Reviewercommentsas#ReviewerCommentsforJPEM-D-13-Inthispaper,theauthordevelopedanovelmethodbasedonamorphologicalfilterandsignalcomplexitymeasurefromaneweddy-currentsensor.Theresultsshowthattheproposedmethodiseffectivefordetectingqualityproblemswithrollerbearings,showingthehighsensitivityinresolveweaksignals.Itmaybeextendedtotheproblemsofsignalprocessinginaccelerationbasedmethod.However,moreeffortsshouldbegiventomakecommentsonthemethodusedtosupportthemethodsinuse.InadditionpleaseimprovetheEnglishtoreducetyandgrrerrors.AlsopleaseaddresstheconcernsComment1:Changethe"Differentfromsignalsintheprocess…"into"UnlikesignalsTheauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment2:PleaseconsiderthefollowingsuggestionforBearingdefectiveinspectionysavitalroleinbearingqualitycontrol.Unlikesignalsintheprocessofconditionmonitoringandfaultdiagnosis,thesignalcharacteristicofdefectivebearingsismuchweakeranddifficulttobefiedthroughtheaccelerationbasedtechniques.Inthispaper,anovelsystemisdevelopedtoinspectautomaticallythesmalldefectsofrollerbearingsforon-linequalitycontrol.Ratherthanusingaccelerationbasedtechniquesthesystememploysahighsensitiveeddycurrentsensortomeasurethediscementprofilesoftheouterraceforhighsignaltonoiseratio.Furthermore,amorphologicalfilterisusedtoenhancethefeaturesignalwhichissubsequentlymeasuredbyKolmogorovcomplexitymeasure.Bothsimulatedsignalsandmeasureddatashowthatthissystemisabletodiagnosedefectsincludingabnormalsurfaceroundness,waviness,misalignedraceswhicharetypicalqualityproblemsinbearingmanufacturinglines.Theauthors’Answer:Thanksalotforthereviewer’ssuggestion.CorrectedComment3:Inthefirstparagraphinintroduction,inline4,"inspectionmeasurescanbeclassifiedintotwostepstoavoiddefects",shouldbe"avoid";inline8,"causedbymanufacturingerrororabrastivewear",shouldbe"abrasive".Theauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment4:Inthefirstparagraphofsection2.1,line3andline5,andsection2.2,line8,"elestic"shouldbe"elastic".Theauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment5:RevisethefirstsentenceofthethirdTheauthors’Thesentencehasbeenchangedasfollow:Unlikesignalsintheprocessofconditionmonitoringandfaultdiagnosis2,3,4,5,thesignalcharacteristicofdefectivebearingsisquiteweak.Comment6:CorrespondingpreviousworksusingnewsignalprocessingmethodsshouldbeTheauthors’Answer:Thankstothereviewer'sadvice,wehavejoinedthefollowingsentencesandreferencesintheintroduction.PartAseriesofmethodoftheextractionofweaksignalhasbeenbroughtout,suchasthestochasticdifficultyofthelatesignalprocessing.503,2014.pp.1773-1785,2011.Yaguo,Lei.,Jing,Lin.,Zhengjia,He.,“Applicationofanimprovedkurtogrammethodforfault1749,2011.Haiyang,Liu.,Weiguo,Huang.,Shibin,Wang.,Zhongkui,Zhu.,“AdaptivespectralkurtosisfilteringbasedonMorletwaveletanditsapplicationforsignaltransientsdetection,”MechanicalSystemsandSignalProcessing.,2014.PartAsanindexinthetime,theKolmogorovhasbeenfoundtobeaneffectivetoolforsignalysisandconditionassessmentinabearingsystem11.TheKolmogorovcanbeusedtoextractcharacteristicswhicharethenusedtoevaluatebearingqualityandtotracetosourcesRuqiang,Yan.,“ComplexityasaMeasureforMachineHealthEvaluation,”IEEETransactionsonIAM,Vol.55,pp.1327-1334,2004.Part