ML20084R628

Discusses Status of GESSAR-II Review Re Probabilistic Fire Analysis.Related Info Encl

ML20084R628
Person / Time
Site:	05000447
Issue date:	01/23/1984
From:	Azarm M, Boccio J, Ruger C BROOKHAVEN NATIONAL LABORATORY
To:	Bari R BROOKHAVEN NATIONAL LABORATORY
Shared Package
ML20084R627	List: ML20084R628
References
FOIA-84-175 NUDOCS 8405220645
Download: ML20084R628 (25)

Text

- . , , , , , . - - .-. - . , , . , . - . ..

i .

} ,,

BROOKHAVEN NATIONAL LABORATORY O '

MEMORAN'DUM Opac<n.2 + h 4 m y m n,,r% sv%o.p p 34 w w x.-e n 7.~- m e , +

DATE: January 23, 1984 To: R. A. Bari FROM: M. A. Azarm, C. Ruger, and J. L. Boccio

SUBJECT:

Status of GESSAR Review: Fire ment1),(Our review prepared of the by the GESSAR General ElectricII'Company, Fire andhas Flood External reached Event Analysis docu-an impasse.

Prevalent to. this. state of, affairs is the fact that (1). the document is not a "sta'nd alone" study of the risks associated with fires within the plant,'and (2) we have not, as yet, received from the utility any responses to our ini-tial inquiries that were submitted to NRC by letter dated Dec. 6,1983. -

This lack of completeness and' utility responsiveness notwithstanding, the purpose of this memo is to amplify further our initial findings. As has been the case in our review of the LIMERICK SARA, the team assigned to review

~"GESSAR II .- Fire" had a dual function. One is to critique those determints-

' tic fire models empToyed in the analysist the other, to assess- the probabilis-

• tic methods and data employed for detennining the frequency of fire-induced initiating events. The results of this effort are then reported to those within the Risk Evaluation Group assigned to review the bottom-line risk num-bers, viz, core-melt frequency. For completeness, the scope of our review alsoentailedthegatheringofin{grmationfromacompaniondocument-the GESSAR II Fire Hazards Analysis.l 1 ,

Probabilistic Fire Analysis ~- A~ Preview At the outset, we must state that with the incorporation of many fire-protection features within the generic plant, the core-melt frequency from fires is expected to be comparatively lower than those from other plants. Pri-marily, this is due to strict compliance to requirements set forth in Appendix R to 10 CFR 50 in Section 9.5.1 of the Standard Review Plan (SRP). However, corroboration of this expectation, through quantitative analysis, has been hampered from the lack of completeness of the document under review. Al so, in certain respects, the document does not contain the necessary elements or steps currently associated with what is perceived to be the state-of-the-art in probabilistic fire-risk assessment. This, coupled with the prevailing large uncertainties in fire-risk analysis, precludes making any viable judg-ments or appraisals on the cited document.

The scope of'our review is, therefore, limited to (1) generic discussions on specific fire protection features designed into the GESSAR Il plant; (2) their potential impact on fire-risk analysis; and (3) to the extent possible, a summary appraisal on the existing GESSAR fire-risk analysis.

8405g645840417 A PDR k Ly04-175 x

i . .

e hwwr.g G ggggp: L z. rw.mawaupemenw&4%4 w*wwinwr .

Several probabilistic fire-risk studies have already been perform-ed (3-6) . In almost all of these studies, fire is one of the major contribu-tors to the overall plant risk, dominated by fire occurrences in a few areas where electrical cabling and equipment of redundant shutdown divisions are l

either mutually located or do not meet specific fire protection requirements.

Thus,. the methodolog.ies. developed for probabilistic fire-risk assessment basi-I cally address single enclosure fires, and the contribution of risk due to a l fire spreading beyond the rated barriers has been judged to be negligible com-l , pared to single enclosure fire risk.. .

For a gingle enclosure fire, three stages of fire propagation are usually considered (H. First stage growth is construed as damage to components in the immediate. vicinity of the, fire. ; source.which usually.causes.the initiating.

event (reactor scrant and other transients). The second stage consists of fire growth to adjacent unprotected cable raceways, separated from the initial fire by the minimum separation criteria (5 ft. vertically and 3 ft. horizontally). ~

Third stage fire growth represents. fire of sufficient severity and duration to damage mutually redundant shutdown methods separated by at least 20 ft. or else protected by rated barriers (usually 1/2 hr. to. I hr. rated blanket). -

For those plant designs and areas, where cabling and/or the equipment used for redundant shutdom methods are located .the. third sta'ge. of. fire growth within ~

• these s~

are~es"usu'aiTf yteTds'the dominant' ftre-rtsk contributors. ~"

In the GESSAR II Plant design, cab 11ng and equipment . associated with the redundant shutdown methods are s (in most cases) separated by three-hour rated ,

barriers. (There are a few exceptions that are discussed in detail in the fire hazard analysis repart.) Minimizing fire growth through efficient utili-zationi of these barriers has'three : impacts on. enfsting fire-risk assessment methodoTogiess vtz,;// vcb .; - mv .<.o - * .

1) conceivably, fire risk is reduced because the contribution of third stage fire growth is lessened. However, the contributions of Stage 1 and 2 fire growth may now be comparable or even more than the Stage 3 fire growth.. Hence, it is important to consider all the areas of the plant where a- fire- can cause an initiating event (first stage growth) ..

2) The third stage- fire growth (now defined as fire spreading through three-hour rated barriers and penetrations) necessiates consideration for multi-enclosure fire spread and thus differs from the various fire PRA's performed thus far for single enclosure fire growth.

3) A fire in high fuel load areas, contiguous to critical areas con-taining safety systems, may impose a threat to plant safety. Al-though the probability of fires,with enough severity and duration to cause barrier failures may be small, their inclusion in prob listic fire analysis for plant design, such as GESSAR II, canno.abi- t be disregarded.

. . . ~ . ,

I s u . . ' '

l .

. var MwwmeAmsuslags414$shthadiggi$,undergt%giscumstancasgequipr ga,%g',,~, y, The first~ can tie comp 11eFwith ff air areas the above factors be considered.

l

! of the plant thatsmay induce an initiating event or contain safety-related l equipment or cabling, are included in the study. The' remaining can be ad-l dressed by boundjng analysis on barrier failures (7) and construction of an adjacency matrixt8) to identify the critical multi-enclosures fire scenarios. .

Once the critical multi-enclosures fire scenarios are identified, the detailed probabilistic modeling can be performed by means of various methodologies such

as those given in References 9 and 10.*

l l The GESSAR Probabilistic Fire Analysis report appears not to have includ-ed these three factors systematically. Hence,. questions as.to completeness of the study and efficiency of the numerical results prevails. Note that the ad-i ditional elements, discussed above, have yet to be implemented in conventional fire probabilistic risk studies and that further research in this area may be e requirede However,- given the existing. method' ologies...some of. the concerns.

mentioned above could have been addressed for GESSAR II plant, at' least through a bounding analysis. . -

' Specific Discussion -

Our specific comments and questions on the GE report on probabilistic fire analysis for GESSAR II are given below. These comments are on the

~ screening analysis.used; them fire frequencies! employed; fire propagation and suppressfon models' incorporated; and',the u'q' antiffcation of ff re-induced core

~

melt probability _

e Screening Analysis Althougir a GE screening analysis has identified: six critical fire areas there is TfttTe discussion.and; almost no documentation as to how the

! # analystr?was'perftirmedS(ItHs? stated thatthe areas identified througEthe ---

screening analysis are basically the same as those suggested' by the NRC.) In addition, the consequence of fire in the diesel generator building is consid-ered negligible under the assumptionr.that the diesel generator catches fire whfie in the standby mode. However,. actuarial data onr fire in nuclear power plants indicates that the frequency of a. diesel generator catching fire per demand is about 7.4x10-4.. Hence, a fire scenario starting with Loss of Off-Site Power (LOSP) and imposing demands on the diesel generators may result in a diesel generator fire while it is not in a standby mode. Given the exis-tence of three diesel generators in GESSAR II plant and the frequency of LOSP of about 0.22/ year, a frequency of 4.8x10-4/ year can be assigned to this fire scenario (LOSP with loss of one diesel).

• These two references may also provide some insight into determining fire-barrier effectiveness.

. Jherefore, BNL cannot address the' adequacy of the screening analy-WNf sN Tu? M19P86E5SP FP p5 test withaar cJaJ.i.JancLs rededngefurther Aec ayes @.g' umentation on this subject.

s Fire Frecuency The GESSAR report states that the annual fire frequency for the crit-ical fire areas are estimated based on the actual fire data in nuclear power plants (ll). However, no discussion on either the data used or on the esti-l mation process employed is provided in the report. Reference (12) lists data

! on building fire frequency based on the actual data given in Reference (11).

These estimations are given in Table 1. It has been our recommendation that the data count for self-ignited for the use of propercable fires splices, be~ reduced overrated by aretardant and flame factor ofcables three to g.

Furthermore, the resultant frequency for self-ignited cable fires can be specialized to a specific fire zone in a building by de ratio of the actual

"' ' combust:ible Wefght or"krei~ in the Yoom to the~ total combustible weight or area -

in the building.

~

It is also not clear how the annual fire frequencies for the ' critical areas are estimated in GESSAR II Probabilistic Study, Specifically, the p-factor, defined in the study as the ratio of combustible area to the room area, may only be justified for the transient fires if one takes into account the critical distances given in Table 2-2 of the GESSAR II Fire Probabilistic

~

• reportDAgafrsMthouitfbrther documentation' explaining how. the fire frequen- ,,o cies are estimated and what is the logical interrelation between the p-factor and the fire frequency in the critical areas, we will avoid any judgment on l the accuracy of fire frequency estimates, l

! s' Zone Specific Comments s, m . Control Roons gg . ~( ,. .,

The scenario considered for the control room starts with a fire in a panel which is confined to a cabinet. Propagation to other cabinets is judged to be negligible. It is. stated that the cabinet design given in Figure A il-I lustrates features which mitigate the spread of fires. Hi mver,. the report does.not contain Figure A. It is also stated that there is 0.2 chance of cable ignition. It is not clear to us what tables are considered and how the 0.2 chance was estimated. There is no indication which cabinets are critical and how the p-factor was ' estimated In conclusion, we recommend for the control room fire a bounding analysis be performed (similar to the one done in Reference (4)), taking into account the plant-specific fire protection features.

Control Eouipment Room The same comments as above are; applicable.

1 l

i

l.

~' enungen *, <eGwes ETectricaT Eouipment Room (Control Buf1 ding)pwrm +rmanciswmg.~

A fire scenario, starting with a fire in a termination cabinet, dis-abling a safety division and then propagating to other divisions is consider-ed. The following assumptions are made for quantification of chis scenario:

1.. It is assumed that it takes 10 minutes for a fire to disable a division of safety cables.

'2. The probability for fire to propagate 20-feet to another division is assumed.to be 0.5.

3. The propagation of fire through the termination cabinet, a three hour rated wall, 50 ft. of separation and another three hour

.. -, ,. -; -+ "-rated- wet + is. assumed = to be 5.0x10-6, Given the assumption that there are no exposed cables in this room, ,

the above estimates, although judgmental, may be appropriate. However', the event tree developed for this fire scenario attempts to combine the two elec-trical equipment rooms together. This combination results in an unur-estimation of the core taelt frequency. The proper estimates can be obtained if two similar event trees for each electrical equipment room fire scenario be

. .s . ; constructed and quantifiede..In addition, the,p-factor. used for these areas ,

can be neither.justiff ed nor' understood.

Cable Tunnel i

l The cable tunnels are located at each side of the control room. Di-visions. I and 4 are located onr one side of the control room, while Divisions 2 and 3' are Tocated on the other side CabTing in Divisions 3. and 4 runs

throughEcondtrits emoedded dthint three-hour resistant concrete exterior walls. -

-The analysis. perfonned for the cable tunnels resembles that of the Electrical Equipment Room with the exception that the probability of suppression for the first 10~ minutes is assumed. to be 0.9. Hence, our critique of thi's fire zone is similar to that given for Electrical Equipment Room.

