Computed tomography based investigation of critical defects and fatigue failure analysis of very high cycle fatigued Ti-6Al-4V produced by laser powder bed fusion

This article aims to predict the crack-initiating defects and to estimate the fatigue limit of additively manufactured Ti-6Al-4V specimens with Murakami ’s empirical area -parameter model. For this purpose, fatigue specimens are manufactured by laser powder bed fusion (LPBF) using different process parameters and selected heat treatment methods. The specimens are scanned using xCT in two different manners: first full in reduced resolution to scan the complete geometry, and second the ROI in high resolution.. Afterwards, the specimens are tested cyclically in the very high cycle fatigue (VHCF) regime. Crack-initiating defects are analysed post-mortem using scanning electron microscopy (SEM). The defect population is examined in the xCT data, in which crack-initiating defects are identified. Using the empirical area -parameter model, the effect of defects on the potential fatigue location is determined. An outlook is given on the performance of the empirical area -parameter model on the prediction of crack-initiating defects and fatigue strength of additively manufactured Ti-6Al-4V specimens. • Lack-of-fusion defects dominate fatigue failure in LPBF Ti-6Al-4V. • El-Haddad fatigue-limit estimates are feasible but show high scatter due to defect size variability. • xCT can image large and small defects even when defect size approaches voxel size. • Low-contrast small defects appear in xCT but hinder precise quantitative analysis. • Defect morphology affects failure, requiring methods beyond the area -parameter model.