Date of Award
Open Access Thesis
Rolling element bearings perform an essential role in most all rotating machinery. To prevent degradation to machine performance, unforeseen costs and unexpected system failure, bearing fault diagnosis and prognosis are used. Acoustic Emission (AE) introduces high sensitivity, early and rapid detection of cracking, and real time monitoring that can alarm once cracking is noticed.
The purpose of this research is to nondestructively monitor the crack growth in rolling element bearings in a marine environment and to determine the acoustic emission parameters which embody crack initiation and propagation. The intellectual merit lies in:
1. the signal alarm developed from an AE data pattern recognition method,
2. the damage quantification procedure based on intensity analysis parameters, and 3. the specially made rotating machine test bed to simulate a bearing in use on a submarine.
The gap in current literature addressed a shortage of data and findings on acoustic emission signal alarm notification and use of shipboard machinery parameters for acoustic emission monitoring of rolling element bearings. Four rolling element bearings were tested in a specially made rotating machine test bed at various load and rotation cycles to exemplify shipboard machinery operation at various depths. Acoustic emission data classification was done through pattern recognition and neural network software (NOESIS). All AE data was clustered using k-means unsupervised method and the lowest correlated features were selected for pattern recognition.
It was concluded that the clustering method used successfully classified crack initiation and propagation. Useful AE parameters for classifying crack initiation and propagation are amplitude, initiation frequency, absolute energy, frequency centroid, peak frequency, and signal strength. With use of intensity analysis, it was determined that the intensity at crack initiation and propagation is much higher than at the final section where failure occurred. Acoustic emission is suitable for remote monitoring of bearing degradation. With the use of signal alarms based upon the clustering method and parameters discussed, one can be notified when a crack is initiating and propagating, and prepare for failure of the bearing. The ability to be notified when cracks are initiating and propagating will prevent unexpected system failure and reduce maintenance cost.
Feirer, B. L.(2020). Damage Evaluation of Rolling Element Bearings for Shipboard Machinery. (Master's thesis). Retrieved from https://scholarcommons.sc.edu/etd/5662