Metadata- and Ontology-based Data Provenance Collection in an Aircraft Digital Twin
Keywords: Provenance, LLM Supporting Knowledge Graph Generation, Digital Twin, Research Artifacts
Abstract: This paper lays out initial analytical results of ongoing research and development activities about data provenance collection in the digital twin of an aircraft. A case-study-based approach to provenance collection is being applied to a data and metadata lake of an aircraft digital twin, to assess the increase in heterogenous data manageability through metadata integration. The research context and motivation are described in the introduction, including provenance in an aircraft digital twin. Next, a problem definition is formulated and approaches discussed based on relevant literature. The working architecture of a demonstrator is then presented (Data Provenance and Metadata Demonstrator). In its current version this demonstrator is implemented as a python/SPARQL/owlready script for metadata structuring, coupled with an example ontology for provenance. It is based on a pipeline from metadata to provenance (Data Collection, Metadata Extraction, Provenance Collection, Provenance Analytics). Most implementation activities have so far focused on the first three phases, with the interplay between the python script and the ontology occurring in Provenance Collection. This has allowed to identify, test and implement some of the technology for automating knowledge graph creation for provenance. This prototype is being further expanded to cope with robustness issues caused by the nesting of metadata, tackled through recursion, and by mismatches between incoming metadata and the demonstrator’s knowledge base, tackled by calls to an LLM.
The final section draws some conclusions and describes potential future work including: further grounding of the problem definition with respect to digital twins; case studies for CAD models; analytics; visualization.
Submission Number: 9
Loading