Hybrid Intelligent Model for Predicting Corrosion Rate of Carbon Steel in CO₂ Environments

ABSTRACT This study aims to construct a prediction model for the internal corrosion rate of offshore pipelines in CO 2 environments, with the intention of providing effective corrosion prediction and protection strategies for the oil and gas industry. By conducting investigative analysis and integrating CO 2 corrosion experimental data, principal component analysis (PCA) was employed to extract the primary influencing factors, which were used as input variables for the support vector regression (SVR) model with corrosion rate as the output variable. The particle swarm optimization (PSO) algorithm was utilized to optimize the hyperparameters of the model, enhancing prediction accuracy. The results indicate that the first eight principal components account for 95.9% of the cumulative contribution, and the optimized SVR model achieved a correlation coefficient (R 2 ) exceeding 0.90. Compared to other models and optimization methods, the SVR model optimized with PCA and PSO effectively predicts the corrosion rate of offshore pipelines, offering theoretical support for corrosion protection.