Research On Robust Control For The High Stiffness Magnettic Suspension System

The beam of the gantry moving type CNC machine tools is for the research object,uses the magnetic levitation technology to reduce friction with the rail, to ensure that thetop of the beam suspension in the stationary rail and achieve precision processing. Thepurpose of this thesis is to eliminate or weaken the influence of various externaldisturbances on the suspended stiffness according to the proposed effective controlmethods for keeping air-gap stiffness based on the working principle of the beam, whichensure the stability of the beam suspension and the suspension of highly accuracy ofair-gap stiffness.First, the portable numerical control machine structure is introduced in detail,expounds the mechanism of suspension beam, with single magnet for example built themathematic model of the system. Due to the mathematical model derived being nonlinear,the local linearization method on the balance point is chosen to gain the linear voltagecurrent control model and the linear current control model.Second, because errors exist between the linearized model and the system modeling,Therefore, the H_∞mixed sensitivity control method to design H_∞controller was selected.Mixed sensitivity optimization algorithm compromised the existing contradictions betweenthe three of the anti-interference ability, tracking ability and robust stability, to improve theperformance of the system to get the desired effect, is an effective way to design robustcontroller. Compared with the classic PID control method, the H_∞robust control has betterinterference rejection capability, thus the system of suspension stiffness was improved.Finally, due to the load disturbances, parameter changes and many other uncertainfactors leading to the change of air gap displacement, the suspension stiffness of thesystem may be weakened. To further improve the robust performance and stiffness of thesystem, the neural network H_∞robust controller is used in this thesis. When the system isconfronting with parameter perturbation and external disturbance, the neural networkrobust control system can achieve good performances and avoid or weaken the steady-state error of the system. Thus, the method improve the stability of system and meet the systemachieve high precision requirements.Simulink simulation study shows that, basing on the voltage model of magneticsuspension system, PID controller, H_∞robust controller and neural network H_∞robustcontroller were designed respectively. Compared with the classic control method, H_∞robust control method make the robust performance and anti-jamming performance of thesystem better, and neural network H_∞robust control methods was adopted to improve thesuspension stiffness of a single electromagnetic suspension system, the organiccombination of the two control methods can satisfy the design requirements of thesuspension system.