ML20079D321
| ML20079D321 | |
| Person / Time | |
|---|---|
| Issue date: | 06/30/1991 |
| From: | Basdekas D, Leung V, Mazetis G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | |
| References | |
| REF-GTECI-130, REF-GTECI-NI, TASK-130, TASK-OR NUREG-1421, NUDOCS 9107080230 | |
| Download: ML20079D321 (34) | |
Text
.
NU R EG-1421 I
Regu atory Ana ysis :?or tae Resolution 0:? Generic Issue 130:
1 Essen:ia Service Wa:er System Fai.ures at Mu :i-Uni: Si:es l
i U.S. Nuclear Regulatory Commission 4
Ollice of Nuclear Regulatory Resemch V. I.eung, D. liasdekas, G. Mazeris p " " %,
o
$e i
- W 1 8#28 92o63o 1421 R PDR i
.., - _... - - _. ~.... -, - -...
l l
f t
AVAILABILITY NOTICE I
Availabihty of Referonco Malonals Cited in NRC Dubhcations Most rjocuments cited in NRC pubhcet.ons will be avadablo from one of tho following j
bources:
1 1.
The NRC Pubhc Document Room, 2120 L Stroot, NW. Lower Lovel Washington, DC
[
20555 t
2.
The Supor.ntendent of Documents, U.S. Government Pr,nting Offico P.O. Dox 37082, i
Washington, DC 20013 7082
{
3.
The Nationat Technical information Service, Spongfield, VA 22161 f
I Although the listing that follows represents the majonty of documents cited in NRC pubhca-tions, it is not intended to be orhaustivo.
j Roferenced dccuments available for inspection and copying for a foo from the NRC Pubhc Document Room includo NRC correspondonco and internal NRC memoranda; NRC Office of inspection and Enforcement bulletins, circulars, information noticos. inspection and investi-gation noticos; Licenseo Event Reports; vondor reports and correspondonce; Comm:ssion papers; and applicant and hconsoo documents and correspondence.
t The following documents in the NUREG sones are available for purchase from the GPO Sales Program; formal NRC staff and contractor reports, NRC sponsored conference proceed-ings, and NRC booklets and brochures. Also avallable aro Regulatory Guidos NRC regula-tions in the Code of fedoral Regulations, and Nuclear Regulatory Commission !ssuancos.
Documents available from the Netional TechnicalInformation Sorvice enclude NUREG senos f
reports and technical reports prepared by other federal agencios and reports prepared by the Atomic Enorgy Commission, forerunner agoney to the Nuclear Regulatory Commist. ion, Documents availabio from public and special technical horarios include all open hterature items, such as books, journal and periodical articlos, and transactions, f odoral Register noticos, federal and stato logislation, and congressional reports con usual'y be obtained from those Obrenos.
t Documents such as thesos, dissertations, foreign reports and translations, and non-NRC conference procood,ngs are evaliab!o for purchaso from the orgenization sponsonng the f
publication cried, t
6 i
Single copies of NRC draf t repons are avadable froo, to the extent of supply, upon wntten roguest to the Offico of information Resources Management. Distribution Secton, U S.
Nuc! car Rogutatory Commission. Washington, DC 20555.
Copios of inoustry codes and standards used in a substantrvo manner in the NRC regulatory proccas are maintained at tho NRC Library, 7920 Norioth Avenuo. Dothesda, Mary:and, and are avaltable there for reference use by the pubbc. Codes and standards aro usually copy-
{
righted and may be purchased from the originating organization or, if they ero Amencan l'
National Standards, from the Ameriven National Standards instituto, 1430 Broedway, i
New York. NY 10018.
P v,.,e-m,
,-n-,.-m,n
,.,,,,,-,,,,-,,,,,,,,,,-,,,,,,r,-,
--+,,ww,,-,
,,,r.,,r
,--n,,,,,, -,,.
-,,--,,,,,-m,+m,,o,,.n,v--,,,--~w
i suiuso-mi Regulatory Ana ysis for the i
Resolution of Generic Issue 130:
Essential Service Water System l
Failures at Multi-Unit Sites 4
i Manuscript Completed: March 1991 D.ite Published: June 1991 V, Ixung. D. ItasdcLas, G. Maretis
\\
Division of Safety issue Resolution OITice of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555 i
s.....
r
, _ - _.. _,.. _,,,... - _ _. _ ~. _. _..,,... _....., _ _ _ _ _, _ _ _ _. _. _ _, -. _. _ _ _ _. _.. _ _. _ _.. _ _ _ _ _ _ _ _
AllSTRACI'
'the essent!al service water systern (liSWS)is required to pres.cnts the regulatory analysis for (11-130, lissental provide cooling in nuclear power plants during normal Senice Water Systern l'ailures at Multi lfrill Siles."'the operation and recident conditiota. 'the l'.SWS typically risk ieduction estimates, cost /lsenefit analyses, and other supports component eooling water heat enhangers. con.
insights gained during this ef fort have sJ.own that tainment spray heat exchartiers, high pressure injection irnpleinentation of the recommendations will Signifi.
pump oil coolets. cmcrgency diesel generators, and ausd-cantly reduce risk and that these improvements are war.
iary l'uilding ventilation coolers. l'ailure of the IISWS ranted in accordance with the twLfit rule,10 ClIt function could lead to s,cvere consequences. Tuis reiwt 50.10Ha)(3).
hi Nt titi!(;_1421
b CONTENTS t
$ 'U),*t Abstract......................................................................................
iii i
I !3 cc u t ive S u m ma ry.............................................................................
vii 1 S t a t e m e n t of t h e Pr oN e m....................................................................
2 Objective..................................................................................
3 3 Al t e rn a t iv e lt e sol u t io n s.......................................................................
4 3.1 Al t e r na t ive 1 -- N o Ac t io n................................................................
4 3.2 Alternative 2 -install Additional Cros> tic...................................................
4 3.3 Alternative 3-Provido !!!ccincal Power Cross Connectioe....................................
4 l
3.4 Alt ernative 4 -- Provide Se parat e intake St ruct ur e............................................
4 r
3.5 Alternative 5-Modtfy Technical Specifications R equirerr.cnts..................................
4 3.6 Alternative 6-Provide Independent RCP Seal Cooling System................................
4 3.7 Alternatise 7-Combine Alternatives 5 and 6 (IS Changes and independent RCP Seal Coohng)....
5 4 Tec h n i cal l 'ind i n gs......................................................................
6 4.1 Cor e Da mag e Fr eq u e ncy Analysis........................................................
6 4.2 Dos e Co n s e q u e nce Analysis..............................................................
Il 4.3 Cos t A n a lyn s...........................................................
11 5 Val u e/im pact Analysis................................................
16 5.1 Al t e r na t iv e I - N o Act ion..........................................................,.
16 5.2 Alternative 2 -install AdJitional Crosstic...............................................
16 5.3 Alternative 3-Provide lilectrical Power Cross-Connection....................................
16 5.4 Alternative 4 - Provide Separate Intake St: uctur e,...........................................
16 5.5 Alternative 5-Modify Technical Specifications itequirements.................................
17 5.6 Alternative 6-Provide Independent RCP Seal Cooling System...............................
18 Alternative 7.-Combine Alternatives 5 and 6 (l'echnical Specification Chantes and Inde 5.7 it CP Seal Cwling).......................................................pe nde n t 18 5.8 U nce rtain ty Analysis.................................
19 5.9 1.ife lixtension Considerations 19 6 D eci si o n I t at iona l e......................................................................
20 6.1 llackfit R ule Considerations.......................................
20 6.2 Plan t. Specific Considera tions..........................................................
20 6.3 Relationship to Other Generic Safety Issues...................
21 6.4 Conclusion and R.itionale.....
21 7 Implemen tation............................
23 8 R e f er e nces...........................
24 v
NURl!G-1421 l
Tables liS.1 liest 115timate Cost /llenefit itatios ($ /I'erson Item).............................
niii 4.1 Operational States of Multi. Unit Sites....................................
7 4.2 State Dependent l oss-of. Service. Water Initiating livent l'requencies............................
8 4.3 Conditional Core Damage Probabilities of the Sequences...............................
10 4.4 Core Damage l'requeng f rom Individual Sequences...............................
10 4.5 Cor e Damage l'r equ ency - Summary.................................................... 10 4.6 l ' ail u r e M od e Classificat ion.....................................................
12 4.7 C D l' lled uct ion foi At t e ma t ives........................................................ 12 4.8 lienefits of Proposed Alternatives (l'erson.it em)............................................ 12 4.9 Ilest listimate Costs of Proposed Altematives ($ Per itcactor)...............
13 4.10 Dir ect Cost listimates ($ Per iteactor)................................
13 4.11 Cost Offsets for Proposed Alternatives ($ per Itcactor).................
14 4.12 O n sit e Con.;equ e n ce s..............................................
15 4.13 Itante of listimates for the Total Cost and the Net Cost ($).....
15 5.1 liest-listimate Cost /llenefit llatios ($/l'erson-Item)...........
17 5.2 liest-listimate Cost /ilenefit llatios ($/Perwn-Item) for 20-Year 1.icense llenewal 19 NUlt!!G-1421 vi
EXECUTIVE SUMM AltY
'lhis repirt provides supjutting information, including a
'!he results of the analysis todicate that the tore damare value/ impact analysis, for the Nudcar iterulatory Com-frequeng (( I)l')ansmp from i SW sptem failure is esu-rnission's (NitC's) resolutain of Generic luue 14),"I's.
matcJ to be 1.5215 01 per acastor year. 'the staff cum.
sential Seruce Water System I adures at Multi Unit incJ sescn pisuble alternatnes to hm cr the Cl)ll, and fitts." this iwue addresses the omettm regardmg the esumated that the potential reJuctions in Cl)l' rante essential seruce water (I SW) system at multi unit pres-from 1.37E 05 to 9.131L 05 per reactor Scar. A detaded sur:nd water reactor sites having Iw o liSW troins per unit dcwnpuon of rmdchnr and auumptions used in the with a croutic capabihty (14 reactor uniu at 7 sites).
analysis are presented m NUlti U /Cib 5526. "Analym ot Typical emnponents cooled by the l{SW >ptem under Risk Iteduction Measures Apphed to Shared lisential normal and accident mnditions are the component coob Scruce Water Systems at Mult: Unit Sites."
ing heat exchangers, contamment spray heat ewhanrea, high-pressure mjection pump oil cmlers, emergency die' Acost benef;t endaation of the possiblealternatisesinde scl rencrators, and auuhaty bmlJmg s entilation eoolers.
cates that cost eff'; cove optnins are annhdde. O= a
'the 1:SW system is also used for coohng the reactor rnore of these alternatnes h oc the potential for signifF coolant pomp (ItCP) stals, typically indirectly ua the cantly r educing the ink frorn hws of I SW. Table I S.1 is a component adag water system (CCWS)or the thatrinE summary of the best.cstimate cost / benefit rataw for each pumps.
of the alternatnes examin?d. Companson of the best-
'llu.s generie issue was initially idennhed as a result of the estimate cost /beneht ratios f or all the alternatives arainst safety evaluation related to the bmiting ";ondition for a guidelme cost / benefit ratio of $1(KK) per person tern operation (if0) relaxation program for Ily ron Unit 1.
