Application of Clustering Algorithm to Blast Furnace Expert System

Focusing on the complex, nonlinear and great hysteresis characteristics in the iron making process of blast furnace, the mathematical model of metallurgical can not satisfy the requirement of guiding operation, so the expert system based on the blast furnace experts' knowledge and experience developed rapidly. However, because of the limitation of the experts' knowledge and experience, this paper applied clustering algorithm to the development of blast furnace expert system. Based on the clustering algorithm, this system developed not only operational furnace profile management model, but also two methods of gas flow distribution pattern recognition to cross thermometric and top infrared imaging. This new system had achieved satisfying results when applied in blast furnaces of WISGCO, Jinan steel and Ansteel.