Artificial Neural Network (ANN) analysis on carbon dioxide corrosion

Artificial neural network (ANN) method for the data processing of carbon dioxide corrosion is proposed on the base of plenty of experimental data. A detailed model for predicting corrosion rate is presented, which does not need all kinds of materials and environment parameters. Only some important parameters are necessary, that is, carbon content of mild steel, carbon dioxide partial pressure, temperature, fluid velocity. The result shows that predicted value is generally consistent with actual value. It is proved that the ANN technology is applicable for the corrosion rate evaluation of complicated system, and it also presents a new approach to evaluating remain life of equipment and developing the expert system.