PSI - Issue 66

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

531

7

80

3kHz 5kHz

70

60

50

10 RMS value, R (mV) 20 30 40

0

-10

With crack a = 4.0 mm Uncharged With crack a =4.0 mm Hydrogen precharged

With crack a =2.0 mm Uncharged With crack a =2.0 mm Hydrogen precharged

Without crack and dent

With dent Without crack

Fig. 8. Voltage values for each specimen from ECT.

3.3.2 Hammering test (HT) Fig. 9 shows the hammering test results for a fixed-end beam at both ends, as shown in Fig. 3. The horizontal axis of the graph represents frequency, while the vertical axis represents peak intensity via sound pressure. Notably, a peak consistently appeared in the 8–9 kHz range across all specimens, likely due to fixture, hammer, or component vibrations, as it remained unaffected by cracks. Therefore, the focus here is on peaks outside the 8–9 kHz range at point B. For the uncharged specimen, a peak at 5 kHz was observed for the crack-free specimen ( N = 0 cycles). However, for the specimen with a crack length of a = 0.778 mm ( N = 70,000 cycles), the 5 kHz peak diminished, with a prominent peak emerging at 3.1 kHz. Similarly, the specimen with a longer crack length of a = 2.04 mm ( N = 120,000 cycles) also showed a peak at 3.1 kHz, suggesting that crack presence alters the material's vibration characteristics, influencing the impact sound testing results. No significant difference in frequency response was observed between points A and C for different crack lengths, but this may be due to differences in location-specific vibration characteristics, making it difficult to distinguish between crack lengths at these points. Fig. 10 presents hammering test results for hydrogen-precharged specimens, including the crack-free specimen data from Fig. 9 for comparison, as hydrogen presence alone does not affect detection results. At N = 4,000 cycles, as with uncharged specimens, the 5 kHz peak disappeared, and a broad peak appeared in the 3–4 kHz range. In contrast, at N = 5,000 cycles (crack length of approximately 2.04 mm), a prominent peak emerged at 3.1 kHz. Although the crack length for N = 4,000 cycles is unknown, linear interpolation from Fig. 5 estimates it at approximately 1.93 mm. The exact crack length is unknown, but the absence of a peak at 5 kHz suggests that the crack may have already formed at N = 4,000. Based on the behavior in the 3–4 kHz range, which is similar to that of the crack-free specimen around N = 4,000, it can be inferred that these peak shifts correspond to the transition from crack-free to crack initiation. These findings suggest that crack detection is possible in both uncharged and hydrogen-precharged specimens, as crack propagation occurs even in shorter specimens. According to Fig. 8, ECT struggled to detect cracks when they reached half the material thickness. However, in hammering test, the critical crack length at which peaks emerge in hydrogen-precharged specimens may be smaller. For uncharged specimens, cracks were detectable at a length of 0.78 mm or more, while for hydrogen-precharged specimens, cracks were detectable at lengths of 2.04 mm or longer. No crack-specific peaks appeared below 1.93 mm in hydrogen-precharged specimens, suggesting that crack detection is more challenging in these materials, likely due to reduced crack opening during propagation. Due to the small cycle difference between 4,000 and 5,000, the extent of crack growth remains uncertain. Furthermore, the crack length of 1.93 mm was estimated by linearly interpolating crack lengths at N = 5,030 and