基于广义局部曲率模态信息熵和BP神经网络的结构损伤识别方法

兰州理工大学学报 ›› 2021, Vol. 47 ›› Issue (1): 129-135.

基于广义局部曲率模态信息熵和BP神经网络的结构损伤识别方法

项长生^1,2, 原子¹, 周宇³

1.兰州理工大学土木工程学院, 甘肃兰州 730050;
2.兰州理工大学西部土木工程防灾减灾教育部工程研究中心, 甘肃兰州 730050;
3.安徽建筑大学土木工程学院, 安徽合肥 230601

收稿日期:2019-11-30 出版日期:2021-02-28 发布日期:2021-03-11
作者简介:项长生(1976-),男,安徽安庆人,硕士,副教授.
基金资助:
国家自然科学基金(51868045)

Structural damage identification method based on local generalized curvature modal information entropy and BP neural network

XIANG Chang-sheng^1,2, YUAN Zi¹, ZHOU Yu³

1. College of Civil Engineering, Lanzhou Univ. of Tech., Lanzhou 730050, China;
2. Western Center of Disaster Mitigation in Civil Engineering of Ministry of Education, Lanzhou Univ. of Tech., Lanzhou 730050, China;
3. School of Civil Engineering, Anhui University of Architecture, Hefei 230601, China

Received:2019-11-30 Online:2021-02-28 Published:2021-03-11

摘要/Abstract

摘要： 针对曲率模态对振型节点较不敏感且无法定量估计损伤的问题,在广义局部信息熵的基础上引入曲率模态,推导出广义局部曲率模态信息熵的公式,并建立相应的损伤指标.利用有限元软件Midas civil建立一简支梁桥损伤模型,提取并处理该简支梁的动力参数,将一阶曲率模态和广义局部曲率模态信息熵分别作为神经网络的输入参数,对损伤进行识别并对比两种参数的识别结果,以此来研究测点数量对指标精确度的影响.研究结果表明:广义局部曲率模态信息熵作为神经网络的输入参数能较好地定位并定量损伤,在靠近振型节点处指标的识别精度高于曲率模态,当测点数量为33时,识别精度最高.

关键词: 广义局部曲率模态信息熵, 曲率模态, bp神经网络, 损伤定位, 损伤定量

Abstract: In view of the fact that curvature mode is less sensitive to vibration mode nodes and can not estimate damage quantitatively, a formula of mode information entropy for generalized local curvature is deduced, based on generalized local information entropy and by introducing conception of the curvature model. The corresponding damage index is established therefore. A simple supported beam is analyzed by using the finite element software Midas civil. The dynamic parameters of the simply supported beam are extracted and processed. The first-order curvature mode and the local generalized curvature modal information entropy are used respectively as two input parameters for the neural network to identify the damage. The conclusion of two parameters having been identified are compared with each other so as to study the influence of the number of measure points on the identification accuracy of each index. The results show that the generalized local curvature mode information entropy as the input parameter of the neural network can locate and quantify the damage much better. The accuracy of indicators near the vibration mode node is higher than the result by the method of the curvature mode. When the number of measuring points takes 33, the identification accuracy becomes the highest.

Key words: generalized local curvature modal information entropy, curvature modal, BP neural network, damage location, damage quantification

中图分类号:

TU312
U447

项长生, 原子, 周宇. 基于广义局部曲率模态信息熵和BP神经网络的结构损伤识别方法[J]. 兰州理工大学学报, 2021, 47(1): 129-135.

XIANG Chang-sheng, YUAN Zi, ZHOU Yu. Structural damage identification method based on local generalized curvature modal information entropy and BP neural network[J]. Journal of Lanzhou University of Technology, 2021, 47(1): 129-135.

参考文献

[1] 宗周红,褚福鹏,牛杰.基于响应面模型修正的桥梁结构损伤识别方法[J].土木工程学报,2013,46(2):115-122.
[2] ZHOU Y,DI S,XIANG C,et al.Damage detection for simply supported bridge with bending fuzzy stiffness consideration[J].Journal of Shanghai Jiaotong University (Science),2018,23(2):308-319.
[3] 周宇,狄生奎,项长生,等.基于柔度曲率梯度的刚架拱桥损伤识别[J].应用基础与工程科学学报,2017,25(6):1211-1221.
[4] LI H,BAO Y,OU J,et al.Structural damage identification based on integration of information fusion and shannon entropy[J].Mechanical Systems and Signal Processing,2008,22(6):1427-1440.
[5] 罗德河,郑东健,胡德华,等.基于传递率函数与信息熵的混凝土坝裂缝损伤识别指标[J].南水北调与水利科技,2017,15(5):141-145.
[6] 孙增寿,韩建刚,任伟新.基于曲率模态和小波变换的结构损伤位置识别[J].地震工程与工程振动,2005,25(4):44-49.
[7] 袁旭东,周晶,黄梅.基于静力位移及频率的结构损伤识别神经网络方法[J].哈尔滨工业大学学报,2005,37(4):58-60.
[8] YANG Z,CHEN X,JIANG Y,et al.Generalised local entropy analysis for crack detection in beam-like structures[J].Nondestructive Testing and Evaluation,2014,29(2):133-153.
[9] 毛福超. 基于广义局部信息熵的钢筋混凝土梁损伤识别研究 [D].兰州:兰州理工大学,2017.
[10] 刘国华,吴志根.引入信息熵理论的砼结构损伤动力识别新思路[J].振动与冲击,2011,30(6):162-171.
[11] KAMINSKI P C.The approximate location of damage through the analysis of natural frequencies with artificial neural networks[J].Journal of Process Mechanical Engineering,1995,209:117-123.
[12] 张开银,孙峙华,邹晓军,等.桥梁结构损伤识别的曲率模态技术[J].武汉理工大学学报,2004,28(6):855-858.
[13] 李忠献,齐怀展,朱劲松.基于模态曲率法的大跨度斜拉桥损伤识别[J].地震工程与工程振动,2007,27(4):122-126.
[14] 杜永峰,张冬兵.曲率模态和神经网络在损伤识别中的应用[J].公路交通科技,2007(11):77-80.
[15] 郑猛,陈恩利,史玉红.基于曲率模态和神经网络的结构损伤诊断[J].石家庄铁道学院学报,2005,18(1):29-32.
[16] 李兆,唐雪松,陈星烨.基于曲率模态和神经网络的分步损伤识别法及其在桥梁结构中的应用[J].长沙理工大学学报(自然科学版),2008,5(2):32-37.
[17] PANDEY A K,BISWAS M,SAMMAN M M,et al.Damage detection from changes in curvature mode shapes[J].Journal of Sound and Vibration,1991,145(2):321-332.
[18] SHANNON C E.A mathematical theory of communication[J].Bell System Technical Journal,1948,27(3):379-423.
[19] 张浩,韩乃杰,刘凯,等.基于宏观应变测试的受弯构件微损伤识别研究[J].重庆交通大学学报(自然科学版),2018,37(11):15-20.
[20] 李宏男,高东伟,伊廷华.土木工程结构健康监测系统的研究状况与进展[J].力学进展,2008,38(2):151-166.
[21] 王嵘冰,徐红艳,李波,等.BP神经网络隐含层节点数确定方法研究[J].计算机技术与发展,2018,28(4):31-35.