An Improved Finite-Set Model Predictive Control Based on Discrete Space Vector Modulation Methods for Grid-Connected Three-Level Voltage Source Inverter

This paper proposes an improved finite-set model predictive control (FS-MPC) to reduce current ripples and balance the neutral-point voltage for grid-connected three-level voltage source inverter. The proposed FS model predictive current control is based on the discrete space vector modulation (DSVM) to synthesize various voltage vectors. The proposed DSVM method enables to calculate all of the virtual voltage vector values immediately without the lookup table. To overcome the excessive computation burden problem, the optimal region determination strategy is presented in order to use only minimum number of voltage vectors for the cost function calculation. To balance the neutral-point voltage, the proposed offset voltage injection (OVI) method based on the FS-MPC is introduced in this paper. Various offset voltage values are obtained and an optimal offset voltage value is determined using the MPC algorithm. Since the proposed control method is based on space vector diagram without being restricted by the switching states, it can be easily applied to a multilevel system or other applications. The simulation and experiment results are presented to validate the effectiveness and performance of the proposed method using the grid-connected three-level T-type voltage source inverter.