A particle‐based stochastic model for bidirectional chloride ingress in concrete exposed to seawater

Abstract Chloride‐induced reinforcement corrosion is one of the causes of the deterioration of reinforced concrete structures exposed to deicing salts and marine environments. Reliable predictions of chloride ingress are essential for the assessment of service life, as they provide the basis for the evaluation of corrosion risk. Most chloride ingress models assume unidirectional diffusion under semi‐infinite boundary conditions, whose validity may be questionable for thin structural members exposed to chloride penetration from opposing surfaces. Furthermore, analytical formulations often involve boundary conditions that limit their applicability predominantly to unidirectional diffusion problems. This study presents a numerical approach to modeling bidirectional chloride transport using a particle‐based model. The model describes diffusion as a random process governed by probabilities and allows for the simulation of interfering diffusion fronts originating from opposing boundaries. This interference leads to thickness‐dependent transport behavior, resulting in deviations from predictions made with conventional unidirectional models. The model accounts for particle diffusion through layered material, distinguishes between skin and internal zones, and introduces primary and secondary age‐related changes in the diffusion coefficient of concrete. Surface boundary conditions are defined through concentration matrices, allowing flexible representation of the environment. To assess the stability and accuracy of the model, a sensitivity analysis of key numerical parameters was conducted, and guidelines for their selection are presented. The bidirectional numerical approach was validated on the basis of long‐term field data from concrete specimens exposed to the marine environment. The results show that the use of simplified diffusion models for two‐sided exposure can lead to significant inaccuracies, which underlines the importance of numerical approaches that account for finite geometry and heterogeneous material properties in service‐life assessment.