论文部分内容阅读
以4200轧机轧制钢板的实测数据为基础,利用Matlab神经网络工具箱,建立了中厚板轧机宽展的RBF神经网络预测模型。通过分析宽展的影响因素,结合传统的数学模型,确立了网络的输入层参数,并对宽度系数spread进行试验调整,确定了最佳的网络结构形式,提高了模型的预测精度。通过实例比较了RBF模型与BP模型的预测效果,并且分析了不同参数下RBF神经网络逼近精度。结果表明,RBF神经网络模型有较好的收敛速度和预测精度,能更好地适用于中厚板轧机宽展模型。
Based on the measured data of 4200 rolling mill plate, a RBF neural network prediction model of wide rolling mill was established by using Matlab neural network toolbox. By analyzing the influencing factors of wide spread and combining the traditional mathematical models, the input layer parameters of the network are established, and the spread of the width coefficient is experimentally adjusted to determine the best network structure and improve the prediction accuracy of the model. The prediction results of RBF model and BP model are compared by examples, and the approximation accuracy of RBF neural network with different parameters is analyzed. The results show that the RBF neural network model has better convergence speed and prediction accuracy, and can be better applied to the wide-gauge plate mill model.