Process control of blast furnace ironmaking poses a great challenge because of its great complexity. As a main indicator of thermal state in blast furnace, silicon content in hot metal must be maintained at an appropriate level. In the present work, silicon content is taken as the output variable and the subspace identification method is used to identify the model between input and output variables. The identified model is then used for prediction so that future information of silicon content can be obtained. With the predictions of silicon content, predictive control of the ironmaking process becomes feasible. Other practical issues like dimension reduction of input variables and data preprocessing are also discussed.
展开▼
