ML20055C955

Informs Commission of Recommended Resolution Re Need for Primary Sys High Capacity on C-E Plants W/O PORVs

ML20055C955
Person / Time
Site:	Palo Verde, Waterford, San Onofre
Issue date:	06/28/1990
From:	Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO)
To:
References
REF-GTECI-084, REF-GTECI-A-45, REF-GTECI-DC, REF-GTECI-NI, TASK-084, TASK-84, TASK-A-45, TASK-OR, TASK-PII, TASK-SE SECY-90-232, NUDOCS 9007020274
Download: ML20055C955 (21)

Text

fjate '

%c y v/f*/ n e n n n a n o n 'initiga n **

POLICY ISSUE June 28, 1990 SECY-90-232 (Inf0Mation) _

EgI: The Commissioners Eggg James M. Taylor Executive Director for Operations Subiect: EVALUATION OF THE NEED FOR PRIMARY SYSTEM HIGH CAPACITY MANUAL VENTING CAPABILITY ON COMBUSTION ENGINEERING (CE) PLANTS WITHOUT PORVs (GI-84)

Purpose:

This paper is to inform the Commission of the staff's recommended resolution of the subject issue regarding the installation of manual high capacity primary system venting capability, and its relation to the Backfit Rule.

Summary: There are six nuclear power plants (Palo Verde 1, 2, 3, San Onofre 2, 3, and Waterford), all of which have Combustion Engineering Nuclear Steam Supply Systems, that have been built and are -

operating without Power Operated Relief Valves (PORVs) or any other primary system manual high t capacity venting capability. Generic Safety Issue 84 was established to determine if such -

high capacity manual venting capability should be required for these six plants.

O(VI Rg As part of the staff's resolution of this issue, conventional cost-benefit analyses were performed and they show that the risk reduction this capability would provide does not outweigh the costs associated with its installation.

However, the resolution process has also inentified a number of qualitative factors which support the installation of this capability.

Since the cost-benefit aspects associated with this issue do not satisfy the tests of the Backfit Rule, the staff is not recommending installation of high capacity manual primary  !

-system venting capability on the subject plants.

Contactf x H /$. Woods , RES 't40TE : TO BE MADE PUBLICLY AVAILABLE I92-3908

\ IN 10 WORKING DAYS FROM THE DIsTE OF THIS PAPER U Qh

/ /{QQ f ________sL____ - -

~

• :- r ;es j m J 4 v '

r- j

~ .

.)

The Commissibners 2  ;

Bowever, consistent with Mr. Chilk's-Gune 24, ,

iets memorandum to Mr. stelle - (M840602), the staff,is bringingLthis matter to the attention of the Commission.

Backgroundt, In the.early 1980's, CELsubmitted its.SystenL40 6 design for staff review. During the course of the review, the staff. questioned theLomission of ,

Power ' Operated Relief Valves - (PORVs) . .f rom . the design. CE maintained that with theLlarger  ;

primary system and pressuriser volume of the~

systen 80 design, the safety valves would not be:

challenged during design basis overpressure events, and hence PORVs were_not needed to

, prevent safety valve challenges (the_ usual L reason PORVs were installed).

I L In Wertinghouse-designed plants undergoing R p licensing at that time, PORVs were being relied u:- upon in steam generator tube rupture (SGTR) safety analyses to perform the safety-related ,

function of depressurising the primary system to equalise primary-to-secondary pressure, thus stopping the primary-to-secondary leak and-keeping calculated offsite doses below 10CFR100 guideline values. In the design of.. San Onofre Units 2 and 3 and Waterford,-this depressurizing function:(for the SGTR safety analysis) was-performed by a safety-related auxiliary -

pressuriser spray system, and it was,later shown that no high capacity manual primary systen venting / capability was needed for the SGTR-analysis.to meet regulatory limits. For CE's l old systen to design and for the design of Palo L Verde Units 1, 2, and 3, this depressurization l function (for the SGTR safety' analysis) was performed by the safety-related reactor coolant systen vents.- [It should be noted that these vents consist of two redundant-one-inch lines (with one-half-inch orifices) from the pressure vessel head, which are not of high'enough capacity torprovide the feed-and-bleed cooling capability discussed in this paper. These vents were provided as a result of TMI action item II.B.1.] The results of CE's analysis demonstrated that no high capacity manual venting capability was Leaded to meet regulatory requirements. Hence, CE maintained that PORVs were not required to meet any NRC regulations.

However, the staff remained concerned that without PORVs, the plant would have no diverse

4* 0 -' i

~

, .l.. ,

j , a The commissioners' 3

< ,c means-for removing decay heat directly>froa the <

primary system via the feed-and-bleed method. 1

m. Feed-and-bleed, which involves injecting cool

< water into the oore'via the high pressure injection pumps and' removing stoaa via the PORVs, was noted as an important means-of decay j

4 heat. removal in the event steam generator .)

cooling is lost. The staff performed ]

assessments of feed-and-bleed capabilities in- y all1 operating U.S. PWRs, and has found a H

, substantial variation in such capabilities l depending upon installed equipment, operator j

,' training, and the' existence of proper- i procedures. However, the' staff has concluded  ;

it that all plants, with the exception of'the six  ;

L' CE plants without PORVs, have some' capability :l

, for feed-and-bleed cooling due to the existence i p of PORVs or other primary system manual high -

capacity venting capability in their designs.

