A conceptual vision for a bio-intelligent manufacturing cell for Selective Laser Melting. (August 2021)

• Record Type: Journal Article
• Authors: Wegener, K.; Spierings, A.B.; Teti, R.; Caggiano, A.; Knüttel, D.; Staub, A.
• Is Part Of: CIRP journal of manufacturing science and technology. Volume 34(2021)
• Page Start: 61
• Page End: 83
• Publication Date: 2021-08
• Subjects: Biologicalization -- Additive manufacturing -- Powder bed fusion -- Artificial intelligence -- Quality management system; Manufacturing processes -- Periodicals; 670.5
• DOI: 10.1016/j.cirpj.2020.11.009
• Languages: English
• ISSNs: 1755-5817

Abstract

Abstract: State of the art additive manufacturing (AM) technologies such as Selective Laser Melting (SLM) are traditionally uncontrolled processes, as typically pre-set processing parameters like laser power, scan strategy and speed and many others are used for production. Additionally, the process is influenced by a large number of parameters which today are only covered by statistics. By this such manufacturing systems are not able to actively react to changing conditions as biological systems can do: Higher biological systems are characterized by the availability of a variety of sensor data coming from partly redundant and failure-tolerant sensor systems and intelligence that allow for continuous adaptation and learning. Furthermore, biological systems self-maintain their health and exchange information with other systems to improve their effectiveness. Hence, the adaptation of basic concepts of biological systems to technical systems, and in this case to AM-technologies, seems to be highly beneficial in order to overcome current shortcomings. The paper presents a visionary concept of an SLM system integrating functionalities and behaviors of the operator, and embeds it into a conceptual framework for an intelligent SLM process chain that is capable of self-optimization. It gives an overview on different monitoring technologies and tools of artificial intelligence that need to be used for this concept, and highlights possibilities to apply machine learning approaches in different time scales along the SLM process chain. A rollout to other manufacturing systems can easily be derived from this concept.