406

HVDCConverterTransformers

DesignReview,TestProcedures,AgeingEvaluationandReliabilityinService

JointWorkingGroup

A2/B4.28

February2010

PAGE

20

JWGA2/B4.28

HVDCConverterTransformers

DesignReview,TestProcedures,AgeingEvaluationandReliabilityinService

Members

MilanSaravolac(FR)(Convenor),GilsonMachadoBastos(BR),NarinderS.Dhaliwal(CA),SergioE.Santo(BR),ThomasFreyhult(SE),YannyH.Fu(NL),JanHajek(SE),

A.C.Hall(UK),PeterHeinzig(DE),UgoPiovan(IT),VijayPrasher(IN),ShirasakaYukiyasu(JP)

CorrespondingMembers

MichaelBridger(NZ),PeterW.Christensen(DK),RogerCormack(ZA),JohnF.Graham(BR),CarlosPeixoto(BR),LiuZehong(CN)

Copyright?2010

“OwnershipofaCIGREpublication,whetherinpaperformoronelectronicsupportonlyinfersrightofuseforpersonalpurposes.Areprohibited,exceptifexplicitlyagreedbyCIGRE,totalorpartialreproductionofthepublicationforuseotherthanpersonalandtransfertoathirdparty;hencecirculationonanyintranetorothercompanynetworkisforbidden”.

Disclaimernotice

“CIGREgivesnowarrantyorassuranceaboutthecontentsofthispublication,nordoesitacceptanyresponsibility,astotheaccuracyorexhaustivenessoftheinformation.Allimpliedwarrantiesandconditionsareexcludedtothemaximumextentpermittedbylaw”.

ISBN:978-2-85873-093-3

CONTENTS

Section Description Page

Summary 3

Introduction 3

FailureSurveyandAnalysis 4

FailureDescription 7

FailureSurveyDiscussion 14

GuidelinesforConductingDesignReviewsforHVDCConverterTransformers…19

SystemData 20

InsulationDesign 21

ReviewofTestRequirements 22

PolarityReversal 22

InfluenceofOilConductivityandPolarizationTimeonPR 24

PolarityReversalDuringDCLineFaults 26

InvestigationofPossibleImprovementsofthePRTest 29

PresentKnowledgeAboutOilElectricalConductivity 30

RecommendationsRegardingthePRTest 31

Conclusion 31

Bibliography 32

Appendix 33

SUMMARY

ThisisareportoftheCIGREJWGA2/B4-28,whichhasbeeninitiatedinordertocontinueCIGREactivitiesonlookingintoperformanceandreliabilityinserviceofHVDCConverterTransformers.

Findingsofthereliabilitystudiescoveringtheperiodfrom1971to2009arepresentedinthereport(thelasttwosurveysshowanencouragingdecreaseinthefailurerate).AnextractfromtheDesignReviewGuideisincludedaswell,focusingonthesystemdata(bothACandDC)andspecificaspectsoftheinsulationdesign.TheJWGbelievesthattheDesignReviewGuidewhichisbasedontheexistingdocumentproducedbytheCIGREWG12.22willprovideausefultoolforensuringhigherreliabilityofHVDCConverterTransformersinservice.

TheJWGreviewedexistingstandardtestspecificationswhileplacinganemphasisonthePolarityReversal(PR)testandtheeffectoftheoilconductivityandpolarizationtimewithaviewofrepresentingtherealserviceconditions.Asummaryoftheanalysisispresentedinthereport,indicatingthatifanoiloflowerconductivity(e.g...=10-14S/m)isusedduringthetestandanoilofhigherconductivity(e.g.=10-13S/m)isusedinservice,thenthePRtestwillnotbeeffectiveasthedielectricstressduringthetestwouldbelowerthaninservice.Thetestvoltage,1.25timeshigherthaninservice,wouldcorrespondtoamaximumdielectricstressequaltoonly36%ofthestressthatwouldbeexperiencedinservice.

Eventhoughapossibleimprovementhasbeenidentified,theJWGconcludedthattheexistingtestshouldnotbemodifiedatthisstageduetothelogisticalandpracticalconstraints.AtthesametimetheJWGrecommendedthatthepriorityshouldbeassignedtomeasuringoilconductivitythroughoutthetransformerlifecycleandevaluatingtheeffectoftheoilconductivityduringthedesignanddesignreviewstage.FurtherCIGREactivitieswererecommendedinthisrespect.

