Premium
End Temperature Prediction of Molten Steel in LF based on CBR–BBN
Author(s) -
Feng Kai,
He Dongfeng,
Xu Anjun,
Wang Hongbing
Publication year - 2016
Publication title -
steel research international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.603
H-Index - 49
eISSN - 1869-344X
pISSN - 1611-3683
DOI - 10.1002/srin.201400512
Subject(s) - reliability (semiconductor) , refining (metallurgy) , artificial neural network , bayesian network , process (computing) , interval (graph theory) , computer science , data mining , artificial intelligence , materials science , mathematics , metallurgy , power (physics) , physics , quantum mechanics , combinatorics , operating system
To improve the control level about the end temperature of molten steel in ladle furnace (LF) refining, a combined method of case‐based reasoning (CBR) and Bayesian belief network (BBN) has been proposed to predict the end temperature of molten steel in LF. The evaluation of the reliability of cases is conducted by applying BBN for the assessment which the CBR method lacks in. A BBN is established based on the actual production process of LF refining. Statistical interval is then determined for each node, and finally the probability of each case in the case base is computed to evaluate the reliability. On such basis, the case retrieval algorithm in CBR is revised using the reliability of cases and the prediction accuracy of the revised algorithm is compared with ordinary CBR and back propagation neural network (BPNN). The results show that CBR–BBN has higher prediction accuracy than ordinary CBR and BPNN in the prediction about the end temperature of molten steel in LF refining.