Text

Vogtle GSI-191 Resolution Plan and Current Status, NRC Public Meeting, November 5, 2015, Revised Meeting Handouts
	ML15320A087
Person / Time
Site:	Vogtle
Issue date:	11/05/2015
From:	; Southern Nuclear Company
To:	; Office of Nuclear Reactor Regulation
References
	Download: ML15320A087 (94)
	v • d • e

NOVEMBER 5, 2015VOGTLE GSI-191 RESOLUTION PLAN AND CURRENT STATUSNRC PUBLIC MEETING 2AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 3PURPOSE OF MEETINGObtain staff feedback on the overall GSI-191 resolution path for VogtleDiscuss proposed plant modificationsDiscuss use of design basis (vs. best estimate) inputs in the evaluation 4VOGTLE PLANT LAYOUTWestinghouse 4-loop PWR (3,626 MWt per unit)Large dry containmentTwo redundant ECCS and CS trainsEach train has an RHR pump, a high head pump, an intermediate head SI pump, and a CS pumpSI and high head pumps piggyback off of the RHR pump discharge during recirculation.Maximum design flow rates:RHR 3,700 gpm/pumpCS 2,600 gpm/pumpTwo independent and redundant containment air cooling trains 5STRAINER ARRANGEMENTTwo RHR and CS pumps each with their own strainerEach GE strainer is similar with four stacks of disksRHR strainer: 18-disk tall, 765 ft2, 5 fttallCS strainer: 14-disk tall, 590 ft2, 4 fttallPerforated plate with 3/32" diameter holesRHR BCS BRHR ACS A 6PLANT RESPONSE TO LOCASPlant response includes the following general actions: Accumulators inject (breaks larger than 2 inches)ECCS injection is initiated from the RWST to the cold legs via RHR, SI, and High Head pumpsContainment spray is initiated from the RWST via CS pumps RHR pumps switched to cold leg recirculation at RWST lo-lo alarmCS pumps switched to recirculation at RWST empty alarmCS pumps secured no earlier than 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months after start of recirculation, and probably before 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months depending on pressure and dose rateRHR pumps switched to hot leg recirculation at 7.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months 7AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 8SCHEDULE UPDATEMILESTONEEXPECTED COMPLETION DATECurrent StatusDevelop containment CAD model to include pipe weldsCompleteCompleteConduct meeting with NRC3rdQuarter 2013CompleteModify probabilistic risk assessment (PRA) to include Strainer and Core Blockage events4thQuarter 2013CompletePerform Chemical Effects testing1stQuarter 2014CompletePerform thermal and hydraulic modeling of RCS, Core, and Containment conditions1stQuarter 2014CompletePerform Strainer Head Loss and Bypass testing to establish correlation for range of break sizes2ndQuarter 2014StrainerBypass Testing -CompleteStrainer Headloss Testing -Using 2009 Vogtle testAssemble base inputs for CASA Grande2ndQuarter 2014Complete -Using NARWHAL instead of CASAGrandeEvaluate Boric Acid Precipitation impacts3rdQuarter 2015Plan on using PWROG WCAP-17788Finalize inputs to CASA Grande3rdQuarter 2015In Progress, Using NARWHAL instead of CASAGrandeComplete Sensitivity Analyses in/for CASA Grande4thQuarter 20152ndQuarter 2016Integrate CASA Grande results into PRA to 1stQuarter 20163rd Quarter 2016Licensing Submittal for VEGPTo be established through discussions with NRC -tentatively September 2016Projected 1stQuarter 2017, require input from: SE on STP Pilot Project and SE on WCAP-17788; which are not yet available.

9WHAT WE HAVE LEARNEDContainment sprays only actuate for the largest hot leg breaks under best estimate conditionsThis results in unsubmerged strainers for many breaks because RWST is left with unused waterWide variation in post-LOCA water levels and sump chemistryAluminum corrosion is greater with sprays operatingPlanned modificationsReduce height of RHR strainers by ~6 inches (remove 2 disks)Change procedures to continue RWST drain down to just below empty level set point when CS does not actuateAllows reduction of Tech Spec min water level by 2% (~1 ft) (increased operating margin)Equivalent to ~2 inches in containment pool levelCombination of these three modifications results in submerged strainers and reduces risk 10RWST LEVELS AND ALARMS(ECCS to Recirculation)(CS to Recirculation)El. 220'-0" Bottom of TankEl. 274'98%93%29%8%(Proposed Change)51.45'(Existing) 11SUBMERGED STRAINER RESULTSBefore ModificationsAfter Modifications~4.97 ftSBLOCA RecircValves open = ~4.75 ftLBLOCA Best Estimate = ~6.6 ftLBLOCA Max = ~8.2 ft~4.425 ftSBLOCA RecircValves open = ~4.5 ftLBLOCA Best Estimate = ~7.9 ftLBLOCA Max = ~8.7 ftAll heights are measured from floorSBLOCA long term = ~5.3 ftSBLOCA long term = ~3.5 ft 12OVERVIEW OF METHODOLOGY Resolution plan modified based on South Texas Project (STP) pilot plant methodologyNo longer planning to perform head loss testing to develop a rule-based head loss model (as presented to the NRC in November 2014)2009 Vogtle head loss testing will be used to determine which breaks contribute to risk quantificationBreaks with debris quantities greater than tested (mass/SA)*will "fail"Breaks with debris quantities less than tested (mass/SA) will "pass"Will continue to quantify risk by evaluating GSI-191 failures