ML20206U694

Discusses Final Rept on PRA for EQ-TAP 5.b.Task 5.b of EQ Tap Closed

ML20206U694
Person / Time
Issue date:	04/04/1996
From:	Dummer A NRC (Affiliation Not Assigned)
To:	Marsh L NRC (Affiliation Not Assigned)
Shared Package
ML20206U672	List: ML20206U667 ML20206U680 ML20206U687 ML20206U694 ML20206U699 ML20206U705 ML20206U711 ML20206U717 ML20206U724 ML20206U729 ML20206U733 ML20206U743
References
FOIA-99-82 NUDOCS 9902170279
Download: ML20206U694 (16)

Text

-- - - _ _ ___ . _ _ _ ,

m......_

.VC L.ca n n mu u . s n i

.; g/ . '  ; .c 3 Wa8HINGfoM,o, 3044 M0t .

g ATT ACllMENT. 3 f Q Q'

' Aprill.

l n, ,

j H[MORANDUM TO: Ledyard D. Marsh, Chief N Plant Systems Branch .

i Olvision of Systems Safetyd..and' Analysis i George T. Hubbard, Chief pn l

THRU:

BOP Systems Section [/. '

Plant Systems Branch- M-  !

DivisionofSystems'SafetdndAnalysis

(

fROM: Ann H. Dummer .

BOP Systems Section ,

Plant Systems Branch ,

i Division of Systems Safety and Analysis .

j SUBJEC1. fINat. PEPORT ON PRA FOR E0-1AP 1ASK S.b '

f In 1992. LO bcgan to attract attention i P the :itense Renewal and

} Environmental Review Project Directorate staff presented their Draft Branch  :

j Techniul Position (81P) on E0 for License Renewal to the ACRS andl l

performed by Sandia National Laboratories (Sandia) resulted in failures or The B1P marginal insulation resistance for some types of qualified cables.

1 I

focused on the difference between qualification standards for older and ll newer plants (IEEE 323/1971 and 1974), advocating re-qualification or upgrade of a I

n l The staff was directed to evaluate t'e l

]

E0 components for license renewal. adequacy of E0. and the BTP f

' as an " operating reactor issue."

As for all issues EQ.

facing a potential backfit,' the staff pe Probabilistic Safety. Assessment Branch (SPSB) staff to be inadequate.

performed a preliminary risk scoping study which concluded that, if the  ;

I reliability of environmentally qualified components is reduced by the presence of a harsh environment encountered under accident (.onditiens, the proba of core damage would significantly increase. addition to the BTP a I l issue.

The scope of the preliminary risk scoping study was limited to core damage prevention, considering internal events only, and to in-containmen equipment, with emphasis on cables.The major objective of the preliminary Attachment 1 was to identify electric equipment that must function in accident-induced harsh environments and that could be major contributors to core damage.

Emphasis was placed on cables because cables are not routinely rep receive minimal maintenance, if any.

risk scoping analysis are (1) E0 failures could have significant risk impact if electric component reliabilities are reduced in the presence of a harsh environment, (2) the magnitude of the impact on core damage frequency is 9902170279 990211 PDR FOIA KOHN99-82 PDR ,

ioi vu v cu1 l l

2 Ledyard i' Marsh

'pecific. and (3) lack of reliability data bases and limit ations in cJrrent probabili. tic risk assessment models resultod'in significant unc these preliminary results.the staf f made recommendations for furthei evaluation of t EQ.

In parallel with the preliminary risk scoping study, the staff developed a task .ction plan on E0 which included a task item identified as a 'ftnal PRA."

Based on the preliminary risk scoping study, the1he stafmain f had enough intent of theconcern with the risk significance of E0 to pursue further work.

"further work" was to determine whether data existed that could be #

perform a more accurate PRA.

EQ, Argonne National Laboratory (Argonne)/,Cundet contract to the NRC, performed additional work on the risk impact of E0 and produce report in October 1993.

