Stochastic Analysis of the Multi-dimensional Effect of Chloride Ingress into Reinforced Concrete

For many reinforced concrete structures corrosion of the reinforcement is an important problem since it can result in expensive maintenance and repair actions. One mode of corrosion initiation occurs when the chloride content around the reinforcement bars exceeds a critical threshold value, which is the subject of the present paper. Typically, the chloride concentration is obtained by solving Fick's diffusion partial differential equation assuming chloride flow into an infinite half space. However, when the concrete structure is relatively thin or when a rebar is situated at a corner of the structure, the assumption seems to be inaccurate. In this study, comparisons are made between analytical models based on infinite domains and series expansion solutions as well as numerical models based on FEM (Finite Element Method). As the parameters governing the problem are random in nature, MCS (Monte Carlo Simulation) is applied to find the probability distributions for the time for initialization of corrosion using the different methods. Comparisons are made between treating the random quantities as random variables or as spatial and temporal distributed random fields. The random fields are modeled by the EOLE (Expansion Optimum Linear Estimation) approach. A bridge pier in a marine environment is considered to exemplify the results.For many reinforced concrete structures corrosion of the reinforcement is an important problem since it can result in expensive maintenance and repair actions. One mode of corrosion initiation occurs when the chloride content around the reinforcement bars exceeds a critical threshold value, which is the subject of the present paper. Typically, the chloride concentration is obtained by solving Fick's diffusion partial differential equation assuming chloride flow into an infinite half space. However, when the concrete structure is relatively thin or when a rebar is situated at a corner of the structure, the assumption seems to be inaccurate. In this study, comparisons are made between analytical models based on infinite domains and series expansion solutions as well as numerical models based on FEM (Finite Element Method). As the parameters governing the problem are random in nature, MCS (Monte Carlo Simulation) is applied to find the probability distributions for the time for initialization of corrosion using the different methods. Comparisons are made between treating the random quantities as random variables or as spatial and temporal distributed random fields. The random fields are modeled by the EOLE (Expansion Optimum Linear Estimation) approach. A bridge pier in a marine environment is considered to exemplify the results