Intelligent Control System of Blast-furnace Melting Efficiency

The paper suggests an approach to creating intelligent blast furnace performance control system. This system is based on the analysis of statistical data on the operation of a blast furnace over several years. Cluster analysis is proposed as the main approach to big data processing. Cluster analysis reveals hidden relationships in data. In addition, this approach allows to determine the operation mode of the blast furnace despite a large set of input parameters. Kohonen neural network was chosen for clustering. In this paper, the results are blast furnace operating modes by blast parameters. The article also presents the concept of an expert system. An expert system based on cluster partitioning data allows to evaluate the efficiency of the furnace according to the criterion of maximum productivity and minimum coke consumption. Besides, this system allows to rely on calculations when planning the work of the workshop for a given time period