ML070440491
| ML070440491 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 01/07/2007 |
| From: | Duke Energy Corp |
| To: | Office of Nuclear Reactor Regulation |
| References | |
| EA-06-294, IR-07-006 | |
| Download: ML070440491 (27) | |
Text
{{#Wiki_filter:" Duke Oconee Nuclear Station r_ En er g y Regulatory Conference Foreign Material Found in the Unit 3 Reactor Building Emergency Sump Suction Headers NRC Region II Office Atlanta, Georgia January 17, 2007 1
a Duke rtl Energy Duke Participants
- Bruce Hamilton, Oconee Site Vice President
- Mike Glover, Oconee Engineering Manager
- Larry Nicholson, Oconee Safety Assurance Manager
- Steve Capps, Oconee MCE Engineering Manager
- Jason Patterson, Oconee MCE Engineering Supervisor
- Steve Nader, Duke PRA Engineering Supervisor
- Mike Barrett, Duke PRA Principal Engineer
- Graham Davenport, Oconee Regulatory Compliance Manager
- Bob Meixell, Oconee Regulatory Compliance Engineer
- Corey Gray, Oconee Regulatory Compliance Engineer 2
. " Duke
,_Ene rgy Agenda
- Introductions
- Opening Remarks
- EeCS Pump and Suction Piping Overview
- Debris Transport Testing
- ECCS Room Temperature Analysis
- PRA Highlights and Results
- Closing Remarks 3
a Duke r_ Ene rgy Opening Remarks .* Oconee implemented an aggressive project to upgrade RB emergency sumps (GL 2004-02)
- Modifications included extensive internal piping inspections using specialized equipment (FOSAR)
- Legacy foreign material was discovered in both Unit 3 containment sump ECCS suction headers during a May 2006
.scheduled refueling outage
~ Discovered prior to installing new sump strainers
- Debris discovered previously in the Unit 2 sump, and during a subsequent shutdown on Unit 1, affected only one train 4
a Dulce
, _Energy Opening Remarks - cont.
- Units 1 and 2 debris not safety significant due to robust ECCS design and single affected train
- Unit 3 debris required further analysis due to discovery in both suction headers
- The presence of foreign material in Oconee systems does not meet Duke expectations
- This issue is not indicative of current Oconee performance
~ FME programs have been improved and strengthened ~ Greater emphasis placed on individual FME accountability
- Duke's evaluation of the Unit 3 condition shows it to be of low safety significance (Green) 5
a Duke Unit 3 ECCS Pump Room and rtllEnergy Suction Piping Arrangement 12" U3 DECAY HEAT SUCTION / FROM R.C. SYSTEM "3 8" '3C " "3A"
- , LOW PRESSURE" - LOW PRESSURE LOW PRESSURE INJECTION PUMP INJECTION PUMP INJECTION PUMP C AND MOTOR AND MOTOR AND MOTOR R.B.
14" M.O. G.V. SPRAY 3LP-8 l - PUMP "3B",
-:>~~
14'M.O.G.V. I i V -7 i L- -------T+T-----,
~ ~- ~ffr: l L ,""
REACTOR BLDG I ' N_ EMERGENCY SUMP 1';-; U : 18" UNIT 3 EMERGENCY SUMP RECIRC. liNE
~~~f4=t=1~~~~~~:=~~~~~~===:l~:~
. I 1% SLOPE . r-
-4 ~
+<::""-f~::=
,I 1'6"* 14"ECC.OEO. V/>J..VE ROOM c:
I 3° 3 1. 33'
------ 't BEACIOB-lll.OO'----l--'----l-+-+-'---------------~ .s! t--t+-t-- l----II-t---+---t--------1 -++
\)~~~§f!it~~~;_tt:-::~5~~'--t---_::"B::._l"TRAJN !
