ML20082D252

Suppl 1 to Quantitative Info on Revised Fire Analysis for Limerick Generating Station,Severe Accident Risk Assessment

ML20082D252
Person / Time
Site:	Limerick
Issue date:	11/30/1983
From:	NUS CORP.
To:
Shared Package
ML20082D194	List: ML20082D192 ML20082D252
References
4161-S-01, 4161-S-1, 4161-S01, 4161-S1, NUDOCS 8311220358
Download: ML20082D252 (25)

Text

_ _ - _ - .-.. --

t QUANTITATIVE INFORMATION ON REVISED FIRE ANALYSIS FOR LIMERICK GENERATING STATION Supplement 2 to ISS SARA, NUS Report No. 4161 i

Prepared for

' PHILADELPHIA ELECTRIC COMPANY

!O i

November 1983 NUS CORPORATION 910 clopper Road Gaithersburg, Maryland 20878 F

O

. - -. .. ,. .--....._~,-.,.....,__-,,y.~,_--_ - - -

Supplement 2 QUANTITATIVE INFORMATION ON REVISED FIRE ANALYSIS 1 FOR LIMERICK GENERATING STATION

1. INTRODUCTION The fire analysis presented as part of the Limerick Generating Station (IGS) Severe Accident Risk Assessment (SARA) (NUS, April 1983) was performed on the basis of fire protection measures described in Revision 1 of the Fire Protection Evaluation Report (FPER) (PEco, 1981). However, since the analysis was performed there have been several major changes in these fire protection measures. These changes are described in Revision 4 (PECo,1983) of the FPER and have been submitted in response to the Limerick Generating Station Safety i Evaluation Report (SER) (NRC, 1983).

Supplement 1 to IES SARA (NUS, July 1983) provided a qualitative discus-sion of the impact on SARA due to the design changes together with an approxi-

> mate estimate of the resulting reduction in the total core melt frequency due to fires.

The purpose of this report is (1) to explain the basis on which certain important events, modeled as an integral part of the fire analysis, have been requantified to reflect the new design; (2) to briefly describe the method used for analyzing the uncertainties associated with the revised fire sequences, which have been refined since the original EARA study was submitted; and (3) to report the detailed results of the revised study including the impact on over-all estimates of annual core melt frequency and public risk.

l

2. REQUANTIFICATION OF EVENTS MODELED IN SARA FIRE ANALYSIS i 2.1 CONDITIONAL PROBABILITY OF FIRE GROWTH STAGE 3 (EVENT E)

The most significant impact on the fire analysis, resulting from the design changes, was in quantifying the conditional probability of disabling l all methods of plant shutdown (defined in Table 1) as a result of fires in I the 13-kV switchgear room, the static inverter room, the safeguard access area, the CRD hydraulic equipment area, and the general equipment area (fire areas 2, 20, 44, 45, and 47). The fire protection design as assessed in the original SARA study relied on the separation of cables and equipment associated l with two of the designated shutdown methods (see Table 1) by a minimum horizon-tal distance of 20 feet, or alternatively, where such horizontal separation was not maintained, cables associated with one of the shutdown methods were protected by -hour or 1-hour fire barriers.

The present design provides physical separation of redundant shutdown j methods by establishing 20-foot-wide combustible free zones between cables and components associated with redundant shutdown methods at strategic loca-tions. These zones are created by enclosing cable trays that pass through such zones with 1-hour rated fire barriers. Alternatively, where such I'

1 I

{

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

?

separation is not maintained between cables associated with redundant shutdown O methods, cables associated with one shutdown method will be provided with 3-hour rated ffte barriers. Additionally, a fixed, manually initiated, water curtain suppression system is located within each combustible free zone to prevent propagation of postulated transient combustible fires through the combustible free zones.

As discussed in Supplement 1 of the SARA study, the above design changes significantly reduce the likelihood of a fire breaching the above-mentioned fire protection features and thereby disabling all methods of plant shutdown.

Such an event is termed fire growth stage 3 and its conditional probability

! (given a fire has been initiated within a fire area) is represented in the fire growth event trees by Event E.

