Contextualization for Generating FAIR Data: A Dynamic Model for Documenting Research Activities.

The digitization of technologies in product manufacturing results in the availability of large amounts of process and product data. To gain knowledge from this data and fully leverage its potential, its structuring and semantically annotation is essential. This allows preserving the context of data generation and makes the data machine-readable and interpretable. Contextualization is the key to generating FAIR (Findable, Accessible, Interoperable, Reusable) data. The documentation of research activities and provenance of generated data is usually achieved by protocols. However, there is often a tension between the desire to document data generation in a structured, semantically rich form and the need to design research and process parameters flexibly as experimental conditions change. To resolve these contradictions, a dynamic model is described that allows to document research activities and implemented into a knowledge and research data management system to resolve these contradictions. The model allows a formal, semantic representation of research steps, parameters and gathered data, while also providing flexibility in the generation of protocol templates and individual experiments through the reuse of semantic building blocks. The approach is carried out within the context of a large collaborative research center, showcasing its use in managing and providing data for heterogeneous research tasks, documentation, and data types across interdisciplinary projects.