中文摘要：

将人工神经网络优化方法引入电器优化设计的研究中,提出动态神经网络的新方法.在充分利用人工神经网络强大的非线形逼近能力和较高效率的优势同时,研究了逼近性能与优化空间、样本采集、计算精度提高等动态有序变化的策略.通过对非线形振荡函数寻优和500kV GIS盆式绝缘子电器产品的结构优化设计,证明文中提出的动态神经网络方法是可行、有效的.为复杂结构、多变量电气产品的优化设计提供了一种新的高效优化设计方法.

英文摘要：

Artificial neural network （ANN） is introduced to optimize the electric apparatus, and a novel dynamic artificial neural network （DANN） has been proposed. Based on the application of the nonlinear approaching ability and high efficiency of the ANN. the strategies of the approaching ability, searching space, sample collection and computation precision have been investigated. The validity and the feasibility of the proposed DANN have been verified by optimizing a nonlinear surge testing function and the disc insulator of a 500kV GIS. Moreover, a novel optimizing method with higher efficiency for complex structure and multivariable cases has been provided.