应用BP神经网络的单机架可逆式冷轧机弯辊力预设定

针对单机架可逆式冷轧机生产过程中，由于不同热轧生产厂的热轧板形原料性能、板形差别较大、从而造成弯辊力预设定值和实际指标相差较大的问题，利用现场生产工艺数据丰富，采用人工智能BP神经网络的方法，通过大量数据输入进行训练，大大的提高了生产过程中弯辊力预设定值的准确性。

Bending Force Presetting with BP Neural Networks for Single Stand Reversing Cold Mill

During the manufacturing process of single stand reversing cold mill, due to the hot rolled plates produced by different hot rolling plant are different in raw material performance and plate profile, there are big difference between bending force presetting value and the actual value. Abundant manufacture technology datum has been used, artificial intelligence BP neural network method has been adopted, and training has been performed by inputting large quantities of datum, which greatly improve the accuracy of bending force presetting value during the manufacturing.