.] '

i Both the staff,and the industry performed j probabilistic analyses on the relative risk reduction that could be afforded by installing i PORVs on the system 80 design. The staff results showed that there was a positive core melt frequency reduction of 2E-05'per reactor per year by installing PORVs, whereas the

, industry results:showed a small negative effect ,

on core melt frequency due to installing PORVs. l Differences were attributed primarily.to I j assumptions on operator response to initiate cooling.

Based on studies. documented in NUREG-1044 L (December, 1984) the staff concluded that PORVs l should be required on the six CE plants.

However, at that time,'the staff was considering the need for generic decay heat removal ,

improvements such as a dedicated decay heat L removal system for all wperating LWRs (USI A-45,

" Decay. Heat Removal Requirements"). Therefore,

=

the issue of requiring PORVs on the CE Systen 80 P design was deferred (as Generic Issue 84) ll pending the resolution of USI A-45.

In 1988, the staff resolved USI A-45. The conclusion reached was that the uniqueness of l- plant designs, coupled with cost-benefit considerations, did not allow for any specific generic improvements (i.e., a dedicated system could not be justified). However, the insights gained from limited scope PRA's on six plants i

-*- ,. .m w -- - -

W." w.

J

, En  !

+ :s

^);

• J '

The Commissioners 4 '

x s supporting the A-45 resolution showed-that there  !

could be relatively. inexpensive plant-specific s cost beneficial. improvements-to the decay heat removal function as a1 result ofta systematic examination. As such, the staff: resolution of A-45 called for a systematic examination of

~

every operating plant's decay' heat removal 4 system for severe accident vulnerabilities, an examination that'could be performed as part'of 1 p the ongoing IPE process. The IPE generic letter  ;

was modified to reflect this resolution, and to" emphasise the need for utilities to pay

~,' particular attention to decay heat removal i vulnerabilities during the conduct of their IPE.

However, the resolution-of U8I A-45 did not explicitly address the PORY issue for the six CE L h- plants.

Discussion: The question now facing the staff is to re-command either: 1) that these six CE plants now be required to install a manual high capacity:

venting capability similar.to.that which exists on all other operating PWRs in-the U.S., based primarily-on certain qualitative factors, or 2) that a manual high capacity venting capability not be required, based on the fact that it.is not required to meet any of the commission's regulations and that conventional cost benefit analysis cannot support such a requirement. '

Based on the constraints of the Backfit Rule,

~

the staff concludes that option (2) is.the only viable choice.

However, the staff notes that previous commission direction requested that such matters be brought to the Commission's attention (Memorandum to Victor Stallo, Jr., from.

Samuel J. Chilk, Secretary to the Commission, M880602, June 24, 1988):

"The Commission recognizes that there are

- m, unique situations-where the staff may identify an improvement that could significantly reduce the risk of a severe accident, but which, in the staff's judgement, may not satisfy all the tests of the Backfit Rule. The staff is directed to bring these issues to the attention of the Commission. The resolution of these issues then-would rest with the Commission, which has the ultimate authority and

- - - - - - - -- -~ - ' ' -

4 4;p .,

[O -h {>

s j LThe Commissioners 5 responsibility to make regulatory policy-decisions.*

-q Accordingly, Enclosure-1 to this paper! discusses

• in more detailfthe GI-84 resolution, including  ;

the quantitative and qualitative factors '

considered. In-addition, the qualitative  !

factors ~that could be used to-argue in favor of l requiring that the six CE plants.be required to

+

install a manual high capacity venting capability are summarised below:

o , l

1) The new CE design (System 80+) as'well as the EPRI-ALWR Requirements Document contain-

. primary system high-capacity manual venting '!

', capability, bringing into question the previous CE conclusion regarding the negative'effect on core melt frequency of installing PORVs. In fact,-the EPRI-ALWR Requirements Document I specifically lists feed-and-bleed operation as-a ,

function-to be performed by its safety Depressurisation and Vent System. In addition,

~

one plant'(Palisades). replaced their two existing PORVs in late-1989 with two larger PORVs to: enhance feed-and-bleed cooling capabilities.

L 2) Although the staff _has not required that l:

'~

the' feed-and-bleed mode be provided as a design ,

basis safety system, and neither the staff nor the industry have developed detailed performance.

criteria for'such a~ system, nevertheless all u PWRs have developed procedures for the use of whatever feed-and-bleed cooling capabilities their existing equipment can provide. .

3) The staff believes present PRA t methodologies may not completely quantify the DER system reliability improvements resulting from this type of diversity for the following reasons:

a) The diverse feed-and-bleed mode provides DHR system reliability improvement in situations that usually involve multiple equipment failures coupled with human errors resulting from failure to understand the nature of the event in progress. Such " common cause" events are likely to be assigned a very low probability in most PRAs. Others are unforeseen and thus are not modeled in any PRA; L

. - -. -- ~ - - _ - - - . _ . _ - . . - - - - --

e z r. ..

z.: ,

,,  ; The_ commissioners s 1 1

' An- extensive number of common-cause-related

~

b) 1 '

events-that are theoretically includedLin-any,PRA may be: eliminated by the truncation '

process used to eliminate low' frequency events so that the total number of events, considered.in the PRA is' manageable. The' I total sum of all such DER system reliability;  ;

improvement that~is not reflected in the;PRA~

' result may be adequate to justify 1 installation of primary system high capacity .

manual venting capability,on a value/ impact 1 ratio' basis.

-These unquantifiable limitations of present PRA technology make it inadvisable to rely v exclusively on'it in making decisions regarding reliability of the vital' DER safety system. A more reliable design would be a primary system. q that-cannot be Hbggglgd gpH'by excessive  !

unremoved decay heat pressurising it beyond the l point ~where existing-SI-pumps can supply cooling

- o water adequate to prevent-core damage, if the l cooling functions of the secondary system are j lost.

4) Manual-high capacity venting capability provides an accident management 4 measure to mid- <

in preventing direct containment heating during high pressure core' melt scenarios. NUREG-1150 studies of both Surry and Sion have shown manual high capacity venting capability lowers the contribution of early containment failure to total risk.

t

The staff has also considered continuing the evaluation of this issue via the IPE process.

- However, very little information has- surf aced over the past several-years (since the NUREG-1044 study was completed) that would change _

substantially the risk' reduction that would be

-u predicted for the installation of manual high capacity venting capability by a PRA-based study. Thus, although the staff expects that the utilities operating these six plants will pay special attention to the decay heat removal function during the conduct of their IPE (just as will all other utilities), the staff does not expect that their IPE's will show that a diverse manual high capacity venting capability is needed.

_ _ _ . ~ . . _ . _ . _ _ . _ . . _ ._._ _ __ ._

m C .., o 1

[ <

LThe1 Commissioners- 7- a i

~

Conclusion t In summary, the staff- concludes that the six CE plants in question continue to meet all of the '

E commission's. rules.and regulations without installation of a manual high. capacity venting i capability. . The staff also concludes-that b '

existing PRA studies and calculated-cost-benefit -;

I ratios do,not support backfitting these six' 'l plants with such a capability due to the high-  :

cost ( ~ $9 million per plant) and the~several hundred person-rem dose to-plant workers = 1 associated with installation-o? such venting capability, and due to the potve.zial negative  !

P effects on plant safety.if th6 dew venting capability were installed utilising PORVs prone to leaking, sticking open, or other operational problems. 'The staff therefore recommends no action be taken on these six plants. In a letter to the commission dated June 12, 1990,

.the ACRs has agreed with this recommendation (see Enclosure 2).

-l However,.in keeping with the previously quoted commission directive, this paper brings to the Commission's attention this unique situation -i where the staff has identified an-improvement '

that could potentially reduce the risk of a-severe accident significantly even though it cannot be quantitatively demonstrated that it satisfies all.the tests of the Backfit Rule.

This-lack of a quantitative demonstration h results from the large uncertainty in the calculated 2E-05 per reactor year core damage frequency reduction associated with this improvement'. The large uncertainty is due to the. inability of present methods to consider the many low frequency events which involve multiple failures and human errors, and due to the present uncertainty in the contribution of direct containment heating phenomena to the +

potential for containment failure during these events.

The staff judgmen. that significant safety enhancement could be achieved by installation of a manual high capacity venting capability is based on the considerations that such a system would provides 1) a diverse means for decay heat removal; and 2) additional capability to prevent high pressure melt scenarios, which could lead to early containment failure from direct containment heating phenomena or which could lead to steam generator tube failures.

H

,m.=

jf -

s 0

The commissioners' S ,

I l~ Also, such-a venting system would provide the option for future upgrading.to automatic a L

L initiation, which would provide safety .

ll enhancements for ATWS scenarios,.and which could. -)

provide reduction in challenges to the pr!. mary )

~ system safety valves'.

.j

~

l , .,/ .)

