Fatigue of AM TiAl6V4: Probabilistic Modelling of the Size Influence as a Transfer Method From Test Specimen to Part

ABSTRACT

This paper investigates the fatigue behavior of additively manufactured TiAl6V4 by comparing two specimen geometries with different highly stressed volumes. A fracture mechanical approach based on the El‐Haddad–Smith–Topper model is combined with defect analysis through X‐ray computed tomography. A combined generalized extreme value and generalized Pareto distribution framework correlates defect size distributions with failure probability across all stress levels. Results show that larger highly stressed volumes increase failure susceptibility due to higher likelihood of critical defects, reducing the fatigue stress range at 50% failure probability from 882 to 611MPa. For topology‐optimized parts, components with localized stress concentrations are less susceptible to defects than those with homogeneous stress distributions. The highly stressed volume approach can yield misleading results for such geometries. The methodology enables transfer of laboratory results to components by integrating stress states, defect statistics, and volume effects.