多层及高层框架结构地震损伤诊断的神经网络方法

本文提出了强震后多层及高层框架结构地震损伤诊断的神经网络方法。文中在提出有结点损伤的梁柱有限元刚度矩阵的基础上，建立了有结点损伤框架结构的有限元模型。通过完好结构和有损伤结构的有限元分析，获取二者应变模态差值作为损伤标识量，并输入径向基（RBF）神经网络进行训练，得到了框架结构结点损伤诊断的神经网络系统。数值仿真分析结果表明，此神经网络可以对多层及高层框架结构结点各种程度的损伤做出成功诊断。

Seismic damage diagnosis of frame structure using artificial neural network

in this paper, a new method to diagnose the seismic damage of frame structure using artificial neural network is presented. Based on establishing the stiffness matrix of the columns and beams with joint damage, the finite element model of the damaged frame structure is set up. The strain mode is acquired by analysing the finite element model of the intact and damaged structure. The difference of the strain mode between the intact and damaged structure is chosen as a damage indicator and is used to train the RBF neural network, which can diagnose the damage of frame structure. The result of the numerical computation shows that the RBF network can diagnose the seismic damage of frame structure successfully.