Adaptive Feedback Linearization and Tracking for Induction Motors

Abstract

We discuss an approach to the control of voltage fed induction motors, which takes explicitly into account the nonlinear multivariable nature of the model and the presence of uncertainties in the motor inductances. Exact input-output linearization techniques are used to design a controller for motor torque and flux. The parameter adaptation is then proposed to overcome their chief drawback, the fact that they rely on an exact cancellation of the system nonlinearities in order to get linear input-output behaviour. The linearization objective is perturbed by the presence of model uncertainties which appear linearly. A direct adaptation scheme which depends on the output’s strong relative degree, yields asymptotically bounded tracking i.e. provides bounded state evolutions, and outputs which track asymptotically the desired trajectories. The method has been tested in simulation, on a high power induction motor model.