Impact du chargement de traction et de la fissuration sur l'évaluation du coefficient de diffusion des chlorures dans des tirants en béton armé

Chlorides in de-icing salts or sea salts can penetrate concrete and reach the reinforcement. It can lead to a big issue in reinforced concrete since chlorides cause corrosion of the steel reinforcement, leading to severe structural deterioration. To quantify the resistance of concrete to chloride penetration, chloride diffusivity in concrete can be measured using diffusion or migration tests. However, in the literature, these tests are carried out on sound concrete, i.e. concrete without reinforcement and cracks. However, real concrete structures generally have reinforcement, are cracked and undergo mechanical loading. The chloride diffusivity measured in the laboratory is therefore different from that measured on concrete as it might be found in a conventional structure.The first objective of this research project was to develop a device to measure chloride diffusivity in a reinforced concrete tie-specimen, representing a bridge reinforcement and its coating, held under mechanical loading and cracked. To achieve this, numerical and experimental optimisations were used to adapt the classical migration test found in the literature to include steel reinforcement and a greater specimen thickness. A device has also been developed to maintain the tensile load. This showed that the presence of a crack increases diffusivity and that the more open the crack, the higher the diffusivity. Similarly, loading causes an increase in diffusivity, even if it does not lead to cracking.The second objective of the research project was to determine the effect of different crack parameters on diffusivity numerically. Diffusion transport has been added to an existing lattice-particle model, providing a good cracking representation. Simulations were then carried out by progressively adding tortuosity and constrictivity to the crack, as well as reinforcement in the concrete. Finally, the use of the numerical model improved the results obtained experimentally by providing a better description of the crack inside the specimen.