This paperdetailsthefatiguelifepredictionmodelforweldingcomponentsbasedonhybridintelligent
technology.Wemakeuseofthecapabilitiesandadvantagesofroughsettheory,particleswarmopti-
mization (PSO)algorithmandBPneuralnetworkforestablishingofthefatiguelifepredictionmodel.
Firstly,roughsettheorywasusedtodealwiththeoriginalfatiguesampledata;theminimumfatigue
feature subsetwasobtained.Secondly,improvedPSOalgorithmwasusedtooptimizetheinitialweighs
and thresholdsoftheBPneuralnetwork,whichresolvessuchproblemsaslocalextremumandslow
convergencethatexistinthetraditionalBPneuralnetwork.Atlast,minimumreducedsubsetwasin-
putted intotheoptimizedBPneuralnetworktoconstructthenovelfatiguelifepredictionmodelfor
welding componentsbythecontinuoustrainingandadjusting.Sampledataofthetitaniumalloywelded
joints wasusedtoverifythecorrectnessandvalidityofthenovelfatiguelifepredictionmodel,simu-
lation resultsshowthatthefatiguelifepredictionmodelproposedinthispaperhasbetterfaulttoler-
ance, higherprecision,andcan fitting fatiguelifevaluemoreaccuratelythantraditionalBPmodel.
Consequently,themodelbasedonhybridintelligenttechnologycanprovideaneffectivenewapproach
to predictthefatiguelifeofweldedjoints.
Paperdetailsthefatiguelifepredictionmodelforweldingcomponentsbasedonhybridintelligent นี้เทคโนโลยี Wemakeuseofthecapabilitiesandadvantagesofroughsettheory, particleswarmopti-algorithmandBPneuralnetworkforestablishingofthefatiguelifepredictionmodel mization (PSO)ประการแรก roughsettheorywasusedtodealwiththeoriginalfatiguesampledata, theminimumfatigueคุณสมบัติ subsetwasobtained ประการที่สอง improvedPSOalgorithmwasusedtooptimizetheinitialweighsและ thresholdsoftheBPneuralnetwork, whichresolvessuchproblemsaslocalextremumandslowconvergencethatexistinthetraditionalBPneuralnetwork.Atlast,minimumreducedsubsetwasin-putted intotheoptimizedBPneuralnetworktoconstructthenovelfatiguelifepredictionmodelforเชื่อม componentsbythecontinuoustrainingandadjusting Sampledataofthetitaniumalloyweldedข้อต่อ wasusedtoverifythecorrectnessandvalidityofthenovelfatiguelifepredictionmodel, simu-resultsshowthatthefatiguelifepredictionmodelproposedinthispaperhasbetterfaulttoler เครื่องดูด-ance, higherprecision, andcan เหมาะสม fatiguelifevaluemoreaccuratelythantraditionalBPmodelดังนั้น themodelbasedonhybridintelligenttechnologycanprovideaneffectivenewapproachการ predictthefatiguelifeofweldedjoints
การแปล กรุณารอสักครู่..
