During long-term service, railway concrete box girders are subjected to the combined effects of mechanical loading, freeze–thaw cycles, and salt attack, posing significant challenges to their durability. To overcome the limitations of conventional diffusion models in handling complex boundaries and multiple coupled factors, this study develops a modified chloride diffusion model incorporating load-induced damage, freeze–thaw effects, material deterioration, and chloride binding. A two-dimensional numerical simulation method based on cellular automata is established to simulate chloride diffusion in railway concrete box girders under complex environmental conditions.
A 32 m prestressed concrete simply supported box girder is selected to investigate chloride diffusion in regions with different geometric boundary conditions. A chloride concentration of 0.17% at the reinforcement surface is adopted as the criterion for corrosion initiation to predict the time to initial reinforcement corrosion in different regions. The results show that geometric boundary conditions significantly influence chloride diffusion. Multidirectional diffusion regions, particularly the junctions between the top or bottom slabs and the webs, as well as corner regions, exhibit significantly faster chloride ingress than ordinary straight regions. Compared with ordinary straight regions, the minimum time required to reach the critical chloride concentration in corner regions can be reduced by nearly 50%. The minimum diffusion times in two typical junction regions are 36.2 and 30.8 years, respectively.
This study reveals the chloride ingress characteristics and durability differences among different regions of railway concrete box girders under complex environmental conditions, providing a reference for durability design, localized protection, and service life prediction. Future research could further consider the coupling between unsaturated moisture transport and chloride migration during freeze–thaw cycles. In addition, long-term field monitoring data could be used to validate and refine the model, thereby improving its applicability and accuracy in predicting the service life of concrete structures under complex engineering conditions. The work entitled “Chloride diffusion and life prediction of railway concrete box girders based on cellular automata” was published on Journal of Railway Science and Technology (published on Aug. 4, 2026).
DOI：10.1016/j.jrst.2026.08.002