Electromagnetic suspension: new results using neural networks

Due to nonlinear magnet characteristics, controllers for magnetically suspended vehicles are usually designed by using linear small perturbation models. This assumption of linearity poses serious constraints on the operating range (allowable airgap) of the suspended objects. A brief account of using neural network (NN) concepts in the formulation of the nonlinear airgap-force characteristics of an experimental electromagnetic suspension system is presented. The main advantage of using the NN is the designer's ability to identify the magnet-force characterstics through experimental data (learning).<
>