A Multilevel Prediction Model of Carbon Efficiency Based on the Differential Evolution Algorithm for the Iron Ore Sintering Process

The iron ore sintering process is the second most energy-consuming process in iron and steel production, where the main energy consumption results from the combustion of carbon. In order to optimize the utilization of carbon, it is necessary to predict the carbon efficiency first. Thus, this paper develops a multilevel prediction model to predict the carbon efficiency. First, the comprehensive coke ratio (CCR) is employed as an index to measure the carbon efficiency. Based on the mechanism analysis, the sintering parameters affecting the CCR are determined. Next, an improved fuzzy C-means (FCM) clustering method is used to automatically identify the optimal number of working conditions for the sintering parameters. Then, the least-squares support vector machine (LS-SVM) submodels are established for different working conditions. To improve the generalization ability of the model, a differential evolution algorithm is developed to optimize the parameters and the weights of the LS-SVM submodels. Finally, the multilevel prediction model for the CCR is established by aggregating LS-SVM submodels. The simulation results based on actual run data demonstrate that the prediction accuracy of the multilevel prediction model is higher than that of a backpropagation neural network model and an FCM-LSSVM model, and the proposed modeling method satisfies the requirements of the actual sintering production.