Modeling of Corrosion Pit Growth in Buried Steel Pipes

A review of the published literature shows that although intensive research has been conducted on corrosion of steel in soils, accurate prediction of corrosion pit growth remains a serious challenge. This study aimed to develop a new model for predicting the maximum pit depth in buried steel pipes and verify it with data obtained from actual corrosion measurements in the field. A method was developed that integrates advanced data mining techniques, shape descriptive modeling, and evolutionary polynomial regression in deriving the underlying relationships between corrosion-influencing factors and model parameters. It was found that the area effect $\psi$ is closely related to the ion content of the soil (${\mathrm{Na}}^{+}{,\mathrm{K}}^{+}$, ${\mathrm{Mg}}^{2+}$, ${\mathrm{Cl}}^{-}$, and ${\mathrm{SO}}_4^{2-}$) and that the correlation between model parameter $K_u$ and soil properties varies among soils with different aeration. It was also found that the developed predictive model exhibits superiority over existing models in quantifying multiple-phase corrosion growth. The proposed method can effectively correlate model parameters with the main contributing factors, which enables researchers and practitioners to accurately predict the maximum corrosion pit depth in buried steel pipes.