Article(id=1149744609946944231, tenantId=1146029695717560320, journalId=1146123302524792850, issueId=1149744594201523010, articleNumber=null, orderNo=null, doi=10.3969/j.issn.1672-6073.2024.03.011, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1702483200000, receivedDateStr=2023-12-14, revisedDate=1709913600000, revisedDateStr=2024-03-09, acceptedDate=null, acceptedDateStr=null, onlineDate=1752050076232, onlineDateStr=2025-07-09, pubDate=null, pubDateStr=null, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1752050076232, onlineIssueDateStr=2025-07-09, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1752050076232, creator=13701087609, updateTime=1752050076232, updator=13701087609, issue=Issue{id=1149744594201523010, tenantId=1146029695717560320, journalId=1146123302524792850, year='2024', volume='37', issue='3', pageStart='1', pageEnd='177', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1752050072478, creator=13701087609, updateTime=1753780147518, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1157001054903685255, tenantId=1146029695717560320, journalId=1146123302524792850, issueId=1149744594201523010, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1157001054903685256, tenantId=1146029695717560320, journalId=1146123302524792850, issueId=1149744594201523010, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=82, endPage=89, ext={EN=ArticleExt(id=1149744610370568943, articleId=1149744609946944231, tenantId=1146029695717560320, journalId=1146123302524792850, language=EN, title=Improved Prediction Model of Metro Wheel Wear Based on BiLSTM Network, columnId=1152669335454658940, journalTitle=Urban Rapid Rail Transit, columnName=Academic Discussion, runingTitle=null, highlight=null, articleAbstract=

In order to address the issue of longterm dependence caused by the extended time span of wheel wear data and improve the prediction accuracy, an improved BiLSTM metro wheel wear prediction model is proposed by optimizing Bidirectional long shortterm memory network (BiLSTM) with Sparrow search algorithm (SSA). Firstly, the hyperparameters of the BiLSTM algorithm, such as the number of neurons, iteration count, input batch size, and learning rate, are optimized using the SSA. This optimization process is conducted within a specified range to obtain the optimal values of these hyperparameters. This optimization process aims to obtain the optimal parameter values. Subsequently, the SSABiLSTM network model is constructed using these optimal parameter values to predict and analyze wheel wear. Tread wear and flange wear are taken as the research objects, and the measured historical wear data of wheel No.1 of the metro’s carriage # 1 are used as inputs to metro and validate the model, and compare the prediction results with those of MLP, LSTM, BiLSTM and SSA-LSTM models. The results show that the improved bidirectional long short-term memory network model has higher wear prediction accuracy, and the mean absolute percentage error (MAPE) of tread wear is reduced by 13.28%, 10.32%, and 1.47%, and flange wear by 9.5%, 0.46%, and 0.02%. The wear of the No. 1 wheel of the same metro No.2 and No.4 cars is predicted and compared with the measured wear data. The average absolute percentage error of tread wear is 1.34% and 1.42%, respectively, and the average absolute percentage error of rim wear is 0.18% and 0.19%, respectively. The results confirm that the model exhibits strong generalization capabilities. The wheel wear prediction model based on improved BiLSTM network (SSA-Bi-LSTM) has high prediction accuracy and good generalization, which provides theoretical support for the intelligent management of metro wheelsets and prolongs wheel service life.

, correspAuthors=Aihua ZHU, 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=Aihua ZHU, Yang BAI, Tangbo BAI, Yali WANG, Caisheng ZHANG, Anyan LI), CN=ArticleExt(id=1149744628619985378, articleId=1149744609946944231, tenantId=1146029695717560320, journalId=1146123302524792850, language=CN, title=基于改进 BiLSTM 网络的地铁车轮磨耗预测模型, columnId=1152669335668568445, journalTitle=都市快轨交通, columnName=学术探讨, runingTitle=null, highlight=null, articleAbstract=

针对地铁车轮磨耗数据时间跨度较长引起的长期依赖问题,为了进一步提升预测精度,提出一种将麻雀搜索算法(sparrow search algorithm,SSA)优化双向长短期记忆网络(bidirectional long short term memory, BiLSTM)的改进 BiLSTM(SSABiLSTM)网络模型,用于地铁车轮磨耗预测。首先,利用麻雀搜索算法对双向长短期记忆网络算法的神经元个数、迭代次数、输入批量和学习率等超参数在给定范围内进行寻优,得到参数最优值;然后,以参数最优值来构建改进 BiLSTM 网络模型,对车轮磨耗进行预测分析;最后,以车轮踏面磨耗和轮缘磨耗作为研究对象,将某地铁1车厢1号车轮的现场实测历史磨耗数据作为输入,对该模型进行训练及验证分析,并与多层感知机(multilayer perceptron, MLP)、LSTM、BiLSTM 以及 SSALSTM 模型的预测结果进行对比。研究结果表明:SSABiLSTM模型的车轮磨耗预测精度更高,与LSTM、BiLSTM 以及SSALSTM 网络模型相比,踏面磨耗的平均绝对百分误差(mean absolute percentage error,MAPE)分别降低了 13.28%、10.32%、1.47%,轮缘磨耗分别降低了9.5%、0.46%、0.02%;分别对同一地铁2号、4号车厢的1号位置车轮磨耗进行预测,并与磨耗实测数据进行对比,踏面磨耗的平均绝对百分比误差分别为1.34%、1.42%,轮缘磨耗的平均绝对百分比误差分别为0.18%、0.19%,验证了本文所提模型具有良好的泛化性,为地铁轮对智能化管理提供理论支持,延长车轮使用寿命。

, correspAuthors=朱爱华, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=RW2+NNRZo6KuIPyyOmgAhQ==, magXml=P02wio6l6beKziCqNB6DDQ==, pdfUrl=null, pdf=HQE6iAni9tenMwAkNR+iRg==, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=vq14gfzUoPTAZs9XM+cvhQ==, mapNumber=null, authorCompany=null, fund=null, authors=

朱爱华,女,教授,博士,主要研究方向为车辆关键系统服役性能与可靠性评估,

, authorsList=朱爱华, 白杨, 白堂博, 王雅莉, 张财胜, 李安琰)}, authors=[Author(id=1154040266974552752, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=zhuaihua@bucea.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1154040267104576179, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040266974552752, language=EN, stringName=Aihua ZHU, firstName=Aihua, middleName=null, lastName=ZHU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040267217822388, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040266974552752, language=CN, stringName=朱爱华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044, bio={"content":"