Auxiliary Building (Electrical Equipment Room and Zone 1 Corridor)

The analysis is not clear at all. There are two auxiliary electrical equipment rooms and two corridors of concern (Zone 1 and 2). It is not clear why only the Zone-1 corridor and one of the auxiliary Electrical Equipment Rocas are considered. There is some Division 4 cabling and equipment in addi-tionr to Division 1 equipment in this area which is not considered. The initi-ating event (fire induced transient type) is not known. The scenario for which the suppression failure probabilities are estimated is also not known.

l In conclusefon, for these two fire zones in the auxiliary building, further documentation on the component Tayout and methodological approach is required.

I _ _ _ . _ . _ _ _ _ - . _ . _ _ _ _ _ _ _ _ _ _ _ _ _ . ,_____._._ _ _ _

Mb - ~ -Qefermfhf'sffhnhAdiksT/rMfwD N""**YW*'h Deterministic fire growth modeling is used in the GESSAR II Fire External Eyent Analysis to determine the occurrence of secondary fires to initiate and subsequently grow when fire propagation is indicated. The resultant fire growth times then serve as input to the probabilistic methodology from which the failure probability of fire suppression is factored into the accident sequence progression. .

The specific deterministic fire growth model used is the computer code COMPBRN(14,151. This code. is' a synthesis of simplified, quasi-steady unit models resulting in what is commonly referred to as a zone approach model.

Our general evaluation of th9 (leterministic models employed in COMPBRN appears in the Limerick SARA review p3f ,

. , ~ ..a.~ ..

,:. . m..,-, . . . . . . , . . . . .

The application of this deterministic model in the GESSAR II Fire Exter- .

nal Event Analysis is somewhat difficult to review. No COMPBRN calculations have been performed specific to the GESSAR II plant geometry. Instead, total reliance for deterministic calculations is placed on COMP 8RN computer runs performed for the Zion (3) and Limerick (4) PRA's. This lack of a self-contained document complicat.es the review process.

. . , , .Notwithstanding, this segment of,the overall_ review consists of two parts'. The' firs't contafns some generaT" comments regarding the completeness of the analysis while the second concerns the application of the calculational results to the GESSAR II scenario. .

l General Discussion A signfffcant porticr, of the electrical cabling in the GESSAR II plant is

! , routed $ througis conduit embedded ts concrete as.fn the Cable Tunnel. Fires,can l occur in areasL surrounding these conduits and depending on the external heat i

load, combustible vapors may be produced inside the conduit due to the pyroly-sis of the cable insulation. These vapors will not ignite near the heat source since they are physically separated from the ignition source by the concrete conduit However, these combustible vapors carr migrate to adjacent fire areas through the conduit penetrations. At conduit termination points,.

which are likely to be within an electrical cabinet, these vapors can ignite when provided with an igaition source such as a spark from equipment in the cabinet This form of inter-area fire- propagation has not been considered in the GESSAR analysis and can lead to investigation of additional scenar. .

A somewhat similar concern is the inter-area propagation of smoke. Areas such as the- Control Room and the Zone 1 Corridor, which contains the remote shutdown panel, have been assumed to remain operational during fires in other areas. Therefore, consideration should be given to the po'ssibility of smoke propagation to these areas from fires in adjacent areas rendering them inhab-itable thereby affecting the success probability for achieving safe shutdown.

l l _ _ _ _ . , . _ _ _ _

~ l There is.an inconsistency regarding the consideration of transierit com-

• MMDusWWe#ftres-9MM11desGE8SNtWyetswwTaMe4.>0ndfestestthattsigsrtiteantwW rp ? '

transient combustibles are present in the Cable Tunnel and Auxiliary Electri-cal Equipment Room. However, no analysis are included to account for fires initiating with transients such as an oil pool fire.

Deterministic Model Application As notied earlier,_ the GESSAR II analysjs calculations contained in the Limerick SARA \ 4) relies heavilyHowever, analysis. on assessment and there ap-pear to be many inconsistancies between the physical situations analyzed in I the Limerick and GESSAR plants which precludes drawing any meaningful conclu- l sions based upon the identical analyses e'mployed.

For example, in the Control Room and Control Equipment Room, the GESSAR , 1 analysis states "the probability of propagation of the fire beyond the panel l tio"cabTes"wa's' assessed tar be 'G.04J.". 'This' appears'to-be taken frorn -a judgs e mental assessment in the Limerick SARA for the Safeguard Access Area, where panel fires that propagate beyond the confinement of the panel and ignite ad- .

jacent cable raceways prior to suppression are considered. Since none of the reported panel fires propagated in'such a manner, an upper bound was assumed in that one in five reported fires propagates. A further five-fold reduction in this upper-bound value was assumed to account for the difficulty of igni-tion of flame-retardant cable insulation _

m, m;.m g p m3 .u.- m ~ y .. .

In the GESSAR Control Room, where this assessment is applied, the situa-tion is different since there are usually no exposed' cables in the Control Room. It is not clear if exposed cables are actually present, if the cables considered are assumed to be internal to an adjacent panel, or if panel-cable raceway propagation was taken to be a conservative estimate of panel-panel propagation. . The Tatter case would yield a highly overconservative

. .probabfTity.2 ; g. g .jj _ y. ,,g g; ,

It is also unclear what rationale was employed for concluding that panel propagation has an higher probability than cabinet propagation. The failure

~

to include the referenced Figure A adds to the uncertainty.

Paper is considered as a transient combustible in the GESSAR control room It is stated that COMPBRN calculations from the Limerick SARA indicate that paper fires are incapable of igniting cable insulation._ The calculation in' the Limerick analysis considered 2 pounds of paper,1 foot in diameter, lo-cated 10 feet below a cable tray. Again, it is not obvious what the relation-ship is between this situation and that of a control room paper fire adjacent to a panel or cabinet. In fact, such a scenario was considered in the Lim-erick analysis, and is included in Table 2-2 of the GESSAR analysis, in which this paper fire is found to be capable of damaging the internal components of a cabinet when within 1 ft. of the cabinet wall. It is unclear why the cable damage analysis was cited rather than' this cabinet analysis.

%&Q , jnMgQQyfgg#ggggggym*wo-a cites a COMPBRN calculation from the Limerick PRA which results in a 0.4 fail-ure probability to suppress a fire in 10 minutes. This 10-minute growth time represents the time interval between the fire self-igniting in a cable tray and spreading to redundant cable trays located at a distance of 5 ft. verti-cally or 3 ft. horizontally from the initial fire. The failure probability to

suppress is taken from the Limerick PRA suppression model (Figure 4-4).

~

l . Since Table 2-I'of the GESSAR analysis indicates that there is no exposed l cable insulation in the Electric Equipment Room, the relevance of the cable l tray fire growth time calculation to the cabinets and panels actually existing in the room is unclear. The Limerick SARA review evaluates the models used in '

the COMPBRN code and indicates that the 10 minute cable tray fire growth time is over conservative.

l .

.. ~,

~. . . .. .. .~ . .

Judgmentally, we feel the GESSAR II plant to be relatively free from fire -

risk. Strict adherence to the cri.teria and guidelines found in Appendix R to l 10 CFR 50 and SRP 9.5.1 are the basis for this qualitative appraisal. Sub-stantive proof, based upon analysis, is however wanting.

If the GESSAR II analysis considers barrier failure to have a finite

'probabfTity#as a ~resuiesof', ffres-beintrinitfated wfthinr critical areas, then

~

the analysis should also include barrier failure from fires that are initiated in those non-critical areas which are contiguous to critical areas of concern.

l The. use of the p-factor should be justified. Considerations where the initiating event frequencies of fires, within specific areas, are reduced by i , factors either weighed by the- amount of combustibles in the area or by their-l . occupancy, must. start with a commom basts.. That is, the actuartal data em-played must also take fuel weight or surface area into consideration before it

~

I is applied specifically.

Increasing the fire resistance of barriers and walls by composite con-struction with fire-resistant materials places greater import on penetration seal effectiveness. This aspect should have been included in the' analysis.

Specific fire scenarios discussed within the report do not always corre-l spond to the particular physical situation. Strong emphasis has been placed l upon fire situations analyzed in other plants, and, at times, yield inconsis-

! tencies as to what is damaged and by what fire-induced stress mechanism.

In short, and with the limitations of the state-of-the-art in fire-risk assessment in mind, we cannot quantitatively appraise the GESSAR II analysis.

l l

l C

REFERENCES ,

1. General Electric Company, "GESSAR II Fire and Flood External Event Analysis," October, 1983. -

2. General Electric Company, "GESSAR II, 238 Nuclear Island Doc. 22A7007, Rey, 2., Appendix- 9A.

3. Commonwealth Edison Co., " Zion Probabilistic Risk Safety Study," 1980.

4. Philadelphia Electric Company. " Severe Accident Risk Assessment, Limerick Generating Staton," Report No. 4161, April 1983.

"- '< - ~ Se - Indiafr Point Probabrilistic"Safetye Study";< Power-Authority of the'St' ate of - -~

NY, Consolidated Edison Company of NY, Inc., Spring 1982, Chapter 6.

6. K.N. Fleming, et al., "A Methodology for Risk Assessment of Ma;I Fires and Its Application to an HTGR Plant," GA-A15402, July 1979.

t

7. Dennis L. Berry, Earl E.' Minor, " Nuclear Powei Plant Fire Protection -

Fire-Hazard Analysis," NUREG/CR-0654, SAND 79-0324, September 1979.

4 'g wantrn % +; & vs 48> w w - + n.m . .

v .

8. Donald A. Duke, "A Systematic Approach to the Identification and Pro-tctfon From Fire of Vital Areas Within Nuclear Power Plants,"

SAND 82-0648, October 1982.

9. A Conceptual Approach Towards a Probability Based Design Guide on Structural Fire Safety," Workshop Report, CIB W14 Workshop " Structural 1983.

- chr+y-w,e NEfra Safy yt fmm January,7

.: . w ueU@. ; vmrM sw

+ - -

Y> -

' 10. Sven Erik Magnusson and Ove Petersorr, " Rational Design Methodology for Fire Exposed Load Bearing Structures," Fire Safety Journal, 3 (1980/81) 227-241,

11. " Nuclear Power Experience," Division of Petroleum Information Corporc~

tionr, Denver, Colorado, December 1981.-

12 Kazarians, M., et al. " Fire Hazard and Failure Model," to be presented at ACTA' Seminar, Palo Alto, Ca., April 1983.

13. Azarm, M.A. et-al., "A Preliminary Review of the Limerick Generating Sta-tion Severe Accident Risk Assessment," Volume I: Core Melt Frequency, NUREG/CR-8NL-NUREG Draft, Brookhaven National Laboratory, August 15, 1983.

l .

. o

• O o

w ec.m m a;-r# A w r% + w c h REFERENCES (Cont'd)

Ik. Siu, N.O., "Probabilistic Models for the Behavior of Compartment Fires," _

School of Engineering and Applied Science, University of California, Los Angeles, Ca.. NUREG/CR-2269, August 1981.

15. Siu, N.0., "COMPBRN - A Computer Code for Modeling Compartment Fires",

School of Engineering and Applied Science, University of California, Los Angeles, Ca., UCLA-ENG-8257, August 1982.

9

, 9 2 D b 4 .A -

s ; I.

g  ;

g_

E,8 r_p .

p .

b

-h/Y M'f,fbhfh } hhhh MJ2dpihr$dMajjgkg'4^y ?b 3ce 'ig; , y . .

w,_ mf g _ _

e l

g . '*

BROOKHAVEN NATIONAL LABORATORY b MEMORANDUM DATE: Janua ry 25, 1984

-To: R. A. Bari FROM: I. A. Papazoglou

SUBJECT:

GESSAR Internal Flood PRA Review

' This meno summarizes the review comments on the GESSAR Internal Flood An-alysis to date. A brief preliminary review was conducted by BNL to identify and -raise issues, concerns, and questions pertaining to the analysis. It is inoortant that GE responds to these concerns before February 8,1984 so as to

. allow BNL enough time to evaluate the GE inputs for inclusion into the final draft report.

