Improved Model Predictive Control With New Cost Function for Hybrid-Inverter Open-Winding PMSM System Based on Energy Storage Model

Conventional model predictive control (C-MPC) usually leads to considerable torque and current ripples since only one voltage vector is applied. In addition, the C-MPC applied in the hybrid-inverter) driven open-winding permanent magnet synchronous motor (OW-PMSM) suffers from complex tuning work of weighting factors and iterated evaluation work of all potential vectors because the capacitor voltage constraint in cost function is related to the switching states. In this article, an improved three-vector MPC based on energy storage model is proposed for the HI-OW-PMSM system. Firstly, a new prediction model of dc-link capacitor voltage based on energy storage model is proposed thus the capacitor voltage constraint is decoupled to the three-phase switching states. Furthermore, a novel restructured cost function without weighting factors can be obtained. Accordingly, by using the new prediction model and the novel cost function, the voltage reference can be easily obtained and a three-vector based prediction strategy without sector division can be carried out to achieve optimal voltage vectors selection and duty cycle calculation. Therefore, the system performance is improved and the calculation burden is also reduced. Finally, comparative experimental studies are carried out to validate the effectiveness of the proposed improved three-vector MPC method.