ADAPTIVE NEURAL NETWORK CONTROL OF ELECTROMAGNETIC SUSPENSION SYSTEM

This research describes the controlled design and implementation of magnetic levitation application. The highly nonlinear electromagnetic suspension (EMS) system is hardly and limited the system control subjected to prescribed stability of system. Due to the nonlinear dynamics of system, the linearization of the nonlinear EMS plant is obtained with linear model by using linear approximations. An attraction force about the prescribed nominal operating point of current and air gap position is chosen for linearization. Linear state feedback control, adaptive neural network control and hybrid of linear state feedback and adaptive neural network control are applied for controlling nonlinear dynamics and parametric uncertainty of plant system. In adaptive neural network control structure, radial basis function (RBF) employs to approximate the nonlinearity and uncertainty of EMS plant due to unstructured modelling. The system stability and adaptation is proofed by using Lyapunov's method. The results of proposed control performance are clearly explained comparatively in practical implementation by experiments.