ML13205A016: Difference between revisions

From kanterella
Jump to navigation Jump to search
(Created page by program invented by StriderTol)
(Created page by program invented by StriderTol)
 
(2 intermediate revisions by the same user not shown)
Line 1: Line 1:
{{Adams
#REDIRECT [[BSEP 13-0070, Response to Request for Additional Information Regarding Voluntary Risk Initiative National Fire Protection Association Standard 805]]
| number = ML13205A016
| issue date = 07/15/2013
| title = Brunswick, Units 1 and 2, Response to Request for Additional Information Regarding Voluntary Risk Initiative National Fire Protection Association Standard 805
| author name = Hamrick G T
| author affiliation = Duke Energy Progress, Inc
| addressee name =
| addressee affiliation = NRC/Document Control Desk, NRC/NRR
| docket = 05000324, 05000325
| license number =
| contact person =
| case reference number = BSEP 13-0070, TAC ME9623, TAC ME9624
| document type = Letter
| page count = 47
| project = TAC:ME9623, TAC:ME9624
| stage = Response to RAI
}}
 
=Text=
{{#Wiki_filter:SDUKE George T. HamrickVice President ENERGY, Brunswick Nuclear PlantDuke Energy ProgressP.O. Box 10429Southport, NC 28461o: 910.457.3698 July 15, 2013Serial: BSEP 13-0070U.S. Nuclear Regulatory Commission ATTN: Document Control DeskWashington, DC 20555-0001
 
==Subject:==
 
Brunswick Steam Electric Plant, Unit Nos. 1 and 2Docket Nos. 50-325 and 50-324Response to Request for Additional Information Regarding Voluntary RiskInitiative National Fire Protection Association Standard 805 (NRC TACNos. ME9623 and ME9624)
 
==References:==
: 1. Letter from Michael J. Annacone (Carolina Power & Light Company) to U.S.Nuclear Regulatory Commission (Serial:
BSEP 12-0106),
LicenseAmendment Request to Adopt NFPA 805 Performance-Based Standard forFire Protection for Light Water Reactor Electric Generating Plants (2001Edition),
dated September 25, 2012, ADAMS Accession NumberML12285A428
: 2. Letter from Michael J. Annacone (Carolina Power & Light Company) to U.S.Nuclear Regulatory Commission (Serial:
BSEP 12-0140),
Additional Information Supporting License Amendment Request to Adopt NFPA 805Performance-Based Standard for Fire Protection for Light Water ReactorElectric Generating Plants (2001 Edition),
dated December 17, 2012, ADAMSAccession Number ML12362A284
: 3. Letter from Christopher Gratton (USNRC) to Michael J. Annacone (Carolina Power & Light Company),
Request for Additional Information Regarding Voluntary Risk Initiative National Fire Protection Association Standard 805(TAC Nos. ME9623 and ME9624),
dated May 15, 2013, ADAMS Accession Number ML13123A231 Ladies and Gentlemen:
By letter dated September 25, 2012 (i.e., Reference 1), as supplemented by letter datedDecember 17, 2012 (i.e., Reference 2), Duke Energy Progress, Inc., formerly known as CarolinaPower & Light Company (CP&L), submitted a license amendment request to adopt a new risk-informed performance-based (RI-PB) fire protection licensing basis for the Brunswick SteamElectric Plant (BSEP), Unit Nos. 1 and 2.During the week of April 8 through April 12, 2013, the NRC conducted an audit at BSEP tosupport development of questions regarding the license amendment request.
On May 15, 2013(i.e., Reference 3), the NRC provided a set of requests for additional information (RAIs)regarding the license amendment request.
That letter divided these RAIs into 60-day, 90-day,www.duke-energy.com U.S. Nuclear Regulatory Commission Page 2 of 3and 120-day responses.
In subsequent telephone calls with the NRC Project Manager forBSEP, the following modifications were agreed to regarding the RAI response schedule shownin the May 15, 2013, letter:" The 60-day RAI responses will be submitted by July 1, 2013 (i.e., 60 days following theMay 2, 2013, clarification call that was conducted with the NRC). These responses wereactually submitted by letter dated June 28, 2013. Probabilistic Risk Assessment (PRA)RAls 1A, 1B, 1C, 1D, 1F, 1G, 11, 1K, 1N, 10, 1P, 1Q, 1R, 4,5,9,10,17, and 18, whichwere included in the set of 60-day RAls, will be addressed in a separate submittal dueby July 15, 2013 (i.e., 60 days following the date of the letter).* The 90-day RAI responses will be submitted by July 31, 2013 (i.e., 90 days following theMay 2, 2013, clarification call). Fire Protection Engineering RAI 1, which was included inthe set of 60-day RAls, will be addressed as part of the 90-day RAI responses.
* The 120-day RAI responses will be submitted by August 30, 2013 (i.e., 120 daysfollowing the May 2, 2013, clarification call). PRA RAI 1 H will be addressed as part of the120-day RAI responses, rather than with the 60-day RAI responses.
The response to RAI 1 K is not being provided with the enclosed 60-day RAI responses.
Theresponse to RAI 1 K will be provided as part of the 90-day RAI responses which will besubmitted by July 31, 2013.A tabulation of the individual RAls and the planned response submittal dates is provided inEnclosure
: 1. Duke Energy's responses to the set of 60-day PRA RAls (i.e., due by July 15,2013) are provided in Enclosure 2.This document contains no new regulatory commitments.
Please refer any questions regarding this submittal to Mr. Lee Grzeck, Manager -Regulatory
: Affairs, at (910) 457-2487.
I declare, under penalty of perjury, that the foregoing is true and correct.
Executed on July 15,2013, 2013.Sincerely, George T. Hamrick
 
==Enclosures:==
: 1. Revised Response Schedule to NFPA 805 Request for Additional Information
: 2. Response to Request for Additional Information Regarding Voluntary Risk Initiative National Fire Protection Association Standard 805 U.S. Nuclear Regulatory Commission Page 3 of 3WRM/wrmcc (with enclosures):
U. S. Nuclear Regulatory Commission, Region IIATTN: Mr. Victor M. McCree, Regional Administrator 245 Peachtree Center Ave, NE, Suite 1200Atlanta, GA 30303-1257 U. S. Nuclear Regulatory Commission ATTN: Mr. Christopher Gratton (Mail Stop OWFN 8G9A)11555 Rockville PikeRockville, MD 20852-2738 U. S. Nuclear Regulatory Commission ATTN: Ms. Michelle P. Catts, NRC Senior Resident Inspector 8470 River RoadSouthport, NC 28461-8869 Chair -North Carolina Utilities Commission P.O. Box 29510Raleigh, NC 27626-0510 Mr. W. Lee Cox, Ill, Section Chief (Electronic Copy Only)Radiation Protection SectionNorth Carolina Department of Health and Human Services1645 Mail Service CenterRaleigh, NC 27699-1645 lee.cox@dhhs.nc.gov Enclosure 1Page 1 of IRevised Response Schedule to NFPA 805 Request for Additional Information Revised Response ScheduleSection Title Question Number(s)
Submittal Date60-Day Response
-Non-PRAProgrammatic 1, 2, 3, 4, 5, 6, 7Safe Shutdown Analysis 3, 4, 6, 7, 8,10,12 July 1, 2013Fire Modeling 1A, 1E, 1F, 1G, 1H, 2A, 2B, 5A, 5B60-Day Response
-PRAProbabilistic Risk 1A, 1B, 1C, 1D, 1F, 1G, 11, 1N, 10, 1P,Assessment IQ, 1R, 4, 5, 9, 10, 17, 18 July 15, 201390 Day ResponseRadiation Release 1, 2, 3Fire Protection Engineering 1, 3,4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,15, 16, 17, 18, 19, 20,21Safe Shutdown Analysis 1, 2, 5,9,11, 13,14 July 31, 2013Probabilistic Risk 1J, 1K, 1M, 2, 3, 6, 7, 11, 12, 13, 14,Assessment 15, 16Fire Modeling 1B, 2C, 5C120 Day ResponseFire Protection Engineering 2Safe Shutdown Analysis 15Probabilistic Risk 1E, 1H, 1L, 8 August 30, 2013Assessment Fire Modeling 1C, 1D, 11, 2D, 3, 4, 6 Enclosure 2Page 1 of 43Response to Request for Additional Information Regarding Voluntary Risk Initiative National Fire Protection Association Standard 805By letter dated September 25, 2012, as supplemented by letter dated December 17, 2012, DukeEnergy Progress, Inc., formerly known as Carolina Power & Light Company, submitted a licenseamendment request (LAR) to adopt a new risk-informed performance-based (RI-PB) fireprotection licensing basis for the Brunswick Steam Electric Plant (BSEP), Unit Nos. 1 and 2.During the week of April 8 through April 12, 2013, the NRC conducted an audit at BSEP tosupport development of questions regarding the license amendment request.
On May 15, 2013,the NRC provided a set of requests for additional information (RAIs) regarding the licenseamendment request.
That letter divided the RAIs into 60-day, 90-day, and 120-day responses.
In subsequent telephone calls with the NRC Project Manager for BSEP, the following modifications were agreed to regarding the RAI response schedule shown in the May 15, 2013,letter:* The 60-day RAI responses will be submitted by July 1, 2013 (i.e., 60 days following theMay 2, 2013, clarification call that was conducted with the NRC). These responses wereactually submitted by letter dated June 28, 2013. Probabilistic Risk Assessment (PRA)RAIs 1A, 1B, IC, 1D, 1F, 1G, 11, 1K, 1N, 10, 1P, 1Q, 1R, 4,5,9, 10, 17, and 18, whichwere included in the set of 60-day RAIs, will be addressed in a separate submittal dueby July 15, 2013 (i.e., 60 days following the date of the letter).* The 90-day RAI responses will be submitted by July 31, 2013 (i.e., 90 days following theMay 2, 2013, clarification call). Fire Protection Engineering RAI 1, which was included inthe set of 60-day RAIs, will be addressed as part of the 90-day RAI responses.
* The 120-day RAI responses will be submitted by August 30, 2013 (i.e., 120 daysfollowing the May 2, 2013, clarification call). PRA RAI 1H will be addressed as part of the120-day RAI responses, rather than with the 60-day RAI responses.
Duke Energy's 60-day response to the PRA RAls is provided below. The response to RAI 1 K,which is not provided herein, will be provided as part of the 90-day RAI responses which will besubmitted by July 31, 2013.Probabilistic Risk Assessment (PRA) Requests for Additional Information PRA RAI 1AClarify the following dispositions to fire F&Os and supporting requirement (SR) assessments identified in Attachment V of the LAR that have the potential to impact the FPRA results and doappear to be fully resolved:
a) F&O 1-8 against ES-Al (Not Met), ES-A2 (Cat 1/11/111),
ES-A3 (Not Met), and FQ-A2(Cat 1/11/111):
Attachment 8 of BNP-PSA-085 shows in a table whether, and in some cases, howinternal event initiators were addressed in the FPRA. Describe how equipment, whosefire-induced failure could cause initiating events, was matched to the appropriate plantresponse models (i.e., internal events accident sequences).
Given the cited sensitivity study results, justify treating the cited initiators as fire-induced failure of equipment Enclosure 2Page 2 of 43following a plant trip rather than using the internal events plant response modelsassociated with internal event initiators.
ResponseThe fire probabilistic risk assessment (FPRA) was developed from the internal events PRA inwhich equipment, whose failure could cause an initiating event, was already matched to theappropriate plant response models. During component selection for the FPRA, these initiating events were reviewed for susceptibility to fire-induced
: failure, as documented in Attachment 8 ofBNP-PSA-085, BNP Fire PRA -Component Selection, and the FPRA added the fire initiator tothe initiating event logic. As described in Attachment V of the LAR for finding and observation (F&O) 1-8, significant tracing of the logic during the peer review confirmed that the initiating event logic was appropriately OR-gated with the system logic, in most cases. However, certainspecific gates associated with inadvertent safety relief valve (SRV) opening were identified asexceptions and were subsequently corrected, as described in Attachment V of the LAR for thedisposition of F&O 1-8. F&O 1-8 also cited a sensitivity study requested for two other initiators (i.e., loss of direct current (DC) power Al and loss of offsite power (LOOP)).
For the loss ofoffsite power, consequential LOOP was found to have been omitted in several locations resulting in additional cutsets and a slightly higher conditional core damage probability (CCDP)than assuming the subsequent failure of the equipment.
Consequently, logic for fire-induced LOOP was added to the fault tree, where appropriate, as described in Attachment V of the LARfor the disposition of F&O 1-8. However, based on the results of the sensitivity study, no changewas made for the loss of DC power, because both DC initiators and response model events arealways adjacent in the fault tree, so there was no difference in cutsets.PRA RAI lBb) F&O 1-9 against ES-Al (Not Met), ES-A4 (Cat 1/11), and FQ-A2 (Cat 1/11/111):
The disposition to this F&O indicates that the independence of High-Pressure CoolantInjection (HPCI) and RCIC is a source of uncertainty.
Explain how the dependency between HPCI and RCIC was accounted for in the FPRA, including a discussion onuncertainty as appropriate.
ResponseContrary to what may be conveyed by the stated disposition of F&O 1-9 in Attachment V of theLAR, the results of the multiple spurious operation (MSO) Expert Panel review, as documented in Attachment 3 of BNP-PSA-085, BNP Fire PRA -Component Selection, did not identify adependency between HPCI and reactor core isolation cooling (RCIC), and therefore no suchdependency is represented in the FPRA. The source of uncertainty is more accurately identified as the reactor pressure vessel (RPV) water level at which the quality of the steam would causean overspeed trip of either the HPCI turbine or the RCIC turbine, depending on which MSO ispostulated, rather than to suggest concurrent failures of both HPCI and RCIC.A clarification to the disposition of F&O 1-9 in the Attachment V will be included in the LARupdate.
Enclosure 2Page 3 of 43PRA RAI ICc) F&O 1-14 against PRM-B4 (Cat 1/11/111):
This F&O indicates that cable tracing was not performed in some cases. In areas wherecable tracing was not performed, identify the assumptions made about possible planttrips and fire induced failures.
