Micro-milling digital twin for real-time tool condition monitoring

Micro-milling is a promising manufacturing process due to its wide material compatibility with work materials and ability to fabricate complex micro features. However, the monitoring of tool condition is challenging because of the stochastic wear behavior these micro tools exhibit, which can drastically change due to instances of tool chipping. This paper presents the development of a high-fidelity digital twin of a micro-milling machine. High throughput data communication over OPC UA was used to communicate instantaneous sensor data to a high-performance computer. Using a trained deep learning model, we demonstrate real-time tool condition prediction over OPC UA. The deep learning model used was found to be in good agreement with post-machining tool wear measurements. The presented work highlights the importance of developing sound machine communication protocols and lays the foundation for integrating machine learning with micro-milling in real-time. © 2024 The Author(s)