Induction machine condition monitoring with higher order spectra. II. Variable frequency operation and automated diagnosis

Abstract

This paper is the second of a two part submission investigating the use of higher order spectra (HOS) as an induction machine condition monitoring tool. Part I investigated the fundamental aspects of this technique and examined fault detection with a sinusoidal supply to the induction machine. Part II goes on to extend the techniques introduced previously to include the condition monitoring of variable speed induction machine operation. The fault conditions investigated in Part I are re-examined for a variable speed, inverter-fed induction machine. It is shown that the previously introduced techniques are entirely general and extend to nonsinusoidal, variable speed induction machine operation without modification. The second part of this paper then goes on to describe the automation of the HOS based predictive frequency method of induction machine condition monitoring. The design, development and testing of a software based induction machine condition monitoring tool is described. The tool removes the need for experienced operator judgment and automatically diagnoses common fault conditions regardless of the machine load, speed or supply type. In addition, the software is constructed from inexpensive, readily available components and is based on the previous work of parts I and II of this paper.