朱爱华,女,教授,博士,主要研究方向为车辆关键系统服役性能与可靠性评估,

"}, bioImg=null, bioContent=

朱爱华,女,教授,博士,主要研究方向为车辆关键系统服役性能与可靠性评估,

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)])]), Author(id=1154040267318485687, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, 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=1154040267461092026, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267318485687, language=EN, stringName=Yang BAI, firstName=Yang, middleName=null, lastName=BAI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040267536589502, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267318485687, language=CN, stringName=白杨, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)])]), Author(id=1154040267616281281, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, 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=1154040267704361665, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267616281281, language=EN, stringName=Tangbo BAI, firstName=Tangbo, middleName=null, lastName=BAI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040267784053445, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267616281281, language=CN, stringName=白堂博, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)])]), Author(id=1154040267855356615, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, orderNo=3, 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=1154040267930854089, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267855356615, language=EN, stringName=Yali WANG, firstName=Yali, middleName=null, lastName=WANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 Beijing Subway Operation Co., Ltd. Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040268014740171, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040267855356615, language=CN, stringName=王雅莉, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 北京市地铁运营有限公司 北京 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266869695146, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=2, ext=[AuthorCompanyExt(id=1154040266878083755, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266869695146, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Beijing Subway Operation Co., Ltd. Beijing 100044), AuthorCompanyExt(id=1154040266886472364, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266869695146, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 北京市地铁运营有限公司 北京 100044)])]), Author(id=1154040268069266125, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, orderNo=4, 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=1154040268144763600, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040268069266125, language=EN, stringName=Caisheng ZHANG, firstName=Caisheng, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040268220261073, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040268069266125, language=CN, stringName=张财胜, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)])]), Author(id=1154040268291564244, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, orderNo=5, 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=1154040268425781976, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040268291564244, language=EN, stringName=Anyan LI, firstName=Anyan, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1154040268501279450, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, authorId=1154040268291564244, language=CN, stringName=李安琰, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)])])], keywords=[Keyword(id=1154040268975235809, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, orderNo=1, keyword=metro), Keyword(id=1154040269042344676, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, orderNo=2, keyword=wear prediction), Keyword(id=1154040269126230756, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, orderNo=3, keyword=sparrow search algorithm), Keyword(id=1154040269189145317, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, orderNo=4, keyword=bidirectional long short time memory networks), Keyword(id=1154040269256254182, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, orderNo=1, keyword=地铁), Keyword(id=1154040269331751656, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, orderNo=2, keyword=磨耗预测), Keyword(id=1154040269398860522, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, orderNo=3, keyword=麻雀搜索算法), Keyword(id=1154040269482746604, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, orderNo=4, keyword=双向长短期记忆网络)], refs=[Reference(id=1154040271739282216, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=33, issue=4, pageStart=44, pageEnd=49, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=郭涛, 王宁, 刘永强, journalName=石家庄铁道大学学报(自然科学版), refType=null, unstructuredReference=郭涛, 王宁, 刘永强. 车轮踏面常规磨耗对高速列车动力学性能的影响[J]. 石家庄铁道大学学报(自然科学版), 2020, 33(4): 44-49., articleTitle=车轮踏面常规磨耗对高速列车动力学性能的影响, refAbstract=null), Reference(id=1154040271798002473, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=33, issue=4, pageStart=44, pageEnd=49, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=GUO Tao, WANG Ning, LIU Yongqiang, journalName=Journal of Shijiazhuang Tiedao University (Natural Science Edition), refType=null, unstructuredReference= GUO Tao , WANG Ning , LIU Yongqiang . Influence of wheel wear on dynamic performance of high-speed train[J]. Journal of Shijiazhuang Tiedao University (Natural Science Edition), 2020, 33(4): 44-49., articleTitle=Influence of wheel wear on dynamic performance of high-speed train, refAbstract=null), Reference(id=1154040271844139818, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2006, volume=261, issue=11/12, pageStart=1253, pageEnd=1264, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=BRAGHIN F, LEWIS R, DWYER-JOYCE R S, journalName=Wear, refType=null, unstructuredReference= BRAGHIN F , LEWIS R , DWYER-JOYCE R S , et al . A mathematical model to predict railway wheel profile evolution due to wear[J]. Wear, 2006, 261(11/12): 1253-1264., articleTitle=A mathematical model to predict railway wheel profile evolution due to wear, refAbstract=null), Reference(id=1154040271894471467, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2016, volume=366/367, issue=null, pageStart=146, pageEnd=156, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=TAO GQ, WEN Z F, ZHAO X, journalName=Wear, refType=null, unstructuredReference= TAO GQ , WEN Z F , ZHAO X , et al . Effects of wheel-rail contact modelling on wheel wear simulation[J]. Wear, 2016, 366/367: 146-156., articleTitle=Effects of wheel-rail contact modelling on wheel wear simulation, refAbstract=null), Reference(id=1154040271953191724, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2018, volume=38, issue=4, pageStart=462, pageEnd=467, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=王雪萍, 张军, 马贺, journalName=摩擦学学报, refType=null, unstructuredReference=王雪萍, 张军, 马贺. 高速列车车轮踏面磨耗预测方法的研究[J]. 摩擦学学报, 2018, 38(4): 462-467., articleTitle=高速列车车轮踏面磨耗预测方法的研究, refAbstract=null), Reference(id=1154040272003523373, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2018, volume=38, issue=4, pageStart=462, pageEnd=467, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=WANG Xueping, ZHANG Jun, MA He, journalName=Tribology, refType=null, unstructuredReference= WANG Xueping , ZHANG Jun , MA He . Prediction method of wheel wear of high speed train[J]. Tribology, 2018, 38(4): 462-467., articleTitle=Prediction method of wheel wear of high speed train, refAbstract=null), Reference(id=1154040272062243630, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2023, volume=26, issue=10, pageStart=109, pageEnd=115, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=张梓鸿, 赵正阳, 王文斌, journalName=城市轨道交通研究, refType=null, unstructuredReference=张梓鸿, 赵正阳, 王文斌, 等. 基于卷积神经网络的城市轨道交通轨道几何不平顺指标预测[J]. 城市轨道交通研究, 2023, 26(10): 109-115., articleTitle=基于卷积神经网络的城市轨道交通轨道几何不平顺指标预测, refAbstract=null), Reference(id=1154040272116769585, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2023, volume=26, issue=10, pageStart=109, pageEnd=115, url=null, language=null, rfNumber=[5], rfOrder=7, authorNames=ZHANG Zihong, ZHAO Zhengyang, WANG Wenbin, journalName=Urban mass transit, refType=null, unstructuredReference= ZHANG Zihong , ZHAO Zhengyang , WANG Wenbin , et al. Prediction of urban rail transit track geometry irregularity indicators based on convolutional neural networks[J]. Urban mass transit, 2023, 26(10): 109-115., articleTitle=Prediction of urban rail transit track geometry irregularity indicators based on convolutional neural networks, refAbstract=null), Reference(id=1154040272175489842, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=39, issue=11, pageStart=80, pageEnd=84, url=null, language=null, rfNumber=[6], rfOrder=8, authorNames=王慧君, 胡定玉, 方宇, journalName=测控技术, refType=null, unstructuredReference=王慧君, 胡定玉, 方宇, 等. 基于梯度提升决策树的车轮轮缘厚度磨耗预测[J]. 测控技术, 2020, 39(11): 80-84., articleTitle=基于梯度提升决策树的车轮轮缘厚度磨耗预测, refAbstract=null), Reference(id=1154040272238404403, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=39, issue=11, pageStart=80, pageEnd=84, url=null, language=null, rfNumber=[6], rfOrder=9, authorNames=WANG Huijun, HU Dingyu, FANG Yu, journalName=Measurement & control technology, refType=null, unstructuredReference= WANG Huijun , HU Dingyu , FANG Yu , et al. Prediction of wheel flange thickness abrasion based on gradient boosting decision tree[J]. Measurement & control technology, 2020, 39(11): 80-84., articleTitle=Prediction of wheel flange thickness abrasion based on gradient boosting decision tree, refAbstract=null), Reference(id=1154040272297124660, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2007, volume=32, issue=12, pageStart=20, pageEnd=22, url=null, language=null, rfNumber=[7], rfOrder=10, authorNames=王文健, 陈明韬, 刘启跃, journalName=润滑与密封, refType=null, unstructuredReference=王文健, 陈明韬, 刘启跃. 