Prediction and Analysis of Reinforcement Corrosion in Simulated Concrete Pore Solution Based on Neural Network

The corrosion of reinforcement has always been a problem to be solved in the field of architecture. In this paper, the corrosion characteristics of chromium alloy steel under different pH conditions are studied. The impedance characteristics and equivalent circuit are predicted by neural network model. First of all, in simulated pore solution with different pH values, the characteristics of Nyquist impedance spectroscopy of the whole chromium alloy under passivation stage and the damaged passivation film of reinforcing bars under initial corrosion stage have been found. Then, according to the difference of impedance characteristics under different pH values, different equivalent circuits have been established and chi(2) values of different equivalent circuits under different chloride ion concentration have been calculated. By fitting the electrochemical parameters of the equivalent circuit with chi(2) values, the equivalent circuit model which can be predicted by neural network has good consistency with the equivalent circuit which can be predicted by chi(2) values.