PSI - Issue 82

Hang Su et al. / Procedia Structural Integrity 82 (2026) 131–137 H. Su et al. / Structural Integrity Procedia 00 (2026) 000–000

133

3

across both HCF and VHCF regimes. Emphasis is placed on quantifying the influences of microstructure and defect characteristics on crack initiation, early growth and fatigue life distribution under identical loading conditions. Through comparative analyses of S–N data, fractographic features and micro/nano-scale observations, the paper aims to elucidate the transition in fatigue mechanisms between defect-dominated and microstructure-governed regimes, providing insights into fatigue reliability and optimization strategies for advanced titanium alloy components in long-life applications.

Nomenclature σ a

applied stress amplitude number of cycles to failure

N f

stress ratio, the ratio of minimum stress to maximum stress

R

2. Mechanical performance The S–N data obtained under fully reversed loading ( R = –1) reveal clear differences in fatigue performance between the AM and CP titanium alloys, as shown in Fig. 1. For the AM alloy, the S–N curve exhibits a distinct two-slope or duplex feature, indicating a transition from the HCF to the VHCF regime. The first slope corresponds to surface or near-surface crack initiation under relatively high stress amplitudes ( σ ₐ > 400 MPa), while the second slope, characterized by a rapid drop in fatigue life, is governed by defect-induced internal crack initiation. This “stepwise” trend is typical for AM materials containing process-related void-type defects such as LoF pores and gas porosity (Du et al., 2021; Pan et al., 2020a; Qian et al., 2020). As the applied stress amplitude decreases below ~300 MPa, fatigue failures are predominantly initiated from internal defects, resulting in large life dispersion and a non convergent endurance limit up to 10⁸ cycles.

Fig. 1. Four groups of S-N data for additively manufactured and conventional titanium alloys under R = –1.

In contrast, the CP titanium alloy with equiaxed α+β microstructure (EM) presents a relatively linear S–N relationship with a smaller slope and broader data scatter. The fatigue life distribution spans over several orders of magnitude, reflecting the inherent variability of microstructural crack initiation mechanisms (Crupi et al., 2017; Heinz and Eifler, 2016). Although no distinct inflection appears within 10⁸ cycles, a gradual reduction in the slope suggests that subsurface initiation begins to compete with surface damage as the number of cycles increases. The fatigue limit for the CP alloy is higher than that of the AM alloy, and the fatigue data show less sensitivity to stress