Issue 37

G. Beretta et alii, Frattura ed Integrità Strutturale, 37 (2016) 228-233; DOI: 10.3221/IGF-ESIS.37.30 231 length that oscillates between 1 and 10 grain diameters, depending on the material, the stress gradient ahead of the notch, etc. This procedure must be repeated for all the directions given by ϴ and ϴ 1 and the minimum value of all the obtained λ( ϴ , ϴ 1 ), will be the predicted notched fatigue limit, λ 0 N :          0 1 min , N (2) The values of ϴ , ϴ 1 for λ 0 N will provide the predicted crack initiation point at the notch and the predicted crack direction along the first grains (Stage I). For a full explanation of the microstructural model applied to a circular notch under proportional biaxial loading, please see the reference [13]. E XPERIMENTAL RESULTS AND PREDICTIONS he experimental fatigue limits for the AISI 304L stainless steel and the predictions with the microstructural model are shown in Fig. 3. There are results for three hole diameters, d=1 mm, d=2 mm and d=3 mm. Besides, the experimental data for smooth specimens, already published [14], and its theoretical curve (ellipse quadrant) is plot. It is seen that the predictions for d=2 mm agree quite well with the experimental data, while for d=1 mm and d=3 mm the agreement is not so good. More experimental data is required to get a reliable evaluation of the model. Figure 3 : Comparison of experimental and predicted fatigue limits for the AISI 304L stainless steel. Crack initiation point and crack direction during the Stage I Figs. 4, 5 and 6 show examples of cracks emanating from the hole for tension-compression, pure torsion and in-phase biaxial loading (σ y ∞ = τ ∞ ), respectively, for applied stresses close to the fatigue limit. The crack initiation point was, in the majority of the tests, located close to the point of maximum principal stress, which according to Fig. 2, means a value of ϴ =0º for the axial tests, ϴ =45º for the torsion tests and ϴ =31.7º for the biaxial tests (σ y ∞ = τ ∞ ). The crack direction along the first 400 µm (5 average grains long), considered as the Stage I of the crack growth, was also studied. The experimental directions were close to the maximum principal stress direction, which means values of ϴ 1 =0º for axial tests, ϴ 1 =45º for torsional tests and ϴ 1 =31.7º for biaxial tests (σ y ∞ = τ ∞ ). With respect to the model, an approximately similar minimum value of λ is obtained for several combinations of ϴ , ϴ 1 . It means that the predicted notched fatigue limit, λ 0 N , is associated with several initiation points and crack directions. Certainly, this is not fully correct, as the experimental values of ϴ , ϴ 1 are quite precise and with low scatter. Further work needs to be done in the model in order to discriminate directions and correctly predict the crack initiation point and the Stage I crack direction. T