Was an "exclusionary approach" used that assumes cablefailure in any areas where the presence of cable cannot be ruled out?ResponseThe decision for which cables to trace was based on particular equipment rather than onparticular areas. Consistent with the guidance in NUREG/CR-6850, cable selection and circuitanalysis was performed for equipment identified during component selection and added to thatalready existing for safe shutdown equipment.
As documented in Table 4 of BNP-PSA-085, BNP Fire PRA -Component Selection, the main feedwater system; condensate system;circulating water system; and turbine control system were not included for cable tracing, and theFPRA assumes these systems to be failed for every fire scenario.
Practically any area mightcontain cables that were not traced, either because the system is considered failed in the FPRAor because the associated piece of equipment was never credited for the PRA or for safeshutdown.
As described in Attachment 10 of BNP-PSA-080, BNP Fire PRA -Quantification, the likelihood of a plant trip due to fire was evaluated on an area basis and considered both the equipment located in the area and the equipment that could be affected by traced cables that traverse thearea. The evaluation considered the possible effects on the plant given a fire in the area but notthe likelihood of a fire in that area or the assumed status of the system in the FPRA. One ofthree possible conditional trip probabilities (i.e., 1.0, for near certain; 0.1, for reduced likelihood; or 0.01, for not likely) was assigned.
To accommodate operator discretion, no area wasassigned a conditional trip probability of zero, even if the area contained no equipment important to plant operation.
PRA RAI IDd) F&O 1-19 against FSS-Al (Not Met):The disposition for this F&O explains that the ZOI associated with a 143 kilo-watt (kW)heat release rate (HRR) (75th percentile) transient fire was used in all fires areas, exceptthe turbine building where a ZOI for a 317 kW HRR (98th percentile) fire was used. Thedisposition provides the basis for this lower HRR as existing and planned administrative
: controls, plant experience, and insights from a bounding sensitivity study. Provide furtherjustification for the use of 143 kW transient fires, given that both 143 kW and 317 kW aretaken from the same HRR distribution.
Include further description of the administrative controls used in the different areas for managing transient combustibles, the results ofreviewing plant experience and records of violations of transient combustible
: controls, other key factors for this reduced fire size, and the results of the bounding sensitivity study referred to in the disposition.
Also, confirm that 143 kW and 317 kW HRRs werethe only transient fire sizes used in the FPRA.
Enclosure 2Page 4 of 43ResponseWhere the zone of influence (ZOI) for a 143 kW HRR was used to identify targets for thetransient walkdowns, the FPRA used the same target sets for both the 75th percentile and the98th percentile fires. Effectively, no two-point HRR distribution was used for these scenarios.
Bycomparison, in the turbine building where the ZOI used for the walkdowns was based on a317 kW HRR, the affected components for the two-point HRR distribution were different for lessthan 10% of the transient ignition sources.
Therefore, the only transient fire sizes used in theFPRA are a single-point 143 kW HRR and a two-point 143 kW and 317 kW HRR.As stated in Attachment V of the LAR for the disposition of F&O 1-19, additional transient walkdowns have been performed using a 317 kW HRR, but those results have not yet beenincorporated into the FPRA. To support the use of a lower HRR in specific areas, an evaluation was performed, as documented in BNP-PSA-086, BNP Fire PRA -Fire Scenario Data,Attachment 25, to determine a reasonably realistic and bounding HRR for transient combustibles.
These results will be incorporated into FIR-NGGC-0009, NFPA 805 Transient Combustibles and Ignition Source Controls
: Program, which, post-transition to NFPA 805, will beused to limit the placement of transient combustibles and ignition sources near equipment andcables. During interaction with the Peer Review Team, an unofficial bounding sensitivity studywas performed to estimate the risk associated for a transient with a larger ZOI in a particular area by using the CCDP for a hot gas layer with 10% of the transient frequency.
: However, theseresults are not available because the sensitivity study was not maintained.
The results of a new sensitivity study will be submitted in the 120-day responses with the resultsof other sensitivity studies.PRA RAI IFf) F&O 1-26 against HR-G1 (Cat I), FSS-B2 (Cat II), and HRA-C1 (Cat II):Describe how the Human Reliability Analysis (HRA) was performed for alternate shutdownfollowing MCR abandonment.
Include in this description:
: i. Identification of events or conditions that prompt the decision to transfercommand-and-control from the MCR to the alternate shutdown station.
Clarifyhow the loss-of-control due to fires in the MCR or Cable Spreading Room wasmodeled.ii. Explanation of how timing was established (i.e., total time available, time until acue is reached, manipulation time, and time for decision-making) and which fireor fires were used as the basis for the timing. Include in the explanation the basisfor any assumptions made about timing.iii. Discussion of how different core damage end-states defined by theAbandonment HRA Event Trees presented in Attachment 10 of BNP-PSA-084 were incorporated into the FPRA, given that some sequences resulted in earlyand others resulted in late core damage.iv. Description of how the feasibility of the operator actions supporting alternate shutdown was assessed.
Enclosure 2Page 5 of 43v. Justification for assuming that continuous communication and coordination willoccur during implementation of OASSD-02 by the different operators at theirdifferent locations.
Include consideration of actions that require taking offheadsets or the unavailability of phone systems.vi. Description of how the impact of complexity on coordination of actions andoperator performance in OASSD-01 and OASSD-02 was addressed.
vii. Description of the treatment of potential dependencies between individual
: actions, including discussion of operator actions that can impact the actions ofother operators.
ResponseF&O 1-26 concerns the use of point estimates for CCDP and conditional large early releaseprobability (CLERP) rather than a detailed task analysis for main control room abandonment (MCRA). As described in Attachment V of the LAR, the disposition of F&O 1-26 involved the useof a HRA to develop a detailed Human Event Probability (HEP) for the control roomabandonment scenario.
This HEP was used to obtain more realistic estimates of core damagefrequency (CDF) and large early release frequency (LERF) for MCRA based on the results of aConsolidated Fire and Smoke Transport (CFAST) computer model of the time required to reachthe main control room (MCR) habitability threshold values for temperature and visibility, asdescribed in Attachment 16 of BNP-PSA-080, BNP Fire PRA -Quantification.
: i. In crediting operator actions for alternate shutdown following control room abandonment, the decision to transfer command-and-control from the MCR to the alternate shutdownstation is prompted by postulated MCR habitability limitations.
In particular, thesehabitability limitations correspond to the threshold values for temperature and visibility inthe hot gas layer, as identified by NUREG/CR-6850.
No operator action is otherwise credited for recovery of loss-of-control due to fires in the MCR or the Cable Spreading Room.ii. To establish the timing for MCRA based on habitability, the CFAST computer modelconsidered electronic equipment fires and ordinary combustible fires within the Unit 1and Unit 2 MCR and electronic equipment rooms with different configurations of controlroom heating, ventilation, and air conditioning (HVAC) and boundary doors. To establish a frequency for MCRA, the convolution of these times conservatively assumed theboundary doors are closed and 1% unavailability for the control room HVAC.After the decision to abandon the MCR, the system window timing for individual humanfailure events (HFEs) was taken from the OASSD-02, Alternative Safe ShutdownProcedure:
Control Building, procedure or Modular Accident Analysis Program (MAAP)calculations referenced in Attachment 10 of BNP-PSA-084, BNP Fire PRA -HumanReliability
: Analysis, (and in the individual HFEs). The only cues that are important in thisanalysis are the initial cue to evacuate the control room, the time it takes to reach a newstep in the procedure, or waiting for verification from another operator before performing an action. Train B will be the safe shutdown train for fires in the control building, excluding battery rooms, and this procedure is written assuming the equipment will beavailable as only random failures will result in the unavailability of Train B equipment andthe probability of that equipment failing is much smaller than the probability of the Enclosure 2Page 6 of 43operator's failing to perform the procedure.
For this reason the procedure is primarily written as a set of steps that do not have a recovery action. The manipulation/travel times were obtained from information gathered during a walkthrough of the OASSD-02procedures with the operators.
Any assumptions made in terms of timing were based ongathered plant simulator/walkdown information.
iii. The CCDP of the operators failing to perform the alternative safe shutdown (ASSD)actions in the OASSD-02 was calculated by summing the probability of reaching eachindividual core damage end state, without consideration for whether the sequenceresulted in early or late core damage. Given a fire that results in abandonment, 12 sequences have an end state that is expected to result in core damage. Separatefrom the rest of the FPRA, the total CDF for MCRA was determined as the product ofthis CCDP and the MCRA frequency determined from the CFAST habitability results.
Asdescribed in Section 3.6.1 of BNP-PSA-080, BNP Fire PRA -Quantification, the risksassociated with MCRA was added (i.e. similar to the treatment of risks from the multi-compartment analysis) to the results of the FPRA to obtain the Fire CDF and Fire LERF.iv. Operators train on the ASSD procedures once every two years in order to remainfamiliar with the ASSD actions and equipment.
Time requirements are stated in theprocedures, and the ability to perform the actions within the stated times was part of thewalk-through with operators to ensure the feasibility of the actions.
Also based on thetime window evaluations presented in the HRA calculator, which included travel pathtime considerations, simulator timing information, and assumptions based on knownsimilar actions at BSEP, all actions were deemed feasible.
: v. Because Section A of 0ASSD-01, Alternative Safe Shutdown Procedure Index, providesguidance for the use of alternate and diverse means (e.g., hand-held portable radios,plant paging system) to support continuous communication and coordination, removal ofthe headset or unavailability of the sound-powered phones was not considered to be aconcern.vi. The impact of complexity on coordination between operators and the evaluation of theseinteractions is addressed by the MCR Abandonment Timelines.
These timelines showwhen one operator, working in one procedure, must perform an action prior to anotheroperator beginning a step in another procedure.
If an operator is required to wait foranother operator to successfully perform an action in the procedure, then the time ittakes the first operator to completely finish their action will be considered as the delaytime (Tdelay) for the second action. The coordination of actions in operator performance isalso addressed in the fault trees that show how if one operator fails to perform an actionthen the goal of multiple
: actions, such as starting an emergency diesel generator (EDG),will be failed.vii. The treatment of dependency between individual actions is addressed in the fault treesthat show how if one operator fails to perform an action then the goal of multiple actions,such as starting an EDG, will be failed. This dependency is also treated in the timelines.
If an action appears before another action in the timeline then the first action must besuccessful in order for the second one to occur. The structure of the OASSD-01 andOASSD-02 procedures is such that there is only one recoverable action per unit, in whichexit from OASSD-01 and entry into OASSD-02 provides an opportunity for an intervening success between the actions.
Consequently, a zero dependency level is applied, and Enclosure 2Page 7 of 43these two HFEs are placed under an AND gate. All other actions appear under ORgates, because the failure of any single action other than these will result in a coredamage sequence.
PRA RAI IGg) F&O 1-30 against FSS-Al (Not met):Describe the approach and assumptions used to model fires in open and closedcabinets, and the sensitivity study on motor control centers (MCCs) presented inSection 4.8.3.1 of the LAR. Include in this description:
: i. Confirmation that walkdowns were performed to determine open and closedcabinets.
ii. Given an MCC cubicle fire, identification of the cubicles in the MCC assumed tofail.iii. Explanation of why the sensitivity study shows no impact on Unit 1 LERF andALERF, and Unit 2 CDF, ACDF, LERF, and ALERF, while showing an increasein Unit 1 CDF and ACDF.ResponseFire is not generally postulated to propagate from closed cabinets, as described in Attachment 3of BNP-PSA-080, BNP Fire PRA -Quantification.
Consequently, fires in closed cabinets locatedoutside the MCR were assumed not to contribute significantly to fire risk, and that risk is notquantified.
: However, open cabinets and closed cabinets located in the MCR and the remoteshutdown panels were quantified as "self' fires where the contents of the cabinet made up thetarget-set for the scenario, as described in Section 3.2.4.3 of BNP-PSA-080.
For open cabinets, the risk associated with fire scenarios involving targets within the zone of influence is alsoquantified in a manner similar to other ignition sources.As described in Section 3.6.5.4 of BNP-PSA-080, closed MCCs were assigned a probability of0.1 to account for the fraction of time that the fire is postulated to propagate outside the closedMCC. This probability directly reduced the Scenario Event Frequency (SEF) for all scenarios associated with the closed MCCs. The target-sets for these scenarios were developed in thesame manner as other ignition sources.
The sensitivity analysis in Section 4.8.3.1 of the LARpresented results with the probability of 0.1 removed, effectively treating fires in the closedMCCs as always propagating outside the MCCs.i. The majority of closed cabinets and closed MCCs were identified as being closed in thesource walkdowns as documented in Attachments 6, 11, and 17 of BNP-PSA-086, BNPFire PRA -Fire Scenario Data. As described in Attachment 6 of BNP-PSA-086, equipment qualification package (i.e., QDP 67) was also cited as a basis for determining about a dozen MCCs to be closed. Four additional MCCs were identified as being closedbased on a previous evaluation, which included a walkdown, as documented inAttachment 25 of BNP-PSA-080.
Enclosure 2Page 8 of 43ii. MCC fires were modeled as the entire MCC, and not down to the individual MCC cubiclelevel. No individual MCC cubicle was assumed to fail from the fire. The failures inducedby a MCC fire were based on cables identified in the zone of influence for the entireMCC.iii. As documented in BNP-PSA-095, BNP Fire PSA -Sensitivity
: Analyses, the sensitivity study shows that the results are largely insensitive to the assumption that closed MCCsare open 10% of the time. This assumption was applied to only 34 of 2401 sourcesevaluated in the FPRA. For Unit 1 LERF, Unit 1 ALERF, Unit 2 CDF, Unit 2 ACDF, Unit 2LERF, and Unit 2 ALERF, these sources simply were not significant enough to cause achange in the base value when expressed to two significant figures.
