Eventos Anais de eventos
COBEM 2019
25th International Congress of Mechanical Engineering
A Simulation-Driven Deep Learning Approach for Condition Monitoring of Hydrodynamic Journal Bearings. Part II: Diagnostics of Ovalization Faults
Submission Author:
Stephen Ekwaro-Osire , Texas , United States
Co-Authors:
Ozhan Gecgel, Stephen Ekwaro-Osire, João Paulo Dias, Gregory Bregion Daniel, Diogo Stuani Alves, Helio Fiori de Castro
Presenter: Stephen Ekwaro-Osire
doi://10.26678/ABCM.COBEM2019.COB2019-0706
Abstract
Bearings are one of the most crucial components of rotating assemblies that have a great role in longevity. Although several studies have been done on rolling element bearings and gears, for journal bearings, it is less clear how some of the damage scenarios (e.g. wear, ovalization) influence the frequency range of the measured vibration data. Moreover, most studies rely on the collection of large amounts of training data from physical experiments or from the field, which is often associated with high costs in test-rig building and instrumentation. In order to address these issues, in this research, simulated vibration databased machine learning algorithm is applied to journal bearings with several different levels of ovalization fault conditions. First, a numerical model was developed to simulate the fault conditions and generate the dataset. Secondly, deep learning method called Convolutional Neural Network (CNN) is trained with training dataset and used for predictions of the testing dataset. The preliminary accuracy results of his framework showed promising results for training the machine learning algorithms with simulated data to be later used on real applications and predictions of faults in rotating machinery.
Keywords
journal bearings, CNN, condition monitoring, Deep learning, Ovalization