基于BP神经网络的钢轨磨损量预测[J]. 润滑与密封, 2007, 32(12): 20-22., articleTitle=基于BP神经网络的钢轨磨损量预测, refAbstract=null), Reference(id=1154040272372622133, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2007, volume=32, issue=12, pageStart=20, pageEnd=22, url=null, language=null, rfNumber=[7], rfOrder=11, authorNames=WANG Wenjian, CHEN Mingtao, LIU Qiyue, journalName=Lubrication engineering, refType=null, unstructuredReference= WANG Wenjian , CHEN Mingtao , LIU Qiyue . Prediction of wear volumes of rail steel based on BP neural network[J]. Lubrication engineering, 2007, 32(12): 20-22., articleTitle=Prediction of wear volumes of rail steel based on BP neural network, refAbstract=null), Reference(id=1154040272431342391, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2013, volume=30, issue=8, pageStart=15, pageEnd=20, url=null, language=null, rfNumber=[8], rfOrder=12, authorNames=王平, 王彩芸, 王文健, journalName=机械设计, refType=null, unstructuredReference=王平, 王彩芸, 王文健, 等. 基于PSO混合算法优化BP网络在钢轨磨损量预测中的应用[J]. 机械设计, 2013, 30(8): 15-20., articleTitle=基于PSO混合算法优化BP网络在钢轨磨损量预测中的应用, refAbstract=null), Reference(id=1154040272498451259, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2013, volume=30, issue=8, pageStart=15, pageEnd=20, url=null, language=null, rfNumber=[8], rfOrder=13, authorNames=WANG Ping, WANG Caiyun, WANG Wenjian, journalName=Journal of machine design, refType=null, unstructuredReference= WANG Ping , WANG Caiyun , WANG Wenjian , et al. Application of BP ANN optimized by hybrid algorithm based on PSO in wear volume prediction of rail steel[J]. Journal of machine design, 2013, 30(8): 15-20., articleTitle=Application of BP ANN optimized by hybrid algorithm based on PSO in wear volume prediction of rail steel, refAbstract=null), Reference(id=1154040272578143037, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2022, volume=54, issue=6, pageStart=1720, pageEnd=1731, url=null, language=null, rfNumber=[9], rfOrder=14, authorNames=王美琪, 王艺, 陈恩利, journalName=力学学报, refType=null, unstructuredReference=王美琪, 王艺, 陈恩利, 等. 基于恒等映射多层极限学习机的高速列车踏面磨耗预测模型[J]. 力学学报, 2022, 54(6): 1720-1731., articleTitle=基于恒等映射多层极限学习机的高速列车踏面磨耗预测模型, refAbstract=null), Reference(id=1154040272636863295, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2022, volume=54, issue=6, pageStart=1720, pageEnd=1731, url=null, language=null, rfNumber=[9], rfOrder=15, authorNames=WANG Meiqi, WANG Yi, CHEN Enli, journalName=Chinese journal of theoretical and applied mechanics, refType=null, unstructuredReference= WANG Meiqi , WANG Yi , CHEN Enli , et al. High-speed train tread wear prediction model based on i-ml-elm[J]. Chinese journal of theoretical and applied mechanics, 2022, 54(6): 1720-1731., articleTitle=High-speed train tread wear prediction model based on i-ml-elm, refAbstract=null), Reference(id=1154040272691389250, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=1997, volume=9, issue=8, pageStart=1735, pageEnd=1780, url=null, language=null, rfNumber=[10], rfOrder=16, authorNames=HOCHREITER S, SCHMIDHUBER J, journalName=Neural computation, refType=null, unstructuredReference= HOCHREITER S , SCHMIDHUBER J . Long short-term memory[J]. Neural computation, 1997, 9(8): 1735-1780., articleTitle=Long short-term memory, refAbstract=null), Reference(id=1154040272745915203, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2023, volume=46, issue=1, pageStart=77, pageEnd=84, url=null, language=null, rfNumber=[11], rfOrder=17, authorNames=吕开云, 邱万锦, 龚循强, journalName=东华理工大学学报(自然科学版), refType=null, unstructuredReference=吕开云, 邱万锦, 龚循强, 等. 基于长短期记忆神经网络的城市交通速度预测[J]. 东华理工大学学报(自然科学版), 2023, 46(1): 77-84., articleTitle=基于长短期记忆神经网络的城市交通速度预测, refAbstract=null), Reference(id=1154040272800441156, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2023, volume=46, issue=1, pageStart=77, pageEnd=84, url=null, language=null, rfNumber=[11], rfOrder=18, authorNames=LYU Kaiyun, QIU Wanjin, GONG Xunqiang, journalName=Journal of East China University of Technology (Natural Science), refType=null, unstructuredReference= LYU Kaiyun , QIU Wanjin , GONG Xunqiang , et al. Urban traffic speed prediction based on long short-term memory neural network[J]. Journal of East China University of Technology (Natural Science), 2023, 46(1): 77-84., articleTitle=Urban traffic speed prediction based on long short-term memory neural network, refAbstract=null), Reference(id=1154040272859161414, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2023, volume=6, issue=8, pageStart=1, pageEnd=15, url=null, language=null, rfNumber=[12], rfOrder=19, authorNames=WANG LY, DAI L M, journalName=Advanced theory and simulations, refType=null, unstructuredReference= WANG LY , DAI L M . Chaotic time series prediction of multi-dimensional nonlinear system based on bidirectional LSTM model[J]. Advanced theory and simulations, 2023, 6(8): 1-15., articleTitle=Chaotic time series prediction of multi-dimensional nonlinear system based on bidirectional LSTM model, refAbstract=null), Reference(id=1154040272934658888, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[13], rfOrder=20, authorNames=薛建凯, journalName=null, refType=null, unstructuredReference=薛建凯. 一种新型的群智能优化技术的研究与应用:麻雀搜索算法[D]. 上海: 东华大学, 2020., articleTitle=一种新型的群智能优化技术的研究与应用:麻雀搜索算法, refAbstract=null), Reference(id=1154040273035322186, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[13], rfOrder=21, authorNames=XUE Jiankai, journalName=null, refType=null, unstructuredReference= XUE Jiankai . Research and application of A novel swarm intelligence optimization technique[D]. Shanghai: Donghua University, 2020., articleTitle=Research and application of A novel swarm intelligence optimization technique, refAbstract=null), Reference(id=1154040273098236748, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, doi=null, pmid=null, pmcid=null, year=2022, volume=162, issue=null, pageStart=108070, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=22, authorNames=WANG M Q, JIA S X, CHEN E L, journalName=Mechanical systems and signal processing, refType=null, unstructuredReference= WANG M Q , JIA S X , CHEN E L , et al . Research and application of neural network for tread wear prediction and optimization[J]. Mechanical systems and signal processing, 2022, 162: 108070., articleTitle=Research and application of neural network for tread wear prediction and optimization, refAbstract=null)], funds=[Fund(id=1154040271626036007, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, awardId=L211007, language=CN, fundingSource=北京市自然科学基金项目(L211007), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1154040266718700198, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=1, ext=[AuthorCompanyExt(id=1154040266722894503, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044), AuthorCompanyExt(id=1154040266731283112, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266718700198, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044)]), AuthorCompany(id=1154040266869695146, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, xref=2, ext=[AuthorCompanyExt(id=1154040266878083755, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266869695146, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Beijing Subway Operation Co., Ltd. Beijing 100044), AuthorCompanyExt(id=1154040266886472364, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, companyId=1154040266869695146, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 北京市地铁运营有限公司 北京 100044)])], figs=[ArticleFig(id=1154040270443242250, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 1, caption=LSTM network structure, figureFileSmall=L1DOBPZ747wRuwO/ctpC+A==, figureFileBig=eyBnDibD+BuyD8N3qvFb4g==, tableContent=null), ArticleFig(id=1154040270531322636, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 1, caption=LSTM 网络结构, figureFileSmall=L1DOBPZ747wRuwO/ctpC+A==, figureFileBig=eyBnDibD+BuyD8N3qvFb4g==, tableContent=null), ArticleFig(id=1154040270590042894, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 2, caption=Structure of BiLSTM algorithm, figureFileSmall=e1wOo+zGeAWo+Idf8r59zQ==, figureFileBig=2NhQotNblX2LWf0B1NM04A==, tableContent=null), ArticleFig(id=1154040270665540368, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 2, caption=BiLSTM 算法结构, figureFileSmall=e1wOo+zGeAWo+Idf8r59zQ==, figureFileBig=2NhQotNblX2LWf0B1NM04A==, tableContent=null), ArticleFig(id=1154040270711677714, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 3, caption=Algorithm flow diagram of improved BiLSTM network, figureFileSmall=z4RuhZ+HKU0gqHA+xeGpww==, figureFileBig=RfA5lgK2eTEVc0bXDjissw==, tableContent=null), ArticleFig(id=1154040270787175188, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 3, caption=改进 BiLSTM 网络算法流程, figureFileSmall=z4RuhZ+HKU0gqHA+xeGpww==, figureFileBig=RfA5lgK2eTEVc0bXDjissw==, tableContent=null), ArticleFig(id=1154040270854284054, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 4, caption=Improved BiLSTM network model parameter optimization, figureFileSmall=mVJTqcNxfgV/DbRmIluWeA==, figureFileBig=4pC9ksryhYrW688LMRd46g==, tableContent=null), ArticleFig(id=1154040270933975832, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 4, caption=改进 BiLSTM 网络模型参数优化, figureFileSmall=mVJTqcNxfgV/DbRmIluWeA==, figureFileBig=4pC9ksryhYrW688LMRd46g==, tableContent=null), ArticleFig(id=1154040270980113178, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 5, caption=Comparison of prediction results of different