Low-Speed Active Disturbance Rejection Control Based on Disturbance Foresight Compensation for Multi-Unit Distributed Permanent Magnet Arc Motor

This article proposes a model predictive speed control with disturbance foresight compensation (DFC-MPSC) based on generalized proportional integral iterative learning observer (GILO) to reduce the speed fluctuation caused by periodic end torque, cogging torque and flux harmonics existing in multi-unit distributed permanent magnet arc motor (arc motor). A novel position- and velocity-based iterative learning law has been developed and extended to the general form of a generalized proportional integral observer. Compared to existing observers, this approach effectively addresses the trade-off between high-frequency periodic disturbance observation and noise suppression. Its higher-order formulation, based on the position equation, significantly reduces the impact of speed measurement noise on observation accuracy, particularly at low speeds. GILO can also improve the observation performance of the generalized proportional integral observer for periodic disturbances in sinusoidal form. Theoretical analysis has verified the stability of GILO. Iterative learning can record periodic disturbance sequences, therefore the observations of disturbances and their future performance can be embedded into the prediction model according to the prediction horizon, which enhances the steady-state and dynamic performance of MPSC, especially at very low speed. The comparatively experimental results have validated the effectiveness of the proposed strategy.