Process Monitoring of Melt Pool Spatter at Melt Pool, Layer and Part Scales

Emission of melt pool spatter is a significant concern for laser powder bed fusion AM. When spatter particles land on unfused material, large lack of fusion-type defects can result. This research involves in-process monitoring of spatter generation by machine learning analyses of high speed videos at the melt pool scale, infrared videos of entire builds, and individual images of fused layers. The combination of these is allowing determination of spatter counts and trajectories as spatter is emitted from melt pools, tracking of spatter particles as they are captured by the flow of argon across the build, and identification of spatter after it has landed on powder. This talk will present results showing changes in spatter counts and trajectories as a function of process parameters, the ability of argon flows to direct spatter away from unfused regions, and the correlation of spatter seen in fused layers with build conditions.