李瑞,范玉刚,张光辉.KLPP特征约简与RELM的高压隔膜泵单向阀故障诊断[J].机械科学与技术,2023,42(8):1332-1339KLPP特征约简与RELM的高压隔膜泵单向阀故障诊断李瑞1,2,范玉刚1,2,张光辉1(1.昆明理工大学信息工程与自动化学院,昆明650500;2.昆明理工大学云南省人工智能重点实验室,昆明650500)摘要:为此提出基于核局部保持投影(KLPP)和正则化极限学习机(RELM)的高压隔膜泵单向阀故障诊断方法。首先,提取单向阀振动信号的时域、频域、时频域特征,构建多域特征集;然后,通过KLPP算法对构建的多域特征集进行维数约简;最后,建立基于RELM的故障诊断模型,用于识别单向阀运行状态。实验结果表明,基于多域特征的故障诊断方法检测精度高于单域特征识别方法;KLPP约简多域特征集,可以有效消除信息冗余;建立的RELM故障诊断模型识别精度达到98.89%,能够有效识别高压隔膜泵单向阀故障类型。关键词:单向阀;故障诊断;核局部保持投影;正则化极限学习机中图分类号:TN710.1;TH165.3文献标志码:ADOI:10.13433/j.cnki.1003-8728.20220076文章编号:1003-8728(2023)08-1332-08CheckValveFaultDiagnosisofHigh-pressureDiaphragmPumpwithKLPPFeatureReductionandRELMLIRui1,2,FANYugang1,2,ZHANGGuanghui1(1.FacultyofInformationEngineeringandAutomation,KunmingUniversityofScienceandTechnology,Kunming650500,China;2.YunnanKeyLaboratoryofArtificialIntelligence,KunmingUniversityofScienceandTechnology,Kunming650500,China)Abstract:Thesingle-domainfeaturecannotfullyreflecttheoperatingstateofcheckvalveofthehigh-pressurediaphragmpump,andthehigh-dimensionalfeaturesetcomposedofmulti-domainfeatureswillproducedimensionaldisasters,andtheinformationredundancyleadstolowrecognitionaccuracyofthefaultdiagnosismodel.Tothisend,afaultdiagnosismethodforcheckvalveofhigh-pressurediaphragmpumpbasedonKLPP(Kernellocalpreservationprojection)andRELM(Regularizedextremelearningmachine)isproposedinthispaper.First,thetimedomain,frequencydomainandtime-frequencydomainfeaturesofcheckvalvevibrationsignalarerespectivelyextractedtoconstructamulti-domainfeatureset.Then,dimensionalityreductionisperformedontheconstructedmulti-domainfeaturesetthroughtheKLPPalgorithm.Finally,afaultdiagnosi...