Limitations of Smeared Crack Models for Dynamic Analysis of Concrete

Performance prediction of concrete structures under explosive loadings or impact is an essential part of the research that is being performed within TNO. One of the current research topics is the explosive safety of tunnel structures. In the context of this research we evaluate the capabilities and limitations of concrete material models in LS-DYNA. The evaluation focuses on the CSCM concrete model and in particular the damage and failure characteristics of the model under single and sequential compression and tensile loading. Like many existing concrete models, the CSCM uses a smeared crack approach to model the reduction in strength of damaged concrete. It will be shown that the smeared crack approach has an intrinsic limit that places a restriction on the minimum size of an element. Furthermore, it is predicted that the built-in fracture energy regularization further aggravates the situation. The regularization algorithm tries to maintain a constant fracture energy. When elements have a size that is smaller than the limit size, the fracture energy of the total structure is increased which causes non-physical behavior. The predictions are confirmed by analyses on a tunnel structure as well as analyses on concrete cylinders under tension and compression. In contrast to the established minimum width, high dynamic loads or very local loads such as explosions or impact require a very fine mesh that can accurately describe the stress state and the shockwaves that are induced during these events. Using a reference load of a BLEVE explosion, the desired element size is derived and it will be shown that the desired element size is far smaller than the lower limit of element size. The consequences of the conflicting restrictions on the element size by the material model and the dynamic loading are illustrated by the tunnel structure analysis.