KEYWORDS

HVDC,ConverterTransformers,ReliabilitySurvey,DesignReviewGuide,TestSpecifications,PolarityReversal,OilConductivity

INTRODUCTION

Everyyear,SCB4(AGB4.04)collectsreliabilityperformancedataonHVDCsystemsincommercialservice.TheinformationiscompiledandpresentedattheCIGREParisbiennialtechnicalsessionsandisusedforplanning,design,constructionandoperationofHVDCsystems.

AconcernwithregardstoreliabilityofHVDCconvertertransformerswasraisedduringtheWGB4.04technicalpresentationattheCIGRE2000meeting.Followingthis,aJointTaskForce(JTF)B4.04/A2-1“AnalysisofHVDCThyristorConverterTransformersPerformance”wasestablished.TheJTFfirstproceededwithasurvey,bydirectlycontactingtheutilitiesusingHVDCsystems.Thepurposeofthesurveywastocollectuptodatefailureinformationontheconvertertransformers.Theresultsofthesurveyindicatedthattheconvertertransformerfailureswerethemostsignificantcontributortotheforcedenergyunavailability(FEU),eventhoughthemajorfailureswereonlyconfinedtoafewprojects.TheJTFpublisheditsreportsin2004inELECTRANo.212[1]andtheCIGREBrochureNo.240[2].

TheJTFB4.04/A2-1raisedthefollowingissuesinitsreport.

Theimportanceofperformingdielectrictestsinaspecificsequence

WhetherornotthePDacceptancelevelshouldbefurtherreducedfrom<300pC

TestprocedureoftheACappliedvoltagetestonvalvewindingswithpartialdischargemeasurement

MethodofdeterminationofharmoniclossesandextralossescausedbyDCbiascurrent

Shouldthetemperaturerisetestbecomearoutinetestandshoulditsdurationbeextended

WhatkindofacceptancecriterionoftheDGAresultsshouldbeusedfortemperaturerisetest

Methodofdeterminationofwindinghotspottemperaturerisesandtheuseoffiberoptictechnology

Thenecessityandconditionofperformingoilflowelectrificationtest

DuetotheseconcernstheSCB4proposedinitiationofanewA2/B4JWGwitharecommendationtocontinueactivitiesinthisrespect,whilefocusingmoreonimprovingthedesignreviewprocessandreviewingthestandardtestprocedureswithaviewtoensurehigherreliabilityinservice,sothatHVDCremaintobevalidandreliablealternativetothemodernHVACsystems.

Followingtheinaugurationin2004,thenewJWGA2/B4-28quicklyconvergedtoanagreementthatthehighestpriorityshouldbeplacedonthedesignreviewguideasitwasbelievedthatthisdocumentwouldprovideasolidplatformandeffectivetoolforreviewingspecificaspectsofHVDCapplicationsandinteractionsbetweenConverterTransformersandassociatedsystems,bothACandDCalike.IthasbeendecidedtobasethisdocumentontheexistingdesignreviewguidecreatedbytheCIGREWG12.22,whichwaspublishedintheCIGREBrochureNo.204in2002[3],whileexpandingonitfurtherinordertoaddressspecificaspectsassociatedwiththeHVDCConverterTransformerapplication,servicerequirementsandelectricalenvironment.

FAILURESURVEYANDANALYSIS

TheperformanceofHVDCconvertertransformerswasfirstreviewedbyJTF12/14.10.TheJTF12/14.10surveyedalltheHVDCsystemsoperatingfromcommissioningupto1990.Forthepurposeofthesurvey,thefailuresweredefinedasactualandpreventive.Afailurewasconsideredtobeactual,ifremovalofaunitfromservicewasrequiredduetothedamageoftheactivepart.Afailurewasconsideredaspreventiveiftheunitdidnotactuallyfailbutwastakenoutofservicetorepairactivepartsfollowingdiagnostictestingsuchasdissolvedgas-in-oilanalysis(DGA),highinsulationpowerfactor,orfailureofsimilarunit(s).Apartfromthegeneralpertinentdata,adescriptionofthefailurewasrequested,togetherwiththeapparentcause(s)assuspectedbytheutilityand/orbythemanufacturerandtherelation,ifany,withthefactorytype/routinetestsperformedonthatunit.Thesurveydidnotcoverexternalbushingflashoversrelatedtopollution.

