PSI - Issue 42

Paulo Mendes et al. / Procedia Structural Integrity 42 (2022) 1752–1761 1759 Paulo Mendes, Rita Dantas, José A.F.O Correia, Nicholas Fantuzzi, Abílio M.P. Jesus / Structural Integrity Procedia 00 (2019) 000 – 000 Figure 4 shows all the P-S-N curves with a probability of survival of 97.7% ( = 97.7% ) for the nominal stress case comparing them all also with the design fatigue curve presented in DNVGL-RP-C203 (DNV GL, 2016). The assessment of these curves was explained before. For this probability of survival ( = 97.7% ) in this case, the Bayesian inference also shows the most fitting distribution curve.

Figure 4 – Comparison of all the = 97.7% P-S-N curves from the different methods for the nominal stress case. Analyzing all the data and results obtained from the statistical analysis, the following main conclusions can be drawn: • The fatigue life extension ratio tends to decrease when the stress level increases; • Bayesian method is the most accurate model to predict the S-N curves for the nominal and hot-spot stress case because the Bayesian inference provides the closest results to the experimental data; • The negative DIC values is easy to understand for the logarithm of the data narrow the scale, resulting in a small variance. Smaller DIC values indicate a better- fitting model and therefore it’s possible to conclude that the nominal stress case presents a better-fitting model when compared to the hot-spot stress case; • Eurocode 3 doesn’t provide any S -N curve for hot-spot stress ranges, therefore the S-N curve with a detail category ∆σ C = 71MPa can be underestimated for this case; • ASTM standard doesn’t provi de guidelines to the assessment of S-N curves for a probability of survival different from 95% (p s = 95%); • The estimation of the shape and scale parameters for the two parameters Weibull distribution through a Bayesian approach could be done; • The three-parameter Weibull model developed by Castillo & Fernández-Canteli (Castillo & Fernández Canteli, 2009) or another statistical distributions could also be used to derive P-S-N curves and compared. Acknowledgements This research work is supported by the following grants: base funding - UIDB/04708/2020 and programmatic funding - UIDP/04708/2020 of the CONSTRUCT - Instituto de I&D em Estruturas e Construções - funded by national funds through the FCT/MCTES (PIDDAC); project grant (UTA-EXPL/IET/0111/2019) SOS-WindEnergy - Sustainable Reuse of Decommissioned Offshore Jacket Platforms for Offshore Wind Energy by national funds (PIDDAC) through the Portuguese Science Foundation (FCT/MCTES); ; AARM4 - High Strength Steels in Metalmechanics 4.0 (POCI-01-0247-FEDER-068492) funded by national funds through the PT2020/COMPETE; and,