Parameter Hierarchical Identification for Magnetic Bearing Control System Using Frequency Response Testing Method

Active magnetic bearings which support rotors without any mechanical contact or lubrication are nowadays widely used in high-speed precision machinery. This paper presents a detailed description of the modeling methods for an active magnetic bearing used in magnetically suspended control moment gyroscopes. A combination of analytical modeling, model updating, and identification was employed to obtain an accurate model. A digital frequency response testing scheme was developed to get the frequency response data of each component in the active magnetic bearing control system. Model identifications based on parameter hierarchical identification principle were applied to obtain the actual model of the power amplifier module and the AMB-rotor module. The accuracy of the model identification was also verified. The experimental results show that model identification method presented in this paper is feasible and effective. The whole modeling and validation process renders an accurate model of a low order for designing a robust control system.