for number of strainers in operationWill use consensus models/design basis inputs for parameters in the GSI-191 Risk-Informed Software*Mass/SA = Mass of debris per unit strainer area 13PILOT PLANT COMPARISON (STPVS. VOGTLE)DifferencesPhysical ECCS trainsStrainer configurationContainment Spray SetpointStrainer designStrainer surface areaModeling SoftwareBreak size and orientation samplingStrainer head loss test protocolChemical effects Core blockageTime-dependent comparison to failure criteriaMethod for calculating CDF and LERFSimilaritiesPhysicalLarge dry containment4 Loop Westinghouse NSSSLow density fiberglass insulationTrisodiumphosphate bufferModelingBreak-specific ZOI debris generation Unqualified coatingsDebris transportPast head loss testing to establish debris limitsPenetration testingMass balance of debris on strainer and core 14SOFTWAREBADGER (Break Accident Debris Generation Evaluator)A computer program that automates break ZOI debris generation calculations using CAD softwareNARWHAL (Nuclear Accident Risk Weighted Analysis)A computer program that evaluates the probability of GSI-191 failures by holistically analyzing the break-specific conditions in a time-dependent manner 15ANALYSIS FLOWCHARTNARWHAL Break Analysis(Strainers)NARWHAL Break Analysis(Core)Strainer Head Loss TestAcceptable MitigationWCAP-17788Strainer PenetrationTestCDFLERFAcceptable Risk: Submit LARNot BoundedBoundedBoundedNot BoundedRG 1.174 Acceptance GuidelinesUnacceptable Risk: Implement Refinements or Modifications 16AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 17DEBRIS GENERATIONInsulation and qualified coatingsAutomated analysis with containment CAD model using a tool called BADGERCalculate quantity and size distribution for each type of debrisPartial breaks from 1/2 inch to double-ended guillotine breaks (DEGBs) for all Class 1 welds in containmentZOIs consistent with deterministic approachUnqualified coatings and latent debris quantities are identical for all breaksNukon17DQualified Epoxy& IOZ with Epoxy topcoat4DInteram Fire Barrier Material11.7D 18INSULATION AND QUALIFIED COATINGS QUANTITIESBADGER database contains 28,434 breaks at 930 weld locationsBreaks evaluated at each Class 1 ISI weldPartial breaks evaluated in 2 inch increments for smaller break sizes, 1 inch increments for larger break sizes, and 1/2 inch increments for largest break sizesPartial breaks evaluated in 45°increments around pipeDEGB evaluated at every weldDebris quantities vary significantly across the range of possible breaks, and are calculated for each breakNukon: 0 ft3to 2,229 ft3Qualified epoxy: 0 lbmto 219 lbmQualified IOZ: 0 lbmto 65 lbmInteramfire barrier: 0 lbmto 60 lbm 19NUKON DEBRIS GENERATED 20UNQUALIFIED COATINGSTypes of Unqualified Coatings at VogtleInorganic Zinc (IOZ)AlkydEpoxyIOZ, alkyd, and epoxy coatings are assumed to fail as 100% particulateSize distribution for degraded qualified coatings and failure timing no longer being utilized because of the reliance on 2009 head loss testing 21UNQUALIFIED COATINGS LOCATIONSCoatings that fail in upper containment would have a reduced transport fraction for breaks where containment sprays are not initiatedCoating TypeUpper Containment Quantity (lbm)Lower Containment Quantity(lbm)Epoxy1,6021,127Alkyd059IOZ2431Total1,6261,217 22CONTAINMENT POOL WATER LEVELSump pool depth is evaluated on a break-specific basis using conservative inputs to minimize water levelThe evaluation considers break location and size for the appropriate contribution of RWST, RCS and SI accumulators to pool levelPlanning to remove 2 disks from each RHR strainer to reduce overall heightModifying procedures for breaks that don't activate containment spray Continue RWST drain down to just below empty level set pointAllows reduction of Tech Spec RWST min water level by 2% (~1 ft) to increase operator marginStrainers are submerged for all scenarios 23DEBRIS TRANSPORTUsing logic tree approach defined in NEI 04-07 consistent with industry developed methods for deterministic closureBlowdownWashdownPool fillRecirculationErosion 24TRANSPORT -WASHDOWNContainment Sprays OnAll fines (fiber and particulate) washed to lower containmentRetention of small and large pieces caught on gratings estimated based on Drywell Debris Transport StudyWashdownto various areas proportional to flow splitContainment Sprays OffAssumed 10% washdown for fines due to condensation and 0% for small pieces 25CONTAINMENT PRESSURE COMPARISONCurves illustrate the basis for containment spray actuation logicFor analysis purposes (NPSH and gas voiding) containment pressure conservatively reduced to saturation pressure when pool temperature is >212 °F and atmospheric when pool temperature is 212 °F or lessAssumption to use for containment spray actuation is not obvious for most breaks/sizes 26FIBER TRANSPORT FRACTIONSTO ONE RHR STRAINER* This pump lineup was evaluated for different break locations. The transport fractions shown are the bounding values for an annulus break near the strainers.Debris TypeSize1 Train w/ Spray2 Train w/ Spray1 Train w/o Spray2 Train w/oSpray*NukonFines58%29%23%12%Small48%24%5%2%Large6%3%7%4%Intact0%0%0%0%LatentFines58%29%28%14%Strainers in Operation2412 27PARTICULATE TRANSPORT FRACTIONSTO ONE RHR STRAINERDebris TypeSize1 Train w/ Spray2 Train w/ Spray1 Train w/o Spray2 Train w/oSprayInteramFines, Smalls58%29%47%23%UnqualifiedEpoxyParticulate58%29%47%23%UnqualifiedIOZParticulate58%29%16%8%Unqualified AlkydParticulate58%29%100%50%Qualified EpoxyParticulate58%29%23%12%Qualified IOZParticulate58%29%23%12%Latent dirt/dustParticulate58%29%28%14%

