Comparison between experimental and numerical results of mixed-mode crack propagation in concrete: Influence of boundary conditions choice

In order to experimentally validate concrete damage models and better characterize the concrete behavior during mixed-mode crack propagation, multiaxial tests are developed. The goal is to perform rich and discriminating tests by using state of the art techniques. The loadings are applied using a hexapod controlled by a 3D displacement system and the cracking state is analyzed via digital image correlation. To probe its consistency the experimental results are compared to numerical simulations with a nonlocal (i.e., gradient-based) damage model. The importance of using accurate boundary conditions while performing the numerical simulations is underlined. The numerical simulations are congruous to the experimental results only when full-field boundary conditions are prescribed, and only in that case it is possible to evaluate the validity or more generally the parameter sensitivity of the damage/fracture model under investigation.