A Method for Predicting the Temperature of Steel Billet Coming Out of Soaking Furnace Based on Deep Belief Neural Network

Accurately predicting the billet discharge temperature through the process parameters of the even-heat furnace is the key to improving the billet rolling quality and reducing the energy consumption of the even-heat furnace, and it is also a major problem faced by the academia and the industry at present. Aiming at this problem, combined with the powerful feature learning and representation capability of deep belief neural network, a deep belief neural network-based billet temperature prediction method for even-heat furnace is proposed to realize the complex nonlinear mapping between the process parameters of even-heat furnace and the billet temperature, and the accuracy and superiority of the algorithms proposed in this paper are verified through the comparison and analysis of the actual production data and existing prediction methods.