PSI - Issue 2_A

Timothy Crump et al. / Procedia Structural Integrity 2 (2016) 381–388 Crump / Structural Integrity Procedia 409 (2016) 000–0 0

387

(a)

(b)

(c)

Fig. 8. Resulting warped stress plots for: (a) Gc =3Jm -2 ; (b) Gc = 0.3Jm -2 ; (c) Gc = 0.03Jm -2 .

4. Conclusions A modelling approach to dynamic crack propagation has been presented here and applied to a well-established experiment using Homolite-100. The approach, referred to as XCZM, combines the mesh independency of the eXtended Finite Element Method, allowing for the crack to cross elements, with a phenomenological rate-dependant cohesive law to represent the fracture process at the meso-scale. A 2D DCB geometry was first considered under a static loading to enable comparison with an available analytical steady state solution (2D) and observed experimental speed. The speed obtained via XCZM was found to be in good agreement with the observed analytical and experimental speed of ~0.3Cr. The loading was then changed to consider prediction of the point of crack branching. This was also found to be in accordance with the experiment at the maximum crack speed of approximately 0.72Cr. The DCB geometry was then extended into 3D to observe the bifurcation angle due to rate-dependant loading. This was also found to be in good agreement with the observed half-branching angle of 4 o . The results demonstrate that XCZM in 3D is able to produce energetically stable macro-crack propagation and bifurcation, lending support for this meso-scale approach to be used in global structural modelling assessments in other materials due to the use of a geometrically independent, rate-dependant phenomenological law. Further work will be to extend XCZM to propagate two cracks within the same model, to be able to observe the full macro-crack branch.