PSI - Issue 57

Mathias Euler et al. / Procedia Structural Integrity 57 (2024) 298–306 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

302

5

Fig. 3. Tests on continuous rail welds with travelling wheel load: (a) test setup (Photo: MPA Stuttgart), (b) detail of tested specimen

Fig. 4. Component fatigue resistance curve by regression line with fixed slope parameter m = 3: (a) travelling wheel load, (b) stationarily pulsating wheel load – Remark 1: In case of stationarily pulsating wheel load, F stands for the maximum load (ratio of minimum and maximum load: 0.05). Remark 2: I n case of travelling wheel load, one travel over girder’s midpoint is one cycle . Remark 3: Girders of different span L were tested.

The component fatigue resistance curves of the test girders for detail #1 are shown in Figure 4. A similar diagram is obtained for detail #2. The achieved number of stress cycles refer to the first visible crack through the weld that can be detected by non-destructive testing. It can be recognised that the numbers of stress cycles to failure, that are achieved in the tests with travelling wheel load, are about one decade lower than those of the tests with stationarily pulsating wheel load with respect to comparable wheel load levels. The non-proportionally multiaxial fatigue loading has been identified as the reason of this observation. The fractographic investigations revealed saw-tooth shaped surfaces of the fatigue cracks. Based on this observation and the findings from Senior & Gurney (1963) and Patrikeev (1983), a fatigue damage hypothesis is developed in (Euler, 2017) assuming that the crack initiation and crack growth take place at preferred planes in the welds of the investigated constructional details: