ART2-Based Approach to Judge the State of the Blast Furnace
Tieqiang Sun, Yixin Yin, Shengli Wu, Xuyan Tu · 2006
The complicate chemical reactions inside the blast furnace and many parameters affecting its working procedure during the process, it is very hard to judge the state of the blast furnace by traditional techniques, ART (adaptive resonance theory) network accommodate these requirements through interactions between different subsystems, automatically detect clustering and form classes of the data structure. This paper proposes the factors of affecting the state of blast furnace; the model of ART2 for judging the state of the blast furnace is established; the state of the blast furnace is classified four sub-states: good, better, notice, bad. When a batch of new data is collected, the state of the blast furnace can be predicated by the ART2 neural network and achieves high veracity