nerc load generating reserves reliability control . martinez nerc load...  nerc...

Click here to load reader

Download NERC Load Generating Reserves Reliability Control . Martinez NERC Load...  NERC Loadâ€Generating

Post on 05-Jun-2018

215 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • NERCLoadGeneratingReservesReliabilityControlStandards

    CarlosMartinezAdvancedSystemsResearchers,Inc.

    cmartinez@asresearchers.comJune1011,2015Washington,DC

    DOE/OETransmissionReliabilityProgram

  • MISOdelivered30samples/secondPhasorGriddataundercurrentDataConfidentialityAgreement(DCA).Algorithmsandvisualsreportsusingdata.

    UniversityofIllinoiscontinuewithalgorithmsresearch,improvements,andvalidations.

    FieldTestdemonstrationatMISOonholdbecausedeliveryofrequestedStateEstimatordataandOnelinediagramsrequireadditionalDCAswithMISO.MISOLegaliscurrentlydefiningnewDCAsrequirements.

    CoordinatewithMISOthedeploymentoflatestalgorithms,visualizationandAutomaticReportsatMISOusingtheirPMUdataandinfrastructure

    ExecutionofproposeFieldTestplanatMISOtogettheirfeedback.

    Risk FieldTestdelaysbecauseofMISOdefinitionofadditionalDCAsfortheremainingdatarequestedforvalidations.

    AutomaticReliabilityReports GARRProject

    2

  • OverallProjectObjectives

    ResearchandReportInteractionsandCorrelationsBetweenPrimary,Secondary,TertiaryDemandSupplyControl

    Performance

    NERCInterconnectionsPrimaryResponse

    SustainabilityPerformanceMetricsDefinition,TrackingandValidation 201014

    NERCInterconnectionsRegionsPrimarySecondaryTertiary

    ControlPerformanceTracking&AnalysisDuringBAL2&BAL3StandardsFieldTrials 201014

    IndustrySurveyandSpecificationforaGenericIndustryWideTool

    forAnalysisTrackPredictFrequencyResponsefor

    Interconnections,BAs,PlantsandGenerators

    3

    COMPLETED COMPLETED

    INPROGRESS

    PENDING

  • FY15Deliverables,RiskFactorsandFollowOnWork

    4

    FY15 Deliverables:a) Completed paper on industry utilization of CERTSNERC tools, accepted for IEEE publication. b) Finalreport for NERC Staff and Reliability Subcommittees including research results and recommendations onNERC Interconnections and two Regions primary, secondary and tertiary control performance during theBAL02 and BAL03 2010 to 2014 Field Trials, including accurate and validated primary responsesustainability performance metrics. c) Functional and database specification for an industrywide tool foranalyzing and tracking Frequency Response for Interconnections, Balancing Authorities, plants andgenerators.

    Risk Factors: Phasor data quality increases substantial development time for data alignment and extrapolations Lack of consensus on frequency events selection criteria and its MW Loss originate discrepancies anddebates for assessing and comparing demandsupply control performance metrics.

    Delay from stakeholders feedback reviewing preliminary reports could delay final deliverables

    Followon Work to Consider for FY16:The project will be completed during FY15. For an effective transfer of projects results to industry: Present and get feedback from NERC Staff, Reliability Subcommittees, and Regions on the final researchresults report, and Frequency Response analysis and tracking tool functional specification.

    Coordinate with PNNL on the implementation and deployment of additional functional specificationcapabilities in their Frequency Response Tool (FRTool2).

    Coordinate and participate with EPG on the monthly analysis and validation of NERC interconnectionsmonthly frequency events report require for BAL003 Frequency Response Performance.

  • NERCInterconnectionsandRegionsPrimarySecondaryTertiary

    ControlPerformanceDuringNewBAL2BAL3StandardsFieldTrials 20102014

    5

  • Background EventsTypicalFrequencyTracesandBAL2,3FieldTrialDataCollected

    6

    BAL00320102014 FIELD TRIALFREQUENCYEVENTSDATAIntercon Eastern Western ERCOT HydroQuebecYear 2014 2013 2012 2011 2010 2014 2013 2012 2011 2010 2014 2013 2012 2011 2010 2014 2013 2012Events 263 177 68 114 139 44 61 69 64 84 49 94 74 101 123 55 90 57

    Withdrawal ReboundArrestingPeriod RecoveryPRIMARYRESPONSE(STANDARD:BAL003)

    SECONDARYTERTIARYRESPONSES(STANDARDS:BAL001,002,004,006)

    StandardsandResearchObjectives

    Inertia

    Governo

    rs

    AGC+

    Reserves

    TracesofSame1SecondMedianforGenerationEvents(148,58)

    Unkno

    wn

    Reason

    s

    DataSources:201020141secondaveragesfrom10Samples/SecondPhasorFrequencyData,andSCADA4SecondData

  • 7

    NERCInterconnectionsPrimaryFrequencyResponsePerformanceTrends

    OBSERVATIONS:a)Eastern,ERCOTincreasingFRMtrends,Western,Hydrodecreasingb)FRMtrendsareequivalentusingBAL003eventsandNERCALR112FrequencyResponseMetric

