Article(id=1148011761636209522, tenantId=1146029695717560320, journalId=1146119989267898375, issueId=1149298831252079541, articleNumber=null, orderNo=null, doi=10.7654/j.issn.2097-1974.20240204, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1638460800000, receivedDateStr=2021-12-03, revisedDate=1659801600000, revisedDateStr=2022-08-07, acceptedDate=null, acceptedDateStr=null, onlineDate=1751636933001, onlineDateStr=2025-07-04, pubDate=1713974400000, pubDateStr=2024-04-25, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1751636933001, onlineIssueDateStr=2025-07-04, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1751636933001, creator=13701087609, updateTime=1751636933001, updator=13701087609, issue=Issue{id=1149298831252079541, tenantId=1146029695717560320, journalId=1146119989267898375, year='2024', volume='47', issue='2', pageStart='1', pageEnd='106', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1751943794309, creator=13701087609, updateTime=1754895895552, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1161680841353552315, tenantId=1146029695717560320, journalId=1146119989267898375, issueId=1149298831252079541, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1161680841353552316, tenantId=1146029695717560320, journalId=1146119989267898375, issueId=1149298831252079541, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=25, endPage=31, ext={EN=ArticleExt(id=1148011761866896256, articleId=1148011761636209522, tenantId=1146029695717560320, journalId=1146119989267898375, language=EN, title=Rocket Engine Fault Detection with Attention based Recurrent Neural Networks, columnId=1154057566893105509, journalTitle=Missiles and Space Vehicles, columnName=Propulsion, runingTitle=null, highlight=null, articleAbstract=

Focusing on the main working phase of liquid rocket engine, with the aid of multivariate non-linear time series analysis, and based on Dual Stage Attention Based Recurrent Neural Networks (DA-RNN), a new time series analysis tool, Convolutional Dual Stage Attention Based Recurrent Neural Networks (CDA-RNN), is proposed, by which a fault trend prediction model is established. Compared with LSTM, DA-RNN, etc, this model shows higher prediction accuracy. Combined with autocorrelation analysis of the prediction residual, a quantitative basis of fault detection is proposed after introducing failure confidence probability. Using hot test data with weak fault to validate the model, result shows that the CDA-RNN model enables robust weak fault muti-parameter detection in unsteady working process. This strategy is so effective that it calls for direct engineering application.

, 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=Wanxuan ZHANG, Zhe LU, Jian ZHANG, Wei XUE, Nan ZHANG), CN=ArticleExt(id=1148011769076904397, articleId=1148011761636209522, tenantId=1146029695717560320, journalId=1146119989267898375, language=CN, title=基于注意力机制循环神经网络的液体火箭发动机故障检测, columnId=1154057567014740326, journalTitle=导弹与航天运载技术, columnName=动力系统, runingTitle=null, highlight=null, articleAbstract=

针对液体火箭发动机主级段工作过程,采用多变量非线性时间序列分析理论,在两级注意力机制循环神经网络(Dual Stage Attention Based Recurrent Neural Networks,DA-RNN)的基础上,提出一种新型时序分析工具——卷积两级注意力机制循环神经网络(Convolutional Dual Stage Attention Based Recurrent Neural Networks, CDA-RNN),从而建立故障趋势预测模型。通过对预测残差进行自相关性分析并定义故障置信概率,提出了故障检测量化依据。利用发生微弱故障的热试车数据进行验证,结果表明,CDA-RNN模型对非稳态工作段微弱故障多参数检测具有良好鲁棒性,该方法十分有效,具有直接应用价值。

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=yCjMkLIlmoshFPpdCdpG0g==, magXml=2KMy6pbydq8qN8YtMseLrg==, pdfUrl=null, pdf=ar5AaulyXlzXgIykyNGnSw==, pdfFileSize=2929464, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=EcNXO2K0emP5gGcoOBP5QA==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=5zWzWG2Th+SCbprr0oYm8Q==, mapNumber=null, authorCompany=null, fund=null, authors=

张万旋(1994—),男,博士,工程师,主要研究方向为动力系统总体设计及健康管理。

卢哲(1992—),男,工程师,主要研究方向为机器学习和深度学习。

张箭(1986—),男,高级工程师,主要研究方向为动力系统总体设计。

薛薇(1981—),女,博士,高级工程师,主要研究方向为动力系统总体设计及健康管理。

张楠(1960—),男,博士,研究员,主要研究方向为动力系统总体设计及健康管理。

, authorsList=张万旋, 卢哲, 张箭, 薛薇, 张楠)}, authors=[Author(id=1197274440502395743, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, orderNo=0, 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=1197274441068626787, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274440502395743, language=EN, stringName=Wanxuan ZHANG, firstName=Wanxuan, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Beijing Aerospace Propulsion Institute,Beijing,100076, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1197274442754736996, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274440502395743, language=CN, stringName=张万旋, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=北京航天动力研究所,北京,100076, bio={"content":"

张万旋(1994—),男,博士,工程师,主要研究方向为动力系统总体设计及健康管理。

"}, bioImg=null, bioContent=

张万旋(1994—),男,博士,工程师,主要研究方向为动力系统总体设计及健康管理。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])]), Author(id=1197274443618763622, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, 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=1197274444457624427, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274443618763622, language=EN, stringName=Zhe LU, firstName=Zhe, middleName=null, lastName=LU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Beijing Aerospace Propulsion Institute,Beijing,100076, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1197274445384565614, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274443618763622, language=CN, stringName=卢哲, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=北京航天动力研究所,北京,100076, bio={"content":"

卢哲(1992—),男,工程师,主要研究方向为机器学习和深度学习。

"}, bioImg=null, bioContent=

卢哲(1992—),男,工程师,主要研究方向为机器学习和深度学习。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])]), Author(id=1197274446097597296, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, 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=1197274448333161333, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274446097597296, language=EN, stringName=Jian ZHANG, firstName=Jian, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Beijing Aerospace Propulsion Institute,Beijing,100076, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1197274449121690487, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274446097597296, language=CN, stringName=张箭, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=北京航天动力研究所,北京,100076, bio={"content":"

张箭(1986—),男,高级工程师,主要研究方向为动力系统总体设计。

"}, bioImg=null, bioContent=

张箭(1986—),男,高级工程师,主要研究方向为动力系统总体设计。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])]), Author(id=1197274449834722170, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, 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=1197274450476450683, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274449834722170, language=EN, stringName=Wei XUE, firstName=Wei, middleName=null, lastName=XUE, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Beijing Aerospace Propulsion Institute,Beijing,100076, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1197274452179338110, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274449834722170, language=CN, stringName=薛薇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=北京航天动力研究所,北京,100076, bio={"content":"

薛薇(1981—),女,博士,高级工程师,主要研究方向为动力系统总体设计及健康管理。

"}, bioImg=null, bioContent=

薛薇(1981—),女,博士,高级工程师,主要研究方向为动力系统总体设计及健康管理。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])]), Author(id=1197274452397441919, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, 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=1197274452707820418, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274452397441919, language=EN, stringName=Nan ZHANG, firstName=Nan, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=Beijing Aerospace Propulsion Institute,Beijing,100076, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1197274452967867268, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, authorId=1197274452397441919, language=CN, stringName=张楠, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=北京航天动力研究所,北京,100076, bio={"content":"

