PSI - Issue 37

Patrick Yadegari et al. / Procedia Structural Integrity 37 (2022) 500–507 P. Yadegari et al. / Structural Integrity Procedia 00 (2019) 000 – 000

506

7

3. Fatigue life calculation of component tests To validate these new estimation methods for the application with the local strain approach, the results of fatigue tests of component-like specimens are recalculated. These tests were performed with alternating tensile-compressive stress ( R = -1) as well as with a predefined tension mean stress ( R = 0). The same four high and ultra-high strength steels were analysed and, among others, notched specimens under plane or circumferential bending and two kinds of tube specimens under impulse pressure or axial load were tested. More information on the manufacturing and geometry of the specimens as well as the test conditions can be found in Straub et al. (2021). Since the main purpose of this experimental investigation was to determine the particular endurance limit, only the few test results in the range of low-cycle fatigue with N < 5 • 10 5 will be considered. According to the procedure of the "Guideline non-linear" and thus the local strain approach, the cycles to failure are calculated for the component tests with both damage parameters. For the characterisation of the cyclic material behaviour, the described estimation methods based on the measured ultimate tensile strength are used. The component influencing factors such as surface roughness and stress gradient at failure-relevant points as well as the measured residual stresses on the surface of the component-like specimens, if available, are taken into account for the fatigue life calculation. A comparison to the experimentally determined number of load cycles is presented in Figure 4, whereby the calculated fatigue life N approx is plotted on the ordinate and the corresponding experimental number of cycles N exp is plotted on the abscissa. For data points in the upper left triangle, the calculated fatigue life turns out to be higher, which puts the estimate on the uncertain side. A data point located in the lower right triangle is conservatively estimated because the matching experimental fatigue life is higher than the calculation would predict. For a probability of failure of 50 %, the optimum of this diagram corresponds to a uniform, thin distribution of the data points around the bisector. As can be seen in Figure 4, the data points are evenly distributed around the angle bisector, whereby the scatter band is only narrow. The cycles to failure calculated according to P RAM are in most cases slightly more conservative compared to P RAJ . With the colour sorting of the investigated materials it can be seen that the materials 100Cr6 and X3CrNiMoAl13-8-2 are in general estimated more conservatively over the entire lifetime range. The data points of both conditions of X40 are on the slightly unsafe side, but here no measurements of the expected life-shortening tensile residual stresses could be performed and taken into account in the calculation.

Fig. 4. N - N -Diagram of the calculated component fatigue life compared to the experimental results.