Computational methods for the detection of wear and damage to milling tools. (October 2022)

• Record Type: Journal Article
• Authors: Ninevski, Dimitar; Thaler, Julia; O'Leary, Paul; Klünsner, Thomas; Mücke, Manfred; Hanna, Lukas; Teppernegg, Tamara; Treichler, Martin; Peissl, Patrick; Czettl, Christoph
• Is Part Of: Journal of manufacturing processes. Volume 82(2022)
• Page Start: 78
• Page End: 87
• Publication Date: 2022-10
• Subjects: Condition monitoring -- Milling tool damage -- Time-series sensor data; Production management -- Data processing -- Periodicals; Manufacturing processes -- Periodicals; Procestechnologie; Productietechniek; Production -- Gestion -- Informatique -- Périodiques; Fabrication -- Périodiques; Manufacturing processes; Production management -- Data processing; Periodicals; 670.5
• DOI: 10.1016/j.jmapro.2022.07.030
• Languages: English
• ISSNs: 1526-6125

Abstract

Abstract: The current paper presents a new computational approach to detect wear and damage to milling tools' cutting edges. The proposed approach is independent from exact information on tool-workpiece interaction conditions and only requires that they remain constant for compared milling operations. Additionally, the approach was thoroughly tested on time-series data obtained from an industrial-scale milling process, instrumented by commercially available instrumentation equipment, during which 18 identical parts were milled. The time-series data contains the bending moments in the x and y directions as well as the torque and tension acting on the milling tool. Some measures used are systematic in nature, based on shape, rotation and work needed for milling, whereas others are statistical in nature, describing the change in the distribution of the data. All of the measures proposed in the current work are relative and mutually invariant, meaning they address different information content of the data independently. A comparison of the mentioned measures with the real-world damage evolution of the milling tool's cutting edges for multiple produced parts yielded consistent results and suggests a high potential for practical tool damage detection in industrial production.