Had moresignificant figures been used the resulting percentage change may have, in some cases,shown a non-zero
: increase, but the conclusion that the results are largely insensitive tothe assumption would not have changed.PRA RAI 11i) F&O 1-38 against LE-G2 (Not Met), LE-F3 (Not Met), UNC-A1 (Not Met), FQ-E1 (NotMet), FQ-F1 (Not Met) combined with F&O 4-18 against QU-E3 (Cat I), QU-A3, UNC-A4(not Met), and FQ-A4 (Cat 1/11/111):
Explain how parametric data uncertainty was propagated and the state of knowledge correlation (SOKC) was evaluated for fire CDF and LERF. Identify fire-PRA-specific parameters (e.g., hot short probabilities, fire frequencies) that can appear in FPRAcutsets and how they were correlated.
NUREG-1855 states that all basic events(regardless of system) must be correlated if their failure rates for a given failure modeare derived from the same data set. Therefore, if SOKC was applied only to basic eventswithin the same system, provide a justification.
ResponseThe BSEP FPRA quantifies individual cutsets for each fire source scenario.
These cutsets werecombined into a single cutset file, as described in Attachment 38 of BNP-PSA-080, BNP FirePRA -Quantification.
Basic events (BEs) were renamed in a manner to allow random failures, fire induced failures, and fire induced hot shorts for the same component to have unique nameswithin the same combined cutset file. Once combined, the cutset basic events were reviewed todetermine if appropriate uncertainty parameters were used for each basic event, and additions were made where appropriate.
The UNCERT (i.e., part of the Electric Power Research Institute (EPRI) Computer Aided FaultTree Analysis (CAFTA) Suite) program was run to evaluate the combined cutset file. Based on aMonte-Carlo sampling
: approach, the mean CDF/LERF was generated along with a probability distribution histogram, which can be found in BNP-PSA-080, Attachment
: 38. The meanCDF/LERF calculations, based on the data uncertainty, were consistent with the CDF/LERF, based on the point estimated solution.
The UNCERT solution for the parametric uncertainty can evaluate the impact of the SOKC forcorrelated data if a single uncertainty parameter is applied to multiple basic events, as is thecase for reliability data using "type coded" values. No other basic event uses a singleuncertainty parameter for multiple components.
Because the mean point estimate CDF/LERF Enclosure 2Page 9 of 43results match those of the CDF/LERF values from the UNCERT solution, it is concluded thatthere are no correlated random failure events requiring SOKC additions to frequency.
For the fire PRA, various events in the combined solutions were reviewed for evaluation forSOKC. The following table, taken from Table 3.2.5-1 in BNP-PSA-080, Attachment 38,discusses the identification of fire data types reviewed for SOKC.Area of Uncertainty Discussion
: 1. Fire ignition frequency The BSEP fire scenarios are based on single ignitionsources.
Therefore, there are no correlated ignitionfrequencies within an individual cutset, precluding SOKCoccurrence concerns.
: 2. Non-detection probabilities A generic non-detection probability is used in quantifying the scenario frequencies.
Multiple detectors are notcredited, so that for individual scenarios, there is nocorrelated data.3. Non-suppression There is no correlation between various types ofprobabilities suppression, in that they are uniquely different.
: 4. Heat release rate severity See Item 1. In addition, the source target relationship isfactor/split fraction based on a single distance that is used to calculate theHRR severity factors.
The split of the generic HRRs isquantified as two individual scenarios, precluding anycorrelated data in a single cutset.5. Circuit failure probabilities With the exception of basic events where the hot shortprobability of 1.0 is used, cutsets including the samecomponent type and failure mode with the same hot shortprobabilities are assumed completely correlated.
TheUNCERT code does not address this correlation, so ananalysis showing the potential change in CDF/LERF issupplied below.Based on the conclusions of the above table, only those combinations of events with hot shortfailure probabilities less than 1.0 are evaluated for SOKC.The merged cutsets were reviewed to identify correlated hot short failure combinations abovethe truncation of 1.OE-9 for CDF and 1.0E-10 for LERF. The types of events, presented in thecutsets requiring SOKC consideration, are spurious operation of air-operated valves (AOVs)and motor-operated valves (MOVs). A SOKC multiplier, based on the number of correlated events in a cutset and calculated using the standard deviation for the hot short probabilities, wasdeveloped and multiplied by the CDF/LERF Fussell-Vesely (F-V) contribution of each SOKCcombination to calculate the increase in CDF/LERF due to SOKC.No cutset was found with correlated events from multiple systems.
Only correlated failureswithin the same systems were present within the cutset solution, so that the SOKC multipliers affect one system only. This is expected, given cutsets are based on loss of functions, andfunctions typically have redundant components at the system level. Also, this may be due in partto the limited number of altered events that are used in the BSEP FPRA. A more extensive use Enclosure 2Page 10 of 43of hot short probabilities with altered events could potentially result in correlated events fromdifferent systems existing within the same cutset.PRA RAI 1Kk) F&O 2-14 against FSS-D7 (Cat I):Clarify whether information from the System Health Reporting and System Notebookprocesses, or other sources, shows data for more than 1 year to confirm that the FireDetection and Suppression Systems have not experienced "outlier behavior."
If only1 year of data was used, justify why this is sufficient.
ResponseThe response to this question will be provided as part of the 90-day RAI responses which will besubmitted by July 31, 2013.PRA RAI INn) F&O 4-14 against FSS-E3 (Cat I), FSS-H5 (Cat I), FSS-H9 (Cat 1/11/111),
UNC-A2(Cat I/Ill):Explain how uncertainty was treated with respect to CDF, LERF, ACDF, and ALERF.Clarify the extent to which statistical quantification of uncertainty was used to evaluatefire CDF, LERF, ACDF, and ALERF. Identify significant fire scenarios where uncertainty was characterized qualitatively.
For these scenarios, explain (per Supporting Requirement QU-E4) how the FPRA is affected by these sources of uncertainty.
ResponseAs outlined in the response to PRA RAI 11, parameter uncertainties are assigned to basicevents in the FPRA model to account for the aleatory uncertainty, based on randomness ofelements in the FPRA, for fire induced failures.
As described in Section 3.2.4 of Attachment 38to BNP-PSA-080, BNP Fire PRA -Quantification, uncertainty parameters were propagated toobtain an uncertainty distribution on the calculated CDF and LERF using Monte Carlo methods.There was no appreciable contribution to risk associated with parametric uncertainty or state ofknowledge correlation.
In that regard, no evaluation of aleatory uncertainty was performed forACDF or ALERF, as it would not be expected to provide new insights.
The risk importance of epistemic uncertainties, based on level of knowledge of FPRA elements, is best assessed via sensitivity analyses by assuming an alternative outcome or method foreach modeling issue or combinations of issues. Thus, the risk importance of a given epistemic uncertainty was assessed by calculating the change in CDF or LERF using an alternate modeling assumption.
The procedure outlined in NUREG/CR-6850 for the Fire PRA uncertainty and sensitivity analysis (i.e., Task 15) was used to identify the important epistemic uncertainty issues associated with the BSEP Fire PRA.BNP-PSA-095, BNP Fire PSA -Sensitivity
: Analyses, provides the following quantitative sensitivities for CDF, LERF, ACDF, and ALERF:0 Ignition Frequency from NUREG/CR-6850 versus Supplement 1 to NUREG/CR-6850, Enclosure 2Page 11 of 43* Removal of credit for control power transformer affect on alternating current (AC) circuitfailure probabilities, and" Treatment of closed MCCs ignition sources as opening for fire.Engineering Change (EC) 89666, FPRA NUREG/CR-6850 Appendix L Sensitivity Study forBNP Main Control Boards with In-Cabinet Incipient Detection, provides quantitative sensitivities for CDF, LERF, ACDF, and ALERF for the use of incipient detection in the main control boards(MCBs) (i.e., Unit 1 only).As listed in Section 3.6.6 of BNP-PSA-080, sources of model uncertainty and relatedassumptions were qualitatively characterized by category, with the associated impact on theFPRA but usually without specific linkage to a particular significant fire scenario listed inSection 3.6.3 of BNP-PSA-080.
Category Item Conservatism Ignition Ignition Frequencies are based on data in YesFrequency the Fire Events Database that includes Clearer and more detailed"potentially challenging" fires and not collection of generic data mayactual observed fires. reduce ignition frequency (IGF).Ignition Method of apportioning a plant IGF Yes/NoFrequency between several pieces of equipment Can vary depending ondoes not take into account equipment factors involved.
: history, maintenance practices, standbyvs. active, etc. By this method the sameequipment in the same configuration attwo different plants would have different ignition frequencies.
Source HRR Bounding values from NUREG/CR-6850 Yeswere typically used for the 98th percentile It is not likely that the actualfile based on the HRR case. For a limited source configurations couldnumber of sources (e.g., cabinets) these support the default HRRs.values were adjusted based on firemodeling insights.
Source HRR Closed cabinet treatment for MCCs. NoBSEP assumes certain MCCs are closed The data for the guidance issources;
: however, guidance indicates that interpreted conservatively.
At480V MCCs can experience energetic best, only a small portion offaults which can create openings to the MCC fires would lead tosupport fire growth. To account for this, an open cabinet situation.
an additional factor of 0.1 is applied to thescenario specific ignition frequency (SSIF) to account for the fraction of thetime the MCC stays closedSource HRR Oil fires involving oil are assumed to Yesinstantly spill and instantly ignite to This method is conservative, maximum HRR with no fire growth. especially for treatment oflarge quantities of oil.
Enclosure 2Page 12 of 43Category Item Conservatism Source HRR No credit is given for the Yesprofile incipient/smoldering stages of fire growth. Allowing for these phases willprovide more time for manualsuppression credit.Source HRR Most fires use a 12 minute ramp to peak Yesprofile HRR, 8 minutes at peak, and 19 minutes Consideration of other factorsdecay. No methodology for consideration will likely shorten duration orof combustible loading and other factors reduce peak HRR for mostis provided, scenarios.
Target Selection Use of multi-point fires is not based on Yestarget importance.
More than two points or moretargeted selection of the twopoints may reduce CCDP.Target Selection Fire modeling target selection only uses Yessingle point fires. Using multi-point fires for firemodeling may reduce CCDPsTarget Selection The MCB probabilistic safety analysis Yes(PSA) panel was evaluated using a single Multi-scenario approach mayfire impacting the full panel. It is possible result in a less conservative to separate fire scenarios in this panel estimate of risk.into multiple scenarios Target Selection The identification of target sets is based Noon the flame and heat products of the fire. However, vulnerable Damage due to smoke is not quantified in components are typically this analysis.
contained in cabinets/panels.
The enclosure provides someprotection against smokedamage for fires originating outside the enclosure.
Vulnerable components inMCR cabinets or remoteshutdown (RSD) panels arealready failed in "self" firesand failure of components inMain Control Boards ismitigated by incipient detection.
Enclosure 2Page 13 of 43Category Item Conservatism Target Selection "Sensitive" electronics (e.g., integrated Nocircuits employing pin-grid arrays but not However, sensitive electro-mechanical
: devices, and not solid electronics are typically state components in power applications, contained in cabinets/panels and not discrete solid state components) and the enclosure will provideare not failed due to heat damage. some amount of protection from external fires. For firesinside MCR cabinets or RSDpanels, failure of thecomponents is alreadyassumed in the "self" fires.For Main Control Boards, thefailure of sensitive electronic is mitigated by incipient detection.
For sensitive electronics notcontained in enclosures, there is a strong possibility the cables to the components are already failed in thescenarios even though theyare not the limiting failure forthe component.
Target Selection Change Package BNP-0224 (BNP-PSA-No080, Attachment
: 37) identified three While the assumption (i.e.raceways (i.e. 2FVDI/CA, 2FVG2/CA and that failure of these raceways2FCP/CB) required for breaker in HGL scenarios is notcoordination that are not in the Brunswick significant) is potentially non-Cable Management System (BCAMs).
No conservative, it is likely thatlocation information is given for these other failures in the HGLraceways.
No basic events are failed for scenario will also fail thehot gas layer (HGL) scenarios based on respective power supplies ifthese raceways the raceways are in locations where a HGL is plausible.
Target Selection The coordination study, Change Package Yes/NoBNP-157 (i.e., BNP-PSA-080, The actual cable length couldAttachment
: 13) ignores cable drop length, be shorter or longer thanbut credits the entire length of the actually used in theendpoint cable tray. It is assumed the coordination study. Thislength difference between the drop and difference is expected to beadditional lengths of partial tray runs is within the uncertainty of thenegligible, coordination study methodand application in the PRAmodel.Damage Time Target damage is based on 625°F, Noalthough a small portion of cables are A lower damage threshold willThermoplastic.
Note that this is consistent provide less time forwith the temperature used for generating suppression.
the ZOls.
Enclosure 2Page 14 of 43Category Item Conservatism Damage Time Target damage does not credit conduit.
YesMore time should be available for suppression due to theuse of conduit.HGL possibility This calculation assumes an HGL is not Noin Electrical possible in the Electrical Tunnel and Pipe If a HGL is possible, theTunnel (ET) and Tunnel. calculated CDF value couldPipe Tunnel (PT) be non-conservative.
Time to HGL See target selection and damage time Yesitems Extending the time to the firsttray igniting will extend thetime to HGL formation andprovide more time forsuppression.
Time to HGL The cable tray growth model in YesNUREG/CR-6850 has limited applicability and appears to be conservative whenapplied outside the limits.Time to HGL Fire spread within a cable tray is offset by Yes/Nodecay. Can vary depending onfactors involved.
Non- Suppression data and fire burn times do YesSuppression not differentiate between time required to Once a fire is controlled, thecontrol and time required to completely damage set is no longersuppress a fire. increasing, which may occurlong before the fire issuppressed.
Non- When the presence of a wet-pipe or CO2  No.Suppression system is indicated in an NFPA 805 Change package BNP-0186Location, it is assumed that the system is (i.e., BNP-PSA-080, effective in suppressing any fire in the Attachment
: 34) identifies NFPA 805 Location.
locations where suppression systems may not be effective in preventing formation of aHGL. No credit was taken forthese systems.Circuit Analysis Detailed circuit analysis to determine Yesprobabilities of spurious BEs due to cable Further detailed circuitfaults is performed based on risk analysis may reducecontribution.
probability of some spuriousBEs.HRA HRA Screening
-No credit is given to Yesoperator manual actions (OMAs). Detailed analysis of OMAsmay reduce risk.HRA Many ex-control room actions are not Yescredited through screening analysis.
