GreenRetroTwin

In practice, the integration of existing machines into new, digitally planned production systems fails due to missing digital representations and the heterogeneity of historically grown data inventories. Without a structured digital description, existing machines can neither be integrated simulatively in virtual commissioning nor reliably assessed with regard to their reusability. Added to this is the fact that existing VC models are typically not maintained after real commissioning and therefore diverge significantly from reality over time. An economically viable, standardized methodology to address these challenges does not yet exist.

The project develops methods to analyze, structure, and transfer existing plant data into interoperable digital models. To this end, ISW is designing an information architecture based on the Asset Administration Shell (AAS) and the Digital Product Passport (DPP). Through semantic translation layers and classification procedures, heterogeneous legacy data is processed and converted into standardized, simulation-capable models that can be integrated into virtual commissioning processes. A further focus lies on automatic model adaptation. Since machines change due to modifications or wear, ISW is investigating a combination of AI-based methods and human-in-the-loop approaches to systematically feed these changes back into the digital models. Validation is carried out using concrete retrofit scenarios with real plant data from the project consortium. The developed solutions will be integrated into existing data and simulation platforms and are aimed particularly at small and medium-sized enterprises that have so far had limited access to digital retrofit technologies. The project thus directly contributes to the goals of the GreenTech Innovation Competition: existing plants are digitally upgraded rather than replaced, saving both resources and energy.