Text

Staff Evaluation of Oyster Creek Individual Plant Evaluation (IPE)
	ML060550282
Person / Time
Site:	Oyster Creek
Issue date:	08/02/1994
From:	Dromerick A; NRC/NRR/DLPM/LPD1
To:	J. J. Barton; GPU Nuclear Corp
	Reckley W, NRR, 415-1323
Shared Package
ML060550281	List: ML060550282; ML060550283; ML060550285; ML060550287; ML060550290;
References
	NUDOCS 9408080092
	Download: ML060550282 (107)
	v • d • e

{{#Wiki_filter:4. 0, A- ? -t: UNItE

STATE

'NUCLEAR REG UATO COMMISSION o

WAi4INl9t0N,.t

,/ Au °*

egust e~94!.

Docket No. 50-219 Mr. John J. Barton Vice President and Director CPU Nuclear Corporation 0 Oyster Creek Nuclear Generating oteaitilon- -.. Post Office Box 388 Forked River, New Jersey 08731 -. K: Dear Mr. Barton:

SUBJECT:

REVIEW OF OYSTER CREEK NUCLEARAGCNERATING STATION INDIVIDUAL PLANT EXAMINATION (IPE) SUBMITTALC(TAC NO. M74443) Enclosed is the staff's evaluati on of GPL J:tWjear Corporation's (GPUN) Oyster Creek IPE for internal events and Anternab flood. The evaluation package consists of: a Staff Evaluation'Aleport (SEA (Enclosure 1); and contractor Technical Evaluation Reports (TERs) for 'ethor.ontend .back-end, and human reliability analysis reviews (Enciosures and 4). Based on our review, we conclude tlhat GPUIJthas imet, the intent of Generic Letter 88-20, and we do not recommend that-'ICXfurther review be conducted. However, our review identified a deficiency lack of treatment of pre-initiators) in the human reliabil ity anaays i port:ionof the IPE which may limit the IPE's usefulness in other applicattions,. In addition, GPUN plans to address a number of potential operator mitt4ition actions during its accident management development phase, spelfical.ly-the need for the interconnection between the fire protection water stystem,a4the tdrywell spray system. We would also:like to mention thatn GPUN di(ot explicitly state that they plan to maintain their ProbabillstlCLRisk 4A$"'f.essment'(PRA) "living." The staff notes that a "living" PRA dould.enhance plant, safety and provide additional assurance that any potentially unrecognized vulnerabilities would be identified and evaluated during, the'11 ie f the. plant. 94080804092 940802 PDR ADOC, 0500021 P PDR' -m f -tSi'> '> 0' ~ ' - : f

Mr..lohn J. Barton By this letter we are closing TAC No. S Al

Pr

1) 1

.nc 1osjres.

1. Staff Evaluation Report

2. tI R (Front-fnd)

3. UR (Back-Fnd)

4. TER Human Reliahil itty Analysis cc w/pnc 1 oswre i Sf-'

nflxl page P.I JRj.8jjL.LI. w/enclosures 1. 2, 3, 4: Docket Ii le NRC S Local PDRs PDI-4 Plant ADromerick Riernan f lois, RES

JRidgely, RES

JChuong, RES 2-M744434i nc ry exander Wdrmerick, Sri Project Manager ojict Director'ate 1-4 V i tion of%~tor Projec ts - I1/

I fite b I ic a r ha~ttr Regulation i P.] 51RO VJ 2Q.tL SV Va rq a JC ilIvo SNorr i % OGCI ACRS ( 10)

JRogcje, RGI w/pnclosure I' O

tI t lA:PDI-4/ PM:POI-4 D:_Q_-4_: NAM( SNri rnk c k: bP '1S to% E i< -l DATE _\\ _/94 5.// 94 j_______: !.N .iLA P4 A S f ri s i X ' I k~ I t o _I _ OFFICIAL RECORD COPY Document Name: G:\\DROMERIC\\M74443.CLO:-:: I I I q t' ' , : - ' 4 i , I t I .. ~ !l

1"11.1=

=i-

Mr. John J. Barton 2 - By this letter we are closing TAC No.. 174443.-. S i fricer. '..',:i '.9. /g.#.<gc Alexa derW.Dromer c Sr. Project Manager r.j: D~rectorate 1-4 Ohisin df ReActor Projects I/II Offic. M4f Nuc1ear Reactor Regulation

1. -vff Evaluation Report

2. TER (Front-End)

3.

TER (Back-End)

4. TER Human Reliability Analysis cc w/enclosure I:

See next page i

Mr. John J. Barton Vice President and Director cc: Ernest L. Blake, Jr., Esquire Shaw, Plttman, Potts & Trowbrldge.: 2300 N Street, NW. Washinqton, DC 20037 Regional Administrator, RegiOn I U.S. Nucleatr Regulatory Commission : 475 Aliendale Road King of Prussia, Pennsylvania 19406 KWR Licensing Manager GPU Nuclear Corporation I Upppr Pond Road Pdrsippany, New Jersey 07054 M ayor ltacey Township 818 West Lacey Road forked River, New Jersey 08731 1 1icens ing Manager oyster Creek Nuclear Generating Station Mail Stop: Site Emergency Bldg. I Post Office Box 388 forked River, New Jersey 08721 -Oyvter.Creek, Nuclear dehera;t in'g Station Resident ihspector c/Qt~'~'Ii.§ Nucleae Regulatory Commission PodOU".ffice Box'445 ./. V r FovrK.d ovr, New Jersey 08731 KehtToch Chief New 4eriey Department of

E"iirbnwetal Protection BuO eru of Nuclear Engineering CN41

~Tireon, NewJersey 08625 Ja. k S. Wetmore TM!.icensin§ Manager cPU Nuc ear Corporation Post IOffice Box 480 Middletown, Pennsylvania 17057

22..

i~I 11 1 i-ko 0 1 Nl-~

I e

,-'R D: ENCLOSURES I f f

. 0 .5.',,#s 1. X .:,. s. i E . i ,'1 .f X.,.,5,},D, STAff EVAlUATION OF t9£ QYSTER CRfEK INDIVIDU.AL PLANf :£XAMINATtON n 0;.(IPr@;,,, ( INTEItNAl: EYE+ONlY} 7

, ,," : ' f ' X 0

f.,;, f ,'.:d. .:.' ! tA, 0 ff fi,: f;. i, w. t; d .s', 7 iC0S' : -'S:: f t: X,' X; j-i S s i,.. S X,. v 0 a;\\e 0 .s, . J. . f, :'..';, S 0 \\ t i, SS,; . D.rYSH. ;:- II

TABLE OF CQNTENT5 i EXECUTIVE

SUMMARY

BACKGROUND.. 4

11.

STArF'S REVIEW.............. ..5

1.

licensee's IPE Process,.;J4.. 4 5

2.

Front-End Analysis.....

4.

6

3.

Back-End Ahalyis. t ...... 9

4.

Human Factor Consideratiof t. 12

5.

Containment Performance lmprovement5 (01)..;.,;... 15

6.

Decay Heat Removal (OHf) Evalujtion....,,...,,, 17

7.

Generic Safety Issues7. 4. *18

8.

Licensee Actions and Commitments from the t 18 III. CONCLUSION...... 19 APPENDIX OYSTER CREEK DAtA

SUMMARY

SHEEt...21 fi-
sc

EXECUTIVE 'oiMMAR The NRC staff completed Its review of the `JInternal event portion of the Oyster Creek IPE submittal and associated documetitation which includes GPU Nuclear Corporation's (GPUN/llcensee) responses to sttaff generated questions and request for additional information. The lirensee's [PE is based on a'Level 1. a '2 Probabilistic Risk Assessment (PRA) consistent With Generic Letter 88-20,'>Appendix'l. The PRA was performed by PLG Inc., with the support from.other4.eIonsul~tants. GPUN personnel familiar with: detail design, controls, procedures','hand systems maintained involvement in the development, analysis, and technici1 ireviews of the Oyster Creek PRA models. The Oyster Creek IPE did not identify any. severe accident vulnerabilities which the licensee defined as any core damage sequence that exceeds IE-4 per reactor year, or containment bypass that exceeds lE-6 per reactor year. The iPE did, however, take credit for a number.of modifications that were installed during the 14R refueling outage. These include the interconnection to the combustion t:rbine generators at the-6adjacent Forked River Site; hard piped containment vent system; ahd operator training for manual initiation of the containment spray system. The IPE estimated the total mean core damage'frequency (COF) from internal events including internal flood as 3.96E-6/yr. Dominant initiating events and their percent contribution (%) to COF Include loss of offsite power (32.8X.). turbine trip (13.1%), and reactor trip (7.7%). IPE importance measures identified failure of electromatic relief :Vlves (EMRV) to close as the largest component contributor to total CDFT(48%). the significance of this contributor stems from tie success; criteria"-.which requires (for many accident initiators) opening and subsequent. closing o.f up to 4 of!S EMRVs. Essential AC power bus failures had also been found -tobe an important contributor (37%) to core damage. In addition, a numbereof the OC IPE dominant sequences involve loss of DC power. DC power is required to'.a'ctivate the isolation condenser, and open the EMRVs to allow for yessel injedlon with-the low pressure firewater system. The OC IPE found a relatively low.station blackout induced core damage frequency of 7.7E-7/yr. The IPE basis for this low frequency primarily stems from utilization of isolation condensers and the firewator system as a source of mak'ulp. System activation does not require AC power nor long-term DC power for extended operation. The staff review noted, however, that the IPE analysis did not specifically model recirculation pump,:seal loss-of-coolant accident (LOCA). This assumption substantially reduces the significance of SBO as a contributor to core damage. Thisjfinding is not consistent with NUREG-1032 "Evaluation of Station Blackout :Accidents at Nuclear Power Plants." Unlike other boiling water reactors (BWRs)j>-yster Creek does not have steam driven makeup capability during station blackout and,' therefore, must rely entirely on natural circulation for core coo,.ling (analogous to pressurized water reactors (PWRs) with steam qgenerators),'*' A pump seal LOCA under these conditions could disrupt natural circulation -and compromise decay heat 1 n mise decay'heat

removal, Although the licensee provided rderenceS to support its position on seal LOCA, the issue remains openmand underjstaff cbnsideration (as a possible generic issue.sepakate from Generic Issue 23).

Because the issue is being addressed separately for BWRs, the staff f P ) review team did not pursue this aspect furthe.-. All modelled operator actions were found to contribute 21% to core damage. The IPE, however, did not perform a pre-niitflator human event analysis. Generic Letter 88-20 requested that licensees examine maintenance and surveillance practices as part ofrtheir* effo~rt to identify potential vUlnrrabilities. These areas are plant-specific and require an examination of routine personnel activities to uncover potential maintenance errors. The staff finds the lack of pre-initiaitor event.'analysis a weakness in the licensee's IPE. which may limit the usefulness of the IPE for future regulatory applications. The Oyster Creek IPE takes substantial credit (50%) for in-vessel recover) following core damage. For low pressure sequences, vessel breach is prevented by injection through condensate, control rod'drive system, or through the use of core spray. supplied by the fire protection system. The IPE also assumes that vessel failure will result in a guaranteed containment failure. for many sequences involving extensive core damage', the IPE did not credit any operator actions. The licensee indicated its conce nnthat the potentia' for adverse effects (which could result from operator mItigation action), could exceed the perceived benefit. For example. in response to staff questions on sequences involving recovery of electrical power, the.7licensee stated that prompt action t.o vent containment without proper *Accident 'Management Guidelines' could result in an earlier source term release thin if no action was taken. Other issiues associated with accident progression.also remain open, e.g., the consequence of activation of Irywell spraysswith corium in the drywell The licensee stated that it plans to postpone f'tther evaluation of potential operator mitigation action to the accident.-. anagement development phase "when better tools will be available (MAAP4).' In response to containment performance Improvement (CPI) program recommerndatons, the licensee considered a plant modification to provide water from the fire protection system to the drywell sprays and has concluded that this modification is not cost beneficial. The licensee has taken the position that the containment will always fall when the reactor vessel fails. This position tay have masked the true potential benefit from enhanced drywell sprays. Other licensees have concluded that havingjthe drywell sprays will significantly reduce the probability of dryvie1l liner melt-through. The ,icensee has stated that it is unclear how operator actions will affect the accident progression, and they intend to evaluate the effects of potential operator actions when appropriate tools (MAAP4) become.available. the staff recommends that the licensee continue to eva'10'ate the need for drywell sprays as part of its accident management'progrnm e4viluatlon.' Based on the review of the Oyster Creek IPE5,,bmtta1 and associated documentation, the staff concludes thatthe i.Hcensee met the intent of Generic letter 88-20, this conclusion is based on the following findings: (1) the IPE is complete with respect to the informat1qn requested In Generic LettPr 2

88-20 and associated guidance dotument H~aEG1335: (2) the front-end systerbs analysis, the back-end containment-perform nce analysis, and the portion of human reliability analysis performed (post. rittator events) are technically sound and capable of identifying plant-sp fi ktvulnerabillities to severe accidents: (3) the licensee employed a vi* 1e means (documentation reviews and walkdowns) to verify that the IPE refl-cted the current plant design and operation; (4) the PRA which formed the basis0of the IPE had been peer reviewed; (5) the licensee participated fully, in thb liE process consistent with the intent of Generic Letter 88-20: (6) the liiensee approprlately evaluated Oyster Creek's decay heat removal&-(DHR) function for vulnerabilities consistent with the Intent of the USI A-45...resolution; and (7) the licensee responded appropriately to recommendations stevming.from the CPI program. It *,hould be noted that the staff's revieW.primarily focused on the licensee's ability to examine Oyster Creek for severe'accidentvulnerabilities. Although rertain aspects of the IPE w-re explored inpmore detail than others, the review is not intended to vaildate the accuracy of the licensee's detailed findings (or quantification estimates) whtdh ::stemmed from the study. 3. @

I. ALK.GROUND On November 23, 1988. the NRC iAssued GeondrltLetter 88-20 which requires licensees to conduct an Individtual Plant..xtamination (IPE) in order to identify potential severe accident'vulneiabilities at their plant and to report the results to.the Commission. ::Through the examination process. a licensee is expected to: (1) develop an overall appreciatinn of severe accident behavior: (2) understand the most.. likely severe accident sequences that could occur at its plant: (3) gain ai:.more quantitative understanding of the overall probabilities of core damage and fission product releasps: and (4) if necessary, reduce the overall probabililty of core damage and radioactive material releases by modifying, where appropriate, hardware and procedures that would help prevent or mitigate seveqr'accidents. As stated in Apppndix D of NUREG-1335, thhe.'IPE submittal guidance document, all IPEs are to be reviewed by'NRC teamsAto determine the extent to which each licensee's IPE process met the intent of'Generic Letter 88-20. The IPE review itself is a two-step process, the first step, or "Step I" review, focuses on completeness and the quality of the submittal. Only selected IPEs are investigated in more detail under a second4step or "Step 2" review. The decision to go to a "Step 2" review is primarily based on the ability of the licensee's methodology to identify vulneOTabllities,,and the consistency of the licensee's IPE findings and conclusions with previous PRA experience. A unique design may also warrant. a' Step 2to better understand the implication of certain IPE findings and conclusionsi,.As part of this process, the Oyster Creek IPE only required a 'Step.IA review~t f On Auqust 14. 1992. GPU Nuclear Corporation :(GPUN) submitted the Oyster Creek IPE in response to Generic Letter 88-20 and associated supplements. (Oyster Creek is a General Electric BWR-2 Mark I single-unit plant with isolation condensers.) The IPE submittal was based.OQn a Level I PRA. and a Level 2 PRA onsistent with Generic Letter 88-20, Appendix 1. The IPE submittal contains the results of an evaluation of Internal events. including internal flooding. The licensee plans to provide a separate submittal on findings stemming from the IPE for external events (IPUEE). The staff will review the IPEEE separately, within the frameworkprescribed in Generic Letter 88-20, Supplement 4. A, part of its review, the NRC contracted with Science & Engineering Associates, Inc. (SEA), Scientech Inc./"Energy Research Inc., and Concord Associates to review the front-end analysli,=' the bick-end analysis, and the human teliat 'ity analysis, respectively.."'A's review is documented in NRC-04-91-066 task 8 report, "Oyster Creek Nuclear Power Plant IPE: Front-End Review." Scientech's review is documented' i'nSCIE. NRC-212-92, "Oyster Creek Individual Plant Examination Back-End Techni 'al Evaluation Report.' Concord's review is documented in CA/TR 92-019-08, i'echnkcal Evaluation Report: Cyster Creek Nuclear Generating Station IndividualIPlant Examination Assessment of Human Reliability Analysis, Document-Only.," On July 27. 199J, the staff sent a request for a6Jitional information to the 1icensee. The licensee responded to the staff's request in a letter dated Octnber 1, 1993. In addition, the licensee" in a letter dated July 3. 1993, 4 Umm I

provided to the staff a feasibility study fo0'mplementatibn of a portable DC generator. This report'cdocuments findings and concluslo 1s which stemmed from the NRC review. Specific numeric'l results and other 'insights taken from the licensee's IPE submittal are listed in th'e'Appendix to this Staff Evaluation Report. II. STAFF'S RE.YjVILW I. Ljinj_ e s IPE Process The Oyster Creek IPE submittal of August i4, 1992, describes the approach taken by the licensee to confirm that the IPE represents the as-built and as-found plant. In addition to detailed document 'reviews, plant walk-throughs were performed by members of the licensee's PRA team (consultants and plat personnel) for familiarization with plant/system operations, equipment layout for origin and susceptibility to floods, and&'946ntainment walk-throughs for information to be used for the back-end analy'is t On the basis of review of the information submitted with the IPE, the I'taff concludes that the licpnsee's walkdowns and documentation reviewt.onstitute a viable process for confirming that the IPE represents the as-buil tand as-found plant. The IP[ submittal contains a summary description of the licensee's IPE procrss, the plant personnel participation'in.the process, and the subsequent in-house peer review of the final product. 1he staff reviewed the licensee s deocription of the IPE program orga-ization, composition of the peer review

teams, and peer findings and conclusions.

The ;staff notes the considerable participation of the GPUN personnel in virtually all aspects of the IPE through technology transfer, Jdel development, reviews, data collection, and requantification of the models with 'plant-specific data. In addition to the IPE team, other GPUN and plant organizations !were involved to insure that the models accurately portrayed the p ant; Although GPUN did not indicate its intentions of maintaining a "living IPRA," the tubmittal stated that, GPUN rrcognizes the potential benefit of the PRA and i tspotential use in future eva 1 u a t ion. As part of the IPE process, GPUN establishedian`independent review team which consisted of personnel from all appropriate organizations including engineering, operations. training, and an independent safety engineering group. This review was in addition t-o internal:'.reviews performed by the GPUN consultants Based on the review of the IPE sUbmittal and associated documentation, the staff concluded that the licensee's peer review process provided reasonable assurance that the IPE analytic techniques had been correctly applied and documentation was accurate, The submittal defined 'vulnerability" as "any ore damage sequence that exceeds IE-4 per reactor year or containment bypass that exceeds IE-6 per reactor year." The fundamental contributors to risk and risk-important accident scenarios were determined by delineating the sequence characteristics and evaluating their importance on the basis of rtheir respective contribution S 11

to core damage frequency and release cat e ry frequency. No plant vulnerabilities were identified and, therefore no potential enhancements were identified to specifically address vulnerbilties. 'he licensee probed the quantitative resultt by performing: (a) an uncertainty analysis; (b) a sensitivity stindy on several key variables; and (c) an importance analysis to identify the most important systems to plant safety. The sensitivity analysis concluded that changes to data or assumptions do not have a significant effect on the overall results. The results of the importance analytis provided-^3 basitsfor the identification of possible low-cost improvements..; The staff finds the licensee's WEt procese capable of identifying severe accident risk contributors (or vulnerabil ies) and that such capability is consistent with the objective of Generic Letter 88-20.

2.

Frr9n-Lj_ Ana ly1 Thp staff examined the IPE front end analysis for completeness and consistency with acceptable PRA practices. The licensee capitalized on insights stemming from the Oyster Creek PRA Level,] study, AtIEG-ll50, and several other PRAs of plants with similar designs. The Level I IPE involves a "plant model" which integrates the system and human action analysis (and assoclateddata), anddelineates accident progression from the initiating events (lEs) to plant damage states. Event tree sequence diagrams (ESDs) were used to identify,availlable succ-ss paths needed to mitigate accident initiators, and to identify subsequent system failures, translating them into rules. Plant-specifilc. analysis and transient assessment reports, in combination with the plant procOdures, served as the basis for the [SO. Top events in the ESDs were-sequenced by initiating events and intersystemn dependencies. Event sequences explicitly represent support systems, front-line systems, human responses, and dependencies. Functional success criteria and specific system succets$criteria for each major plant safety function with respect to.elach IE citgoty are clearly and appropriately described, The dominant accident 'sequence0..groups and their contributions to the core damage frequency are identified along with the contributing important systems. The front-end IPE analysis used the large event tree/small fault. tree methodology which treats dependencies on the:event tree as split fractions rather than through the logical linking of.fAult trees. The licensee used the latest modification of this method in which.:the event trees are replaced by logic diagrams, i.e., tables of rules. Thus, no event trees were explicitly presented in the Oyster Creek submittal. The licensee's IPE submittal identified 28 1nitiating event groups for Oyster Creek. These groups were further categorized into three broad groups: (1) general transients (15 initiating events); (2) loss of coolant accidents, small LOCAs (6 initiating events) and (3): -.large LOCAs (7 initiating events). Initiating events were determined by using'- A' master logic diagram which identifies the various plant functions thatfG could fail and lead to a plant 6

trip. these groups weie reviewed agains vious PRAs and industry studies, plant operational experience, and' thelFin &1'afety Analysis Peport (FSAR). The IPE identified and analyzed plant-specjfic Initiators. These included: interfacing system LOCA, loss oflintake chAh el flow to the intake structure, loss of Turbine Building Component Cooling`Water (TBCCW), unisolated steamline breaks and large pipe breaks inside contajrilMent,:and internal flooding. In response to staff questions on success criteria, the licensee stated that only RELAPS/RETRAN computer codes had been usedMn the-development of thermal hydraulic analysis in support of the Level 1 analysis. Further, core damage is defined as water at the top of active1fuel and decreasing. MAAP had not been used to develop Level I success critriat.14 The IPE analyzed front-line systems and maior support systems including but not limited to AC/DC vital power, service/circulating water, and instrument dir The IPE provides a clear descriptlon:Vf the top events considered; tU,, surcess criteria; the support systems required: the systems' cornfiguration, operation, testing, maintenanceiand technical specifications assumptions; and the systems' boundaries. The system analysii s task utilized the fault tree approach to logically combine the basic events and failure probability in order to derive the split fraction values used In the plant model. A comprehensive analysis of system dependendies was performed and included support to support, support to front-line, and front-line to front-line system. In order to develop plant-speciffic IE frequencies, a Bayesian update of generic (twx, IE data was performed utilzing plant-specific information. The data sources used were clearly identifiedlin the IPE. The staff notes that the licensee made an effective useX of both generic and plant-specific IE data. Further, the IPE submittal provides a detailj.ed discussion of the dependencies between lEs and mitigating system ;(inc.lud I t~front-line and support systems), and clearly presents how each IEqgroup afficts the split fractions used in the model. A Bayesian update process was also used to develop the IPE's systems' database. A generic database encompassing the cumulative experience from a Liroe population of nuclear power plant s watcombined with a comprehensive plant-specific database containing moref thin"lO years of Oyster Creek experience. The update was performed using..the data analysis module of the kISKMAN program. Plant-specific features.were considered in selecting the appropriate generic distributionsitn order'tb obtain a "coherent" integration and upda' ng of the database. As recomniended idn NUREG-1335, the IPE made extensive use of plant-specific data. Systems -and. components such as emergency core cooling pumps, batteries, diesel generators, electric buswork and breakers, service water pumps, Instrument air, primary containment isolation, Automatic Depressurization System (ADS) valves, and other components were quantified using plant-specdfic (mainly post-1982) data. The common cause failures (CCFs)4were analyi0c in two categories, The first category includes sharing of common components, effects of floods, and human errors during test and maintenance. The second category includes design errors, construction errors, procedural def1,fiencles, and unforeseen 7

environmental variations. Common caus$01Vebts were incorporated into the system analysis in order to identify thetCCF mechanism. the quantification of the CCF factors was'accoiplished by a AdItlple g eek letter (MLG) methodology, consistent with NUREG/CR-478O. "Respon4ing to tht staff's request for additional information, the litensee lIsted 50 common cause failure events and their associated contribution to core datmge. The submittal contains the top 100 most.probable core damage sequences in Appendix C, Table C.5-1, of the Level r,.eport in accordance with the reporting guidelines in NUREG-.1335. These 0oo highest frequency sequences account for 82% of total core: damage frequency. The IPE derived a point estimate mean of 3.96 E-6/year for a total CDU.' An uncertainty analysis identified the 5th and 95th percentile faS .31E-6/year and 9.82E-6/year, respectively. Among the dominant accident sequences, about 20.8% (7.69E-7/year) of the total CDF was contributed by the los$;sof all AtCpower (station blackout) with failure of an EMRV to reclose, Turbine 'trip with loss of all DC power contributed 7X of the total CDF' (2.59E-I/yr), and reactor trip with the loss of all DC power contributed 5.7% of thetAtal C0F.(2.IE-7/yr), Other dominant sequences included: inadvertent MSIV closure with loss of all DC power (3.3%); loss of offsite power events (LO$P) with-EMRV closure and core spray failures (3,2%); loss of TBCCW with EMAY.V-,'closure and core spray failures (2.8X); and large below core LCCA with spray failure (2.6%). the dominant lEs include: loss of offsite power (32.8% of total COF); turbine lrip (13.1X); reactor trip (7.7%); MSIV, ;6losure (7.7%); and total loss of feedwater (5.7%). The IPE did not find nticipated transients without scram (ATWS) as a significant contributor'to the total CDF, based on credit taken for plant modifications for ATWS prevention and mitigation and the incorporation of operator recovery actions in the emergency operating procedures (EOPs), The IPE performed an importance analysistthat showed that EMRV failure to close contributes most to total COF (48%)j Essential AC power bus failures contribute 37% and DC power failures abouit 33%. This importance measure percent COF is that percentage resulting from the summation of the frequency of all sequences involving the 'lop events,,and it represents the percentage decrease in the COF that would result if the top event or system failure could be made zero. A sensitivity sttidy was performed oh several key variables in the study: LOSP events recovery; EMRV failures to close; and recovery of containment heat: removal (including recoVery of DC power and containment spray). The analysis concluded that changes to data or assumptions do not have a significant effect on the overallVresults4 A number of the Oyster Creek IPE dominant,.sequences involve loss of DC power. (Oyster Creek has only 3-hour battery capacity).r These sequences and associated contribution to core damage InIlude:Aturbine trip with loss of all DC power (7%): reactor trip with loss`^of DC power (5.7%); and inadvertent MSIV closure with loss of DC power (33t)`,DC power is required to activate the isolation condenser and open'the EMR4s. to allow for' vessel injection with the low pressure firewater system., 8 8 '..'