Further detailed analysis ofDetailed HRA Analysis is performed only HFEs may reduce HEPs forfor selected HFEs based on quantified some events.importance.
Enclosure 2Page 15 of 43Category Item Conservatism HRA HFEs performed outside the control room Yesare assumed to be unsuccessful if they For large compartments, require traversing or performing an action small fire scenarios, andin a compartment with a fire. actions with long performance time available, theenvironmental effects of a fireare limited and some creditcould be given.Quantification Power dependencies for spurious Yesoperation.
In some cases, the powermay not be available in thetime period needed to supportsome fire induced spuriousevents.Quantification Utilizing Min-Cut Upper Bound YesTools approximation with large numbers of 1.0Basic Events causes overestimation ofresults.PRA RAI 10o) F&O 5-13 against QU-D2 (Not Met), QU-F3 (Cat I), FQ-E1 (Not Met), and FQ-F1 (NotMet):The disposition to this F&O indicates that a review was performed on FPRA modeling toconfirm that no inconsistencies were created between sequence and system modeling, or between the FPRA and how the plant is operated.
This discussion of this review is notapparent in the cited documentation (BNP-PSA-085).
Describe this review and itsconclusions, and identify where it is documented.
ResponseDocumentation of this review is distributed through different sections of BNP-PSA-085, BNPFire PRA -Component Selection, with conclusions consisting of statements of acceptability ordescriptions of the required changes.Section 3.3.1.4 of BNP-PSA-085 describes the review of the PRA Internal Events modelAccident Sequence Notebook (i.e. BNP-PSA-029, PRA Model Event Tree and AccidentSequence Delineation) and Level 2 Accident Sequence Notebook (i.e. BNP-PSA-049, PRAModel Sections 7-9 Level 2 Analysis).
These PRA notebooks address the plant response andthe event trees developed for that response.
The review determined that the FPRA will bemaintained as part of the Internal Events PRA Model of Record and concluded that the use ofthe same PRA models as for the internal events sequence quantification ensures thatinterdependencies are modeled consistently and appropriately.
Because the BSEP PRA usesfunctional tops for the event trees, and functions are modeled by initiating events and by systemmodels, a detailed sequence by sequence review is not required and would provide no benefit.The fire events cannot change the modeled functions, and the conclusion is always that theaccident sequence logic is adequate because the functions are always the same: reactivity
: control, RPV integrity, inventory
: control, pressure
: control, decay heat removal, and containment integrity.
Enclosure 2Page 16 of 43Section 3.3.1.4 of BNP-PSA-085 also describes the review of applicable initiating events (i.e.,BNP-PSA-032, PRA Model Appendix C Initiating Events Assessment) and the relevance to thefire model (i.e., BNP-PSA-085, Attachment 8). Consideration of possible additional initiating events that might be unique to the FPRA is documented in Table 3-2 of BNP-PSA-085.
Attachments 3 and 4 of BNP-PSA-085 document the review and disposition of various fire-induced MSOs postulated by plant and industry personnel to have potential impact on mitigation functions and systems.
This review resulted in certain FPRA model changes as documented inAttachments 9 and 12 of BNP-PSA-085 and included the creation of a simplified bypass eventtree for a main steam isolation valve (MSIV) MSO, as described BNP-PSA-085, Section 3.3.1.4and Attachment 13.For the Level 2 review, a detailed review of the containment isolation is performed inBNP-PSA-085, Attachment
: 6. Subsequently the review of the fire impact on the Level 2 accidentsequences and phenomenological events was performed as documented in Attachment 13 ofBNP-PSA-085.
PRA RAI IPp) F&O 5-15 against QU-F2 (Cat 1/11/111),
QU-F3 (Cat I), QU-D6 (Cat I), QU-D7 (Not Met),FQ-E1 (Not Met), and FQ-F1 (Not Met) combined with F&O 5-16 against LE-F1 (NotMet), LE-F2 (Cat I), LE-G3 (Not Met), UNC-A1 (Not Met), FQ-E1 (Not Met), and FQ-F1(Not Met):Describe the assessment performed to determine the significant risk contributors andrisk importance events and failures for CDF and LERF. Clarify how the insights fromimportance analysis were used to review the correctness and reasonableness of theFPRA modeling.
ResponseAs described in response to PRA RAI 11, the individual scenario cutsets were merged into asingle cutset file and Section 3.4 of Attachment 38 to BNP-PSA-080, BNP Fire PRA -Quantification, lists, by cutset probability, the top contributors to both CDF and LERF. Priorto merging the cutsets, the correctness and reasonableness of the FPRA modeling werereviewed based on detailed cutset reviews for individual scenarios, such as that documented in Attachment 39 of BNP-PSA-080.
Based on the merged cutsets, the risk contributors and risk importance events and failureswere assessed, as described in Section 3.3 of Attachment 38 to BNP-PSA-80, and includedrankings in the following categories:
* Fire Compartments
* Fire Scenarios
* Fire Accident Sequences
* Containment Failure Types (i.e., LERF only)* Operator Actions* Fire Induced Equipment Failure Modes" Random Component Failures" Systems Enclosure 2Page 17 of 43* Component Type FailuresFor each of these categories other than Containment Failure Types, the top contributors wereranked for both CDF and LERF according to the percent contribution to risk. For Containment Failure Types, the assessment only considered the contributions to LERF. The importance ranking results in each of these ranking categories are generally used in addressing whichportions of the FPRA model need further refinement.
Insights from importance analysis were used to review the correctness and reasonableness ofthe FPRA modeling by comparing the results against what is normally understood about plantresponse.
Attachment 38 of BNP-PSA-080 provides the following insights for reviewing thecorrectness and reasonableness of the FPRA model:* The MCR and cable spreading rooms dominate risk contributions from FireCompartments;
* Scenarios involving fixed ignition
: sources, rather than transient combustibles, are majorcontributors;
* All transient and station blackout (SBO) sequences for fire result in either loss of makeupevents or loss of decay heat removal events that result in a loss of makeup;* With emergency power blackout associated with the dominant cause of core damage,failures of fail-safe containment isolation valves may not contribute as much to LERF aspresented;
* Control room abandonment for habitability is one of the more important operator actions;* Lack of a specific method for evaluating fire-induced instrument faults is evident in theresults;* Random component failure rankings show test and maintenance unavailability isimportant to fire risk;* The plant system that contributes most to fire risk is the AC power system, with* AC breakers as contributing components.
PRA RAI IQq) F&O 5-18 against LE-G2 (Not Met), LE-F3 (Not Met), LE-G4 (Not Met), UNC-A1 (NotMet), UNC-A2 (Cat I/Il/Ill),
FQ-E1 (Not Met), and FQ-F1 (Not Met):These F&Os note that uncertainty and importance analysis was not performed for fireLERF. Describe the sources of uncertainty and results of importance analyses of fireLERF.ResponseThe sources of model uncertainty and related assumptions which are listed in the response toPRA RAI 1 N are considered applicable to LERF. Of these, the most significant area of epistemic uncertainty with regard to LERF is in the area of circuit analysis, specifically as related tospurious operation of containment isolation valves. At the time of the LAR submittal, two typesof failures were not addressed by guidance in NUREG/CR-6850.
These include the treatment oflow voltage instrumentation loops and the likelihood of grounding or clearing hot shorts in DCcircuits.
Consequently, each of these is treated with a value of 1.0 in the BSEP FPRA. The Enclosure 2Page 18 of 43resulting cutsets are dominated by signal failures causing valve spurious operations or primarycontainment isolation valves (PCIVs) remaining spuriously open, even though their design is tofail safe in the closed/isolated position.
A more realistic assessment of these affects wouldgreatly reduce LERF. No additional area of uncertainty was found that is unique to the fireFPRA that is not already addressed for internal events.The sources of aleatory uncertainty were evaluated for LERF, and a detailed results analysiswas performed for LERF and documented as Attachment 38 to the BNP-PSA-080, BNP FirePRA -Quantification.
This analysis includes evaluation of parametric uncertainty for acombined LERF solution and various importance evaluations for the same solution.
In particular, parametric uncertainty was evaluated for the following:
(a) Fire scenario event frequencies (i.e., initiators),
(b) Component failure probabilities (i.e., random faults and hot short probabilities),
(c) Component maintenance unavailability, (d) Human error probabilities, (e) Common cause failures, and(f) Recovery Actions (i.e., main control room abandonment from environmental causes)As described in response to PRA RAI 11, the individual scenario cutsets were merged into asingle cutset file. Merging the cutsets required differentiation between the naming convention ofbasic events that are due to fire failures, hot shorts, or random failures.
UNCERT (i.e., the EPRI CAFTA Suite) was run using the cutset file. Random failure uncertainty parameters were based on the internal events data analysis.
Ignition source error factors werebased on NUREG/CR-6850, Supplement 1, Chapter 10. The hot short probability error factorswere based on NUREG/CR-6850 Tables 10-1 and 10-2. Other event error factors, such asthose for HRAs, were based on their specific data analysis.
Based on a Monte-Carlo samplingapproach, the code determined the mean LERF and associated uncertainty distributions.
Each mean LERF from the uncertainty simulation compared favorably with its associated meanpoint estimate LERF, providing confidence in the published risk results.
As taken fromBNP-PSA-080, Attachment 38, the uncertainty histograms for the frequency density distribution and the statistics for the confidence internals are provided for each of the units risk metrics inthe following figures and tables.
Enclosure 2Page 19 of 43Figure 3.2.4-3.
BSEP Unit 1 FPRA LERF Parametric Uncertainty Histogram
[BNP-PSA-080, Rev. 3, Attachment 38]M-.o 4O~95%- :115E-M# wroln 10M41E-7 5E-71,9.5 zbs Ib5 4.E- 51E-511 --.- -Freueqc / PodayTable 3.2.4-3 -BSEP Unit 1 FPRA LERF -Parametric Uncertainty Statistics
[BNP-PSA-080, Rev. 3, Attachment 38]5% Confidence Mean 95% Confidence Point EstimateMean5%Median95%Standard Deviation SkewnessKurtosis3.98E-069.95E-072.54E-061.10E-054.08E-064.09E-061.02E-062.59E-061.16E-055.56E-067.68E+009.92E+014.20E-061.04E-062.63E-061.21E-05 Enclosure 2Page 20 of 43Figure 3.2.4-4.
BSEP Unit 2 FPRA LERF Parametric Uncertainty Histogram
[BNP-PSA-080, Rev. 3, Attachment 38]'a4-E-7 5E-7 6E-71E-6ZEE6 31E6 4E-6 5E1-6 E-61.E-521E5 31E5 41E5 51-5I
/ PobabRyTable 3.2.4-4 -BSEP Unit 2 FPRA LERF -Parametric Uncertainty Statistics
[BNP-PSA-080, Rev. 3, Attachment 38]5% Confidence Mean 95% Confidence Point EstimateMean5%Median95%Standard Deviation SkewnessKurtosis1.49E-066.20E-071.19E-063.12E-061.51E-061.51E-066.28E-071.21E-063.23E-061.32E-061.24E+013.64E+021.54E-066.40E-071.22E-063.33E-06SOKC was evaluated for LERF cutsets as part of the uncertainty analysis.
The UNCERTprogram results above demonstrated that there is no apparent increase in risk over the pointestimate risk, so it was concluded that there is no significant data correlation from type codeddata events.
Enclosure 2Page 21 of 43However, the potential for non-type coded data events specific to the fire analysis wasexamined.
The following areas of uncertainty were examined for data correlation asrecommended by the December 2011 FPRA Peer Review:Table 3.2.5-1:
Identification of Fire Data Types Reviewed for State of Knowledge Correlation
[BNP-PSA-080, Rev. 3, Attachment 38]Area of Uncertainty Discussion
: 1. Fire ignition frequency The BSEP fire scenarios are based on single ignition sources.Therefore, there are no correlated ignition frequencies within anindividual cutset, precluding SOKC occurrence concerns.
: 2. Non-detection A generic non-detection probability is used in quantifying theprobabilities scenario frequencies.
Multiple detectors are not credited, so thatfor individual scenarios, there are no correlated data.3. Non-suppression There is no correlation between various types of suppression, inprobabilities that they are uniquely different.
: 4. Heat release rate See Item 1. In addition, the source target relationship is basedseverity factor/split on a single distance that is used to calculate the HRR severityfraction factors.
The split of the generic HRRs is quantified as twoindividual scenarios, precluding any correlated data in singlecutsets.5. Circuit failure With the exception of basic events where the hot short probability Probabilities of 1.0 is used, cutsets including the same component type andfailure mode with the same hot short probabilities are assumedcompletely correlated.
The UNCERT code does not address thiscorrelation, so an analysis showing the potential change in LERFis supplied below.The state of knowledge correlation impact on the CAFTA solved cutsets was assessed by useof a multiplier to the combined correlated failure probabilities.
For BSEP, no more than twocorrelated hot short failure events showed up in a single cutset. The multiplier for two correlated failures can be expressed as:((E(,A))2
+Var G))(E(A) )2Where:E(A) = the expected value or mean of (A)A\ = the failure rate being evaluated for the correlated failure data.Variance was calculated from the standard deviation (Sigma) using the following formula.Var(A) = A2x (e 2-1)
Enclosure 2Page 22 of 43BSEP cutsets include basic events with only the hot short probabilities of 0.6 and 0.3. Using asigma of 0.35 (i.e., Table 3.2.2-2 of BNP-PSA-0080, Attachment 38), the variances are 0.047and 0.012, respectively.