28CONTAINMENT SPRAY AND POOL pHCombining conditions in a non-physical way to bound release and precipitationContainment spray from the RWST will use a maximum pH for acidic conditions associated with the minimum RWST boron concentration.The design basis maximum containment pool pH is used for the sprays during recirculation and the containment pool to calculate chemical release. Lower pH values/profiles are used for aluminum solubility to account for lower TSP concentrations, higher boric acid concentrations, and pH effects due to core release and radiolysis.

29CHEMICAL EFFECTSOverviewChemical precipitate quantities are determined for each breakCorrosion/Dissolution ModelDissolution from insulation and concrete is determined using the WCAP-16530 equationsCorrosion, dissolution and passivation of aluminum metal is determined using the equations developed by Howe et al. (UNM)SolubilityNo credit will be taken for calcium solubilityANL solubility equation (ML091610696, Eq. 4) will be used to credit delayed aluminum precipitation Once temperature limit is reached, all aluminum will precipitate Aluminum is eventually "forced" to precipitate for all breaksPrecipitate SurrogatesWCAP-16530 Calcium PhosphateWCAP-16530 Sodium Aluminum Silicate 30MAXIMUM DEBRIS GENERATEDBounding quantities of Nukon, Interam and qualified coatings for DEGB in Loop 1&4 SG compartmentDebris typeQuantityNotesNukon2,229 ft3Including all size categoriesInteram40 lbm30% fiber and 70% particulateQualified coatings249 lbmIOZ and epoxyUnqualified coatings2,843 lbmIOZ, alkyd, and epoxyLatent fiber4 ft315% of total latent debris; 2.4 lbm/ft3Latent particulate51 lbm85% of total latentdebrisMiscellaneous debris2 ft2Total surface area of tape and labels 31DEBRIS QUANTITIES AT ONE RHR STRAINERDebris Type2009 Test QuantityBounding Hot Leg Break(two trains with CS)Bounding Cold Leg Break(2 trains,with CS)Bounding Cold Leg Break (single trainwith CS)Nukon106.1ft3337.3 ft3333.6ft3667.3ft3Latent fiber3.9 ft3*1.2 ft31.2 ft32.32 ft3Interam290.3 lbm*12.0 lbm17.4 lbm34.8 lbmQualified coatings696.8 lbm*27.3 lbm72.2 lbm144.4 lbmUnqualified coatings2874.1 lbm*824.5 lbm824.5 lbm1648.9 lbmLatent particulate52.7 lbm*14.8 lbm14.8 lbm29.6 lbmSodium aluminum silicate89.1 lbm~55 lbm~55 lbm~102 lbmCalcium phosphate52.8 lbm~53 lbm~53 lbm~108 lbm* These tested quantities exceed currently estimated values for all breaks under all equipment combinations at Vogtle 322009 STRAINER HEAD LOSS TESTINGTesting consistent with the NRC March 2008 GuidanceTank test with prototypical 7-disk strainer moduleTotal area of 69 ft2Walls and suction pipe arranged consistent with plant strainerBounding RHR strainer approach velocity for runout flow rate (4,500 gpm)1 inch submergence for vortex observationsAlionTest Facility 332009 TESTING DEBRIS LOADS Nukondebris quantity based on 7D ZOIChemical precipitates quantity from WCAP-16530The following debris surrogates usedNukon and latent fiber: NukonCoatings: Silicon carbide (4 to 20 micron)Latent particulate: Silica sand w/ size distribution consistent with NEI 04-07 Volume 2 (75 to 2000 microns)Interam fire barrier: Interam E-54A 342009 TEST PROCEDUREDebris introduction consistent with the NRC March 2008 GuidanceFor thin-bed testing, all particulate added first followed by small batches of fiber finesFor full-load testing, fiber and particulate mixture added in batches with constant particulate to fiber ratioChemical debris batched in lastHead loss allowed to stabilize after each chemical addition 352009 TEST RESULTSDebris LoadThin-bed test head loss (ft)Full-load test head loss (ft)Fiber + Particulate0.6315.462After calcium phosphate31.656.57After sodium aluminum silicate32.6011.80Note:1.Equivalent bed thickness of 0.625 inches, added in 5 fiber only batches, each 1/8" equivalent thickness2.Equivalent bed thickness of 1.913 inches, added in 4 batches, each 0.478" equivalent thickness3.Each chemical separately added in 3 equal batches 36APPLICATION OF 2009 RESULTSTotal conventional debris plus calcium phosphate head loss will be applied at the start of recirculationTotal aluminum precipitate head loss will be applied when temperature decreases to solubility limitHead loss will be scaled as a function of the average approach velocity and temperature based on the results of the flow sweeps performed at the end of the thin-bed and full-load testsResults will be extrapolated to 30 daysBreaks that exceed the maximum tested fiber quantity, particulate quantity, or chemical precipitate quantity will be assigned a failing head loss value 37STRAINER ACCEPTANCE CRITERIARHR and CS Pump NPSH MarginThe NPSH margin is calculated using break-specific water level and flow ratesMinimum NPSH margin is 16.6 ft at 210.96F and acontainment pressure of -0.3 psigStructuralStrainer stress analysis is based on a crush pressure of 24.0 ft for the RHR strainers and 23.0 ft for the CS strainersGas void2% void fraction at pump inlet 38FIBER PENETRATION TESTINGMultiple tank tests were performed at Alden in 2014 for various strainer approach velocities, number of strainer disks and boron / buffer concentrationsNukon prepared into fines per latest NEI GuidanceA curve fit of the test data will be used to evaluate maximum fiber penetration for a 16 disk strainer (RHR) and 14 disk strainer (CS).