In evaluating the probability of Event E the range of possible mechanisms were first identified. They are as follows:

1. Failure to extinguish the fire before cable insulation, protected by a fire barrier which is enveloped by fire, undergoes thermal degradation resulting in electrical failure.

2. Failure to extinguish the fire before cable insulation, protected by a 20 feet horizontal separation from the fire, undergoes thermal degradation resulting in electrical failure.

3. Fire barrier is missing or is ineffective.

For the design assessed in the original SARA study, mechanisa.a (1) and p(/ (2) were the most significant means of fire progressing to the third fire

growth stage. However, this is no longer the case for the revised design for

! the following reasons.

. First, in order for the fire to propagate via mechanism (1), the combus-tible loading and available air supply must be sufficient to support a fire of a severity and duration capable of penetrating a fire barrier having a 3-hour rating (qualified against ASTM E119 test method). The only installed combustible present in the Limerick safety-related fire areas with any potential to supporting a large fire is cable insulation. No credible transient combus-l tible fire in these areas would give rise to conditions which approached the

! ASTM E119 3-hour test conditions.

Multitray cable fire tests have been conducted by Factory Mutual Research Corporation (FMRC) (Sumitra, 1982) and Sandia National Laboratories (Krause, 1982) which involved EPR/Hypalon cable, which is used extensively at LGS, together with other types of cable insulation, which also pass the IEEE 383 j flame test. Cable tray loadings were approximately 40 percent by cross section, which is generally the maximum loading at LGS (PECo,1982) . The test steady-state fire durations were considerably less than 3-hours even when no attempts were made at suppressing the fire and the air supply was unrestricted. Tempera-tures in the region of the fires did exceed those specified the ASTM E119

~ _ ..

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

, fire barrier test method for short durations. However, the integrated time /

temperature history of the fires, which is the standard measure for relating

, fire severity with the fire resistance rating of barriers (NFPA 1976), is well below the fire severity under the 3-hour AS'IM E119 time / temperature curve.

The tests described above were conducted with only a fraction of the

total combustibles located within an ISS fire area. It is possible that had there been additional cable trays the fire may have transferred to another location and continued to burn for longer than the test fires. However, the tests do indicate that any given section of cable raceway protected by a fire barrier would not be exposed to severe temperatures for a period sufficient to cause failure of the barrier. It should be noted that at many localities i within IAS the combustible loading and available air supply will not be suf-ficient to support fires as severe as those observed during the tests.

l It is therefore concluded that cables enclosed by the 3-hcur rated fire barriers are extremely unlikely to fail even in the absence of any attempts to suppress the fire.

Secondly, there is no longer a combustible pathway in the 20-feet-wide zones which protect redundant cable raceways from being damaged by a fire.

, Thus, in order for the fire to propagate via mechanism (2), the heat from a

, fire on one side of a zone must be sufficiently intense to cause cable damage

on the opposite side of the zone. Both radiant and convective heat transfer 4

mechanisms were examined to determine whether this was possible. The minimum critical heat flux for electrical failure has been determined by Lee (1981) and ranges from 9 kW/m2 to 17 kW/m2 for EPR/Hypalon cables. Conservatively,

assuming that the radiant energy release from a cable fire is twice that j observed during the FMRC tests involving EPR/hypalon cable (Sumitra,1982),

calculations have shown that the maximum heat flux at a horizontal distance of 20 feet from the fire is below 9 kW/m2 No reduction in heat flux due to

} smoke effects was included in the calculation. Thus, cable damage will not 4

occur due to radiant heat transfer. The stratification of hot gases generated by cable fires is recognized as a potential means of compromising the protection offered by physical separating redundant cable raceways (Berry, 1983). This

m,Jhanism maybe particularly significant in relatively small enclosures with low ceilings where there is limited opportunity for fire plumes to be cooled by mixing with the surrounding air. However, where 20 foot separation is used as a means of protecting redundant raceways at ISS, (fire areas 44, 45, and 47), the associated enclosures used are relatively large; the ceiling heights are approximately 30 feet and the minimum floor area is 9000 square

! feet. The fire tests involving EPR/hypalon cable, which were in large enclo -

2 sures (Sumitra,1982), indicate that plume temperatures cool rapidly with

! increasing height above the fire such that at 20 feet and 25 feet above floor level certerline temperatures were 2300C and 1660C respectively. Since the minimum surface temperature at which EPR hypalon cable undergoes degradation is 2950C (Lee,1981), connective heat transfer across the 20 foot wide combus-I tible free zones at ISS is considered to be an unlikely means of cable failure.

