Review of Recent Methods and Learning Techniques in Prediction of Tool Life

Abstract

Abstract The reliability of the cutting tool is a key factor in accomplishing the objective of the machining processes. The methods and learning techniques to predict the tool life have been studied for several decades on basic machine tools, including milling, turning, and also drilling as a part of a smart manufacturing system. The current state of the cutting tool is used for modeling the tool’s life. The undesired condition throughout machining, such as excessive vibration, will affect the tool life directly due to the tool failure, including chatter, wear, and breakage. The prediction of tool life using Tool Condition Monitoring Systems (TCMS) is fundamental for manufacturing productivity and also cost-effectiveness. It is implemented by using advanced sensors for data acquisition and determining the process parameters as well as vibration, sound, and cutting force are collected, either directly or indirectly. Signal processing methods and Artificial Intelligent (AI) classifiers are used to obtain the current state of cutting tools during machining processes. This paper reviews various methods and learning techniques of TCMS in predicting the tool life for different machining processes. An experimental tool life prediction was performed by using K-mer signal recognition as feature extraction for sound signal data collected by the microphone for the CNC turning process on alloy steel. Then, Support Vector Machine (SVM) is used to show the accuracy of the method. The experimental results have shown that it is feasible to apply K-mer signal recognition as feature extraction and SVM as machine learning to predict the tool life with 95% accuracy.