张楠(1960—),男,博士,研究员,主要研究方向为动力系统总体设计及健康管理。

"}, bioImg=null, bioContent=

张楠(1960—),男,博士,研究员,主要研究方向为动力系统总体设计及健康管理。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])])], keywords=[Keyword(id=1197274454926607242, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, orderNo=1, keyword=multivariate time series), Keyword(id=1197274456755323787, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, orderNo=2, keyword=attention mechanism), Keyword(id=1197274457074090894, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, orderNo=3, keyword=recurrent neural network), Keyword(id=1197274457380275088, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, orderNo=4, keyword=convolution neural network), Keyword(id=1197274457816482706, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, orderNo=5, keyword=autocorrelation analysis), Keyword(id=1197274458466599828, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, orderNo=1, keyword=多变量时间序列), Keyword(id=1197274459074773911, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, orderNo=2, keyword=注意力机制), Keyword(id=1197274460723135385, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, orderNo=3, keyword=循环神经网络), Keyword(id=1197274461587162012, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, orderNo=4, keyword=卷积神经网络), Keyword(id=1197274462530880412, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, orderNo=5, keyword=自相关性分析)], refs=[Reference(id=1197274490972455881, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2022, volume=48, issue=3, pageStart=79, pageEnd=86, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=刘子俊, 冯勇, 陈景龙, journalName=火箭推进, refType=null, unstructuredReference=刘子俊, 冯勇, 陈景龙, 等.基于多源数据的液体火箭发动机智能异常检测[J]. 火箭推进, 2022,48(3): 79-86., articleTitle=基于多源数据的液体火箭发动机智能异常检测, refAbstract=null), Reference(id=1197274492398519242, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2022, volume=48, issue=3, pageStart=79, pageEnd=86, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=LIU Zijun, FENG Yong, CHEN Jinglong, journalName=Journal of Rocket Propulsion, refType=null, unstructuredReference=LIU Zijun, FENG Yong, CHEN Jinglong, et al.Intelligent anomaly detection of liquid rocket engine with multi-source data[J]. Journal of Rocket Propulsion, 2022,48(3): 79-86., articleTitle=Intelligent anomaly detection of liquid rocket engine with multi-source data, refAbstract=null), Reference(id=1197274492700509131, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2021, volume=47, issue=4, pageStart=52, pageEnd=58, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=王珺, 吕海鑫, 陈景龙, journalName=火箭推进, refType=null, unstructuredReference=王珺, 吕海鑫, 陈景龙, 等.液体火箭发动机健康状态智能检测方法[J]. 火箭推进, 2021,47(4): 52-58., articleTitle=液体火箭发动机健康状态智能检测方法, refAbstract=null), Reference(id=1197274492935390156, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2021, volume=47, issue=4, pageStart=52, pageEnd=58, url=null, language=null, rfNumber=[2], rfOrder=3, authorNames=WANG Jun, LYU Haixin, CHEN Jinglong, journalName=Journal of Rocket Propulsion, refType=null, unstructuredReference=WANG Jun, LYU Haixin, CHEN Jinglong, et al.Intelligent detection method of liquid rocket engine health state[J]. Journal of Rocket Propulsion, 2021,47(4): 52-58., articleTitle=Intelligent detection method of liquid rocket engine health state, refAbstract=null), Reference(id=1197274493707142093, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2020, volume=28, issue=2, pageStart=33, pageEnd=38, url=null, language=null, rfNumber=[3], rfOrder=4, authorNames=邓晨, 薛薇, 郑孟伟, journalName=计算机测量与控制, refType=null, unstructuredReference=邓晨, 薛薇, 郑孟伟, 等.基于改进ARMA模型的火箭发动机稳态工况过程实时故障诊断方法研究[J]. 计算机测量与控制, 2020,28(2): 33-38., articleTitle=基于改进ARMA模型的火箭发动机稳态工况过程实时故障诊断方法研究, refAbstract=null), Reference(id=1197274494420173774, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2020, volume=28, issue=2, pageStart=33, pageEnd=38, url=null, language=null, rfNumber=[3], rfOrder=5, authorNames=DENG Chen, XUE Wei, ZHENG Mengwei, journalName=Computer Measurement and Control, refType=null, unstructuredReference=DENG Chen, XUE Wei, ZHENG Mengwei, et al.Study on real-time diagnosis method of the main stage working condition of rocket engine based on improved ARMA method[J]. Computer Measurement and Control, 2020,28(2): 33-38., articleTitle=Study on real-time diagnosis method of the main stage working condition of rocket engine based on improved ARMA method, refAbstract=null), Reference(id=1197274495015764943, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2005, volume=20, issue=6, pageStart=1088, pageEnd=1092, url=null, language=null, rfNumber=[4], rfOrder=6, authorNames=杨雪, 张振鹏, 向世勇, journalName=航空动力学报, refType=null, unstructuredReference=杨雪, 张振鹏, 向世勇, 等.基于自回归模型的自适应阈值算法[J]. 航空动力学报, 2005,20(6): 1088-1092., articleTitle=基于自回归模型的自适应阈值算法, refAbstract=null), Reference(id=1197274496701875152, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2005, volume=20, issue=6, pageStart=1088, pageEnd=1092, url=null, language=null, rfNumber=[4], rfOrder=7, authorNames=YANG Xue, ZHANG Zhenpeng, XIANG Shiyong, journalName=Journal of Aerospace Power, refType=null, unstructuredReference=YANG Xue, ZHANG Zhenpeng, XIANG Shiyong, et al.Adaptive threshold algorithm based on autoregressive model for fault detection of LRE in ground test[J]. Journal of Aerospace Power, 2005,20(6): 1088-1092., articleTitle=Adaptive threshold algorithm based on autoregressive model for fault detection of LRE in ground test, refAbstract=null), Reference(id=1197274496928367569, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[5], rfOrder=8, authorNames=聂侥, journalName=基于过程神经网络的液体火箭发动机故障预测方法研究, refType=null, unstructuredReference=聂侥. 基于过程神经网络的液体火箭发动机故障预测方法研究[D]. 长沙: 国防科学技术大学, 2017., articleTitle=null, refAbstract=null), Reference(id=1197274497955972050, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[5], rfOrder=9, authorNames=NIE Yao, journalName=Investigation on fault prediction methods based on process neural network for liquid-propellant rocket engines, refType=null, unstructuredReference=NIE Yao. Investigation on fault prediction methods based on process neural network for liquid-propellant rocket engines[D]. Changsha: National University of Defense Technology, 2017., articleTitle=null, refAbstract=null), Reference(id=1197274498593506259, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=http://arxiv. org/abs/1704.02971v1, language=null, rfNumber=[6], rfOrder=10, authorNames=QIN Yao, SONG Dongjin, CHE Haifeng, journalName=null, refType=null, unstructuredReference=QIN Yao, SONG Dongjin, CHE Haifeng, et al.A dual-stage attention-based recurrent neural network for time series prediction[EB/OL].(2017-04-07)[2017-08-17]. http://arxiv. org/abs/1704.02971v1., articleTitle=A dual-stage attention-based recurrent neural network for time series prediction, refAbstract=null), Reference(id=1197274499235234772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=1997, volume=9, issue=8, pageStart=1735, pageEnd=1780, url=null, language=null, rfNumber=[7], rfOrder=11, 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=1197274499906323413, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=http://arxiv. org/abs/1706.03762v2, language=null, rfNumber=[8], rfOrder=12, authorNames=VASWANI A, SHAZEER N, PARMAR N, journalName=null, refType=null, unstructuredReference=VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[EB/OL].(2017-06-12)[2017-06-19]. http://arxiv. org/abs/1706.03762v2., articleTitle=Attention is all you need, refAbstract=null), Reference(id=1197274501596627926, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=143, pageStart=113082, pageEnd=null, url=null, language=null, rfNumber=[9], rfOrder=13, authorNames=LIU Yeqi, GONG Chuanyang, LING Yang, journalName=Expert Systems with Applications, refType=null, unstructuredReference=LIU Yeqi, GONG Chuanyang, LING Yang, et al.DSTP-RNN: a dual-stage two-phase attention-based recurrent neural networks for long-term and multivariate time series prediction[J]. Expert Systems with Applications, 2020143:113082., articleTitle=DSTP-RNN: a dual-stage two-phase attention-based recurrent neural networks for long-term and multivariate time series prediction, refAbstract=null), Reference(id=1197274502460654551, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=http://arxiv.org/abs/1703.07015, language=null, rfNumber=[10], rfOrder=14, authorNames=LAI Guokun, CHANG Weicheng, YANG Yiming, journalName=null, refType=null, unstructuredReference=LAI Guokun, CHANG Weicheng, YANG Yiming, et al.Modeling long-and short-term temporal patterns with deep neural networks[EB/OL].