Application of Clustering Methods for Online Tool Condition Monitoring and Fault Diagnosis in High-Speed Milling Processes
| Autor: | Gan Oon Peen, Amin J. Torabi, Xiang Li, Beng Siong Lim, Meng Joo Er |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
0209 industrial biotechnology
Engineering Computer Networks and Communications Feature extraction 0211 other engineering and technologies 02 engineering and technology Fault (power engineering) Machine learning computer.software_genre 020901 industrial engineering & automation Wavelet Electrical and Electronic Engineering Tool wear Cluster analysis Reference model 021103 operations research business.industry Process (computing) Condition monitoring Computer Science Applications Control and Systems Engineering Artificial intelligence Data mining business computer Information Systems |
| Zdroj: | IEEE Systems Journal. 10:721-732 |
| ISSN: | 2373-7816 1932-8184 |
| DOI: | 10.1109/jsyst.2015.2425793 |
| Popis: | Tool condition monitoring (TCM) is a necessary action in a high-speed milling (HSM) process. As a worn milling tool might irreversibly damage a workpiece, there is a vital demand for a TCM system to evaluate the tool wear progress, or equivalently the resultant surface roughness, nonintrusively. To build up a condition monitoring system for HSM processes, sensor signals are to be utilized to form a reference model that reflects the performance of the system. Therefore, a desired reference model has to apply more efficient feature extraction and artificial intelligence techniques to be more repeatable and generalizable. This paper illustrates the performance of clustering techniques on high-speed end milling experimental data. Studied clustering methods are applied to the wavelet features of force and vibration signals to illustrate the repeatability of their results. It is shown that clustering methods can coarsely capture the status of the process and can be applied for fault diagnosis and TCM purposes. It is also discussed how the application of clustering methods may improve the performance of existing reference models toward the more efficient utilization of available experimental data and to develop easily generalizable reference models. Finally, a possible application of clustering results is discussed comparing with state-of-the-art papers. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|