Multistep Model Predictive Control for Permanent Magnet Synchronous Machine

Jianqiao Zou,Wei Xu,Yi Liu,Chaoxu Mu

DOI: https://doi.org/10.1109/apec.2017.7930744

2017-01-01

Abstract:Due to heavy computational burden, the predictive step of finite control set model predictive control (FCS-MPC) is usually chosen as one. However, long predictive horizon yields in general a better closed loop performance than short horizon. In order to improve the control performance of model predictive control (MPC) and avoid long calculating time, the multistep model predictive control (MMPC) is applied to permanent magnet synchronous machine (PMSM), with assumption that the input voltage of PMSM is continuous without constraints. Unfortunately, due to output voltage limit of inverter, the optimal voltage vector solved by MMPC may exceed the output capacity of inverter. Therefore, according to the optimal voltage vector of MMPC without constraints, one modulation strategy is proposed to help handle this constraint problem. Finally, both simulation and experiments verify the effectiveness of this control strategy.