Abstract
Open science has the capacity of boosting innovative solutions and knowledge development thanks to a transparent access to data shared within the research community and collaborative networks. Because of this, it has become a policy priority in various research and development strategy plans and roadmaps, but the awareness if its potential is still limited in industry. Additive manufacturing (AM) represents a field where open science initiatives may have a great impact, as large academic and industrial communities are working in the same area, enormous quantities of data are generated on a daily basis by companies and research centers, and many challenging problems still need to be solved. This article presents a case study based on an open science collaboration project between TRUMPF Laser- und Systemtechnik GmbH, one of the major AM systems developers and Politecnico di Milano. The case study relies on an open data set including in-line and in-situ signals gathered during the laser powder bed fusion of specimens of aluminum parts on an industrial machine. The signals were acquired by means of two photodiodes installed co-axially to the laser path. The specimens were designed to introduce, on purpose, anomalies in certain locations and in certain layers. The data set is specifically designed to support the development of novel in-situ monitoring methodologies for fast and robust anomaly detection while the part is being built. A layerwise statistical monitoring approach is proposed and preliminary results are presented, but the problem is open to additional research and to the exploration of novel solutions.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplemental material
All data are available at https://www.ic.polimi.it/open-data-challenge/
Notes
1 https://ec.europa.eu/info/research-and-innovation/strategy/strategy-2020-2024/our-digital-future/open-science_en (last access: 14/07/2021)
2 https://freemelt.com/join-the-game/ (last access: 13/07/2021)
3 https://openadditive.com/wp-content/uploads/2021/01/PANDA_Machine_Control.pdf (last access: 13/07/2021)
4 https://ammd.nist.gov/ (last access: 13/07/2020)
5 https://www.ic.polimi.it/open-data-challenge (last access: 13/07/2020)
6 The data set described in this article is available at the following link: https://www.ic.polimi.it/open-data-challenge/
7 https://www.skills4am.eu/ (last access: 16/07/2021)
Additional information
Funding
Notes on contributors
Marc Gronle
Marc Gronle is the current head of the team on sensors, optics and electronics development within the additive manufacturing division of TRUMPF Laser- und Systemtechnik GmbH.
Marco Grasso
Marco Grasso is an Assistant Professor in the Department of Mechanical Engineering of Politecnico di Milano. His research interests include machine learning and statistical learning techniques for in-situ monitoring of advanced manufacturing processes, with a focus on metal additive manufacturing.
Emidio Granito
Emidio Granito is a former student of Politecnico di Milano. He contributed to the research presented in this paper in the framework of his MSc thesis. Frederik Schaal is the former head of the team on sensors, optics and electronics development within the additive manufacturing division of TRUMPF Laser- und Systemtechnik GmbH.
Frederik Schaal
Frederik Schaal is the former head of the team on sensors, optics and electronics development within the additive manufacturing division of TRUMPF Laser- und Systemtechnik GmbH.
Bianca Maria Colosimo
Bianca Maria Colosimo is a Full Professor and Deputy Head of the Department of Mechanical Engineering at Politecnico di Milano. Her research interests include big data mining & machine learning for advanced manufacturing process monitoring, modelling and control, with a focus on innovative additive manufacturing and 3D bioprinting solutions.