Article(id=1276896715483513644, tenantId=1146029695717560320, journalId=1276577754012160025, issueId=1276896661737701828, articleNumber=null, orderNo=null, doi=10.3724/j.gyjzG26031308, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1773331200000, receivedDateStr=2026-03-13, revisedDate=null, revisedDateStr=null, acceptedDate=null, acceptedDateStr=null, onlineDate=1782365500565, onlineDateStr=2026-06-25, pubDate=1779206400000, pubDateStr=2026-05-20, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1782365500565, onlineIssueDateStr=2026-06-25, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1782365500565, creator=13701087609, updateTime=1782365500565, updator=13701087609, issue=Issue{id=1276896661737701828, tenantId=1146029695717560320, journalId=1276577754012160025, year='2026', volume='56', issue='5', pageStart='1', pageEnd='264', issueExtLink='null', onlineDate='null', pubDate='1779206400000', pubDateStr='2026-05-20', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1782365487751, creator='13701087609', updateTime=1782367237543, updator='13701087609', preIssue=null, nextIssue=null, articleTotal=null, ext={EN=IssueExt(id=1276904000968589318, tenantId=1146029695717560320, journalId=1276577754012160025, issueId=1276896661737701828, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1276904000968589319, tenantId=1146029695717560320, journalId=1276577754012160025, issueId=1276896661737701828, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null, downloadFileDto=null}, startPage=176, endPage=186, ext={EN=ArticleExt(id=1276896715697423150, articleId=1276896715483513644, tenantId=1146029695717560320, journalId=1276577754012160025, language=EN, title=Deep Learning-Based Imputation of Multi-Source Missing Data for Offshore Wind Turbines, columnId=null, journalTitle=Industrial Construction, columnName=null, runingTitle=null, highlight=null, articleAbstract=

To address the problem of missing multi-source monitoring data of offshore wind turbines caused by sensor failures or communication interruptions under harsh operating conditions, this paper proposes a novel imputation model based on a multi-head gated residual network. This method achieves collaborative fusion of supervisory control and data acquisition (SCADA) data and structural vibration monitoring data through feature concatenation, employs a gated residual network to extract deep nonlinear coupling features, and uses a multi-head parallel output architecture for the independent reconstruction of these two heterogeneous data types. During the training stage, a dynamic masking mechanism combined with a hybrid loss function is adopted to enhance the model’s adaptability to complex aerodynamic operating conditions. Validated with field data from a 10 MW offshore wind turbine, the proposed model achieved a high coefficient of determination under training conditions, enabling accurate reconstruction of missing multi-source data. In generalization tests for non-training periods, although the coefficient of determination of the model’s predictions fluctuated slightly, the model still effectively captured the overall trends of monitoring signals. Notably, the degradation in generalization performance for vibration data was less pronounced than that for SCADA data, demonstrating its greater stability. The proposed method can significantly improve the completeness and reliability of multi-source monitoring data for wind turbines and holds considerable potential for engineering applications.

, authors=null, authorsList=Dong LI, Yizhen LIAO, Yuan SANG, Zeyu LI, Bo YAO, Hongbing CHEN, authorCompany=null, 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, fund=null), CN=ArticleExt(id=1276896719036089153, articleId=1276896715483513644, tenantId=1146029695717560320, journalId=1276577754012160025, language=CN, title=基于深度学习的海上风机多源缺失数据的填补方法, columnId=1276896715768726319, journalTitle=工业建筑, columnName=工程运维技术, runingTitle=null, highlight=null, articleAbstract=

针对海上风电机组在恶劣工况下因传感器失效或通信中断引发的多源监测数据缺失问题,构建了一种基于多头门控残差网络的填补模型。该方法通过特征级联将数据采集与监视控制系统(SCADA)数据和结构振动监测数据进行协同融合,并引入门控残差网络提取深度非线性耦合特征;同时设计出多头并行输出结构,实现对两类异构数据的独立重建。训练阶段采用动态掩码机制与混合损失函数,以增强模型对复杂气动工况的适应能力。基于某10 MW海上风电机组实测数据的验证结果显示,该模型在训练工况下具有较高的决定系数,能够实现多源缺失数据的精准重构。在面向非训练时段的泛化测试中,模型预测的决定系数虽略有波动,但仍能有效捕捉信号整体变化趋势;尤其对振动类数据,其泛化性能衰减幅度小于SCADA数据,展现出更强的稳定性。该方法可显著提升风机多源监测数据的完整性与可靠性,具备良好的工程应用前景。

, authors=

李栋,副研究员,主要从事大跨度空间薄膜结构、组合结构、海上风电、结构风工程等方面研究,

, authorsList=李栋, 廖毅桢, 桑源, 李泽宇, 姚博, 陈洪兵, authorCompany=null, correspAuthors=null, authorNote=null, correspAuthorsNote=
陈洪兵,教授,主要从事工程结构损伤检测、服役安全智能诊断,钢-混凝土组合结构缺陷无损检测技术和多尺度多物理场仿真等领域研究,
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=/1sIbNb7HwklqEXr/3Qf+Q==, magXml=sjvceVbR+9FEL19TwTB7VQ==, pdfUrl=null, pdf=BXjsLdoo6b+qiKghuhdlLg==, pdfFileSize=11687872, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=E5TzAxcHHML7dg+n39dF1g==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=UmeJ5Wmv8KSkQXxobtzlXA==, mapNumber=null, fund=null)}, authors=[Author(id=1276896719665234767, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=dongli@fzu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1276896719921087313, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896719665234767, language=EN, stringName=Dong LI, firstName=Dong, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896720009167698, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896719665234767, language=CN, stringName=李栋, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio={"content":"