models with measured data, figureFileSmall=Rqcn+8qmwUBnQSG+4NNyjg==, figureFileBig=DCNw5ZuIyq/wAigl1+nfFw==, tableContent=null), ArticleFig(id=1154040271038833436, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 5, caption=不同模型的预测结果与实测数据对比, figureFileSmall=Rqcn+8qmwUBnQSG+4NNyjg==, figureFileBig=DCNw5ZuIyq/wAigl1+nfFw==, tableContent=null), ArticleFig(id=1154040271080776478, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 6, caption=Predicted single wear results for different wheel tread surfaces, figureFileSmall=gICMreAt/vjOLx4w1FEHwA==, figureFileBig=hC3TW2gYfRll6o+9PQHh6Q==, tableContent=null), ArticleFig(id=1154040271131108128, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 6, caption=不同车轮踏面单次磨耗预测结果, figureFileSmall=gICMreAt/vjOLx4w1FEHwA==, figureFileBig=hC3TW2gYfRll6o+9PQHh6Q==, tableContent=null), ArticleFig(id=1154040271181439777, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Figure 7, caption=Predicted results of flange wear for different wheels, figureFileSmall=vMoW5YKEHVOsodT7WvxruA==, figureFileBig=bFwFLb6Bo6ytyEcOsLONvQ==, tableContent=null), ArticleFig(id=1154040271240160034, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=图 7, caption=不同车轮轮缘磨耗预测结果, figureFileSmall=vMoW5YKEHVOsodT7WvxruA==, figureFileBig=bFwFLb6Bo6ytyEcOsLONvQ==, tableContent=null), ArticleFig(id=1154040271286297379, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Table 1, caption=Comparison of MAPE、RMSE、MAE、 ${R}^{2}$ for wear prediction of different models, figureFileSmall=null, figureFileBig=null, tableContent=
轮对参数 模型 MAPE RMSE MAE R2
踏面磨耗 MLP 0.1433 0.1226 0.0910 0.7027
LSTM 0.1456 0.1172 0.0881 0.7285
BiLSTM 0.1160 0.0784 0.0633 0.8784
SSA-LSTM 0.0275 0.0357 0.0216 0.9748
SSA-BiLSTM 0.0128 0.0132 0.0009 0.9965
轮缘磨耗 MLP 0.0073 0.2550 0.2123 0.4205
LSTM 0.0105 0.3361 0.3038 0.4050
BiLSTM 0.0056 0.1922 0.1612 0.6708
SSA-LSTM 0.0012 0.0409 0.0299 0.9751
SSA-BiLSTM 0.0010 0.0132 0.0115 0.9845
), ArticleFig(id=1154040271365989156, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=表 1, caption=不同模型磨耗预测的 MAPE、RMSE、MAE、 ${R}^{2}$ 对比, figureFileSmall=null, figureFileBig=null, tableContent=
轮对参数 模型 MAPE RMSE MAE R2
踏面磨耗 MLP 0.1433 0.1226 0.0910 0.7027
LSTM 0.1456 0.1172 0.0881 0.7285
BiLSTM 0.1160 0.0784 0.0633 0.8784
SSA-LSTM 0.0275 0.0357 0.0216 0.9748
SSA-BiLSTM 0.0128 0.0132 0.0009 0.9965
轮缘磨耗 MLP 0.0073 0.2550 0.2123 0.4205
LSTM 0.0105 0.3361 0.3038 0.4050
BiLSTM 0.0056 0.1922 0.1612 0.6708
SSA-LSTM 0.0012 0.0409 0.0299 0.9751
SSA-BiLSTM 0.0010 0.0132 0.0115 0.9845
), ArticleFig(id=1154040271424709413, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=EN, label=Table 2, caption=Prediction results of wheel position 1 of cars 2 and 4, figureFileSmall=null, figureFileBig=null, tableContent=
轮对参数 车轮编号 MAPE RMSE MAE
踏面磨耗 2 车 1 号车轮 0.0134 0.0879 0.004 0.9889
4 车 1 号车轮 0.0142 0.095 0.0071 0.9858
轮缘磨耗 2 车 1 号车轮 0.0018 0.0447 0.0341 0.9773
4 车 1 号车轮 0.0019 0.0704 0.0568 0.9683
), ArticleFig(id=1154040271470846758, tenantId=1146029695717560320, journalId=1146123302524792850, articleId=1149744609946944231, language=CN, label=表 2, caption=2 车、 4 车 1 号位置车轮预测结果, figureFileSmall=null, figureFileBig=null, tableContent=
轮对参数 车轮编号 MAPE RMSE MAE
踏面磨耗 2 车 1 号车轮 0.0134 0.0879 0.004 0.9889
4 车 1 号车轮 0.0142 0.095 0.0071 0.9858
轮缘磨耗 2 车 1 号车轮 0.0018 0.0447 0.0341 0.9773
4 车 1 号车轮 0.0019 0.0704 0.0568 0.9683
)], attaches=null, journal=Journal(id=1146119287908970507, delFlag=0, nameCn=都市快轨交通, nameEn=Urban Rapid Rail Transit, nameHistory1=null, nameHistory2=null, issn=1672-6073, eissn=, cn=11-5144/U, coden=null, periodic=1, language=CN, oaType=0, 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=P1a+Ek/jjjHf/jufN8hKnw==, journalPrice=null, startedYear=null, abbrevIsoEn=Urban Rap Rail Trans, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1756899235814, createdBy=null, updatedBy=18614031015, firstLetterCn=U, firstLetterEn=U, subjectCode=Engineering, subjectName=工程, subjectCodeEn=Engineering, subjectNameEn=null, picCn=P1a+Ek/jjjHf/jufN8hKnw==, picEn=Ml/zvLJzYsu7AhRo5kg9VQ==, jcr=null, cjcr=null, exts=[JournalExt(id=1158448607830827008, language=CN, name=都市快轨交通, nameHistory1=null, nameHistory2=null, managedBy=中国科学技术协会, sponsoredBy=中国技术经济学会, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=北京市西直门外北京交通大学机械工程楼D905室, officeZip=100044, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.urt.cn, createdTime=1754125271042, updatedTime=1754125271042, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://dskgjt.ijournals.cn/dskgjt/site/menu/20200615141511001, submissionAuthorUrl=http://dskgjt.ijournals.cn/dskgjt/author/login, submissionEditorUrl=http://dskgjt.ijournals.cn/dskgjt/editor/login, submissionReviewUrl=http://dskgjt.ijournals.cn/dskgjt/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1158448607931490305, language=EN, name=Urban Rapid Rail Transit, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1754125271065, updatedTime=1754125271065, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://dskgjt.ijournals.cn/dskgjten/site/menu/20241120152535001, submissionAuthorUrl=http://dskgjt.ijournals.cn/dskgjt/author/login, submissionEditorUrl=http://dskgjt.ijournals.cn/dskgjt/editor/login, submissionReviewUrl=http://dskgjt.ijournals.cn/dskgjt/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1146123302524792850, websiteList=[Website(id=1148243202399789480, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123302524792850, 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/dskgjt/CN, language=CN, createTime=1751692112779, createBy=18614031015, updateTime=1753522796719, updateBy=18614031015, name=都市快轨交通-中文站点, tplId=1146099689490845704, title=都市快轨交通, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155921926926229709, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202399789480, code=articleTextType, value=kx, createTime=1753522863370, updateTime=1753522863370, creator=18614031015, updator=18614031015), WebsiteProps(id=1155921926905258186, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202399789480, code=banner, value=null, createTime=1753522863365, updateTime=1753522863365, creator=18614031015, updator=18614031015), WebsiteProps(id=1155921926896869577, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202399789480, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=AqFiyT4buOvVaL3FTbHG1g==, createTime=1753522863363, updateTime=1753522863363, creator=18614031015, updator=18614031015), WebsiteProps(id=1155921926917841100, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202399789480, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753522863368, updateTime=1753522863368, creator=18614031015, updator=18614031015), WebsiteProps(id=1155921926913646795, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202399789480, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753522863367, updateTime=1753522863367, creator=18614031015, updator=18614031015)]), Website(id=1155921810454626336, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123302524792850, 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/dskgjt/EN, language=EN, createTime=1753522835604, createBy=18614031015, updateTime=1753522835604, updateBy=18614031015, name=都市快轨交通-英文站点, tplId=1146101810881728533, title=Urban Rapid Rail Transit, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155922050809192658, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155921810454626336, code=articleTextType, value=kx, createTime=1753522892906, updateTime=1753522892906, creator=18614031015, updator=18614031015), WebsiteProps(id=1155922050792415439, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155921810454626336, code=banner, value=null, createTime=1753522892902, updateTime=1753522892902, creator=18614031015, updator=18614031015), WebsiteProps(id=1155922050788221134, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155921810454626336, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=AqFiyT4buOvVaL3FTbHG1g==, createTime=1753522892901, updateTime=1753522892901, creator=18614031015, updator=18614031015), WebsiteProps(id=1155922050804998353, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155921810454626336, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753522892905, updateTime=1753522892905, creator=18614031015, updator=18614031015), WebsiteProps(id=1155922050796609744, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155921810454626336, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753522892903, updateTime=1753522892903, creator=18614031015, updator=18614031015)])], journalTitle=都市快轨交通, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=Urban Rapid Rail Transit, journalPhotoCn=P1a+Ek/jjjHf/jufN8hKnw==, journalPhotoEn=Ml/zvLJzYsu7AhRo5kg9VQ==, journalFirstLetter=U, 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=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/dskgjt/CN/10.3969/j.issn.1672-6073.2024.03.011, detailUrlEn=https://castjournals.cast.org.cn/joweb/dskgjt/EN/10.3969/j.issn.1672-6073.2024.03.011, pdfUrlCn=https://castjournals.cast.org.cn/joweb/dskgjt/CN/PDF/10.3969/j.issn.1672-6073.2024.03.011, pdfUrlEn=https://castjournals.cast.org.cn/joweb/dskgjt/EN/PDF/10.3969/j.issn.1672-6073.2024.03.011, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于改进 BiLSTM 网络的地铁车轮磨耗预测模型
收藏切换
PDF下载
朱爱华 1 , 白杨 1 , 白堂博 1 , 王雅莉 2 , 张财胜 1 , 李安琰 1
都市快轨交通 | 学术探讨 2024,37(3): 82-89
收起
收藏切换
都市快轨交通 | 学术探讨 2024, 37(3): 82-89
基于改进 BiLSTM 网络的地铁车轮磨耗预测模型
全屏
朱爱华1 , 白杨1, 白堂博1, 王雅莉2, 张财胜1, 李安琰1
作者信息
  • 1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044
  • 2 北京市地铁运营有限公司 北京 100044
  • 朱爱华,女,教授,博士,主要研究方向为车辆关键系统服役性能与可靠性评估,

