Development of a Corrosion Model for Prediction of Atmospheric Corrosion of Mild Steel

Abstract

Corrosion is a naturally occurring phenomenon commonly defined as the deterioration or loss of functions of materials (usually metals) due to the effect of the environmental factors. According to the nature of environment, corrosion can be broadly categorized as corrosion in atmosphere, corrosion in water, corrosion in sea, corrosion in soil etc. Among these types, corrosion of steel in atmosphere is a problem of great interest. Since steel is the most extensively used structural material in industry, and it is well accepted fact that the cost of material deterioration in an atmospheric environment is enormous. Corrosion of metal in atmosphere is inevitable but is controllable with the aid of proper corrosion management systems. For the implementation of a proper corrosion management system it is necessary to study the corrosive nature (corroivity) of the operation environment of a material. Development of a relationship between corrosivity and environmental variables such as relative humidity, temperature, salinity etc. which is known as Corrosion modeling is widely used method use for the evaluation of corrosivity of atmosphere. This paper describes the work carried out to formulate a model for the prediction of corrosion of mild steel under Sri Lankan atmospheric conditions. For this purpose, a model was proposed which is based on published literature on corrosion modeling. The proposed model was calibrated by the data obtained from field exposure tests which were conducted in four different locations in Sri Lanka. The chi-square goodness of fit test has been used to find out the performance of model. The model showed good performance with goodness of fit at 95% significance level. Finally, the model was validated with different set of data and the prediction performance of this model shows a good capability on forecasting of the rate of corrosion of mild steel in different atmospheric conditions in Sri Lanka.