The goal of the project "VisionAdapt - Imaging Shavings Shape Detection for the Automation of High Dynamic Turning Processes using Process Adaptation" is the technology transfer of autonomous environment detection for shavings shape detection for High Dynamic Turning (HDT) process. Turning is a widely used machining process in all industrial sectors. Due to the material removal during continuous cutting, long or critical shavings shapes can occur. These shavings are often difficult to remove, necessitating interruption of production and manual intervention. This also increases the risk of damage to the tool or workpiece. The occurrence of critical shavings prevents automated and reliable production, especially in individual production. In principle, shavings breaking behavior can be influenced by the geometry of the shavings cross-section, in addition to the material properties, cutting speed, and shavings breakers on the tool.
The project aims to automatically detect critical shavings using ML-based imaging process monitoring. The existing system collects image data of the resulting shavings in parallel with the manufacturing process. Machine learning techniques detect critical shavings, enabling automated removal, eliminating the need to actively interrupt the process.
The developed concept will be implemented in a demonstrator, which will be used in a real-life use case in a turning-milling center. The developed approach will later be transferred to other machines, materials, and components. The goal is to develop a real-time-based process analysis and adaptation for HDT processes, which will enable later transfer to industry.
Partners
- Leibniz Universität Hannover, Institut für Fertigungstechnik und Werkzeugmaschinen (IFW) - Technische Universität Braunschweig, Institut für Konstruktionstechnik (IK) - DMG Mori AG, Manager Corporate Development & Worldwide Institutions