Acoustics and vibration based milling diagnostics using supervised machine learning approaches

Abstract

In industrial machining processes, tool failures may result in losses in surface and dimensional accuracy of a finished part, or possible damage to both the work piece and the machine. Consequently, tool condition monitoring has become essential to achieve high-quality machining as well as cost-effective production. Moreover, cutting tool degradation may vary considerably under different operation conditions and materials behaviour. Therefore, real time identification of the tool state during machining is critical. In this study the acoustic and vibration responses of the cutting tool and of the workpiece material are measured with the help of a microphone and a piezoelectric accelerometer. Many milling experiments were conducted for different conditions and tool failures, from which features were calculated in time and frequency domains and then split into two groups: 80% for the machine learning model training, 20% for the test phase. Then several supervised machine learning approaches have been applied and compared. Results of our study are implemented in a real tool wear predictive maintenance framework.