Numerical results are compared to experimental data on concretes exposed in field conditions for 19 years and subjected to wetting/drying cycles in seawater after a couple of curing days. They were obtained thanks to a hydration model assuming fully hydrated concretes (19 years of curing time) and a reactive transport model in saturated conditions taking into account precipitation/dissolution of minerals and their kinetics and adsorption of ionic species on C-S-H. Numerical results show surprisingly rather good results, especially for concretes with fly ash. For OPC, although model considers average bulk porosity modifications, the experimental apparent diffusion coefficient increases much more. For concretes with silica fume, numerical results also show underestimations of total chloride content also experimental apparent diffusion coefficients are constant. Analysis of hydration calculations show that concretes are not fully hydrated after 19 years. Additional chloride content from reactions between anhydrous phases and chloride ingress may appear.