Attachment 2. The objective of this work was (a) to invnstigate the availability and extent of adequate reliability data for electrical equipment somponent*. .n harsh ensfronment5: (b) to in.m h,al* 01her potentia' sources of information on reliability data in harsn envirnnments, such as i.he results (c) io assess of testing programs sponsored by the government or indust ry: f whether this information can be used With'PRA technique! to par orn a defendable assessment of risk impacts associated with E0 issues; and (d) identify areas where more data are needed as well as approaches of obtaining such data (with special focus given to cable systems).

The literature review performed by Argonne showed that no reliability data bases esist for The the performance of electrical components in harsh investigation of other potential sources of related environrents. work on cable information included: test tests data from the THl-2 accicent:

performed at Sandia on cables, electrical systems sponsored by EPRI:

penetrations terminal blocks, pressure switches, pressure transmitters, and radiation monitors; and tests performed for NRC for. solenoid operated valves.

From the review presented in this work, Argoi.ne concluded that the available ,

informativa cannot be used with PRA techniques to perform a defendable <

assessment of risk impact associated with E0 issues btcause there is not enough data to obtain defendable failure rates for the different components cable systems. This conclusion is also supported by the insights of equipment qualification experts on the use of equipment qualification data to derive quantitat ive measurements of reliability, Experts interviewed as part of the Argonne work agreed that qualification data cannot be used for reliability because a limited number of samples was tested for qualification. The issue of reliability has not been addreAsed in current qualification practices. Although qualification tests are intended for design verification and prevention of common cause f ailures due to harsh environment, l such tests' focus on successful testing and failures were never reported in i Qualification reports. Therefore, no conclusion about reliability of cnmponents in accident conditions can be made using qualification data because the available data only reflect successful tests and do not reflect all test I c,>crience.

I 1

s

Ledya-d B. Mar -

Uhen the none,:.tence of EQ reliability date was_ confirmed by Argonne, the i

staff put PRA ..c,rk *on hold" until Brookhav W Hational Laboratory's (BNL) comprehensive 10 literature was c:mpleted? .The'$taf f hoped 'this literature l

review would oncover additional However, information' 'orddatasthat the BNL literature-review provided no could be used to continue work in PRA. a additional ins.ght into the use of PRA for L[p:ek .x,, 4 tocompletetheactionplanitemenPRA,thO,Uff{reviewedthepreviously discussed activities, with focus on Sandia's'1989' publication, NUREG/CR-5313.

"[quipment Qualification (E*)-Risk Scoping Study." This report was used as a reference in both the staff's preliminary risk: scoping study and Argonne'sThe purpose of study. A summary of this report is included in Attachment 3.

this study was to assess whether any historical EQ issues, related to the implementatien of the EQ rule, that were Thesubjects study providesof pattinsights NRC research and a appear to have any significant impact on risk.

systemat individual plants, ic f rwework for ;xamining equipment survivability issues fo important equi,iment that could be affected by accident environments.

(

in NUREG/ER m 13, the authors first developed a list of candidate risk significant < wipment that must function in accident-inituced harshthey Sbcond, environment s and whose failure would be risk significant.

identified time cnmponents (cables, solenoid operators, etc.) for which harsh environment reliabilities might differ substantially from the reliability These two values based on normal operation conditions! employed:in past PRAs.

l t

activities provided a qualitative data base to supportdrisk ' Riskanalysis importanceof the i

candidate equipment operations in a harshjenv,r.opment.;letely unreliable and:

analyses were performed by assessing the"linpact'ofitcomp perfectly reliable equipment on ri;k (risk achievement and risk reduction).

l This practices.

approach circumvented numerous Constraints imposed by on a plant-specific PRA to identify risk important EQ components for condition l

monitoring. such an evaluation could either include all systems modeled for

(' the plant-spec ific PRA or be limited to the components identified in NUREG,CR-5313.

In order to perform a quantitative risk assessment for EQ, it is necessary to have adequate reliability data for equipmert operations in harsh environments.

