Use this URL to cite or link to this record in EThOS:
Title: Induction machine condition monitoring with higher order spectra
Author: Arthur, Neil
Awarding Body: University of Aberdeen
Current Institution: University of Aberdeen
Date of Award: 1998
Availability of Full Text:
Access through EThOS:
Full text unavailable from EThOS. Please try the link below.
Access through Institution:
In a variety of industrial sectors the condition monitoring of induction machines is an important aspect of any condition based maintenance regime. To date, no panoptic means of assessing induction machine health from a single machine parameter exists, further, little work has been performed on the condition monitoring of inverter fed induction machines, an area of increasing importance. In addition, the signal processing tool Higher Order Spectra has been the subject of a great deal of research, and these tools have certain properties which make them ideal for application in a condition monitoring environment. However, limited published work exists in this area, with even less material describing Higher Order Spectra as an induction machine condition monitoring tool in a quantitative fashion. This thesis reconciles these two anomalies, and describes the application of Higher Order Spectra to induction machine condition monitoring. A number of induction machine fault conditions are analysed theoretically, and the subsequent effect of these faults on induction machine vibration described using simple system theory. Experimental results are presented and it is shown that Higher Order Spectra make the optimal diagnosis tool for these fault conditions. It is further shown that this diagnosis method is independent of the machine load and supply. Further, in the case of the inverter supply condition, this diagnosis is independent of the machine speed and in the majority of cases, the magnitude of the induction machine fault can also be identified using this technique. This allows a predicted time to catastrophic failure to be identified and the scheduling of maintenance in an appropriate and optimal manner. Finally, the entire diagnostic method is combined in an automated, software based diagnostic tool based on the previous analysis. It is shown that this tool provides diagnostic performance close to that for ideal induction machine condition monitoring and represents a relatively simple and inexpensive method of monitoring machine health independent of machine supply, load and speed.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Maintenance regime; Fault conditions; Health Electromechnical devices Electronic apparatus and appliances Signal processing Information theory Pattern recognition systems Pattern perception Image processing