A Deep Learning Approach to Prognostics of Rolling Element Bearings
Keywords:Bearing Degradation, Long short-term memory, Feature Extraction, Prognostics, Degradation assessment
The use of deep learning approaches for prognostics and remaining useful life predictions have become obviously prevalent. Artificial recurrent neural networks like the long short-term memory are popularly employed for forecasting, prognostics and health management practices, and in other fields of life. As an unsupervised learning approach, the efficiency of the long short-term memory for time-series predictive purposes is quite remarkable in contrast to standard feedforward neural networks. Virtually all mechanical systems consist mostly of rotating components which are by nature, prone to degradation/failure from known and uncertain causes. As a result, condition monitoring of these rolling element bearings is necessary in order to carry out prognostics and make necessary life predictions which guide safe and cost-effective decision making. Several studies have been conducted on effective approaches and methods for accurate prognostics of rolling element bearings; however, this paper presents a case study on rolling element bearing prognostics and degradation performance using an LSTM model.
How to Cite
Open access licenses
Open Access is by licensing the content with a Creative Commons (CC) license.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.