Large Language Model-Based Formalization of Engineering Requirements Into Ontology-Constrained Knowledge Graphs

Requirements engineering plays a central role in mechanical design, yet technical requirements remain predominantly expressed in natural language, limiting traceability, validation, and computational reasoning. This work presents an ontology-constrained pipeline for transforming natural-language engineering requirements into industrial ontologies foundry (IOF)-grounded knowledge graphs enriched with QUDT-based quantitative semantics. The pipeline decomposes text blocks into individual prescriptive clauses, extracts structural slots and constraint atoms through a typed intermediate representation, normalizes quantitative expressions via QUDT unit and quantity-kind grounding, and instantiates IOF-compliant Web Ontology Language (OWL) instance graphs, referred to as assertional boxes (ABoxes). The transformation is implemented as a hybrid neuro-symbolic workflow that combines large language models (LLMs) with typed intermediate representations, rule-based post-processing, and description-logic reasoning. Evaluation on a formula SAE rules corpus, intentionally selected to stress quantitative constraint handling, shows good structural reliability in the evaluated setting, and also highlights important end-to-end limitations. Decomposition accuracy was 63.74%, indicating that clause segmentation remains a major source of propagated error. Conditional on correctly decomposed requirements, slot-level extraction achieved a macro accuracy of 94.50%; however, the stricter row-level criterion requiring all structure and quantity slots to be correct was 69.38%. Quantitative constraint identification reached 97.64% precision and 96.88% recall. Normalization coverage was 98.78%, with residual errors primarily attributable to quantity-kind disambiguation. At the graph level, 93.33% of grounded artifacts passed all ontology-conformance checks. These results indicate that ontology-constrained LLM pipelines can support requirements formalization, while still requiring human oversight for decomposition-sensitive, logically complex, or compliance-critical use cases. The code and data used in this study are openly available in the ontology-req-pipeline GitHub repository.