Study Of The Strip Thickness Intelligent Control Methods

With the development of science, the quantity and the quality have gradually become the first target of strip production. Thickness precision is one of the most important indexes for evaluating equality of hot strip rolling production. It has more practical value for the current research in strip thickness precision control.Based on the deeply understanding of the basic concepts and principles of conduct of the Gauge Control system, this paper designs the Neural Network Predictive Control strategy of hot rolling mill thickness gauge feedback AGC And through simulation, it analyses and validates the control strategy adopted by the performance.First this paper studies the gauge control system on the basic principles deeply. Because of the process, the thickness gauge feedback AGC has the pure time delay, which causes the dynamic performance and stability of system dropping; it establishes the model of system, and the thickness gauge feedback AGC and Smith predictor separately. Because the thickness gauge feedback AGC belongs to feedback system, the control action always falls behind the interruption, it can not achieve good control impact; Smith predictor can eliminate the influence of time delay to the stability of thickness gauge feedback AGC, reduce the overshoot when the model is exact. But Smith predictor slow down the adjustment of time, so it takes RBF neural network and speed up the adjustment of time obviously. But Smith predictor depends on the exact model, the performance of the system will greatly reduce, and the stability of the system will be affected when the model is not exact. Therefore it must look for other ways to control the system to achieve better performance.Based on the analysis above, in this paper neural networks predictive control strategy is imported. First it introduces the basic principles of neural networks and predictive control, then introduces the basic characters, algorithm principles and specific steps of neural network predictive control/Strategy, and designs the neural network predictive control strategy of the thickness gauge feedback AGC systems and makes simulation. The simulation proves that the applications of neural network predictive control have made a very good effect.The experiment proved that the design of the neural network predictive control can eliminate the influence of time delay to the stability of thickness gauge feedback AGC, and increase the performance of the gauge control system; it has strong application and a certain value.