I The licensee's IPE station blackout analySl did not:address recirculatioh pump seal LOCA, although the staff identified gross seal failure as i potentially dominant core damage sequencei n station blackout accidents at nuclear power plants (NUREG-1032). Oyster Creek, for example, does not have a steam driven makeup system available during station blackout (unlike other BWRs). A LOCA during station blackout would compromise decay heat removal by degrading natural circulation between the reactor core and isolation condenser. In response to staff questions the licensee stated that loss of coolant through the recirculation pump seals'would be "insignificant" on loss of pump seal cooling, a condition which I:ould exiit during station blackout. Although the licensee provided references to support its position, the issue remains open and under independent staff ::consideration (as a possible generic issue separate from Generic Issue 23)i' Because' the issue is being considered separately for BWRs, the staff (IPE) review. team did not pursue this aspect further. The staff notes, howevier, that t he Oyster Creek IPE analysis is sensitive to assumptions associated with 'recirculatlon pump seal failures (i.e.. impact the estimated core damage frequency by more than an order of magnitude). the [PE's flooding analysis was divided into.two parts. In the first part, effects were addressed in the rules'and modules of the mitigating systems analyses. In the second part, flood source and equipment location data were compiled and catalogued and only components.that were deemed significant to plant risk were analyzed. The flooding analysis considered the effects on components (including electricall of beinglisubmerged, sprayed, or exposed to condensing steam. The calculated flood-induced COF Is 2.08 E-7. Approximately 78i of the flood-inlduced COf is due to floods in the turbine building, with the remaining due to f loods,'in the reactor building. Based on the IPE description and licensee responses to questions, the staff finds the licensee's IPE methodology clearly described and justified in its submittal. Based on the staff's review oft'the front-end analysis and the staff's finding that the analytical techniques used are capable of identifying potential core damage vulnerabilities, the:staff concludes that the IPE front-Pnd andlySis meets the intent of2-Generic Letter 88-20.

3.

Back-4fnd Analysis-i The staff examined the licensee',s'back-.end'analysis'for completeness and consistency with the guidance specified' in eneric Letter 88-20, Appendix 1. The Oyster Creek consultant, PLG.Incorporated,'used the RISKMAN methodology to quantify the event trees and version 7.03 of."MAAP-3.0B. The analyses conformed to Electric Power Research Insti.tUtes (EPRI's) recommendations related to selected model parameter values`.` MAAP was not used to investigate in-vessel recovery under damaged core conditions. The licensee, through PLG, had EQ1' Engineering Consultants perform a plant-specific containment structural analysis t-odevelop containment failure pressure, temperature, and location insights, The mean ultimate containment failure pressure was determined to be 134 psig..The staff found the approach 9

consistent with Generic Letter 88-20, AppendX I (Guidance on the Examination of Containment System Performance). Three issues are unique at Oyster Creek. First, the torus was strengthened in the 1980's. This resulted in about a 25% increase in pressure capacity to a best estimate limit of 153 psig. Second, the :sand,,normally between the drywell shell and the concrete wall at the ,rywell floor elevation, has been

removed, Corrosion has occurred at this: lcatWon which has reduced the structural integrity to about 8 psi below the.drywell head flange leakage pressure.

Finally, Oyster Creek has a 1-foot thick, 6-.inch high curb at the liner-drywell floor interface. The volume of the sump and within the curb is sufficient to contain all of the estimated corlum volume. This reduces the liner melt-through probability by approximately 50%. Thus, there are no wetwell failures, and drywell failures are at the drywell floor location due to over pressure with a small contribution from liner melt-through. the translation of the Level I accident sequences into Level 2 Containment Event Tree (CET) and accident release charact~ristics was performed by mapping each of the accident sequences into Plant Damage States (PDS). The PDS were defined by the condition of the plant at the end :of the Level I analysis. the PDS considers the rea'ztor pressure (high or low), drywell floor conditions (wet or dry), containment integrity (intact, bypassed, failure within a few hours of event initiation, or falls: later), status of active systems (containment vent, suppression pool cooling, drywell sprays, and water to cool debris), and status of reactor building (isolated, firewater system in the reactor building, and standby gas treatment system (SGTS) operability).

However, the reactor building and SGTS effectiveness was assumed to he zero based on the dominant containment failure mode being a catastrophic breach of containment.

The licensee reduced.the suggested screening criteria identified by an order of magnitude to ensure consideration of Sequences which could be important to containment Integrity and risk. k;',Th e licensee has listed all of the Level 2 sequences with a frequency equal'-to :or greater than I E-10 (49 Sequences), exceeding the NUREG-1335 screening uidelines, The licensee identified 19 P0S which, were mapped into seven key plant damage states (KPDS). The KPDSs were used as the entry states to the CET. The CET models the core degradation, vessel.Failure, "containment behavior, and reactor building behavior. The CFT was developed to resemble the Peach Bottom NUREG/CR-4551 accident progression event trees. The quantification of:the CET for each KPOS was carried through a number of split fractions defined for each top event, The results were used to define CET end-states bins which were subsequently used to develop source term categories. The sourceiterm was evaluated using a source term event tree (STET). The STET considered six questions: drywell spray availability; reactor pressure at time of vessel failure; condition of containment (intact, vented, early or late failUre); containment failure mode (leak or gross); availability of pool scrubbing; and availability of reactor building mitigation. The results of the STET '.'ere grouped into six key release categories (KRC) based on similarities of containment failure, timing, and mitigative features. The source.terms for' the KRCs were calculated by selecting representative sequences and using MAAP to model the behavior and I0

release of 12 radionuclide groups. The t0i0ig of the release was based on the estimated containment failure time from the::initlation of the accident, as follows: Early (E) - 3 hours or less after veisel failure, Late (L) - More than 3 hours after'yessel failure. Sensitivity studies concerning accident phenomenology were not performed. Instead, the licensee stated in response tothe staff's request for additional information, a combination of parameters wertechosen from those recommended by EPRI, to give a conservative response in source term released. Substantial credit (50%) is taken for in-vessel recovery following core damage. This is partially a result of the-licensee's definition of core damage. (Core damage is defined as witer at'the top of active fuel and decreasing.) For low pressure sequences, vessel breach is prevented by using the condensate system, control rod drive'system, or fire protection system through the core spray system. The assumption was made that vessel failure will result in a guaranteed containment fa11ore. For many sequences involving extensive core damage, no credit was given to operator actions. The licensee indicated its concern that the potential foriadverse effects (which could result from operator mitigation action) cousld`exceed the perceived benefit. For example, in response to staff questions 'O 'sequences involving recovery of electrical power, the licensee stated that prompt action to vent containment without proper "accident management guidellnesl.could result in an earlier source term release than if no actlon was taken. Other issues associated with accident progression also remain open, e.g, ' the consequence of activation of drywell sprays with corium in the cirywell. The licensee stated that it plans to postpone further evaluation of potential operator mitigation action to the accident management development phase "when better tools will be available (MAAP4)." The accident management program is a key element in closure of severe accident concerns, and the staff recommends that the licensee address these issues within that framework.' The licensee considered the effects of containment temperature and pressure on the elastomer seals. These seals are used for the drywell head flange and equipment and manway hatches. For all of'th"" potential accident sequences considered, the temperature and pressure profi es are expected to result in no or little leakage. This result is based onitheir consultant's analysis (EQE) and agrees with the results of analysis discussed in NUREG/CR-5565, NUREG/CR-4944, NUREG/CR-4096, and NIJREG/CR-4064. the licensee also examined the failure of containment isolation. The modeling of containment isolation failure is based onva4fault tree model. The fault tree incorporates modeling of automatic contaihment isolation valves that penetrate containment and are open to the containment atmosphere (e.g., vent and purge lines) as well as potential containment bypass lines whose system pressure is less than 90 psi, larger than 1l-nch in diameter, and contains non-manual isolation valves. The fault tree considers automatic and manual isolation signal failures and component and common cause failures. 0W.-MIP'1 10 now .N 11 A,

,h-

{ r:

The licensee employed a process to understand quantify severe accident pr(qression. The process lead to a determihatlon of conditional containment failure probabilities and containment failure modes consistent with the intent of r'Ir.-eric Letter 88-20, Appendix, 1. The following tables show the conditional containment failure probability as a function of failure location and timing, respectively.: Contolnment failure iocations.

Drywell 42.3%

(Liner:1Melt-throug9h 17%)

WetwelI 0

^ Bypass 7.3%

Intact

.50. 4% LQntainment Failur',1'lmiings

Early 15.9%
Late 26.4%
Bypass 7.3%

a Intact (following vessel breach) 0.0%

No Vessel Breach 50.4%

Of particular interest is that the probability-of containment failure is zero if react.or vessel failure is prevented and one'if reactor vessel fails. This is dute to the fact that the recovery of electric power was not considered once coro: d{mage commenced. Therefore, there was no recovery of containment heat Comonval or drywell sprays. ihte rrocess of determination of conditional c 6tainment failure probabilities arid containment failure modes was consistent 'wIth the intent of Generic Letter 88-20. Appendix 1. The dominant contributorsto containment failure were found to be consistent with insights from other, analysis of similar designs. The licensee characterized containment performance for each of the CET end---tates. The licensee considered.the failure of containment seals and containment isolation failures. The staff's X` Pew 'did not identify any significant problems or errors in thq back-en~dinalysis. The overall assessment of the back-end analysis"As that the licensee has made reasonable use of probabilisitic techniques in performfl"g'.the back-end analysis, and that the techniques employed are capable :f identifHing plant vulnerabilities. Based on these findings, the staff concludes that the licensee's back-end IPE re 1is consistent with the intent of Generic Letter 88-20.

4.

l`.1l4msn FactQor__Q~nsAidti I:' r a

n.

The licensee acknowledged three type:; of human-errors, pre-initiator human events associated with errors during, routine activitles (such as valve misalignment) leaving equipment disabled ("GroUp A"), initiating human events associated with errors, causing a plant abnormal condition ("Group B") and 1?2

N

.wI 1'

1 am:u

I post-initiator human events associated wltfherrors during operator response to an abnormal condition, i.e., an 1initiator#. *Group C"). The 1PE did not perform a pre-initiator 'han event anTlysls. the rationale provided by the IPE and in the llcensee's responses for this approach is that: (a) usually few pre-initiator events are. identified during a Human Reliability Analysis (HRA); (b) typically they are not significant contributors to core damage frequency; and (c) the frequency of ipre-initiator events is captured in the basic equipment failure rates and, hence, there could be double counting of failures if a separate analysis was performed. These types of errors have been shown to be dominant contributors In other studies and are not necessarily part of the basic equipment failure rate (e.g.. NUREG, 1150 analyses). In:addition`,'Generic Letter 88-20 requested the licensees tu examine maintenance and surveillance practices as part of their effort to identify potential vulnerabilitlies. Generally most plants have administrative controls for preventing system unavailability due to test and restoration activities. The pr'Icess by which these controls are implemented. however, determines whether thei' are practices creating the potential for-Vla'ing a :,ystem in an undetecte' disabled state (resulting in equipment ravailability on demand). WhilE the staff agrees that a portion of pre.. initiator events can be captured when performing a Bayesian update (provided anplp operational data is available), unlies routine personnel activities are examined as part of the IPE HRA, 'Such instances of potential errors may not be i.incovered. The staff finds the lack of pre-initiator event analysis a wpaknpss of the licensee's HRA. which may limit the usefulness of the IPF for f a t re regulatory applications. i on s Initiator huaian events were analyzed as part of the IE analysis consistent with acceptable PRA practices. Post-initiator human events were extehsively analyzed. They were further distinguished tr human events associated with response-type actions and to human events associated with recovery-typeaictions. Response-type actions include those human actions performed in response to the first level directive of the WOPs, such as reading instrumentatiOn to determine reactor water level status or maintaining reactor water level with different systems. Recovery-type actions include actions performed to recover from a specific failure or fault, i.e., cross connecting electrical busses following loss of offsite power (proceduralized action) or gaging alfailed instrument air relief valve (non-proceduralized action). In order to identify post-.nitiator human events, the licensee examined the FuPs, system instructions, and off-normal e'Vent procedures associated with the accident sequences delineated and:the systenis modeled. Further, discussions were held with plant operators on the interpretation and implementation of plant procedures to identify and understand the specific actions and the specific components manipulated when responding to the accident sequences modeled. 88-20The licensee employed the Success Likeiihood Index Methodology (SLIM) to quantify post-initiator events. The licensee's evaluation was based on 13,

eliciting the control room operators' judgement of each action analyzed. Important factors influencing human perforiance (for example, the type and location of plant procedures, operator co6mUnicatlon, location of required actions, effect of annunciators and alarms -,and the timel available versus the time required to perform the needed human itlon) iwere considered in the analysis. Piant-specific performinc shaping factors were used in the calculation of the human error probabilitj iHEPS). The HkA dealt extensively with the issue o Xccounting for the effects of multiple operator actions and the.dependencies among human actions. A .confusion" performance %haping factor wasi cluded in the quantification of oarh human error to account for dependenctes among steps of an individual task. Further, the IPE performed a thorough 0sensitivity to multiple operator actions analysis that included a quantitative and a qualitative portion. The quantitative sensitivity re-estimated the.CDO by increasing the HEPs to determine their relative iT"ortance. The qualitative sensitivity reviewed the time available versus the time required for -an action and crew changes, for all dctions in a scenario. This sensitivity"did not identify any dependent act ions that were treated as independent 'The staff also notes that the licensee used a sound approach to-address initiple operator actions, and that the study had been used for,lanning plant modifications in a wiy that allowed a better understanding of various operator.Interactions. The post-initiator quantitative :results and4.the insights derived fror the analysis are also discussed in a:clear and concise manner in the IPE. A tital of 34 functionally different operator actiohis-(and a total of 66 individual actionv) were modeled in the IPEL The total,; human errs contribution to (DF is 21X. The IPE lists (Tables 2.14 and Ž.1V,4)-and discusses the most important human actions In the context of their contribution to the total core damaqQ frequency. It should be noted that 'fo Individual (or combination of) human action(s) were dominant in the Oyster Creek APt. The most important human action to CDF is initiation of contatnfment cooling (2,76%), followed hy failure of manual core spray function core':spray (2.70) and recovery of DC power (2.50)). An improvement regarding operator trainingg for initiating the containment spray system was identified and implemented&.-n addition, one procedural and several operator training improvements werep, dentified that are under review and consideration by the licensee. In suImmary, the staff finds the HRA methodology described in the licensee's sUb:nittal supports the quantitative understuhding of the overall probability of core damage during plant opertions, as wellas an understanding of the contribution of human actions to that probabtility. Therefore, the staff find: the licensee's assessment of human rellabillty capable of discovering severe accident vulnerabilities from human errors and consistent with the intent of Generic Letter 88-20. The staff notes 'tnat, the licensee used a thorough, systUiatic. and traceable post-Initiator event human analysis. However, the staff finds the lack of pre-initiator human *Vdnt analysis a weakness of the licensee's HRA which may have an impact to the, IPE'.s usefulness in other applications, The staff encourages the llct#fla to consider pre-initiators explicitly in its HRA in any future revisionsof its PRA. 1 4

5 n a As a rpoilt of the Containment Performance Improvement Program. r~~ornm;~4tion, for lmprovements were made for licensees to consider as part ,0th' IP! proreneS. These reconmendations weretg dentifled in Generic Letter 88 20, S'upolement 1. Each of these proposed Itprovements Is discussed , sp.r AtPl)y below. I) A _hirlp.' The licensee has proposed installation of an 8-inch hAr d Ine( vent from the torus Olr space tofhe stack. Venting is initiated rirectly before containment pressure reaches 3.0 psig >-rr"-ponding to the containment spray Start signal and ADS actuation loqfil .etpoints) and again before torus pressure reaches the primary ',rtainmont pressure Imit, as lirected by, yste-Creek's [OP. [hi, l pnfe" has suggested an alternate StrAtegy to protect the 'irrJency.corp cooling system ([(CS) pu ps from a loss of net positive 'w t Inn hoad (UPSH) in lieu of ventingrCntainmernt, the licensee ii u'lng suppression pool coolIIIg Wlth the restdup) heat renroval (WHP) 4r1 tsiing the wetwel I sprays to reduse the suppression pool water ItXrf!I J r P After a sufficient redliction in pool water temperature. the iryf? I sprays would be used to reduce' the drywell temperature. 1hi' ot...'d prevent a potential loss of NPSH by venting containment. the '-rnv ,)rposed considerinq this procedfre during the preparation if e Iijont management guidelines, and thould be considered as part of I. d(f; ident Managemsent Program.. '2) An d.1 irpa.' t-LY-p.r.eIon or drVwefl sprjvf: Pr;v i onS 1r using the fire peotection 4Ostem pumps aligned to 'upply t!cth 'livisions of the core spray'system hAve been provided. The fire r ) eAction system consists of two diesel driven'2000 gpm pumps. The 'F1, ri I rivfn pump has its own DC power supply, and it can be startei "i v -onnection of the fire protection system Im the Core Spray (S) i s hy means of a 4-inch (which'reduces to 3-inches) li ne !ff)m *r 12-inch fire mafn ring hleader. -The fire protection system c-in Io be used to provide make-up lto the isoation condenser. Each isolation condenser is provided 'tire protoit1on system water through a 6 ifl(h I ir 'ron the 12 inch flre natn rfln eader NI( provision exists for using the fire pretaction system with th, irywoll s;prays. The licensee hat concludod:that this capability is not .? t u)nefic ial for the followinti reasont. tirst, those sequences where drywell s prays could be beneficial represent only 8.75% of the total (orp Iamance frequency. Second,: the flow r~te at the nozZle would not dovs;eIop A (ull spray pattern, but would 'ruh out of the,pray nrzi;ls.' Without a full spray pattern, the fission product scrubbing would be qrr.iat 11. ro-duced. And finally. `without a folly developed spray, the 'l.4tility to cool the containment shell lgreatly reduced," f uehe'.t iori, "it is highly likely that fire.protection water exiting the h!ilv In Ihe vessel left by the exiting coritum would provide a comparable f1eq?,op ft containment shell cool:ing.T this last rgument is only true if wte-r slxts the hole sufficiently beforethe corium reaches the Is f, IIt WI -...... -.' - , 1 ; . I 11 "I -. 1111, :,

atE: t he ^. n Thi drywell liner and that the drywel it flOSded at the llner, This depends on the melt progression-and vehdlfailure assumptions. Given that the water will exit the reactor vesSOA throUgh the failure location And that a pool of water will overlay th.'..tor`utn. drywell sprays coulIJ still be important. As discusstd in NU0tE9/R-5978. the drywell sprays could be important if: (a) the water poo0ltcould not be kept subcooled. (b) there is excessive, late, release ot e.arse aerosols from residual fuel in the reactor vessel directly to the drywell atmosphere; or (c) there is extensive revaporization of depofited fission products from the reactor Coolant system after reactor vessl.1failure. Furthermore. NIUREG/CR-5861 states that flooding containment prior to core relocation onto the bottom head can signifi-cantlyAde00y or prevent vessel failure. The liu#nsvee has taken the position that the containment will always fail when the reactor vessel fails. This position may have masked the true potential benefit from enhanced dryvell sprays. Other licensees heio conolusded that having the drywell sprays will significantly reduce thp probabilIty of drywel1 liner melt-throtigh. The licensee has stated that it is 1unclear how operator,actions wlW.l affect the accident progression, and they intend to evaluate' the effects of potential operator actions when appropriatietoolst; '(`P4) become available, The ,taff recommends that the Iicentlie continua0to evaluate the need for drywell sprays as part of its acctident Ma n*ement program evaluation, .n nh. ( ' ;.,i R

eoresSurIzaLion r

1 lyJ: The IPE sub ittal-itated that the licensee would Consider procurement of a prrtable generator, based&on its cost effectiveness. Ihe Stat ion batterIes will provide DC powvirfor a minimum of 3 hours, Howpver, in a letter dated July',2,. 1993, the licensee stated that: (I) portable DC generators were not-readily avai.lable: (Z, for extended N-tation blackout conditions, portable AC generators, to be used for battPry charging, could readily be obtained through an outside supplier; .4nd (3) providing a portable DC-power supply was not cost-effective, WBafd on their analysis, the I ccnsee stated that portablo generators werp not procured at this time but will be reconsidered during preparation of the accident management guidelines. This need for alternate pcwer supply at the site willbe teviewed as part of the Arr d sdent Managemqent Program.

4) tnhrater.

jriLLsLib QiS fre(ii j EPis), R: Th.eilCensee has incorporated Rpvifion 4 of the BWROG EPGs. ti:.s.d on this review, the staff concludes that thelicensee has responded to IhN CPI Program recommendations, has searched fo.';vulnerabilities associated with containment performance during severe accidents, and its evaluation is monsistent with the intent of.:Generic Letter 88.2Oand associated Nopplement 1. However, certain aspects of the analysis are to receive further (nslIderatI or a% part of the licenske s Accident Management Program. I I,,' I

6, QHR Evaluation

-c In accordance with the resolution of Unresolyed Safety Issue (US!) A-45, the licensee performed an examination of Oyster'Creek to.dentify decay heat removal (DHR) vulnerabilities.

Theo:,maluatlo considered various combinations of reactor vessel Inventory makeup' and decay h.t removal rejection pathways. The analysis took miniinal credit for human re6Cvery actions. The plant features listed below were cons Idei&46in the lIcensee*s evaluation of the Oyster Creek decay heat remoVkil funci t i (1) The normal path for decay heat :remova {ff jjlvey Y the feedwater system and main condenser. The success citeria fQK this path require that main steam isolation valves (MS'`sd i are open :nd th~t the main condenser and the support systems are available. :The X qutred support systems include instrument air system for cont:ol of Zhefeedwiter regulating valves, 4160 VAC system for the feedwalter and th t condonsate pumps,:120 VAC feedwater control power, and.12;5 VDC fOr the instrument and logic. The turbine building closed coollig water (tCC) is also required for pump and lube oil coolers., (2) The df'cay removal path through the isolation condenser can be utilized following reactor isolation transients -where either the main condenser i, unavailable or MSIVs are closed. The Success criteria for this path require initiation of one ofjtwo isolation condensers, followed by the 'tjcress(ul long term shell side makeup water. The emergency makeup

waiter for the long term operati1on, due to-the boll-off of the shell side inventory, can be provided by Wither the t ondensate transfer system or thre fire protection water system.

The WhIgh pressure makeup on the eventual loss of the reactor coolant sy$~ n ` inventory can be provided via the control rod drive hydraiul it sys, t.; (3) Decay heat may be transferred through coo'lant discharged Into tie containment. The discharge ma) involvee-apipe break (in the event of a LOCA), or through the operation of' relief or safety valves. The decay heat is removed from the containment vifih e spray/emergency service w.dter system and transferred to thejintaki canal.. (4) Upon failure of the contalnmentspray/eOergency-.trvice water system, decay heat may be transferred to the contd ii nt and outside atmosphere through the hardened vent system. The recovery of containment heat removal isvwell documented in the submittal. The overall contribution of loss of decay heat?.emoval to CDF had been found to be 3.96X. Based on the process that the licensee used to Search :for DHR vulnerabilities, and review of plant-specific features the staf finds the licensee's DHR evaluation to be consistent with the82intent of eneric Letter 68-20 and resolution of USI A-4S. E

7.

Generic Safety Issues As part of the tPE submittal, the'lltenseE proposed resolution of several generic issues including U51 A-17, ".System Interaction in Nuclear Power Plants;" USI A-47, "Safety Impl~cat-ins) oUTitohtrol Systert:" Generic Issue (GI)-10I, "BWR Water Level Redundancy;15a R4iGI-lOS, lntei acing System LOCA at BWRs." However, USIA-17, GI.-lOl,0 and>`1O5 w'ere resolved by.taff with no new requirements. Accordingly, :the l'cesee's proposed resolution of these issues was not reviewed in detail.: :The,:rei#ew of the licensee's response to Generic Letter 89-19, "Request for Action"gelated to Resolution of Unresolved Safety Issue A-47," addresses.US!A-47 resY1lution.

8.

Liwqsee Actions and Commitments Frau the IPE The licensee used the [PE processito.identi$fy plant and/or procedural modifications. The IPE took credit for several modifications that the licensee installed during the 1411 refuelinho "outage. These include installation of a hard piped contaInment vent system; operator training for manual initiation of the contlinrnent spray Wsystem; and installation of interconnection to the combustio'-tturbine generators at the adjacent Forked River Site. The combustion turbine interconnection will make it possible to supply power from the combustion turbines directly to non-essential 4160 V bus IA and emergency loads of essential 4160' V.uses IC and ID via cross-tie. Purchasing a portable power generator.and developing procedures for recovering offsite or onsite power were i detified asf dditional improvements for coping with station blackout. While the.proceddrtirdevelopment is underway, the licPnsee plans to evaluate the purchasing- 0an-:additional AC generator before the )SR refueling outage.. The.staff recoies the licensee's intent to address station blackout events $.bk the A.ticonnection to the two combustion turbines and recovery of AC power..procedue development. IPE findings indicate that there are a number of additional 'low-cost" improvements whirh could enhance overall. reactor safety. These planned actions include: Development ofanemergency procedure for Loss of Offsite Power. o Development of an emergency procEdure for Loss of DC Power. o Increased training on,the impotance of the core spray system. o Changes to maintenance schedullg for the core spray system to improve downtime. ) Programs instituted.tb reduce blockage and fouling of the isolation condensers. o Modifications to implement the'>-Reactor Overfill Protection System. o Consider the development of specific guidance, training, and procedures for reactor overfil 0 .ransients. o Increased emphasis i'n. trainin gn key. operator actions as defined by the IPE. o Consideration of alternate contXAnment heat removal capability to maintain minimal NPSI,'as part of Accident Management. o Alternate water supplyj,,for dr sprays (Accident Management). 18,;.,.'wi.

n.. '. ';

M_01_ _11_

1. -- 1:1. I I I

Although the NRC review did.not examihe th merits of the above recommendations in detail, the staff note ithat the licensee is applying PRA/IPE findings to enhance plant-safetyj.: ;T` .taff, therefore, finds the licensee's actions reasonable. Ill S.L -'V..Tu1 the staff finds the licensee's WE submittal;for internal events including internal flooding is consistent with the inf rmation requested in NURIEG-1335. Based on the review of the submittal, th: litcensee's response to questions and associated information, the staff::finds the§,.l~icensee's IPE conclusion that no fundamental weakness or severe accident Vulnerabilities exist at Oyster Creek to be reasonable. The staff notes that: (I) GPUN personnel participated&in virtually a,11 aspects of the IPE through technology transfer, model.developrnen*t' reviews, data collection, and requantification of the models with.plnt-specific data. In addition to the IPE team, other GPUN and plant og0anizations were involveo to insure that the models accurately reflect th~eas-bult, as-operated plant. (2) The licensee established an.:independ44'review team which consisted of personnel from all appropriate organizations including engineering, operations, training, and an indepondent::safety. engineering group. This review was in addition to internal reViews performed by the GPUN consultants and provicles assurance tha tthe. WPE analytic techniques had been correctly applied and..document.at was accurate. (3) The front-end IPE analysis is complete6,with respect to the level of detail requested in NUREG-1335. In addition, the analytical techniques were found to be consistent with ot~herzNRC reviewed and accepted Probabilistic Safety Analyses (PSAs).X (4) The back-end analysis addressed the most important severe accident phenomena associated with Mark I conta..nments. No obvious or significant problems or e~rrors were idantified. (5) The HRA allowed the licensee to deve l.:anu.:understanding of the contribution of human errors$ .to COF a4ntc'ontainment failure probabilities. However, lack: of analyis of pre-initiator events is a limitation of the licensee's IPE. (6) [he employed analytical techniques An the front-end analysis, the back-end analysis, and the HRA are. capable of identifying potential plant-specific vulnerabilities. (7) The licensee's IPE process searched fo DHR vulnerabilities consistent with the USI A-45 (Decay Heat Removal, oilability) resolution. (8) The licensee responded to CPI.Program re ommendations which include searching for vulnerabilities associatd with containment performance 19 ffi. ftS,.S 0,

I -¶ 4 during severe atcidents. Howeversthe licensee plans to address a number of issues in its follow-.on at. dent management program. Based on the above findings, the, staff concludes that the licensee demonstrated an overall appreciation of severe accidents. has an understanding (if the most likely severe accident sequenche.s that could occur at the Oyster Creek facility, has gained a quantitativenunderstanding of core damage and fission product release, and responded appropriately to safety improvement opportunities identified during the process'.- The staff, therefore, finds the Oyster Creek [PC process acceptable in mee`t.ing the intent of Generic Letter 88-20, rhe staff, however, finds the lack of analysis of pre-initiator human events a weakness of the licensee's IPE that may limit its usefulness in other applications. The staff encourages the l4,iensee to improve its HRA by 'including pre-initiators in any-future revisions of its PRA. The staff also notes that GPUN did not explicitly steat Ii-that they plan to maintain their PRA "1living." The staff notes that -.a "'.iving0 PRA could enhance plant safety and provide additional assurance that,any potentially unrecognized vulnerabilitie't would be identified and evaluated durln t life of the plant., Principal Contributors: ErasmiaLois John Ridgely Jin Chung flate-2 A... 0y~~~ Sf;; '.i 0

l.'.

