A prediction approach of concrete carbonation based on the inverse Gaussian process and Bayesian method

Abstract

Concrete carbonation is one of the major factors causing the deterioration of reinforced concrete structures, and accurately predicting carbonation depth is of great significance for the safety assessment of the structure. This study aims to develop a prediction method of carbonation behavior by incorporating multi-source information using the Bayesian method. First, in the proposed approach, the inverse Gaussian process is used to model the evolution process of carbonation depth, which can capture the temporal variability and the monotonicity of the deterioration phenomenon very well. Then, a proper prior for model is determined by absorbing the knowledge of the existing empirical carbonation model. To fuse the accelerated data and field inspection data, the Bayesian inference is performed to update the posterior distributions of model parameters by Gibbs sampling technique. Finally, a practical case is performed to illustrate the validity and accuracy of our proposed approach.