(2017-03-21)[2018-04-18]. http://arxiv.org/abs/1703.07015., articleTitle=Modeling long-and short-term temporal patterns with deep neural networks, refAbstract=null), Reference(id=1197274502909445080, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=http://arxiv.org/abs/1809.04206v1, language=null, rfNumber=[11], rfOrder=15, authorNames=SHI Shunyao, SUN Fankeng, LEE Hungyi, journalName=null, refType=null, unstructuredReference=SHI Shunyao, SUN Fankeng, LEE Hungyi. Temporal pattern attention for multivariate time series forecasting[EB/OL].(2016-09-12)[2019-09-18]. http://arxiv.org/abs/1809.04206v1., articleTitle=Temporal pattern attention for multivariate time series forecasting, refAbstract=null), Reference(id=1197274503622476761, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2019, volume=null, issue=8, pageStart=3144, pageEnd=3158, url=null, language=null, rfNumber=[12], rfOrder=16, authorNames=ZHU Xiaobin, CHENG Yuqiang, WU Jianjun, journalName=IEEE Access, refType=null, unstructuredReference=ZHU Xiaobin, CHENG Yuqiang, WU Jianjun, et al.Steady-state process fault detection for liquid rocket engines based on convolutional auto-encoder and one-class support vector machine[J]. IEEE Access, 20198:3144-3158., articleTitle=Steady-state process fault detection for liquid rocket engines based on convolutional auto-encoder and one-class support vector machine, refAbstract=null), Reference(id=1197274504528446426, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2002, volume=21, issue=1, pageStart=1, pageEnd=4, url=null, language=null, rfNumber=[13], rfOrder=17, authorNames=朱恒伟, 王克昌, 陈启智, journalName=推进技术, refType=null, unstructuredReference=朱恒伟, 王克昌, 陈启智. 液体火箭发动机地面试车故障检测的自适应阈值算法[J]. 推进技术, 2002,21(1): 1-4., articleTitle=液体火箭发动机地面试车故障检测的自适应阈值算法, refAbstract=null), Reference(id=1197274506264888283, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2002, volume=21, issue=1, pageStart=1, pageEnd=4, url=null, language=null, rfNumber=[13], rfOrder=18, authorNames=ZHU Hengwei, WANG Kechang, CHEN Qizhi, journalName=Journal of Propulsion Technology, refType=null, unstructuredReference=ZHU Hengwei, WANG Kechang, CHEN Qizhi. Adaptive thresholds algorithm for fault detection of liquid rocket engine in ground test[J]. Journal of Propulsion Technology, 2002,21(1): 1-4., articleTitle=Adaptive thresholds algorithm for fault detection of liquid rocket engine in ground test, refAbstract=null), Reference(id=1197274506940171228, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, doi=null, pmid=null, pmcid=null, year=2012, volume=48, issue=2, pageStart=1739, pageEnd=1747, url=null, language=null, rfNumber=[14], rfOrder=19, authorNames=CHUNG Weiho, CHEN Chiaoen, journalName=IEEE Transactions on Aerospace and Electronic Systems, refType=null, unstructuredReference=CHUNG Weiho, CHEN Chiaoen. Detecting number of coherent signals in array processing by Ljung-box statistic[J]. IEEE Transactions on Aerospace and Electronic Systems, 2012,48(2): 1739-1747., articleTitle=Detecting number of coherent signals in array processing by Ljung-box statistic, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1197274439575454555, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, xref=null, ext=[AuthorCompanyExt(id=1197274439583843163, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=Beijing Aerospace Propulsion Institute,Beijing,100076), AuthorCompanyExt(id=1197274439592231772, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, companyId=1197274439575454555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=北京航天动力研究所,北京,100076)])], figs=[ArticleFig(id=1197274463940166561, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 1, caption=Structure of LSTM, figureFileSmall=p6LJJ/BYhKDSmg0vaufHTw==, figureFileBig=EcNXO2K0emP5gGcoOBP5QA==, tableContent=null), ArticleFig(id=1197274465538196386, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图1, caption=LSTM结构, figureFileSmall=p6LJJ/BYhKDSmg0vaufHTw==, figureFileBig=EcNXO2K0emP5gGcoOBP5QA==, tableContent=null), ArticleFig(id=1197274467429827493, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 2, caption=Structure of sequence to sequence model, figureFileSmall=aIWaCL+ZWzq898akTin7KQ==, figureFileBig=ANHbEITlSE7r3BZPoiyejQ==, tableContent=null), ArticleFig(id=1197274468264493990, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图2, caption=编码器-解码器结构, figureFileSmall=aIWaCL+ZWzq898akTin7KQ==, figureFileBig=ANHbEITlSE7r3BZPoiyejQ==, tableContent=null), ArticleFig(id=1197274469921244071, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 3, caption=Encoder: attention for feature dimension, figureFileSmall=kClSlJql0x7ZZE6Q8uj8DA==, figureFileBig=P2DDtUOS524lw8c3R3EKDA==, tableContent=null), ArticleFig(id=1197274470579749802, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图3, caption=编码器:特征维度注意力机制, figureFileSmall=kClSlJql0x7ZZE6Q8uj8DA==, figureFileBig=P2DDtUOS524lw8c3R3EKDA==, tableContent=null), ArticleFig(id=1197274471536051114, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 4, caption=Decoder: attention for time dimension, figureFileSmall=4/ZWmqPdqC17sVWNCfu9PA==, figureFileBig=NRBoNhYlvqBRRKB4i8UTPw==, tableContent=null), ArticleFig(id=1197274472257471403, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图4, caption=解码器:时间维度注意力机制, figureFileSmall=4/ZWmqPdqC17sVWNCfu9PA==, figureFileBig=NRBoNhYlvqBRRKB4i8UTPw==, tableContent=null), ArticleFig(id=1197274472458797997, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 5, caption=Oxygen turbine pump speed and adaptive threshold, figureFileSmall=QAc0X6HCv5U1C6tNongZSw==, figureFileBig=dOLOp37J/IzIJdGyzxwjFw==, tableContent=null), ArticleFig(id=1197274472865645487, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图5, caption=氧涡轮泵转速及自适应阈值, figureFileSmall=QAc0X6HCv5U1C6tNongZSw==, figureFileBig=dOLOp37J/IzIJdGyzxwjFw==, tableContent=null), ArticleFig(id=1197274475143152562, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 6, caption=Gas generator pressure, hydrogen turbine inlet pressure and adaptive threshold, figureFileSmall=9amAmD0J/HR2TSyVPxjGdQ==, figureFileBig=M29gbWnyNme9non2/TZ1Cw==, tableContent=null), ArticleFig(id=1197274475847795634, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图6, caption=燃气发生器压力、氢涡轮入口压力及自适应阈值, figureFileSmall=9amAmD0J/HR2TSyVPxjGdQ==, figureFileBig=M29gbWnyNme9non2/TZ1Cw==, tableContent=null), ArticleFig(id=1197274476552438708, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 7, caption=Process of fault detection strategy based on CDA-RNN, figureFileSmall=kuTQF598bNJ5N92NcDfHxA==, figureFileBig=6VrY76YkCyQT24TcGBGuOA==, tableContent=null), ArticleFig(id=1197274476766348214, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图7, caption=基于CDA-RNN的故障检测策略流程, figureFileSmall=kuTQF598bNJ5N92NcDfHxA==, figureFileBig=6VrY76YkCyQT24TcGBGuOA==, tableContent=null), ArticleFig(id=1197274477018006456, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 8, caption=Train set, figureFileSmall=Oi3wXA8+Oqs/8p+hLvk6sg==, figureFileBig=KvEsHSTVczAm7MIL+Pjkdg==, tableContent=null), ArticleFig(id=1197274477248693178, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图8, caption=训练集, figureFileSmall=Oi3wXA8+Oqs/8p+hLvk6sg==, figureFileBig=KvEsHSTVczAm7MIL+Pjkdg==, tableContent=null), ArticleFig(id=1197274478880277436, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 9, caption=A1 test normalized oxygen turbine pump speed, figureFileSmall=ShyNZmq+8vmbxnOyLjZLCw==, figureFileBig=O58aoSvCKx0xKvAIzvia5Q==, tableContent=null), ArticleFig(id=1197274479161295806, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图9, caption=$\mathrm{A}1$次试车归一化氧涡轮泵转速, figureFileSmall=ShyNZmq+8vmbxnOyLjZLCw==, figureFileBig=O58aoSvCKx0xKvAIzvia5Q==, tableContent=null), ArticleFig(id=1197274479836578752, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 10, caption=A2 test normalized oxygen turbine pump speed, figureFileSmall=GBBfWgKaBCpNAqtDSyX+lA==, figureFileBig=2YfqNpjkk4atlyiC2PsY2Q==, tableContent=null), ArticleFig(id=1197274480629302208, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图10, caption=A2次试车归一化氧涡轮泵转速, figureFileSmall=GBBfWgKaBCpNAqtDSyX+lA==, figureFileBig=2YfqNpjkk4atlyiC2PsY2Q==, tableContent=null), ArticleFig(id=1197274481585603521, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 