AUMMRYNSEEX Qiitir..+/- 1iTERNt1ALENT$)JflR~tfi Total core damage frequency:(CDF) ppipt estimate:, 3.69 E-6/Year Initiating event importanCe to tot a'OF: o Loss of offsite power' 32.8% o Turbine trip 13.1% o Reactor trip 7.7% o MSIV closure 6.9% o Total loss of feedwater 5.7% o Loss of condenser vacuum 4.0% o Loss of TBCCW 4.0% o Loss of intake structure 3.3% o Electric pressure regulator.fa,,' 3.2% o Large below core inside cont. L A 2.9% O Dominant core damage sequences and corntribution to CDF: o Station blackout with failure of.-' U an EMRV to reclose '20.8% o Turbine trip with loss of all.DCpower

70%

o Reactor trip with loss of all DCG'power 5.7X o Inadvertent MSIV closure with -loss of all DC power 3.3% o LOSP with EMRV failure to cl~ose and core spray failure 3.2% o Loss of TBCCW with falIures, of -E',RV close and core spray 2.8% o Large below core LOCA with cor -pray:, failure 2.6% o RWCU overpressurization with core.spray failure 2.0% o Loss of intake flow wilth.EMRVf,'fflure and core spray failures 2.0% o Loss of condenser vacuum lth lo$s (f all DC power 1.8% a Operator actinn importance t1o total CDF:.,i, o Initiation of Containment Cool.i'hg'

2.76%

o Core spay (Manual initiate or injection with fire protection) 2.70% o Recover of DC Power 2.50% o !Recover Offsite Power: 2.20% o Initiation of IC makeup: 1.51% o Containment Venting 1.47% u Manual initiation of ADS '1.23% o Initiation of Boron injection (ATWS) 1.22% 21 z ':q-:'.-

MP o Level and Power Control Pollowln'g.ATWS 1.08% o Control Post Trip RPV Level 1.03% System importance to total ;tF. O EMRV closure 48% o 4160 VAC essential bus t0 37% o 4160 VAC essential bus IC 37% o 125 VDC bus C 33% o 125 VOC bus B 31% o Recovery from LOSP 26% o Core spray 21% o Reactor scram 6% o 4160 VAC bus IA o 4160 VAC bus IB 4% Conditional containment failure probablity given core dai O Orywell 42.3%

(:,tLiner Melt-through mage:

17%) o Wetwel 0 I .0% Bypass

7. 3X o

Intact (Vessel Breach Prevented)/ ::. 50.4% Important plant hardware and-,plant characteristics regarding containment performance: o 8-inch hardened torus:,vent. _- o 6-inch high, 1-foot thick drywelV floor curb at the drywell liner. O Two isolation condensers, operate with only opening one DC powered valve. o torus structural strength incre .edby 25% due to installation of straps. o Liner corrosion at the liner-sahd (which has been removed) - concrete interface (reducers strength by about 8 psi). O Alternate water supply to reactor6vessel and isolation condenser. Modifications the IPE took credit: o Interconnection to the"combustlon'iturbine generators at the adjacent Forked River Site. v Bard piped containment' vent system, o Operator training for manual initiation of the containment spray system. Significant PRA findings:. o IPE importance measures identifiead failure of electromatic relief (EMRV) to close as the 'largest component contributor to total COF (48%). The significance of'this contributor stems from mitigation success criteria which'Tequires (f.r many accident initiators) opening and subsequent tlosing of yp to 4 of 5 EMRVs. 22

o Losses of offsite, power are significant.cootributions to CDF; the planned modification to use.tdependent offsite power source will help mitigate the effects of 'loss of offsite power event. o The plant is highly dependentnb,'OC power; battery monitoring and maintenance will continue to 'lmportant. o The licensee installed an 8-'-,£h hard vent to reduce containment pressure. The an'al~yis,

Nowefv, showed that containment venting could result in inadequate NP$H, for the RHR pumps, an effect that can be alleviated by reducing the suppression pool temperature with sprays before venting.:

Potential improvements under evaluio n: : o Integrated loss of offsite power and station blackout procedure which includes cross-tieing buses and alignment of the alternate AC capability. o Loss of all DC power proceduk'oand'a portable power generator for the essential loads.- O Training in the containment tPray system and changes in the 1ireventive maintenance on"the containment spray and emergency service water, o Post trip reactor feedwatercontrfol (Reactor Overfill Protection System (ROPS)). o Alternate containment heat remoYal capability to maintain minimal NPSH (as part of Accident:Man aement). o Alternate water supply f.or:d.rw11 sprays (as part of Accident Management). Information has been taken from the Oyster Creek IPE and has not been validated by the NRC staff,. 23

'F. eNC OS2R ()YJ1[f (R(IK I NO I VI UAL PAt~ EXAMINATION 1[AINIICAL EVALUATIONREPORT (FRONI-IND) i

1. -

-., - I., I*

,,,

w - ,_ _ . , dvl imm 5 - I I _-- -- ---- S :;\\I o.! tL i~ I \\ O)yste~r C-reek Nuclear PwrPatI~K O ( 'ntractor'I'echoiec11 IEv~i ttioi Rep~ort INRC-.04.9 I-040, V~ijk.8 Boi'iard WVcbwr, I)NVIO la00~ic-I. lr!. Anidrew W~oIfori D)NV T0 fiuica. hwc, kit1111 Dar~inh, ieI &ifgnrbgAssoc4iates, Iuic. perrormed rr. Science idEgo~~i So~ts Inc.. I~e~ed foi-th INuIClAr Romillitorv Con n4siom I a ,, ~ 4

1-4-; .1___ _-_-_1__ - I. qI I ll C.. \\ C V. 1 III Ot I114 Appr1a IA RcuI It' h 1 11.1 .2 Plant~~~~) I Ij otmct adu I -le:ror kI A.:c N-foIlhocouogc foId Iv I ti r Jnof )1n Rf PlantrVo~nci ti. II 2 Ci ofvidrl( Divertze Mcpmk ot D) I I* ~.1kXI l .\\ IN AN) )(ONC1A.IP)N lit

h CIt""I( O t~ieIfl 4Subfli-i~tiif (~$ )ter ~er k, tIhe I.susle raI'.4'! 1in iiu i u to thw Suhnilittial Ofniy1/4 Iio, it WI! Ito fie (CN( is slic v ,'I;-Kciw - I iw r eVc' w~h i't I Of t4hI rcV&1 i~ I idotiii fIV ivoie, related r0 ii h 'c (V M an to 10upipK I hes I dtn~ 1fil NW( Flew k\\ I 1w f( I t ned13/4 o'Iewtitt f (l'e hniv A nc C !)N nerowi ctt 7 cn'-ic\\~ t-Inc iSFA ) Iti reVIVWcrs follo d the pr,~c~ ci.d ) I i l cc icr Wt-Yc'A' I.. Iujie'" (Ii IN"' U IJ1and [ejChj Spe I hi. :I,-( n 1 W hrs,hitt111ral to Sk A iD Sephti ~r I hW N fhulltl was thejet In r'lA riiI)NV, DNV II gUD W~roek 61,~O.tolivr. I. 992. FIie(\\kcn ()clohc I ncdt ()'I-,Ic IX. tie Icsvtew toctl'dc oil, a del.1tect rcv Ow of the mtihrnittaItO develop an Hils' ';Mhil I O fi'l!l.f. lIm. *md -iipport syitembs. and to dpintify apparew 0I dtCICnI`Cs. it' ;IM lin Owi tntirtri:,ill vrnl pnwv-e, (If the I PE U; c1

tw i rpoo the prelimnary review \\&v i~t,

'lenI\\ pe t< ;lrn in) hr 1[SAR that sdouald he c,(,inlsuIiMd'f~t-

lIi1Mii, irliaio aId Jd.uI roitioll.

w t n'cj":( o in!' riIItat0 IIn cc h( I P1T. suhrnit ~IL' )n (h)ct '. er 1I 4 and 15~. 0 92, I he lates~t U~pdated Final Skf Analys k Report ('S AR) and 'I e-NAM~I1 S4peci fi~kat lnn" (T ech S Pe Cs for OC)CN Werc r~wd, Thik revic"w was PCrfrd ii NW' NRR wiim., upti tO ale docu'twm at ion pro vtidd by,~h NRR projne Ina wimipr. TIhC locuLs 'I hi, rov\\ict i\\ko ((4 pa imi a hatter Unde rstanding o4 varot plant system~s,. p1sinm design. and -ol

I - __ - I so Nil ACTIVITY RESULT Recerve Oyster Creek Nuciea Power Plant tPE Submit.al Receive FSAR list of itets Of Interest Technical Sp6cficatcns - ntDesign = te C N t ts intedc sues Poe~PanhE~blt~ [i!obe Humarq Fators, and: Resa1ed ack-End Review Complete Data Sheets vew R r to NC 4 7 - f I_ Incorporate Review Corrnents on Draft Recort Final Reviewv Report :o NJRC Figure 1. SEA Step i Review for Oyster Creek Nuclear Power Plant Unit 1 Front-End IPE

I w; \\kt 1 IttIt k .titti N~ A c IIIIhcr J3 a I't ml W c i,~ of 11w~! I eIP I u m t I ( c I c Ie 1 k III: w~in~ l-II,~ ~%~~ ckt I sK~ I rt 1 a.: tIhI ~.b~ t d tu tidtI 5lI %it (Arop rjti( i thIeIo Ilk. 'I hc X ) N( .\\4 S I'i)K iI~l I. p A ls in t tI< I \\ 'k \\A compicl -il ill I mtLilw j~,,,f oL(9 I.'And( rkev Is-eti I I .iiiiI I it- .'it liut\\ I i IP jwn tie I i/liela Iteview IR

U X~

1p1tf NIT SOINItCk I CLIit I~ \\tI I. ). N 5i. lhdto' I I I Awp ;t I itt cin~ a e VI IJ~ oA t PNV6 Inc Ifet tlt teP nil V Vetic c I 4 \\ t i MIrfi Iiit'd t i It-'4: i1 IItjinIItii i f eM 6% apprdcti. nria ht itcpiitlcd It; sc .1'"J l' ; 1 1rit I ii' Ii fit'even fiOetvc iai her thas th i n o I I uk I il. th 'SI f I,, R I~et Srti tl' tth iiiiuisii tjtstos.dlIpidciii ic heti test 'tI hcPP AVited perl oullieds rules. nitewsn Ve'ohtcO' I rI ri6 ot 0tyII) i~i lid inhe Il rII tits tatt2., ul It

Ie em eiu the IabI I I re t(,

Ih I Ifct vr6 r.pmcI~c ttwl Owies ~ti II thCi~'k si ui t i e et c rec 1 ar>i'ovicdI110 ifo the 1] p fltcfl I I aI l \\I I SvI III1 Io sic: i Ii I I fI4I I I I t IIIIII I I alL 2 sp I aIC rulstm b w.crdl mod u a1r act lis .'i iop lIt~ nto tic systemi fault trees,, Crtict Cttcv'I '.1'ed ws aI se nlimie lit0otitilC, tiiI1" Xf)PCCaIiCII ecvcnt trees. COrnllfon6 ciwsva f4Ihir were

.: ..11Y At tile svSteii l \\ ci BI~ i Ii oge nei.: an11d, PI,-p~ i ditnt n

wee n trord, into hie fauilt Ire 'Ill ItIianrill,,m ncf1w c(ut' daiuiagc'l fr'equlency v ia th large ov vfl t tesww~ Larrivd ion us~,ini' 'thetilt n) 1uhiic. t Imiinplemcdet by thc RI.SKAIAN softwAlre iiacge.,\\ A int11cifairit\\ aHiilv1 s w'tN t1 fr `1 ii h 11ii.,or ci iit lorst h.C t

~II)cla. I~ia. ~ e ~i r aIit'a n ran l vi i;II II 'IT_ bie (' li liIv initial criiin , I ix n. llY w May-cs. I Thi' on I c 0111;w I : -_\\1 t ( It klctai ( kt\\'Itrm~i an kio,WArk I peo cdl'd 11I nni. I I1w I ~~'I hie 1 a11 ( crg cp or tr'( acaalii a I i i I I I I W I.; oc I. I 'mjlmu iI~aII'li mid PS.A,. ih'I Ik, aC%%VIa did imt fiilm I Ii.%tiri of sJI1inuir piit or n list of' PSAs~ of simiilar plai~~~.N. -!!k-I "W\\ 1< 2 I'.!va( pkin IS, Mun Mih Po0it I UnIV-1 JNMP II The I [S'AI ,1~, Whi! Ic'a~i iC~ltI: ~ Ct i.Nt;S pri cait IIr I Vri,:ih~ dArea, of rtcacit r. '\\' i.i I\\ k' at he,afnic vinui e: h~eLrhcas ndfaio havc titken pI.i(~k I a i Ii i c oer VII hie cours~e (,ic Onime Ot~allC(d comparison wotihd nai hN' tHi iwi ;t dl ificicce howecin OCNGS andi NMP. r that OC\\G~S has two) isolmimui (4Il(IC.Ci' ,In c N Al I I haN to0tii. '4 I

tf

~t/(III4 v\\i viIe~JM, no(ue LoNs il~wl

,Ii \\ *ICN atf~ iIII)Rwtii\\ ~QCcf

vr ~f B Wt f\\Iscp~I1:iat~IQ.I~i n I

\\ aivf ,Je,, afoL lcd! 'II I 1\\0.icr I. 'c 1w Ht i~ Ut.' Ift Ii tiert( bittl do rCCwt' D)( pwc I tr, t ,, h t t pr t,ur CS V I

O it I111c 1

n I ItlI`V 1v l l 4-t 0 I I f ;t,~(' I r' t ~iH .1 Ifl'SfcCII(I o rdIfa~~ V 4, 11t.. C ',r / I i Int' s a k a)I Isi alk oh 0 I Ii h olttifl C'S i p'-r

ii 4 c cttiiht~i'~i l~f ItII'i V1U0 h Cu ti Ifi~c I-Coll )Ck Itd IttI OW.i iio H *:mI I W I v;I oIN (In' finictl) Jul P~llrps al ~~C (~Iymer ii~iUClliitior Clitm IN jtltvd it\\

li Kin'('iitha' iijim i etirce1, ii wag' Jdcntificd&fiAt the kvI i;eaCil itr rcOVIC\\ arc i(XiCI/)i 'loq (u.~cnlce'. and slalHofl black-out SC(JUCtnces 4IS11111 IWofC ihiaii 1111ce htoii'. C . (:( NITA( TOR REIVxIW FI; I'N )I NG'S II.1 I Uo t\\ie anrd Id(enltirfcfl ioniof I I'll I ns i ghtA ThI N jsI I vIcw 't ft on r fi ndi ngs, nclUdirip a SurnzfIlk&' (I IE sL trenvth ii' nd weaknessc, 'I lie ft V. Ci III t ns.mt tredss eaich work arp~ txpli Rify in th corder thev Jppcor in the VMW

rbc i I k, 1w i ii 4r tSWefr ouydo h ~drnilaii[ io vci dbtel klot-.i ii Hicl ~.r i.nt'drtdaaItte~rdn~tod~ud~ines of Nt 1R lG( 1 335 Si nce Ow tt'~n i ~\\wth s ~Ytwt)docutnenctatuim and a cross-tetcrem c iikc , hlr.I I\\ q, fl i'rIitC 1h ers,~s rei'Prece Tc rubkwtt helpfl )I I. fu rIo Lorn I tI ,,i Ii J il v,' Hc. dv tu'criftm tit mnoftiwhodologY (O s~ifl nlwa uttsss C ,m a n Iu f, I II the II I Itmra V ;port. Thel 1 vie.Wv( illid MCCe a.s [CO%,Idti.I 1 ) Ih t N k< i I- \\t t I,p ui bir mir (Iiscremcfie lIs L itecl t'lllh,Ly)iJ5lht~ik hls re~view%. lith IC doieltmLI I jut, proidud(t( c' ~~ritled briefl y In SeIo fth 0(PN\\ A; )etildCif 1[a11 11CCI VI i (I~ih nf' h~hWerneul i lie qwianitf, iv~tion (-ohncue.

rtnncrulami'..si addc ii into hye faul t rav:,1iodcI, Ii~l~~einponient depunidefe te

.sti h,)J a. 4'11 IC 'n ~i'il ll ien St'~t I 1w or iazinteflinee, kVc tnudWithin. cach .sv~~eriu, hul nut) lk loss, t u I s f~c' c 1\\ Ic I I)I W I it i I plied "'I feaaC ~~I il te, qluarititicilnlfm pnxkm.' A I LIfII CLi`1t V antImdvsi wa.~ tvi~ormicd. hI canchd.iofl theO, pnthodoliogy used in (lie M( N( WE4lii .%Ibihftit tl is consistent wvith thle riethod47 idelt e h ~neric Letter 98-21) and NUM Ik(G. Vcrlv Itlte /c* dxriplion j'S provided re'gardilfg the a lol quaificationt of d/tc,odel. infact, the vivirrtimi" uFl erimft ot' restilts d06S 114Ot eapp'a Fi.('ure 2-1 olr in Hue( tc vt of.f .'ii, m, 2. .6j

OldI r~ I 4 1I 4 w slit-,1ialrth1-i ptV) tilt A1 te o dOwilllic fltdholl ilid litII" iiw tinits slafftIi~~h the, a~i~tjjII I; 13dvpo pln.1h ni"Oti~iid t' CInn.11L It'lt fl' ( f~e~ii'n I~aniNi~nUl, rttrgnc~Oc~rtn~!I~it~in~c, 1&' l)~ ;nd I iccI I it-i )i l 111W', OtI.1[ t I QLktL'LI ilod (Iiie t11 R 'ie.wkir~e te 1 ~taff mir ver t\\ tht.liclim

cc~r~aI~~d~k\\~'~ tt f~tIiti' m

owIi~ii 11 i ~iiict*~ i'~ i

'\\~t~1~*

iri 1,s 1i1WO'1Vi1. ofen Rjh Ai c'vwani.rt kiant ctuvinkcci I it ( treVK ic4. flet v~Ii.'l l otiX I Uris imi 'dootminO irkood iender.i I ~C nra ii iu i 0ivi IkCloent 1t6 ( 'Cr C> (.OtigraOfl~fil Jm)(a l t' e rpCI l1. With rcvi ows ot Ih iVflt

Sacqtow i a.rS(SI s)wi 4Pf uI:

depll ncii" II2(LJIIlII (1Ii) ve iity thc VMihduiy 61f ihe plant modolse. I hP is 11 (dl c o;rmugh andl woulmd Clmsuirt fIn tp flant mod)(els' reprtsent thle ils-bt iliii. u'-ise.114-dl'c phaili. Thei FKM s are %,cry h~elpll(,in ocmenthig plant responrM' to accidIeiit II. 14 h1 turnal Flood ing Methodology lHic ( )(iN IS Hood( anlsio' d~d~ ntwo wpton cornpkttcI as part or athe ci flNt A m which I OC A minjating cnt prpgedhr h the basc plant model, ihc other pctiiida a sparatc screening' anal~qk of spccfl flooIng event's. 01.)()d! ' nI whiich a-ppcar-explicii:y III the Lg~v.4 I 1PRA, flood effcois arc addit-sed Ill lt(ii Ic' llodu~icii for Ilhe inin g'aling sy 01 1rns.ir:I'l ysc~9~r.th iecliliani'IM lscs, Ii I'(d mi1k 7 m l I

I i itb 1c, 1 iil N~ v ito di fl~ Ru,1(itV Ij Mfng 1ort j 'mcI Ittt dn tw l c 11c t ih i.l !,If!~ !:ct Hi od-Inl till-.!r ne Btijlditi4~, '11)i (liew,c mairtjit IL: t'cu I O mitr: Pt.:w ci~:~ St:iv,1fitiii c i inlrilu.r Ic C) If'. -d6.- rnood wlcrc V \\o 1 I tImc I -- I ki ii. i,1 TlB I' i i I 'c I 1,! CIi O~i:111i II wd I i'sI the M n rch i9, ahzmbr,,tt, I'u' t i,v A'" vk q ' i;,' tV,:',r A'1 w, ir PItY,'.r IP/anti. Vo0lultie A~~n1I~u~ ta he tict'1u(vnitCl tili wut ~~~~ \\crc fioiii1twnned P,,ccrdiril: to I[t!e n Limhesr O%'Cin ITIS 1w dPPI1C;ibc () Ct 'I lit, rc ic wi-.' IS 'llCC I I Kin 1ll'( to Kind II 5(Pj)UIrat siifmmifryli~f rcsiii 1 for the int!ernal imiflo c 11cc is ithichli rv i nchiuded( in the LOACA, anrlayvis, fir L.O-CAs outg~de of' conl hinm ieai I ha Ii or id gellict, lthe ann lysis wnais. 11 IIrogb 1) though, dic~tilt to foillow. WVithouitt flie beneflit of' a plIant l ((ir, it is dif'rieult, to, gain a clear undeii, ndingWo the spat jal ashpects of' lthe flood analvsis. We are not convyinced: that afll ooditng sources or writer propagation effects linic( been con.sidered. Based on, thbdeAtails` prVioded,~ the central conclusion namecli Ihat thtre i.s nit significant (threat( from loternal Dodlng,:- seems reasonable. We (14) 11m e~er ole c(liniti8

-in I-I II 111111 I I Ila

I 1/2,( *((1,)Pi 7, / 2 2 C'/,It PR fh'hng iim/ eaidexcqiic'ii Iav if oi e I I a H in . a Atiteiff 111)1 'iC riiri I A uijtjfJjjlauWun in ('j, I'Ciii fr 'p biluf flu P ila! 1,,;,>;' lure IiA('i (Iii dpAre'i:aJiIuIg I/IIV j,' I/Wi eeiij/Hjt'lif flint' 'till l)i (Ii r(,/ ! fullil' ol f/:c oPt' A/u'I1' /iiu1i/". A PPiflPlIifll (peraic)r ji, Iu'tI is tnt lxi lila/v tweu/e ii Ii' e'/llli / \\ 'H Ni

r. 'o':,h,' c ii': it ui ACrC(ePiCul ijl Au ri ICt ('liCe' tA / e, utue/ed I

. jib hi'iluite' I,, 11.1.1.5 1 iilitv Peer Ihvk F i 1w inlependenl peer ie' I( nrprvided in\\ppcndi 1) iI the I .evel I PR A I hc F' F.: N wv Ic wc I ii sing IwO pira lie! d(ort onc A tfldepondenr in -house review ' n"i F:' 'Q 1w in e\\i'rflal wiisiiltfliit, both rcvio S I(Ok pkicc cirl in F9) I 'ilL' i: F Ci; ii V IIiVOF \\CF ii I lit'. 1 .ev.oi I PRA 1'h iinip tIi(! Oil ni tic tR'CiI,I 1iiN Ii 'iI ii it Ill: I 'Vii', j ) I II;ieJ on the 'oti 'iient litd in Appendix I). 11W group jt'iie 'i Ni('tf an [',\\ .!Q, 1 'ri'iiiii'l who \\VCUC hOt Thuiliar with I RAs Occurre(F iiiestiv.11 lilt F.,, hi I 'i.'rrii I l:>l)) level. A tutd In the PtA "it i o$m'1ime. difficult to direc'1v linA (lie /5/) 14'i(Ii i/ic' /U 0, iflo(ICIS rules. t/.iis. provid further juslifictifwi for f/ic rcdcr In ('I.' i.'re' I/u/I they jul/v ii,,dcrsiand Ihe peonI ruls fiks qt)icr d*ai focusliag solely ot ihe 12*1) Flit' c ' icuii.t rc view was performed by Dr. I)vid II JOhnSon, i consultant from P1.G '11w if 1w c'' ient S I ndiiitC thtit III r(VkW WLl detrn kd. It k Worth riot I rig (hat I lie indcp&'riden t CL'vi('wer was uko the proJcVIL niiariagr for *PIG 'S contribution of' the I.ei'i I I'RA. I)

I I. I.2 I<6' %i ol Ncciw III S4cc vI III Io~4 :III wid, M ki AiiHillvsi It, 1.2. I Inintialing LKN4ciii Ivvick% I ~ 1~ II;, H iml iil, C \\ ntl il' a v arrld ol /1I reComfn t~hCd h S ctIo I~. It ofIicI4.f PCC lic '.uiv.II&Ii1)11Ohi(htIot iTo~rlitdn ~ ii C 'fl tt u i 1(1git, nIt, 4:1, tic11 c ihdsi luH: (i \\ u\\ I \\hC II dI II I I/J I II k .I{i(et I ve' malbot~ti rcl SC Illav becoueIuIc Ur.Ic pI~i IHI iI Ilic ('sitII l OYaf i UPl t4i1 1 Io wiala S dHilf I lo t~i Iut 11CC Ill t cCI 1 v a 11 I cfC ~I s I 1 I Is on'I~t flielsi k (IC t1t Jiiint ',a to~ I rN':ivic I : imiis 1cpItIII aII aI~w Is t -C1iuI" col 111titi t mt h ip~ SCd forit'l( 4 tieim pos i II)

ittdii c

I: 1)11t I 1' i\\ I od I cL'(% It bakfYun 4 t~pr o the I s~k IllodicI na, YVYcnms Al kII H I)JI'C)~ itt 0 iIN'u w a'l Val 11 d

I rtCV,~*

('tio oilcrat iflll pwo durk.IIL m.,lall.i% Hie ()NC ;S I Rt<A re )I. Infaigcct gop.rt~ duitinc cvenutK At thi lecvd. I IItItI '.J IcIII) I rdlt 101, I I IIAny events wver~e $CCp~lid.ut'fiOting kimfuiciatiom. 1filed ii (11 tic Inforirllima H[ Hn.rlIc)ltI ill SCC1) 1n '1.. ~),Clev~4 wrecrud Thle U 'N(S 111IlK appears~ it~) ot ilistify scre~ening out ot $Cyersl ini(latbigi (welitSnfil I. I CLIwater linLW IbrLaks outideSI coltaiiflflniet

2.

