Abstract
The nonlinear modelling of the magnetic levitation system is performed considering the dynamics of both the electric and magnetic circuits. The linear coefficients of the system are identified explicitly because they are important for the detail design of the control system. Other characteristic including nonlinearity is modelled by using Neural network. In order to perform this method in the actual system, the teaching signal is made by adding the disturbance signal to the control voltage signal. Numerical simulation results show that the linear coefficients of the system are accurately identified.
