A Software Architecture for Real-Time Digital Twins in Machining

Digital twins are increasingly recognized as a key enabler for optimizing complex systems by providing virtual representations that are continuously updated with real-world data. To unlock their full potential, real-time capabilities are essential, ensuring that decision-making, monitoring, and control can be performed with minimal latency. However, previous work has mainly concentrated on offline analysis or loosely coupled synchronization, leaving the systematic design of scalable and reliable real-time digital twins insufficiently addressed. In this paper we introduce a reference architecture for implementing real-time digital twins, covering requirements such as data acquisition, integration, synchronization, and processing, and highlighting enabling technologies that support low-latency and dependable operations. To validate this approach, we developed a prototype framework and applied it to a machine tool case study, demonstrating how real-time synchronization and decision support can be achieved in practice. The results show that our framework enables consistent, high-frequency interaction between physical and virtual assets, providing a foundation for advanced applications in monitoring, control, and optimization of manufacturing systems.