Enabling Scalable Provenance-Based Data Management for Digital Engineering Workflows: A Semantically Grounded Approach for HPC Scale CFD

As High-Performance Computing (HPC) enables the rapid generation of simulation results within aerospace workflows, engineering teams increasingly face challenges related to model credibility, artifact traceability, and digital reusability. This work introduces a provenance-first architecture that captures prospective and retrospective provenance across Model-Based Engineering (MBE) workflows, applied to computational fluid dynamics (CFD) on HPC systems. The approach considers digital processes and artifacts from the knowledge engineering perspective. It is not sufficient to focus on effective, auditable storage of modelling information; turning that information into actionable knowledge requires that the sense, meaning, and context of the modelling activities and artifacts be preserved and made useful to both humans and computers. The work also introduces the data integration elements necessary to capture provenance of artifacts used in modelling as a chain-of-custody process. Using the standard PROV (provenance) ontology extended via a new application ontology, we demonstrate that it is possible to create an explicit, formal specification of the conceptual CFD workflow. We demonstrate this approach through real-world CFD workflows. The resulting system not only provides artifact-level chain-of-custody but also enables future reasoning across digital threads using semantic graphs. This work supports digital engineering compliance with DoD 5000.97 and contributes a scalable reference architecture for provenance-aware data governance in digital twin ecosystems.