Fatigue behaviour of L-PBF Inconel 718: influence of surface condition and mean stress with statistical evaluation of S–N curves

Laser-powder bed fusion (L-PBF) manufactured Inconel 718 is used in fatigue-critical applications, and the combined effects of surface condition, build orientation, and mean stress are not fully understood. This study investigates the fatigue behaviour of heat-treated L-PBF Inconel 718 with surface states (as-built, sandblasted, and machined), build orientations, and load ratios. Constant-amplitude fatigue tests and evaluated using Kohout–Věchet and Basquin regression, supported by statistical S–N curve comparison, Haigh-diagram-based mean-stress assessment, and Murakami √area-based defect-severity evaluation. Results indicate that fatigue behaviour is influenced by surface condition and crack-initiation-site severity, while build orientation has a minor influence when low porosity and optimized heat treatment are achieved. Machined specimens exhibit the highest fatigue strength, whereas as-built surfaces lead to reduced fatigue life due to surface or near-surface crack initiation sites. Sandblasting provides a moderate improvement, although this may result from reduced surface asperity severity, modified surface morphology and residual stresses, which were not quantified but remains inferior to machining. Haigh-diagram evaluation and comparison with mean-stress models show that tensile mean stress reduces fatigue strength, in the lower-cycle regime, with a diminishing influence towards higher cycle counts. Murakami-based evaluation shows that fatigue life is not governed by defect size alone, but by the combined effect of defect size, defect position, and applied stress level. These findings suggest that L-PBF Inconel 718 can approach fatigue-performance levels for conventionally manufactured material under post-processing conditions, while emphasizing that fatigue design under tensile mean stress requires consideration of surface condition, near-surface defect severity, and mean-stress sensitivity.