1) In the GESSAR Internal Flood Analysis, two water sources were considered; they are potential cracking or rupture of pipes, and leakage from seals and nlands. Provide the rationale why internal flood due to naintenance of _ equinnent was not considered in the analysis. Explain in nore detail how the effects of ~ draining'the suppression pool and/or the condensate storane tank are evaluated for the various potential flood areas.

2') On p.3-6 of the Internal Flood Analysis, it is stated that the instrumentation of the drywell floor drain leak detection system is fed from an uninterniptable power source. Provide additional discussion on the "uninterruptable" power source and on how the value of 2x10-3 is de-rived.

3) Industrial data were cited as a basis for assuming certain flood frequencies in different parts of the plant. Indicate the section of the GESSAR SAR from which the information is derived.

E.

o; 9 liero to R. A. Bari from I. A. Papazogl ou

.. s, January 25, 1984 l# Page 2

4) In the event that there is a pipe rupture at elevations above the botton level of the building, water will drain into the botton level through stairways and cable trays, etc. Discuss why the cascading of water from higher elevation which could result in potential ca,non cause failure of systems at lower levels is not considered in the GESSAR analysis.

.5) In .the Internal Flood Analysis, GE evaluated the scenario of flooding in the diesel generator building leading to a renual shutdown. Uhat is the inpact to core damage for a scenario with the simultaneous occurrence of a loss of offsite power and a flooding event in one of the diesel generator buil di ngs.

6) , Provide a-discussion of the locations of safety systen components, instru-ment panels, and electrical nanels with respect to flood height in the GES-SAR-design.

IAP/dn 4

4

w -~

6Cl BROOKHAVEN NATIONAL LABORATORY MEMORANDUM DATE: Janua ry 24, 1984 TO: R. A. Ba ri FROM : {, A, papazoglou

SUBJECT:

GESSAR Seismic PR A Review The purpose of this neno is to sumnarize the review status of the GESSAR seismic _ event PRA to date and to highlicht the. major concerns identified durica the review thus far. This review is based on two documents that were submitted to the NRC at two different times: GESSAR II Seisnic Event Analysis, Seotenber 1983(1) and GESSAR II Seismic Event Uncertainty Analysi s , December 1983.(2) On the . 9th and 10th of January 1984, a technical infomation meetina was held between RNL and GE to discuss clarifications of the reports and suDplement information needed from GE.

Comments presented herein reflect the benefits of that meetino. This neno is oroanized into two sections. Section 1.0 presents the summary of the GESSAR systems analysis review and was nrepared by BNL. Section 2.0 contains comments that are relevant to the seismic hazard and the seismic fragility analyses and was prepared by Jack A. Benjamin and Associates, a consulting firm retained by BNL for this study.

1.0 SYSTEM ANALYSIS In the prelininary review of the systens analysis of the GESSAR seismic event PRA, BNL examined the event tree and fault tree models that were develooed in the reports to describe the plant responses to an earthquake.

. System camponent and human actions contained in these models were evaluated with respect to their reasonableness and completeness. Since an earthquake is a ootential cause for disablina redundant trains or systems, attention was also focused on the treatment of dependent failures between systems. As a w

Meno to R. A. Bari fran I. A. Papazoolou Janua ry 24, 1984 Pace 2

~

result of this oreliminary review effort, the following areas of conern have been identified.

Critical Connonent List -

Since the GESSAR II desiqn is of a generic nature, it appears that previous seismic PRAs would provide a good startina point to identify a list of critical connonents for consideration in addition to those'which may be

. plant-soeci fic to GE SSAR. Reviev of Table 3-25, p.81, of Reference 1 indicated that onl.y a linited number of conoonents are included in the

- a nal ys i s. RNL conciled a similar list of critical cannonents based on previous seismic event PRAs and on a review of the internal event PRA systen f ault trees. This list is presented in Table 1.1 and it only contains conconents that are not addressed in Table 3-25. A discussion should be provided in 'the report to establish why these canoonents are excluded fran the analysi s.- If they are believed to have large capacity factors, a discussion should be included to exclain why the GESSAR components would exhibit such capacities in licht of previous analyses and the basis of assurance that the plant to be built would consist of components that are being characterized in the GESSAR analysis.

Instrunentation .

Also appearing in Table 1.1 are instrunent panels and display i ns trunentation. It is obvious that instrumentation in General is vital to the. operation as well as the safe shutdown of the plant. Failure of instrunent panels could lead to the loss of systen function. Sinil a rly ,

~

' failure of various sensors, such as level, or pressure, dependinq on the failure nodes and the numbe" of failures may also result in loss of systen functions. It is suggested that additional infornation be provided to address the failure of instrumentation due to seismic events and its effect uoon systen performance.

4

Meno to R. A. Bari from I. A. Papazoolou Janua ry 24, 1984 Pace 3 Disolay instrunentation failure is not considered to result directly in systen failures. Oftentines its principle function is to provide the operator with confirmatory information. In sone instances, operator actions as prescribed by the procedures reauire information from display instrunentation.

~ A detailed discussion should be furnished on the likelihood of ca,non cause display instrumentation failures due to an earthquake and the potential impact upon operator actions given the occurrence of these failures, h:

Relav Chatter The relay chatter phenanenon has not been included inthe GESSAR anal.ysis.

The effects of relay ' chatter can he sunnarized into three different catego ri es. The first case concerns relays that chatter in an earthouake but do not alter the systen state throuch breaker trips after the seisnic event.

Its inpact' upon the availability of a systen is considered to be ninimal. The second case concerns relays that chatte'r in an earthquake resultino in' breaker trips; however, resets of these breakers are located in the control roan and can be actua.ted by the operator to restore a systen'. A successful systen operation in this case is predicated on the recognition by the coerator that the systen is triboed off line and on the manual reset action of the operator.

Lastly, in the event that relays chatter resulting in breaker isolations, resetting of relays may have to be done at local panels away from the control roan. Moreover, prior to resetting, careful diagnostic procedures will have to be followed to ensure that indeed no faulted conditions exist. For instance, the in-plant ' electric circuit will be under this category.

Infornation should he provided by GE to address (1) the effects of relay chatter upon the availability of the GESSAR systens and (2) the modeling of subsequent operator action to recover breaker isolations.

Hunan Errors i- In the two GESSAR reports, it aopears that no consideration was given to the modelina of the increased stress on the operator as a result of an earth-l i

9 f!eno to R. A. Rari from I. A. Paoazoglou January 24, 1984 .

Page 4

- quake. Subsequent to the onset of a reactor transient or an ATNS, a number of oce'rator actions have been assumed in the GESSAR seismic event trees, a dis-cussion .should be provided to justify vhy the same human failure probabilities were used in light of a seismic event.

Event Tree Bill perforned a preliminary review of the GESSAR seismic event trees and f; requests that additional infoination be provided in the folla ing areas in order to facilitate the review,

1) In the develonnent of the GESSAR seismic event tree, the 3 diesel cal-

-non node failure is modeled explicitly, Figure 4.1.(1) How is the 2 diesel conron node failure (divisions 1 and 2) nodeled in the an-alysis?

2) It- aDoears that the hardware deoende.nces between the LPCI and the RHR

- systens are not considered. Provide a discussion on the treatnent of dependence between the low pressure core injection and the RHR systens and how it is nndeled in the event tree.

3) In the January neetina with GE, BNL questioned the definition of the Eng and ESCB functions soecifically with respect to their NOT-event definition. It appears that the NOT-event definitions of these functions in Fiqure 4.1 is not consistent with the definitions provided in Figures 4.4 and 4.5.(1) GE should furnish a clar-ification on these event definitions.

4) In the seismic ATWS event tree, Fiqure 4.3,(1) the level control function by .the reactor operator is not included; provide a discussion to support its anission.

5) As noted in the BNL review of the GESSAR internal event PRA, GE has assuned that in the event of an inadvertent ADS, low pressure ECCS is L-

Meno to R. A. Bari from I. A. Papazoglou Janua ry 24, 1984 Page 5 adequate in nitiaatina an AT4S if level control is naintained. It ap*-

pears that a similar assunction is also nade in the develonnent of the ATWS event tree. Explain in detail (i) why is there no degradation in the human reliability to control water level, (ii) the procedure that the operator has to follow, and (iii) how much time is available to perforn the task.

Fault Tree -

, 1) In Fiaure 4.12, a 50% value is specified for the failure of the shroud support and a 5", value is assioned for the hydraulic control unit.

Provide a discussion on how these valces are used in t,he seismic quantification and the basis for their derivation. If an internal GE docunent or calculation is referenced, a copy is requested for re-

~

vi ew.

2) In the GESSAR seismic fault trees, it is noted that failures of nunos and power divisions are nodeled as independent events, that is wi th no correlation. For instance, in Fiqure 4.11, failure of RHR ounns A and B are considered to he independent basic event. Sinila rly, the los s

-of power divisions l _and 2 are also independent. Provide a detailed discussion to show that in the event of an earthquake, these ounos or the different power divisions would not be sub.iected to common cause failures and that the assumption of independence is adequate and re-asonable for the GESSAR analysis.

2.0 SEISMIC HAZARD AND FRAGILITY This section docurents the status of our review of the GESSAR Il probabilistic seisnic' risk assessnent (PRA). Ile have read and studied

~ References 1 and 2-which docurent the results of the analysis. Dr. John Reed

. attended a neetinq at the General Electric Company (GE) in San Jose, California on 9 January 1984 At this neetinq GE discussed the calculations for this ca,ponent. He have reviewed these calculations. He also have

fieno to R. A. Bari fra, I. A. Papazoglou Janua ry 24, 1984 Pane 6 received a cooy of questions prepared by the USNRC which have been subnitted to GE for written response.

This section is organized into oeneral coMnunts, seisnic hazard analysis connents, and seisnic fraaility analysis connents. The status of our review and information needed to ca,plete air work is presented below.

General Comments References 1 and 2 do not provide sufficient information to perforn a

.conDiete review of the GES3AR II seisnic event analyses. Because References 1 and 2 are generic and do not apoly to a speci fic site or plant (i.e. , as conDared to past PRAs such as the ones conducted for Zion, Indian Point, and Linerick), the ultimate purpose and intended use of the GE inalysis is not clear. Based on a' creliminary review, the ,results do not envelop hazard and f racili ty data fra, PR As - suhni tted to the llSMRC to date. We request that GE state their chilosophy concerning how the seismic PRA analysis will be applied to specific plants. LThe ultinate use of References 1 and 2 should be defined by GE in order for us to determine if the intended objectives have been a ch ieved.

Many of the questions fornally asked by the-USNRC express the sane

~

concerns that we have. IIe have not repeated these questions and assume that they will be answered in the near future. Question 720.150 is particularly inportant, since this question addresses the safety factors for duration, danping,_ and . inelastic energy absorption. It appears to us that the duration factor is the same as the factor used to sh ft the hazard curves fran peak to i

ef fective ground notion. This concern should be addressed by GE in response to Question 720.150. .

Because the structural capacities are apparently hiah, the oroblen of desion and construction errors becones very important. In a practical sense this consideration could dominate the results of the analysis. Since GE has not generally included the effects of design and construction errors in their analysis, they should_ state why this issue of electrical conoonents also is t

~

"eno to R. A. Bari fran I. A. Pacazoqlou Janua ry 24, 1984 Pace 7 not addressed. GE should verify that this is not a problem for GESSAR II.,

Intergranular stress corrosion crackino has been a problem for GE plants in the cast. Is this problen pertinent to GESSAR II plants, and what effect will this problen have on the seismic capacity of oicino? Finally, what are the capacities of block walls planned for GESSAR II olants? Are block walls to be located near any safety-related eauionent?

