A Multi-Time-Scale Prediction Method for Blast Furnace Gas Utilization Rate Based on Causality Analysis

Blast furnace operating parameters regulate the gas utilization rate (GUR), and different operating parameters affect the GUR on different time scales. However, the existing methods only analyze and model the prediction on a single-time scale, and the accuracy of the model needs to be improved. In this paper, a multi-time-scale prediction method of blast furnace gas utilization rate based on causality is proposed. Firstly, according to the production characteristics of the blast furnace, burden and blast supply are the operation means to regulate the blast furnace on different time scales. Then, this paper presents a blast furnace operation parameter extraction method fusing correlation coefficient and information entropy, and selects two most informative parameters to characterize the operation. On this basis, the causal analysis method is used to prove that the selected operation parameters affect the development trend of GUR. Finally, the long and short time scale prediction models of GUR are established, which are verified by the data of the actual blast furnace production process, improving the prediction accuracy of the GUR efficiently.