3902004006008001000120014001600180005000100001500020000Cumulative Fiber Penetration (g)Cumulative Fiber Addition (g)FIBER PENETRATION RESULTS 40IN-VESSEL EFFECTS Transport/accumulation of fiber to the core that penetrates RHR strainers is dependent on break location and flow path and will be based on WCAP-17788Values for core blockage and boron precipitation acceptance criteria will be based on WCAP-17788 41EX-VESSEL EFFECTSA bounding evaluation performed for all componentsExisting evaluations will be updated in accordance with WCAP-16406-P-A,"Evaluation of Downstream Sump Debris Effects in Support of GSI-191" Revision 1 and the accompanying NRC SER 42LOCADMA bounding calculation performed to cover all scenariosExisting evaluations will be updated in accordance with WCAP-16793-A,"Evaluation of Long-Term Cooling Considering Particulate, Fibrous, and Chemical Debris in the Recirculating Fluid" Revision 2 and the accompanying NRC SER 43AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 44GSI-191 ACCEPTANCE CRITERIAAcceptanceCriteriaMethod for AddressingDebris exceeds limits for upstream blockage,e.g. refueling canal drainBoundinganalysis -part of transport calculationStrainer head loss exceeds pump NPSH margin or strainer structural marginBreak-specific analysis based on tested debris limits and maximum tested head lossesGas voids from degasification or flashingexceed strainer/pump limitsBreak-specific analysis based on head loss, pool temperature, etc.Pumps fail due to air intrusion from vortexingBounding analysisPenetrated debris exceeds ex-vessel wear and blockage limitsBounding analysisPenetrated debris exceeds in-vessel fuel blockage and boron precipitation limitsBreak-specific analysis based on penetration testing and flow splitsDebris accumulation on cladding prevents adequate heat transfer Bounding analysis 45OVERVIEW OF MODELINGGSI-191 phenomena evaluated in a holistic, time-dependent manner with an evaluation tool called NARWHALDeveloped using object oriented designTracks movement of water and debrisEvaluates each break with incremental time steps up to 30 daysEvaluates range of breaks to determine which ones pass/fail the strainer and core acceptance criteria Uses top-down LOCA frequencies to calculate conditional failure probabilities for small, medium, and large breaks 46OVERVIEW OF MODELINGNARWHAL prototype has been used for preliminary risk-informed evaluations by several plantsVersion 1.0 is currently under development Software requirements, design, implementation, V&V, and user documentation prepared under ENERCON's Appendix B QA programNARWHAL's object oriented design allows users to model unique plant-specific geometry with a common executable fileSimplifies software maintenance Significantly reduces potential for software errors 47OVERVIEW OF MODELINGRWSTReactorVesselBreakCoreECCSCSSpray NozzlesContainmentCompartmentsSump PoolECCS Strainers 48OVERVIEW OF MODELING 49OVERVIEW OF MODELINGNARWHAL integrates models and inputs from design basis calculations and GSI-191 tests/analyses:Water volumes, temperature profiles, and pH from design basis calculationsDebris quantities from BADGER database Location-specific transport fractions from transport calculationChemical effects calculated using WCAP-16530 and Howe modelConventional and chemical head loss from 2009 strainer testing (scaled for temperature and flow)NPSH and structural margin from design basis calculationsDegasification calculated using standard physical modelsTime-dependent penetration using data fit from 2014 strainer testingCore failure calculated using WCAP-17788 model and limits 50OVERVIEW OF MODELING1. Select unique break location, size, and orientation2. ZOI debris quantities from BADGER (insulation and qualified coatings)3. Debris quantities outside ZOI (unqualified coatings, latent, miscellaneous)4. Blowdown transport to containment compartments and sump pool5. Washdown transport from containment compartments to sump pool6. Pool fill transport from sump pool to strainers and inactive cavities7. Recirculation transport from sump pool to strainers8. Debris penetration through strainers9. Debris accumulation on core 51OVERVIEW OF MODELING2. Corrosion/ dissolution of metals, concrete, and debris in sump pool1. Corrosion/ dissolution of metals, concrete, and debris by CS4. Formation of chemical precipitates3. Precipitate solubility limit5. Recirculation transport from sump pool to strainers6. Debris penetration through strainers7. Debris accumulation on core 52OVERVIEW OF MODELING1. Strainer Head Loss (CSHL + Conventional HL + Chemical HL)2. Degasification gas void fraction3. Pump NPSH margin4. Strainer structural margin5a. Does HL Exceed NPSH or structural margin?Yes7a. Fail strainer criteriaNo6a. Pass strainer criteria5b. Does void fraction exceed strainer or pump limits?Yes7b. Fail strainer criteriaNo6b. Pass strainer criteria 53OVERVIEW OF MODELING1. Core debris accumulation2b. Does quantity exceed blockage limits?Yes3. Fail core criteriaNo4. Pass core criteria2a. Does quantity exceed boron limits?YesNo 54OVERVIEW OF MODELINGAnalytical results showing whether a given break passes or fails are highly dependent on the assumptions and models used to evaluate the breakFor example, if sprays are not initiated for a given break: A smaller fraction of debris is washed down to the containment pool Corrosion/dissolution is reduced (unsubmerged materials)A larger fraction of debris in the pool is transported to the RHR strainersIt is not always obvious what conditions are bounding 55INPUTSInputValuesSensitivityLOCA FrequencyBased on NUREG-1829/PRANeedto address uncertaintyDebris GenerationBased on consensus modelsNoneDebris TransportBased on consensus modelsNoneContainment TempDesign basisNonePool TempDesign basisNoneContainment SprayActivation and DurationCompeting effects (e.g. washdown, strainer surface area, corrosion)Needto run for activation and durationPool Volume/LevelBased on consensus modelsNonePoolpHDesign Basis maximum for corrosion, minimumfor precipitationNoneECCS FlowRatesDesign basis (same for all breaks)NoneAluminum and Calcium CorrosionWCAP-16530 and UNM equationsNoneAluminum PrecipitationANL equationNonePenetration2014 testingNeedto address uncertaintyHead LossMaximums Extrapolated and scaled from 2009 tests final valuesNone 56DIFFERENCES (STPVS. VOGTLE)Physical ECCS trains (3 vs 2)Strainer configuration (3 combined vs 4 separate)Containment Spray SetpointStrainer design (flow control vs no flow control)Strainer surface area (~1800 ft2vs 765 ft2) 57DIFFERENCES (STPVS. VOGTLE)ModelingRisk Informed Software (CASA Grande/RUFF/FiDOEvs BADGER/NARWHAL)Break size and orientation sampling (Search algorithm vs uniform)Strainer head loss test protocol (flume vs tank)Aluminum and Calcium precipitate quantity (bounding test vs break-specific analysis)Aluminum passivation (not credited vs credited using Howe paper)Core blockage (RELAP5-3D vs WCAP-17788)Failure criteria (bounding analysis vs time-dependent comparison of head loss with NPSH, gas void, and flashing) GSI-191 risk quantification (critical break size frequency vs conditional failure probability entered into PRA) 58AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 59INTERFACE WITH PRAThe change in core damage frequency (CDF) and change in large early release frequency (LERF) due to issues related to GSI-191 will be determined using:LOCA frequencies from NUREG-1829 or the Vogtle PRAEquipment configuration logic from PRA model of recordGSI-191 conditional failure probabilities from NARWHALFinal risk calculation will be performed using the Vogtle model of record with GSI-191 conditional failure probabilities 60INTERFACE WITH PRA1. PRA identification of accident scenarios and equipment configurations that are risk-significant to GSI-1912. NARWHAL evaluation of GSI-191 phenomena for each risk-significant scenario/configuration to determine CFPs for each strainer/core failure basic event3. PRA quantification of CDF & LERF with GSI-191 failures4. PRA quantification of CDF & LERF with no GSI-191 failuresOutput to submittal documentation5. Calculate CDF & LERFYes6. Meets RG 1.174 Criteria?7a. Identification of analytical refinements to analysisNo7b. Identification of potential plant modifications (physical or procedural) 61METHODOLOGY FOR USING LOCA FREQUENCY TO CALCULATE CONDITIONAL FAILURE PROBABILITYPartition PRA categories (Catk) into size ranges (SRi)Calculate probability of a LOCA occurring in each size range for every PRA category (P(SRilCatk))Calculate probability of a LOCA occurring at each weld (Weldj) within each size range (P(WeldjlSRi))Calculate probability estimate for success criteria failure (SCF) at each weld in each size range (P(SCFlWeldj,SRi))Calculate probability estimate for success criteria failure at every PRA category (P(SCFlCatk))

62METHODOLOGYFOR USING LOCA FREQUENCY TO CALCULATE CONDITIONAL FAILURE PROBABILITYPSCFCat=PSRlCatPWeldSRPSCFWeld,SRFrom NUREG-1829 (or Vogtle PRA)From NUREG-1829 (or Vogtle PRA) using top-down methodologyFrom NARWHAL 63ILLUSTRATION OF PRA CATEGORIES AND LOCA FREQUENCY SIZE RANGES 64EQUIVALENT DEGB DIAMETERNUREG-1829,Section3.7:"ThecorrelationswhichrelateflowratesandLOCAsizecategoriestotheeffectivebreaksizesineachPWRandBWRsystemaresummarizedinTable3.8.Thebreaksizecorrespondstoapartialfractureforpipeswithlargerdiametersthanthebreaksize,acompletesingle-endedruptureinpipeswiththesameinsidediameter,oraDEGBinpipeshavinginsidediameters12timesthebreaksize.AllpanelistsusedthesecorrelationstorelatetheirelicitationresponsesdeterminedfortheeffectivebreaksizestotheappropriateLOCAsizecategory.ItisimportanttostressthatbreakscanoccurineitherLOCAsensitivitypipingornon-pipingsystemsandcomponents."

