Project: KI-sy Twin

Many existing industrial plants have hardly been digitally developed to date – a situation that is increasingly becoming an obstacle. Outdated documentation, a lack of standards, and manual processes make maintenance, process optimization, and the use of digital tools difficult. The result is high operating costs, isolated individual solutions, and considerable organizational effort. Companies are faced with the task of preparing their brownfield facilities for the digital future – ideally without having to build their own IT infrastructure.

In the KI-sy Twin project, Fraunhofer IFF, Fraunhofer IOSB-INA, and industry partners are working on a practical approach: the automated creation of digital twins of existing plants – with the help of artificial intelligence and the Asset Administration Shell (AAS).

The goal is to fully automate the collection, structuring, and digital mapping of plant data. The resulting digital twin complies with common industry standards, reduces manual effort, and enables seamless integration into modern data rooms—without isolated applications or high investments.

The AI-sy Twin system automatically captures information such as type plates, technical documentation, and structural relationships within the plant. This data is transferred to the AAS in accordance with standards. Eliminating manual intervention reduces sources of error, decreases the effort required for data maintenance, and ensures seamless digitization.

A key advantage is that the digitized information can be passed on to external partners, such as service providers or manufacturers, in a standardized manner via data rooms. This creates new opportunities for process improvement and better collaboration along the value chain.

A particular focus is on older existing plants, which have hardly been digitized to date. AI-supported processes—such as the automatic reading of type plates—make data capture much easier. Even complex structures such as flowcharts can be mapped in a standards-compliant manner. The result is a comprehensive, interoperable digital twin that serves as the basis for data-driven applications.

Fraunhofer IFF contributes its expertise from previous projects such as ViPro and ProcessAssist and provides a real-world test environment. As part of AI-sy Twin, this existing facility is being modernized in collaboration with industry partners to create a hybrid demonstrator with both brownfield and greenfield elements. Fraunhofer IOSB-INA is contributing its many years of experience in the field of standardized data models, in particular the administration shell, and in the field of machine learning.

In this realistic environment, new technologies can be tested and further developed under industrial conditions. The project thus demonstrates not only the technical feasibility but also the economic benefits of digitalization with AI and the administration shell.

• Fully automated creation of digital twins according to the AAS standard for existing plants
• Simplification of digitization through AI-supported data collection
• Structured mapping of complex plants (e.g., according to flowcharts)
• Minimization of manual documentation efforts
• Interoperability through the use of established industry standards
• Seamless connection to industrial data rooms
• Establishment of a standardized data interface for operators and manufacturers
• Practical tests and further development based on a real brownfield plant