AI-based wear monitoring of sliding bearings in dynamically operating machinery
In drivetrains for renewable energy applications, e.g. wind and hydro power drives, sliding bearings are susceptible to wear. As of today, no method for a reliable wear prediction is existent due to multivariant wear behavior and stochastic operating conditions of sliding bearings in renewable energy applications. Promising, physics-based approaches have been developed and validated. However, these models are only accurate if the operating conditions of a specific machine. Furthermore, these numerical approaches are still very time consuming, which makes their direct integration towards condition monitoring impractical. First machine learning models show promising results under stationary operating conditions. For a dynamic operation, which is commonly observed in reality, suitable models have not been developed nor validated, due to the missing data and physical understanding of the underlying processes.
The aim of this project is to develop a method, which combines information from sensor measurements and information from physics-based sliding bearing simulations, to predict wear for dynamically operated sliding bearings with the use of machine learning methods.
01.09.2022 – 28.02.2023
Bundesministerium für Bildung und Forschung (BMBF)
Ministerium für Kultur und Wissenschaft des Landes Nordrhein-Westfalen (MKW) im Rahmen der Exzellenzstrategie von Bund und LändernCopyright: © PTJ
ERS - Exploratory Research Space