They all The staff, Argonne and DNL searched for such reliability data.

concluded that available information and data cannot adequately supportThe lack of quantitative risk assessments of EQ issues. fact that qual.ification tests were and no reliability testing has been performed for equipment in a harsh Sandia developed a method in NUREG/CR-5313 to evaluate risk environment.

significant equipment operations which could be used on a plant-specific basis, but plant-specific analysis is outside the scope of the EQ TAP risk activities. Therefore, no more work should be performed on PRA for the EQ-1AP.

I

" *='==F- =-.-w, a

+

1

' d u . Ca' & ? ' y;;;,,7+ . ,*,r.y9,,

j -c t+?qt, - ?. ,ye u qq, , ,. 4.; .y .. ,

~.+ + ;ys,:- . , .

- s '3(4, G.. c...;, W %, . ,' '. ,

Led /ard B. Marsh ,

.v.. .

Ilm insights from the This cle. Ink 5.b of the [Q 1ask Action-Plari.

prelimin:ir f risk scoping study, the /irgonne study, and huREG/CR-5313 should be reviewed for the. Status Review, lask.6cof.;,the.[0:.r s .

g. lask Action Plan.

I Attachmert*,: As stated a

1 5

l 1

)

3 i

a .

I I

i j

}

i i

1 i

,1 t

4 i

i l

a 4

4 h

1 M

d l

4 9

. ;! V? -

Oe" .;

. -s . e - .

i w

o AliA(nM[4' i I

" Risk Impact of Environmental Qualification Requirements for-Electrical )

Equipment at Operating Huclear Power Plants," .l g

. flicholas I . / Saltol.?SPSB/NRR: '

MarchQ0;&lHaQ;;(.; o,14 ' m, ,

'%44 :s n. -

(0NCtV51045

• Core damage frequency estimates for.both.PWR and BWR plants could increase significantly if electrical equipment reliabilities are. reduced due to the presence of a harsh environment.

important risk contributors could Current PRA perceptions regarding:iabilities are reduced due to the

' t hange if electrical equipment rel presence of a hart.h environment -

• The magnitude of core damage freQuenty impc.'t is plant specific.

• Due to the lack of reliability data bases and the limitations in current DRA models, an accurate assessment of the risk associated with harsh s'avironments is not possible at this time.

LPLUF IC. [0VlPMENT IDENTIFIED Electrical components that support risk signifi. cant operationsmin. harsh ,

^

environments: 4 .. , .

• Solenoid and motor operators.inside containment <

.- to ' .

• Steam generator level detection'circuitsfin PWRs These devices / systems, including cables, connectors, penetrations, and transmitters are susceptible to thermal degradation of electronics and age degr dation of seals with subsequent moisture. intrusion.

IMPOR} ANT SE0VE'l[13 Risk-important core damage sequences and related in-containment components .

facing harsh environments were identified.

PWRs

• Large anc medism LOCAs affect MOVs required to open 15-18 .'.surs into the accident to provide hot leg recirculation.

• Small and transient-induced LOCAs affect SG level detectors which impact  :

AfW operation and affect PORV solenoid and block valve operators:

therefore contributing to failure of " feed and bleed" operation after AfW-failure.

t L __

.'-- jff,kkfi "~

'Y ,

i

,,:p -u: -r..u.u c - . .;

' ~ . d,* (

l l

O

.D i

,j%%.' t IM- ,.

i

• liitt-rmodlate and small LOCAs followed by random fat!:ce rf 7 : 4Hect l SRV solenoid operators. l 1ransient with loss of suppression pool cocling (tw u quence) a ffects j hMV and M51V solenoid operators and M51V t>ypas', valve. rnator operators. I 4

'l' war. subsequently found by Argonne (see Argonne Draft letter Report j

" identification. Characterization, and Evaluation of Risk Important Accident l

Scenasi:,, Related to EQ lssues." Decembor, 1993) that the-TW sequence may not be as significant when credit-is taken for high pressure injection with a 7 cnr.teni rod drive (CRD) pump, i a

l i

i

\

t i

i f

A J

4

ate W +

Al l J is'*i', t :

