Issue 78

T. Nasri et alii, Frattura ed Integrità Strutturale, 78 (2026) 275-297; DOI: 10.3221/IGF-ESIS.78.17

Specimens and test conditions All specimens were extracted from the same 30 mm-thick rolled plate such that the specimen mid-plane, in the through thickness direction, was located 10 mm below one plate surface – rather than at the exact plate mid-thickness (15 mm) – in order to avoid both near-surface effects (rolling-induced residual stresses, cooling-rate gradients) and potential core segregation. Prior to testing, all faces were mechanically polished using SiC abrasive papers (320, 600, 800, and 1200 grit), followed by diamond pastes of 6, 3, and 1 µm. This procedure removes surface residual stresses and enables direct optical observation of crack propagation.

Figure 3: Specimen geometries (mm). (a) Smooth specimen ( ௠௔௫ = 300 MPa, = 0.1). (b) Holed specimen, type A (crack ⊥ ஽ ). (c) Holed specimen, type B (crack ∥ ஽ ). Identical hole geometry (90 × 35 × 5 mm, Ø10 mm, ௧ = 2.3, ௠௔௫ = 200 MPa) for both types. Flat dogbone specimens were machined with the loading axis parallel to the rolling direction and tested at max  = 300 MPa and R = 0.1 ( t K ≈ 1.1). These specimens were used to identify the Paris constants C and n for the reference material. The identified values - C = 1.23 × 10 ⁻⁶ m/cycle·(MPa √ m) ⁻ ⁿ and n = 1.77 – constitute the level-1 reference of the hierarchical identification strategy (see "Parameter Identification" below). Type A specimens were machined with the loading axis parallel to the rolling direction (  ∥ D R ). A central hole of 10 mm diameter introduces a stress concentration factor t K = 2.3. The crack initiates at the hole edge perpendicular to the loading direction and propagates perpendicular to D R . The relevant grain size for the enriched model is 154 µm TL d d    , with an initial crack length 0 A a  = 35.4 µm. In agreement with Chen et al. [14,15], type A cracks preferentially interact with recrystallised/non-recrystallised interfaces (high-angle boundaries, θ > 30°), leading to pronounced oscillations in d a /dN . Type B specimens were machined with the loading axis perpendicular to the rolling direction ( D R   ). The hole geometry is identical to that of type A specimens ( t K = 2.3). The crack propagates parallel to D R ; the relevant grain size is L d d   = 395 µm, with an initial crack length 0 B a  = 31.5 µm. For the smooth (dogbone) specimens, a single specimen was tested under a step-loading protocol ( max  = 200 MPa, then 250 MPa, then 300 MPa, each level applied after the absence of measurable damage at the previous level over several million cycles), with the 300 MPa level used to identify the reference Paris constants. For the holed specimens, three Type A specimens were machined, of which two were fully monitored in fatigue (crack initiation and propagation), and two Type

279