MR J J es M.: ylor secutiv Director '

)

for operations-Enclosurest

, 1. Background and Discussion of GI-84 Resolution *

.2. . Letter from Carlyle Michelson (ACRS) to the Honorable  :

.Kenneth M. Carr, dated' June.12, 1990 DISTRIBUTION:

Commissioners-

-OGC OIG

.GPA REGIONAL OFFICES EDO-ACRS ACNW ASLBP

-ASLAP SECY

m

o o..-

I

.4 1

ENCLOSURE 1-

'Backcround and Discussion'of GI-84 Resolution '

i

1. BACKGROUND All Westinghouse,' Babcock and Wilcox,. older combustion-Engineering (CE) plants,_and newer CE plants (System 80+), ,

have one or more PORVs or other primary system high capacity l~

manual 1 venting capability. However, the six most recent CE 3410-MWt and 3800 MWt plants now in operation do not have j PORVs or other such primary system high capacity manual-venting capability. Possible safety implications of this design difference were highlighted by an ACRS memorandum to o the commission in October, 1983 #3 which resulted in the l initiation of GI-84.

1 I

i A review of the need for a rapid depressurization capability

'on CE plants was carried out by the staff. Results of the.

review-were published in NUREG-1044 N) and reported to the Commission in 1984 03, and are summarised below in Sections 1.1 to 1.4.

L 1.1 Compliance With current Reculatory Reauirements The staff considered "I the CE plants' ability: 1) to -

depressurize and thereby limit offsite releases following a Steam Generator Tube Rupture (SGTR) event by use of the Auxiliary Pressuriser Spray (APS) system; and 2) to provide Low Temperature over-Pressure- (LTOP) protection during cold shutdown operations utilizing the safety valves in the Shutdown cooling System (SDCS). Many other plants utilize primary system high capacity manual venting capability to mitigate these events.  ;

The staff concluded that, with the exception of certain i single failure vulnerabilities identified in the APS,-the CE plants without PORVJ meet current regulatory requirements.

Corrective actions were subsequently taken to increase the APS reliability to an acceptable level with respect to the single. failure vulnerabilities. Mitigation of a single SGTR with either a primary system high capacity manual venting

-- capability or with the APS results in acceptable offsite radiological consequences that are essentially the same.

Further, mitigation using the APS has the advantages of providing a controllable depressurization technique and of adding fluid to the reactor coolant system. Multiple tube E ruptures--as either an initiating event or as a consequence -

L of other accidents--were concluded to be sufficiently low in probability that they need not be considered as design-basis accidents.

i l

16 s . q m;

u. *:

, i

' 0: Enclosure L2 y

The staff also concluded'"3 that SDC8 safety valves provide acceptable LTOP protection during cold shutdown.

4 1.2 Canabilities Beyond tho' Current Reaulatory Reauirements

, The staff also evaluated "3 the capabilities of the CE- 1 i designed plants without PORVs to mitigate-the consequences I of multiple-failure accidents that are beyon6 the current d regulatory requirements, including multiple steam generator a

< tube ruptures, total loss of all feedwater events, small J break-LOCA without high pressure coolant injection, and pressurized-thermal shock events. The staff concluded "3 that the existing. systems should be able to mitigate the spectrum of multiple-failure accidents considered. However, ,

there are limitations associated with the mitigation systems.

The capability.of the-APS to depressurize the RCS depends on the presence of a steam space in the pressuriser and on a number of operator actions. If the pressuriser insurge rate becomes excessivs, the rate of depressurization from the APS is significantly reduced. Also, if the pressuriser becomes-water solid, the APS ils unable to depressurize the system at all. A properly sized and reliably powered primary system high capacity manual venting capability would be capablelof reducing system pressure without these limitations. Also, although the staff has confidence in the deterministic L assessment of the APS-system, the staff recognises the fact I

that there may be malfunctions that could render the system unable to control plant pressure.

Similar limitations can be expressed regarding the decay heat removal systems. Analyses performed by CE showed that L the condensate system'is able to supply sufficient steam generator feedwater to avoid core uncovery. However, the

, condensate system relies on offsite power, and the steam generators themselves must be able to remove decay heat. In

.the event of loss of all feedwater, the steam generators may become unable toLremove decay heat, and a suitably sized End properly operated primary system high capacity. manual li' venting capability could remove decay heat and avoid core damage. Similarly, if installed in the forat of a PORV, this capability could keep the pressuriser safety valves from lifting and prevent an unisolable LOCA should one become stuck open..

On balance, although the staff recognizes that the existing systems can mitigate a number of multiple-failure accident scenarios that are beyond the current regulatory requirements, there are considerable uncertainties in this l ability, and a properly sized primary system high capacity l

l c

m e W3 7..

o :e s i

,fEnclosure? 3 manual venting capacity could provide additional reliability  !

