From Asset Administration Shells to Knowledge Graphs: Rule-Based Semantic Lifting for Cross Domain Reasoning in Cyber-Physical Systems

Abstract

Cyber-physical systems (CPS) integrate heterogeneous engineering models from domains such as process, mechanical, and electrical engineering. The heterogeneity of these models and their underlying representations makes semantic interoperability difficult, particularly when engineering information must be interpreted consistently across domain boundaries. The Asset Administration Shell (AAS) provides a standardized, domain-independent metamodel for representing asset information and supports semantic annotation through references to external vocabularies. Although AAS defines structural relationships between model elements and supports RDF serialization, it does not provide an ontology-aligned representation that enables knowledge graph-based reasoning across heterogeneous engineering domains. To address this limitation, we propose a rule-based semantic lifting framework that transforms AAS Submodel Templates and instances into OWL/RDF knowledge graphs while preserving their structural semantics. The resulting knowledge graph supports declarative querying and rule-based reasoning over integrated engineering information without introducing additional domain-specific ontologies. The framework is demonstrated through an industrial case study that integrates process and electrical engineering models within a model-based systems engineering workflow. The evaluation shows that the proposed approach enables machine-interpretable cross-domain integration, semantic traceability, and the derivation of execution-level bindings from process-level specifications.