Article(id=1245389862614975093, tenantId=1146029695717560320, journalId=1241701559352995854, issueId=1245389858412282468, articleNumber=null, orderNo=null, doi=10.13197/j.eeed.2024.0205, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1662307200000, receivedDateStr=2022-09-05, revisedDate=1675526400000, revisedDateStr=2023-02-05, acceptedDate=null, acceptedDateStr=null, onlineDate=1774853681471, onlineDateStr=2026-03-30, pubDate=null, pubDateStr=null, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1774853681471, onlineIssueDateStr=2026-03-30, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1774853681471, creator=13701087609, updateTime=1774853681471, updator=13701087609, issue=Issue{id=1245389858412282468, tenantId=1146029695717560320, journalId=1241701559352995854, year='2024', volume='44', issue='2', pageStart='1', pageEnd='232', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1774853680470, creator=13701087609, updateTime=1774854277127, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1245392361031840387, tenantId=1146029695717560320, journalId=1241701559352995854, issueId=1245389858412282468, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1245392361031840388, tenantId=1146029695717560320, journalId=1241701559352995854, issueId=1245389858412282468, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=38, endPage=49, ext={EN=ArticleExt(id=1245389863399309947, articleId=1245389862614975093, tenantId=1146029695717560320, journalId=1241701559352995854, language=EN, title=Structural damage identification based on self-training semi-supervised neural network, columnId=null, journalTitle=Earthquake Engineering and Engineering Dynamics, columnName=null, runingTitle=null, highlight=null, articleAbstract=

A structural damage recognition framework based on a self-training semi-supervised neural network (SSNN) is proposed to solve the problem of insufficient labeled data in structural damage identification. The framework utilizes the multilayer perceptron (MLP) neural network for semi-supervised training by the self-training method. The data samples with high confidence are selected from the unlabeled data to make pseudo labels, expanding the training set. Normalized frequency change ratio and damage signature index are employed as input features of neural networks to identify structural damage. Firstly, the theory fundamentals of semi-supervised self-training learning are introduced. Secondly, the procedure of structural damage identification based on self-training semi-supervised learning, including neural network construction, damage characteristic extraction, and classifier evaluation, is introduced. Finally, the proposed damage identification method is illustrated by numerical simulation of a spatial truss and experimental data of a three-story frame. The results show that the self-training semi-supervised method can expand the labeled sample data by selecting samples with higher confidence from unlabeled data, thus providing sufficient labeled data for damage identification. Under the insufficient labeled data conditions, the SSNN performs better than MLP. Compared with MLP, SSNN increases the identification accuracy by 4% and 9% under the single and two positions damage locations, respectively.

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Shiqiang QIN, Rui YANG, Sheng SU), CN=ArticleExt(id=1245389874983977961, articleId=1245389862614975093, tenantId=1146029695717560320, journalId=1241701559352995854, language=CN, title=基于自训练半监督神经网络的结构损伤识别, columnId=0, journalTitle=地震工程与工程振动, columnName=, runingTitle=null, highlight=null, articleAbstract=

为解决结构损伤识别中标签样本不足的问题,提出一种基于自训练半监督神经网络(self-training semi-supervised neural networks,SSNN)的结构损伤识别框架,该框架利用自训练半监督方法对多层感知机(multilayer perceptron,MLP)神经网络进行训练,从无标签样本中挑选置信度高的样本标注伪标签,扩大样本训练集,并采用归一化频率变化率和损伤特征指数作为输入特征,用于识别结构损伤。首先,介绍自训练半监督学习的基本理论和方法;其次,从神经网络构建、损伤特征提取、分类器评估等方面,给出结构损伤识别流程;最后,通过空间桁架的数值案例及3层框架的试验数据,验证所提出的损伤识别方法。结果表明:自训练半监督学习能够从无标签样本中选取置信度较高的样本,为损伤识别提供更充足的有标签样本;在标记样本不足的条件下,SSNN比MLP神经网络的损伤识别效果更好;相较于MLP神经网络,SSNN在单一位置损伤工况下,识别准确率提升约4%,2个位置损伤识别准确率提升约9%。

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=FvKe+ot+Neqn5t5AoJO/+Q==, magXml=ECMCq4DB1uQugOFfy1bRGQ==, pdfUrl=null, pdf=u3WgJfGtrdwjxFH5dSA0JA==, pdfFileSize=6716029, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=1d7A9U4DYJukpbdx5wze/Q==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=odBbw7/RswtBpZVHtjks1Q==, mapNumber=null, authorCompany=null, fund=null, authors=

秦世强(1987—),男,副教授,博士,主要从事结构健康监测研究。E-mail:

, authorsList=秦世强, 杨睿, 苏晟)}, authors=[Author(id=1245389875424378886, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=shiqiangqin@whut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245389875512459286, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389875424378886, language=EN, stringName=Shiqiang QIN, firstName=Shiqiang, middleName=null, lastName=QIN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245389875646677023, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389875424378886, language=CN, stringName=秦世强, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070, bio={"content":"

秦世强(1987—),男,副教授,博士,主要从事结构健康监测研究。E-mail:

