论文部分内容阅读
热连轧过程中，为了提高轧制力预设定精度，提出一种新的修改轧制力模型参数的方法·利用ＢＰ 神经网络对以往的大量生产数据进行训练、预测·对ＢＰ 神经网络的预测结果利用最小二乘法，回归出轧制力模型中的温度相关系数ｍ １ 和变形速度相关系数ｍ ３·现场生产实验证明，应用修改后的轧制力模型系数，提高了轧制力预设定精度，从而使头部厚度精度有较大提高·对于象本溪钢铁公司热连轧厂这样的老企业，这种新方法更具有在线应用的可行性·
In the process of hot strip rolling, in order to improve the presetting accuracy of the rolling force, a new method to modify the parameters of the rolling force model is proposed. The BP neural network is used to train and predict the mass production data in the past. Prediction results Using the least squares method, the temperature dependence coefficient m 1 and the deformation velocity correlation coefficient m 3 in the rolling force model are regressed. · The field production experiments show that the revised rolling force model coefficients improve the rolling force pre-setting The accuracy of the head, so that the head thickness precision has greatly improved for such as the Benxi Iron and Steel Company hot rolling plant such an old business, this new method is more feasible for online applications ·