Model Predictive Current Control With Reduced Complexity for Five-Phase Three-Level NPC Voltage-Source Inverters

Finite control-set (FCS) model predictive control (MPC) is an attractive control technique, which exhibits fast response, intuitive principles, and multivariable control. However, the application of MPC on multilevel and multiphase traction systems is challenging due to the heavy complexity level. This article proposes a reduced-complexity MPC scheme for five-phase three-level (3L) voltage-source inverters (VSIs). In the proposed method, the control set consists of one zero vector and 80 virtual voltage vectors (
$\text{V}^{3}\text{s}$
), which are designed from the eligible 113 basic voltage vectors to eliminate the auxiliary (
$x$
–
$y$
) subspace harmonics. Then, the deadbeat (DB) solution is adopted to reduce the enumeration process from 81 to only four vectors during every sampling interval. Furthermore, the hysteresis control is adopted to balance the NP voltage without weighting factors. The proposed control scheme is compared with the existing FCS-model predictive current control (MPCC) schemes and the conventional proportional–integral (PI) controller with space-vector pulsewidth modulation (SVPWM). Simulation and experimental results verify that the proposed scheme achieves superior steady-state performance and fast dynamic response without parameters or weighting factors tuning, and it can be implemented in a low-cost digital signal processor (DSP) TMS320F28335 for a five-phase 3L-VSI prototype.