Adaptive control of non-linear plants using neural networks-application to a flux control in AC drive system

Application of the backpropagation neural networks to a self-tuning adaptive control of unknown, nonlinear and feedback linearizable plants is examined. The control structure analysed in the paper is based on a method recently reported in literature with some suggested modifications, which are verified in the simulation experiments. Neural networks are employed to build a model of unknown, nonlinear system which is used to synthesise a control input. Self-tuning adaptive control algorithm is then applied to a stator flux control of an induction motor with three phase stator windings and short circuited rotor winding.