PSI - Issue 19

Jennifer Hrabowski et al. / Procedia Structural Integrity 19 (2019) 267–274 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

271

5

with gap made of CHS is 45 for a ratio t 0 /t i = 1.0 and 90 for t 0 /t i = 2.0. Values in between can be interpolated, so that for the series investigated, a detail category 71 for t 0 /t i = 1.6 is relevant.

Fig. 5. Fatigue test results of CHS K-joints.

In the evaluation of the  -N f diagram in Fig. 5, it is noticeable that the test points are far above the red design line (DC71 according to EN 1993-1-9 (2010)). The fatigue tests resulted in two outliers (marked with an arrow). The reason for this early cracking will be clarified by a fracture surface analysis. The statistical evaluation of the 12 remaining tests results in a fatigue strength at 2 million load cycles with freely calculated inverse slope of m = 3.8 at 95% survival probability of  N/mm². Based on the given inverse slop m = 5.0, then even an outlier is still above the notch class 90 (green). This is two detail classes higher than EN 1993-1-9 (2010) and by Zhao et al. (2001) predict. One possible reason is seen in the welding. As already mentioned, the weld execution has a decisive influence on the fatigue strength. For the CHS-joints, contrary to the recommendations (Fig. 6a), the weld was not passed around the brace. But as shown in Fig. 6b), the welding starts at the crown heel and runs in one pass over the saddle and the crown toe over the gap and over the crown toe and saddle and ends at the crown heel of the second brace. The welding of the 10 mm thick braces is executed with a root pass and three top layers with a weld preparation according to EN 1090-2 (2018).

Fig. 6. Weld execution for CHS-joints a) recommendation b) implementation.

The advantage is that in the gap region no start- or stop position of the weld is needed. With this procedure, the highly stressed gap area is bridged and reinforced by the multi-layer weld. This results in a shift of the crack initiation