Rolling Mill Based On Neural Network Shape And Gauge Decoupling Control

Shape and thickness are two most important factors of strip quality. The AGC system can guarantee the precision of the vertical thickness and the AFC can guarantee the precision of horizontal thickness. But there exits a coupling phenomenon between the strip thickness, which is controlled by AGC system, and strip shape which is controlled by AFC system. In conventional control, these interactions are omitted. With the requirement of further enhancement of product quality, this cannot be neglected any more, so, decoupling control of strip shape and thickness is urgent.On the basis of analyze of coupling system, this dissertation adopts the method of computer simulation. The chief research work of the decoupling system is as follows:In the aspect of modeling, this dissertation sets up the mathematics model of the coupling system completely and systematically on the theoretical analyzes of system.In the aspect of system control, this dissertation use compensator design a decoupling system first because this model is ease to understand and reliable in engineering. The results of system simulation show this decoupling control system can achieve the control demands.Mainly, his dissertation uses intelligence control to solve the coupling of strip thickness and shape. The results of system simulation show this neural decoupling control system can achieve the control goal.