The Research On Crown Closed-Loop Control Mathematical Model

With the development of morden industry,user's requirement for high strip shape quality is increasing day by day. Crown is one of the important criteria of strip shape quality evaluation. In the rolling, the crown closed-loop feedback control system is an important means of control compliance to ensure that the crown of along the length of. the strip is qualified. Therefore, accuracy of the crown closed-loop control model directly impact on the quality of the steel plate shape. Plate crown and shape control system is a complex industrial control system,and the rolling elements have stronger non-linear, time-varying and mutual coupling, so that the mathematical model established according to the traditional theory is difficult to achieve the desired control effect.So, the the research that solving crown closed-loop control model in my strip production has very important significance of academic.This paper makes a crown adjustment model based BP neural network. By using the model, research the relationship between bending force adjustment and the corresponding crown adjustment at different rolling conditions,and the transfer coefficient of bending force on the crown,what expressing the relationship, implicit in the neural network model. Therefore, in the election of the network model's input parameters,consider the factors of transfer coefficient of bending force on the crown.Secondly, use the train samples and test samples to train and test the network, the calculated values of crown adjustment neural network model is very similar to measured values by analysis, and show that crown adjustment BP neural network model have high prediction accuracy.And on this basis, establish crown closed-loop feedback control model based on BP neural network. Crown gauge installed at finishing export, quickly detecte the strip crown actual value in real-time.Under its crown bias with the target crown in accordance, adjust the bending force of the frames to made a steel crown coincided with the target. It is practically applicable to domestic steel 1780mm hot rolling production line,collecting field data and analyzing practical application system. The results show that the application of convexity closed-loop feedback control system based on BP neural network has significantly improved crown shape quality, and prove the correctness and practicality of the model.