Lime Addition Calculation Model of Converter Based on Mechanism and Back Propagation Neural Network
Aiming at the smelting conditions of converter with high scrap ratio,with the calculation of converter alkalinty as theory basis,and the primary and secondary data collection system of steel plant converter as the data source,and by using the designing data screening rules,data prepocessing and calculating loss funiction to improve the calculation accuracy,the lime addition calculation model based on mechanism and back propagation neural network with structure of 15-3-50-5-1 was established.The calculated results were compared with the real values and the calculated results of the traditional model.The results show that the average relative error between the lime added quality predicted by the mechanism+BP neural network model and the real value was only 10.7%,which was about 7%lower than the relative error between traditional calculation model calculated result and real value,indicating that the new model had small error and high precision in predicting the lime addition.
BP neural networkbig dataconverterlimeintelligent joining model
万方数据
维普
