Template-bayesian approach for the evaluation of melt pool shape and dimension of a DED-process from in-situ X-ray images

• DOI: 10.1016/j.cirp.2021.03.011
• Publication Date: 2021-04-15T23:42:09Z
• Authors: Adrian Lindenmeyer; Samantha Webster; Michael F. Zaeh; Kornel F. Ehmann; Jian Cao
• Keywords: Deposition

Directed Energy Deposition (DED) is a highly localized process in which metal powders are added to a laser-induced molten pool. The shape and size of the melt pool ultimately determine the local cooling/solidification rate and, thus, the material's microstructure and properties. To study melt pool shape, in situ X-ray imaging techniques have been used. However, the data afterwards typically are manually analysed, which creates a bottleneck in understanding fundamental phenomena in DED. Here, a promising method to automatically extract melt pool shape and dimensions from in situ X-ray DED melt pool images using templates and Bayesian reasoning is proposed.