Enhanced Predictive Model Based Deadbeat Control for PMSM Drives Using Exponential Extended State Observer
Fengxiang Wang, Dongliang Ke, Xinhong Yu, Dongxiao Huang
Abstract
In order to improve robustness and dynamic tracking performance, this article proposes a deadbeat control method based on an enhanced predictive model and an exponential extended state observer (EDB-EESO). By incorporating parameter mismatches and load perturbation as lumped disturbance, extended state-space realizations of permanent magnet synchronous motor (PMSM) speed and stator current mathematical model are presented. To achieve higher robustness under the disturbance, the exponential extended state observer (EESO) is proposed based on the extended state-space realizations. Combining the EESO, the enhanced predictive model is designed to improve overshoot and settling time caused by the system inertial link. To eliminate the effect of time delay, a two-step predictive control method is adopted in the enhanced predictive model. Furthermore, the cost function of EDB-EESO is evaluated by utilizing the deadbeat control method and the stability of EDB-EESO is proved. The experimental results validate the strong robustness and excellent dynamic tracking performance of EDB-EESO in PMSM drives compared with conventional methods.