Using this information, the SOKC multiplier for two combined hot shortprobability events was calculated as follows using the equation above.Table 3.2.5-2:
Calculation of State of Knowledge Correlation Multiplier
[BNP-PSA-080, Rev. 3, Attachment 38]Hot Short Prob. (HSP) Standard Deviation SOKC SOKCVariance Mlile de(Lambda)
(Sigma) Multiplier Adder0.6 0.35 0.047 1.130.13*(LERF) 0.3 0.35 0.012 1.13The merged cutset were reviewed to identify correlated hot short failure combinations above thetruncation of 1.OE-10 for LERF. The potential increase in LERF if the SOKC multiplier wasapplied is the product of the baseline LERF, the F-V importance of affected
: cutsets, and theSOKC adder. The following tables summarize the hot short event combinations and potential risk increase from SOKC.Table 3.2.5-5:
BSEP Unit 1 LERF -Impact of State of Knowledge Correlation
[EC 92418]HSP Event 1* Event 2* Combination LERF AdditionF-V Contribution to LERFImportance (of 4.08E-6)0.6 CAC1AOV-CO-V1O_A1 CAC1AOV-CO-V9_A1 3.14E-03 1.28E-08 1.67E-090.6 MSSlAOV-CO-F003_Al MSSlAOV-CO-F004_Al 1.69E-03 6.90E-09 8.96E-10Total LERF Addition 2.56E-09* Suffix of "_A#" is added during cutset merging to identify altered events. (0.06%)Table 3.2.5-6:
BSEP Unit 2 LERF -Impact of State of Knowledge Correlation
[EC 92418]HSP Event 1" Event 2* Combination LERF AdditionF-V Contribution to LERFImportance (of 1.51E-6)0.6 CAC2AOV-CO-Vl 0Al CAC2AOV-CO-V9_A1 1.40E-01 2.11 E-07 2.75E-080.6 MSS2AOV-CO-F022AAl MSS2AOV-CO-F028A_Al 1.38E-02 2.08E-08 2.71 E-090.6 MSS2AOV-CO-F022BAl MSS2AOV-CO-F028B_Al 1.38E-02 2.08E-08 2.71 E-090.6 MSS2AOV-CO-F022CAl MSS2AOV-CO-F028C_Al 1.38E-02 2.08E-08 2.71 E-090.6 MSS2AOV-CO-F022D_Al MSS2AOV-CO-F028D_Al 1.38E-02 2.OBE-08 2.71E-09Total LERF Addition* Suffix of" "A#" is added during cutset merging to identify altered events.3.83E-08(2.54%)The SOKC analysis showed minimal impact on the LERF solutions and small impact on theUnit 2 LERF solution.
All the contribution came from the control room complex, for which Enclosure 2Page 23 of 43adding credit for the first few steps of the abandonment procedure to secure power wouldcorrect these spurious events. This suggests that removing conservatisms from the analysisshould be prioritized versus making SOKC additions.
Analysis of FPRA Important LERF Contributors:
The merged FPRA cutsets were analyzed to rank several categories of contributors to LERF.The following categories were reviewed and presented in BNP-PSA-080, Attachment 38:0000000SFire compartments important to risk,Fire scenarios important to risk,Accident sequence types important to riskContainment failure types important to risk,Operator actions important to risk,Fire induced equipment failure modes important to risk,Component types important to risk, andFailure groupings important to riskAn abbreviated summary of the results is provided here for the top contributors.
Fire Compartment Importance The following tables rank the fire compartments by contribution to risk for each for the plantunits and risk metrics.
The flag events identifying the fire compartments were used to assemblethe ranking.Table 3.3.1-3 -BSEP Unit 1 FPRA LERF -Fire Compartment Risk Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Cumulative Percent LERF Probability NFPA-805 Location Description Name Contribution Contribution FLFC210 63.40% 63.40% 2.59E-06 1.00E+00 CB-05 -U1 Cable Spreading Rm 23' ElevFLFC230 98.50% 35.10% 1.43E-06 1.00E+00 CB-23 -Control Rm 49' ElevTable 3.3.1-4 -BSEP Unit 2 FPRA LERF -Fire Compartment Risk Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Cumulative Percent LERF Probability NFPA-805 Location Description Name Contribution Contribution FLFC230 81.50% 81.50% 1.23E-06 1.OOE+O0 CB-23 -Control Rm 49' ElevFLFC211 95.20% 13.70% 2.07E-07 1.00E+00 CB-06 -U2 Cable Spreading Rm 23' ElevFLFC402 96.50% 1.30%, 1.96E-08 1.00E+00 TB2-01C/D/E/F/G/H
-U2 Turbine Building (TB)Equipment Areas: 20ftFire Ignition Source Scenario Importance The following tables rank the fire ignition source scenarios by contribution to risk for each unitand risk metric. The ignition source contribution to hot gas layer is not included in the risk Enclosure 2Page 24 of 43ranking but rather by an event for the total contribution of all sources in the compartment to hotgas layer. The percent contribution column is representative of the Fussell-Vesely importance measure.Table 3.3.2-3 -BSEP Unit 1 FPRA LERF -Fire Scenario Risk Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name Cumulative Percent LERF Frequency Fire Source Description Contribution Contribution
%FC210_BHGL 58.2% 58.2% 2.4E-06 2.4E-06 FC210 HGL%FC230_4801_B75 73.9% 15.7% 6.4E-07 6.4E-07 1-11A -NODE HY3: CTRL BLDG 125V DCDISTRIBUTION PANEL 11A%FC230_4801_B98 79.1% 5.2% 2.1E-07 2.1E-07 1-11A -NODE HY3: CTRL BLDG 125V DCDISTRIBUTION PANEL 11A%FC210_4525_BFM 83.0% 3.9% 1.6E-07 4.7E-05 1-COM-C -480V UNIT SUBSTATION COM-C%FC230_MCRA 85.5% 2.5% 1.OE-07 1.OE-07 FC230 -LOSS OF MCR HABITABILITY WITHFAILED ABANDONMENT
%FC230_4747_B98 87.1% 1.6% 6.4E-08 6.4E-08 1-H112-P606
-NODE JG5, JH9: RADIATION MONITORING CABINETTable 3.3.2-4 -BSEP Unit 2 FPRA LERF -Fire Scenario Risk Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name Cumulative Percent LERF Frequency Fire Source Description Contribution Contribution
%FC230_4819_SELF 23.5% 23.5% 3.6E-07 1.9E-05 2-12A-HZ3
-NODE HZ3: CTRL BLDG 125VDCDISTRIBUTION PANEL%FC230_4807_BFM 32.0% 8.5% 1.3E-07 1.3E-07 2-H12-P609
-NODE JM6, JH3: RPS TRIPSYSTEM A, MAIN CONTROL ROOM PN%FC230_4845_BFM 40.2% 8.3% 1.2E-07 3.2E-07 2-H 12-P608 -NODE JC5, JC6, JC7, JP7, JP8, JP9:POWER RANGE NEUTRON%FC230_MCRA 47.1% 6.8% 1.0E-07 1.OE-07 FC230 -LOSS OF MCR HABITABILITY WITHFAILED ABANDONMENT
%FC211_4572_BFM 52.5% 5.4% 8.2E-08 3.7E-05 2-2L -480V UNIT SUBSTATION 2L BUS%FC230_4873_BFM 57.8% 5.3% 8.1E-08 1.3E-07 2-H12-P617
-NODE JEG: RHR A RELAYVERTICAL BOARD%FC211_4568_BFM 62.9% 5.1% 7.7E-08 3.7E-05 2-COM-D -UNIT SUBST COMMON 'D'%FC230_4829_BFM 66.2% 3.3% 5.OE-08 1.3E-07 2-H12-P618
-NODE JHO: RHR RELAY VERTICALBOARD%FC230_4878_B98 69.4% 3.2% 4.8E-08 4.8E-08 2-H12-P630
-NODE JE1: REACTORANNUNCIATOR CABINET%FC230_4814_B75 71.8% 2.4% 3.6E-08 6.4E-07 2-XU-30 -NODE H61: DIESEL GEN 4 ESS LOGICCABINET%FC230_4897_SELF 72.9% 1.1% 1.6E-08 2.OE-05 2-12B -NODE HZ5: CTRL BLDG 125VDCDISTRIBUTION PANEL Enclosure 2Page 25 of 43Table 3.3.2-4 -BSEP Unit 2 FPRA LERF -Fire Scenario Risk Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name Cumulative Percent LERF Frequency Fire Source Description Contribution Contribution
%FC230_4806_B98 73.9% 1.1% 1.6E-08 1.6E-08 2-XU-47 -NODE JS8: TERMINAL CABINET FORRIP%FC230_4835_B98 75.0% 1.1% 1.6E-08 1.6E-08 2-XU-48 -NODE JS9: TERMINAL CABINET FORRIP%FC230_4876_B98 76.0% 1.1% 1.6E-08 1.6E-08 2-XU-46 -NODE JL9: BOP LOGICANNUNCIATOR CAB FOR UA-23 & 24%FC230_4877_B98 77.1% 1.1% 1.6E-08 1.6E-08 2-XU-52 -NODE JW5: ANNUNCIATOR LOGICCABINET FOR UA 25 & 27Fire Accident Sequence Contributions Accident Sequences were reviewed by major event types. All transient and SBO sequences result in either loss of makeup events or loss of decay heat removal events that result in a lossof makeup. The specific sequences are not identified because the sequence flag events are notused during FPRA quantification.
: However, the cutsets contain sufficient information to identifysequence contributions based on the failed events. The following tables provide a breakdown ofthe sequence contributions for both units for LERF.Table 3.3.3-3 -BSEP Unit 1 FPRA LERF -Accident Sequence Type Risk Ranking (>0.25%)[BNP-PSA-080, Rev. 3, Attachment 38]Accident Sequence Percent LERF CommentType Contrib.Transients 87.7% 3.6E-06 Includes control room abandonment, loss of makeup, loss of decay heatremoval, and LOOP or loss of emergency power.MSIV Loss of Cooling 11.8% 4.8E-07 Bypass sequenceAccident (LOCA)Station Blackout 0.4% 1.4E-08 SBO does not include non-LOOP sequences with emergency bus failures.
Enclosure 2Page 26 of 43Table 3.3.3-4 -BSEP Unit 2 FPRA LERF -Accident Sequence Type Risk Ranking (>0.25%)[BNP-PSA-080, Rev. 3, Attachment 38]Accident Sequence Percent LERF CommentType Contrib.MSIV LOCA 56.5% 8.5E-07 Bypass sequenceTransients 40.0% 6.0E-07 Includes control room abandonment, loss of makeup, loss of decay heatremoval, and LOOP or loss of emergency power.Station Blackout 2.5% 3.7E-08 SBO does not include non-LOOP sequences with emergency bus failures.
Small LOCA 1.0% 1.6E-08 Associated with RPV Head Vent MSOsPlant damage states (PDS) are not specifically identified, as the flag events are set to trueduring quantification.
The standard
: suggests, as an example, to identify plant damage stateimportance as a means of calculating LERF contribution.
While this may be important formodels that calculate LERF through PDS split fractions developed in models external to thefault tree solution, it is not relevant to the BSEP LERF model that has all the PDS contributors explicitly modeled.
Therefore, the importance factors to LERF can be evaluated withoutidentifying PDS flags.Containment Failure RankingUnit 1 LERF is dominated by a single scenario associated with the Unit 1 cable spread room hotgas layer. For this scenario, the calculation method does not provide insights into thecontributors to LERF. It is assumed that a containment isolation failure occurs from one or moreof many possible penetrations.
: However, the dominant cause of core damage is associated with emergency power blackout.
Therefore, it is highly likely, that the LERF would be less thanthe presented value due to isolation valves closing in the safe state after power is lost. Thisspecific contribution to LERF is an area of uncertainty.
Similarly, the analysis for control roomabandonment does not identify specific contributors to LERF. For both cases, the scenario ispresented as the cause of LERF and not the specific type of containment failure.The relative importance of types of containment failures contributing to LERF are provided in thefollowing tables.Table 3.3.3-5 -BSEP Unit 1 FPRA LERF -Containment Failure Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Accident Sequence Percent LERF CommentType Contrib.U1 Cable Spread Rm. 58.2% 2.37E-06 The HGL scenario assumes a 100% contribution to LERF due toHGL containment isolation faults. This is an area of uncertainty.
PCI Failures
-Fire 26.6% 1.09E-06 Containment isolation failures associated with CAC and DW drainInduced systems.MSIV LOCA 11.8% 4.80E-07 Bypass sequence from direct valve cable hot shorts or loss of isolation signal. Value includes random failures.
MCR Abandonment 2.5% 1.03E-07 The failure of control room abandonment assumes a 10% contribution to LERF.
Enclosure 2Page 27 of 43Table 3.3.3-6 -BSEP Unit 2 FPRA LERF -Containment Failure Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Accident Sequence Percent LERF CommentType Contrib.MSIV LOCA 56.5% 8.5E-07 Bypass sequence from direct valve cable hot shorts or loss of isolation signal. Value includes random failures.
PCI Failures
-Fire 33.2% 5.0E-07 Containment isolation failures associated with Containment Induced Atmospheric Control (CAC) and drywell (DW) drain systems.MCR Abandonment 6.8% 1.OE-07 The failure of control room abandonment assumes a 10% contribution to LERF.Level 2 Phenomena 2.5% 3.8E-08 Includes over-pressure phenomena such as hydrogen deflagration, steam explosions, direct containment
: heating, and vapor suppression failures.
Includes containment breach from induced ISLOCA, RPVpedestal attack and missiles.
PCI Failures
-Non-fire 1.4% 2.2E-08 Associated with RPV Head Vent MSOsOperator Actions Important to FPRA RiskThe FPRA cutsets were reviewed to determine and rank important operator actions.
Thisincludes post initiation type (i.e., Cp) and non-procedural type (i.e., Cr) actions that are part ofthe emergency operating procedure (EOP) network.
The only alternate shutdown action that ismodeled is control room abandonment.