65METHOD FOR ADDRESSING DEGB FREQUENCIESNUREG-1829 provides LOCA frequencies vs. break flow rates, which are converted to equivalent diameter break sizesDouble ended guillotine breaks (DEGBs) have a flow rate (area) twice as large as the pipe cross-sectional areaEquivalent diameter for DEBG is calculated using pipe inner diameter: DDEGB,EqDpipeDpipeassumed to progress to DEGB for debris generation (bounding debris quantities)Frequency associated with DEGB "tail" assigned to Dpipebreak size to calculate conditional failure probability within each size rangeNUREG-1829 only provides frequencies up to 31" breaks, so frequencies corresponding to primary loop pipe DDEGB,Eqvalues is determined through log-log extrapolation29" 41.0"31"43.8" 66METHOD FOR ADDRESSING DEGB FREQUENCIES 67EQUIPMENT CONFIGURATIONMost likely situation would be that all equipment is available and fully functionalEquipment failures due to non-GSI-191 related issues can have a major effect on GSI-191 phenomena (debris transport, flow splits, temperature and pressure profiles, etc.) There are many possible equipment failure combinations (RHR pumps, containment spray pumps, charging pumps, SI pumps, fan coolers, etc.)At Vogtle, GSI-191 effects can be reasonably represented or bounded for most equipment failure combinations by the cases where all pumps are running or a single train failure 68EQUIPMENT CONFIGURATIONLBLOCAAll equipment available bounding or reasonably representative for: No pump failures: ~79%1 or 2 CS pump failures: ~18%1 RHR pump failure bounding or reasonably representative for other equipment failure configurations: ~3%The values above are being used to determine which configurations to investigate with NARWHALConditional failure probabilities for GSI-191 basic events (strainer and core failure) will be manually entered into the PRA model of record to calculate CDF~97%

69CONDITIONAL FAILURE PROBABILITIESCoreconditional failure probability (CFP) higher when all pumps are operatingStrainerCFP higher with 1 train operatingAdditional scenarios will likely need to be evaluated (1 CS pump failure, 1 RHR pump failure, etc.)

70PRA MODEL CHANGESVogtle PRA model modified to incorporate conditional failure probabilities (CFPs) for GSI-191 strainer failures and core failures with associated initiating events and equipment configurationsGSI-191 "failure" defined using PRA success criteriaStrainer failure defined as failure of RHR pump/strainerCS strainer failures are calculated in NARWHAL, but not used in PRAGSI-191 failures binned into following CFP groups (separate CFPs defined for each LOCA category and each pump state)Core failuresRHR Strainer A failures (without Strainer B or core failures)RHR Strainer B failures (without Strainer A or core failures)RHR Strainer A and B failures (without core failures)Each core and strainer CFP represented in PRA with a basic event that is combined with LOCA initiating event and pump failure logic to represent pump state associated with CFP 71PRA MODEL CHANGES 72AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 73UNCERTAINTY QUANTIFICATIONDraft RegGuide 1.229 requires uncertainty quantification for risk-informed GSI-191 evaluationUncertainty quantification also required by RegGuide 1.174Two different approaches that could be used to quantify uncertaintySimplified approach using sensitivity analysisStatisticalsampling of input parameter distributions and propagating uncertainties Consensus inputsand models are considered to have no uncertaintyVogtle inputs that require uncertainty quantificationLOCA frequency valuesPenetration modelContainmentspray activation and durationPossibly others 74UNCERTAINTY QUANTIFICATION USING SENSITIVITY ANALYSIS 75UNCERTAINTY QUANTIFICATION USING STATISTICAL SAMPLING1. Select equipment failure configuration2. Sample all random input values3. Select break location, size, and orientation5. Compare to GSI-191 acceptance criteria and PRA success criteria to identify strainer or core failures9. Output to PRA model4. Evaluate GSI-191 phenomena at each time step for 30 days6. Sufficientbreaks evaluated to estimate CFP?No7. Sufficient random samples to estimate CFP PDF?YesNo8. All significant equipment configs. evaluated?YesNoYes 76AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 77SUBMITTAL DOCUMENTATION DISCUSSIONPlanning to follow the format, content, and depth of the August 2015 STP LARGL 2004-02 Response following Staff