A Novel Analysis of Furnace Condition Based on Biclustering

In the blast furnace iron-making, furnace condition is directly related to the production efficiency and blast furnace life. Currently, the analysis of furnace condition is mainly based on the artificial experience. Further progress does not exist in the analysis of furnace condition due to the difficulty of the mechanism modeling. In this paper, the biclustering algorithm is applied to the analysis of furnace condition, which is more suitable for the analysis of high dimensional data and has the ability to deal with large-scale data. First, the rank Order-Preserving Submatrix (r-OPSM) algorithm based on hierarchical search is proposed, which improves the robustness of OPSM algorithm to noise and is used for correlation analysis among multiple blast furnace conditions. On this basis, the novel concept of intermediate furnace condition is pointed out, which lays a theoretical foundation for further analysis of the state of the furnace.