Saliency-based Dynamic Encoderless Operation of Two Induction Machines in Parallel Configuration

Abstract

The concept of encoderless control applied to dual motor drives is gaining more and more attention in the last few decades. Mostly, available research focuses on acquiring rotor information (flux or position) by means of motor models or flux observers supported by current sensors. Due to the non-linear switching characteristics of the inverter, such models tend to become unstable around zero frequency. In this region, the extraction of machine spatial saliencies is meaningful. This paper extracts the rotor-slotting saliency of each motor individually using a pulse excitation strategy applied to dual motors that are equipped with three current sensors in total. Out of these sensors, two current sensors are attached to two phases of one motor, and the third sensor is attached to the second motor. In addition, a signal processing method is employed to compute the torque and flux level of each motor. Experimental encoderless results measured on a dual motor test-bench show excellent dynamic performance even at zero mechanical speed as well as accurate torque-sharing calculation.