Experimental Investigation and Machine Learning Prediction of Corrosion Rate and Inhibition Efficiency of ASTM A36 Steel in NaCl Media

Abstract

This work aims to investigate the effects of corrosion factors including variations in inhibitor type, inhibitor concentration, and immersion duration towards on the corrosion rate and inhibition efficiency of ASTM A36 steel in a 3.5% NaCl solution. The research was revealed from experimental data and machine learning approach. Corrosion rate measurement was performed by adding different inhibitors which are sodium nitrite and ascorbic acid inhibitors. Each sample was immersed for 7, 14, and 21 days in different inhibitor concentrations of 250, 500, and 750 ppm.. Weight loss was obtained to calculate the corrosion rate, then it was used to get inhibition efficiency. The experimental data was collected as the dataset to develop corrosion rate and inhibition efficiency models. The dataset was trained using several regression algorithms, and the selected model was evaluated by the highest R² value. The best performance of predictive corrosion rate model was 1.00 from XGBR and 0.86 for the inhibition efficiency prediction model obtained from RFR. The most significant factor affecting corrosion rate was inhibitor concentration, while immersion duration had the strongest effect on inhibitor efficiency. Therefore, the provides a comprehensive understanding to explain the relationship between corrosion control conditions in ASTM S36 steel.