欢迎来到天天文库
浏览记录
ID:58307027
大小:717.35 KB
页数:7页
时间:2020-05-20
《非线性随机振动分析的概率密度演化方法-论文.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、第49卷第2期西南交通大学学报Vol.49No.22014年4月JOURNALOFSOUTHWESTJIAOTONGUNIVERSITYApr.2014文章编号:02582724(2014)02022007DOI:10.3969/j.issn.02582724.2014.02.006非线性随机振动分析的概率密度演化方法彭勇波,李杰(同济大学土木工程防灾国家重点实验室,上海200092)摘要:为深入探讨概率密度演化方法对于非线性随机振动分析的适用性,考察了随机地震动作用下一类硬弹簧Duffing振子的非线性响应,对概率密度演化方法与经典非线性随机振动分析进行了比较研究.结果表
2、明:在弱非线性水平,概率密度演化方法与混沌多项式展开、MonteCarlo模拟的解答一致;在强非线性水平,数值求解误差、人为截断误差放大,概率密度演化方法与混沌多项式展开解答在MonteCarlo模拟解附近上下波动,表明概率密度演化方法与经典非线性随机振动解答在均方特征意义上是等价的.关键词:概率密度演化方法;混沌多项式展开;MonteCarlo模拟;随机地震动;KarhunenLoève分解中图分类号:O324文献标志码:AProbabilityDensityEvolutionMethodofNonlinearRandomVibrationAnalysisPENGYongbo,
3、LIJie(StateKeyLaboratoryofDisasterReductioninCivilEngineering,TongjiUniversity,Shanghai200092,China)Abstract:Inordertorevealtheapplicabilityoftheprobabilitydensityevolutionmethodinnonlinearrandomvibrationanalysis,acomparativeresearchoftheprobabilitydensityevolutionmethodandtheclassicalnonlinea
4、rrandomvibrationanalysiswascarriedoutbyinvestigatingthenonlinearresponsesofaclassofrandomlybasedrivenDuffingoscillatorsusingtheprobabilitydensityevolutionmethod(PDEM),theadaptivepolynomialchaosexpansion(APCE)andtheMonteCarlosimulation(MCS).Aphysicallybasedstochasticgroundmotionmodelwasemploye
5、d,andrepresentedbyaKarhunenLoèveexpansionintheapplicationoftheAPCE.ThisdiscreterepresentationcanbeviewedasaprojectionofthephysicalvectorspaceintotheGaussianvectorspace.Numericalresultsrevealthatthesolutionprocessesofthethreeapproachesareidenticaltoweaklynonlinearsystems,whiletheyareapproximat
6、elyidenticaltostronglynonlinearsystemsthougherrorsresultedfromnumericaltechniquesandartificialtruncationsareamplified,indicatingthatthesolutionofthePDEMisequivalenttothatoftheclassicalnonlinearrandomvibrationanalysisinthemeansquaresense.ThePDEM,moreover,goesastepfurtherthantheclassicalnonline
7、arrandomvibrationanalysissincetheprobabilitydensityfunctionofresponsesandthedynamicreliabilityofsystemscanbesimultaneouslyprovidedbythePDEM.Theothermethods,however,needmuchmorecomputationaleffortstoobtainhighorderstatisticsofresponses.K
此文档下载收益归作者所有