Jing,Wang.,Guanghua,Xu.,“ApplicationofimprovedmorphologicalfiltertotheextractionofTheoreticalComment7:Section2.1andsection2.2havethesameTheauthors’Answer:Thankstothereviewer'sadvice,wehavechangedthetitleofsection2.2as“Measuringbearingdefectsbymeansofthemorphologicalfilter”Comment8:Fig.1(a)"shows"….,ratherthan"describes".Inaddition,thefigurequalityshouldbeTheauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment9:Inthefirstparagraphofsection2.1,line3andline5,andsection2.2,line8,"elestic"shouldbe"elastic".Theauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment10:Ln42,p2,changingwhile"While"into"However"willmakemoreTheauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment11:Inthethirdparagraphofsection2.2,"Inthispaper,weutilizeanaverageweightedcombinationofopen-closingandclose-openingoperation",pleaseexinwhychoosethismethod.Theauthors’Answer:Theopeningoperationcansmooththesignalfrombelowbycuttingdownitspeaks,andtheclosingoperationcansmooththesignalfromabovebyfillingupitsvalleys.AsshowninFig.7,thediscementsensorsignalisclosetothesymmetricalshapeaftertheoperationofremovingmean.Thusweutilizeanaverageweightedcombinationofopen-closingandclose-openingoperationinthispaper.Fig.7.Signalsoffivetypesofbearingsa)Qualityqualifiedbearings.b)Abnormalroughnessonouterraceway.c)Abnormalroughnessoninnerraceway.d)Bruiseonouterraceway.e)Bruiseoninnerraceway.Comment12:Insection2.3,pleaseaddmoreexnationaboutthecomplexitymeasurealgorithmtomakeitclearer.Theauthors’Comment13:Fromthegraphs,itlooksthatthemaintrendofsignalscanberemovedbyaconventionallowpassfilterwhichisefficientandreliable.Whatisthekeybenefitforusingthemorphologicalfilter?Theauthors’Fig.9. Resultsofmorphologicalfiltera)Theoriginalsignal.b)Theperiodicsignal.c)Theimpactsignal.WiththeproblemofthesamefrequencyAsshowninFig.9,theimpulsivesignalswhichareintensityrelatedtothequalityofbearingshavethesamefrequencywiththemaintrendofsignals.Thelowpassfiltermayfilterusefulcompositionsofsignals.MorphologicalfiltercanhelptoidentifydifferentqualityAsshowninFig.7,signalsofbearingswithabnormalroughnesstrendtocosinewaveformclasses,whilesignalsofbearingswithbruiseproblemstrendtotrianglewaveformclasses.Morphologicalfiltercanhelptoidentifydifferentqualityproblemstosomeextent.Morphologicalfilterhashighcodeexecutionefficiency.Therunningtimeisabout0.36susingaThinkpadT410icomputer,whichcanbeusedforon-lineoroff-linemonitoring.Comment14:Line18,P3"tocharacterize…"ratherthanTheauthors’Answer:Thanksforthereviewer’ssuggestion.CorrectedComment15:Line17,ReferencesarerequiredforthekeyfactsdescribedthisTheauthorsAnswer:8（12）Comment16:HowthedigitalizedsignalcanbemanipulatedasstringsforcalculatingtheKolmogorovcomplexityvalues.Morereferencesordescriptionrequired.Theauthors’找合適的參考文獻(xiàn)即可TheKolmogorovcomplexityhasbeenfoundtobeaneffectivetoolforsignal ysisandconditionassessmentinabearingsystem.Asanindexinthetime ,theKolmogorovhasbeenfoundtobeaneffectivetoolforsignalysisandconditionassessmentinabearingsystem11.TheKolmogorovcanbeusedtoextractcharacteristicswhicharethenusedtoevaluatebearingqualityandtotracetosources11.Ruqiang,Yan.,“ComplexityasaMeasureforMachineHealthEvaluation,”IEEETransactionsonIAM,Vol.55,pp.1327-1334,2004.Comment17:WhatisthetypeofnoiseinthesimulatedTheauthors’Thecomplexityofthesignaliscloselyrelatedtothecompositionofsignal.Thenoisedeterminesthecomplexityofthesignalinthesimulation.Theusednoiseiscolorednoise,containingdifferentspectrumstructures.Asforthewhitenoise,regardlessofitsintensitychange,theresultisverifieditscomplexityisessentiallythesame.Fig.3SimulationsoftheIntheoriginalmanuscript,theauthorprovidesresultsofthesimulationsignalandthecomplexityasFig.3.Thenoiseusedcomefromanactualrun-to-failuretestmeasuredbyanaccelerationsensor[].Afterreviewingandcarefullyysistheopinionofthereviewer,authorsthinkthatthepictureisnotasgoodastodescribethecomplexityofthesignalanditmayconfusereaders.Thereforeweuseanotherfigureinthispaperasbelow.Thosedatasetsareconstructedwithdifferenttypicalsignalssuchassinusoidal,sinusoidalwithamplitudemodulation,sinusoidalwithfrequencymodulation,andwhitenoise,andtheyareusedtotesttheverificationofcomplexity.