Implementation of Finite-State Model Predictive Control for Commutation Torque Ripple Minimization of Permanent-Magnet Brushless DC Motor

Abstract

A strategy based on finite-state model predictive control is proposed for permanent-magnet brushless dc motors (BLDCMs) to reduce commutation torque ripple. The main contribution is a detailed description of the algorithm design process applied to BLDCM for commutation torque ripple minimization, which points out that the optimal conduction status is directly selected and the exact duration of each conduction status is not required in control process. This method proposes a unified approach for suppressing commutation torque ripple over the entire speed range without distinguishing high speed and low speed and overcomes the difficulties of commutated-phase-current control, avoiding complex current controllers or modulation models. A discrete-time noncommutated-phase-current predictive model of BLDCM during commutation is established. According to the predefined cost function, the optimal switching state is directly selected and applied during the next sampling period so as to make the slope rates of incoming and outgoing phase currents match in the course of commutation, thus ensuring the minimization of commutation torque ripple. The simulation and experiment results show that the proposed method can effectively reduce commutation torque ripple within the whole speed range and achieve good performance in minimizing commutation torque ripple in both dynamic and steady states.