A novel single-neuron smith-predictive AGC system

Thickness accuracy is the important quality index of the hot rolling strip. Therefore, the automatic gauge control (AGC) is always the important function that the control system of hot rolling trip completes firstly. Due to the complicated rolling process, nonlinearity and time-variation of the controlled system, the simple AGC system can´t obtain desirable effect. In the paper, a new single neuron-Smith prediction control system of AGC is presented. Stability analysis and the simulated experiments on this system are carried out. The simulated results indicate that the new system has the advantages of quick response, low overshooting, small steady-state error and good anti-interference ability. New control system improves the control performance of system with time delay and the accuracy of strip thickness.