文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 工程科技 > 电子/电路 > A neural network application for reliability modelling and condition-based predictive maintenance
侵权投诉

A neural network application for reliability modelling and condition-based predictive maintenance

IntJAdvManufTechnol(2005)25:174–179DOI10.1007/s00170-003-1835-3

Chang-ChingLin·Hsien-YuTseng

Aneuralnetworkapplicationforreliabilitymodellingandcondition-basedpredictivemaintenance

Received:17March2003/Accepted:16June2003/Publishedonline:2June2004 Springer-VerlagLondonLimited2004

Abstract

A neural network application for reliability modelling and condition-based predictive maintenance

decisionsontheuse

A neural network application for reliability modelling and condition-based predictive maintenance

ofmachineryarebasedon

A neural network application for reliability modelling and condition-based predictive maintenance

historicaldataand

A neural network application for reliability modelling and condition-based predictive maintenance

commonsense.However,carryingout

A neural network application for reliability modelling and condition-based predictive maintenance

aneffectivepredictivemaintenanceplan,

A neural network application for reliability modelling and condition-based predictive maintenance

informationaboutcurrentmachineconditions

A neural network application for reliability modelling and condition-based predictive maintenance

mustbemadeknownto

A neural network application for reliability modelling and condition-based predictive maintenance

thedecision-maker.Inthispaper,anewmethodofobtainingmaintenanceinformationhasbeenproposed.Byinte-gratingtraditionalreliabilitymodellingtechniquestime,onlineperformanceestimationmodel,informationsuchashazardrateandmeantimebetweenfailurescanbecalculated.Essentially,thispaperpresentsaninnovativemethodtosynthesiseinformation(suchasvibrationsignals)withhigh(likereliabilitystatistics)toformarigorousforbettermachinemaintenance.KeywordsCerebellarmodelarticulationcontroller·Neuralnetwork·Predictivemaintenance·Weibullproportionalhazardsmodel

systemtodeterminethehealthstatusofmachinery,namely,cere-bellarmodelcontrollerneuralnetwork-basedma-chinemodel(CMAC-PEM).InordertoverifyourtheWeibullproportionalhazardsmodel(WPHM)hasalsobeenimplemented.Additionally,abearingdeteriorationexperimentwasconducted.ItsdatawereusedtotestboththeCMAC-PEMandtheWPHMmethodologies.Theircomparisonresultsandanalysesaregiveninthispaper.

2Literaturereview

Therearethreemaintypesofmaintenance,namelyimprove-mentmaintenance,preventivemaintenance,andcorrectivemain-tenance.TheeffortsofaretoreduceoreliminatetheneedforBycontrast,cor-rectivemaintenanceistherepairexecutedafterfailureoccurrenceandpreventivemaintenancedenotesallactionsin-tendedtokeepequipmentingoodoperatingconditionandtoavoidfailures[2].

Condition-basedmaintenance(CBM)isdynamicpreventivemaintenanceinpractice.Markovandsemi-MarkovmodelshavebeenthepreferredinsimulationandevaluationofCBM[3–5].OtherlikeMonteCarlomodelling[6,7]oranarti cialintelligenceapproach[8],havealsobeenpro-posedbyseveralresearchers.

Asamodellingtheprocessofage-dependentfail-ures,thehazardmodellingiswellknowninmain-tenanceproportionalhazardsmodel(PHM),de-velopedbyCox[9],extendsthetraditionalhazardmodeltoincludecorrelativeinformationsuppliedbydiagnosticvariables.Ithasbeenwidelyusedinmedicaldiagnostics,andwas rstap-pliedtoengineeringreliabilityproblemsbyJardineandAnder-son[10].Kobbacyetal.[11,12]proposedaheuristicapproachforimplementingaproportionalhazardsmodeltoschedulethenextmaintenanceintervalonthebasisoftheequipment’sfullconditionhistory.

Albus rstproposedCMAC[13,14].Itiscapableofveryfastlearning,andcontainscertainfeaturesofinterpolationandap-

1Introduction

Itiswellknownthat99%failuresareprecededbysomeindicators[1].condition-basedpredictivemaintenanceisprobablytheeconomicalwaytomaintainmachinery.Itsideaistoallowthedeterminationofmachineryhealthinareal-time,onlinefashion.Assuch,faultscanbepre-dictedbeforetheytakeplace.Maintenancecanthenbescheduledasneeded.Reportedbene tsofpredictivemaintenanceincludereduceddowntime,lowermaintenancecosts,andreductionofunexpectedcatastrophicfailures.

Theobjectiveofthispaperistocombinetraditionalrelia-bilitymodellingmethodswithvibration-basedmonitoringtech-niquesandarti cialneuralnetworktechnologiesinanintegrated

C.-C.Lin·H.-Y.Tseng(u)

DepartmentofIndustrialEngineeringandManagement,St.John’s&St.Mary’sInstituteofTechnology,

499,Sec.4TamkingRoad,Tamsui,Taipei251,Taiwan,R.O.C.E-mail:hytseng@mail.sjsmit.edu.twTel.:+886-228013131ext.6564Fax:+886-228013143

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email