In-process digital monitoring of additive manufacturing: Proposed machine learning approach and potential implications on sustainability

Abstract

Additive Manufacturing (AM) technologies have recently gained significance amongst industries as well as everyday consumers. This is largely due to the benefits that they offer in terms of design freedom, lead-time reduction, mass-customization as well as potential sustainability improvements due to efficiency in resource usage. However, conventional manufacturing industries are reluctant to integrate AM within their established process chains due to the unpredictability of the process and the quality of the final parts that are printed. Conventional manufacturing process have the advantage of decades of research in developing process knowledge and optimization, which culminates in accurate process predictability. This gap in process understanding is one that AM will need to cover in a short time. AM does have the benefit of being a digital manufacturing process and with the adoption of advanced Artificial Intelligence (AI) and Machine Learning (ML) techniques in production lines, there may not have been a better industrial age for its implementation. This paper presents a case for actively developing AM processes using ML. Then a method for in-process monitoring of the printing process is presented and discussed. The main benefit from using the proposed system is an increase in the efficiency and final quality of the parts printed, as a result of which there is an increased efficiency in resource usage due to preventing material loss due to failed builds and defected parts.