Algorithmic implementation of a generalized cohesive crack model

Abstract

Smeared fictitious crack models can be regarded as generalized cohesive crack models. The classic fictitious crack models, i.e. the fixed crack, multiple fixed crack, rotating crack and microplane model, are based on different assumptions for the orientation of developing cracks. A smooth transition between the extreme cases, the fixed crack and the rotating crack model, is provided by the adaptive fixed crack model. In this approach, the critical direction of failure is uniquely identified based on Mohr's hypothesis. Thus, the critical direction depends on the character of the failure criterion and the type of loading. The numeric implementation of the adaptive fixed crack model has given rise to some subtle questions. It is shown that even for a classical fixed crack concept, the algorithmic tangent stiffness may have to include components of crack rotation, depending on the imposed strategy for the global equilibrium iteration scheme.