Seismic Hazard Analysis Connents In the GESSAR II seisnic PRA a best-estinate seisnic hazard curve was developed for the highest seismicity GESSAR II sites, (0.4 of Ref.1).. The best-estimate curve is considered by GE to be a realistic, nedian-centered, unner-bound seismic hazard curve. Further, GE expects that the GESSAR seisnic hazard curve will bound site-soecific curves at a na.iority of the notential GESSAR sites at the 50 percent crobability level (p.13, Ref.1). The GESSAR II Seismic Event Analysis (Ref.1) was followed by the GESSAR II Seismic Event Uncertainty Analysis (Ref.2), which included an evaluation of the uncertainty in the seismic-hazard. An. initial review of these r.eports has been perforned.

Connents on these reports- are niven below.

The best-estirate seismic hazard curve developed by GE was based on an envel oping approach. In developing the best-estimate envelope hazard curve, the results of recent utility-sponsored studies and a U.S. Geoloaical Survey study that evaluated nround shaking hazards for the contiquous U.S. were used.

The envelope curve selected by GE is considered to be a best-estinate of the extreme values of the best-estinate curves at potential GESSAR sites. It should be noted that the sites which are potential locations for a GESSAR facility are not clearly defined in either GE report. General Electric Company should state what sites the seismic PRA analysis is applicable to.

In the approach used to develop the best-estinate hazard curve, the results ~ of four recent PRA studies were used. They are the Indian Point, Zion, Oyster Creek, and Linerick PRAs. The actual decree to which the U.S.

Geolooical Survey study results were used is not explicitly stated in Reference 1. The best-estinate GESSAR hazard curve was then subjectively

W Meno to R. A. Bari frcn I. A. Papazoglou Janua ry 24, 1984 Page 8 taken as the enveloce of the best-estinate airves of the above lis'ted four studies. Further, the GESSAR hazard curve was defined to have an effective acceleration truncation value nf 0.95q.

The basis for the 0.95c acceleration cutoff is not supported. GE should orovide a basis for this cutoff value.

There is no evidence provided in the GESSAR reports to support the statement that the GESSAR best-estinate hazard curve is in fact an upper-bound, or an upper-bound that will not he exceeded by 80 percent of the best-estinate curves at ootential GFSSAR sites. The arbitrary selection of the four PRA studies used in the GE study, and the sub,iective nanner in which the GESSAR hest-estinate hazard curve was sel'ected, raises innortant questions

- about the develoonent process and the full neaning of the results. It is not clear, in a probabilistic sense what the GESSAR hest-estinate curve reoresents, other than an enveloce of the four hazard curves considered in the

. study. ,

With respect to the uncertainty analysis for seismic hazard, a number of ,

conceras are raised about the nethodoloay, data base, and ultinately the final resul t s. In evaluating the uncertainty in the seismic hazard otrves, GE elected not to conduct a study that systenatically addresses each of the sources of uncertainty in hazard assessment. Rather, they used the results of a published study that polled a group of experts on their estinate of the annual frequency of occurrence of earthquake ground notion levels at various nuclear power plant sites. The results of the expert opinion survey were the basis for making uncertainty estinates.

With respect to the survey itself, several auestions as to its adeauacy

-are ra ised. The experts were given some data on the seismicity in the region surrounding each plant site and asked to provide estinates of the annual

-frequency of exceedance of ground shaking at each site. Within this fo rnat, it is a difficult task for the experts to rationally and consistently provide probability estinates for rare events. An alternative approach is to provide

- each. expert with the ooportunity to break the nroblem into more tangible parts (i.e. , seismic source _. attenuation, etc.) allowinq for a more systematic

Peno to R. A. Bari fran I. A. Papazoolou Janua ry 24, 1984 Page 9 evalution, which is less prone to overlookina significant sources of uncertainty and is more easily perceived by the experts in a orobabilistic sense.

The study used .in the GE analysis was oublished in February 1975, and undoubtedly performed in 1974 In the last ten years, consideranle work has Deen done.in the area of seisnic hazard assessnent, incluNinq solicitation of expert opinion, geologic and scientific investigation, etc. Consecuently, the use of the 1975 Okrent study as a basis for uncertainty estinates is seriously que stioned . It should also be pointed out that the number of experts used in the study (7) was relatively snall. Also, the degree to which those who participated in 'the survey can be considered probabilistic seismic hazard experts for the entire ll.S. is questioned. -

Me believe that each of the individuals who Darticipated in the survey is a recoanized exDert in one or possibly more areas of seismic hazard evaluation. However, none of the experts can Se considered, nor would they clain, to be experts in all the areas required to nake probabilistic hazard evaluatidns. These areas include, probability, statistics, seisnolony, geoloay, ground notion attenuation, local and regional tectonics of each site ,

being investigated , etc. The extensive range of expertise required to make probabilistic seismic hazard assessments is one of the prinary reasons for breaking the hazard assessnent into integral parts, allowing the experts to deal with each part of the analysis individually.

The results of the expert ooinion survey were used as a basis to estinate the coefficient of variation of the annual frequency of exceedance of levels of ground shakina. A preliminary review indicates that the results (Table 2-3, Ref.2) for accelerations less than 0.50q are reasonable in that the uncertainty estinates are consistent with resnect to previous site-specific studies and expert opinion surveys. However, at higher acceleration values, which is the region that da7inates core melt frequency estinates, the un-certainty values are too snall. This also partially explains the reason for the relatively narrow distribution on seismic core melt freauency (Fiqure 4-2, Re f. 2 ) .

"eno to R. A. Bari fron I. A. Papazoglou

. Janua ry 24, 1984 Page 10 The following list sunnarizes our initial concerns with reoard to the GES-SAR seismic hazard analysis. He reauest that GE respond to these c'oncerns.

. The potential GESSAR sites are not identified in References 1 and 2.

. The nethodolony used to deternine a best-estinate hazard curve is probably not adequate to reet GE's ch,jective to nroduce an upper-bound, best-estinate hazard curve for potential GESSAR sites.

. No evidence is provided to suoport the state,ent that the r dSSAR best-estinate hazard curve is expected to bound nore than 80 percent of the best-estinate curves for potential GESSAR sites.

. The basis for the 0.950 acceleration truncation value for the hazard curves-is not orovided.

. The 1975, study hv Okrent does not apoear to adequately orovide a basis

.for the seismic hazard uncertainty estinates. A nunber of concerns're-lated to the expert opinion survey were raised above.

The uncertainty estinates at effective acceleration levels greater than 0.50g appear to be low.

Seisnic Fraaility Analysis Comments The basis for the fragility analysis is past PRA studies and data which GE has obtained. The information documented in References 1 and 2 is not sufficient to perforn a critical review of the GESSAR II seismic fragility an-alysis. Median capacities are apparently based on generic calculations using the sane basic procedures used in past PRAs subnitted to the USNRC (i.e. , the so called Zion method). Ne are not able to judae the adequacy of the median values without studying the calculations performed by GE. The variabilities used in the analysis were based entirely on results from past studies, and specific values for structures, components, and equipnent were- not developed

- by GE . The followina cannents and concerns are based on review of References 1 and 2 and the calculations for pipina which were provided.

m e, .

t. (;

1 i; j g--q\.-

a

}\  %.?

'C ' Meno to R. A. Rari fron I. A. Papazoglou y

, _' y . January 24, 1984 F; Page 11 c.

u; ,, . ,-

s \ ,

The coef cie'nts of variation assuned in Reherences 1 and 2 for the fragility anal /Ms are generally low compareds to, past PRA results. It is stated in the [alculations for piping ttyat leaer values were used because of

~

additional desi('n considerations to be olace upon GESSAR plants. It is not obvious why this is so. In fact, because of the nature of GESSAR and the generic analysis performed) the' uncertainty should be greater rather than less cor.ipared to analyses for specific plants. It is reauested that GE provide the bases for the coefficients of variation on the structural capacities assuned

~

in the analysis. .

As stated above, there is not sufficient 'information provided in Ref-erences 1 and 2 to- complete the review. At the neetina with GE on January 9th conolete calculations- for Reference 1 were reauested for all structures listed in Tables 3-2 thr ouch 1-13, the eiqht camponents in Table 3-19, and the eleven components in Table 25. These calculations are needed to conolete the review and to deternine the adequacy of the analysis.

The following additional infornation is also requested. The naqes cited

~

refer to References 1 and ? as indicated:

Reference 1 pg.24: . References 9,10, and 14 do not appear to be correct. What are the correct reference numbers?

09.35: What is the basis for the strenath margin of 1.3 given at the bottom of the pace?

pg.38: What is the basis of the 1.2 value assumed for Fss?

04.40: What are the data which substantiate a value of beta equal to 0.2 for inelastic analysis?

9 i..

P Peno to R. A. Rari from I. A. Papazogicu Janua ry 24, 1984 Pace 12 90.42: was the 10 per* cent increase cited on this oace also assumed for struc-ture capacities? If yes, then 'ere all analyses for the structures perforned usino the tine history analysis nethod rather than a re-sponse soectrun analysis nethod?

pg.50: The basis for the 30 cercent increase for the ef.fect of dynamic yield stress as conoared to static yield stress is requested. Note that the explanation aiven in the calculations for pipinq for this factor is not clear. -

p9.50: - Justification should be provided why the dancina narqin for structures

, is aoolicable to the capacities for conoonents and equionent, since at the eauionent failure level the supportina structure may be uncracked and still elastic. -

Reference 2 .

pg.3: The basis for the equations for and ' is reauested.

Table 2-3: What are the logarithmic standard deviation values corresponding to each coefficient of variation used in the analysis?

The calculations for piping provided by GE were reviewed. It has been shown in past. PRAs the supports are generally weaker then piping. However, in the GE calculations it is stated tha't support failure due to seismic loads is precluded since supports have been qualified for more severe loads. The basis for not considering failures of the piping supports is requested. Considera-tion should be qiven to supports which are desiqned essentially to resist only seisnic loads. Also, cons.ideration should be given to support hardware which is designed for AISC requirenents as opposed to ASME criteria (i.e. , at the buildina/ support interf ace).

In order to neet that deadline we need to have all calculations by January 31,-1984 and the response to other questions and concerns by February 8,1984

f. . :. .

Meno to R. A. Bari from I. A. Paoazoglou Janua ry 24, 1984 Page 13 References

.1. 'GESSAR II Seismic Event Analysis, General Electric Company, Sept. 1983.

2. - GESSAR II Seismic Event Uncertainty Analysis, General Electric Company, Decenber 1983.

Table 1.1 BNL Critical Component List Condensate Storace Tank Diesel Oil Storace Tank Diesel Oil Day Tank Burned Pipe Service Water Puno Horizontal RHR Puno -

. Diesel Generator Control Panels Cable Trav dc Rus Battery Rack Diesel - Generator Heat Vent '

RPV Support Reactor Internal s.

Instrument Panels----------

Of splay Instrumentation----l-- see Instrumentation IAP/dn 8

, t GENER AL @ ELECTRIC NUCLEAR POWER systems DMslON GENEQ.At M/C 682, EE;;*(K CO)WANY e ' 75 .uRTNEa %EMEMFN-222-83 408 925-5040

• SAN JCsE C AMCWA U '25 JN F-086-83 December 2, 1983 -

U.S. Nuclear Regulatory Commission Office of Nuclear Reactor Regulation Washington, D.C. 20555 .

Attention: Mr. D.G. Eisenhut Division of Licensing

SUBJECT:

IN THE MATTER OF 238 NUCLEAR ISLAND GENERAL ELECTRIC STANDARD SAFETY ANALYSIS REPORT (GESSAR II)

DOCKET NO. STN 50-447 APPENDIX 15E - STATION BLACK 0UT CAPABILITY Attached please find a draft of new GESSAR II Appendix 15E pertaining to station blackout capability. This appendix concludes that the GESSAR II station blackout capability exceeds ten (10) hours. The assessed capability assumes credit for operator actions that are straightforward and where means exists to enable.the operator to execute the action. Where features and/or equipment are not present, potential design improvements are recommended.

It is anticipated that upon completion of NRC review a formal amendment on the GESSAR II docket will be submitted. This is anticipated to occur in early 1984 If there are any questions on the information provided herein please contact J.F. Quirk at (408) 925-2606 or J.N. Fox of my staff at (408) 225-5039.