Improved Prediction Model of Metro Wheel Wear Based on BiLSTM Network
Aihua ZHU1 , Yang BAI1, Tangbo BAI1, Yali WANG2, Caisheng ZHANG1, Anyan LI1
Affiliations
  • 1 School of Mechanical-electronic and Vehicle Engineering Beijing University of Civil Engineering and Architecture Beijing 100044
  • 2 Beijing Subway Operation Co., Ltd. Beijing 100044
doi: 10.3969/j.issn.1672-6073.2024.03.011
文章导航
收藏切换

针对地铁车轮磨耗数据时间跨度较长引起的长期依赖问题,为了进一步提升预测精度,提出一种将麻雀搜索算法(sparrow search algorithm,SSA)优化双向长短期记忆网络(bidirectional long short term memory, BiLSTM)的改进 BiLSTM(SSABiLSTM)网络模型,用于地铁车轮磨耗预测。首先,利用麻雀搜索算法对双向长短期记忆网络算法的神经元个数、迭代次数、输入批量和学习率等超参数在给定范围内进行寻优,得到参数最优值;然后,以参数最优值来构建改进 BiLSTM 网络模型,对车轮磨耗进行预测分析;最后,以车轮踏面磨耗和轮缘磨耗作为研究对象,将某地铁1车厢1号车轮的现场实测历史磨耗数据作为输入,对该模型进行训练及验证分析,并与多层感知机(multilayer perceptron, MLP)、LSTM、BiLSTM 以及 SSALSTM 模型的预测结果进行对比。研究结果表明:SSABiLSTM模型的车轮磨耗预测精度更高,与LSTM、BiLSTM 以及SSALSTM 网络模型相比,踏面磨耗的平均绝对百分误差(mean absolute percentage error,MAPE)分别降低了 13.28%、10.32%、1.47%,轮缘磨耗分别降低了9.5%、0.46%、0.02%;分别对同一地铁2号、4号车厢的1号位置车轮磨耗进行预测,并与磨耗实测数据进行对比,踏面磨耗的平均绝对百分比误差分别为1.34%、1.42%,轮缘磨耗的平均绝对百分比误差分别为0.18%、0.19%,验证了本文所提模型具有良好的泛化性,为地铁轮对智能化管理提供理论支持,延长车轮使用寿命。

地铁  /  磨耗预测  /  麻雀搜索算法  /  双向长短期记忆网络

In order to address the issue of longterm dependence caused by the extended time span of wheel wear data and improve the prediction accuracy, an improved BiLSTM metro wheel wear prediction model is proposed by optimizing Bidirectional long shortterm memory network (BiLSTM) with Sparrow search algorithm (SSA). Firstly, the hyperparameters of the BiLSTM algorithm, such as the number of neurons, iteration count, input batch size, and learning rate, are optimized using the SSA. This optimization process is conducted within a specified range to obtain the optimal values of these hyperparameters. This optimization process aims to obtain the optimal parameter values. Subsequently, the SSABiLSTM network model is constructed using these optimal parameter values to predict and analyze wheel wear. Tread wear and flange wear are taken as the research objects, and the measured historical wear data of wheel No.1 of the metro’s carriage # 1 are used as inputs to metro and validate the model, and compare the prediction results with those of MLP, LSTM, BiLSTM and SSA-LSTM models. The results show that the improved bidirectional long short-term memory network model has higher wear prediction accuracy, and the mean absolute percentage error (MAPE) of tread wear is reduced by 13.28%, 10.32%, and 1.47%, and flange wear by 9.5%, 0.46%, and 0.02%. The wear of the No. 1 wheel of the same metro No.2 and No.4 cars is predicted and compared with the measured wear data. The average absolute percentage error of tread wear is 1.34% and 1.42%, respectively, and the average absolute percentage error of rim wear is 0.18% and 0.19%, respectively. The results confirm that the model exhibits strong generalization capabilities. The wheel wear prediction model based on improved BiLSTM network (SSA-Bi-LSTM) has high prediction accuracy and good generalization, which provides theoretical support for the intelligent management of metro wheelsets and prolongs wheel service life.

