EACS 2016 paper - TOOL WEAR STATE CLUSTERING IN MILLING BASED ON RECORDED ACOUSTIC EMISSION
Any type of content formally published in an academic journal, usually following a peer-review process.
EACS 2016 Paper No. 201
It is widely accepted that tool wear has a direct impact on a machining process, playing a key part in surface integrity, part quality, and therefore process efficiency. By establishing the state of a tool during a machining process, it is possible to estimate both the surface properties and the optimal process parameters, while allowing intelligent predictions about the future state of the process to be made; thus ultimately reducing unexpected component damage. This state estimate can be achieved by implementing a variety of in-process monitoring techniques, and observing the development of selected data features as the wear state of the tool progresses. This paper explores the use of a principal component analysis (PCA) along with a multi-class support vector machine (SVM) to cluster a set of tools’ wear states, based upon sampled acoustic emissions (AE) released during ball-nosed milling of Titanium-5Al- 5Mo-5V-3Cr (Ti-5553).