Very truly yours, k

'Glenn

/[/

. SfferW, Ma. a er Nuclear Safety & Licensin Operation Attachment cc: F.J. Miraglia (NRC)

C.C. Thomas (NRC)

R.M. Ketchel (GE-Washington Liaison Office)

L.S. Gifford (GE- Bethesda Liaison Office)

R. Villa (GE) f  % 'N

. 9 . 3 D

GESSAR II 238 NUCLEAR ISLAND l

i f%. D, YW '

- ,t % !  !

L.I :I bu*{  ; I i APPENDIX 15E STATIGN BLACK 0UT CAPABILITY e

mi i , i

=  !

APPENDIX 15E CONTENTS SECTION TITLE PAGE 15E APPENDIX 15E - STATION BLACK 0UT CAPABILITY 15E.1 INTRODUCTION AND CONCLUSIONS 15E.1.1 Introduction 15E.1.2 Conclusions 15E.2 DEFINITION OF STATION BLACK 0UT 15E.3 INDICATION OF STATION BLACK 0UT 15E.4 INSTRUMENTATION REQUIREMENTS 15E.5 PLANT RESPONSE FOLLOWING A STATION BLACK 0UT 15E. 5.1 Areas-15E.5.1.1 RCIC Room 15E. 5.1.2 Remote Shutdown Panel Area 15E.5.1.3 Suppression Pool 15E.5.1.4 Drywell 15E.5.1.5 Control Room-15E. 5.1. 6 Fuel Pool 15E.5.2 Energy Supplies -

15E.5.2.1 Pneumatic Supply 15E.5.2.2 125 VDC - Bus E 15E.5.2.3 125 VDC - Bus F 15E. 5.2.4 125 VDC - Bus G 15E . 5.2. 5 125 VDC - Bus H ISEA ATTACHMENT A TO APPENDIX 15E - ACRS QUESTIONS

• PERTAINING TO AC/DC POWER SYSTEM RELIABILITY 15EA.1 DC RELIABILITY 15EA.2 GRID RELIABILITY

~

15EA.3 DIESEL GENERATORS 15EA.4 LOW POWER TESTING / SIMULATED LOSS OF 0FFSITE POWER 15EB ATTACHMENT B TO APPENDIX 15E - RCIC ROOM HEATUP DURING A STATION BLACK 0UT 15EB.1 PURPOSE 15EB.2 INTRODUCTION 15EB.3 MODELING AND ASSUMPTIONS 15EB.4 INPUT PARAMETERS 15EB.5 RESULTS AND DISCUSSION

I APPENDIX 15E TABLES TABLE TITLE PAGE 15E-1 Variables Assessed For Station Blackout Assessment 15E-2 Power Supplies to Instruments Needed For a Blackout APPENDIX 15E FIGURES FIGURE TITLE PAGE 15EB-1 RCIC Room Temperature Response Following a Station Blackout

~ 15EB-2 RCIC Room Temperature Response Following a Station Blackout - $ensitivity to High Water Temperature 15EB-3 RCIC Room Temperature Response Following a Station Blackout - Sensitivity to Low Steam Leakage Rate N , , .

' I o

'15E.1 INTRODUCTION AND CDNCLUSIONS 15E.1.1 Introduction This appendix is provided to demonstrate that the GESSAR II design has substantial capability to prevent a core damaging event well beyond the two-hour value recommendedb 'y NUREG-0626 and assumed in the Probabilistic Risk Assessment (Section 150.3).

~

Attachment A contains respon'ses to pertinent questions on station blackout of interest to the staff. These are addressed in more detail in other parts of this appendix.

15E.1.2 Conclusions The GESSAR II station blackout capability exceeds ten (10) hours. The

. assessed capability assumes credit for operator actions, that are straight-forward and where means exist to enable the operator to execute the action.

Where features and/or equipment are not present, potential design improvements are recommended. These operator actions and potential design improvements are summarized below:

1. Operator Actions

a. Manual RPV Water Level Control with RCIC.

b. Shift of RCIC pump suction to the condensate storage tank.

c. Vessel depressurization with SRVs to about 200 psig. Maintain vessel pressure above 150 psig with manual SRV control.

2. Potential Design Improvements

a. Provide manual logic override of the RCIC suction transfer signal and test line closure signal from the control room.

s n

b. Provide Enhanced Water Level Instrumentation (currently under review for Appendix 1D).

c. Provide alternate power supply to RCIC gland compressor.

An ongoing evaluation of the 125 VDC battery capability is in progress.

However, if necessary to ensure 10-hour capability, emergency DC . bus cross ties, or larger battery capacity, or other methods will be identified.

In addition to the above-actions, the following contingency actions could be taken to provide even longer duration capability are:

1. Provide override capability for the RCIC room high temperature isolatfori logic to be used if room temperature exceeds about 150 F.

2. Extend SRV pheumatic supply by replacing air bottles if depleted. .

A connection outside the fuel building would be more convenient.

15E.2 DEFINITION OF STATION BLACK 0UT p

Station blackout refers to the total loss of both off-site and on-site a.c.

electrical power. In draft information pertaining to proposed Regulatory Guides, the NRC consulta..ts refer to " Emergency AC" loss in addition to offsite power loss. This could be interpreted as the Division 1 and 2 Standby Emergency Diesel Generators. Both HPCS and RCIC operate at high pressure and can be considered redundant water sources available for maintaining core cooling during design basis assumptions that assume a single failure (i.e.,

such as a D-G) . This configuration is believed to be adequate to comply with the proposed regulatory requirements. For purposes of this assessment,

however, a failure of the HPCS diesel generator has been assumed in addition to loss of offsite power and the division 1 and 2 diesel generators thus providing a more severe impact on plant systems and the station battery.

A one-line diagram of the GESSAR II design is shown in Figure 8.3-1. Three divisions of 6.9 kv on-site power are provided; two by standby emergency .

diesel generators (in addition to preferred and alternate off-site power sources); the third by an off-site power source and a separate and diverse diesel generator dedicated to division 3 eletrical power. Division 3 supports the High Pressure Core Spray (HPCS) system and all of its supporting auxiliaries. .

The GESSAR II design also includes a steam turbine driven Reactor Core Isolation Cooling System (RCIC) which operates in an emergency independently of a.c.

electrical power. This system is designed to provide high pressure makeup to the RPV during isolation events and would thus be initiated automatically ,

during a postulated blackout event. The plant response with RCIC alone has been reviewed, and the duration capability of the GESSAR II plant in excess of ten hours has been verified. This configuration is consistent with the station blackout Jefinition in the Probabilistic Risk Assessment (Section 15D.3).

In:the evaluation certain assumptions have been made:

o No Loss of Coolant Accident (LOCA), stuck open relief valve (SRV) or failure to scram concurrent with the station blackout is considered.

l l

l 1

1

i .

o In evaluation of equipment, some capability beyond environmental qualification Ifmits has been assumed. In assessing the ultimate

'ailure capability of equipment the judgement of senior General Electric engineering personnel has been relied upon to provide guidance. Such judgements are explicitly call out in the following sections.

o Operator actions are identified where adequate time and skills would be expected to be available to a typical operating plant staff. No extra-ordinary actions on the part of the operator are only assumed; rather, uppstraightforward, simple actions are allowed.

o No credit for off-site assistance from a utility maintenance crew using portable electric generators or batteries has been assumed for this assessment even though this possibility may exist within the time frame of interest. Such capability might be considered by an applicant to improve the restoration time for on-site emergency a.c. power if the situation warranted.

i s G . 3 .TMoi C #T\od OF S TAW ** ts'A C WOVT The station blackout event is daracterized by a loss of all off-site power (preferred and alternate feeders) and a loss of divisions 1, 2 and 3 of on-site a.c. Power. As noted in Section 1D.2.3.33 of the assessment against Regulatory Guide 1.97, the class 1E power distribu-4 tion systan monitcrs voltage on the three 6.9 kv a.c. buses and the four 125 V d.c. tau =a. This indication is displayed on panel P900 in the main control ro m. A potential station blackout event would be

first noticed by the plant operators by a dange in the control roca

li@ ting whid would alert him to evaluate both the plant and the electrical distribution systes status. By observation of the loss of bus voltage on the 6.9 kv talses "E", "F" and 'G" and the breaker l position- for incming voltage to these buses, the operator would be alerted to the sresence of a potential blackout event. Voltage indicr. tion on the d.c., buses E, F, G E H would assure the operator that l power is available to control the event. l 1

Prior to coM_WLg the various operatcr acticms needed to mitigate a blackout event, the operator must distinguish between a short duration event and a prolonged blackout. A short duration event would be one in whi& restoration of an off-site or on-site a.c. power source would occur Erior to develognant of conditions requiring the operator actions 1

7-d w -re-- r --me - - - , - - ---. . - - .-.s -t-- +-w,- --,m--+-+.e++---ww== --we- - - w---* - -=~=wwv=w--+,w --w-e~-u-w---~**me---,-

• m --e vwee--

defined later in this suppleant. Minimizing the time to recognize this i event is important so that the potential drain on the batteries is controlled.

Upon recognition of the a.c. power source failure, an anv414=ry operator would be sent to and of the diesel generatz roms to attempt a manual I start. Simultaneously, the witrol rom operator abould atteelt to start and diesel fra the main control rom. In addition, the systen ~

dispatcher would be contacted by the shift supervisor to detemine the status and likelihood of off-site power restoration. Accomplishment of these activities in addition to those related to controlling vessel water level and pressure is expected to take about 30 minutes.

S us recognition of a station blackout event and the initiation of any bl'ackout specific operator actions is expected to be delayed for about 30 minutes.

ss E .4 2 bi sTfus ae Art-Amota n.sht p-asM sNT f Instrtunentation required to monitoc plant status during a blackout event has been selected fra a review of the type A through E variables discussed gf wdix

.=ic .1D whid is the response to Reg Guide 1.97 require-monts. This list has been atxpented slightly to account for specific variables su& a room tanperatures and certain valve and breaker position indications whid are needed to deteminc plant conditions.

tSE-l Table 3lEpsts the variables considered and whether w not they are needed for the blackout sequence. The basis for selection generally is based on the need fac the operatcc to follow Dnergency Procedure Guide lines (or take other actions which may later be established) during the period of interest. As such type A variables are identified as need-ing indiation during the blackout event while variables whidi are more representative of monitoring core damage oc breaks of the reactcc coolant boundary oc effluent release are em::lu&d. m i55-7 2 Table 4=fpowsthepowersuppliesintheCESSARgdesignforthe instrtments needed. All indications needed to follow the blackout event are oc will be powered from 125V d.c. sources, coMd

, 'Ihe applicang6 },:covide d.c.A b ek u p .b power to the condensate storage tank lwel indicatcr and to ensure local control rom tanpera-ture indication as available.

1

• S m ce_ a\ u se s s h e - , m 3 -fv o - a e,o1\ M shbn L,l. o. c k o d % t cw w AM m en sb d e.n y, loa , o. s e b b up at e d -boei. 3 gig,f p m y (tKSPdo.25 FOLLONuJG A STR O Q SLAC% CUT

'Ibe key plant areas whidi could potentially effect the ability of the plant desigri to acm-nrhte a station blackout are:

o RC C room o Renote shutdown pmel area o Supsression pool and containnent o DONell o Control rom o Fuel pool In addition noMectrical a.c. plant energy sug: plies will be consumed

- and need to be addressed to assess the plant capability. 'Ihese are:

o Pnetsnatic Air Stpply Systen (ADS) o D.C. Power Distribution Systen

'Ihese areas and energy supplies will be discussed in subsequent 5 -

i o ve mw+.1 sections. An estimate of limiting condition, design or operatec actions needed are noted in eadi.

