8 the in-plane constraint (Figs 6b and 7b). These changes are accompanied by the wide variations of the buckling behaviour and crack extension rate [6, 7]. An extremely high level of the global inplane constraint corresponds to the conditions of transverse plane strain, when in the course of crack extension the outer specimen boundaries x = ± W0 are straight and fixed (u(N)f = 0 in Fig 5b). A sufficient level of the proximity to this conditions is achieved in testing the specimens of the basic geometry (H0 = 0.1W0). In this case, the SST resistance for the Al-alloy D16AT is presented by the angle ψn = (1.38 ± 0.1) degrees. This reference value is about half of the lower-limiting value ψc = (3.0 ± 1.5) degrees determined for the material in question in [3] with the use of ASTM/ ISO Standard test method [9, 10]. In comparison with the critical angle ψc, the novel characteristic of ductile tearing ψn is much more consistent and reproducible quantity. Figures 6b and 7b show that the SST resistance, as measured by the angle ψn, contrary to fracture initiation resistance ψni, is very sensitive to changing the specimen geometry and size. In the whole, our results [1-7] demonstrate that the СТОА-ψ depends on the combined effects of many variables, namely, PD geometry and its size, geometry and type of an original stress raiser, boundary restraints, buckling behaviour, crack extension rate, as well as technique used for evaluating this fracture parameter. In particular, one can see (Fig. 9) that the angle β associated with the different pairs of distinct fracture events strongly depends upon the choosing a neighboring pair. For the MR(T)- 0.1-1.0 specimen of the basic geometry this angle closely correlates with the CTOA- ψ. Test data for MR(T) specimens give a simple, inexpensive and yet accurate estimations of fracture initiation stress σNi for shallow cracks originating from a typical stress raiser. The large distinctions between the lower-limiting values of this stress for the PD of different size observed in Fig. 10 call into question the very meaning of the commonly used characterisation of the net-section stresses in terms of the flow stress σf = 0.5(σ02 + σUTS). The characteristic values σNt, σNi and σNs of the stress σN have the advantage of relating directly to the actual state and position of advancing crack tips. Besides, they relate closely to such customarily quantities as the yield strength σ02 and ultimate tensile strength σUTS obtained under uniform straining of the standard smooth specimens. (a) (b) Figure 9. A fragment of the test record (a) displayed earlier in Fig. 4 and schematic presentation of a transition from the one-quarter crack profile s(x)t1 to the next profile s(x)t2 for the spontaneously arrested crack that are shown together with two through-life fracture curves s(x)s and s(x)n related to states s and n (b). 5. General remarks Our experimental results demonstrate that there is not worth striving for longer use of the M(T) specimen (Fig. 1a) at least in standard crack extension tests. This geometry reflects
RkJQdWJsaXNoZXIy MjM0NDE=