General purpose digital twin framework using digital shadow and distributed system concepts

Digital twin (DT) is an emerging concept in the Industry 4.0 era. It integrates intelligence into industrial processes. The broadness of DT's concept allows for multiple definitions and designs to be labeled as DTs. However, general-purpose DTs are still out of reach due to the lack of consensus, tool dependency, and the focus on system integration. In this paper, a general-purpose DT framework is proposed by utilizing distributed system concepts. Data acquisition, system modeling, and intelligence applications are modularized in order to enable framework generality. Digital shadow (DS) is implemented as a core component of DT, which provides a real-time virtual replica of its physical counterpart. DT applications utilize DS models for recommending intelligent decisions. A systematic modeling approach is designed by differentiating between DS models and DT applications. The proposed approach defines the model flow based on its roles and objectives. The proposed DT framework is applied to a semiconductor production use case. For this system, the proposed DT provides insights into the production system's behavior and estimates product completion time. Operators dispatch workin-progress products using DT recommendations based on the equipment setup and the production system's status.