validated on a T-NPC LCL-filtered inverter by both the simulation and experimental results.