4.3.3 An empirical study of vibration based bearing fault diagnosis methods
- Event
- 20. GMA/ITG-Fachtagung Sensoren und Messsysteme 2019
2019-06-25 - 2019-06-26
Nürnberg, Germany - Chapter
- 4.3 Machine Learning und Signalverarbeitung
- Author(s)
- K. Pichler, C. Hesch, C. Kastl - Linz Center of Mechatronics GmbH, Linz (Österreich), T. Ooijevaar - Flanders Make, Leuven (Belgien)
- Pages
- 355 - 361
- DOI
- 10.5162/sensoren2019/4.3.3
- ISBN
- 978-3-9819376-0-2
- Price
- free
Abstract
This paper presents an empirical study in which universally applicable fault diagnosis methods are used to evaluate vibration data of bearings. The data were acquired on two different test beds: a gear box test bed containing various bearings at different health states, and an accelerated life time (ALT) test bed to degrade a bearing and introduce an operational fault. Features are extracted from the raw data of two different accelerometers and used to monitor the actual health state of bearings. For that purpose, feature selection and classifier training is performed in a supervised learning approach. For testing the proposed approach, cross validation is applied to the data. The results of the gearbox test bed show that the classification accuracy data increases with the revolution speed of the bearing. Furthermore, the data of a high-end sensor allow higher classification accuracy than the data of a lowcost sensor. The results of the ALT test bed show that the same features that were identified in the gearbox test start to change significantly when the bearing degrades.