Issue 48

V. Giannella et alii, Frattura ed Integrità Strutturale, 48 (2019) 639-647; DOI: 10.3221/IGF-ESIS.48.61 642 For Test I (Fig. 3), the crack propagation takes place in a direction orthogonal to the static load. On the contrary, the absence of a static loading would lead to a crack propagation direction orthogonal to the cyclic loading. This means that a transition of the crack propagation direction from orthogonal to the cyclic loading to orthogonal to the static loading must take place for a static load in between 0 kN and 24 kN (always considering a cyclic load of ±8 kN). Test II (Fig. 4) was scheduled for better understanding such a transition behavior. In order to investigate the range of low static loads, the static and cyclic load were initially set up in Test II (Fig. 4) to 3 kN and ±24 kN, respectively. The crack propagated mostly in the initial notch plane for this load step. Increasing the static load to 12 kN in the next load step, crack directions did not change significantly its initial pattern (Fig. 4b). Figure 3 : Experimental crack paths for Test I. (Courtesy of Christian Kontermann, TU Darmstadt, Germany) (a) (b) Figure 4 : Experimental crack paths for Test II: (a) load step 1; (b) load step 2. (Courtesy of Christian Kontermann, TU Darmstadt, Germany) N UMERICAL CALCULATIONS umerical simulations were performed with the BEASY DBEM code [19] and with the FRANC3D FEM code [2]. Cruciform specimens here considered have a non-uniform thickness and therefore the models presented in the followings are three-dimensional. N