3° 58 ,25' 18" UNIT 3' EMERGENCY SUMP RECIRC. liNE 4'-1 71ll" I 26'-{) lYI6"
A Duke r .Ene rgy Overview II Duke analysis methodology similar to the NRC SDP Phase III, using many of the same inputs II Additional details incorporated in the event tree:
~ Transport testing results ~ Room heat-up analysis results ~ Plant specific reliability and unavailability data 7
a Duke 3A ECCS Header: Input Differences
' . Ene rgy Summary NRC Duke Flat washer passes through 3LP 100% Same Washer Transport into 3A LPI Pump 100% Washer Transport into 3A LPI Pump based on flow split with 3A RBS Pump 3A LPI Pump Failure 100% when washer Same enters pump Seal Failure assumed to be 30% when Same washer enters pump 8
,tI1Energy " Duke 3B ECCS Header: Input Differences Summary NRC Duke Wrench Jaw Passes through 3LP-20 100% Same (LB LOCA)
Wrench Jaw always passes to 3B LPI Pump Same 3B LPI Pump Failure 100% when wrench Same jaw enters pump Seal Failure probability 30% when wrench Same jaw enters pump Wrench Jaw passes through 3LP-20 100% Wrench Jaw passes through 3LP-20 25% for for MB LOCA with both LPI trains in MB LOCA with both LPI trains in service service Wrench Jaw passes through 3LP-20 100% Wrench Jaw passes through 3LP-20 45% for for MB LOCA with one LPI train in service MB LOCA with one LPI train in service 9
,til .a Duke Energy Common Input Differences Summary NRC Duke Unavailability for LPI Pumps assumed to be Unavailability for LPI Pumps assumed to be 2% 1.2% for 3A & 3B; 0.26% for 3C based on plant records MB and LB LOCA Effects Combined MB and LB LOCA Effects Separated 10
a Duke
,111Energy Testing / Analysis Overview
- Impact of debris on nuclear safety is contingent on transport
- Flow testing preferable to computational fluid dynamics analysis to determine transport probability
- Transport dependent on flow rate and piping configuration
- Room heat-up analysis performed to determine ambient temperature effects from postulated ECCS pump seal failure
>- Analysis model confirmed by plant data 11
a Duke r_Ene rgy Debris Transport Testing
- Alden Research Laboratory conducted debris transport testing at various flow rates on a wrench jaw and flat washer
- Testing was conducted between November 16,2006 and December 4, 2006
- Test loop built to model physical arrangement of portions of Unit 3 ECCS suction piping
~ Included wedge-shaped recess to represent horizontal guide for suction isolation gate valves 3LP-19 & 20 12
a r_ Ene rgy Duke Test Loop Model "3B" "3C" "3A" r LOW PRESSURE " LOW PRESSURE LOW PRESSURE
. INJECTION PUMP INJECTION PUMP INJECTION PUMP
. .' AND MOTOR .... AND MOTOR AND MOTOR 14" M.O. G.V.
A.B. SPRAY ~ i ~ .
... ~-8 ) PUMP "3B"
"'C~
II r I MODEL now ___ :'_. .~ .*. -l _ _ _ _ _ <;: BEACIO.a.aI..QG,_ _-L~~~Wj;:...====l---------- t
~r
" Duke r_ Energy Debris Transport Testing - cont.
- Debris chosen for testing (adjustable wrench jaw and flat washer) consistent in size, shape, and mass of recovered objects
- Flow was varied corresponding to postulated LB and MB LOCA conditions
- Results for the 2 %" flat washer transport tests:
~ Washer transport initiated between 2500 and 3200 GPM ~ 9 flow tests performed; washer passed through valve 4 times ~ Analysis does not credit capture of washer in valve guide 14
a
,tilEnergy Duke Debris Transport Testing - cont.
- Results for the adjustable wrench jaw transport tests:
y Jaw transport initiated at flows above 2800 GPM y 25 flow tests performed at ~ 3000 GPM; jaw passed through valve 2 times y 20 flow tests performed at ~ 3500 GPM, jaw passed through valve 9 times y Analysis does credit capture of wrench jaw in valve guide 15
Containment Sump Temperature Analysis
- For cases in which a pump seal leak is assumed, LOCA accident conditions directly affect ECCS room temperatures once on RBES recirculation
~ Break size / location determines temperature of RBES contents
- Containment sump temperature analysis performed to determine sump temperature at the time of debris transport for various LOCA sizes
- For MB LOCAs, ECCS sump fluid temperatures are not high enough to lead to steaming in the ECCS rooms concurrent with an assumed pump seal leak 16
" Duke r_ Energy ECCS Room Heat-up Analysis
- Engineering firm Hawks, Giffels & Pullin (HGP), Inc. analyzed room temperature profiles following postulated ECCS pump seal failure
- Analysis utilized R T3 software to calculate peak room temperature conditions - extensive HGP and Duke experience
- For MB LOCAs, concluded that ECCS room atmospheric conditions will not result in failure of other available LPI pumps and associated equipment for any postulated mechanical seal leak
- For LB LOCAs, concluded that ECCS room atmospheric conditions will not result in failure of other available LPI pumps and associated equipment for postulated mechanical seal leaks up to 25 gpm 17
A Duke PRA Highlights: r_Ene rgy Methods and Key Assumptions 3A ECCS Header
- The flat washer is assumed to pass through valve 3LP-19 100%
of the time
- Washer transports into suction of3A LPI Pump based on flow split with 3A RBS Pump (different from NRC)
- 3A LPI Pump failure probability assumed to be 100% when washer enters pump
- Seal failure probability assumed to be 30% when washer enters pump 18
a Duke r_Ene rgy PRA Highlights: Methods and Key Assumptions - cont. 3B ECCS Header
- The wrench jaw is assumed to pass through valve 3LP-20 100%
of time for LB LOCA
- The wrench jaw always transports to the 3B LPI pump
- 3B LPI Pump failure probability assumed to be 100% when wrench jaw enters pump
- Seal failure probability assumed to be 30% when wrench jaw enters pump 19
a Duke PRA Highlights:
,"Ene rgy Methods and Key Assumptions - cont.