李栋,副研究员,主要从事大跨度空间薄膜结构、组合结构、海上风电、结构风工程等方面研究,

"}, bioImg=null, bioContent=

李栋,副研究员,主要从事大跨度空间薄膜结构、组合结构、海上风电、结构风工程等方面研究,

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896720302768980, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896720399237974, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720302768980, language=EN, stringName=Yizhen LIAO, firstName=Yizhen, middleName=null, lastName=LIAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896720730587991, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720302768980, language=CN, stringName=廖毅桢, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896720789308249, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896720852222811, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720789308249, language=EN, stringName=Yuan SANG, firstName=Yuan, middleName=null, lastName=SANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896721170989916, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720789308249, language=CN, stringName=桑源, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896721259070302, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896721590420320, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896721259070302, language=EN, stringName=Zeyu LI, firstName=Zeyu, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896723087786849, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896721259070302, language=CN, stringName=李泽宇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719413576518, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=2, ext=[AuthorCompanyExt(id=1276896719421965127, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China), AuthorCompanyExt(id=1276896719468102472, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083)])]), Author(id=1276896723146507107, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896723242976101, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723146507107, language=EN, stringName=Bo YAO, firstName=Bo, middleName=null, lastName=YAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3China MCC5 Group Co.,Ltd., Chengdu476300, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896723326862182, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723146507107, language=CN, stringName=姚博, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3中国五冶集团有限公司,成都476300, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719535211338, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=3, ext=[AuthorCompanyExt(id=1276896719543599947, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3China MCC5 Group Co.,Ltd., Chengdu476300, China), AuthorCompanyExt(id=1276896719551988556, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3中国五冶集团有限公司,成都476300)])]), Author(id=1276896723398165352, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=hongbingchen@ustb.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1276896723754681194, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723398165352, language=EN, stringName=Hongbing CHEN, firstName=Hongbing, middleName=null, lastName=CHEN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896724098614123, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723398165352, language=CN, stringName=陈洪兵, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719413576518, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=2, ext=[AuthorCompanyExt(id=1276896719421965127, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China), AuthorCompanyExt(id=1276896719468102472, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083)])])], keywords=[Keyword(id=1276896724610319212, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, orderNo=1, keyword=offshore wind turbines), Keyword(id=1276896725029749613, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, orderNo=2, keyword=data imputation), Keyword(id=1276896725394654062, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, orderNo=3, keyword=gated residual network), Keyword(id=1276896725772141423, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, orderNo=4, keyword=multi-source data fusion), Keyword(id=1276896725843444592, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, orderNo=1, keyword=海上风电机组), Keyword(id=1276896726204154737, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, orderNo=2, keyword=数据填补), Keyword(id=1276896727873487730, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, orderNo=3, keyword=门控残差网络), Keyword(id=1276896727944790899, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, orderNo=4, keyword=多源数据融合)], refs=[Reference(id=1276896742113149840, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2024, volume=17, issue=null, pageStart=3392695, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=WANG F X, ZHANG S, HOU Y Z, journalName=Journal of Selected Topics in Applied Earth Observations and Remote Sensing, refType=null, unstructuredReference=WANG F XZHANG SHOU Y Zet al. Extraction of offshore wind turbines in china by combining multispectral and SAR image data [J]. Journal of Selected Topics in Applied Earth Observations and Remote Sensing202417:3392695., articleTitle=Extraction of offshore wind turbines in china by combining multispectral and SAR image data, refAbstract=null), Reference(id=1276896743404995473, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=46, issue=2, pageStart=624, pageEnd=632, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=蔡彦枫, 李晓宇, 任宗栋, journalName=太阳能学报, refType=null, unstructuredReference=蔡彦枫,李晓宇,任宗栋,.面向空中风力发电系统的风能资源分析[J].太阳能学报202546(2):624-632., articleTitle=面向空中风力发电系统的风能资源分析, refAbstract=null), Reference(id=1276896743878951826, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=null, issue=1, pageStart=29, pageEnd=37, url=null, language=null, rfNumber=[3], rfOrder=2, authorNames=张珺辉, 董绍华, 叶迎春, journalName=新技术新工艺, refType=null, unstructuredReference=张珺辉,董绍华,叶迎春.基于LabVIEW的海上风机齿轮箱故障诊断系统设计[J].新技术新工艺2025(1):29-37., articleTitle=基于LabVIEW的海上风机齿轮箱故障诊断系统设计, refAbstract=null), Reference(id=1276896745548284819, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2020, volume=138, issue=null, pageStart=106587, pageEnd=null, url=null, language=null, rfNumber=[4], rfOrder=3, authorNames=LEI Y G, YANG B, JIANG X W, journalName=Mechanical Systems and Signal Processing, refType=null, unstructuredReference=LEI Y GYANG BJIANG X Wet al. Applications of machine learning to machine fault diagnosis: a review and roadmap [J]. Mechanical Systems and Signal Processing2020138: 106587., articleTitle=Applications of machine learning to machine fault diagnosis: a review and roadmap, refAbstract=null), Reference(id=1276896746357785492, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2019, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[5], rfOrder=4, authorNames=TAO S Y, QIAN Z, PEI Y, journalName=null, refType=null, unstructuredReference=TAO S YQIAN ZPEI Yet al. Wind turbine failure detection based on SCADA data and data mining method [C]// Renewable Power Generation Conference. 2019., articleTitle=Wind turbine failure detection based on SCADA data and data mining method, refAbstract=null), Reference(id=1276896746773021589, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2022, volume=8, issue=null, pageStart=834, pageEnd=847, url=null, language=null, rfNumber=[6], rfOrder=5, authorNames=ABORGEL T, SHEHATA A S, KOTB M A, journalName=Energy Reports, refType=null, unstructuredReference=ABORGEL TSHEHATA A SKOTB M Aet al. Heavy lift semi-submersible ships utilization in offshore wind turbines industry [J]. Energy Reports20228: 834-847., articleTitle=Heavy lift semi-submersible ships utilization in offshore wind turbines industry, refAbstract=null), Reference(id=1276896747246977942, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=380, issue=null, pageStart=124829, pageEnd=null, url=null, language=null, rfNumber=[7], rfOrder=6, authorNames=ZHANG W W, VATN J, RASHEED A, journalName=Applied Energy, refType=null, unstructuredReference=ZHANG W WVATN JRASHEED A. Gearbox pump failure prognostics in offshore wind turbine by an integrated data-driven model [J]. Applied Energy2025380: 124829., articleTitle=Gearbox pump failure prognostics in offshore wind turbine by an integrated data-driven model, refAbstract=null), Reference(id=1276896748048090007, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2024, volume=35, issue=null, pageStart=14747, pageEnd=null, url=null, language=null, rfNumber=[8], rfOrder=7, authorNames=HUSSAIN M, MIRJAT N H, SHAIKH F, journalName=Journal of Selected Topics in Applied Earth Observations and Remote Sensing, refType=null, unstructuredReference=HUSSAIN MMIRJAT N HSHAIKH Fet al. Condition monitoring and fault diagnosis of wind turbine: a systematic literature review [J]. Journal of Selected Topics in Applied Earth Observations and Remote Sensing202435:14747., articleTitle=Condition monitoring and fault diagnosis of wind turbine: a systematic literature review, refAbstract=null), Reference(id=1276896748492686232, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2023, volume=287, issue=null, pageStart=115862, pageEnd=null, url=null, language=null, rfNumber=[9], rfOrder=8, authorNames=GOROSTIDI N, PARDO D, NAVA V, journalName=Renewable Energy and Power Quality Journal, refType=null, unstructuredReference=GOROSTIDI NPARDO DNAVA V. Diagnosis of the health status of mooring systems for floating offshore wind turbines [J].Renewable Energy and Power Quality Journal2023287: 115862., articleTitle=Diagnosis of the health status of mooring systems for floating offshore wind turbines, refAbstract=null), Reference(id=1276896750141047705, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2022, volume=20, issue=null, pageStart=2172, pageEnd=2183, url=null, language=null, rfNumber=[10], rfOrder=9, authorNames=ENCALADA-DAVILA Á, TUTIVÉN C, MOYÓN L, journalName=Renewable Energy and Power Quality Journal, refType=null, unstructuredReference=ENCALADA-DAVILA ÁTUTIVÉN CMOYÓN Let al. Main bearing fault prognosis in wind turbines based on gated recurrent unit neural networks [J].Renewable Energy and Power Quality Journal202220: 2172-2183., articleTitle=Main bearing fault prognosis in wind turbines based on gated recurrent unit neural networks, refAbstract=null), Reference(id=1276896750577255323, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=44, issue=null, pageStart=101949, pageEnd=null, url=null, language=null, rfNumber=[11], rfOrder=10, authorNames=ZHAO J H, SHEN X D, LIU T J, journalName=Sustainable Energy, Grids and Networks, refType=null, unstructuredReference=ZHAO J HSHEN X DLIU T Jet al. Probabilistic imputation of missing offshore wind speed based on conditional diffusion models [J]. Sustainable Energy, Grids and Networks202544: 101949., articleTitle=Probabilistic imputation of missing offshore wind speed based on conditional diffusion models, refAbstract=null), Reference(id=1276896751005074332, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2020, volume=139, issue=null, pageStart=106610, pageEnd=null, url=null, language=null, rfNumber=[12], rfOrder=11, authorNames=QU F M, LIU J H, MA Y J, journalName=Mechanical Systems and Signal Processing, refType=null, unstructuredReference=QU F MLIU J HMA Y Jet al. A novel wind turbine data imputation method with multiple optimizations based on GANs [J]. Mechanical Systems and Signal Processing2020139: 106610., articleTitle=A novel wind turbine data imputation method with multiple optimizations based on GANs, refAbstract=null), Reference(id=1276896755539116957, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=2, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[13], rfOrder=12, authorNames=LIAO W L, BAK-JENSEN B, PILLAI J R, journalName=Journal of Modern Power Systems and Clean Energy, refType=null, unstructuredReference=LIAO W LBAK-JENSEN BPILLAI J Ret al. Data-driven missing data imputation for wind farms using context encoder [J].Journal of Modern Power Systems and Clean Energy2022(2):000894., articleTitle=Data-driven missing data imputation for wind farms using context encoder, refAbstract=null), Reference(id=1276896756000490399, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2024, volume=299, issue=null, pageStart=112114, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=13, authorNames=FAN Y W, FENG C L, WU R, journalName=Knowledge-Based Systems, refType=null, unstructuredReference=FAN Y WFENG C LWU Ret al. Multiscale-attention masked autoencoder for missing data imputation of wind turbines [J]. Knowledge-Based Systems2024299: 112114., articleTitle=Multiscale-attention masked autoencoder for missing data imputation of wind turbines, refAbstract=null), Reference(id=1276896756432503712, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2024, volume=300, issue=null, pageStart=112176, pageEnd=null, url=null, language=null, rfNumber=[15], rfOrder=14, authorNames=RADEV B, KASI V R, BINGI K, journalName=Knowledge-Based Systems, refType=null, unstructuredReference=RADEV BKASI V RBINGI K. Enhancement of Texas wind turbine power predictions using fractional order neural network by incorporating machine learning models to impute missing data [J]. Knowledge-Based Systems2024300: 112176., articleTitle=Enhancement of Texas wind turbine power predictions using fractional order neural network by incorporating machine learning models to impute missing data, refAbstract=null), Reference(id=1276896756860322722, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[16], rfOrder=15, authorNames=WANG B, ZHOU B, ZOU M H, journalName=null, refType=null, unstructuredReference=WANG BZHOU BZOU M Het al. Research on condition monitoring of offshore wind turbine gearbox based on K-means clustering and extreme learning machine [C]// International Conference on Artificial Intelligence and Power Engineering. 2023., articleTitle=Research on condition monitoring of offshore wind turbine gearbox based on K-means clustering and extreme learning machine, refAbstract=null), Reference(id=1276896757237810083, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2019, volume=187, issue=null, pageStart=106190, pageEnd=null, url=null, language=null, rfNumber=[17], rfOrder=16, authorNames=JU S H, HUANG Y C, journalName=Ocean Engineering, refType=null, unstructuredReference=JU S HHUANG Y C. Analyses of offshore wind turbine structures with soil-structure interaction under earthquakes [J]. Ocean Engineering2019187: 106190., articleTitle=Analyses of offshore wind turbine structures with soil-structure interaction under earthquakes, refAbstract=null), Reference(id=1276896757678212004, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=99, issue=null, pageStart=103706, pageEnd=null, url=null, language=null, rfNumber=[18], rfOrder=17, authorNames=GAO B, LI C, ZHOU F L, journalName=Marine Structures, refType=null, unstructuredReference=GAO BLI CZHOU F Let al. Shaking table tests of offshore wind turbine systems with a suction bucket foundation in sandy seabed subject to earthquake and wind loads [J]. Marine Structures202599: 103706., articleTitle=Shaking table tests of offshore wind turbine systems with a suction bucket foundation in sandy seabed subject to earthquake and wind loads, refAbstract=null), Reference(id=1276896759351739302, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2025, volume=159, issue=null, pageStart=104625, pageEnd=null, url=null, language=null, rfNumber=[19], rfOrder=18, authorNames=BAI H, ZHANG J H, SUN K, journalName=Applied Ocean Research, refType=null, unstructuredReference=BAI HZHANG J HSUN Ket al. Nonlinear long-term vibration response prediction of offshore wind turbines under full operating conditions based on deep learning [J]. Applied Ocean Research2025159: 104625., articleTitle=Nonlinear long-term vibration response prediction of offshore wind turbines under full operating conditions based on deep learning, refAbstract=null), Reference(id=1276896759771169703, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2018, volume=42, issue=11, pageStart=102, pageEnd=108, url=null, language=null, rfNumber=[20], rfOrder=19, authorNames=赵洪山, 刘辉海, 刘宏杨, journalName=电力系统自动化, refType=null, unstructuredReference=赵洪山,刘辉海,刘宏杨,.基于堆叠自编码网络的风电机组发电机状态监测与故障诊断[J].电力系统自动化201842(11):102-108., articleTitle=基于堆叠自编码网络的风电机组发电机状态监测与故障诊断, refAbstract=null), Reference(id=1276896760194794408, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=22, pageStart=47, pageEnd=50, url=null, language=null, rfNumber=[21], rfOrder=20, authorNames=张舒翔, 刘晓彤, 郭旭峰, journalName=电工技术, refType=null, unstructuredReference=张舒翔,刘晓彤,郭旭峰,.基于SCADA数据驱动的风电机组状态监测[J].电工技术2022(22):47-50., articleTitle=基于SCADA数据驱动的风电机组状态监测, refAbstract=null), Reference(id=1276896761012683689, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=2026, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[22], rfOrder=21, authorNames=王凤, 刘伊彤, 邹, journalName=系统工程理论与实践, refType=null, unstructuredReference=王凤,刘伊彤,邹著,.考虑突发关联状态的风力机组故障预警方法[J/OL].系统工程理论与实践2026[2026-03-09]. https://link.cnki.net/urlid/11.2267.n.20240929.0955.009., articleTitle=考虑突发关联状态的风力机组故障预警方法, refAbstract=null), Reference(id=1276896761457279915, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, doi=null, pmid=null, pmcid=null, year=null, volume=2024, issue=01266, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[22], rfOrder=22, authorNames=null, journalName=*福建省自然科学基金面上项目, refType=null, unstructuredReference=*福建省自然科学基金面上项目2024J01266)。, articleTitle=null, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)]), AuthorCompany(id=1276896719413576518, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=2, ext=[AuthorCompanyExt(id=1276896719421965127, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China), AuthorCompanyExt(id=1276896719468102472, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083)]), AuthorCompany(id=1276896719535211338, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=3, ext=[AuthorCompanyExt(id=1276896719543599947, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3China MCC5 Group Co.,Ltd., Chengdu476300, China), AuthorCompanyExt(id=1276896719551988556, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3中国五冶集团有限公司,成都476300)])], figs=[ArticleFig(id=1276896728783651700, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.1, caption=Model architecture, figureFileSmall=1z8GY+EcAbEoBeKiZO7ClQ==, figureFileBig=E5TzAxcHHML7dg+n39dF1g==, tableContent=null), ArticleFig(id=1276896729211470709, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图1, caption=模型框架, figureFileSmall=1z8GY+EcAbEoBeKiZO7ClQ==, figureFileBig=E5TzAxcHHML7dg+n39dF1g==, tableContent=null), ArticleFig(id=1276896729970639734, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.2, caption=Flowchart of model training strategy, figureFileSmall=aBBfKScPaFtVfNVcA0Zssg==, figureFileBig=xQrpDnu3DsiRvoLKNZ3Bww==, tableContent=null), ArticleFig(id=1276896730037748599, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图2, caption=模型训练策略流程, figureFileSmall=aBBfKScPaFtVfNVcA0Zssg==, figureFileBig=xQrpDnu3DsiRvoLKNZ3Bww==, tableContent=null), ArticleFig(id=1276896732046820216, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.3, caption=Data acquisition equipment and measured curves, figureFileSmall=FtQxROA548wpiFRlXol1lQ==, figureFileBig=QiCU3ka8ZuFPJrS4Yn0QbQ==, tableContent=null), ArticleFig(id=1276896732118123385, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图3, caption=采集设备与所得曲线, figureFileSmall=FtQxROA548wpiFRlXol1lQ==, figureFileBig=QiCU3ka8ZuFPJrS4Yn0QbQ==, tableContent=null), ArticleFig(id=1276896732189426554, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.4, caption=Correlation matrix of 120 feature parameters, figureFileSmall=NOWKNfznCIbsaq0v+jDn8w==, figureFileBig=P4IsVlMzkAEE5/whSFp/xg==, tableContent=null), ArticleFig(id=1276896732478833531, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图4, caption=120项特征参数相关性图, figureFileSmall=NOWKNfznCIbsaq0v+jDn8w==, figureFileBig=P4IsVlMzkAEE5/whSFp/xg==, tableContent=null), ArticleFig(id=1276896732592079740, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.5, caption=Correlation matrix of 20 feature parameters, figureFileSmall=TpOCTPxdOsVlHIOT+/f2ug==, figureFileBig=Ed4OJyhIlvd7+JFm18hSVw==, tableContent=null), ArticleFig(id=1276896732667577213, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图5, caption=20项特征参数相关性, figureFileSmall=TpOCTPxdOsVlHIOT+/f2ug==, figureFileBig=Ed4OJyhIlvd7+JFm18hSVw==, tableContent=null), ArticleFig(id=1276896732906652542, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.6, caption=Comparison of predicted and observed values of the SCADA system, figureFileSmall=992GwhQnySxZwMvMf/yJtw==, figureFileBig=f3HXQV+oxZt2xSs1LAPFaw==, tableContent=null), ArticleFig(id=1276896732998927231, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图6, caption=SCADA系统的预测值和观测值的比较, figureFileSmall=992GwhQnySxZwMvMf/yJtw==, figureFileBig=f3HXQV+oxZt2xSs1LAPFaw==, tableContent=null), ArticleFig(id=1276896733305111424, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.7, caption=Comparison of predicted and observed values of the vibration system, figureFileSmall=o9hQbvmvsZVIy0tf3goK9g==, figureFileBig=ZkX9yIdnqR0Pbmlb4g8Orw==, tableContent=null), ArticleFig(id=1276896733388997505, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图7, caption=振动系统的预测值和观测值比较, figureFileSmall=o9hQbvmvsZVIy0tf3goK9g==, figureFileBig=ZkX9yIdnqR0Pbmlb4g8Orw==, tableContent=null), ArticleFig(id=1276896733745513346, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Fig.8, caption=Comparison of predicted and observed values for non-training periods, figureFileSmall=UcYxnS3fs1VG+l279aBguw==, figureFileBig=4Oja3jKUXHyH2mxtdDWG7A==, tableContent=null), ArticleFig(id=1276896733837788035, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=图8, caption=非训练日期的预测值和观测值比较, figureFileSmall=UcYxnS3fs1VG+l279aBguw==, figureFileBig=4Oja3jKUXHyH2mxtdDWG7A==, tableContent=null), ArticleFig(id=1276896733892313988, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 1, caption=

