ICF13A

4 It should be emphasized that in all specimens under consideration tearing occurred by cracking parallel to the rolling direction in the plane inclined at 45° to the specimen surfaces. The regime of slant tearing starts from the very instant i0 of crack nucleation and continued up to the complete separation of a specimen by a growing crack at the instant f (Figs 2 and 3). Note that sudden formation of two short cracks near the points n (Fig. 1) was always preceded by a clearly distinct drop in test records at the attaining the state t0 (see Fig. 3b). As can be seen from Fig. 2, the decrease of PD height 2H0, when accompanied by the decrease of stress raiser length 2c0, ensure realization of the fully-controllable discontinuity of the slant crack extension process [7]. Very orderly dips on the softening branch of the diagram reflect a cyclic variation in the crack profile geometry. Each cycle consists of five test events (Figs 3a and 4). Here. 2v(L) is the displacement of a grips fixture that was measured in a synchronized manner with the crack extension, displacement 2v(m) of the points m (Fig 1) and force P (frequency 5 Hz). The diagrams in Figs (2a, 3 and 4) were obtained on the MR(T) specimens tested without the use of anti-buckling guide plates. Buckling of thin plates is a competitive failure mechanism resulting from the elastic compressive stress acting parallel to the crack growth line. Before tests of the MDR(T) specimens (Table 1) they were lightly clamped between two anti-buckling guide plates. 3. Fully-developed crack profiles Generally, stable crack extension can be seen as interplay of concurrent processes jointly represented by seven through-life fracture curves on the so-called Integrated Fracture Diagram (IFD) [6, 7]. Of these curves, we consider in some detail only a curve describing the geometry of a fullydeveloped crack in a broken-down specimen. In fact this is one branch (lower-right) out of four branches of a centre crack profile (see Fig 5a). Further on, this branch of a crack, having the key importance for our approach, will be designated as n-relationship of the IFD. The input data for determination of n-relationship are distances 2s(x)n between the lower and upper borders of a fully-developed crack (Fig 5a). Of course, at points x = ± cf this distances is taken to be zero. By comparing data presented in Figs 6a and 7a, one can find that crack profiles in the specimens of the small and large width are qualitatively similar. An unexpected result is a significant drop in s(x)n values for the MR(T)-0.5-10.0 specimen containing the shortest stress raiser. This is the single open hole of diameter 2r0 = 2mm. (a) (b) Figure 2. Comparison of test records for the MR(T) specimens of width 2W0 = 120mm and different height 2H0 (a) shown together with test records for the MDR(T)-1.0-1.0 specimens having the similar PD geometry, but different length 2c0 of the original stress raisers with the identical curvature of theirs tips s(b).

RkJQdWJsaXNoZXIy MjM0NDE=