PSI - Issue 38

Isabel Huther et al. / Procedia Structural Integrity 38 (2022) 466–476 Huther et al./ Structural Integrity Procedia 00 (2021) 000 – 000

474

9

6. Fatigue behaviour after an overload 6.1. Fatigue test conditions More than 60 fatigue tests were performed in tension with a stress ratio R = 0.1 on welded specimens after having applied the overload. On three welded specimens with an overload corresponding to 1 YS, 1.2 YS and 1.56 YS respectively, the fatigue tests are performed step by step, at the stress range level ∆ = 197 , to study the variation of the residual stresses versus fatigue cycles. 6.2. Fatigue results The fatigue results were plotted on Fig. 8, (a) for the tests without overload and with an overload of 1 YS, (b) for the tests with an overload of 1.25 YS and (c) for the tests with an overload of 1.56 YS. The S-N curves were established by the Basquin's formula (3), with the mean curves and design curves at a non-failure probability of 95% with a confidence interval of 75%. The fatigue strength (FAT) at 2M cycles for a non-failure probability of 95% with a confidence interval of 75% is also calculated. All the results are synthetized in Table 2. = − ∙ ∆ (3) For the as-welded joints without overload, the FAT obtained corresponds to FAT 63 which is similar to the FAT defined by the IIW fatigue recommendations in the weld detail 523. An overload of 1 YS on the welded joints is not of effect on fatigue strength (Figure 10a). For the overload of 1.25 YS and 1.56 YS, the FAT values obtained are 74 MPa and 99 MPa respectively, the improvement factor is 1.2 and 1.57 at 2 million of cycles. And the S-N curve slope parameter (m) increases from 3 to 6 for the overload of 1.56 YS.

Table 2. S-N curves parameters [6] m

log C

stdv

FAT

∆  mean

A.W. 1 YS

3

11.981

0.164

79

63

1.25 YS 1.56 YS

3.1 4.5

12.430 15.584

0.141 0.099

93

71

120

100

a

b