Forecast for the thermal state index of pellet mine based on artificial neural network

Abstract

Three models of artificial neural network were established to predict the thermal state indexes (RDI, RI, RSI) of iron ore pellets. Initial input factors of the networks were found according to the pellet mine theory. Then sensitivity analysis was used to quantify the importance of each input variable and reduce the networks' input dimensionality. At last, minimum sets of input factors of networks were found to improve the accuracy of network prediction. Simulation results show that the prediction models meet the actual engineering application requirement.