Statistical Correlation Analysis of Surface Roughness of Micromilled 316L Stainless Steel Components Fabricated by FDM–FFF Hybrid Manufacturing

This study evaluates the surface roughness of micromilled 316L stainless steel parts fabricated via fused filament fabrication (FFF) and sintering, establishing statistical links between additive manufacturing and post-machining parameters. The surface roughness of the final part is affected by both 3D printing and micromachining parameters. The presented work has direct practical relevance because micromilled 316L stainless steel components are frequently used in applications such as lab-on-a-chip (LOC) devices and micro-electro-mechanical systems (MEMS), where fatigue behavior and the rheological behavior of fluid flow play critical roles. Both fluid flow and fatigue performance of micromilled components are highly dependent on surface integrity, including surface roughness, residual stresses, and microstructure. Specimens were produced using a 3D printer, under controlled layer thicknesses, raster angles, and fabrication directions, followed by a sintering process for the 3D-printed parts. The sintered parts are then micromilled at varying cutting directions (Angle Cut). Surface roughness (Ra) was measured with a profilometer, generating 34 experimental datasets analyzed through correlation and regression modeling. Cutting direction (Angle Cut) exhibited the strongest positive correlation with Ra (r = 0.486, <i>p</i> = 0.004), followed by layer thickness (r = 0.326, <i>p</i> = 0.060), whereas raster angle and fabrication direction had minimal influence. The multiple linear regression model accounted for 33.5% of Ra variance (R² = 0.335, <i>p</i> = 0.0158), highlighting that fine-layer deposition and alignment of tool paths with filament orientation significantly improve post-machined surface quality. Results confirm that additive-induced anisotropy persists after sintering, affecting chip formation and surface morphology during micromilling. The novelty of this work lies in its integrated hybrid framework, linking metal FFF process parameters, fabrication direction, and machining outcomes through a unified statistical approach. This foundation supports machine-learning-based prediction and hybrid process optimization in metal FFF systems, providing guidance for high-quality additive–subtractive manufacturing.