Making knowledge graphs work for smart manufacturing: Research topics, applications and prospects

Wan, Yuwei, Liu, Ying ORCID: https://orcid.org/0000-0001-9319-5940, Chen, Zheyuan, Chen, Chong, Li, Xinyu, Hu, Fu and Packianather, Michael ORCID: https://orcid.org/0000-0002-9436-8206 2024. Making knowledge graphs work for smart manufacturing: Research topics, applications and prospects. Journal of Manufacturing Systems 76 , pp. 103-132. 10.1016/j.jmsy.2024.07.009

PDF - Published Version Available under License Creative Commons Attribution. Download (18MB)

Abstract

Smart manufacturing (SM) confronts several challenges inherently suited to knowledge graphs (KGs) capabilities. The first key challenge lies in the synthesis of complex and varied data surrounding the manufacturing context, which demands advanced semantic analysis and inference capabilities. The second main limitation is the contextualization of manufacturing systems and the exploitation of manufacturing domain knowledge, which requires a dynamic and holistic representation of knowledge. The last major obstacle arises from the facilitation of intricate decision-making processes towards correlated manufacturing ecosystems, which benefit from interconnected data structures that KGs excel at organizing. However, the existing survey studies concentrated on distinct facets of SM and offered isolated insights into KG applications while overlooking the interconnections between various KG technologies and their application across multiple domains. What specific role KGs should play in SM towards the aforementioned challenges, how to effectively harness KGs for these challenges, and the essential topics and methodologies required to make KGs functional remain underexplored. To explore the potential of KGs in SM, this study adopts a systematic approach to investigate, evaluate, and analyse current research on KGs, identifying core advancements and their implications for future manufacturing practices. Firstly, cutting-edge developments in the challenge-driven roles of KGs and KG techniques are identified, from knowledge extraction and mining to techniques for KG construction and updates, further extending to KG embedding, fusion, and reasoning—central to driving SM ecosystems. Specifically, the KG technologies for SM are depicted holistically, emphasizing the interplay of diverse KG techniques with a comprehensive framework. Subsequently, this foundation outlines and discusses key application scenarios of KGs from engineering design to predictive maintenance, covering the main representative stages of the manufacturing life cycle. Lastly, this study explores the intricate interplay of the practical challenges and advantages of KGs in manufacturing systems, pointing to emerging research avenues.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Engineering
Publisher:	Elsevier
ISSN:	0278-6125
Date of First Compliant Deposit:	1 August 2024
Date of Acceptance:	18 July 2024
Last Modified:	01 Aug 2024 10:50
URI:	https://orca.cardiff.ac.uk/id/eprint/171004

Actions (repository staff only)

Edit Item