TheJTF12/14.10-01publishedin1994areportinELECTRANo.155[4]summarizingtheperformanceofHVDCsystemsupto1990.

Table1.Summaryoftransformerfailures(1972-1990)

Category->

A

Bushings

B

ValveWindings

C

ACWindings

D

StaticShields

ELTC

F

Core&Magn

Shields

G

InternalConn.

Mechanical11

3-Itaipu

3-Itaipu

2_Dumr1-Hamil

1-Comerf1-Oklaun

Dielectric13+2P

1-Blackw1-IFA

2P-IFA

4-Itaipu

5-Itaipu

1-S.Butte

1-CU

Thermal5+35P

1-EelRiv11P-EelRiv

4-EelRiv

24P-

Itaipu

Ind.Curr.2+51P

2-Nelson1P-Nelson2P-Hamil

48P-Itaipu

Operator1

1-Inga

Unknown1

1-PIE

Total33+88P

6+2P

7

6+11P

8

3

2+51P

1+24P

Arecommendationwasmadethatfollowingaspectsshouldbecoveredindetailinthedesignreviewpriortofabrication:

Valvesidebushings

Currentsinducedinthecore

Harmonicstrayfluxes

Shortcircuitforces

Windingturntoturninsulation

Windingcablesinter-stripinsulation

Loadtapchangers

Staticshields

Inaddition,theJTF12/14.10alsoinvestigatedtherelationshipbetweentestandservicestressesasafunctionofconductivityratioforHVDCconvertertransformersandsmoothingreactors[5].TheJTF12/14.10recommendedthatdurationofthelongtermdctestshouldbeextendedtotwohoursandthatonehourappliedactestshouldbeperformedwithpartialdischarge(PD)measurement.TheserecommendationswereeventuallyincorporatedinthenewversionoftheIECStandard61378-2issuedin2001.

AsthecontinuedSCB4reliabilitysurveypointedtoafurtherincreaseinFEUduetotheHVDCConverterTransformerfailuresanewJTF(B4/A2)wasestablishedattheCIGRE2000Sessionwithamandatetoanalysethesystemperformancedatacollectedannuallywithanobjectivetoidentifydominantcausesunderlyingthegrowingtransformerfailurerate.

ThenewJTFB4.04/A2-1firstproceededwithasurveybydirectlycontactingtheutilitiesusingHVDCsystems.

Table2.Summaryofthetransformerfailuresfrom1991to2002.

SummaryofTransformerfailures1991to2002

Category->

A

Bushings

B

ValveWindings

C

ACWindings

D

StaticShields

ELTC

F

Core&Magn.

Shields

G

InternalConn.

Mechanical9+8P

1-H.Hnsh11-Itaipu

1-PIE-Ex

1P-Itaipu1P-F.Skan

1-Radsn

2-C.C.

2-Kontek2P-F.Skan1-Radsn

1P-K.Skn2

1P-NR2

1P-PIE-UG

1P-

Skagrk

Dielectric36+10P

1-Skagrk2-Gezhba

3-Chnd-Pg2-Chatgy1-NR1

6-NR2

1-Meanum2-Itaipu

10-R.Dadri

4P-Chatgy

1-NR1

1-R.Dadri1-R.Dadri*

1-F.Skan1-Gzhba1-R.Dadri

1-Haenam*2P-K.Skn2

1P-Vyborg

1-C.C.

3P-C.C.

Thermal2+3P

2-NR2

1P-C.C.

2P-SwPol

Ind.Curr.

1+3P

1-H.Gate2P-H.Gate

1P-K.skn2

Operator1

1-IPP

Unknown7+1P

1-S.Pond1-V.Smith1P-Madws

1-S.Butte

1-PIE-Upg

1-EelR

1-Haenam1-Welch

Total56+25P

8+3P

28+6P

5+0P

1+1P

12+2P

1+9P

1+4P

*Failureinvolvedmorethanonecategory

Thesurveyalsoincludedfailuredescriptionsprovidedbythereportingsystems.TheJTFB4.04/A2-1submitteditsreport[2]in2004witharecommendationforanewJTForJWGA2/B4tobeformedtofurtherinvestigatetheissuesmentionedaboveandthatthescopeshouldincludeconsiderationofdetailedtestrecommendationstorelevantIECandIEEEcommittees.

