Implementing self- autonomic properties in self-coordinated manufacturing processes for the Industry 4.0 context

• Published in: Computers in industry 2020-10, Vol.121, p.103247, Article 103247
• Main Authors: Sánchez, M., Exposito, E., Aguilar, J.
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Autonomic computing; Computer Aided Engineering; Computer Science; Industry 4.0; Machine Learning; Process mining; Self-coordination; Self-supervising
• ISSN: 0166-3615; 1872-6194
• DOI: 10.1016/j.compind.2020.103247

•This paper shows the use of Everything mining, Autonomic computing and IoE to add self-* properties to manufacturing systems.•An autonomic supervisory system was built with the purpose of allow self-supervising of manufacturing processes.•The supervisory system uses two everything-mining techniques: process mining and big data mining.•The process-mining gets useful insights from the manufacturing process, that is used to detect failures.•The data mining builds a predictive model that is used to predict whether or not a product will fail the quality control test. Industry 4.0 requires high levels of autonomy in order to guarantee the manufacturing processes to achieve production goals. For this, it is needed high levels of coordination, cooperation, and collaboration, such that the manufacturing process’ actors can communicate and interoperate. A previous paper proposed three autonomic cycles of data analytics tasks for self-coordination in manufacturing processes. In this paper, we implement one of these autonomic cycles, allowing self-supervising of the coordination process. This autonomic cycle is designed using the MIDANO’s methodology, and implemented and tested using an experimental tool that was developed to replay the production process event logs, in order to detect failures and invoke the autonomic cycle for self-healing when needed.