Probabilistic and Statistical Modeling of Chloride-Induced Corrosion for Concrete Containing Metakaolin

Abstract

This paper presents new models for computing the chloride-induced corrosion period based on combining Monte Carlo simulation and statistical analysis methods. The study is divided into two parts. The first part utilized Monte Carlo simulation to predict the probability of corrosion (1–200 years) for concrete containing metakaolin (MK) with different mixture proportions and concrete covers. Also, the effect of different mixture parameters (percentage of MK, binder content, and water-to-binder ratio) on the probability of corrosion was studied in this part. The second part adopted statistical analysis to develop models predicting the chloride-induced corrosion period (time to reach 10% probability of corrosion initiation) in various MK mixtures and identified the most significant factors affecting this period. Design charts were also developed using statistical analysis to facilitate and simplify the prediction of the chloride-induced corrosion period. The results showed that the probability of corrosion decreased as the percentage of MK increased. Also, using a lower W/B ratio or higher binder content in MK mixtures improved the effectiveness of MK to reduce the probability of corrosion. The results also showed that the most significant factor affecting the chloride-induced corrosion period was found to be MK replacement, W/B ratio, and binder content respectively, in order of significance. The developed models and design charts in this paper will help designers and engineers to better understand the influence and the importance of various mix design parameters of MK mixtures on the chloride-induced corrosion period estimation.