PSI - Issue 75

D. Tousse Tchamassi et al. / Procedia Structural Integrity 75 (2025) 450–456 Tousse Tchamassi / Structural Integrity Procedia (2025)

455

6

3.2.2. 3D observation with X-ray tomography Fig. 6 shows longitudinal and transverse cross-sections calculated from the reconstructed images obtained after X ray tomography analysis of the specimen of the undeformed material tested at ±1.0%. A projection of all cracks on the plane perpendicular to the applied load is also shown. Multiple crack initiation from the side surface was confirmed (Fig. 6e). This shows that the fracture event limiting the fatigue lifetime was not crack initiation. Even when a crack was detected at the extensometer attachment location, the main (deeper) cracks (located in front of the broken lines in Fig. 6d-e) were found far from the two extensometer attachments (which location along the specimen axis is indicated with continuous lines in Fig. 6d-e). This further confirms that the extensometer attachments did not trigger the main cracking events and thus did not drive the LCF lifetime. Moreover, the selected sections showed that the fatigue cracks were generally not perpendicular to the loading axis, as they were not contained in the sections. In LCF tests, crack initiation is generally expected to be driven by local accumulation of plastic deformation along planes tilted with respect to the loading axis, and not by the maximal principal stress. The present results at the macroscopic scale agreed with this mechanism.

Fig. 6. Reconstructed 3D image of the specimen of undeformed material fatigue tested at ±1.0%. (a, b): selected transverse sections crossing the two main cracks; (c,d): projections on two perpendicular planes containing the loading axis; broken lines indicate the sections in (a, b); continuous lines indicate the location of the extensometer attachments. (e) Projection of all sections on a plane perpendicular to the loading axis, Dark grey regions show the cracks that significantly propagated during the test. 3.2.3. Observation of surface cracks The side surfaces of unbroken specimens showed a large amount of transgranular microcracks nucleated from persistent slip bands (PSBs) (Fig. 7). As expected in LCF, crack initiation occurred by slip localization into PSBs and was neither affected by the specimen polishing procedure nor by the extensometer attachment method (Fig 7a). Surface microcracks were tilted with respect to the loading axis, in agreement with macroscopic observations. Many crack deviations were observed (Fig. 7b), with a mean spacing close to 20 µm, i.e., close to the size of larger parent austenite grains and martensite packets. This suggests that microcrack nucleation and propagation phenomena were linked to plastic deformation mechanisms and that high-angle boundaries hindered early crack propagation. Thus, the size and orientation of microcracks was consistent with those of slip planes (inside individual crystals) in the material, as shown by PSBs in Fig. 7b.