Mesoscale Modeling of Chloride Penetration in Unsaturated Concrete Damaged by Freeze-Thaw Cycling: Fid-Bau Portal

Mesoscale Modeling of Chloride Penetration in Unsaturated Concrete Damaged by Freeze-Thaw Cycling

Wang, Licheng / Ueda, Tamon

For concrete structures exposed to frost attack, cracks, or microcracks induced by freeze-thaw cycling can format interconnecting flow paths and allow more water or chloride ions to penetrate into the bulk concrete. It will subsequently facilitate further deterioration of concrete structures and accelerate the corrosion of embedded reinforced steel bars. Moreover, in reality most concrete structures are rarely fully saturated, so that chloride transportation in unsaturated concrete must be studied with respect to the water moving process in order to cover the real existing service conditions. In the current work, a numerical simulation method based on the mesoscale composite structure of concrete, named the lattice network model, is established to analyze the penetration property of concrete; especially the effects of microcracking induced by freeze-thaw damage on the unsaturated flow behavior are investigated. In the mesoscale model, concrete is treated as a three-phase composite material consisting of coarse aggregates, mortar matrix, and interfacial transition zone (ITZ) between the aggregate and the mortar matrix. The diffusivities of each phase, (i.e., water and chloride diffusion coefficients) is separately characterized and quantified in terms of the published test results. The unsaturated flow theory for capillary water absorption and chloride transport is employed to simulate the ingress of water and chloride ions into concrete. It is found that the water absorption and chloride penetration are substantially influenced by the frost action, and the cumulative absorbed water and chloride penetration depth are increased with the increase of freezing-thawing cycles (FTCs). Furthermore, the numerical predictions about water absorption and chloride profiles are compared with the experimental measurements. The comparisons indicate that numerical predictions agree very well with the test data.