Prediction model for furnace cylinder activity based on GRA-SSA-LSSVM

Since entering the new century, China's iron and steel technology has developed rapidly, and the issue of blast furnace life has been increasingly valued by researchers. Furnace cylinder activity is considered to be the key factor affecting the life and stable anterograde of blast furnace, so reasonable prediction of furnace activity is of great significance to guide blast furnace production and improve blast furnace smelting efficiency. This paper proposes a furnace cylinder activity prediction model based on the GRA-SSA-LSSVM model. The pre-treated production data is used to obtain the quantified activity index of blast furnace cylinder by relevant formulas. Then the feature variables were selected by grey relational analysis combined with mechanism analysis. In order to further improve the prediction accuracy of the model, a furnace cylinder activity prediction model based on Sparrow search algorithm (SSA) optimized least square support vector machine (LSSVM) was proposed based on the above content, and the effectiveness of the predictive model was simulated and analyzed. The results show that the accuracy of predicting the cylinder activity using this method is high, which can meet the needs of guiding the actual production.