"}, bioImg=null, bioContent=

秦世强(1987—),男,副教授,博士,主要从事结构健康监测研究。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245389875327910908, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, xref=null, ext=[AuthorCompanyExt(id=1245389875336299517, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1245389875344688127, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070)])]), Author(id=1245389875722174502, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245389875843809330, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389875722174502, language=EN, stringName=Rui YANG, firstName=Rui, middleName=null, lastName=YANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245389875990609985, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389875722174502, language=CN, stringName=杨睿, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245389875327910908, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, xref=null, ext=[AuthorCompanyExt(id=1245389875336299517, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1245389875344688127, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070)])]), Author(id=1245389876120633421, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1245389876300988510, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389876120633421, language=EN, stringName=Sheng SU, firstName=Sheng, middleName=null, lastName=SU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1245389876410040423, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, authorId=1245389876120633421, language=CN, stringName=苏晟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1245389875327910908, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, xref=null, ext=[AuthorCompanyExt(id=1245389875336299517, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1245389875344688127, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070)])])], keywords=[Keyword(id=1245389876569423986, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, orderNo=1, keyword=structural damage identification), Keyword(id=1245389876737196162, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, orderNo=2, keyword=semi-supervised learning), Keyword(id=1245389876850442381, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, orderNo=3, keyword=self-training), Keyword(id=1245389877001437331, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, orderNo=4, keyword=pseudo label), Keyword(id=1245389877110489241, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, orderNo=5, keyword=neural network), Keyword(id=1245389877202763939, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, orderNo=1, keyword=结构损伤识别), Keyword(id=1245389877307621547, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, orderNo=2, keyword=半监督学习), Keyword(id=1245389877471199414, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, orderNo=3, keyword=自训练), Keyword(id=1245389877601222845, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, orderNo=4, keyword=伪标签), Keyword(id=1245389877710274758, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, orderNo=5, keyword=神经网络)], refs=[Reference(id=1245389882202374646, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2017, volume=34, issue=5, pageStart=59, pageEnd=64, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=赵少杰, 唐细彪, 任伟新, journalName=铁道工程学报, refType=null, unstructuredReference=赵少杰,唐细彪,任伟新. 桥梁事故的统计特征分析及安全风险防控原则[J]. 铁道工程学报201734(5):59-64., articleTitle=桥梁事故的统计特征分析及安全风险防控原则, refAbstract=null), Reference(id=1245389882328203774, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2017, volume=34, issue=5, pageStart=59, pageEnd=64, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=ZHAO Shaojie, TANG Xibiao, REN Weixin, journalName=Journal of Railway Engineering Society, refType=null, unstructuredReference=ZHAO Shaojie, TANG Xibiao, REN Weixin. The statistical characteristics analysis of bridge accidents and prevention principle of security risk[J]. Journal of Railway Engineering Society, 2017, 34(5): 59-64. (in Chinese), articleTitle=The statistical characteristics analysis of bridge accidents and prevention principle of security risk, refAbstract=null), Reference(id=1245389882424672778, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=41, issue=4, pageStart=145, pageEnd=156, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=骆勇鹏, 王林堃, 廖飞宇, journalName=地震工程与工程振动, refType=null, unstructuredReference=骆勇鹏,王林堃,廖飞宇,. 基于一维卷积神经网络的结构损伤识别[J]. 地震工程与工程振动202141(4):145-156., articleTitle=基于一维卷积神经网络的结构损伤识别, refAbstract=null), Reference(id=1245389882516947468, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=41, issue=4, pageStart=145, pageEnd=156, url=null, language=null, rfNumber=[2], rfOrder=3, authorNames=LUO Yongpeng, WANG Linkun, LIAO Feiyu, journalName=Earthquake Engineering and Engineering Dynamics, refType=null, unstructuredReference=LUO Yongpeng, WANG Linkun, LIAO Feiyu, et al. Vibration-based structural damage identification by 1-dimensional convolutional neural network[J]. Earthquake Engineering and Engineering Dynamics, 2021, 41(4): 145-156. (in Chinese), articleTitle=Vibration-based structural damage identification by 1-dimensional convolutional neural network, refAbstract=null), Reference(id=1245389882617610771, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2015, volume=35, issue=6, pageStart=71, pageEnd=77, url=null, language=null, rfNumber=[3], rfOrder=4, authorNames=李延强, 张阳, journalName=地震工程与工程振动, refType=null, unstructuredReference=李延强,张阳. 基于支持向量机与分层遗传算法的斜拉桥全结构损伤分步识别[J]. 地震工程与工程振动201535(6):71-77., articleTitle=基于支持向量机与分层遗传算法的斜拉桥全结构损伤分步识别, refAbstract=null), Reference(id=1245389882722468379, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2015, volume=35, issue=6, pageStart=71, pageEnd=77, url=null, language=null, rfNumber=[3], rfOrder=5, authorNames=LI Yanqiang, ZHANG Yang, journalName=Earthquake Engineering and Engineering Dynamics, refType=null, unstructuredReference=LI Yanqiang, ZHANG Yang. Step-by-step damage detection for cable-stayed bridge based on support vector machine and hierarchic genetic algorithm[J]. Earthquake Engineering and Engineering Dynamics, 2015, 35(6): 71-77. (in Chinese), articleTitle=Step-by-step damage detection for cable-stayed bridge based on support vector machine and hierarchic genetic algorithm, refAbstract=null), Reference(id=1245389882827325987, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2019, volume=32, issue=11, pageStart=1, pageEnd=20, url=null, language=null, rfNumber=[4], rfOrder=6, authorNames=孙利民, 尚志强, 夏烨, journalName=中国公路学报, refType=null, unstructuredReference=孙利民,尚志强,夏烨. 大数据背景下的桥梁结构健康监测研究现状与展望[J]. 中国公路学报201932(11):1-20., articleTitle=大数据背景下的桥梁结构健康监测研究现状与展望, refAbstract=null), Reference(id=1245389882927989290, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2019, volume=32, issue=11, pageStart=1, pageEnd=20, url=null, language=null, rfNumber=[4], rfOrder=7, authorNames=SUN Limin, SHANG Zhiqiang, XIA Ye, journalName=China Journal of Highway and Transport, refType=null, unstructuredReference=SUN Limin, SHANG Zhiqiang, XIA Ye. Development and prospect of bridge structural health monitoring in the context of big data[J]. China Journal of Highway and Transport, 2019, 32(11): 1-20. (in Chinese), articleTitle=Development and prospect of bridge structural health monitoring in the context of big data, refAbstract=null), Reference(id=1245389883028652593, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=28, issue=4, pageStart=2621, pageEnd=2643, url=null, language=null, rfNumber=[5], rfOrder=8, authorNames=FLAH M, NUNEZ I, BEN CHAABENE W, journalName=Archives of Computational Methods in Engineering, refType=null, unstructuredReference=FLAH M, NUNEZ I, BEN CHAABENE W, et al. Machine learning algorithms in civil structural health monitoring: A systematic review[J]. Archives of Computational Methods in Engineering, 2021, 28(4): 2621-2643., articleTitle=Machine learning algorithms in civil structural health monitoring: A systematic review, refAbstract=null), Reference(id=1245389883133510202, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2019, volume=19, issue=24, pageStart=5501, pageEnd=null, url=null, language=null, rfNumber=[6], rfOrder=9, authorNames=CHUN C J, RYU S K, journalName=Sensors, refType=null, unstructuredReference=CHUN C J, RYU S K. Road surface damage detection using fully convolutional neural networks and semi-supervised learning[J]. Sensors, 2019, 19(24): 5501., articleTitle=Road surface damage detection using fully convolutional neural networks and semi-supervised learning, refAbstract=null), Reference(id=1245389883271922243, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=55, issue=9, pageStart=1705, pageEnd=1713, url=null, language=null, rfNumber=[7], rfOrder=10, authorNames=黄发明, 潘李含, 姚池, journalName=浙江大学学报(工学版), refType=null, unstructuredReference=黄发明,潘李含,姚池,. 基于半监督机器学习的滑坡易发性预测建模[J]. 浙江大学学报(工学版)202155(9):1705-1713., articleTitle=基于半监督机器学习的滑坡易发性预测建模, refAbstract=null), Reference(id=1245389883364196936, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=55, issue=9, pageStart=1705, pageEnd=1713, url=null, language=null, rfNumber=[7], rfOrder=11, authorNames=HUANG Faming, PAN Lihan, YAO Chi, journalName=Journal of Zhejiang University (Engineering Science), refType=null, unstructuredReference=HUANG Faming, PAN Lihan, YAO Chi, et al. Landslide susceptibility prediction modelling based on semi-supervised machine learning[J]. Journal of Zhejiang University (Engineering Science)202155(9): 1705-1713. (in Chinese), articleTitle=Landslide susceptibility prediction modelling based on semi-supervised machine learning, refAbstract=null), Reference(id=1245389883477443149, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2016, volume=42, issue=4, pageStart=128, pageEnd=133, url=null, language=null, rfNumber=[8], rfOrder=12, authorNames=徐菁, 曲丽敏, 卢翠萍, journalName=兰州理工大学学报, refType=null, unstructuredReference=徐菁,曲丽敏,卢翠萍. 大跨度空间网格结构的健康监测系统[J]. 兰州理工大学学报201642(4):128-133., articleTitle=大跨度空间网格结构的健康监测系统, refAbstract=null), Reference(id=1245389883615855190, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2016, volume=42, issue=4, pageStart=128, pageEnd=133, url=null, language=null, rfNumber=[8], rfOrder=13, authorNames=XU Jing, QU Limin, LU Cuiping, journalName=Journal of Lanzhou University of Technology, refType=null, unstructuredReference=XU Jing, QU Limin, LU Cuiping. Health monitoring system with large-span space-netted structure[J]. Journal of Lanzhou University of Technology, 2016, 42(4): 128-133. (in Chinese), articleTitle=Health monitoring system with large-span space-netted structure, refAbstract=null), Reference(id=1245389883741684318, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2016, volume=23, issue=3, pageStart=446, pageEnd=457, url=null, language=null, rfNumber=[9], rfOrder=14, authorNames=SATPAL S B, GUHA A, BANERJEE S, journalName=Structural Control and Health Monitoring, refType=null, unstructuredReference=SATPAL S B, GUHA A, BANERJEE S. Damage identification in aluminum beams using support vector machine: Numerical and experimental studies[J]. Structural Control and Health Monitoring, 2016, 23(3): 446-457., articleTitle=Damage identification in aluminum beams using support vector machine: Numerical and experimental studies, refAbstract=null), Reference(id=1245389883875902052, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=20, issue=4, pageStart=1675, pageEnd=1688, url=null, language=null, rfNumber=[10], rfOrder=15, authorNames=ZHANG Z M, SUN C, journalName=Structural Health Monitoring, refType=null, unstructuredReference=ZHANG Z M, SUN C. Structural damage identification via physics-guided machine learning: A methodology integrating pattern recognition with finite element model updating[J]. Structural Health Monitoring, 2021, 20(4): 1675-1688., articleTitle=Structural damage identification via physics-guided machine learning: A methodology integrating pattern recognition with finite element model updating, refAbstract=null), Reference(id=1245389884026897002, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2019, volume=119, issue=null, pageStart=100, pageEnd=119, url=null, language=null, rfNumber=[11], rfOrder=16, authorNames=ROGERS T J, WORDEN K, FUENTES R, journalName=Mechanical Systems and Signal Processing, refType=null, unstructuredReference=ROGERS T J, WORDEN K, FUENTES R, et al. A Bayesian non-parametric clustering approach for semi-supervised structural health monitoring[J]. Mechanical Systems and Signal Processing, 2019, 119: 100-119., articleTitle=A Bayesian non-parametric clustering approach for semi-supervised structural health monitoring, refAbstract=null), Reference(id=1245389884148531825, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2003, volume=259, issue=2, pageStart=323, pageEnd=343, url=null, language=null, rfNumber=[12], rfOrder=17, authorNames=WORDEN K, MANSON G, ALLMAN D, journalName=Journal of Sound and Vibration, refType=null, unstructuredReference=WORDEN K, MANSON G, ALLMAN D. Experimental validation of a structural health monitoring methodology: Part I. novelty detection on a laboratory structure[J]. Journal of Sound and Vibration, 2003, 259(2): 323-343., articleTitle=Experimental validation of a structural health monitoring methodology: Part I. novelty detection on a laboratory structure, refAbstract=null), Reference(id=1245389884270166648, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2003, volume=259, issue=2, pageStart=345, pageEnd=363, url=null, language=null, rfNumber=[13], rfOrder=18, authorNames=MANSON G, WORDEN K, ALLMAN D, journalName=Journal of Sound and Vibration, refType=null, unstructuredReference=MANSON G, WORDEN K, ALLMAN D. Experimental validation of a structural health monitoring methodology: Part II. novelty detection on a gnat aircraft[J]. Journal of Sound and Vibration, 2003, 259(2): 345-363., articleTitle=Experimental validation of a structural health monitoring methodology: Part II. novelty detection on a gnat aircraft, refAbstract=null), Reference(id=1245389884379218558, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2020, volume=239, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=19, authorNames=XU D J, LIU P, CHEN Z C, journalName=Engineering Fracture Mechanics, refType=null, unstructuredReference=XU D J, LIU P, CHEN Z C. A waveform clustering method for damage mode identification for composite laminates under hygrothermal environment[J]. Engineering Fracture Mechanics, 2020, 239: 107290., articleTitle=A waveform clustering method for damage mode identification for composite laminates under hygrothermal environment, refAbstract=null), Reference(id=1245389884475687553, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2011, volume=181, issue=7, pageStart=1187, pageEnd=1211, url=null, language=null, rfNumber=[15], rfOrder=20, authorNames=MAHARAJ E A, D’URSO P, journalName=Information Sciences, refType=null, unstructuredReference=MAHARAJ E A, D’URSO P. Fuzzy clustering of time series in the frequency domain[J]. Information Sciences, 2011, 181(7): 1187-1211., articleTitle=Fuzzy clustering of time series in the frequency domain, refAbstract=null), Reference(id=1245389884597322376, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2007, volume=29, issue=2, pageStart=174, pageEnd=184, url=null, language=null, rfNumber=[16], rfOrder=21, authorNames=DA SILVA S, DIAS M Jr, LOPES V Jr, journalName=Journal of the Brazilian Society of Mechanical Sciences and Engineering, refType=null, unstructuredReference=DA SILVA S, DIAS M Jr, LOPES V Jr. Damage detection in a benchmark structure using AR-ARX models and statistical pattern recognition[J]. Journal of the Brazilian Society of Mechanical Sciences and Engineering, 2007, 29(2): 174-184., articleTitle=Damage detection in a benchmark structure using AR-ARX models and statistical pattern recognition, refAbstract=null), Reference(id=1245389884706374286, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2017, volume=32, issue=5, pageStart=361, pageEnd=378, url=null, language=null, rfNumber=[17], rfOrder=22, authorNames=CHA Y J, CHOI W, BÜYÜKÖZTÜRK O, journalName=Computer-Aided Civil and Infrastructure Engineering, refType=null, unstructuredReference=CHA Y J, CHOI W, BÜYÜKÖZTÜRK O. Deep learning-based crack damage detection using convolutional neural networks[J]. Computer-Aided Civil and Infrastructure Engineering, 2017, 32(5): 361-378., articleTitle=Deep learning-based crack damage detection using convolutional neural networks, refAbstract=null), Reference(id=1245389884811231892, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2022, volume=135, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[18], rfOrder=23, authorNames=SHIM S, KIM J, LEE S W, journalName=Automation in Construction, refType=null, unstructuredReference=SHIM S, KIM J, LEE S W, et al. Road damage detection using super-resolution and semi-supervised learning with generative adversarial network[J]. Automation in Construction, 2022, 135: 104139., articleTitle=Road damage detection using super-resolution and semi-supervised learning with generative adversarial network, refAbstract=null), Reference(id=1245389884911895193, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2020, volume=191, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[19], rfOrder=24, authorNames=HUANG F, CAO Z, GUO J, journalName=Catena (Giessen), refType=null, unstructuredReference=HUANG F, CAO Z, GUO J, et al. Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping[J]. Catena (Giessen), 2020, 191: 104580., articleTitle=Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping, refAbstract=null), Reference(id=1245389884999975584, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=41, issue=9, pageStart=93, pageEnd=96, 107, url=null, language=null, rfNumber=[20], rfOrder=25, authorNames=霍士伟, 郭圣明, 唐宇波, journalName=舰船电子工程, refType=null, unstructuredReference=霍士伟,郭圣明,唐宇波. 基于自训练半监督学习的战场态势评估模型[J]. 舰船电子工程202141(9):93-96, 107., articleTitle=基于自训练半监督学习的战场态势评估模型, refAbstract=null), Reference(id=1245389885104833185, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=41, issue=9, pageStart=93, pageEnd=96, 107, url=null, language=null, rfNumber=[20], rfOrder=26, authorNames=HUO Shiwei, GUO Shengming, TANG Yubo, journalName=Ship Electronic Engineering, refType=null, unstructuredReference=HUO Shiwei, GUO Shengming, TANG Yubo. Battlefield situation assessment model based on self-training semi-supervised learning[J]. Ship Electronic Engineering, 2021, 41(9): 93-96, 107. (in Chinese), articleTitle=Battlefield situation assessment model based on self-training semi-supervised learning, refAbstract=null), Reference(id=1245389885201302179, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2010, volume=24, issue=3, pageStart=415, pageEnd=439, url=null, language=null, rfNumber=[21], rfOrder=27, authorNames=ZHOU Z, LI M, journalName=Knowledge and Information Systems, refType=null, unstructuredReference=ZHOU Z, LI M. Semi-supervised learning by disagreement[J]. Knowledge and Information Systems, 2010, 24(3): 415-439., articleTitle=Semi-supervised learning by disagreement, refAbstract=null), Reference(id=1245389885306159781, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2019, volume=4, issue=1, pageStart=17, pageEnd=null, url=null, language=null, rfNumber=[22], rfOrder=28, authorNames=BOUZENAD A E, EL MOUNTASSIR M, YAACOUBI S, journalName=Inventions, refType=null, unstructuredReference=BOUZENAD A E, EL MOUNTASSIR M, YAACOUBI S, et al. A semi-supervised based K-means algorithm for optimal guided waves structural health monitoring: A case study[J]. Inventions, 2019, 4(1): 17., articleTitle=A semi-supervised based K-means algorithm for optimal guided waves structural health monitoring: A case study, refAbstract=null), Reference(id=1245389885390045865, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=5584, pageEnd=5591, url=null, language=null, rfNumber=[23], rfOrder=29, authorNames=PEI Z X, LIN R H, ZHANG X B, journalName=null, refType=null, unstructuredReference=PEI Z X, LIN R H, ZHANG X B, et al. CFM: A consistency filtering mechanism for road damage detection[C]∥2020 IEEE International Conference on Big Data (Big Data). Atlanta, GA, USA. Piscataway, NJ, USA: IEEE, 2021: 5584-5591., articleTitle=CFM: A consistency filtering mechanism for road damage detection, refAbstract=null), Reference(id=1245389885494903469, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2017, volume=8, issue=1, pageStart=355, pageEnd=370, url=null, language=null, rfNumber=[24], rfOrder=30, authorNames=TANHA J, VAN SOMEREN M, AFSARMANESH H, journalName=International Journal of Machine Learning and Cybernetics, refType=null, unstructuredReference=TANHA J, VAN SOMEREN M, AFSARMANESH H. Semi-supervised self-training for decision tree classifiers[J]. International Journal of Machine Learning and Cybernetics, 2017, 8(1): 355-370., articleTitle=Semi-supervised self-training for decision tree classifiers, refAbstract=null), Reference(id=1245389885582983858, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=1978, volume=20, issue=2, pageStart=93, pageEnd=100, url=null, language=null, rfNumber=[25], rfOrder=31, authorNames=ADAMS R D, CAWLEY P, PYE C J, journalName=Journal of Mechanical Engineering Science, refType=null, unstructuredReference=ADAMS R D, CAWLEY P, PYE C J, et al. A vibration technique for non-destructively assessing the integrity of structures[J]. Journal of Mechanical Engineering Science, 1978, 20(2): 93-100., articleTitle=A vibration technique for non-destructively assessing the integrity of structures, refAbstract=null), Reference(id=1245389885679452853, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=1979, volume=14, issue=2, pageStart=49, pageEnd=57, url=null, language=null, rfNumber=[26], rfOrder=32, authorNames=CAWLEY P, ADAMS R D, journalName=The Journal of Strain Analysis for Engineering Design, refType=null, unstructuredReference=CAWLEY P, ADAMS R D. The location of defects in structures from measurements of natural frequencies[J]. The Journal of Strain Analysis for Engineering Design, 1979, 14(2): 49-57., articleTitle=The location of defects in structures from measurements of natural frequencies, refAbstract=null), Reference(id=1245389885792699064, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2003, volume=25, issue=1, pageStart=57, pageEnd=67, url=null, language=null, rfNumber=[27], rfOrder=33, authorNames=KIM J T, RYU Y S, CHO H M, journalName=Engineering Structures, refType=null, unstructuredReference=KIM J T, RYU Y S, CHO H M, et al. Damage identification in beam-type structures: Frequency-based method vs mode-shape-based method[J]. Engineering Structures, 2003, 25(1): 57-67., articleTitle=Damage identification in beam-type structures: Frequency-based method vs mode-shape-based method, refAbstract=null), Reference(id=1245389885880779453, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=1998, volume=210, issue=1, pageStart=91, pageEnd=115, url=null, language=null, rfNumber=[28], rfOrder=34, authorNames=LAM H F, KO J M, WONG C W, journalName=Journal of Sound and Vibration, refType=null, unstructuredReference=LAM H F, KO J M, WONG C W. Localization of damaged structural connections based on experimental modal and sensitivity analysis[J]. Journal of Sound and Vibration, 1998, 210(1): 91-115., articleTitle=Localization of damaged structural connections based on experimental modal and sensitivity analysis, refAbstract=null), Reference(id=1245389885960471233, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=1997, volume=205, issue=1, pageStart=1, pageEnd=18, url=null, language=null, rfNumber=[29], rfOrder=35, authorNames=FARRAR C R, JAMES G H III, journalName=Journal of Sound and Vibration, refType=null, unstructuredReference=FARRAR C R, JAMES G H III. System identification from ambient vibration measurements on a bridge[J]. Journal of Sound and Vibration, 1997, 205(1): 1-18., articleTitle=System identification from ambient vibration measurements on a bridge, refAbstract=null), Reference(id=1245389886052745925, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, doi=null, pmid=null, pmcid=null, year=2009, volume=3, issue=1, pageStart=1320, pageEnd=1327, url=null, language=null, rfNumber=[30], rfOrder=36, authorNames=FIGUEIREDO E, PARK G, FIGUEIRAS J, journalName=Algorithms, refType=null, unstructuredReference=FIGUEIREDO E, PARK G, FIGUEIRAS J, et al. Structural health monitoring algorithm comparisons using standard data sets[J]. Algorithms, 2009, 3(1): 1320-1327., articleTitle=Structural health monitoring algorithm comparisons using standard data sets, refAbstract=null)], funds=[Fund(id=1245389882030408175, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, awardId=51608408, language=CN, fundingSource=国家自然科学基金项目(51608408), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1245389875327910908, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, xref=null, ext=[AuthorCompanyExt(id=1245389875336299517, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1245389875344688127, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, companyId=1245389875327910908, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学 土木工程与建筑学院,湖北 武汉 430070)])], figs=[ArticleFig(id=1245389877970321617, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 1, caption=Structural damage identification flow chart based on SSNN, figureFileSmall=1q4auQ1abp2joBeLSBvbrQ==, figureFileBig=1d7A9U4DYJukpbdx5wze/Q==, tableContent=null), ArticleFig(id=1245389878096150750, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图1, caption=基于SSNN的结构损伤识别流程图, figureFileSmall=1q4auQ1abp2joBeLSBvbrQ==, figureFileBig=1d7A9U4DYJukpbdx5wze/Q==, tableContent=null), ArticleFig(id=1245389878368780530, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 2, caption=Schematic diagram of space truss structure, figureFileSmall=bgpWkGp2oSiCR6aKgIuYfw==, figureFileBig=aqnjlOy+/IOQrw2VNY4vpA==, tableContent=null), ArticleFig(id=1245389878507192571, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图2, caption=空间桁架结构示意图, figureFileSmall=bgpWkGp2oSiCR6aKgIuYfw==, figureFileBig=aqnjlOy+/IOQrw2VNY4vpA==, tableContent=null), ArticleFig(id=1245389878637216007, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 3, caption=Single position damage location results of space truss, figureFileSmall=LHx+RTRt9huJWuQAtnJnHQ==, figureFileBig=V9tC8XRXXa/V44Vf1F+WPg==, tableContent=null), ArticleFig(id=1245389878742073616, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图3, caption=空间桁架的单位置损伤定位结果, figureFileSmall=LHx+RTRt9huJWuQAtnJnHQ==, figureFileBig=V9tC8XRXXa/V44Vf1F+WPg==, tableContent=null), ArticleFig(id=1245389878846931225, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 4, caption=Confusion matrix for single position damage of space truss, figureFileSmall=1vKlktKWwwBkCBm9UXpeFA==, figureFileBig=7IKuoKbn2jCKRxBq7t/NSg==, tableContent=null), ArticleFig(id=1245389878964371748, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图4, caption=空间桁架单位置损伤识别的混淆矩阵, figureFileSmall=1vKlktKWwwBkCBm9UXpeFA==, figureFileBig=7IKuoKbn2jCKRxBq7t/NSg==, tableContent=null), ArticleFig(id=1245389879081812268, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 5, caption=F1 score for damage detection of space truss with single position damage, figureFileSmall=3iwaKO4dUHczEgVWxy3gqA==, figureFileBig=i3ChX0HAkfKS39dCwNCeFQ==, tableContent=null), ArticleFig(id=1245389879253778746, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图5, caption=空间桁架单位置损伤识别F1分数, figureFileSmall=3iwaKO4dUHczEgVWxy3gqA==, figureFileBig=i3ChX0HAkfKS39dCwNCeFQ==, tableContent=null), ArticleFig(id=1245389879387996483, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 6, caption=Variations in error for damage detection of space truss with single damage, figureFileSmall=cmOP2tuSTpxDx6RugU0jgQ==, figureFileBig=bPj15HIkHIcD3psmDbMSgw==, tableContent=null), ArticleFig(id=1245389879505437007, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图6, caption=空间桁架单位置损伤识别误差变化, figureFileSmall=cmOP2tuSTpxDx6RugU0jgQ==, figureFileBig=bPj15HIkHIcD3psmDbMSgw==, tableContent=null), ArticleFig(id=1245389879652237654, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 7, caption=Confusion matrix for two position damages of space truss, figureFileSmall=cDZ0AoprDHXTLXEePrSIZA==, figureFileBig=fAoYcjTHTn5e93Z4v4jcpw==, tableContent=null), ArticleFig(id=1245389879765483870, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图7, caption=空间桁架2个位置损伤识别的混淆矩阵, figureFileSmall=cDZ0AoprDHXTLXEePrSIZA==, figureFileBig=fAoYcjTHTn5e93Z4v4jcpw==, tableContent=null), ArticleFig(id=1245389879912284522, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 8, caption=F1 score for damage detection of space truss with two position damages, figureFileSmall=6mryXRM9KUGDbJuhgjvREw==, figureFileBig=zYr+bw8NTANngl28sYB95A==, tableContent=null), ArticleFig(id=1245389880033919345, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图8, caption=空间桁架2个位置损伤识别F1分数, figureFileSmall=6mryXRM9KUGDbJuhgjvREw==, figureFileBig=zYr+bw8NTANngl28sYB95A==, tableContent=null), ArticleFig(id=1245389880147165561, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 9, caption=Test layout of three-story frame[30], figureFileSmall=FHjN1UDj7Dxgrvh6DubYKw==, figureFileBig=cxZuidCTuSOylt6ew842pA==, tableContent=null), ArticleFig(id=1245389880264606080, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图9, caption=3层框架试验布置[30], figureFileSmall=FHjN1UDj7Dxgrvh6DubYKw==, figureFileBig=cxZuidCTuSOylt6ew842pA==, tableContent=null), ArticleFig(id=1245389880382046598, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 10, caption=The first three modes of the three-story frame under different damage conditions, figureFileSmall=DqgAxgezdp/h+Id+FWecFg==, figureFileBig=nRKZ1pfb+Frojgt/MxWlsQ==, tableContent=null), ArticleFig(id=1245389880461738377, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图10, caption=不同损伤工况下3层框架的前3阶振型, figureFileSmall=DqgAxgezdp/h+Id+FWecFg==, figureFileBig=nRKZ1pfb+Frojgt/MxWlsQ==, tableContent=null), ArticleFig(id=1245389880566595983, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 11, caption=Damage localization results of three-story frame, figureFileSmall=r0WN/ep69L9HFSg/t7rTVw==, figureFileBig=YkDwfUfK40VRyCAYW2jCsQ==, tableContent=null), ArticleFig(id=1245389880663064982, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图11, caption=3层框架损伤定位结果, figureFileSmall=r0WN/ep69L9HFSg/t7rTVw==, figureFileBig=YkDwfUfK40VRyCAYW2jCsQ==, tableContent=null), ArticleFig(id=1245389880755339674, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 12, caption=Confusion matrix for damage identification of three-story frame, figureFileSmall=YqTSuqiQCi8FonNfy3ZTzA==, figureFileBig=pgReHRBX57K8r00ug/oeAg==, tableContent=null), ArticleFig(id=1245389880889557409, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图12, caption=3层框架损伤识别的混淆矩阵, figureFileSmall=YqTSuqiQCi8FonNfy3ZTzA==, figureFileBig=pgReHRBX57K8r00ug/oeAg==, tableContent=null), ArticleFig(id=1245389881044746665, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Fig. 13, caption=F1 score for damage identification of three-story frame, figureFileSmall=4b3QA/F/5T7fIt+ETS506w==, figureFileBig=FPrpOdF2cqpklJd1WawDvQ==, tableContent=null), ArticleFig(id=1245389881170575793, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=图13, caption=3层框架损伤识别的F1分数, figureFileSmall=4b3QA/F/5T7fIt+ETS506w==, figureFileBig=FPrpOdF2cqpklJd1WawDvQ==, tableContent=null), ArticleFig(id=1245389881292210617, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Table 1, caption=

