Fatigue performance of Ti6Al4V lattices: relative density as a partial quantitative predictor

• Fatigue strength drops by up to 95 % as relative density decreases to 12.5 %. • Normalized fatigue strength still depends on relative density in gyroid lattices. • Effective stress fails to fully explain fatigue strength loss at low densities. • Micro-CT scans revealed RD and wall thickness deviations from as-designed geometry. • FEA captured key aspects of plastic collapse, validating quasi-static behavior. Lattice structures realized through additive manufacturing have garnered increasing interest within both academia and industry in recent years. Various factors, including unit cell topology, base material, heat treatments, and relative density, significantly influence the overall behaviour of these architectured structures. This study specifically examines the compressive mechanical behaviour of solid-based gyroid lattices made of Ti6Al4V alloy through Laser Powder Bed Fusion (PBF-LB) technique. Specimens with four different relative densities were produced to investigate the impact of this parameter on the compressive behaviour (quasi-static and fatigue); furthermore, each relative density category included two sets of specimens to evaluate the effect of annealing and Hot Isostatic Pressing (HIP) as post-processing techniques. Micro-CT scans, microstructural, postmortem and finite element analyses were included to further evaluate the failure mechanisms and explain the observed experimental results. Furthermore, the behaviour documented in the present analysis has been correlated with a wide fatigue dataset retrieved from literature in an effort to dig deeper into the behaviour of these structures. The results, together with the retrieved dataset, allowed for a more comprehensive understanding also considering aspects such as yielding effect, surface roughness and notch mechanics. It has been proved that the use of optimized process parameters and cheaper heat treatments is able to match the beneficial effects expected by HIP. Furthermore, easy-to-use methodologies to account for the reduction in strength due to the change in relative density presented in the literature, such as effective and normalized stress, have been considered to evaluate their accuracy, but also their limitations.