Signal Stream Clustering for Tool-Revolution-Level Tool Condition Monitoring in Milling Process



Published Mar 26, 2021
Si Jie Phua Xiang Li Wee Keong Ng Beng Siong Lim Weixiang Zhong Junhong Zhou


The researches in tool condition monitoring often collect large amount of sensor signal data from experiments to study the complex tool condition relationships with signals. In order to provide new light into this process on a real-time basis, it is critical to identify and detect abnormality at the lowest resolution possible so that the wear beha- vior on each flute within a tool revolution can be clearly shown. A signal stream clustering method is developed to separate numerous tool-revolution signals into similar groups, each representing a specific set of corresponding events. In our expe- riment, the 1000 tool-revolution signals in force signal stream are grouped into 5 clusters. These clusters in turn provide a visual mean to assess the tool condition at the most detailed level. In addi- tion, the clusters also enable complex tool condi- tion relationships to be established from the sig- natures of each set of events.

How to Cite

Phua, S. J. ., Li, X. ., Keong Ng, W. ., Siong Lim, B. ., Zhong, W. ., & Zhou, J. . (2021). Signal Stream Clustering for Tool-Revolution-Level Tool Condition Monitoring in Milling Process. Annual Conference of the PHM Society, 1(1). Retrieved from
Abstract 124 | PDF Downloads 86



CBM, condition monitoring, applications: industrial, applications: manufacturing

(Aggarwal et al., 2003) Charu C. Aggarwal, T. J. Watson, Resch Ctr, Jiawei Han, Jianyong Wang, and Philip S. Yu. A framework for clustering evolv- ing data streams. In VLDB, pages 81–92, 2003.
(Aliustaoglu et al., 2009) C. Aliustaoglu, H. M. Er- tunc, and H. Ocak. Tool wear condition monitoring using a sensor fusion model based on fuzzy infer- ence system. Mechanical Systems and Signal Pro- cessing, 23(2):539–546, 2009.
(Amer et al., 2007) W. Amer, R. Grosvenor, and P. Prickett. Machine tool condition monitoring us- ing sweeping filter techniques. Proceedings of the Institution of Mechanical Engineers, Part I (Journal of Systems and Control Engineering), 221(I1):103– 117, 2007.
(Bagnall and Janacek, 2005) A. Bagnall and G. Janacek. Clustering time series with clipped data. Machine Learning, 58(2-3):151–178, 2005.
(Chung and Geddam, 2003) K. T. Chung and A. Ged- dam. A multi-sensor approach to the monitoring of end milling operations. Journal of Materials Pro- cessing Technology, 139:15–20, 2003.
(Hong et al., 2006) G. S. Hong, J. Sun, Y. S. Wong, M. Rahman, and Z. G. Wang. Effective training data selection in tool condition monitoring system. International Journal of Machine Tools and Manu- facture, 46(2):218–224, 2006.
(Keogh and Lin, 2005) E. Keogh and J. Lin. Cluster- ing of time-series subsequences is meaningless: im- plications for previous and future research. Knowl- edge and Information Systems, 8(2):154–177, 2005.
(Li et al., 2006) Xiang Li, Junhong Zhou, Hao Zeng, Yoke San Wong, and Geok Soon Hong. An intel- ligent predictive engine for milling machine prog- nostic monitoring. 2006 4th IEEE International Conference on Industrial Informatics, pages 1075– 1080, 2006.
(Li et al., 2007) Weimin Li, Liangxu Liu, and Jia- jin Le. Clustering streaming time series using cbc. In ICCS ’07: Proceedings of the 7th in- ternational conference on Computational Science, Part III, pages 629–636, Berlin, Heidelberg, 2007. Springer-Verlag.
(Lin et al., 2003) Jessica Lin, Eamonn Keogh, Ste- fano Lonardi, and Bill Chiu. A symbolic rep- resentation of time series, with implications for streaming algorithms. In In Proceedings of the 8th ACM SIGMOD Workshop on Research Issues in Data Mining and Knowledge Discovery, pages 2– 11. ACM Press, 2003.
(Rehorn et al., 2005) A.G. Rehorn, Jin Jiang, and P.E. Orban. State-of-the-art methods and results in tool condition monitoring: a review. International Jour- nal of Advanced Manufacturing Technology, 26(7- 8):693–710, 2005.
(Rodrigues et al., 2008) Pedro P. Rodrigues, Joao Gama, and Joao Pedroso. Hierarchical clustering of time-series data streams. Knowledge and Data Engineering, IEEE Transactions, 20(5):615–627, 2008.
(Sun et al., 2008) Jie Sun, Yoke San Wong, Hong Geok Soon, Mustafizur Rahman, and Zhigang Wang. Identification of feature set for effective tool condition monitoring - a case study in titanium machining. 4th IEEE Conference on Automation Science and Engineering, CASE 2008, pages 273–278, 2008.
(Tansel et al., 2005) I. N. Tansel, W. Y. Bao, N. S. Reen, and C. V. Kropas-Hughes. Genetic tool moni- tor (gtm) for micro-end-milling operations. Interna- tional Journal of Machine Tools and Manufacture, 45(3):293–299, 2005.
(Zhu et al., 2008a) K. P. Zhu, G. S. Hong, and Y. S. Wong. A comparative study of feature selection for hidden markov model-based micro-milling tool wear monitoring. Machining Science and Technol- ogy, 12(3):348–369, 2008.
(Zhu et al., 2008b) K. P. Zhu, Y. S. Wong, and G. S. Hong. Noise-robust tool condition monitoring in micro-milling with hidden markov models. In Soft Computing Applications in Industry, pages 23–46. Springer, 2008.
Technical Research Papers