Single position damage setting of space truss

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况损伤位置损伤程度样本数量/个
D11#(10%~90%)E1100
D22#(10%~90%)E2100
D33#(10%~90%)E3100
D44#(10%~90%)E4100
D55#(10%~90%)E5100
), ArticleFig(id=1245389881430622657, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=表1, caption=

空间桁架的单一位置损伤工况设置

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况损伤位置损伤程度样本数量/个
D11#(10%~90%)E1100
D22#(10%~90%)E2100
D33#(10%~90%)E3100
D44#(10%~90%)E4100
D55#(10%~90%)E5100
), ArticleFig(id=1245389881543868873, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Table 2, caption=

Two position damage setting of space truss

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况损伤位置样本数量/个损伤工况损伤位置样本数量/个
S11#,2#80S61#,4#80
S21#,3#80S72#,5#80
S31#,4#80S83#,4#80
S41#,5#80S93#,5#80
S52#,3#80S104#,5#80
), ArticleFig(id=1245389881694863825, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=表2, caption=

空间桁架的2个位置损伤工况设置

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况损伤位置样本数量/个损伤工况损伤位置样本数量/个
S11#,2#80S61#,4#80
S21#,3#80S72#,5#80
S31#,4#80S83#,4#80
S41#,5#80S93#,5#80
S52#,3#80S104#,5#80
), ArticleFig(id=1245389881816498649, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=EN, label=Table 3, caption=

Test damage condition of three-story frame

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况编号损伤等级样本数量/个
E11~10-0.22k110
11~20-0.44k110
E221~30-0.22k210
31~40-0.44k210
E341~50-0.22k310
51~60-0.44k310
), ArticleFig(id=1245389881908773346, tenantId=1146029695717560320, journalId=1241701559352995854, articleId=1245389862614975093, language=CN, label=表3, caption=