" Collection and Evaluation of Existing Reliability Data-for' Electrical I Equipment Performance in a Harsh Environment - A Litorature Review."

fl. A. Hanan;ind C. P,..Innos ";

Argonne'Hattonal'taborAlory l0ctober,;1992-The Inis document summarizes E0 tect' rest',t s for a variety uf components.

f9RJ work on cables; l

l authors reviewed reports on lHi 2 equipment' performance; Sandia tests of cables, penetrations, terminal blocks pressure switches.

in Research Center pressore transmitters. and radiation monitors; and franti .

testinq of solennid operated valves.

l f ron> the TMI-2 work. Argonne concluded that most coulpmant failures were predominantly a result of moisture intrusion that generally occurred at the cler'cical penetration to a device. ,

i1cotified the following Based un review of previous test resulti. tFe autho'" is 4 result of an 10 i mrponents that could have increased f ulvo 1 ' -

.irti9nt ha.sh environment:

+ ables

- electrical penetration assemblies

• pressure switches ..;;:,.1( p p

+ solenoid operated valves Appendir i summarized expert opinion on the subject of elicitation for severe environment reliability of equipment.. The experts interviewed agrend that qualification data cannot be used for reliability because a limited number of sampics was tested for qualification. The issue of reliability has not been addiessed in current 4.alification practices. Qualification tests are l due to intended for design verification and prevention of common cause f ai ure Since qualification focused on successful testing and h ..h environment.

doromontation of the equipment, failures during qualification testing were Therefore, no conclusion atiout neve, reported in qualification reports. ,

rel ability of components in accident conditions can be made using '

qualification data because the available data only reflect successful tests ani! 10 not reflect all test experience.

i

U ! At.ilMf fl' :

HU4fG/CR-5313, "Eauipment Qualific.ation (EO).Rist scoping study" L . D. ':as t a rd e t al.

n<

4

% nd i a .'ta'. i nn a l L abo rM -r !j

.lanuary 1989

}

MTpfpf the authen- first developed a list of candidate risk significant equipment j that must f unction in accident induced harsh environments and whose failure would be risk significant. $0cond, they ddent(fled those components (cables.

solenoid operators. etc.) for which harth'envir'onment rollabilities might dif f er sute.tantially from the reliabilltf valtiot based on normal operation condition. nmployed in past PRAs. These two activities provided a qualitativo 4

data base to justify harsh environment parametric' risk M.hievement analysis -

for the c,unildate equipment operations. This project approach circumvented j

numerne nstraints imposed by currnnt (0 and PRA prat tices.

1 l

Historically, equipment qualification hu been 6anterned with loss of coolant _'l accidents (tJCAs), main steam line broso. (M$ lim) and high energy line breaks l l

(H[LBs). vRAs, on the other hand, aro concernbd with 6 much more diverse sot of accident sequenceh.

, Current fu calculations assume early loss of equipment function during the accident Iquipment failure due to harsh environments may not be immediate, allowing .ome initial core cooling. For exampic, in both the TW BWR sequence 4

(transient with loss of suppression pool cooling) and the transient-induced or small br,ak LOCA PWR sequences equipment reliability data for steam, l

t empera t ur.. . and pressure conditions would be useful for PRA, but early However, l

radiation accident conditions are not assumed for these sequences, j much of the available [ qualification) data that characterizes the accident t

performance for solenoid and motor operators includes a 200 Mrad radiation exposure prior to simulating the steam / temperature COntillions, i

G.NQ %Lon>

1

- from a PRA perspective, the EQ issue of whether correct accident Also not f

' acceleration techniques have been used is not risk significant.

i risk significant are: simultaneous versus sequential accident simulations, presence of oxygen in the test chamber, simulation of beta with gamma.

l a from a PRA perspective, much safety-related equipment inside, containment need not function in an accident radiation environment; the remaining equipment needs to function during radiation conditions for a few hours to a few days.