(morv nlow blockage initialing eventL%

3.,a k a ie ait ( I10) or rinsitrimen tation". pnetratiobk: @ RWCLJ bot tonm hid ~piping Iivilli lth ImI toil III the vessol.

\\ A AI\\\\ ciic I I,', IIr

,iiI,dIhtpnef1$xi Rt

()*(N eomeiuidvs Iha S.icuIN IrOLl Ict I ("(11jtic pnwt k~ cifia~be h of operatitg~tg il hLtrOfcIjing, i llticldi n.g O le Ion (rol romiin. ~fII Ii Inr )i)II I req IL re v..e III Ia I iorI II os.4Ih ak ele hd~ Iwgel r roo0nis. 1111 I est' 11lai1, res ir fitt Il1ON( L'tCA prt)tildtI as itietiirige dkt Veaionitsd )Ic .01%'1 ISJ 10. ii: I' 1 ~' h I II Ii

a himi v Iu p r li m IIII k

X i 1h nli i:IAci Of1l;tt,~tr 111iIiill i. cx'nti i's d 'c scr Ibc (I, ' ection 4.0 of thed I evc I l'IR.A I ~ uI c I I hrisIoI ~ot hc xlctt~dokvii '111(1 coltih'dchifinSl 111.1d In III 1 1'A 110( if dtic xci p pln ~ pcctl IL, I j )czcircl, Wasl tile2 s.11i 'Ii [IIJ 1 11 "rt! I., 'pitii N O ~ponctlail tav uepor~ilr ~ ~ d met hod s btased O11tti he 3;x IIiiC1li'iIit I pn IhNthl IiY.a( in'I otx he develOpmenrt,ol ~a prior, dk~tiri)Liillt br 2il iiii it iI kt'intI-k itth I art tvnt pohhti~

--bihuly,

'e v IVI1 ph hilttviN I id 1n ( I'di('ti d s it{ iji i ~ I \\n'IC Ilkiitau vn..~e ~~o f3V('W w('id (jluaititif cd UNIil' -t pki '~i2;lin: t1'r(Jrd1~c '1( i"itInit)(1iil Is tdt'-Cf frorti opCratiny' experlecte iepoitctl ti

1 t\\IW P Nw.1ll IM\\Yk dl~t:I xxccuedfreN ceI e~h%*

dillerenees inI planiti~i hio 1l't Ifi atC ii %kc (I-c rike(' it omi ( )vsIC C(rcck Scrai.)Data and trasiment evenlt rcpoitis r i lJtctors, or similar inforimilion regaircing the distribullon or generic daita, aire riot providedd iil tile, I PE)'. This i's imfportanlt.inrorm

ntIjon, tince 044"ro'r fatctiirs can heaivily influencoi thle flita po int ust itlia I tse(d ill tilL' COFi(~ftliltOl

'I hr 'p antif FC;i oil mOf IT' frequcoc ier is rC-$Orible., T he su0brittal i A~cs effective Ud ill h' tihi VC1e rIL. diild I II Iln pc c Ific d ataI.1 The1 I 6 6A d'4t1a isrot-refe~renced, lhowlever-. Also), ~I so V. ) cck of I k-ev IE LC oss or O)ffile Po We shwdL C~ i rrori factor of approxinmatcl% e,

I k ;I clIr 1154 Io ItIis 4e n rt 9& 't IIIIitfhi g k4' flp t'rnI I' j! i rl cktIrl I) III Il L. T p; Ca III. IsII ld:II; (liE ri lili tites~ i in ill ist~ i~iici ji of~i vefig tiiii mi4ldl~ ('spjet('t. (,(I M IairLtor ni Ic'os IlIall The lu' ijh c'rrhr-f' r, t"Wk(I~h to rdl'(liv'( ilt'it, oanl %,Sih'. I'iujn l I'i'r I I, il l i1%- io 2, J) r((id cO 41 iiimm idi olue ' 6.21>2. r'atlhcr thllt 3 i-:a111 s ice th1w v. iltii used in the -1PV tirthermiOre, if C ith pkintr vvidencet iil I itot I a if wi, Wj h a c ri-or Factor oif 31, the~ I 10S P (tFivIlienev jul1p to~ 8,)I-2. s int rice 1 I. ) mil kinitor (IE 11iniiltts thk WEI fain -( niany.a oter ) e'xAmiiriainu of' flit' prior distribluuiion an1d p)ian~t-p('cif1c, evidienuct nillv Ie~~rnc N!j c vic

s c u l' a ( ( t I Ip

In w rhint.4 t' Ip Hila Cl ai 1 11

K!-1!v WIlow' k(ji'.\\ M11011i Iit0l I l

jI)V'tcl (tc'k htCCi)7ffd.9prytrIt iS fi6ll IIH1 ui I~tiWli"L ItiI) ' 'i hi iow Ha ics r I li'cst 1UR rI IkelIy ODt bei1k ducL tO Ow Ic lit ti II\\ 'ot !I Ill suiron1ari'. (lit' inlitialinlg L'vt'nt (iI - 14 have Iteni dor v MrTI I, combinat~ioni o41 lf vocr4'it.111d gi' I hCiftfa I I li' Soil rmt liiivv hccri (idtt ifj Ct'd I 111 fist ~i .aIthlotigh Som i( of? fl t h thora(tIItIir '1v implclul oI' inlitiattinlg eVoits on fro nt-lhnv and tippoI~i(rt 5yVit('l1s hasliv A he~t icl f(It'crijion(11 if' nwl( procss f'or scrcnlng.inlthittig ev cn is f'romn conrsidrt I'o n shollild II. 1.2.2 he of lront-fine:hn Suppocr't $Sy.06 n~vi Thre b ilk im?* ini 0mt-Iine and.guppBrrs sY',aeiA:l~ Were: Anti y~d i n dtet II 1 I ronl -I it'I S 'st c MN5 vli lor I'E ott'0i1(0111

I i I I I i I 'S 1,1% it I a IL ti)crcs Ia i n vt-f(A1S -I Indk .h j d OncI miw )"I.1 ati, ~ d I ihli 1 III 0m CIo e o lig \\ and: 'm ctI Sa et vHal Ref Ar .'i ion rem ( srv adc'crip lio ol 1he top events cutinmd'.rcd, n uc critct~ ecrnemr pvidlcd f tmihe Klacs (if miost success c~riteria. Also provided arc; supports systmscquircd. sy'sterns supported., c mlu: i a i~l an op raion peiodic testingA f~u ircmcnts; m'Taintcnance, operaloi c on.Ih potem nial it causer an I1V-appcahle cchclScii~inhocl assumplions, split1 fraction (dchmti!lowl cW( tll n CaIuse analvsis, anIIdrestilik. The6 discossions 'arc, complete and tc) the Ilevel o rllIII (Ic.dLgtICa It'r w% Vt\\k, AL1l tf'C6't afr' ecflCosed, ttIia snot a reqtiirclcntn I13

II.cl a i'\\ clotcd n il~ Il C\\ t'It Op i'.c li I Iicj viim! lo!': modi-I. i 1w pIn Inw I I ho C o v~0 Ic I I Ie CI Idbr iII In dl*' 11011i Ci,'I I' IV tV tcn is pr'ov dk di l ~ ~ h irl. 't pae > I i&lne hI at cr.i ktid ipfi~i~cjlt SCdp)bkW of OfX lul flP

\\ Iton 'iii i'l tm O'.n

I cni' ar. th

' cowro rh om 1< A d th~illuI I i flll o St V Isn. cx a mpk Is Ci 1w I,.1 I ihr I Ih'i PR r if aIllodls c'C Itidetsif n o O y~fstedcI(rcdonexamnple tr hlic I n I .P X \\'I. it c lIt' Iii1 'I otIŽ I 10'dlt Ev nt. t 'llt roitvu is f~' .0 )I I

i.

d 1 eo o n 1111cd. hr' I W I If in'iJ' lt it I 160. 'V Itdipe~ I2%~ Pf'o" ttuIde RiY h ' wth 4 11 I' 1' ii ta, a' i c i 1?l~t n s 'i w irfrirktmo. fi~l~ fo r icreu pin ti A I \\Vmijicc

ad t he it drli, f~pm ~~~OAa in unit ing vvent A,

lhasedl on thle rvvic%~ it is cotncluuded 1114l al ipporiant,, fronit-line aind support sv~,tems1 req iiii red fo r prwei ention or core damage are m~odoled in'th :~OCNGS JE except ror II VA(C aidruci rc ptinmp Seal coolinig. F'uurther *jusif'iCttion fbrlnt m,~Vodielling these sy'stemns should 11123 S.st erv Depe'ndencies and( Support Sysilem 'Ihe r )'N\\(iS IN:Ivrpef'ri'ind ;i(a pcc)v nlssrf~su cedni' and s~~n 'V'.IVeIIi' lirter -~V".tC'l kCfpnliC0 'uct0S, wero treiliod in thfiro gr0UpS:- N1u111irrl to (4) ipport. snppoiN41

IlI I c Y Inc. J1~v~d.~-e iretitnj~n c, '\\h. ii' U \\ ii it f' W 'C 'II!, In 'n ggJA.H wi IJ'/ d t~ frfiOi 4b'

j~I~

Ix le picpr. (ft' I C' I,'C

,f~

(:~t~tI MP'i IRA s/thiuldti I'd' (/itPW full ri n tIi r' i '~ iii' tiilii I(Iti h~t Ihr InI t (at('d (lepeiuklcnief. bwi wecu pilant w64vir :n ~ilit .1 lcI'.rl hIcti'l( dIl 1-folIq.i'tfet itilirner. (le Jocfickcn :,w~ 1 ~ re ideleififited. othetr Illalii 11ie ~ i6 111i I 1A ( alld~ ITCirT ni icIcolg as previous.1 (fiscuPs.%d. I 101 I-2. i 4 .co tlic.. I, t(111o I Cale 0 f4J a i h lKdl IrepiI I l' II' cIIl ' Ow ---ocm'. mfidoCin r.in a rn I I i 6 ictly I Wolfli~ f Il~ Ii ke i1 I.I )d l t I ,IIN IV ir I 1C' svni1r CAHWSC~. to Ocn)C(1mCcii iIIVl I 11de Arld Atr pndel Nifaure arc 'iv in1% ItI~~l ll h. fl Ow Tcv \\kt Illn a clucstioning SmntC rCpardifIp til colillintmn 11SL o tI1Cs(0 IvM1iP' he~ ;!e 1 Ift 1 ) wcd 1

y*lrnjjf (I?.pcfl(.ldt ftira c ds 1~ the J

bnwn miuiliple (rck Icui te ( 7L mctIIhi~ do,c tipcd mdi rel Iil ncI il PL(-G. The OC I PI t rdi~~ ia'vtiv upon re ecrinced inacri d s tt"II n)C n mt:01od.Ii . pro ccdllres~ and d~~~ ~ ~ h ra er.A ~uctc' nc dcC o!I 1-~ ~f1 oIx J" Icr ( rvck-sp~ic w N(;L: dmiikasl 13 g~' t ln tho O(.) RA sc Il4 i1

U V tt I I I t i I I t L b i C 1 e i i lt f l U d n i t i I 1 t i r , a i oI titl I,u at *C ~ ~ t r1 O! k ek ; Lii 1 i v t t'i i'-' I p I*;~g ~ ' c i ~ Z ( ) c o I u ok a t d ' u lfl t I ll t yct r II!l jj ~ j~ I r vp rw lo dt i ( ( CvNfl-IU% b i4t t1n dv It. to l t tttv c(- 'I ~,Itcri I ! I li 1w n t u r f thet ride s-n kl d Ufle4 l~p~ c i i t, E~ l i i i ~i t, ~(ft c lm lE i caIi.S E li I t t a ( 'I )l j it t l e c zi iI y V rj r fih , tf si ~ltii i a ltc c oif c(EIII nIII E EII t ; oIII III. -, cf l .o I ~ Ie w do le I i lic 1' c m lot t c r ie ; m n'th d o lO y a~,~ ti c x H itr 11 1 c v ent ir et'.Efv ril n ix ~ I h; h' - I 'A I ( ( i i I I bi I C I Et II )ti I I IIo flfI's m f fi~ t ~ c ft t i vuko, " i tep resent IN, c' ti t t. I v :, h ~ I~t '- i tn r fli n t )c 1 t t U f I k~ I Ilk Ic ~ I i I Ic t ( I(Ct tt w I cVih ~ cd b thed coi s miuc tio3 of viI I' n "c' I lit c hi it "t ( IS I ) IS J w rv u c d 0 I C t f t3 Wcco" V lh, ~ .,Ivailathlc Ioi flmtijr t i1w IlC~t % i itII ta(II f~ events and Mtb c u n y m F~aIlures. T he ESI), aire reviex~ ed and it a IJa c I int I ruh>. 1 I S is ImPortt'mnt Io nlote i~ t h~ le ft M )s Proiw d e ',Itll t% ~ hfC h III tIcr r a nd t tc v u ft ir no n R A in l t rC k A, T he p idnil mi od d r I' ll e filie 'it c t he titt I 101.'p05C'cnt toen Of1 thle p'la nt aI. ff d l cl d,And iu t uf eo I icriI. ihv rule stw o are 1on e1 0 :i ~ Il [I I t I no_ mm ___.-.-ft

!A' IP I t it I Ic t.jg pI twilt prtoloeo; providtd'i tilt Ibn44't ar t ir'li oalc(e 4A cril trili. I I)' r tcr ris r ' tit.11 S~,ilf II. lC (i iti (rit c ri a ':ic i j.i~ hI% O I lI i~ allIi ii4i( ' t '-i(I (Iru uII W tIW NiWcen fh-frj9el. a' III it, fo l w i-t fit ,ti, e, I; I, OiYltI i I .ft~ riiffliei' r*. ~ *',*, '~ ~ ' ~. au'i 'AI wow fi M Iikh I 6fpf l! i t' !fl /a I 'a 'fa 'YZ ~ /u .1tdV.J1.? h'.qItta a a~(~a apaa~a' '~ Pa Na~11101utu111111( csvh'tlt enIr thk41iiWdk(rty hi E)III&% paIrticilarlv sl hoc 'at 0 0!In 1fontuurvd1Vi optrri icki0~~~,tI riseIradirIh iaa I dwo f o ~t) l) fi if)g lettf IrtIIIta1CiIt $tinl expflk C itly I) clc 1 .1ie :wT' and 11lrhiflic Irptp. Top C veffi t Iis4~c ~prp, (r tho IFU'N udent iffid I hr 's Ink ldc re~Iclor' frIp, urlbileic I), (2S' co'flOy C ofdker VaUcuumf, iurbifn bylato~ cwmitii of !vot I %tt.i, vr Itv Ic \\'L ` i uArt' m, de~ '4viw th..In khktl nSl fl)a if %1tcafl io;I I loo, condeni-atc :und:

( )f'\\ N~p~i I~ . ~fthy v~e.. t~n~u~~ *'?tfi~ A1t IIth II Iit. 11 I I~ IIe fitd[e d a y ((m nhev ha IIIotl, c itii I l II( It III I V IIIII I III v c ;I M o I L. c I I cok WC I ojti ce I N'atc ifinh I I 'N:odt r I V e~~t.tiy ~e t t~nict~~iy. ',fA~~ i4 t~ I cn~ ;f~ If itlitim dllp ttCn.fr~tit-)'trp ra t '& t Itl

I

¶p;p¶ IV hvAre ' I rik w1(II, thi a~)J1ear'gI chro,~ rint I I reflt~C if ftf c ( Ii u, m ~t f r'hme(v f(rly tIaIk Ii loll~ 3% (phoritfIicd it~oth-n i. AI I 3/4 Iv I if 1 o11 ('thti' ie01.. io Ittic v on~.d flux ~ip i ( taile. 'Ill 4" o il, Ii ;, I I '4Iimii h bc III' ld I fil(I Ia 1w sti'e~s criteria fore eorc spray, In, 111c fill', (iffer-14 froil that .t%313CI II (iP ISA It. Whiik tile FS"ARI Chipter, 15 111IIIl$t5 reuIre twio riain jpt1lpj) aiId rile Ioonslt'r puIIIItp b operatble' if IPJ-, asgurncgthAt 'ohl h itr id~)l )OtLrPf' 'Ifit g'ellrIal I r~itrisie't moodule dncve not accr~teItgiyo h clcht~lpump Se:,vaN rimmsi~im r1.I genter~id tranlsient. TIh~at is there Is no0 congldezi~lon of recirculation punmp %eUil ILOCAs ocemiring (huring the mitigation portio of A genj661sInlent. If thet rtL'ircuaiotinf j1IIja N sreIviu rIW I hO seak utr COC)y njecition (proiphabl rrom I11w ( It 1) M ftn'1e I fir h~v cmil roIMd leitktgto wifiltc~opi~ng hWL ti pm lpcit HA O/earing votilitig 114C' S %ei sIlue in f tIoth tlw~e moef 111E coling, ii re

n. thiv M'iiI4% %kill Atil evvi if, (flit I1.-

MIIUU I

p1) I11 I I,ea k vIl ripv. ifwI(K,' (JCI'eftIV coivRt~t~f~ pumt~f sit O)v 1Ii'r f rcck. M. I I ~ r ' rteek lea1~ ikohj~l)im lwIfl'A OfuJ'UsJ c ~,trni and thintraery too dic cwvS %I i III IV

, C(re catiingi dlIW8 (it d(reciI. In16~h~etm h

fi'eveol. lend therefforv leicsc tlermie~ (ilc %t:if aire no.aftScto%~ nd 8.2. 1.2 (of lit* R.X rid ict. 14 lea lonlg tert i Inkeilep !toztl)eycf~sel i not rjulired when cooling witli the isola( ionr lhS-ite I.9,.7 (if th KA o yse ri ae~~ielkgc remulting frOm ioss (if svcal ICl)olfling

ir ccepIlabk for two helarsqbrwmver, hL7h t jir4t1v 'gnariig te bung ternm r(t(qI m etiicitn(, ffir,eat coiolitig or fihe coue 6,qu110R r te of r-cal cooling. F'or example, if aI seal k';,k trat of 101) gplii Is 11A.411ned. (IntlVSc invento~ry of 200 gal per inch is I I sIe mI II IiI I I) ci Inee ofS I walei or miorr:i IIyI VedAbcd)V tlwe4Jo) of thle Core, then " it hotel eA ki-g1)it) 11to

. vess~el filie II),) of (lie Nw wIT i un1a1r inuhou JO) h'iur~., 'I'ic seJ c ;III h. is".1 ated bY ct61o il~ e isol"f Ion, ~nV' int* cireila Jt imn ;ipeing. 6ett this I-vq ci rv.% I'4(V ~II~I~~)I 01 th ' 14(FI1I find electrieldi power This' mod~ (ille Is conslideCred de'ficienfl -iit ~ tooit nof addrem (h O po~imibiiy tor a po(Is.t i rmiJew.clIA) via ther recircilafirln pum ktaI~ N, I) 1.c II I ( )('A vi-Ccirkinr of bhtcikk i f %%tled4*1IC1As h4e k;k ncstl-below %0hich ilj' iismd~je lo etcvenicorvIyJgs i ! ;Ot CXCf"4cdc;RP promitrc vlowv cmrc %prayv fi '~:'l~ tumpret Omfuoff head bcforom0c the ont rridps,. An vquii&rlon( hlt'ksbev forr S111111 I()'M S is not 911tedcc Iii (le IIl Ul~~tiI "K~uhm f 1 %IV KA typlcally consider seVeral stmtill LOC(A break flze.s for stcirm/wale cra~~ne %teum flashing provides niore e'fficient dupiressoi r~i~t ion. The folo I:I (e,;ht 11ons are' modclec4 ii thc snul tO.A 11w Iytik: rceactor atrip. condenisaci'. Ivethtin ct AD, ), cmv yn'*ray. fre wirtr yem pbt blawkup to core s~pray, j.mid MNS IV iii ~sI lit j'

f-is Aa101 fA* I r,.IIi I lit '!LI n ' i~.ttF I eAJj% j iL1 1 A-Jt l4LW, A, ~I ' ll, IA~./~2 A ii fif (c ru h'~s 11110it les Io c s AItIj IbCA hcvs hItlui 1 rules are frinimwialonaI vwrrect a iid ;Ccomititf irit short terni v;i)hl Ckiponio4WA eoj AII 1 C nit t virving _lfwatimis. Ji:r ~ I if 0\\ 1 n,1it' 01I1c sor trift pla-nt ves pNfC! to I "Igt.9 CAqi: Dic to the design W' ilic I~~' -~ i~ Io tI. IIwr tc ri re II I I ik tTrt Iofpsi Imal i()i) I-Or at larg IX.O A O\\thwt 'ii! ~ alh!v. k l'\\ tctloo(d ;IhmlvC lith IOp (f i I

I iYCr11 IrrW fAP)

.air th1i sr rcson. tfie 11I'l niti cled 'I 'in i-ivi of hifc 1.( )( 'A: aboive cotu m.40 Ieo d~rjc, depwidingz on the break loc.it tiI 11t1 If.pt"I I Al' A ltvedw"aten' litne IJ)CA 611tsilde-eontillIinnint is notl addits~sed, as 1i,(rtt1141i1cd in Sectimi 11.1.2.1 of. tlis 1_0veW. II. I'ilId itN tIe II ol%,liwnp (ntIino Sy Im~ r0ilpr % itt, condetvat %tra;ctak PIAI~tu'ii 'd 'p1. V. l I fi-0' water en jetio40ii (hackupl 16 torec ~pray). An inicrta~cig il' v t f I Ln I() ,c p ~~tia ll fteNWU I% nIleI Jn hclzar c leaks below \\ 'I1'A. 1c'-vi% (of' I hie ruleks, mod ittes for 1:Lir g eo LOeshh,`nodo ICknic or incon~sitencies for thye IF',, idvintified. The reedlwater line l1OCA Aw a' no ftddregsed. I,III t. Tel III I,()C(. kvshpmvkPv MmIt)i~le 'I liii' Imohule ilmudclN iliv long, WmrlJ plant resports to~ both 1tn1 and lirge I OCAM. All Ii )(A uliutI1tI0ir e llCLit 01111/h s tioddtie. The out tof Ih luIng( temLOCA rcsponsc modleOll islil u ht tiw iIm~ xc iodtillc which is in turn analy."0ed tby the,PS modulc. Thc long term L OCA motitutululte twhcs'cs thec per'onimarnec of nhtc f,(Iitowi ng sy~stoJl b1::primary containifrict ivolatiin, ioflit!:IIC IIIIII Cu dinp I1)co0ftuIi nil nt %icnting, reactor buildin 1isoltion, end standby gas treatment. I 1tcme IVStetl1 pim tarily pirovide containment czoofing vnn iboakn functions. The rules modufle k itti Imii lycmrrect, no diclfrldesck wvrc. (mend.

mi~rt-inlilli~ll I Iv~it RcI1)EvaI Jcc~ovkqr Nlloclf4. .1 I: )dI e I n II he ON )~ter C(reek I P1`!'Y u(rdr~ htcvr

f. long termi Conta iii i c ill IW.rl W11Wri!

