论文部分内容阅读
讨论一种动态神经网络──Elman回归神经网络的结构和算法．基于这一网络结构提出了非线性时变工业对象─-直流电弧的神经网络建模方法，并与用其他方法为对象建立的模型进行了比较，结果证明回归网络模型能够很好地适配该工业对象，显示了动态神经网络在工业对象建模中的良好应用前景．
Discusses a dynamic neural network ─ ─ Elman regression neural network structure and algorithm. Based on this network structure, a neural network modeling method of nonlinear time-varying industrial object DC arc is proposed and compared with the model established by other methods. The results show that the regression network model can well fit the Industrial objects, showing the dynamic application of dynamic neural network in the industrial object modeling.