• Hydronen control equipment inside contair. ment would be exposed to risk

' signi f ic ant radiation conditions substantially less than rurrently em;)1oyed during [Q t est ing.

iisa mmii mimam

I - . . . . . . , ,

Q'py:j- ^u

l 4 ,

N6

• "no- I f the ieportance of the acc{ m it rarflation dose is (m ' ~phas12ed. The current overcon$crvatism in assumed radiation during 10 'o. ting may actually impact risk adversely (see SG level detector enrale).

• Ior 'te steam generator level transmitters the Authors nnte that PRAs do not udel the transmitter harsh environment reliability in determining the auxiliary feedwater system reliability. Degraded transmitter o-rings may produce a common-cause susceptibility to moisture degradation.

The folle, tng issues have potenttal appliCal.ility to safety-related Components with potn.tial risk significance. Addi,tional investigation may by warranted:

1. Adequate scaling / protection of safety-related circuits from moisture int naion/ condensation ef fects.

2. Choue of ambient environments as a'hasit for equipment qualification.

3. Humidity agir; affects.

4. Choue of accident envirenments frer equipment qualification.

5. Use of alternative test approaches ;.tultiple sample reliability assn'. ment and/or fragility testing) to complement current qualification type testing.

6. An nierated aging methods f or nuclear station batteries.

EELU99 WOW l

lhe authors evaluated a number of previous studies on PRA and/or E0 and developed the following " Historical PRA insights:" l

\

NUREG 1150. " Reactor Risk Reference Document," draft February 1987 The intersection of an elevated environment and a significant risk reduction Interval or risk increase interval provides the possibility of a significant E0 Impact on pRA results. Systems designed to mitigate consequences of an 4

accident (i.e.. containment integrity fystems and fission product scrubbing mechanisms) which are of less importante to core damage potential than risk

' mit igat ion are the very systems which could see the most severe environmental conditions after the onset of core damage. Calculations based on single component failures do not provide a complete perspective regarding the d

potential for common-cause equipmont failure to impact risk.

NUREG/CR 4]44, "Importance Ranking Based on Ag hg Considerations" Components within safety systems that had many moving parts were the most risk signifuant components susceptible to aging effects. These included NOVs in the emergency injection systems and sources of emergency power and signal processing. l NSAC/36. "Importance Ranking in Equipment Qualification" l This industry document concluded that actions such as plant modifications or operating procedure changes may be a more cost effective method for achieving risk reduction than proving qualification for safety-systems of an older plant.

I

!1

4

~3p@ @.2 #

+

HUREG/CR-)H2, " Identification of EquipmJnt:and. components Predicted as

) Significant Contributors to Core Damage'

- Ihe purpo:,e of this study was Lo' identify) components, predicted to Equipment be-significant contributors to dominant severe; accident 7 sequences.

' identified by this screening process indudes power operated relief valves (PORVs), motor-operated valves (MOVs), solonoid operated valves (SOVs), main steam isolation valves (MSIVs), electrical cables, connectors, and limit

' switchos.

I NUREG/CR-4537, " Electrical Equipment Performance in Accidents" lhls study as designed to determine the performance of safety The results of therelated study

equipment under severe accident conditions.

' indicated that safety relief valve actuation assemblies and main steam Isolation valve solenoid control assemblies are' risk significant equipment items with the potential of seeing environmental stre n in excess of qualifIcat1on levels, SPLClf IC i WIPMENI ANDf COUfNCC) ele e. fscer.al.p.t 1 m 1 Deterders and PORVs and AstgLI.tif.d Block Valves NRC requirements that [Q testing be based on " instantaneous release" of part

] of the enre inventory have potentially increased the feedwater-related scram i

" rates for PWRs. The impact of this HRLrequirement.on PWR feedwater scram rate shnold be further investigated.to provide a basis for possible j

modificatinn of the NRC requirement.

l

! Steam gmmrator level transmitters are an example of aquipmont currently qualified for accident radiation conditions even though PRA does not model their use after core melt. From a PRA perspective if auxiliary feedwater to J

