Predicting the Corrosion Rate of Medium Carbon Steel Using Artificial Neural Networks

Medium carbon steel is commonly used in waterfront structures, i.e., ports, and piers, where it is surrounded by very aggressive environmental conditions. Thus, it is very susceptible to different forms of corrosion. This work proposed an artificial neural network (ANN) model to predict corrosion rate of tempered medium carbon steel in environmental conditions close to these conditions where it is commonly used. Tafel analysis was used to determine the corrosion rate of the heat-treated samples. Optical microscope was used also to examine the morphology of the surface after tempering process. Eleven different tempering temperatures between 400 to 600 degrees C, and three holding times 45, 90, 135 min were selected. Over the whole set of experimental data, the results show that the proposed ANN can achieve an excellent classification accuracy of around 92.63%. Therefore, the proposed model has promising potential application to predict medium carbon steel corrosion at different tempering conditions.