Digital Twin of Micro-Milling Process for Micro-Tool Wear Monitoring

Abstract

Abstract This paper presents a novel digital twin of the micro-milling process that can indirectly monitor the micro-tool wear progression by making inferences from the real-time and simulated variables of the micro-milling process. With its wear monitoring service, the digital twin is regarded as the new approach in the field of tool wear monitoring (TWM) systems. The dynamics of the micro-milling process are simulated by using physics-based models, such as spindle motor, spindle controller, and cutting torque models with real-time data from the actual micro-milling machine. The advantage of the proposed digital twin is that the wear monitoring can adaptively adjust to the main machining parameters, such as feed rate and spindle speed. Therefore, exhaustive training of the models is not needed whenever the machining parameters change. The performance of the digital twin in monitoring the wear progression has been evaluated through several slot micro-milling experiments of the stainless steel workpiece. The evaluation and analysis of the experiment result concluded that the proposed digital twin could detect wear progression abnormality by using the estimate discrepancy. Furthermore, the wear severity could be recognized using the final wear estimation value.