PSI - Issue 57

Mehdi Ghanadi et al. / Procedia Structural Integrity 57 (2024) 386–394

391

6

Mehdi Ghanadi et al./ Structural Integrity Procedia 00 (2023) 000 – 000

1

1

log

log

log C N −

  =

m m (3) The nominal stress range versus the numberof fatigue life cycles, S-N curve, of several test specimens, as indicated in Tab.1, is shown in Fig. 5(a). FEA for each test specimen is also carried out and the calculated effective notch stress range is shown in Fig. 5(b). The probabilistic model, considering the most fitted Weibull shape parameter, is then performed, and the result is presented in Fig. 5(c). Comparing these graphs reveals that the probabilistic simulation framework for the fitted value of the shape parameter  reduces the variation in fatigue test data compared to the conventional nominal stress method.

Figure 5. (a) Nominal stress range (fatigue test data), (b) Effective notch stress range and (c) Equivalent stress range ( 12  = ) versus fatigue life It should be noted that the thickness effect for plates thicker than a reference value, which is 25 mm according to the IIW recommendations, is taken into account by multiplication of FAT class by a thickness reduction factor ( ) f t , Eq.(4). In this formula the values of effective thickness eff t and thickness correction exponent n are calculated based on the IIW recommendations (Hobbacher, 2016).

( ) ( ) ref n eff t t =

f t

(4)

Variations of all these three stress ranges, mentioned in Fig. 5, together with the modified nominal stress range, considering the thickness reduction factor in Eq.(4), are plotted against several plate thicknesses, see Fig. 6. The plots are also normalized with respect to their values at reference thickness, Fig. 6(b). Overall, the results show that the stress ranges increase as plates become