热轧轧制力小波神经网络预报模型的研究

Research on Prediction of Rolling Force for Hot Mill Based on Wavelet Neural Network

为了进一步提高热连轧精轧机组轧制力的设定精度,采用小波神经网络建立轧制力预报模型,并采用改进的快速BP算法来训练网络。仿真结果表明:建立的轧制力预报模型的预报值与实际值之间的相对误差在±6%以内,且学习算法收敛速度快。

The wavelet neural networks were used to predict the rolling force in order to improve the preset accuracy of rolling force for the hot rolling mill, by using the improved BP algorithm to train the networks. The simulation results demonstrate that the relative error of the rolling force by wavelet neural network is about ± 6% and the convergence speed of this algorithm is fast.