-to-

tsu.5.I Ave.as is.u.s.1.I pg nere n~-

CL. Itamaan for Concern o Room temperature increase without area cooling could cause a loss of RCIC control due to equignent failure.

o Isolation and turbig trip due to leak detection systen trip. (Trip setpoint aggrox.170 F) could prevent RCIC fran operating .

volve.s leecnw s o

Staan damage line on drainA.may restart fail after air supplyAexhausted causing systen i

b. Plant Haapanna o Approx.122 F in 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (w/ CST suction) o Approx.133 F in 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (w/SP suction) , See AM achmant B o Asprox.101 in 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (w/10 lb/br stemn),

Cribir al t' - ---e-J n Limitation E Differential Coil Approx.170 F water tenp.

Macpetic Speed Senacc 225) ,

capability Instrumentation 212'F >12 hrs C. Aasmed Oparator Actions o Manual switch of RCIC suction to CST at about 30 min.

o Override RCIC hi@ tenp isolation if roan temp > apgrox.150 F (not expected) o Manual RN level control of RCIC to avoid L8 trip and restart.

d. Potential M~14 Hearf emm/Actiona 4

o Ensure werride capability exists for RCIC roan isolation signal.

o Ensure override capability fac RCIC suction transfer.

o Provide logic changes to permit low flw RCIC injection. Requires override capability on test line to CST to obtain flow split between Tr return and vessel.

-IW

155.5 . t . 'L Asee:j n- ,+. ni+rwn panel Area

4. Reason for Ctm mrn o RCIC electronim cnuld fail if area tanperature exceeds 150 F.

o Accesa needed if control roan tacomfortable er electronics erratic.

b. Plant Response o Not evaluated, but very little heat source ince Renote Shutdown Station pane t1 @brgized tmtil control tr er switch is thrown.

o Expect area temperature to renain

<150 F for 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br /> C. Anamned hraene Adians None.

d. Potential Modifications / Actions None.

l 8

l tL-

l 15G.5.t.3 me -.mainn pnn1 A '

A.Raaman for Concern o

Hig) 175 7 suppression and reduced pool tenperature lube oil coolina could cause to g,cJ c, NPSH limits (approx.

j o High supg ession pool level causes suction transfer.

o High contairunent air temperature may cause erratic RPV indication.

o High suppression pool tenperature and level increases containnent loads,

b. Plant Response

. twat T T L Notes: Tcbased on Tsp + judgnent

_ 3p (hrs) 4(F) (g) 7 (ft) based on Table 15D.2-2 T@ calculation 1 135 100 +2 5 190 175 +5 T hility >10 hours.

10 220 220 9 weir 15 225 225 9 weir 20 230 230 9 weir Instruments qualified to 185 F; capability likely to 250 F.

C, A--a omrator Aceinns o Manual switchover back to CST within 1 hr. eliminates potential NPSH problen.

o Maintain vessel pressure below heat capacity tenperature limit per EEGs

- ensure written gocedures contain heat capacity tenperature limit curve - may need to exceed heat ca p city taopetature limit slightly after approx. 6 hrs, but acceptable because no additional depressurization required. Ctmsistent with EEGs.

d poenneini Mndifications/ Actions o Ensure manual override capability for RCIC suction transfer

-nL

GG SS AL E 2.3 8 NO cusOE T3t Aw 0 ATTAC N M1ENT A To Appswoi V 15 E AC lt S GLUSSi\0NS Psg_TAimNG To A C /D C POWstt. SY STE M assut A G t L WY

\S E A.I O C a.st.t 4su \ rY Ouestiorn The PEC Staff has issued a report (ICREG-0666) on the reliability of d.c. power systen in which a 2-train d.c. systen found to meet mininun NRC requirenents was evaluated. As a result, the d.c. power systen was identified as a potentially high contributor to core melt.

'Ihe applicant could be asked what his assessment of his d.c. systen is and what consideration he has given to the recommendations of lbREG-0666.

Response: We do not favor the use of such a minimum systen as considered in IUREG-0666, Ebr example, it has a single bus tie breaker with too much potential for The G SSAQ E d.c. cross-connection ign allowed common cause failure. Gurgoriginal catability with .dlal cross-tie breakers and double key inter 1ccks. GE d+ M*+ h c" 33- " "*'o-

" S *b M

c. o #Cw W c, \N if c o.w N skom 4-(p a.,$ 4 i do4 J nok cowhe *k k C

% ct . c. ss c.a.pabs p 4% Mghd A+$ -

The following is provided in response to the recommendations in NUREG-0666:

(1) Prohibits certain design and operational features of the d.c.

power systen such as use of a tie breaker which could compromise 3C divisional independence. As noted above, GESSARpcomplies although we believe cross-connection capability is appropriate for specific conditions during shutdown and occurrences which require last resort flexibility (such as station blackout). ESSAR s four t

safety-related batteries, each of which has two chargers so that charger maintenance does not require use of cross-connections nor cause draw-down on the battery.

(2) Addresses testing and maintenance activities. These are accom-plished by the applicant. We agree with these r m nda-E tions, and the GESgdesign allows their implenentation.

(3) Requires staggered test and maintenance activities to minimize the potential for human error related common cause failure. This is controlled in the field, but we agree that these actions are appropriate.

(4) Requires design and operational features to be adequate to maintain reactoc core moling in the hot standby mndition follcuing the loss of any one d.c. power bus and a single independent failure of any other systen required for shutdown cooling. Although we cannot disagree with the intent of this recommendation, a judgnent as to what features are needed should be tenpered with an assess-ment of the reliability of the d.c. Power loads and sources. We have concentrated on maintaining full separation and independence between division 1 and division 2 d.c. systens to provide this reliability. ' .

6With four independent d.c. systens and with three inde ndent a.c. sfsk h N @:N $$sYan

" pability mwh wj O.4 N U R.s - 0 666 acow a. o^,

For example, a potentially adverse capability loss would follo.i frcan the loss of both RHR systens, but the suppression pool can

store decay heat for several hours, during which it may be possible to recover active decay heat r eoval.

t SEA. 2. Grid Reliabilitv Onaeime What is the applicant's assessment of grid reliability and what procedures exist for restoring offsite power to the plant in the event of this loss.

Response: 'Ihe grid is the responsibility of the applicant, and we assume he will meet the NRC requir ments in this area. On loss of nor:ral preferred offsite power, there is autmatic transfer to the alternate offsite power source and, if necessary, to the onsite diesel generators.

Restoring preferred power is accomplished mcnually by the control rom operatoc. '1he specific procedures for restoration of power in the switchyard or transmission systens would be developed by the applicant.

Station Riackout Analysis cuestiont What are the results of the applicant's station blackout I

analyses? Has the applicant made a best-estimate analysis of the accident sequence and evaluated what might be done to improve the plant, or has a conservative analysis been made with a mre melt assumed after sane specified degradation of the battery?

Ws evaNahow asyndJ No b o& #1N\ov' t Resnonser A - Our best-estimate analys s to the extent that it is complete is the primary subject of this supplement. We have identified potential systen-design and procedural improvenents, and we wd l % pl *med Mw we on eencure**** ko~ +b e N " C +b a+ +k*1 s aks s 9a kom a s s \ v-*. 4 k.a. t 1 s w< .

3o-

~

, i Our proL@ilistic risk assessment considered station blackout capability in a conservative manner (core cooling lost in two hours due to battery depletion and loss of RCIC control). We believe the more realistic treatment considering automatic and manual d.c. load shedding shows a substantially longer capability.

15GA.3 Diesel Generators Ouestient What is the applicant's assessment of his diesel generator i syst s? To what extent has LER and operating experiences been us2d to improve the design?

Response Our HPCS diesel generator has undergone extensive testing '

)

(including 300 tests without failure) which has been &cumented for the l l

NRC. Frcm this testing and from field experience we have high confidence ,

in the design. Extensive review of the design specifiation, the installation design and the auxiliary syst e design for the larger diesel d e-n sh.4e.s l generators (division 1 and 2) N high availability from these mits. .

ISEA. 4 Low Power Testinc/Shnolated Ioss of Offsite Power l  !

Onactiont Has the applicant performed low power testing and a simulated )

loss of offsite pwer test? If so, what are the results and what has the l applicant learned?

Eesponse:

Tk i s s s n vss e ov .s Al,43 of' %.a A pph c.Q ,

6

-n-i

- _ - - _ .._,_-_.,____,__..___.--.--___________.m.,

.. - I 15E .5. h +

3 ag mv=11 OL. Itamaan for Cancern o Hi@ drywell t:e@erature could cause RN level instrtsnent reference leg boiloff.

o Hi@ drywell temperature mi@t exceed qualification levels for drywell equipment.

o Higt drywell tagerature could cause SRV solenoid failure. '

h. Plant Itasponsa Approx. 135 P during plant operation

<270 F prior to depressurization at 30 min. Capability: ,

unlimited -

<200 F after depressurization to 200 pi l Drywell equipnent qualified for >300 F ,

C- Aamumed Operator Acticms o Depressurization to approx. 200 psi to limit drywell haatup, o Maintain gressure >118 pi to avoid reference leg flooding, o Maintain R W water level apsrox. + 20" on Enhanced Level Instrument.

-o, . Itecommanded Modifications / Actions o Enhanced water level instrtsnent (ELI) compensates for drywell and i contairuneet tanperature effects. (Previously reconsnended. See S W Wft A p p W ID.)

'f 5

15 5.5. t . 5 omtrol par =

vereew4

4. Reason for Cenern o Hic $1 control rom tanperature could cause computer / microprocessor controls to fail.

o High tangerature could make the control rom uninhabitable.

b 21AntJkEEmat o 3GCC flocc section heat sinks expected to prevent Capability:

heatup above 1057. miimited.

o H eidity could become tmo m fortable but not '

minhabitable.

Microprocessors (ELI, ERIS, etc.) unreliable above approx.105 F but backup information is available at Renote Shutdown Station (RSS) .

C. Assuured Operator Actions o Transfer control to renote shutdown station (RSS) if cont.rol room beames tminhabitable. (not expected)

d. Potantial Modificatiang/ Actions Nee.

l l

l l

l 6

- I V~ ~

IS E. S. I . 6 wpi nel _. --

O. Raasan for Cnevwrn o Loss of fuel pool cooling could cause fuel pool to boil away, le. JW mt. Responsa o Approx. 14 hrs to boiling Basis: Judgnant Erobably longer with o Approx. 77 hra to fuel uncovery less hot fuel Capability >75 hrs.

C. Anmunad Ouerater Adinrur None, but SRV air bottle replacement (see Imetmatic supply) could be hampered by fuel building environment after approx.14 hrs.

d,PotentialModificatieng/ Actions Consideration of moving extra air bottles to corridor outside fuel building. Not re.plired for station blackout.

t d

16-

_ . , - , _ _ . , _ _- - - . . -.r,.,-

15 D.5.1 Enev% S w p p b __

15 E. 5. 5.1

- , T1y 4 an.i==H e .cien1y G- Ma;1gg sou. ons of consungtion o ADB/SRV -

o Drywell and containment vacutan breakers b Entdanted naration (5000 @ available)

SRV Depressurization aggrox. 50 actuations 9 8 G/ actuation = 400 &

Ongoing SRV use approx.

2 x x8 = 240 CFH Leakage 9 1 CFH/ valve x 8 valves = 8 CFH Di Vacuum Breakers approx. 9 x 2 VB = 4 CEE gg t'+=1 apgrox. 250 CPH 250

= 18 bra. Capacity >18 hrs C. Operator Actions to Extend naration o Air bottle replacunerat after depletion possible if necessary (not expected) .

o Rotate use of ADS /SRV valves to penait time for acetanulators to radiarge and give preference to Division 2 ADS /SRV values.

o Monitcr SRV position indiation to indiate need for switch to other values (valves close when air sigply lost).

d,PqtantialModifications l None

)

4

-\~1-

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

tss.s.2.t

. , _ _, p l is wr - = = E

4. flaiat sources of corunmption See Table 8.3-6

b. Est.l. mated Duration (1950 anp hours (AH), 2 hr) 11CIC Gland Congressor Modification (see belw) delete 58A Capacility Shed load apsrox 35A (sea belw) >

• hrs.

Steady state load apsrox. 251-58-35 = 158A C. Ormraene AcHenn fa Extend Duration o Shed the follwing loads (at apsrox. 30 min.)