metro  /  wear prediction  /  sparrow search algorithm  /  bidirectional long short time memory networks
朱爱华, 白杨, 白堂博, 王雅莉, 张财胜, 李安琰. 基于改进 BiLSTM 网络的地铁车轮磨耗预测模型. 都市快轨交通, 2024 , 37 (3) : 82 -89 . DOI: 10.3969/j.issn.1672-6073.2024.03.011
Aihua ZHU, Yang BAI, Tangbo BAI, Yali WANG, Caisheng ZHANG, Anyan LI. Improved Prediction Model of Metro Wheel Wear Based on BiLSTM Network[J]. Urban Rapid Rail Transit, 2024 , 37 (3) : 82 -89 . DOI: 10.3969/j.issn.1672-6073.2024.03.011
轮对作为列车的关键部位之一, 不仅在钢轨上作纯滚动,同时受到曲线离心力的作用,轮轨间产生复杂的横向和纵向蠕滑力,导致车轮出现踏面磨损和轮缘磨耗等问题 [ 1 ] 。通过对车轮磨耗进行预测,可以制定合理的轮对维护计划, 减少维修成本, 降低事故和故障发生的风险, 提高列车的安全性和可靠性。
针对轮轨磨耗预测问题, 很多学者基于理论分析和数值计算, 提出了摩擦功等多种轮轨磨耗预测模型 [ 2 - 4 ] ,这些研究主要建立在动力学和摩擦学等理论基础上, 无法解决实际运行过程中环境和不确定因素对磨耗的影响。近年来随着人工智能的不断发展, 人工智能算法被应用于轨道交通领域, 如卷积神经网络 [ 5 ] 、 支持向量机等。并在磨耗预测方面取得了一定的进展。王文健等 [ 7 ] 利用 $\mathrm{{BP}}$ 神经网络对钢轨磨损量进行预测。在此基础上,王平等 [ 8 ] 在 BP 神经网络(back propagation neural network, BPNN)的基础上利用粒子群算法进行优化, 提升对钢轨磨损预测的准确性。 王美琪等 [ 9 ] 针对单一踏面磨耗模型无法应对复杂工况下磨耗定量计算的问题, 提出基于恒等映射的多层学习机模型, 该模型可以计算不同速度、里程和曲线半径等参数对踏面磨耗的影响。
虽然人工神经网络在车轮磨耗预测上的应用受到越来越多的关注, 但由于车轮磨耗数据的时间跨度较长, 利用神经网络进行预测会引起长期依赖问题, 即当时间序列较长时, 网络会遗忘较早的信息, 从而导致梯度消失 [ 10 - 12 ] 等问题的出现。为提高预测效果并且更好地解决上述问题, 本文提出利用双向长短期记忆网络(bidirectional long short term memory, BiLSTM) 来解决车轮磨耗数据存在的长期依赖问题, 并改善了数据利用率低、特征获取不充分等问题, 提高预测的准确性。
为提高搜索精度、收敛速度、稳定性和避免陷入局部最优值, 本文在双向长短期记忆算法的基础上引入麻雀搜索算法, 提出一种融合麻雀搜索算法与双向长短期记忆算法的改进 BiLSTM 网络模型。首先,利用麻雀搜索算法 [ 13 ] (sparrow search algorithm, SSA) 优化 BiLSTM 的神经元个数、迭代次数、输入批量和学习率等超参数, 从而获得一组最优的超参数来发挥模型的最佳性能。然后, 建立 SSA-BiLSTM 混合预测模型, 以某地铁连续 6 年的车轮磨耗实测数据为样本, 对车轮踏面磨耗和轮缘磨耗数据进行学习和预测, 并且与多层感知机(multilayer perceptron, MLP)、LSTM、 BiLSTM 以及 SSA-LSTM 模型进行对比分析, 说明该模型的优越性。最后,对同一列车 2 号、4 号车厢的 1 号位置车轮磨耗数据进行学习预测, 进一步验证模型的有效性和适用性。本文采用的改进 BiLSTM 网络模型可以实现对车轮磨耗的有效预测, 为地铁运行安全评估提供参考。
MLP 是一种前馈神经网络, 由多个全连接层组成, 应用于分类、回归和特征提取等任务。递归神经网络(recurrent neural network, RNN)是一种具有循环连接的神经网络模型, 它通过前一时刻隐藏层的状态来更新当前隐藏层的状态, 这样的自连接隐藏层结构会产生依赖问题。LSTM 继承了 RNN 隐藏层采用的自连接形式, 通过引入三种门结构来处理由长期依赖问题导致 RNN 存在的梯度消失和梯度爆炸等问题 [ 14 ] , LSTM 的网络结构如 图 1 所示。
图 1 中, $\mathbf{i}\text{、}\mathbf{f}$$\mathbf{o}$ 分别表示输入、遗忘以及输出门, ${\mathbf{i}}_{t}$$t$ 时刻输入门输出, ${\mathbf{f}}_{t}$$t$ 时刻遗忘门输出, ${\mathbf{o}}_{t}$$t$ 时刻输出门输出, ${\mathbf{c}}_{t}$ 为记忆单元, ${\mathbf{x}}_{t}$ 为输入向量, ${\mathbf{h}}_{t}$ 为输出向量, $\mathbf{\sigma }$ 为激活函数, ${\mathbf{\theta }}_{f}\text{、}{\mathbf{\theta }}_{i}\text{、}{\mathbf{\theta }}_{o}$ 均为上一时刻输出 ${\mathbf{h}}_{t - 1}$ 与该时刻输入 ${\mathbf{x}}_{t}$ 组成的向量。
输入门是用来控制流入记忆单元的数据大小, 计算式为
${\mathbf{i}}_{t}= \sigma \left({{\mathbf{W}}_{xi}{\mathbf{x}}_{t}+ {\mathbf{W}}_{hi}{\mathbf{h}}_{t - 1}+ {\mathbf{b}}_{i}}\right)$
式中, ${\mathbf{W}}_{xi}$${\mathbf{W}}_{hi}$ 为输入门的权重矩阵; ${\mathbf{b}}_{i}$ 为输入门的偏置向量。
LSTM 的核心为遗忘门, 用来决定历史信息中需要保留的数据,所以前一时刻记忆单元 ${\mathbf{c}}_{t - 1}$ 对当前时刻记忆单元 ${\mathbf{c}}_{t}$ 的影响可表示为
${\mathbf{c}}_{t}= {\mathbf{f}}_{t}\odot {\mathbf{c}}_{t - 1}+ {\mathbf{i}}_{t}\odot {\widetilde{\mathbf{C}}}_{t}$
${\mathbf{f}}_{t}= \sigma \left({{\mathbf{W}}_{xf}{\mathbf{x}}_{t}+ {\mathbf{W}}_{hf}{\mathbf{h}}_{t - 1}+ {\mathbf{b}}_{f}}\right)$
${\widetilde{\mathbf{C}}}_{t}= \tanh \left({{\mathbf{W}}_{C}{\mathbf{x}}_{t}+ {\mathbf{W}}_{C}{h}_{t - 1}+ {\mathbf{b}}_{C}}\right)$
式中, $\odot$ 代表 Hadamard 积运算; ${\mathbf{W}}_{xf}$${\mathbf{W}}_{hf}$ 为遗忘门的权重参数; ${\mathbf{W}}_{C}$ 为遗忘门的权重矩阵; ${\mathbf{b}}_{j}$${\mathbf{b}}_{C}$ 为遗忘门的偏置参数; ${\widetilde{\mathbf{C}}}_{t}$ 为当前输入的记忆单元状态。
将当前输入特征 ${\mathbf{x}}_{t}$ 和前一时刻输出 ${\mathbf{h}}_{t - 1}$ 传递给激活函数 $\mathbf{\sigma }$ ,得到激活函数值 ${\mathbf{\sigma }}_{t}$ ,并计算当前时刻输出 ${\mathbf{h}}_{t}$ ,计算式为
${\mathbf{h}}_{t}= {\mathbf{o}}_{t}\odot \tanh \left({\mathbf{c}}_{t}\right)$
${\mathbf{o}}_{t}= \sigma \left({{\mathbf{W}}_{xo}{\mathbf{x}}_{t}+ {\mathbf{W}}_{ho}{\mathbf{h}}_{t - 1}+ {\mathbf{b}}_{o}}\right)$
式中, ${\mathbf{W}}_{xo}$${\mathbf{W}}_{ho}$ 为输出门的权重参数; ${\mathbf{b}}_{o}$ 为输出门的偏置参数。
BiLSTM 神经网络模型是通过两个相互独立的 LSTM 网络构成, 通过正序和逆序的方式将序列信息分别输入到网络, 使每个 LSTM 单元都可以获得前向和后向的信息。其中,前向 LSTM 的输出为 ${\mathbf{Z}}_{t}^{f}$ ,后向 LSTM 的输出为 ${\mathbf{Z}}_{t}^{b}$ ,其计算式为式(1) $\sim$ 式(6)。该网络是将前向 LSTM 和后向 LSTM 的输出向量(即提取后的特征向量)进行拼接后形成的向量作为最终输出。 最终输出向量计算式为
${\mathbf{y}}_{t}= \left\lbrack {{\mathbf{Z}}_{t}^{f}\cdot {\mathbf{Z}}_{t}^{b}}\right\rbrack $
BiLSTM 的模型是使 $t$ 时刻得到的特征数据具有前后时刻的信息,与单个 LSTM 结构相比,特征提取效率和性能更佳。模型结构如 图 2 所示,其中, ${\mathbf{M}}_{t}$ 为权重矩阵, ${\mathbf{Z}}_{t}$ 为输出向量。
假设由 $n$ 只麻雀组成的种群表示为
$\mathbf{X}= \left\lbrack \begin{matrix}{x}_{1}^{1}& {x}_{1}^{2}& \cdots &{x}_{1}^{d}\\{x}_{2}^{1}& {x}_{2}^{2}& \cdots &{x}_{2}^{d}\\\vdots &\vdots &\vdots &\vdots \\{x}_{n}^{1}& {x}_{n}^{2}& \cdots &{x}_{n}^{d}\end{matrix}\right\rbrack $
式中, $d$ 为优化问题的维数; $n$ 为麻雀总数, ${\mathbf{x}}_{n}^{d}$ 为矩阵第 $n$ 行第 $d$ 列的元素。因此,麻雀的适应度值 ${\mathbf{F}}_{x}$ 可表示为
${\mathbf{F}}_{\mathbf{x}}= \left\lbrack \begin{matrix} f\left({{x}_{1}^{1}\;{x}_{1}^{2}\;\cdots \;{x}_{1}^{d}}\right)\\ f\left({{x}_{2}^{1}\;{x}_{2}^{2}\;\cdots \;{x}_{2}^{d}}\right)\\\vdots \\ f\left({{x}_{n}^{1}\;{x}_{n}^{2}\;\cdots \;{x}_{n}^{d}}\right)\end{matrix}\right\rbrack $
