在低速轧钢时，由于测厚仪测量值的长时间滞后，传统的后馈控制效果不好。
At a low steel rolling speed, the traditional AGC feedback control is not very good due to a long time delay caused by the thickness measurement.
创造性思维的培育，应实现从静态思维向动态思维、从单层面思维向多层面思维、从后馈思维向超前思维的转变。
The cultivation of creative thought should realize the transformation from static state thinking to dynamic state thinking; from single layer thinking to multiple layer thinking.
提出一种基于参考误差的投影算法来训练网络权值，训练后网络输出能逼近期望的前馈力矩，并从理论上证明跟踪误差的收敛性。
The neural network trained by projection algorithm with the reference error is used to approximate the desired feedforward compensation. Convergence of the tracking error is proved.
应用推荐