3 1

i i

- - - - - - - . .. ..-.-., --,- -, _ _ _ ._ ,,- _ . ..~.--, _ _ ,,,.,n,._,,--_-n _ n_ ,,. ,-_g..,. -mv.,.ny,.n..,,---,.n--m,,nwn

'The third mechanism identified (3), " fire barrier is missing or ineffective,"

may occur for a variety of reasons such as:

1. Not installed initially

2. Removed and not replaced

3. Damaged and not repaired

4. Cables incorrectly routed in unprotected cable trays Each of these is considered to be a plausible event and has been known to occur at other LWR plants. Furthermore, such occurrences are difficult to detect and may go uncorrected for long periods of time. The data recorded in the Licensee Event Reports (LERs) for such events is probably incomplete and the population from which the data is derived (i.e., the total quantity of fire barrier installed in U.S. light-water reactors) is not readily definable.

Thus, there is no alternative but to assign a probability to the failure mecha-nism based on judgment that the occurrence of a missing or ineffective fire blanket within the range of influence of a fire is a reasonably rare but plausible event. A probability of 10-3 is assigned.

For fire areas, such as the 13-kV switchgear room (fire area 2), which are relatively small, the probability of Event E in the fire growth event trees is 10-3 since most of the fires are located in a position where they could potentially damage protected cable raceways. However, as in the original SARA analysis, for large fire areas, namely, the safeguard access area, the l

CRD hydraulic equipment area and the general equipment area (fire areas 44,

! 45, and 47), an additional factor was incorporated into Event E. This factor represents the fraction of fires occurring within the fire area located in (x,,s) ' the vicinity of protected cable raceways. For fire area 44 the fraction is

' O.5 and for fire areas 45 and 47, the fraction is 0.25. The basis for assign-ing these values is discussed in Cnapter 4 of the SARA study.

2.2 UNAVAILABILITIES OF SYSTIIMS DAMAGED BY FIRE L In addition to the reanalysis of the fire propagation probabilities, some of the damaged system unavailabilities have been reevaluated in a more refined manner by modifying and requantifying system level fault trees at the component level. This became necessary since random equipment failures which are unrelated to the fire damage but occur coincidently are a greater influence on the overall outcome of the revised study than they were previously. These revised estimates of system unavailability are reflected by the changes to probabilities of Events B, D, and F in the fire growth event trees (Figure 1-7) .

2.3 CONTRIBUTION FROM NONSIGNIFICANT FIRE AREAS In the original SARA study, the point estimate annoal core melt frequency from fires was 2.3 x 10-5 of which eight fire areas contributed over 95 percent.

The contribution from all remaining (nonsignificant) LGS fire areas was shown by conservative analysis to be approximately 10-6 Since the revised fire protection design has substantially reduced the contribution from the eight significant fire areas, the estimate of the contribution from all other areas 4

.becomes more important. As a consequence, a less conservative re-evaluation of those areas was performed taking into account the following factors which l

( were not previously inc,luded:

i -

1. Reduction iC dable fire frequency due to improved Limerick design

2. Likelihood;of dire suppression prior to loss of all systems served by cable oc ecaipment in the fire area

3. Revised estir.at,es of damaged system unavailabilities The results of the reevaluation indicate that the total contribution to i annual core melt fraquency from fires in these nonsignificant areas is approxi-l mately 2 x 10-7

-3. UNCERTAINTY V.NALYSIS In order to provide a better understanding of the uncertainties araociated with individual fire sequence and overall core melt frequencies, a mort mathe-l matically rigorous analysis was performed than that previously described in

! the SARA study.

Briefly, the model used for the uncertainty analysis was developed by formulating algebraic equations to represent each of the significant fire sequences identified by the fire study. Fire frequencies, fire growth prob-i abilities and degraded system unavailabilities were input as separate variables which were assumed to be distributed lognormally and were assigned judgmental error factors.

