Research on Disturbance Rejection of Active Magnetic Bearing System Based on Generalized Integrator-Extended State Observer
摘要
Active magnetic bearing (AMB) rotor systems require strong anti-disturbance capabilities for stable operation due to their susceptibility to external disturbances and parametric uncertainties. The Active Disturbance Rejection Control (ADRC) framework handles these issues by aggregating various disturbances into a "total disturbance," which is estimated and compensated for by an Extended State Observer (ESO). However, the conventional ESO haslimitations in effectively suppressing high-frequency vibrations caused by rotor imbalance.
PurposeThis study aims to overcome the aforementioned limitation by proposing an enhanced ADRC strategy that incorporates a Generalized Integrator-based Extended State Observer (GI-ESO) to improve the suppression of both low- and high-frequency disturbances in AMB rotor systems.
MethodsA generalized integrator (GI) was integrated into the disturbance estimation loop of the ESO. The resonant frequency of the GI was precisely tuned to match the rotor's rotational frequency using a second-order generalized integral frequency-locked loop (SOGI-FLL). A frequency-domain analysis was performed to assess the impact of different parameters on the stability of the closed-loop system. The proposed method was validated through simulations and experimental acceleration tests from 0 to 6000 rpm.
ResultsSimulation results confirmed that the proposed GI-ESO achieves superior suppression of both low- and high-frequency disturbances compared to the conventional ESO. Experimental validation demonstrated that the GI-ESO reduces the vibration amplitude of the AMB rotor system by approximately 17.71% relative to the standard ESO.
ConclusionsThe proposed ADRC strategy with GI-ESO effectively enhances the anti-disturbance performance of AMB rotor systems, particularly in suppressing high-frequency vibrations caused by imbalance. The method demonstrates practical efficacy and represents a significant improvement over conventional ESO-based approaches.