Network layer settings for missing data imputation

, figureFileSmall=null, figureFileBig=null, tableContent=
网络层参数网络层参数
输入层

N(特征+掩码)

丢弃层0.15
全连接层768层归一化Standard
激活函数0.01初始学习率1.0×10-3
门控激活Sigmoid最小学习率1.0×10-6
GRN模块数8最大迭代轮次1500
输出头全连接层256批次大小1024
Vib输出层50权重衰减1.0×10-4
SCADA输出层41Adam 优化器β1=0.9, β2=0.999
), ArticleFig(id=1276896734169138053, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表1, caption=

缺失数据填充网络层设置

, figureFileSmall=null, figureFileBig=null, tableContent=
网络层参数网络层参数
输入层

N(特征+掩码)

丢弃层0.15
全连接层768层归一化Standard
激活函数0.01初始学习率1.0×10-3
门控激活Sigmoid最小学习率1.0×10-6
GRN模块数8最大迭代轮次1500
输出头全连接层256批次大小1024
Vib输出层50权重衰减1.0×10-4
SCADA输出层41Adam 优化器β1=0.9, β2=0.999
), ArticleFig(id=1276896734605345670, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 2, caption=

Physically derived feature variables

, figureFileSmall=null, figureFileBig=null, tableContent=
变量类别衍生变量名称数量
总计10
气动能量因子风速平方 (v2)1
角度连续化编码风向正余弦、偏航角正余弦、偏航误差正余弦(sinθwd, cosθwd)6
向量交互特征风速-风向耦合分量 (v⋅sinθwd, v⋅cosθwd)2
基准参考量归一化参考风速 (vref)1
), ArticleFig(id=1276896734995415943, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表2, caption=

物理衍生特征变量

, figureFileSmall=null, figureFileBig=null, tableContent=
变量类别衍生变量名称数量
总计10
气动能量因子风速平方 (v2)1
角度连续化编码风向正余弦、偏航角正余弦、偏航误差正余弦(sinθwd, cosθwd)6
向量交互特征风速-风向耦合分量 (v⋅sinθwd, v⋅cosθwd)2
基准参考量归一化参考风速 (vref)1
), ArticleFig(id=1276896736715080584, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 3, caption=

Selected feature parameters

, figureFileSmall=null, figureFileBig=null, tableContent=
参数名称单位参数名称单位参数名称单位
功率曲线风速m/s机舱内部温度有功功率kW
瞬时风速m/s机舱外部温度无功功率kW
30 s平均风速m/s轮毂内部温度电网A相电压V
10 min平均风速m/s发电机定子绕组温度电网B相电压V
转子转速m/s变流器冷却介质温度

电网C相电压

V
发电机转速m/s左内循环入口温度

电网A相电流

kA
主轴转速m/s机舱内部湿度%

电网B相电流

kA
塔筒前后振动位移m桨叶角度A(°)

电网C相电流

kA
塔筒侧向振动位移m桨叶角度B(°)

总有功发电量

kW·h
绝对风向桨叶角度C(°)
机舱绝对位置转子方位角编码器值(°)
), ArticleFig(id=1276896737151288201, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表3, caption=

