Neural Network Model On Corrosion Behavior Of Carbon Steel In Simulated Acid Rain And Its Application

Human has a long history on using steel, after the industrial revolution in eighteenth century carbon steel palys an important role in construction industry, manufacturing and people's daily life. But when carbon steel exposed in natural environment, it's easily to be corroded by the influence of various natural factors. The study of steel corrosion has always been an important issus in corrosion area. Acid rain is an important factors in metal corrosion which in atmospheric environment. It could significantly accelerate the metal corrosion rate. Neural network is the mathematical system of imitate human brain's structure and function, it's able to build a non-linear relationship between input information and output information.And the neural network has a broad prospect in the application of corrosion.In this dissertation, the study was focused on the corrosion behavior of Q235 and 35steel in simulated acid rain. By the electrochemical experimental, it's corrosion current density in different acid rain solution was determined. The electrochemical experimental data was inputed neural network to establish corrosion model. This model was improved and used to forecast and analyze the carbon steel's corrosion behavior in acid rain. The results showed that: when the pH was increased, the corrosion potential of Q235 and 35steel were gradually shifted in positive direction, and current density was decreased. Integrated film was formed on the carbon steel surface and partly protected parent metal when PH exceed 4.5, and this situation would cause corrosion rate rapid decline. With the concentration of NO₃^- and SO₄^2- increased ,the corrosion potential of Q235 and 35steel were gradually shifted in positive direction, the polarization resistance decreased and increased the corrosion current density increased; Cl^- were made the corrosion potential shifted in negative direction and polarization resistance decreased, so increased the corrosion current density. In this three kinds of corrosive anions, the effect of Cl^- with corrosion was slightly greater than SO₄^2-, and NO₃^- the effect of NO₃^- is least.Although the direct effect of cations in simulation acid rain solution with corrosion behavior were relatively small. But their influence could be reflected by effect PH. In order to study the relationship between the major cations in simulation acid rain solution and PH , the neural network model of Ion concentration-PH was build. This model was used to forcast the relationship between the cation concentration and PH, and find the Ca²⁺ and SO₄^2- was the main factor of neutralization.