Review Guidance (will have some pointers to risk informed summary)Risk Quantification and SummaryDefense in Depth and safety marginRequests for exemptions (if 10CFR50.46(c) rule is not finalized)10CFR50.46(d)GDC 19 (need being evaluated)GDC 35GDC 38GDC 41GDC 50.67 (need being evaluated)License amendment requestTech Spec markupsTech Spec BasesFSAR markupsList of commitments 78QUALITY ASSURANCEGSI-191 calculations are being revised to be safety related under vendor Appendix B QA programsHead loss testing in 2009 was conducted and documented as safety related under vendor Appendix B QA program Penetration testing was conducted and documented in 2014 as non-safety related following work practices established by vendor Appendix B QA programPRA has undergone an industry peer review per RG 1.200 and the ASME/ANS PRA Standard for CDF and LERF (RA-Sa-2009) 79CONCLUSIONSModels used to analyze GSI-191 phenomena at Vogtle are consistent with methods accepted for design basis evaluationsPreliminary results indicate that risk associated with GSI-191 is very low 80SCHEDULE UPDATEMILESTONEEXPECTED COMPLETION DATECurrent StatusDevelop containment CAD model to include pipe weldsCompleteCompleteConduct meeting with NRC3rdQuarter 2013CompleteModify probabilistic risk assessment (PRA) to include Strainer and Core Blockage events4thQuarter 2013CompletePerform Chemical Effects testing1stQuarter 2014CompletePerform thermal and hydraulic modeling of RCS, Core, and Containment conditions1stQuarter 2014CompletePerform Strainer Head Loss and Bypass testing to establish correlation for range of break sizes2ndQuarter 2014StrainerBypass Testing -CompleteStrainer Headloss Testing -Using 2009 Vogtle testAssemble base inputs for CASA Grande2ndQuarter 2014Complete -Using NARWHAL instead of CASAGrandeEvaluate Boric Acid Precipitation impacts3rdQuarter 2015Plan on using PWROG WCAP-17788Finalize inputs to CASA Grande3rdQuarter 2015In Progress, Using NARWHAL instead of CASAGrandeComplete Sensitivity Analyses in/for CASA Grande4thQuarter 20152ndQuarter 2016Integrate CASA Grande results into PRA to 1stQuarter 20163rdQuarter 2016Licensing Submittal for VEGPTo be established through discussions with NRC -tentatively September 2016Projected 1stQuarter 2017, require input from: SE on STP Pilot Project and SE on WCAP-17788; which are not yet available.

81AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 82BACKUP SLIDES 83EXAMPLE CALCULATIONDetailed physical calculationsBreak Location 11201-049-16-RB (10.5" DEG hot leg side break in annulus near surge line)1 train operating with containment sprays activatedTransport fractions: MaximumRHR flow rate: 3,700 gpm(design flow)CS duration: 6 HoursDesign basis temperature profile -Max Safeguards, with pool temperature drop to 90°F at Day 30Minimum water level profileSump pH used for corrosion: 8.1 Containment spray pH during Injection used for corrosion: 4.71pH used for solubility: 7.6Submerged Aluminum Area: 278.7 ft2Unsubmerged Aluminum Area: 741.3 ft2Submerged Concrete: 2,813 ft2Aluminum Metal Release equations -WCAP 16530 Aluminum Solubility Equation -ANL Aluminum Solubility EquationAluminum Solubility: Timing Only 84BREAK LOCATION 85POOL LEVEL 86FIBER TRANSPORT TO ONE RHR STRAINER (1 TRAIN W/SPRAY)*Transport fraction takes into account time-dependent transport (CS strainer active for 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months )**Combined fraction of fines due to erosion and intact piecesDebris TypeSizeDG Quantity (ft3)TransportFraction*Quantity(ft3)NukonFines9.558%6.1Small32.77%**1.5**Large14.84%**0.6**Intact15.90%0.0Total72.98.2LatentFines12.558%7.8Total85.416.0 87PARTICULATE TRANSPORT TO ONE RHR STRAINER (1 TRAIN W/SPRAY)Debris TypeSizeDG Quantity (lbm)TransportFractionQuantity(lbm)InteramFines, Smalls121.258%70.3UnqualifiedEpoxyParticulate2,602.058%1,509.2UnqualifiedIOZParticulate25.058%14.5Unqualified AlkydParticulate32.058%18.6Qualified EpoxyParticulate7.558%4.4Qualified IOZParticulate2.458%1.4Latent dirt/dustParticulate170.058%98.6Total2,960.11,717.0 88CHEMICAL PRECIPITATE QUANTITIES 89ALUMINUM CONCENTRATION AND SOLUBILITY 90HEAD LOSSClean strainer head loss: Bounding value of 0.162 ftat 4,500 gpmConventional head loss Head loss due to chemical precipitates* Values scaled to 16 disk RHR strainer area** Corrected to 3700 gpmand 90°F*** Applied at switchover to hot leg recirculationHead Loss TypeTransported QuantityMaxTested Quantity*Head