3层框架的试验损伤工况

, figureFileSmall=null, figureFileBig=null, tableContent=
损伤工况编号损伤等级样本数量/个
E11~10-0.22k110
11~20-0.44k110
E221~30-0.22k210
31~40-0.44k210
E341~50-0.22k310
51~60-0.44k310
)], attaches=null, journal=Journal(id=1241699397025705997, delFlag=0, nameCn=地震工程与工程振动, nameEn=Earthquake Engineering and Engineering Dynamics, nameHistory1=null, nameHistory2=null, issn=1000-1301, eissn=null, cn=23-1157/P, coden=null, periodic=1, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=/vhigrb0Ie5dk9EvytPyQg==, journalPrice=null, startedYear=null, abbrevIsoEn=Earthquake Engineering and Engineering Dynamics, journalRemark=null, publicationField=null, createdTime=1773973805912, updatedTime=1773974475511, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=E, firstLetterEn=E, subjectCode=Engineering, subjectName=null, subjectCodeEn=Engineering, subjectNameEn=null, picCn=/vhigrb0Ie5dk9EvytPyQg==, picEn=mP4JjBszibypE3yTy4J1ag==, jcr=null, cjcr=null, exts=[JournalExt(id=1241702205619753886, language=CN, name=地震工程与工程振动, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1773974475531, updatedTime=1773974475531, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://dzgc.paperonce.org/#/Login, submissionEditorUrl=https://dzgc.paperonce.org/#/Login, submissionReviewUrl=https://dzgc.paperonce.org/#/Login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1241702205665891231, language=EN, name=Earthquake Engineering and Engineering Dynamics, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1773974475542, updatedTime=1773974475542, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://dzgc.paperonce.org/#/Login, submissionEditorUrl=https://dzgc.paperonce.org/#/Login, submissionReviewUrl=https://dzgc.paperonce.org/#/Login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1241701559352995854, websiteList=[Website(id=1241702575498646432, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1241701559352995854, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/dzgcygczd/CN, language=CN, createTime=1773974563717, createBy=18614031015, updateTime=1773974590635, updateBy=18614031015, name=地震工程与工程振动-中文, tplId=1146099689490845704, title=地震工程与工程振动, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1241703043805270973, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=articleTextType, value=kx, createTime=1773974675370, updateTime=1773974675370, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043771716538, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=banner, value=null, createTime=1773974675362, updateTime=1773974675362, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043830436800, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=grayFlag, value=0, createTime=1773974675376, updateTime=1773974675376, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043759133625, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=logo, value=https://castjournals.cast.org.cn/joweb/dzgcygczd/CN/file/pic?fileId=C6sfoMWhaI5iihfhixVjYQ==, createTime=1773974675359, updateTime=1773974675359, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043843019714, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=minRunFlag, value=0, createTime=1773974675379, updateTime=1773974675379, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043801076668, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/dzgcygczd/CN/file/pic, createTime=1773974675369, updateTime=1773974675369, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043838825409, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=silenceFlag, value=0, createTime=1773974675378, updateTime=1773974675378, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043780105147, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1773974675364, updateTime=1773974675364, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043813659582, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=themeColor, value=null, createTime=1773974675372, updateTime=1773974675372, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703043822048191, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575498646432, code=themeStyle, value=null, createTime=1773974675374, updateTime=1773974675374, creator=18614031015, updator=18614031015)]), Website(id=1241702575569949602, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1241701559352995854, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/dzgcygczd/EN, language=EN, createTime=1773974563735, createBy=18614031015, updateTime=1773974605208, updateBy=18614031015, name=地震工程与工程振动-英文, tplId=1146101810881728533, title=Earthquake Engineering and Engineering Dynamics, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1241703068497138631, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=articleTextType, value=kx, createTime=1773974681257, updateTime=1773974681257, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068480361412, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=banner, value=null, createTime=1773974681253, updateTime=1773974681253, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068513915850, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=grayFlag, value=0, createTime=1773974681261, updateTime=1773974681261, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068471972803, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=logo, value=https://castjournals.cast.org.cn/joweb/dzgcygczd/EN/file/pic?fileId=C6sfoMWhaI5iihfhixVjYQ==, createTime=1773974681251, updateTime=1773974681251, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068530693068, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=minRunFlag, value=0, createTime=1773974681265, updateTime=1773974681265, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068492944326, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/dzgcygczd/EN/file/pic, createTime=1773974681256, updateTime=1773974681256, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068522304459, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=silenceFlag, value=0, createTime=1773974681263, updateTime=1773974681263, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068484555717, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1773974681254, updateTime=1773974681254, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068505527240, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=themeColor, value=null, createTime=1773974681259, updateTime=1773974681259, creator=18614031015, updator=18614031015), WebsiteProps(id=1241703068509721545, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1241702575569949602, code=themeStyle, value=null, createTime=1773974681260, updateTime=1773974681260, creator=18614031015, updator=18614031015)])], journalTitle=地震工程与工程振动, weixinUrl=null, journalUrl=https://dzgc.paperonce.org/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Earthquake Engineering and Engineering Dynamics, journalPhotoCn=/vhigrb0Ie5dk9EvytPyQg==, journalPhotoEn=mP4JjBszibypE3yTy4J1ag==, journalFirstLetter=E, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/dzgcygczd/CN/10.13197/j.eeed.2024.0205, detailUrlEn=https://castjournals.cast.org.cn/joweb/dzgcygczd/EN/10.13197/j.eeed.2024.0205, pdfUrlCn=https://castjournals.cast.org.cn/joweb/dzgcygczd/CN/PDF/10.13197/j.eeed.2024.0205, pdfUrlEn=https://castjournals.cast.org.cn/joweb/dzgcygczd/EN/PDF/10.13197/j.eeed.2024.0205, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于自训练半监督神经网络的结构损伤识别
收藏切换
PDF下载
秦世强 , 杨睿 , 苏晟
地震工程与工程振动 | 2024,44(2): 38-49
收起
收藏切换
地震工程与工程振动 | 2024, 44(2): 38-49
基于自训练半监督神经网络的结构损伤识别
全屏
秦世强 , 杨睿, 苏晟
作者信息
  • 武汉理工大学 土木工程与建筑学院,湖北 武汉 430070
  • 秦世强(1987—),男,副教授,博士,主要从事结构健康监测研究。E-mail:

Structural damage identification based on self-training semi-supervised neural network
Shiqiang QIN , Rui YANG, Sheng SU
Affiliations
  • School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China
doi: 10.13197/j.eeed.2024.0205
文章导航
收藏切换

为解决结构损伤识别中标签样本不足的问题,提出一种基于自训练半监督神经网络(self-training semi-supervised neural networks,SSNN)的结构损伤识别框架,该框架利用自训练半监督方法对多层感知机(multilayer perceptron,MLP)神经网络进行训练,从无标签样本中挑选置信度高的样本标注伪标签,扩大样本训练集,并采用归一化频率变化率和损伤特征指数作为输入特征,用于识别结构损伤。首先,介绍自训练半监督学习的基本理论和方法;其次,从神经网络构建、损伤特征提取、分类器评估等方面,给出结构损伤识别流程;最后,通过空间桁架的数值案例及3层框架的试验数据,验证所提出的损伤识别方法。结果表明:自训练半监督学习能够从无标签样本中选取置信度较高的样本,为损伤识别提供更充足的有标签样本;在标记样本不足的条件下,SSNN比MLP神经网络的损伤识别效果更好;相较于MLP神经网络,SSNN在单一位置损伤工况下,识别准确率提升约4%,2个位置损伤识别准确率提升约9%。

结构损伤识别  /  半监督学习  /  自训练  /  伪标签  /  神经网络

A structural damage recognition framework based on a self-training semi-supervised neural network (SSNN) is proposed to solve the problem of insufficient labeled data in structural damage identification. The framework utilizes the multilayer perceptron (MLP) neural network for semi-supervised training by the self-training method. The data samples with high confidence are selected from the unlabeled data to make pseudo labels, expanding the training set. Normalized frequency change ratio and damage signature index are employed as input features of neural networks to identify structural damage. Firstly, the theory fundamentals of semi-supervised self-training learning are introduced. Secondly, the procedure of structural damage identification based on self-training semi-supervised learning, including neural network construction, damage characteristic extraction, and classifier evaluation, is introduced. Finally, the proposed damage identification method is illustrated by numerical simulation of a spatial truss and experimental data of a three-story frame. The results show that the self-training semi-supervised method can expand the labeled sample data by selecting samples with higher confidence from unlabeled data, thus providing sufficient labeled data for damage identification. Under the insufficient labeled data conditions, the SSNN performs better than MLP. Compared with MLP, SSNN increases the identification accuracy by 4% and 9% under the single and two positions damage locations, respectively.

