PSI - Issue 75

Ralf Glienke et al. / Procedia Structural Integrity 75 (2025) 474–488

481

8

Ralf Glienke et al. / Structural Integrity Procedia 00 (2019) 000 – 000

on the top or bottom side of the specimen, starting from the mill scale. As a result of the cutting process, a localised hardening in the cut edge with a low surface roughness was observed. This contributes significantly to increasing the crack initiation life, which can also be visualised by the flat S-N curve (see Fig. 7e). The 3 rd crack location was directly observed at the cut edge. The blast-cleaning process also shifted the crack location of S355-OF-BC (#10) towards the cut edge. For the laser-cut specimens, (#08) the failure location was not shifted due to blast-cleaning process. Failure of the base material specimens (from #11 to #16) without post-treatment was detected starting from the mill scale or at the corner. Regardless of the blast-cleaning treatment, the flat bars with central hole failed entirely at the edge of the hole, where the greatest notch effect is present.

Fig. 5. Location of typical crack initiation during fatigue tests on the specimens of series #01, #05, #09, #11 and #17

3.5. Fatigue test results Following, the results are discussed based on the statistical evaluation according to the Eurocode 3 background document (ECCS (2018)) for the test-related slope parameter m and the probabilities of survival Ps = 50 % and Ps = 95 %. The improvement in fatigue strength is calculated using the factor k Imp = Δ  C,Imp /Δ  C , which corresponds to the ratio of the mean fatigue resistance (P S = 50 %) of the post-treated specimens to the reference (mill scale). This corresponds to the interpretation of Ahola et al. (2023) and Günther et al. (2009). A positive value for k Imp means an improvement in fatigue strength due to blast-cleaning, while a negative value is associated with a reduction. The ratio m BC /m represents the relative change in the slope parameter of the S-N curve due to post-treatment. A flatter S-N curve is linked to a longer crack initiation phase (m BC / m > 0). 3.5.1. Findings on transverse loaded butt-welded joints The statistical evaluation of the fatigue test results of butt welds of plate thickness t = 20 mm in the as-welded condition (#01, Fig. 7a) provides a characteristic fatigue strength (P S = 95%) of Δ  C = 111 N/mm² for the test-related slope parameter m = 3.7. For the constant slope parameter m = 3, the transverse loaded butt welds can be classified into DC 100 (m 1 = 3). The reference DC 90 (m 1 = 3) for butt-welded joints in plates and flat steels, is based on the evaluation of 2,843 test results from 199 test series (Feldman et al. (2019)), whereby the weld toe angles or excess weld metal were documented only for a very small number of tests. The base materials and welding processes, which include many manual processes, are often also not documented. DC 90 must therefore be understood as the lower limit. For the higher weld seam quality of submerged arc welding with a serial character of the production, the classification in the next higher DC 100 is possible from the authors point of view. This was also validated based on the effective notch stress concept and fracture mechanics as well as component tests (scale effect), see Glienke et al. (2024b). The findings can also be confirmed for the specimens of plate thickness t = 40 mm (#05, m = 3.9, Δ  C = 116 N/mm², Fig. 7d), whereby the effect of size k S = (25/t) 0.2 appears to be too conservative. With the aim of testing in the finite life region, both the as-welded specimens (#01, #05) and the specimens with blast-cleaning treatment (#02, #06), were tested at R = 0.3 downwards from the yield strength. The blast-cleaning treatment has a positive effect on the fatigue strength for both plate thicknesses. This results in improvement factors of k Imp = 1.36 (#02) and k Imp = 1.31 (#06). The slopes of the S-N curves also become flatter (m = 7.2 and m = 6.6). Significant increases in fatigue life can be recognised only from stress ranges below Δ  = 200 N/mm². However, it was not possible to confirm fatigue strengths as high as those of Shojai et al. (2023) or Hensel et al. (2019), who carried out the fatigue tests at R = 0.1. In the following, the authors decided to consider the scale effect (superimposition of residual stresses and load stresses) on small specimens within the scope of the fatigue tests on blast-cleaned butt welds.