PSI - Issue 66

Yamato Abiru et al. / Procedia Structural Integrity 66 (2024) 525–534 Author name / Structural Integrity Procedia 00 (2025) 000–000

530

6

3.2. Fracture morphologies Fig. 7 presents SEM images of the fracture surfaces of a specimen fractured under a load of P max = 20 kN. Both uncharged and hydrogen-precharged materials exhibit minimal surface roughness, with no significant changes observed. Tanaka et al. (2007) previously reported that flat facets appeared during fatigue testing of hydrogen precharged JIS-SCM435 steel, with uncharged specimens displaying numerous slip bands over a broad area surrounding the cracks. In hydrogen-precharged specimens, however, slip bands were localized near the crack tip, indicating slip concentration. In a prior study at P = 40 kN (Abiru et al. 2024), distinct rough fracture surfaces were noted in uncharged materials, whereas the fracture surfaces of hydrogen-precharged materials appeared flat. As shown in Fig. 7, some uncharged materials displayed pronounced surface roughness on a macroscopic scale. Although no clear facet formation due to hydrogen was observed, it is likely that hydrogen influenced crack growth and accelerated propagation, as indicated by the crack growth trends in Fig. 5.

Fig. 7. Fracture surface morphologies and crack propagation illustration for uncharged and hydrogen-precharged specimens ( P max = 20 kN).

3.3. Nondestructive flaw detection 3.3.1. Eddy-current testing (ECT)

Fig. 8 presents the ECT results, with the voltage representing the peak value measured between 130 and 150 mm from the left end of the specimen. Higher voltage values were observed at 5 kHz compared to 3 kHz, as the RMS voltage is lower at reduced excitation frequencies. This difference is due to the increased sensitivity of eddy currents at higher frequencies. No significant voltage changes were observed for specimens without cracks or dents, for those with dents but no cracks, or for specimens with a crack length a of approximately 2 mm (half the wall thickness), whether uncharged or hydrogen-precharged. Therefore, voltage changes due to cracks were undetectable for 2 mm crack lengths, indicating that the eddy-current method is primarily effective in detecting external cracks. Comparing voltage values for a crack extending fully through the thickness (4 mm, reaching the outer surface), the uncharged specimen displayed a reading of 71.4 V at 5 kHz, while the hydrogen-precharged specimen showed 16.7 V at the same frequency. This suggests that hydrogen-induced effects, potentially causing fewer openings at the crack tip, may reduce eddy-current flow. Such differences in crack propagation patterns could lead to variations in detection sensitivity.