Deep learning for process monitoring of additive manufacturing

A Concept for the Prediction of Material Porosity of Additive Manufactured Components by Deep Learning. Material porosity of components produced by additive manufacturing (AM) such as Laser Powder Bed Fusion (L-PBF) and Laser Directed Energy Deposition (L-DED) is related to process parameters, e.g., layer thickness and build-up rate. To enable the in-situ process monitoring of AM, deep learning is a promising solution, in which heterogeneous dara sets such as process parameters, CAD models and thermal images of layers can be used as training data. The trained model can predict the porosity of components manufactured with AM in-situ. © Carl Hanser Verlag GmbH & Co. KG