11, caption=B1 test normalized gas generator pressure, figureFileSmall=Cm+Y1Y1TnAO0cKTGK1p94A==, figureFileBig=8VOWY3X0jY+ugd0GjvSIDw==, tableContent=null), ArticleFig(id=1197274483225576386, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图11, caption=B1次试车归一化燃气发生器压力, figureFileSmall=Cm+Y1Y1TnAO0cKTGK1p94A==, figureFileBig=8VOWY3X0jY+ugd0GjvSIDw==, tableContent=null), ArticleFig(id=1197274484144128963, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 12, caption=Physical significance of fault confidence probability, figureFileSmall=z8a/HHSCEFf4S9BEkmXOtQ==, figureFileBig=jUlpyFddQwJh0ytmHxpBkw==, tableContent=null), ArticleFig(id=1197274485037515716, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图12, caption=故障置信概率的物理意义, figureFileSmall=z8a/HHSCEFf4S9BEkmXOtQ==, figureFileBig=jUlpyFddQwJh0ytmHxpBkw==, tableContent=null), ArticleFig(id=1197274485750547397, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Fig. 13, caption=B1 test fault confidence probability, figureFileSmall=SyaRMdwEPmj8SNpH6mUCxA==, figureFileBig=jQHDxcVqvgn2gYSEAcrMzA==, tableContent=null), ArticleFig(id=1197274486425830342, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=图13, caption=B1次试车故障置信概率, figureFileSmall=SyaRMdwEPmj8SNpH6mUCxA==, figureFileBig=jQHDxcVqvgn2gYSEAcrMzA==, tableContent=null), ArticleFig(id=1197274488283907015, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=EN, label=Tab. 1, caption=Prediction error of each model in test set, figureFileSmall=null, figureFileBig=null, tableContent=
模型no预测(Al/A2)${Pg}$预测(B1)
MAEMSEMAEMSE
LSTM46.84/50.803827/4716失效失效
编码器-解码器LSTM45.62/45.783771/38020.042570.003425
DA-RNN44.41/44.583690/35440.042310.003383
CDA-RNN44.26/42.233437/31540.040940.003118
), ArticleFig(id=1197274489093407688, tenantId=1146029695717560320, journalId=1146119989267898375, articleId=1148011761636209522, language=CN, label=表1, caption=各模型测试集预测误差, figureFileSmall=null, figureFileBig=null, tableContent=
模型no预测(Al/A2)${Pg}$预测(B1)
MAEMSEMAEMSE
LSTM46.84/50.803827/4716失效失效
编码器-解码器LSTM45.62/45.783771/38020.042570.003425
DA-RNN44.41/44.583690/35440.042310.003383
CDA-RNN44.26/42.233437/31540.040940.003118
)], attaches=null, journal=Journal(id=1146118917132496903, delFlag=0, nameCn=导弹与航天运载技术(中英文), nameEn=Missiles and Space Vehicles, nameHistory1=null, nameHistory2=null, issn=2097-1974, eissn=, cn=10-1807/V, coden=null, periodic=1, language=CN, oaType=否, 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=DdfacUcWE1ibGHaCsbhL8w==, journalPrice=null, startedYear=null, abbrevIsoEn=Miss Space Veh, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1753780023753, createdBy=null, updatedBy=13701087609, firstLetterCn=M, firstLetterEn=M, subjectCode=Engineering, subjectName=工程, subjectCodeEn=Engineering, subjectNameEn=null, picCn=DdfacUcWE1ibGHaCsbhL8w==, picEn=jBog3LY3UHfAAYfYEqIBHg==, jcr=null, cjcr=null, exts=[JournalExt(id=1157000535829205111, language=CN, name=导弹与航天运载技术(中英文), 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=1753780023772, updatedTime=1753780023772, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=null, submissionAuthorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, submissionEditorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, submissionReviewUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, submissionCeEditorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, submissionAeEditorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, option={"copyright":""}), JournalExt(id=1157000535871148152, language=EN, name=Missiles and Space Vehicles, 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=1753780023782, updatedTime=1753780023782, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=null, submissionAuthorUrl=https://journal.ids.fzyun.cn/auth/realms/journal/protocol/openid-connect/auth?client_id=journal-ddht-author&redirect_uri=https%3A%2F%2Fddht.portal.founderss.cn%2Foauth%2Fcallback&response_type=code&scope=phone+openid+email+profile&state=fefa73ea-30bb-4b68, submissionEditorUrl=, submissionReviewUrl=, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1146119989267898375, websiteList=[Website(id=1148243202378817956, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146119989267898375, 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/ddyht/CN, language=CN, createTime=1751692112774, createBy=18614031015, updateTime=1753519037132, updateBy=18614031015, name=导弹与航天运载技术-中文站点, tplId=1146099689490845704, title=导弹与航天运载技术(中英文), delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148620442723704855, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202378817956, code=articleTextType, value=kx, createTime=1751782053881, updateTime=1751782053881, creator=18614031015, updator=18614031015), WebsiteProps(id=1148620442694344724, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202378817956, code=banner, value=null, createTime=1751782053874, updateTime=1751782053874, creator=18614031015, updator=18614031015), WebsiteProps(id=1148620442677567507, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202378817956, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=gfUyzanfTuxQ2yc+L/MeWA==, createTime=1751782053870, updateTime=1751782053870, creator=18614031015, updator=18614031015), WebsiteProps(id=1148620442711121942, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202378817956, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751782053878, updateTime=1751782053878, creator=18614031015, updator=18614031015), WebsiteProps(id=1148620442706927637, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202378817956, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751782053877, updateTime=1751782053877, creator=18614031015, updator=18614031015)]), Website(id=1155906039850618895, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146119989267898375, 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/ddyht/EN, language=EN, createTime=1753519075604, createBy=18614031015, updateTime=1753519075604, updateBy=18614031015, name=导弹与航天运载技术-英文站点, tplId=1146101810881728533, title=Missiles and Space Vehicles, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155907112585777347, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155906039850618895, code=articleTextType, value=kx, createTime=1753519331356, updateTime=1753519331356, creator=18614031015, updator=18614031015), WebsiteProps(id=1155907112556417216, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155906039850618895, code=banner, value=null, createTime=1753519331349, updateTime=1753519331349, creator=18614031015, updator=18614031015), WebsiteProps(id=1155907112552222911, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155906039850618895, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=gfUyzanfTuxQ2yc+L/MeWA==, createTime=1753519331348, updateTime=1753519331348, creator=18614031015, updator=18614031015), WebsiteProps(id=1155907112577388738, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155906039850618895, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753519331354, updateTime=1753519331354, creator=18614031015, updator=18614031015), WebsiteProps(id=1155907112564805825, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155906039850618895, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753519331351, updateTime=1753519331351, creator=18614031015, updator=18614031015)])], journalTitle=导弹与航天运载技术(中英文), weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=Missiles and Space Vehicles, journalPhotoCn=DdfacUcWE1ibGHaCsbhL8w==, journalPhotoEn=jBog3LY3UHfAAYfYEqIBHg==, journalFirstLetter=M, 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/ddyht/CN/10.7654/j.issn.2097-1974.20240204, detailUrlEn=https://castjournals.cast.org.cn/joweb/ddyht/EN/10.7654/j.issn.2097-1974.20240204, pdfUrlCn=https://castjournals.cast.org.cn/joweb/ddyht/CN/PDF/10.7654/j.issn.2097-1974.20240204, pdfUrlEn=https://castjournals.cast.org.cn/joweb/ddyht/EN/PDF/10.7654/j.issn.2097-1974.20240204, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于注意力机制循环神经网络的液体火箭发动机故障检测
收藏切换
PDF下载
张万旋 , 卢哲 , 张箭 , 薛薇 , 张楠
导弹与航天运载技术 | 动力系统 2024,47(2): 25-31
收起
收藏切换
导弹与航天运载技术 | 动力系统 2024, 47(2): 25-31
基于注意力机制循环神经网络的液体火箭发动机故障检测
全屏
张万旋, 卢哲, 张箭, 薛薇, 张楠
作者信息
  • 北京航天动力研究所,北京,100076
  • 张万旋(1994—),男,博士,工程师,主要研究方向为动力系统总体设计及健康管理。

