美加停电调研报告英文Word下载.docx

美加停电调研报告英文Word下载.docx

《美加停电调研报告英文Word下载.docx》由会员分享，可在线阅读，更多相关《美加停电调研报告英文Word下载.docx（16页珍藏版）》请在冰豆网上搜索。

dollars）.

2.TheProcessandCauseoftheAugust14Blackout

Part1:

theBlackoutBeganinOhio

Phase1ANormalAfternoonDegrades:

12:

15EDTto14:

14EDT

NorthernOhiowasexperiencinganordinary

Augustafternoon,withloadsmoderatelyhighto

serveairconditioningdemand,consuminghigh

levelsofreactivepower.WithtwoofCleveland’s

activeandreactivepowerproductionanchors

alreadyshutdown（Davis-BesseandEastlake4）,

thelossoftheEastlake5unitat13:

31EDTfurther

depletedcriticalvoltagesupportfortheCleveland-

Akronarea.Detailedsimulationmodeling

revealsthatthelossofEastlake5wasasignificant

factorintheoutagelaterthatafternoon—

withEastlake5outofservice,transmissionlineloadingswerenotablyhigherbutwellwithinnormal

ratings.AfterthelossofFE’sHarding-

Chamberlinlineat15:

05EDT,thesystem

eventuallybecameunabletosustainadditional

contingencies,eventhoughkey345kVlineloadings

didnotexceedtheirnormalratings.Had

Eastlake5remainedinservice,subsequentline

loadingswouldhavebeenlower.LossofEastlake

5,however,didnotinitiatetheblackout.Rather,

subsequentcomputerfailuresleadingtotheloss

ofsituationalawarenessinFE’scontrolroomand

thelossofkeyFEtransmissionlinesduetocontacts

withtreeswerethemostimportantcauses.

At14:

02EDT,DaytonPower&

Light’s（DPL）Stuart-

Atlanta345-kVlinetrippedoff-lineduetoa

treecontact.Thislinehadnodirectelectrical

effectonFE’ssystem—butitdidaffectMISO’sperformance

asreliabilitycoordinator,eventhough

PJMisthereliabilitycoordinatorfortheDPLline.

OneofMISO’sprimarysystemconditionevaluation

tools,itsstateestimator,wasunabletoassess

systemconditionsformostoftheperiodbetween

12:

15and15:

34EDT,duetoacombinationof

humanerrorandtheeffectofthelossofDPL’sStuart-

AtlantalineonotherMISOlinesasreflectedin

thestateestimator’scalculations.Withoutan

effectivestateestimator,MISOwasunabletoperform

contingencyanalysesofgenerationandline

losseswithinitsreliabilityzone.Therefore,

through15:

34EDTMISOcouldnotdetermine

thatwithEastlake5down,othertransmission

lineswouldoverloadifFElostamajortransmission

line,andcouldnotissueappropriatewarnings

andoperationalinstructions.

Causes:

Thefactthatthestateestimator

wasnotrunningautomaticallyon

itsregular5-minuteschedulewas

discoveredabout14:

40EDT.The

automatictriggerwasre-enabled

butagainthestateestimatorfailedtosolvesuccessfully.

Thistimeinvestigationidentifiedthe

Stuart-Atlanta345-kVlineoutage（which

occurredat14:

02EDT）tobethelikelycause.This

lineiswithintheDaytonPowerandLightcontrol

areainsouthernOhioandisunderPJM’sreliability

umbrellaratherthanMISO’s.Eventhoughit

affectselectricalflowswithinMISO,itsstatushad

notbeenautomaticallylinkedtoMISO’sstate

estimator.

Thediscrepancybetweenactualmeasuredsystem

flows（withStuart-Atlantaoff-line）andtheMISO

model（whichassumedStuart-Atlantaon-line）

preventedthestateestimatorfromsolvingcorrectly.

At15:

09EDT,wheninformedbythesystem

engineerthattheStuart-Atlantalineappeared

tobetheproblem,theMISOoperatorsaid（mistakenly）

thatthislinewasinservice.Thesystem

engineerthentriedunsuccessfullytoreachasolution

withtheStuart-Atlantalinemodeledasin

serviceuntilapproximately15:

29EDT,whenthe

MISOoperatorcalledPJMtoverifythecorrectstatus.

AftertheydeterminedthatStuart-Atlantahad

tripped,theyupdatedthestateestimatorandit

solvedsuccessfully.TheRTCAwasthenrunmanually

andsolvedsuccessfullyat15:

41EDT.

MISO’sstateestimatorandcontingencyanalysis

werebackunderfullautomaticoperationand

solvingeffectivelyby16:

04EDT,abouttwominutes

beforethestartofthecascade.