' the' steam generators is being used to maintain core cooling, then the steam generators must boil dry prior to core melt. Thus the transmitters are not needed after core melt when the accident radiation conditions would occur, for many plants these same steam generator level detectors are used to ihe setpoints for initiate a feedwater trip during normal operation.

al,owable variation in the steam generator level during normal operation are partially based on the accident accuracy of the steam generator level

! transmitters. Manufacturers' tests indicate that the dominant contribution For to

! transmitter inaccuracy is caused by the accident radiation exposure.

example, Barton specified for its transmitters a 3X inaccuracy during steam / thermal exposures but a 10% inaccuracy during radiation exposures.

Hence. NkC's instantaneous radiation release requirement currently controls feodwater trio setpoints. [liminating this regulatory requirement might allow greater icvel. licxiallity during alant operatinn with respect to steam gene events (nr PRA accident sequences.

small break steam and temperature 1he primary environments of concern art condittons, Chemical sprays, radiation, and hydrogen burn conditions are not applicable from a PRA perspective since they occur after initiation of core damage. Poor scaling against moisture intrusion would have to be considered a

I' m;n 4 e, .

. s. 4 4 ,. 4 '

mp, ~

Isr m ry concern. 50V and MOVs would only b e W h9d ta fonctIon for'a few h er', of high dose rate exposures. Confirming cah uht it.nt regarding actual eco:pment radiation exposure conditions may be w r4nt M .

EWp SRVs aj,d MSIV1 In the TW sequence (transient with loss of suppression pool cooling). J containment cooling can be restored by opening.the MS!Vs in the harsh cnvironment caused by high suppression pool temperaturos and pressures. )

fnvironmentally-induced failure of>the H5IVs: leading to .fallure to restore the pou r conversion system is currently'not modelled by PRA analysis.- [ven though the environments surrounding the HSIVs have WortenPd as the IW sequence progresses, PRAs currently calculate that recovery becomes more probable (i.e., PRAs currently do not account

• for< environmental. f ailures of the MSlVs),

Of primary concern for the HSIVs would be the solenoid operators, the limit 1 switches, the cabling, the electrical connections and seals, the electrical penitrations, and the motor operated bypass valve operator and controls.

10 obtain a bounding estimate for the potential' impart of MSiv failure on cora damage frequency, the authors assumedinon-recovery of the power conversion sy' tem for those sequences which could produce a harsh environment in contefament, this resulted in, at most,'a fartor nf two increase in base case core damage frequency in the plants analyzed.

Of concern for the SRVs would be the solenoid operators, the cabling, the ennnections, the electrical penetrations, and the valvo position indication lhere is a potential impact on equipment operability from humidity effects, connection interfaces, and seniing techniques, In addition, there is devices, evidenra that accident equipment reli4bility differs from normal operation equipment reliability. Therefore, R[LIABill1Y testing of solenoids under accident conditions may be worthwhile. '

llPLl. RCIC. and HPCS oumn PRAs in BWRs, the HPCI, RCIC, and HPCS pumps are located outside containment.

cur.ently assume f ailure of these. pumps when suppression pool temperatures en ced 210 240'f ,

for BWR/4s, for non-station blackout sequences, operability of HPCl/RCIC pumps during high suppression pool temperature conditions is only risk significant if the harsh suppression pool conditions concurrently impact the reliability of the HSIVs and the SRVs. Core' damage frequency contributions from TW sequences (transient with loss of, suppression pool cooling) can become i significant and substantially alter current PRA perceptions regarding the risk I significance of the TW sequence when all these systems (IIPCl/RCIC, SRV and MSlV) are af fected by common-cause high suppression pool / containment i temperatures.

l The current design basis fQ requirements for HPCI or HPCS pumps generally lack PRA significance. Demonstration that HPCI, RCIC, and HPCS pumps would be reliable when they pump high temperature water from the suppression pool I _ _ _ - _ - _ - _ _ _ _ _ - - - _ _

i

.o .