-from defense in depth.-

s 1.3.Erpbabilistic Risk Assessment (PRA)

To' determine some quantitative measure of the change in safety that would result from the addition of such .

capability, the staff asked CE for information necessary to estimate this change in'a probabilistic way.") The staff has

. reviewed CE's. responses and requested Sandia National Laboratory to perform an independent analysis. In addition,.

the staff has performed its own probabilistic assessment.-

All three studies included a. quantitative analysis of the g i loss-of-main-feedwater event, including the loss of main l feedwater caused by loss of offsite power. Steam generator L tube rupturas (SGTR) were considered quantitatively in the l

CE.and Sandia analyses; only the staff analysis included a quantification of the benefits-from additional high capacity manual venting for ATWS sequences. External' events, fires,

! and flood were not considered in any of the studies.-

1^

Several additional potential. benefits from the addition of ,

primary system high capacity manual venting capability were-not quantified by CE, tho' staff, or Sandia. . These include the possible limitation of challenges to the safety. valves and.the ability to depressurise the RCS while a core melt'is in progress. This latter potential benefit would decrease the probability of failure of the steam generator tubes from steam' overpressure when the core slumps into the lower reactor vessel plenum, and could affect the direct containment heating (DCH) related risk as discussed in Section 3.2.

1 The staff's best-estimate calculations showed that if l primary system high capacity manual venting capability were t installed on CE plants, the core damage frequency for loss L of heat sink and ATHS sequences would be reduced from about 6E-05 per reactor year to approximately 4E-05 per reactor year. These are mean value estimates that combine the results of ATWS and non-ATWS sequences (the non-ATWS sequences are the principal contributors).')

The staff judgment that significant safety enhancement could be achieved by installation of a manual high capacity venting capability is based on the considerations that such a system would provide: 1) a diverse means for decay heat removal; and 2) additional capability to prevent high pressure melt scenarios, which could lead to early containment failure from direct containment heating phenomena or which could lead to steam generator tube failures. Also, such a venting system would provide the l

.w , - -

e .

Enclosure' 4 optionJfor future upgrading to' automatic initiation,'which would provide safety enhancements for'ATW8 scenarios,cand which could' provide reduction in challenges to the primary system safety valves.

However,iit remained for the analyses described in Section 1.4 below to show whether such safety enhancement would be cost effective. '

1.4 Value/Immact Analysis The staff performed an evaluation to determine if the installation of primary system high capacity manual venting capability'in represent CE plants lacking a cost-beneficial such safety capability w)ould improvement.(

The costs and other considerations discussed in Section 3.4 L of NUREG-1044 (2) are used to aid in assessing the valus/ impact (V/I) ratio and safety importance of installing primary system high capacity manual venting capability in the CE design. The V/I ratio is the quotient of the safety H

' benefits (values in terms of averted risk (person-rem) to'  !

costs (impacts)).- The averted risk resulting from a reduction in core melt frequency of'2E-05 per reactor year )

is 880 person-rem for SONG-2. The primary system high ,

capacity manual venting capability cost per plant (after l operation)'is $9.2 million. The resulting V/I ratio is 96 _

person-rem per $1 million.- Thus the resulting v/I ratio provided on3y a marginal V/I when compared to the unofficial L guideline most often utilized in keeping radiation releases as low as is reasonably achievable [ALARA, defined in Appendix I to Title 10 of the Code of Federal Regulations y part 50-(10 CFR 50)) which is a ratio of 1000 person-rem I averted per $1 million (i.e., the reciprocal of $1000 per person-rem). That is, the 96 person-rem per $1 million V/I ratio associated with the addition of primary system high capacity manual venting capability to these six plants is significantly outside the guideline for acceptability of at least 1,000 person-rem per $1 million.

~

It must also be recognised"that the V/I, as defined above, does not consider that the sene population is subject to the combined risk reduction from Units 2 and 3. Based on this interpretation, the combined potential risk reduction of .

1760 person-rem per two unit reactor. site, when considered independent of the V/I ratio, represents an important safety benefit. If the same reduction is assumed to apply for all six subject units, then the total risk reduction to the U.S.

population is estimated to be 5280 person-rem, a significant safety benefit.

e.

n .

i* 20 -

+ j

.W' Enclosure 5 j The above shows that a small but positive V/I ratio would result from the installation of primary system manual venting-capability at SONG 8-2 and -3.N3.high-capacity However, if :

l the values (benefits) discussed in the CommissionLPaper with-

which this discussion is enclosed could be quantified, a higher-V/I ratio would result. Therefore, recognizingithe limits inherent in the quantitative V/I ratio and based on a potential' reduction in core melt frequency of 2E-05,per reactor year and a potential site-specific risk reduction of 1760' person-rem, the staff found that-the. installation of

. primary system high capacity manual venting capability would

. provide a significant safety benefit..

2. PREVIOUS POSITION (NUREG-1044, N) December 1984):

n As a result of the considerations discussed (j)above (Section 1), in 1984 the staff published NUREG-1044 which concludedt  ;

1 uThe.overall results of the staff evaluation concluded that-there is a not benefit in adding PORVs to the current'CE .i plants without them. Although the quantitative portions of the evaluation (cost-benefit, regulatory requirements) do not support.this conclusion, other factors (beyond design-

-basis events, imponderable events, engineering judgement, large uncertainties in PRA results) that entered into the .

decision-making, when combined with the quantitative- a evaluation,= lead to this conclusion."

