Evaluation of variational phase-field models for dynamic brittle fracture

During the last decades various phase-field model have been developed in the variational approach to model dynamic fracture of brittle solids. Complex fracture phenomena such as crack branching, dynamic crack instability and fragmentation, etc., can be captured with a minimal list of physical parameters. However, an objective evaluation of existing results is missing and in particular, the issues of mesh size dependence and length scale sensitivity have been rarely studied. Aiming to bridge the above gap, this paper presents an extensive evaluation of commonly used variational phase-field models for dynamic fracture to address the question: how well can we predict dynamic fracture with current PFMs. Several popular PFMs in the literature are considered, implemented within a single code platform and used to simulate six dynamic fracture problems of homogeneous brittle solids. Mesh convergence and length scale sensitivity are systematically analyzed. Quantitative evaluation of the branching angle, crack tip position and/or velocity with respect to available experiments, and comparison against the results of other approaches, are both presented.