NL-14-0927, Response to RAI Regarding License Amendment Request for Transition to 10 CFR 50.48(c) - NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generating Plants. Part 1 of 3

Response to RAI Regarding License Amendment Request for Transition to 10 CFR 50.48(c) - NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generating Plants. Part 1 of 3

ML14189A144
Person / Time
Site:	Farley
Issue date:	07/03/2014
From:	Pierce C Southern Nuclear Operating Co
To:	Document Control Desk, Office of Nuclear Reactor Regulation
Shared Package
ML14189A142	List: NL-14-0927, Response to RAI Regarding License Amendment Request for Transition to 10 CFR 50.48(c) - NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generating Plants. Part 1 of 3
References
NL-14-0927
Download: ML14189A144 (39)

Text

Charles R. Pierce Southern Nuclear Withhold from public Regulatory Affairs Director Operating Company, Inc. disclosure per 10 CFR 2.390 40 Inverness Center Parkway Post Office Box 1295 Birmingham, AL 35242 Tel 205.992.7872 SOUTHER COMHAN N Fax 205.992.7601 COMPANY July 3, 2014 Docket Nos.: 50-348 NL-14-0927 50-364 U. S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D. C. 20555-0001 Joseph M. Farley Nuclear Plant Response to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c) - NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generating Plants Ladies and Gentlemen:

By letter dated September 25, 2012, the Southern Nuclear Operating Company (SNC) submitted a license amendment request (LAR) for Joseph M. Farley Units 1 and 2 (Ref. TAC NOS. ME9741 and ME9742). The proposed amendment requests the review and approval for adoption of a new fire protection licensing basis which complies with the requirements in Sections 50.48(a) and 50.48(c) to Title 10 to the Code of Federal Regulations (10 CFR), and the guidance in Regulatory Guide (RG) 1.205, Revision 1, Risk-Informed, Performance-Based Fire Protectionfor Existing Light-Water Nuclear Power Plants.

By letter dated December 12, 2012, the Nuclear Regulatory Commission (NRC)

Staff requested supplemental information regarding the acceptance of the license amendment (Adams Accession No. ML12345A398). SNC provided the requested information by letter dated December 20, 2012. The NRC staff subsequently completed the acceptance review by letter dated January 24, 2013, (Adams Accession No. ML13022A158).

By letter dated July 8, 2013, the NRC Staff formally transmitted a request for additional information (RAI) related to the referenced license amendment. SNC's responses to these RAIs are being provided by three submittals. By letter dated September 16, 2013, SNC provided the first set of responses. By letter dated October 30, 2013, SNC provided the second set of responses and by letter dated November 12, 2013, SNC provided the remaining set of responses.

U.S. Nuclear Regulatory Commission NL-14-0927 Page 2 By letter dated March 28, 2014, the NRC Staff formally transmitted the second round of requests for additional information related to the referenced license amendment request. By letter dated April 23, 2014, SNC provided the 30 day response to the second round of RAIs. By letter dated May 23, 2014, SNC provided the 60 day response to six of the eight remaining RAIs.

Revised Attachments S, V, and W are provided as attachments to this letter. The updated total plant, Fire PRA and delta risk values are provided in the updated Attachments V and W. These attachments provide the response to PRA RAIs 06.a.01 and 35. The revised Attachment S reflects the updates associated with the RAI responses. Revisions to Attachments C and G will be provided in a separate submittal. Along with other revisions, a new modification item has been added to Attachment S for installation of the next generation Westinghouse shutdown seals for the reactor coolant pumps on both Unit 1 and Unit 2.

Following installation of the shutdown seals, additional refinements to the Fire PRA model may need to be incorporated into the Fire PRA model to verify the validity of the reported change in risk on as-built conditions. An additional item has been added to address additional operations guidance. Attachment S, Modifications and Implementation Items, and Attachment W, Fire PRA Insights, contain sensitive information and should be withheld from public disclosure under 10 CFR 2.390.

The No Significant Hazards Consideration determination provided in the original submittal is not altered by the RAI responses provided herein.

If you have any questions, please contact Ken McElroy at (205) 992-7369.

Mr. C. R. Pierce states he is Regulatory Affairs Director of Southern Nuclear Operating Company, is authorized to execute this oath on behalf of Southern Nuclear Operating Company and, to the best of his knowledge and belief, the facts set forth in this letter are true and correct.

Respectfully submitted, C. R. Pierce Regulatory Affairs Director CRP/jkb/lac adsub c ando ed before me this 3 LCday of .,2014.

4r 7

Notary Public My commission expires: / i6

U.S. Nuclear Regulatory Commission NL-14-0927 Page 3

Enclosure:

1. Response to Probabilistic Risk Assessment RAI Attachments: Attachment S- Modifications and Implementation Items Attachment V- Fire PRA Quality Attachment W - Fire PRA Insights cc: Southern Nuclear Operating Company Mr. S. E. Kuczynski, Chairman, President & CEO Mr. D. G. Bost, Executive Vice President & Chief Nuclear Officer Ms. C. A. Gayheart, Vice President - Farley Mr. B. L. Ivey, Vice President - Regulatory Affairs Mr. D. R. Madison, Vice President - Fleet Operations Mr. B. J. Adams, Vice President - Engineering Mr. R. R Martin, Regulatory Manager - Farley RTYPE: CFA04.054 U. S. Nuclear Regulatory Commission Mr. V. M. McCree, Regional Administrator Mr. S. A. Williams, NRR Project Manager - Farley Mr. P. K. Niebaum, Senior Resident Inspector - Farley Mr. J. R. Sowa, Resident Inspector - Farley Alabama Department of Public Health Dr. D. E. Williamson, State Health Officer

Joseph M. Farley Nuclear Plant Response to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)

NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generating Plants Enclosure 1 Response to Probabilistic Risk Assessment RAI

Enclosure 1 Response to Probabilistic Risk Assessment RAI Farley PRA RAI 35 Section 2.4.3.3 of the NFPA 805 standard incorporated by reference into 10 CFR 50.48(c) states that the PSA approach, methods, and data shall be acceptable to the AHJ, which is the NRC. Regulatory Guide (RG) 1.205, "Risk-Informed, Performance-Based Fire Protection for Existing Light-Water Nuclear Power Plants," identifies NUREG/CR-6850 as documenting a methodology for conducting a Fire PRA (FPRA) and endorses, with exceptions and clarifications, NEI 04-02, "Guidance for Implementing a Risk-Informed, Performance-Based Fire Protection Program Under 10 CFR 50.48(c)," Rev. 2, as providing methods acceptable to the staff for adopting a fire protection program consistent with NFPA 805. RG 1.200, "An Approach for Determining the Technical Adequacy of Probabilistic Risk Assessment Results for Risk-Informed Activities," describes a peer review process utilizing an associated ASME/ANS standard (currently ASME/ANS-RA-Sa-2009, "Addenda to ASME/ANS RA-S-2008, Standard for Leveil/Large Early Release Frequency Probabilistic Risk Assessment for Nuclear Power Plant Applications") as one acceptable approach for determining the technical adequacy of the PRA once acceptable consensus approaches or models have been established. In a letter dated July 12, 2006 to NEI (ADAMS Accession No. ML061660105), the NRC established the ongoing FAQ process where official agency positions regarding acceptable methods can be documented until they can be included in revisions to RG 1.205 or NEI 04-02.

Section 2.4.4.1 of NFPA 805 states that the change in public health risk arising from transition from the current fire protection program to an NFPA 805 based program, and all future plant changes to the program, shall be acceptable to the AHJ, which is the NRC. RG 1.174, "An Approach for Using Probabilistic Risk Assessment in Risk Informed Decisions on Plant-Specific Changes to the Licensing Basis," provides quantitative guidelines on CDF and LERF and identifies acceptable changes to these frequencies that result from proposed changes to the plant's licensing basis and describes a general framework to determine the acceptability of risk-informed changes.

As stated on page B-1 of Appendix B of PRA-BC-F-1 1-004, "Fire PRA Logic Model," the new Westinghouse Shutdown Shield (SDS) was installed in fall 2010.

The internal events PRA (IEPRA), upon which the FPRA is based, takes credit for the SDS (failure rate of 0.0271/demand), limiting the leakage rate to 2 gpm where the faces of the SDS seal components remain in contact. The assumed leakage rate is increased to 19 gpm if the SDS actuates but the pump shaft continues to rotate if not tripped in a timely manner. Finally, if the SDS does not actuate at all, "existing" (Westinghouse Owners Group (WOG) 2000 or Rhodes Model) seal model leakage rates are applied. Given the July 26, 2013, 10 CFR Part 21 notification by Westinghouse concerning defects with the SDS performance, provide a sensitivity evaluation that removes all credit for the SDS package, including both probability and consequences as appropriate. Provide revised estimates of CDF, LERF, A CDF and A LERF, including non-fire hazards for CDF and LERF, as a result of removal of this credit. Should this result in any changes to conclusions regarding the transition satisfying RG 1.174 risk/A risk guidelines, address any changes that will be made to accommodate this.

E1-1

Enclosure 1 Response to Probabilistic Risk Assessment RAI When performing this analysis, include the composite effect from all previous re-evaluations, including any synergistic effects, specifically including the following:

a. From the LAR and the December 20, 2012 LAR Supplement, sensitivities related to the electrical cabinet fire severity method (Section V.2.1) and use of control power transformer (CPT)

(Section V.2.3; also response to PRA RAI 08.a).

b. From the RAI Responses dated September 16, 2013 (ADAMS Accession No. ML14038A019):

L PRA RAI 01 .a - Removal of credit for VEWFDS in the MCR (also PRA RAI 01.01) ii. PRA RAI 15.a - Revised seismic CDF based on 2008 USGS data iii. PRA RAI 28.k - Validity of current Ignition Bin 15 fire frequencies

c. From the RAI Responses dated November 12, 2013 (ADAMS Accession No. ML13318A027):

i. PRA RAI 07.e - Use of 0.1 CCDP for MCR Abandonment ii. PRA RAI 17.d- Turbine Building Collapse iii. PRA RAI 33.c - Revised MCR Abandonment analysis (also RAI PRA 33.c.01)

RESPONSE

The response to the Reactor Coolant Pump (RCP) seal model was provided in SNC's letter dated May 23, 2014. SNC plans to install the next generation shutdown seal from Westinghouse as a corrective action resolution for the issues identified in the Part 21 notification. An installation item has been added to Attachment S for installation of these next generation shutdown seals.