e y ~, .# .

j

?* ~ -,

,p fwater tr+:oraturns up to 385'I) Coulb tubstantially rnduru core damago irnquenti"sforsomeplanti, P.yR Conta.Lnfgst fans . . .; , ,

i The authors recommend addittonal PRA analytes to determine whether PWR containment fan operation is sufficiently. risk significant to warrant substantial equipment qualification attention.. Lf an Coolers are assumed to f all either because of cable f ailure in the harsh radlallon And Steam condition. or because of Clogging by core debr{s an(f Aerosoll, If fan operation la deemed to be tufficiently ritk significant, then consideration of 4

humidity aqing effects and appropriate lubrication anti maintenance activilles -

may ho appropriato, INR_MKLPWR. 319tLB AllgdA d I a t i on tip n i t _o r s barrior via an increased signal for radiAtton levels.wlthln lho Cont,ainm available to the

high rani,o radiation monitor is one if the few signaitContainment area radiation

- operator to indicate loss of ceramic integrity.

?

levels of the order IL43 R/h or greator would suggest.that '"

the fuel ceramic barrier u being lost via core melting. , ,

The largest risk impact due to operation of tilgh' Range Radiation Monitors is for the early' release accident sequences." for ently release scenarios.

containment failure follows core molt by at most a foW hours. There does not j

appear to be a risk basis for requiring a factor of two accuracy for.the radia'. ion monitor in the full range of 1 R/h to l[d R/h. Radiation monitor signals in the low ranges are below those that indicato significant core molt

' and hence would Itkely not initiate emergency response decision making.

e i

I n.+

i

i ,

t ody a ri: 51. Marsh *$*

i lhii (b.,es lask S.b of the E0 lask Action Plan, lhe insights from the i prellmoury risk scoping study, the Argonne study, and NUR(G/CR 5313 should be caviewN f or the status Reviewi Ink 6 of the (Q 146L Action Plan, 3

Attachment 5: As 51ated d b f': ^

cye+ ,

' ,3t' -

;;mps .m. + , ,

kUt; , - ' '

Uil{}, ? ,

4 1

4, ,

l 4

l l

j 0151 Rittui 10N , SPLB LQ I110 CGratton l J1atum PShomanskt

! N$ altos 1 Gilolahan Alhadant ,

JVora, RES .

3 , ; 4,g.m ,

SAggarwal. RES ..; .j ;- ,

DOCUM(Ni NAM [: G:\E0\PRANOTES

• sce i.

previous concurrence

.r.iv, . e w .. .... n.c ni, wie... in in, w. c.c.n uje su ech-nvenei.. . . con, with .n.cha ninntiawre = . . cop Or r Ir.I SPLlt:0SSA SC:SPLD:DSSA SCtSPSStDSSA NAMl ADummer* Cllubbard MIack' DAlf 04/03/96 9///96 d

04/03/96 __

OfflCIAL RETORD COPY i i sm

( __

e ~.

p:..

' :(<:, '

, bu y, ,<ay n ;j;ygaqg p. p -

ggc. y

' t,5" ,.,

t

^'y. '- Q h.' h ' - ,

'{ -

Lerlyard B. Marsh , .

This closes Task 5.b of the EQ Task Action Plan. The insights from the preliminary risk scoping study..the Argonne' study, and NUREG/CR-5313 should be reviewed for the Status Review, Task 6 of the EQ Task Action Plan.

Attachments: As stated l

DISlRIBUTION: SPLB EQ File ,,, .e

• ~

CGratton ,

JTatum l PShemanski l

NSaltos i GHolahan AThadant JVora, RES SAggarwal, RES DOCUMtNT NAME: G:\EQ\PRANOTES o = We ce$

see previous concurrence indicate in the bon C= Cow w/o attachment / enclosure DCopy with attachment / enclosure W n t o r ec e 6 ve e c opy of this docuwat ,

I SC:SPLB:DSSA . SC:SPSB:DSSA Off!CE SPLB:DSSA GHubbard* JF1ack*

NAME ADummer*

04/04/96 04/03/96 DATE 04/03/96 Off101AL RECORD COPY

~ _ _ _ _ _ _ - _ _ _ _

4

,y .

