Bridging conceptual models and numerical domains: a semantic integration approach for federated co-simulation in MBSE

Semantic integration of distributed modeling and simulation services is a major challenge in Model-Based Systems Engineering (MBSE), especially when linking high-level architecture models with detailed numerical simulations. Existing methods often fail to bridge conceptual models and simulation tools, which limits their support for cross-domain co-simulation. In this paper, we present a new approach that combines SysML v2 with ontology-based semantics to create a unified co-simulation platform. A graph database serves as a semantic layer, storing and managing relationships between system architecture elements and simulation artifacts. By federating the Asset Administration Shell (AAS) standard into this knowledge graph, users can query simulation metadata alongside product design information. We also offer an interface to extract component details from SysML v2 repositories in SPARQL-ready form. Our framework enables dynamic configuration and validation of co-simulation scenarios through formal semantic mappings. A representative case study demonstrates how this solution simplifies data integration and enhances traceability in federated engineering environments. These results highlight the potential of semantic technologies to advance simulation-driven MBSE and pave the way for future tool integration and digital engineering infrastructures.