Research On Modeling For BOF Steelmaking Based On Mechanism And Intelligent Method

Basic oxygen furnace (BOF) is the most vital steelmaking method. To meet the increasing requirements of more strict quality and richer varieties and to deal with characteristics of nonlinear and strong coupling in this complex physical and chemical process, the related measuring methods and control models attract much attention. A mechanism model of weight of hot metal charged which is used to guide charging process is proposed based on thermal equilibrium and material balance to meet needs of project, while a model of TFe content of endpoint slag based on extreme learning machine (ELM), and a hybrid model connecting support vector machine (SVM) with exponential model of endpoint carbon content are established to improve predicting accuracies.During steelmaking process, increasing the ratio of scrape charged is a very important method to reduce production costs. The model calculates weight of hot metal and max weight of scrape charged based on target of each heat, requirement of melting process and information of hot metal and scrape, and it is restricted to thermal equilibrium and material balance. Simulation results show that the model can accurately calculate the weights charged and can be used to guide charging process.TFe content of endpoint slag determines the weight of tapped steel and affects accuracy of the built model of weight of hot metal charged. Based on ELM, a model with smaller error than traditional experience model is proposed.The exponential model of carbon content is a very valuable experience model. Unfortunately, lack of effective method of determining the exponent limits its application. Based on off-gas analysis data, a hybrid model of SVM and exponent model is built with a better performance on test data.