Statistical modelling of the influential factors on chloride penetration in concrete

It is widely acknowledged that reinforcing steel corrosion is the most common deterioration mechanism in reinforced concrete structures and that chloride penetration is one of the processes that induce such a mechanism. Although chloride ingress in concrete has been thoroughly investigated, it is still difficult to foresee, in a quantitative manner, the performance of a given concrete mix regarding chloride penetration over time. This work intends to provide simple analytical models to the prediction of concrete's performance with respect to chloride penetration, conveying it in the form of a diffusion coefficient. For that purpose, a comprehensive sample of 493 case studies were collected from 43 publications. The statistical modelling is performed using a multiple linear regression. This approach allowed identifying explanatory variables to this phenomenon and its transposition to mathematical models. This paper presents three analytical models to characterize the resistance of concrete to chloride penetration. A general model to estimate chloride diffusion coefficients is proposed, as well as two specific models, one for natural exposure conditions and the other for migration tests. The physical meaning of the variables considered in the models was discussed, allowing a comprehensible description of chloride penetration in concrete.