Where the state of knowledge for two or several input variables was identi-cal, for example cable fire frequencies within all fire areas, a dependence was introduced between the variables. Having constructed distributions for

! each variable, these were then propagated through the algebraic expression l

using the computer code SPASM. Further discussion on this topic is given in l Chapter 11 of the SARA study.

The resulting distribution of annual core melt frequency due to fires is approximately log-normal with a median of 1.4 x 10-6 and an error factor of 8.5 (the original SARA study indicated a median of 8.7 x 10-6 with an error factor of 10). Error factors on individual. fire sequence frequencies were determined to be approximately 20. These results are based on error factors for the input variables as follows:

Fire frequencies: 10 Fire propagation probabilities:

Fire growth stage 1 to 2 (cable) 2 Fire growth stage 1 to 2 (panels) 10 Fire growth stage 2 to 3 10 Damaged system unavailabilities 3-10 O

. - . . . .. = - - -

4. - RESULTS OF REVISED FIRE ANALYSIS The revised contributions to the core melt frequency due to fires are summarized in Table 2. The values, presented in this table, with the exception of the contributions from the Auxiliary Equipment Room, are derived from the fire growth event trees which are included as Figures 1 through 7. The analysis of the revised Auxiliary Equipment Room design is discussed in Supplement 1 of SARA.

The design changes to the LGS fire protection measures reduce the overall

. contribution from fires to the annual mean core melt frequency from 2.3 x 10-5 to 3.4 x 10-6 While the analysis of the old design indicated that the majority of the fire-induced core melt frequency came from fir t which them-

! selves damaged all means of plant shutdown (fire growth stage 3), the assess-

' ment of the present design indicates that a higher proportion cf the contribu-tion from fires comes from a combination of fire damage and random equipment failures (fire growth stages 1 and 2) . Furthermore, the contribution from j fires is now divided between Class I accident category--failure of inventory 4

make-up and Class II accident category--failure of long-term heat removal, whereas previously the only significant contribution from fires was to Class I.

A description of the dominant fires sequences is given in Table 3.

A summary of the contributions to the IAS annual core melt frequency from internal and external events, which incorporates the results of the revised i fire study, is given as Tables 4 and 5.

4 As mentioned above, fire sequences contribute to accident Classes I and

! II, and as such are contributors to latent cancer fatalities. Figure 8 shows

' CCDFs for latent-cancer fatalities from fire initiating events alone. Figure 9 shows the median CCDF for total latent cancer fatalities and the contributions from internal, fire, and seismic accident sequences. Figure 10 shows the lower, median, and upper estimate CCDFs for total latent cancer fatalities

from internal,' seismic, and fire initiating events. ,

4 The improved fire protection design reduces the predicted average latent cancer fatalities from all initiating events to 1.2 x 10-3 per year (lower estimate); 1.0 x 10-2 (median); and 1.1 x 10-1 per year (upper oound). The corresponding values from the original SARA study (Table 12-9) were 2.3 x 10-3; 1.2 x 10-2 and 1.5 x 10-1 All the above values are obtained by integrating the areas under the CCDFs.

5. REFERENCES Berry, 1983. The Relationship of Fire Protection Research to Plant Safety. Transactions of the Eleventh Water Reactor Safety Research Informa-tion Meeting, NUREG/CP-0047.

Lee, J. L. 1981. A Study of the Damagability of Electrical Cables in Simulated Fire Environments, NP-1767, Electric Power Research Institute, Palo i Alto, California.

6 l - -, ~ _ _ . . _ _ . _ _ _ _ _ . _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _

. Krause, F. R. 1982. Burn Mode Analysis of Horizontal Cable Trays, Q NUREG/CR-2431.

NFPA, 1976. Fire Protection Handbook, 14th ed. National Fire Protection Association.

NUS, April 1983. Limerick Generating Station Severe Accident Risk

_Assessmes.t (SARA) .

1 NUS, July 1983. Impact of Plant Design Changes on Limerick Generating Station Severe Accident Risk Assessment, Supplement 1 to IGS SARA.