T~Ihr" I oril (

IvC*C(

IVC IIPP I o i LfltI0 intip Ions ter vessel rncr111 valbIcI-In' s 11rvccs fuI 'ado cl'.itarrtimlcht, iWith. ~ontui t1 shrge of decay heat it) ctirurt. vcrrtrhiledN chall cnging~ the stretngtb of th rr

cnannnt, III IItvc I" I I a.'o il Im It r

I (.(clcd hI I i t hr nrdiII IC irICIcldc I I termgIC rccovey (f (Ic power I'Ll LIIx I ap av I' v1,t, i it inl rtimnci air rrecovc6iy t torus vent, and long ternm RI I\\' NVn lb~~h~so Ii "P% d 1ing1P If~U~,Mp-rt mlodulcs are crealed ifor !InCIW al1 !InN,it1 a1n I 1 )( 'A rc,,ponc T[he otitput. from (hInd~ ~ hndrce into tile PDI S denciiI fsloi l nd 1,114. IllfI(IIII. kI' is gilnct ii,iii co-rect, So d hCwII s wr on I I. 1.2.0~ 11)Imuin.Int sUII(lociev! I III C I I I lii arI1CI I ILII'III wnc fo OCG _37 F/5 1h E eorsC. nii ('[hi tll I1, t WI)ol L'Olftfihtltiofl5 t 0~ pki;nt &1hrrrage slate cnrh~o~tC~ A UICI a a a eo wsqeCCS I gie In h VC rl r O~~pciit ntiatr, sytml I a Wa~ ~la parngaslnmnnt conltributors., IlttiC~~iarttcq for each of the top 2() m~CarIM11~ )S i providcd inApplcnidix C. Thles 'ontrtibutc 2oabu 62% of thc total CDF. hIlo di 'iii'finT II, ate: Imos of CJ) ire k ~'r (Stion.113 16kut Turbinei Trip,' Reactor Trip. MiSI% 1woI

IIlt, I

I'ml 1.w M4* I Idwiicr,.. Losgs~j 0of 11ndenite K iumtin Hitti I los of, TBICCW. III6LIC IO (WII lt~III ?1 f I) I ritr rIW l'1rge Other couiti llutrors are-I I, a

1;4rtak c NI;i ! II,.ich I t i' P rL'S ,ninle II' fl4 I i ow, mlt ri: ()(A bbv I tc ' a n t ilov wl K cot ihu~siiO~t te DF ~ 0CtO-ta t iC Re I C f-V a I I ii i; lmlm;'( ip whiirlC 0! 48%u)e LsF.ft iic MCpirb,~~otiht n7 t1w() I 'IMhiL~m UIif~lilhiiiCld 10 ;Ihow 33%O/f r th ' coE.I 1it IP 1AkC (crdit 6IIu ali ifl!ilc L.O\\l LiSeliicirgrv0fIte

0101,

~ffit ~)hut Ieil on i itrofne isnt eI, Ink) A, lieurIira iii liier MI a I tin.1 ii '1ni I I i timit. A lit ionrs titoeatriirSWi r~i(st 'llacin iao 1iiia iinen d i' nta i1,1on ilt core spr ayC, i rey D p I-r, VC( Ir Ct of Isite (' Lniim lol Of It rlireti). indtt h? o e oopSneIator iwill ortm iote f IS fis(K ~i)3pn'. vh I II II v ot I I Al N,1 I V o w e h i

s ~ig rIi i ic a t i.'bacatltt IVA dIWSrca~lor c~oolnt toe lmcap o I tw boLts (vqnlivrdonl

[& i smnalWlI OC A) 'which rqt"t ior low prestsirc coolant I ii jcA o il Tlic cont intncd h acat reject ion to thet toi rect dtad one wcContainincot cootling The absecnce of corltainnlflhi~t cool N ~nJ~ o1~o NPS IlI for thie corc spri.N til e s. iii It~ a omncuefailure i h ~ docs not uppcar explicitly in the topl se(Im.1lce's. It appears inl file (op) Sequtweie ba ist k l vct'ly~ uwiisked. The moi~deling or tfie C)1:111111C~ 11 tOi Ii i.V-of the I)GS is athrcfungdeto the change it) s-a nrihie imno ti, Ufi I I

fur ih ( ( f en frII IKI FK 4 to VIl )F D lDti1rthcrwlorc1 thle var rail (ldvsLrile iwle ci cus '3-l t I hlVl~ si nt 1;uut 1,1 s lsit nin jiar. thtr~srtlh in thale (lesil(.. 11 he xll l! 'i check het iCllil§tix l 1)1~us : I.: u m .hd.o t ls fsqune r\\ c i iI'n it I o fIc I1RA (I )c I iluc i tv lt;) I 014,.ascs,.we Could not re p)rtdhcc vthe c id t 1d 1q-ili vIi~tc pr~it~id l in (hc ex t. Ih TI1,_I~ f~~l$ t SCtOO .(fandApenl. ( I 'k)A wortr'.3/41ilt'(i. p'ift estittlatc Value ' d sh1tid'Kh oLi I d 'O'O"r SpCICCk calculations I'i SC I f l!,q i~' i,;- l Ii111 htit cil:tI0i.ti i Jd tIC is' /t a Tte 1he dit.ere twe III (t i;t1IIt,! ;: 1ill Ic T illin (;uoted in litl (Plis cIaust i tIhe-tincluded sotim Successs evcnts wit '1 ih I ,hJlii Ic's, Ownit 1,0 - xvc 's~vd I. o Tei .' evontf I is'. 4 ' (si stI rl) !.). fo Il low di tilL iurc ot both cucici t'c

i1 us

,Ijtl, guiticratoms (shit ie*.xn. hII, aJ'ck.l 3h hee .11 of il I ( V Io act,,s.!rt<, litll *uat io l thle} Ri:llhotifl lof' Al t l ptilet f Fhe E litvlR l to Clos I \\ttuttIttts to rcccvcr (fje ,c fai.Ic laving flo source brt reactot Ic vo,, I ii i~ikitp.lk'cup l Reactoi if lein I d, the, Iop o aCtIve fuel, I'I-LU :s .iCt l(:opcral iotn. I-ue fiiilkre. tk~ s'rncd to occur shortly t hem .e,... r -scenairio umrming is 37iintsfo~lzto core d(lanuig. The reco\\vcr\\ tofl oll;stn power event (rncan rruAHere in f tilurc.258) includes credit lt-thec future cmncrgcncyi' powcr ste& thl 1Pxorkcd 1River Site, ad'acent It) 2 )1 - I12 'I I'1 Tul rbinle trip I. followed by indep at failures of' both dfivisions (,I' f)C power for 3 hours. Lmqos, of, DC, poWcrvdiqvibles all 4 160 VAC switchgeat. ID~s maliy start andle rup,~ butcf~ b londed onto. hus~cs. resulting itn a stltion blackout. Reactiror m iint postiblc: howvcvcr,.MSIVs will Calose On tle loss -of. Dcpon w a ty A

vacSne will cycle. cvelttuallvs (leplet ing rector inventor LInIi Ifuel 61ij, un6 vered.

Scenario tI)nlin.,from II tfi Csire dlatmiage iOp I ly -tit26 ntilvs.2 b' 'f 'N' I

2 21./ Tlit~ t hro tk denitscl fit~ ~V '~nhro elx Opf t ha thcit

l. ow~i't~i'

'eitit Io&.cp thatl the H.: I, Loys o h ot utsie poY Cir wii h -NI IR fajrc to d ose mid tiIllo ut c c tic ri ,C'clH. l)G t'iaq ~qt jsul sitirt ap&d Ioad, 4ind Core injection is ah1ioncki Ih iIiIhe (1 NsIi i iips d i:Pre6stre i's reduce 1) w Fh. I: Ni wC of \\0hich fails to C.Lo~. to~sry~prne it) Ildpnet ailh CkDRt injectli~n is, unable tot0~~o 1, thle IIR V andt(ti 11IVitwovry resulI S 'I Ilk, stItIiiii j IaI l)t'tivj(ICS.111 CXCelICletr1 [reseflt.0tIf O( ilti ith rCspc~t to the 1'r4)n-e(id anl'I..livi leI' of, detail is iflcn ilttIy dti(ii contributors. IL. I.27 [ron t ind Itnd lBuck-Ed ntefie I'lle Cie 'C /I.cvc In2tterfaIce, was ;ccomphl ichd through ii. Ct of Plant111 Dtliavae St ates WIWI. Puk-mo lc\\ l -lc weci (Isd ill place ot evcnt Irec to deifi~,tea~l. possible PIDSs. These modt~ules iii wIii poss Whe Le.vel I /Lcvel 2 interfac6. ~tich ats:0AIA nmitI cooling, andi toruis venting~. Vlw P'DS (j(Ianilfcatioll was performed using (h.RSAAN44~smlrticth proccss followved 1( patfy thle c wc damlagc fricquency. h PSscre-ehidt-dritrf vUsed a trunfcat ion vlco I 0. ;11t htupf1i.f SC(jtICflCCS. wcrc combined flor simrp]i fjcmtI~ th truncation limit bccane izcro. LImvlcsive coimkining of, PDSs was, phrforpdWI keep 6,be2urnbor of distinct stacs elo thle c~ lmi o 50 Tisws erore b cn Whattn etorowfrequency CfldstattCs into Cmoiiitllevt ciidslateS. No l'uriher inrormtld l"I prwted4m -4i ow, the comsolidiitlon was peirliorimed. For example, tlie cutoff limnit f6,r.qequIcfl, uantifIcation was fomnid ;,s a1 ftttiihot e it) a 1i.1v p ti iIta of seqltIcetts,

IIlt nafew l W et CRfli. I tOM havn b~titf-l rtoi h ( Iii q' tarli t cet wei';'! tots( sc~ SrvL~fle Olut. ft j1O...tIe th;at c^i)ii',Oleljtiatiil' !. Id \\e U imsed No I ic s.tltiCents st ineoc¢rc(ly Cltgrii(e(l all( it hihis dc s no()t

Ippear to he the case.

TIhe S'creening crtierfr. sii J& lSltnt with NURE i-1 3 1i I hLI kl v:i.s tor gr (MIpIng i'p C i not:, pro)Vided is'l1w I PE' This slou ld he provided )IM, htt mini Pllint I xitnatw St'tes explicitly c-on!idMtepd aprt tnr reactotl Caid iotamlin'ill Th stt On. '[he IxvCF I kqcqltlccs contain..al j ary information t'oLt the P)S analIsis, therefore no s .tyestmA "We added,ir.PDr modies. th Iw 11f.t does not aiddress the timling' ethe fJhn i fa ilure wNith respeCt to (ic re t1 Cr corc tuooh Cn ipnilnnt 11y hitgh lnrattlic.. In othel words; th lie 'l. n1 e' nii aiddres.s.. uponl l oss I). (Ct CoO linI g tI tht-Coflt111111110nt aind 101t1oru. WichI' IfAilureI t (:curs ltinst comra ltiment (hu ctlo CrptMr~ lr Core cooling equ i pment du(. to h te nit pcriit lue R osh orl 1'.d(ifLquutC NPS IllidS important piysical Iinsight should be adIICSCI

1. 1.2.8 Nlt11ti-Untit (.7olm~idertirjlons
f Ov te r ('rcek Nuticlea- (;cnerat in! PStation IS sings i il facIli-el'iI-Unit considerations atc not 11.1.3 Rl(view or ftie 1'F, Quantitative Proce6ssl
f The O(: NIG IPIE used thc RISKMANcodc packag ferscequc Uanification and plant d amgc staie ;inalvsis. A dcSleription of how the codc produced qutlitative and quantitativc results is not provided. D)ciailS are provided in Section 5.4 of the PRA wiN broad refcrcnces to the integrated uIantilicalion proecss.

Informalttivc details pertalining to t1ffquantification. process should he provi(ded. A truncation Iimit of IE-13 k.reported I rth mtablc of top 10( scqucnccs. Appendix C of the L evel I. [IRA. No trlincciofl dlimits we portcd in thc systems analysis. A rntricauion 1init of 5 F-l0() was usjed in the plnnt dinmage tAt4 screcning Analysis.

^

3"5' S

II,,, ~ 011;h rrll:i'l 9S;(1ht 'rŽ 1i1lWIeUn o 'il~ ll l't I~ 'i~I nil n' c\\ nI3/4 ~l~np icIU ailure data. intl iuuman), vcrrs~e~tnipttotemdlNo mod)I. (l'I The ovt'r II11J aproachIt cr~ift Awdl (1t~edpc and 110.ddenlceklie It r~jcIl4kv 1ii 111h1 oi'~ \\%Tler plotted lbr 104 I)~(lfh& fcluct.c and the t()I six plantl I t: 1' 0 I C I k k c I, '1hw ptilm 6-tilintac f~ h !Ft~(c~ /v. Tlhe C.I )F (hu"It btmitn' I I v S ?I N lc~~l'?l.6 (taken Irliol ,it [it U'C VI~ ()f.I, RA) I IcI' ( ( t 100 mlIf dnnaiill sAqnnce ale r~v(widcd i Afp tI~d d i X of the P~R&\\ hI'iv (111aIII ih'icatmol process is: ValidI. "l.o~1 I(1 15proce(du re arv scaN1ltefrvd I iii'oiighouti thel PRiA, It Nv'otild be CoNfilt fteedCaII jCc'ncenttrmtcd itito.a shigle I. 1. 3. I Qiiliiit ife itII ion1 of, die Im1pact Orn11 ntgs~c C yitcms vin Compoirwiit F~ailutres Sp~lt trlict!Il'n M'R(lianti e separately. in he. CDP 6akt1Aation., i36calse comotlpleni-te. ci intormtatiot) is lon'Iu inltih process. thCe cliintificatign of CDP'- i fullly integrated only by uw, oif Ilhe A sei~iivit St~lywas perlformed on scvcra.kyvrabei heud:lsotoficpwr r'ecoVery. FN lll\\\\V taz lutes to close, anfd recovery of contaiinm erntbc removal (including recovery' of N' ) owcr atn confammttelnt spraty). An analy~is was miadp. rergarding Ithe sens-itivity to datai \\tn k\\.3-WOOli 1 11) ion5. 1T1' anallse cokelude 1hI-t ehltnge to, d1iato leun~ iegaic 26 4-.

tc I. l c I I I I IIt I IICt I Cret t1tn01I tsi tf.0 Ct (&i C) ii4) vC( ;'I s Ig III ICitai c II cct t il I I II Wi I r~)I i ' hr lns. I,,c no1 t rue for Ow heasoz~,, the l7mkw '4rw Attrdv Conctl(idvs dima relax~ingi filvom'piions Im i l(r th lNIR~Vs would samo 1w.ht l( ~llq VoflcIl~itsPmu fiIo mu tthta10%/r inceromws in.L fnlt t' trhlIfsprnd Itc 5%1 incvrease inl ttal [1..1.11 ul Tree Com11ponient iFailurt1 Data hI Ivt unii. fit, C( )( NG P~R A (f alsel`~ %WLI (VIpd"U.ing II BavciflYSdl Update proLcvSNo t il it I c'..1I I 1ula;1 I \\c c jC ifi. rom Il iarl 0~uai n of II I IL' iI ' J l tN W \\iI Il I I kI) tI hrII N II Ii I it "pcci tie dilbaw fihiat thdjut rpent ~ Of~tna.Over I O ) c'ars of prljliI

111Iuiii~ucr I I

.\\ C )CrI('flem Iat Oyste r C. reok rht CuIWn 601ion (I iSC LI S th s I IIIuItI> 1I.111Wo 1:11u1t I c oip)el aticdia eg~d sions a re presemicd for pflamt-Nreciltinc nC-!1C-. andl( co~~rnmon cit s eC cipnntfiur4t III'Ial-S(C ific ~ I atI In-pcf~cd air aa ucs aa 'The IPI I for ()C( iS miade extensivC tiue ofor& ) I~O'aiIIIr lteLCCS ai ,tn~ ~:/t'tI ~unaa ah i is ot pnnt~eonennde d by NilRERG-1I33$ Iv:

iaIs s

iiplnt*pc iti ata ncue;crcrclcyrtec Coolinrg p)unmps. bat Icries d rc' I )2crur:t

r.

et: I)W hSvork, and hrN cr All of ~thcse, ~oneonts. cxccpr electrical) huswr \\kfk / anIv c Inl :Iddtt ion, scrvijce wate~r pulmps, f~rmtl~r rmr Coldfi et i solat tonl A I)S vafl \\'C, anrid other compf~t wc~cuific r i ~daa Inporients wlrn aspccificdad ('C I ICi Ic a tI All i.~cnencw data is listed In J'swblc 43-8 of 'Ih evl I PRA No description is providedl lregarding (tl:, generic (lalta used, Erroo-fivictors 'ar o presented, cither.. T[he onlyV inrormaiiion providlcd is aI tefcrenicc to PLG's dAtatnsi for nuclear-,reactor PRAs. Since I

i'itv%

5(1111ipoilmit til hiirle r~ltte. itit Iiied phtAtAp40 daii tis I&'uI ev~

~Z% UkIilqjtitit oil foilti iv' 1.ita1 "ithbolt I dvIescrlptloi or, thie (istI.MWI~tllie%6.- '~,ntimn ( a vc I );ia I.. I hr ;ij"Prkn 4 i W

'.II rL-dI (111-ltwkt depi p d'n Ien I6illtte-the clnol iiifp c t:ck Ivtei I ii 11 i1k dc%

",rc'

11nl Ilett hc III i(

I ihL OGC'1 F. rwItkj I.ivtili ulpoul IcICICI)( cd I~lixct ix. I 'lo clII n'IV it t it1w' ilitc. phnt spI it iN I baku~ o 'il lsc 0 (It'ii t me I rit w I' *1~ i iI01wt L

tcf,

('nd bI 1w plan V-i~ o I doejinWthainct itdutvi r\\ wriC i~tlovi(I ! ( "tr I I

t.

I, 1 sv Ic tst:) 0410 c hc I di %idmil f:~iL~ ~ l 0odificat mit ni to 1110 1n11(,\\'tI11na (cVc'11~

II' j, !

tn t K flL Iw i.snmmarm I eIttmI ~h% L I ii II paar Ime v v~ Ic\\i,i -i ,rtI, I i Ir rm i vte I emII tR ti) I I.. 'MeArA o( )t vtld imSiII OT'ummll t.sll-i h

nit Ilia ibft niw 1"'\\\\rI I (I oystr ('rId)

A.&~n$ft (n41I'iep mmv S~a II (ICl(ICH ,Ir *'111at01 "ist~. t, i l (ai .. was t he I I G, penfcric ccflhfT1fledi0 (1ittbiso'".T Ia Se cminid o cafI p lit: ~\\itathihiba bL' II bit II ',cd. 6r mc n trined IL. 1.4 itv itw oAilii' II.: Apfprowach to R~Ci im hCL 11141 Ntctihodology fiwt Idenrti'icaZionf o.r. PuInt r 1 1)i IraIit ies Hi. CitI iCCii' i 'tl ei ead

g. ou r rcvc' CUCI PL ',S methodoiopy to Iamdelli lv planII\\'t I Iem[inemhi I Iiv1.

v Thle OCAN(3S; I PE ptcscnts ~ iar 4~cflflitijW6l of a vuinerahi itv ais ",inv come (furnail't sel I(me ht excceds I Ii,4,per reIasqelc cptc year, or any ~n aihmc~it bypa ss sequence thint exlcedsk I ii-6 per vyear I. No ~~en~~~c eefudAcrigt theAIPE submnittaI's (IL'l1ifIioni. Thciii' er accept thp conflCu I~l adin ovd~nb~ exist at. Oyster Creek. Itis PossibleV. tising~ th(e OCN(;S criteria, to bneacmoetContribute to.99 or thle (1)1F. Lilli vei lit) vtiIlierahilities would be Identfiecd by thrirnum ricaI criteria.

I I I I.I.. ~ I ' h n t 1 n p r o i t i u i s ~ t n P h N n O e M o di<l e uto n. % I ) v e n j t w. n I o f ion c n c r ~ n e p ~ IVr 'o o f po l nI)( n w I vaerg(h p reIncyb ON qr si i 4 ptir r IleeP wktor o fII'ld~l t a n n a n om t iren6 ". co r ~ osp a s \\'f te I '. ( 1 i can e d to ntai rec~~ ~ ~ d ~ n o r th s c ois, p a v Sytm n n p M ~o(dif'ion ns (fit r 15 ) toRpkien he k a'iOr O vert'ill Prowteci nn \\ 1 m I n ' I IO;' th e Iee41 'a o prae t '4 s ~ f c ~ i a e. train min . and p o,.c dt11(m t~ wi m orin overflhl tioingI t[j Inc r' aet C1ii'ophisi in'6rt n inirg bn.Kgy'opoerator aet I'los iI" d~efIined~ liv te 11II-I h eso 1(1W% Cost ii di la(i n ho Od ~ Ppj~6l~dton of W EI ' intwigh ts.. [ H i i mpact i t hest ph im cd miodlti k-cd iwn w ere tn t Iif&Jr c ht(fIlhe St101Io g Idle -ndclik mc s ofi) I OltcCIII mcP wv r, the ptircha:se of a portalble, I)C' ~g(itm ff3 4tove) should he t' vid I ted, h' P r a tA cse credit f or the 116, iln -pIl thtnd m o iftd ns (im le en te after 4R ): t se~ of the~ UOrkcd Riets te Qraftpntatc Ac powve r. 2, A hard-piped o ai m p v nt sy t )

3. I'ovisionis for an a I ninaIy iita4 ~containmcnt spray system.

11.1 I I Pt Iiriion 14diahilit:' of' 1)1i i ~& n h vlIT '/ AS kx:phtnt divingea qefC wiih I'l I or the W(C4)taI H h: Ilaec ~rrdt'.flic~ Cont 'ji ~ tA6l 9it (%I: t 1W oa cluat cr I K eIia a~c Ol.(Cr is~ va~i I1lW.OfAU' t cclare thc ksc lose I i ia a i ra cii r th~IIhe v i. ie ifrIkttCi1~II d wudolcv is niot consis ei~t ll ~Ai ~I I'l defIIinitio nr tit' M)!IR. Iv foiY h tlt) nCn~tifCl ea eaia.mm atI I C I~m~' a t ilt DIM r a~atv If I.I.2 11KI1IaUM(Itr ()IkeMIR \\Idl or ir1I1 t11 ~1;m1 aflC'11 Sfa(n\\/l mer)cfl(V Servtec~c

1e f1M\\

r va~lalel 1 I. I kic C an I I n ineaIaI Ie\\t C y.t in )tca tie ikaflIMd fiv outip a MR he ~t lI tI ~Cl ('crcii~d~i.d ii he i~~. ilng it ~4 spport require fr o)peration1. II IC i.te a CC I I I itIIn pt IInI I seal C il Inres, 1ah a gcneral LrantiS~vt hns riot becn adequatc Iv

aa dIr s'wd Jll aril, I iaafleets hie tv'%ottioii of DIl4R,1SsueS.

The (")vSIeI ("rek ta" tat v has s(vra niue fau re w ItIa r~ct )1

itliIt, Il no, hi~i pi es%ure I ( (

,y~en Ind ni, I~pesl ~lu Ivit~ .PCl) I) 11( WJIR &KLigfli ~v Orc~im'dliivel h1oIh 'tstIf l 'tf\\ nul 'I P n : Iddilionl it)L ow ji Orit spraYI System. et' k no htrt (2k sytm CoigIpoke byV S pej/01Ai he onvrc u~ I Rcdcitcdt cL(1 I lI :inile-tt VcoI11 ingu A~ikbl Ill. OVERIALLI IXAiuAIATION' 'A ND1 CONCLMU ION' haw sw~ tni~di ~Sd (I1 ii I~vc1<2 PRA forI -itrnl evn' Tllc co('mall Ihc iiri inC 1(" k,(I fa ll "(,t LI svsti fal:,inaykkl f~vi boh intlne an uppnit vsrcrusw. J li 111f~ti~inn 2v15Iiieid~ e t vrul p~ri~ an PA Tor II`R p~lats. howkc~ei I lac I rni~d. suIch asIi *nsintd in r k'-wilr tlne litwe ow utmlC a IIJ111n III 'I 1wlie IFreI uei~~ecdrvdui~~reeasonable approach. airlthup~h litc cmi ~s I ti 'vscins;iuavsisp InioflIS

lnFlatO, nfaultrccwre "W

lr wrii fo 10r most wVtetm fai IiiC l F1,1if to n w-1 c nd su1pport Sytemis were~(,dckd wiujth ythe possiblc exception of' recircnikitioni inI sIa ct I I mv oos of, I I \\ A\\( room cooling i~ fl 'i c lUdd, based on FS AR analvscs A &(onut'nat ion of' pt ant specific and genriidtaweird:`4~d to quantif'y basic c vent prov;.1! ints. ( rrnm cas fIlucwere ev alu~iled jiSing. thMultiple Crock Letter Method ;rand are propawiletd correct I'y through i le modl,; somne 'dat4~cc~nqorin~g:.OCCurredI which mua~ ki

III ~ I ki if In Ji '. T)I 11iff N Ierh'rmc vIuIIw the 1 Iu

0. 11iLthod ilj l In

.11 \\ "tll ~YriKt vvk f~ tlI ~i t6~~t~h~ li ~~cdn LlbC hn'wevr. it I-, lwn In) :".iiwiwon. it is ndr Opifliofl thdt tiit' OC"NOiS;T.E1 good Lee I PRA. wilhi minotli ke'iiilvws r to Ie IoI ~~sues rmtis~cIif thklO)6~e~ 1wiI help o0 corr-cil %vcak lie ",Ill I.shiIil2 I [Ipimlloiwhi (l(o1Wi 406k Itry Orprfloojitid it. mippoi~rI OC N(;S'~ conlulo 0. Ii v N: 'I ilt It tI w Imilijy ;ftirthlr fowilo.0,00d StiIhtstjatilc thiN k w. ;\\i 41i

-'A 2:

Sq 00.

t-4, A. .A X : . ~~~~~~~~~~~~~~~~~A, ',t:;;';;,,,1:..,.I; I f-If I UiSUNI 3 2.$-.v OY:1r PJ R~ [.k.fK 0 r491VlOUAI PLAI fAMU4AtI1ON I FUIMKllAI iV;AIUA1IJO~g+;Tpb RJ (1I3A(,K -[-NDU)-e t~~~. 'A tuS i

'. ',"tS ' '.;'

' ' "s;t"'i...

OYSTER CRfEEK: IND)IVIDUAL PLANTrEXAiNJINTJON. TF[CIIINICAL. EVAIAJATION:.~OT W 11 Amrnamoori~ Prepared for thre LI S Nuiclear Regulatory ~,Conimi~sson, U.nder Contract NRC-O5-')1I 0698'~O3 March 9,~*jI994 s C llN'r84CH,, ind ~I I 82 IPairkiawun Dri C Rock 'le, MArylad2~ L j&2 c22&.

INDI VIIUAL.IIATID~NU4TO T[K.: CI IN I C A I FE'VAL IJA T IO N "EUP S. P1 cptrc r I' Ii Amrtwr(IxdrIa-- Sofeiv Aimlvlsis G roup V TrchbruLAI RO' ov~ A puat, Mo0h ntn;idl NfixfinR"Jsc ~~ MOer: Miat ij-cr, .Acicv Analy>snk G rou p M:,naier4 laey Anli~aysis Goroup SCIENTFCH., INC Miti-ch 9, 1994

TA Itl p () O (NTZ1t$I I O \\'I ktC ([ONRRt '1~WV )N R Ic vac: and IdCI i11ficalimn of II Elo inghtk 2 I I icncr.11 Reww t 1tI Pt U ca -tnd

And, iyfical Arcl$l 2 I II ( '

I CWe~ (: I II I Isnpt o,'uifclopedend cieq I2I In -nnltd lBack-end I)endck 2 22 vqwncrwc w~ith Sign Ificath ProbaIlIty,,'. 2 2 I.nt~u ides and fTnmntng. 7 I 21 ( onI,11inn110 I ohaiton.1a I IhT '1e

12.

$~vstcmflk unan Recwsc.n~ D tominant C'ontributri t 1PC Insight 2 2 I Imnpact oil I :quwpmilent Re~av KvdIucig, Probal'nln1tv of'C(ore D amageo T~o 2 I I I I )efilliti('n of \\ ullwobTility. 2II2 Plant Imiprovcmcnis 2 Rv 1 spons t CP I IProgra IfiRco PnaI~~ 22 1IN. sti cnvt hs.Ind Wetak nessscs 22 I I IT F' Sr enIg Ihs j I (A\\I:AH.\\1I AL\\'IAT~IION 0%sict Cieck III". li c-k-Illd RO' ICX% I' g I .1

1 41 I I II S12 I 3)

cv I mm 1 /M Ill k II" I

L) ;

I.I INTH OI.'ILC HION ( )vrici ( rek I fln id u VId Plant F-Aamninatioh (IE)!3 'knd ~ubniiifal 11 'Thi's tcthlnical ciahaid ori teport compilhes %kit h the rcttrnetso b .s~ Nucicat, Regidator\\~ Commin ri it coinlrAciol r icrk mdet Allr Step I rcvie~ s, and Adtipts the NR Step I Review objct nves. \\k Im ch Inhilde IC he kloWrnn! Ic dt-ci lit tic If' the I P I submittal provides Uhe lkvely 'Cdetai requested In the (Twcdntmcc D~ocument" NIAOR:6(-]I3.1~

I isessthestrngt~ an th wekneseso'the IPIIE submittal oeaprehlmnary I~~ of td iOsiion0 abuhe WIP Suhb~llittill, baseCd (i) 1w imited Step I review~

nco-.pletche I le11 EvOkiation' IMCASumrnuarr Sheet lii ectn n 2 f h TI~R.we umnari~urfin in~an bie'fly describe the Oyster I(rc Il s111,1i1n11a ;tIis it pentainn to the work-Ircqirrnn otiedftil'heConitactior tack, or (kr I iach tin 'It I f SIOnh'cLAMii nn2 ~'orresponds.0, a spCOi j W c~ reureetInSb oe n \\i 22 ,cl'c out "It AsC "Strctit Of the Oyster (re btitl tpti and we~knecSses lIn Section '\\Nv piecs'n it oii evaluation of the Oyster Ct6k, I PrI 6vcr1I aq wclI based on thre Step i rc\\,ie\\ A Hcridcd to O ilr rcpori Is thle I PF [.Evaluat ionStn~ lhe' t, Which W.C c0111ntpLetd on tIL e( V\\I cI h.-aetc I I c-~ i, I <cIA I 4T mmc iPn IMnr

2.

