Natural corrosion in reinforced concrete structures

Sammanfattning: Among the treats to the durability of concrete structures, corrosion of the reinforcement bars is undoubtedly the most common one. Corrosion damages impair safety and durability of infrastructure, and assessment of the safety is challenging due to the complex nature of the corrosion process. Furthermore, research on the topic often requires adapting results from short-time laboratory tests, where corrosion of the reinforcement bars is induced using impressed current, to the reality of existing structures. The use of impressed current results in differences in type and distribution of corrosion products. Naturally corroded specimens are hence the necessary bridge between the knowledge acquired from artificially corroded specimens and the application to real structures. This works investigates the structural effects of natural, chloride-induced corrosion in reinforced concrete structures. Specifically, three research questions were investigated. First, the bond and anchorage of naturally corroded plain bars was studied using 3-point bending and pull-out tests. The tests were designed to be applied to specimens taken from a decommissioned bridge from the 1930s. The bond capacity of plain bars was observed to be significantly higher than in results obtained from laboratory tests on similar bars. Significant factors influencing the effect of corrosion damages on the bond were casting position and presence of stirrups. Finite element analyses were used to further investigate the bond-slip behaviour of the tested specimens. The results highlighted the effect of the loss of bond at yielding on the structural behaviour of the specimens. The second question looked into the characteristics of the corrosion products and the surrounding concrete; this is relevant to assess corrosion damages in existing structures. Neutron imaging and X-ray computed tomography were used to obtain qualitative and quantitative data on corrosion damages in a naturally corroded specimens, including iron to rust ratio. Comparison with an artificially corroded specimen showed differences in distribution of the corrosion products. Finally, possible correlations between transversal cracks and corrosion damages was investigated. A dataset was compiled from experiments available in literature. In the selected studies, corrosion of the steel reinforcement resulted from exposing laboratory specimens, pre-cracked in 3-point bending, to chloride environments. No clear correlation between surface crack width and corrosion characteristics was found, but corrosion pits were shown to likely appear in the proximity of transveral cracks. To conclude, this work highlights the complexity of the corrosion process and argues that a thorough understanding of the material and environmental characteristics influencing this process is necessary to properly assess existing structures. Tests on naturally corroded structures are a fundamental step towards acquiring this knowledge.