式中, $f$ 表示为适应度值。
探索者位置更新函数为
${X}_{u, v}^{r + 1}= \left\{\begin{matrix}{X}_{u, v}\cdot \exp \left({-\frac{u}{\alpha \cdot {N}_{\max }}}\right),{R}_{2}< {R}_{\mathrm{{sa}}}\\{X}_{u, v}+ {QL},{R}_{2}\geq {R}_{\mathrm{{sa}}}\end{matrix}\right.$
式中, $r$ 为迭代次数; ${N}_{\max }$ 为迭代过程的最大次数; ${X}_{u, v}$ 为第 $u$ 只麻雀在第 $v$ 维的位置; $\alpha$ 是随机常数; ${R}_{2}$ 为预警值, ${R}_{\mathrm{{sa}}}$ 为安全值; $Q$ 为常数; $\mathbf{L}$ 是元素为 1 的 $1 \times d$ 的矩阵。
${R}_{2}< {R}_{\mathrm{{sa}}}$ 时,表明该区域安全; 当 ${R}_{2}\geq {R}_{\mathrm{{sa}}}$ 时,表明该区域危险需前往安全区域。追随者会抢夺探索者找到的食物, 其位置更新式为
${X}_{u, v}^{r + 1}= \left\{\begin{array}{l} Q \cdot \exp \left({-\frac{{X}_{\text{worst }}- {X}_{u, v}^{r}}{{u}^{2}}}\right), u >\frac{n}{2}\\{X}_{\mathrm{P}}^{r + 1}+ \left|{{X}_{u, v}- {X}_{\mathrm{p}}^{r + 1}}\right|\cdot {A}^{+ }\cdot L, u \leq \frac{n}{2}\end{array}\right.$
式中, ${X}_{\mathrm{P}}$${X}_{\text{worst }}$ 为探索者当前的最优和最差位置; $\mathbf{A}$ 是元素为 1 或 -1 的 $1 \times d$ 的矩阵,且 ${\mathbf{A}}^{+ }= {\mathbf{A}}^{\mathrm{T}}{\left(\mathbf{A}{\mathbf{A}}^{\mathrm{T}}\right)}^{-1}$ ; 当 $u > n/2$ 时,表示第 $u$ 个加入者需到其他区域寻找食物。
假设种群中 10%到 20%的个体为预警者,预警者的位置更新式为
${X}_{u, v}^{r + 1}= \left\{\begin{array}{ll}{X}_{\text{best }}^{r}+ \beta \cdot \left|{{X}_{u, v}^{r}- {X}_{\text{best }}^{r}}\right|& {f}_{i}> {f}_{g}\\{X}_{u, v}^{r}+ K \cdot \left(\frac{\left|{X}_{u, v}^{r}- {X}_{\text{worst }}^{r}\right|}{\left({{f}_{i}- {f}_{w}}\right)+ \varepsilon }\right)& {f}_{i}= {f}_{g}\end{array}\right.$
式中, ${X}_{\text{best }}$ 是目前的最优位置; $\beta$ 是控制步长的参数; $K$ 为常数; ${f}_{i}$ 为当前适应度值; ${f}_{g}$${f}_{w}$ 分别为适应度值中的最优和最差值; $\varepsilon$ 为常数,其意义是避免分母为 0 。
${f}_{i}> {f}_{g}$ 时,代表麻雀处于区域边缘,被捕食者攻击的风险较高,其中 ${X}_{\text{best }}$ 代表最佳位置。当 ${f}_{i}= {f}_{g}$ 时, 代表出现危险。
为了更好地对地铁车轮磨耗进行预测, 以 BiLSTM 算法为基础模型引入 SSA 优化算法, 建立改进 BiLSTM 网络的地铁车轮磨耗预测模型。
对于 BiLSTM 算法来说, 超参数(如神经元个数、 迭代次数、输入批量和学习率等)的设置对网络的拟合能力非常重要, 但这些参数需根据训练结果的优劣进行手动调节, 导致模型的训练效率和准确性较低。而 SSA 算法与其他优化算法相比, 具有精度高、设置参数少、收敛速度快等诸多优点。本文通过 SSA 算法进行全局搜索, 得到 BiLSTM 的最优超参数, 包括迭代次数、学习率以及隐含层的神经元数等, 并将该结果传递到 BiLSTM 网络中, 从而提高 BiLSTM 网络寻找最优超参数的效率, 提升模型的预测结果。改进 BiLSTM 网络算法流程如 图 3 所示。
SSA 算法优化 BiLSTM 相关参数步骤如下。
步骤 1: 首先对车轮磨耗数据进行清洗并归一化, 划分出训练集和测试集。
步骤 2: 初始化 SSA 及 BiLSTM 的相关参数, 其中包括 SSA 算法的种群数量、迭代次数等, 以及 BiLSTM 的批大小(batch size)等。
步骤 3: 给定预警值,并将预警值与安全值进行比较, 根据式(10)对探索者位置进行更新。
步骤 4: 根据式(11)对追随者位置进行更新。
步骤 5: 根据式(12)对预警者位置进行更新。
步骤 6: 完成种群位置更新后, 计算新位置的适应度值, 将两个适应度值进行比较, 并更新全局中的最优与最差位置信息。
步骤 7: 判断是否达到最大迭代值, 若未达到重复执行步骤 3 , 满足条件则终止循环, 输出最优超参数, 并将该结果传递到 BiLSTM 网络中, 从而建立改进 BiLSTM 网络预测模型。
步骤 8: 将测试集输入模型, 输出结果, 结束运行。
本文选择某地铁车辆连续 6 年的车轮实测数据对模型进行训练验证。每次测量包括 6 节车厢、每节车厢 8 个车轮的踏面磨耗、车轮滚动圆直径、走行公里数、轮缘厚度等相关参数。其中, 选择车轮滚动圆直径、轮缘厚度和走行公里数作为输入参数对车轮踏面磨耗进行预测, 以车轮滚动圆直径、车轮踏面磨耗和走行公里数作为输入参数来对轮缘厚度进行预测。
由于环境原因导致数据存在误差, 为了减少误差, 使用多次测量并取平均值来减少随机误差, 然后使用 3 次样条插值法对数据进行平滑处理, 从而获得满足要求的磨耗数据。在数据输入到网络之前首先要进行归一化处理, 归一化也可称为标准化, 其目的是为了解决由于不同变量之间量纲存在差异的问题, 并且进行归一化时还可以加快模型的训练速度。本文采用常见的MinMaxScaler 归一化方法对数据进行处理, 经过处理后输入数据将处于 $\left\lbrack {0,1}\right\rbrack$ 之间,对应的计算式为
$\text{ xstd }= \frac{x -{x}_{\min }}{{x}_{\max }- {x}_{\min }}$
式中, xstd 表示该数据完成归一化; $x$ 为原数据, ${x}_{\max }$ 为数据最大值, ${x}_{\min }$ 为最小值。
在对数据进行归一化后, 将样本数据的 80% 作为训练集对模型进行训练, 剩余 20%作为验证集检验模型的预测效果。为了对比使用 SSA 优化后的 BiLSTM 模型在车轮磨耗预测中的有效性和准确性, 本文将改进 BiLSTM 网络与 MLP、LSTM、BiLSTM 以及 SSA-LSTM 算法进行比较。
为了对模型的预测效果和准确度进行分析, 从网络性能评价指标中选取 4 种指标对预测效果进行验证: 平均绝对百分误差(mean absolute percentage error, MAPE)、均方根误差(root mean square error, RMSE)、 平均绝对误差(mean absolute error, MAE)、决定系数 ${R}^{2}$ 。前三种值越小代表着预测输出越接近真实值,决定系数表示模型与数据的拟合程度, 值越大代表模型拟合效果越好。评价标准的计算式分别为
$\text{ MAPE }= \frac{{100}\%}{s}\mathop{\sum }\limits_{{i = 1}}^{s}\left|\frac{{\widehat{y}}_{i}- {y}_{i}}{{y}_{i}}\right|$
$\text{ RMSE }= \sqrt{\frac{1}{s}\mathop{\sum }\limits_{{i = 1}}^{s}{\left({\widehat{y}}_{i}- {y}_{i}\right)}^{2}}$
$\mathrm{{MAE}}= \frac{1}{s}\mathop{\sum }\limits_{{i = 1}}^{s}\left|{{\widehat{y}}_{i}- {y}_{i}}\right|$
${R}^{2}= 1 -\frac{\mathop{\sum }\limits_{{i = 1}}^{s}{\left({\widehat{y}}_{i}- {y}_{i}\right)}^{2}}{\mathop{\sum }\limits_{{i = 1}}^{s}{\left({\bar{y}}_{i}- {y}_{i}\right)}^{2}}$
式中: ${\widehat{y}}_{i}$ 代表预测值, ${y}_{i}$ 代表真实值; ${\bar{y}}_{i}$ 代表真实值的平均值; $s$ 为数据总量。
在模型进行训练之前, 先利用 SSA 优化算法对 BiLSTM 网络中的超参数进行训练, 得到超参数的最优值。对优化模型的训练参数设置如下: 麻雀的种群数量设置为 20 , 探索者在麻雀中的比例为 20%, 预警值设置为 0.9 , 学习率的选择范围为 $\left\lbrack {{0.0001},{0.01}}\right\rbrack$ , 迭代次数的取值范围为 [ 10 , 200 ] ,隐含层的神经元取值范围为 $\left\lbrack {1,{100}}\right\rbrack$ ,训练次数为 30。在模型的训练过程中, SSA 算法对 BiLSTM 算法中的隐含层神经元个数、 学习率以及迭代次数不断进行调整, 以寻求最佳参数, 当适应度不再发生改变时结束训练,此时得到的值即为 BiLSTM 网络的最优参数。参数优化结果如 图 4 所示, 隐含层的神经元个数分别为 90、94 , 迭代次数为 178 , 学习率为 0.0016 , 适应度为 0.0003 , 此时模型的训练效果达到最佳。
在 SSA 完成参数寻优后, 使用改进 BiLSTM 网络模型对磨耗数据进行预测, 并分别与 MLP、LSTM、 BiLSTM 以及 SSA-LSTM 算法进行对比。不同模型下车轮踏面磨耗预测和轮缘磨耗预测的 MAPE、RMSE、 MAE、 ${R}^{2}$ 对比如 表 1 所示。
