Prediction of the CNC Tool Wear Using the Machine Learning Technique

Abstract

The Fourth Industrial Revolution (4IR) has been attracting increasing attention. In particular, the concept of a smart factory has been developed in the manufacturing industry. As a part of the ongoing extensive research on the computer numerical control (CNC), a large number of studies have explored the methods of noise and vibration data collection through attaching various acoustic and vibration sensors to CNC machines. The present study is aimed to analyze the effect of equipment operating conditions on the tool wear based on the CNC machine data using the machine learning approach. The collected data include x-, y-, and z-axis force components and the moving speed of the CNC equipment. Data characteristics and distribution of the collected datasets were analyzed through the exploratory data analysis. Specifically, random forest, an extreme gradient boosting model, and support vector machine were used for the data analysis to analyze the effect of CNC equipment operating conditions on the tool wear. The results of this study are expected to contribute to determining the optimal CNC equipment operating conditions, thereby improving the quality and preventing machine damage.