( )N'Ir.A(Thi iiz"IE V EINDifJ4: 2.1 1{evie(% ;s d Identification of 1ir P t Itisbi I .i II I S I is s IfLI I I t e ii accordance Witl 1 o I 2.1.1e (neral Review of IPF Bulk-Fnd Anac1;sAI Process 2.1l.1I.1 C ( omp~letcnelss 'I ht t s(,I rectk Indjividual Plant lF arnination (IPE) B~End iubmittal is essentially complcete It. hNI)CCI o tWie level Of dcAil recqjuested.ii NJRI R 4G

Thc lPI Fsubmittal mects the NfK(

Wtstjipi e s.1 ct i:on screeinin criteria descrbcd in cieneni* Lettr :XX-20. and surnmarraes how\\ thih vx.; dim in (Ose Table S. 1, page of.thdeL 2 PRA11 2.1. 1.2 I)escriotiln,.Itfcli onlind Consi~ten I lii I1 irier In di (Ib gv (iised is clcar v de.scibed o iti i1ection iA riu tifed lhe appi, och 1(di!.,,cd v-01n1i.S"nVK WithIicne;icI.e: ter lte g-2O. pnd\\tx I

2. 1. 1.3 Prc Use

, I ~fo [It' As. iii,tr it-nSmbsection I

p.

page1 - I of' theiPtI.0subtttl rport, "Tlhe analys's of (Oster (reek coit W1 ncn1 pc llifi nuance as accominplished in he cont extending t(he Level I st udy t o x.%\\'l -', dellined In Nl JRW/CR-23() INUR The LeveI cadcl luantioicanon led to t itnni tisn i 19 planl damage SIMats ('L ) with cucncyQk 10' per year or greater lfo Tlue L.e el 2 analyvsis; these PDSs were condensed into-atet ofsevcn key plant damage states K '1 ).S li whicl containment event:it tre to(TV sRwrceeopcd Representative sequences \\v\\ ci selvctie lIr each KfIIs. MAAIA3 OIB, v :7 0t,.was'used to calculate severe accident CM cur online an1d containmient loads for each..f hc reptraiive scquences 2.1.1.4 I'eer Review 2f IPI' Independent p),er review of the Oyster Crek TPE by t~mheber Indcpendent In House Review (;roup (III:IRG) is discussed in Subse~titon D3,2 4Appendix D. of the Level 2 PRA T he HItMl(i rc\\i-e'. commlents and tiheir disposition arc dcscrib! in Subsection D)3.3.: Of the nine 11IIRG conmmients listed, only one lcd to a textual rcV sion t was concluded that the other IIIIRG comments could bc adequately addressend by res ng only to the rcviewcrs who made then) It appears that the Oyster Creck IPE dji8 recive adeuate and appropriate peer revie-N '4, ,-SS S:. I I'S. 444 '4A t Jt ' -f1 I ' -I 7., I ' 1 d 1!7 1 ;!",~

I I I k - -, ,,' =

2. i.2

( otiIainmnent A nnlysis/CharacecrizAtlon

2. 1.2. I Fromn-end, Ihtck-en(L Ijenendonc k I II: Intlr face betNCt'cn the Level I system analysjix.adie Level 2 containmment analysis i onsist ol a SCe t)f' plant datae states (P s),' asdiscussein Section. 5 of the Level 2 IPRA A PJ)S is the resuilt ofonc or morc of a number ofphysical con n.,which were analyzed in Level I ilclu(dInQ the bd1lowing. prcssurc insid6th:e rearcttd t

iime of vessel breach, prescnce oft %alcr on the drywell floor, containment psprt9.o'uaO ' tty,4tatus, availability of water to cool the co(ic dbris. suppression pooi COolift atnd cofltAliit enting Other secmndary condition. Wtre idk.o considered The Level :.I analysis of Oysteufcn k -nded. in principle, at the onset of sieuficant core damage. Which was defindd as thes i vhen. :-..the top of the active fuel is i(ovelt%'rd. and vessel water level is continuing tb dro he Loevel I event trces identified I O plant damagi: SlateS wilil a frequency o.IOE-,8 per yfaaot greater. For the purpose of' I vcl a halvsi'., these were reduced to a set of~seVeinkey pi' d'amage states, selected on the basis of the (neriI etter reporting criteria Each cfthe seven Kebss received consideration during the slubs'eqntlL't I C'cl 2 analyses It appears il-tt pr per.t~nt waas taken offiont-end to back-end (i'pnml.ncei. and. ovelallJ, the analysis.-front~cnd ,-krjd dependency is logically anld clei ~l' '2. '.2.2 .eui e alic es wit Ii sivnifihttic lro abilii 1 Accident seI(utitieces

ith a significant probability of murrcnce were evaluated as notedin the 1 evel 2 I'hk A. Sectlris 5 and 8-A number bf KPbsit a higher annual frequency rate than I 01:- Xveai rcceived further eonsideratioh, Using thWe FTntanment event tree (C'lj These K 11)S; were a factor of 10 lowerthil the NRC f*cnccfrequency criteria." and are listed I ble i (S n pIage X-5 of the lcevel Z PRA) The Tevents uJsed to fbrther analyze the KI DSI \\Ci e CelcetC(l to address it-vessel core degradation-the:potential for in-vessel reco\\er-thLe plienomena associated with cx-vessel progrsSionItainnent integrity challenges.

cortioiultsl fiilirut. it. timing, and thc;cff6, tivenessn -twr sareguard systems to mitigate ollsite e lealses

lc ('iT eS nd-states were binned logeethO, atnumber of release categories. Iecause only ona (

was actually devcloped4 ithd tO be q ied for each KPDS, and therefore the (I I bItranching prohblailities (split fraction ) vared, iiiOst caes:, for each KPDS The (CT1 utsed in the Level 2 PRA wvas developedGto-r1 eble th&NUREG. 150 (and NUREG/CR-45S I accident progreCssion cvent trees (APEiT) developed fothe Peach Bottom examination I How:ever, Iewer events were analyzed duing the Oyltrreek IPE (The examiners asked 145 questions at leach Blotiom, as compared W'ith 16 at Oyter Creek): The Oyster Creek IPI sibmittal explains th1cat answers to many odihe questions asked at Peach Bottom were "implicit in the defnittions of the Oyster Creek, plant damagc statel !and therefore not included in the Oyster ('reck (C1T. (In actuality. however, somc'tf thc qucstihs whose answers were considered implicit we-cre asked as parl ofthe Ovstcr'C:reek CIT; V r:. / } l. ph*,-.. Ihe. )oster ('reek (lI., shown in l igure ti: page 7 if the Level Z P}RA. See Subsection 2 I ; Or a further desc'ription of C LT :top vCentsW 0 - i ~ t ~. l~ ' ~i i;t c i (hslo~ I0. lcc

ak1: 1 kek Im-

-o I (')SlS~lt~:!ll 14(:k jlkl<ClEW v\\;' +XS W~ d h fRizslatchli; I 1)il

'I he 4t1ani fitet Ion of he ('FTr fIor each KPrS wiasI irr IC I uhil a numLIIIr II sf)II, (I', I tr 1I' defined fhr ea;tL1 hI[pp1iCAtIeI ( 1 top CVLfrlt Thepfk ess tarpears reasoriable. hut It 1' 1iitS ilr 1 1, ftl l(mk illch natilre of each split hiaction. the trmin gy enpl.vyed, and the split I actiwn l', itii., 'se i a Ie U()- I(if thc Ivel ': I'RA) ehe final rli j f the quantification of the (. I seIlqIcnl ci ate not Qi('1 nI 'e rLsLIt.< %veiC used o:definc the tle~sc categories and to calcuhlic Othe tI Ve1l.ncies CIreItlvTh Iicading release fract onf 4iscutsed in SuIbeCtion 2 1 0

2. 1.2.3 Failtire 1Node't and Tuin, I he )vsici ('reek containment failure qhkractertion is descrnhed in Section (i otl lth I c" eil 2 1}1R A. anl it is detailed in an I:(A Ifinginecring calcation reportl. whhich appcndCLJ d th liic 1 evl 2 P R A Piwe S or, the LQL I ngiineering reptortit rS thc f6ilowirg lii !t;'t [wtlt'(;,

t

i ilrL.t tltii

/'j' t ht't b ltusd luiluirel~ pfh iuth t\\' dwci g initi ' n'i c 1...' ts i, t '. ,i( li i '-I t u 1 al rw f:111 h Ittw~ nvu cI A ht p~ ~ v ei. h e i sh e d

p.

I f 1 t it e st A lerlt' d 1014. I it A iik-ii '`11 "w;.d Ihini AI lids I the Xcs IC iN, ibs di-1cdlit etipc 5sin ban I.:}ilti' of Ow (I'rN\\%oi'l iwa:il (Wftlp itt 4 -:VA I>. tihai~i : thti \\cit ~imef Ili II),tl? d i-X viikl0 I. O;sijp tipp %Ni nt hiatl wi~tq (,I eich oltlihe tail'td c IIl eC exanuned' the rohabtyoffailur e was calculated as a fiuncition oI Intl al pr 'ssIiCre t'it0bhin containment metal tenripcerats ranging front 30(i' 1 to 1J201) 1* A'\\ hov--normal failure model was Uced to "formi othe An O\\'erall tincertaint \\ ab i in t c imted.], ustinlu material-strcigth, and motdelijnancrtaintnes, %vlihe ra railonale for sciAL'sli) h eli ha;is ofthe overall csttmtrivte w.ka ,$ t giVt P III ['lhe contll;linnient capacities of each failure mode apcratures of 3(l)'o r and 7o(w' v ite sholk Ii in Iables t-I and t-2 of the Ievel 2 PwA The failemode with the least containment 1pt C eSIIL' li1'8 that o1 leaae Through 11hC bolted drywell:had ngc connection. 12 1 psig tt 3096 I. after 7i) iouirs I lowever, as 70 ho1irs is a rebativCly lIin-for accident progression, the subiscquenti f,,il,,,.e mode-the nmctirieranc failUIc of` ,drywlhhie (I34-psig also at _O00A+-l vas mole fica nI (i om the source tertml viewpeint). 'I he linler melt-through (a conscquence ofdirect cont6at ofeth containment shell with t' el deIris \\aS also analy'zed 'h'lle results appearlrasonable when compared with NLURFG'R-s'4.2 (Table 1 0- 1 0 ol'thc Lc'vel 2 PKA). A diffrence factor of cited in the Oyster Creek results is attribiled to a 6-inch-high and I - foot-wide cur!) o;n, thc drywedlfloor. The analysis of this difterenc fion) NI 'UlRc/CR.$1:3 the appears to have bten adequt 2.1.2.4 (ontainrienm Isolationp6 Ji Itapre ('ont1a;i1nCIen ISolat1ion failure is considered part o pf g!aht damnage state One of the prinmar\\ conditions takn; into account before binning plan, aioa.states is the "('onlainmet lhlessurc ( i isctr I i't k 11 '1 Haick-kIndtt ReCi'5 t *N t;'. It I

ktitinditdr lnnegil. S Sa i " s This condit'tonadresc(the ",tainm nt isolation failures andi ptmefltlidi nnit;aintieln! hvipcs. as well as e ~rk or ae cont 46mnt failures A s i aIt c'(d on ace I i-f the ie F' subilit ,cak$ were not. sl addi esse!d In this 1!LJ\\ Blecause the t.nistetCreek lihnt is C-ntinuoUSl onitlored for o\\vLen cinttril 1t insure inerling, plant.. operatrni

taf Two1d8,bo alerted to. such leaks and would rcspond accordingly This ix reasonable and0tonsigtoiit Idi thor hack-end asscssnients of f

IcIlitie sih ineried containnientS 2.L2.5 Si/lliiiian I es)OIse I I I K ( I 1 Insidedrs po &ssibIle methods for afresting the core tjnder the top event \\V13 of (1l:.I Lach Kil )PS i..c eentd t identifV the ones with reasonablo recly potential Ilo er enl \\.Liniher I I in the (FT (tlrnic renc ~ Crcw~Vent ntainmnent in (orc Dhamage 1 en ios M n1 eICIed tihe itntiClonal venting.OtrhiupprIes. pool air spacer SucceSs of0 thIls et ent means ti hll at lihe vCnI nlow capacity is adeluate AQd bt"4ing.t)rm containment failure would hec l irvw1kd Dt'pend i )on the KiPt), thI probblit Y Ws top event ocur'ig (proil I\\o fil.iict 1, \\c,,ii) mitcs fom 1 M1)17 to j ti aS ne\\ld htl~ the probhiliN nthat h.*.crc Vents die- ~ls trina ent does not change lot ol lpre-oi tLi dam.e o(ntins ln htle Ievel Linal yisistothopojst~O damage" venting in the Level ana ,Indition to wtwCel (tonlS).vcnt;ing.:enting

i1e drywel is available and is pio.edcii.1acti ated. for the reresentativrte sn I0ppS NJAU (note Section 1 () 1 on in rdi nc10d ic 2s)

Abunl, ha,1ll o, (Ile ("DI": is allocaled to 'No 'es4sMl Breachi rsulting in no radtonuclidc Ieleases \\Vcsel hIi u(a(hi (aIter rcoe danmage) is prvciinted by either it ducing firc protection water \\k hcn et vesscl is undt l low prssi re or prviding surtictent %c6trol rod drive hydraulic system f1lov wheIIn IhC VsCSCl Is Linder high pressure. For both Vesel injection modes, operator action is ftiiet cd ( oi c paue 1 0-2 ol' the I.evel 2 PRAAi IThus iu a pt ars that this operator-controlled oli II Lirimici tio has a ket bea rilip on the radi6logical release profile for Oyster Creck. 2.1.2.0 Radionliclide Reflea e Ch(iracterioat0n 'I'lrk radiontclide release clhatacterization is described in'Sectlon 1;1 of the Level 2 portion of the sumniital Release categories (qualitative descriptions of th`'ntainment 'event tree end-state ilrisj and associated source terms (riantitative cescriptions;6thc CET end-statc bins, including release tuinirg and rclease Cractions) are Generated As'an aid in defining the release categorics. a source tcirm event tree (STIT) is used, as statetl dnthe subr The purpose of such a trcc is to del'ine the different rleasewcategoris.for.WhJt,th' thesoUrmrcharacteristics could be suifflicicntly diflerent to warrant a separate sourzc. termdefdin n " _y , X, -0. A. t i' A .,,.li.' <, 'f S Ih 15l [.1 is shomni in lzigure 11 1; The characteriztion of the radionuclide releases firom the iontaninent ate It hitnction of tic seven to:p cevents iI t cT ST A seven-character eCd-state v 1,1 ' l k O I M Al ' A;,;,

M , r %., idenrt fier ws deN cloped, as describccl on iage%12 and i 1.1 [he se-nS 1I1. T t(i;' p e.*trh V k.tudi t, cither a plani danage state group or tj tatus of crtain ('1. 1p eVentr 'I h rnlcl i (t hillmlinm' (CItA' Scqjucntces to rclease! categ6driesrfthed in Tables I I - I and I I-' Itc thCeIII A the Itecase catl orv biinning are sho.3n in.Tahbb 3 fi:.ach relcase caleyor-kkas asslned to an rnvelOpnit release catcgory "accordingr to e this f fcn nservative condensation " 1omm this. Si\\ k) treleawe categories (K R(c) were,einerated, t nIptiiled and conservative chafacteriwation ofthle feleases to the cnvirmnment Th source trs fr these KRCs WerM calculated Iby SelCCrInL rep;.>cnaiihvc sequences and using.AAP to motel the behavior and release of 12 radionuclide y'f ips as lisled on pave I 1-5 of the eve l 2 PR The analyss iprocess andi assumptions are ilcscried lVr each KRC starting op pate i1-6 lTh:ftble 1 *42 page I- -1l the fr(e!len(A (of C. I KR( is lisidtcd, wvhile in 'ahle 1I -5, page 7-:A, ,5ot term information (release iractiotNs and' releart imics) for cach.KK( is proqved: in~ ddxable1 l-6, page II - iX suininiarizes the htr onh oflo sl within the vesscl tid contait ~it the nd (of the NIAAI' runs I he (diss rinlinatlor interrogatories Ai(Ito.defT thras categories were the fillo-inu Reactor coolant system. pressre M'essd hreach' Dlryv,.ell sprays a,41010e'

'oie dCirnage arrest6d in,-vesd; TIinrle "f contl ainruent f~aildrc-.-

.Suec of containment failure 2 mm( ontainnieni bypassd' :, S ulppressionI pool s I bhiflg prior-{ tOntainnient failure') Acc rdIInt-rt Iiga: ion ii reactOr bijtitg (euer lctter 8-2( 1 stat el that the 15(lloswing shud be repo'rted "an! ulbncr tional sequenice li,.K h.ls a core darna e fre(qluencv greater thani lx, IO' reactor year and that leads to c nitainment ailmIre M hCilih can result in a radioactive rcleasc m ~nitude greater than or equal lo BWkR- ' it l\\VR-.J release catevories of WASik V.400" The Oyster (reek IPF suhmittal mrceis ihis reporlu! rctqiremiient See Tabl 9-Ipage o-51or 's1 r lie radionliclide characterization is well dcVeIoo lnd portraed in the submittal It appears to le re(rsonrahlc assessriient of radionuclide transport dnd releasc 2.1.3 Quantitative. Core Danmagt EtiJsiate

2. 1.3.1 Severe Accidentu Proifrssion The accident progression analysis. perfored with cAAP computer code, is discussed in Setion 9) ol'the l.cvel 2 PRA (Note the lin itationg'ind assumptions used in the MIAAIP analysis.

as described in Section 4,3 of the Level¢2 PRA ) P results arc discussed and presented in flpmli s and abIles for the following KP DSs:' a Low Pre.s.ure Station Blackout with stuck-Oictn Relief' Valve (lI)LW)

I rehi Pressure Station t lacktout (Nt1.0W 0

cr (tAek 11t1 PI ck-ldlziicvlevicw X 6 X Rc I I 'NI.i; h n j

I attge 1)1A I.(X7A with No Core4 Spray O1M0 Y 'I t t X!i,e T ri, A tW'S wvith,S IC f4aiilur (i): r Ractor Wtet Cleanu) RV ( l0) Sy,$stei Fa$uc in'll Pesurc Reducing Stitt' ((J\\t 1/4-1ypas.Cgequece o s.o;S of I ei}d waIcr with Failuire o6FCranm ftzohrge a oluc (SD)\\) 1to Isolate (NiJAL; ) - 1paSequcnSe ,e Station B3lackout with. SD tailulr6r 1to~lati l-W 03ypass Sequence 4 ("iscs~ion nf' thle plenlncnolo11gcal Unccrtiainfies.0 oCccident progression rould not be 2.1.3.2 I)nminant (Contributors: ConkiwtencN with'4VE Insilhts I able I - p* vc I I 0, of tilhe Level 2. 3RA compareteul Oyster okCreekand NURF(G-1 150 leach H olton 1,, the dominant contribulors t0 theie These results and those of the lIt ipartri Ik I I. are given in Table 1 hcc <where itca e that the carly containment ililures at . )\\stetr (Cieek are les' than otic-third of those at itt atrick or l'each ]tom of( eispoliill\\. ()vstfer ( reek has a signiicantlyL higher At ofnovessel breach than does t:~, (ic t 1'e.ich Roti ton I lowevr, the OY9icr re h0rtrienmet~t is assutmed to al)'v s fai In ;pdIi hc(-; iwi M)ii At C( )VccveN is assulmed Tahle I. ( Contajitielt failurC i6 f P(lage Or ( )F; ( ompatrisor 1to Fitzpatrick 1ITiCAnd lPech Bo NUREG-150 Results 4 es, 'tbi

4trCe-l ( 'tiin" itailureI l

itzpatri. k ,<......h....cni/ l ()vstcrkIreePj .L ~ t~ .O J 1i)1

cr vearJ

-~ i0X 10::; 4 3 2xlI0' na; __o:____ l3 I.ate [at ilmit_: e2 0 ;

1. 0 264

__t2.5 ~.l O0 No Vessl Iicach O 4 04 2.1.3.3 (Ctliar;tctcrizai(in offConta nrit ierfo i ce. Thic containment perforniancc observed during the Oyster Cr;klPIEwas charactrizcd using containment event trees The top events of these tvent treesp~t discussed in Subsection 7.2 and listed in I ahk 7.1 of the l.evel 2 PRA The CEr i*hironoI6Cty sdels core degradation vessel failure, contaimnmnt behavior, and reactor building behavioir T first top event analyzed waTc one t hat w ti0d ( occur in the entry state From thc. froTthend (i h KPL)S) The next t op five CVCIents COriS!i'.Il of pheltlonIea that could occuirjirtif the btein ore daria~e began until vessel OY0 i)!l'

t'k 114 1:kF.l 1lc.:<fft

'^0 .:¢ tKaslil~lx

UrvaO, ITOl,

L. I

lb *~'lrr~n~nnl t hee e r.nt v~rc ea~.l b~ch64ft~ aI~e failure. contr nrlcnit 1o01rn1111ate ; itorw-tcrmI containmenft rep.s4tO preve$Ibontainnient failure 1w CktabliSh1nl a~lipat e lm~ i bed coolring and to itrefOi.o 'lcOitaininent heat, T~hese events \\,,crc (Ncirltr truc o)f containmeirnt venting, incidert 6fcofttal nremarninF intit late, an e tinment leak areas 'e Aq'it to to n the O or Cree i. I : It cIrilk ed o s-tLudy phenhomena th1c lhe reactor building indegrtrN and t he. I'

,. v N'

I hc hLIIr I nIwr to reduce an 94ts ou'rte tsifice and Iculd filed As shIaow in hen I uMmryi Table 2. i helc I contain fhe top en-dessed f ontanment h Ir;;'. ; III at r;nrVIIt loadinp scS aktilated LIng t itht MAAP comptite Code 2.1I.3.4 fItjI t no C ,Iit!dftBeair atl.\\ t d: ~ st X ' btc ;nvpin't and loer acidhrrtesi on equipriwhtaimn hc at.io could be nts \\^ A. lin worlo cc 1 .th apw I ffl< nt enljn ilcci&*lt~ 0 of Cilpt1mn renInl Iniac aca 2.1.4 lIrdticiri2 lProlbalility of CprileIanal.rh'inPodc ce~ 2....4. I I Iefmiltrt nfi nlnerab ilitV AS notle in Subsection 3.2, pa e 3-2, o( IPP, i(S)Uhittil pril, hA vulnerhil I\\l s fefneci it i I core damrage sequence l)Jtrthat cedue l~.qr rcto? an or any contAinment h, pass seL ILiII it~ larpe early containment failure sequencethat exceodx - pce;rractor year (d t fr';in.i f. .vtilnewhiliieic f(tr the Oister i ren k nuJceAtl pwier pan W'ih respect to plant imiptovenicnis, Su)cfttltS! Su

ntIRInd, paM o thh submittal repoo c\\plarmed nir 'Because oft Ie relativel, low frequencies ASS66atld Wit. the various containiment t>ur.

Modes, no specilic hardwate modifications &r chan ge t64kilting procedures beyond those 2.1k. ti -'1~ levef. PRAlii~el will~s'~ he us: as a ma idenfi led in the level I ana lykis are plAnn'ed at hl tIm Th ee RAwl eueda ao input to thre development of aeccdcnt matnageme)nt guidefinc Althouph no hack *cnd mromets are, Olanned ~sc~h um a doe address., front-end 1111tl~llf !1X*i1pro t svcnit ctdrlcri q:U miiv~atta~l O"Zll*, M i;t,{,:i itnprmxmnent s* \\vhich might affect niitigation "of hecnqences of hack-end events Note Se-ction 2 I 1 below prlant Imi ,cuuett planined as part at h h d ,VtItI Sstelin [his s n is Walied A V80l4b inhe A tnc important lPt hack -end

.7

2. 1.\\

Ivl CIiit, ork II litrlD I~,r inrstekr itrta t.I I iil Nt'iCllesqnt~a xe-2104e-eeo ver " ',1'( tMrar~ rln

Tabe hl 2.-(),itvr CretAk CE o 't~DsrPfions I I ('r [ ~, f I ) c mvI a P) fl it I~c%.~i~f n (1. 1~ Flff State IS

UNr, K

Os ~ar 9krty i1 ~SitcksOpen Prior to V'cse fHresach Smn 1i ak Pr~ t: tot 0Iao VV->. J~, iiContainrnent Irtc rirtje nlBech K LSmall Leak Atc r (a 'nortg*ir iien 1 g to Top~ I.Pvent JJ Buit In dd t IrJ4t C'e iie~ (Containment ItAct LaWc ee Bec

I2Small Itxeak Area~ ifc~nAh~t FA6ik n TopiEentt 1 No Hydroynilc l~lAucr1niiinPrdctninoth ca 5,-4 pRecleor Buildi Not 8

yca ed at asswssmcnt. deimnp "dirtv Ve nting,' that Is, vcd~alter~aAeThis i addr s5cd under CFA: 'I1o I-x[ent II., dkSCLIed on paq 7.7 n ~~ I O2ihc Level 2 PRA F~or the KNc Plan:t uamage State (KPDS) O.IAU (notjert oana hm~t'Mtinx the TIIEP. Appendix), :1w vent rlp ik Judged to lie e~fhcivc (9I114 of th ietesN vaUe gi~ed frert the Level I "clean" Clly~ 1 ,4 'l i:~ II Umm.

