Strip flatness and gauge PID neural network control based on wavelet network

Strip rolling is a very complicated nonlinear process. Automatic flatness control (AFC) and automatic gauge control (AGC) are nonlinear, interacted and coupled each other. A novel PID neural network control method of AFC-AGC is developed, in which Smith predictor is designed by using two wavelet networks whose structures and parameters are identical. Because inputs of network are independent of time delay order, the sensitivity of dimension for wavelet network is reduced. This method combines PID neural network with Smith predictor constructed by two wavelet networks and it can meet robust request. Simulation results indicate that the system can eliminate periodic disturbance quickly and has better response performance.