Control Strategy for Hybrid Magnetic Bearing Based on Improved Cascaded Reduced-Order Active Disturbance Rejection Controller

Abstract

During the startup of the hydraulic turbine generators, the hybrid magnetic bearing support system exhibits displacement fluctuations, and the nonlinearity and strong coupling characteristics of the magnetic bearings limit the accuracy of rotor modeling, making traditional control methods difficult to adapt to parameter variations. To suppress startup disturbances and achieve a control strategy with low computational complexity and high precision, this paper proposes a five-degree-of-freedom hybrid magnetic bearing control strategy based on an improved cascaded reduced-order linear active disturbance rejection controller (CRLADRC). The front-stage reduced-order linear extended state observer (FRLESO) reduces the system's computational complexity, enabling the system to maintain stability during motor startup disturbances. The second-stage reduced-order linear extended state observer (SRLESO) further enhances the system's disturbance estimation accuracy while maintaining low computational complexity. Furthermore, the disturbance rejection and noise suppression capabilities are analyzed in the frequency domain and the stability of the proposed control method is proven using Lyapunov theory. Experimental results indicate that the proposed strategy effectively reduces displacement fluctuations in the hybrid magnetic bearing support system during motor startup, significantly enhancing the system's robustness.