structural damage identification  /  semi-supervised learning  /  self-training  /  pseudo label  /  neural network
秦世强, 杨睿, 苏晟. 基于自训练半监督神经网络的结构损伤识别. 地震工程与工程振动, 2024 , 44 (2) : 38 -49 . DOI: 10.13197/j.eeed.2024.0205
Shiqiang QIN, Rui YANG, Sheng SU. Structural damage identification based on self-training semi-supervised neural network[J]. Earthquake Engineering and Engineering Dynamics, 2024 , 44 (2) : 38 -49 . DOI: 10.13197/j.eeed.2024.0205
近年来,因工程结构事故导致人员伤亡和经济损失的案例屡见不鲜。2016年,江西省泰和老大桥在施工作业时发生垮塌,造成3人遇难,5人受伤[1];2018年3月15日,美国佛罗里达州立大学附近一座正在施工的人行天桥发生倒塌,直接造成的经济损失约1 420 万美元。因此,采用科学的方法对工程结构进行健康监测,从而及时掌握工程结构的性能损伤,防止突发性事故的发生,就尤为必要[2]。结构损伤识别是结构健康监测的核心技术。近十年,随着机器学习和人工智能理论的不断发展与应用,基于机器学习的结构损伤识别成为当前结构健康监测方向的研究热点[3],并取得了大量成果。文献[4-5]综述了基于机器学习的结构损伤识别方法。
目前,按照是否利用已知样本数据的标签信息,可以将机器学习分为监督学习和无监督学习两大类[6-7]。监督学习通过学习有标签样本的信息,建立输入数据特征和标签信息的映射关系,用以预测新的样本数据。基于监督学习的结构损伤识别已有大量研究成果。徐菁等[8]提出基于时间序列自回归模型和浅层BP神经网络的损伤识别方法,以网壳结构为研究对象,研究表明:该损伤识别方法具有较高的准确性。SATPAL等[9]基于一阶模态振型,结合支持向量机对悬臂梁以及两端固支梁进行损伤识别,通过对比平均误差百分比与相对误差百分比来判断支持向量机的识别效果。基于监督学习的损伤识别方法的一大挑战是缺乏有标签的样本数据,导致训练的模型预测精度不高[10-11]。无监督学习仅使用无标签数据,从样本数据特征出发,达到聚类或降维的效果。它能在一定程度上避免有标签样本难以获得的问题,在结构损伤识别领域也得到一定的关注。例如,WORDEN等[12]和MANSON等[13]采用无监督的方法识别机翼可能潜在的损伤,该方法将结构损伤识别简化为离群值检测问题。然而,该方法仅区分受损状态和未损坏状态,未能进一步确定损伤发生位置。为了对损伤进行更准确的定位,研究者将聚类的思想引入损伤识别中,将具有相同或相似特征的样本划分为同一类别。XU等[14]提出一种基于波形聚类的复合材料损伤识别方法,通过聚类结果能够得出各种损伤模式的特征。MAHARAJ等[15]和DA SILVA等[16]将模糊聚类分析方法引入到损伤识别当中,使用模糊C均值聚类算法对损伤进行归类,这种基于模糊聚类分析的损伤识别方法比非模糊聚类的算法具有更好的适用性。然而,上述基于聚类的无监督方法对于数据的分布类型有较高的要求,获取聚类特征也较为困难。综上所述,基于监督和无监督机器学习的结构损伤识别均已取得一系列成果。然而,基于监督学习的结构损伤识别,难以获得数量充足且能够满足模型训练要求的有标签样本数据[17-18]。对于无监督学习,尽管在模型的训练和测试阶段中不需要已知损伤类别的样本作为模型的输入,但由于缺乏损伤类别的先验知识的引导,无监督机器学习的建模准确性难以保证[19]
实际工程问题中,相较于数量庞大的无标签样本,有标签样本难以获取。这主要是由于人工标记样本的成本十分高昂[620]。为更好地利用这些数据,半监督学习近年来受到广泛关注。在损伤识别领域中,SHIM等[18]提出采用半监督学习的方法来识别结构损伤。半监督学习(semi-supervised learning,SSL)是一种能够有效利用无标签样本信息辅助训练的机器学习模型,通过同时使用无标签样本和有标签样本,可有效提高模型的精度和泛化性能[21]。SHIM等[18]应用半监督学习的思想,使用生成对抗神经网络来提高训练图像的分辨率,用于路面的损伤识别,获得更准确的结果。然而,上述方法仅输出损伤面积,没有对损伤具体位置进行进一步研究。BOUZENAD等[22]采用K-means聚类的方法来识别管道损伤,引入半监督学习的方式,利用有标签样本和无标签样本信息计算聚类中心和半径,来确定损伤预警阈值,获得较好的识别效果。然而,该方法本质上仍属于无监督聚类方法,仅识别管道是否损伤,并不能对损伤进行准确定位。
针对上述问题,为有效地利用无标签样本,提高模型预测精度,准确地定位损伤,本文以自训练半监督学习为模型整体训练框架,采用多层感知机(multilayer perceptron,MLP)神经网络为分类器,提出一种基于自训练半监督神经网络(self-training semi-supervised neural network,SSNN)的结构损伤识别模型。自训练方法是一种适应性强且简单的半监督学习方法[23-24],它通过在训练中使用带有伪标记的高置信度样本来提高性能。本文通过数值模拟和试验数据,验证该方法在损伤识别中的可行性和有效性。
半监督学习的目的是联合使用有标签数据集和无标签数据集,获得一个稳定的机器学习模型。自训练算法具有结构简单、预测精度高、应用范围广等优点,是典型的半监督机器学习模式。自训练半监督学习算法的基本思路如下:首先用初始有标签数据集训练初始分类器,对无标签样本进行分类;然后根据伪标签选择策略将置信度较高的样本挑选加入有标签数据集,扩大训练样本数据并重新训练分类器;最后继续重复上述训练分类、标注选择及扩充再训练的过程,直至有标签数据集不再扩大,即无法再选择出伪标签样本;此时获得更为精确稳定的分类器,并对测试集样本进行预测与评估。
挑选高置信度的伪标签样本是半监督学习流程的关键步骤,自训练算法利用基本分类器的概率估计来选择高置信度样本,以神经网络为例简要介绍本文的伪标签选择策略。当采用神经网络作为算法的基本分类器时,神经网络的输出层通过Softmax函数,将多个输出限定在[0,1]之间,并且满足概率之和为1;Softmax函数可以表示为
式中:zi为输出层第i个节点的输出值;n为输出层中节点个数,即分类的类别个数。式(1)计算值为神经网络对每个分类输出的概率值。
通过Softmax函数计算可得某样本数据x对应在每个分类上的概率(p1p2,…,pn),以此作为置信度的评价标准,即
式中:p1p2,…,pn为对应在1~n分类每个类别的概率值;m为概率最大值所对应的分类类别;θ为置信度阈值,一般取为0.75;该阈值在半监督学习中已得到验证[24]。自训练算法利用模型自身的置信预测,当训练过程中pmθ时,为未标记的数据生成伪标签,并添加至有标签样本集,从而实现训练集的扩充。
为更好地利用自训练算法来提升损伤识别的效果,需要选择表现优秀的机器学习模型作为自训练算法的基本分类器。神经网络已经广泛应用于损伤识别领域,它是由大量的、简单的处理单元(称为神经元)广泛地互相连接而形成的复杂网络系统,它反映人脑功能的许多基本特征,是一个高度复杂的非线性动力学习系统。
本文以MLP神经网络作为基分类器,MLP在单层神经网络的基础上引入一到多个隐藏层。隐藏层位于输入层和输出层之间,在MLP中通过大量的自适应学习可产生高度非线性的输入、输出对应关系。MLP的输出节点表达式为
式中:yiyj分别为输入节点和输出节点;wij为输入层中节点i与隐含层中节点j的连接权值。wij随学习进度变化的表达式为
式中:k为迭代次数;δj为隐含层中节点j的局部梯度;φ为学习率参数;β为动量常数。
神经网络隐藏层所采用的激活函数是修正线性单元(rectified linear unit,ReLU)函数,其表达式为
ReLU函数是一种神经网络中常用的激活函数,相较于传统激活函数有计算简便、实际收敛速度较快、有效避免梯度消失等优点。神经网络的梯度优化器采用自适应矩估计(adaptive moment estimation,ADAM),ADAM算法是对随机目标函数执行一阶梯度优化的算法,该算法基于适应性低阶矩估计。
ADAM算法与传统随机梯度下降不同。其不同于传统随机梯度下降以单一的学习率更新所有权重,而是通过计算梯度的一阶矩估计和二阶矩估计,为不同的参数设置相应的自适应学习率,从而减少误差,提高优化效率。ADAM算法公式为
式中:gt为随机目标函数的梯度;mt为梯度的一阶矩估计量;vt为梯度的二阶矩估计量;β1β2为超参数。ADAM算法中常给定超参数β1 =0.9、β2 =0.999。
对变量mtvt进行修正:
使用以上偏差修正后的变量,将模型参数中每个元素的学习率按元素运算重新调整后,使目标函数自变量中每个元素都分别拥有自己的学习率。
ADAM的参数更新公式为
式中:α为学习率;ε为了维持数值稳定性而添加的常数,通常取10-8
结构损伤特征的提取和选择是结构损伤识别的关键步骤。因此,选取敏感的结构损伤特征,能够准确、高效地进行损伤诊断。固有频率和振型等模态特性已被广泛应用于基于数据驱动的结构损伤识别。ADAMS等[25]首次使用频率变化来检测结构损伤;随后,CAWLEY等[26]对其进行了改进和扩展。由于固有频率的全局性,它并不是损伤定位的最佳特征。有研究表明,损伤状态与健康状态下的频率变化率是关于损伤位置的函数[27]。因此,以归一化频率变化率(normalized frequency change ratio,NFCR)和振型损伤特征指数(damage signature index,DSI)作为损伤特征(X[28]进行损伤识别,损伤特征X可表示为
式中,NFCR为由频率变化率(fractional frequency change,FFC)计算而来,其表达式为
式中,N为固有频率的模态阶数;第i阶的FFC表示为
式中,fuifdi为结构未损伤和损伤状态下的第i阶模态频率。
模态振型包含的结构空间信息对损伤定位很有价值,而且不容易受到环境影响[29]。因此,引入DSI作为损伤识别的特征向量。DSI是一种由振型和频率组成的复合损伤特征;研究表明,它与损伤位置有关[29],定义为
式中,ΦuiΦdi为结构未损伤和损伤状态下的第i阶模态振型向量。
为评价机器学习分类器的效果,引入F1-Score(F1分数)作为评价指标。在分类问题中,对某一类别的查准率定义为正确预测为该类别的样本数量占全部预测为该类别的样本数量的比值,查全率是正确预测为该类别的样本数量占全部真实标签为该类别的样本数量的比值,查准率P和查全率R可分别定义为
式中:Ntp为某一类别被正确预测的样本数量;Nfp为错误预测为某一类别的样本数量;Nfn为某一类别被错误预测的样本数量。
查准率和查全率是一对矛盾的度量。一般来说,查准率高时,查全率往往偏低;而查全率高时,查准率往往偏低。例如,当Ntp一定时,P越大,根据式(13),则Nfp越小;由于样本总数为定值,故Nfn越大,根据式(14)可知,R就越小。理想状态下的分类模型有较好查准率的同时也有较好查全率,因此更常用的度量分类器质量的评价指标是F1分数。F1分数又称平衡F分数,其定义为查准率和查全率的调和平均数,具体计算公式为
根据F1分数的定义可知,每个类别的F1分数均是一个取值为[0,1]区间的数值,类别的F1分数越大,分类器在该类别上的分类效果越好。对于多分类问题,可以对每个类别计算F1分数,进而将所有类的F1分数合并起来考虑,引入宏平均F1(Macro-F1),即将所有类别计算所得的F1分数取平均值。本研究中采用分类的准确率和Macro-F1来评估损伤识别分类的效果。
已有研究表明增加神经网络的深度有助于提升神经网络的效果[18]。因此,设计5层神经网络,包含输入层、输出层和3个隐藏层;3个隐藏层的神经元数量分别设计为[25,50,25],对于输入以及输出层的神经元数量,分别由训练输入样本的特征向量维数和分类的类别数目决定。选择ReLU函数作为每个隐藏层的激活函数;同时,采用ADAM算法作为对模型参数进行梯度下降的优化器。采用上述的5层神经网络作为自训练半监督学习的基分类器,基于SSNN的结构损伤流程如图1所示,其基本步骤如下。
1)对结构损伤的原始数据进行处理,通过工作模态分析提取模态频率及振型向量,并且按照式(10)~式(12)计算损伤敏感特征NFCR和DSI
2)按照损伤类别均匀划分初始训练集(有标签样本Dl)、扩充训练集(无标签样本Du)及测试集(T)。
3)利用初始训练集对MLP神经网络模型进行训练,得到初始分类模型。
4)利用训练得到的初始分类模型对扩充训练集中的无标签样本进行分类,从分类结果中选择置信度较高(pmθ)的样本作为伪标签样本加入初始训练集,并再次对MLP神经网络模型进行训练。
5)重复上述训练模型、对无标签样本添加伪标签、选择伪标签样本及重新训练模型的过程,直至通过pmθ挑选出的伪标签样本集为空集⌀,即有标签的训练集不再扩大,可得训练后的最终分类模型。
6)利用最终分类模型对测试集中的样本进行分类并评估分类器效果。
