(Dr. Thomas Kurfess, advisor)
"Remote Condition Monitoring of Rolling Element Bearings with Natural Crack Development"
Rolling element bearings are used in a variety of devices, including rotating machinery in the manufacturing industry. Bearing failures account for a large percentage of breakdowns in such machines and can lead to costly downtime. Rolling element bearing failure is probabilistic in nature. Because of this, maintenance programs based on set schedules or past performance are not completely effective in preventing failures during operation. Techniques in condition monitoring have been investigated for decades with the aim of determining the actual status of the bearings and replacing them based on that information.
Research in the area of condition monitoring has been ongoing at the Georgia Institute of Technology, addressing prognostics as well as diagnostics. Previous work includes the implementation of the High Frequency Resonance Technique and Adaptive Line Enhancer for processing vibration data from bearings with artificially induced cracks. Life tests generating natural damage with ball bearings have also been carried out.
The work detailed in this thesis further explores the diagnostic capabilities
of the existing experimental setup and enhances the condition monitoring
scheme. Life testing is extended to straight roller bearings, using
conditions of high speed and load to generate real damage. Data from
multiple sensors gathered over the life of the bearings provide information
both before and after damage occurs. A remote monitoring system is
developed and implemented as an enhancement to the bearing test stand.
The previous motor control, data acquisition and data processing programs
are redesigned to allow for fully automated, continuous operation.