PSI - Issue 57

Gloria Hofmann et al. / Procedia Structural Integrity 57 (2024) 452–460 Hofmann, G.; Bartsch, H.; Kuhlmann, U.; Feldmann, M.

454

3

As a visualization of the test results depending on the stress range and the plate thickness in three-dimensional space would be rathercomplex, the test results were normalized to 2 Mio. cycles along the S/N curve with a slope of m = 5 for the ground-flushed butt-welded joints and with a slope of m = 3 for the unmachined butt-welded joints. Despite the filters mentioned above, the test results of both the machined and the unmachined specimens still show a wide scatter as a function of the plate thickness. This can be explained by the many test series having different foci and thus very inhomogeneous conditions. It also shows that other factors also have a noticeable influence on the fatigue strength of the butt-welded joints. However, it is difficult to extract these influencing factors as the test results were gathered from different sources. In the respective test series, different emphases were placed, such as welding imperfections or material. This alone makes comparability difficult. In addition, not all the necessary factors are known for all test series.

2.2. Weld reinforcement removed

After grinding, the geometry of the butt joint is more like that of a simple plate, and as explained by (Braun, et al., 2023) this not only improves the fatigue resistance but also affects the slope of the S/N curve. As with unwelded components, the test results tend to conform to a slope of m = 5. Therefore, the evaluation of the thickness effect for this detail is given with a slope of m = 5. The test results collected in the database and filtered according to the previous chapter are shown in Fig. 1.

550

test results prEN 1993-1-9 EN 1993-1-9

450

350

250

150 Δσ transformed to 2 Mio. cycles [MPa]

50

0

20

40

60

plate thickness t [mm]

Fig. 1: Stress range (transformed to 2 Mio. cycles, m = 5) for ground flush butt-welded joints, (Hofmann, in preparation).

Due to the large scatter and the low data availability for thicker plates, the plate thickness influence on the fatigue strength cannot be statistically confirmed. However, (Braun, et al., 2023) show that the curve of DC 112 for plate thicknesses t < 25 mm and the respective reduction for higher plate thicknesses, both for the current and the upcoming standard, is set quite conservatively and an improvement of the DC per se may be possible.

2.3. As welded

Also, for butt-welded joints as welded, a large scatter field of test results can be seen. However, for this detail, test series were carried out which explicitly focused on the influence of the plate thickness. In Fig. 2 the series that explicitly investigated the plate thickness are highlighted. Among them are also the own test results which were realized within the project (Feldmann, et al., exp. 2024) as well as the collection of data of this project. Since in our experiments all parameters were the same except of the thickness, only the influence of the plate thickness was explicitly investigated, varying the thickness between 10 mm, 20 mm, 50 mm and 80 mm. All other influencing factors were kept as constant, as far as possible. This involved cutting all test specimens with a width of 130 mm along the rolling direction from a steel grade of S355ML. The applied axial stress for all specimens were 200 MPa with a stress ratio of R = 0,05. A visual inspection of the specimens before the tests concerning the reinforcement showed that all details could be classified as DC 90 according to the rules of both, the current and the upcoming Eurocode. According to the upcoming Eurocode (prEN 1993-1-9, 2023), though, butt-welded joints may have a maximum misalignment of < 5% of the plate thickness. Some of the test specimens with 10 mm plate thickness,