筛选后的特征参数

, figureFileSmall=null, figureFileBig=null, tableContent=
参数名称单位参数名称单位参数名称单位
功率曲线风速m/s机舱内部温度有功功率kW
瞬时风速m/s机舱外部温度无功功率kW
30 s平均风速m/s轮毂内部温度电网A相电压V
10 min平均风速m/s发电机定子绕组温度电网B相电压V
转子转速m/s变流器冷却介质温度

电网C相电压

V
发电机转速m/s左内循环入口温度

电网A相电流

kA
主轴转速m/s机舱内部湿度%

电网B相电流

kA
塔筒前后振动位移m桨叶角度A(°)

电网C相电流

kA
塔筒侧向振动位移m桨叶角度B(°)

总有功发电量

kW·h
绝对风向桨叶角度C(°)
机舱绝对位置转子方位角编码器值(°)
), ArticleFig(id=1276896737595884426, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 4, caption=

Prediction Performance metrics of the proposed model

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
发电机转速0.9949±0.01260.0468±0.00050.5504±0.2074
电网A相电流0.9921±0.00080.0229±0.00042.2317±0.4207
有功功率0.9917±0.00180.0289±0.00072.8296±0.3301
电网B相电压0.9874±0.00080.0310±0.00020.1164±0.0425
C1均方根值0.9757±0.00540.0114±0.00038.0789±0.2105
D1平均绝对值0.9725±0.01220.0097±0.00018.9481±0.3117
A2主频率0.9237±0.00540.0214±0.000213.701±0.4384
), ArticleFig(id=1276896738002731915, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表4, caption=

模型预测性能指标

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
发电机转速0.9949±0.01260.0468±0.00050.5504±0.2074
电网A相电流0.9921±0.00080.0229±0.00042.2317±0.4207
有功功率0.9917±0.00180.0289±0.00072.8296±0.3301
电网B相电压0.9874±0.00080.0310±0.00020.1164±0.0425
C1均方根值0.9757±0.00540.0114±0.00038.0789±0.2105
D1平均绝对值0.9725±0.01220.0097±0.00018.9481±0.3117
A2主频率0.9237±0.00540.0214±0.000213.701±0.4384
), ArticleFig(id=1276896738430550924, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 5, caption=

Comparative performance metrics of models

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
发电机转速(LI)0.9521±0.02340.0812±0.00121.8420±0.4125
电网A相电流(LI)0.9315±0.01870.0654±0.00215.8741±0.6542
C1均方根值(LI)0.8423±0.03150.0354±0.001515.421±1.0245
D1平均绝对值(LI)0.8214±0.04210.0287±0.001816.874±1.2451
发电机转速(RFR)0.9103±0.00000.1245±0.00003.1254±0.0000
电网A相电流(RFR)0.8842±0.00000.0912±0.00008.9412±0.0000
C1均方根值(RFR)0.7154±0.00000.0512±0.000022.147±0.0000
D1平均绝对值(RFR)0.6841±0.00000.0421±0.000025.412±0.0000
), ArticleFig(id=1276896738497659789, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表5, caption=

模型对比性能指标

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
发电机转速(LI)0.9521±0.02340.0812±0.00121.8420±0.4125
电网A相电流(LI)0.9315±0.01870.0654±0.00215.8741±0.6542
C1均方根值(LI)0.8423±0.03150.0354±0.001515.421±1.0245
D1平均绝对值(LI)0.8214±0.04210.0287±0.001816.874±1.2451
发电机转速(RFR)0.9103±0.00000.1245±0.00003.1254±0.0000
电网A相电流(RFR)0.8842±0.00000.0912±0.00008.9412±0.0000
C1均方根值(RFR)0.7154±0.00000.0512±0.000022.147±0.0000
D1平均绝对值(RFR)0.6841±0.00000.0421±0.000025.412±0.0000
), ArticleFig(id=1276896739244245902, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=EN, label=Table 6, caption=

Generalization performance metrics of the proposed model

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
C1均方根值0.9681±0.00070.0184±0.000311.0604±0.1416
A1平均绝对值0.9347±0.00140.0256±0.000222.6520±0.3095
D1主频率0.9332±0.00170.0238±0.000317.4078±0.1439
电网B相电压0.9548±0.00060.0403±0.00037.2971± 0.0795
功率曲线风速0.8923±0.00170.0671±0.000521.7703±0.3011
瞬时风速0.8703±0.00210.0744±0.000718.5450±0.2591
), ArticleFig(id=1276896741341397903, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, language=CN, label=表6, caption=

模型泛化性能指标

, figureFileSmall=null, figureFileBig=null, tableContent=
参数R²RMSEMAPE/%
C1均方根值0.9681±0.00070.0184±0.000311.0604±0.1416
A1平均绝对值0.9347±0.00140.0256±0.000222.6520±0.3095
D1主频率0.9332±0.00170.0238±0.000317.4078±0.1439
电网B相电压0.9548±0.00060.0403±0.00037.2971± 0.0795
功率曲线风速0.8923±0.00170.0671±0.000521.7703±0.3011
瞬时风速0.8703±0.00210.0744±0.000718.5450±0.2591
)], attaches=null, journal=Journal(id=1276576551509397524, delFlag=0, nameCn=工业建筑, nameEn=Industrial Construction, nameHistory1=null, nameHistory2=null, issn=1000-8993, eissn=null, cn=11-2068/TU, coden=null, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=KHR9x8UxgSX2gqrJN8gfTg==, journalPrice=null, startedYear=null, abbrevIsoEn=Industrial Construction, journalRemark=null, publicationField=null, createdTime=1782289167526, updatedTime=1784021757656, createdBy=18614031015, updatedBy=13041195026, firstLetterCn=G, firstLetterEn=G, subjectCode=Engineering, subjectName=null, subjectCodeEn=Engineering, subjectNameEn=null, picCn=KHR9x8UxgSX2gqrJN8gfTg==, picEn=glrMkaIg87Oz10PZrZOnSg==, jcr=null, cjcr=null, exts=[JournalExt(id=1283843561321894544, language=CN, name=工业建筑, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1784021757679, updatedTime=1784021757679, createdBy=13041195026, updatedBy=13041195026, submissionGuidelinesUrl=, submissionAuthorUrl=https://www.scicloudcenter.com/IC, submissionEditorUrl=https://www.scicloudcenter.com/IC, submissionReviewUrl=https://www.scicloudcenter.com/IC, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1283843561376420497, language=EN, name=Industrial Construction, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1784021757692, updatedTime=1784021757692, createdBy=13041195026, updatedBy=13041195026, submissionGuidelinesUrl=, submissionAuthorUrl=https://www.scicloudcenter.com/IC, submissionEditorUrl=https://www.scicloudcenter.com/IC, submissionReviewUrl=https://www.scicloudcenter.com/IC, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1276577754012160025, websiteList=[Website(id=1276578024992669908, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1276577754012160025, 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/gyjz/CN, language=CN, createTime=1782289518831, createBy=18614031015, updateTime=1782290419260, updateBy=18614031015, name=工业建筑-中文, tplId=1146099689490845704, title=工业建筑, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1276581942745231958, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=articleTextType, value=kx, createTime=1782290452896, updateTime=1782290452896, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942715871827, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=banner, value=null, createTime=1782290452889, updateTime=1782290452889, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942778786393, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=grayFlag, value=0, createTime=1782290452904, updateTime=1782290452904, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942703288914, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=logo, value=https://castjournals.cast.org.cn/joweb/gyjz/CN/file/pic?fileId=ObTw7qIsmguDunMFeBGvBA==, createTime=1782290452886, updateTime=1782290452886, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942795563611, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=minRunFlag, value=0, createTime=1782290452908, updateTime=1782290452908, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942736843349, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/gyjz/CN/file/pic, createTime=1782290452894, updateTime=1782290452894, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942787175002, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=silenceFlag, value=0, createTime=1782290452906, updateTime=1782290452906, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942724260436, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1782290452891, updateTime=1782290452891, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942757814871, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=themeColor, value=null, createTime=1782290452899, updateTime=1782290452899, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581942766203480, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578024992669908, code=themeStyle, value=null, createTime=1782290452901, updateTime=1782290452901, creator=18614031015, updator=18614031015)]), Website(id=1276578025097527513, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1276577754012160025, 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/gyjz/EN, language=EN, createTime=1782289518856, createBy=18614031015, updateTime=1782290415485, updateBy=18614031015, name=工业建筑-英文, tplId=1146101810881728533, title=Industrial Construction, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1276581912684655179, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=articleTextType, value=kx, createTime=1782290445729, updateTime=1782290445729, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912659489352, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=banner, value=null, createTime=1782290445723, updateTime=1782290445723, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912722403918, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=grayFlag, value=0, createTime=1782290445738, updateTime=1782290445738, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912634323527, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=logo, value=https://castjournals.cast.org.cn/joweb/gyjz/EN/file/pic?fileId=ObTw7qIsmguDunMFeBGvBA==, createTime=1782290445717, updateTime=1782290445717, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912739181136, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=minRunFlag, value=0, createTime=1782290445742, updateTime=1782290445742, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912676266570, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/gyjz/EN/file/pic, createTime=1782290445727, updateTime=1782290445727, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912730792527, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=silenceFlag, value=0, createTime=1782290445740, updateTime=1782290445740, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912667877961, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1782290445725, updateTime=1782290445725, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912693043788, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=themeColor, value=null, createTime=1782290445731, updateTime=1782290445731, creator=18614031015, updator=18614031015), WebsiteProps(id=1276581912709821005, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1276578025097527513, code=themeStyle, value=null, createTime=1782290445735, updateTime=1782290445735, creator=18614031015, updator=18614031015)])], journalTitle=工业建筑, weixinUrl=null, journalUrl=http://gyjz.ic-mag.com/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Industrial Construction, journalPhotoCn=KHR9x8UxgSX2gqrJN8gfTg==, journalPhotoEn=glrMkaIg87Oz10PZrZOnSg==, journalFirstLetter=G, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false, interPubPlatform=, interPubPlatformUrl=null), detailUrlCn=https://castjournals.cast.org.cn/joweb/gyjz/CN/10.3724/j.gyjzG26031308, detailUrlEn=https://castjournals.cast.org.cn/joweb/gyjz/EN/10.3724/j.gyjzG26031308, pdfUrlCn=https://castjournals.cast.org.cn/joweb/gyjz/CN/PDF/10.3724/j.gyjzG26031308, pdfUrlEn=https://castjournals.cast.org.cn/joweb/gyjz/EN/PDF/10.3724/j.gyjzG26031308, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, previewStatus=0, delFlag=0, hasFullText=1, orderTime=1779206400000, fullTextJson=null, articleText=null, reference=null)
收藏切换
基于深度学习的海上风机多源缺失数据的填补方法
收藏切换
PDF下载
李栋 1 , 廖毅桢 1 , 桑源 1 , 李泽宇 2 , 姚博 3 , 陈洪兵 2
工业建筑 | 工程运维技术 2026,56(5): 176-186
收起
收藏切换
工业建筑 |工程运维技术 2026 , 56 (5) : 176 -186
基于深度学习的海上风机多源缺失数据的填补方法
全屏
[Author(id=1276896719665234767, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=dongli@fzu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1276896719921087313, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896719665234767, language=EN, stringName=Dong LI, firstName=Dong, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896720009167698, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896719665234767, language=CN, stringName=李栋, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio={"content":"

