半无头轧制工艺下的精轧宽展预测

PREDICTION OF LATERAL SPREAD IN FINISHING STANDS OF SEMI-ENDLESS ROLLING

为了提高半无头轧制精轧宽展的预测精度,采用BP神经网络方法代替传统数学模型进行宽展预测,并采用改进的BP神经网络算法提高学习速度,与传统数学模型方法得到的预测精度作比较,结果表明采用此方法预测精度更优。

In order to improve the predictable precision of lateral spread in finishing stands of semi-endless rolling, the BP neural networks is used to predict the lateral spread instead of the traditional mathematical model. An improved algorithm of BP neural network is used to improve learning speed. The results show that the predictable precision of BP neural networks is better than that of traditional mathematical models.