The current modeling of the SDS performance in the PRA for NFPA 805 reflects the original PRA modeling guidance for the Westinghouse shutdown seal (WCAP-17100-P/NP Rev. 1). PRA modeling guidance specific to this next generation (Generation Ill) SDS is in the development and review process for submittal to the NRC. SNC expects the RCP SDS risk benefit using this model to be consistent with the benefit currently being determined using the previous WCAP-1 7100-P/NP, Rev. 1 model. SNC will monitor Westinghouse and PWR Owners Group efforts relative to issuance of this modeling guidance and NRC review.

The updated total plant, Fire PRA and delta risk values are provided in the updated NFPA 805 LAR Attachments V and W. These attachments also provide the response to PRA RAI 06.a.01. A revision to Attachment S of the LAR is also provided to reflect updates associated with the cumulative RAI responses.

E1-2

Joseph M. Farley Nuclear Plant Response to Request for Additional Information Regarding License Amendment Request for Transition to 10 CFR 50.48(c)

NFPA 805 Performance Based Standard for Fire Protection for Light Water Reactor Generatinq Plants Attachment V- Fire PRA Quality

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Southern Nuclear Operating Company Attachment V Fire PRA Quality V. Fire PRA Quality I 3229 Pages Attached l

Page V-I I 0$

I Rev 0jI Page V-1

Southern Nuclear Operating Company Attachment V - Fire PRA Quality V.1 Fire PRA Quality In accordance with RG 1.205 position 4.3:

"The licensee should submit the documentation describedin Section 4.2 of Regulatory Guide 1.200 to address the baseline PRA and application-specific analyses. For PRA Standard"supportingrequirements"important to the NFPA 805 risk assessments,the NRC position is that CapabilityCategory II is generally acceptable. Licensees shouldjustify use of CapabilityCategoryI for specific supportingrequirements in theirNFPA 805 risk assessments, if they contend that it is adequate for the application.Licensees should also evaluate whether portions of the PRA need to meet CapabilityCategory//I, as describedin the PRA Standard."

The FNP Fire PRA has undergone a RG 1.200, Revision 2, Peer Review against the ASME PRA Supporting Requirements (SRs) by a team of knowledgeable industry (vendor and utility) personnel. The review was conducted by the Westinghouse Owners Group in October 2011 under LTR-RAM-II-12-007, "Fire PRA Peer Review against the Fire PRA Standard Supporting Requirements from Section 4 of the ASME/ANS PRA Standard for the Farley Nuclear Plant Fire Probabilistic Risk Assessment" in accordance with NEI 07-12 as endorsed by RG 1.200 Rev 2. The conclusion of the review was that the FNP methodologies being used were appropriate and sufficient to satisfy the ASME/ANS PRA Standard RA-Sa-2009. The review team also noted that NUREG/CR-6850 methodologies were applied correctly.

The summary of the peer review findings exhibited the following statistics for the evaluation of elements to the combined PRA Standard. For the FNP Fire PRA, 88% of the SRs were assessed at Capability Category IIor higher, including 8% of the SRs being assessed at Capability Category I1l. The FNP Fire PRA had an additional 5% of the applicable SRs assessed at the Capability Category I level. The Fire PRA was found to not meet 7% of the applicable SRs.

The Westinghouse Peer Group concluded that the Farley Fire PRA is consistent with the ASME/ANS PRA Standard and supports risk-informed applications. As a result of the peer review and the fire risk evaluation process the FNP Fire PRA has undergone additional model refinements. These refinements were made consistent with the methodologies that were reviewed during the FNP Peer Review.

This attachment provides a detailed assessment of each of the findings identified by the Peer Review team. Table V-1 lists each finding and provides the FNP disposition of the finding. Table V-2 lists each SR that was MET at a Capability Category I level and the disposition by FNP as to why that Capability Category level is acceptable.--Thiistable would only include SR, rot PrFeViu*ly idenfientable V 1. RIn the ase of FlNP al Gapability Category I leevel Sl had a finding written againet them and are therfor n ludedI in Table V/ 1.

The Peer Review Reports (LTR-RAM-II-12-007) will be available for NRC review.

I Page V-2 IRev Ojl I Rev Oji Page V-2

Southern Nuclear Operating Company Attachment V - Fire PRA Quality V.2 Sensitivity of Fire PRA Methods During the development of the FNP Fire PRA, SNO applied an alternate analysis method to NUREG/CR-6850. SNO determined that the approach taken by this alternate method provides a more realistic representation in the FNP Fire PRA model.

V.2.1 Electrical Cabinet Fire Severity Methodology The Electrical Cabinet Severity Factor Methodology has been removed from the Farley Fire PRA. As such, the sensitivity on this method is no longer required.

NUIREG!CZIR 6850 provides, recommended Heat Release Rate (HRR) diStributions for several different ignition sources based on experimenta! data and expert judgment.

NUREG!CIR 69850.also provides fire ignition frequencies developed froM a reVieW Of the EPRI Fire Events Database (FEDB) that classified firc events as cithcr potentially cahalleninqig or non challengqing. These two elements arc not explicitly coordinated to ensure that the HRR is coensistent with industr; experience. Despite this, NJRE=GiCR 6850 provides recommen~ded HRR values based on fie geei c abinet cases However, for a given cabinet care, the recomnmended treatment is the same wt respect to HRRs. As there is wide variation in the physical characteristics, of fires in the FEOB3 even within a cabinet case, the fire frequenci*es derived from that data set do not necessarily mnatch up with the HRR recom~mended in NUREG/CR 6850. lndustr,' effort implementing the NUREGiICR 6850 mnethodology have discoered instanecs where the prescrbe tAreatmet results in predictions that are inconsistent w~ith the industr The inco-nsistency between the basis for the prescribed fire frequency values and the HRR treatment methodology provided in NUREGIGIR 6850 was, adderesed via a reie of the industry fire events. The modified treatment involves the application o-f a fire frequencay adjustment factor. The factors are effectively a conditional probabilityt repflect the fraction of fimres tha;t are predicted to-exhibit bhehavior consirstent with that described in NUJREG!CR 6850, Appndcs and G. This NUREGIGIR 648-50 behav&'or is inte-nded to addres nagrat the assumFed 12 minute growth rate, theHR distibution, and the behavior Of fire suppression activities. The manri hich this conditional probability is developed for elecr-,ica! cabinets is such that it inherentl i ncludes some credit forF fire suppression. if this flacrtonr is applied in an analyris, no additional credit forF supeso can be applied to eliminate the consideration o porssil fiedaae othe firsit nearby target. Inherent in this treatment is the assumptio that a postulated aggressive fire represents a potential threat to nearby targets unless9 -Afe-ature prccluder, such effects. Aftributes that were considered when evaluating the potential for fire damage to nearby targets- -are- struc-tural fe-ature assroc-iated with the fire ignition sourcGe (panel enclosrUe) and spacing between the fire ignRition source and the nearest target of interest. in some cases, the cabinet itSelf Ma" be sealed as defined by guidanre provided in NUREGIGIR 6850 and FAQ 08 0012. In this case, fire would not propagate beyond the ignition source (as defined byth enclosurUe). Use34 of thte El-cftrical C.Mabint Fire S9e'.eity Methodology at Farley Nuclear Plant did noet compromirse the fire moedeling that was, pe~formed in the field. in other words, the factor moedifies the fraction of events for wAhic.h fire mo-deling is applied I Page V.3 IRev Oji I Rev 0jl Page V-3

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Southern Nuclear Operating Company Attachment V - Fire PRA Quality

!* AI i-tinner flflRRIflIe Tire nrOn~1o~TInfl DeVOnCI me TICE~T e~Tt2rnm miner C.ifl :3D01'! orertit Tor

...............-. ~.

FriinFnrrsinnrnnlentIat the eFGent timino thews medemeodi'lared GaAanneIn riteGI v*r. aapa rn*a anstRt Wait the evena Ant.; .a.tt..a ala-an andl, th Ifarife *k ' rnnamna ya%*1**. GaGkt-~ler U CIMOV -M-M-1-1-V Ir For 0 Morwrlau V V r"O tf V.21. FP Application The F=NP Fire PRA utilized thirs mnethodology for many of the scenarios that have an ignRition Soure OR the Bin 15 catego\". Table V.2 1 provides the conditi*oal p-rbabilipg Of propagatior definedfoh* each type of source, a, applicable. For additiona! fmation relating to the use and implementation of this method, refer to F=NP Srenar*o Developmcnt Report (PRA' B3C F 41 014) section 11 and refeerrne 17 (Supplementa!

Fire PRA' Methods, Rey 4).

Tabl V2 SU MMARY OF. BIN 15A CONDITIOINALI PROBABILITIES CoIVnd"itio nal P1robability Gr Il Crnditnal ProIbability Cabinet Type Proepagating Fire Proconted in ProGpagating Fire Uced ino SupplemCental Ctthodr FNP.-AnayE-Switchgearc!Load Ccnters 2 2rm-o04 F0 Motor-Con-,trol1 Centters 4~QF-Q4 2.,OE 01 Low Voltage Cabinetl; ME-2  !.~EOF=0 Note 1 lnverters arc not treated 6eparately. !%Wtad, they Wre grouped with MCGc in this 1.12 1 2 Sensitivity; an 1c ISDOf EoCtirica' Cabinet FireG SeVerity

"/'L! ..... JI.L _ J nRis methou is GonSIuerca

_'_ :J--*-- J not enaorsecaOorMe oevelopmenm

---

J[ ---- J .... J {--*LL-- J of a -rwe ir-'w., anu a t;OIR~lt~l~l~-V J;i,~;;II1 Hai; ILtI ff ,LOWlt; P4V.,1 %7j,;tO; GUUL- tVI use~t Ml~mtHiU.

ne ~~ensw~ir~.' in~,v~i~ rerno'!eR me eonnmon~i nrnn~niirn..' worn me ~nnhic.anIe a.a' h n.nI ata. t.t. T, raat.fa y ~aA. aan,,nla nasa', a~ r;at. a'~fl A ra t nflA,4,nAaaalk A nra A-+a~.