李栋,副研究员,主要从事大跨度空间薄膜结构、组合结构、海上风电、结构风工程等方面研究,

"}, bioImg=null, bioContent=

李栋,副研究员,主要从事大跨度空间薄膜结构、组合结构、海上风电、结构风工程等方面研究,

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896720302768980, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896720399237974, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720302768980, language=EN, stringName=Yizhen LIAO, firstName=Yizhen, middleName=null, lastName=LIAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896720730587991, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720302768980, language=CN, stringName=廖毅桢, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896720789308249, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896720852222811, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720789308249, language=EN, stringName=Yuan SANG, firstName=Yuan, middleName=null, lastName=SANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896721170989916, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896720789308249, language=CN, stringName=桑源, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1福州大学土木工程学院,福州350108, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719275164483, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=1, ext=[AuthorCompanyExt(id=1276896719287747396, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1College of Civil Engineering, Fuzhou University, Fuzhou350108, China), AuthorCompanyExt(id=1276896719338079045, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719275164483, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1福州大学土木工程学院,福州350108)])]), Author(id=1276896721259070302, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896721590420320, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896721259070302, language=EN, stringName=Zeyu LI, firstName=Zeyu, middleName=null, lastName=LI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896723087786849, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896721259070302, language=CN, stringName=李泽宇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719413576518, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=2, ext=[AuthorCompanyExt(id=1276896719421965127, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China), AuthorCompanyExt(id=1276896719468102472, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083)])]), Author(id=1276896723146507107, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, 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=1276896723242976101, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723146507107, language=EN, stringName=Bo YAO, firstName=Bo, middleName=null, lastName=YAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3China MCC5 Group Co.,Ltd., Chengdu476300, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896723326862182, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723146507107, language=CN, stringName=姚博, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3中国五冶集团有限公司,成都476300, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719535211338, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=3, ext=[AuthorCompanyExt(id=1276896719543599947, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3China MCC5 Group Co.,Ltd., Chengdu476300, China), AuthorCompanyExt(id=1276896719551988556, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719535211338, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3中国五冶集团有限公司,成都476300)])]), Author(id=1276896723398165352, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=hongbingchen@ustb.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1276896723754681194, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723398165352, language=EN, stringName=Hongbing CHEN, firstName=Hongbing, middleName=null, lastName=CHEN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1276896724098614123, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, authorId=1276896723398165352, language=CN, stringName=陈洪兵, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1276896719413576518, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, xref=2, ext=[AuthorCompanyExt(id=1276896719421965127, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China), AuthorCompanyExt(id=1276896719468102472, tenantId=1146029695717560320, journalId=1276577754012160025, articleId=1276896715483513644, companyId=1276896719413576518, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083)])])]
李栋1 , 廖毅桢1, 桑源1, 李泽宇2, 姚博3, 陈洪兵2
作者信息
  • 1福州大学土木工程学院,福州350108
  • 2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083
  • 3中国五冶集团有限公司,成都476300
通讯作者:
陈洪兵,教授,主要从事工程结构损伤检测、服役安全智能诊断,钢-混凝土组合结构缺陷无损检测技术和多尺度多物理场仿真等领域研究,
Deep Learning-Based Imputation of Multi-Source Missing Data for Offshore Wind Turbines
Dong LI1 , Yizhen LIAO1, Yuan SANG1, Zeyu LI2, Bo YAO3, Hongbing CHEN2
Affiliations
  • 1College of Civil Engineering, Fuzhou University, Fuzhou350108, China
  • 2Future City College / Institute of Urbanization and Urban Safety, University of Science and Technology Beijing, Beijing100083, China
  • 3China MCC5 Group Co.,Ltd., Chengdu476300, China
出版时间: 2026-05-20 doi: 10.3724/j.gyjzG26031308
文章导航
收藏切换

针对海上风电机组在恶劣工况下因传感器失效或通信中断引发的多源监测数据缺失问题,构建了一种基于多头门控残差网络的填补模型。该方法通过特征级联将数据采集与监视控制系统(SCADA)数据和结构振动监测数据进行协同融合,并引入门控残差网络提取深度非线性耦合特征;同时设计出多头并行输出结构,实现对两类异构数据的独立重建。训练阶段采用动态掩码机制与混合损失函数,以增强模型对复杂气动工况的适应能力。基于某10 MW海上风电机组实测数据的验证结果显示,该模型在训练工况下具有较高的决定系数,能够实现多源缺失数据的精准重构。在面向非训练时段的泛化测试中,模型预测的决定系数虽略有波动,但仍能有效捕捉信号整体变化趋势;尤其对振动类数据,其泛化性能衰减幅度小于SCADA数据,展现出更强的稳定性。该方法可显著提升风机多源监测数据的完整性与可靠性,具备良好的工程应用前景。

海上风电机组  /  数据填补  /  门控残差网络  /  多源数据融合

To address the problem of missing multi-source monitoring data of offshore wind turbines caused by sensor failures or communication interruptions under harsh operating conditions, this paper proposes a novel imputation model based on a multi-head gated residual network. This method achieves collaborative fusion of supervisory control and data acquisition (SCADA) data and structural vibration monitoring data through feature concatenation, employs a gated residual network to extract deep nonlinear coupling features, and uses a multi-head parallel output architecture for the independent reconstruction of these two heterogeneous data types. During the training stage, a dynamic masking mechanism combined with a hybrid loss function is adopted to enhance the model’s adaptability to complex aerodynamic operating conditions. Validated with field data from a 10 MW offshore wind turbine, the proposed model achieved a high coefficient of determination under training conditions, enabling accurate reconstruction of missing multi-source data. In generalization tests for non-training periods, although the coefficient of determination of the model’s predictions fluctuated slightly, the model still effectively captured the overall trends of monitoring signals. Notably, the degradation in generalization performance for vibration data was less pronounced than that for SCADA data, demonstrating its greater stability. The proposed method can significantly improve the completeness and reliability of multi-source monitoring data for wind turbines and holds considerable potential for engineering applications.