- The wrench jaw is assumed to pass through valve 3LP-20 25%
of the time for MB LOCA events with both LPI trains in service (different from NRC)
~ Near the middle of the range seen in the flow tests (8% - 45%) ~ Intended to account for RBES flows that are not constant but increasing at some rate depending on break size
- The wrench jaw is assumed to pass through valve 3LP-20 45%
of the time for MB LOCA events with 1 LPI train in service (different from NRC)
~ Higher RBES flows for single train in service biased result towards high flow case 20
a Duke r_Ene rgy PRA Highlights: Methods and Key Assumptions - cont.
- Pump Seal Failure Consequences - MB LOCA events:
~ Analyses show that the sump temperature is low enough to preclude adverse environmental impact ~ Flooding occurs only when LPIP sump system failure occurs, lE-02 failure probability assumed 21
a Duke r_ Ene rgy PRA Highlights: Methods and Key Assumptions - cont.
- Pump Seal Failure Consequences - Large LOCA
~ Adjacent pump can experience consequential failure if the pump that ingests debris has a seal failure AND
.:. There is enough flashing fluid hot enough to result in unfavorable environmental conditions, OR
.:. The leaking fluid causes a flooding concern Seal Flow Rate Sump Temp < 211 0 F Sump Temp> 211 0 F
< 25 GPM
>25 GPM (40 GPMMAX) 22
a Duke r_Energy PRA Highlights: Methods and Key Assumptions - cont.
- Medium LOCA pipe break initiator from Oconee PRA is 3.6E-05/year
- Large LOCA initiator from Oconee PRA is 4.5E-06/year
- Initiating events consistent with NUREG/CR-5750 y Adjusted for the Oconee capacity factor applied to the Oconee PRA initiating event frequencies (0.9 - different from NRC)
- Plant specific data for LPI Unavailability used (different from NRC) y 1.2% unavailability for each 3A & 3B LPI Train y 0.26% unavailability for the 3C LPI Train y Based on actual historical data from MR program 23
a Duke ,_Ene rgy PRA Results: Conclusions
- Risk impact is low (Green).
>- ~CDF is 2.7E-07/year for Medium LOCA cases r--;>> >- ~CDF is 6.0E-07/year for Large LOCA cases r-..J >- Total ~CDF is 8.7E-07/year
- Dominant risk contribution is from LB LOCA events.
- Duke risk assessment contains several conservative assumptions.
24
a Duke r_Energy PRA Results: Summary of-Conservatisms
- Conservative Risk Analysis Assumptions:
y Pump failures are assumed to occur 100% of the time debris is ingested y The washer always passes through 3LP-19 for all LOCAs y The wrench jaw always passes through 3LP-20 for large LOCAs y No credit taken for recovering LPI from maintenance. 25
A Duke PRA Results: r_Energy Summary of Conservatisms - cant.
- Conservative Risk Analysis Assumptions (cont.):
y LB LOCA failure probability of 0.5 for pump failure
.:. Some LB LOCA events will have sump temperatures that will not be high enough to cause flashing in the ECCS pump rooms
.:. Some LB LOCA events will result in seal leaks < 25 gpm which, as stated with the MB LOCA events, do not result in an adverse environment in the ECCS pump rooms
.:. No credit taken for forced ventilation or realistic initial room temperatures 26
" Duke r_Ene rgy Closing Remarks
- The Unit 3 debris issue is of very low safety significance based on L\CDP of 8.7E-07.
- Duke clearly recognizes the vital safety function performed by the ECCS sump suction.
- The conditions resulting in this finding do not meet Duke expectations
~ Improvements have been made and will continue to be made with FME processes and procedures.
27}}