首页 | 官方网站   微博 | 高级检索  
     

基于神经网络的结构损伤诊断的探讨
引用本文:阿肯江·托呼提,张长江,严跃成.基于神经网络的结构损伤诊断的探讨[J].新疆大学学报(理工版),2006,23(3):355-361.
作者姓名:阿肯江·托呼提  张长江  严跃成
作者单位:新疆大学建筑工程学院 新疆乌鲁木齐830008
摘    要:利用神经网络技术对框架结构的损伤位置和损伤程度进行了研究,用振动模态分析技术对组合结构、框架结构的动力参数进行计算,得出组合结构的各种单损伤情况和双损伤情况的各种组合的频率、竖向位移、水平位移变化数据,将这些数据和损伤位置、损伤程度之间的映射关系作为网络训练样本,用神经网络技术进行数值模拟,提出了修正隐含层节点数公式和优选振型向量的方法.说明了用低阶振型和经优选后的振型向量结合的方法进行组合结构、框架结构损伤诊断的可行性,并且用这个方法成功地解决对称结构难于判断对称损伤的难题.

关 键 词:损伤识别  频率  神经网络  对称结构  隐含层节点数  组合结构
文章编号:1000-2839(2006)03-0355-07
收稿时间:2005-11-06
修稿时间:2005年11月6日

Damage Diagnosis of Structure Based on Neural Network
Akenjiang TUOHUTI,ZHANG Chang-jiang,YAN Yue-cheng.Damage Diagnosis of Structure Based on Neural Network[J].Journal of Xinjiang University(Science & Engineering),2006,23(3):355-361.
Authors:Akenjiang TUOHUTI  ZHANG Chang-jiang  YAN Yue-cheng
Affiliation:College of Architecture Engineering, Xinjiang University, Urumqi, Xinjiang 830008, China
Abstract:In this paper,neural network technology is used to solve the position and the extent of of damaged frame structure.Vibrate modal analysis technology is used to calculate the dynamic parameter of frame structure and combined structure,draw a series of frequency and vertical displacement and horizontal displacement,the relation of this number is used to train as network sample.Use this number to numerically simulated all kinds of one damage identify and two damage identify in a combined structure and a frame structure,draw some conclusion,how to choose the best number of implication layers and how to optimize mode of vibration vector approach and it was feasible that use low order frequency and certain mode of vibration or their derived quantity incorporation came by way of structural impair position,extent of damage.
Keywords:damage identify  frequency  neural network  symmetrical structure  the number of implication layer  combine structure
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号