Application of improved local models of large scale databasebased online modeling to prediction of molten iron temperature of blast furnace

Norio Kaneko, Shinroku Matsuzaki, Masahiro Ito, Haruhisa Oogai, Kenko Uchida

    Research output: Contribution to journalArticle

    11 Citations (Scopus)

    Abstract

    The large scale database-based online modeling called LOM is the one of local modeling method. This method has been developed to apply the just-in-time modeling for the blast furnace by us. In this paper, we propose two new types of local models in LOM to improve the prediction performance. One is used weighted multiple regression model as a linear local model of LOM. The other is used on-line Bayesian learning model as a nonlinear local model of LOM. In order to compare the prediction performance of the two types of local models in LOM, we evaluate the prediction performance by using the real process data of the blast furnace.

    Original languageEnglish
    Pages (from-to)939-945
    Number of pages7
    JournalISIJ International
    Volume50
    Issue number7
    DOIs
    Publication statusPublished - 2010

      Fingerprint

    Keywords

    • Blast furnace
    • Just-in-time modeling
    • Mutual information
    • On-line bayesian learning
    • Prediction
    • Process control
    • Sequential monte carlo
    • Weighted multiple regression

    ASJC Scopus subject areas

    • Mechanical Engineering
    • Mechanics of Materials
    • Materials Chemistry
    • Metals and Alloys

    Cite this