论文部分内容阅读
采用BP神经网络方法代替传统的数学模型预测精轧机组轧制带钢的宽度变化，以提高热轧带钢的宽度精度，并进行了不同网络结构的比较研究。结果表明，BP神经网络方法优于传统数学模型方法，其预测值与实测值的标准差减小了51．9％。
The BP neural network method was used to replace the traditional mathematical model to predict the width variation of the strip in the finishing mill and to improve the width precision of the hot strip and to compare the different network structures. The results show that the BP neural network method is superior to the traditional mathematical model method, and the standard deviation of the predicted value and the measured value is reduced by 51.9%.