    卢哲(1992—),男,工程师,主要研究方向为机器学习和深度学习。

    张箭(1986—),男,高级工程师,主要研究方向为动力系统总体设计。

    薛薇(1981—),女,博士,高级工程师,主要研究方向为动力系统总体设计及健康管理。

    张楠(1960—),男,博士,研究员,主要研究方向为动力系统总体设计及健康管理。

Rocket Engine Fault Detection with Attention based Recurrent Neural Networks
Wanxuan ZHANG, Zhe LU, Jian ZHANG, Wei XUE, Nan ZHANG
Affiliations
  • Beijing Aerospace Propulsion Institute,Beijing,100076
出版时间: 2024-04-25 doi: 10.7654/j.issn.2097-1974.20240204
文章导航
收藏切换

针对液体火箭发动机主级段工作过程,采用多变量非线性时间序列分析理论,在两级注意力机制循环神经网络(Dual Stage Attention Based Recurrent Neural Networks,DA-RNN)的基础上,提出一种新型时序分析工具——卷积两级注意力机制循环神经网络(Convolutional Dual Stage Attention Based Recurrent Neural Networks, CDA-RNN),从而建立故障趋势预测模型。通过对预测残差进行自相关性分析并定义故障置信概率,提出了故障检测量化依据。利用发生微弱故障的热试车数据进行验证,结果表明,CDA-RNN模型对非稳态工作段微弱故障多参数检测具有良好鲁棒性,该方法十分有效,具有直接应用价值。

多变量时间序列  /  注意力机制  /  循环神经网络  /  卷积神经网络  /  自相关性分析

Focusing on the main working phase of liquid rocket engine, with the aid of multivariate non-linear time series analysis, and based on Dual Stage Attention Based Recurrent Neural Networks (DA-RNN), a new time series analysis tool, Convolutional Dual Stage Attention Based Recurrent Neural Networks (CDA-RNN), is proposed, by which a fault trend prediction model is established. Compared with LSTM, DA-RNN, etc, this model shows higher prediction accuracy. Combined with autocorrelation analysis of the prediction residual, a quantitative basis of fault detection is proposed after introducing failure confidence probability. Using hot test data with weak fault to validate the model, result shows that the CDA-RNN model enables robust weak fault muti-parameter detection in unsteady working process. This strategy is so effective that it calls for direct engineering application.

