Combination Model Based on CBR and SVM for BOF Oxygen Volume Calculation

Abstract

Oxygen blowing volume control is very important in Basic Oxygen Furnace (BOF) steelmaking. A combination model based on information theory and artificial intelligence technology is proposed for oxygen blowing volume calculation. The combination model is composed of Case-based Reasoning (CBR) model and Support Vector Machine (SVM) model. In CBR model, the mutual information is introduced in case retrieval step to determine the weights of attributes. In SVM model, the mutual information is adopt to distinguish the importance of input variables by setting a different weight to each variable. The CBR model is viewed as experience based model and the SVM model is viewed as data based model. To model the oxygen blowing volume accurately, CBR model and SVM model are combined. Tests on a 180 ton BOF data are implemented to validate the effectiveness of the proposed method.