Self-tuning control based on RBF neural network observer in suppression of imbalance vibration of magnetically suspended flywheels

High resolution earth observation requires high precision attitude control of satellites, because the chatter of satellites can decay the resolution of the earth observation. Magnetically suspended flywheels (MSFW) with the advantages of no contact, no frication, high precision and long life, are the ideal actuators of high precision attitude control of satellites. But there still are several disturbing forces and torques in MSFW which affect the attitude control precision. Aimed at the main disturbing sources, the rotor imbalance, a rotor dynamic model is built and the error of traditional method in suppression of imbalance vibration is analyzed. A RBF neural network observer is taken to identify the rotor imbalance, and a self-tuning control based on the observer is presented to eliminate the imbalance vibration. Simulation results demonstrate that the RBF neural network observer can observe the rotor imbalance and the self-tuning control can eliminate the imbalance vibration significantly.