Prediction Of Silicon Content In Blast Furnace Hot Metal Based On Particle Swarm Optimization Extreme Learning Machine

In the process of blast frunace smelting,reasonable furnace temperature is one of the key factors to keep the blast furnace production stable and straightforward.Due to the complexity of the blast furnace smelting process and the limitation of current technical level,it is difficult to accurately grasp the temperature inside the furnace.Therefore,the prediction technology of the furnace temperature is of great significance to the operation of the blast furnace.In the process of ironmaking in blast furnace,the thermal state of blast furnace is usually characterized by the silicon content of molten iron.Establishing a reliable prediction model of silicon content in molten iron has important guiding significance for blast furnace operators and has important theoretical research value.In this paper,the silicon content of blast furnace hot metal is used to indirectly predict the furnace temperature.The main work includes:(1)Aiming at the problem of insufficient prediction efficiency and precision of silicon content,a method combining principal component analysis(PCA)and particle swarm optimization(PSO)improved extreme learning machine(ELM)is proposed to predict silicon content in hot metal of blast furnace.Because there are many factors affecting the silicon content of molten iron,and the factors affect each other,the input variables that affect the silicon content are subjected to dimensionality reduction through principal component analysis.The particle swarm optimization algorithm is used to optimize the weight and threshold of the extreme learning machine.The root mean square error is used as a fitness function to build a predictive model.The extracted principal component was used as model input,and the molten silicon content was used as a model output.Finally,the extreme learning machine algorithm and particle swarm optimization extreme learning machine are compared.The experimental results show that the improved prediction model improves the accuracy of silicon content prediction.This method can provide a new idea and method for the production operation of blast furnace.(2)Under the Qt platform,combined with the VTK graphic visualization toolkit,the blast furnace 3D visualization system was developed.The system can not only draw the silicon content prediction results in two dimensions,but also dynamically display the blast furnace 3D model.Through the research on VTK visualization method,it provides reference for developers to configure VTK in Windows/MinGW development environment and visualization research.