shows that all the alternatives are cost beneltetal except ITW system support from it) ton Unit 2 ua the enustte Alternatne 4, which entmis a separate intale structure.
lhe terulatory analysis used these cmt/ benefit calcula-betacen the two units was not available whde Unit 2 was tions for conudermg a popmed resolunon to GI-130.
under constrastion. To support the ICO relasation pro-
.lhe proposed resolupon is a mmbmanon of Alternative 6 rram, flyron Umt I performed a pn habdistic nsk aucss-ment (PRA) of the !!SW system. 'lhe inurhts denved (or 6a) plus Alternatoe 5 to paniJe a backup means of fmm that study indicated that the core damage frequene)
RCP teal coohng plus additional liSW terhmcal specife canons aM emergency procedmes.
arismg from the unavadabihty of a two-train (one pump per tratn)l!SW sptem could present a sigmf acant ink to the p3blic health and safety, particularly if one !!SW
.lhe cost /bt nefit ratios were ab.o calculated for heense pump aom the adjacent unit na an !!SW sptem crosstie
' C"'* dI I"I "" "M U"" I "'"" "I 2U lf d' S' "I d 'C *di"I"E is not available.
pinut hie of 50 years. A comptnson of the results shows that the cmt/ benefit ratios for all analped backfit alter-At malti unit sites,crosstics ate usually prouded betw een natives a,e considerably lower for extended plant hfe.
the !!SW sutems of the adpent umt to enhance opera-INen so, Alternatne No. 4, Separate intake Structure, tional fleubdity; however, the technical specihcations stdl remams appreciably higher than the $1.tK)0 per (IS) for these plants have sypically not placed any oper.
person-tem ruidehne at a cost / benefit ratio of $2.285 per abthty requirements in the aJ;acent umt's I;SW system, perwn scin.
puticularly during shutdown noJes 5 and 6.
GIinterest to the decision makers on this rencric issuc
'this regulatory analysis is partly based on a mothfied are the insights and news available in related PR A reliabihty analysis performeJ by Rrookhaven National dmumentation in the open htcrature. 'lhe PR A w irk laboratory (HN!_) for the llyron plant. The pR A model available in NURl!G-1150 (plus supporting documenta-was mothheJ to reflect the multFurut configuranon and tion)is an extensive source of risk analyses information the assumption of hanng an 1:SW system fadure as an for an tmderstandmg of liSW s ulnerabihties. An examin.
iniualmg event for the accident sr quence. Also, it was ation of the NURIL-ll50 documentation of the three determined that a more recent value for RCP sealloss-of-PWRs studied intheates that the analyst considered the coolant accident (1 OCA) probability, based on the data liSW redundancy for two of the three pWRs large enough developed in NURiiG-1150 "Reattor Risk Refercrce that a cornplete loss of !!SW as an event-imtiator was Document"(1987), should be established for the present deemed not credible (eight pumps available in Sequoyah, analysis. A model was des cloped to incorporate the prob-Units I and 2). None of the five plants in the abthly u an RCP seal 1.OCA as a function of time and NURICG-il50 study is af fected by GI-130, however, it leakage rate of the RCP scal. In addinon, both short-and should be noted that one of the PWRs (Zion) identified long-term recovery actions that mqht affect the fmal the scruce water wrtnbution to rak to be substantial outcome were exammed.
(approumatdy 1.51!-4/RY). 'lhis contribution for Zion vu NtlRl G-1421
l I
was approumately 429 of the total core damare from all
'the staf f proposes to rei.oh e Gene ne issue 130 by issuing causes.
a generie letter, under 10 Cl It 50 $4(f), to the htensees and appheants of the 14 plantwveh edin this evaluation.
Another Pita woik availab!c in the open hte.rature is
'lhe e $ntcnt of the retiene let er will address both the NSAC-l4s," Service Water Systerns and Nuclear Plant pres entive and nuticatis e aspests of the propoicd resolo-Safety," dated hiay 1990. Although it is only a compilation hon as discussed in Chapter 6.'t he implementation phase of earlier Pita results performed by the industry f or six of Generic issue 130 will be closely coordmated with that plants, it is useful to note that a greater appreciation of of Generie issue 23, which deals with the itCP seal reli-the.cmce water splem's contnbution to plant nsk has abdity for both nor mal operation and acciJent conditions.
moved the industry to initiate a program to improse sen-
'iuidance for resohing that generie issue is proposed in ice water performance.'lhe limited guidance avadable in
- 1) raft Iterulatory Guide 1)G-100'. "lteactor Coolant NS AC-14S is a step in the right direction. 'the wide tante Pump Seals." Wlule awaitmg completion of pubbe its iew of Cl)l's (frorn low service water) oser the six plants and comment on 1)raf t Itepulatory Guide 1)G-100S, the studied surgests large variabthty in plant specific 1;SW backup scal cooling portion of Alternative 7 oce Chapter configurations. 'lhe averare Cl)F (from low semce b) may be defened.The reason for allowing the defertal water) for sa plants was 6.5f t:-05/ItY, with a contnbu-of this adJitional proicti,oo ictates to the cather deseb tion range of 2.331504/ItY to" negligible." hiany detads opment and promu1 ration of 10 01'11 50.63, the station of these su Pit As are not includcJ in NSAC-148 and, blackout rule, which was based on an assumption regard-thcr(fore, must be used only with treat caution. 'lhe mg the inarmtude of itCP seal leakce dunng a station oserall mempe is that the scruce water system prouJes blackout esent. While it was lef t to GI-23 to vahdate that an impor; ant safety function that could be a 5,ubstantul assumption, GI-130 n ale tused on a seal i OC A mo& I contnbutor to escrall plant risk. 'this messtre tends to scry smular to G1-23, but d.fferent from the leakare le7d addeJ crcJence to the GI-130 cor.clusions.
assumption m 10 C1'It 50 63.
Table I:S.1 Itest.1: stimate Cost Itenilit Itatios ($JI4: son.itt m)
Total Cost llent fit
'l otal Cost!!te nt lit Without Astited With Astritd Alternathe Omite Omts Onsite Costs l.
No Action 2.
A ddtuonal Cresstic 433 23s 3.
Electrical Cioss<onnection W
Note 1 4.
Separate Intake Structure 3847 3n51 5.
Technical Speelfication hiodifications 4 ProceJures 25 Note 1 6
Iliph Pressure Pump for itCP Seal Cooling d62 hs4 ba. 17trewater for 'lhermal Harner Coalmp 37 Note 1 7.
Comhnun:>n b 4 5 756
$14 7a. Combination 6a, 5 39 Note l Ste 1: inaJas mcora onme cwe. nm.nea m a m i mi mon N UltliG - 1421 un
- l. STATEMENT OF Tile l'ROllLEM Generic Issue 130,"lissential Senice Water System 1: ail-A suney of operational esperience (itefs. 3 and 4)shows urcs at Multi Unit Sites," was identified in 1986 Otefs.1, that a number of ddferent components in the liSW 9 s.
- 2) as a result of the flyron Unit I vulnerabihty to core-tem may fail to perform their intended function in a damage sequences in the absence of the availandity of variety of ways. Ilowever, review of operating expenence llyron Unit 2 (not operational at the time), llecause of the has indicated that there are dominant failure modes for considerations in heensing flyron Units 1 and 2 and the the 17.SW system associated with failures of certain com-immediate need to make a third essential service water ponents. Such fadures have invoh ed the tras chng sct eens (11SW) purnp avadable to llyton Unit i via a crosstie with or other common cause problems at the intake structure one of the two llyron Unit 2 liSW pumps, the llyton th.it leading to partial or complete loss of the water supply.
I concern was treated as a plant specific issue. Ilowever,
'the liSW pumps and their electrical power supply are the Ilyron phnt specific issue raised questions relative to other important contnbutors to the partial or total loss of multi-unit sites that have only two 1.SW pumps with a the !!SW system. All liSW systems at the GI-130 multi-crosstic capabdity between them.
unit sites are safety estems, and their designs are plant-pf;c,,ah plant :;pce6c equip.- ent, emne ca;9W'y, l'ourteen units at seven sites that have the baue livron nd BW operabdity needs for successf ul acciJent mitipa-liSW configuration were evaluated as part of this issue.
Hon operations.
'lhese multi unit sites have two liSW pumps per unit (one A comprehensne review anu evaluation of the operating per train)with a sharing of one of the two pumps with the expenence with liSW sy>tems has been performed and is other unit via a en sstic betw een the two units,livaluation reported in NUlt!!G/CI(-55 6 (itef. 3).1 xcludmg 9s.
of other design configurations of liSW sptems m light-tem forhng (sediment, biofouhng, corrouon, erosion),
water reactors (LWils), including those of smgle unit the total number of plant evenir involvmg a possible r.ites, will be per formed under GI-153, *! oss of I!ssential completelossof the!!SW funcuonwas 12(Itef.3, Appen-Service Water in I. Wits."
dix 15). System fouling data w ere noted but excluded frorn the current analysis because of the earlier resolution of
!! should be noted that the enteria for the success of the Genene issue $1, "Irnproung the Itchabihty of Open
!!SW systems in providmg adequate coohng capability Cycle Ser ice Water Splems"(see also the discussion in during normal, accident, anJ pmt-accident cenditions are Chapter 6 of this report).'lhis penod of data retties al was design-specific and depend on tht plant configuration, calculated to be 667 reactor-years for PWits.
the capacities of the liSW pumps, and equipment de-In PAR one esent invohed a ec.nplete loss of liSW at pendencies on 11SW coohng. Although the enteria for San Onofre, Umt 1. At 100% power, a shaf t on the oper-success may be as vaned as the 11S% sptems, this evalu-ating salt water coohng (SWC) pump sheared because of ation assumed a genenc set of success entena as a repre-vibration. This event then led to the additional fadure of sentative model for putposes of quantifymg the eveu.s the norrnal standby pump (discharge valve failed to open) leading to possible core-damage acc;Jents. lhese genene as well as the fadure of a second auuliary standby pump cnteria are discussed below and apply only to mulu umt (lost prime). 'this led to a complete low of liSW flow for sites having two liSW pumps per plant with a croshe
.dmut 15 mir.utes, at w hich time a fourth pump was manu-capabihty between them, alh crossconnected from the travehng screen wash sys-tem to estabbsh coohng water flow.
