Scan-based fatigue prediction of DED-Arc additively manufactured steel components

The Directed Energy Deposition process using an electric arc (DED-Arc) enables the manufacturing of customized steel components for civil engineering applications, yet fatigue performance is governed by the characteristic wavy as-built surface resulting from the layer-wise material deposition. Quantifying these notch effects resulting from the surface topography is an essential step for a reliable fatigue analysis of DED-Arc components. In this paper, we present a novel framework for reproducible scan-based fatigue analysis, linking the as-built surface topography to fatigue life prediction without the need for geometric reverse engineering of the 3D-scanned surfaces. To demonstrate its applicability, the workflow is demonstrated on high-strength low-alloy (HSLA) steel DED-Arc fatigue specimens with distinctive as-built surface topographies, resulting from different DED-Arc printing parameters. Full-field stresses of the 3D-scanned components were computed using the Finite Cell Method (FCM). Local stress concentration factors (SCF) were extracted using multiple variants of the Theory of Critical Distances (TCD) and validated against the results from uniaxial tensile fatigue testing. The volumetric TCD evaluation achieved the highest conformity between the location of the numerical stress concentrations and the experimentally observed fracture locations. Fatigue life was predicted using a local strain model and a regression model with the SCFs resulting from TCD evaluations as scaling factors to the nominal stress ranges from testing. The regression model showed high agreement, with predictions falling within the 3.0 scatter band of the experimental fatigue lives, while the local strain model provided reasonable agreement for fatigue life N < 5 ⋅ 1 0 5 cycles. A more regular as-built surface topography consistently resulted in lower local SCFs and higher fatigue life. The proposed workflow enables reproducible and transferable fatigue life predictions for DED-Arc as-built components across varying surface topographies.