为验证本文提出的基于SSNN的损伤识别方法,采用空间桁架进行数值分析,如图2所示。空间桁架总跨度8 m,主桁间距和高度均为1 m,桁架材料密度为7.85×103 kg/m3,所有杆件均采用圆形截面,截面面积为0.002 m2。桁架杆件弹性模量的初始值为2.1×105 MPa,采用ANSYS建立空间桁架的有限元模型,桁架由102个单元和32节点组成。经计算,桁架前3阶固有频率分别为19.836、29.791、53.587 Hz。将空间桁架分为5个子结构,通过改变子结构的弹性模量(E1~E5),模拟结构不同位置的损伤,如图2所示。
对于单一位置损伤识别问题,设置5个损伤工况,每个损伤工况各模拟100个损伤样本,各样本损伤程度在(10%~90%)E之间随机折减,如表1所示。利用前3阶模态频率和振型,按照式(10)~式(12)计算损伤指标,构建数据集。
样本数据集以5%∶5%∶90%的比例划分为初始训练集、扩充训练集(无标签数据)以及测试集,以此模拟现实中缺乏有标签样本进行网络训练的情况。训练集和扩充训练集的数据均匀分布在5个损伤类别之中。具体而言,采用25个有标签的训练样本,25个无标签的扩充训练样本进行SSNN的训练,使用训练后的模型对450个测试样本进行分类并计算测试准确率来评估分类器效果。同时,采用25个有标签样本训练MLP神经网络作为对比,2种方法在训练时采用相同的超参数,根据机器学习中的常用取值,分别取学习率为0.002,正则化项权重为0.000 2。为避免机器学习算法的随机性和偶然性,2种方法均进行100次重复训练。同时,为加快收敛速度,在输入神经网络时采用Z-Score标准化对原始数据集进行预处理。
MLP神经网络和SSNN方法对空间桁架的损伤定位结果如图3所示。图中,x轴表示一阶频率的归一化频率变化率NFCR1y轴表示特征DSI1。D1~D5表示不同位置的损伤工况(实心圆),错误表示识别错误的损伤样本。由图3可知,MLP神经网络和SSNN方法在空间桁架损伤数据集上重复训练100次的平均分类准确率分别为94%和98%,即SSNN准确率提升4%;此外,MLP神经网络将部分工况D2的样本错误标记为工况D4,其原因是D2与D4属于对称位置的损伤工况,在训练样本较少的情况下更易混淆。然而,同样的损伤样本在SSNN中被更好地区分和定位,这说明SSNN方法通过自训练半监督的方式扩大训练样本量,可以更好地识别损伤样本。
2种分类模型在测试集上的分类结果的混淆矩阵如图4所示。混淆矩阵的每一列代表预测类别标签,每一行代表数据的真实类别标签。由于空间桁架案例设置5种单损伤工况,即5个分类问题,在测试集上每个类别的样本数量均为90。由混淆矩阵的定义可知,矩阵的主对角线元素值均为90时,分类效果达到最佳,因此混淆矩阵的分类元素值越向着对角线集中,代表分类效果越好。由图4可知,SSNN方法在工况D2与D3上分类正确的样本数量相较于MLP有明显的提升。以工况D3为例,MLP将其中的9个样本错误预测为D5,而在SSNN方法中这些错误分类被很好地预测。
为进一步评估损伤识别分类器的效果,图5绘制了按式(15)计算得到每个损伤工况类别的F1分数和总体的F1分数。图中x轴表示5种损伤工况以及所有工况的宏平均,y轴表示对应工况分类的F1分数,MLP和ST分别表示MLP神经网络和SSNN方法。由图5可知,在工况D2~D5下,SSNN方法分类的F1分数高于MLP神经网络的F1分数,这表明SSNN方法在工况D2~D5上的识别效果优于MLP神经网络。此外,SSNN方法的宏平均F1分数相较于MLP神经网络也要更高,即SSNN方法在测试集上损伤分类的整体效果优于MLP神经网络。这是因为,与MLP方法相比,SSNN方法通过不断挑选高置信度的伪标签样本加入训练集中,充分利用在MLP方法中未使用到的无标签数据集的信息,使得SSNN方法获得更好地预测效果。
为考查模态参数测试误差对所提方法的影响,在模态参数识别时添加不同比例的白噪声作为误差。采用不同噪声水平下的模态识别结果制作数据集,采用SSNN方法识别结构损伤;同时,采用MLP神经网络识别结果作为对比。图6绘制了在不同模态识别误差下,2种方法在空间桁架单一位置损伤识别的准确率变化趋势图。由图6可知,随着模态识别误差不断增大,2种方法的准确率均不断降低,分别从93%和97%降至82%和89%,这表明模态识别的误差对2种方法识别结构损伤具有一定的影响。同时可以看出,在不同的误差水平下,SSNN方法相较于MLP神经网络均处于较高的损伤识别率;并且随着误差百分比的增大,SSNN方法相对于MLP神经网络的优势不断增大,从4%扩大至7%。这表明SSNN方法相较于MLP神经网络具有更高的抗噪性能和鲁棒性。
实际工程结构的损伤通常是多个位置同时发生,为考查SSNN方法在多位置损伤识别中的表现,数值算例进一步研究2个位置损伤识别问题。对于2个位置损伤问题,假定空间桁架的5个位置产生2个位置的随机损伤,损伤程度在对应的位置弹性模量(10%~90%)E之间随机折减,表2给出了具体损伤工况的设置。
空间桁架的2个位置损伤共设置10个分类,测试集中每一类的样本数量均为20个,共计200个测试样本。图7展示了2种方法在测试集上分类结果的混淆矩阵。以工况S10为例,在MLP神经网络中,将2个样本误判为工况S7和3个样本误判为工况S9;这些错误在SSNN方法中均被正确的识别。同样的,SSNN方法在工况S1~工况S3以及工况S7~工况S10上的损伤识别效果明显优于MLP神经网络。进一步统计各工况损伤识别结果可知,2种损伤识别方法在空间桁架2个损伤数据集的识别准确率分别为82%和91%;即SSNN方法准确率比MLP神经网络提升9%。
图8绘制了每个损伤工况以及总体的F1分数,对于大部分2个位置损伤工况,SSNN方法在这些工况上的F1分数高于MLP神经网络的F1分数。此外,SSNN方法和MLP神经网络方法的宏平均F1分数分别为0.91和0.82,表明SSNN方法损伤定位效果更好。
空间桁架的损伤识别数值案例表明,在2个位置损伤识别中,SSNN方法的识别准确率相较于MLP神经网络有9%的提升;而在单一位置损伤识别中,尽管MLP神经网络已经达到较高的识别准确率(94%),采用SSNN方法仍有4%的提升(达到98%)。因此,在相同条件下,SSNN方法在损伤识别的效果优于MLP神经网络。在损伤位置的识别上,SSNN方法通过自训练半监督算法为无标签样本添加伪标签,筛选置信度高的样本并将其添加到训练过程中,能够较明显的提升预测精度,从而更好地识别损伤位置。
为考查基于SSNN的损伤识别方法在实测数据中的表现,使用Los Alamos National Laboratory发布的3层框架结构的试验数据进行研究[30]。该3层框架结构由铝柱和铝板通过螺栓连接而成如图9所示。各层铝板的结构尺寸为305 mm×305 mm×25 mm,4根铝柱的结构尺寸为177 mm×25 mm×6 mm。
采用电动激振器在基底层沿着结构中心线横向施加白噪声激励,并使用布置在每一层的加速度传感器收集响应,采样间隔为3.1 ms,对应采样频率为322.58 Hz,如图9所示。对于每种结构损伤工况,考虑到激发和结构性质的变化,进行10次振动测试。每次测试持续时间25.3 s,得到的加速度响应可用于提取和计算模态参数。
本文从Figueredo的数据库[30]中选择3层框架的3种损伤工况,以验证SSNN方法在损伤定位中的有效性,如表3所示。试验中通过折减铝柱的刚度来模拟实际损伤,所选工况包括第一至三层柱刚度折减的结构损伤(分别用E1~E3表示)。根据ZHANG等[10]的计算,每种损伤类别包含2种损伤程度:①某层楼板的一柱刚度降低87.5%,即层间刚度ki降低22%(i =1,2,3);②某层楼板两柱刚度降低87.5%,即层间刚度ki降低44%(i =1,2,3)。
利用随机子空间法识别结构模态参数,并进一步计算损伤特征值。图10比较了结构损伤前后的前3阶模态振型。图中,E0表示未损伤状态,研究时已排除由刚体运动引起的模态。可以看出,在不同的损伤情况下,每个振型都有明显的振型变化。因此,振型变化信息可以用来评估结构的损伤。由于样本数据较少,数据分配为20%的有标签数据作为初始训练集,20%的未标记数据作为扩充训练集,用于自训练半监督学习,60%的数据作为测试集;同时采用MLP神经网络识别的损伤结果作为对比。
图11对比了MLP和SSNN方法对3层框架的损伤定位结果。如图11所示,x轴与y轴分别表示降维后的特征一、特征二,E1~E3表示损伤类型,错误表示识别有误的损伤样本。所有的损伤样本数据被绘制在一个二维特征空间中,与损伤位置对应的损伤工况用不同颜色的实心圆标记。由图11可知,训练好的神经网分络分类器在工况E2的损伤定位错误结果,在SSNN方法中被准确定位,这说明在试验数据中SSNN方法的性能优于MLP神经网络,具有实际意义。
图12展示了2种方法在3层框架公开数据集损伤识别的混淆矩阵。由混淆矩阵的定义可知,图中SSNN方法能够纠正在MLP神经网络中被误判为工况E3的2个损伤样本(实际为工况E2)。这表明SSNN方法在工况E2、工况E3的损伤识别上相较于MLP神经网络有着更优秀的效果。
根据2种方法的分类结果,按照式(15)可以计算得到每个损伤工况类别的F1分数以及总体的宏平均F1分数,其结果如图13所示。在工况E2、工况E3下,SSNN方法的F1分数高于MLP神经网络的F1分数。对于总体宏平均-F1分数而言,SSNN方法的宏平均-F1分数相较于MLP神经网络也要更高,因此,SSNN方法在测试集上损伤分类的效果优于MLP神经网络。
本文提出基于SSNN的结构损伤识别框架,利用NFCR和DSI作为损伤指标输入MLP神经网络,并采用自训练半监督的训练方式解决损伤识别中有标签样本不足的问题,用于结构损伤识别。通过空间桁架的数值案例和3层框架的试验案例验证所提出SSNN方法的有效性,得到以下结论:
1)采用NFCR和DSI作为损伤指标,能够准确地反映结构的损伤位置,将其用于训练MLP神经网络的输入特征,能够取得较好的损伤定位效果。
2)基于SSNN的损伤识别方法能在有标签样本较少的情况下,取得较好的损伤识别效果。在初始训练集样本相同的情况下,相较于MLP神经网络,SSNN方法的损伤识别效果更好,单一位置损伤识别准确率提升约4%;2个位置损伤识别准确率提升约9%。
3)SSNN方法为结构损伤识别的样本扩充提供新思路,并很好地解决损伤识别中有标签样本不足的问题。基于SSNN的损伤识别方法相对于只采用有标签样本的MLP神经网络来说,能够利用无标签样本进行辅助扩充训练,在有标签样本较少的条件下,可以提高分类模型预测精度和泛化性能。
  • 国家自然科学基金项目(51608408)
参考文献 引证文献
排序方式:
[1]
赵少杰,唐细彪,任伟新. 桥梁事故的统计特征分析及安全风险防控原则[J]. 铁道工程学报201734(5):59-64.
ZHAO Shaojie, TANG Xibiao, REN Weixin. The statistical characteristics analysis of bridge accidents and prevention principle of security risk[J]. Journal of Railway Engineering Society, 2017, 34(5): 59-64. (in Chinese)
[2]
骆勇鹏,王林堃,廖飞宇,. 基于一维卷积神经网络的结构损伤识别[J]. 地震工程与工程振动202141(4):145-156.
LUO Yongpeng, WANG Linkun, LIAO Feiyu, et al. Vibration-based structural damage identification by 1-dimensional convolutional neural network[J]. Earthquake Engineering and Engineering Dynamics, 2021, 41(4): 145-156. (in Chinese)
[3]
李延强,张阳. 基于支持向量机与分层遗传算法的斜拉桥全结构损伤分步识别[J]. 地震工程与工程振动201535(6):71-77.
LI Yanqiang, ZHANG Yang. Step-by-step damage detection for cable-stayed bridge based on support vector machine and hierarchic genetic algorithm[J]. Earthquake Engineering and Engineering Dynamics, 2015, 35(6): 71-77. (in Chinese)
[4]
孙利民,尚志强,夏烨. 大数据背景下的桥梁结构健康监测研究现状与展望[J]. 中国公路学报201932(11):1-20.
SUN Limin, SHANG Zhiqiang, XIA Ye. Development and prospect of bridge structural health monitoring in the context of big data[J]. China Journal of Highway and Transport, 2019, 32(11): 1-20. (in Chinese)
