Tool condition monitoring using Random forest and FURIA through statistical learning

Abstract The introduction of machine learning techniques and their application in various fields has made the task of humans easier. One such application is tool condition monitoring. The increase in demand for high-quality products at low costs forces the manufacturing Industry towards a defect-free environment. The present work deals to realize the tool wear and its condition using machine learning techniques. The machining parameters such as spindle speed, feed, and depth of cut were selected for the study. The vibration signals were captured with good and defective tools under different operating conditions. The vibration signals were processed and the required statistical information was extracted. The extracted features were then classified using the various Machine learning (ML) models such as random forest, fuzzy unordered rule induction (FURIA), and Hoeffding tree for predicting the tool condition. Among the considered ML model, the random forest algorithm produced the maximum classification accuracy of 93.65%.