Additive-manufactured metal parts show fatigue behavior controlled by geometry, material state, and manufacturing defects. Predicting fatigue life therefore requires metrics that capture stress distribution, defect severity, and microstructural response under cyclic loading. Guidance from the National Institute of Standards and Technology and standards developed by ASTM International support using combined mechanical and defect-based predictors rather than single-number allowances. Markus Seifi National Institute of Standards and Technology emphasizes the need to quantify both defect populations and local stresses to improve life prediction.
Mechanical and loading metrics
The classic stress-based approach remains central: S-N curves and maximum cyclic stress predict fatigue for large smooth specimens. For printed geometries, however, nominal stress often underestimates local conditions. Fatigue crack growth metrics from fracture mechanics, such as the stress intensity factor range Delta K, outperform S-N methods when small cracks or defects dominate initiation. Load ratio and mean stress effects are particularly relevant for metals produced by powder bed or directed energy deposition, because residual stresses alter effective cyclic loading.
Defect and microstructure metrics
Defects uniquely produced by additive processes mean that defect size and location are strong predictors. Quantitative metrics like the projected defect area or equivalent circular defect size correlate with life in lab studies. Porosity fraction and the largest internal void size measured by X-ray computed tomography directly influence initiation life, while surface roughness controls early crack nucleation on external surfaces. Microstructural descriptors such as grain orientation, cellular substructure, and phase distribution determine crack propagation rates; this is why metallographic measures and orientation maps are commonly reported alongside mechanical tests.
Relevance, causes, and consequences
Causes of variability include process parameters, powder quality, and post-processing such as heat treatment or machining. The consequence of ignoring combined metrics is non-conservative design: unexpected early failure can lead to safety risks and economic losses, especially in aerospace and medical implants where territorial certification regimes impose strict requirements. ASTM Committee F42 ASTM International and research at NIST recommend integrating non-destructive evaluation, probabilistic defect statistics, and local stress computations into certification workflows. Cultural adoption of these complex metrics varies by industry and region, with heavily regulated sectors moving faster toward rigorous, evidence-based prediction methods.