Research On Prediction Model Of Silicon Content In Blast Frnace Hot Metal

The most important production link in iron and steel smelting is blast furnace ironmaking.The internal variable distribution is complex,a large number of physical and chemical reactions and structural closure,which makes it difficult for operators to directly measure the blast furnace temperature.In the production process,the change of the blast furnace temperature can be expressed indirectly by the change of the silicon content of the molten iron.Therefore,it is essential to establish an accurate and reliable blast furnace molten iron silicon content prediction model to regulate the blast furnace temperature and ensure the smooth operation of the blast furnace.First,the blast furnace ironmaking was introduced to establish a model closer to the actual blast furnace production.Based on expert experience and grey correlation analysis,seven variables that are highly correlated with the silicon content of the molten iron were selected as input variables of the model,and the collected data were preprocessed to lay the foundation for the establishment of the model.Secondly,genetic algorithm was introduced to optimize the extreme learning machine,which could generate input layer weights and hidden layer thresholds at random when predicting silicon content in molten iron.Through the comparative analysis of the simulation results,the prediction accuracy of the improved prediction model is slightly higher than that of the single extreme learning machine prediction model.On this basis,the PSO-GA algorithm was formed by combining the particle swarm algorithm and genetic algorithm,with particle swarm algorithm as the main and genetic algorithm as the supplement.This method combined the advantages of the two algorithms,and established a PSO-GA-ELM-based prediction model for silicon content in blast furnace molten iron.Simulation results verify the validity of the model.Figure 19;Table 4;Reference 50...