Design and implementation of a MRAS sliding mode observer for the inverse of the rotor time constant of the induction motor

The paper presents an estimator for the inverse of the rotor time constant of the induction motor. Estimation is done using a sliding mode observer under the assumption that the stationary frame fluxes are known. With measured voltages, currents, and with known fluxes and speed, the paper develops an MRAS-type sliding mode observer. This works well under ideal conditions; however, if the speed is inaccurate or if the magnetising inductance saturates, the accuracy suffers. The paper develops a model for the saturated induction motor and attempts to also estimate the saturation level. In the second part, an observer with only a single set of feedback terms is developed. This is transformed into a sensorless observer by feeding it with a speed estimate (assumed inaccurate). The error in the estimated rotor time constant can be reduced by operating the motor drive with a low ratio of i_d to i_q.