Sensorless Identification of Machine Saliencies in Induction Motors in the Presence of Periodic Mechanical Disturbances

Sensorless speed control of AC motors has been intensively investigated over the last few decades using multiple strategies. In particular, stable operation around zero speed can only be guaranteed so far by the so-called signal injection methods, which extract motor saliencies to acquire a rotor flux or rotor position.Research on issues of induction motor saliency extraction at first put into service is quite limited but necessary. In certain induction machine drives, cogging torques or torque harmonics are unavoidably present, causing undesired effects such as vibrations, noise and losses. Several torque harmonics possess a mechanical periodicity and magnitude, which is not known when the sensorless drive is put into service. Such harmonics can hampering thus the identification of the motor saliencies, especially at start-up. In this sense, this paper proposes a method to estimate machine saliencies for sensorless speed control under presence of torque harmonics without any position sensor. Thus, the method identifies and compensates simultaneously cogging torques and motor saliencies in an iterative scheme, ensuring smooth sensorless speed operation based on rotor slotting saliency.Experimental results extracting machine saliencies from an induction motor that shows rotor-related cogging torques will be provided.