Neural Network And Robust Used In Controlling The Thinkness Of Rolling Mill

With the development of science, the quantity and quality are the first targets of strip production. But in analyzing the dynamic characteristics of rolling mill, we find that the system exists nonlinear uncertainty and slow time-variation, which are caused by nonlinear mode and random disturbance. This leads to the approximation of the practical system of the model and parameter and it is difficult to construct the accurate mathematical model. Therefore, the routine controlling method isn't used any longer. The control method of neural network and robust is brought forward according to the problems as above. The object is identified by linear model and the neural network's learning ability and nonlinear map ability for the control method. The method applies neural network's learning ability and nonlinear map ability to the design of controller in order to obtain the precise control of output for the uncertain and nonlinear system. Using the robust feedback controller to guarantee the course continuity and the system stability. The influence that is caused by external disturbance has been avoided. This method can not only make neural network and routine controller learn from other's virtues to offset their weakness and exert their good qualities, but also reduce the effect of identified precision to the characteristic of the control system. This paper focuses on the feasibility and stability of the neural network and robust controller is applied to position inner loop. The study of simulation indicates that the effectiveness of the controller and the feasibility of the whole project. Actually, the arithmetic is greatly affecting the precision of controller and identification. We put forward an improved method for the disadvantage of local point with which the study ratio can adjust according to the error camber and enhance the rate of the constringency and the accuracy of study of the neural network. Now the method is going to be test in automatic control of computer of steelworks of tan-shan.