1 r f-a n,.

A aCKID Da~a Caa4a 0at *.sl ~ .a.

-a,-.-

Rey 0, PRA BCG- F 12 001. A summa,.*r* f the review pr ,o i oded below, The base Fire PRA does noet include credit for manual sprsio because iti inherently credited in the electirial cabinet factor method However, since this sensitivity is removing the conditional probability, mnanual suppression cr.edit can b ev taken, feasiblo.

The sensitiity,analysis-focused on the effects on the Unit 1 Train A and B analysi This, is based on the Unit 1 and Unit 2 moedel yielding s*Fimiar results. TableV22 Mrnuidiag the reR.Iitez of thp ApnnitiviP-I Page VA I

I Rev Ojl Oji Page V-4

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Southern Nuclear Operating Company Attachment V Fire PRA Quality -

Table V.22 Case Resultant CDF Deita-GDF c~ag GQF--Iase 5-24E45 ~ -

CODF Elec;4t Ca;b FS ACOF Elect Cab ESW SeA64" ~ 4.-03 -2 E- 1 46-06 4-:77%

LERF -3ase 4-26E-06--

LERF El!ect Cab FS Sg4 345EE-06 206%

ALERE Baco 444E--97--

ALERE Ele--tct ab E.S T~~hla X/*.L 2) cyk, Rh) W  ;..w Ltt *14A. *'

t tke,  ;*

L fiMe

• CDFl=

. Or,t 7 rME= . Qr%, p~r n ~r%+.mI *i.. inn~+ ~%1'rlt= ,.J U 0J 01=

QC: Q&.,*

• L~ flJ~.fl~tUY ~JU ij Ut. flAil VtP~YU *i fltfl I USAI t V tU Ut..7 U.AIU~UA .U.flt.I I Ult. I Ufl*. I I IS.y

^ r% r L- I *A*

(Jt-II: lil r.IJk11I AAttAIA + V EI 449I IA 4

~

A -'If- A C z

-r A

L. I Alit It 1 11 11 A

L.

AI ILL-Ai L- L-IL iT~A it4 Ai I Z*

1I CA 1 r!nr*

.IA... -~ 0r1 fl~ -.I.I.A -Z... A .4nr- nl- Z--.1 I-M1r- ... L-:-L :..... LL..-

IwtUITS WeU J.OeI= We "' - "'t-1 fdl~t 116 ONI¶ 'E. ;Iv tP I01 L1=rtr,= wVii.ii rtfiuut W contribution of tho full poWer inernal events and seismirc risk. The delta ILE"RF for this 6ensitiVitY is 5.131E 0:7.

PFo the sensitivity analysis, based on; the tota! plant GDF= of 9. 19E 05, the delta CDF=

should be in the Region; 1 range as identified by RG 1.174 fo1racIeptability. However, the dotaDF for this sensitiVity is just outside the Region 11range. Based On the total plant LERF of .1OE 06, the .L,-RF= should also be in the Re-o .. range. TheR*LERF for this sensiitivity is within the Region ,1range.

As stated previously, t-he- onioalprobability Of propagation; Of electrical cabinet fire us~ed in the base Fire PRA. is not endorsed by the NRC. In order to meet the scheduled LA.R 6ubmittal date, the sensitivity study was conducted over the period of only a few weeks. initially, all pan+el farto* s were set to 1.0, and oGF +A DF.LERF RALER*,-,F .

incrGeased markedly. With selected refinements on top scenarios, and limited additional 4

taFý1 wLu rlt~wrsncyt. +k,ý UU~U ;nm U1.ll~-. e, %Atir- ^- n -r~r.r+c,n

' nk p SOr., ithprm kirn hr A kr.rar hplopn~

I f, r, V.2.2 Sensitivity Analysis for Use of Generic Ignition Frequency NUREG/CR-6850 provides generic ignition frequencies as identified in task 6. These frequencies have since been reviewed and updated as part of FAQ 08-0048. The updated generic frequencies are documented in NUREG/CR - 6850 Supplement 1, Chapter 10. NUREG/CR-6850 Supplement 1 states that if the analyst uses these new generic frequencies then a sensitivity study must be done. The sensitivity is only required for bins, as defined in Task 6 of NUREG/CR - 6850, that have an alpha value of less than or equal to 1 in EPRI 1016735. This is discussed in more detail in note 10 of NUREG/CR - 6850 Supplement 1.

Page V-5 Rev Rev 0jiOji Page V-5

Southern Nuclear Operating Company Attachment V - Fire PRA Quality As discussed in Farley Fire PRA Task 1 & 6, Plant Partitioning and Fire Ignition Frequency, PRA-BC-F-1 1-0094, the generic frequencies from NUREG/CR - 6850 Supplement 1 were Bayesian updated to reflect plant specific data. Table V. 2-13 includes the results of the sensitivity. For the purposes of this sensitivity, the Unit 1 Train A and B CDF/LERF risk model is used. Similar results/insights are expected for Unit 2 CDF/LERF.

Table V.2-13 Ignition Frequency Sensitivity Resultant Delta Case CDFILERF CDFILERF Change 6.59E-056724E-CDF - Base 05 1.1 7 E -0 4 7 4 6 = 5 .1 1E -7 8 64 %

CDF - IGF Sensitivity 06 052.22E--06 78Y4 ACDF - Base <01g.40E 0Q

<01 ! 20Ei 05 -2.23E- 155921%

ACDF - IGF Sensitivity 053770E-06 4.81 E-006-.26&

LERF - Base 06 8.19E-0Q 40G- 3.38E-LERF - IGF Sensitivity 06 O64434E--

ALERF - Base <014E 0:7 0

<011 ofir z0E*-1.1 1E- 1021Y6-44%

ALERF - IGF Sensitivity 056-6E--

Note 1: The <0 delta is defined in Table W-6.

It should be noted that an increase in CDF/LERF is expected based on the type of sensitivity being performed. One of the ignition frequency Bins, Bin 15 - Electrical Cabinets, is included as part of this analysis. The ignition frequency for Bin 15 had a decrease of approximately a factor of 2 between NUREG/CR-6850 and NUREG/CR-6850 Supplement 1. The Farley Fire PRA results are mostly driven by scenarios that have a Bin 15 ignition source, so it follows that an increase is seen in this sensitivity.

Table V.2-13 shows that the total fire CDF results are 1.1 7E-047..46E 05 or a total plant risk of 1 .459GE-04E--5 for CDF, which includes the contribution of the full power internal events and seismic risk. The delta CDF for this sensitivity is -3.67E-051.2-

05. Table V.2-13 shows that the total fire LERF results are 8.19E-063.1OE 06 or a total plant risk of 8.52E-0634.E406 for LERF, which includes the contribution of the full power internal events and seismic risk. The delta LERF for this sensitivity is -2.20E-051.0!E-06.

For the sensitivity analysis, based on the total plant CDF of 1.45E-0494Q*-Q0 the delta CDF should be in the Region III range. Hewever-tThe delta CDF for this sensitivity is withinjust-eutside the Region III range. Based on the total plant LERF of 8.52E-063.36E 06 the delta LERF should alse-be in the Region II range. The delta LERF for this sensitivity is within the Region IIrange.

The e!eGtrical cabinet fire severity factor sencitivity analysis shews that remoin the severity factorre6ult6 in estimated AGIDFIALERF values Gioest te a pac l

I Rev 0j.1 Page V-6

Southern Nuclear Operating Company Attachment V - Fire PRA Quality criteria defined On RG 1.174. HoweverF, there sensitivity estimrates include emnbedded analytical conser.'atisms, SUch as crediting futur~e plant modifications for the hypothctic~al GOMpliant case. it i6 believed that when these analytical conscrwatisms are removed, the delta Fr*s estimates wil! be within the accGeptable regulator; limits. Hoee, naeodan~e With recent dierction provided by the staff (leciFe fromn Chardes E. Moulton to AlexandeF R. Klein, August 10, 2012, ML= 122200600), Southern replacin. the 2012 indutry consensus elecrrical cabinet method with themethod

k A k KIHQCfý-/fý prvov T ~-IS- C20AQ "rk *S.S f^

Flu , ri ..

AA;#;

flx t% " 1.45 VTVf Al ;

zo PC rf^

11 " Iy A +

a po 4;

Though the ACDF/ALERF are within the acceptable guidelines of RG 1.174, a review of the contributors by fire area for this sensitivity is provided below. This review includes the Defense in Depth actions currently identified in the Fire Risk Evaluations for areas with high risk contributions.The ignition frequency sensitivity study also shows that Using oldcr, more conserIative EPR Ifie frequen1y estimates results in A1 DF!4AI

• RF values slightly above the acceptance criteria defined by RG 1.174. Additional work may lower the AGDFDALERF values estimated this sensitivity Ai study. However, sinae SNC irs Feestablishing the baseline and delta risk estimates while replacing the electirial cabine model, no additional work is pe~fGFmed at present to justif' the resUlts of thirs sensitivity study.

Table V.2-2

__________ Defense in Depth Actions _______

Fire Area DID Action DID Description Fire Area Desrio Description The Fire PRA model assumes a transient fire HRR of 317 kW.

