Abstract：Concrete carbonation analysis is of vital importance for the long-time performance assessment of concrete structures. In this work, the theoretical and empirical concrete carbonation models in current use are investigated. By adding model correction terms to the deterministic model and correcting the potential bias in the deterministic model, the probabilistic carbonation model is constructed. The posterior statistics of the model parameters are computed through the Bayesian updating approach based on the field test data of concrete carbonation. The corresponding time-variant probability analysis for the concrete carbonation process is established based on the developed probabilistic model. The influence of the model uncertainty on the concrete carbonation failure probability is taken into consideration. Furthermore, the above process is applied to a simply supported concrete slab to evaluate the carbonation performance of this concrete structure.
