Industrial Big Data in an Industry 4.0 Environment: Challenges, Schemes, and Applications for Predictive Maintenance

• Published in: IEEE access 2017-01, Vol.5, p.23484-23491
• Main Authors: Yan, Jihong, Meng, Yue, Lu, Lei, Li, Lin
• Format: Article
• Language: English
• Subjects: Big Data; Communications systems; Data mining; Data models; Data processing; Feature extraction; Industrial applications; Industrial big data; Industries; Industry 4.0; Life prediction; Machining; Manufacturing processes; multiple invisible factors; multisource heterogeneous data; Plant reliability; Predictive maintenance; structuralization and characterization; System reliability
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2017.2765544

Industry 4.0 can make a factory smart by applying intelligent information processing approaches, communication systems, future-oriented techniques, and more. However, the high complexity, automation, and flexibility of an intelligent factory bring new challenges to reliability and safety. Industrial big data generated by multisource sensors, intercommunication within the system and external-related information, and so on, might provide new solutions for predictive maintenance to improve system reliability. This paper puts forth attributes of industrial big data processing and actively explores industrial big data processing-based predictive maintenance. A novel framework is proposed for structuring multisource heterogeneous information, characterizing structured data with consideration of the spatiotemporal property, and modeling invisible factors, which would make the production process transparent and eventually implement predictive maintenance on facilities and energy saving in the industry 4.0 era. The effectiveness of the proposed scheme was verified by analyzing multisource heterogeneous industrial data for the remaining life prediction of key components of machining equipment.