PSI - Issue 42

Matěj Mžourek et al. / Procedia Structural Integrity 42 (2022) 457 – 464 Matěj Mžourek / Structural Integrity Procedia 00 (2019) 000 – 000

462

6

Fig. 5. Fatigue limit as a function of HV, based on Garwood et al. (1951).

3.5. Residual Stresses Axial and tangential residual stresses were measured on 2-4 specimens representing each of the unnotched specimen configurations after the experiment. Relaxation due to mechanical loading is expectable. Measurements were thus taken in the position of the minimal diameter, as well as 8 mm away from the beginning of the transitional radius from the minimal diameter, where the virgin state of the material is assumed. It has been revealed that large compressive residual stresses are present on the specimens (up to -505 MPa), and their reduction due to loading has been estimated to be between 20 % and 46 % if the values in the critical section and out of it were compared. Residual stresses superimpose a mean component on the loading cycle, which is expected to have a large impact on estimated fatigue properties. The detected stresses however seem to be quite uniform across individual tested specimen geometries, and any correction of the S-N curves based on these stresses is expected to be identical. No action to account for the residual stresses has thus been taken at this point. 4. Evaluation of experimental results and discussion 4.1. Size effect Obtained critical volumes via the V-model for a range of values of n are plotted in Fig. 6. a). Values for the notched specimens are orders of magnitude lower than for the unnotched specimens due to the stress gradient at the notch tip. The V ( n ) plots of specimens A02 and A04 feature a steep drop at values corresponding to the inverse of the stress intensity factor n = 1/ K t of the respective specimens. This is due to the elongated central part featured on specimen types A02 and A04, see Fig. 2. When n > 1/ K t , volume satisfying the stress condition Eq. (2) is concentrated near the transitional radius of these smooth specimens. When n < 1/ K t , the stress threshold is lower than the nominal stress, and the whole volume of the elongated central section of specimens A02 and A04 satisfies condition (2). Even small variations of n can thus affect the obtained values of the critical volume significantly - up to the point that individual curves in Fig. 6. a) change their order in the vertical direction depending on the value of n . The correction by which the model given by Eq. (1) shifts an S-N curve vertically is different for various values of n , including even its direction in specific cases. The A-model mimics this behavior. The influence of the h crit parameter on the obtained volumes is observable in Fig. 6. b), where the stress threshold parameter n = 0.971 (this value was found to be optimal via regression analysis, discussed lower) is kept constant. The depth parameter h crit does not influence the obtained critical volumes of the notched specimens, as with such a high value of n , volume satisfying stress condition (2) is localized closer to the surface of the specimen than is the h crit depth. The h crit parameter was found to have little effect on obtained results if only results from the four smooth A01-4 specimens were available ( Mžourek et al. (2023)). While the parameter does affect the obtained volumes, when n is