Research On Adaptive Inverse Control For Suspension System Of Bearingless Induction Motors Based On Nonlinear Filters

Compared with other types of motors, the bearingless induction motor(BIM) has many advantages such as compact structure, no friction, easy flux-weakening control, high-speed and super high-speed operation, etc. For this reason, it has been applied to various hi-tech fields including agricultural production machinery, agricultural power generation equipment, semiconductor industry, life sciences,and flywheel energy storage. But as a multi variable, nonlinear and strong coupling system, the BIM has a complex electromagnetic relationship. Moreover, the complex coupling not only exits between the rotating force and levitation force of the BIM, but also in the radial levitation force between two degrees of freedom. In order to realize the stable suspension and controllable rotation of the rotor, improve the control precision, the nonlinear dynamic decoupling control must be carried out. So it has important theoretical and practical significance to consider the adaptive inverse decoupling control of the BIM.Supported by the National Natural Science Foundation of China under grant 61174055 titled “Adaptive inverse decoupling control and digital technology for high-speed magnetic suspension motorized spindles used in numerically-controlled machine tools” and some related foundations, theoretical and experimental research around the control of BIM is developed.First of all, the structure of torque winding and suspension winding is analyzed, and then the suspension principle of the BIM is studied. With the aid of Maxwell stress tensor method, the mathematical model and motion equation of two degree of freedom BIM are established, which lays a theoretical foundation for the following simulation studies.Then, based on the independent control theory of torque and suspension force, and combining nonlinear adaptive filter and inverse system control method, the adaptive inverse decoupling control strategy for the suspension system of the BIM by using the nonlinear adaptive filter is proposed. By adopting the nonlinear adaptive filter, the adaptive model and inverse model of the levitation system are established respectively. The inverse model is copied and connected before the corresponding system in series to obtain the inverse controller by making use of the algorithm of variable step size Least Mean Square(LMS) to adjust the weights on line. This method has no need to identify the torque winding air gap flux, and the use of nonlinear filter can solve the problem of accurate modeling, and the coupling problem between variables in the modeling process. The simulation results show that the control strategy achieves the dynamic decoupling of the suspension system.At last, the hardware control system is designed based on dSPACE for the BIM, and then the experimental platform is built. The design principle diagrams of the main circuit and the control circuit of the peripheral expansion hardware system are proposed. Based on this control system, the on-line experimental procedure is described in detail, and the experimental results are analyzed.