Philadelphia Electric Company, 1981. Fire Protection Evaluation Report, Revision 1.

Philadelphia Electric Company, 1982. Final Safety Analysis Report l Limerick Generating Station. Revision 12 1

Philadelphia Electric Company, 1983. Fire Protection Evaluation Report, Revision 4.

, Sumitra, P. S. 1982. Categorization of Cable Flammability; Intermediate Scale Fire h ats of Cable Tray Installations, NP-1881, Electric Power Research Institute, Palo Alto, California.

i i

i J

i 7

l i

Table 1. Systems or Components Associated l

. I with Shutdown Methods l l

i Shutdown method System or component Method Aa RCIC ADS RER train A RHR-SW train A ESNS train A Standby diesels A and C Method Ba HPCI ADS i RHR train B RHR-SW train B ESWS train B Standby diesels B and D Method Ca (equivalent to alternate ADS method A)b RHR trains A and C RHR-SW train A ESWS train A Standby diesels A and C Method Da (equivalent to alternate ADS method B)D RHR trains B and D RHR-SW train B ESWS train B Standby diesels B and D aAs defined in the IES Fire Protection Evaluation Report (Revision 4) bAs defined in the IES Fire Protection Evaluation Report (Revision 1)

(Table 1 of ISS - SARA, Supplement 1) 8

\

Table 2. Point Estimates of Core Melt Frequency for Revised Analysis of Fires at LGS Fire area Transient Percent zone initiator Cable combustible Panel Total contribution 2 4.9-7 1.1-7 6.8-7 1.3-6 38.1%

(2.4-6) (5.9-7) (3. 2-6) (6.2-6) 20 3.7-9 1.0-9 2.4-8 2.9-8 0.9%

(5.0-8) (1.5-8) (3.8-8) (1.0-7) 22 7.8-8 2.5-7 NA 3.3-7 9.7%

(6.1-8) (1.9-7) (NA) (2.5-7) 24 NA 1.3-7 2.7-7 4.0-7 11.8%

(E) (1.0-7) (1.6-7) (2.6-7) 25 E E E E E%

(E) (2.6-7) (1.0-7) (3.6-7) 44 4.0-7 4.0-8 2.5-7 6.9-7 20.3%

(4.1-6) (4.1-7) (1.5-6) (6.0-6) 45 1.4-7 1.9-8 5.6-8 2.2-7 6.5%

(4.7-6) (6.6-7) (1.0-6) (6.4-6) 47 1.3-7 1.9-8 1.1-7 2.5-7 7.4%

(1.2-6) (1.8-7) J5.2-7) (1.9-6) 1.2-6 5.7-7 1.4-6 3.2-6 (1.3-5) (2.4-6) (6.5-6) (2.2-5)

Contribution from all other fire areas 2.0-7 5.9%

(1.0-6)

Total annual core melt frequency from fires 3.4-6

( 2. 3-5) values in parentheses are for old design.

(Revised Table 4-6 of LGS SARA)

O '

Table 3. Dominant Contributor to Core Melt O Frequency due to Fires Sequency / event Description Point estimate

TyQN resulting A fire in the 13KV switch- 8.2-7 gear room leads to reactor from fire in fire area 2 isolation and in combination with random equipment fail-ures leads to failure of long term heat removal i

, TyQUV resulting A fire in the 13KV switch- 3.6-7 from fire in gear room leads to reactor fire area 2 isolation and in combination with random equipment fail-ures leads to failure of inventory make-up systems TyQUV resulting A fire in the safeguard 4.1-7 4 frce fire in access area leads to reactor fire area 44 isolation and in combination j with random equipment fail-ures, results in failure of inventory make-up systems O TFQUV resulting from fire in A fire in the control room leads to reactor isolation 3.6-7 fire area 24 and in combination with

random equipment failures results in failure of inven-tory make-up systems i TyQUX resulting A fire in the safeguard 2.9-7 from fire in . access area leads to reactor fire area 44 isolation and in combination with random equipment fail-ures results in all high pressure injection systems and timely ADS actuation failure i-10

