An Improved MRAS for Rotor Time Constant Updating in Induction Motor Drives Utilizing Dot Product of Stator Current and Rotor Flux

Dot product of stator current and rotor flux can be chosen to form a model reference adaptive system (MRAS) to update the rotor time constant online in indirect-vector-controlled induction motor drives. Such a functional candidate is attractive due to the fact that its steady-state value is immune to stator resistance variation and dead-time effect. However, the stability of the estimation system is still not clarified. In this paper, the small-signal model of the whole system is established, paving a way to explore its stability and improve its performance further. It is clearly shown that both inappropriate selection of stator resistance and adaptation gain in the MRAS would result in instability phenomenon when the motor is operated at low-speed regenerating mode. In order to address this problem and hence guarantee the stable operation, the stator resistance used in the rotor flux estimator is designed to vary according to the signs of the $q$ -axis commanded current and stator frequency. Besides, the adaptation gain in the MRAS is designed based on the stability analysis using the root-locus method. The validity of the proposed theory is demonstrated through simulations and experiments.