PSI - Issue 28

6

Author name / Structural Integrity Procedia 00 (2019) 000–000

Kris Hectors et al. / Procedia Structural Integrity 28 (2020) 239–252

244

sections. Finally a fatigue life estimation is obtained. 4. Case Studies 4.1. Two-level block loading comparisons

In order to make a quantitative comparison between the lifetime predictions of the previously introduced damage models, they are first compared to experimental data available in literature. Dattoma et al. (Dattoma et al. 2006) performed 50 tensile-compression fatigue tests with load ratio � � �1 on a long cylindrical test specimen made from 30NiCrMoV12 steel to obtain the material S-N curve. The mechanical properties of the material are the yield strength � � ��� ��� , the ultimate tensile strength � � 10�� ��� and the fatigue limit � � ��1��� . The S-N curve that was obtained for the material conforms to equation 10 when � � � (Dattoma et al. 2006). � � �1�� � � ��� � �� �10� Dattoma et al. also performed two-level block loading tests for 18 different load sequences using six different load levels. After applying a predetermined number of cycles at the first load level, the second load level was started and held until failure occurred. Table 1 shows the details of the high-low and low-high load sequence experiments respectively. Table 2 shows the lifetime predictions of the different damage models in absolute values and also relative to the experimental values. This last information is also visualized in figure 2. Table 3 summarizes the statistical analysis of the lifetime predictions. Several conclusions can be drawn from the results shown in tables 2 and 3. As mentioned in the introduction, lifetime predictions using Miner’s rule have been reported as conservative for high-low loading sequences and non-conservative for low-high loading sequences. The same is observed here, but also for the other considered models as they tend to be conservative for high-low loading sequences and vice versa. Nonetheless, the results show that none of the models consistently produces either conservative or non-conservative predictions even when considering only one type of loading (i.e. high-low or low-high).

Figure 2: Comparison of experimental results given in (Dattoma et al. 2006) and predictions by the studied damage accumulation models