An Intelligent Selection Strategy for Burden Distribution Based on Classification of Blast Furnace Condition

摘要

B0615. However, due to the low transparency of this process, operators adjust the burden distribution according to their own experiences, which makes it difficult to keep optimal production indexes for a BF. In this paper, an intelligent selection strategy of burden distribution is devised according to the energy-consuming index, where carbon-monoxide utilization rate (CMUR) is taken as the energy-consuming index. First, state variables are chosen based on the calculated Pearson correlation coefficient among each state variable. Then, the chosen state variables are divided into five clusters by k-means clustering algorithm. Finally, the values of CMUR in each condition cluster with different burden distribution are counted. According to the statistics, the burden distribution with the highest CMUR is judged to be the optimal one. Simulation results based on industrial data show that the proposed strategy selects the optimal burden distribution and provides a reasonable guide for the operators.