欢迎来到天天文库
浏览记录
ID:16071073
大小:410.65 KB
页数:6页
时间:2018-08-07
《大坝安全诊断的溷沌优化神经网络模型》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、万方数据第27卷第8期2006年8月岩土力学RockandSoilMechanicsVbl.27NO.8Aug.2006文章■号l1000--7598--(2006)08—1344--05大坝安全诊断的混沌优化神经网络模型曹茂森1,一,邱秀梅2,夏宁1(1.河海大学土木工程学院,南京210098.2.山东农业大学水利土木工程学院,泰安271018)摘要t为了提高大坝变形的预测精度,采用小波变换和分形理论对大坝位移观测数据的非线性动力学特性进行了分析,揭示了其具有低维混沌动力特性,这为大坝变形预测模型的建立提供了理论依据和先验知识。基于低维混沌动力特性,设计了能
2、捕获大坝位移观测数据全局动力特性,兼具神经网络模型结构优化和动力机制时新的混沌优化神经网络大坝变形预测模型。在工程实例中,由多个度量指标组成量化评价体系,对模型预测性能进行综合评价,结果表明,所建模型比传统BP神经网络和ARMA模型具有更高的预测精度。关键词。大坝位移;低维混沌:动力特性;小波变换;混沌优化神经网络中图分类号tTv698;O233文献标识码:AAchaos-optimizedneuralnetworkmodelfordamsafetymonitoringCAOMao—senl一,QIUXiu—mei2,XIANin91(1.CollegeofC
3、ivilEngineering,HohaiUniversity,Nanjing210098,China;2.CollegeofWaterConservancyandCivilEngineering,ShandongAgriculturalUniversity,Tai’an271018,China)Abstract:Damdeformationpredictionisimportantfordamsafetymonitoringandhasbecomeafocusofincreasinginterestinrecentyears.Inthisstudy,onth
4、ebasisofnonlineardynamicpropertyanalysisoftheobservationsofdamdisplacements,anovelmethodologyisproposedtoestablishdamdeformationpredictionmodelwithimprovedpredictionprecision.Firstly,thedynamicpropertiesofobservationsofdamdisplacementsarestudiedbycombinedwavelettransformwithfractal,
5、andtheresultsrevealthatdamdisplacementspossesscertainlOWdimensionalchaoticcharacter.Thisprovidestheoreticalfoundationandtranscendentalknowledgeforrelationalestablishmentofdamdeformationpredictionmodel.Moreover,derivedfromthelOWdimensionalchaoticcharacter,achaos—optimizedneuralnetwor
6、kmodelfordamdeformationpredictionisconstructed,whichisnotonlycapableofcapturingthedynamicpropertiesofobservationsofdamdisplacementsbutalsoofimplementingthemodel’Sstructuraloptimizationanddynamicmechanismrefreshing.Finally,inthepracticalapplicationofdamdeformationprediction,themodelp
7、erformanceisquantificationallyassessedbymultipleindices.Theresultdemonstratesthatchaos—optimizedneuralnetworkmodelholdshigherpredictionprecisionthantheconventionalbackpropagation(BP)neuralnetworkandARMAmodels;andtherefore,itispromisingfordamsafetymonitoring.Keywords:damdisplacements
8、;lowdimensionalchao
此文档下载收益归作者所有