International audienceThe objective of this study is to investigate damage-temperature-stress level-permeability interactions in structural concrete. The tests are performed on hollow cylindrical concrete specimens, subjected to compressive loading and temperature up to 150 °C. The results emphasize that at stress levels lower than 80% of the peak stress, the variation of permeability is small and it is slightly influenced by the stress. As a matter of fact, the permeability under load is smaller than the permeability measured after unloading. As the load exceeds 80% of the peak stress, micro-cracking increases rapidly, causing an increase of the permeability and a greater sensitivity to the applied load, i.e. a noticeable difference between the permeability measured under load and after unloading, the first becoming greater than the latter. In the post-peak phase the increase of permeability is much larger due to significant crack width growth. The increase of permeability with the applied load seems to be greater with temperature, inducing further alterations of concrete and dilation of the porous structure of the material. Finally, the experimental results seem to agree with the format of coupled evolution of the permeability due to damage and temperature assumed by Gawin et al. [D. Gawin, C.E. Majorana, B.A. Schrefler, Numerical analysis of hygro-thermal behaviour and damage of concrete at high temperature, Mechanics of Cohesive-Frictional Materials 4 (1999) 37–74.]

Extraction of a crack opening from a continuous approach using regularized damage models. ABSTRACTCrack opening governs many transfer properties that play a pivotal role in durability analyses. Instead of trying to combine continuum and discrete models in computational analyses, it would be attractive to derive from the continuum approach an estimate of crack opening, without considering the explicit description of a discontinuous displacement field in the computational model. This is the prime objective of this contribution.The derivation is based on the comparison between two continuous variables: the distribution if the effective non local strain that controls damage and an analytical distribution of the effective non local variable that derives from a strong discontinuity analysis. Close to complete failure, these distributions should be very close to each other. Their comparison provides two quantities: the displacement jump across the crack and the distance between the two profiles. This distance is an error indicator defining how close the damage distribution is from that corresponding to a crack surrounded by a fracture process zone. It may subsequently serve in continuous/discrete models in order to define the threshold below which the continuum approach is close enough to the discrete one in order to switch descriptions. The estimation of the crack opening is illustrated on a onedimensional example and the error between the profiles issued from discontinuous and FE analyses is found to be of a few percents close to complete failure.[ ] U

No abstract