Research On Modeling And Stability Control Of Flexible Rotor With Magnetic Bearing Based On Robust Control

With the development of modern industry,requirement for the speed of rotating machinery is increasing.High speed brings about much performance improvement to the device,such as power density promotion and simplified structural size.Due to the limitation of the traditional bearing with mechanical contact mode,high speed will leads to serious wear and the bearing life can't be guaranteed.In such a situation,it is difficult to meet the needs of industrial applications.The active magnetic bearing(AMB)is a novel bearing with good performance,which breaks through the constraints of traditional bearings.The AMB has greatly improved the performance of rotating machinery with its non-contact features and has been widely used,including magnetically suspended high-speed motors,magnetically suspended compressors and gas turbines,etc.What's more,in fluid rotating machinery,the energy conversion mode of coaxial direct drive makes the rotor slender.Finally,the rated operating speed of the rotor is above its critical bending speed due to high speed reqirement and coaxial direct drive,and the rotor becomes flexible.However,the flexible rotor with AMB would encounter dramatic resonance when it passes the critical speed during starting or stopping,and the rotor amplitude increases sharply,eventually destabilizing.This will cause destruction to the mechanical structure and security incidents.The resonance vibration of the flexible rotor can be effectively suppressed by AMB with adjustable bearing stiffness and damping.The AMB is an inherently open-loop unstable system.The controller determines the performance of the system.Moreover,there are many uncertainties in the AMB system which deteriorate control performance.The control of rotor resonance vibration brings more difficulties to the high performance controller design.The classical PID control is simple and easy to implement.However,it can't handle complex rotor dynamics issues and take the uncertainties into account directly,especially when dealing with rotor flexible modes,it's difficult to obtain desired performance.To this end,the robust control theory was introduced and the study on modeling and stability control of flexible rotor with magnetic bearing based on robust control was conducted in this dissertation.The following research topics were explored.The flexible rotor dynamics model was developed by the finite element method and rotor dynamic was analyzed.The rotor model was updated based on the genetic optimization algorithm,and the modal truncation method was used to reduce the model order.The linear dynamic model of the AMB was developed,and the displacement stiffness and current stiffness of the AMB were determined based on the experimental data.The power amplifier and sensor model were developed by a combination of theoretical modeling and system identification method.Finally,the system model was verified by the frequency sweep test.The main uncertainties in the AMB system were classified,including current stiffness,displacement stiffness,mode shape,modal frequency and rotor damping.Based on the v-gap metric tool,the relationship between each parameter uncertainty perturbation and the v-gap metric value was obtained and the influence of each uncertainty on the closed-loop stability of the system was revealed,which lay foundation for the non-conservative specification of uncertainties in the robust controller design.Based on the robust control theory,the robust H_?and?synthesis controllers were designed and simulation analysis was carried out.The results showed that the?synthesis control had excellent robustness to the modal frequency perturbation and could further improved the closed-loop modal damping level of the system.At the same time,some selection experiences and criteria of weighting functions were given from the perspective of engineering experience,aiming to solve the problem about selecting weighting functions.A flexible rotor test rig with AMBs was constructed and the H_?and?synthesis controllers were applied to the test rig.Experiments were carried out on the flexible rotor under each controller,such as lifting test,static suspension,sensitivity function measurement,compliance function measurement and speed increase test.The experimental results showed that the?synthesis controller has good vibration suppression effect on the first bending mode vibration of the rotor,and the supercritical operation of the rotor was achieved,which verified the correctness and effectiveness of the designed?synthesis controller.Aiming to solve the stochastic problem of weighting function selection in the robust control,the artificial neural network identification method based on confidence interval was proposed to estimate the system uncertainty boundary,and the low-order weighting transfer function was fitted to obtain the non-conservative specification of the model uncertainty.Based on the identified weighting function,the?synthesis controllers were designed and verified by performance test.The results showed that the?synthesis controllers based on the identified weighting function can effectively suppress the rotor resonance vibration,and verify the effectiveness of the identification metheod.