公路混凝土橋梁損傷評(píng)估技術(shù)研究.pdf

1、鄭州大學(xué)碩士學(xué)位論文公路混凝土橋梁損傷評(píng)估技術(shù)研究姓名：劉建民申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：水工結(jié)構(gòu)工程指導(dǎo)教師：李清富韓菊紅2003.5.28囂州丈掌1學(xué)硬士學(xué)位論文AbstractTheassessmentandpredictionofdamageinbridgesisanecessarypartnbridgeconservation、maintenanceandtechnicalreconstructionoftodaywhichiSa

2、lsooneofthemostheatedproblemsthatarebeingdiscussedbyscientistsinengineeringareaandscholarsathomeandabroad，andalongwiththencrementofoldbridges，theproblemattractsmoreaRentionCombiningwiththecharacterofbridges，thethesisgive

3、sasystematicandthoroughstudyonsometechnicalproblemsofassessmentofdamageinconcretebridges，whichincludes：(1)Integrating砸氆theresearchresultsinengineering，asystematicanalysisforthecauseofdamagedconcretebridgesinourcountrywas

4、giveninthethesiswhichwillconsequentlyleadtosomepertinentproposalstowardsplan、design、constructionandadministrationofbridgesaswellaslayafoundationfortheestablishmentanddecisionmakingoftheassessmentsystemofdamagedbridgesf2A

5、ccordingtoitscharacterbridgesystemcanbedividedintoseveralmultilevel—systemsandthusahierarchyassessmentmodeliSestablishedOnthebasisoftheseresearchworksacomprehensivegradings37stemforassessmentfactorsisdiscussedanddevelope

6、d，beingpreparedfortheneuralnetworkassessmentofdamagedconcretebridgesf3)Basedonthehierarchyassessmentmodel，aneuralnetworkmethodtoassessthedamageinbridgesiSputforwardThescoresofassessmentfactorsbeingmadetheinputvalueandthe

7、scoreofdamageassessmenttheoutputValue，amultilevelassessmentmodeforbridgescameintobeingwhichinturncouldcome協(xié)thefinalresultofthebridgeasawholeandgetthedamageratingaccordingtothe委yenstandard下hemethod。makinguseofneuralnetwor

8、k輪learnfromexitingassessmentexamplescouldundergoassessmentinsteadofexpertgroupwiththerelationshipbetweenassessmentresultandassessmentfactorsbestexpressedinthenetworkAsaresultitcouldnotonlyreduceassessmentworkloadandsaver

9、esources，butalSOaccumulateexperts’experienceandgetridofhumanfactorsinthecourseofassessmenttomaketheres稍tsoundobjectiveSimultaneouslytheestablishedsystemiScharaclerizedbylearningofitselgwhichcancompleteassessmentthroughsi

10、mulatingexperiencedassessmentmechanismwhenthereiSnoexpertorhasnotgotenoughexpels、Toconcludefromthegivenexamples，eachmodel，afterspendinglittletimeontrainingthemselvesfromsampledata，couldassessthedamagedegreeforexistingbri

11、dgesusingtrainedweightedvalueSandthreshotds(4)Consideringsuchactualdi伍cultiesaslackingofengineeringdataanddiscrepancyinmanaginglevel，fourmethodsfordamageassessmentareproposedinthethesisbasedonexpertImowledge，whichareweig

12、htedaveragemethod、fuzzycomprehensiveevaluationmethod、generalizedquantificationmethodandsetpairanalysismethodtomakeusefulsupplementstotheneuralnetworkmethodDisadvantageouseffects，causedbydissimilarityofsubjectivefactorssu