Transient Current plant transient controls are Electrical Controls unlikely to support sufficient 1-034 Penetration Room material to be introduced into the Train A room that would generate this level of HRR Electrical 1-035 Penetration Room Train B Main Control Room Automatic suppression is installed Diesel Diesel Generator Automatic in specific fire areas throughout Generator Fire Rooms and Suppression the plant, some of these systems Area Switchgear rooms are not credited in the Fire PRA Outdoor Units 1 and 2 XFMR Area Start-up XFMRs A

_ I__and B V.2.3 Sensitivity Analysis for Use of Control Power Transformer The Farley Fire PRA was originally developed using the recommended hot short probabilities from NUREG/CR-6850. ML14086A165, "Supplemental Interim Technical Rev 0jl Page V-7

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Guidance on Fire-Induced Circuit Failure Mode Likelihood Analysis" was published on April 23, 2014. This guidance provided updated data for the recommended hot short probabilities to be used in a Fire PRA. The guidance is provided in ML14017A135, "Enclosure 1 - Supplement to Interim Guidance Pending Publication of Expert Elicitation Results." These updated hot short probabilities have been incorporated into the Farley Fire PRA and therefore, the sensitivity analysis for the credit of a control power transformer is no longer required.

NU_**!EGGI 63860. M,_,_,_l.r mr*-..n.erm...dr . ,__ he* th A~eb, fitier h.o.t. fOFrP*r T*.mpne. 1. t h G-OntAin A GGIA#01 MOWAF tFAROAFMA-F !GPT-ý WithiA th'S GiFnUit FA-449FARGOD T-ASk 10 Elf NUREGIGR 6850. F=9F GOFGUOtS that GGntaiR a GPT, the additieRal faGtOF fGF thee best.

estimate of the hOt 61100 i6 a faGtGF of 2 IoweF than th06e GiFGUffitS Withn' 't :4 P:Fý T-hi6 seRsitiyity aRalysis quaRtifies the impacA eR the Fesults of the hOt 61494 pMbability wor4a tn innrpaaia by -Anmia app-ndipci U=IWP rng: the nwrnmap-A Mf th*R ARRIOnAtonn all ha ckrr+ rrhýk*l*#;ýr_ %&jýrýArUhlýd fnr hjn+h tu car_ of a*muota jw a thjnrQ rjn0+2ýn;n CD:r and,. these..,,," C . r IFsk m Ga"ses deb;~ ,4p;*-eb..... the. ito;,i;.ty . ..... te ,., ,u<^e. gI! -th thn. . 0.... ... thes"e rases .. th--e* he.h~. . ... f ,, b ,and* blyw..... sst.9 . .............

a~ ~ ~ ~ ~ ~ fth~neu ~~~~~~* estvt s thneut ~ay  ! i. nG h P Fee fti sliiiy h W~i 1t~iRACDF Fisk-bned is used. Si *iaFeuti*sgt an e en etd e ll tH9,inal 6et of sGeRaFies. MfteF tHe sen6my" analysis waS GOMP19tett tHIBFe were no a d-d-w t ie mal st-r-Urotsural r-,hange6mad-e- to the fault tFee, an! . efinernents. Th96e F8fiR8M9Rt6 Gaursed the GI)FILERF to be slightly di&FPOýt. UnFn A.A.1hat or., feuRd OR the data-pFeseRted heFe. SiRGe the updateG Fnade WeFe eRly rGeRaFie level FefiReffleRts that did Rot have aR eveFall e#ert eR the aFialysis, the seFisitivity aRalysis eFigiRaIly peFfGFFned

A A k ki F_

t7vd"b am tto ", uOts 0 1-t t, a F"t-P IF4k*. I;

....s senswtivity was GAIV FeV19We_ GA 40.9 --ase analysis anti not the delta

• lFISK. IIoa6eu

_rL.: A B

OR the 5.;95% IRGFease it is jus;mea that the deltaRSK Will RE)t 1RGFea6eFneFe tRaR t

+ A fk f +k .4 N  ; 11 4k; 0% +ý ;f  ; -F D t'-% 1 17 A ar"%Mrl ON 0 M V ca r 0 00 ri W c1t2ow-F rivou or Wr ca v This GenGlurien is based on hve key points. 1) All hot sheFt pr-ebab aIsties weFe doubled, when fGF the PUFPeses of this SeA6*tiVity only the het sheFt ffebabil4ves r.FeditiRg the GPT would FequiFe doubling; and 2) The delta Fi6k GOM i .Aet Of SGeRaFiGr*fFOF:R the base aRaly6is, theFefQF8 using the peFGeRt i;.R.G.FAaA_ý_#OM the total GQF= ir, abeundiRg rimsm -it f4r tha clizitn roRk ýmnnnt

1. m*ll nn*n nn tn* imnq*T *n m~~~~~~~~~~~~~~ l. * *l* I -- V * -i.

• l~l-l 1.H*

.1 -- - -- -A!li -1. W -1l .. i j ....

r_ D R I

I Rev Oil Page V-8

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition CS-BI The Farley breaker Closed The PRA components are not explicitly discussed in a The Farley circuit analysis coordination coordination calculation. Calculation SE-C051326701- calculation, SE-C051326701-002, documentation was 002 is titled NSCA components; however, informal has been updated to address all identified to be review has determined that PRA components are coordination concerns. This update incomplete based on addressed. The information to determine the status of identified two panels that were the Farley Fire PRA coordination for PRA components consists of informal found to not be coordinated; all components being queries and spreadsheets. Supporting Requirement other panels were dispositioned as credited. CS-Bl-01 Category II requires all buses credited in the being acceptable. The two panels Fire PRA to be analyzed for proper over current are N 1R19L00504 and coordination and protection. N2R1 9L00504. Calculation PRA-BC-F-1 1-003 (Cable Selection and Revise calculation SE-C051326701-002 to formally Detailed Circuit Analysis) has been validate that PRA buses are addressed for proper updated to address this coordination and incorporate results into the Fire PRA coordination issue. Based on these model ad needed. conclusions these two panels have been failed in every scenario for the Farley Fire PRA. Associated Circuits Analysis Common Power Supply And Common Enclosure calculation, SE-C051326701-002 has been updated to reflect this update. The Farley Component Selection Report, PRA-BC-F-002, has also been updated to reflect the inclusion of these panels to the UNL list, see Page F-24 and F-1 80, Appendix F.

CS-B1 The Farley breaker Closed E-068 identifies cases where the cable lengths of An analysis was completed that coordination electrical loads were credited to demonstrate selective reviewed the panels that credited calculations use cable coordination for the Cable Spreading room. This cable length as part of the length as part of the assumption is only valid for the Appendix R fire where justification for coordination. The justification for proper the equipment and cables are assumed damaged for entire function of these panels was coordination. This is the entire fire area. Supporting Requirement CS-B1-01 then failed for any fire that impacted not a justifiable Category II requires all buses credited in the Fire PRA to the cable within the identified I

I Rev Ojl Page V-9

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition disposition for use in be analyzed for proper over current coordination and length. Once the length the Fire PRA. protection. requirement was met the function of that cable was the only function Analyze impacted PRA buses for proper coordination failed. See calculations, Associated and incorporate results into the Fire PRA model. Circuits Analysis Common Power Supply And Common Enclosure calculation, SE-C051326701-002, and PRA-BC-F-1 1-003 (Cable Selection and Detailed Circuit Analysis) for more information. A modification is also scheduled to improve coordination for six additional 125VDC load distribution panels per unit. For further information on the modification of these panels see Plant Modifications Committed in Table S-2 of Attachment S.

FQ-A3 Appendix L of NUREG- Closed A non-suppression probability of 3.04E-5 is used for the The Farley MCR analysis has been CR/6850 had been Main Control Room (NSP-0401* basic events). A updated to accurately apply the incorrectly applied to review of the Scenario Development report, the non-suppression factors as the Main Control Board Summary Report, and the MCR Report did not locate appropriate to the Main Control scenarios in the Farley the justification of this probability. Based on discussion Board scenarios. The Farley Fire Fire PRA. The ignition with the Farley team, the values were derived from Scenario Report discusses the frequencies have since NUREG/CR-6850, Attachment L. A review of that scenario development process for been updated to Attachment did not support a NSP below 1E-4 under the the Main Control Board and the use accurately apply best of circumstances. A NSP of 2E-2 (similar to other of Appendix L in section 13.1.2 of Appendix L. NSP events) would make MCR fire the highest. PRA-BC-1 1-014.

contributor to plant risk.

Re-evaluate the NSP used for the Control Room and document the evaluation clearly in one of the reports.

FQ-C1 Possible combination Closed A review of cutsets for different sequences found Every COMBO event is evaluated I

I Rev 0jI Page V-1 0

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition events were missing multiple HRA combinations that are not being replaced and incorporated in the fire PRA. An from the cutset results by a COMBO* event and do not appear to be evaluated updated dependency analysis was for the Fire PRA model. for dependence. One such combination is lOP- completed after the peer review MS032-IH-F and OAR_B_1-----H-F which has a findings were addressed in the combined failure probability of approximately 7E-5. A model. The latest results of the review of the HRA Calculator package supplied shows dependency analysis can be found that no evaluation was performed for this combination of in the Human Reliability Analysis for events. Other HRA combinations could also be missing, Fire Events, PRA-BC-F-11-016.

particularly with new operator actions added for the fire scenarios. HRA dependence could significantly increase cutsets since the rule file makes HRAs independent unless the events are replaced by an evaluated combination.

Review the FPRA cutsets without recovery (all events set to screening values) to ensure that all important combinations are evaluated.

FQ-D1 The CCFP for Farley Closed In Section 3 of the Farley Nuclear Plant Summary The Farley Fire PRA has continued Fire PRA was much Report, Farley reports a CDF of 9.65E-05/year and a to evolve and be refined throughout greater than what the LERF of 1.92E-5/year. This yields a Conditional the analysis. Currently the CCFP is FPIE number was. Containment Failure Probability (CCFP) of 1.99E-01. at a much more reasonable value After continued For the FPIE PRA, the reported CDF was of the order of based on the final CDF and LERF refinement the Fire 3.5E-05/year and the reported LERF was of the order of results. The results and insights PRA CCFP has 2E-07/year. This translates to a CCFP of about 4E-03. related to CDF and LERF can be decreased to a more This is a significant difference, especially when found in the Farley Summary reasonable value as considering that the leading contributor to LERF for the Report section 3 of PRA-BC-1 1-compared with the FPIE PRA, SGTR, is not applicable for fire. This yields 017.