The following tables present the important operatoractions for LERF (i.e., contributing
> 1%) for each of the two units. All actions are performed inwhole or in part away from the control room unless otherwise indicated by bold text.Table 3.3.4-3 -BSEP Unit 1 FPRA LERF -Human Action Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name* Percent LERF Description Contrib.OPER-VALVECLOSE 21.40% 8.73E-07 FAILURE TO CLOSE VALVE GIVEN FAILURE OF SIGNALOPER-MCRA 2.52% 1.03E-07 FAILURE OF MAIN CONTROL ROOM ABANDONMENT OPER-FPS1 2.46% 1.00E-07 FAILURE TO ALIGN FIREWATER FOR COOLANT INJECTION FLOW (ONEUNIT)OPER-DEPRESS 1.46% 5.96E-08 FAILURE TO MANUALLY INITIATE AND ALIGN LOW-PRESSURE SYSTEMS* Bolded text indicated control room action Enclosure 2Page 28 of 43Table 3.3.4-4 -BSEP Unit 2 FPRA LERF -Human Action Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name Percent LERF Description Contrib.OPER-FPS1 20.70% 3.13E-07 FAILURE TO ALIGN FIREWATER FOR COOLANT INJECTION FLOW (ONEUNIT)OPER-MCRA 6.82% 1.03E-07 FAILURE OF MAIN CONTROL ROOM ABANDONMENT OPER-4160X 2.45% 3.70E-08 FAILURE TO ALIGN POWER FROM OPPOSITE UNITOPER-SWRHR-O 2.24% 3.38E-08 FAILURE TO LOCALLY OPEN THE DISCHARGE VALVES FOR RHR INJECTION OPER-RCICEXT 2.04% 3.08E-08 FAILURE TO EXTEND RCIC OPERATION BY MANAGING HCTL ANDDEFEATING TRIPSOPER-480X 1.73% 2.61E-08 FAILURE TO CONNECT UNIT 1 SUBSTATIONS ES AND E6OPER-SWRHR-C 1.71% 2.58E-08 FAILURE TO LOCALLY CLOSE SW VALVES FOR FW INJECTION Fire Induced Component Failures Important to RiskThe FPRA cutsets were reviewed to determine and rank important fire induced component failures.
The following tables present failure modes contributing greater than 2% to LERF.Failure events are sorted by Fussell-Vesely importance (i.e., % contribution).
Bolded events arethose failure modes that are spurious events. These events have values of 1.0, typically indicating that the sum of hot short probabilities for multiple control cables is approximately 1.0or that detailed circuit analysis was not performed.
Additional circuit analysis may provideimproved risk results.
A specific method for evaluating instrument faults has not been identified at this time.Table 3.3.5-3 -BSEP Unit 1FPRA LERF -Fire Induced Component Failure Mode Ranking (>2%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name* Percent LERF Prob. Description Contrib.ACPOBKR-CO-AU9_Ti 61.80% 2.52E-06 1.OOE+00 CIRCUIT BREAKER AU9 FAILS TO REMAIN CLOSEDACPOBKR-CO-AV4_Ti 58.80% 2.40E-06 1.OOE+00 CIRCUIT BREAKER AV4 FAILS TO REMAIN CLOSEDICCiPIT-HI-N21A_T1 24.50% 1.00E-06 1.OOE+00 PRESSURE TRANSMITTER B21-PT-N021A FAILS HIGHICClPTT-HI-N21B Ti 24.50% 1.O0E-06 1.00E+00 PRESSURE TRANSMITTER B21-PT-NO21B FAILS HIGHADS-CHANAB_T1 23.90% 9.75E-07 1.OOE+00 SPURIOUS OPERATION OF ADS CHANNEL A OR BICC1LTr-HI-N017A_T1 21.20% 8.65E-07 1.OOE+00 REACTOR WATER LOW LEVEL #1 LEVEL TRANSMITTER B21-LT-N017A FAILS HIGHICC1L'II-HI-N017C T1 21.20% 8.65E-07 1.00E+00 REACTOR WATER LOW LEVEL #1 LEVEL TRANSMITTER B21-LT-N017C FAILS HIGHICCtPTT-LO-N002A_Tt 21.20% 8.65E-07 1.OOE+O0 DRYWELL HIGH PRESSURE TRANSMITTER C71-PT-NO02A FAILS LOWICClPTr-LO-NOO2C_T1 21.20% 8.65E-07 1.00E+00 DRYWELL HIGH PRESSURE TRANSMITTER C71-PT-NO02C FAILS LOWICCiLTn-HI-NO24BT1 6.09% 2.48E-07 1.00E+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-N024B FAILS HIGHICCiLTT-HI-N025B_T1 6.09% 2.48E-07 1.OOE+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-NO25B FAILS HIGHDCPlBDC-LP1AP_Ti 4.34% 1.77E-07 1.OOE+0O FAILURE OF DCP 125V DC SWITCHBOARD 1A BUS P Enclosure 2Page 29 of 43Table 3.3.5-3 -BSEP Unit 1 FPRA LERF -Fire Induced Component Failure Mode Ranking (>2%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name* Percent LERF Prob. Description Contrib.CACiSOV-CO-V49_T1 2.84% 1.16E-07 1.00E+00 SOLENOID VALVE CAC-V49 TRANSFERS OPENCAClSOV-CO-VSOTl 2.84% 1.16E-07 1.00E+00 SOLENOID VALVE CAC-V50 TRANSFERS OPEN* Bolded text indicated control room actionTable 3.3.5-4 -BSEP Unit 2 FPRA LERF -Fire Induced Component Failure Mode Ranking (>2%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name* Percent LERF Prob. Description Contrib.DCP2BDC-LP4A_Ti 34.10% 5.15E-07 1.00E+00 FAILURE OF DCP 125V DC DISTRIBUTION PANEL 4AACPOTFM-LP-SAT2_Ti 25.30% 3.82E-07 1.OOE+00 TRANSFORMER SAT #2 FAILURE NO POWERICC2LTT-HI-N024A_T1 25.00% 3.78E-07 1.OOE+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-N024A FAILS HIGHICC2L1-r-HI-N025ATi 25.00% 3.78E-07 1.OOE+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-N025A FAILS HIGHRCI2TME-HI-N021B_T1 19.40% 2.93E-07 1.00E+00 TEMPERATURE ELEMENT E51-TE-N021B SPURIOUS OPERATION RCIZTME-HI-N022BT1 19.40% 2.93E-07 I.QOE+00 TEMPERATURE ELEMENT E51-TE-N022B SPURIOUS OPERATION CACZAOV-CO-VlO A1 14.00% 2.12E-07 6.OOE-01 CAC-V10 FAILS TO REMAIN CLOSEDCAC2AOV-CO-V9_A1 14.00% 2.12E-07 6.OOE-01 CAC-V9 FAILS TO REMAIN CLOSEDICC2LT--HI-N024B_T1 14.00% 2.12E-07 1.00E+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-N024B FAILS HIGHICC2LTT-HI-N025B_T1 14.00% 2.12E-07 1.OOE+00 REACTOR LOW LEVEL TRANSMITTER B21-LT-NO25B FAILS HIGHSRV2CBL-DCP-NORM_Ti 13.80% 2.08E-07 1.00E+00 NORMAL DCP SUPPLY CABLES BETWEEN DP 4B AND SRV RELAY PANELDAMAGED BY POWER dueICC2PTr-HI-N21D_Ti 11.70% 1.77E-07 1.00E+00 PRESSURE TRANSMITTER B21-PT-NO21D FAILS HIGHCAC2SOV-CO-V49_T1 10.70% 1.62E-07 1.00E+00 SOLENOID VALVE CAC-V49 TRANSFERS OPENCAC2SOV-CO-VS0_T1 10.70% 1.62E-07 1.OOE+00 SOLENOID VALVE CAC-V50 TRANSFERS OPENICC2LSO-NO-N037_T1 9.71% 1.47E-07 1.OOE+00 Reactor Fuel Zone Level Recorder B21-LR-R615 (B21-LT-N037)
FailsSRV2SOV-CC-FO13A_Ti 9.27% 1.40E-07 1.OOE+00 SOV FOR ADS SRV B21-FO13A FAILS TO OPENSRV2SOV-CC-FO13B_T1 9.27% 1.40E-07 1.00E+00 SOV FOR NON-ADS SRV B21-FO13B FAILS TO OPENSRV2SOV-CC-F013C T1 9.27% 1.40E-07 1.OOE+00 SOV FOR ADS SRV B21-FO13C FAILS TO OPENSRV2SOV-CC-FO13D_Ti 9.27% 1.40E-07 1.OOE+00 SOV FOR ADS SRV B21-FO13D FAILS TO OPENSRV2SOV-CC-FO13E_Ti 9.27% 1.40E-07 1.OOE+00 SOV FOR NON-ADS SRV B21-FO13E FAILS TO OPENSRV2SOV-CC-FO13HTi 9.27% 1.40E-07 1.OOE+00 SOV FOR ADS SRV B21-FO13H FAILS TO OPENSRV2SOV-CC-FO13J_Ti 9.27% 1.40E-07 1.OOE+00 SOV FOR ADS SRV B21-F013J FAILS TO OPENSRV2SOV-CC-FO13L_T1 9.25% 1.40E-07 1.00E+00 SOV FOR ADS SRV B21-FO13L FAILS TO OPENSRV2SOV-CC-F013K_T1 9.09% 1.37E-07 1.00E+00 SOV FOR ADS SRV B21-FO13K FAILS TO OPENDCP2BDC-LP12A_Ti 8.23% 1.24E-07 1.00E+00 FAILURE OF DCP 125V DC DISTRIBUTION PANEL 12AMSS2AOV-CO-F003_TI 7.19% 1.09E-07 1.OOE+00 RPV HEAD VENT VALVE B21-F003 SPURIOUSLY OPENSMSS2AOV-CO-F004_Ti 7.19% 1.09E-07 1.00E+00 RPV HEAD VENT VALVE B21-F004 SPURIOUSLY OPENSCAC2AOV-CO-V216_T1 6.17% 9.32E-08 1.OOE+00 CAC-V216 FAILS TO REMAIN CLOSEDCAC2AOV-CO-V7_Al 6.07% 9.17E-08 6.OOE-01 CAC-V7 FAILS TO REMAIN CLOSEDACPO0KR-CO-AV4.T1 5.72% 8.64E-08 1.OOE+O0 CIRCUIT BREAKER AV4 FAILS TO REMAIN CLOSEDFIRE-LOCOND 5.69% 8.60E-08 1.00E+00 FIRE INDUCED LOSS OF CONDENSER or CWS or TCS Enclosure 2Page 30 of 43Table 3.3.5-4 -BSEP Unit 2 FPRA LERF -Fire Induced Component Failure Mode Ranking (>2%)[BNP-PSA-080, Rev. 3, Attachment 38]Event Name* Percent LERF Prob. Description Contrib.ACPOBKR-OO-AXI_TI 5.65% 8.54E-08 1.OOE+O0 MCC CIRCUIT BREAKER AX1 FAILS TO CLOSERCI2PPS-SA-NOI2A.Ti 5.65% 8.54E-08 1.OOE+O0 PRESSURE SWITCH E51-NO12A SPURIOUS OPERATION RCI2PPS-SA-N012C_T2 5.65% 8.54E-08 1.OOE+00 PRESSURE SWITCH E51-NO12C SPURIOUS OPERATION ACPOBKR-CO-AZiTi 5.49% 8.29E-08 1.OOE+00 CIRCUIT BREAKER AZi FAILS TO REMAIN CLOSEDACPOBKR-CO-AZ5.Ti 5.38% 8.13E-08 1.OOE+O0 CIRCUIT BREAKER AZ5 FAILS TO REMAIN OPENACPOBKR-OO-AIO_Ti 5.32% 8.04E-08 1.OOE+00 CIRCUIT BREAKER A10 FAILS TO CLOSERCI2PPS-SA-N19B.T1 5.29% 7.99E-08 1.OOE+00 PRESSURE SWITCH ES1-NO19B SPURIOUSLY ACTUATEDRCI2PPS-SA-N19D_T1 5.29% 7.99E-08 1.00E+O0 PRESSURE SWITCH E51-NO19D SPURIOUSLY ACTUATEDSWS2MOV-OC-V37.T1 5.26% 7.95E-08 1.OOE+00 MOTOR-OPERATED VALVE SW V37 FAILS TO REMAIN OPENSWS2PPS-SAP3213LT1 4.92% 7.43E-08 1.00E+00 PRESSURE SWITCH PS3213 SPURIOUS OPERATION FAILS LOW ISOLATESHEADERICC2PTT-HI-N21B_T1 4.88% 7.37E-08 1.OOE+00 PRESSURE TRANSMITTER B21-PT-NO21B FAILS HIGHHPC2PPS-SA-N12A_T1 3.97% 6.OOE-08 1.00E+00 PRESSURE SWITCH E41-NO12A SPURIOUSLY ACTUATESHPC2PPS-SA-Nl2CTi 3.97% 6.OOE-08 1.OOE+00 PRESSURE SWITCH E41-NO12C SPURIOUSLY ACTUATESACPOBKR-OO-2AC6_Ti 2.46% 3.72E-08 1.OOE+00 CIRCUIT BREAKER FROM SAT #2 TO 2C (2-AC6) FAILS TO CLOSEACP1BAC-LP-32A_T1 2.21% 3.34E-08 1.OOE+00 120V AC DISTRIBUTION PANEL 32A FAILURE (NO POWER)ICC2PTT-HI-N09A_T1 2.21% 3.34E-08 1.OOE+O0 PRESSURE DIFFERENTIAL TRANSMITTER B21-PDT-NO09A FAILS HIGHACP2BAC-LP-2D_T1 2.03% 3.07E-08 1.OOE+00 120V AC DISTRIBUTION PANEL 2D FAILURE (NO POWER)IAN2SOV-OCSV5481_Ti 2.01% 3.04E-08 1.OOE+00 SOLENOID VALVE SV 5481 TRANSFERS CLOSED* Bolded text indicated control room actionSystem Importance RankinqThe FPRA cutsets were reviewed to determine and rank important systems to fire risk. Theranking is based on the basic event three letter prefix for the system and includes both randomfaults and fire-induced faults. It should be understood that feedwater and main steam are linkedfault trees to condensate, so that this contribution should be considered to include thosesystems.
The following tables present the results of the analysis.
Table 3.3.7-3 -BSEP Unit 1 FPRA LERF -System Importance Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]System Percent LERF SystemContrib.
PrefixAC Power System 65.26% 2.66E-06 ACP*Instrumentation and Control Pseudo-system 31.82% 1.30E-06 ICc*DC Power System 5.62% 2.29E-07 DCP*Main Steam System 5.59% 2.28E-07 MSS*Containment Atmospheric Control System and 5.21% 2.12E-07 CAC*Hardened Wetwell Vent Enclosure 2Page 31 of 43Table 3.3.7-3 -BSEP Unit 1 FPRA LERF -System Importance Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]System Percent LERF SystemContrib.