During normal operation, one liSW pump per unit pro-A detaded examination of the loss of I!SW events indi-vides adequate cooling to systems such as component cates that a number of events occurred in hhdes 5 and 6 coohng water (CCW), reactor coolant pump (ItCP) scals, and air conditioning and ventilauon. 'lhe second liSW (shutdown), and seme of them may not hase been a com-pump per unit is assumed to be in a standby mode nor-plete loss ef liSW in ter ms of total stoppage et liSW flow, mally. llecause of load sheddmp (isolation of nonessenual esen though the i.SW system might have been declared equipment), one 1:SW pump per unit is awumed capable inoperable, of handhng accident and couldown heat loads. lypical
'the difference m the !!SW sptern between power and equipment cooled by the liSW under these conditions are shutdown operation is pnrnardy the actual heat load and the CCW heat exchangers, containment spray heat ex-equipment affected by the loss of !!SW. In adJition, the changers, diesel pencrators, and auuhary building venti-actual administrative requirements imposed by the tech-lation coolers. With one plant in power operat on and the nical specifications also difier and make these two opera-second plant in the shutdown or refuchng mode of opera.
tional modes more distmtt.
tion, the criteria assurne onc !!SW pump can provide adequate coohng to shut down the operatmp plant To calculate the initiatirp esent hoquency for loss of through the crosstic connection shou!d the need aro.c.
lihW. the total operatmp 1 SW-sptern years for all 1
N Ultl.(i - 1C 1
pWils, %7 reactor years, was divided into two parts as tive time fractions before calet'ating the core darnare follows:
frequeng (C111') salues, as discussed in Section 4.1.1.
487 reactor years at power Should a lost liSW splem function fail to be recovered.
ISO reactor years at s.hutdown the r,sulting core damage accident could lead to signifi-cant :uk to the public.'lhe most dominant sequence is the I:inally, the respective losses of allI!SW frequencies w cre
- cactor coolant pump seat loss-of-coolant acc dent (l(CP-calculated to be 1.111-~03 per reactor year at power, IDCA). 'this specific sequence is the subject of GI-23, 3.211-412 per reactor year at shutdown with one purup
- lteactor Coolant Pump Sea' 17ailures" (Itef. 5). GI-130 running and one at standby, and 2.911-O' per reactor year estimated the total core Cl)l? attributable to the loss of at shutdown with one pump rtnning and the other in l'.SW for seven two unit sites (Section 4) and the cost maintenance.'lhese numbers then were weighted for the ef fectiveness of ses eral alter native mahfications (Section various operational states of each unit and their respec-
- 5) that could lower this Cl)l7 NIJit!!G-1421 2
l
- 2. OlUECTIVE
'the purpose of the Gencric issue 130 program is to evalu-issued ruidance to the staf f(itef. 6) set ting a goal for CDl' ate the safety or adequ.ny of a two. pump 1:SW < rem in of less than 1.0ll-04/l(Y from all contnbuton. To meet existing multi unit l' Wit power plant sites, and secondly, such a raal, the staff has aimed for the benchmark that a to examine the cost effeeth eness of allernathe measures smgle contributor to the Cl)F be no more than 10% of the fer reducing the overall vulnerabihty to !!SW system fail-abos e sugrested value, or no more than 1.Oli-05/l(Y 'the ures.
application of the safety yoal guidance and the objectives of previously resolved USis to 01-130 was limited to Probabilistic methods were used to ast. css the CDF, the using them as general guidelines to the deenion process potential reduction in risk of the mothfications, and their described in Section 6. l(irid application of such a quanti-cost effectiveness. 'lhe overall objective for rc solution of tative objective to define an absolute requircment was not 01-130 is that contnt ution from loss of the !!SW system made, Smcc the liSW vulnerability ssue is only a fraction should be a snnll percentage of the total CDF from all of the f atal r.mtribution to risk from all causes, the cut-causes.
rent safety goal ruidance that the overall mean frequency L
of a large release should be less than 1 in 1,(XK),(KK) per I
l'or USI A-45, the staff recommended that the frequeng 3 car is not directly usable in this case. 'lhis is panly be-of events related to decay heat ternmal f.tilure Icading to cause an oserall probabihstic nsk analysis (PI(A) consid-core damage should be : educed to a lever (around enny all causes was not in the scope of GI-130. Ilowever, 1.0ll-5/l(Y) that the probabdity of such an accident in the consistent with current pohn ruidance in 1(cferences 6 next 30 years wculd be about 0.03 oased on a population and 7, a judgment was made that, in light of the safety of around 110 plants. A similar core damare objective roals and available knowledge, the recomrnendation to (1.0l!-5/l(Y)was noted m USI A-44 *Statian lilackout."
backfit selected design and operational improvements to
'lhese objectives are also cons.htent with the recently reduce nsk from liSW failures is wananted (Section 6).
4 i
I 3
NUl(IIG-1421
- - - - = - -
0
l
- 3. Al,TERNATIVE RESOI.UTIONS
'lhere w cre several alternatives consider ed for the r esolu-passing the intake. 'ihe common mode failure of these tion of Genene issue 130. 'these alternatives sre de-screens may occur as a loss of the common inlet or com-senbed below.
mon water source.'the whole intake structure or screens could be af fccted by events such as floodmg or frecting.
3.1 Alternative 1-No Action
'the alternative considered here is a completely separate intake structure and swing pump serving as a redundant With this alternative, there would be no new regulatory intake sourte of liSW water. It may be located on the requirements. Consistent with existing regul.tions, this same water source, but in a phpicalb separate location.
alternative does not preclude a licensee, or un applicant An alternative design, which would provide additional for an operatinglicense, from proposing to the NitC staff independence and diversity, would be to install the addi-design changes intended to enhance the reliability or op-tional intake structure c. a physically separate water crabdity of the l.SW system and its comsments on a sourec (c.g.. pond or lake). 'the separate intake structure plant specific basts.
alte nativeincludesthe tructure screens,associatedmo-tors, valves, and piping. A r.,mg EoW pump would also be made available to either ucit with redundant electrical 3.2 Alternative 2-Install Additional power supplies. common mode failure considerations ar e Crosstic assumed to play a primary role in the desirn and installa-tion of the new structure (such as heated spaces in areas
- lhe liSW systems of the seven multi unit sites analynd of the country subject to f reeting conditions).
um, r 01-130 are cross connected through pipe connec-tions and isolation vidves. '!his arrangement allows the Modify Tecimleal 3.5 Alternative 5.degh.o. aentS operator of one unit to utilite the 1;SW cooling capacity of Spee.l.l leat. ions the other unit under most circumstances. In most cases, the crosstie isolation valves can be remotely operated. A in operating modes 5 and 6 (shutdown and refuchng, hardware failure to open the isolation valves, should the respectively), the status of liSW pumps is uncertain be-need arise, could result in advcrse conditions. A parallel cause technical specifications (13) typ;cally do not re-cross-connection could reduce the possibihty of this kind quire that the liSW pumps be operationalin these shut-of failure, and in addition, would allow for more flexibie down modes. 'this alternative invohes imposing an mamtenance options.
explicit operabihty requirement on at least one of the IISW pcmps of a unit while in modes 5 and 6 to provide 3.3 Alternative 3-l'rovide Electrical b dup for the other unit ESW vystera An additional improvement is testmg the unit crosstie valves to provide l'oWer cross-Connection greater assarance of operability, thereoy reduemg the hardware failure assumptions on the crosstie valves. Also, In general, the electrical power supphes to the ESW this alternative includes additional credit for improve-trains are separated and have no cross-connection capa' ments in emergeng procedur es for recovering from a loss bility, i.e., the Train A liSW pump cannot be powered of service water accident.
from electrical Train 11 (or Diesel 11). 'this alternative investigated the possibility of crosstics between the elec-trical trainsof the unit with respect to the operation of the 3.6 Alteritat,ive 6-Proviu two ESW pumps (rrains A and it).'the cross-connection independent RCl' Seal Cooling of electrical power supply of other electric d comp ments, g gggg3 y
such as motor-operated valves, was not considered as part of this alternative because of their less significant poten-
'lhis alt rnative provides an independent water supply tial contribution to risk as observed m the operational and 'hsNhution sy> tem for backup cooling of the ItCP expenence failure data.
seah in case of liSW loss. Preventing an ItCP seal failut e and, hence, a small break losvof coolant accident (IDC A) would remove a substantial risk contributor as-3.4 Alternative 4-l'rovide Separate sociated with this issce. lhis alternative is ako a consid-llitake Strueture crabon in Generic issue 23, itcactor Cooiant Pump eei S
l'ailorec(itef.5) Anobjectiveof theiesolution of GI-23
'the intake structure is usually a single structure divided is to reduce the probabihty of seal failure, thus making it a into separate bays by concrete walls. 'lhete are a number relatively small contributor to total core damage of screens installed to prevent large foreign objects from frequeng.
N Uit110-1421 4
3.7 Alternative 7-Cornhine reduction. ilie conibination af Alternatives 5 anit 6 Alter:1atives 5 ant! 6 (TS Changes n nicly technical spenfication (hantes regarding liniits
"" '#"" E "I" ' E '" " ' ""' " $ '*
"""E""'"*""E' all(I Ill(lepell(leni llCl' Seal crations, crosstic testing requirernents, and procedures
(,00lillg) iniprovernent combined with an independent itCI' seal Under this alternative, a cornbination of two or rnore cooling systern, could be espected to result in a inore alternatives discussed atxwe could result in greater sisk substantial CDI reduction and still be cost effective.
5 N Ult i:(i -1421
-..1
4, TECilNICAL FINDINGS
'lhe llrookhaven National Iaboratory (llNI.) evaluation to the respective operating states of the two umts derived of failurcs of the liSW system at multi umt sites included from a faul; three analysts (Itcf. 3).
a determination of the frequency of initiation of loss of the ESW m. tem, core damage frequency from liSW fail-4.1.2 ESW System and 1(CP Seal 1.OCA ure, dose consequence analysis, and cost / benefit analysis.
14pegyen.
1hc detailed evaluation is found in NUlti!G/ Cit-5526 (itef. 3).
'the event tree established in iteferer.cc 2 indicated that the small 1.OCA caused 17 ItCP seal failure and auxillary feedwater (Al'W) systerr. failure are the dominant acci.
4.1 Core Damage Frequency Analysis dent sequences. It was dedded to use a more recent model for seal 1.OCA probability. 'the ItCP seal failure pro es areased on the miel developed in itefer-The core damage vulnerability mused by the f ailure of the ence 8, w hich provides the probabihty of sealleakage as a liSW system enay be estimated by developing a full +eale fun on pf the leak rate and clapsed time af ter the loss of Pil A model, includmg initiating event fr equency catego-sealem hnp rics, event tree and fault tree analyses, and mcorporation of support System dependencies.1he Pita model was A s mplihed recovery model was also developed by llNI.
then modified to reflect various plant operating configu; in Itefuence 3 for the sequences relative to the failure of rations to analyre the consequences of the loss of liS%
the lWW systemJthe termery M! W%4 num-function in each operapng glate 05 shown m 1 Ata 41.
