PSI - Issue 83

Amir Hossein Mirzaei et al. / Procedia Structural Integrity 83 (2026) 239–245

240

heterogeneity, process-induced defects, rough surface features, and residual stress states that are not typically present to the same extent in wrought materials (Li et al., 2016). Consequently, fatigue properties may vary markedly with process selection and post-processing, and robust characterization often requires extensive experimental campaigns. In conventional fatigue characterization, S–N testing is typically performed across several stress amplitudes to establish design-relevant fatigue metrics. However, the limited number of available specimens per AM build and the cost associated with repeated testing across parameter sets motivate accelerated approaches capable of reducing experimental time and specimen consumption while still providing reliable fatigue-strength estimates (Milner et al., 2021). Alongside bulk AM parts, AM has enabled widespread adoption of architected lattice structures, where global behavior is controlled by unit-cell topology and geometric design parameters. Triply periodic minimal surface (TPMS) lattices, and in particular gyroids, are of special interest because their continuous surface-based topology can provide an advantageous balance between load-bearing capability and interconnected porosity. A governing design variable for such lattices is the relative density (RD), which strongly affects the effective load-bearing area and therefore influences both quasi-static response and fatigue behavior. Recent studies on Ti-6Al-4V gyroids have demonstrated that RD changes can significantly modify compressive fatigue performance and that simple normalization strategies may not fully eliminate RD dependence across datasets (Foti et al., 2025). Furthermore, RD related geometric choices interact with manufacturability and defect sensitivity. For instance, thicker gyroid designs can increase stiffness and compressive strength, yet reduced fatigue strength may still occur when surface quality and internal defect content are unfavorable (Mahmoud et al., 2020). Consistently, fatigue studies performed at increasing design porosity levels (decreasing RD) report a clear reduction in fatigue strength; gyroid lattices can outperform alternative topologies at the same RD, but fatigue failure remains highly sensitive to surface-connected defects and partially fused powder particles (kaya et al., 2025). One promising accelerated fatigue-testing approach is the step-loading (SL) method (Nicholas, 2002), in which a single specimen is subjected to constant-amplitude blocks at incrementally increased stress until failure, and a fatigue-strength metric is inferred from the applied loading sequence. Because step-loading is intrinsically a variable-amplitude history, the interpretation of SL outcomes must account for the potential influence of prior blocks on subsequent damage evolution, especially in defect-sensitive AM materials. To the authors’ knowledge, the applicability and reliability of SL-based accelerated testing for AM gyroid lattices, particularly regarding how RD affects agreement with conventional S–N-derived fatigue strengths, has not yet been systematically assessed. In this work, the SL procedure originally proposed for fatigue-limit estimation is evaluated for PBF-LB Ti-6Al-4V gyroid lattices produced at different relative densities, with SL-derived fatigue-strength estimates compared directly against reference values obtained from conventional S–N data at the same life level (Foti et al., 2025). The remainder of the paper introduces the SL approach, describes the material, manufacturing route, and fatigue-test program, and then discusses the accuracy, limitations, and RD-dependence of fatigue-strength estimation in architected gyroid lattices. 2. Theory and concept In the step-loading (SL) method, fatigue strength is estimated by subjecting a specimen to successive blocks of constant-amplitude cyclic loading with progressively increasing stress. The test starts from an initial stress level, σ 0 , selected below the expected fatigue strength, and each stress level is maintained for a prescribed number of cycles, N b . If the specimen survives the block, the maximum stress is increased by a constant increment, ∆ σ , and the next loading block is applied. This sequence is continued until failure occurs during the final block at stress level σ f , as schematically shown in Fig. 1. In this way, one specimen is used to probe several stress levels within a single test, which makes the method attractive as an accelerated alternative to conventional constant-amplitude fatigue testing. Following the formulation proposed by Nicholas (2002), the fatigue strength associated with the adopted block length can be estimated from the position of failure within the last block. If failure takes place after cycles in the final block, the corresponding fatigue strength, , is obtained from: -1 f s f b N N      = +     (1)