点击次数：

论文类型：期刊论文

发表时间：2014-01-10

发表刊物：NEUROCOMPUTING

收录刊物：SCIE、EI、Scopus

卷号：123

期号：,SI

页面范围：415-423

ISSN号：0925-2312

关键字：Basic oxygen furnace; Case-based reasoning; Support vector machine; Mutual information; Conditional entropy

摘要：The control of oxygen blowing volume and coolant addition amount is very important in the production of Basic Oxygen Furnace (BOF). In this paper, for oxygen blowing volume and coolant addition amount calculation, a combination model is proposed based on information theory and artificial intelligence technology, which is composed of Case-based Reasoning (CBR) model and Support Vector Machine (SVM) model. For the former, mutual information is introduced to determine the weights of attributes and make the retrieval step more valid than traditional method. For the latter, mutual information is adopted to distinguish the importance of input variables by setting a weight for each variable. The CBR model is considered as experience-based model and the SVM model is a data-based model. To solve the control problems of oxygen blowing volume and coolant addition amount, CBR model and SVM model are combined by conditional entropy. Tests on a 180 t BOF data substantiate the effectiveness of the proposed model. (C) 2013 Elsevier B.V. All rights reserved.