On the Prediction of Atmospheric Corrosion of Metals and Alloys in Chile Using Artificial Neural Networks

Most metals and alloys exposed to the environment suffer deterioration due to the effects of atmospheric corrosion. This study presents results obtained for the corrosion of carbon steel, galvanised steel, copper and aluminium exposed to the environment for a period of 3 years, at 9 different sites around Chile. Mathematical models based on artificial neural networks are used to evaluate the corrosion of the metals and alloys as a function of meteorological variables (relative humidity, temperature and amount of rainfall), pollutants (chloride and sulphur dioxide) and time. The advantages of these models in predicting corrosion is also shown in comparison to traditional statistical regression models when considering the dependence of corrosion as a function of time alone.