Bayesian Belief Network to Assess Carbonation-Induced Corrosion in Reinforced Concrete

Abstract

Corrosion of reinforcing steel is a major cause of deterioration of reinforced concrete structures. Two primary causes leading to corrosion of reinforcing steel are, respectively, carbonation and chloride contamination of the concrete cover. Estimating the time to corrosion initiation through modeling enables a qualitative and quantitative assessment of the influence of several variables on the mechanism involved. However, many of these variables are subject to stochastic uncertainty, which can be represented through either frequentist or Bayesian probability. This paper proposes a probabilistic-based model for carbonation-induced corrosion based on a Bayesian belief network. The proposed model is compared with a more traditional probabilistic-based technique, Monte Carlo simulation. Both methodologies are illustrated through two case studies and similarities and differences are highlighted.