This paper presents the results from validation of models for prediction of chloride ingress in concrete exposed in de–icing salt road environment. Three models including the simple error–function complement (ERFC) model, the DuraCrete model and the ClinConc model, were evaluated using the measurement data collected from both the field exposure site after over ten years exposure and the real road bridges after 25–30 years in service. The sensitivity of input parameters in each model is analysed. The results show that, among different input parameters, the age factor is the most sensitive one. The simple ERFC model significantly overestimates chloride ingress. The DuraCrete model, if the input parameters are properly selected, may give a reasonably good prediction, otherwise often underestimates chloride ingress. The ClinConc model in general gives fairly good predictions for chloride ingress in de–icing salt road environment with heavy traffic at high speed.
Keywords: cement, chloride ingress, concrete corrosion, de–icing salt, diffusion, concrete durability, modelling, road environment, service life, model validation, chloride penetration, road de–icing, age factor, heavy traffic, traffic speed