TheJWGA2/B4-28conductedthreetransformerfailuresurveyscoveringtransformerfailuresfrom2003to2008.ThesurveyquestionnairewassimilartothatusedbyJTFB4.04/A2-1.Inthefirstsurveyconductedin2005,atotalof14projectsresponded.Therewere12actualfailuresintheperiod2003to2004,outofwhich11failureswereassociatedwiththevalvewindingsandtheremainingonewiththeline(AC)winding.Inaddition,therewere45preventivefailuresreported.Inthesecondsurveyconductedin2007,atotalof29systemsresponded.Therewasatotalof15actualfailuresin2005and2006.Inadditiontherewere21preventivefailuresreported.

ThethirdsurveywasconductedinFeb2009,atotalof34projectsresponded.Therewere9actualfailuresand8preventfailuresreportedin2007and2008.

Table3.-Summaryoffailuresreportedin2005,2007and2009surveys

SummaryofTransformerFailures2003-2008

Category

->

A

Bushings

B

ValveWdgs

CAC

Wdgs

D

StaticShields

ELTC

F

Core&Magn.Shields

G

InternalConn.

Mech

1

1-BassLink

Dielect

4P-Longqu

6–Rihand

1–NR1

1P-Chezu

29+14P

6–Chan.Ph

1P-NR1

1P-

4-ChandBtB

5-Itaipu

2P-NR2

3G_Chang

1P-Skgrk3

1-Itaipu

2-VizagI

1-NR2

1-CU

2P-3G-

2P-Itaipu

Guang

1-Skg3

1-Vyborg

Thermal

4P-EGAT

1P-Zhengp

5+56P

2P-GezNanq

2P-NR1

1P-TSQ

1P-Baltic

4P-TNB

7P-Vizag1

2-Lingbao

2P-NR2

6P-SwePol

1P-RTE

7P-Sasram

1-GG2

1P-

13P-Chand

1-Nanq

3G_Chang

BB

1-TianGuan

1P-NR1

2P-NR2

1P-3G_G

Ind.Curr

0+1P

1P-SwePol

Operator

1+0P

1-KSkan1

Unknown

0+2P

1P-Apollo

1P-Apollo

Total

36+73P

5+6P

18+7P

3+4P

0+0P

3+12P

0+10P

7+35P

*Failureinvolvedmorethanonecategory

Table3.showsthefailuredataforthesurveysfrom2003to2008.Therewasatotalof36actualfailuresand73preventivefailures.

FailureDescription

Allreportingsystemswererequestedtosubmitdescriptionsofthefailures,theoutagetimeandthecorrectiveactiontaken.Thefollowingdescribesthefailuresbycategories.

Bushings(5Actual+6Prevent)ThreeGorges-Longquan-4Prevent

SF6gasleakinthebushing.Oneunitwasreplacedwithspareand3unitswererepaired.Theoutagetimeforreplacingtransformerwas48hours.

Gezhuba-Nanqiao:2Prevent

Thebushingshowedsignsofoverheatingduringoverloadtest.Thebushingswerechanged.Thefailureswereoneateachendonthelink.

Outagetime:121hours

LingbaoBtB:2Actualfailures

Bushingfailureonthevalvewindingside.Outagetime:288hrs.

Thelastshieldingofvalvebushingpossiblydidnotcontactwithgroundreliably,anamountofspace

chargeaccumulates,leadingtodischarge.OnApr.24,2007@04:24,thevalvesidebushingofphaseC(Yabushing,voltageis220kV)exploded,the“groundingcurrentprotectionofprotectionandcontrolsystem”,“gasprotection”,“pressurerelease”protectionoperatedcorrectly.

Outagetime:261hours

Gezhuba-Nanqiao:1Failure

NanQiaostationPoleIIphaseCtrippedbecauseofthemaintankgasrelayactedbutpressurereliefnotactedin21,Dec,2008.Afterthetroubleunitwasdisassembled,theinternalconnectionbetweenwindingandbushing2.1wasfounddamaged.Thedamagedtransformerwastransportedtolocalmanufacturerforrepair.Afterinvestigation,itwasconcludedthatthefailurewascausedbybushingqualityitself.

Outagetime:91.5hours

Gui-Guang2-1Failure

OneY-ConvertertransformerACsystemsidebushingblastingduringoperation,andthepoletripped.Theunitwasreplacedwithspare.

