PSI - Issue 57

Yevgen Gorash et al. / Procedia Structural Integrity 57 (2024) 611–617

616

6

Y. Gorash et al. / Structural Integrity Procedia 00 (2023) 000–000

d

c

a

b

Fig. 7. Cross-sections of USF testing sample no.37 (duration of 166m cycles at stress amplitude of 264 MPa) obtained with CT-scanning showing welding pores: a) big internal Ø = 450 µ m; a) small internal Ø = 120 µ m; c) medium internal Ø = 180 µ m; d) small sub-surface Ø = 70 µ m.

d

c

a

b

Fig. 8. Cross-sections of USF testing sample no.45 (duration of 81.8m cycles at stress amplitude of 239 MPa) obtained with CT-scanning showing welding pores: a) small internal Ø = 150 µ m; a) big open Ø = 480 µ m; c) big internal Ø = 750 µ m; d) small internal Ø = 110 µ m.

small sub-surface pore (see pore in Fig. 5 with Ø = 75 µ m) to 100 million cycles for a relatively big pore sitting away from the surface (see the pore in Fig. 6 with Ø = 470 µ m which is 640 µ m away frow surface). Another prove of the large variation of pore size and location can be demonstrated with CT-scanning results of two randomly picked samples cut from the plate ID no. ZL805 shown in Fig. 3. Figure 7 shows the cross-sections with pores in sample no.37 that lasted 166 million cycles at stress amplitude of σ a = 264 MPa. Crack in this sample has initiated on the big internal pore in Fig. 7a with diameter of Ø = 450 µ m. On other hand, Fig. 8 shows the cross sections with pores in sample no.45 that lasted 81.8 million cycles at stress amplitude of σ a = 239 MPa. Crack in this sample has initiated on the big open pore in Fig. 8b with diameter of Ø = 480 µ m. Compared to the sample no.37, sample no.45 has similar stress amplitude level (10% di ff erence) and pore diameter (7% di ff erence), but twice shorter fatigue life. This can be explained by fact that sample no.37 had both phases of crack initiation and crack propagation with internal crack front, whilst sample no.45 had only phase of crack propagation with external crack front. To address this issue in the analysis of experimental results, a Fatigue Performance Parameter – FPP ( P fp ) is applied to the fatigue data for welded samples to quantify the quality of the material: P fp = ln N · σ a , (1) where N is a number of cycles to failure and σ a is a nominal stress amplitude. Welds made of currently preferred steel S355JR + AR and the candidate steel S275JR + AR were tested at ultrasonic frequency of 20kHz in two conditions – polished surface and pre-corroded. The obtained experimental data was aggregated and processed using a FPP ( P fp ) as shown in Fig. 9 to identify a better performing grade. Comparison of P fp for parent materials with polished surface demonstrated a significant advantage of 12.8% for S355JR + ARgrade which is attributed to a higher yield strength resulting in a longer time to crack initiation. However, this advantage reduces to just 4.4% di ff erence in P fp for welded samples with polished surface, where the cracks mostly initiate from the welding pores. This observation proves that internal welding defects eliminate any benefits of higher yield strength by a sharp reduction of the time to crack initiation. But the most important finding is that the comparison of welded samples in the 1 month pre-corroded condition shows a slight benefit of 1.8% for the candidate grade S275JR + AR. It should be noted that in pre-corroded samples the cracks mostly initiate from the corrosion pits on the surface, that proves a higher damaging e ff ect of corrosion on the fatigue performance compared to the negative 5. Discussion