Application of Deep Belief Network in Prediction of Secondary Chemical Components of Sinter

Sintering process is one of the important processes of steel smelting. Sintering is the main raw material for blast furnace ironmaking. Predicting the chemical composition of sinter is the key method to optimize the sintering process. At present, the chemical composition prediction algorithms of sinter are all the main chemical components for sinter. With the improvement of algorithms and computer technology, the prediction of the main chemical composition of sinter has achieved more and more good results and it is difficult to obtain a substantial increase. Therefore, Prediction of secondary chemical composition of sinter is of great importance to optimize the quality of sinter. In view of the above problems, Deep Belief Network algorithm is applied to predict the secondary chemical composition of sinter, and a prediction model of secondary chemical composition of sinter with Deep Belief Network is established by deeply analyzing the technology mechanism and characteristics of the sintering process. First, the Deep Belief Network structure and parameters are designed. The unsupervised greedy algorithm is used to pre-train the model, and the BP neural network is used to supervise the inverse trimming weights to optimize the whole model. Finally, simulation and verification. The simulation results show that the error between the predicted value and the actual value is small and the prediction accuracy is high, which shows the validity of Deep Belief Network in predicting the secondary chemical composition of sinter.