Sensitivity and conditioning issues in speed sensorless control of induction motors

This paper addresses sensitivity and conditioning issues associated with the problem of estimating the rotor speed and electric parameters of an induction motor based on measurements of stator voltages and currents, and their potential implication in the sensorless control of induction motors. The authors show that for ill-conditioned speed and parameter estimation problems, subset selection techniques can be used to stabilize the estimates resulting in estimators with significantly better performance that results in improved performance for the sensorless controller. Simulations are used to illustrate these concepts.