An improved Multi-objective Finite Control Set Model Predictive Control for Grid Connected Three-level Neutral-Point Clamped Inverter

This paper presents an improved multi-objective finite control set model predictive control (MO FCS-MPC) for grid connected three-level neutral-point clamped (3L-NPC) inverter. The FCS-MPC is based on the prediction of the grid current and DC-link capacitor voltage used the discrete-time model of 3L-NPC inverter. A cost function is defined to compromise the control objectives: control of grid current, balance the DC-link capacitor voltage and reduce the switching number. The cost function is evaluated each sampling period of all switching states (SSs) of the inverter (27 SSs of 3L-NPC) to select the optimal SS that will be applied in the next sampling period. The computational burden will increase exponentially with the system complexity. To overcome this issue, we proposed to select the sector of the inverter output voltage's reference using its angle. Consequently, the cost function is evaluated only 4 instead 27 and 6 times in the standard and efficient FCS-MPCs, respectively. Its performance improvements in comparison with the efficient MO FCS-MPC are validated with simulation results.