论文部分内容阅读
为了改善国内某钢铁厂炉卷轧机的轧制力模型的预报精度,提出将结合热模拟实验建立的传统轧制力模型计算值作为Elman神经网络的一个输入项,将传统数学模型预报的轧制力与实测轧制力的相对误差作为此神经网络输出项的方式构建网络模型,通过大量的在线数据分析,这种将神经网络与传统数学模型相结合的方法明显地改善了轧制力的预报精度。该神经网络模型可为以轧制力为主要控制目标的炉卷轧机的过程自动化系统提供可靠的模型参数。
In order to improve the prediction accuracy of the rolling force model of a domestic steel mill, it is proposed to combine the calculated values of the traditional rolling force model established by the thermal simulation experiment as an input term of the Elman neural network to predict the rolling force predicted by the traditional mathematical model Force and measured rolling force relative error as the output of the neural network to build a network model, through a large number of online data analysis, the neural network and the traditional mathematical model combined significantly improved the prediction of rolling force Accuracy. The neural network model provides reliable model parameters for a process automation system of a rolling mill with rolling force as the main control target.