Application of Large scale Database-based Online Modelling on Blast Furnace Operation
| Authors: | Ito Masahiro, Nippon Steel Corporation, Japan
Matsuzaki Shinroku, Nippon Steel Corporation, Japan
Ogai Harutoshi, Waseda University, Japan
Mori Kenichi, Waseda University, Japan
Uchida Kenko, Waseda University, Japan
Saito Shinichi, Nittetsu Elex Corporation, Japan
Sasaki Nozomi, Nippon Steel Corporation, Japan |
|---|
| Topic: | 6.2 Mining, Mineral & Metal Processing |
|---|
| Session: | Steel Mills, Sintering, Furnaces and Converters |
|---|
| Keywords: | Database, Dynamic modelling, Nonlinear models, On-line control, Regression estimates, Steel manufacture, Variables, Validity |
|---|
Abstract
A practical method of JIT modelling for a blast furnace process data whose characteristics are very complicated physical phenomena and stiff nonlinear process are studied. The proposed method is composed of severely selecting process data variables by stepwise method and searching the past similar data in quantized topological space constructed by the selected variables. The effectiveness of this method is confirmed. It is also confirmed that it is sufficiently possible to apply this method to online use because the calculation time is quite short for searching the past similar data and estimating the future.
