Model-Based Detection of Rotor Faults Without Rotor Position Sensor—The Sensorless Vienna Monitoring Method

Abstract

This paper deals with an extended approach of the Vienna Monitoring Method (VMM), which is a model-based technique to detect rotor asymmetries in the squirrel cage of an induction machine. The conventional VMM requires the measured voltages, currents, and the signal of a rotor position sensor. The novel scheme presented in this paper alternatively works without a rotor position sensor. In particular, for low-inertia drives, accurate estimation of rotor position is required. The rotor-fault-related double-slip-frequency torque modulation causes a speed ripple with the same frequency. Consequently, low-inertia drives are exposed to higher speed ripples. In this case, it is not sufficient to estimate the mean value of the actual speed, only. Even the speed ripple has to be acquired to benefit from the accuracy of the employed models of the VMM. The proposed technique evaluates the signatures of an inherent rotor fault in order to determine the rotor position signal. The speed ripple can be obtained from the torque modulations that the models already compute. This way, an accurate rotor fault detection technique without rotor position sensor can be realized.