Coke Quality Prediction Model Based on DE-RBF Neural Network

摘要

In view of the difficulty in on-line detection of coke quality, which seriously affects the automation of coking production process, a coke quality prediction model based on differential evolution algorithm optimized RBF neural network (DE-RBF) is proposed. Firstly, the factors influencing coke quality are determined and the input dimension of neural network is reduced by principal component analysis. Secondly, the weight of RBF neural network is optimized by differential evolution algorithm, and then the improved algorithm is used to predict coke quality. Finally, the model is verified by experimental simulation. The results show that the model is of higher prediction ability and practicability.