PSI - Issue 13

1624 Yuri Petrov et al. / Procedia Structural Integrity 13 (2018) 1620–1625 Yuri Petrov/ Structural Integrity Procedia 00 (2018) 000–000 5 loading resulted in a considerably larger scatter of the values and no curve could be built. The obtained results qualitatively correspond to experimental results from Ravi-Chandar and Knauss (1984c).

Fig. 3. Numerically obtained − data for a hypothetical material for quasistatic loading(a) and pulse loading (b). This way, whole variety of experimentally observed cases of the − dependence can be obtained using single approach based on the incubation time concept: specimen shape dependent curves, nonexistence of the curve for the pulse loading and a common − curve for the quasistatic loading. Scattering of the SIF values is considered as a feature related to a discrete ("quantized") character (Petrov, 1996) of the crack propagation process. The size of the scatter determines if it is possible to construct a classic − curve. 4. Conclusion Instabilities related to the dynamic crack propagation process are studied in this paper. Stress intensity factor – crack velocity dependence ( − relation) is discussed. This relation is regarded as a material property within a classic framework of dynamic fracture. However, uniqueness and existence of this dependence was doubted in a number of experimental works: dependence of the − curve on specimen shape and way of loading was observed in works by Kobayashi and Dally (1977), Kalthoff (1983), Ravi-Chandar (1984c). In particular, classic − curves were observed in quasistatic experiments (see e.g. Dally (1979)), while no − relation was observed for the case of pulse loading. Various types of the − dependence were obtained numerically using the incubation time fracture criterion. Scattering of the stress intensity values is regarded as a natural feature of the crack propagation, related to discreteness of this process. It was found that pulse loading of samples results in a large scatter of the values, while scatter for