A Design Methodology of Multi-level Digital Twins

This paper proposes a design methodology of Digital Twins enabling multi-level simulation of equipment in a manufacturing plant. In this context, the Digital Twin covers a central role where it can perform analysis of the current state of the plant and, more importantly, prediction regarding its future state. This requires the use of complex models for all the nodes that compose the entire production, to obtain a more accurate estimation of future equipment states. Furthermore, most of the commercial tools provided by different vendors do not consider this dimension of the problem and allows to perform simulations of the plant with a very high level of abstraction, or with the use of statistical approximation. On the other hand, several physical process simulators allow to model and simulate single equipment, but without considering the production line perspective. Multi-Level modeling considers different levels of abstraction of the same model, allowing to switch from a model to another. This paper proposes a design flow methodology based on multi-level approach, that allows to obtain a unique environment where physical and production simulators are integrated automatically. The entire design flow is validated with a real use case scenario. The obtained results show different simulation strategies using multi-level approach with different synchronization granularity.