FPIE. inconsistent results.

While the current results may be correct, Farley needs to look at the contributors to LERF to explain the basis for the high CCFP with respect to the FPIE PRA CCFP.

I I Rev 0j1 Page V-11I

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition Farley should look at sequences where the fire not only causes core damage but also directly affects containment integrity. Two likely candidates are sequences that lead to a new ISLOCA scenario and sequences that lead to containment isolation scenarios.

FQ-E1 The Farley Fire PRA Closed The summary report lists and describes significant The Farley Summary report documentation did not contributors to core damage and LERF. The back includes additional details accurately address the references require consideration of analysis issues describing the types of reviews that types of reviews that which are not described in the report as having been were completed on the Farley Fire were performed during done. For example, the back references require a PRA. The type of review and the the scenario cutset review the results of the PRA for modeling consistency, detailed cutset reviews are review sessions. a review of results to determine that the flag event described in section C. 1 of settings, mutually exclusive event rules, and recovery Appendix C in the Summary Report, rules yield logical results, a review of contributors for PRA-BC-F-1 1-017.

reasonableness and a review of the importance results for reasonableness. Appendix F notes that these were accomplished and typically refers back to Appendix C.

Appendix C does not describe these reviews as being accomplished, nor does it describe the results of the reviews. In addition, back Reference D5 requires a review of non-significant cutsets for reasonableness.

Appendix F states that dominant cutsets were reviewed and those that were reduced in frequency to non-significance as a result of the review constitute the review of non-significant cutsets. This does not satisfy the requirement to review non-significant cutsets. Non-significant cutsets generated in the solution of the model need to be reviewed to confirm that their frequency is not underestimated due to modeling errors.

Expand the discussion of model solution and review in the summary report to indicate that required review items have been accomplished.

FQ-F1 Level of detail Closed The documentation of the FPRA results does not The Farley Summary report has I

I Rev 0jl Page V-1 2

Southern Nuclear Operatina Companv Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition describing the risk adequately describe the top risk contributors such that it been updated to reflect the insights significant scenarios is clear why these scenarios, basic events, and human by reviewing the top contributors for was identified as not actions are dominant. Based on other findings (FQ-D1- CDF and LERF. This describes the being sufficient in detail 01 and FQ-E1-01), it is not clear that the Farley team fire induced impacts as well as the understands the bases for these top scenarios. Results random failures. The resolution of presentation is important for PRA acceptability, this finding is found in Appendix C Understanding of the PRA results is necessary for of PRA-BC-F-1 1-017.

performing any RI application to support the plant Provide more detailed discussions of the fire impacts and results to represent a strong understanding of the fire scenarios.

FSS-A2 Target set definition in Closed FNP is missing the basis for not including targets A review of the full room burnout Fire Zones (FZs) that outside the fire compartment for full room burnout scenarios was completed that do not have fire rated scenarios. For full room burnout scenarios, all targets in looked for open boundaries to the boundaries on all sides the fire compartment are included. However, there is no adjoining FZs and the possible as it relates to documented basis for not including targets outside the interactions that could take place.

scenarios that are fire compartment for full room burnout scenarios. If the For some particular fire areas a classified as full room compartment has an opening to an adjacent scenario was postulated that would burnouts. compartment, it was not verified that targets in the fail all targets within the fire area.

adjacent compartment would be outside of the ZOI of all However, in most cases it was the ignition sources in the compartment analyzed for full determined that there was no room burnout. ignition source near the open boundary that would impact targets See F&O PP-B3-01 (F) for a possible resolution. in an adjoining FZ. The Farley Fire Scenario Report includes discussion of the scenario development process for these specific cases in section 3.1.1 of PRA-BC- 11-0 14.

FSS-B2 The Main Control Closed An office workstation fire scenario is discussed in the The Farley Main Control Room Room Abandonment documentation, but is not fully justified. The workstation Abandonment Calculation includes IRev 0#l Page V-1 3

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition calculation identifies fire scenario is potentially the most significant fire the discussion of the workstation the potential for a scenario considered. fire in section B.8 as a sensitivity to workstation fire but the analysis with the results shown does not describe the Provide better documentation of how the workstation fire in Table B-8. NUREG/CR-6850 fire type in significant was modeled and the results of this fire scenario, does not provide any basis for this detail type of fire from an ignition frequency standpoint. Therefore it is not included as one of the potential ignition sources in the base calculation. A review of the sensitivity analysis involving the workstation shows that the analysis is not sensitive to that type of fire given the design of the Main Control Room envelope.

FSS-C1 The Farley Fire PRA Closed Two-point fire intensity model that encompass low The development of fire scenarios does not employ the likelihood, but potentially risk contributing, fire events for the Farley Fire PRA did not use of a two point fire were not used in all cases. Fire scenarios were done identify any instances where further modeling treatment in- with ignition sources characterized with one fire analysis resolution would be gained the development of the intensity, by the treatment as inferred by the fire scenarios, requirements for CC II and CC III.

To reach Capability Category II, use a two-point The implications of retaining the CC intensity model for all ignition sources. I treatment in lieu of refining as described for CC II or CC III is Utility Comment: The development of fire scenarios potentially a higher calculated CDF for the Farley Fire PRA did not identify any instances contribution. The CC I treatment where further analysis resolution would be gained by the inherently will not result in under-treatment as inferred by the requirements for CC II and estimation of fire risk. As such, the CC Ill. The implications of retaining the CC I treatment current treatment is conservative.

in lieu of refining as described for CC II or CC Ill is Provided this treatment does not potentially a higher calculated CDF contribution. The result in masking of risk increases in CC I treatment inherently will not result in under- future applications, further estimation of fire risk. As such, the current treatment is refinements are not considered conservative. Provided this treatment does not result in necessary.

I I Rev 0jl Page V-14

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition masking of risk increases in future applications, further refinements are not considered necessary.

Response: The SR stipulates that a two-point model is required for CC-Il. As you stated in your comment, Farley feels that the one-point model is conservative and justified. This would be viewed as the proposed resolution, but the F&O stands.

FSS-C2 The Farley Fire PRA Closed Ignition source intensity were characterized such that The only readily available reference did not characterize the fire is initiated at full peak intensity and ignition sources for a time dependent growth rate ignition source intensity that are significant contributors to fire risk were not that could be considered in the for a time dependent characterized using a realistic time-dependant fire analysis is 12 minutes as growth rate in the growth profile. Generic methods from the Hughes recommended in NUREG/CR-6850.

scenario development. Associates Generic Fire Modeling Treatments were The treatment would involve a t2 used to characterize ignition source intensity. These growth rate. If a particular generic methods did not incorporate fire growth curves. source/target interaction has a spacing where the target is at the Characterize ignition sources that are significant critical damage spacing threshold, contributors to fire risk using a realistic time-dependant such a treatment may provide some fire growth profile. benefit as successful suppression with that time period would prevent Utility Comment: The only readily available reference target damage. However, if the for a time dependent growth rate that could be target is located well within the considered in the analysis is 12 minutes as calculated damage distance, the recommended in NUREG/CR-6850. The treatment corresponding time to reaching the would involve a t2 growth rate. Ifa particular damage threshold is very short and source/target interaction has a spacing where the target effectively precludes any is at the critical damage spacing threshold, such a meaningful credit for suppression.

treatment may provide some benefit as successful In the case of the Farley Fire PRA, suppression with that time period would prevent target the majority of the target spacing for damage. However, ifthe target is located well within the the dominant risk contributors is calculated damage distance, the corresponding time to such that no meaningful credit for reaching the damage threshold is very short and suppression is available. In other effectively precludes any meaningful credit for dominant risk contributors, the I

I Rev Oil Page V-1 5

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition suppression. In the case of the Farley Fire PRA, the scenario involves high energy majority of the target spacing for the dominant risk arcing fault (HEAF) events were no contributors is such that no meaningful credit for growth time is applicable. The suppression is available. In other dominant risk implications of retaining the CC I contributors, the scenario involves high energy arcing treatment in lieu of refining as fault (HEAF) events were no growth time is applicable, described for CC Il/111 is potentially a The implications of retaining the CC I treatment in lieu of slightly higher calculated CDF refining as described for CC Il/111 is potentially a slightly contribution. The CC I treatment higher calculated CDF contribution. The CC I treatment inherently will not result in under-inherently will not result in under-estimation of fire risk. estimation of fire risk. As such, the As such, the current treatment is conservative, current treatment is conservative.

Provided this treatment does not result in masking of Provided this treatment does not risk increases in future applications, further refinements result in masking of risk increases in are not considered necessary. future applications, further refinements are not considered Response: The Farley modeling was found to be necessary.

consistent with CC-I but did not meet the requirements of CC-Il. The comment provides the basis for stating that the existing treatment is adequate. It does not provide evidence that a time-dependent heat release rate model was used.

FSS-D1 The treatment of Closed The fire modeling tools selected for use are appropriate The Farley Scenario development secondary for evaluating the zone of influence associated with notebook was updated to include combustibles was not individual fixed and transient ignition sources, but do not additional details on how the clearly defined in the provide for estimating fire growth and damage behavior treatment of secondary scenario development for fire scenarios involving ignition and fire spread on combustibles is dealt with during documentation. secondary combustibles. With the generic fire modeling scenario development. Further treatment selected for this fire PRA, there does not information regarding this finding appear to be a way to model fire growth on secondary can be found in section 4.0 of PRA-combustibles. Consequently, the extent of fire BC-F-1 1-014.

development cannot be modeled.

Where secondary combustibles are located within the zone of influence, develop methods for estimating fire I

I Rev 0j1 Page V-16

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition growth on secondary combustibles and the damage caused by this additional fire development.

