Health Monitoring of Multipoint Cutting Tool Insert on Vertical Machining Centre A Machine Learning Approach

Abstract

A cutting tool is the most significant contributor to the material removal process and tool condition monitoring is attracting industry attention. In-process development of any kind of tool fault leads to a reduction in machining accuracy, degradation of surface-finish, and causes interruptions, to name a few. Such faults are untraceable using the conventional condition monitoring approach and need to be addressed smartly. In an attempt to characterize such unknown moments, a framework is proposed in this study. In order to generate data sets, vibration signatures were acquired for faulty and fault-free configurations of a tipped tool, for face milling process. Tool faults such as notch and crater wear, wear at nose and flank, and edge fracture are considered in the current investigation. It is indeed challenging task to deal with raw data collected through dynamic experimentation. This necessitates processing of signal that executes its synthesis, modification, and analysis. First, the signal was represented in time domain and later adapted to wavelet transformed time domain using that assisted denoising of signal. The descriptive statistics obtained were used to quantify variation between the faulty and fault-free classes of the tool. The logic of the decision tree (DT) has assisted the selection of significant features. Further, family of Trees, Bayes, Functions, Rules-based, Lazy, Meta etc. were deployed for training the data and classification model was constructed. Finally, fault classification for test and blind datasets is presented considering the proposed framework. The study focuses on smart condition monitoring to investigate suitability of machine learning approach and recommend the superior model for effective classification.