M of o cry factors that are established based on the To facilitate the present analysis, llNb selected an exist-ing Ily ron Unit 1 PR A model (llef. 2), w hich was already Operating experience data bases regarding liSW systems developed and which examined the !!SW gstem of a consisting mostly of licensee event report (1 lill) submit-single unit (llyton Unit 2 was not operational at the time).
tals wcre scarthed by llN1, and, as confirmed by
'Ihe flyron model was modtfied by llNI. to include the NUltl!G-1275 (Itci'. 4), the duration of I!SW sy stem fail-effects of multi unit configuration and short or long-ute has varied from less than I hout to a few days. 'the term recovery actions. Additianally, the probability of data suggest that there are three (haracteristic time peri-RCP seal LOCA was established, based on a more recent ods of system tecovery. 'the first time period involves pump seal failure rnodel as described in NURi?G/
liSW failures that may be recovered within I hour and CR-4550 (Ref. 8), and incorporated in the present analy-consists of a large majonty of the liSW events (approxi-sis.
mately 70% of the total)flhe second time period,imolv-ing more problematical hardware or other failures, ex-tends up to 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. I' rom 90 to 96% of all events may be 4.1.1 Initiating Event l'requency recovered in less than 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.'the last group of events may take a relatively long time for recovery, and they The frequency of the loss of liSW as an initiating event for generally involve the most serious hardware problems. It multi unit site operations was initially derived from op-is estimated that only about 1% of the events were not crational experience for single unit pWR operationsJihis iccovered by the end of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
loss of liSW initiating ever.1 frequency was then modified to reflect multi unit PWR sites. As the system configura.
4.1.3 llelatise Tiine Fractions tion for various operating states may be different, the respective loss of liSW initiating event frequency for each Since the average time of operation varies with different of these operating states was determined separately. An operating configurations, it is necessary to estimate the approximation method was used, combining the experi-relative time fractions for each operatmg stateflhe rela-ence data with an analytical technique. A multi unit liSW tive time fractions essentially represent the aserage system fault tree was developed similar to the existing length of time of the specific multi unit operating state model of Ilyron Unit 1.'lhis modtfied model represents coupled with the arrangement of the liSW Systems.
the unavailability of the second unit to supply I!SW to the Maintenance or test-rclated outage time of !!SW equip-first unit, given the complete loss of liSW in the first unit.
ment must also be accounted for in the systern's average "the fault tree is provided in Appendix D of Ref erence 3.
time fraction.1hc ESW flow requirement may be satis-Table 4.2 lists the frequency of each opcrating state as an fied through the unit crosstics utilizing the liSW pumps of initiating event.1his frequency was calculated on the the other unit, liased on discussions with utilities, it was basis of the operational experience reflected by the base assumed that the crosstics are used about 10% of the time initiator, and then multiplied by a modifier corresponding durmg the shutdown period.
NURiiG-1421 6
l i
Table 4.1 Operational States of h1utti. Unit Sites Unit 1 Unit 2 Site's ESW Pump l'.SW Pump State Unit 1 1
2 Un;t 2 1
2 la OP R
AOT OP R
AOT Ib OP R
AOY OP R
Sil Ic OP R
Sil OP R
AOT Id OP R
Sil OP R
Sll
!!a GP R
AOT DN R
AOT lib OP R
AOT DN R
Sil lic OP R
AOT DN Sil h1 Ild OP R
AOT DN h1 h1 lie OP R
Sil DN R
AOT lif OP R
Sil DN R
Sil lig OP R
Sil DN Sil h1 Ilh OP R
Sil DN h1 h1 lila DN R
AOT OP R
AOT I!!b DN R
AOT OP R
Sil lile L)N R
Sil OP P
AOT lild DN R
Sil OP R
Sil IVa DN R
'J ' i DN R
AOT IVb DN R
DN R
Sil ivc DN R
$fU DN Sil h1 IVd IP' R
%?
DN h1 M
IVe DN R
hl DN R
Sil IVf DN R
J1 DN R
AOT IVg DN R
S3 DN Sil h1 IVh DN R
5 11 DN h1 h1 OP - Operating.
DN - Shutdown.
R = Pump running.
SI) = Pump in standby.
AOT = Pump in test (allowable outage time).
h1 - Maintenance.
7 NURl!O-1421
i Table 4.2 State. Dependent laiss.cf.Senice Water Initiating Event l'requencies States Unit 1 Unit 2 ESW Unit Initiating Event Pumps Pumps l*requency/ Reactor Year I - Unit 1-Up/2-Up R/AOT R/AOT 1.611- 01 R/Sil 1.41!-02 R/Sil R/AOT 1.211-02 R/ Sit 1.1li-03 II-Unit 1-Up/2 Down R/AOT R/A OT 1.211-02 II/ Sin 1.111- 02 Sil/hl 1.411-02 ht/ht 1.611-01 R/Sil R/AOT 9.7Ii-04 R/Sil 8.91.!-04 Sil/ht 1.111- 03 ht/ht 1.21! 412 111 - Unit 1-Down/2-Up R/AOT R/AOT 2.31!-02 R/Sil 2.111-02 R/Sil R/AOT 2.61!-02 R/Sil 2.311-03 IV - Unit 1-Down/2-Down R/AOT R/A OT 2.311--02 R/Sil 2.111- 0 2 Sil/h1 2.611-02 ht/ht 2.911-01 R/Sil R/AOT 2.611-02 R/Sil 2.3 fi-02 SII/N1 2.911-02 ht/ht 3.211-02 l
l NURin 1421 8
i
- lhe most dominant time fraction is that of the power (l.OSW) frequeng. The failure modes indicated in Tab!c operating arrangement, i.e., both units at power and one 4.4 are based on actual operating experience, liSW train of cact..mit running with the other in standby.
'the base case initiating event frequency was modified to take into account the effects of the partleular alternative 4.1.4 Core Datnage Frequency under consideration. First, the fraction of the initiating event frequency that could be improved by each alterna-The CDF from loss of service water was calculated usin8 tive under consideration was determined using the data the following expression:
listed in Table 4.2. Second, the relative change in the
!!SW system reliability with and without the improvement CDI = r' Aj (state)* P (sequence)' RTi provides an indication of the potential reduction in the i
CDF. Fault tree analyses, which me'uded the logie Where A is the state-dependent initiating event fre-modules or additional component failure rates that reptc-sent the proposed mod,fication, were empkiyed to quency, given that the unit is in this state for the full year, estimate the total system unavailability. 'lhe reliability and RT is the relative time fraction of theith state, while analyses of the improvements were periormed for each P is the ith sequence probability (conditional core dam-state or plant configuration, resulting in a calculation of i
age probability),
configuration dependent initiating event frequencies.
'lhe conditional core damage probabilities for the domi-As notea in Section 3, the following potential improve-nant sequences are summarized in Table 4.3. The sum of rnents w etc analy/ed regarding their capability to provide all the sequences during power operation results in P a cost effective reduction in nsk caused by a I.OSW (power operation) = 1.03Ii4)l, which reficcts the condi-event:
tional probability of core damage given a complete loss of
!!SW during power operations. The corresponding value Additional Crosstie-Reducing the possibility of the for shutdown is P (shutdown) - 2.8211-02. The most malfunction of the cross connection between units.
dominant contributor for all sequeneca, including shut-down,is the RCP seal LOCA: P (scal LOCA) = 6.811 4)2, e
lilectrical Power Cross-Connection-Increasing the which is approximately 65% of P (power operation).
redundancy of the electrical power supplies to liSW pumps.
The core damage frequencies from various accident se-quences are summari/ed in Tables 4.4 and 4.5. The most Separate intake Structure or llay with an AdJitional dominant sequence is the RCP seal 1 OCA: CDF (scal Swmg 1:SW Pump-Increasing the redundancy of LOCA) = 8.8II-05 per reactor-y:ar, which is about 60%
the ultimate heat sink or source of cooling and in-of the total CDF from IISW loss of 1.5I!--04 per reactor.
creasing the availabihty of the F.SW pumps.
year.
Changing technical specification requirements and "the total CDF from loss of liSW (1.5E-04 per reactor.
emergeacy procedu-year) is judged to be substantial compared to the total from all causes (typically in a range of atxiut 1.011-4 to Installation of an independent RCP scal cooling sys-2.0l!-4 per reactor-year). The next section presents tk tem.
results of an examination of different alternatives that could lower this core damage frequency.
A combination of RCP seal cooling system and changes in technical "ifications and procedures.
4.1.5 EITects of Potential Irnprovements on
'Ihe first three alternatives were selected based on con-Core Dainage Frequency siderations regarding the liSW failure mechanisms oh-served in the PWR operating historydata base. A particu-
'!he alternatives for improvements were initially selected lar operating mode when both liSW purnps of the in NURI!G/CR-5526 (Ref. 3) by considering (a) the shutdown plant are inoperable (State lid and h)is a con-dominant failure modes of the !!SW system (4isted in cern since there are no exphcit technical specifications Table 4.4) and (b) the dominant accident sequences con-requirements on the liSW system in this operating mode.
tributing to the relatively high CDF. Since there is no Therefore, the alternative of imposmg adthtional TS re-smgle dominant failure mechanism represented in the quirements was also analy/ed regardmg its effeet on CDF initiating event frequency, a number of different options reduction potential. 'this alternative also considers addi-were considered, includmg combinations of particular tional credit for unit crosstic testing and e'mergency failure modes to reduce the initiating loss of service water procedur es.
4 NURl!G-1421 1
-.~.-
t l
Table 4.3 Conditional Core Damage Probabilities of the Sequences Sequences Conditional Core Damage Probability Power Operations RCP Seal LOCA - P(Scal IDCA) 6.8E-02 Auxiliary Feedwater - Puw 2.3E-02 leng Term AISV - Pvaw 9.111-03 Other Sequences - Poise, 3.2E-03 Total - P(Operation) 1.03E-4)1 Shutdown - P(Shutdown) 2.821!-02 o
Table 4.4 Core Damage Frequency from Individual Sequences initiating Event Core Damage Frequency Sequence Frequency Sequences
'RT Probability-P CDF/R-YR Seal LOCA - P(SL) 13E-03 6.SE-02 8.8E-05 AITV - Puw 1.3E-03 2.3E-02 3.0E-05 1.3E-03 9.1E-03 1.2E-05 long Term - P:Arw Other - Poise, 1.3E-03 3.2E-03 4.211-06 Total Power Operation
- P(Power Operation) 1311-03 1.03E-01 1.3E-04 Shutdown - P(Shutdown)
~;.1 E-04 2.82E412 2.0E-05 TOTA 1.
1.511-04 Table 4.5 Core Damage Frequency-Summary Initiating Sequence Event Frequency Probability Core Damage States
- RT P
Frequency CDF/R-YR I + II 1.30E-03 1.03E-01 1.3 E-04 III 4 IV 7.1E-04 2.82E-02 2.0E-05 TOTAL 1.51!-04 NUREG-1421 10 l
l De most dominant contnbution to the CDF anses from A calculation of the consequences associated with shut.
the failure of the RCP seal upon loss of seal cooling from down operations was also performed. While the use of the unavailability of the ESW.Derefore, the installation power operation release categones for consequence cal-of an independent RCP seal cooling system that would culations at shutdown may appear to overestimate conse-cool the seals in the event of loss of ESW was also evalu.
quences, Reference 3 indicates that the person rem con-ated as a potential improvement. De result. are summa.
sequences are relatively insensitive to the source term.
rued in Table 4.5.