Outagetime:177hours

BassLink:1Failure

Therewasabushingfailureduetomechanicalstressfractureatthechangeindiameterbehindtheoilflange.FractureinRIPpropagatedbythermalcyclingduetovariablepowertransferofinterconnector.Failureoccurredwhenfracturecompromisedvoltagewithstandwithinbushingflange

Outagetime:200.65hours

ValveWindings(18Actual+7Prevent)

Rihand-Dadri-2Actual(2003-2004)

Thefailurehasbeencausedbyaninterturnfaultduetotheweakeningofinterstrandinsulationbetweenparallelconductors.Meanwhilethetransformerwindingshavebeenstrengthenedtothemaximumextentpossiblewithoutcompromisingthedesignparameterslikeimpedance,etc.Therootcauseoffailurewasstillunderinvestigationatthetimeofreporting.

Theunitswerereplacedwithsparesin39hours.

RihandDelhi:4Actual(2005-2006)

Fourtransformersfailedin2006.Threefaultsonthedeltavalvewindingoneonstarvalvewinding.Noexactcauseforfailureisknown.

Outagetime: 72hoursforthreefailuresassparewasavailable.

1152hoursinforthefourthfailure(nospare).

Chandrapur-Padghe-4Actual(2003-2004)

Fourunitsfailedduetoturntoturnfaultinthevalvewindings.Therootcauseoffailurewasstillunderinvestigationatthetimeofreporting.Theunitsweresenttofactoryforrepairs.

Outagetimeswere4359,7464and7464hourswhensparewasnotavailable.Outagetimewas221hourswhensparewasavailable

Chandrapur-Padghe:2Actual(2005-2006)

OnetransformerfailedatChandrapurandwasreplacedbyspare.Outagetime:264hours

OnetransformerfailedatPadghe,nosparewasavailable.Outagetime:5760hours.

Itaipu-5Actual+2Preventfailures

Fiveunitsfailedduetofaultinthevalvewindings.Theunitswererepairedonsite.Thereare4spareunitsonsiteandthesparecanbeplacedinservicein4days.

Arebuildingprogramhasbeenstartedtoreplacetheoldtransformerswithrebuiltunits.Thenewunitshaveredesignedwindings.

Outagetime:24hourseach

NelsonRiver1:2Prevent

2005-Transformerwasover30yearsoldandwasgassing2006-Replacedover30yearsoldtransformerthatwasgassingOutagetime:Alloutageswereplannedaheadoftime

05-199hrs

06-541hrs

NelsonRiver2:1Actual+2Prevent

2005-Replaced2unitswhichhadbeendesignatedforreplacementduetogenericfault. .Outagetime: Alloutageswereplannedaheadoftime

05-181hrs

05-139hrs

2007-AtDorsey,convertertransformerT31Sfailedduetofaultinthevalvewinding.Thetransformerwasreplacedwiththesparetransformer.

Theforcedoutagetimewas576.37hours

Skagerrak3:1Prevent

ChangedtransformerbecauseofpossibilityofCopperSulphidefailureastheoilwasfoundtobecorrosive.

Outagetime:336hours

ACWindings(3Actual+4Prevent)NelsonRiverBipole1-1Failure

TheunitfailedduetoturntoturnfaultneartheneutraloftheACwindings.Thisunitwasover30yearsold.

Theunitwasreplacedwithspare.Totaloutagetimewas689hours.Theextendedoutagetimewasrequiredduetocompactdesignofthebuswork.Boththeoldandnewunitshavetobedisassembledbeforemoving.

NelsonRiverBipole1-1Preventfailure

Thisunithadbeengassingforanumberofyearsandhadbeenkeptinserviceusingon-linedegassifier.Astheunitwasover30yearsolditwasreplacedwithspareandthenotrepaired.Totalplannedoutagetimewas238hours.

NelsonRiverBipole2-2Preventfailures

Thesetransformerswerepartofthestage1equipmentandhadbeenknowntohaveconductorinsulationfailureproblems.Theunitshadbeengassingforafewyearsandwerereplacedwithspareunits.Repairedwerefoundtobenotcosteffective.

Itaipu:1Actualfailure

Unknowncause.Outagetime:24hours

CUPROJECT:1Failure

Transformerhadafaultintapchangerwindingononephaseorintapchangerwindingleads.Faultcurrentsdamagedotherwindings.Arcgasses/productsapparentlyinitiatedafaulttogroundinsidethetankto1ac-sidebushing.

