Research On Adaptive Neural Network Control Of Electromagnetic Suspension System

Compared with other types of rail transit,maglev trains have the advantages of low operating noise,strong climbing ability,small turning radius,pollution-free and low cost,etc.,and have attracted wide attention from all over the world.The development of maglev technology and its industry is of great significance to the solution of my country’s urban traffic problems.Among the current vehicle suspension systems,the electromagnetic suction system is the most mature.Therefore,this article mainly focuses on the electromagnetic suction system for its suspension control research.First,through the mechanical decoupling of the maglev train,the multi-point levitation problem of the system is transformed into a single-point levitation problem,the mathematical model of the electromagnetic suction levitation system is deduced.And based on the linearization of the equilibrium point,the third-order voltage control model and the second-order current control model of the system are established respectively,comparative analysis of the control characteristics of the above two models,and proved the controllability and observability of current control.Finally,a brief introduction to the working principle of the suspension chopperSecondly,in view of the difficulty of traditional control algorithms to effectively solve the problems of maglev trains encountering load changes,system parameter changes,and external interference that lead to deterioration of control performance during operation,this paper designs a new type of adaptive neural network control algorithm,and passes Lyapunov’s law proves its stability,Afterwards,simulation models based on PID control,integral sliding mode control and adaptive neural network control were built in Matlab/Simulink software.And under a variety of working conditions,the simulation results of the three control methods are compared and analyzed.The results show that the controller designed in this paper has strong antiinterference,robustness and model adaptability.Finally,based on the levitation car model of the Maglev Institute,a levitation controller with TMS320F28335 as the main control chip was designed,and the corresponding hardware and software design was completed.And under a variety of typical working conditions,the control effect of PID control and the adaptive neural network control designed in this paper are compared and analyzed.The experimental results show that the controller designed in this paper has excellent control performance.