This is because of interdiction enteria and because of the relatively high contnbution of long lived isotopes to the Table 4.6 shows the relative contnbution to the imtiating long terrn dose. De total consequences for shutdown event frequency of vanous failure modes of the ESW operations were calculated in NUREG/CR-5526 (Ref.3) system, and Table 4.7 shows the reduction in CDP for the to be 3.1E+06 person rem. Elence, the overall benefit various altematives analyzed.
for each attemative considered in terms of averted conse-quences in person rem may be estimated by multiplying the power consequences with the power ACDF and the 4.2 DOSC Consequence A:sysis shutdown consequences w4th the shutdown aCDF, ada.
ing the two products and multiplying by 30 years, the I.or purposes of this study, consequences are measured in assumed Itfetime of the average GI-130 plant, llence:
person tem andbenefitsin person remavetted.Once the CDF and changes in CDF from a potential resolution Totaillenefit - 30 x (ACDFm x 5.5E + 06 +
alternative have been calculated (Section 4.1), the next ACDFso x 3.1.E + 06),
step is to calculate the corresponding consequences in person-rem, and hence, benefits in person-rem averted.
Table 4 8 shows the benefits (or consequence reduction)
,t he reactor safety study (Rel. 9) first atternpted to evalu-n person. rem that was calculated for each proposed al-ate contamment per armance for a number of accident temast sequences. As part of that attempt, a set of radioactive release parameters wss developed corresponding to spe-cific containment f. ilure modes. More recently, the NRC 4.3 Cost Analysis has documented in NUREG-ll50 (Ref.10) a detailed To calculate costs for the various backfit alternatives, assessment of the risk associated with five nuclear power several sources were consulted (Ref. 3). Some cost esti.
plants. NUREG-il50 t epresents the most updated ana.
mates were derived from an NRC-sponsored research lytical framework for the assessment of containment per-report (Ref. I1). Another source was the computer print-formance, meluding source terms and off-site conse.
out for the Energy Economic Data Base (EEDil) and quences. lt was decided to use NUREG-il50 as the basis supporting documents (Ref.12). Still another source was for the evaluation of the seven two-unit sites of this issue, discussions with utilities.
A rnore detailed description of these calculations and their apphcation to this study is given in Reference 3.The An initial overall assumption was that the backfits can be consequence model specific to the Zion site was used as accomplished outside of the entical path. Consultation the starting point of the consequence assessment of th with utility Eersonnel confirmed that this should be possi-seven sites of thts issue because of the availability of its ble. Otherwise, the direct costs will nse substantially, at detailed modeling and evaluation in the NUREG-ll50 the rate of $400K for each day that replacement power ts effort.nc multi-unit sites evaluated in the GI-130 study needed.
would be expected to produce average consequences smaller than those calculated for the Zion site because of For each alternative resolution, the costs noted in Sec-their location and respective population densities within tions 4.3.1 through 4.3.7 were considered.
their evacuation zones. For this rcason, adjustments w ere made to the Zion consequences as discussed in detail in 4.3.1 D.irect Costs Reference 3 and summaraed in the following paragraph.
This cost category includes factory purchases, installa.
A companson was made of the results from Zion with tion, and onsite labor and matenals, but excludes indirect those of the SurTy and Sequoyah plants, and it was con-costs (e.g., engincenng, administrative). It is given in the cluded that the consequences of an ESW induced core first column of Table 4.9 as a best estimate.
damage at a large dry containment plant, typical of the GI-130 plants, would be 47% of the total consequences Table 4.10 shows the best estimate and the range of esti-for Zion, or 8.0E + % person rem. It should be noted that raates in the direct cost:;. Alternative 5 (technical specifi-this is for power operation only and does not take contain-cation modtfications includmg pmcedures and crosstic ment systems recove ry into consideration. When racovery testing) shows zero costs because this item was already actions are taken into consideration, this number is modi-meluded in Column 4 (technical specification costs) of fied to 5.5E + 06 person rem.
Table 4.9.
I1 NUREG-1421
Table 4.6 l' allure hinde Classification Relathe Contribution l' allure Afode to initiating Frequency Intake structure unavailable 35 %
Inss of electrical power supply 35 %
less of ESW pumps 20 %
Other 10 %
Table 4.7 CDF lleduction for Alternatives Alternative ACDF 1.
No Action N/A 2.
Additional Crosstic 1.6011-05 3.
lilectrical Power Cross-Connection 1.4 t!-05 4.
Separate Intake Structure 9.1313-05 5.
Technical Specifications hiodifications and Pmeedures 2.5511-05 6.
Independent 11CP Seal Cooling 7.8213-05 7.
Combination of Alt. 6 + Alt,5 9.1011-05 Table 4.8 Ilenefits of Proposed Alternataes (Person-Hem)
Iow flest liigh Alternative Estimate Estimate Intimate 1.
No Action A Additional Crosstie 739 2,635 4,951 3.
Electrical Cross-Connection 645 2.349 4,467 4.
Separate intake Structure 3,992 14.324 27,004 5.
Technical Specifications hfodifications 1,150 4,141 7,825 6.
Independent 11CP Seal Cooling 3,510 12,870 24.570 7.
Combination of Alternatives S and 6 4,063 14,821 28,211 NUllEG-1421 12
y i
Table 4.9 Bent.Pstimate Costs of Proposed Alternatises ($ Per Reactor)
Column Number 1
2 3
4 5
6 Include include Onsite include NRC Conseq.
Include include Tech.
Cost.
Offset Direct Indirect O&M Spee.
Total Net Alternatives Cost Cost Cost Cost Cost Cost 1.
No Action 2.
Additional Crosstic
$557K
$724K
$1,050K
$1,080K
$1,140K
$627K 3.
Electrical Cross Connection
$50K
$65K
$94K
$128K
$189K
-$246K 4.
Separate Intake Structurc
$29,000K
$38,000K
$55,000K
$55.000K
$55,100K
$52,300K
- 5. Technical Spec. Modifications
$0
$0
$0
$83K
$104K
-$684K 6.
liigh Pressure Pump for RCP Seals
$5,900K
$7,700K
$11,000K Sll,000K
$11,100K
$8,800K 6a. Fire Water for Thermal Barrier Cooling
$200K
$260K
$378K
$412K
$473K
-$ 1,900K Table 4.10 Direct Cost Estimates ($ Per Reactor) 1Aw liigh Alternatives Estimate liest Estimate Estimate 1.
No Action 2.
Additional Crosstie 250K
$57K 1,000K 3.
Electrical Cross-Connection 50K 50K 50K 4.
Separate Intake Structure 7,000K 29,000K 38,(X)0K 5.
Technical Specifications Modifications (see text) 0 0
0 6.
liigh-Pressure Pump for RCP Seals 1,000K 5,900K 15,000K 6a. Fire Water for Thermal Barrier Cooling 127K 200K 273K 43.2 indirect Costs was completed overnight (e.g., excluding the time costs of capital). To arrive at the total operating and maintenance The indirect costs are usually a certain frnction of the (O&M) cost, the annual value was integrated and dis-direct cost. As recommended in NUREG/CR-4627 (Ref.
counted over the remaining plant life (30 years). Alterna.
11),30% was tud (the range is from 25% to 33% for tive 5 (modify technical specifications) was assumed not to engineering and quality assurance costs for in-place struc-involve any O&M costs, Column 3 of Table 4.9 includes tures). Column 2 of Table 4.9 includes this cost compo-this cost component. In calculating O&M costs, a 5%
discount rate was assumed consistent with the NRC ree-nent.
ommended practice.
433 Operating und Maintenance Costs 43.4 Teclinical Specifications Costs Usually, these costs annually equal 3% of total "over-liach alternative involves modifying technical specifica-night" costs. Overnight costs represent the sum of total tions to a certain extent. Accordmg to N URl!G/CR-4627 direct and indirect costs, assuming that the modification (Ref.11) these costs are $18K per reactor for a simple 13 NURiiG-1421
l case and $35K per reactor for a complicated or controver-offset cost of each alternative. He cleanup and replace-sial one. It was assumed that each alternative will result in ment power costs were calculated as follows (Ref.13):
a simple technical snecification change. No choice in-cludes the cost of a public hearing.The fourth column of 1
I
-,u " ~#.m}
costs in Table 4.9 includes this component of cost.
r where: u integrated and discounted cost
=
43.5 NRC Costs C,
cost of cleanup ($100M/yr)
NRC costs include the development and implementation C,
cost of replacement power ($400K/ day)
=
costs. The development costs should be about $11K per reactor for a simple case and $21K per reactor for a discount rate (0.05/yr) r -
complicated one. No case includes the cost of a public hearing. He former figure was chosen here. Operating at remaining plant life (30 yr)
=
costs would be incurred after the resolution's implemen-duration of cleanup / power replacement m =
tation, and they would cover ensuring compliance with the new requirements. The operating costs have to be (10 ')
Y integrated and discounted, since they are recurring. He implementation and operating costs were estimated a!
Table 4.9 shows components of the total cost and the net
$50K per reactor. Total NRC costs would then be $11 A cost for the best-estimate case (the costs are 1.cr reactor).
+ 50K = $61K per reactor. Column 5 of Table 4.9 in.
The net cost is the total cost mmus the cost offset (from cludes the NRC costs. For a technical specification and Table 4.11). If the net cost is negative, the alternative is procedures change, the total NRC costs would be $21K cost beneficial regardless of the cost / benefit ratio. It per reactor (Ref.11).
should be noted that each column in l'able 4.9 subsumes the cost item m the previous column and includes an additional indicated cost component. For instance, Col-4.3.6 Averted Onsite Costs temn 2, ~1nclude Indirect Cost" includes the direct cost and the indirect costs of an alternative.
Averted onsite costs are taken into account as cost offsets (Ta' ' 4.!!) te the calculated cost of the proposed resola-4.3.7 Range of cost Estimates tion alWrnatives, consistent with NRC policy. Table 4.12 lists t!.e averted consequences. It can be seen that the Table 4.13 presents the range of estimates obtained for onsite personnel exposure per accident will be low, com-the total cost (corresponding to Column 5 of Table 4.9) pared to the offsite exposure and other onsite conse-and the net cost (corresponding to Column 6 of Table quences, so this component was not consid: red further.
4.9). The low values were calculated by taking the lowest he numbers are from NUREG/CR-3568 (Ref.13) ns estimates in the data of various cost components (mainly best-estimate numbers. Averted onsite exposure wculd direct costs)and carrying the computation through to the be added to the offsite person rem exposure as part of the final number. The high values were calculated by taking benefits, but the effect is negligibly small. For cleanup the highest estimates in the data of the various cost com-and replacement power, the integrated and discounted ponents and carrying the computation though to the final costs are then multiplied by the ACDP to arrive at the number.
Table 4.11 Cost Offsets for Proposed Alternatives ($ per Reactor)
Alternatives Cost Offset ($)
1.
No A: lion 2.
Additionci Crosstie 513K 3.
Electrical Cross-Connection 435K 4.
Separate Intake Structure 2,750K 5.
Technical Specilications Modifications 788K 6.
Independent RC? Seal Cooling 2,340K 7.
Combination of Alternatives 5 & 6 2,730K NUREG-1421 14 l
a Table 4.12 Onsite Consequences T)pe Amount Occupational Doses:
Imncliate:
1,0(X) Person Item long Tei,m 20,000 Person Item Total 21.000 Person Item x 30 yr x $1,000/p-r m - $6.311+ 0S yr lieplacement Power
$1.811 + 10 yr Cleanup
$1.211 + 10 yr Total Onsite Consequer ces
$3.Olia 10h'
this number to be multiplied by ACDF for each alternative.
