Employing spatial and amplitude discriminators to partition and analyze LPBF surface features

Abstract

The surface characteristics of Laser powder bed fusion (LPBF) components are an increasingly important aspect in assessing their performance as well as qualifying them to established standards. Typical LPBF surface features include layer-by-layer staircase edges, laser tracks, various degrees of balling, unfused/partially melted powder, spatter particles, and solidification chevrons. They present differently depending on being located on the component's top, side, up or down skins. Their spatial and amplitude features can also vary widely in magnitude depending on the alloy and process parameters involved. Although a significant body of work exists on characterizing LPBF surfaces, there is still little understanding of the cause-effect relationships they have with LPBF process parameters. Hence, there is limited capacity to tailor/optimize surface characteristics for specific standards/applications, such as fatigue performance, wear resistance, implant-bone biointegration, heat exchanger optimization, etc. Before the causal relationships can be explored, one must understand which LPBF characteristic feature or features are controlling a particular LPBF surface roughness. This work describes a new approach to the multiscale characterization and analysis of surfaces specific to LPBF. The method applies spatial and amplitude discriminators to partition LPBF features into separate topographies, enabling an assessment of their individual contribution to surface roughness. It involved systematically varying the cut-off wavelength over the entire spatial range of the surface and detecting discrete transitions in surface field parameters that identified a change in the dominant surface feature. The cut-off wavelengths where these transitions occurred were then used to partition the surface into separate topographies containing particular surface features. This enabled an independent assessment of their individual contribution to overall roughness. The method was demonstrated on LPBF top surfaces using five process parameter sets spanning from low to high energy density. It was successful in determining the segmentation cut-off wavelengths of each set, enabling an assessment of their individual contribution to overall surface roughness. The results indicate that this method has the potential to enable exploration of the causal relationships between process parameters and specific surface features as well as provide a basis for tailoring LPBF surfaces.