DescriptionThis thesis presents modeling methods for the rebar degradation process and gives a forecast for it by using machine learning approaches. With the widespread use of reinforced concrete materials around the world, rebar has become a key component to bridges and roads. Due to the influence of the external environment, the concrete enclosing the reinforcement will crack, and the rebar embedded will gradually corrode, which may cause serious consequences. Based on the experimental data of reinforcement corrosion, this study analyzed the factors affecting the rebar corrosion. Considering the influence of experimental environmental factors, a two-stage degradation model of rebar degradation is established to evaluate the corrosion rate of steel bar under different concrete crack size and chloride ion concentration. The accelerated life testing model for crack size on concrete and chloride ion concentration is established and its reliability performance under different environmental pressure is evaluated. Machine learning methods are also used to predict the concrete degradation level, and a method of combining machine learning with two-stage degradation model is found. Together this study is useful to understand the rebar corrosion and the influence of environmental factors on its reliability performance. Through effective evaluation and maintenance, it helps to formulate maintenance strategies and prevent safety accidents.