Multi-Vector Model Predictive Control for Permanent Magnet Synchronous Motors With Disturbance Compensation

Abstract

Targeting the weaknesses of conventional model predictive control (MPC), which fails to suppress disturbances effectively and has significant current pulsations, a multi-vector MPC strategy based on improved active disturbance rejection control (ADRC) is proposed in this paper. To enable voltage vector (VV) combinations applied to inverters to cover any direction with variable amplitude, four VVs are selected based on the sector-based vector selection method and the cost function, with the consideration of reductions in the number of switching transitions and common-mode voltage (CMV). In addition, current slopes are introduced for deadbeat control to achieve duty cycle-based control. Furthermore, to overcome the limitations of the conventional speed loop that cannot consider both rapid response and overshoot, an improved ADRC controller with disturbance compensation is designed to optimize the control performance. Dynamic observation and the feedback compensation of disturbances are implemented by utilizing a cascaded extended state observer (ESO). To alleviate the estimation burden of cascaded ESO, partial disturbances are evaluated and subsequently compensated into the controller. The simulation results demonstrate the feasibility of the proposed multi-vector MPC strategy that enables effective disturbance suppression.