PrefixConventional and Nuclear Service Water System 3.50% 1.43E-07 SWS*Reactor Core Isolation Cooling System 3.03% 1.24E-07 RCI*High Pressure Coolant Injection System 2.81% 1.15E-07 HPC*Safety Relief Valves and ADS 2.57% 1.05E-07 SRV*Condensate System 1.95% 7.94E-08 CDS*Reactor Building Closed Cooling Water System 1.10% 4.50E-08 RCC*Table 3.3.7-4 -BSEP Unit 2 FPRA LERF -System Importance Ranking (>1%)[BNP-PSA-080, Rev. 3, Attachment 38]System Percent LERF SystemContrib.
PrefixInstrumentation and Control Pseudo-system 50.14% 7.57E-07 ICC*AC Power System 47.08% 7.11E-07 ACP*DC Power System 38.08% 5.75E-07 DCP*Containment Atmospheric Control System andHardened Wetwell Vent 34.99% 5.29E-07 CAC*Reactor Core Isolation Cooling System 33.61% 5.08E-07 RCI*Safety Relief Valves and ADS 23.83% 3.60E-07 SRV*Conventional and Nuclear Service Water System 20.28% 3.06E-07 SWS*Main Steam System 18.11% 2.74E-07 MSS*Condensate System 6.10% 9.22E-08 CDS*High Pressure Coolant Injection System 5.51% 8.32E-08 HPC*Instrument Air and Nitrogen 4.17% 6.31E-08 IAN*Residual Heat Removal 3.23% 4.88E-08 RHR*Control Rod Drive System (Makeup) 1.92% 2.90E-08 CRD*Turbine Building Closed Cooling Water 1.17% 1.77E-08 TBC*Analysis of Component Types and Failure Types Importance The FPRA cutsets were reviewed to determine the risk ranking of important types of component (e.g., valves and pumps) and types of failure modes (i.e., common cause failures and humanfailures).
The ranking is based on the basic event naming conventions, where the three-letter prefix for the system and includes both random faults and fire-induced faults. The following tables present the results of the analysis.
Enclosure 2Page 32 of 43Table 3.3.8-3 -BSEP Unit 1 FPRA LERF -Component and Failure Type Rankings
(>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Component Type/ Failure Type Percent LERF BE StringContrib.ValvesAll Valves 12.42% 5.07E-07
??????V-*
Pneumatic Valves 8.59% 3.50E-07
????AOV-*
Solenoid Valves 4.55% 1.85E-07
????SOV*Motor Operated Valves 1.54% 6.27E-08
????MOV-*
Main DriversAll Pumps 1.83% 7.46E-08
?????DP-*
Motor-Driven Pumps 1.81% 7.40E-08
????MDP-*
AC Power Components AC Breakers 64.26% 2.62E-06 ACPBKR*AC Transformers 1.55% 6.33E-08
????TFM*DC Power Components DC Distribution Panels 5.06% 2.07E-07
????BDC*Instrumentation and RelaysLevel Transmitters 27.66% 1.13E-06
????LTI*Pressure Transmitters 26.71% 1.09E-06
????PT-*Pressure Switches 5.28% 2.15E-07
????PPS*Temperature Elements 1.11% 4.53E-08
????TME*Failure Type Contributions HRAs -Type Cp/Cr 28.52% 1.16E-06 OPER-*HRAs- Post Processed 25.95% 1.06E-06 XOP-*Common Cause Failures 2.04% 8.31E-08
*-CF*Type A -Pre-Init HRAs 1.16% 4.73E-08
????XHE-MN*
Enclosure 2Page 33 of 43Table 3.3.8-4 -BSEP Unit 2 FPRA LERF -Component and Failure Type Rankings
(>1%)[BNP-PSA-080, Rev. 3, Attachment 38]Component Type/ Failure Type Percent LERF BE StringContrib.ValvesAll Valves 62.02% 9.37E-07
??????V-*
Pneumatic Valves 41.13% 6.21E-07
????AOV-*
Solenoid Valves 23.97% 3.62E-07
????SOV*Motor Operated Valves 17.85% 2.70E-07
????MOV-*
Manual Valves 3.58% 5.41E-08
????XVN-*
Main DriversCompressors 1.11% 1.68E-08
????MDC*All Pumps 1.00% 1.51E-08
?????DP-*
AC Power Components AC Transformers 26.45% 4.OOE-07
????TFM*AC Breakers 17.32% 2.62E-07 ACP?BKR*AC Switchgears/Buses/MCCs 5.06% 7.64E-08
????BAC*DC Power Components DC Distribution Panels 36.47% 5.51E-07
????BDC*DC Batteries 1.10% 1.66E-08
????BAT*Instrumentation and RelaysLevel Transmitters 39.52% 5.97E-07
????LTT*Pressure Transmitters 20.46% 3.09E-07
????PTT*Temperature Elements 20.40% 3.08E-07
????TME*Pressure Switches 18.32% 2.77E-07
????PPS*Group Contributions HRAs -Type Cp/Cr 33.98% 5.13E-07 OPER-*HRAs- Post-processed 27.05% 4.09E-07 XOP-*Common Cause Failures 5.11% 7.72E-08
*-CF*Type A -Pre-Init HRAs 4.39% 6.63E-08
????XHE-MN*
Enclosure 2Page 34 of 43PRA RAI IRr) F&O 6-1 against CS-B1 (Cat II) and CS-C4 (Not Met):It is unclear from the documentation whether the breaker coordination studies forBrunswick Units 1 and 2 are complete.
Section 3.3.1.7 of the LAR states that "shortcircuit and coordination calculations shall be updated as necessary" and it is noted thatthere are several breaker coordination change packages, and revised packagesdocumented in BNP-PSA-080.
Attachment 36 of BNP-PSA-080 states that threeraceways could not be routed. In light of these observations:
: i. Clarify how the breaker coordination study assessed the three raceways thatcould not be routed, given that breaker coordination is assessed based on lengthof cable.ii. Clarify that all panels modeled in the FPRA have been evaluated and whetherthe breaker coordination study is complete.
Responsei. Attachment 36 (i.e., BNP-0218, Page C-1) of BNP-PSA-080, BNP Fire PRA -Quantification, identifies two raceways that could not be routed on drawings.
However,Section 5.d of Attachment 36 describes how a reasonable approximation for the firezones that the raceways traverse was possible based on the raceways adjacent to theraceway in question, the cable start and end equipment, and general plant layoutknowledge.
Attachment 37 (i.e., BNP-0224, Page C-4) of BNP-PSA-080 identifies three otherraceways that could not be routed. While not immediately apparent from the information in Attachment 37, the raceways that are missing information are located in the Unit 2electrical equipment room (i.e., Control Building 49'). The raceway for 2-2A-120V involves a cable running between two panels in adjacent rows in the Unit 2 electrical equipment room (i.e., Control Building 49'). The two raceways for 2-2D-120V involve acable running from a panel in the Unit 2 electrical equipment room (i.e., ControlBuilding 49') to a panel in the Unit 2 cable spreading room (i.e., Control Building 23'),which is where the two other raceways are known to be located.These three raceways are identified as a source of uncertainty in Section 3.6.6 ofBNP-PSA-080, and the risk associated with their assumed failure is qualitatively addressed as a non-conservative assumption (i.e., Section 3.1.3.44) that is likelymitigated in the HGL scenario by other failures for the respective power supplies.
ii. All panels modeled in the FPRA were included within the scope of the breakercoordination study, as described in BNP-0217, which has been completed.
Enclosure 2Page 35 of 43PRA RAI 4Per NUREG/CR-6850 Section 11.5.1.6, transient fires should at a minimum be placed inlocations within the plant physical analysis units (PAUs) where conditional core damageprobabilities are highest for that PAU (i.e., at "pinch points").
Pinch points include locations ofredundant trains or the vicinity of other potentially risk-relevant equipment, including the cablingassociated with each. Transient fires should be placed at all appropriate locations in a PAUwhere they can threaten pinch points. Hot work should be assumed to occur in locations wherehot work is a possibility, even if improbable, keeping in mind the same philosophy.
Describehow transient and hot work fires are distributed within the PAUs at your plant. In particular, identify the criteria for your plant used to determine where an ignition source is placed within thePAUs. Also, if there are areas within a PAU where no transient or hot work fires are postulated because those areas are considered inaccessible, describe the criteria used to define"inaccessible."
Note that an inaccessible area is not the same as a location where placement ofa transient is simply unlikely.
If there are "inaccessible" locations where hot work or transient fires are improbable and these locations are pinch points, provide a sensitivity study todetermine the possible risk increase reflecting the possible size and frequency of fires in theselocations.
ResponseAs described in Section 9.2.4 of Attachment 22 of BNP-PSA-083, BNP Fire PRA -PlantPartitioning and Ignition Frequency, the transient ignition sources are identified by walkdowns, using a specific methodology.
This methodology assumes that transient ignition sources can beplaced anywhere within a compartment;
: however, only those sources with potential targetswithin the specified ZOI actually result in fire scenarios that can be modeled in the FPRA.Consequently, potential targets identified with this methodology encompass "pinch points" asdescribed in Section 11.5.1.6 of NUREG/CR-6850.
: However, if an area has no potential targetwithin the ZOI of any postulated ignition source, that area has no fire scenario for a transient ignition source. As described in Section 9.1.1 of BNP-PSA-086, the ZOI used during thewalkdown is based on thermoset cables, as would be appropriate for the plant, with a largetrash bag on the floor under or near a raceway or cable being the mental model of the mostlikely transient ignition source. In practice, transient ignition sources are not postulated on top ofplant equipment or wedged between cable trays because that is generally not a realistic representation of housekeeping at the plant.Without establishing specific
: criteria, Section 9.2.4 of Attachment 22 of BNP-PSA-083 describes a Locked High Radiation area as "inaccessible."
Because this represents an administrative control with a physical barrier and health threat, a sensitivity study is not considered to berequired.
As described in Section 9.2.4 of BNP-PSA-086, BNP Fire PRA -Fire Scenario Data, cable firesdue to cutting and welding are assigned no target sets because a continuous fire watch with anextinguisher is required by OFPP-005, Fire Watch Program, to be present during hot workactivities and is assumed to extinguish such a fire before it can spread beyond the original tray.Transient fires due to cutting and welding are assumed to involve the same target sets as thegeneral transients.
Enclosure 2Page 36 of 43PRA RAI 5The sensitivity study presented in Section 4.8.3.6 of the LAR removes credit for incipient detection, also known as, the Very Early Warning Fire Detection System that will be installed inthe MCR main control boards (MCBs). Explain why the sensitivity study results indicate nochange (i.e., 0%) in ACDF but relatively significant change (i.e., +48%) in ALERF.ResponseThe change in the ALERF value is predominantly caused by scenario FC230_8951 BCR (i.e.,see EC 89666). This scenario is a fire in panel P601 which is not suppressed and damagesexternal targets.
The frequency of this scenario is increased in the sensitivity study due to theremoval of credit for incipient detection which increases the probability of not suppressing thefire prior to external damage.The scenario is a ALERF contributor due to cables in the ZOI related to the MSIVs. Thesecables are identified as variances from deterministic requirements (VFDRs) because the MSIVsare required to close for RPV inventory control.
The PRA conservatively assumes an openinboard and outboard MSIV in the same line to be a LERF pathway because the status of theTurbine Control Valve and Turbine Stop Valve are unknown.
Removing the MSIV cables fromthe target set effectively reduces the CLERP from 1.0 to nearly zero. The increase in thefrequency for the scenario causes an increase in the ALERF.This scenario does not contribute to ACDF because the target set for the scenario is sufficient tocause a CCDP of 1.0 regardless of whether it contains the VFDR cables for the MSIVs.PRA RAI 9Identify if any VFDRs in the LAR involved performance-based evaluations of wrapped orembedded cables. If applicable, describe how wrapped or embedded cables were modeled inthe FPRA, including assumptions and insights on how these cables contribute to the VFDRdelta-risk evaluations.
ResponseBSEP does not have any VFDR that involves performance-based evaluations of wrapped orembedded cables.Protected cables are modeled in the PRA by removing the associated failures from the targetsets of the respective sources.
While not applicable to VFDRs, the FPRA credited cableprotection when it was unrelated to deterministic compliance but still appropriate for riskreduction.
Cases where cable protection was credited are as follows:Conduit 161L1/BA was assumed to be protected from 1-1L-480V Unit Substation 1L, asdescribed in Section 3.1.3.43 of BNP-PSA-080, BNP Fire PRA -Quantification.
Thisprotection will be achieved through modification, Table S-1, item 5 and will be furtherdiscussed in the response for RAI-SSA-02.
Enclosure 2Page 37 of 43* Four cables in the MCR were assumed to be protected from specific MCR sources, asdescribed in Sections 3.1.3.46, 3.1.3.47, and 3.1.3.48 of BNP-PSA-080.
This protection will be achieved through modification, Table S-1, item 7 and will be further discussed inthe response for RAI-SSA-02.
" Raceways in DG-05 (i.e., Diesel Generator Cell 1) which are encased in Pyrocrete wereassumed to be protected from sources in the area, as described in Section 3.1.3.34 ofBNP-PSA-080.
This will be further discussed in the response for RAI-FPE-15 PRA RAI 10Attachment W of the LAR presents the total CDF and LERF for Units 1 and 2 and specifies theCDF from each of the following contributors:
"Internal Events (including internal flooding),"
"External Flood," "High Wind," "Seismic,"
and "Fire." The seismic CDF (6.2 E-8/yr for Unit 1 and6.5E-8/yr for Unit 2) used in this estimate is low compared to the seismic CDF estimate(1.5E-5/yr) presented in a memorandum from NRC staff dated September 2010 providing updated results for Generic Issue 199 (memo titled: Safety/Risk Assessment Results forGeneric Issue 199, Implication for Updated Probabilistic Seismic Hazard Estimates in Centraland Eastern United states on Existing Plants").
