Linked Data Architecture for Assistance and Traceability in Smart Manufacturing

Abstract

Traceability systems and digital assistance solutions are becoming increasingly vital parts of modern manufacturing environments. They help tracking quality-related information throughout the production process and support workers and maintenance personnel to cope with the increasing complexity of manufacturing technologies. In order to support these use cases, the integration of information from different data sources is required to create the necessary insights into processes, equipment and production quality. Common challenges for such integration scenarios are the various data formats, encodings and software interfaces that are involved in the acquisition, transmission, management and retrieval of relevant product and process data. This paper proposes a Linked Data based system architecture for modular and decoupled assistance software. Its web-oriented approach allows to connect two usually disparate data sets: semantic descriptions of complex production systems on the one hand and high-volume and high-velocity production data on the other hand. The proposed concept is illustrated with a typical example from the manufacturing domain. The described End-of-Line quality assessment on forming machines is used for traceability and product monitoring.

Availability‐guaranteeing maintenance of series machine tools

“…The data assigned to the information views are often heterogeneous and are located in different systems. Linked data technologies 28 are used to link this information semantically (Semantic Web) 29,30 and to query it via SPARQL requests. Information such as 3D geometries, material data, maintenance history, or service documents can thus be obtained from external systems.…”

Enabling Federated Learning Services Using OPC UA, Linked Data and GAIA-X in Cognitive Production

“…This web-oriented approach allows to combine the semantic descriptions of production systems (stored as RDF) and large amounts of production data (stored in a time series database), i.e. from MES or ERP systems and OPC UA servers [12]. The LF uses a JSON-LD inspired data format -a lightweight linked data format that employs a simple hierarchical structure to represent items and their attributes -and an HTTP-based APIs to exchange the data [27] The attribute values are further qualified by a context (organizational unit, simulation run, etc.)…”

“…While specialized time series databases excel in storing vast amounts of measurements, they are limited by their metadata support capabilities. This predicament calls for a flexible approach that merges Linked Data principles with straightforward data formats and APIs for time series data, laying the groundwork for creating semantic digital twins within the manufacturing domain, as outlined in the LinkedFactory (LF) proposal [12], which advocates for portable SPARQL queries over contextualized time series data in industrial settings.…”

Increasing Resilience of Production Systems by Integrated Design

“…Regarding connectivity and data exchange, standardized interfaces and data models should be established to support transparency and to complete traceability within a single production system as well as along the supply network. To accomplish this, the standard "Open Platform Communications Unified Architecture" (OPC UA) can be combined with linked data approaches [4] for modeling system structures, components, processes, and products as well as for linking them to related operating data.…”