Phase2FE’sComputerFailures:

14:

14EDTto15:

59EDT

Startingaround14:

14EDT,FE’scontrolroom

operatorslostthealarmfunctionthatprovided

audibleandvisualindicationswhenasignificant

pieceofequipmentchangedfromanacceptableto

aproblematiccondition.Shortlythereafter,the

EMSsystemlostanumberofitsremotecontrol

consoles.Nextitlosttheprimaryservercomputerthatwashostingthealarmfunction,andthenthe

backupserversuchthatallfunctionsthatwere

beingsupportedontheseserverswerestoppedat

54EDT.However,foroveranhournoonein

FE’scontrolroomgraspedthattheircomputersystems

werenotoperatingproperly,eventhough

FE’sInformationTechnologysupportstaffknew

oftheproblemsandwereworkingtosolvethem,

andtheabsenceofalarmsandothersymptoms

offeredmanycluestotheoperatorsoftheEMS

system’simpairedstate.Thus,withoutafunctioning

EMSortheknowledgethatithadfailed,FE’s

systemoperatorsremainedunawarethattheir

electricalsystemconditionwasbeginningtodegrade.Unknowingly,theyusedtheoutdated

systemconditioninformationtheydidhavetodiscount

informationfromothersaboutgrowingsystem

problems.

Cause:

Lossofthefirstservercausedanauto-pagetobe

issuedtoalertFE’sEMSITsupportpersonnelto

theproblem.Whentheback-upserverfailed,it

toosentanauto-pagetoFE’sITstaff.Theydidnot

notifycontrolroomoperatorsoftheproblem.At

15:

08EDT,ITstafferscompleteda“warmreboot”

（restart）oftheprimaryserver.Startupdiagnostics

monitoredduringthatrebootverifiedthatthe

computerandallexpectedprocesseswererunning;

accordingly,FE’sITstaffbelievedthatthey

hadsuccessfullyrestartedthenodeandallthe

processesitwashosting.However,althoughthe

serveranditsapplicationswereagainrunning,the

alarmsystemremainedfrozenandnon-functional,

evenontherestartedcomputer.TheITstaff

didnotconfirmthatthe

alarmsystemwasagain

workingproperlywiththe

controlroomoperators.

Anothercasualtyofthelossofbothserverswas

theAutomaticGenerationControl（AGC）function

hostedonthosecomputers.LossofAGCmeant

thatFE’soperatorscouldnotrunaffiliated

powerplantsonpre-setprogramstorespondautomatically

tomeetFE’ssystemloadandinterchange

obligations.AlthoughtheAGCdidnot

workfrom14:

54EDTto15:

08EDTand15:

46EDT

to15:

59EDT（periodswhenbothserverswere

down）,thislossoffunction

doesnotappeartohavehad

aneffectontheblackout.

Phase3ThreeFE345-kV

TransmissionLineFailures

andManyPhoneCalls:

05EDTto15:

57EDT

From15:

05:

41EDTto15:

41:

35EDT,three345-kV

linesfailedwithpowerflowsatorbeloweach

transmissionline’semergencyrating.Theseline

tripswerenotrandom.Rather,eachwastheresult

ofacontactbetweenalineandatreethat

hadgrownsotallthat,overaperiodofyears,it

encroachedintotherequiredclearanceheightfor

theline.Aseachlinefailed,itsoutageincreased

theloadingontheremaininglines（Figure5.5）.As

eachofthetransmissionlinesfailed,andpower

flowsshiftedtoothertransmissionpaths,voltages

ontherestofFE’ssystemdegradedfurther（Figure

5.6）.

Overgrowntrees,asopposedto

excessiveconductorsag,caused

eachofthesefaults.Whilesag

mayhavecontributedtothese

events,theseincidentsoccurred

becausethetreesgrewtootallandencroached

intothespacebelowthelinewhichisintended

tobeclearofanyobjects,notbecausethelines

saggedintoshorttrees.Becausethetreeswereso

tall（asdiscussedbelow）,eachoftheselines

faultedundersystemconditionswellwithinspecified

operatingparameters.Theinvestigationteam

foundfieldevidenceoftreecontactatallthree

locations,includinghumanobservationofthe

Hanna-Junipercontact.Evidenceoutlinedbelow

confirmsthatcontactwithtreescausedtheshort

circuitstogroundthatcausedeachlinetotripout

onAugust14.

Tobesurethattheevidenceoftree/linecontacts

andtreeremainsfoundateachsitewaslinkedto

theeventsofAugust14,theteamlookedat

whethertheselineshadanypriorhistoryofoutages

inprecedingmonthsoryearsthatmighthave

resultedintheburnmarks,debarking,andother

vegetativeevidenceoflinecontacts.Therecord

establishesthattherewerenopriorsustainedoutages

knowntobecausedbytreesfortheselinesin

2001,2002,and2003.2

Phase4138-kVTransmissionSystem

CollapseinNorthernOhio:

39to16:

08EDT

AseachofFE’s345-kVlinesintheClevelandarea

trippedout,itincreasedloadinganddecreased

voltageontheunderlying138-kVsystemserving

ClevelandandAkron,pushingthoselinesinto

overload.Starti