LossCorresponding to Max Tested Quantity(Unadjusted)Head Losswith Flow and TemperatureCorrection**Fiber16.0ft3109.9ft35.46 ft4.41 ftParticulate1,717.0 lbm3,914.5 lbmCalcium Phosphate9.6 lbm52.8 lbm1.11 ft0.90ftSodium Aluminum Silicate74.5 lbm89.0 lbm5.24 ft4.23 ftExtrapolation to30 days2.13 ft***1.72 ft 91Short Term: Long Term: VOID FRACTION AT RHRPUMP 92DEBRIS TRACKING100% capture of particulate and precipitate at strainerPenetration modeled based on test data curve fitsAll fiber is treated as fines 93CORE ACCUMULATION 94SUMMARYCriteriaAcceptanceCriteriaExample ResultsExamplePass/FailNPSH Margin16.6 ft11.3 ftPassStrainer Structural Margin24.7 ft11.3 ftPassPartial SubmergenceHead loss lessthan 1/2 submerged heightFully submergedPassGas Void Fraction2% at the pump0.24% at the pumpPassDebris LimitExceeds what was testedNot exceededPassAssumed core limits (to be replaced with WCAP-17788 criteria)75 g/FA forHot Leg Breaks7.5 g/FA for Cold Leg Breaks32 g/FAN/APass NOVEMBER 5, 2015VOGTLE GSI-191 RESOLUTION PLAN AND CURRENT STATUSNRC PUBLIC MEETING 2AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 3PURPOSE OF MEETINGObtain staff feedback on the overall GSI-191 resolution path for VogtleDiscuss proposed plant modificationsDiscuss use of design basis (vs. best estimate) inputs in the evaluation 4VOGTLE PLANT LAYOUTWestinghouse 4-loop PWR (3,626 MWt per unit)Large dry containmentTwo redundant ECCS and CS trainsEach train has an RHR pump, a high head pump, an intermediate head SI pump, and a CS pumpSI and high head pumps piggyback off of the RHR pump discharge during recirculation.Maximum design flow rates:RHR 3,700 gpm/pumpCS 2,600 gpm/pumpTwo independent and redundant containment air cooling trains 5STRAINER ARRANGEMENTTwo RHR and CS pumps each with their own strainerEach GE strainer is similar with four stacks of disksRHR strainer: 18-disk tall, 765 ft2, 5 fttallCS strainer: 14-disk tall, 590 ft2, 4 fttallPerforated plate with 3/32" diameter holesRHR BCS BRHR ACS A 6PLANT RESPONSE TO LOCASPlant response includes the following general actions: Accumulators inject (breaks larger than 2 inches)ECCS injection is initiated from the RWST to the cold legs via RHR, SI, and High Head pumpsContainment spray is initiated from the RWST via CS pumps RHR pumps switched to cold leg recirculation at RWST lo-lo alarmCS pumps switched to recirculation at RWST empty alarmCS pumps secured no earlier than 1.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months after start of recirculation, and probably before 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months depending on pressure and dose rateRHR pumps switched to hot leg recirculation at 7.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months 7AGENDA8:30 -Introductions/Purpose of Meeting8:45 -High Level Overview of Vogtle Closure Strategy &What we have learned9:00 -Review of Inputs and Assumptions10:00 -Explanation of Modeling11:30 -Interface with PRA12:15 -Uncertainty Quantification12:45 -Lunch13:45 -Continued Discussion of Inputs and Assumptions14:30 -Continued Discussion of PRA and Uncertainty15:00 -Submittal Documentation15:45 -Closing Remarks 8SCHEDULE UPDATEMILESTONEEXPECTED COMPLETION DATECurrent StatusDevelop containment CAD model to include pipe weldsCompleteCompleteConduct meeting with NRC3rdQuarter 2013CompleteModify probabilistic risk assessment (PRA) to include Strainer and Core Blockage events4thQuarter 2013CompletePerform Chemical Effects testing1stQuarter 2014CompletePerform thermal and hydraulic modeling of RCS, Core, and Containment conditions1stQuarter 2014CompletePerform Strainer Head Loss and Bypass testing to establish correlation for range of break sizes2ndQuarter 2014StrainerBypass Testing -CompleteStrainer Headloss Testing -Using 2009 Vogtle testAssemble base inputs for CASA Grande2ndQuarter 2014Complete -Using NARWHAL instead of CASAGrandeEvaluate Boric Acid Precipitation impacts3rdQuarter 2015Plan on using PWROG WCAP-17788Finalize inputs to CASA Grande3rdQuarter 2015In Progress, Using NARWHAL instead of CASAGrandeComplete Sensitivity Analyses in/for CASA Grande4thQuarter 20152ndQuarter 2016Integrate CASA Grande results into PRA to 1stQuarter 20163rd Quarter 2016Licensing Submittal for VEGPTo be established through discussions with NRC -tentatively September 2016Projected 1stQuarter 2017, require input from: SE on STP Pilot Project and SE on WCAP-17788; which are not yet available.