Análise da propagação da corrosão em estruturas de concreto armado: revisão e avaliação dos modelos matemáticos existentes na literatura

Abstract

This study presents a comparative analysis of empirical models from the literature used to predict the corrosion rate of reinforcement in reinforced concrete structures. Nine models were selected based on their recurrence in technical and scientific studies, and are presented along with their respective formulas and variables. The evaluation was conducted using an experimental database comprising 199 data sets obtained from both laboratory and field tests, as reported in the literature. The empirical models were applied to this database, and their results were compared with the corresponding experimental values. The statistical analysis was based on the mean value, standard deviation, and coefficient of variation of the ratio between the experimentally observed values and those calculated by the models, enabling an assessment of both the average prediction tendency and the variability of each model. Graphs were prepared to visually represent the performance of each model. The study also highlights the characteristics, limitations, and application domains of the analyzed models, emphasizing the importance of a careful selection of equations in durability assessments and service life prediction. It is concluded that model performance varies significantly, and that a combined analysis of the three statistical indicators is essential to evaluate the reliability of corrosion rate estimations in reinforced concrete structures. The accuracy of the models depends on environmental conditions, and a probabilistic error analysis is fundamental to enhance the reliability of predictions.