Two-step Model Predictive Power Control for a Single-Phase NPC Inverter with Fixed Switching Frequency and Optimal Switching Sequences

Traditional model predictive power control (MPPC) has outstanding dynamic and static performance. However, this kind of control method is difficult to be applied into single-phase system and suffers from variable switching frequency and heavy computational burden. Thus, this paper presents a novel two-step MPPC for a single-phase neutral point clamped (NPC) inverter with fixed switching frequency and optimal switching sequences (OSS). A virtual axis is formed by using second-order generalized integrator (SOGI), in order to realize active and reactive power estimation of the inverter system. The proposed MPPC method is based on two-step model prediction, which involves two cost functions. The control objective of the first cost function is to regulate the power from the inverter system to the grid, while the control objective of the second cost function is to balance the neutral-point voltage. The most suitable control vectors are selected and their effective time are determined by solving these two cost functions. Then, specific patterns of switching sequence are set up to achieve symmetric PWM control signal. Compared to traditional MPPC method, the proposed 2-step MPPC method has constant switching frequency and reduced computational amount. Simulation and experimental results verify the effectiveness and excellent performance of the proposed control method.