M, 2.I.5 ReiponseA to (CT Progi w q (,enetin L.etlet "XX20,- 2 iu0 pJkment N8 o; I,:lio t be.bing recommendations of the (onaintlen:t Performanc Improventi Pfrtain inig to the Mark I containmn nl Alternate watcr suppty fir: injection leartr pressurc vcsoel t4iprcssuri.B..nsystem relial)ilitv enhancement b1 mneijencS pr(cedurcs and traingn K It ectv.rn 4 11 the sullimittal report, theC. rcc4Cn 4tiOns arc addressed An alternate water supply tbr ii in placc and credit is taken for it in the IPi;2t SpeciflcAlly it :s CWU i0 that injection is timely enough and of .ufiiclent quan tity toprvcnt. feaslch SThis is addrcssed in top event VB in the CV tlAn altcrnatc water Mipplyf;r ill, spray uas considered to not he .t ,:tCsefetivC, althou)~gh here *re %iiions where, if water was prhwdcd to the cirv, nonncoolable core &dehr,7i14 it WO rnidte the conseqtuences of the accident Re¢tctt prcrekre c 0 ystem rcJiahtlitv enhancement is acco 11iibs 0h1dyb provyidht n ¶itce connection scheduled f r the I 4t reticldinP outage TI9 OKII r i lkelihood fff an ctecndcd sta:'.:II blacko oti thereby imp0oh`,d epr l tcliability (Ip i has implemented thc iIW RvDn 4 Si( and thev are rcflected in the 11l 2,2 lf[I: Mregathis and Wgilkncrsft 2.2.1 In: rSrengl: I I hI II: suimittal appears to hwii. AddricsCmt ofthe important and relevant ptittionia In en jfiienf idntail t44s1 ofteing se vvc accident hazards. such as liner nmelt.throuih, are *ysiteaticai;:;ddressed IThe resuljts of the l~l~ at Oy£ct~ C^reek atc emiparcd in surncir' detail with the NI JRi( i-I I ( results at Peach IBottom l ndW difrcne% are well documented 3 1-The back-end analysis is robVst, fi.C iM w,in ed in a *ay lhat protects the results froni the 6.impact of chAngie t hy u ie h result ofa frorit-cnd analysis) A frontrend analyfis showlid noi sgri~fie ntly*bc nditional calculations of the hIack -cnd r &.41, i Is

2.2.2 lIl; X eakncircs l:.; : A dicu~ston of ihc impact of cverac ddent cquprnntt chavior could not he loctted in the suhmittal 2 Although the probabilities of s"queiicc oecur ieet reasonable throughotit the submittal, the sources of such cstimates'Are oft lcd. For cxample CET split fractions probahiltticK arc not wclJ de:in:d I TheC Oyster Crcek IP£ seems to fou o he t!tsian containment characteristics that lihave carlv, detrimcntal cffccts orn helth; And, 6, carly rcleases to the environment

.argc, early rlclasos are impornt.u ctidt 4

nliy. :Snmal1 fraction of the probable .i'ldent events that ghould be considrl d nd. tub qticnt containmcent responses that m ve tip a back-end assessment W in itrftcts And consequences is v,, t )Vts ee.lI ~ c.1 ml Me',' if 0 lf1 I -it~c I

t

' :t 0' ; D",N': 'S,'^'n i .,0 'f i;

I

,'s-,.,ii .li~ W k kv,, 5'

3, OVIKRA I. fA'A IUATIO As. isci-is-.ed Iin Sect ion 2, thtA IPF.1 subittLPtad% 8 n Oun back-end informnation. Which (ont riblif',, to the resoltjtion of seeeacdrtvusblt ~c at Oi~igter Creek A lai~ ciT(tof the haick-erId por-tiOns ofteIT 46mitirJA wel WIten and, directed, to addtcssinlp (,l'nctic Llettf~.( i,~uos 1The jiue-M 1 olnrfhhough is addressed wel'1 conurew te ith i~ an attractiv". cresign fcaturecA~ prnwcat t6ii VF repoii Ini oininian.- oor c)ncernA ~about the 4ubW61tat '11'W Ihnthdlgv ap cars to gerlto'Olpding conse~quence assessment and L~l muI )aI I to) safeiv d o~aAOtat~ a~idrstanding theivulnrerabilitics of tht, containment and teIrnat thm cqR1et OPs. and phenomenolop ha~v onl co nta6 rimenl performaricc I'licer appears to 1w no dirtan f~k~t neinti&s This COUpled %kith a lik f enitvivank~s caen he ov)rnOtconcu nsadCould give the `ikr(1frv imnprcsi' n of the slate of i"knowkidse of coniainment vuln,.irahiu'fies A htpgc frat lion (i h oedr~ fcu~c eut n nil vessel breac.h. that Is i1 ~ itci i k a rresteI d in-vowte There apppars'to~'6~no d Is c tIs NI IIIor of ItIf f nc ii a i i ii l~ in thii conclui,,ofl~forfthe.senii'ti-vity-of this reotlt to rmajor Hit quIlnt ificaution ofthe(Fctsdtnbetrd ,:A' I l~'5l! tu~~k ii k i Ii~c' ~ ~ 12 ~~vi~qit/N~Ir~l J. 1,4 1 1~ Il!,I evo

4. I (ericral Public I tlidtis Corporati& 'Qyerck l ividual Plant Lxamination Repo Atqgws l 19 2 2 (iI'( i,Nuclear (tcorporations Qand PLO tn kO9 qrcek Probabiistic Risk Assssment (I 've'l!J I,VOk I through 6. November A 8Aeiwican Nticlear Society arid Intduute o.f ERJt caI and Electronics Enginceis.'1KA l'(occdtijrs (ildide A (Juide to ihcPerfbm fProbabilisic Rick Assessments fot Nuclcar Powe~fr P'lants" ":pre:pared 1the U S4clen Regulatory Commission. K il~l(( KR-20 0 \\'ols I and 24rJanuy 19) j' 'S 77.,k'\\< '.? '

~

"i 4 AS tnffTc 777' 77777 O I0~ e ,tc

lI.~w'a z;VI4%

It......... t'inI N~ tI*

f.

g;v

00 0Xi 'fi4, rt,"

,;4

lilt: FAllJ Fl)§A'sD ~~M R-$lETS iPNN'R H Hck-crid Faict, Plaril Name Oyster (reek ( aitt nriict Type \\Itirk I 1cselce ofa (itrV\\cll floorconcrete6urba thI he erin th sandbed region. 2S p~ter nt Increased structdural capability -bf tor.pics.tfa backlit pcrfonmed, and WI nctreased containimenitcooling Waa~illy Skas W-1 m the NPSHI limits with 'I a it) (i drvwell presm'r. None t I trck I Niinumer or Plant Daage Slates: 1tlirit.e1 ('ColiiiieM Failure Pressure Additional RadimOiuide Trnpr n ~tnk titurc Suppression Pool scrubbins asCurmeld Howe rrtb. 44 Contamc inent failure modes appear to preclude the possibility of containment fajluitsubi Reactor Building in ligation does l ot appear to be creditieDsin tN Comiifioinal Probability That Thc Con;ain id No Is; fi d A value couild not be found in the IPE strmtabi tj umed to be very lowv because Of thl Charact1ristil of the Mark, I inerte ontantr wc ~ik t1,it t Filnd R ' ler k lv Icv si n I / arch I), I,, i0S

f i; E;VALUAATION ANDI)AT)A S.t.ARV. SI* FTT Ii)portanlitI hnsighis, Includiii Uniqlue Safety Fcaturcsm!' Piesence of a (drwvell floor concrete rurb! t thiinning Vth liner.in the sandbed region Ilt! Arf lic.' release occurs 2 hors t rclcasc is causcd by a l'vp.rs scenlrio, 2> percent incrcased structural caphity of tors as a result of a backfit p1r InMrmc(d and an increased coniainmncnt'c"Oolzng bio t.' i at a result of improving the N4 It nillIS %m witII a rise in drywcll prcssu:e III pt Ce rf ened I i.t ! iprovee i Nonu ibplemntued, huIt thccontainmn6i0t codlin6 ca Iycascd as a remslt of I)mplovir !he NPSI I llirllitsv with a ris iryw'cltpr Simi pliled O(yster Creek Cli~rx ; Kev Plaknt I)arniap I requency Ldar ly Palre.1 B La re No Vessel S tat per H eart B__ I reach I I\\ I 1131-o 0001. N I FW I I~~b( 3 69 74t 7I:.-7 00 80 81I ,N .KCI _. 7:k-7 L, 0 1 r 0 0 M.IAI.!_ S I 6 3-0)w ~ I t (JA ) 6 R 4 I -7

0)

....... ot00 @O.! 0.8 0 0,..... J j.*-. NJMIIV I 0 I 0 0 O0 I! A 2 RL i~r iiI fNMal ch 9. I 994 I )'le: It' iek III, iHt. i-L kL'IvI e'10 .4

OYSI [NCL68U$~r ((R:CRCEK I NO I VjUAL PLIANT-lom1NAION, TECHNICAL. EVALOAT ION tPR (HUMAN RELIABIlIT V NL~S 4 444X I

TEC}INICAL EVALUATION REPORT OYSTER CREEK NUCLEAR GENERAT IG STATION INDIVIDUAL PLANT.EXA=MIEION: ASSESSMENT OF HUMAN RELIABILTANALYSIS DOCUMENT-ONLtY:K-"', M.G0. f-eck P. M. Iaas [~~repared ft'or' ,,> : "' U.S. Nuclear Regulatory Cominsiion:,: Office of Nuclear Regulatory RetSAch. Division of Safety Issue Reslutio: Draft, Novecr 1992nb Final, Fcbruary, 1994 1 1-4 CONCORD ASSOCIATES.1C, Systems Performaonce E rg i 725 Pcilissippi Farkway Knoxville TN` .,,37,932; Contract No.W;NRC-.91-069 9.> Task Ordoro.S ) D.......,tS -f..... I " 'I f 4 Li &(,,'I I'D - I i

TABLE OP CNTTS

t. INTRODIUCTION 1.1 Stcp 1 1IRA Rcvicw Appmac.h...........

1,2 Oystcr Creck IPE IRA ApprFacl.

2. CONT'RAcroR REVIEW FINDINGS,`,

2,1 Work Requirement 1 1

2. I.1 W R I.I.I :
*: ,.,,,:^t^ 4

> -? 2.1.2 WVR I 12 k***4 2.1.3 W R 1.1.3.., 2;1.4 WR I 1. I:4....... ,.,.0

i..

2.1.5 W\\R I.L.5.. 2.2 Work Requirement 1.2.. 2.2.2 WR 1.2.2 ; 4.... 2.3 Work Requirement.] 3 2.3.2 WR 1.3.2. 2, 2.3.3 W R 1.3.3 2.3.4 WR 1.3.4 2.4 Work Requirement I1.4 A 2.4.1 WR 14.1.1 2.4.2 WR 1.4.2....... 2.5 Work Rcqtuircmcnt 2.0

3. OVFwRALL EVALUATION AND CONCLUJSIONS 4w..,.

4. IPT IEVALUATION AND DATA SUMMA8Y $SI1f I'P DATA

SUMMARY

SHEETS (I-(JNIAN RIELIA 11r1Y) .i v-i4 i "0 /', C, REFI~R ENCES.4* j Y'4 - ~ ~~ ~ ~ 0 'a ?--',;; X ~.

w,,............

1 4 4 4 10 I I I I 12 12 12 12 13 13 1 3 1 3 14 1 5 I S I 5 16 17 IX 22 __5_ Www" m~ -'ON"-

This technical evaluation report (ThR) t is: s uini p thedocumentation only revicw of thc I luniari Reliability Analysis portion of 6c Oyster uclea Gcnerating Station Individual Plant Fxarnination ([PE) submittal to th toryCommission (NRC). The body of tie report consists of four secti on, per thritions of thc Task Order: (1) this Introduction, which provides a brief summary of hiaroach to this Step I review and of the Oystcr C(reck IPE HRA approach; (2) Contractor R.0 deindings, adtailed documentation of findings tor cach work requirement specified ,in. the'sk rder (3) overall Evaluation and (:onctusiotns, which summarizes the imporan flndin desult sfrom thc review, and (4) the NRC summary data sheets. 1.1 Step I lIRA Review Approach Thc documcntation-only review approachi forOysterrk EH. involves the following six stCps. illustrated in Figure 1, These sC tspl esttpociyps 2'through 4, are interactive and iteralive, hut follow this general progressisott.. (I) Scoping Review - an overview of he.,entire IPS'$tbmittit.,Read summary sections, plant dcscripcions, the major HRA-pertinent tcti6sandresulI sections. Skin/scan thc entire submittal, including appendices ar'id eaieidf cnd and back-end analyses. Identify thc basic approach used for thie, HRA.and-t 0thc:*nization of the HRA documentation, including any obvious mijor omissions. ldentlf table pfatures of the plant, the ovcrall IS'ES approach, or the IRA approach that deserspecial attention. Identify and obtain refcrcnces that may nced to bc re~vievd..orch e; obvious points of interface with front-end and back-end analysis. Review d.sW ptionS ofIPM,4R.A team qualifications.' (2) D)etailed Review of I RRA Sectiontis. a d-t l d i ~ewand assessment of the primary IRA scction(s) of the submittal.: Ths involvest1 thorough (re)rcading of descriptions (of methcdlology, noting assumptions., data'or, 4 and other important aspects of the analysis, and annotating any questions, potentcln pblemwareas, missing information, or issues for further investigation. Secon j.i~nv.-a.. companson of information and documentation found in, the subrmi alL.it IH$ m e -RAm hodology/appmach to the information/ documentation qt0iremeitified in accepted HRA approaches used in other PSAs. For example, sinc the Oyst$.. Creek P E used a Success Likelihckd Index based methodology (SLIM),-;th -conpma& involved reviewing the information contained in the submittal rcgarding-t'he major; ste in4the SLIM approach as described in NUREG/CR-3518 and 4016 (Refh :. and. 2)ina1y, the detailed review involves an attempt to "track" the complete assessment of aw key operator actions through the IIRA process described in thc submittal. By tracldng.wc mean' simply identifying that tihC sulbmittal contains sufficient infonnation';to -cluarly delineate "'ft li .S M' O,,

mcthodology, major wassumpt ions 1, m.esuch as performance shapihg factors, data sources, and refcvrnce for tu .uittiyc and quantitative assessmcnt of human actions. Therc is no attq t nuntitative analysis. (3) Rcsponse to Work Requiremenit_ ;asisssmt of speific issues identified in thc Task Order work rcquiremcnts. This is,.n iten!~bYlmin asstssment responding to each work reqluiremcnt. Thc focus is identificad on of and. wweaknesses of the llRA portions of thc submittal and insights

gardlngtarcsults or :potential arcas of improvement. Any questions that :reuire inpu t from the licensee arc identified.

This step includes compnction of hN

ahts, which is Work Requirement 2 in the Task Order.

(4) Interface with Front-End and Bac-End Reviewrs:. two-way exchange of information and discussion of issucs. The focus is on HRA.cts of front-end or back-end analysis, hut includes a general exchange ofi iformatio findings. The interaction takes place informally throughout thc review, but prmarily t completion of the overview in Step I above, and again after completion of Steps 02 a 3 as writing of the TER begins. More lomlal interaction occurs during ;the closingeeting of NRC staff and IPE review contractors in Step 6. (5) Prepare (he TER: - devclop and weite lhis Lthi;cal evaluation report. This involves: preparation of a draft reportdocumenting )kUl work accomplished, findings, and conclusions, internal technical rcviewyvenifylngfindlings and conclusions and compliance with Task Order Requircments: editoria reind: printing. (6) NRC Slwff and Contractor Meeting - held aftif$bmittal of the TERs from contractors to revicw findings and conclusionns ad Ona ie'lsdns for the licensee (if any). 1.2 Oyster Creek PE BIRA Approach' ; The Ovster Creek IPE consists of Level I and 2: Probbilistic Risk Assessment (PRA) without cvaluation of external events.- Thc PRA's::.methodolotyemploys the "large event tree - small fault trec" approach. Thc PRA is innovative in that iih`ogic',of the plant model is entered as logic statem:nts or "modules" that can be ditectly linikdiminadng the need for support states. Specific operator actions arc identified by. the' analysts.bid on review, of operating procedures, systcin analysis and develo ment of plant, model, and,lnorporated into the system analysis for ys stem split fractions and plant model, i The IRA approach described in the submittal, essentdally dircted at quantifying human error probability (HEP) estimates, was, performe Using a Success:Likelihood Index (SLIM) based methodology. This method relies heavily.,o)n, the, us ofoperator. input in evaluating human actions. The submittal provides details of perf omaishaping, factors used, the structured opcrator survey formal and process for dtrmnauon of PSF values, and the process for evaluiation of lIE!'. ~' ' - - A.,Z U

2 It WI .NDINGS The subsections below ~drs xlcty t$~je.ec f~cwr currcts spCcified 'I'hc Semgt,,S"cD~tn5 iloVa f bfthrc work recluircrticritss ci in thc Task Order. Forloach ite; ttm to identify notable points asbt thc subrilittal, hoth strcngths and weakn sshs, and thts as to how thc submittal might tx. improved with regard to thc specific work tedid-nt Al dthe overall intent of Cicnenc Letier 8X-20. Information obtained from the licci-se )NAre 1e tNRC questions has been factored into this fj na I rcrport.. r i 2.1 Work Requirement 1.1 Peerfrrm a gnP.g re iew of the human reliability analysis.

S.;:

... i 2.1.1 WR 1.1.1 The [PE submittallis etially complete with respect to the type of information and level of detail-reqUete in the PE Submittal Guidance l)ocument NURE-1.3435. List any, obvious omisSions. NUR:?EG-. -I. ous X lable 2-1 lists thc major items identifiedni: NUOi*33S peninent to HRA that were checked, Thc following arc the findings forithis work it en e -,t. of pat_,',. (1) ieneral Methoxiology. The plant modevl eltt~4eoped by combining the response of plant svsterms with operator functions asprovided in pl'antprocdures (EOPs and abnormal response 1)rxce(]Lircs) to represent the intcgratied plantrcspts. These operator functions are included as top Icvel events, Models for most of iese f un~t qUire operation of systems or components. System models required to support th op etalso include imponant operator actions (including many of those described adXve which i systcm operability. They aze dxcumented in the system notcbooks. Specific ope:ittatoractiodpt identired by the analysts were cvaluated, arid the reults wetr incorporated-iftwA he .ve system fault trces or sequence event trees. The overall (descriptioti of the-HRA. efon rtin Sealp 6.of the PRA (Level I) report provides a clear understanding of the general methodolo approach to addressing human actions within thfe ll.. 'Thec model of human interactions used f the evaluation of HEPs splits the response into three phases: identification, diagnosis and response. The actions of operators were classified as skill, rule and knowledge based4aedin4 and evaluated accordingly. The SLIM-based method was used tiO valuttei ;perator actions in the IPE. Input pertinent to perfornancc shaping factors (PSs) wa d ObtI from operators. The submittal provided reasonably detailed descriptions ot~thei.s.nticturtd st onnairc used to obtain operator input. PSks used were described and justifie. The c tion of HEPs based on input was outlined in the submittal. Only post-cvent human errors w evaluated for the IPE. (2) Inforimation Asscmbly. A listing of rference P of similar plants, including Pcach Bottom (Ref. 3). that were reviewed for thevOyster -rek A was provided in Section 1.6 of 44 f 0 j,~~~~~~-..,, JQ I< X w. 0' ' ', ' ,,.'0 j' > ' ' f J $ ' / ".. 04 OWN I I I I .4

lbc2-1 NUREG-iV$1kAjtC~tdw 11.1 ,NURliGi 1335 Rl. IrRERfNCE tM'OkMA1 R W4ENT. TO: IIRA 2.1.1 Genetal Methodology Condise des An of HR effort and how it is inlegrated with e lPE tasks/analysis. 2.1.2 Information Assembly ia l:2.2 L st of r:fetcncd PRAs, insights regarding -IRA,' hunja4, M ormanX

2.1.23 Conc.iie kischp on of plant docurrentation I 0usedfor I-IRA i~frat onm concise discussion of the pr:cess usc t&Conflm that the HIRA represents conditionsinA th bilt, as-operated plant.

2:,-, h, 2.4 pD~o tohe4 &walkthrough

activity, d___

noltid ing I4.iclist participation.

2. 1.3 Accidcnt Sequencc De$6ripttiono ess for assuring. human actions D)CeineCation cotsidIrdi jnldatlnR events and accident scequence delineation speiaMist involvement.

?4 S9s~teml Analysis Dcsription of ippss for assuring that the impacts -o~f huiman acri0~ arincluded in systems analysis,

.proce*ss (or in'fc.fath~g ;R;A.

2.1.5 QuvIantifical)rio Proccss I 2.15 41 PARAK`,e.imrnon, cause analysis.

2. 1.5.3 aTypes.,fhumn failures considered in the IPE;':a categorizttln and.concise description cxist.

2.:.54 Li.t1 f huLantreliabi lty data and.timc available for V. actions: data sources clearly Iideatii f 4"d i Mf ed. alist of errors considered. c.ito t I results of screening,

:W 2.1.S5 List ot:A data obtained from plant experience an od/process for obtaining data, list Of generi6

1di, 2.lS.6 Concisede ipt on. of method by which R'EuPsaar jqun

, including break down such as - taskJ.analysjs., aptehniqu s for combining. ,_______ pr.obabilities,+ ag lng dependencies, etc. 5

Tahlc 2-1 NIJREGI33S HR 1etfsChcedWR 11-1 =- NtJkl.(- 1335 R FJFRE NC!.. Th$RMA0N ~RTNENT7 TI IRA

2. 1.t Front-End RCsuits Scrcening Proccss and:: ; -' 1 '

Humi.n t tiont to important sequences arc c16;rly idGnt+f4 A eoncisw definition of vuinerabi~tiIs provided, along with a discussion of citeria u identify vulnerabilities. A listing of vflnterabilit H provided, with clear definition of iAt thum n performance. Underlying o rete d vulnerabilities arc idcntlied. 2i.6 S enqt that. Were it not for low human error rat, fery actions,: would have bcen above the apl1abk core damage frequency sdrecning citaare identified and discussed. 2 1.'7IAny h performance issies peninent to U;Is wor. Gt& ildenti(fied and discussed as atppropriate. BackY,'nd Submiittal npact-X ptor action on containment response

r. 1dcnyt~fiflcti~ns assumed to be accomplished by Qcrator reasnably cxpected to be f:a~tcomplshc~~cr tdie severe accident conditions ex ted ipment accssibil i ty, survivability.

infitormaton a>p.41abiity, etc. have becn considered. Crtical humn6 ctions havc been identified and included inyj nt-ces and quantitative iIRA

a,^SSc

mer1L 2.3 I Spcific Safety rFcaturesi;. l0 and Potential Improvemcnts

PI' Any aiI-ra related aspects of unique tdomftetyfeatures arc di.wsscd, inluding anythatsbited. in significantly lowering typically highequency core melt sequences.

,Hurhan related ptential improvcmens - procedures, -ining,. etc.- bSponse to vul nerabi iftics arc 'clearl~y:identifidand discussed. 2.4 IPF. Inec Utility Team and rnal Rcview The submittAl deprbcs the utility staff participation

andJ 4involvem in the: 1-IA. An independent in-

.hotic revie A was conducted.

t o

6 . 2? 'C A WY.-.

.i MOMMIN

/Y the subrrilittAll The methodology used frhl y 4 on Min thc TMI I fRA (Ref. 4) and is a rcfi nemcnt of' that aaI~ lia dcumcntazion to acouirc IIRA information was identified. It includcd plant:e >pr ures, emergency operating proccdurcs. (EOPs), and surveillace and: mahitenticdurcs, A.dctailed description was provided for each action to be analyied byn #t pcl ld ilg plant conditions and other constraints. The plant operators evaluate th~PSP~ by $,pen the "PRA Human Action Survey ornn." Thc survey proccss is a structrdth&tvaluatc the performance shaping factors, The survey form used was providd ite.-t1I .well. as dcuriled information on thle IPS[' brcakdown and linkagc to the survcy~ N or4nat sttnr Humran Action Walkdowns" were pcrfor ytean bers responsible for evaluating operator actions with experienced operator: personne y were conducted to familiarize themselves with the operator actions modeled as.well 'eri fy operator action survey forms. The SLIM-based evaluation process used plant:oprator input from the survey form to evaluate IISKs whiich were convcrted to the succcss likelihood inhx values. The survey process and infonration collection appear to be well structured, (3) Accident Seitience Dlelineation.i Tehnnr nn the plant design and supporting cAlculainons arc combined with abnormal response and ?pocdures to form the basis af thc l vcni Sequence Diagrams. Specific operator actions rddt'd to prevent degradation of plant conditions are identified by the analysts du ri.ng dvelopin t^&nd evaluation of Event Scquence Diagrams. A 'D)ctailed Human Action D nscrion'i A p edb RAanalysts and vcrific</ moidlified by lIRA walk;down. Dctails of each operator. ad j:were provided in Appendix Hi of the submittal Incorporation *(f operator actions intushJk i discusse d in Section 2. 1.3 of this TVER. (4) System Analysis. The System analysis is ,esribe detlon 5 of the Oyster Creek I-IRA (OCPRA) Level I report. System dcscription's are appritiely detailed and comprehensive. System notehox)ks were developeid for each sysitem 'anal;e

SA ttmai-y.of the contents waus included in the submittal, and notebooks are, ptovided irA dix Fto the-submittal. Included in each notebook are the important opcrator actions AfMor th sten, operation. In addition to routine information on major components andinstrumentatk, fthe notebooks include information oll systcm dcpcndencies and intcrfaces, testing aricdmaiaritoei et'ch nical specifications, system opcration, modeling assumptions, and success priteria, NrOpAror actions are incorporated into the PRA in appropriatc system fault trees. Documentation.ofofin fault trees are provided in the sYstem notcolxoks.

Documentation appears-to'.be suffckt t,c support a detailed evaluation, if one. were necessary. Thc incorporation of EOP 1 ,teps in:sy$tem Modes was addressed above, (5) Omuantif'ictionProccss. 'Huma Interartieons -(.,;.wergrouped into three major classifications for quantification. depending onpthe time a! wh the action occurs in the accident s(c-entrio. "Group A" H-is occur prioir tothe initiat r vent, ae the result of human errors durtng maintenance, testing, or calibration. .cdvtiics:"lB -IHs fare thosc that result in initiating events. These arc captured in thea',,ti`aevenilfeqtuencies obtained from plant operating experience. 1herefore, Group Bt n thc IPE IRA analysis. c - the ,PE H RA analysis. e t

"GCroup C" arc broken down into two sub-cl aOsltt(C oCta r~tlons pdrforrned in rcsponsc to procdiurcs, particularly Emergency Operatn 6Ptedtl'- and (CR) recovery qctions in responsc to unavailability of a safetyjfunction, -hich mayhmay not be procedurali7ed. CP cvcnits appar as hcadings in the event trcs or a, basic v;In system or functional fault trees. Type (CR evcnts are separatcly addcd to thj modellng initial quantification and are addressed at the acciden: sequcncc cutset level

e Or si eqimn failur'-

?. a Group A Ills crror frequencics were considered tobe rd in-the basic equipment failure rates for misalignmcnt or failure to restore systemsu T itil states that this failure mode is not a large contributor to system failure. The.ubma £4haertain Group A crrors werc included in system models, but no details were provided Ths is the subject of a request for adIditional information in section 2.2. 1. The quantification process used for GroupC Hls 'i.Oyster Creek WPE is described in cons;Jerahbc detail. I-or each operator action, a fairly detailedvdescription of plant conditions and other constraints was provided to the operatort TevSLlad evaluation process uses plant operator input for cvaluating Performance Shapingl cto-PA ). ';Selection of PSFs is justified in thc submittal. Conversion of these PSFs 6o te esn Iidex value is accomplished by use of weighting factors based on the class 6 dtiot1 dwledgecor skill based) for each "phase" of identification, diagnosis and resp c. Tel:lihodindex value is converted to error prohability for cach behavior imodel phas'e-using rentc actions to "calibrate" the SUccCSS LikelfihuKi Index for each action phase. The sur'cy sheets completed by the operatorsi are nt~ru d to a lcvel of detail and with questions intended to reduce the variability.of the subjective nst s All inputs were analyzed to provide: a data spread for statistical analysis f.or cstimaiipg., the uncertainty of the values There is a concise surnmary of the common cause-analysis rovided in Section 5.3.3.3 of the PKRA Lxvcl I report in the submittal. The submittal states'pro-initiator human errors arc not c'I.sidcrel boecalise they arc captured in the component fte-data :analyzed. Pre-initiator (Group A) human crrors *uc discussed in section 21.2.1 o 0his*ort, and a request for additional information on their treatment is provided in at scto. o o causc events are a subset of thilese pre-initiator ermors. (6) Front-End Results and Screening Process.the Eultt defines vulnerability as any core dlamlage sequence that exceeds I.0t E-4 per rMotor ear1 ny containment bypass sequence or large early containment failure that exceeds l,0 E pet r year. No vulnerabilities were identified. A structured review was petforned to identify po tl low cost improvements. The resuhls of level I and 2 PRAs, as well as contnbutori to sy$ehavailability and operator action error rates were reviewed. io

4.