IMS panel H13-P669 (NSPS) - 25A from lEPS inverter Bnergency li@ ting (fuel building) - 10A .

d.PotentialModifications o Power RCIC gland ocapressac from an alternate source, o Delete 125 VDC emergency lighting systen except for control tuilding or move to Bus J.

o Provide Boergency crosstie capability with dual crosstie breakers and double key interlocks if needed for longer duration.*

o Prwide larger capacity battery if needed for longer & ration *.

I

• 1he capability of this battery with load shedding is being evaluated. If

. the estimated duration is less than about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, the addition of crossties oc expanded battery size will be reviewed to determine the optimum configuration for achieving a 10-hour capability.

15 E . 5. 2 . 3 1 M vrr - me p g M;- ry g A . Major 5:am== of car ==+4nn See Table 8.3-7 -

b . Estimated Duration (1500AH,2hr) ~

Shed Loads approx. 40A (see below) Capability Steady State Load = 175 -40 = 135A > " hrs.

C. Operator k+4nna to Extend Duration Shed the following loads at approx. 30 min.

- MIS panel H13-P670 (NSPS) -25A

- Doergency lighting -15A d,potentialModifications N.

o Delete 125 VDC emergency lighting in auxiliary building o "rovide larger capacity battery if needed for longer duration"

• Ihe capability of this battery with load afiedding is being evaluated. If the estinated duration is less than about'# hours, the addition of croesties oc expanded battery size will be reviewed to determine the optiman configuration for adileving a 10-hour capability.

-t9-

c - .

Is s.s. 2. 4

, - - . A 'y g in une ami a g mwn of canni=.+4nn See 222.e 8.3-8 Estimated Duration (400 AB, 6 hr)

Shed Loach = 25A (see bel w) car =hility SS load = 78 -25 = 53A >

• hrs.

Operator Aceinns to w and Duration i

Shed the feliming load at approx. 30 min.

MIS Innel H13-P671 (NSPS) -25A potent ial Mnairicatinna

+

Larger capacity battery if needed for longer duration.

/

• The capability of this battery with lead shedding is being evaluated. If the estimated diration is less than about!## hours, the addition of crossties oc +-dr.d battery size will be reviewed to determine the optimian confic31 ration fee achieving a 10-hour capability.

9 to -

. i 150,5.1 5 ns; vrr ni= n Maior sources of consmptim See Table 8.3-9 Estimated Duratim (425 AEi, 2 hr)

Load Shed = 25A m=hnity SS IAnd = 100 -25 = 75A >

• nrs.

Dparator ActiCE2 fd2 Extend Duration Shed the following load at approx. 30 min.

Shed 196 Penal H13-P672 (NSPS) -25A Potential Modifications None I

10 t

'the capability of this battery with lead edding is being evaluated. If the estinated duration is less than about hours, the addition of crossties or aparded battery size will be reviewed'to detemine the optinna configuration for achieving a 10-hour capability.

I 4

e

-tt -

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

nae es. i s E- t vmwEES gapen PCR STMIDN BUCEDUT ASSESSENT l

RG 1.97 RG 1.97 Discussion Needed in Black-Variahlm h Catamry M out Carw e en?

naaetivity centrol Neutron Flux A,B 1 1D.2.3.1 No*

(value, rata, trend)

Control Rod Position B 3 1D.2.3.2 No*

Boron Concentration B 3 1D.2.3.3 No (sampls)

cne. cnni4no ,

! Coolant Level in the A,B,C 1 1D.2.3.4 Yes Reactoc (value, trend)

Maintaining Reacts Coolant

.sva*== intmarity

RG Pressure A,B,C 1 1D,2.3.5 Yes (value + alarm)

Drywell Stanp Level B,C 3 ID.2.3.6 No (value + alarm)

Drywell Pressure B,C,D 1,2 1D.2.3.7 No i

! Primary Contaiment E 1 1D.2.3.8 No Area Radiation C 3 iSuppression Pool A,C,D 1.2 1D.2.3.9 Yes

Water Level Maintaining Containment Tntaarity l

lPrimaryContainment . B 1 1D.2.3.10 Yes**

i Isolation valve Position (Excluding Qwck Valves)

, Primary Containnent A 1 1D.2.3.ll Yes

, Tenparature

• A2WS plus blackout is not considered in this study. Failure to scram can be inferred frcan abnormal water level and pressure respnse.

• • Plus RCIC minin n flow.

11-

I naLE ass t S E-1 l

)

VARIABug Komm RR STMIM BUCKCUT ASSESSMT (Csb 1 RG 1.97 m 1.97 Discussion Needed in Black-Variahla Tvem Cata mrv Submaction tw* hamtw?

Mkin&A4ning Contairment Tntmarity (Chr+ 4 n M I

JPrinary Contaiment A,B,C 1 1D.2.3.12 Yes  !

Pressure (value, rate, trend,

+ alam)

Drywell/Containnent A,C 1 1D.2.3.13 No Hydrogen Concentration (value)

Secondary Containytnt C,E , 2 1D.2.3.14 No Area Radiation (value)

Secondary Containment C,E 2 1D.2.3.15 No

. Noblo Gas Effluent l Primary Containnent C 3 ID.2.3.16 No ,

Noble Gas Effluent (Suppression Pool A,D 1,2 1D.2.3.17 Yes

! Tenperature

Drywell Air Tenperature A,D 1,2 1D.2.3.18 . Yes i .

iFuel Cladding Barrier hi+ neina

CoolantRadiation WA WA 1D.2.3.19 -

l (value + ala m)

Coolant Ganna C 3 1D.2.3.20 No (1 sample /6 hours) results within 72 he syntam aparation Main Stase Line Isolation D 2 1D.2.3.21 No Valve Imakage Control System Pressure Contaiment Spray Flow D 2 1D.2.3.22 No

-Z3-

TABLE ess 15 E-l 1

VNtIABLE Mme KR STMEM BLAOCOUT ASSESENT ( CedwM)

. s i RG 1.97 N; 1.97 Discussion Needed in Black-var 4=h1= Tvm category Atlaschim out sequence?

i l Syste Operation

' ttw+4ma

. Residual Best Remmal D 2 1D.2.3.22 No

! (NHL) Syste nw *

! RER Service Water nw D 2 1D.2.3.23 No .

Low Pressure Coolant D 2 1D.2.3.22 No

, Injection Systen nw Reactcr Core Isolation D 2 1D.2.3.24 Yes Celing Systen nw

RCIC Roan Temp. - - -

Yes

! Control Roan Temp. - - -

Yes I Hi@ Pressure Coolant D 2 1D.2.3.24 No Spray System nw ..

Core Spray Systen Ew D 2 1D.2.3.24 No

' Standby Liquid Control D 2 1D.2.3.25 No*

l System (EG) Fim

SLG Stocage Tank Level D 3 1D.2.3.26 No i SRV Position D 2 1D.2.3.27 Yes Feedwater n w D 3 1D.2.3.28 No

! CST Level D 3 1D.2.3.29 Yes ESF Celfng Water nw D 2 1D.2.3.30 No ESF Cooling Water Tupperature D 2 1D.2.3.30 No i

Hi @ Radioactivity Tank Level D 3 1D.2.3.31 No j Bnergency Vent Desper Position D 2 1D.2.3.32 Yes l'

j Stan@y Burgy Status D 2 1D.2.3.33 ,Yes*

i *IDCluding breRher position.

- 'L A -

O

-.-,-.--..,.-..~.,.,- .,,-,,,,,,,..,__,-----,,~,..,,x,,-- . , , , - - - . . ~ , - - - , , - - . - - , . - , , - - . - - -

, , . - l TABLE M 15 E -l

' VMIABU!S MSESSED POR STM3DN BUCEDUT MSESSIENT ( Cow 4% wo e- cN BG 1.97 BG 1.97 Discussion Needed in Black-vari =hla Tvm cataaory subsection gg c==nm?

, Effluent Monitoring SG15 Ventilation Mw Rate E 2 1D.2.3.34 No E 3 1D.2.3.34 No lOtherVentilationMowRates Particulate / Halogen E 3 1D.2.3.35 No Release (sample)

Erwirons Radioactivity E 3 1D.2.3.36 No Monitoring Metecrolog( E 3 1D.2.3.37 No

' Post-Accident M H ng E 3 1D.2.3.38 No (sample) 9

- LS -

TreLE sus 15 E- 2.

IOfER SUPEIES TO INSHUMS EBED RR A BLACEDUT Conttol Room Power varbhla indim*ar Sutx21v Available? 3 2,33 R W Lsvel B21 D623A 120 Inst. Bus A Yes 1 R623B 120 Inst. Bus B RW Pressure B21 R623A 120 Inst. Bus A Yes 1 R623B 120 Inst. Bus B Yes 1 Suppression Pool Water Level P50-9500A,B 125 VDC Yes 3 Pri. Containnent Isol. Valve Indication RPS Yes Position Li@ts

' Pri. Containment Tenperature T41-RR613A,B 125 VDC Yes 3 Pri. Containment Pressure T41-RR618A,B 125 VDC Yes 3

! Suppression Pool Tenperature P50-R600A,B 125 VDC Yes 3 Drywell Air Temperature T81-RR611A,B 125 VDC Yes 3 RCIC Flow E51-R606 RPS Yes RCIC Roan Tenperature E31-R608 RPS Yes Control Roan Tuoperature - -

Yes 5 SRV Position Indicating 125 VDC Yes Li@ts GT Level By appliccnt By applicant Yes 2 Doergency Vent Dumper Position Indicating 125 VDC Yes 3 Li@ts Stan &y Energy Status 619 kv AC Voltmeters Source Yes DC Voltmeters Source Yes Air P53-9606A,B 125 VDC Yes 3

. l

• • L $ ~

. I Notas to Tabla 1-2-15 5 - 1

1. Enhanced Water Level Instrtanent to be powered from d.c. power, i

2. D.C. Power to be provided by appli ant.

3. Power Supply frca 125V d.c. to Reactcc Island Logic Panels P881 or M82.

4. EAhaust air measurunent may be tmreliable. Local thennaneter to be supplied by applicant.

o

.1 *

[

f 4

4 d

b

( - L 1-

_ . - _ . _ - _ . _ = -

GESS Att-3E 7SS N O C u tA R. TSt AND ATTACHHEVT 6 To -

AP PsuOn V LSE p.c:t c R.c o x wsATVP DUR.wo A ST Am op S t A CW- OVT 4

l 515 8. L pu g. PO SE eHa c%eb ne purpose of this momesendangis to document the results of analysis performed by i

Containment and Radiological Engineering on Reactor Core Isolation Cooling System (RCIC) room temperstare response during a station blackest for tese GESSAR "It .

gdiana He results indicate that a station blackout imposes no threat to the speration of RCIC with the RCIC room temperature reaching 122'F 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> into the trcasient, well below the polat above which RCIC performance would be degraded.

Sensitivity results for some of the most important parameters are also gives.

1 1508.2 4INf10 DUCTION A station blackout ressits la loss of all 'A.C. power (both of fsite and oasite so:rce s) , initiating reactor isolation and scran. For this analysis all three diesel generators of a RTR plant are assumed inoperative, i.e., no Emergency Core Cooling System (ECCS) pumps are available: this leaves the battesy operated RCIC as the only system available for sore sooling. nu s, it is essential that the RCIC remains cpe ra tiona l.

An important requirement for the proper functioning of the RCIC is that  ;

tho RCIC room temperstare be asistained below the equipment operational limit.

ne loss of all A.C. power also means the loss of lashting, anziliary equipment cpe ra tion, area IVAC and drywell fan coolers, resniting in a drywell hostsp. At sono point reactor depressurization will be initiated to reduce the heat impat to the drywell, although the reactor is assumed to be depressurised only to the point sufficiently above the RCIC shutoff pressure so that operation of the RCIC can be j caintsiand.