(要加到Fig.3.Lempel-ZivindexvaluesofdifferentsimulationofThanksforthereviewers'valuablesuggestionwhichhasmadeanimportantComment18:ThetitleofFig.3isnotcorrect,pleaseTheauthors’Answer:Thanksforthereviewer’ssuggestion.ThetitleofFig.3hasbeenchangedwith“Simulationsofthecomplexity”.Comment19:Moredetailshouldbeprovidedforthesensor.Especiallyhowdifferencefromanormaleddycurrentsensorthatmakesitmoresensitiveandaccurate.Theauthors’Answer:Thanksforthereviewer’sTheelasticdeformationisquitesmall,andthevalueisintherangeof0.1to20Duetothevibrationtyisverysmall;therearealmostnooutputsignalsfromaccelerationsensor.Thankyouverymuchforthereferees’preciousopinionandourteamisreadytobuyanaccelerationsensorfromNSKforthenextexperimentTheeddycurrentsensorisusedtodetectthetinydeformationofouterring.Comparedwithaccelerationsensor,thismethodismoresensitiveandaccurate.However,thecommoneddycurrentsensorscanachievethecorrespondingdetectionresultsaslongasthedetectionrangeiswithintherangeof0.1to20microns.I'mverysorrybecauseoftheauthors’inappropriatedescribebringtheconfusiontothereviewer.Comment20:AcomparativeresultshouldbeprovidedtoconvincetheproposedmethodismoreTheauthors’Answer:Thanksforthereviewer’sTherearequiteanumberofpapersaboutthefaultdetectionofbearings,whiletheresearchofproblemsofbearingqualityisless.Themaintestingmethodisthroughthedetectionofthestaticgeometrysizeofeachcomponentofbearings.Wealsotrytousethevibrationaccelerationsensorandtheacousticemissionsensor;resultsshowthatthevibrationtyisverysmall;therearealmostnooutputsignalsfromaccelerationsensor,whichcannotbeusedforthedetectionoftheproblemofthebearingquality.Itisfoundthattheacousticemissionsensorcanbeimplementedtodetectthelubricationstateofbearings,whileitcanalsonotbeusedforthedetectionoftheproblemofthebearingquality.Theauthorswouldliketothankthereviewerfortheirinsightfulcommentsandusefulsuggestionsthathelptoimprovethequalityofthiswork.Reviewer#2:Thispapermainlydiscussesbearingqualityevaluationbasedonmorphologyfilterandthekolmogorovcomplexity.Thispaperissomewhatinteresting,butitneedtobefurtherimproved.Thecommentsaregivenbelow.Comment1:Inintroduction,themainbearingdefectsevaluationmethodsintime brieflyintroduced.Othermethod,suchasmethodsinfrequencyandtime-frequency ,shouldbeintroduced.Thenewlituresaboutfaultdiagnosisbasedonmorphologyfilterandcomplexitywhicharepublishedinthisjournalorotherjournals,shouldbedetailedinthefirstsection.Theauthors’Answer:Thankstothereviewer'sadvice,wehavejoinedthefollowingsentencesandreferencesintheintroduction.PartAseriesofmethodoftheextractionofweaksignalhasbeenbroughtout,suchasthestochasticdifficultyofthelatesignalprocessing.503,2014.pp.1773-1785,2011.Yaguo,Lei.,Jing,Lin.,Zhengjia,He.,“Applicationofanimprovedkurtogrammethodforfault1749,2011.Haiyang,Liu.,Weiguo,Huang.,Shibin,Wang.,Zhongkui,Zhu.,“AdaptivespectralkurtosisfilteringbasedonMorletwaveletanditsapplicationforsignaltransientsdetection,”MechanicalSystemsandSignalProcessing.,2014.PartAsanindexinthetime,theKolmogorovhasbeenfoundtobeaneffectivetoolforsignalysisandconditionassessmentinabearingsystem11.TheKolmogorovcanbeusedtoextractcharacteristicswhicharethenusedtoevaluatebearingqualityandtotracetosourcesRuqiang,Yan.,“ComplexityasaMeasureforMachineHealthEvaluation,”IEEETransactionsonIAM,Vol.55,pp.1327-1334,2004.Part Jing,Wang.,Guanghua,Xu.,“ApplicationofimprovedmorphologicalfiltertotheextractionofButitisnotclearwhetherenoughorbatchbearingsareusedtoobtaintheresultinthesetwotables.Howmanybearingsusedshouldbeexinedclearlyasnowitisveryvague.Theauthors’Answer:Thankstothereviewer'spreciousremind,andwehavechangedthemanuscriptasfollow:Duetothesupportoftheproject,testbearingsinthispaperare8306madebytheLYC .Thecompositionofsamplesisasfollow.Thenumberofbruisetestbearingsis20,and10bearingshavethebruiseontheinner