Research on the Prediction of Corrosion Rate of Industrial Circulating Cooling Water

To maintain stable production operations, it's crucial to promptly assess the corrosion trends in industrial circulating cooling water quality. A prediction method of corrosion rate based on Extreme Gradient Boosting (XGBoost) algorithm is proposed for predicting the corrosion rate of industrial circulating cooling water. Given the numerous and nonlinear characteristics of water quality parameters, Kernel Principal Component Analysis is first used to reduce the dimensionality of water quality parameters. Based on the XGBoost model as the prediction model, the parameter selection of the XGBoost model is combined with Bayesian optimization algorithm to find the optimal parameter combination. The results indicate that the established XGBoost model exhibits strong learning and predictive capabilities, making it suitable for predicting the corrosion rate in industrial circulating water quality.