Model predictive control of three-phase voltage-source converters with improved tracking performance

The conventional finite control set-model predictive control (FCS-MPC) has not been completely embraced by the power industry because of large tracking errors and the high required sampling frequency. Therefore, the double-vector-based model predictive control (DVB-MPC), which enfolds the deadbeat control (DBC) theory, has gradually crept into researchers’ horizons in recent years. In the universal DVB-MPC (UDVB-MPC) scheme, selecting two operation vectors in a single sector is a major obstacle to achieving satisfactory tracking performance. There is still a biggish error between the synthesized output voltage and the reference value. To alleviate this issue, this paper proposes an improved double-vector-based model predictive control (IDVB-MPC) scheme. The tracking errors are reduced dramatically by adding two extra vectors in adjacent sectors to the candidate set and dividing a sector into eight zones. Then the power fluctuation is reduced and the grid-connected current quality is improved. To verify the effectiveness of the proposed scheme, the simulation and experiment comparisons between UDVB-MPC and IDVB-MPC are carried out. The results show that IDVB-MPC reduces the quantitative tracking errors, the THD of the output currents, and the quantitative active power ripple by about 50%, 20%, and 35% with UDVB-MPC as a benchmark, respectively. • An algorithm based on model predictive control and space vector modulation. • The proposed scheme can be applied to three-phase voltage-source converters. • Using a low-complexity procedure to reduce tracking errors. • The simulation and experiment are performed to verify the improved tracking performance.