摘要
Asanefficienttoolinhandlinguncertainissues,Dempster-Shaferevidencetheoryhasbeenincreasinglyusedinstructuralhealthmonitoringanddamagedetection.Inapplications,however,Dempster-Shaferevidencetheorysometimesleadstocounter-intuitiveresults.Inthisstudy,anewfusionalgorithmofevidencetheoryisputforwardtoaddressvariouscounter-intuitiveproblemsandmanagethereliabilitydifferenceoftheevidence.Theproposedalgorithmcomprisesthefollowingaspects:(1)Dempster’scombinationruleisgeneralizedbyintroducingtheconceptofevidenceullage.ThenewruleallowsclassicalDempster’sruleandcanresolvecounter-intuitiveproblemscausebyevidenceconflictandevidencecompatibility;(2)areliabilityassessingmethodbasedonaprioriandposteriorknowledgeisproposed.Comparedwithconventionalreliabilityassessment,theproposedmethodcanreflecttheactualevidencereliabilitiesandcanefficientlyreducedecisionrisk.Numericalexamplesconfirmthevalidityandutilityoftheproposedalgorithm.Inaddition,anexperimentalinvestigationonaspatialtrussstructureiscarriedouttoillustratetheidentifiedabilityoftheproposedapproach.Theresultsindicatethatthefusionalgorithmhasnostrictrequestontheaccuracyandconsistencyofevidencesourcesandcanefficientlyenhancediagnosticaccuracy.
出版日期
2019年02月12日(中国期刊网平台首次上网日期,不代表论文的发表时间)