Hot Metal Temperature Prediction based on AdaBoost Model and SVM Model
Based on the actual production data of No.3 BF in an ironmaking plant,the primary data of the furnace is pro-cessed because of its problems of value duplicating,missing or abnormality,and the AdaBoost model and SVM model are used to predict the hot metal temperature.The processing results prove that the prediction accuracy of AdaBoost model is better than that of SVM model with R2 reaching 0.878 and prediction accuracy reaching±5℃ for 85.21%of predictions,demonstrating that AdaBoost model can satisfy the actual needs of the furnace production.Based on the data warehouse sys-tem of BF ironmaking,the data connection is built up between the FineBI front-end tool and the application of hot metal temperature prediction technology to form a front-end interface that is composed of the modules of characteristic data,cor-relation analysis,prediction result,model evaluation and prediction curve.This interface providesa visualized display for the application of hot metal temperature prediction technology.
blast furnacehot metal temperature predictionAdaBoost modelSVM modeldata warehouse