Issue 48

C. Santus, Frattura ed Integrità Strutturale, 48 (2019) 442-450; DOI: 10.3221/IGF-ESIS.48.42 447 where th K  is the threshold stress intensity factor range and fl   the (plain specimen) fatigue limit range. When the FS parameter is considered at the fretting hot-spot, as discussed in the Introduction, there are two directions of maximum shear stress amplitude, at an angular distance of 90°, shown in Fig. 4. The direction of maximum FS parameter is close to the one that, between these two perpendicular planes, undergoes the highest maximum normal stress. Despite the usual evidence of inward crack orientation, the outward direction is more aligned with the tensile stress induced by the bulk load, while the inward direction more significantly experiences the normal compressive stress produced by the contact pad. For this reason, the stress analysis at the hot-spot can lead to a misleading crack direction prediction. Figure 4 : (a) Critical plane analysis at the fretting hot-spot, the two perpendicular maximum shear stress amplitude directions can drive the FS prediction either Inward (b) or Outward (c) depending on the maximum normal stress. Following the Point Method (PM) of the TCD, the stresses can be evaluated at the location L /2 along any notch bisector. For the present fretting geometry, a vertical line starting from the hot-spot can be considered instead of the bisector, Fig. 5 (a). The results obtained with this assumption are not graphically reported for the sake of brevity. Although the stresses are lower than those obtained at the hot-spot, between the two normal directions with maximum shear stress amplitude, the outward direction still experiences the highest FS parameter, thus again the crack orientation prediction turns out to be inaccurate. Better results are obtained if the L /2 stress point is considered, not just at a fixed location below the hot- spot, but following the (potential) crack plane, shown in Fig. 5 (b), which better identifies the stresses along the crack line before its formation. When this stress evaluation point is placed below the contact pad, rather than along any other outward direction, the normal stress is more compressive. However, the shear stress amplitude at two perpendicular directions is no longer equal, whereas it is higher along the inward direction. For this reason, this latter modelling approach is more promising, and thus followed in the test analysis reported hereafter. Figure 5 : Multiaxial applications of the Point Method to the fretting problem: (a) the considered point is at L /2 and perpendicularly below the surface, (b) the point is still at L /2 but following the (possible) critical plane orientation. Short non-propagating cracks were observed at the runout specimens of the first series (as is), with an initial very shallow inclination. If the maximum FS parameter is searched at the fretting hot-spot, as discussed above, the predicted crack orientation is outward. On the other hand, the negative inclination agrees with a high shear stress amplitude, however, the normal stress is less tensile in this angular region. Finally, the SWT prediction is very close to the perpendicular direction, and thus again not in good agreement with the evidence, (see Fig. 6 (b)). After an initial propagation of a few microns, the orientation is more perpendicular to the fretting surface, and the crack angle can be assumed as approx. α = 30°. According to the PM analysis with no fixed position, introduced in Fig. 5 (b), this angle is in agreement with the