FSS-D7 The Fire PRA credits Closed SR FSS-D7 requires credited fire suppression systems Supporting documentation has been the in cabinet C02 to be installed and maintained in accordance with included in the Farley Fire Scenario system installed at applicable codes and standards, and the credited report to further discuss the in Farley. There was no systems must be in fully operational state during plant cabinet C02 suppression system documentation operation. These requirements are not met, but fire and the associated test and provided to support the suppression systems are still being credited. As noted inspection procedures that are availability of this in the Conclusions section of Document # 0005-0012- credited in the Fire PRA. It has also system. 002-002-04 (Hughes Associates), "The other main been identified that the system does concern with the systems installed at FNP is the periodic require modifications, such as maintenance and subsequent corrective action. Firstly, mechanical equipment and the plant procedures for inspection, testing and detection upgrades, to be made to maintenance (ITM) do not address a few key activities make the system operable as required by NFPA 12. Secondly, the prioritization of designed. This is found in section work orders sometimes results in extended impairments 8.1.1 of PRA-BC-F-1 1-014.

(e.g., observed CR / work request tags over two years old), which negatively affects the fire protection program objective to maintain working systems." Credit is being taken for fire suppression systems that do not meet the requirements of FSS-D7 for taking this credit.

Verify that credited fire suppression systems are installed and maintained in accordance with applicable codes and standards and demonstrate that credited systems are in a fully operable state during plant operation.

FSS-D7 The non-suppression Closed In Tables 13-1 through 13-12 the equation eA(-lambda*t) The application of the MCR probability that was was used to calculate the non-suppression probability abandonment non suppression originally used to for MCR abandonment scenarios. The control room probability has been re-evaluated calculate the MCR lambda value from Table P-2 was selected. The time, t, using the floor value of 1.OOE-03 for abandonment was obtained through the CFAST runs and plugged into all bins that are determined to reach frequency was un- the equation. In scenarios in which the time to the abandonment threshold. The conservative based on abandonment was greater than 25 minutes a nominal results of this review are identified I

I Rev Oj# Page V-17

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition direction provided in NSP of 0 was selected. A NSP of 0 should not be in section 13 of PRA-BC-F-1 1-014.

Appendix P of NUREG- assumed for these cases. Instead, it is suggested to run CR/6850. the CFAST cases longer than 25 minutes such that the analysis can credit a larger time with no abandonment conditions reached (i.e. ifthe case is ran to 60 minutes with no abandonment conditions reached, t can be credited up to 60 minutes) and still use the e(Iambda*t) equation to calculate NSP.

Additionally, the MCR equipment rooms are normally unoccupied and NSP should be associated with the electrical equipment room vs. the control room. Ifthe control room lambda is used, a basis should be developed why the control room lambda is more appropriate than the electrical cabinet lambda. If the control room lambda basis has been justified, then a sensitivity analysis should be performed using the lambda of electrical fires. This calculation can be non-conservative.

FSS-D8 The Farley Fire PRA Closed Note 8 associated with SR FSS-D8 suggests The Fire PRA was first developed does not look at the consideration of the time available to suppress a fire without credit for suppression or time available for a prior to target damage and specific features of physical detection, the target set for a given suppression system to analysis units and fire scenarios under analysis that scenario was based on the ignition successfully suppress a might impact suppression system activation and source type. Further in the analysis fire before target coverage. Such consideration is not documented. credit for the existing detection and damage. Credit is taken for automatic fire suppression in some suppression, and in some cases scenarios without consideration of the factors required plant modifications, systems were under this SR. credited. For these cases where the credit was taken the target set Perform an analysis that considers the time available to was not changed based on the time suppress a fire prior to target damage and the specific to suppression or distance to target.

features of the PAUs and fire scenarios under analysis Instead a conservative approach to determine what impact they have on suppression was taken to leave the original system activation and coverage, target set included in the Fire PRA I

I Rev 0jl Page V-1 8

Southern Nuclear Operating Company ScnAttachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition along with the failure rate of the suppression system, therefore not requiring a review of damage time vs. suppression time. This is found in section 8.0 of PRA-BC-F-1 1-014.

FSS-E3 The Farley Closed Supporting requirement E3 asks to provide a mean The documentation has been documentation did not value of, and statistical representation of, the uncertainty updated to include discussions address the uncertainty intervals for the parameters used for fire modeling the related to the uncertainty for fire related to the use of fire fire scenarios. Farley performed fire size and heat modeling. See Table D-1 of the modeling for the fire release rate selection in accordance with NUREG/CR- Farley Fire PRA Summary report, scenarios. 6850 and/or applicable FAQs. However, the methods for PRA-BC-F-1 1-017.

developing the statistical representation of the uncertainty intervals and mean values currently do not The associated SR was exist. However, this is not reported in the dispositioned as CC I which is documentation. judged to be sufficient given the two concerns noted.

In the documentation, explain that it is understood that methods for developing the statistical representation of the uncertainty intervals and mean values currently do not exist.

Utility Comment: This specific F&O was issued against a technical element and the indicated resolution involves a documentation clarification. This documentation clarification will be implemented.

FSS-F1 The exposed structural Closed Section 2.11 of the FNP Summary Report A review and analysis was steel evaluation was (FNPSummaryReport-final.pdf) claims that, "The completed of the structures at not originally performed Structural Steel Evaluation performed to evaluate the Farley for both units to determine as part of the Farley potential for fire to impact structural steel capacity which the amount of exposed structure Fire PRA. could impact fire compartment boundaries is steel that is susceptible to fire documented in the FNP Fire PRA Report PRA-BC-F damage and ultimately leading to a 014, Rev. 0, Fire Scenarios Report [5]." This building collapse. The analysis documentation was not found in the referenced report. concluded that there is a potential Include in the Fire Scenarios report the structural steel for this scenario to occur in the I

I Rev 0jl Page V-19

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition evaluation identified in final Summary Report and Turbine Building. This scenario has update self-assessment. been added and is accounted for in the total plant risk and delta risk calculations. This is found in section 10.5 of PRA-BC-F-1 1-014.

FSS-G6 The Farley Fire PRA Closed The Multi-compartment analysis identifies several areas The Farley Fire PRA MCA analysis MCA analysis was where further evaluation is required. This evaluation has been completed with all incomplete at the time has not been completed to either screen the zone or scenarios being evaluated. The of review with many develop a fire scenario based on multi-compartment fire. HGL/MCA report has been updated open items. A screening of the multi-compartment scenarios were to show the final results for the done, those that were screened out were not included in analysis. This is found in the quantification. The multi-compartment scenarios Attachment B of PRA-BC-1 1-015.

flagged for further evaluation are in Table 3-1 of the Multi-Compartment Analysis. Further evaluation is still being worked on, so these scenarios have not been included in quantification. Given the current CDF, the MCA could increase risk above 1E-4/yr.

Complete the MCA to either quantify the PAUs where a fire could spread to an adjacent PAUs or screen the PAUs for MCA FSS-H1 Non-Fire PRA targets Closed For fire scenarios considered during the peer review The scenario development were removed from the walkdown, the nature and characteristics of the damage database has been re-populated database. Leading to target set were different in three different sets provided with all target set information, inconsistencies for review, including the computer printout of the fire targets specifically modeled in the between the scenario scenario summary and two sets of walkdown notes. Fire PRA and those that are not.

development sheets One consistent set of documentation should be The scenario printout sheets found and what was identified maintained in a retrievable format. in Appendix A of the Fire scenario in the field. development report contain all Include all relevant target sets in the computer-based targets identified during the walk documentation and handle by disposition those targets down phase regardless of the that are not risk significant for a particular scenario, relationship to Fire PRA components. This is found in Appendix A of PRA-BC-F-1 1-014.

I I Rev Ojl Page V-20

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition FSS-H5 The Farley Closed The generic fire modeling tool referenced in the Fire The documentation has been documentation did not Scenario Report, Reference 6 (Hughes generic updated to include discussions address the uncertainty treatment) is used for generic treatment of ignition related to the uncertainty for fire related to the use of fire sources as an approach to bound many scenarios, but modeling in response to SR FSS-modeling for the fire its use does not provide uncertainty treatment on a fire E3. See Table D-1 of the Farley scenarios, scenario basis. Fire PRA Summary report, PRA-BC-F-1 1-017.

Provide uncertainty evaluations at least generically for those scenarios that use the generic treatment tools and on a case by case basis for the sources that use additional detailed fire modeling to further describe the scenarios used.

Utility Comment: This specific F&O is inconsistent with F&O FSS-E3-01. The indicated resolution for FSS-E3-01 states in part that the analysis documentation should be enhanced to note that methods for developing the statistical representation of the uncertainty intervals and mean values currently do not exist. However, F&O FSS-H5-01 then asks to undertake evaluations to address uncertainty. These this latter F&O should be revised so that it is consistent with FSS-E3-01.

Response: The F&Os address the specific SR requirements. The response to F&O FSS-E3-01 may be used to justify the treatment of uncertainty for FSS but the F&O documents compliance with the standard and as such remains.

IGN-A7 Newly installed Closed During the walkdown - ignition sources (specifically The Farley Fire PRA has been in potential Ignition electrical cabinets) were found in the plant that is not development for some time. The sources were identified listed on the list of ignition sources for the particular ignition source walk down and in the field that were PAU. Specific examples include N1 R1 L0001 in the scenario development were some of not included as part of cable spreading room and N1R15AO02X and the first tasks that were completed the original scenario N1 R5A003X in the switchgear room. A walkdown as part of this analysis. A qualitative IRev 0jl Page V-21

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition count for this bin.

IGN-A7 The Farley scenario Closed During the walkdown of the Bravo 4160 Switchgear The Farley Fire PRA has been development did not room, it was observed that the Foxtrot Switchgear was updated to accurately correct the accurately account for split between 2 PAUs. The switchgear had a count of scenario development to account the frequency split 15 vertical sections. PAU 335 had a count of 8 for the ignition source split between between the two fire switchgear vertical sections and PAU 343 had a count of the two fire zones of the SWGR zones as it was 8 vertical sections. This is a clear example of room. The ignition source count of identified in the field. inadequate PAU boundary. the SWGRs has not been changed to reflect the accurate number of Recommend that the PAU such that the Foxtrot cubicles. This change would result switchgear is contained in one PAU and the count of the in a non-significant impact to the entire switchgear should be 15 vertical sections. In total plant ignition frequency based cases where ignition sources have been split between on the total count for Bin 15. The PAUs the count should be verified correct. ignition frequency for the scenarios related to the SWGRs are accurately represented. These updates can be found in the Farley Scenario report, Appendix A, PRA-BC-F- 11-014.

