ISBN 9783843955508

978-3-8439-5550-8, Reihe Apparatedesign

Alexander S. Behr
Automated Ontology Development for Catalytic Process Research Data Management

211 Seiten, Dissertation Technische Universität Dortmund (2024), Softcover, A5

Zusammenfassung / Abstract

Digitalization is accelerating catalysis research, which is an interdisciplinary field, producing diverse and complex data. The FAIR guidelines address the need to improve the data quality and thus the data value chain. To achieve this, data must not only be stored in a structured manner but the foundations for new data standards must also be created. Ontologies, which can be used to model conceptual knowledge in a structured and machine-readable way, provide a basis for this. Ontology-based knowledge graphs are therefore a key technology for storing implicit knowledge in a machine-readable and FAIR manner, thereby significantly improving the data value chain in catalysis research. With a focus on semantic modeling of research data in the field of catalysis and related sciences, this work presents various tools for a pipeline for semantic data enrichment and applications of the resulting, mainly automated, workflows. First, an overview of existing ontologies related to catalysis science is provided. In contrast to existing ontology portals, metadata on the ontologies is also included here, which, for example, shows the proximity to certain subdomains of catalysis research. This enables a more detailed classification of the ontologies, which also shows that catalysis science has not yet been sufficiently modeled by ontologies. Natural Language Processing (NLP) is therefore used to develop methods that automate and facilitate the creation and expansion of ontologies. Since ontologies can model a high degree of semantic complexity, the modeling of arbitrary catalytic reactions in a new ontology developed in this work, Reac4Cat, is then presented. Finally, to demonstrate the advantages of ontology-based modeling, automated process simulations are carried out using the example of a biocatalytic process. For this purpose, a knowledge graph for laboratory and simulation data is created automatically, based, among other things, on the previously developed Reac4Cat ontology.