Subspace Identification of Unstable Transfer Function Model for a Magnetic Levitation System

This paper deals with the identification of an unstable transfer function model for a magnetic levitation system (Maglev) by a subspace identification method. The closed loop nonlinear model equation of the magnetic levitation system under PID control is simulated and a second order unstable transfer function model is identified. PID controller settings are calculated based on the identified model. Using the PID controller on an experimental Maglev system, an improved transfer function model is obtained. A PID controller is designed based on the model, and the performances of the above PID controllers are evaluated on the experimental system. The controller designed on the model identified on the experimental data works better than that obtained from nonlinear or linear model equation.