offshore wind turbines  /  data imputation  /  gated residual network  /  multi-source data fusion
李栋, 廖毅桢, 桑源, 李泽宇, 姚博, 陈洪兵. 基于深度学习的海上风机多源缺失数据的填补方法. 工业建筑, 2026 , 56 (5) : 176 -186 . DOI: 10.3724/j.gyjzG26031308
Dong LI, Yizhen LIAO, Yuan SANG, Zeyu LI, Bo YAO, Hongbing CHEN. Deep Learning-Based Imputation of Multi-Source Missing Data for Offshore Wind Turbines[J]. Industrial Construction, 2026 , 56 (5) : 176 -186 . DOI: 10.3724/j.gyjzG26031308
风能已成为全球能源结构转型的关键支撑,其清洁、可再生特性驱动风电装机规模持续扩张1-2。然而,海上风机长期处于高湿、盐雾及波浪耦合荷载等严酷环境,主轴承、齿轮箱与发电机等关键传动部件故障风险显著升高3-5。统计表明,发电机与齿轮箱故障合计占风电机组非计划停运原因的比例超过2/36。加之海上风电场多位于远海区域,受限于可达性差、作业窗口期短等因素,常规巡检与人工诊断手段难以及时响应,导致运维成本急剧攀升7-8。鉴于此,充分挖掘既有监控与数据采集(SCADA)系统监测数据,构建高效的状态评估与早期预警机制,已成为提升海上风电机组运行可靠性与全寿命经济性的核心路径69
目前,风力发电机故障预警研究多基于SCADA系统数据,或依赖振动与应变测量数据10。在状态监测领域,基于图结构的学习与预测方法展现出重要潜力。Zhao等11结合图卷积网络与多头注意力的条件扩散模型,能够捕捉风速数据的时空相关性,并通过概率化填补区间提升数据可靠性。Qu等12将数据排列重构为旅行商问题,优化了生成对抗网络框架,显著提高了大规模风机数据集的填补精度。Liao等13采用深度卷积网络的上下文编码器,有效建模非线性特征与多属性关联,在波动性风电数据中表现优异。Fan等14设计多尺度注意力掩码自编码器,通过局部化Transformer层提取时序与机组间依赖关系,在降低计算成本的同时实现高精度填补。Radev等15将双向长短期记忆网络(LSTM)的风参数预测与分数阶神经网络相结合,在多类运行场景下提升了功率预测准确性。Wang等16通过结合生成对抗填补网络与XGBoost算法,有效解决了SCADA数据缺失问题,并构建了高精度的失控预警模型;其创新性引入迁移学习技术,使模型能快速适配新机型,显著降低了计算与部署成本。
此外,振动响应是评估海上风机结构动力特性、保障其安全稳定运行的关键参数,准确预测振动响应对风电机组长期服役性能至关重要17-18。目前,海上风机振动响应预测主要依赖数值模拟与有限元分析。尽管此类方法能较精确地模拟结构在复杂荷载下的动力行为,但其计算资源消耗大、耗时长的特点,限制了在风机全寿命周期分析与实时监测中的应用。因此,基于机器学习与深度学习的模型逐步应用于海上风电机组振动响应预测。Bai等19采用多输入递归BiLSTM网络,开发了一种混合深度学习和模拟框架,实现比传统模拟快745倍的计算速度,同时保持91.94%的预测准确率,为耗时数值方法提供了高效替代方案。
在SCADA数据驱动的状态监测与故障诊断方面,相关研究亦取得重要进展。赵洪山等20基于堆叠自编码网络构建风电机组发电机状态监测模型,通过重构误差变化趋势实现异常预警与故障诊断。张舒翔等21改进Inception v1结构,提出1D_Inception v1模型,有效提取SCADA数据空间多尺度特征,实现高精度状态监测。王凤等22融合变分模态分解、卷积神经网络、多尺度图注意网络及灰狼优化算法,构建面向大型风机集群的故障关联状态预测与识别方法,显著提升多状态故障预警的自动化水平与识别精度。上述方法为利用SCADA数据开展精细化状态监测提供了可行技术路径。
然而,当前研究仍存在一定局限性。当海上风机传感器损坏或数据传输中断时,由于远海维修困难,难以及时修正缺失数据,可能对风机安全运行构成威胁。尽管已有多种数据填补方法被提出,如基于生成对抗网络、扩散模型、Transformer结构等,但现有方法尚未实现SCADA系统数据与振动监测数据的深度融合与相互补充。由于振动传感器成本远高于SCADA传感器,多数风机振动传感器布设数量有限,若能利用已有SCADA数据对缺失振动数据进行有效填补,不仅可实现利用SCADA数据等效替代振动数据进行动力响应分析,还能显著降低运营成本,提升风机系统整体可靠性与智能化水平。因此,发展面向多源数据融合的缺失数据填补与状态预测方法,是未来研究的重要方向。
针对风电机组监测数据中因传感器失效或通信链路中断导致的缺失问题,提出一种基于多头门控残差网络(Multi-Head Gated Residual Network)的通用填补框架。该框架对SCADA及振动监测数据进行统一特征构造与归一化处理,依托6组GRN模块串联构建深层特征提取主干,并设置多头输出分支并行完成振动与SCADA特征的联合重构。训练环节引入自适应掩码、噪声扰动、混合损失函数及L2约束等机制,以增强模型在非理想工况下的稳健性与泛化迁移能力。进一步采用K折交叉验证对集成模型进行训练,结合N-1诊断策略与多场景鲁棒性测试,系统评估模型在多种缺失模式下的数据还原效能。本方法面向风力发电机组状态监测系统中常见的采集缺失场景,可为设备健康评估与故障早期预警提供高质量、高连续性的数据基础。
为应对海上风电机组多源传感器数据缺失问题,本研究构建了一个基于深度学习的通用数据填补框架。该框架旨在通过捕获SCADA系统与振动监测系统数据间的复杂非线性关系,实现对缺失值的高精度重建。本章系统阐述模型的理论基础,包括核心模块门控残差网络(Gated Residual Network, GRN)的结构原理、多头输出机制的设计思路、动态训练策略与损失函数优化方法,以及模型性能的评估指标体系。
构建的缺失数据填充模型基于多头门控残差网络架构,包含公共特征提取主体及针对振动与SCADA数据的独立预测分支,其整体结构如图1所示。网络主体部分由8个门控残差块(GRN Block)串联而成,每个模块内部通过双路映射实现特征筛选与残差学习。其主路径依次包含全连接层、LeakyReLU激活函数(负斜率为0.01)及丢失层(Dropout),而门控路径利用Sigmoid函数调节信息流向,最终经残差加和与层归一化处理输出。主体网络隐藏层维度统一设置为768,双分支预测头各包含一个256维的全连接子层。输入层接收由原始特征与掩码向量拼接而成的2×N维向量。为了提升模型的泛化能力与收敛稳定性,设置丢失率为0.15,最大迭代轮次为1500,并采用初始学习率为0.001的余弦退火衰减策略,最小学习率设为1.0×10⁻⁶。模型训练的批次大小为1024,权重衰减系数为1.0×10⁻⁴,Adam优化器参数β₁=0.9、β₂=0.999。网络层的具体参数设置见表1
门控残差网络是所提出模型的基本构成单元,旨在从高维、不完整的数据中有效提取特征,同时缓解深层网络训练中常见的梯度消失问题。GRN模块巧妙地结合了残差学习与门控机制,允许信息选择性地通过网络层,从而增强模型对复杂模式的表达能力。其内部结构如图1(b)所示。
对于输入特征向量 x,GRN模块包含两条并行的计算路径:主路径负责对输入进行非线性变换以提取深层特征;门控路径则负责生成一个动态的调节信号,用于控制主路径信息的流出量。主路径的计算过程可表示为:
y=DropoutW2δ(W1x+b1)+b2
式中:y 为主路径的输出向量;δ 为LeakyReLU激活函数,用于引入非线性;W₁ 和 W₂ 分别为第一个和第二个全连接层的权重矩阵;b₁ 和 b₂ 为对应的偏置向量;Dropout层用于在训练过程中随机抑制一部分神经元,以防止模型过拟合。
门控路径通过一个独立的全连接层,并应用Sigmoid激活函数生成一个门控向量 g
g=σ(W3x+b3)
式中:g 为门控向量,其每个元素值介于0~1,起到信息流量“阀门”的作用;σ 为Sigmoid函数;W₃ 与 b₃ 分别为该全连接层的权重矩阵和偏置向量。
最终,主路径的输出 y与门控向量 g进行逐元素相乘,实现信息的选择性通过。相乘后的结果与原始输入x 通过残差连接相加,以促进梯度流动和特征复用,最后再经过层归一化(Layer Normalization)处理,得到GRN模块的最终输出 z
z=LayerNorm(x+gy)
式中:⊙ 为Hadamard积,即逐元素相乘;LayerNorm为层归一化操作,用于稳定训练过程。该结构使模型能够根据输入x自适应地调节信息传递强度,从而更灵活、更有效地学习特征间的非线性依赖关系。
考虑到振动监测数据与SCADA数据在物理含义、量纲及分布特性上存在显著差异,若采用单一的输出层同时预测两类特征,可能导致模型难以兼顾各自的精度需求。为此,本文在GRN主干网络之后设计了多头并行输出结构,分别针对振动特征与SCADA特征设置独立的预测头,如图1(d)所示。
主干网络的输出(即经过多层GRN模块提取的高阶特征)被同时馈入两个预测分支:振动预测头(Vibration Head)与SCADA预测头(SCADA Head)。每个预测头均由一个全连接层、一个LeakyReLU激活层和一个最终的输出层构成。这种解耦设计允许每个分支独立地学习专属于该类数据的映射函数,从而提升预测的专业性与准确性。模型最终输出与输入维度相对应的预测结果,其中振动头输出重构的振动特征,SCADA头输出重构的SCADA特征。
模型采用自监督学习范式进行训练,其核心思想是通过重构人为构造的缺失数据,使模型掌握特征间的内在关联。训练过程中的关键环节包括动态掩码生成、数据增强及多目标损失函数的设计,流程如图2所示。
在每个训练批次中,首先基于原始数据批次 T 为每个样本随机生成一个损失掩码 M。该掩码用于标记当前批次中需要被模型预测的特征位置,即模拟的缺失特征。随后,对数据进行序列化增强以生成增强后的输入 X,首先,根据损失掩码 M 将待预测的特征值置零,以模拟数据缺失的真实场景;其次,在已知特征中随机丢弃部分信息,迫使模型不过度依赖单一特征,提升其泛化能力;最后,向数据中添加微小的随机高斯噪声,以模拟传感器测量误差和信号扰动,从而增强模型对输入扰动的鲁棒性。增强后的数据 X 被输入多头GRN模型,从而得到预测值。
损失函数的设计是训练的核心,它兼顾了预测精度与模型泛化能力。总损失由振动头损失与SCADA头损失加权求和构成:
Ltotal=λvibLvib+λscadaLscada
式中:LvibLscada分别为振动预测头和SCADA预测头的损失;λvibλscada为用于平衡两类任务重要性的权重系数。每个预测头的损失函数Lhead进一步细化为多个目标的组合:
Lhead=iMaskedwiLδyi,y^i+αyi-y^i+βy^1
式中:Masked 表示由损失掩码 M 指定的需要计算损失的特征索引集合;wi为对应第i个特征的权重,用于调节不同特征对损失的贡献;Lδ为Huber损失;α为控制平均绝对误差(MAE)项权重的超参数;βL1正则化项的权重,此项仅应用于振动预测头,用于促进预测值的稀疏性,鼓励模型捕捉更本质的振动模式。
Huber损失是一种结合了均方差(MSE)和平均绝对误差(MAE)优点的鲁棒损失函数,其定义为:
Lσ(y,y^)=12y-y^2            y-y^δδy-y^-12δ其他        
式中:y为真实值;ŷ为预测值;δ为一个超参数阈值。Huber损失综合了均方误差对小误差敏感与平均绝对误差对异常值鲁棒的优点,当预测误差小于δ时采用MSE,误差较大时则转为MAE,从而在保证梯度稳定性的同时,提升模型对噪声数据的鲁棒性。
模型采用Adam优化器进行参数更新。为加速收敛并避免陷入局部最优,引入了带重启的余弦退火(cosine annealing with restarts)学习率调度策略。训练初期设置学习率预热(warm-up)阶段,有助于稳定训练过程。同时应用梯度裁剪(gradient clipping)技术,将梯度的L2范数限制在预设阈值内,以防止梯度爆炸。
为全面评估模型预测性能,选用决定系数(R2)、均方根误差(RMSE)和平均绝对百分比误差(MAPE)作为核心评价指标,同时引入标准差(σ)辅助分析数据离散程度。
决定系数衡量模型对真实值变异性的解释能力,其值越接近1,表示拟合效果越好:
