Process fault diagnosis method for cement clinker preparation process based on BO-XGBoost

This paper proposes a diagnosis method based on Bayesian Optimization (BO) Exterme Gradient Boosting (XGBoost) model for cement clinker preparation process faults, which mainly addresses the shortcomings of untimely processing and low sensitivity in the fault diagnosis system. To realize the intelligent diagnosis of preparation process fault. Firstly, the key parameters associated with the process fault are determined based on the mechanism of the preparation process of the material, and secondly, the data of the associated key parameters are pre-processed by outlier rejection, filtering and normalization; then, the parameters of the XGBoost model are optimized based on the Bayesian method and the production data to establish the process fault diagnosis model of the preparation process; finally, the actual data are combined with multiple models and other models optimized by Bayesian to achieve Finally, the validity of the model is verified by comparing multiple models and other models implemented by Bayesian optimization. The experimental results show that the BO-XGBoost diagnostic method presented in this paper has higher diagnostic accuracy.