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

2056

4

From Figure 4 it can already be seen, that all results fall into QL B, if not differently intended. Hence, a further distinction within QL B was made. A quality level “B/2” was introduced. The limit values of this quality level were defined as half of the limit values of quality level B. Not all results of the geometry measurements are displayed here. However, it can be summed up, that most of the results fell into QL B/2 if not differently intended. Detailed information and all results can be found in Dilger et al. (2018).

Figure 4: Measured values for the axial misalignment, Cat 2: Disk laser welds, Cat 3: Electron beam welds, a: Axial misalignment not intended, b: Axial misalignm ent of 0.15∙t intended

4. Fatigue testing

Fatigue testing was carried out on servo-hydraulic testing rigs according to EN 50100 with a stress ratio of R = 0.1 and constant amplitude loading. The failure criterion was breakage of the sample or a load cycle nu mber of N = 5∙10 6 . For the statistical evaluation of the data, linear regression was used, run-outs were not considered in the regression. Additionally to the calculated inverse slope m, a statistical evaluation with a fixed slope of m = 3 in the high cycle fatigue region was conducted. This supplementary evaluation was done, so that the results are better comparable to current standards on the notch stress concept, which recommend a slope of m = 3. The position of the regression line with a fixed slope was defined in such a way that it passed through the center of gravity of all broken samples of each S-N curve. However, an evaluation with a fixed slope leads to an overestimation of the fatigue strength for lower numbers of load cycles and an underestimation of the fatigue strength for higher numbers of load cycles if the calculated slope of the same data set is m > 3. To compare the results of the fatigue tests to IIW FAT-classes, which are mainly based on arc welded joints, the stress range at 2∙10 6 load cycles is displayed in Figure 5 and Figure 6. Furthermore, a mean stress transformation from R = 0.1 to R = 0.5 was done. A mean stress sensitivity factor for welded components according to FKM (2002) was used. The mean stress sensitivity factor is defined as M = 0.3 between R = -1 and R = 0 and M/3 between R = 0 and R = 0.5. A uniform scatter band of 1:1.5 according to Haibach (2005) was used to calculate a probability of survival of 97.5%. The results of the S-N curves with no intended axial misalignment show that the calculated slopes are all higher than the fixed slope of m = 3. The results of the S-N curves show relatively little scatter to each other. No effect of the base metal as well as the sheet thickness can be observed. Comparing the experimental data to the FAT classes that are suggested for arc welded joints, it can be seen, that FAT 71 is reached in all cases. Even FAT 80, which is usually intended for butt joints that are welded from both sides, is reached. Reason for this can be found in the good quality of the welded joints that was also found in the weld geometry characterization.