R2=1-i=1nyi-y^i2i=1nyi-y¯2
式中:n为样本总数;yi为第i个样本的真实值;y^i为第i个样本的预测值;y¯为所有样本真实值的均值。
均方根误差反映预测误差的总体水平,它对较大误差敏感,计算式为:
RMSE=1ni=1nyi-y^i2
平均绝对百分比误差以百分比形式呈现相对误差,便于直观理解,计算公式为:
MAPE=100%ni=1nyi-y^iyi1ni=1nyi-y^i2
RMSE与MAPE的值越接近于0,表明预测精度越高。需注意,MAPE在真实值接近零时可能失真,因此适用于目标值始终为正且远离零的场景。标准差σ用于描述数据波动范围,可辅助判断预测误差相对于数据自身变异的程度。
为构建高精度数据填补模型,需对原始SCADA系统与振动监测系统的多源数据进行系统化处理,包括数据对齐、特征提取、物理衍生特征构建及特征筛选降维。本章详细阐述数据来源、融合方法、衍生特征设计及特征选择过程,为后续模型训练奠定基础。
本研究基于中国某海上风电场项目,项目所处海域风能资源丰富,且大规模采用了便于施工的吸力筒导管架基础和大容量风机。本文采用的数据取自该项目的一台10 MW海上风力发电机。利用其监控与数据采集系统(SCADA)和独立的振动监测系统,采集并存储了该机组在 2023 年运行过程中的多维度数据,涵盖运行工况、环境参数以及机组机械与电气设备的功能状态。2023 年下半年的运行数据,涵盖齿轮箱、发电机、变压器等关键部件的温度、压力、功率、风速、状态等参数,共 874 项原始特征,采样频率为 10 Hz。振动监测系统同步采集了24 项特征,包括塔筒顶部倾角、振动、位移及加速度等参数。传感器沿塔筒高度布置于5个截面(A ~ E),数据采集时段为2023 年3月。采集风机转速以及风轮状态的设备如图 3(a)、(b)所示,收集所得的功率曲线风速过程线与主轴转速过程线如图3(c)、(d)所示,这些曲线直观地展示了数据的连续性与可靠性。
整体数据的采集时长为2023年3—12月。后续模型训练选取其中一天时间靠近的63840个样本进行训练,模型采用两阶段随机比例进行训练集、验证集以及测试集的划分。第一阶段为测试集划分,采用函数方法在全量数据中随机抽取15%的样本作为独立测试集,剩余部分作为训练集与验证集。第二阶段则使用五折交叉验证对剩余数据进行划分,每一折训练中,模型再次将剩余数据随机分为80%的训练集和20%的验证集。上述随机划分并非按照时间顺序进行,同时严格保证同一样本不会同时出现在测试集和训练集中,不存在样本重叠。
由于SCADA与振动系统原始采样频率不一致,需进行数据对齐与特征融合。首先,对所有振动通道信号采用零相位四阶巴特沃斯带通滤波器进行滤波,以分离主要动态响应并抑制噪声与基线漂移。滤波后信号按1 s间隔(对应SCADA系统10个采样点)分割为不重叠片段。对每个片段提取时域与频域特征。时域特征包括均方根值(RMS)、峰度(Kurtosis)和平均绝对值(MAV),计算公式如下:
Kurtosis=1Ni=1N(xi-x¯)41Ni=1N(xi-x¯)22
MAV=1Ni=1Nxi
式中:N为采样点数;xi为滤波后的离散信号点;x¯为该段信号的均值。
频域特征通过对每个片段进行快速傅里叶变换(FFT)获得频谱序列 Xk),并提取主幅度(MainAmp)及其对应主频率(MainFreq):
X(k)=n=0N-1x(n)e-j2πNkn    k0,N-1
MainAmp=maxkXk
MainFreq=fsNargmaxkXk
式中:n为时间序列索引;k 为频率索引;j为虚数单位;fs 为振动信号的采样频率。
该流程每秒为每个振动通道生成5个特征(RMS,Kurtosis,MAV,MainAmp,MainFreq),振动传感器布置于塔筒的A~E五个截面,每个截面安装2个传感器,共计10个通道,由此构成50维振动特征集。
原始SCADA数据中的角度变量(如风向、偏航角)存在0°~360°的周期性数值突变,且这些变量与风机的气动载荷呈复杂的非线性关系。为增强模型对复杂气动工况的辨识能力,使其能够捕捉推力与转矩对风速的非线性响应,构造风速与风向正、余弦的耦合交互项,形成2个向量特征,以增强模型对风能矢量空间分布的表征能力。最后,保留归一化参考风速作为基准量。这10项衍生特征构成了一个富含物理先验信息的特征空间,有助于模型捕捉因气动载荷不对称而引发的异常振动模式。
原始SCADA数据含874项特征,其中包含大量冗余或弱相关变量,直接输入模型可能导致过拟合与计算负担。因此需进行系统化特征筛选。首先剔除变化不明显的特征(如恒定值或近恒定值)及传感器存在但无有效数据输出的通道,初步保留120项特征。这些特征覆盖机组关键运行状态。
为量化剩余特征间的线性相关程度,本研究采用皮尔逊相关系数进行分析。皮尔逊相关系数 ρ 衡量了两个变量 XY 之间线性相关的强度和方向,其定义为:
ρ X,Y=i=1nXi-X¯Yi-Y¯i=1nXi-X¯2i=1nYi-Y¯2
式中:XiYi分别是变量 XY 的第i个观测值;X¯Y¯ 分别是 XY 的样本均值。系数的取值范围为[-1, 1],绝对值越接近1,表示线性相关性越强。
基于初步筛选的120项特征,绘制了相关性热图,如图4所示。为便于观察,从中选取了高中低相关系数共20项特征进行细节展示,如图5所示。
在进行特征选择时,首先以发电机转速,即评估失控风险的关键指标为目标,筛选出与其显著相关的特征。然后,对于任意两个特征之间皮尔逊相关系数绝对值超过0.95的冗余对,仅保留其中一个与目标变量更相关或更具物理意义的特征。通过上述步骤,最终从120项中精选出31项关键SCADA特征,如表2所示。这些特征全面涵盖了风速、转速、温度、电压、电流、振动、桨距角、风向等关键维度。
将筛选出的31项SCADA特征、经处理得到的50维振动特征以及10项物理衍生特征进行合并,构成一个包含91维特征的融合特征向量(表3)。该向量集成了机组运行参数、结构振动响应以及物理机理信息,为模型提供了丰富而全面的输入。
由于各特征的量纲和数值范围差异巨大,为消除其对模型训练的不利影响并加速梯度下降收敛,采用最小-最大归一化方法将所有特征值线性映射到[0, 1]区间:
x'=x-min(x)max(x)-min(x)
式中:x为原始特征值;x'为归一化后的特征值;min(x)和max(x)分别为该特征在数据集中的最小值和最大值。经过上述处理,最终获得标准化后的91维特征数据集,用于后续的模型训练与验证。
采用经预处理后的SCADA与振动监测数据对多头门控残差网络模型进行训练。为评估模型预测能力,分别从两个系统中选取关键特征参数进行对比分析。SCADA系统选取发电机转速、电网A相电流、有功功率及电网B相电压;振动系统选取均方根值、平均绝对值及主幅值。将各参数的模型预测值与实际观测值进行对比,结果如图6图7所示。
图6展示了SCADA系统四项关键参数的预测曲线与实际观测曲线的对比。可见:发电机转速、电网A相电流、有功功率的预测曲线与实际曲线高度吻合,模型表现出优异的拟合能力;电网B相电压虽波动性较大,但预测曲线仍能准确捕捉其整体变化趋势。图7展示了振动系统三项参数的预测结果。相较于SCADA数据,振动信号包含大量高频尖峰,波动更为剧烈。模型预测在尖峰处略显平滑,但对信号的整体趋势及关键波峰波谷的捕捉依然精准,表明模型对非线性、非平稳信号具有较强的泛化能力。
为进一步量化模型性能,采用决定系数R²、均方根误差RMSE和平均绝对百分比误差MAPE作为评价指标,结果如表4所示,表中为进行10次独立随机抽取形成的平均值以及标准差值。SCADA系统4项参数的R²均大于0.98,RMSE低于0.05,MAPE小于3%,表明模型对SCADA数据具有极高的预测精度。振动系统三项参数的R²均大于0.92,RMSE低于0.03,MAPE除主幅值略高于10%外,其余均小于10%,同样验证了模型的有效性。
为深入验证多头门控残差网络在多源数据重构任务中的性能优势,本研究选取线性插值法(LI)与随机森林回归算法(RFR)作为对比基线。其中,线性插值法基于时间邻域连续性假设,系工程实践中处理传感器数据缺失的常用基础方法;随机森林则作为经典机器学习模型,具备对多变量非线性映射关系的拟合能力。实验部分严格遵循前文所述的数据划分策略,以发电机转速、电网A相电流、C1均方根值及D1平均绝对值4项关键参数为对象进行填补效果评估,并统计10次独立重复实验所得评价指标的均值与标准差,以保证对比结果的稳健性与统计意义,结果如表5所示。
表45所示量化评估结果可知,所提填补模型在4项关键参数的重构任务中均取得最优性能。在发电机转速的填补中,该模型R2达0.9949,相较于线性插值法(R2=0.9103)提升约9.29%;均方根误差由0.1245降至0.0468,表明该架构对气动工况所致动态波动具有更强的捕捉能力。针对随机性更为显著的振动数据,随机森林回归算法在D1平均绝对值项上的R2仅为0.6841,反映出传统插值策略在处理高频结构响应时的局限。相较之下,所提方法在同一参量上取得R2=0.9725的重构精度,较随机森林回归提升约18.39%,验证了门控残差块对非线性耦合特征的有效提取。依托多头并行输出机制实现的协同重构框架,所提模型不仅在SCADA数据填补中展现出较高的复原精度,亦在复杂振动特征的还原中表现出较强的鲁棒性,可为海上风机多源监测数据的完整性提供技术支撑。
为了评估模型在实际应用中面对未知数据的泛化能力,选取了非训练日期的SCADA与振动监测系统非重叠时段数据进行预测分析。结合图8的预测曲线与表6(表中为进行10次独立随机抽取形成的平均值以及标准差值)的性能指标可以看出,模型在非训练数据集上依然能够较好地跟踪原始信号的整体变化趋势,具备优秀的泛化性能。从总体评价指标对比来看,相较于3.1节中模型在训练数据集上的表现,模型对非训练数据的预测结果呈现出决定系数R2整体略微下降、平均绝对百分比误差MAPE有所上升的趋势。这符合深度学习模型在面对分布存在差异的未知工况数据时的常规表现。
进一步以在表4表6中均被选作评估参数的电网B相电压和C1均方根值为例进行量化分析。对于SCADA系统中的电网B相电压,其R2由训练时段的0.9874下降至非训练时段的0.9548,MAPE则由0.1164%上升至7.2971%;对于振动系统中的C1均方根值,其R2由0.9757微降至0.9681,MAPE由8.0789%上升至11.0604%。数据对比及图像走势表明,尽管面对非训练日期的离散数据,模型在局部突变或极值处的拟合精度略有降低,但各项核心评价指标仍保持在较高水平,说明模型针对非训练日期的数据依然表现出良好的重构与预测功能,但在复杂非平稳工况的细节特征捕捉方面仍具有一定的提升空间。
此外,综合对比非训练数据中两类系统的预测指标可以观察到,模型对振动类型数据的泛化性能整体优于SCADA系统数据。在非训练时段下,代表振动系统特征的参数性能指标下降幅度相对较小。另外,从图8的对比曲线中可见其预测值与真实值的重合度更高。这表明本文所提出的多头门控残差网络框架,在融合多源特征并进行独立分支预测后,能够更加深入地学习到风电机组结构振动响应的深层物理规律。因此,模型在面对非训练日期的振动系统数据时,功能衰减更小,预测结果相对更为准确稳定。
针对海上风电机组多源传感器数据在复杂工况下易发生缺失的问题,本文提出了一种基于多头门控残差网络的通用数据填补框架。通过对真实10 MW海上风机SCADA数据与振动监测数据的特征融合与模型验证,主要得出以下结论:
1)构建的多头门控残差网络模型能够有效提取SCADA系统与振动监测系统数据间的高维非线性物理关联。引入的动态掩码与混合损失函数策略,在无需完整标签数据的前提下,实现了对异构监测数据的高精度自监督联合重构。
2)在训练集数据表现上,模型对SCADA关键特征和振动特征均展现出优异的拟合能力。SCADA参数决定系数总体大于0.98,振动参数决定系数大于0.92,能够精准捕捉低频运行状态与高频结构动态响应的整体规律与关键波峰波谷。
3)对比实验显示,所提模型在发电机转速与振动参量重构中,R2较线性插值、随机森林分别提升9.29%与18.39%,验证了门控残差结构对非线性耦合特征提取的有效性,为后续泛化性评估提供参照。
4)在非训练日期的泛化能力验证中,模型对SCADA数据的预测整体决定系数略有下降,平均绝对百分比误差有所上升,局部极值拟合存在提升空间,但依然能够准确追踪信号变化趋势,保持了良好的预测功能。
5)模型对非训练日期的振动系统数据表现出比SCADA数据更小的功能衰减,表明该框架在挖掘深层结构动力学特征方面具有更强的鲁棒性,可为海上风电机组的全寿命周期状态评估提供高连续性的数据基础。