- Table 4. Annual Core Melt Frequency 5th 95th Point j Percentile Median Percentile Estimate Internal 2.4-6 9.2-6 6.0-5 1.5-5 External

seismic 2.2-9 3.3-7 2.7-5 5.7-6 Fire (revised 1.7-7 1.4-6 1.2-5 3.4-6

study)

Other NEGLIGIBLE Total 4.0-6 1.8-S 7.8-5 2.4-5 1

I l (Revised Table 12-1 of IGS-SARA)

I k

i l

l l

t th 4

3 11

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

Table 5. Accident-class frequencies

, Initiating- Annual frequency event type St Median 95% Point estimate CLASS I

, Internal 1.2-6a 9.0-6 9.2-5 1.2-5 Fire -- -- --

2.5-6 Seismic 1.3-9 1.7-7 1.7-5 3.2-6 CLASS iib Internal -- -- --

9.6-7 Seismic - -- -

5.0-8 Fire -- -- --

9.3-7 CLASS IIIb Internal -- -- --

1.09-6 Seismic 2.7-12 1.8-8 4.9-6 9.2-7 CLASS IV Internal 2.2-8 1.3-7 8.1-7 1.25-7 O Seismic 2.9-13 2.1-9 6.7-7 1.3-7 U

CLASS IS Seismic 8.0-14 7.6-9 7.0-6 1.2-6 CLASS S Internal 1.0-9 1.0-8 1.0-7 2.7-8 Seismic 1.9-21 3.2-11 2.5-6 4.1-7 al.2-6 = 1.2 x 10-6 bA dash means that a separate uncertainty analysis was not performed for these contributions.

(Revised Table 12-6 LGS-SARA) 12

• ,-,ww--,- , - -- w--vm-,we,-----c-,- ~w---e---ao n- w wn-m---,~-,-w-v-* e-s- wnemwenwwne ww w--,,--,ve-rwnwnw-w--w.e---~rw,re-wnw , - -

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

O A B C D O E F O

Fire .Uridamaged Fire suppressed Undamaged Fire suppressed Undamaged in systems fx: fore damaging systems before dan.klng i systems cable /TC/ panel mitigate unprotected mitigate or otected mitigate accident given raceway accident given raceways accident given FGS1 FGS1 (Failure gives FGS2) FGS2 (Failure gives FGS3) FGS3 l Core Annual sequence S' I9"*"*Y us OK i i

OK 5 i i

OK 5.5 x 10-5/ 10 x 10-3 1.3 x 10-5j 1.0/1.0/.M LO 1.8 x 10-3 CM 5.5x10-8 1.3x10-8 7.2x10-8 l

j 7.9 x 10-3 CM 4.3x10-7 1x10-7 5.7x10-7 i I

i i NA given C = 1.0/2 x 10-5 CM - -

3.6x 10-8 1

1 1

FGS = Fire growth stage 4.9x10-7 1.1x to-7 6 8x10-7

TC = Transient combustible j Total annual core-melt frequency = 1.3 x 10-6 i

i d

Figure 1. Fire-growth event tree for fire area 2 l (Revised Figure D-9 of LGS-SARA)

i I

t 1

l

A

()

B C D G

E F J

Fire Undamaued Fare suppiessed Unden.vy:d Fue suppressed Undamaged in systeins in: fore elim sying systems t)efore den.wjing systems catste/TC/ panel nistiqaie unprosecical mitigate pr otecteil mitegate acculent given raceway accident given facew.sys accident given FGS1 1-GS1 (Failuie uives FGS2) FGS2 (Failute gives FGS31 FGS3

'f Annual sequence sus I' *9 8'"CY OK OK i l i

OK i 6.0 x 10-5/ 1 x 10-3 1.7 x 10-5j

.02 ,,

, 1.2x10-g 3.4 x10-10 3.5x10-10 l 4

4.4 x 10-4 -.4/.4/.04 CM I

r .

t

.4 x 104 CM 2.3x10-9 6.4x10-10 1.7x10~9 3.2 x 10-6/3.2 x 10-6/ 4.9 x 10-5 CM 1.9x10-10 5.4x10-10 2.2x10-8 FGS- Five growth stage 3.7x10-9 1.0x10-9 2.4 x10-8 TC - Transient conninastitale

' * " ~

• Because of the evaluation of event E, the prot 2 ability of event C

, is not included in the evaluatiosi of tiie seqiseisce freeps 3 icy.