RCIC initially draws water from the Condensate Storage Tank (CST). Rowever, as '

actomatic evitehover to the suppression pool as the water 11-l l

i .

l i

soarso could 'oeste if the CST water level drops too low or the suppression pool water povol rises above a* sertain potat. Sines the suppression pool heats up as a result

)f SEY disaharges and subsequent remotor depressurisation, and simes the desiga

empercture for the RCIC pump is 140*F, a massal switch book to the CST from the a:ppression pool as the RCIC water sourse la required whom the pool temperature lspyroaches140*F. Simee the time period when the RCIC takes saation from the

sappression pool is relatively short (about 30 minates) sompared to the transiest beriod Cf interest (up to 20 hours2.314815e-4 days <br />0.00556 hours <br />3.306878e-5 weeks <br />7.61e-6 months <br />), the impact sa RCIC room temperature la assuming

' :bat RCIC draws all water from the CST is insisaificant.

j CIS .,3 40DELING AM ASSUMPTIONS h

Io cedel the RCIC room temperature response, thermodynamis properties of steam and

'; air is the room are evaluated based on mass and energy belasses. Nest soareas and l hest sicks were seasidered. In addition, some stem has leaked into the room throssh jths RCIC turbine gland seal. The room is cesservatively assumed to be isolated from lthe cdj ccent rooms.

1 IEsst Sources - The following heat sourses are modeled:

i l* Steam Pipes - there is a six inch steam pipe systrom of the RCIC i

terbine, 60 f t long, with three inches of isemistion, with the pipe

temperature assumed equal to the resetor stem temperature of $52*F

. sader normal operating conditions, and 388'F af ter resetor (

depressurisation to 200 pois): and a sistosa inek eskaast steen pipe

}

doomstream of the RCIC turbine, 40 ft loss, with two inches of '

l isostation, with pipe temperature at 250*F because steen pressure donastream of the turbine is held at 25 pois.

l ,

!* Water Pipes - two maissalated water pipes, one asetion pipe and the l other discharge pipe, with diasastoms of 8"133 ft and 6"136 f t, ,

e:rry water from the water soeroe and~ inj oet it into the remotor. As contioned previously, the water sourse may be either the CST or the i f

stypression pool, thus the water temperature may vary from the CST

{'

to:perature of 90*F sp to the suppression pool taperature. Depending on the RCIC room temperature at a partiesler time, these water pipe s

! csy be either heat sourses or heat sinks.

i le Tarbine - the RCIC turbine is isosisted. The turbine temperature is takes as the average upstream sad downstrom steen temperatures.

Small portions of turbine that are not lassisted are not modeled.

• RCIC Pump - the RCIC pump weighs 6400 lba and is not inesisted. As in tho ease of water pipes, the RCIC pump may bosome a heat sink depending on the room temperature and the water. temperature.

I:st Staks - The following host slaks are modeled:

\

i I

l l

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

i , a ',* .. .

O Comerete Walls, Floor sad ceiling - the walls are 26 ft tall, with -

cidths varying from 18 ft to 31 f t. Thieknesses very from 1 ft to 3 l ft. These straetsres were conservatively assumed to be insulated on I the outer surface.

o Tarbine Base Plate - it weighs p001he and is asiassisted.

O Room Cooler - it weighs 20001km and is alasalated.

O As mentioned previoasty, the water pipes sad RCIC pump bosome heat j slaks if the RCIC room temperature is higher than the RCIC water ,

i temper stare.

l Analytical Assumptions - The to11owing assumptions were made is the analysis, with jtstifications for these assumptions gives subsequently:

'* Air and stem are uniformly mixed at all times.

l

'* Air behaves lik; an ideal gas.

i

!* No condensation on strastarsi surf aces.

!* The RCIC room is isolated from the satsonadings. .

i

!O Beat saaduation is one dimensional through strustares and walls.

1

, Simso the pericd of interest is severs 1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, steam 1eaked into the room has

s fficient time to diffuse sad six with air, therefore, the amiform sizing assumption l1sastodapproximation. Al so, since only low pressures and temperatures are icaesamtered, the ideal gas law holds tras for air.

, Assumptions of no condensation on strustarat surf aces is conservative becanse the l froc-convection heat transfer coef fiefeat used la tae absence of acadessa tion is

, smaller than the condensing heat transfer coef fisient. Isolating the RCIC room is

! cacther senservatism, because mass sad energy are preventd from leeving the room .

jthrstghsondsstion, soavection and radiation. Flaally, the oms-dimensional heat I

! condaation assumption is oorrest essept at the eormers of the walls, but the impact

' is ne gligible.

O84 i,INFFT PARAMETERS i The fellowing initial senditions and key parameters were used la the analysis 1

0 Initial room temperature was 90'F.

• Steam leakage rate was 70 lbm/hr.

i r

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

i i  :

! l i

l.

l' No resetor depressurisation for the first 30 instes (as the operator j 'is trying to determine e opriate actions) and the resetor was cooled l l dera at 100*F/k::. I

,' Temperature of RCIC water was 90*F, which is the teshaisal

spoeification CST temperstare, beessee the RCIC saa take asetion from ,

j tho espyression pool for only a short period of time and the operator i ci11 evitch the aseton back to the CST as the pool approsehes 140*F. ,

l5G8 5 i ERSULTS Ale DISCUSSIONS .

A timeshare computer program has been developed to carry out the salesistions  ;

dessrib

d above. t i

lThe RCIC room temperature response following a station bleekost is gives in Figure 1588 -L '

Q Tha temperatste ineresses rapidly during the first host of the transient, then

/

!the rete of taeresse levels off seeequently. The room temperature rises to 119'F at

'oight hosts of transions and 122*F at twelve hosts of transient. '

i 15 s 2. show a. d issa-3 the sensitivity reesits at high water temperature and low steen

Figarcep _g ^ __

!1ookago rate, respective;y. With the water temperature at 140'F the RCIC room l temperature rises to 133*F at twelve hours, while at the steam leakage rate of 10 I line/hr (which sorresponds te new turbine gland seal eendition) the room temperature jreschcs only 1018F at twelve hoses. The high sensitivity to the steam leakage rate lis das to the large latest heat of stees which is released spos eoadensing la the ,

RCIC rces. The sensitivity study also indientes that there is no impact of reestor  !

jc2sidoon rate on the RCIC room temperatste resposes. '

iTh3 above rossits indicate that the RCIC room temperature twelve hosts following a ,

i statico blackout to be substantially below the equipment gsaffication limits of 212'F j

lfar the first six hours and 150*F between sia and twelve hours following a station i bl askott . This shows that proper operation of the RCIC saa be maintained for many ihsdrs daring a station blackout to provide adequate sore coolias.

k i

I l

l l

i t

l I

• m8

( ene i'

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

l l

• l t al i

'l y

• ; 1

. I . ...

' } '* d : '

. , : e.

. %d . . - - - -

1 .

i 7% i .

LL Y%

i l l 0 I rJ 1 C l'

!  ! , g i

~'

i l 1 l e I j ,

.  ; I* I 10 g W I

.o.

fy .L '

(

i .

l (i J  :

l.

l .I .* ,y

, s . =

I

)

! L,4. ' L.,1 =-. . _ _ .

e <

P 6 g A*

w l

. . l

i. . . . .

e , . . i

l. , l l ,

l 4 i i *

. 4.  !  !, .i. . ...! l .-

. 4 .

i f, a . e. . . .

... .1 e.

W -w ... .. , .

I,

. b. g . ., . .g g

i.

g. .~

y gO .. i

• .g '

3. . , ,y .

~

6 L . . . .

. j e . .

. i

..e4 .

_...L

..I J,

. :Q/,, . .. ..'1.

, I e di m. ..., .. W

. .I . . , . . . .

.1 .._.

..{'

Q u t .

i ..

{. .

. .n p.

4g .

m ,

_l

-3 .- ,

o. . ;.

e .. . .. , . . .,

..).

.... 1 I

1. 3.L. . . .. . ,. . . . . . . . . . . .. . . .

. . . 1 .., . . .. . ... 1 i

_ .4

l. ..

1

.4 ..

.. s e >+

w . . . .....

i . 6 i

1 . .. ... . .. .. ...

.._ 1. .. 1 .... ... ... 4

.-. in.

g

. ..1. .

l. 4 _.

. h .

.. .4. _ g43

.s I

C ., .

o .. .

gg 9 .. . . 9 .. ...

... ... o .. .

g 4

3, . .

. ..g. .

.. . q. Cr" . . . . . -I

,.... ... l

... . I i . ... . . . .. . _

l .. -

.- = . . . . .

ye ' vam vaid e ,

l s

1(-

, . s I * '

l -

l

[.

a l

l

_f ..

p) ....q. . . ;l . .

• I *

• E. ( ,\.N l . .

e

.a ,

~ ' ~ ' *

• Y

=

O O1 i

l

'w @ iP l

IO l' t r

I

' ' I- > 3 . ..., C ..

I; -

lI l j

_ . 1.31 '

3

- s 'p , .i ,

1tl 1

r ,I  %

q g .

I i

a. I .

. ..i.... . _ . . . . .. .

M = 1 I

[

"b -

,([ .

t

j. . . . . . . . . . . .

7 . _

bk i ims: .

$ 1I_. -

G4-~

e

• 1d'  ! - . . .N~ -

""I .i.t 2. l i

.-... d . . . . . . - . . - - - .

'y 0-v-y- I I i i

. . . . o ..._._

d l h I- g s f .

d ' ' l Q.. a.

1 ' -

_ ..: - . . .m .\. $.

2! - ' - - - * - -

l l

r -- T.

m, . .

,' _.. . a ___

i I .

g l 48 '

5 l. -

5 8  ; , i ,

0

- 3

$  ! f

'h ,!

u. .

. . . . . .W ...

I .

i ,

• 1 l l l 1 I  ! ll

~

g I o l 0 ip 6 i o -

o '

.. j . . u. . . . .,.

r,, . . . . , ,,.

e ..

,c,,, ..

, ..,,,,, .. o- .

-j-- r - );,"" f 1 4 g 'V .A 7 h w * * -

6

. I . ,

e. ....e.e * ..

. e.. .u. a e . .. a.e ed4. e . . . . . .

. l I 7 ==*

i o*...

e .G

. [' !

. 6 l I j t- . .. .. g . .

i

..'p, i

e

.- i .

.%3 b

g . .,,. i N M k,,

,'_g, O?

• I I

I. l .

4 i .

ElI _0

... .. .t . .. ,I

. . .e l C ' 3, l

i O p t *

• _y 44

! . 3 '! '

,ma h 'k y 8

11 lg.

e.

l- .

9 . . e i , . e. l . . . .

__g u M

• gy'g . y.

1 .i

. .. . 4...

A.

. . . . . . .4 . ....

..6

, 9... g w

, , e. - .

3', g" . . .!. . . .

s . . .. .. .

3 . . .

Y W9 .

I

. .. . 4 6

Q. ..

g ..

.. .  ; 4

.O. .

. . .. y.. , .. 9 w

.... . 4

. .y. . .. d.

t

- s. ..

... 1 m

• =

"Y  ;  ; i k*

. e gM

. .3

. . 6 e

_g ,

. . . .. g 9 .

\

W 4

- 9 b '

d.

p ,g ,

,.. g l

. l 6

. ..i ..

. i

,k.

w y (/} .< 4 . . s . .

s

.i ..

s g ,..

,,,,,._.< A .. . . . .

M

.... .. .. i w . J. . .

. e } , . .. < . ..

. s f

. -t .

. . e 4

.3., .

. . ..e....

.9 9..

I e

.4 . 4 ..< .... ...

.k

)

.4g.. . .

, . .i ... < . . . . . .

0 -

,p,,

g ...

.... ...i

.. i t

... .w

(

... . .. .- . 4

.< < ....i C

.. ......, ..qg . . .D D

.. 4 ..., .-

. .e

. . . al4 ig .

. q p ...

es I

. . .e i

... .=

== . ...

0

....i ee . .

. 6 hp TM

~

M *I"*' YN '

. . i .

. . ..- . . . . ....., . . . ...i s ..

. s.

e. . . i

.Mm. m .s i m .w w W WM

.