Adaptive Deadbeat Predictive Current Control for PMSM With Feed Forward Method

The steady-state error and torque ripple caused by the mismatch or variation of parameters in the deadbeat predictive current control of permanent magnet synchronous motor are presented in this paper. An improved adaptive deadbeat current predictive model only related to inductance parameters is proposed. The influence of inductance variation on the system stability margin is quantitatively analyzed. When the variation is less than 50%, the system performance can be improved obviously. When the variation reaches 50%, the system will oscillate and cannot operate stably. The feed forward control strategy is introduced to improve the stability margin of the system, and the inductance disturbance of the system oscillation is expanded to 67%. Simulations and experiments are carried out for traditional deadbeat current predictive model, adaptive deadbeat predictive model, and adaptive model with feed forward control. Results show that under various working conditions, the effectiveness of the proposed method in eliminating the steady-state error caused by parameter disturbances is verified. The current distortion and torque ripple are also restrained. Both robustness and steady-state performance of the system are improved.