Table 4.13 Itange of Estimates for the Total Cost and the Net Cost ($)
Total Cost Net Cost 1.ow Ilest liigh I.uw Ilest Iligh Alternatives I' stimate Estimate l' stimate Estimate Estimate Estimate l.
No Action 2.
Additional Crosstic 550K 1,140K 2,000K 37K 627K 1,500K 3.
lilectrical Cross-Connection 173K 189K 205K
-262K
-246K
-230K 4.
Separate intake Structure 14,000K 55,100K 72,000K ll.000K 52,300K 69,000K 5.
Technical Specifications Modtfications 48K 104K 171K
-740K
-684K
-617K 6.
liigh-Pressure Pump for 11CP Seal Cooling 2,000K 11.100K 29,0(X)K 1,200K 8,800K 28,200K 6a. Fire Water for 'lhermal llarrier CmW 318K 473K 624K
-2,000K
- 1,900K
.- 1,700 K 15 NLIltliG-1421
- 5. VALUE/ IMPACT ANALYSIS De value/ impact (V/I) methodology for analyzing the 5.3 Alternative 3-Provide Electrical various alternatives examined under this study is based on Power Cross-Connection the regarements of the backfit rule (10 CFR 50.109)and related implementing guidance contained in References 13,14, and 15. One of the primary considerations here is One of the observed contributors to the unavailability of the derivation of cost / benefit ratios for each alternative the ESW system is related to the s cliability of the electri.
evaluated in terms of cost in dollars per person rem cal power supply and control system. liased on the data in averted, which may be compared to a guideline such as Reference 3, the loss of the electrical power supply from
$ 1,000 per person rem.His quantitative guideline is one various causes was relatively high; however, the recovery of the elements considered in the decision-making proc.
times associated witn these events indicate a relatively ess. Deterministic considerations on the merits of a pro-fast average recovery observed during losses of the ESW posed alternative resolution are also a part of the decision system.
with respect to a given alternative (Section 6). In the folknving sections a description of each alternative and the results of a value/ impact assessment are presented, in general, the electrical power supplies to the ESW Table 5.1 summarizes the results of this assessment for trains are separated and have no cross connection capa-the various alternatives analyzed, bility, i.e., Train A IISW pump cannot be powered from electrical Train 11.This alternative therefore investigated the implementation of a cross connection between the 5.1 Alternative 1-No Action electrical trains of the unit with respect to the operation of the two ESW pumps (l' rains A and II). The cross-con-Under this alternative there would be no new regulatory nection of electrical power supply for other electrical requirements. Consistent with existing regulations, this components, such as motor-operated valves, was not con.
alternative does not preclude a licensee, or an applicant sidered as part of this because of their less significtmt for an operating licenee, from proposing to the NRC staff potential for risk contribution as observed in the opera.
design changes intended to enhance the reliability or op, tional data. It is envisioned t hat the electrical power cross-erability of the ESW system and its components on a connection would be an exclusively manual operation.
plant-specific basis. Table 5.1 summarizes the results of Ilowever, the possibility of adverse interactions between i
this assessment for the various alternatives analyzed.
electrical Trains A and il, such as the inadvertent transfer of faults from one train to the other, and hence the loss of both trains, make this alternative of questionable value, 5.2 Alternative 2-Install Additional Even if this contribution to possible adverse interactions Crosstie between trains is set aside, the CDF reduction is not significamt because of the relatively fast recovery ob.
8'
'" E"*"
De ESW systems of the seven multi unit sites analyzed under GI-130 are cross-connected through pipe connec-yatio, without takmg into account the potential adverse Weram ns Ma ahername, was caWaM Me W tions and isolation valves. His arrangement allows the per pers n rem, and if the averted onsite costs are taken operator of one unit to utilize the ESW cooling capacity of the other unit. In most cases, the crosstic isolation valves pto asun!, the net cost becomes negative, i,c., resulting in a net sanngs.
can be remotely operated. A hardware failure to open the isolation valves, should the need arise, could result in adverse conditions. A parallel caoss-connection could re-i duce the possibility of this kind of failure, and in addition-5.4 Aliernative 4-Provide SeI)arate would allow for more flexible maintenance options. The effects of the isolation valve failures on the CDF were not Intake Structure large because of the relatively low observed isolation valve failure rates, indicating that other hardware compo-A review of the failure nodes of the intake structure nents are more significant in reducing the overall system indicates that one of the observed liSW fatture me:ha.
unavailability. He core damage frequency reduction of nisms is the failure of certain intake com;xments such as this alternative was estimated to be 1.6 E-05/RY.
traveling screens or strainers. This type of failure within the intake structure stops or restricts the flow of cooling The cost / benefit ratio for this alternative was calculated water to the plant. A separate intake structure, hicated to be $433 per per son-rem, or 5238 per person-rem taking either on the same body of water or a different water into account averted onsite costs.
source, would make a backu p cooling capability available.
NUREG-1421 16
Table 5.1 Best-Estimate Cost /Denefit Itatios ($/ Person-Item)
Alternathen Total Cost / Benefit Net Cost /llenefit 1.
No Action 2.
Additional Crosstie 433 238 3.
tilectrical Cross-Connection 80 4.
Separate Intake Structure 3847 3651 5.
Technical Specifications htodtfications 25 6.
Ihgh-Pressure itCP Seal Cooling 862 6S4 6a. Fire Water for'1hermal llarrier Cooling 37 7.
Combination of 5 and 6 756 574 7a. Combination of 5 and 6a 39
' Including averted onsite costs results in a net cost savings.
The intake structure is usually a single structure divided
$3,651 per person-rem taking into account averted onsite into separate bays by concrete walls. 'lhete are a number costs.
of screens installed to prevent the intake blockage by large foreign objects.'lhe collapse or plugging of these 5.5 Alternative 5-Modif}> Technical screens may occur as a common mode failure because of the common inlet or common water source. 'the whole Specifications Requirements intake structure could also be affected by events such as Certain operating modes, Modes 5 and 6 (shutdown and flooding or freezing.
refueling modes respective'y), were examined with regard to specific i equirements in the technical specifications. In The alternative considered here is a completely separate these operating modes, the reactor is in shutdown condi, intake structure serving as a redundant intake source of tiou and the status of its ESW pumps is uncertain. 'lhe ESW. It may be kicated on the same water source, but in a technical specifications do not require that any of the separate location. An alternative design, which would liSW pumps be operationalin these modes. An implicit pluvide additional independence or diversity, would be to requirement is imposed on the liSW trains through the install the additional intake structure on a separate water explicit requirement to operate the residual heat removal source. Naturally, there are sites where this would not be system to remove decay heat.
fea.tibic, in essence, the operator of the unit in shutdown may utih7e the unit's own liSW pumps to provide the neces-The separate intake structure alternative includes the sary heat removal function but may just as well decide to structure, screens, and the associated motors, valves, and use the unit crosstics to supply ESW flow from the other piping. A swing ESW pump would be made available to unit. In the absence of any requirements on the !!SW either unit with redundant c!cctrical power supplies.'Ihis pumps, both pumps could be maintamed or made inoper-arrangement is intended to reduce the 1.?obability of two able at the same time. Although thts is not a universal failure mechanisms: one invohing electric, ' supply fail-practice, certain modelmg assumptions were made based urcs and the other involving operating failures of the n mformation g thered from plant sites representing a liSW pumps.'Ihc additional liSW pump would be a swing more conventional practice myolymg the admmistrative pump serving either unit depending on the needs of both control of crosstic use, and the f!SW pump maintenance units.This combination of a separate intake structure and schedule. In the basic analytical model it was assumed additional swing pump with redundant electrical power that the simultaneous shutdown of both I!SW pumps supplies would affect a large fraction of the initiating could occur only randomly.
event frequency related to the failure mechanisms involv-ing the intake, the ESW pumps, and their power supplies.
.!he unavailability of the Unit 2 ESW pumps to provide backup for the Unit 1 !!SW system may be reduced by The calculated reduction in CDF associated with this imposing an explicit operability requirement on at least alternative was 9.13E-05/itY 'the respective cost / bene-one of the liSW pu nps of Umt 2 while the latter is in fit ratio was calculated to be $3,847 per peison-rem, and Modes 5 and 6. An additional improvement is the testmg
{
17 NUlti!G-1421 l
1
of theunit to-unitcrossticvalvestoprovidegreaterassur-2.
Dedicated water storage tank with capacity to last at ance of operability. Also, this alternative includes addi.
least 8 to 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, tional credit for improvernents in emergency procedures for recovering from a 1 OSW accident. The resulting CDF 3.
Independent of ac powcr, nonseismic, diesel-driven calculations indicated that the CDF would be reduced by
- pump, 2.551kOS/RY, '!he respective cost / benefit ratio for this alternative was determined to be $25 per person.ren, 4.
No support system emling required, and and,if the averted onsite costs are taken into account, the net cost becomes negative, i.e., results in a net cost 5.
Once through itCP seat heat temoval.
- savings, Other design alternatives may also be considered that use arrangements different from the high pressure pump in-5.6 Alternative 6-Provide jection. One less costly alternative would provide flow Indepnident RCP Seal Cooling through the RCP thermal barrier heat exchangers by 3,3gg connecting the firewater system into the CCW lines. Most firewater systems have one diesel-driven firewater pump,
'lhe technical findings reported here in Section 4 and in which usually is independent of the liSW system.
Reference 3 indicate that the major contributor to the ESW-related component of CDF comes from the failure
'Ihe CDI. reduction for th.is alternative mvolving a high-of the RCP seals following a loss of USW. Specifically, the pressure seal cooling sptem was calculated to be RCP seal LOCA sequence contributes about 60% of the 7.8211-05. The respective cost / benefit ratio for this alter-total CDF attributable to liSW failures, llence, if the nathe inmlvmg a high-pressure seal cooling system was likelihood of a LOCA induced by an RCP seal failure may c leulated tobe $S62 perperson rem,or$684 perperson-be reduced, a proportionatcly sigmficant reduction in rem if the averted onsite costs are taken into account. Ihc CDF may be achieved.
cost / benefit ratio for this alternative involving a connec-tion to the fire water system for thermal barner cooling
'Ihis alternative provides for a dedicated seal cooling sys-was calculated to be $37 per person tem, or,if the averted tem that would continue to provide heat removal capalill-onsite costs were taken into account, this alternative ity after a loss-of-ESW event. The cooling requirements would result in a net cost savings.
of the RCP seals are relatively small, and a single small-capacity high pressure pump capable of delivering 50 to 5.7 Alternative 7-Combine 100 gpm was judged to be sufficient. The pump may be driven either by an electric motor or, for electrical inde-Alternatives 5 and 6 (Technical pendence from the point of view of other accident scenar-Speelfication CllangeS and ios (such as station blackout) a diesel driven pump option Independent RCP Seal Cooling) may also be considered.
