PSI - Issue 43

Kamila Kozáková et al. / Procedia Structural Integrity 43 (2023) 178–183 Kamila Koza´kova´ / Structural Integrity Procedia 00 (2023) 000–000

182

5

0.5

Experiment

Predictions

r = 0.2 mm

0.4

0.1 length parameter, l [mm] 0.2 0.3

0

10 6

10 7

10 8

10 9

10 10

number of cycles to fracture, N f [-]

Fig. 6. The length parameter

Fig. 7. Lifetime predictions and experimental data

3 mm di ff er from the experimental curve, especially in areas beyond 10 7 cycles. In this area, the lifetime predictions of the notched specimens are not satisfactory. Real experiments showed earlier fatigue failure than expected. For this reason, notch surfaces were observed under a scanning electron microscope. It turned out that notches with notch radii r = 0 . 8 mm and r = 3 mm were made in a di ff erent way than notches with radii r = 0 . 1 mm and r = 0 . 2 mm. Notch surfaces of notches with radii of r = 0 . 1 mm and r = 0 . 2 mm were smooth (see Fig. 8a, b). These notches were machined with knives that had the same radii as the notches. Therefore, the predictions of notched specimens with radius r = 0 . 1 mm fit almost perfectly with the experimental data. On the contrary, notches with radii r = 0 . 8 mm and r = 3 mm were made with a knife with a smaller radius than the radii of the produced notches (see Fig. 8c-f). These figures show the steps created during the production of these large notches with a small knife. These steps and dimples cause stress concentrations. The axial stress concentration at the root of the notch led to a shortening of the lifetime of these specimens. Thus the process of machining of the notches can negatively influence the experimentally obtained fatigue data in comparison to the predictions. The mismatch is apparent, especially in the area of gigacycle fatigue, where small details are crucial. Similar study of the surface treatment was made by Jambor in [9], where shot peening was used to enhance the fatigue life of the notched specimens made of the same alloy. As described above, many theories using average stress ahead of a notch tip have been developed for fatigue predictions of the e ff ects of notches and other concentrators. These theories are based on knowledge of average stress over a distance ahead of a stress concentrator. In this paper new method for fatigue lifetime predictions of notched specimens was presented. Contrary to previous theories, this method takes into account the dependence of averaging distance l on number of cycles to fracture. The method uses experimental data of smooth and notched specimens and corresponding axial stress distribution for the determination of length parameter l . The length parameter can be subsequently used for lifetime predictions of notched specimens. It was proved from the calculation that length parameter l decreases with the increasing number of cycles. For small notches (with smooth surface), the predictions are very similar to experimental data. In the case of bigger notches, machined with a small knife tool, the way of notch production negatively a ff ected the fatigue lifetime of the specimens in comparison to the predicted lifetime based on theoretical stress distributions. When designing notched components, it is necessary to pay attention to the way of notch machining, especially in the area of gigacycle fatigue. 4. Conclusions