multivariate time series  /  attention mechanism  /  recurrent neural network  /  convolution neural network  /  autocorrelation analysis
张万旋, 卢哲, 张箭, 薛薇, 张楠. 基于注意力机制循环神经网络的液体火箭发动机故障检测. 导弹与航天运载技术, 2024 , 47 (2) : 25 -31 . DOI: 10.7654/j.issn.2097-1974.20240204
Wanxuan ZHANG, Zhe LU, Jian ZHANG, Wei XUE, Nan ZHANG. Rocket Engine Fault Detection with Attention based Recurrent Neural Networks[J]. Missiles and Space Vehicles, 2024 , 47 (2) : 25 -31 . DOI: 10.7654/j.issn.2097-1974.20240204
建立实时、可靠、精确的液体火箭发动机故障检测模型对运载系统健康监控十分重要。发动机是复杂强耦合热力系统, 用来表征发动机正常工作的特征量之间存在相关性, 特征量当前值与其历史值存在回归关系, 故障会破坏这种相互关系。根据特征之间相关性进行故障检测的算法例如主成分分析法(Principal Component Analysis, PCA)[1],其原理是利用样本的主成分向量还原样本, 通过对重构误差来判断系统是否发生故障。利用特征过去与未来回归关系进行故障检测的算法主要有移动平均自回归(Auto-Regressive Moving Average, ARMA)[2-4]模型。但是,上述方法均具有一定局限性, PCA方法利用静态数据进行分析, 没有将发动机动态特性纳入考量。ARMA模型只能对单一特征的时间序列进行分析。PCA与ARMA作为线性模型, 只能对平稳序列建模, 对于起动、关机、变工况等非平稳过程建模能力较差。针对线性检测模型的不足,聂侥[5]利用过程神经网络建立发动机参数预测模型, 但是该神经网络的基础结构是全连接层, 将带来计算量大、长序列梯度消失等问题。
两级注意力机制循环神经网络(Dual Stage Attention Based Recurrent Neural Network, DA-RNN)是Yao等[6]在2017年提出的时间序列分析模型,由两个长短期记忆(Long Short-Term Memory, LSTM)[7]神经网络按照编码器-解码器(encoder-decoder)架构组成, 主要用于多变量时间序列预测。其中, 在第1层LSTM中引入特征维度注意力机制[8],在第2层LSTM中引入时间维度注意力机制, 因此可以捕捉同一时刻不同特征、同一特征不同时刻之间的相互关系。基于注意力机制思想, 研究者们对其进行了诸多改进[9-11]。本文利用注意力机制,提出一种新型神经网络一一卷积两级注意力机制循环神经网络(Convolutional Dual Stage Attention Based Recurrent Neural Networks, CDA-RNN), 从而实现发动机多参数、非线性过程故障检测的目标。
LSTM是一种循环神经网络(Recurrent Neural Network, RNN), 通过设计记忆存储单元并引入3种门限(遗忘门、输入门和输出门),解决了长序列学习困难以及梯度消失和梯度爆炸的问题, 其结构如图1所示。
每个LSTM单元函数相同, 利用上一时刻隐层状态${h}_{t - 1}$${c}_{t - 1}$以及本时刻状态${x}_{t}$对下一时刻隐层状态进行计算, 并在最后一个单元预测下一时刻状态, 每个单元计算过程为
${i}_{t}= \sigma \left({{\mathbf{W}}_{i}\left\lbrack {{h}_{t - 1};{x}_{t}}\right\rbrack +{b}_{i}}\right)$
${\widetilde{c}}_{t}= \tan h\left({{\mathbf{W}}_{c}\left\lbrack {{h}_{t - 1};{x}_{t}}\right\rbrack +{b}_{c}}\right)$
${c}_{t}= {f}_{t}\cdot {c}_{t - 1}+ {i}_{t}\cdot {\widetilde{c}}_{t}$
${o}_{t}= \sigma \left({{\mathbf{W}}_{o}\left\lbrack {{h}_{t - 1};{x}_{t}}\right\rbrack +{b}_{o}}\right)$
${h}_{t}= {o}_{t}\cdot \tan h\left({c}_{t}\right)$
式中$W$为系数矩阵;$b$为偏置常数;$\sigma$为sigmoid函数。
DA-RNN是针对多变量时间序列预测问题提出的一种深度学习模型, 采用两个LSTM神经网络构成编码器-解码器结构,如图2所示。
定义输入序列为
${x}^{k}= \left\lbrack {{x}_{1}^{k},{x}_{2}^{k},{x}_{3}^{k},\cdots ,{x}_{T}^{k}}\right\rbrack ,1 \leq k \leq n $
式中$n$为特征数量;$T$为时间窗口长度。
目标序列为
${y}^{k}= \left\lbrack {{y}_{1}^{k},{y}_{2}^{k},{y}_{3}^{k},\cdots ,{y}_{T}^{k}}\right\rbrack ,1 \leq k \leq n $
目标是建立模型$f$${y}_{T + 1}$:
${y}_{T + 1}= f\left({{x}_{T},{x}_{T - 1},\cdots ,{y}_{T},{y}_{T - 1},\cdots }\right)$
在编码器中引入特征维度注意力机制, 如图3所示。
长度为$T$的序列注意力层softmax层注意力加权后的输入序列
编码器注意力机制捕捉第$k$个特征与$t - 1$时刻隐层单元的相关性, 注意力评分为
${e}_{t}^{k}= {v}_{f}\tan h\left({{\mathbf{W}}_{f}\cdot \left\lbrack {{h}_{t - 1};{s}_{t - 1}}\right\rbrack +{U}_{f}\cdot {x}^{k}+ {b}_{f}}\right)$
编码器注意力权重为
${\alpha }_{t}^{k}= \frac{\exp \left({e}_{t}^{k}\right)}{\mathop{\sum }\limits_{{k = 1}}^{n}\exp \left({e}_{t}^{k}\right)} $
由此得到注意力加权的输入序列:
${\widetilde{x}}_{t}= \left\lbrack {{\alpha }_{t}^{1}{x}_{t}^{1},{\alpha }_{t}^{2}{x}_{t}^{2},\cdots ,{\alpha }_{t}^{n}{x}_{t}^{n}}\right\rbrack ,1 \leq t \leq T $
$t$时刻编码器隐层单元为
${h}_{t}= f\left({{h}_{t - 1},{\widetilde{x}}_{t}}\right)$
在解码器中引入时间维度注意力机制, 如图4所示。
解码器注意力机制捕捉第$i$时刻的编码器隐层单元与$t - 1$时刻的解码器隐层单元之间的相关性,注意力评分为
${l}_{t}^{i}= {v}_{d}\tan h\left({{\mathbf{W}}_{d}\cdot \left\lbrack {{d}_{t - 1};{s}_{t - 1}^{\prime }}\right\rbrack +{U}_{d}\cdot {h}_{i}+ {b}_{d}}\right)$
解码器注意力权重为
${\beta }_{t}^{i}= \frac{\exp \left({l}_{t}^{i}\right)}{\mathop{\sum }\limits_{{i = 1}}^{T}\exp \left({l}_{t}^{i}\right)} $
得到中间矢量为
${\mathbf{c}}_{t}= \mathop{\sum }\limits_{{i = 1}}^{T}{\beta }_{t}^{i}{h}_{i},1 \leq t \leq T $
$t - 1$时刻中间矢量与$t$时刻目标序列拼接,用一个全连接层更新目标序列, 可得:
${\widetilde{y}}_{t}= \widetilde{w}\cdot \left\lbrack {{y}_{t};{\mathbf{c}}_{t - 1}}\right\rbrack +\widetilde{b}$
