Indexed by:期刊论文

First Author:Zhao J.

Correspondence Author:Zhao, J.; Research Center of Information and Control, Dalian University of Technology, Dalian 116024, China; email: zhaoj@dlut.edu.cn

Co-author:Zhang X.,Wang W.

Date of Publication:2010-12-01

Journal:ICIC Express Letters

Included Journals:EI、Scopus

Document Type:J

Volume:4

Issue:6 A

Page Number:2069-2074

ISSN No.:1881803X

Abstract:The accurate prediction to the gas amount consumed by hot blast furnace (HBF) is one of the important tasks for the gas scheduling in steel industry. In this study, a prediction model based on least support vector machine (LSSVM) is developed by considering the characteristics of gas consumption flow and the on-line requirement in practice. Since it is difficult to determine the optimal parameters of the model by some common methods, a fast cross-validation (FCV) method combined with the gradient based optimal algorithm is proposed in this paper to realize the on-line optimization to the hyper parameters. The simulation results with the practical blast furnace gas consumption data of HBF in Shanghai Baosteel show the modeling method is time-economized, while the on-line prediction performance are superior to the standard SVM and the neural network. ICIC International ? 2010 ISSN 1881-803X.

Translation or Not:no