Research On Control Methods Of Magnetic Levitation Systems

With the growth of demand for high speed in industrial production and human life, magnetic technology has become the focus in the academic circle. The magnetic levitation system is one of the most important platforms to study and research control theory. This thesis is based on the GML magnetic experimental equipment provided by GUGAO company based on MATLAB, in which system control algorithm will be designed and analyzed in detail. The main work is as follows:At first, the structure and work principle of magnetic levitation system are introduced in the thesis. On this basis, the thesis establishes the nonlinear math model of magnetic levitation system. And linearized the nonlinear math model of system by the method of balance point expansion, and then have the linear model of magnetic levitation.Next, the thesis designs and simulates several controllers of magnetic levitation:at first, the thesis designs and simulates PID controller, fuzzy controller and adaptive fuzzy PID controller of magnetic levitation, and analyzes the control performance, advantages and disadvantages of the three types of controllers. Simulation results show that:adaptive fuzzy PID control is relative to the simple PID control to improve the dynamic response performance of the system, relative to the fuzzy control is to improve steady state performance of the system; the system's comprehensive performance has been improved under the control of adaptive fuzzy PID.Then, fuzzy algorithm and neural algorithm is used together in this thesis. The thesis designs and simulates fuzzy neural network controller of magnetic levitation, and analyzed the control performance. Finally use PSO algorithm and IPSO algorithm to optimize the fuzzy neural network control algorithm, simulate the magnetic levitation, simulation results show that the fuzzy neural network control based on PSO is relative to the fuzzy neural network control to improve the dynamic response performance, but there is oscillation stable. The fuzzy neural network control based on IP SO is relative to the first two to improve both the dynamic response performance and the steady state performance. The comprehensive performance is perfect.After that, in the actual debugging, the experimental platform of magnetic levitation is built successfully first, and then introduce PID controller, adaptive fuzzy PID controller and fuzzy neural controller based on IPSO into the magnetic levitation. The experiment results show that the designed in this thesis can achieve ideal control effect.Finally, the thesis summarizes the work done, and put forward the prospect of further research directions and problems.