PSI - Issue 13

Ann-Christin Hesse et al. / Procedia Structural Integrity 13 (2018) 2053–2058 Ann-Christin Hesse et al./ Structural Integrity Procedia 00 (2018) 000 – 000

2057 5

Figure 5: S- N curve results at n = 2∙10 6 load cycles in comparison to FAT classes as results of the linear regression (varying slopes of the S-N curves)

Figure 6: S-N curve results at n = 2∙10 6 load cycles in comparison to FAT classes as results of the linear regression with a fixed slope of m = 3

Figure 8: S-N curve results of the welds with an axial misalignment of 0.15∙t at n = 2∙106 load cycles in comparison to FAT classes a s results of the linear regression with a fixed slope of m = 3

Figure 7: S-N curve results of the welds with an axial misalignment of 0.15∙t at n = 2∙106 load cycles in comparison to FAT classes as results of the linear regression (varying slopes of the S-N curves)

Figure 7 and Figure 8 show the cumulated results of the samples with an axial misalignment 0.15∙t in comparison to FAT classes. It is apparent, that the axial misalignment in the samples leads generally to lower fatigue strengths. This effect is based on two major factors: On the one hand, the connecting area between the two sheets is lower; on the other hand, the test set-up lead to a superimposed torque, as testing was carried out using additional sheets in the clamping of the samples to compensate the axial misalignment. This clamping situation is more severe than the situation in usual applications. The axial misalignment of 15% of the sheet thickness is not covered in the standardly used FAT-classes. Jonsson et al. (2016) give advice when it comes to axial misalignments that are higher than the ones that are covered by the FAT classes. In Jonsson et al. (2016), FAT 63 is advised if the axial misalignment is ≤ 0.15∙t and thus within quality level C. FAT 45 is advised if the axial misalignment is ≤ 0.25∙t and thus falls into qualit y level D. One of the eight investigated S-N curves did not reach FAT 63. One reason for this could be the fact that the average axial misalignment of this S-N curve was slightly higher than intended, see Figure 4. Another reason for this could be the fact, that clamping of the samples led to a superimposed torque. However, the use of FAT 56 seems a conservative recommendation.