10 -6 10 -5 10 -4 10 -3 10 -2 d a / d N 2 4 6 8 10 20 40 DK ef f [MPa Öm] R=-1, M(T) R=-0.5, M(T) R=0.1, C(T) R=0.5, C(T) R=0.7 C(T) R=0.8 C(T) 6013-T6 T(L) base material W=50mm B=4mm 30 Hz laboratory air 10 -6 10 -5 10 -4 10 -3 10 -2 d a / d N 2 4 6 8 10 20 40 DK ef f [MPa Öm] R=0.1, C(T) R=0.7, C(T) 2024-T3 T(L) base material W=50mm B=4mm 30 Hz laboratory air 4.1 eff 8 6.810 K dN da D = - 3.7 eff 7 1.110 K dN da D = - 2 eff 0.55 0.33R 0.12R K K + = + D D ÷ ø ö ç è æ - - - - = D D 2 eff 2.097 0.811R 0.297R 1 1 1 R 1 K K Figure 5: Effective parent material FCP curves fitted by power laws. Keff was calculated by the empirical formulas on top of the figures taken from [18] (2024-T3) and [8] (6013-T6). Rt in the Keff- K-relationships given on top of Figure 5. The fatigue crack growth rate da/dN is then estimated by the simple power laws shown in Figure 5. The predicted FCP curves are very close to the data measured with the welded specimens, Figure 4. This means that our assumption was right and that the apparent improvement of FGP properties at low loads and load ratios displayed in Figure 4 is solely caused by compressive residual stresses present in the crack tip region of the C(T) specimens. FCP IN TRANSVERSE WELD SPECIMENS To get pronounced residual stress effects on da/dN it was essential to keep K and R constant at relatively low values ( K < 15 MPaÖm, R = 0.1) during the tests of the transverse weld (TW) C(T)-specimens. Because of the very low loads at the end of the tests (< 1 % of machine capacity), R and K deviated slightly from the constant values at a/W > 0.75. In Figure 6 the FSW fatigue crack propagation results of two TW specimens were therefore normalized with the respective fatigue crack propagation values of the parent material at the same nominal loading. The da/dN-prediction was calculated as described in the previous section. Also for this specimen geometry there is a favorable agreement between experimental and predicted values. This indicates that acceleration, deceleration and final acceleration of the crack crossing the weld is mainly caused by residual stresses and not by varying resistance to FCP of the different weld microstructures. CONCLUSIONS On the basis of the effective stress intensity factor range Keff-approach and a simple residual stress intensity factor estimation it was shown that the differences in FCP behavior of welded specimens and parent material (2024-T3 and 6013-T6) are almost completely caused by residual stresses and not by the different resistance to FCP of the various weld microstructures. ACKNOWLEDGEMENTS The authors wish to thank Mr. U. Fuchs, and Mr. H.-W. Sauer for their support in welding and in the experimental investigations and Mr. F. Palm from the EADS Corporate Research Center, Munich, for kindly providing some of the FSW joints.
RkJQdWJsaXNoZXIy MjM0NDE=