Thetransformerwasmanufacturedin1976andwasinservicefor30years.OutageTime:247hours

Tian-Guang:1PreventFailure

OnetransformerwasreplacedbythesparetransformersincetheDissolvedGasAnalysis(DGA)Problem.ThefaultissuspectedtobeintheACwinding.

Outagetime:49hours

LoadTapChanger(LTC)(3Actual+12Prevent)BalticCable:1Prevent

Duetohighnumberoftapchangeroperationstheoilvalvewascloggedbecauseofdirtyoil.Thelinktrippedonlowoilpressure.Onlineoilfilteringsystemhasbeeninstalledinordertoreducetheoilcontamination.

Outagetime-39hours.

SwePol:6Prevent

All6unitshavebeengassingsincetheverybeginning.AswasstatedinthelastsurveyaspareunitwasorderedandtheunitwiththehighestgassingratewasexchangedforthenewspareinOctober–November2004.DeficiencieswerefoundintheOLTCselectorswitchcontacts,howeverthiswasatthetimeandstillisnotconsideredtobetheonlyfault.Thisledtothedecisiontochangethecontactstonewsilverplatedonesinallunitsinbothstationsintheautumn2005.AstiffeningofthemountingoftheOLTCsweremadeatthesametimeasaprecautiontoavoidfrettingonthecontactsduetovibrations.Atthesametimethetransformeroilwasdegassedandapassivatoragainstsulphurcorrosionwasadded.NormalrevisionoftheOLTCswasalsomadeduringtheoutagetime(afterabout90,000operations).

Outagetime:309hours

3G-Changzhou:2PreventFailures

AbnormallyhighgassingintheOLTC.Thetransformerwasreplacedbyaspareunit.Outagetime:72hours

InJanuary2007outage,therepairedY/Ytransformerwasputintoserviceagaintotesttheoperationcapability.InFebruarythetransformerC2H2increasedbeyondtheChinesestandard.In2008February,thetransformerwasreplacedbyspareone.

Outagetime:72hours

3G-Guangdong;2Prevent

AbnormallyhighgassingintheOLTC.Thetransformerwasreplacedbyaspareunit..Outagetime:72hourseach.

Skagerrak3:1Failure

PhasetophaseflashoverinthetapchangerdiverterswitchofaconvertertransformerinDenmark.Thetransformerwassentforrepair.Allthreeloadtapchangerwindingsandthediverterswitchwerereplaced.

Outagetimewas7366hours

Vyborg:1Failure

OLTCdiverterwasfaultedonphaseB,HVCU4,RS(rectifier):.DiverterprotectionhadoperatedwhentheOLTCswitchingoccured.

Outagetime:22.5hours

KontiSkan1:1ActualFailure

Thetapchangerwassetinmanualonthemarshallingboxintherectifier,thiscausedthecontrolofthetapchangertogotolocalcontrol,whenthemarshallingboxwasputbackinautomode,thecontrolremained

thetapchangerinlocalcontrol,whenthepowerwasraisedthecontrolwasswappingbetweentwocontrolmodeswhichresultedinmorethan600tappingordersinonehour,whichcausedhightemperatureandoverpressureinthetapchangertank,withtripfrompressuredevice.Aseparationplateonthetapchangercoverblewoff.

Outagetime:210hours

CoreandMagneticShields(10Prevent)ThailandMalaysia-8Prevent

Thecoregroundingresistorwasfoundtobeburntononeunit.Allunitswererepairedbysolidlygroundingthecore.

RTE-France:1Prevent

Contactbetweenmagneticcircuitscreatedahotspotinthetank.Thetransformerwasexchangedwithspare.Thetransformerwasrepairedonsite.ADCvoltagetestwasperformedaftertherepair.

After6monthsintoservicethistransformercontinuedtocreategasinoil.Itwasmadeanoiltreatmentoneyearafterthereturntoserviceofthetransformer.Now,gasinoilisquitestable.

Outagetime:27hours

Apollo:1Preventfailures

Bridge1-Thetransformerwasgassingandwastakenoutforinspection.Nofaultwasfoundandthetransformerwasputbackinservice.

