Prediction Of Steel Temperature In Ladle Furnace Based On Neural Network

Ladle furnace have become an integral part of modern refining processes.The use of the LF not only improves the quality of the steel,but also increases the variety of steel.The accurate prediction of the temperature of the molten steel in the LF has great significance for the continuous casting production and the quality of the cast steel connected to the next process,and it has favorable effects in terms of cost control and resource conservation.The prediction of molten steel temperature has always been a topic of concern to many scholars.Aiming at the problems of molten steel temperature prediction,this paper analyzes the smelting process of LF and establishes the mechanism model of LF molten steel temperature prediction and the intelligent model with extreme learning machine as the core algorithm.In the mechanism model,some parameters cannot be accurately obtained,and the prediction accuracy is difficult to guarantee.Due to lack of mechanism guidance and excessive reliance on data,the generalization performance of intelligent models is poor.In view of the above problems,by analyzing the advantages and disadvantages of the mechanism model and the intelligent model,the mechanism model is combined with the modeling idea of the intelligent model,the intelligent method is used to determine the parameters that cannot be accurately obtained in the mechanism model,and then the mechanism model is used to predict the molten steel temperature.That is to establish a hybrid model of temperature prediction,and use the improved artificial bee colony algorithm to optimize the intelligent model,effectively improving the accuracy of the intelligent model.Aiming at the problem of fault error in actual data,this paper uses CED-based gross error detection method and soft measurement modeling method to detect the gross error data.This method can effectively find the gross error in the data set and Get an accurate soft measurement model.The data simulation experiment of a 100-ton LF refining furnace from a domestic steel mill shows that,The molten steel temperature prediction model established in this paper can meet the production demand,and can accurately predict the molten steel temperature,which can effectively guide the accurate control of temperature in actual production.