| Active magnetic bearing (AMD) is a new type supporting part, more and more attention was paid on it in the science and enterprise circle for its merits with no friction, without lubrication, long service life and so on. It was a typical mechatronics product and involved electromagnetic, mechanics, rotordynamics, control theory and computer science. The emphasis of the dissertation was placed on the digital controller of the AMB and the main contents listed as follows:Firstly, on the basis of the principle of the magnetic bearing, analyzed the magnetic force of the 16-pole radial bearing in single-degree-of-freedom, deduced the nonlinear relationship between the magnetic force and control current, rotor's displacement. From which the magnetic force formula under two-degree-of-freedom was presented.To the inherent characteristics of the AMB, different control algorithms were presented under different scopes. In the linear scope the non-complete PID algorithm was adopted to eliminate the static error and improve the dynamic performance. The relationship of the controller's parameters and the system performance was simulated. And in the nonlinear scope the fuzzy logic algorithm was adopted to suit for the nonlinear characteristics, the performance was improved greatly from the simulated results.In the whole bearing system, to a complex mechatronics product, all kinds of factors can influence the system performance, such as the power amplifier's bias current, the magnetic pole position and the sensor's fixing position and so on. In order to discover these influences, the liftoff track and the external disturbance track were simulated based on the system nonlinear mathematical model.The hardware system based on TMS320VC33 was presented at last, and tuned the control system with the non-complete PID algorithm. The rotor's liftoff tracks were presented under the single-degree-of-freedom and the two-degree-of-freedom, and compared the results between digital controller and analog controller.
|
PDF Full Text Request