资源描述:
《基于静动力凝聚及扩展卡尔曼滤波的连续梁桥损伤识别》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、第32卷第7期Vol.32No.7工程力学2015年7月July2015ENGINEERINGMECHANICS156文章编号:1000-4750(2015)07-0156-08基于静动力凝聚及扩展卡尔曼滤波的连续梁桥损伤识别何浩祥,吕永伟,韩恩圳(北京工业大学工程抗震与结构诊治北京市重点实验室,北京100124)摘要:以卡尔曼滤波算法为代表的物理参数识别方法在结构损伤识别方面得到广泛应用,但限于状态方程的复杂性,大部分应用集中在具有平动自由度的剪切型建筑结构模型,且一般需要较完备的激励和响应信息。利用静力
2、凝聚方法消去转动自由度以建立力学模型,并提出考虑Rayleigh阻尼的动力凝聚方法,实现了较复杂连续梁桥有限元模型的等效简化。针对桥梁检测及健康监测需求,提出了利用锤击产生自由振动的激励方式进行连续梁桥扩展卡尔曼滤波在线损伤识别方法,从而克服了传统方法需要复杂激励信号的不足。以一座三跨连续梁桥为例进行了仿真分析,识别出了不同位置的刚度和阻尼物理参数,参数识别结果具有较高精度和抗噪性,收敛速度快,证明该方法有效可行。关键词:扩展卡尔曼滤波;损伤识别;静力凝聚方法;连续梁桥;健康监测中图分类号:U448.21;
3、TU317文献标志码:Adoi:10.6052/j.issn.1000-4750.2014.01.0005DAMAGEDETECTIONFORCONTINUOUSGIRDERBRIDGEBASEDONSTATIC-DYNAMICCONDENSATIONANDEKFHEHao-xiang,LÜYong-wei,HANEn-zhen(BeijingLaboratoryofEarthquakeEngineeringandStructuralRetrofit,BeijingUniversityofTechnolog
4、y,Beijing100124,China)Abstract:Asaclassicalmethodofphysicalparameteridentification,ExtendedKalmanfiltering(EKF)algorithmiswidelyusedinstructuraldamagedetection.However,duetothecomplexityofthestateequations,mostapplicationsfocusontheshearstructuremodelwitht
5、ranslationaldegreesoffreedomconsideredonly,andthecomprehensiveexcitationandresponsesignalsaregenerallyneeded.Recognizingthattherotationaldegreesoffreedomcanbeeliminatedbythestaticcondensationmethod,thispaperproposesthedynamiccondensationmethodconsideringRa
6、yleighdamping,inordertoestablishtheequivalentandsimplifiedmodelforcontinuousgirderbridges.Accordingtotherequirementsofbridgeinspectionandhealthmonitoring,theonlineandconvenientdamagedetectionmethodbasedonExtendedKalmanfilteringispresented,andthefreevibrati
7、onisgeneratedbyonehammeraction,sothatthedeficiencyofthetraditionalmethodneedingcomplexexcitationinformationisovercame.Takingathree-spancontinuousgirderbridgeasanexample,thecorrespondingstiffness,thedamagelocationanddegreeandthedampingparameterwereidentifie
8、daccuratelybytheproposedmethod,showingthattheproposedmethod,withfastconvergencespeed,isfeasibleforthecaseofdynamicsignalwithhighnoise-signalratioandpracticalengineering.Keywords:extendedKalmanfiltering;damage