A Prediction Method for CO Utilization Rate Based on KPCA Contribution Plots and ELM Algorithm

Abstract

CO utilization rate is an important indicator reflecting the operating state and energy consumption of blast furnace (BF). To this end, a CO utilization rate prediction method based on KPCA and ELM is proposed. Due to the non-linear characteristics of blast furnace data, the KPCA contribution map method is used to filter 49 input variables, and 37 variables with large contributions such as hot air pressure are obtained as input variables, which solves the variable redundancy problem that the ELM algorithm cannot solve. Then, the ELM method was used to establish a prediction model for the utilization rate of CO in a blast furnace. According to the actual data on site, the mean square error and coefficient of determination are used to evaluate the validity of the prediction results of the ELM model and the ELM model. The experimental results show that the method proposed in this paper has stronger model generalization ability, and can more accurately predict the blast furnace CO utilization rate, and provide timely and effective decision support for subsequent blast furnace operation optimization and energy saving and emission reduction.