ATTACHMENT 4 POSSIBLE USES OF PROBABILISTIC RISK. ASSESSMENT TO GAIN RISK INSJGHTS' ENVIRONMENT L IFICATION BYi, .

1. id .

GARETH PARRY DIVISION OF SYSTEMS SAFETY AND ANALYSIS SENIOR LEVEL' ADVISOR ON PROBABILISTIC. SAFETY _ ASSESSMENT There are two potential uses of probabilistic risk assessment (PRA) to gain risk insights in to environmental qualification (EQ). One use of PRA would be to use it to categorize and prioritize equipment with respect to its being on a condition monitoring list. While all the potentially vulnerable equipment may not typically be modeled in the PRA directly (instrumentation for example), it would be relatively easy to find the right " hooks" (basic events or gates) in the model with which to associate the relevant equipment. For this purpose, the aim would be to identify equipment that has aged or deteriorated so that it no longer would survive the environment for which it is qualified, in this case the categorization or ranking would be performed using the scenarios for which the equipment is required rather than those for

(,

which the E0 stresses are exceeded. ,

Another use is to use the PRA to explore:the" appropriateness of the EQ design basis accident scenarios. To do this'. requires using the PRA to identify whether there are realistic scenarios that can result in the environment becoming harsher than that assumed in the EQ tests. Or, put another way, are the EQ test design basis accidents bounding in terms of the environmental stresses expected in the range of scenarios identified by the PRA. One '

example of this would be concerns associated with flooding, i.e., equipment designed for steam and high temperature environments becoming submerged  ;

because, for example, of containment ' flooding in BWRs. To address this issue

! it would be necessary to identify whether there were any sequences in which EQ

- equipment were to become submerged, and whether any credit is taken for their operability.

There are two issues to look at; the role of the equipment in preventing core damage, and the role of the equipment in determining the containment failure probability and therefore the radionuclide release and public risk. The flooding example above is likely to be more of an issue for risk than core damage, because it only becomes a concern when a lot of water is put into containment, either directly (containment flooding in a BWR) or indirectly (for a PWR as an example, through a hole in.the RCS, followed by failure to go i to sump recirculation, necessitating maintenance of cooling by refilling the RWST). i i

11 is likely to be the case that it would not be possible to use many PRAs directly without modification. Many BWR PRAs do not model containment flooding as a core damage prevention strategy but they may include.it in the l

1 i

. &J ' 4. ? i F' .

- ,' *2/ NNf;,, ) * *

. m w., , v ,

Gus C. tainas 2 ,

tevel 2 portion of the PRA. 'Also,' for PWRs,'there are few examples in which the PRAs model RWST refill as a prevention strategy. These are typically very low frequency scenarios, and, in addition, it.is likely that the impact on risk may be minimal since they take aflong time to develop, and are probably  ;

t not LERF contributors,

~

it should be noted further that, for the assessment to be appropriate, the PRA model should reficct the way'the plant yould be operated in accordance with , ,;

the E0Ps, and standard plant practices.f:In~PRA a particular success strategy .&

u!

istypicallyadoptedforrepresentationl41nithe;eventitree.vSo,'for-some.

Westinghouse PWRs PRAs,' for small LOCAsrtthe success- path is that the .,

operators cooldown and depressurize the; reactor-(using AFW as the heat sink) .

to establish RHR, and minimize the break flow. .Many PRAs, however, model the  !

sequence of events as continuation of SI until the RWST is depleted, followed i by switch over to high pressure recirculation, and heat removal via the RHR heat exchangers. In either case, the AFW would not be required for more than a few hours, and EQ concerns about the: steam. generator level sensors may be minimal. # .  ;

j l

l' g, '-

.o 9