The prediction is quite well in the shear lip developing area. Static effects have not been incorporated in the calculation. 0.001 0.01 0.1 1 10 100 1 10 100 ∆K (MPa√m) da/dN (µm/cycle) da/dN-shear lip corr. Pmax=25 kN Pmax=35 kN Pmax=65 kN Pmax=95 kN start shear lip from here complete slanted Figure 3 Experimental fatigue crack growth rate results and prediction for AA5083 at R=0.1 Underloads in AA 2024 Constant ∆K tests with constant ∆K underloads were performed in AA 2024. The underloads were high enough to lead to development of shear lips, while the ∆K before and after the underloads were to low for it. All Kmax were the same to avoid plasticity induced crack closure at the loading transitions. A retardation in crack growth rate was measured after the underloads. The retardation was larger for higher ∆K of the underloads and for more underload cycles. The shear lips vanished quickly after the underloads, but the retardation of the crack growth rate lasted about three times as long, i.e. if there is an effect of the shear lips on ∆K, it extends outside the area where shear lips are really present. The same calculation is applied as on AA 5083, only now, in this situation of decreasing and vanishing shear lip width, ∆K is not assumed to be influenced by the real shear lip width at the crack tip, but by a mean of the shear lip width extending over some mm crack length after the tip. Calculations taking the actual shear lip width at the tip as a measure for ∆K, and calculations taking the mean shear lip width over 3 mm after the tip were applied and compared with a real test with underloads, see figure 4. The 3 mm was estimated using a potential drop measurement of the affected crack length by crack closure in this case. A possible explanation of the change in slope. The following explanation is rather speculative at the moment. It is found that in AA 5083 the change in slope starts at 0.1 µm/cycle in air and at 0.2 µm/cycle in seawater [2]. The change in slope is independent of Kmax or ∆K. Earlier investigation on AA 2024 in vacuum did result in a change of slope at about 0.004 µm/cycle (for R=0.3). These results indicate that the reason of the change in slope is not due to mechanically causes alone, but that it more
RkJQdWJsaXNoZXIy MjM0NDE=