Digital twins are already part of modern production strategies – yet only a fully integrated Digital Twin System (DTS) unlocks their full potential. It consists of three closely connected components: a real machine, its virtual counterpart, and bidirectional data exchange that links both worlds in real time. This enables continuous state acquisition, immediate detection of deviations, and active intervention in the process. Such a DTS has been developed and researched at IFW Hannover. The goal is to go beyond mere monitoring of manufacturing processes and to enable truly adaptive control of production. To achieve this, the system integrates the machine control and a technological simulation through a real-time interface, creating a closed control loop between the physical and digital domains.
System Architecture
At the center of the setup is a five-axis machine tool (DMG MORI Milltap 700) equipped with a Siemens Sinumerik 840D sl control. It continuously provides axis positions, currents, spindle speeds, and tool parameters, which are transmitted to the virtual machine. The virtual machine is implemented in the technological simulation IFW CutS, developed at IFW. It uses a cartesian multi-dexel model to accurately represent workpiece geometries and engagement conditions. The digital data flow runs between the real and virtual machine: a continuous bidirectional connection through which process information is exchanged. If the simulation detects a deviation, it can immediately adapt control variables – such as the feed rate – to adjust the real process. Latency analysis revealed an average reaction time of only 34 ms, enabling real-time intervention.
Tool Breakage Detection in Under 30 ms
The DTS was applied in the manufacturing of orthopedic implants. A data-driven model integrated into the virtual machine predicts process forces based on tool motion and continuously compares them with the measured forces. As soon as a significant deviation occurs, the virtual machine sends a stop signal to the control system – on average after 28 ms. In this way, tool breakage can be reliably detected and expensive workpieces as well as machine components can be protected.
Adaptive Feed Control Improves Dimensional Accuracy
Within the DFG project EmSim, the system was extended with adaptive control functionality. The virtual machine uses AI-supported process models to predict the expected form error and compares it with the permissible tolerance. If the predicted value exceeds the limit, the feed rate is automatically reduced. In machining trials for implant production, the contour form error was reduced from approximately 50 µm to around 30 µm using this system. This ensures dimensional accuracy even for complex geometries – a decisive advantage for the precise manufacturing of patient-specific implants.
From Digital Shadow to a Decision-Capable System
By coupling machine and simulation, IFW goes beyond the classical Digital Twin. The DTS becomes an active decision-making system that directly influences the machining process. This creates the foundation for adaptive and data-driven manufacturing. The results demonstrate that tangible improvements in process reliability and machining accuracy are already achievable today – and at the same time open up perspectives for future applications in dynamic production environments.
Outlook
The Digital Twin System developed at IFW demonstrates that adaptive and self-optimising manufacturing can already be realised today. Future work focuses on extending the virtual machine with process monitoring using adaptive limit boundaries based on machine learning. This functionality enables limit values to be automatically adjusted depending on tolerances and geometries, allowing critical states to be reliably avoided. In this way, a new generation of learning manufacturing systems is emerging – systems that operate not only transparently, but also proactively and resiliently, marking a decisive step toward intelligent production.
Contact:
For further information, please contact Martin Winkler by phone at +49 511 762 4991 or via email at winkler@ifw.uni-hannover.de.
