Open Access Open Access  Restricted Access Subscription or Fee Access

Online Damage State Clustering of Cutting Tools Using Dirichlet Process Mixture Models

CHANDULA T. WICKRAMARACHCHI, TIMOTHY J. ROGERS, WAYNE LEAHY, THOMAS E. MCLEAY, ELIZABETH J. CROSS

Abstract


Tool condition monitoring is an area of research with the main focus on predicting time to tool failure. Ultimately, the goal in the manufacturing industry is to establish a production line with a predictive maintenance strategy in order to reduce waste. One of the most crucial challenges in tool condition monitoring is the availability of descriptive labels of wear states. It is not possible to directly acquire damage labels whilst the tool is in use. Consequently, collecting machining datasets is expensive and time consuming as the process must be interrupted to obtain damage labels. Furthermore, the labels alone cannot be used as an indicator for when a tool may fail. An online semi-supervised learning method is proposed in this paper that allows the use of incomplete label sets and alleviates the need for pre-labelled training data, therefore reducing the costs associated with data collection. The authors apply Dirichlet process mixture models to an acoustic emission dataset, collected during a turning operation. The Dirichlet process mixture model allows data clustering as it is collected without the need to set the number of possible clusters a-priori. This will enable operators to visualise when the characteristics of the cutting process change during machining, avoiding the need for exhaustive measures for tracking tool wear, or the early disposal of tools, depending on the context.


DOI
10.12783/shm2019/32490

Full Text:

PDF