PSI - Issue 28

Yuri Petrov et al. / Procedia Structural Integrity 28 (2020) 1975–1980 Yuri Petrov/ Structural Integrity Procedia 00 (2019) 000–000 5 various loading pulse amplitudes. Also, experimental data from Ravi-Chandar and Knauss (1984a), is shown in figure 2a. As can be seen from this figure, considerable scattering of the ! values is present for larger crack velocities and thus the experimentally observed scattering was obtained numerically using the developed scheme. The 9 µ s incubation time value provides larger averaged ! values comparing to the experiment. This may be explained by the fact that the crack branching cannot be accounted for since the developed numerical scheme allows only straight crack paths, while branching is experimentally observed for the higher crack velocities. Branching can potentially lead to higher energy dissipation due to increased fracture surface area and thus result in lower energy flux into the vicinity of the main crack. On the other hand, application of lower incubation time value (1 µ s) results in a better fit with the experimental data still describing the SIF scattering. 1979

Fig. 2. SIF – crack velocity relations. Experiment (a), simulations (b,c) and load (d)

5. Conclusions

To simulate dynamic crack propagation an approach based on the concept of incubation time was used. This approach implies spatial and time discretization of the fracture process and thus unstable behavior of the fracture defining values (for example, considerable scattering of the SIF for a moving crack) is regarded as a natural feature of the crack propagation process. Conducted numerical simulations of the crack propagation experiments revealed scattering of the ! values which qualitatively describes the experimentally observed phenomenon. The developed approach does not depend on any relations and conditions containing SIF and thus lets one evade known ambiguity of the corresponding SIF-based fracture models (Ravi-Chandar and Knauss (1984a,c), Dally et al. (1985)).