Parameter Fusion Modeling Method for Hot Strip Rolling Process and its Application

According to the production characteristics of different hot-rolled strip products and the characteristics of products, based on the in-depth study of the hot-rolled mechanism of strip, a modeling strategy combining multiple intelligent modeling methods with mechanism model is proposed, and a high-performance model of hot-rolled process parameters of strip is established. For common steel, neural network is used to modify the mechanism model adaptively; for special rule steel, fuzzy rule algorithm is added based on the output of neural network; for small batch of special steel, case-based reasoning method is added to modify the mechanism model. Finally, the parameter fusion modeling method of hot strip rolling process is formed and successfully applied to 1549mm hot rolling production line of Taiyuan Iron and Steel Group Co., Ltd (TISCO). The practical application results show that the fusion modeling method improves the modeling accuracy, speed and adaptive ability, and achieves good economic benefits.