Unlocking Semantic Interoperability in Industry with Large Language Models
Abstract
The digitalization of industrial processes within Industry 4.0 depends on the reuse of technical product data across systems, organizations, and lifecycle stages.In practice, the main challenge is not often the availability of data, but their heterogeneity, which prevents automated interpretation and semantic interoperability.This paper presents an automated pipeline for incrementally transforming existing heterogeneous technical data into interoperable representations.The approach builds on available industrial data and combines Large Language Models, retrieval-augmented generation, and the Asset Administration Shell to align manufacturer-specific descriptions with standardized vocabularies such as ECLASS.Using fluid power datasheets as an exemplary data source, the pipeline is introduced to demonstrate how heterogeneous engineering data can be made interoperable.The focus is on the pipeline concept itself, which aims to reduce manual mapping effort and support a gradual transition toward Industry 4.0-compliant digital engineering environments, rather than on the processing of datasheets.While demonstrated in the fluid power domain, the approach is transferable to other technical domains facing similar data heterogeneity challenges.
Authors 4
-
Affiliation as printed
RWTH Aachen University , Institute for Fluid Power Drives and Systems (ifas) , Campus- Boulevard 30 , 52074 Aachen , Germany
-
Affiliation as printed
RWTH Aachen University , Institute for Fluid Power Drives and Systems (ifas) , Campus- Boulevard 30 , 52074 Aachen , Germany
-
Affiliation as printed
RWTH Aachen University , Institute for Fluid Power Drives and Systems (ifas) , Campus- Boulevard 30 , 52074 Aachen , Germany
-
Affiliation as printed
RWTH Aachen University , Institute for Fluid Power Drives and Systems (ifas) , Campus- Boulevard 30 , 52074 Aachen , Germany
Cited by 0 stored of 0
No patents citing this paper on Lens.org (checked 2026-10-06).