"The NRC is also in the process of resolving the more L comprehensive generic issue, USI A-45,-" Decay Heat' Removal L Requirements n. Because part of the benefit of~the,PORVs was H predicated on their ability to-provide an alternate decay L heat removal. path (feed-and-bleed), any improvements in decay heat removal capability that might be promulgated as a

. result.of the USI A-45 assessment could reduce the not ,

-benefit of PORVs. Therefore, the staff concludes that the decision regarding PORVs for these CE plants should be deferred a USIA-45."ggincorporatedintothetechnicalresolutionof

3. DISCUSSION OF CURRENT STATUSt 4

3.1.USI A-4) Resolution 3.1.1 Qgar,riction of nosolution When the above previous position was published in 1984, it was anticipated that a generic corrective action would be required by USI A-45 for all plants that would significantly affacc the. risk from fallure of the decay heat removal function and thereby resolve the primary system high

.=s-1

Enclosure:

6 capacity manual venting' capability issue for CE plants. .For -I

~

example,.one of the specific corrective actions that was  !

considered for A-45 was provision of a reliable-feed-and- H bleed cooling' capability for all'PWR plants, which would ]

have: involved installation of primary system high capacity l manual' venting capability on CE -plants.' This was not l selected for generic resolution of A-45 because it would not 1 achieve the desired core damage 1 frequency _ decrease on all q plants, and because there was uncertainty regarding;the L i

operators' reliability in manually-initiating feed-and-bleed t cooling quickly enough to effectively: Prevent. core damage.

Anotherispecific corrective action considered for generic resolution of A-45=was to require a separate and independent decay heat removal system. If such.a system'had been L

required, it would likely have been possible to show that-primary system:high capacity manual venting capability for

' decay heat' removal by feed-and-bleed would not be necessary in addition to'that system. .However, such a system was not i required because it could not be shown to be cost effective -

using accepted methodologies.

l USI A-45 has now been resolved (D (as described below in this D section 3.1.1);withent imposition of a generic corrective action'that significantly affects the primary system-high capacity. manual venting capability issue for CE plants.

The following conclusions were reached under the USI A-45 program:

1) The expected frequency of core damage related to

-failure of the decay heat removal function varies significantly between individual plants, andmat certain plants may be considerably above a staff goal of IE-05 reactor year-(this goal applies to individual sequences and was established a number of years ago to assist resolution of generic issues).

2) The value/ impact ratio of possible corrective

(. actions also varies significantly between. individual e i plants. Although some corrective actions Do achieve an acceptable value/ impact ratio for'cortain plants, none of the corrective actions analyzed (including provision of reliable feed-and-bleed cooling capability with primary system high capacity manual- <

venting capability or a separate and independent decay heat removal system) will simultaneously achieve an acceptable value/ impact ratio (using conventional methodology) and also reach the staff's partitioned core damage frequency goal for generic issue resolutions.

) - A t 4

<l Enclosure .7

3) since all of,the significant USI A-45 results have j

, been found to be highly plant specific, it is not l i

appropriate to propose _a single generio action to be applied uniformly to all plants.

.i

4) For_any specific plant,:to determine the~ level of i DER related risk and the value.and impact'of I proposed corrective' design' changes-(including primary system'high capacity manual venting d capability as part of a feed-and-bleed cooling- O capability),: detailed analyses'of-that specific q~

plant.are necessary.

Based on the~above conclusions, since decay' heat removal E failure-related risks and the effects of particular corrective actions are highly plant specific, the staff has recommended that each licensee be required to perform a limited-scope probabilistic risk type assessment (PRA) for l its plants. This assessment will:be conducted in I conjunction with the Individual Plant Examination-(IPE). I program. Plant specific implementation (including any l corrective actions proposed by the licensee and/or required '

l by the commission) will also be accomplished within the planned IPE activities, and USI A-45 (as a separate generic task) has been terminated. .

)1 3.1.2 Effect of A-45 Resolution on GI-84 l

Thus, the anticipation that USI A-45-resolution would also resolve'GI-84 proved to be-unfounded. U8I A-45 has now been resolved, but the GI-84 question (regarding primary system high capacity manual venting capability on cE plants) has  ;

not been settled.

It might be argued that rejection of the feed-and-bleed enhancement option as the generic resolution of A-45 constitutes an across-the-board rejection of the concept of W Lfeed-and-bleed cooling, but such is not the case. The feed-and-bleed enhancement option was not selected'for generic resolution of A-45 principally because the staff did not believe the generically. desired reduction in core damage risk could be achieved through feed-and-bleed enhancement (each of the four PWR plants erasined was assumed to inititally have some minimal feed-and-bleed core cooling capatility).

This does not imply, however, that absence of any feed-and-b'.eed cooling capability is acceptable. Even though enhancement of the feed-and-bleed option was not found to be justifiable as the generic A-45 resolution, nevertheless, for plants with no capacity for feed-and-bleed cooling, its provision through the addition of primary system high

L- -

.' j 3' -

s L EncloWore , 4 l l'

L onpacity manual venting capability would represent-a significant safety improvement. The significance-of such diverse shutdown cooling capability is discussed more fully in the Commission Paper with which this discussion is enclosed.

