Relating the surface topography of as-built Inconel 718 surfaces to laser powder bed fusion process parameters using multivariate regression analysis

The surface topography of 3D printed parts differs from that of wrought or machined parts of the same material. This work specifically focuses on Inconel 718 specimens manufactured with laser powder bed fusion (L-PBF). Predicting the surface topography of the as-built surfaces based on the L-PBF process parameters is important to manufacture parts with a specific surface topography without costly post-processing. We measure the surface topography of Inconel 718 as-built surfaces manufactured with different bulk and contour laser power, scan speed, layer thickness, and build orientation, and perform multivariate regression analysis with traditional Rand S-parameters, statistical and deterministic parameters, and even combinations of surface topography parameters, to determine which parameters most closely relate to the L-PBF process parameters. The deterministic surface topography parameters derived from a 9-point peak-identification scheme resulted in the best-fit regression equations with the highest adjusted-R2 and, thus, showed the closest relationship to the L-PBF process parameters. In contrast to the R-, S-, and statistical parameters, the deterministic approach considers the actual peaks of the surface topography rather than relying on a discrete number of traces of the surface. This conclusion is different from existing knowledge on the surface topography of as-built L-PBF surfaces, which only considers traditional R- and S-parameters.

Automated summary

This study focuses on the surface topography of Inconel 718 specimens manufactured with laser powder bed fusion. Deterministic surface topography parameters derived from a 9-point peak-identification scheme showed the closest relationship to the L-PBF process parameters, in contrast to traditional R- and S-parameters.