E

a. -

S *e~ - f f /Pu u' ' 00 B < ' 1 X2 II

No listing was provided of sequences: that were" 'it not 'for 1human error rates in rccovCry actions, would have been above the applicable core damnage ruency criteia. nor was any'clear statcment that n(o such sequ ences cxist. A rtequircd hy NULfEG-1335, GSE and other safety, issucse;& haLs rinternal flooding, Loss of Feedwater Control, and alicmatc water supply for6,drywecl sp yesei injection, wcre analyzcd by Oyster Creck, and the resulis are. reported iiijhe WE &ibitial. No vulnerabilitics were identified. Several analysis of these safety issuesinriolved h i' tion9 ~which, were considered ilimrtant enough to have potential improvementtljdentidtQ l I. The alternate drywcll spray source considerj Ipssr Ofipotection diesel water with nManual operated valves., Bcause of high radia'tio m core damage, the required shielding to allow access would make.the rnm!odiflCatipn t prohibitive for the minimal affect on cooling corc debris.

2.

Prxcedurc changes to improve operator rcsponse1 titeI 1.fltioding were recommended. A new reactor overfill prevention systeml.Is to be installd r loss of feedwatcr system control becausC of concerns about operator responses to ilate MSIVs within the allowed tlirlC., (7) flack-End Sbitta. The Containmcnt Event Trecs (CETS consider the influence of the pihysicafl and chemical processes :on changing the tontatnmedtlt,'pressurc and (in the case of containment failure or bypass) on affecting the release of fission'oroducts from the containment. T'he end staitc or the front-end analysis is binned according lant damage sttes and use as input to (Cis. The plant damage state informationii.ncludesfhaillowing categories: physical condition in the reactor coolant system and containmnent at trd ot vessel breach; integrity of primary containment and status of associated active -$ystorns: intogrity of secondary containment and status of a.s.vxiatui active systems. Containment rKxicls include "diriy vening." Thee.aeie4 n, n actions directly modeled in the analysis. The containment analysis used th6 resulti fie l system status as input for the back-end plant damage state. Thercfore, rny hptnah ats wcre indirectly incorporated into the back-end analysis. The results of the fronjt-idn.o im t venting HEPs were used in the back -cld analysis. (X) Specific Safety Features and Potential Improvements. A n-fber of specific safety features of the Oyster Creek plant were discussed in Sectioni t oftthe ubmittaJ Specific procedure changes and modifications were identified as cost cffective and. abing implemented. These include: Containmntl vent modifications and aissocated procctire-revisions. 9 .,'X0: ';:o. 7:~~~ E ,..,.:t . e.

Station, blackout technical i basi '.doum 'idntegrated loss of offfitc powcr procedure to provide: recovery 1off itrjitc;pwtr; :for alignmcnt pnd cross ticing huses to critical. cquipm n an d Gap ond alignment of alterate AC capability. e Loofll pri the integrated loss of offsitc power procedure Reactor ovcrfill prevention' systemic is to,e sta1Ied for reactor overfill transients because of conccrn for operator6.tspOn-6 to.lb te MSTVs within the required timc. 4;,, Imnprovemenis or cnhancernents under consider.tion.'incli;i, tevelopmcnt of a specific procedure nd ntraig on reacor overfill transients. Operator training should emphasgze, ImportaiWt s '.(listed in Section 8. 15) which were identified by the PRA as imForta~ntAin re,`ing core damage risk. (9) If'l. ltilitv Team and Internal Review' While 4theJP'development was supported by a consultint (PLG, Inc.), the submittal states that. on. of 4`' objectives of the study was to build on in-house PRA expertise and dcvclop tools -for, onron*goisk mMigement activities after thc completion of the PRA. CPU sprovided system.hn~y ts: engineers and plant operations personnel is a part of the PRA team. HR.s ali'sts,,r*ihe connactor organization aa well as (il I9 were included on the IPE icam. The initernal rcview process described in ie'i b 'tobe extensive. Multiple engineers and operations personnel iwith, expertise in

ek' design and operation were involved in the reviews. A review of the commen ts spggc.,tat the team provided a thorough review, An outsidc consultant with expertisein PRA nethIogy reviewed the 1PE for technical ictlhtoxis. With regard to the personnel on the,tem

, ho w'n individual was identified as the IIRA revicwcr or as having previous HRA"exporiencoThsubmittal would be strengthened if a thorough review of' the NRA portion ofIthe-IA..were ini'Idd in the rcview process.

2. 1.2 WR 1.1.2 The employed 1IRA methodology s darly described and justified for selection.

4 Section 6 of the Level I report included in the submit~t l:arly descibes the steps performed in the IIRA portion of the [PE. The SLUMm'Tn~htodol,,i.'awell. stablishod and documented JIRA arproach. The SLIM-bascd evaluation.p tssu Oyster Creek uses plant operator input as the basis for PSFs which are. converted.to the-'Css likelihood index value using weighting factors. The success likeliho indx alu 4ivenod to error probability using calibration values from "known HEPs. -'Thr.,are reque5 :cor: `ditional information on the imniecrmentation of the SI.IM-based'methodologyvwhtirceA led in the sections which follow. 0~~~~* XXLXe n

0000t~'4

\\0,Y;.4. ID

j.. 2.1.3 WR 1.1.3 The methodology (nciudtng$hmanactiopr taxonomy) cmpiloyed is capabic of idehtifyihg fig porta h i tiong, and conitins a discumssio of the most important humat actiQns.gd4errors: 'Ihe hunman action taxonomy used in+t!e JIRA

M}kady identifled in the submittal. The model of human interactions uscd for eVthe iluIdividce the rcsponsc itlo three phliaws

identification, diagnosis and respons. Te of operator were classified as skill, nile or knowlcdgc based actions and were cvaiuated adlngly: Details on the human act ions and the qtiantification were provided in Setton 6'and6 Appndix:E of the submittal. UThe submittal stated that procedures were review1.to identify operator actions to be included in thc plant model. One important opratr *act t was not included in the plant model, and 1which is in the EOPs, is containment flodinghis was identified by review of independen review comments for March 27 1111 me.ting ; ppandix D to the submittal. lhe rcsponse to the comment was that the operatlolr cionwlWtrui "to establish or maintain stabli shiutdown conditions." Because -the steps, i: EPs e contrinmcnt flooding would likely hbe carried out by thc operators. IPns fot B Suppression Pool type containments have identi iFed containmcnt, flooding as a sourc of conment failure when core damage and vessel melt through occur after thc torps is oded (lIss of pressure suppression capability). The Licensee indicated in respornse to an NRC qu:sl$¢nthis point, that this potential down-side" of containment flooding had becn. evaluted dan1s not fcluded becausc of its low likcihoxod (if (k'siJ1TtCfl.; : 2.1.4 WVR 1.1.4 The 1EF_ submittal cpoe iableprocems to confirm that the W1E' represents the as-built, as operajlant. 'Technical infornatn on the plant,desgni hn40-pporting calculations are combined with a)normal response and EOP proceures jto fonlvtbasis of the Event Sequencc D)iagrams (WISI). The VSDs werc prcsented to varous GPUN organizations including plant operations, safety analysis, and training departmentsfr revie.The resulting final ESDs were used as the prinmary input in the development ofh lt m-ei In addition, walkdowns were held to varify in or n was$correct. A structured pogranm was provided to prepare detailed descriptions of all Hutman Actions to be analyzed. Plan( walkdowns )y risk assessment personnel, a human factors specl0.list, and plant operators ovcr a 3 day period conFrirned the accuracy of the detailed dosciptiotl$. Thc final chek on asbuilt and asoperated wat p.rovedl by the Independent Review Group. Membvers were chosen for-their expertis<<in plaint and operation. The Indepcndent Review Group reviewed the entire submittatiniltud te s and operator action sections. This process appears to: be a :reasobable ad "tc approach to assuring that the II'E represents the as-built, as-operated pL t 1,

21.5 WR 1.1.5 The JIRA hadf per re tot-;lp assure thc bnalyie lechniques werc correctly applied Thc internal review process dcribcd i the subritnd discussed in Section 2.1.1(9) abxve appears tO h conprchcnsivc, with cxccpton tof thetob, analysis. No individual was ideritifiecd a.s thc IIRA reviewer or as having prvious HR xeence. No other peer-revicw Was identified in the submitwl for the HRA analysis.P, view by qualified IIRA personnel helps provide additional confidencc that the 1.A i ere appropriately applied and results arc correct. Tic submittal would be strengthened y adional information conccring any IIRA review and qualifications or the HRA rewer(): g

~ ~ ~~

~~~N f 2.2 Work Requirement 1.2 Rcvicw the ino1tjfiki;y sequences that could occur at the plant. ..i: -i'/ 0' 2.2.1 WK 1.2.1 The accident sequcen~c approp y considered human actions consistent with other NUIREG.1150 and oOhcrNIW3 cepted PSAs (see table NURE(;-1335 Appendix B). 'The huMIan ctions of Grand Gulf (Re f. 5)we.iad to the OCPRA human actions. TIhc review shows that cquivalent, actions we cdonsid" the OCPRA sequences. Additional hIuMan1 actions were inchlded in the OOPRAKbc i~ideof Ithe additional operator instructjions provided by the new (Rev.; 4) EOPsS As was d in Section 2.1.3 earlicr, a potential discrepancy in the incorporation of POP steps w}s twntified and additional informnation on the process for identifying and including proceduralized.prator actions into the PRA. lPreinitiator (G(roup A) hunlan errors squc asLcacbron error or misalignment of systemris or insltrUmnientation are not mcxicled in the PRA.- Te scubdltal states that "misalignment of systemrs aIrc not mLKJcicdl in. the (PRA\\ sincethescasc,."

of unavailability arc captured in the component failure data,` Pre-initiator hurnan errors r normally considered in PRAs; (e.g., see Grand Gulf (Rcf. 5) and Surry (Ref. 6)P!JRA4.

W it i re that, in general, pre-initiators typically have lcss impact on estimated CDO thanidoo-daf'dgtors, significant contributions from pre-initiators have becn identified in som ,RAs stematic analysis of pre-initiator human errors and contributing factors would provide much'reAter confidence that no important errors leave tken missed. And,: the informatinm gained-O "generic" factors.infuencing human perfornmancc, e.g. procedurcs.o1 administrativc contr'1,may indicate relatively low-cost means for significant improvement. 2.2.2 WR 1.2.2 The accident sequences scrcnedouttbcause of low human error (see NUREG;11335, Section 2.1.6.6ppear appropriate, based on HRA techniques employed. The submittal addrcsscd the importance h buman aosb examining the contribution to core damage for three groupings: (I) all operator actions(2) rto' groupd into nin genenal categories., and (3) top 10 individual operator ctions All modeled operator actions 12 V 0,,., 0','z..' s- '

0

wcrc found to contributc 21% jf toall iofiige. Thc most important groups Of operator action and rcmoval of containment heat. Thc individual opcrator acdons*were tirmhnost important groups andLATWS sequences. and their contribu tion to total core damare rang m l% to 2.76%, Dwcaidd inforrmation about cach operator action is available;in nAppcndlX of thc Lcvel. I Rcport. ITe anal ysis of operator action Col utb~ution.ttorae da mage provides insight into which operator actions arc the most imponrtnt, but( nth f iormation rcquired by NURFEG 1335, Section .1.0.6 was not found, 2.3 Work Requirement. l evkw ~tiuantltatl~ e nature of the 111E submittal. 2.3.1 WR 1.3.1 Thc employd hum eorprobability (HEP) screening valus appear capable of gcrocnhing in ficant humnan errors. Screening or "conscrvativc" vatup;s were used fronly a few sciected operator actions including circulating water system floodi ng" ndlosa ositC power recovery in this 1I'F. In these cases values arc provided-without referencing any ice. c The values appear to be approprilate, but the siiniittle would he strengtheed If thc

.uc. of the values is referenced or additional informlaltion on the technical basi.k for thdse c4 ncs were provided.

While therr were few actions for whichnuecal screening was pcrformied. it should bx noted thtll potenltially significant qualitative screening is perfornled in the process of selecting those huiarnl~l alctions to be evaluated.:Oqcrator acis modeled, including recovery aclions appear to he aippropriatc based on review of similar PRs.' However, the submittal does not provide much inf'orniation on the proess by whichithe specfi ones selected for Oyster Creek. In general, the basis was said to be "required' oprat rctnt s, and abnormal proccdures. 1The submlitIal would be strengthened by' a dis~tAsdunl iepecific rationale, assumptions ;In(d criteria for selection of' actions. 2.3.2 WK 1.3.2 The IPE 'divloped nian error probabilities (IIEMs) for significant human actionS, 'or, provie Wationale for using screening values. With the exception of the screcnting; values`citd jve no numerical screening of I-IFPs typical in many PRAs was. identified from the subot ta review. Actions selectcd for analysis were analyzed directly, and lHENPs were dcveIod ethod used to quantify 1IRIA's are discussed in S.ction 2.3.3. 2.3.3 WR 1.3.3 Sources of generk humanllabllity data used in the IPE were identified and rationale for their ust provided.:Generic human error probabilities (1JEP) data were modified using plant-speciric Performance Shaping Factors OPSF's) as appropriate, and rationale provided for selection of employed PSFs. ,:~4' ,EN .D ,-;-00 'f0.'<,,,: , __ I

I The SIAIM-based evalu ation procc us ea lpr actiohs for the I IRA. Thus the data is ncither "gcneric" fnor -p ant.7pe i lnthe ijsua1 srnsc of those words. There is some mcrit to thc asscrtion that sin:e t pe op rirc from this particular plant, their judgmicnts probably rcflect some degrec of plant s C"ifc pcriencc. On the other hand. the operator judginent primanily specifics the re/u ijt imq'..ofPSFs. The absolute values arc detemlincd by the selected anchor-points, andtheibti ios nol discuss the selection of those anchor values in much depih. The prossr. scle f PSFs and justification of thc OfnCs sselected is reasonably well described in sbtialTh SPs chosen for use along with the process attempt to account for dependenci df e mltiple and successive operator actions. As noted earlier, the process for ciclwrdon

f8wet udgmnt from the operators appears to bc well structured and $systcmc apiedl Convcrsion of these l'Sls to the success likelihood indei( alueis accomplished by usc of weighting factors based on the class of action (rIe, knowege 'or skill based) for each model for action phase of identification, diagnosis and responsc.: -,nAdditional factor to account for the signilicancc of the class of action for theidijagnosisphaseWyas used to increase the value for knowledge based activities.

The submittal prc'vides g ri lovrview of the basis for the weighting factors used. Because of the importance ofth ighing factors in calculation of Success l..ikelihood Index and the ViIEPs, it is teli that Th bttaj would be strengthened by inclusion of a more detailed description of.th2i, asi)?4 a#od 1-r d process used in developin the wTfigting factors used. g S

0..'

'..0 'Thc Success l.ikelihoxd Index value is converte to error prbabihty using reference actions to calihratc" the SUCCeSS likelihoo Index valu for achc aon ,identification, diagnosis and response phases. As indicated above, there is l-itt'le rti provided in the submittal on the selection of reference actions/valucs to calite I' 23.4 WR 1.3.4 The recovry method is rd iedt for recovery acions app:ar justified Three types of recoveries are addrcssed in the.$ubittal system recoveries incorporated into system logic modcls, proxedurally directed rIc=veries, and procedurally directed recoveries. The later two types of recoveries were added: to he plant model following initial quantification and refinement. With exception of "dirty venting"ndiscus n Section 2.1.1.(7), no credit was included for post vessel breach recoveries i'n.the back-end P Methoxds, data, and assumptions used to quantify recovctionsgre clearly and concisely sutlnlianle7d. Information provided includes, adescrption Q h recovery action, amount of tinmc available for the action, manual actions required, proire avaOlability, how the need for action is perceived, cognition class for activitiesj and siucces criteria for :recovery. A concise description of I'S s and their use in the SLIM method was prded., WVhile we did not perform detailed checks to validate numerical estimates`: the RlEP 4,ues overall appear to be reasonable and consistent with other PSAs. Values for'slcgted ortor Actions were compared with prevoiuts PRAs in section 6.3.6.of thcLevel 1 report Jn fand fi to bnit. , 4 a; w2t 4;F 2 f< i' ' M d' 2 'l

2.4 Work Reqi rement 14 r ucing the probalility of core damage or fission.pr.6duct r$ 2.4.1 WR 1.4.1 The APE analysis,-app .to support the licensee's definition of vulnerability, and. that the definition prb%'ides a means by which the identification of potenlial zvulncrabiIits .(as. so r dermed) and plant modifications (safct' enhancemcnts) is made pssible. The IPEt submittal :defines vulnerability. s aIny damage sequence that exceeds 1.0 E-4 per reactor year, or any containment bypass sequencd t,targe ly containment failurc that cxccds 1 0 E-6 per reactor year. No vulnerabilities:were identfied. A structured review was performed to identify potential low cost improvetnents..Th sults of level I and 2 PRAs were reviewed, well as major contributors to systcm unaviila i y IAd operator action error rates. Rcsults of this review are discussed :below. :The oveti'l[ -ss employed 'in the IPE for identifying vulnerabilities and cost effective safety enhknden~es appars to be comprehensive and able to sy-s-tealtically identify cost effectiSVq ty ehhents, 2.4.2 WK 1,4.2 The identifcati ofp provments include human-related plant modifications (e.g., procedur d lng,),.'pnd proposed modifications are reasonably expected to enhance hun ablity and plant safty. Cost effective plant improvmcnts identified de `Ih IPE process and being incorporated arc discussed in Section 8 of the IPE rrt. 'Th re$lts bof 1evel : and 2 PRAs, contributors to svstem unavailability, and operatqr 'ction errorates' were reviewed to identify potential enhanceentcris. No information was pr8vid d.on ~,evaluation of the improvemcnt in the 11'1 results, but it appears that the additior.6al g~ atwcprocedures:should enhance the operator rerfottamcC. Specific cost effentive 6hentenrifld are being implemented including: Containment lVent nodificatohs d td ssodakd procedure revisions. 11,airi-,...... e ::t e 1, tdd. MC..........S Station Blackout technicalV. basis.do fnt 'and integrated loss of offsite power procedure to provide: rectry of offsor onsite power; alignment and cross-ticing buses to critical.equipmt ' nd,;tsrtqp.r d alignment of alternate AC capability. 1 Loss of all DC power prpteurem tobe Oor dinated with the integrated loss of offsite power procedure A new Reactor overfill,:PreveCtion st is,to be installed for reactor overfill transients because of col .i fo t s ses to isolate MSIVs within the required time. 4,4

Imlprovcricnis or cnhancemcents undercosideratdi Developmcnt of specific procodure anid Operator Training should :mphasiro idcntified by thc PRA as imporan jn r the PRA 0 < $,H' g on reactor ovcrfill transicnts. actions listed in Section S. 1.5 Oore damagc risk. Were While no discussion of thc cvaluatdonlforirnpl procedure changes, training cmphasis and Idltl' by the PIRA as contributors to operator e0: 2.5 Work Requirement 2.0 Comp Compicted data sheets are :included in vwas ~provided in thc submittal, the toutd help address problems identified tvailability ::.

444.;4 44,j 44en' 4

4 7 444 ,34 44) 4444-4 4-4 44444 444444 44... 4 44444 444.4 3/4 "4 444 .4 4-444 I I : . I q I r I O I,, :, I I I

3. OVERALL EVA1 lATI AND CONCIlUSIONS On the basis of our review, :

concluded at with; rgard to the IIRA, the suhrjalttal demonstrates that thc licensee used *.asonablr J o meet the intent of (icneri: Lcter XX-2). ()verall, thc iRA methodo1fgyzisg frntification Of important actions, analysis of factors influencing human performne, 4uaidon of human error, assessing the impact of humann error on system responselandict f (::DJ and releases) appears reasonabic and cotsistent with practice in ocher P#AsJAr ble process vas in place to identify px)tcntial human-related irnprovcmenls, Notable wcaknesscs of the submrittaltre`the etotreat pre-initiator errors explicitly and the description of basis for weighting factors and cf L. 'of reference Human Error events to calibratc Ithe SLIM-based HIRA cvaluatlon, 'I: is typfc4 Practicf in PRAs to test pre-initiators such as laintenanscc te~st and: calibration. ertrs :cxp , -The Dsubmittal should include a clear and concise justification for the asscriton thatsuch <o. are negligible and /or are incorporated in componcnl failure data. Thc conversion of PS, Su ~ccss Likclihood Index is accomplished hV iuse of weighting factors for difrtnt types om interaction. A more detailed description of th1c derivation of these weightinjg. factors woui ihave strngthened the submittal. Thc SLIM-ha.eld meilnidology must be "calibrateld! tings ktni n or accepted J£Ps. The submittal would halve becti strengthened if the discutsion and JuWiflcation. was expanded for thle 11I1N tlused for lalibration of the SU1M based, HA. 2 0 V'- 2- < f~~~~

v~S;m

."',:,,.A'%,',,,'X" 0,', :. t '' S , "', ,0"" : 0 ,' N'#, '0' A 0 A X 0. .i:.,,' j ~ :.'

E LA

4. IPEF EVALJUATION A4DDTA i4AY ShIEETS IPE DATA

SUMMARY

$TS Plant Name: Oyster Creek Nuclear C er--;fl Informatinn Assembly List of plants, PSA, or other analyusknownqte employed similar mcthodology. 'M I I (PLO) Ix-Control Roomn actionns treated? List, Ycs, NMultipic actions as: required for redovmrY: able..6.4,.ia-c provides operator actions for sequences an& plant. l,_onjo_ aptI i .',0 P Ifruan Failure D)ata (Generic and, lant, peciic) Y-Analytical ncthod used, e.g., Eprd t l 4ERP, SLIM-MAUD, HCR. TRC. SIAMb ha sed Were the following human crrorste ondc (1) Pre-initiator, e.g., maintenanrc eT i d testing, cquipment calibration, andc restoration, A.isumed to he included in comronent failUr ta. (2) Post-iniatitor procedural9 Yes fall

(3) pos t-intiacrrcvr Control'R -Ex-Control -oom Ycs " i"" "' 'i' Types of human errors 11nsidered comission, commission Errors or omisJloo onl Source of human reliability data - Gecncric Oa ta) X No Simulator Data? ,No Expert Judgmcnt-- Used SLIM mcthod! Opcator inpiutbased on detailed descriptions of operator actions using structured survey ftnoprovide input for PSFs. Most significant, operator Actions$,'%,: The most importntngoups ps.'mtor action were those assciated with cstablishing RPV inWOO c iTmen heat and ATWS sequences. Human. Error contnbution to Gore damge feuency (if known). 21 % i1 t

I Vulnerahilitics aIISOciated with h~un Noric identified PLAN T I MPROVEMENTS'AND UNJR1J S AT URES IsIT'pro)vctmcnt insights stcmmkg from TM Appendix B or submittal rcvicw,$ itb0rs opera or errors and provides following recomrncndations (Section 8.S). .;I; Q Coli1.iidcr spccic tit' crfcto overfill transicnts. Cnonsidcr uraining emphasis 'tatCnsisti¢,succssfu) performance of following actions can red.uce Core Damtigc rsk: Opcritnr injects flrv watfri'thrcIgh C Say tystem during loss of AC powcr and urisola tcd LOCA outsidecontaiditrceents: (4pcra, r inhibits ADSN sancd l.nc4T A Fduring ATVS with FW availahlc and condcncrfaikd dand f V fSV closurc. Opcrator inhibits ADS dwuih. SATWh \\h FW failed and EMERV/SV closurc. Operator manually rc-cneritzs zs AJt1IB und rctarts at kast onc TRCCW' pump following a loss of o bh tr Opc rator trips reactor afte f4ilurT(iigh l Opcraitor srcurcs or isolates cndensat4 tser header to reactor building within I or 2 hours after condens te ^.nsfersply Iine brcak in the reactor building rator tips plunt.k i ng line brcak in the runion Irnplcmcnted human (actor nitpot!7c11 Conuiinmnit Vent modi catloni and istc4 prcdure revisions. \\,- -qt'N,: '20 d 0 ~~~~~'9 r

Station Blackout tcchnlcalbjs Pintegrted loss of crffsitc powr dOct td oV o fand. w procedurc to pmvidc: rccovCroy f.sItpsit power, for alignment and cross-(icing buscs to critical cqluip r*n itr p ad alignmcnt of alternate AC ,capabiliiy. Inss of all tiC power proccdurt £0_4, t wd wi£th the Integrated loss or offsifc power procedure

K'<<j' A nCw Reactor ovcrfilcd peveb ct o sys itq be nstalled, for reactor overfill traficnts bccao t

olat MSIVs within the reiluiral time.- I.nhanccmcnts undcr colsidera t l)cvclopmcnt of specific procidure 4d trA 1n n reactor overfill transients. (operntor Trruining should cmrpair Impo tacions li~tcd in Section R. 1.5 wcre identified by the PRA as imp 31ant inrdtilrg4Ore damaggc risk (listed undcr imylprovcrfient insights aN -ve)- 'Mc alternate drywell spray tsrce nsidert4c~ o f firc protection diesel water with minual operated valvIs, lectnscoigh r odiation from core damagc, thc reqjuired shielding to allow as s would the modificaton cost pmhi'tiive for the minimal affect on cooling 11t1 debris

Prix~cdure changst WmPr*v0et;.P M a0'c~~~ oIntmrnl flooding were recomrnindcd,.

In:{a4, Portithlc D)C generator and qUpreflnt C4 (rupplycScnltial bC loaids IA 7' 0. ,,~4.;X t,,,

I.Imbrcy, D."IAMMATYArAp fi Af~sin Humran Error.Probahilitics Using StructuredExpert Ju4gmn. 1J ~C~51911USNIRC, March 1984. 2Rost, r-..A. ct al., "ApplicatilOn fSII A eto Inezvc Comuc Blased Methods for OlfitinrtA~smit ffurnn Performanct and R eliability.'. NOUREQ/CR-4O0,i br 9$

3. USNRC. "'Analygig of Core Dan"_gI Pqe1Jsto0ntr Events: Peach13ottom Un if 2 NUREG/CR4SSONJV 4 0tb? ~

4. Pickard, L~owe And Oarrick` n~"IhtM~

Ind' Unit I ProbabIlistic Risk Assessmenti pered (,or GPU, ~Niiee4i'December 1986. .5 US NRC, "Analysif of Corce D)aimerudIsfo ten~a )Events, Grand 6ulf. I,` NURIiG/CR-45S3f"I l,' 1

6. US NR C, 'Analygsi 'ot Cre ba4

~~*mIternal Events, Surry, Unit-I," NUR[XWCR.45,5O/~Vl-, j~, I .~~.4 NI'. .9449}}

ML060550282

Contents

Text

SUBJECT:

SUMMARY

SUMMARY

SUMMARY

SUMMARY

Navigation menu