3.2 Updated PRA and V/I Analyses l

L The staff has also considered continuing the evaluation of H L this' issue via the'IPE process. However, with-the possible j L exception of the direct containment heating (DCH) issue.

u discussed below, very little information has surfaced over .

L the past several years (since the NUREG-1044 study was J completed) that would change substantially the risk  ;:

reduction that would-be predicted for the installation of 4 manual high capacity manual venting capability by a PRA-  ;

based study.- Thus, although the staff expects that the utilities operating these six plants will pay special- ,

l attention to the decay heat-removal function during the -

conduct of their IPE-(just as will all other utilities), the staff does not believe that the.IPE analyses will produce a L cost-benefit result outside of the uncertainty band l associated with the results quoted in Section 1.4., and thus the IPE results will not show that a diverse manual high

'. capacity manual venting capability is needed.

The most significant new factor that may be included in any IPE related cost-benefit analyses for installing new high capacity manual venting capability will likely be the decreasing risk resulting from the DCH phenomena. It appears likely that the probability of vessel failure in an-

~

early, high pressure mode that would contribute to DCH risk could be decreased appreciably if the operators initiate early depressurisation using a rather large primary system high capacity' manual venting capability. However, it is doubtful that the inclusion of DCH in IPE analyses will significantly alter the section 1.4 cost-benefit results, although the change wi1.1 likely be in the direction of increased justifi;acion for adding the high capacity manual venting capability (i.e., it would tend to become more cost effective).

4- 9 J 3.3 Sn==ary of Current Status t The staff has concluded that the six CE plants in question continue to meet all of the Commission's rules and regulations. The staff also has concluded that existing PRA studies do not generate cost-benefit ratios that support backfitting these six plants with a manual high capacity manual venting capability. Therefore, the staff recommends no action'be taken on these six plants.

+ .

.. a

.H. (

Rnclosure 9 However, in keeping with the Commission- directive quoted in the Commission Paper with which this is enclosed, that' paper-(including this enclosure) brings to the Commission e s attention the reasons why an added primary system high-capacity manual venting capability would represent an improvement that could significantly reduce the risk of a-severe accident, evas though an added primary system high ','

capacity manual venting. capability does not satisfy all the-tests of the Backfit Rule. '

4. RESOLUTION ALTERNATIVES  ;

The staff considered the two options summarised in the Commission Paper with which this discussion is enclosed. ,

These options are more fully discussed in this section 4. -

a 4.1 Option #1 -

Conclude that No High capacity Manual Venting capacity is Needed.

The advantage of this option is that it is supported and justified by presently available value/ impact analyses consistent with the Backfit Rule. For that reason, the 'l staff has chosen this option. >

4.2 Ontion #2 -

Issue Orders To All Plants'Without PORVs-Requiring Installation of Primary System High capacity ~ Manual Venting capability.

The orders would require installation of primary system high j capacity manual venting capability comparable to other plants of similar design. Capacity and reliability, criteria would not be specified, since they have not been developed for other plants. However, it would be recommended that a blowdown capacity be installed that is " sufficient to enable primary system blowdown to be initiated as late as steam-generator dryout after a loss of all feedwater event, without 1 predicted core-damage." such a capacity is most likely to be compatible with any performance requirements for feed-and-bleed cooling and/or decreasing DCH related l risk that might be developed in the future, without further plant modification. The order would encourage licensees to install equipment of a design whose reliability would consider the total risk due to stuck open PORVs and stuck open SVs.

The orders for installation of-the primary system high capacity manual venting would not be in the form of "show cause" orders, since (as discussed in the Commission Paper with which this discussion is enclosed) the staff has already become familiar with the quantitative arguments such an order would invite from licensees (i.e., showing that the

. - - - -- - .- .- - - =. - . - - -.

7.

. . .. W W

i f i

Enclosure- 10 feed-and-bleed mode is not needed because'it is not justified on a value/ impact basis).

The staff is constrained from recommending this resolution.

.for GI-84 because the cost / benefit ratio that results from- -;

PRA-based. calculations (see Section 1.4) does not provide a L clear justification for requiring primary system-high l  : capacity manual venting. In particular, while the benefit in adding a' annual-high capacity manual venting is substantial (core damage frequency reduction of 2E-05 per j L

reactor year), the maximum costs that can be justified,  !

using $1000 per person-rea as guidance, is approximately $1  ;

million per plant,7whereas a high capacity manual venting system would cost up4 to ten times that amount per plant.