于是$t$时刻解码器隐层单元为
${d}_{t}= f\left({{d}_{t - 1},{\widetilde{y}}_{t}}\right)$
最后计算得到$T + 1$时刻预测矢量为
${\mathbf{y}}_{T + 1}= {v}_{y}\left({{\mathbf{W}}_{y}\cdot \left\lbrack {{d}_{T};{\mathbf{c}}_{T}}\right\rbrack +{b}_{w}}\right)+ {b}_{y}$
文献[12]指出, 利用卷积神经网络可对多变量时间序列进行特征提取, 能够提高模型非线性, 并对时间序列进行滤波, 满足液体火箭发动机多工况及非平稳段检测的需求。因此, 在DA-RNN之前引入一个卷积层, 在时间维度对数据进行滤波和特征提取, 从而建立本文CDA-RNN模型。
采用$k$$n \times w$卷积核沿着输入序列的时间轴滚动,$k$个卷积核滤波器将原特征进行非线性组合并对其进行滤波后,得到长度为$k$的特征向量,卷积层运算按下式给出:
$\begin{array}{l} \bar{x}_{j, t}=\tan h\left(C_{j} \hat{x}_{t}^{1}+b_{j}\right] \\ 1 \leqslant t \leqslant T-w+1,1 \leqslant j \leqslant k \end{array}$
式中${\widehat{x}}_{t}$为卷积核窗口内的时间序列;${C}_{j},{b}_{j}$为第$j$个卷积核矩阵和偏置;${\bar{x}}_{j, t}$为卷积核输出;$w$为卷积核时间窗口长度。
为了验证本文方法对微弱故障的有效性, 采用热试车数据进行验证。编号为$\mathrm{A}$的发动机在连续两次试车时(记为$\mathrm{A}1\text{、}\mathrm{\;A}2$)分别在约${17}\mathrm{\;s}\text{、}{21}\mathrm{\;s}$出现了主要参数凹坑现象, 此时, 发动机参数尚未稳定。分解检查发现, 该现象由导向环组件碰磨导致。图5给出了$\mathrm{A}1\text{、}\mathrm{\;A}2$次试车归一化氧涡轮泵转速${no}$曲线及由自适应阈值法[13]确定的阈值。
编号为$\mathrm{B}$的发动机在某次试车(记为$\mathrm{B}1$)主级段47.1~49.1 s时出现了压力参数凸台现象,以燃气路反应最为明显,试后分析知该现象与燃气四通内部的异常燃烧流动状态有关。图6给出了归一化燃气发生器压力${Pg}$、氢涡轮入口压力$P$itf曲线及其自适应阈值法确定的阈值。
其中, 自适应阈值法的窗口长度与带宽系数根据文献[13]的方法采用同技术状态产品历史正常试车数据确定。结果表明, 对于参数凹坑、参数凸台这种微弱故障情况, 无论如何选择带宽系数, 自适应阈值法都存在误检或漏检现象。
因此, 以参数凹坑和凸台等自适应阈值算法无法检测的微弱故障为研究对象, 基于CDA-RNN进行多参数预测检测, 如图7所示。
本文选取编号为$\mathrm{C}$的发动机4次正常试车(记为$\mathrm{C}1\text{、}\mathrm{C}2\text{、}\mathrm{C}3\text{、}\mathrm{C}4)$主级段$5 \sim {475}\mathrm{\;s}$数据对模型进行离线训练,将A1次$\left({{10}\sim {95}\mathrm{\;s}}\right)$、A2次$\left({{10}\sim {95}\mathrm{\;s}}\right)$、B1次$\left({5 \sim {95}\mathrm{\;s}}\right)$试车数据作为测试集。将4次试车数据拼接在一张图中, 如图8所示。
某型发动机试车测点共97个, C1~C4次试车形成了97个长为188000的特征序列。由于各传感器通道数据量纲差异较大,对数据进行标准化处理。
如果将97个特征全部输入预测模型进行计算, 将消耗大量计算资源, 并且将引入无关特征, 对预测过程造成干扰。因此, 采取相关性分析法进行特征选取, 即对97个特征分别与目标特征进行皮尔逊相关系数计算, 按下式给出:
$\rho \left({x, y}\right)= \frac{\operatorname{Cov}\left({x, y}\right)}{\sqrt{\operatorname{Var}\left( x\right)\cdot \operatorname{Var}\left( y\right)}}$
综合考虑计算量和相关性,针对氧涡轮泵转速${no}$预测,取相关系数0.8以上的特征,得到${no}$的相关特征为 [‘no’,‘Pigv1’,‘qmpf’,‘Pitf’,‘Petf’,‘Pgs’,‘Pgifs’,‘Pgios’,‘Pefs’,‘Pwif’], 针对燃气发生器压力Pg预测, 取相关系数0.7以上的特征,得到${Pg}$的相关特征为$\left\lbrack {{}^{\prime }P{g}^{\prime }}\right.$, ‘Pgifs’,‘Pgios’]。由于该型发动机采用燃气阀对混合比进行阶跃式调节, 主级工作段具有低、中、高3种工况。 因此, 将燃气阀门动作作为输入特征之一, 其每一时刻的取值为-1、0、1,分别对应发动机低、中、高工况。
为验证模型预测效果, 采用工业界常用的时序预测模型LSTM、编码器-解码器LSTM模型(无注意力机制)与DA-RNN、CDA-RNN共4种模型对A1、A2次试车no参数和B1次试车${Pg}$参数进行预测,单步预测时间为${0.2}\mathrm{\;{ms}}$,预测误差如表1所示。
两两比较表1中模型可知: a)编码器-解码器结构提高了预测精度, 这是因为编码器对特征进行了提取;b)注意力机制模型预测精度较高,这是因为注意力机制捕捉了特征之间、不同时刻同一特征之间的关联性;c)卷积层的引入进一步提高了预测精度, 这是由于卷积层对特征做了进一步提取。
正常样本的预测残差服从高斯白噪声分布, 而故障样本预测残差不满足[2],因此,可采取Ljung-Box (LB)检验[14]判断残差是否服从白噪声分布。计算残差自相关系数进行假设检验:
${H}_{0}: {\rho }_{1}= {\rho }_{2}= \ldots ={\rho }_{m}= 0,{H}_{1}: \exists i\text{ s.t.}{\rho }_{i}\neq 0 $
构造$Q$统计量如下:
$ Q\left( m\right)= T\left({T + 2}\right)\mathop{\sum }\limits_{{l = 1}}^{m}\frac{{\rho }_{l}^{2}}{T - l}$
式中${\rho }_{i}$为延迟时刻为$i$的相关系数。如果$Q\left( m\right)>$${\chi }_{m}^{2}\left(\alpha \right)$或者$Q\left( m\right)$对应的$p$值小于$\alpha$,则拒绝原假设${H}_{0}$
采用CDA-RNN模型对A1、A2、B1次试车进行检测,结果分别如图9~图11所示。
图9~图10可知, 在故障阶段CDA-RNN模型预测残差的$p$值在故障期间迅速降到0.05置信度以下, 保持趋于零的低水平。故障恢复后,$p$值迅速回到0.05以上并在较高水平。因此, 该模型能有效检测出碰磨故障和燃烧流动异常微弱故障。由图9残差时序可以看出,无论是${17}\mathrm{\;s}$前的参数非稳态段,还是${80}\sim {95}\mathrm{\;s}$的变推力阶段,模型预测误差未显著区别于平稳段。因此CDA-RNN对于变工况等非线性过程具有较好的检测鲁棒性。这是因为CDA-RNN引入了卷积层,卷积核对一定长度(0.1s)的时序片段进行非线性映射, 能够检测出该样本处于哪种工作状态, 因此取得了较好的非稳态段检测稳定性。
为了使检测结果更加直观,根据$\mathrm{{LB}}$检验$p$值的特性,定义时刻$t$的故障置信概率${P}_{f}$如下:
${P}_{f}\left( t\right)= \frac{{\int }_{t -{\Delta t}}^{t}\max \left({a - p,0}\right)\mathrm{d}t}{a\Delta t}$
式中${\Delta t}$为检测窗口;$a$为置信阈值;${P}_{f}$的物理意义为${\Delta t}$时间窗口内置信度线$a$以下$p$值线以上闭合曲线的面积(如图12阴影部分所示)与长为$a$宽为${\Delta t}$的矩形面积之比。取${\Delta t}= 1\mathrm{\;s}, a ={0.05}$,以$\mathrm{B}1$次试车$\mathrm{{CDA}}- \mathrm{{RNN}}$模型检测为例,得到${P}_{f}$曲线如图13所示。B1次试车故障置信概率在第${47}\mathrm{\;s}$迅速上升至接近1且保持到第${49}\mathrm{\;s}$,依据故障置信概率图,可以为故障检测提供可靠的量化依据。
