Multi-mode smelting method based on the classification system of molten iron

Abstract

The invention is in the field of iron and steel metallurgy, specifically a method and system for determining the amount of alloy added during the converter tapping process. Given that the LSTM neural network has a strong ability to capture nonlinear relationships, the invention builds an alloy element yield prediction model based on the LSTM neural network. Because different alloy elements have different factors that affect their yield, that is, different model input variables, different LSTM models are established for training. Furthermore, the invention uses integer linear programming to combine the yield prediction results to determine the alloy addition amount. This method not only finds the optimal alloy proportioning scheme quickly, but it also improves the component hit rate and the stability of steel products in the converter steelmaking process, obtains the lowest total cost, effectively reduces alloying costs, and has a good application prospect.