AN application of prediction model in blast furnace hot metal silicon content based on neural network

Abstract

Radial Basis Function (RBF) neural network is used to predict the blast furnace hot metal based on its characteristics such as fast convergence and global optimization. As hot metal silicon content had close relationship with furnace temperature, the change of temperature in furnace was reflected indirectly by hot metal silicon content. Newrbe function in Matlab was applied for function approximation. Normalized data of normal production for a long period was used for training and simulation. The results showed that the hitting rate of prediction for silicon content was improved. The application of RBF neural network prediction model in blast furnace could forecast Si-content, judge the trend of temperature and realize the control of blast furnace temperature, which was advantageous to energy saving. Moreover, the model can monitor multi-objects simultaneously and provide guidance for blast furnace process