Leveraging self-supervised learning for vibration data in industrial separators
Publication date
2024-03
Document type
Conference paper
Author
Heuwinkel, Tim
Merkelbach, Silke
Janssen, Nils
Enzberg, Sebastian von
Dumitrescu, Roman
Organisational unit
Fraunhofer Institute for Mechatronic Systems Design IEM
Bergische Universität Wuppertal
Hochschule Magdeburg-Stendal
Book title
Machine learning for cyber physical systems
Part of the university bibliography
✅
Keyword
Machine learning
Industrial separator
Vibration data
Self-supervised learning
Abstract
Industrial separators play a pivotal role in production processes of various sectors such as chemical, pharmaceutical, biotechnology, oil extraction and food industries, with over 3000 distinct applications. Operating these separators involves managing several process parameters as well as discharge and cleaning cycles, which are hard to control mainly due to deficiencies of current physical sensor technology. Recent studies have shown that machine learning can be utilized to detect faults and particle presence in separators via vibration data. However, traditional machine learning methods require domain expertise or vast amounts of labeled data. We propose the use of self-supervised learning to resolve this issue by learning useful representations from unlabeled data, which is significantly easier and cheaper to obtain. An empirical validation on data from a disk stack separator shows that self-supervised learning can improve upon manual feature engineering and supervised approaches in terms of cost, accuracy and data efficiency.
Version
Published version
Access right on openHSU
Open access