本文提出CDA-RNN模型,结合残差$\mathrm{{LB}}$检验对试车中的微弱故障进行检测,结果表明:
a)由于卷积层、编码器解码器结构、注意力机制的引入, CDA-RNN模型能够对起动后非稳态段、 稳态工作段和变推力段的时序数据进行多参数高精度预测, 预测精度高于LSTM等传统模型, 改进了ARMA等线性模型只能对单参数、平稳段样本进行检测的不足。
b)针对预测残差进行LB白噪声检验,能够对转子碰磨、压力凸台等微弱故障进行有效识别, 所定义的故障置信概率能够为故障检测提供量化依据。
参考文献 引证文献
排序方式:
[1]
刘子俊, 冯勇, 陈景龙, 等.基于多源数据的液体火箭发动机智能异常检测[J]. 火箭推进, 2022,48(3): 79-86.
LIU Zijun, FENG Yong, CHEN Jinglong, et al.Intelligent anomaly detection of liquid rocket engine with multi-source data[J]. Journal of Rocket Propulsion, 2022,48(3): 79-86.
[2]
王珺, 吕海鑫, 陈景龙, 等.液体火箭发动机健康状态智能检测方法[J]. 火箭推进, 2021,47(4): 52-58.
WANG Jun, LYU Haixin, CHEN Jinglong, et al.Intelligent detection method of liquid rocket engine health state[J]. Journal of Rocket Propulsion, 2021,47(4): 52-58.
[3]
邓晨, 薛薇, 郑孟伟, 等.基于改进ARMA模型的火箭发动机稳态工况过程实时故障诊断方法研究[J]. 计算机测量与控制, 2020,28(2): 33-38.
DENG Chen, XUE Wei, ZHENG Mengwei, et al.Study on real-time diagnosis method of the main stage working condition of rocket engine based on improved ARMA method[J]. Computer Measurement and Control, 2020,28(2): 33-38.
[4]
杨雪, 张振鹏, 向世勇, 等.基于自回归模型的自适应阈值算法[J]. 航空动力学报, 2005,20(6): 1088-1092.
YANG Xue, ZHANG Zhenpeng, XIANG Shiyong, et al.Adaptive threshold algorithm based on autoregressive model for fault detection of LRE in ground test[J]. Journal of Aerospace Power, 2005,20(6): 1088-1092.
[5]
聂侥. 基于过程神经网络的液体火箭发动机故障预测方法研究[D]. 长沙: 国防科学技术大学, 2017.
NIE Yao. Investigation on fault prediction methods based on process neural network for liquid-propellant rocket engines[D]. Changsha: National University of Defense Technology, 2017.
[6]
QIN Yao, SONG Dongjin, CHE Haifeng, et al.A dual-stage attention-based recurrent neural network for time series prediction[EB/OL].(2017-04-07)[2017-08-17]. http://arxiv. org/abs/1704.02971v1. http://arxiv. org/abs/1704.02971v1
[7]
HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997,9(8): 1735-1780.
[8]
VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[EB/OL].(2017-06-12)[2017-06-19]. http://arxiv. org/abs/1706.03762v2. http://arxiv. org/abs/1706.03762v2
[9]
LIU Yeqi, GONG Chuanyang, LING Yang, et al.DSTP-RNN: a dual-stage two-phase attention-based recurrent neural networks for long-term and multivariate time series prediction[J]. Expert Systems with Applications, 2020143:113082.
[10]
LAI Guokun, CHANG Weicheng, YANG Yiming, et al.Modeling long-and short-term temporal patterns with deep neural networks[EB/OL].(2017-03-21)[2018-04-18]. http://arxiv.org/abs/1703.07015. http://arxiv.org/abs/1703.07015
[11]
SHI Shunyao, SUN Fankeng, LEE Hungyi. Temporal pattern attention for multivariate time series forecasting[EB/OL].(2016-09-12)[2019-09-18]. http://arxiv.org/abs/1809.04206v1. http://arxiv.org/abs/1809.04206v1
[12]
ZHU Xiaobin, CHENG Yuqiang, WU Jianjun, et al.Steady-state process fault detection for liquid rocket engines based on convolutional auto-encoder and one-class support vector machine[J]. IEEE Access, 20198:3144-3158.
[13]
朱恒伟, 王克昌, 陈启智. 液体火箭发动机地面试车故障检测的自适应阈值算法[J]. 推进技术, 2002,21(1): 1-4.
ZHU Hengwei, WANG Kechang, CHEN Qizhi. Adaptive thresholds algorithm for fault detection of liquid rocket engine in ground test[J]. Journal of Propulsion Technology, 2002,21(1): 1-4.
[14]
CHUNG Weiho, CHEN Chiaoen. Detecting number of coherent signals in array processing by Ljung-box statistic[J]. IEEE Transactions on Aerospace and Electronic Systems, 2012,48(2): 1739-1747.
2024年第47卷第2期
PDF下载
416
172
引用本文
BibTeX
文章信息
doi: 10.7654/j.issn.2097-1974.20240204
  • 接收时间:2021-12-03
  • 首发时间:2025-07-04
  • 出版时间:2024-04-25
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2021-12-03
  • 修回日期:2022-08-07
基金
作者信息
    北京航天动力研究所,北京,100076
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/ddyht/CN/10.7654/j.issn.2097-1974.20240204
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