The Research On Intelligent Control Method Of Flatness Fof Hydraulic Bending Roll

How to improve the quality of the plate strip is an important subject confronted by steel manufacturers. At present, Hydraulic Bending Roll is the most common and effective method of shape control, which has been widely applied in the shape control of mill. Therefore, it has become a necessary requirement for the development of shape control technique to research the shape control theory completely and deeply, and establish the precise systematical model of Hydraulic Bending Roll shape control.In this article, the author studied systematically the theory and basic component of the Hydraulic Bending Roll control system, and established the mathematical model of hydraulic bending force control system. Additionally, the author analyzed explicitly the influence of characteristics of system, steady-state error and interference to the system. And then the author established the dynamic simulation model of Hydraulic Bending Roll control system through using Simulink tool of Mablab software, and simulated the hydraulic and bending force, and discussed the problems of influence of component parameters to the characteristics of the system.Now most of the mill Hydraulic Bending Roll control system adopted the traditional PID control algorithm, however, practice proved that it is difficult for the traditional PID control to satisfy the requirement of high precision control. In this article, the author firstly used adaptive Online Genetic Algorithm to optimize the PID parameters, and then used the CMAC neural network and PID composite control strategy, to further improve the tracking performance of Hydraulic Bending Roll Loop to the hydraulic and bending force. Comparing with the traditional PID, the performance indicator of the system has improved greatly; meanwhile it accelerated the response speed of system, and increased the stability of system, gaining a perfect control effect.In order to satisfy the demand of consumers to the high quality of strip products, steel manufacturers need to increase the shape control accuracy furthermore. However, during the practical process of rolling, the control subject has the characteristics of nonlinear, multivariate and strong coupling, so it is difficult for them to use the traditional mathematical method to establish precise shape model. In view of the above problem, the author considered the ANFIS toolbox as a platform, using the intellectual model-building method ANIFS (Adaptive Neuro-Fuzzy Inference System) to realize the fuzzy modeling of bending force and quadratic shape, in addition, the author used lots of rolling data from practice to test the FIS model,as a result, it indicated that the model that using the method of ANFIS to establish has better adaptability, anti-interference and control precision.