PSI - Issue 75

Ralf Glienke et al. / Procedia Structural Integrity 75 (2025) 474–488 Ralf Glienke et al / Structural Integrity Procedia 00 (2019) 000 – 000

480

7

The coverage was determined acc. to SAE J 2277 (2023) by indirect method using a fluorescent marker and a UV light (see Fig. 3 d)). Additionally, coverage was validated by direct method using a 30x magnifying glass, (see Fig. 3e)). All blast-cleaned specimens had a coverage of D = 100 % and a surface preparation grade of Sa 3 acc. to ISO 8501-1 (2007) was determined directly after blasting. Roughness measurements acc. ISO 21920-2 (2022) in conjunction with ISO 8503-4 (2012) were carried out at multiple parallel lines using the 3D optical profiler Keyence VR-6000 and statistical evaluated. In blast-cleaned condition, the surface roughness was determined to be R a = 17.24 ± 1.44 μm and R z = 114.34 ± 8.43 μm (corresponds R y5̅̅̅̅ ). A residual stress state depth profile was determined using the hole drilling method and optical measurement of the surface distortion via electric spackle pattern interferometry acc. ASTM E 837 (2020) and Steinzig et al. (2003a), Ponslet et al. (2003a), Ponslet et al. (2003b), Steinzig et al. (2003b) using the stress analyser system Stresstech PRISM . Several measurements were carried out on blast cleaned surface. Residual stress mean values were calculated for each measurement depth from the individual measurements. The measured two-dimensional residual stress depth profiles, split in tensor components are shown in Fig. 4 a). Compressive residual stresses in the near-surface region down to an orthogonal depth of d = 0.30 mm were verified. The near-surface work hardening effect was examined by Vickers hardness testing acc. ISO 6507-1 (2024) with different test forces, see Fig. 4 b). The surface was slightly ground so that measurements could be made on the plateaus in between the dents from blasting. To ensure that the effects of grinding on residual stresses were minimized, a very light and controlled grinding process was used, employing fine grained abrasive and minimal pressure to achieve plateau areas. In addition, it was ensured that the grinding process has no thermal effects by minimising overheating through water cooling. The following results were determined: 206 ± 8.6 HV 10, 242 ± 9.9 HV 1, 266 ± 9.1 HV 0.3, suggesting a strong work hardening effect close to the specimen surface, judging from the decreasing depth of the indentations with lower test forces. Plastic deformation of the microstructure in the near-surface region was verified by metallographic examination, see Fig. 4 c). The preparation was done based on ISO 17639 (2022) and the section was etched with 3 % Nital (ISO/TR 16060 (2014)). The medium thickness of the near-surface deformation layer was determined as the mean value of separate measurements at different locations of the surface resulting in t = 49 ± 7.9 μm.

Fig. 4. a) Residual stress depth profile of blast-cleaned specimen, b) surface hardness testing and indentations with different test forces on grounded specimens, c) thickness measurement of deformation zone in microsection

3.4. Fatigue tests The fatigue tests were carried out on the test rigs at the University of Applied Sciences in Wismar and the Fraunhofer IGP in Rostock. The specimens were tested in the force-controlled mode of the testing machines until fracture at constant amplitude loading. To derive a detail category, the specimens were tested in the finite life region of the S-N curve according to the load level method downwards from the yield strength. For the series #01 to #06 of the butt welds, a uniform distribution of geometrical characteristics on the load levels was considered. This ensure that the fatigue test results for the butt welds sufficiently represent the quality level B (ISO 5817) for the imperfect shape and dimensions. For the linear misalignment e, max. 0.1 t or max. 2 mm are permissible. Fig. 5 shows the typical crack initiation locations for the different test specimens. The butt welds always failed at the weld transition, so that the crack was able to propagate in depth of the specimen. Crack propagation was mostly in the form of semi-elliptical crack, but also extended surface cracks were observed. Three typical failure modes were observed on the free edges, while in Fig. 5, the failure of specimens from series #09 was observed starting from the corner between the mill scale and the cut edge. In the case of the laser-cut edge (series #07), the crack was observed