L k

i E

e

' *- , 9 ,

a,

Enclosure:

11 References

1. Letter to N. Palladino from J. Ray, HNeed for Rapid . )

Depressurisation capability in Newer Combustion Engineering, Inc. Plants," October 18, 1983 r

2. " Evaluation of the Need for a Rapid Depressurisation 1 Capability for combustion Engineering Plants," NUREG-1044, December, 1984.

l

3. Commission Paper " Power Operated Relief Valves for Combustion Engineering Plants," SECY-84-134, March 23, 1984.

4. " Cost / Benefit Analysis of Add.'.ng Feed-and-Bleed Capability to Combustion Engineering Preactrised Water Reactors,"

NUREG/CR-3421, August, 1983.

5. . SECY 88-260, " Shutdown Decay Meat Removal Requirements (USI l A-45)," September 17, 1988.

1 4.

- - - . . A . . _ . . _ - - - . - _ - - _ - _ - - _ __.____ .___-_

• ~ ~ ^ ' '~~' ^ ^ ^

Q;i$ i-d7 7. v c.

7, 2 .

llE ENCLOSmE 2

{@N d%i_ L a a ne: \'

UNITED STATES NUCLEAR REGULATORY COMMISSION - ,

Th M g . ADVISORY COMMITTEE ON REACTOR SAFEOUARDS

" [h WASHINGTON, D, C. 20808 ;  ;

7 All

[

l.j '

1 June 12,.1990 e ,

4 i ml[

Y4 The Honorable Kenneth M. Carr

. Chairman .

)

4 4 U.S. Nuclear Regulatory Commission '

Washington, D.C. 20555 ,

i

Dear Chairman Carr:

\

SUBJECT:

GENERIC ISSUE-84, COMBUSTION ENGINEERING PLANTS WITHOUT .

POWER OPERATED RELIEF VALVES '

[

y~

During the 362nd meeting of the Advisory ' Committee on Reactor ' -.i c Safeguards, June 7-9,. . 1990, we reviewed the staff's proposed-  !

resolution of Generic Issue-84, "CE Plants Without PORVs." We had y the benefit of discussions ' with members of the NRC staff and

. representatives of Combustion Engineering Incorporated (CE). We  !

n also had the benefit of the referenced document.

Generic Issue-84 (GI-84)~ applies to six operating ~CE units that do ,

not have PORVs or other means - to rapidly reduce reactor coolant 1 pressure through the venting of steam from the system pressurizar.

These are the six CE units designed after 1970: San Onofre Units 2 and.3; Waterford Unit 3; and-Palo Verde Units 1, 2, and 3.' These 3

. units.do have enhanced capability to reduce pressure by means of

, essentially. safety grade auxiliary pressurizer spray. systems.  ;

This generic' issue was established to determine'if the capacity for " feed and bleed cooling" afforded by PORVs should be required ,

for -these units. Most pressurized water reactors can " bleed" through PORVs and " feed" with high pressure makeup pumps as an '

emergency means for removing decay heat from the core. While this-cooling mode is believed to be useful in reducing risk of core overheating in some circumstances, the NRC has not made feed =and bleed capability-a requirement.

'In its report dated December 15, 1981, the Committee expressed concern over a lack of means to feed and bleed in these plants ,

during-its review of the CE System-80 standard plant design. At i the Committee's request, studies of the pros and cons of installa-tion of PORVs on the CE plants were conducted by NRC and CE. These studies indicated ambiguity as to whether there would be a small reduction or a small increase in risk resulting from a backfit addition of PORVs. In 1983, the Committee agreed with a staff decision to incorporate this issue into the then ongoing effort on

P .- , j 4 - -

.. 3 0

t ,1 1  ;

The Honorable Kenneth M. Carr 2 June 12,1990

~ Unresolved' Safety Issue (USI) A-45, " Shutdown Decay Heat Removal .

Requirements.". However,-the ultimate resolution of USI A-45_in 1 1988 did not ' explicitly address the PORV concarn, which was therefore carried on as GI-84. g The staff has now proposed that installation of PCRVs not be J required for - the affected plants. The reasons citod _for this decision'are that such action is not required to meet any of the Commission's regulations, 'and that a cost benefit analysis, as '

called for under ~ provisions of the backfit rule, indicates-such:

system modifications are not justified.- In general, risk analyses, which are an important ingt to the assessment ' of ecst ,and' l benefits, lack the accuracy needed to make decisions about the very small differences in risk, plus or minus, that could be created by addition'of PORVs to these plants. We also note that the risk analyses that have been conducted for resolution of GI-84 were limited in that they'did not include consideration of external events as initiators. Nevertheless, we believe that these analyses -

have been useful and we concur with the staff recomitendations.

Sincerely, j

. Carlyle Michelson

( Chairman l-

Reference:

i- Draft Commission Paper, " Evaluation of the Need for Primary System-High Capacity Manual' Venting Capability on Combustion Engineering ,

(CE) Plants Without PORVs (GI-84)," transmitted by memorandum dated April 27, 1990 from Warren Minners, Office of Nuclear Regulatory

( Research, NRC, to Raymond-F. Fraley, ACRS.

I l

l l

l