参考文献 引证文献
排序方式:
[1]
WANG F XZHANG SHOU Y Zet al. Extraction of offshore wind turbines in china by combining multispectral and SAR image data [J]. Journal of Selected Topics in Applied Earth Observations and Remote Sensing202417:3392695.
[2]
蔡彦枫,李晓宇,任宗栋,.面向空中风力发电系统的风能资源分析[J].太阳能学报202546(2):624-632.
[3]
张珺辉,董绍华,叶迎春.基于LabVIEW的海上风机齿轮箱故障诊断系统设计[J].新技术新工艺2025(1):29-37.
[4]
LEI Y GYANG BJIANG X Wet al. Applications of machine learning to machine fault diagnosis: a review and roadmap [J]. Mechanical Systems and Signal Processing2020138: 106587.
[5]
TAO S YQIAN ZPEI Yet al. Wind turbine failure detection based on SCADA data and data mining method [C]// Renewable Power Generation Conference. 2019.
[6]
ABORGEL TSHEHATA A SKOTB M Aet al. Heavy lift semi-submersible ships utilization in offshore wind turbines industry [J]. Energy Reports20228: 834-847.
[7]
ZHANG W WVATN JRASHEED A. Gearbox pump failure prognostics in offshore wind turbine by an integrated data-driven model [J]. Applied Energy2025380: 124829.
[8]
HUSSAIN MMIRJAT N HSHAIKH Fet al. Condition monitoring and fault diagnosis of wind turbine: a systematic literature review [J]. Journal of Selected Topics in Applied Earth Observations and Remote Sensing202435:14747.
[9]
GOROSTIDI NPARDO DNAVA V. Diagnosis of the health status of mooring systems for floating offshore wind turbines [J].Renewable Energy and Power Quality Journal2023287: 115862.
[10]
ENCALADA-DAVILA ÁTUTIVÉN CMOYÓN Let al. Main bearing fault prognosis in wind turbines based on gated recurrent unit neural networks [J].Renewable Energy and Power Quality Journal202220: 2172-2183.
[11]
ZHAO J HSHEN X DLIU T Jet al. Probabilistic imputation of missing offshore wind speed based on conditional diffusion models [J]. Sustainable Energy, Grids and Networks202544: 101949.
[12]
QU F MLIU J HMA Y Jet al. A novel wind turbine data imputation method with multiple optimizations based on GANs [J]. Mechanical Systems and Signal Processing2020139: 106610.
[13]
LIAO W LBAK-JENSEN BPILLAI J Ret al. Data-driven missing data imputation for wind farms using context encoder [J].Journal of Modern Power Systems and Clean Energy2022(2):000894.
[14]
FAN Y WFENG C LWU Ret al. Multiscale-attention masked autoencoder for missing data imputation of wind turbines [J]. Knowledge-Based Systems2024299: 112114.
[15]
RADEV BKASI V RBINGI K. Enhancement of Texas wind turbine power predictions using fractional order neural network by incorporating machine learning models to impute missing data [J]. Knowledge-Based Systems2024300: 112176.
[16]
WANG BZHOU BZOU M Het al. Research on condition monitoring of offshore wind turbine gearbox based on K-means clustering and extreme learning machine [C]// International Conference on Artificial Intelligence and Power Engineering. 2023.
[17]
JU S HHUANG Y C. Analyses of offshore wind turbine structures with soil-structure interaction under earthquakes [J]. Ocean Engineering2019187: 106190.
[18]
GAO BLI CZHOU F Let al. Shaking table tests of offshore wind turbine systems with a suction bucket foundation in sandy seabed subject to earthquake and wind loads [J]. Marine Structures202599: 103706.
[19]
BAI HZHANG J HSUN Ket al. Nonlinear long-term vibration response prediction of offshore wind turbines under full operating conditions based on deep learning [J]. Applied Ocean Research2025159: 104625.
[20]
赵洪山,刘辉海,刘宏杨,.基于堆叠自编码网络的风电机组发电机状态监测与故障诊断[J].电力系统自动化201842(11):102-108.
[21]
张舒翔,刘晓彤,郭旭峰,.基于SCADA数据驱动的风电机组状态监测[J].电工技术2022(22):47-50.
[22]
王凤,刘伊彤,邹著,.考虑突发关联状态的风力机组故障预警方法[J/OL].系统工程理论与实践2026[2026-03-09]. https://link.cnki.net/urlid/11.2267.n.20240929.0955.009.
*福建省自然科学基金面上项目2024J01266)。
2026年第56卷第5期
PDF下载
72
29
引用本文
BibTeX
文章信息
doi: 10.3724/j.gyjzG26031308
  • 接收时间:2026-03-13
  • 首发时间:2026-06-25
  • 出版时间:2026-05-20
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2026-03-13
基金
作者信息
    1福州大学土木工程学院,福州350108
    2北京科技大学未来城市学院/城镇化与城市安全研究院,北京100083
    3中国五冶集团有限公司,成都476300

通讯作者:

陈洪兵,教授,主要从事工程结构损伤检测、服役安全智能诊断,钢-混凝土组合结构缺陷无损检测技术和多尺度多物理场仿真等领域研究,
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/gyjz/CN/10.3724/j.gyjzG26031308
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