  • 8

    Region1andits5BAsPrimaryFrequencyResponsePerformanceTrends 201214

    Note:RegionandBAsAnonymousNamesBecauseDataConfidentialityAgreements

    OBSERVATIONS:a)SmallBAsFRMareclosertoitsFROwhencomparetolargeBAsb)OutliereventshaveaverysignificantimpactonFRMestimationsusingregressionmodelsc)AllBAsshowincreasingFRMtrendssincestartingBAL03FieldTrial

  • 9

    Region1FrequencyResponseEstimationsComparisonUsingMedianandRegression

    REGION120122014FREQUENCYRESPONSESMEASUREMENTS(FRM)USINGMEDIANANDREGRESSIONMODELS(ExcludingOutlierEvents)

    Region1BAs Year

    FrequencyResponseMeasurement (FRM)

    MeanAbsolutePercentage Error(MAPE)

    Median[MW/0.1Hz]

    Regression[MW/0.1Hz] Median Regression

    BA12012 33.1 32.8 121.9% 121.5%2013 104.9 130.6 122.3% 150.3%2014 94.5 96.7 373.3% 380.9 %

    BA22012 54.2 46.1 204.4% 177.02%2013 66.4 64.2 144.8% 141.3%2014 75.2 71.3 216.1% 205.8%

    BA32012 36.9 33.4 107.6 % 97.7%2013 36.5 38.3 30.4% 32.3 %2014 47.3 42.2 88.5% 79.5%

    BA42012 9.9 7.9 93.1% 87.7 %2013 10.5 11.6 141.7% 156.7%2014 11.2 11.04 167.3% 165.03%

    BA52012 74.7 66.01 188.7% 173.3%2013 140.9 98.63 83.2% 70.7%2014 117.7 118.2 172.9% 173.6%

    OBSERVATIONS: 60%ofFRMestimations

    producelowerMAPEusingtheMedianModel

    LargeBAsFRMestimationsproducelowerMAPEusingtheRegressionModel

  • EasternTimeErrorCorrectionTrend

    NERCInterconnectionsSecondaryControlPerformanceTrends

    NERCIntercons.AverageFrequencyDeviationTrend

    PossibleReasonsforImprovedSecondaryControlTrends: 2005EnergyPolicyActERO

    creation. AdditionalemphasisonNERC

    ControlStandards ParticipationonFieldTrialsfor

    BAL2andBAL3Standards Reductioninnumberof

    EasternandWesternControlAreas(2011=100,2014=75).

    Toolsthatbetterindicatecurrentperformance,suchastheIntelligentAlarmsfromNERCCERTSResourceAdequacyApplication

    InadvertentInterchangeTool,whichgivesBAsaheadsupthattheircontrolmayrequiresomeinvestigation

    Economy(Recession)

    10

    FromNYISO

  • NERCInterconnectionsPrimaryResponseSustainabilityand

    FrequencyEventsMWLossProbabilisticEstimations 20102014

    11

  • Freq

    uenc

    y Tr

    aces

    of M

    edia

    ns fo

    r 201

    4C

    andi

    date

    Gen

    erat

    ion

    Even

    tsReliabilityRisksforEarlyWithdrawalandSustainabilityConditionIdentificationMetric

    12

    EarlywithdrawalcouldimpactSecondaryResponsebyextendingthetimeforreachingpreeventfrequencieswithinthe15minutesrequirebyBAL002(DCS)

    InterconnectionsfrequencycouldgoclosertoUFLSlevelsifconsecutiveeventsoccurduringperiodsofearlywithdrawal.For2014therewereabout1and2percentofeventswithconsecutiveeventsforERCOTandHydroQuebecrespectively

    RisksforEarlyWithdrawalofFrequencyResponse:

    tric

    (anegative metriccouldindicateearlywithdrawalofprimaryfrequencyresponse)

    SustainabilityConditionPerformanceMetric:

  • 13

    NERCInterconnections2014PrimaryResponseSustainabilityTrendsandImpacts

    Nadir=59.954at32seconds Nadir=59.916=FreqCat9seconds

    Nadir=59.872=FreqCat6seconds

    Nadir=59.512=FreqCat5seconds

    (LowestUFLS)=410[mHz](9timesNadir)

    (LowestUFLS)=170[mHz](2timesNadir)

    (LowestUFLS)=450[mHz](4timesNadir) (LowestUFLS)=400[mHz]

    (2timesNadir)

    OBSERVATIONS:a)TheWesterninterconnectionhastheshortestdistancebetweenlowesteventandUFLS(170mHz),b)Sustainabilityisnotseverelyimpactingsecondarycontrol(BAL02,DCS),returntopreeventfrequency

  • ERCOTEventsMWLossProbabilisticModelUsingPowerLaw,LogNormalDistributions

    14

    OBSERVATIONS: ResearchresultsindicateNERC

    InterconnectionsFrequencyEventsMWLosscanbeestimateusingaprobabilisticmodelbasedonPowerLawandLogNormalDistributions.

    ComparisonofinterconnectionseventsactualMWLossestimatedwiththeprobabilisticmodelproduceerrorsrangingbetween1017percent

    ErrorsbetweenactualandprobabilisticMWLossestimationsissmallerforsmallinterconnections

  • QUESTIONSandANSWERS

    15

View more