A New Approach for Modeling Energy Information and Investigation of the Energy Consumption of Furnaces with Various Steel Grades

Accurate prediction of furnace energy consumption in the steel industry by considering different grades has received less attention, despite its importance in energy optimization. In this study, the modeling of steel furnace energy consumption information flow using an artificial neural network (ANN) was investigated for four steel-grade furnaces (A283, KA37, S235, S355). Experimental data, including furnace volume, heating time, average slab temperature, average furnace temperature, and fuel energy consumption, were collected from an industrial unit. Independent variables (furnace volume, heating time, slab and furnace temperatures) and dependent variable (fuel energy consumption) were modeled using a feedforward multilayer neural network (MLP) and presented as a function. The results showed that the performance of the ANN was poor for regression coefficients (R) less than 0.5, acceptable in the range of 0.5 to 0.7, and very accurate for R > 0.8. The agreement between the numerical and experimental results confirms the validity of this method and demonstrates its application in optimizing fuel consumption and reducing waste.