Document Type : Original Article

Authors

• Farhad Mohsenifar ¹
• Gholam Reza Khayati ²

¹ Mechanical Engineering Department, Higher Education Complex of Bam, Bam

² Department of Materials Engineering and Metallurgy, Shahid Bahonar University of Kerman, Kerman

Abstract

In this study, Gene Expression Programming (GEP) was employed to develop a predictive model for the corrosion current density of galvanized steel in a simulated soil solution. Five environmental variables, including chloride, sulfate, bicarbonate, and citric acid concentrations and temperature, were considered as input parameters, while corrosion current density was used as the output. A dataset comprising 32 experimental observations was divided into training and testing subsets, and several GEP models were developed using GeneX-ProTools. The optimal model was selected based on regression coefficients, residual-error distribution, root mean square error, and mean absolute error. Among the developed models, GEP 12 showed the best predictive performance and was used to derive an explicit mathematical expression for corrosion current density. The predicted results showed good agreement with the experimental data for both training and testing datasets, demonstrating the capability of GEP to capture the nonlinear relationships governing the corrosion process. Sensitivity analysis ranked the investigated parameters as follows: temperature > citric acid > sulfate > chloride > bicarbonate. Increasing temperature, chloride, and bicarbonate concentrations generally increased the predicted corrosion current density, whereas sulfate and citric acid exhibited negative sensitivity coefficients within the investigated ranges. Three-dimensional response surfaces further revealed significant nonlinear interactions among the environmental parameters, particularly between temperature and ionic species. Overall, the developed GEP model provides an explicit and interpretable tool for predicting the corrosion behavior of galvanized steel in simulated soil environments and identifying the relative importance of environmental variables.

Keywords

• Gene expression programming
• Sensitivity analysis
• Corrosion