A sensorless speed estimation algorithm for use in induction motor fault detection applications

Abstract

A novel sensorless speed estimation algorithm for use with direct online three-phase induction motors is proposed. Speed information is extracted from the motor current spectrum by tracking the frequency of key components, which vary as a function of motor rotational speed. An important advantage of this technique is that the speed estimation algorithm is independent of motor mechanical and electrical parameters. The algorithm operates via estimating rotor bar number, which is in turn used to determine rotational speed via rotor bar pass frequency detection including a sanity check on estimated speed via comparison with a linear speed estimate based on rated nameplate data. Experimental results are included for a range of three-phase induction motors including motors carrying faults (bearing, rotor, stator and air-gap eccentricity). The results demonstrate the robustness of the algorithm to motors operating with a variety of faults and thus the potential for use of the algorithm in induction motor fault detection and diagnosis applications.