Statistical Characterisation of Chloride Ingress Parameters from Normal and Accelerated Tests

Chloride ingress into concrete is frequently recognized as a main cause leading to important damages in reinforced concrete structures. In natural conditions, chloride penetration into concrete is a very slow process and requires significant time to reach a considerable level of chloride content in RC structures. Hence, this fact leads to the difficulties in long-term lifetime assessment of RC structures under chloride attack. In laboratory conditions, accelerated tests could speed up the chloride ingress rate and give data on mid- and long-term of chloride penetration process. Once the equivalent time in accelerated tests is determined these data could be used for further analysis. This study proposes an approach based on Bayesian Network that allows statistical determination of chloride ingress parameters from normal and accelerated data. These parameters could provide a more reliable prediction of probability of corrosion initiation.