Assessment of deteriorating reinforced concrete structures using artificial neural networks

An artificial neural network was used to assess deteriorating reinforced concrete (RC) structures using periodical inspection data for thermal power plants along the coast of Tokyo Bay arranged by the Tokyo Electric Power Company. In the analysis, the focus is on chloride-induced corrosion damage of RC structures. 13 input variables such as crack width, crack direction, number of cracks, etc. were selected as the inputs to the artificial neural network, and four output variables were chosen as the desired damage levels. Using a successfully trained neural network, a sensitivity analysis determines the influence of a change in each variable such as maximum crack width, area of peeling-off of concrete, exposure of reinforcement, etc., on the damage level