Abstract:  Robustness of traditional PID controller of active magnetic bearings (AMB) was insufficient, and the unbalanced vibration was difficult to be suppressed efficiently. Aiming at these problems, vibration control of AMB-rotor system was executed based on LQG (linear quadratic Gaussian) controller and LMS (least mean square) algorithm. Firstly, the rotor system mode was built by finite element method, and the order of the model was reduced. Secondly, considering the influence of sensor and power amplifier model and parameters, coupled with the rotor system model, the comprehensive model of the system was established. The state observer and state feedback controller were designed by linear quadratic method, and then the LQG controller was finished to build. Finally, after the LMS algorithm was added to the model, the vibration suppression mode of ABM-rotor system was built and solved combined the LQG controller and LMS algorithm. The results indicate that the established state observer can effectively follow the displacement change of the system and realize the displacement output of the following system in 0.1s. The vibration amplitude of the system is reduced by 16.6% after increasing the displacement element of matrix Q without considering LMS algorithm. The unbalanced vibration can be suppressed efficiently after the LMS algorithm is considered, and when the displacement element of matrix Q is 106, the vibration amplitude is reduced by 90% in 1.8s at first order bending critical speed.

Key words:  active magnetic bearing system; vibration control; linear quadratic Gaussian (LQG) controller; least mean square (LMS) algorithm

ZHANG Sheng-guang, ZHANG Xue-ning, HU Wen-ying.Vibration control of active magnetic bearing system based on LQG and LMS[J].Journal of Mechanical & Electrical Engineering,2022,39(2):210-216.