Image-based study on fatigue crack initiation mechanism of Ti-6Al-4V fabricated by laser-based powder bed fusion

Additively manufactured Titanium alloys , specifically Ti-6Al-4V (Ti64) produced through laser based powder bed fusion (PBF-LB), hold great promise for applications in the medical, marine, and aerospace industries. However, PBF-LB Ti64 often exhibits inferior fatigue integrity due to defects introduced during the additive manufacturing process . This study focuses on the initiation mechanism of fatigue cracks, which are pivotal for assessing the fatigue life of as-built PBF-LB components. PBF-LB defects are intricately linked to contour process parameters (e.g., laser power and scan velocity) as well as build orientation. We investigate how these building variables affect the near-surface defect characteristics and fatigue crack initiation . Image-based methods are employed to perform qualitative and quantitative fatigue analyses. The specimens are periodically scanned using micro-CT before and during the fatigue testing, with final fracture surfaces examined using a scanning electron microscope (SEM). A set of efficient image analysis protocols has been proposed to track the crack evolution in 4D along the entire gauge section volume and establish correlations between crack initiation sites and initial defects. Furthermore, a convolutional network-based finite element solver (FEA-Net) has been utilized to directly predict fatigue crack initiation locations. Our findings reveal that the contour building strategies significantly influence near-surface defects and fatigue crack initiation. Importantly, FEA-Net accurately predicts crack initiation sites based solely on cross-section images, showcasing the potential of image-based approaches for advancing our understanding of PBF-LB component fatigue behavior . • A comprehensive image-based study about the fatigue crack initiation mechanism of as-built PBF-LB Ti64. • Contour process strategy and build orientation of PBF-LB specimens are related to defect types and level. • Image-based direct numerical simulation could predict fatigue crack initiation sites.