通常 ${R}^{2}$ 越大, MAPE、RMSE、MAE 越小,该模型的性能越好。由 表 1 可知, 相比于 MLP、LSTM、 BiLSTM 和 SSA-LSTM 模型, SSA-BiLSTM 模型的 MAPE、RMSE、MAE 最低,而 ${R}^{2}$ 最高,踏面和轮缘磨耗 ${R}^{2}$ 的分别为 0.9965 和 0.9845,接近于最佳状态。 不同算法的对比结果表明, 本文的模型在车轮踏面磨耗和轮缘磨耗预测精度方面有较大的提升, 与 MLP、 LSTM、BiLSTM 以及 SSA-LSTM 神经网络相比准确性有明显提高。
图 5 为不同模型的车轮踏面单次磨耗、累计磨耗和轮缘磨耗预测结果与实测数据对比。由 图 5 可知, 随着运行里程的增加, 5 种模型的预测结果差异越来越大, 本文构建的改进 BiLSTM 网络模型所得的预测结果曲线与车轮磨耗真实数据曲线基本吻合, 该算法相较于传统的单一算法具有更高的预测精度, 里程增加时, 其优势更加明显。
为了进一步验证改进 BiLSTM 网络车轮磨耗预测模型的准确性, 同时选取同一列车的 2 号车、 4 号车的 1 号位置车轮磨耗数据进行预测, 并对预测结果与实测结果进行分析计算。 2 车、 4 车 1 号位置车轮的预测评价指标如 表 2 所示, 预测结果如 图 6图 7 所示。 本文利用 1 号车厢 1 号轮训练所得模型对同一列车不同车轮进行预测, 由 图 6图 7 可知, 预测结果与真实数据基本一致, 表明该模型在训练过程中学习到该车车轮磨耗的一般特征和规律, 不仅在特定位置车轮磨耗预测准确性较高, 对同列车不同车轮的磨耗预测结果也有较高的精确度, 说明该模型具有较好的泛化性。
本文通过麻雀搜索算法对双向长短期记忆网络的迭代次数、输入批量、神经元个数和学习率等超参数进行优化, 提出了基于改进双向长短期记忆网络的地铁车轮磨耗预测模型, 为地铁安全及维护提供参考。 得到主要结论如下:
1) 提出改进 BiLSTM 网络算法, 建立基于改进 BiLSTM 网络的车轮磨耗预测模型, 通过某列车实测车轮磨耗数据进行验证, 车轮踏面与轮缘磨耗的平均绝对百分比误差为 1.28%和 0.1%,可以准确地对车轮磨耗数据进行学习并预测, 证明该模型的有效性。
2) 对某地铁车辆的车轮磨耗历史数据预测结果表明, 与 MLP, LSTM, BiLSTM 和 SSA-LSTM 模型相比, 本文所提出的预测模型的踏面磨耗和轮缘磨耗的 MAPE、RMSE、MAE 值最低,踏面磨耗为 1.28%、 ${1.32}\%$${0.09}\%$ ,轮缘磨耗为 ${0.1}\%$${1.32}\%$${1.15}\%$ , ${R}^{2}$ 更加接近真实值,分别为 99.65%和 98.45%, 从而证明了改进 BiLSTM 网络模型具有较高的预测精度。
3) 对同一地铁不同车厢的车轮磨耗数据进行预测,得出预测结果的 ${R}^{2}$ 均达到 ${96}\%$ 以上,表明改进 BiLSTM 网络模型具有较高的预测精度, 验证该模型有较好的泛化性。
  • 北京市自然科学基金项目(L211007)
参考文献 引证文献
排序方式:
[1]
郭涛, 王宁, 刘永强. 车轮踏面常规磨耗对高速列车动力学性能的影响[J]. 石家庄铁道大学学报(自然科学版), 2020, 33(4): 44-49.
GUO Tao , WANG Ning , LIU Yongqiang . Influence of wheel wear on dynamic performance of high-speed train[J]. Journal of Shijiazhuang Tiedao University (Natural Science Edition), 2020, 33(4): 44-49.
[2]
BRAGHIN F , LEWIS R , DWYER-JOYCE R S , et al . A mathematical model to predict railway wheel profile evolution due to wear[J]. Wear, 2006, 261(11/12): 1253-1264.
[3]
TAO GQ , WEN Z F , ZHAO X , et al . Effects of wheel-rail contact modelling on wheel wear simulation[J]. Wear, 2016, 366/367: 146-156.
[4]
王雪萍, 张军, 马贺. 高速列车车轮踏面磨耗预测方法的研究[J]. 摩擦学学报, 2018, 38(4): 462-467.
WANG Xueping , ZHANG Jun , MA He . Prediction method of wheel wear of high speed train[J]. Tribology, 2018, 38(4): 462-467.
[5]
张梓鸿, 赵正阳, 王文斌, 等. 基于卷积神经网络的城市轨道交通轨道几何不平顺指标预测[J]. 城市轨道交通研究, 2023, 26(10): 109-115.
ZHANG Zihong , ZHAO Zhengyang , WANG Wenbin , et al. Prediction of urban rail transit track geometry irregularity indicators based on convolutional neural networks[J]. Urban mass transit, 2023, 26(10): 109-115.
[6]
王慧君, 胡定玉, 方宇, 等. 基于梯度提升决策树的车轮轮缘厚度磨耗预测[J]. 测控技术, 2020, 39(11): 80-84.
WANG Huijun , HU Dingyu , FANG Yu , et al. Prediction of wheel flange thickness abrasion based on gradient boosting decision tree[J]. Measurement & control technology, 2020, 39(11): 80-84.
[7]
王文健, 陈明韬, 刘启跃. 基于BP神经网络的钢轨磨损量预测[J]. 润滑与密封, 2007, 32(12): 20-22.
WANG Wenjian , CHEN Mingtao , LIU Qiyue . Prediction of wear volumes of rail steel based on BP neural network[J]. Lubrication engineering, 2007, 32(12): 20-22.
[8]
王平, 王彩芸, 王文健, 等. 基于PSO混合算法优化BP网络在钢轨磨损量预测中的应用[J]. 机械设计, 2013, 30(8): 15-20.
WANG Ping , WANG Caiyun , WANG Wenjian , et al. Application of BP ANN optimized by hybrid algorithm based on PSO in wear volume prediction of rail steel[J]. Journal of machine design, 2013, 30(8): 15-20.
[9]
王美琪, 王艺, 陈恩利, 等. 基于恒等映射多层极限学习机的高速列车踏面磨耗预测模型[J]. 力学学报, 2022, 54(6): 1720-1731.
WANG Meiqi , WANG Yi , CHEN Enli , et al. High-speed train tread wear prediction model based on i-ml-elm[J]. Chinese journal of theoretical and applied mechanics, 2022, 54(6): 1720-1731.
[10]
HOCHREITER S , SCHMIDHUBER J . Long short-term memory[J]. Neural computation, 1997, 9(8): 1735-1780.
[11]
吕开云, 邱万锦, 龚循强, 等. 基于长短期记忆神经网络的城市交通速度预测[J]. 东华理工大学学报(自然科学版), 2023, 46(1): 77-84.
LYU Kaiyun , QIU Wanjin , GONG Xunqiang , et al. Urban traffic speed prediction based on long short-term memory neural network[J]. Journal of East China University of Technology (Natural Science), 2023, 46(1): 77-84.
[12]
WANG LY , DAI L M . Chaotic time series prediction of multi-dimensional nonlinear system based on bidirectional LSTM model[J]. Advanced theory and simulations, 2023, 6(8): 1-15.
[13]
薛建凯. 一种新型的群智能优化技术的研究与应用:麻雀搜索算法[D]. 上海: 东华大学, 2020.
XUE Jiankai . Research and application of A novel swarm intelligence optimization technique[D]. Shanghai: Donghua University, 2020.
[14]
WANG M Q , JIA S X , CHEN E L , et al . Research and application of neural network for tread wear prediction and optimization[J]. Mechanical systems and signal processing, 2022, 162: 108070.
2024年第37卷第3期
PDF下载
325
127
引用本文
BibTeX
文章信息
doi: 10.3969/j.issn.1672-6073.2024.03.011
  • 接收时间:2023-12-14
  • 首发时间:2025-07-09
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2023-12-14
  • 修回日期:2024-03-09
基金
北京市自然科学基金项目(L211007)
作者信息
    1 北京建筑大学 城市轨道交通车辆服役性能保障北京市重点试验室 北京 100044
    2 北京市地铁运营有限公司 北京 100044
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/dskgjt/CN/10.3969/j.issn.1672-6073.2024.03.011
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