Tension Control Of Galvanizing Line Of Cold Rolled Sheet

Nowadays, galvanization is continuously produced with high speed. It is necessary for the strip in the constant tension to make the production work continuously. The tension control is important control target in relation to productivity and quality. However, there is a great couple between tension and speed on the galvanizing line, which have a serious effect on tension and speed.The thesis's background is the continuous galvanizing line of Benxi Steel Cold Rolled Sheet Co., Ltd.. The tension control system of the line is analysed and mathematic model of tension system of galvanizing line is set up. Then inverse system method, neural network inverse system method and internal model control based on inverse system method are proposed to control tension.For the couple between tension and speed in tension control of galvanizing line, inverse system method is utilized to control the tension system. The inverse system of tension system is gotten by analytic method. Pseudo-linear system is formed by joining the inverse system to the original system. Then linear controller is designed to control the pseudo-linear system. Simulation show the inverse system method can realize the decoupling control between tension and speed.For the shortcut that inverse system method need mathematic model, neural network inverse system method is utilized to control the tension of galvanizing line. The neural network inverse system of original tension system is gotten by stimulating the original tension system and collecting needed signals, which replace the analytic inverse system to control tension. Simulation show the neural network inverse system can realize decoupling control between tension and speed and doesn't depend on the mathematic model of original system.For the possibility of modeling error when neural network is used to set up the inverse system, the internal model control based on neural network inverse system method is proposed to control the tension system.Internal model control can eliminate the modeling error and disturbance by theoretical analysis.Simulation also show that internal model control combined the neural network inverse system can control tension better.