Figure 2. Fire-growth event tree for fire area 20 (Revised Figure D-10 of LGS-SARA) i t

~%  %

v A B C D E F Fire Urxfarnaged Fire suppressed unutanard Fire suppeessed Undamaged in systems tzfore al.unspng systerns before dan.maing systems cattle /TC nusaisie unpiutes.teil emt4: ate in otected mitigate a;culent givi:: e xcw.r acculent given exeway s accident given FGS1 FGS1 (Failuiu mves FGS2) FGS2 (Failure ipves FGS3) FGS3 Annual sequence sus I4"'"CY OK 4

i OK

OK i

1.0 2.3 x 10-4/

0.1 1.8 x 10-3 4.1 x 10-8 1.3 x 10-7 1 7.2 x 10-4 CM i

i N/A (for E = 1.0) CM - -

i 1.6 x 10-4 CM 3.7 x 10-8 1.2 x 10-7 i FGS= Fire growth staue 7.8 x 10-8 2.5 x 10-7

~ ' " " " " ' ' " '

Total annual core-melt frequency - 3.3 x 10-7

Figure 3. Fire-growth event :ree for fire ama 22 (Revised Figure D-11 of LGS-SARA) i i

^

A B C D Fire in Undamaged Fire suppressed Undamaged TC/ panel systems liefore damaging systems ,

mitigate unpsotected mitigate accident given raceway accident given FGS1 FGS1 (Failure gives FGS2) FGS2 Core Annual sequence 3,,

tus freem OK 1

OK I

i 7.2 x 10-5j 1.0/.025 1.8 x 10-3 4

1.8 x 10-3 CM t.3 x 10-7 8.1 x 10-8

' i NA given event C = 1.0/1.1 x 10-4 CM NA 1.9 x 10-7 FGS = Fire 9 owth stage 1.3 x 10-7 2.7 x 10-7 TC = Taansient comtjustihte

_7 Figure 4. Fire-growth event tree for fire area 24 (Revised Figure D-12 of LGS-SARA) i I

t O

A V

i B C D U

E F O .

Fire Undamaged Fire suppressed Undamaged Fire suppressed Undamaged in systems tefore damaging systems before sfamaging systems cable /TC/ panel mitit ute unprotected mitigate swotected mitigate accident given raceway accident given raceways accident given FGS1 FGS1 (Failure uives FGS2) FGS2 (Failure gives FGS3) FGS3 Core Annual sequence Sta frequency OK I

OK OK i

1.7 x 10-4/ 5 x 10-4

) 1.7 x 10-5j

.4/.4/.04 1.0 l 1.3 x 10-3 CM 8.5 x 10-8* 8.5 x 10-8' 2.6 x 10-8 4

l 1

l 4.25 x 10-3 CM 2.9 x 10-7 2.9 x 10-8 2.2 x 10-I i 1.6 x 10-4/1.6 x 10-4/1.0 x 10-6 CM 2.7 x 10-8 2.7 x 10-8 1.3 x 10-8 s FCS = Fire growth sta9e 4.0 x 10-7 4.0 x 10-8 2.5 x 10-7

aC = Transient combustible Total annual core-melt frequency = 6.9 x 10-7

' . Because of the evaluation of event E, the probability of event C is not included in the evaluation of the sequence freepsency.

t Figure 5. Fire-growth event tree for fire area 44 (Revised Figures 4-5, 4-6 and 4-7 of LGS-SARA) t 1

4 1

I \

J A B C D E F Undamaged Fire suppressed Undamaued Fire suppressa.d Undamaged Fire before (famaging systems in systems imfore d. unaging systems cable /TC/ panel miti 9ate unprotected mitigate tu otected mitigate accident givt:n saceway accident given raceways accident given FGS1 FGS1 (Failure gives FGS2) FGS2 (Failure gives FGS31 FGS3 8

t U Annual sequence a- frequency OK P'"*'

