Research On Neural Network Backstepping Control Of Linear Synchronous Motor Magnetic Levitation System

The controllable excitation linear magnetic levitation synchronous motor is a new type of linear motor.It controls the electromagnetic suction in the vertical direction through the excitation current,suspends the platform on the guide rail,realizes frictionless operation,and can meet the high positioning accuracy and high-Speed requirements.The magnetic levitation system has the characteristics of parameter perturbation,multi-variables,and strong coupling.It is a typical nonlinear system,and it is necessary to design a suitable controller to achieve precise control of the feed platform.This subject is funded by the National Natural Science Foundation of China: "Research on the Operating Mechanism and Control Strategy of the Controlled Excitation Linear Synchronous Motor Magnetic Levitation Feed Platform(Project Approval Number: 51575363)",with the magnetic suspension feed platform as the application background.Under certain disturbances,the linear motor magnetic levitation system control problem.The main research contents of the article are as follows(1)Operation mechanism analysis and mathematical model establishment of maglev feed platform for controllable excited linear synchronous motor.By analyzing the working principle of the motor,a mathematical model of the linear motor in the horizontal and levitating directions is established,and the analytical expressions of the horizontal thrust and the magnetic levitation force are derived.The nonlinear problems of the system is analyzed.(2)Design of an Adaptive Neural Network Backstepping Controller.Aiming at the characteristics of multi-parameter perturbation of magnetic levitation system and the susceptibility to external disturbances,an ideal controller was designed based on the backstepping design method.Based on this,the RBF neural network was used to approximate the non-measurable part of the controller,the adaptive adjustment law of neural network weights is designed,and then mathematical proof is constructed by constructing appropriate Lyapunov functions.Finally,the simulation of the adaptive neural network backstepping controller of the magnetic levitation system is compared with the PI controller to compare the steady state and dynamic performance.(3)Hardware circuit and software design of magnetic levitation control system.The hardware circuit design includes DSP minimum system circuit design with TMS320F28377 as the core.Rectifier circuit design,power circuit design,inverter circuit design,IPM power drive circuit design,magnetic levitation height detection circuit design,current detection circuit design and corresponding adjustment circuit design.Software design is mainly for the design of main program and interrupt service subroutine.The main program implements the DSP system initialization and the call and circulation of the interrupt service subroutine;the interrupt service subroutine realizes the realization of each function.