Discrete-time nonlinear controller for induction motor

Use of feedback linearizing techniques with input/output decoupling has been proposed for the control of induction motors. In the papers reported in the literature only the continuous-time control has been analysed, showing that complete linearization and decoupling can be achieved by a nonlinear state feedback control law. Since the state of the induction motor is not directly accessible (rotor flux is not measurable in a standard machine), an observer must be used in the controller. Due to the nonlinearity of the induction motor model, even if the system is stable using state feedback and the observer itself is stable, the closed loop system may be unstable. However, the stability of the closed loop system can be guaranteed if the observer dynamics are set sufficiently fast. The complexity of the nonlinear control law demands for a digital implementation of the controller. This must be carefully considered; a direct implementation of the continuous-time controller may lead to instability of the closed loop system due to the discretization effects, even if the sampling time is fixed to the smallest value compatible with todays technologies. To overcome this problem, the observer must be designed directly in the discrete-time domain. In the paper, the design procedure is reported and the controller implementation is discussed in detail. Simulation experiments are reported, including analog-to-digital conversions, computational delays and pulse width modulation, showing the validity of the proposed approach.<
>