PSI - Issue 5

U. Zerbst et al. / Procedia Structural Integrity 5 (2017) 745–752 U. Zerbst et al./ Structural Integrity Procedia 00 (2017) 000 – 000

750

6

resulting residual lifetimes are not too diffe rent for both cases. A number of random geometries is theoretically “tested” at various stress levels which results in a scatter band of the residual lifetimes. Likewise, an endurance limit (defined for the survival of 10 7 loading cycles in Fig. 6 (a)) can be determined. Such as above, a number of random geometries is “tested” at different stress levels. Based on the results, the probability of component failure before reaching the 10 7 loading cycle limit is determined for each stress level. This finally allows the stochastic determination of the endurance limit. An example of simulated S-N curve data is provided in Fig. 8, an example of the growth of an individual crack in Fig. 9.

Fig. 6: (a) Scheme of the stochastic determination of the finite life fatigue limit and the endurance limit (example: cracks initiation at weld toes); (b) Definition of the parameters statistically taken into account in the analyses (not L and T).

Fig. 7: Multiple crack propagation in a butt weld of S355NL, stress ratio R = -1 (R =  min /  max ); (a) Fracture surface with crack development visualized by annealing and beach marking (b) The same re-plotted and added by the corresponding numbers of loading cycles (c) corresponding results of the simulation using the approach of Madia et al. (2017).

Residual stresses in notched geometries A problem on its own is caused by residual stresses in notched geometries. Residual stresses are introduced during manufacturing, i.e., welding, heat treatment or reshaping. No exhaustive discussion of this topic shall be provided here. The only point to be addressed is that as-manufactured residual stresses may experience redistribution and relaxation during cyclic loading when, as at