Robust deadbeat control of the non-linear magnetic levitation system using adaptive control techniques

摘要

The modern push toward faster and more efficient transportation systems has caused many new techniques to be developed. One of which is a Magnetically Levitated train, or Maglev, where an electro-magnetic control system is used to hover and propel a train at high speeds and efficiency. This has been proven to work with the 'Transrapid' Maglev train in Germany, and the 'Shanghai Maglev Train' in China. The project studied a similar dynamic control system which is applied to operate these trains. This system provides many advantageous control characteristics in a control system sense.An ECP (Educational Control Products) Magnetic Levitation plant was used in the development of this control system. The robust deadbeat controller was designed and simulated firstly, in order to prove its viability. The initial simulations found that a switch controller was required, due to the nonlinear dynamics of the plant. The designed switch controller was further investigated in simulations, and the results showed that the controller was viable and effective, and could be applied to the plant.While applying the designed switch controller to the plant, some real-world problems such as noise and control errors were encountered. Digital signal processing techniques were employed to remove these issues. In the final testing, results were obtained and evaluated from the designed controller and a traditional PID controller. These results showed that the objectives of the designed controller were well achieved, and its performance is consistently 60% better in settling time than that of a traditional PID controller. Other performance results include a 66% better disturbance rejection, and a 30% better bandwidth. Analysis showed that controller designed was a viable alternative to available controllers in the industry.