CLUSTERING-BASED ANOMALY DETECTION FOR PISTON BEHAVIOUR IN LASER SHEET METAL CUTTING MACHINES
Abstract
With the advancement of technology in industrial systems, data-driven approaches have come to the forefront, thereby increasing the importance of digitalization. In this context, monitoring the dynamic behaviours of critical components in machine manufacturing and detecting anomaly conditions have become a necessity within the scope of Industry 4.0. Accordingly, the aim of this study is to investigate the methodological feasibility and industrial applicability of anomaly-related approaches for pistons frequently used in the metal sheet processing industry. For this purpose, the behaviour of exhaust flap pistons in laser sheet metal cutting machines was examined based on pressure and forward motion time data. During operation, slag accumulation on the pistons leads to irregular motion characteristics, which manifest as anomaly conditions in piston movements. To identify such conditions, anomaly patterns in the data collected from the pistons via OPC UA were estimated using unsupervised machine learning clustering methods. Within this framework, the results obtained from K-means and DBSCAN clustering methods were analysed and compared. The findings indicate that monitoring the timedependent behaviours of pistons and detecting anomalies can serve as a preliminary step toward preventive maintenance activities. Consequently, such approaches enable the development of applications aimed at increasing operational efficiency and reducing maintenance costs. Furthermore, integrating machine learning–oriented approaches into the anomaly detection process has the potential to enhance overall system performance and reliability.
Keywords
Ethical Statement
Thanks
References
- Benítez, P., Rodrigues, F., Talukdar, S., Gavilán, S., Varum, H. and Spacone, E. (2019) Analysis of correlation between real degradation data and a carbonation model for concrete structures, Cement and Concrete Composites, 95, 247-259. doi:10.1016/j.cemconcomp.2018.09.019
- Bilgin, T. T. and Çamurcu, Y. (2005) DBSCAN, OPTICS ve K-Means kümeleme algoritmalarının uygulamalı karşılaştırılması, Politeknik Dergisi, 8(2), 139-145.
- Ceyhan, H. and Kasapbaşi, M. C. (2022) Üretim sistemlerinde makine öğrenmesi ile kestirimci bakım uygulaması ve modellemesi, Avrupa Bilim ve Teknoloji Dergisi, 33, 167-175. doi:10.31590/ejosat.1019210
- Chazhoor, A., Mounika, Y., Vergin Raja Sarobin, M., Sanjana, M. V. and Yasashvini, R. (2020) Predictive maintenance using machine learning based classification models, IOP Conference Series: Materials Science and Engineering, 954, 012001. doi:10.1088/1757-899X/954/1/012001
- Çam, M. (2023) Endüstri 4.0: dijitalleşme ve muhasebe, Journal of Politics Economy and Management, 6(1), 11-29.
- Demiralay, M. and Çamurcu, A. Y. (2005) Cure, Agnes ve K-Means algoritmalarındaki kümeleme yeteneklerinin karşılaştırılması, İstanbul Ticaret Üniversitesi Fen Bilimleri Dergisi, 4(8), 1-18.
- Drath, R. and Horch, A. (2014) Industrie 4.0: hit or hype? [Industry forum], IEEE Industrial Electronics Magazine, 8(2), 56–58. doi:10.1109/MEI.2014.2312079
- Dündar, D. R., Sarıçi̇çek, İ., Çınar, E. and Yazıcı, A. (2021) Kestirimci bakımda makine öğrenmesi: literatür araştırması, Eskişehir Osmangazi Üniversitesi Mühendislik ve Mimarlık Fakültesi Dergisi, 29(2), 256-276. doi: 10.31796/ogummf.873963
Details
Primary Language
English
Subjects
Industrial Engineering
Journal Section
Research Article
Publication Date
August 4, 2026
Submission Date
April 22, 2025
Acceptance Date
April 10, 2026
Published in Issue
Year 2026 Volume: 31 Number: 2