IGN-A9 The transient factors in Closed PAU 2321 (Sample Panel Room) has a transient fire The transient ignition frequency the ignition frequency frequency of zero. Similar to the first page of Appendix allocation has been re-visited for the development had B, a storage factor of "low" or 1 should be chosen such Farley Fire PRA based on this identified fire zones that that 2321 has a non-zero transient fire frequency. Right finding. The appropriate changes had a 0 factor which led now 2321 has a non-zero ignition frequency due to a have been made to accurately to a frequency of 0. small number of cable in the area filling Bins 11 and 12. reflect the transient ignitions sources located within each fire A non-zero transient factor should be filled in. zone. These updates were made in Farley Plant Partitioning and Ignition Source Task 1 and 6 report, PRA-BC-F-1 1-003, and carried into the ignition source calculation for the scenario development, PRA-BC-F-11-014.

I Rev 0j1 Page V-23

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition development, and/or review of plant modification is necessary to review of the panels identified ensure the plant FPRA reflects the as built as operated during the peer review walkdown configuration. showed no significant change in the plant CDF. This is based on the fire This issue may be due to new plant equipment that was zones these panels were located in added after the initial ignition frequency walkdown - and the level of scenario nevertheless the fire PRA should be reconciled to development already included in include these new ignition sources. these fire zones. The panels are located in a part of the room that already contains detailed scenarios and the introduction of the new sources are not expected to change the target set of adjoin scenarios.

Section 3.5 of the Summary Report, PRA-BC-F-1 1-017, provides steps that will be taken to account for changes in the plant design that have occurred since the initial Fire PRA development.

IGN-A7 The yard transformers Closed Bins 27-29 have not been filled. Large Yard The Farley Fire PRA task 6 has had been incorrectly Transformers have been incorrectly binned in Bin 23 been updated to accurately binned during the Task ("indoor transformers"). It is clearly stated in represent the transformers located 6 development and NUREG/CR-6850 that large yard transformers are not in the yard to their applicable bins should be moved to part of this count. As a result each large outdoor and have been removed from bin their appropriate bins. transformers (MT, UAT, SuT) should be binned in both 23. The frequency per component Bin 27 (Yard Transformer - Catastrophic) and Bin 28 has been updated accordingly and (Yard Transformer - Non Catastrophic). Additionally, used for the applicable scenarios.

Bin 29, Transformer Yard - Others, should also be filled. See Appendix C of Plant Partitioning and Fire Ignition Since Bin 23 may have been misinterpreted, it is Frequency for Farley Fire PRA, suggested that indoor transformers typically associated PRA-BC-F-1 1-009.

with essential lighting, etc. be looked at for applicability in the FPRA if not already evaluated. Indoor transformers over 45kVA should be included in the IRev 0# Page V-22

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition PP-B3 The Farley Fire PRA Closed SNOC has not provided sufficient evidence that Fire The Farley scenario development did not contain Zone PAUs were evaluated for fire resistance report has been updated to provide sufficient information on capabilities of barriers, nor was there sufficient evidence more details on the scenario scenario development that credited spatial separations were analyzed. development based on the ignition with respect to the Specific examples are cited in PRA-BC-F-1 1-001, source and target identification crediting of fire barriers. Section 2.2, for PAUs that use "natural divisions." The process. This can be found in the document cites that the lack of fire barriers between Farley Scenario Development these PAUs will be evaluated during the MCA. report, PRA-BC-F-1 1-014, section However, the MCA analysis appears to only discuss hot 3.1.1. The impact on the Hot Gas layer issues, and does not consider whether a fire Layer and Multi-Compartment propagates outside of the PAU or ifthere is a zone of Analysis has also been revisited to influence and target damage outside of the PAU. assure that the boundaries of the Another example of where spatial separation is credited rooms have been adequately is Tool Room 0441. represented in the calculation of the volumes.

Full room burnout scenarios are developed and quantified, but without sufficient evidence that fire barriers or spatial separation issues have been evaluated. It appears that specific PAUs are screened from having multi-compartment impacts without consideration of fire propagation or ZOsimpact across spatial divisions.

SNOC has presented a plan to resolve the Fire Zone PAU vs. Fire Area PAU issue. Implementation of this plan is sufficient to address the issues identified in PP-B2 and PP-B3. In the plan, Fire Areas will be treated as PAUs. Particularly, SNOC staff have acknowledged that for "full burn" and "base case" fire scenarios, they will review and document the capabilities of barriers and the appropriateness of credited spatial separations, and will not inappropriately credit barriers or spatial separations for fire scenarios. The plan includes the following:

1. Those APs that have one or more boundaries I

I Rev 0j1 Page V-24

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition that are not physical features or are not rated fire barriers will be identified and a requirement will be added to clarify that this must be recognized in the development of fire scenarios.

There will be confirmation that the results of the above have been observed and documented.

2. Enhance the documentation to acknowledge the crediting of non-rated physical boundaries and provide a basis recognizing that the justification will rely on physical observations during plant walkdowns or through equivalent means as well as general construction methods (masonry block wall, concrete walls, etc.).

3. Address the nature and consequence of anticipated fire events for all APs for which explicit fire scenarios are not developed (base cases) and confirm that the results are appropriate given the boundaries for the AP.

4. Confirm that bounding room burn-out cases are not used for any APs that are not fully bounded by physical fire barriers, and that there is a justification for crediting those physical barriers.

5. Confirm that the resulting analysis does not change (reduce) the level of resolution associated with the existing fire scenarios developed to support the requirements of SRs associated with FSS.

Modify the hot gas layer and multi-compartment analysis (MCA) so that any unnecessary conservatism caused by using a smaller volume artificially caused by an I

I Rev 0jl Page V-25

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition assumed AP boundary are removed.

PP-C3 The Farley Fire PRA Closed Plant personnel have given verbal assurance that plant The Farley Task 1 and 6 report did not contain walkdowns have been performed to confirm the plant identifies the ignition sources sufficient information on partitioning boundaries. It is reasonable to presume that identified in each fire zone. The scenario development the fire protection engineer would perform this walkdown results of the walkdowns are input with respect to the task. In addition, walkdowns were performed to support into a database that contains the identification of fire the Fire PRA ignition frequency task. Furthermore, necessary information related to barriers. some notes were found as further evidence that some Task 1 and 6. This database is walkdowns were performed. However, documentation considered to be the controlled copy of the plant partitioning walkdown is not readily available of the results of these tasks. These for peer review. SR PP-C3 requires documentation of results are found in Appendix D of key or unique features of the partitioning elements for report PRA-BC-F-1 1-009. Section each physical analysis unit. SR PP-B7 requires a 3.1.1 of the Farley Scenario Report, confirmatory walkdown of partitioning elements. PRA-BC-F-1 1-014, describes the process of identifying applicable Include Plant Partitioning walkdown sheets as part of scenarios based on the ignition PRA secondary documentation, and refer to the source, surrounding targets and fire walkdown sheets in PRA-BC-F-1 1-001, Farley Fire PRA barriers.

Tasks 1 & 6, Plant Partitioning and Fire Ignition Frequency. In particular, fire barriers and spatial separations that are credited in fire scenarios should be validated. When where no prior documentation can be found, new walkdowns may be required.

PP-C3 The Farley Fire PRA Closed Fire Zones are identified as Fire PRA plant analysis The Farley Fire PRA documentation did not contain units in PRA-BC-F-1 1-00. Fire PRA staff have has been updated to be consistent sufficient information on expressed that the Fire Areas, not Fire Zones, should be in the naming convention scenario development assessed as the PAUs. However, the Fire Zone PAU throughout the analysis concerning with respect to the form the basis for initial PAU ignition frequency, whole the use of PAU and fire zone. The documentation of fire room bums, and initial screening in later PRA analysis 'rooms' at Farley are considered fire barriers. Fire Zones as PAUs are used consistently and zones, while the fire areas are extensively in the FPRA documentation. There is a considered PAUs. The Task I and disconnect between the PAUs defined in PRA-BC-F-1 1- 6 report, Plant Partitioning and 00 and SNOC staffs statements of what constitutes a Ignition Frequency PRA-BC-F-1 1-I I Rev 0jl Page V-26

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition PAU. This adversely affected the review of the Plant 009, Cable selection and Detailed Partitioning technical element. SNOC desires to call the Circuit Analysis PRA-BC-F-1 1-003, entities that are currently described as Fire Zone PAUs and the Farley Scenario report, as Administrative Partition, and to treat Fire Areas as PRA-BC-F-1 1-014 contain this PAUs. clarification.

F&O PP-B3-01 identifies an acceptable plan to address the technical issues around the definition of PAUs, that Fire Areas, not Fire Zones, form the basis for PAUs.

Fire Zones and similar entities will be identified as "Administrative Partitions" (AP). Since the term "Physical Analysis Unit" or PAU is extensively in Fire PRA documentation to describe Fire Zone PAUs, all Fire PRA documents should be reviewed and revised to call these compartments Administrative Partition.

Furthermore, the term "Administrative Partition" (AP) should be defined in the PP documentation and the APs descriptions (formally, Fire Zone PAUs), should be retained.

PRM-B2 The Farley internal Closed Internal Events PRA peer review exceptions and Table. 1 of Fire PRA logic events finding had only deficiencies have only partially been dispositioned. Development, PRA-BC-F-1 1-004 been partially Table 1 of the Fire Model document (PRA-BC-F-1 1- has been updated to address all addressed in respect to 004_V~a) lists some of the internal events findings, but internal events PRA findings and the impact on the Fire not all. All findings included in the internal events peer their impacts on the fire PRA.

