Analyzing the Time-Delay Correlation Between Operational and State Parameters of Blast Furnace: A DTW and Clustering Based Method

The complexity of the smelting process leads to many difficulties in accurately controlling the reaction of the blast furnace. It takes some time after the adjustment of the operational parameters to cause a change in the state parameters. The identification of time delay correlation between the operational parameters and state parameters of blast furnace is of great importance to optimization of blast furnace operation. In this paper, we propose a DTW and clustering based method to identify the time-delay characteristics between operational and state parameters. We propose to use the DTW algorithm as distance measure and use the k-means algorithm to clustering the pre-processed data to reduce the complexity of the calculation. We then use the Spearman correlation coefficient to measure time-delay correlation between operational and state parameters. We validated the proposed method using data collected from a real life blast furnace. Experimental results show that the proposed method can effective identify the time-delay characteristics.