Also, the CDF provided for internal events(7.9E-6/yr) is much lower than the internal events CDF (4.2E-5/yr) reported in NUREG-1437, Supplement 25, dated 2006, for the BSEP license renewal environmental report. Identify thebases for the internal and seismic event CDFs and LERFs presented in the LAR, and justify theadequacy of these risk estimates for this application.
ResponseThe CDF and LERF for the internal events PRA, as presented in Table W-1 of the LAR, wereestimated from model of record (MOR) 2011, as described in Section 3.2.5 of BNP-PSA-030, PRA Model Sequence Quantification.
By contrast, Section G.2.1 of NUREG-1437 identifies MOR03 as the source that reported CDF. As documented in Attachment 8 of BNP-PSA-068, BNP -PSA BWROG F&O Resolutions, numerous changes have been made to the internalevents PRA model in the intervening years.
Enclosure 2Page 38 of 43BSEP Model of Record Update SummaryMOR REV Unit CDF (freq/yr)
LERF (freq/yr)
MOR04 Unit 1 4.11E-05 6.31E-07Unit 2 4.04E-05 6.29E-07Unit 1 3.59E-05 6.31E-07Unit 2 3.32E-05 6.29E-07M0R06 6 Unit 1 3.34E-05 2.13E-06 (Based on MOR03)Unit 2 3.07E-05 2.13E-06 (Based on MOR63)Unit 1 1.22E-05 2.13E-06 (Based on MOR03)Unit 2 1.1 7E-05 2.13E-06 (Based on MOR03)72007 Self Assessment Unit 1 1.09E-05 2.13E-06 (Based on MOR03)Unit 2 1.09E-05 2.13E-06 (Based on MOR03)MOR201 Unit 1 8.91 E-06 1.08E-07Unit 2 7.79E-06 1.11 E-072010 Peer ReviewMOR2011 10 Unit 1 7.87E-06 5.67E-07Unit 2 7.86E-06 5.65E-07MOR04 addresses findings and observations from 2001 Peer Review along with failure andunavailability data and success criteria update.MOR05 updated HRAs and answered findings and observations from 2001 Peer Review tosupport MSPI.MOR06 used the on-line EOOS-06 model as its starting point. The major enhancement associated with the EOOS-06 model was the incorporation of the ability to cross-tie service airbetween units. Changed MOV and AOV data from time based failure to demand based failure.MOR07 changes included several items: (1) implemented a new diesel room heat analysis, additionally added two independent generators to support DC power; (2) removed safety businitiators; (3) updated HRA events for post initiators and included consideration of improvedbattery life analysis; (4)added new initiating event for loss of intake structure; (5) update EDGdata to 2007; (6) update LOOP to 2007; (7) air operated valve data update; (8) incorporated aconversion from calendar years to reactor years per the PRA standard; (9) added additional water source for circulation water pump seals; and (10) provided logic to credit swing turbinebuilding closed cooling water system for both units.MOR08 involved improved logic for the instrument air system modification and improvedanticipated transient without scram (ATWS) logic, and resolved most of the issues from the2007 Self Assessment.
MOR2010 incorporates updates to the common cause failure analysis for batteries and otherequipment with major revisions to meet Regulatory Guide 1.200 improvements, including revisions to accident sequences, data and initiating event updates, and plant configuration Enclosure 2Page 39 of 43through 2009. The remaining items identified in 2007 Regulatory Guide 1.200 Self Assessment were also resolved.
MOR2011 incorporated completely revised internal flooding and new high wind analysis to meetRegulatory Guide 1.200, and incorporated model changes for multiple spurious logic forNFPA 805 and fire PRA.Table D-1 of the September 2010 NRC memorandum (i.e., ML12335A421, Safety/Risk Assessment Results For Generic Issue 199, "Implications Of Updated Probabilistic SeismicHazard Estimates In Central And Eastern United States On Existing Plants')
presents thesafety/risk assessment results for Generic Issue 199 based on the 2008 United StatesGeological Survey (USGS) Seismic Hazard Curves. In particular, a point estimate for seismicCDF was developed by integrating the mean seismic hazard curve and the mean plant levelfragility curve for each nuclear power plant. However, lacking a realistic plant-level fragility curve for BSEP, that seismic CDF was conservatively based on the high confidence, lowprobability of failure (HCLPF).
By contrast, Table W-1 of the LAR presented seismic CDF andLERF developed from a seismic margins analysis using a 0.16 g earthquake, rather than aseismic PRA, since the Individual Plant Examination of External Events (IPEEE) used a seismicmargins analysis.
Although these values do not reflect the USGS information considered forGeneric Issue 199, the seismic CDF and LERF, as presented in Table W-1 of the LAR, areconsidered the best available information for BSEP.PRA RAI 17American Society of Mechanical Engineers/American Nuclear Society (ASME/ANS)
RA-Sa-2009 describes when changes to a PRA should be characterized as a "PRA upgrade."
Identify any such changes made to the internal events or FPRA subsequent to your most recentfull-scope peer review. Also, address the following:
a) If any changes are characterized as a PRA upgrade, indicate if a focused-scope peerreview was performed for these changes consistent with the guidance inASME/ANS-RA-Sa-2009, and describe any findings and their resolution.
b) If a focused-scope peer review has not been performed for changes characterized as aPRA upgrade, describe what actions will be implemented to comply with the ASME/ANSstandard.
ResponseNo change made to either the internal events PRA or the FPRA since the last respective fullscope peer review is considered to be characteristic of an upgrade as described in ASME/ANS-RA-Sa-2009, Section 1-A.
Enclosure 2Page 40 of 43PRA RAI 18APlease clarify the following dispositions to internal events PRA F&Os identified in Attachment Uof the LAR that have the potential to impact the FPRA results and do not appear to be fullyresolved:
a) F&O 6-8 against SC-C2 (Cat 1/11/111):
Identify software codes other than MAAP that were used to establish success criteria(e.g., GOTHIC),
and describe any limitations of these codes to support success criteriaused in the PRA.ResponseAs described in BNP-PSA-070, PSA Related GOTHIC Thermal/Hydraulic Codes for BNP,GOTHIC was used in room heat-up analyses for the PRA model. Section 6.3.3 of Attachment 3of BNP-PSA-070 evaluated GOTHIC action items, or potential errors, reported to the usercommunity since the release of GOTHIC 7.2b (QA) and concluded that none of the outstanding reported errors impacts or affects the results of the analyses.
PRA RAI 18Bb) F&O 4-5 against SY-A13 (Cat 1/11/111):
This F&O states that failure of feedwater check valve F032A or F032B can lead to flowdiversion that defeats HPCI or RCIC. The disposition to this F&O argues that thesevalves each have a failure probability two orders of magnitude lower than other HPCI orRCIC failures and therefore do not need to be modeled.
Given that check valves fail atapproximately 2E-4/demand, it is not clear why these failures can be dismissed perguidance in SR SY-Al5. Provide further justification for dismissing these failures.
ResponseThe postulated HPCI/RCIC flow diversion path is through either B21-F032A or B21-F032B thenthrough both FW-V10 and FW-FV-177 to the condenser.
F032A and F032B are normally-open stop check valves, while FW-V10 is a normally-closed, motor-operated valve and FW-FV-177 isa normally-closed air-operated valve. As listed in Table 5 of BNP-PSA-004, PSA ModelAppendix B Component Failure Database, the nominal fail-to-close probability of a check valveis approximately 1E-4/demand, while the nominal fail-to-remain-closed rates for an MOV andAOV are approximately 5E-8/hour and 2E-7/hr, respectively.
Over the 24-hour mission time, theprobability of some combination of these valves resulting in the postulated flow diversion isabout IE-15.
Enclosure 2Page 41 of 43PRA RAI 18Cc) F&O 3-3 against HR-E3 (Cat I): F&O 3-4 against HR-E4 (Cat I):Annex E4 of BNP-PSA-034 (Human Reliability Analysis) presents an "Operator Interview Worksheet" form and an "engineering review,"
but no operator interview results.Describe how and where interviews with plant operators and training staff for thepurpose of confirming procedure interpretation in support of the PRA modeling aredocumented.
: Likewise, describe where and how talk-throughs with plant operators orsimulator observations for the purpose of confirming the response models for thescenarios modeled in the PRA are documented.
If these interview or talk-throughs donot exist as part of the FPRA documentation, provide the interview and talk-through results.ResponseProcedure talk-throughs were included as part of the operator interviews for which Annex E4 toAttachment 1 of BNP-PSA-034, PRA Model Appendix E Human Reliability
: Analysis, was usedas both an outline of major topic points and a convenient form for capturing specific information.
After the interviews, the collected information was then added to the respective fields of theappropriate sections (e.g., Scenario Description, Key Assumptions, Procedure and Training, Operator Interview Insights) of the HRA Calculator to document the analysis of the HFE. Theelectronic file for the HRA Calculator is documented as an output of BNP-PSA-034.
The conductof simulator observations and the subsequent assessment of the data are documented inSection E1.2 of Annex El to Attachment 1 of BNP-PSA-034.
PRA RAI 18Dd) F&O 2-3 against HR-12 (Cat I/Il/1ll):
Describe the Human Failure Events (HFEs) screening process.
Explain how HFEs thatwere screened out of the internal events PRA but could impact FPRA results wereevaluated.
ResponseHFEs were not screened out in the internal events PRA. F&O 2-3 concerns a documentation issue related to the assignment of a relatively high screening HEP for use in identifying thoseHFEs to be subjected to more detailed analysis based on their resultant risk significance.
However as described in Attachment 3 of BNP-PSA-034, PRA Model Appendix E HumanReliability
: Analysis, even those HFEs which were not identified for more detailed analysis wereretained in the model in case they should become more important for alternate plantconfigurations or for specific initiating events such as fire.PRA RAI 18Ee) F&O 2-2 against DA-C8 (Cat I):The F&O states that plant specific data concerning standby time is not collected andused in the PRA. Explain how the requirement to determine component standby time Enclosure 2Page 42 of 43(i.e., DA-C8) using operational records is met. Alternatively, justify why meeting thisrequirement at Capability Category II is not needed.ResponseAs described in Section B.2.3 of BNP-PSA-004, PSA Model Appendix B Component FailureDatabase, in response to F&O 2-2, a review of plant specific operating data over a period ofseveral years showed the standby times to be sufficiently balanced among multiple trains of asystem to be appropriately represented by the approximate split of 50% for a two-train systemor 33% for a three-train system.PRA RAI 18Ff) F&O 6-12 against LE-G5 (Not Met):It is not clear what was done to resolve this F&O. Characterization of LERF uncertainty is presented in BNP-PSA-075, but limitations in the LERF analysis do not appear to beprovided in this document or elsewhere.
Clarify what the specific limitations in the LERFanalysis are for this application.
ResponseSince the FPRA is based on the internal events PRA, the LERF analyses have some commonlimitations related to timing and containment damage. As noted in Section 6.7 of BNP-PSA-049, PRA Model Sections 7-9 Level 2 Analysis, minimal or no credit is provided for the recovery offailed equipment and loss of adequate injection at the time of containment failure due to thelimited information on human response and equipment capability under such adverseconditions.
Additional limitations which result from conservative modeling assumptions arespecific to the FPRA and are common between the CDF analysis and the LERF analysis.
Feedwater, Condensate, Circulating Water (i.e., Condenser),
and Turbine Control are assumedfailed for all fire scenarios because the related cables were not routed, as described inSection 3.1.3.42 of BNP-PSA-080, BNP Fire PRA -Quantification.
As also described inSection 3.2.11 of BNP-PSA-080, the truncation limit of 1 E-1 0/yr for LERF was on the edge ofwhat is practical for the FPRA using a ONES solution, because of excessive quantification timeand additional runs that the quantification engine could not solve.PRA RAI 18Gg) F&O 1-22 against IFSO-A4 thru IFQU-B2 (Many SRs are Not Met):For nearly all internal flooding findings presented in Attachment U of the LAR, thedispositions state that internal flooding can have no impact on the FPRA. A number ofscenarios listed in Tables W-2-1 and W-2-2 of the LAR supplement result in LOCAs. Ingeneral, spurious actuations have the potential to cause internal flooding.
Clarifywhether any fire event can result in internal flooding.
If flooding can occur as a result of afire event, then further justify why these F&Os and other internal flooding F&Os can haveno impact on fire CDF, LERF, ACDF, and ALERF.
Enclosure 2Page 43 of 43ResponseAs described in Attachment 3 to BNP-PSA-085, BNP Fire PRA -Component Selection, theMSO review resulted in the addition to the FPRA of some relatively small Interfacing SystemsLoss of Coolant Accident (ISLOCA) scenarios to radwaste or the condenser.
The potential damage due to flooding was considered, but no additional risk due to internal flooding wasidentified to result from any fire event. Consistent with Section F.0 of Attachment 2 toBNP-PSA-035, PSA Model Appendix F Internal Flooding
: Analysis, had there been any riskresulting from fire-induced internal
: flooding, it would be quantified by the FPRA, rather than inthe internal flooding model, and attributed to fire because fire was the initiating event.PRA RAI 18Hh) F&O 6-16 against IFSN-A6 (not Met) and F&O 1-33 against IFQU-A9 (Not Met)Since spurious actuations have the potential to cause spray effects, clarify whether anyfire event can result in spray effects impacting components modeled in the PRA. If so,justify why these F&Os can have no impact on fire CDF, LERF, ACDF, and ALERF.ResponseSpray effects due to the inadvertent actuation of the fire protection system were evaluated inBNP-PSA-035, PSA Model Appendix F Internal Flooding
: Analysis, but the fire-induced actuation of the fire protection system would not be considered "spurious."
Otherwise, no additional riskdue to spray effects was identified to result from any fire event. Consistent with Section F.0 ofAttachment 2 to BNP-PSA-035, had there been any risk resulting from fire-induced sprayeffects, it would be quantified by the FPRA, rather than in the internal flooding model, andattributed to fire because fire was the initiating event.}}

Latest revision as of 12:51, 21 April 2019

Redirect to:

Retrieved from "https://kanterella.com/w/index.php?title=ML13205A016&oldid=1492455"