PRA. review must be included and disposed in the PRM notebook. Disposition of findings could not be verified.

Discussion with Southern Company personnel indicated that some of the findings had not been addressed.

Expand Table 1 of the Fire Model document to include all findings. Describe the impact of the finding on the fire PRA. For those that impact model elements applicable to the fire analysis, describe the resolution in I

I Rev 0jl Page V-27

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status Finding/Observation Disposition sufficient detail to allow a reviewer to conclude the finding has been dispositioned.

PRM-C1 The RCP shutdown Closed The new RCP shutdown seals are included in the fault Fire PRA has been developed seals were not tree model but are not described in Appendix B. based on internal events PRA adequately discussed Appendix B should be revised to describe these new model having model of RCP in the documentation seals and their impact on RCP seal failure flow rate. shutdown seal. Section 2.0, for the Fire PRA model The Fire PRA modeling pertaining to RCP seal failure is Appendix B of Fire PRA logic development, not adequately described in PRA-BC-F-1 1-004. Development, PRA-BC-F-11-004 has been updated to add RCP Revise Appendix B to describe the new shutdown seals shutdown seal modeling.

and their impact.

UNC-A1 The Farley fire PRA Closed Farley presents the CDF results in Section 3.0 of the Appendix D of the Farley Summary provided Train A and B Summary Report. The way the results are presented report, PRA-BC-F-1 1-017, has been CDF results but did not are as an annualize CDF for Train A operating and an updated with a revised parametric define total plant CDF. annualize CDF for Train B operating and both are called uncertainty analysis for both CDF The parametric total plant CDF. There is no discussion as to what these and LERF for Train A and B uncertainty analysis two CDF values meant or a value for the "true" plant individually. The quality of the should be more specific CDF. In Appendix D of the Summary Report, Farley analysis was improved by applying in scope and use a presents the results of their parametric uncertainty the Monte Carlo method with greater sampling size. analysis for CDF. Although not documented, this 50,000 samples. The resulting appears to be for CDF related to Train A Operating only. curves are well behaved and the The parametric uncertainty analysis was performed calculated means show minimal using the Latin Hypercube method with only 1000 difference when compared to the samples. The resulting curve was not well behaved and point estimates.

the calculated mean is well below the point estimate in Section 3. Discussion of how the total plant CDF/LERF is calculated is also As a start, Farley needs to define what the two results, provided in the Summary Report.

annualize CDF for Train A operating and an annualize This describes how the Train A and CDF for Train B operating, mean and a single total Plant Train B results are averaged CDF needs to be presented. This will probably be the together to obtain the total plant average of the original two values. For the uncertainty CDF/LERF.

analysis, Farley needs to document what is covered by the analysis, Train A results, Train B results or both.

I I Rev 0jl Page V-28

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-1 Fire PRA Peer Review - Facts and Observations SR Topic Status FindinglObservation Disposition Farley did run an uncertainty case using 10,000 samples and the results seemed to be better behaved.

Farley is running an uncertainty case with 50,000 samples which is consistent with their FPIE PRA process. The results of this analysis should be presented in Appendix D in the Summary Report instead of the current analysis.

UNC-A1 The Farley Closed Farley did quantify the fire-related LERF for Unit 1 but The Farley Summary report has documentation did not failed to meet the requirements from LE-F2 and LE-F3 been updated to reflect the insights adequately address the from Section 2 which require that "REVIEW contributors by reviewing the top contributors for review of LERF for reasonableness (e.g., to assure excessive LERF. This describes the fire scenarios in the conservatisms have not skewed the results, level of induced impacts as well as the

.analysis to show that plant-specificity is appropriate for significant random failures. The resolution of the appropriate reviews contributors, etc.)" and "IDENTIFY and characterize the this finding is found in Appendix C had been completed. LERF sources of model uncertainty and related of PRA-BC-F-1 1-017.

assumptions in a manner consistent with the applicable requirements of Tables 2-2.7-2(d) and 2-2.7-2(e)." As discussed in F&O FQ-D1-01, the calculated LERF and CCFP indicate that there some potential issues with the LERF calculation.

See F&O FQ-D1-01 (F) and perform the reasonableness reviews after requantifying I Page V-29 I IRev Oji 0jl Page V-29

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-2 Fire PRA- Category ISummary1 SR Capability Topic Status Category FSS-C1NIA CC-I [FSS-Cl-01]

Two-point fire intensity model that encompass low likelihood, but potentially The development of fire scenarios for the risk contributing, fire events were not used in all cases. Fire scenarios were Farley Fire PRA did not identify any instances done with ignition sources characterized with one fire intensity, where further analysis resolution would be gained by the treatment as inferred by the To reach Capability Category II, use a two-point intensity model for all ignition requirements for CC II and CC Ill. The sources. implications of retaining the CC I treatment in lieu of refining as described for CC II or CC III Utility Comment: The development of fire scenarios for the Farley Fire PRA is potentially a higher calculated CDF did not identify any instances where further analysis resolution would be contribution. The CC I treatment inherently will gained by the treatment as inferred by the requirements for CC II and CC III. not result in under-estimation of fire risk. As The implications of retaining the CC I treatment in lieu of refining as such, the current treatment is conservative.

described for CC II or CC III is potentially a higher calculated CDF Provided this treatment does not result in contribution. The CC I treatment inherently will not result in under-estimation masking of risk increases in future applications, of fire risk. As such, the current treatment is conservative. Provided this further refinements are not considered treatment does not result in masking of risk increases in future applications, necessary.

further refinements are not considered necessary.

Response: The SR stipulates that a two-point model is required for CC-Il.

As you stated in your comment, Farley feels that the one-point model is conservative and justified. This would be viewed as the proposed resolution, but the F&O stands.

FSS-C2 CC-I [FSS-C2-01]

Ignition source intensity were characterized such that fire is initiated at full The only readily available reference for a time

• peak intensity and ignition sources that are significant contributors to fire risk dependent growth rate that could be were not characterized using a realistic time-dependent fire growth profile. considered in the analysis is 12 minutes as Generic methods from the Hughes Associates Generic Fire Modeling recommended in NUREG/CR-6850. The Treatments were used to characterize ignition source intensity. These treatment would involve a t2 growth rate. If a generic methods did not incorporate fire growth curves. particular source/target interaction has a spacing where the target is at the critical Characterize ignition sources that are significant contributors to fire risk using damage spacing threshold, such a treatment a realistic time-dependent fire growth profile. may provide some benefit as successful suppression with that time period would Utility Comment: The only readily available reference for a time dependent prevent target damage. However, if the target I Page V-30 I

I Rev Oji Rev.Ojl Page V-30

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-2 Fire PRA- Category I Summary1 SR Capability Topic Status Category growth rate that could be considered in the analysis is 12 minutes as is located well within the calculated damage recommended in NUREG/CR-6850. The treatment would involve a t2 growth distance, the corresponding time to reaching rate. If a particular source/target interaction has a spacing where the target is the damage threshold is very short and at the critical damage spacing threshold, such a treatment may provide some effectively precludes any meaningful credit for benefit as successful suppression with that time period would prevent target suppression. In the case of the Farley Fire damage. However, if the target is located well within the calculated damage PRA, the majority of the target spacing for the distance, the corresponding time to reaching the damage threshold is very dominant risk contributors is such that no short and effectively precludes any meaningful credit for suppression. In the meaningful credit for suppression is available.

case of the Farley Fire PRA, the majority of the target spacing for the In other dominant risk contributors, the dominant risk contributors is such that no meaningful credit for suppression is scenario involves high energy arcing fault available. In other dominant risk contributors, the scenario involves high (HEAF) events were no growth time is energy arcing fault (HEAF) events were no growth time is applicable. The applicable. The implications of retaining the implications of retaining the CC I treatment in lieu of refining as described for CC I treatment in lieu of refining as described CC Il/111 is potentially a slightly higher calculated CDF contribution. The CC I for CC Il/111 is potentially a slightly higher treatment inherently will not result in under-estimation of fire risk. As such, calculated CDF contribution. The CC I the current treatment is conservative. Provided this treatment does not result treatment inherently will not result in under-in masking of risk increases in future applications, further refinements are not estimation of fire risk. As such, the current considered necessary. treatment is conservative. Provided this treatment does not result in masking of risk Response: The Farley modeling was found to be consistent with CC-I but increases in future applications, further did not meet the requirements of CC-Il. The comment provides the basis for refinements are not considered necessary.

stating that the existing treatment is adequate. It does not provide evidence that a time-dependent heat release rate model was used.

FSS-E3 CC-I [FSS-E3-01J Supporting requirement E3 asks to provide a mean value of, and statistical The documentation has been updated to representation of, the uncertainty intervals for the parameters used for fire include discussions related to the uncertainty modeling the fire scenarios. Farley performed fire size and heat release rate for fire modeling. See Table D-1 of the Farley selection in accordance with NUREG/CR-6850 and/or applicable FAQs. Fire PRA Summary report, PRA-BC-F-1 1-017.

However, the methods for developing the statistical representation of the uncertainty intervals and mean values currently do not exist. However, this is The associated SR was dispositioned as CC I not reported in the documentation. which is judged to be sufficient given the two concerns noted.

In the documentation, explain that it is understood that methods for developing the statistical representation of the uncertainty intervals and mean I Page V-31 I Oji I Rev 0j# Page V-31

Southern Nuclear Operating Company Attachment V - Fire PRA Quality Table V-2 Fire PRA- Category I Summary1 SR Capability Topic Status Category values currently do not exist.

Utility Comment: This specific F&O was issued against a technical element and the indicated resolution involves a documentation clarification. This documentation clarification will be implemented.

- All Fire PRAs SRs characterized as Capability Category I were identified as Findings in the Peer Review. Rofer to Table V 1 for id..ti,,catan .ndrcc'-lutin of ,'ndings.

I Page V-32 I IRev Oji Rev 0jl Page V-32