Experimental and numerical simulation of carbonation-induced corrosion in reinforced concretes

Despite the extensive research on carbonation-induced corrosion (CIC), there is still no consensus on the carbonation front location and the reinforcement corrosion onset. In addition, a reliable diffusion-reaction numerical simulation for quantitative prediction of carbonation depth seems to be lacking. In this study, normal and pozzolanic concretes, both with and without reinforcing, were exposed to 5 % CO 2 , 30 ℃, and 65 % RH for up to 12 months. A novel finite element method-based model was simulated using COMSOL Multiphysics software and validated by carbonation depth results of experimental data. Regarding half-cell potential results, consideration of the criterion of significant potential drop instead of the threshold values specified in ASTM C876 allows for determining the onset time of CIC. Among the surveyed approaches for corrosion onset, considering the critical carbonation depth led to a conservative prediction with an average error of about 41 %. However, the presented theoretical models provided a more reliable prediction with an average error of 18 %. By increasing the concrete cover thickness on the reinforcement bars from 1.5 to 2.5 cm, the average initiation time of CIC increased by approximately three times. The section loss rate of reinforcements in 8 % SF and 20 % GGBFS mixtures increased by about 49 % and 86 %, respectively, compared to the control mixture without SCMs. • A novel diffusion-reaction numerical simulation was proposed for predicting carbonation depth. • Five approaches were compared to predict the onset time of CIC. • The criterion of significant potential drop is helpful to assess the onset time of CIC based on the half-cell potential test. • As the cover thickness of RC increased from 1.5 to 2.5 cm, the onset time of CIC increased approximately 3 times. • Applying the critical carbonation depth led to a more conservative prediction for CIC initiation compared to theoretical approaches.