基于振动的结构健康监测技术
|
摘要
介绍了基于振动的结构健康监测与诊断的基本概念与理论,论述了基于振动和神经网络方法的结构损伤检测的原理、策略及其在土木工程中的应用,重点探讨了各种结构损伤检测技术存在的问题及有效的解决方法.考虑到测量噪声及特征参数对实际结构检测的影响,提出了自适应概率神经网络(APNN)和主组分分析(PCAPNN)技术,并用三种概率神经网络(传统概率神经网络TPNN,APNN和PCAPNN)模型对某悬索桥进行了结构损伤检测,结果发现APNN的识别精度最高,PCAPNN次之,TPNN最差.
On the basis of the necessity and urgency of the structural health monitoring and diagnosis,the basic concepts and the principles of the structural health monitoring based on vibration was briefly introduced.The principles,strategies and application for the civil engineering based on vibration and neural networks were described.Some existing problems and solving methods for various techniques were mainly discussed.In view of measurement noise and feature vectors effect on the damage detection in practice,the adaptive probabilistic neural network (PNN) and the principal component analysis PNN are proposed.A suspension bridge is detected by use of three different PNN models.This study shows that the identification accuracy (IA) for damage localization by using the adaptive PNN is the maximum,the IA by using the traditional PNN is the minimum and the IA by using the PCAPNN is between the former two models.
On the basis of the necessity and urgency of the structural health monitoring and diagnosis,the basic concepts and the principles of the structural health monitoring based on vibration was briefly introduced.The principles,strategies and application for the civil engineering based on vibration and neural networks were described.Some existing problems and solving methods for various techniques were mainly discussed.In view of measurement noise and feature vectors effect on the damage detection in practice,the adaptive probabilistic neural network (PNN) and the principal component analysis PNN are proposed.A suspension bridge is detected by use of three different PNN models.This study shows that the identification accuracy (IA) for damage localization by using the adaptive PNN is the maximum,the IA by using the traditional PNN is the minimum and the IA by using the PCAPNN is between the former two models.
引文
[1] 杨大智.智能材料与系统[M].天津:天津大学出版社,2000.
[2] HousnerGW.Structuralcontrol:past,present,andfuture[J].JournalofEngineeringMechanics,1997,123(9):897-971.
[3] DoeblingSW,FarrarCR,etal.Damageidentificationandhealthmonitoringofstructuralandmechanicalsystemsfromchangesintheirvibrationcharacteristics:areview[R].LosAlamosNationalLaboratoryReportLA-13070-MS,1996.
[4] 李宏男,李东升.土木工程结构安全性评估、健康监测及诊断述评[J].地震工程与工程振动,2002,22(3):82-90.
[5] DoeblingSW,FarrarCR.Thestateoftheartinstructuralidentificationofconstructedfacilities[R].LosAlamosNationalLaboratoryReportDRAFT,1999.
[6] SmithSWandBeattieCA.Modelcorrelationanddamagelocationforlargespacestructures:Scantmethoddevelopmentandevaluation[R].ReportNo.NASA-CR-188102,NASA,1991.
[7] SmithSW.Iterativeuseofdirectmatrixupdates:Connectivelyandconvergence[A].Proc.Ofthe33thAIAAStructures,StructuralDynamics,andMaterialsConference[C].1992,1797-1806.
[8] KimHMandBartkowiczTJ.Damagedetectionandhealthmonitoringoflargespacestructures[J].SoundandVibration,1993,27(6):12-17.
[9] LiuPL.Identificationanddamagedetectionoftrussesusingmodaldata[J].JournalofStructuralEngineering,1995,121(4):599-608.
[10] NorrisMA,MeirovitchL.Onthemodelingforparameteridentificationindistributedstructures[J].InternationalJournalforNumericalMethodsinEngineering,1989(28):2451-2463.
[11] LimTW.Structuraldamagedetectionusingconstrainedeigenstructureassignment[J].JournalofGuidance,Control,andDynamics,1995,18(3):411-418.
[12] KimHM,BartkowiczTJ.Atwo stepstructuraldamagedetectionapproachwithlimitedinstrumentation[J].JournalofVibrationandAcoustics-TransactionsoftheASME,1997,119(2):258-264.
[13] BrownDL,AllemangRJ,ZimmermanRJ,etal.Parameterestimationtechniquesformodalanalysis[R].ReportNo.SAETechnicalpaperseries790221.1979.
[14] VenkatasubramanianVandChanK.Aneuralnetworkmethodologyforprocessfaultdiagnosis[J].JournalofAIChE,1989,35(12):1993-2002.
[15] 杨晓楠,姜绍飞,倪一清.基于特征向量简化的复杂土木结构损伤检测[J].沈阳建筑工程学院学报(自然科学版),2003,19(2):81-84.
[2] HousnerGW.Structuralcontrol:past,present,andfuture[J].JournalofEngineeringMechanics,1997,123(9):897-971.
[3] DoeblingSW,FarrarCR,etal.Damageidentificationandhealthmonitoringofstructuralandmechanicalsystemsfromchangesintheirvibrationcharacteristics:areview[R].LosAlamosNationalLaboratoryReportLA-13070-MS,1996.
[4] 李宏男,李东升.土木工程结构安全性评估、健康监测及诊断述评[J].地震工程与工程振动,2002,22(3):82-90.
[5] DoeblingSW,FarrarCR.Thestateoftheartinstructuralidentificationofconstructedfacilities[R].LosAlamosNationalLaboratoryReportDRAFT,1999.
[6] SmithSWandBeattieCA.Modelcorrelationanddamagelocationforlargespacestructures:Scantmethoddevelopmentandevaluation[R].ReportNo.NASA-CR-188102,NASA,1991.
[7] SmithSW.Iterativeuseofdirectmatrixupdates:Connectivelyandconvergence[A].Proc.Ofthe33thAIAAStructures,StructuralDynamics,andMaterialsConference[C].1992,1797-1806.
[8] KimHMandBartkowiczTJ.Damagedetectionandhealthmonitoringoflargespacestructures[J].SoundandVibration,1993,27(6):12-17.
[9] LiuPL.Identificationanddamagedetectionoftrussesusingmodaldata[J].JournalofStructuralEngineering,1995,121(4):599-608.
[10] NorrisMA,MeirovitchL.Onthemodelingforparameteridentificationindistributedstructures[J].InternationalJournalforNumericalMethodsinEngineering,1989(28):2451-2463.
[11] LimTW.Structuraldamagedetectionusingconstrainedeigenstructureassignment[J].JournalofGuidance,Control,andDynamics,1995,18(3):411-418.
[12] KimHM,BartkowiczTJ.Atwo stepstructuraldamagedetectionapproachwithlimitedinstrumentation[J].JournalofVibrationandAcoustics-TransactionsoftheASME,1997,119(2):258-264.
[13] BrownDL,AllemangRJ,ZimmermanRJ,etal.Parameterestimationtechniquesformodalanalysis[R].ReportNo.SAETechnicalpaperseries790221.1979.
[14] VenkatasubramanianVandChanK.Aneuralnetworkmethodologyforprocessfaultdiagnosis[J].JournalofAIChE,1989,35(12):1993-2002.
[15] 杨晓楠,姜绍飞,倪一清.基于特征向量简化的复杂土木结构损伤检测[J].沈阳建筑工程学院学报(自然科学版),2003,19(2):81-84.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心