ICF13B

13th International Conference on Fracture June 16–21, 2013, Beijing, China -7- Simulations 0 76,179,191 t s  0+40 t t s  0 70 t t s   0 100 t t s   (b), synthesized fringe patterns of numerical sample 3.3 (c), crack growth process of numerical sample 3.3 Figure 5. One main crack path Experiments 30 t s  50 t s  70 t s  (a), fringe patterns of experimental sample 4.4 Simulations 0 76,179,191 t s  0+30 t t s   0 50 t t s   0 70 t t s   (b), synthesized fringe patterns of numerical sample 3.6 (c), crack growth process of numerical sample 3.6 Figure 6. Bifurcation (The colour of all images in Fig. 6.a and b has been reversed, since the records are too dark to see.) Before crack starts, the experimental fringe patterns and simulation results show significant similarity. However, when crack develops, the stress distribution of the simulation becomes blurred. There are three possible reasons for this: (1) The mesh density is not large enough near the crack tips. In the simulation, the element size is still large, so the energy of the stress wave released from broken elements is significant. On the contrary, the granularity of the epoxy resin is small, and the energy released from the broken of crystalline grain is small and continuous in real experiments.

RkJQdWJsaXNoZXIy MjM0NDE=