As shown in Table 5.1, most of the analyzed alternatives The single high-pressure pump and diesel would provide have favorable cost / benefit ratios (pr esented as $/ person-flow via the cooling header to the four injection lines (one rem). In these cost / benefit calculv. dons, it was assumed to each RCP scal). It was assumed that the pump and that each of the alternatives (1 shrough 6a) was utilized diesel would not require auxiliary cooling for the lube oil, individually and independent;y from the other alterna-hearings, etc., as the suction flow or air cooling would be tives.
sufficient to provide all their heat removal requirements.
It was also assumed that the return flow from the RCP For the combination, the CDF reduction is calculated seals wot'Id not be recycled. In other words, a once.
w hen two alternatives are combined and utilized together through cooling cycle would be used with a water supply to reduce the risk from the loss of ESW function. The sufficient to last 8 to 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, alternative with the highest ACDF and favorable cost /
benefit ratio was ranked first and served as the starting It is assumed that a dedicated tank will be installed, with a point. This was Alternative 6 (or 6a), the dedicated cool-capacity satisfying s to 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> of seal cooling. After this ing system for the RCP seals. When the next alternative time, additional coohng could be provided by other avail.
was considered, the CDF reduction was calculated with able water supplies, such as the refueling water storage Alternative 6 (or 6a) already incorporated. The combined ta n k.
CDF reduction resulting from the implementation of Al-ternatives 5 and 6 was calculated to be 9.12 x 10 5/RY,and l
In modeling the system, the following assumptions were the respective cost / benefit ratio of $756 per person-rem, made:
or $574 per persm-rem with the averted onsite costs taken into account with a RCP seal cooling system involv-1.
Single high pressure pump,50 to 100 ppm capacity, ing a high pressure cooling system. '!he cost / benefit ratio NUREG-1421 18
for this combination of alternatives with an itCP thermal The results of the uncertainty analyses show a mean value intrict cooling system utilizing the fire water supply was of CDF from loss of E.crvice water of 1.49I1-4 per reactor-calculated to be $39 per person rem, and if the averted year, with a value of 596 and 95cc of 3.991! 5/ItY and onsite costs were taken into consideration a net gain 3.73E-04/l(Y, respectively.
would be achieved (i.e., a negative cost of implementa-U "I 5.9 Life Extension Considerations The regulatory proccas by which ;icense renewal may be 5.8 Uncerta,inty Analys,tS accomplished is currently under development by the NitC. It is envisioned that a license renewal for an addi-This section discusses the sources and treatment of uncer-tional term of 20) ears may be achievable tused on satisfy-tainty for the GI-130 study. Uncertainty is expressed as a ing specific requirements still to be established. llence, to quantitative bounding of the mean value. Uncertainty consider the ef fect oflicense renewal on the results of the arises from the selection of the data base used to deter-evaluation of GI-130,a reanalysis of the cost / benefit ratio mine parameter values, modeling assumptions, and com-parameters for each luckfit alternative was performed.
pleteness of the analysis.
The results of this reanalysis show that the benefits will increase by a factor of 1.67, while the costs, both inc arred and averted, willincrease by a factor of about 1.2 for most Although a complete analysis of all data uncertainties was f the backfit alternatives analyzed, not conducted, uncertainty studies were performed on selected issues that were important to the results. Unter-Table 5.2 summarizes the cos'/ benefit ratios based on a tainty cta were gathered, evaluated, and reported in the license renewal of 20 years or a remaining plant life of 50 form of distributions for these selected issues. 'Ihis data-years. A comparison of these numbers with those listed in gathering and reduction was used to gauge the effects of Table 5.1 shows that the cost / benefit ntios for all ana-the mdisidual data uncertainty on the final CDI results of lyzed backfit alternatives are considerably lower for ex-the analysts.
tended plant life of 50 years vis a vis a plant life of 30
'lhe primary areas of uncertainty exist in the determina-tion of the initiating frequency values, modeling, and Even though all alternatives listed in Tables 5.1 and 5.2 data. Each of these particular areas was addressed and the become more cost effective with life extension, Alterna-final result combines these issues to present the uncer-tive No. 4, Separate Intake Structure, still remains appre-tainty of the CDF. All other parameters were treated as ciably higher than the $1,000 per person-rem guideline at point estimates.
a cost / benefit ratio of $2,285 per person-rem.
Table 5.2 Ilest-Estimate Cost /llenefit Itatios ($/ Person-Item) for 20. Year 1.icense Itenewal Alternatives Total Cost /llenefit Net Cost /llenefit 1.
No Action 2.
Ad6tional Crosstie 271 133 3.
Electrical Cross-Connection 50 4.
Separate Intake Structure 2421 2285 5.
Technical Specifications Modifications 16 6.
liigh-Pressure ItCP Seal Cooling 541 412 6a. Fire Water for Thermal liarrier Cooling 23 7.
Ccmbmation of 5 and 6 474 343 7a. Combination of 5 and 6a 24
' Including averted onsite costs results in a net cost savings.
19 NUIEG-1421
- 6. DECISION RATIONALE
'lhis generic issue was identified as a consequence of the on the face of it, would indicate the need for additional 9yron Unit I evaluation with respect to its vulnerability to risk reduction, core-damage sequences in the absence of a crosstic from the !!SW of Unit 2. 'Ihis configuration existed because We have reviewed this aspect of our evaluation of GI-130 Unit 2 was under construction, and was eventually sup-and have concluded that additional improvements beyond plemented by the crosstic between units. 'lhere are 14 Alternative 7 cannot be justified at this time based on the units at 7 sites having 2 service water pumps per unit (1 following considerations.
per train) with a sharing of 1 pump between units via a crosstic between them, similar to the current crosstic in 6.1 llackiit Rule Considerations the 2 Ilyron units it was decided to focus the attention of this study on these seven two-unit sites because the design When the possibility of additional corrective measures of their IISW sysicm was expected to show the most (beyond Alternative 7) was considered, the resulting re-vulnerable configuration to nsk-significant sequences.
duction in CDF was either too small (i.e., approached
'Ihe remaining 1. Wits will be evaluated under GI-153, diminishing returns), or the cost / benefit ratio too high to "Inss of lissential Service Water in LWRs."
be consistent with toe backfit rule 'lhe examination for added corrective measures focused on those systems that A3 discussed in Chapter 5, most of the alternatives for are dependent on liSW and that performed a role in reducing the risk associated with this issue would be cost-several of the more dominant event sequences. l'or exam-effective in meeting the $1,000/ person rem guideline, plc, the alternative that recommended a design change to Furthermore, the objective of the GI-130 resolution is make the auxiliary feedwater system (Al'WS)independ-that the risk contributions from loss of the liSW system be ent of liSW cooling did produce a modest CDF reduction reduced consistent with the backfit rule's two basic re-(CDF was reduced from 6.lli-05/RY to 4RI-05/ltY).
quirements that Ihe improvement be both a sulstantial liven further reduction is theoretically possible by remov-increase in protection and cost effecure.
ing dependence on liSW of each system and component, one by one, until virtually complete independence is A combmation of potential improvements consisting of achieved. This is the ideal maximum reduction in vulner-the installation of a dedicated RCP seal cooling system ability to loss of service water; however, it is judged that and improvements in techracal specifications with respect going further in this generic plant cedculation is pressing to ESW system operation, including crosstic testing and the i;mits of precision beyond what iswarranted for plant-improvements in procedures, was shown to be capable of specific application to theoc 14 units. In addition, such an reduciag the total CDF by 60% (to 6.lli-05/RY) in a alternative (AFWS upgrade)would be applicable only to cost effective manner. Ilence, this is deemed to meet the some of the plant sites evaluated under thisi' sue; tbree of backfit rule, the seven sites are known to already have AIV systems independent of liSW cooling. In another case, Alterna-
"the overall approach to arriving at the proposed resolu-tive 4,involvmg the installation of a separate intake struc-tion considered both tne numerical results of the cost /
ture and a swing pump to be shared by the two units, was benefit analysis aad the spcctrur" and type of potential determined to be capable of providing a substantial risk improvements available for potential risk reduction for ieduction but was estimated to be not cost-effective.
loss of service water sequences. From the prevention perspective of loss of service water, it would iie desirable 6.2 Plant-S )ecific Considerations I
to choose those alternatives that could reduce the num-her of occurrences of the loss of service water imtiators-As part of the implementation phase of resolving this From the mitiltation perspective,it would be desirable to issue, we recommend that the licensets or applicants of choose those :.lternatives that would help to reduce the the 14 plants evaluated under GI-130 perform a review of consequencen of loss of service water.'lhe proposed reso-their respective plant-specific designs vis-a vis the recom-lution (Alter native 7) was selected to achieve some bal-mendationsof Alternative 7(combinationof Alternatives aneing of bi th these views that is, the improvement,in 5 and 6)and report, pursuant to 10 Cl R 50.54(f), whether technical specifications would assist on the prevention and how these recommendations would be implemented.
side, while the improved emergency procedures and backup seal cooling would provide a blend of both pre-
'!his licensee (or applicant) cffort would take mto consid-vention and mitigation capabihties.
cration the existing plant speedic design features that,in some cases, would be different from those assumed in the The HNL analysis (Ref. 3) shows that, after genene model used in the evaluation of this ;ssue. As a implementation of Alternative 7, there remains a residual result of this effort,it is expected that mdividual beensees component of CDFof 6.lli-05'R Y from FSW loss which, or appheants wdl submit a desenptum of the measures NURiiG-1421 20
1 i
taken as a result of the resolution of this generic issue, ESW (sediment, biofouling. corrosion, etc.). Ilence, considering producing at least a comparabic CDF reduc-there is no direct impact of GI-51 on GI-130.
tion as has been calculated for the Alternative 7 combina-G1-153,"less of Essential Service Water in 1 WRs" tion in the G1-130 generic calculations.
is under review and is expected to be assigned NRC
'!he results of some plant specific PRA evaluations re.
staff resources (Ref.19) for its resolution its ported by EPRI in Reference 16 supports the view that purpose is to assess this issue for all LWRs not plant specific designs incorporating features recom.
already covered by 01-130. Insights gained by the mended by the resolution of this generic issue would evaluation of GI-153 are expceted to be useful in result in significant reductions in CDF, l'or some plants, confirming or supplementing the technical findings of GI-130.
the licensee or applicant may find it desirable or necessary to propose other design icatures, such as providing AFWS cooling independent of ESW, to improve en the 6A Conclusion and Rationale assurance that the rtsk from loss of ESW will result m a small fraction of the total risk for their individual phnts.
