Identification of Key Environmental Factors Affecting Atmospheric Corrosion of Mild Steel Using Machine Learning Techniques

Corrosion of mild steel poses significant economic and structural challenges, with traditional prediction methods often lacking accuracy. This study leverages machine learning techniques, specifically Random Forest Regression (RFR) and correlation analysis (Pearson and Spearman), to enhance corrosion rate prediction. A dataset of mild steel corrosion under Indian atmospheric conditions was analyzed, identifying Minimum Relative Humidity, Temperature Variations, Rainfall, Chloride, and Sulfur Dioxide as key influencing factors. The Random Forest model outperformed conventional methods in predictive accuracy. These results demonstrate the importance of AI-driven approaches for automated corrosion monitoring and predictive maintenance, improving material longevity in industrial applications.