Outagetime:72hours

InternalConnections(7Actual+35Prevent)Chezu-1Preventfailure

Breakdownofinsulationbetweenwindingsandthetankviapressboardsupport.Pressboardcontaminationwassuspectedandwasreplaced.

ThreeGorgesZhengping-1Preventfailure

Thegascontentwashigherthanthecustomer’sstandardsotheunitwasreplaced.

Sasaram,Chandrapur,VizagI-27Preventfailures

Theyhavehadlocalizedoverheatingontheconnectionbetweenvalvewindingandthebushingduetolackofcooling.TheproblemwasnotedthroughDGA.Plannedoutagesweretakentorepairtransformers.Themodificationwasmadetoimprovetheoilflow.Alltransformersofthesamedesignwererepairedonsite.

ChandrapurBtB:4Actualfailures

Allfailuresduetoflashoverbetweentapleads.Oneunitfailedtwiceevenafterithadbeenrepairedpreviously.Thethirdunithadbeenrepairedtwicebeforeandfailedagain.

OutageTime:456hrsin2005

423hrsin2006

Vizag1:2Actualfailures

Thetwofailuresoccurredduetoflashoverbetweenthetapleads.Therepairsweremadeonsite.Oneoftheunitshassincefailedtwicein2007.

Outagetime:Year2005-360hours.

NelsonRiver1:2Preventfailures

2006-RepairedtapleadthatwasoverheatingOutagetime:276hrs.Theoutagewaspreplanned.NelsonRiver2:3Prevent

2006-Thiswasoneofthenewunitsthatwasinstalledthepreviousyear.Theunitstartedgassingshortlyafteritwascommissioned.Detailedinspectionshowedoverheatedconnections.Theunitwasreplacedwithspareandisunderrepairs.

Outagetime: Outagetime67hours.

SwePol-1Preventfailure

Theunithasbeengassingforlongperiodoftimeandwasreplacedafterthesparewasavailable.Thecauseofgassingisstillunderinvestigation.

Tian-Guang:1ActualFailure

Failureofinternalconnectionbetweenwindingandthebushing.Outagetime:672hours

Apollo:1Preventfailures

Bridge2-Duetoabnormalgassinglevel,thetransformerwastakenoutofserviceandinspected.Arcingmarkswerefoundonthebushingconnections.Bushingcontacts(multicontacts)wereremovedtobereplaced.

Outagetime:72hours

3G-Guang:1PreventFailure

TransformerwasgassingandtakenoutofserviceforrepairsinMar2007andreplacedwithspare.Theinternalconnectionbetweenwindingandbushingwasfoundbroken.TheunitwasrepairedandputbackinserviceandisworkingOK.

Outagetime:72hours

NelsonRiver2:1Preventfailure

AtHenday,convertertransformerT42Swasinspectedinternallyandnumberofhotconnectionswerefound.Thiswasanewtransformerbuthasbeengassingsincethedayitwasinstalled.Thetransformerwasreplacedwithspareandisbeingrepaired.Theworkwasdoneduringregularlyscheduledoutage.

FailureSurveyDiscussion

Thefailurerateiscalculatedasfollows:

FailureRate

TotalNumberofUnitsFailed

NumberofUnitsinServiceYearsinService

DuetoinsufficienttechnicaldatamadeavailabletotheJWGforthepurposeofthisanalysisitwasnotpossibletodeterminetheactualcausesofindividualfailuresasrequestedbySCB4,noritwaspossibletoidentifysomegeneraltrends,despitesomelimiteddatavolunteeredbysomemembersoftheJWG.

TheJWGalsodidnotlookintofailuresattributabletotheeffectofthecorrosiveoilasthisphenomenonwascoveredbyadedicatedCIGREWG(A2-32).

Figure1.showstheConverterTransformersfailureratehistorybasedonthecombineddataofsurveyscarriedoutinthefollowingperiods:

1972to1990

1991to2002

2003to2004(reportedin2005)

2005to2006(reportedin2007)

2007to2008(reportedin2009)

ItcanbeobservedintheFigure1.thatfollowingareducedfailurerateintheperiodfrom1991to2002,therewasasignificantincreaseintheperiodfrom2003to2008asderivedfromthecombineddatacollectedthroughthreeindividualsurveys(2003-2004,2005-2006and2007-2008).However,whenthefailurerateisshownseparatelyforthethreeindividualsurveysreportedintheperiod