[5]
FLAH M, NUNEZ I, BEN CHAABENE W, et al. Machine learning algorithms in civil structural health monitoring: A systematic review[J]. Archives of Computational Methods in Engineering, 2021, 28(4): 2621-2643.
[6]
CHUN C J, RYU S K. Road surface damage detection using fully convolutional neural networks and semi-supervised learning[J]. Sensors, 2019, 19(24): 5501.
[7]
黄发明,潘李含,姚池,. 基于半监督机器学习的滑坡易发性预测建模[J]. 浙江大学学报(工学版)202155(9):1705-1713.
HUANG Faming, PAN Lihan, YAO Chi, et al. Landslide susceptibility prediction modelling based on semi-supervised machine learning[J]. Journal of Zhejiang University (Engineering Science)202155(9): 1705-1713. (in Chinese)
[8]
徐菁,曲丽敏,卢翠萍. 大跨度空间网格结构的健康监测系统[J]. 兰州理工大学学报201642(4):128-133.
XU Jing, QU Limin, LU Cuiping. Health monitoring system with large-span space-netted structure[J]. Journal of Lanzhou University of Technology, 2016, 42(4): 128-133. (in Chinese)
[9]
SATPAL S B, GUHA A, BANERJEE S. Damage identification in aluminum beams using support vector machine: Numerical and experimental studies[J]. Structural Control and Health Monitoring, 2016, 23(3): 446-457.
[10]
ZHANG Z M, SUN C. Structural damage identification via physics-guided machine learning: A methodology integrating pattern recognition with finite element model updating[J]. Structural Health Monitoring, 2021, 20(4): 1675-1688.
[11]
ROGERS T J, WORDEN K, FUENTES R, et al. A Bayesian non-parametric clustering approach for semi-supervised structural health monitoring[J]. Mechanical Systems and Signal Processing, 2019, 119: 100-119.
[12]
WORDEN K, MANSON G, ALLMAN D. Experimental validation of a structural health monitoring methodology: Part I. novelty detection on a laboratory structure[J]. Journal of Sound and Vibration, 2003, 259(2): 323-343.
[13]
MANSON G, WORDEN K, ALLMAN D. Experimental validation of a structural health monitoring methodology: Part II. novelty detection on a gnat aircraft[J]. Journal of Sound and Vibration, 2003, 259(2): 345-363.
[14]
XU D J, LIU P, CHEN Z C. A waveform clustering method for damage mode identification for composite laminates under hygrothermal environment[J]. Engineering Fracture Mechanics, 2020, 239: 107290.
[15]
MAHARAJ E A, D’URSO P. Fuzzy clustering of time series in the frequency domain[J]. Information Sciences, 2011, 181(7): 1187-1211.
[16]
DA SILVA S, DIAS M Jr, LOPES V Jr. Damage detection in a benchmark structure using AR-ARX models and statistical pattern recognition[J]. Journal of the Brazilian Society of Mechanical Sciences and Engineering, 2007, 29(2): 174-184.
[17]
CHA Y J, CHOI W, BÜYÜKÖZTÜRK O. Deep learning-based crack damage detection using convolutional neural networks[J]. Computer-Aided Civil and Infrastructure Engineering, 2017, 32(5): 361-378.
[18]
SHIM S, KIM J, LEE S W, et al. Road damage detection using super-resolution and semi-supervised learning with generative adversarial network[J]. Automation in Construction, 2022, 135: 104139.
[19]
HUANG F, CAO Z, GUO J, et al. Comparisons of heuristic, general statistical and machine learning models for landslide susceptibility prediction and mapping[J]. Catena (Giessen), 2020, 191: 104580.
[20]
霍士伟,郭圣明,唐宇波. 基于自训练半监督学习的战场态势评估模型[J]. 舰船电子工程202141(9):93-96, 107.
HUO Shiwei, GUO Shengming, TANG Yubo. Battlefield situation assessment model based on self-training semi-supervised learning[J]. Ship Electronic Engineering, 2021, 41(9): 93-96, 107. (in Chinese)
[21]
ZHOU Z, LI M. Semi-supervised learning by disagreement[J]. Knowledge and Information Systems, 2010, 24(3): 415-439.
[22]
BOUZENAD A E, EL MOUNTASSIR M, YAACOUBI S, et al. A semi-supervised based K-means algorithm for optimal guided waves structural health monitoring: A case study[J]. Inventions, 2019, 4(1): 17.
[23]
PEI Z X, LIN R H, ZHANG X B, et al. CFM: A consistency filtering mechanism for road damage detection[C]∥2020 IEEE International Conference on Big Data (Big Data). Atlanta, GA, USA. Piscataway, NJ, USA: IEEE, 2021: 5584-5591.
[24]
TANHA J, VAN SOMEREN M, AFSARMANESH H. Semi-supervised self-training for decision tree classifiers[J]. International Journal of Machine Learning and Cybernetics, 2017, 8(1): 355-370.
[25]
ADAMS R D, CAWLEY P, PYE C J, et al. A vibration technique for non-destructively assessing the integrity of structures[J]. Journal of Mechanical Engineering Science, 1978, 20(2): 93-100.
[26]
CAWLEY P, ADAMS R D. The location of defects in structures from measurements of natural frequencies[J]. The Journal of Strain Analysis for Engineering Design, 1979, 14(2): 49-57.
[27]
KIM J T, RYU Y S, CHO H M, et al. Damage identification in beam-type structures: Frequency-based method vs mode-shape-based method[J]. Engineering Structures, 2003, 25(1): 57-67.
[28]
LAM H F, KO J M, WONG C W. Localization of damaged structural connections based on experimental modal and sensitivity analysis[J]. Journal of Sound and Vibration, 1998, 210(1): 91-115.
[29]
FARRAR C R, JAMES G H III. System identification from ambient vibration measurements on a bridge[J]. Journal of Sound and Vibration, 1997, 205(1): 1-18.
[30]
FIGUEIREDO E, PARK G, FIGUEIRAS J, et al. Structural health monitoring algorithm comparisons using standard data sets[J]. Algorithms, 2009, 3(1): 1320-1327.
2024年第44卷第2期
PDF下载
66
35
引用本文
BibTeX
文章信息
doi: 10.13197/j.eeed.2024.0205
  • 接收时间:2022-09-05
  • 首发时间:2026-03-30
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2022-09-05
  • 修回日期:2023-02-05
基金
国家自然科学基金项目(51608408)
作者信息
    武汉理工大学 土木工程与建筑学院,湖北 武汉 430070
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/dzgcygczd/CN/10.13197/j.eeed.2024.0205
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