基于撓度的鐵路雙線簡支鋼桁梁橋桿件損傷程度識別研究
發(fā)布時間:2018-08-01 16:13
【摘要】:以梁橋節(jié)點(diǎn)最大位移改變率作為損傷程度傷識別指標(biāo),分別采用廣義回歸神經(jīng)網(wǎng)絡(luò)(GRNN)算法和ε-支持向量回歸機(jī)(ε-SVR)算法,進(jìn)行損傷程度識別研究。通過對一座鐵路雙線簡支鋼桁梁橋某桿件的損傷程度識別研究發(fā)現(xiàn):(1)GRNN損傷程度識別模型具有一定的抗噪能力,不具有泛化性。(2)SVR損傷程度識別模型具有很強(qiáng)的抗噪能力和很好的泛化性。(3)以橋梁節(jié)點(diǎn)最大位移改變率作為損傷程度識別指標(biāo)時,數(shù)據(jù)回歸算法不能采用GRNN算法,應(yīng)采用ε-SVR算法。
[Abstract]:The maximum displacement change rate of beam bridge joint is used as the index of damage degree identification, and the generalized regression neural network (GRNN) algorithm and 蔚-support vector regression machine (蔚 -SVR) algorithm are used to study the damage degree identification. Based on the research on the damage degree of a member of a steel truss bridge with two railway lines simply supported, it is found that: (1) the GRNN damage degree recognition model has a certain anti-noise ability. (2) SVR damage recognition model has strong anti-noise ability and good generalization. (3) when the maximum displacement change rate of bridge node is taken as the index of damage degree identification, the data regression algorithm can not adopt GRNN algorithm, but 蔚 -SVR algorithm should be used.
【作者單位】: 河北省電力勘測設(shè)計(jì)研究院土建部;石家莊鐵道大學(xué)工程力學(xué)系;石家莊鐵道大學(xué)大型結(jié)構(gòu)健康診斷與控制研究所;
【基金】:國家自然科學(xué)基金(51278315) 河北省自然科學(xué)基金(E2012210061) 河北省教育廳基金(Z2013034)
【分類號】:U441.4
[Abstract]:The maximum displacement change rate of beam bridge joint is used as the index of damage degree identification, and the generalized regression neural network (GRNN) algorithm and 蔚-support vector regression machine (蔚 -SVR) algorithm are used to study the damage degree identification. Based on the research on the damage degree of a member of a steel truss bridge with two railway lines simply supported, it is found that: (1) the GRNN damage degree recognition model has a certain anti-noise ability. (2) SVR damage recognition model has strong anti-noise ability and good generalization. (3) when the maximum displacement change rate of bridge node is taken as the index of damage degree identification, the data regression algorithm can not adopt GRNN algorithm, but 蔚 -SVR algorithm should be used.
【作者單位】: 河北省電力勘測設(shè)計(jì)研究院土建部;石家莊鐵道大學(xué)工程力學(xué)系;石家莊鐵道大學(xué)大型結(jié)構(gòu)健康診斷與控制研究所;
【基金】:國家自然科學(xué)基金(51278315) 河北省自然科學(xué)基金(E2012210061) 河北省教育廳基金(Z2013034)
【分類號】:U441.4
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 劉濤;李愛群;丁幼亮;;基于小波包能量譜的大跨斜拉橋拉索損傷預(yù)警方法[J];東南大學(xué)學(xué)報(bào)(自然科學(xué)版);2007年02期
2 姜紹飛;張帥;;基于模糊神經(jīng)網(wǎng)絡(luò)的數(shù)據(jù)融合結(jié)構(gòu)損傷識別方法[J];工程力學(xué);2008年02期
3 朱宏平;余t,
本文編號:2158124
本文鏈接:http://sikaile.net/kejilunwen/jiaotonggongchenglunwen/2158124.html
最近更新
教材專著