On the basis of the considerations discussed ia items 6.1 through 6.3 above and the technical findings of this study, including the value/ mpact analysis of Section 5, it is con-i 6.3 RelationshiE to Other Generie cluded that the combination of Alternativc3 5 and 6, the Safety 1ssues modification of technical specifications and procedures a ng na n
an epen nt D seal A number of generic safety issues related to GI-130 are in ng W
n, am pppnate @ rehnm various stages of resolution, including some that have measures. lhes measures pravide a substantia! increase pircady been resolved.Their impact on GI-130 is as fol-m overall protection of the public health and safety and IU "
are cost effective.
o G1-23, " Reactor Coolant Pump Seal Failurcs" Of interc:st to the decision process on this generic issue
-This generic safety issue addresses the same possi-are the insights and views available in related pR A docu.
ble improvements as Alternative 6 and,in part, Al-mentation in the open literature.The pR A work available ternative 7 of GI-130.1he evaluation of GI-23 and in NUREG-1150, " Reactor Risk Reference Document" the Draft Regulatory Guide DG-1008, have been (plus supporting documentation)(Ref.10), is a source of issued for public comment (Ref. 5).
cxtensive information on risk analyses for an understand-ing of ESW vulnerabilitics. An examination of the An objective of GI-23 is to reduce the risk of severe NUREG-ll50 documentation of the three PWRs that accidents associated with RCP seal failure by reduc-were studied indicates that the ESW redundancy for two ing the probability of seal failures, thus making it a of the three l'WRs was considered large enough that a relatively small contributor to total CDF. GI-23 complete loss of ESW as an event initiator was deemed could entail the installation of a separate and inde-not credible (eight pumps available in Sequoyah, Units 1 pendent cooling system for the RCP seals. llence, and 2). None of the five plants in the NUREG-1150 study implementation of the GI-23 resolution could pro-is a GI-130 plarit: hewever, it is worthwhile to note that vide a substantial portion of the proposed GI-130 one of the PWRs (Zion) identified the service water con-resolution. As such, the proposed resolution of tribution to risk to be substantial (approximately GI-130 will be coordinated with the resolution of 1.5E-4/RY). This contribution for Zion was approx-GI-23 (see Section 7. Implementation)-
imately 42% of the total CDF from all causes.
o GI-51, " improving the Reliability of Open-Cycle Another PRA work available in the open literature is Service Water Systems" -This generic safety issue NSAC-148, " Service Water Systems and Nuclear Plant was resolved in August 1989 and its implementation Safety" (Ref.16). Although it is only a cornpilation of began with the issuance or Generic 1.etter 89-13 carher PRA results for six plants performed by the (Ref.17) and Supplement 1 (Ref.18). The GI-51 industry, it is useful to note that a greater appreciation of implementation entails a series of surveillance, con-the service water systern's contribution to plant risk has trol, and test recommendations to ensure that the moved the industry to initi:.te a program to improve ESW systems of all nuclear power plants meet appli-service water performance. The hmited guidance cable licensing guidelines.
available in NS AC-148 is a step in the right direction.The wide range of CDFs frorn loss of service water over the six During the review of the operational experience plants studied suggests large sariability in p ant speafic data for GI-130, credit was taken for corrective
!!SW configurations. 'the average CDF from loss of measures as a result of the GI-51 resolution by ex-service water for the six plants was 6.55E-05/RY, with a ciuding those events that involved fouling of the range of 2.33E-04/RY to " negligible" contnhution.
l 21 NURl!G-1121
. - --. -. - -.. - _. -. - - -.. _ - - ~. - - -
l l
r t
While many details of these six PRAs are not included in system provides an important safety function that could NSAC-148 and therefore must be considered only with be a substantial contributor to overall plant risk tends to great caution, the overall message that the senice ' 'er tend added credence to the 01-130 conclusions.
r 4
i p
I I
L e
l
- 7. IMPl.EMENTATION
'Ihe staff proposes to implemer.t the resolution of of public review and comment on draft llegulatory Guide Generic Issue 130 by i.ssuing a generic Icticr under DG 1008, the backup scal cooling portion of Alternatise 10 CI'll 50.54(f), to the licensees and applicants of the 7 (see Chaptcr 6) may be deferred. The reason 14 plants involved in this evaluation.1he content of the for allowing the deferral of this additional protection generic letter will address both the preventive and relates to the earlier development and promulgation i
mitigative aspects of the proposed resolution as discussed of 10 CI lt 50.63 (the station blackout rule), which was in Section 6.The implementation phase of Gencric Issue based on an assumption regardmg the magnitude of 11CP 130 will be closely coordinated with that of Generic issue seal leakage during a station blackout event. While it 23, which deals with the itCP seal reliability for both was left to GI-23 to validate that assumption, G1-130 is normal operation and accident conditions. Guidance on also based cn a seal l_GCA model very similar to the resolution of that generic issue is proposed in Draft 01-23, but different from the leakage assumption in llegulatory Guide DG 1008, While awaiting completion 10 CI lt f0.63.
23 NUltl!G--1421
- 8. REFERENCES 1.
Memontndum from T, P, Speis, NRC, to 11. L 10.
M. L Ernst et al, " Reactor Risk Reference Docu-Thompson, NRC, " Safety Evaluation Report Re-ment," NURIIG-ll50, Vol.1 (Draft for Com-lated to the LCO Relaxation Program for the Dyron ment), February 1987, Generating Station," January 15,1986.*
11.
fi. Claiborne et al, " Generic Cost Estimates,"
2.
N. Z. Cho et al., " Analysis of Allowed Outage NUREG/CR-4627. Revision 1, Science & Engi-Times at flyron Generating Station," NUREG/
neering Associates, February 1989 CR-4404, IINL-NUREG-51930, June 1986' 12.
" Complete CONCISE Printout for Model 148-PWR-ME (A3-ME) Energy liconomic Data 3.
P. Kohut et al.," Analysis of Risk Reduction Meas' llase (EEDil) Phase IX Update," Volume 2 of 9, urcs Applied to Shared Essential Senice Water I'lan' Set 2 of 3, 7697.900, DGE Report 871102, Systems at Multi-Unit Sites," NUREG/CR-5526,
- 1937, llNI,-NUREG-52225, May 1991, 13.
S.
W.
lleaocrlin et al.,
"A llandbook for 4.
P. Iam and E. Leeds,
- Operating Experience Feed-Value-Impact Assessment, NUR1iG/CR-3568, back Report-Senice Water System Failures and PNL-4646, December 1983.
Degradations," NUREG-1275, Vol 3 November 19SS.
14.
Memorandum from E. S. lleckjord, RES Office Letter No. 2 " Procedures for Obtaining Regula-5.
" Notice of Solicitation of Public Comments on Ge-tmy impact Analysis Review and Support," Novem-ber 18,1988.,
neric Issue 23, ' Reactor Coolant Pump Seal Fail-ure,' and Draft Regulatory Guide: Issuance, Avail-15.
Memorandum from E. S. Beckjord to Distribution, ability, Federal Register, Nol. 56, 56 FR 16130.
RES Office Letter No. 3," Procedure and G uidance A "I I9' for the Resolution of Generic ksues," May 10, P
1988.'
6.
Memorandum from Edward L
Jordan (CRGR:NRC) to Eric S. Beckjerd (RES:NRC),
16.
"Senice Water Systems and Nuclear Plant Safety,"
" Implementation of the Safety Goals," September NS AC-148, (prepared by Pickard, lowe and Gar-6, 1990.*
rick, Inc.), Nuclear Safety Analysis Center and Electric Power Research Institute, May 1990.
7.
" Safety Goals for the Operation of Nuclear Power Plants," Policy Statement, FederalRegister, Vol. 51, 17.
James G. Partlow, "Senice water System Problems 51 FRN 26044, August 4,1986.
Affecting Safety-Related liquipment," Generic Letter 89-13 July 18,1989.*
8.
T, A. Wheeler et al., "Aryalysts of Core Damage 18.
James G. Partlow," Service Water System problems Frequency from Internal Events: Expert Judgment Affecting Safety Related Equipment", Generic Elicitation," NUREG/CR-1550, Volume 2,
Letter 89-13, Supplement 1, Apnl 4.1990.
SANDS 6-20A April 1989.
19.
Picmarandum from K. Kniel, NRC, to C.11. Ader, 9.
" Reactor Safety Study: An Assessment of Risksin NRC, " Request for Prioritization of New Generic U.S.
Commercial Nuclear Power Plan ts "
Safety Issue 'Icss of lissential Senice Water in NUREG-75/014 (WAS11 1400), December 1975.
1/NRs'," May 2,1990.'
' Avaibble in the NR('1%bhc !)ocument Room un<ter Nt !RI Ai-1421.
NUREG-1421 24
. a
F3 c FORM 31s U S. NUCLEAR REGULATORY COMMISSION
- 1. REPOHT NUMBER
( Asstyred by I AC. Add Vol.
(F-89)
S@. Rev., and Acuendum Num-kJ4CM 1102.
t**-
" a'v l W.*
BIBLIOGRAPHIC DATA SHEET (s.s metruccons on ine re-sei NUREG-1421
- 2. TITLE AND SVttilTLL
- 3. D AIL HLPORT PUtSUSHLO Regulatory Analysis for the Resolution of Generic Issue 130: Essential Service uoNys 1
ytAn 1
Water System Failures at Multi Unit Sites June 1991
- a. eN on onANT Nuuacn
- b. Ault4R 6)
- 6. lYPt OF HLPOHf L
V. l.cung, D. Basdekas, G. Maretis Regulatory L PEReOO COVERED tircius+,e Daisal S. PERFORMING 08aAt42ATION - NAME AND ADORLSS (11 NHC, provice D' vision, Ottice or Regm. U. S. Nuc6 ear Regulatory CCurimass.on, and rfull6ng address; 41 contractor, prov6de name and mallmg a:$ dress.)
Division of Safety issues Resolution Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555
- 9. SPONSOHING ORGANZATION - NAML AND ADDRESS (if NRC, type 'Same as atove'; if Cont' actor, provce NHC Divism. Ottco or Hegior, U.S. Nn. clear Regutatory Commession. and maatng access,)
Same as above.
- 10. SUFPLEMLNT ARY NOTLS
- 11. ABSTRACT (200 worcs or less)
The essential service water system (ESWS)is required to provide cmling in nuclear power plants during nocmal op-eration and accident conditions. The ESWS typically supports component cooling water heat exchangers, contain-ment spray heat exchangers, high-pressure injec' ion pump oil coolers, emergency diesel generators, and auxiliary building ventilation coolers. Failure of the ESWS function could lead to severe consequences. ' Itis report presents the regulatory analysis for GI-130," Essential Senice Water System Failures at Multi Unit Sites." 'lhe risk reduc-tion estimates, cost / benefit analyses, and other insights gained during this effort have shown that implementation of the recommendations will significantly reduce risk and that these improvements are warranted in accordance with the backfit rule,10 CFR 50.109(a)(3).
13 ^vu ^esut y sT ATculNT
- 12. xcY wonosioEseneTOns tust wo.d. or ontasas snat un asnt researer.rs m tocatmo ene r. porta Unlimited R SECURITY CLASrOCADON generic issue
' 8 h N8')
service water system Unclassified regulatory analysis cIh,s peporo multi unit sites Unclassified crosstic is Num 0+ "Aue s shared safety systems
]
ie mL PAC FORM 335 (2-69)
?
THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER l
k UNITED STATES NUCLEAR REGULATORY COMMISSION EOcINIiUS,o WASHINGTON. D.C. 20555 nawa. o o OFFICIAL BUS:N;SS PEN ALTY FOR 0Fitv ATE USE S3W) l I IA'41AI19C1901
'J '
g ils J g!!,j g gnticatr m svcs r - 2 2 '.
WasHINGioti DC 2055c h
8 E
E
$s 5:
N.
Du '
gc d'
45 v.$
Y=
h:c
>0 m'
EF I
O 40 sv Qh Du i
b D
a
- - - _ _ - _