Using multivariate grey model and principal component analysis to modeling the blast furnace

Blast furnace ironmaking process (BFIP) can be considered as a grey system due to the complexity. In this paper, a new approach is proposed to predict the silicon content in blast furnace (BF) hot metal based on the multivariate grey model in grey theory. Principal component analysis (PCA) method is also used to deal with the high correlation relationship between different variables of BFIP. With the new variables extracted using PCA technology, multivariate grey models show better performance. Numerical simulations show that the prediction accuracy of BF silicon content with multivariate grey models combined with PCA is remarkably improved compared with typical multivariate grey models.