Issue 48

R. Baptista et alii, Frattura ed Integrità Strutturale, 48 (2019) 257-268; DOI: 10.3221/IGF-ESIS.48.27 264 Figure 6 : Stress intensity factors vs crack length and obtained differences between ASTM and FEM simulations. C m Diff. [%] Diff. [%] ASTM 1.03E-14 4.06 2D Contour Integral 8.70E-15 4.15 -16% 2% 3D Contour Integral 6.90E-15 4.17 -33% 3% 3D XFEM 7.00E-15 4.15 -32% 2% Table 5 : Paris Law parameters determined for each type of FEM simulation, for the HAZ material. Figure 7 : Experimental fatigue crack growth on the HAZ vs FEM simulations using different models. Fatigue crack growth on a CT specimen containing a longitudinal weld When considering welded specimens, Zong et al. [23] have also used this technique to calculate fatigue crack propagation. Unlike the second case presented in this paper, Zong et al. [23] did not considered the full three-dimensional geometry of the longitudinal weld. When considering the weld geometry, the specimen will be subjected to a mixed mode loading, as K II will no longer be zero. This means that the MTS criterion will predict crack kinking, according to Eqn. (2). As one can see in Fig. 8 a), the algorithm is accurately predicting the crack deflection. As mentioned the 2D and 3D have different geometries. The 2D is just an approximation, and in fact contains more material in the simulated weld. Therefore, one can see the 2D model predicts a higher crack deflection. The 3D models predicted very close results, with the crack deflected by 1 mm over a 40 mm length crack, while the 2D models predicted a 2 mm deflection for the same crack length. The