OK 1

i OK 1.2 x 10-4I 15 x M-4 1.7 x 10-5j 4/.4/.04 1.0 6.6 x 10-4 CM 3.0x 10-8

• 4.3x 10~ 9' 6.6x10-9 i ,

^

1.8 x 10-3 CM 8.6x10-8 1.2x10-8 4.8x10 4 1.6 x 10-4/1.6 x 10-4/2.3 x 10-6 CM 1.9x10-8 2.7x10-8 1.5x10-9 FGS = Fire growth stage 1.4x10-7 1.9x10-8 5.6x10-8 TC = Transient coministible Total annual core-melt frequency = 2.2 x 10-7

. Because of the evaluation of event E, the probability of event C is not included in the evaluation of the sequence frespiency, i

j Figure 6. Fire-growth event tree for fire area 45 (Revised Figure D-13 of LGS-SARA) 1 <

t i

4

) ,

d

s k

)w e e--

( (/

A B C D E F Fire Undamaged Fue suppressed Und.sn aged Fire suppressed undamaged in .

systerns before si. unaging systerns before al.sn.sjing systems cd;k/YC/,PJf.2? mitssgate satignotecleal suitujate gn otectual snitigate accident given raceway accident given raceways accident given FGS1 F GS1 (Failuie epv. > FGS2) F GS2 (Failueu anwes i GS3) FGS3 i

Core Annual sequence Sta- frequency sus

' D*"'I OK OK OK 2.5 x 10-4 1.1-4/

.4/.4/.04 1.0 -8 1.7-5/1.1-3 4.3x10-8 1.1x10 4 1.8 x 10-3 CM 7.9x10-8 1.2x10 4 7.9x10 4 1.6 x 10-4/1.6 x 10-4/1.3 x 10-5 CM 1.8x10-8 2.7x10-8 1.4x10-8 FGS = Fire growth stage 1.3x10-7 1.9x10-8 1.1 x 10--7 TC = Transient comtnistihte

, Total annual core melt frequency = 2.5 x 10-7 Because of the evalisation of event E. theinrobability of event C is not includealin the evaluation of the scepsence frespiency.

Figure 7. Fire-growth event tree for fire area 47 (Revised Figure D-14 of LGS-SAR A)

[

10 A

- I Iiililij l iI111lll I

I I 11111l 1

IIlillll 1 I IItill 10-5 _ _

Upper estimate _

10-6 _ _

~ ""

Z 1

[ Median estimate _

g _ _

10-7 - -

Lower estimate Z

~

i l

10-8 _

t _

l - _

l _ _

10-9 I iiisiti! I i i I t iti! I i i tini ! I i i r i ti t! I i i i i tii 10 I 3 4 10 2 5

10 0 10 10 10 l

Number of latent fatalities, N Figure 8. CCDFs for latent-cancer fatalities from fire initiating events (Revised Figure 12_12 of LGS-SARA)

O 10 4

- 1 i i i11111' 1 I I 18111' i I I t ill Ul l l i i l li s l' i i i i l i s_t-10- 5 - _

Internal -

Seismic 10-6 _ _

1. 2 1, -

Fi 7 res _

g - -

e

u. _ -

10_7 __ __

- - i l

10-8 _ _

I 10-9 I I I I IIII! I I I 18111! I I I 11111! I I lillii ll I t ill 2 4 10 5 10 0 10 1

10 103 10 Number of latent fatalities, N O Figure 9. CCDFs for latent-cancer fatalities (Median Estimate)

I i

---g -.-y- -ww-y-w,w----mv-v-----w - -

w, - -

O 10~4 _ i i i,,,,,l ,

,,,,iiil , i iiii,l ,

,,iiii,l i , , , , i i i.

Upper estimate -

Median estimate 10-5 __ __

Lower estimate 10-6 __ _

l- 2 2 O

4 g Fr _ ._

g u.

10-7 -- --

Z 10-8 , . _.

I 10-9 ' ' I Ill Hi I I I ' I 'll l I I lill' ' ' 'llll' ' II H 4

10 0 10 1 10 2 10 3 10 10 5 Number of latent fatalities. N l

1 Figure 10. CCDFs for latent-cancer fatalities from internal, seismic and fire initiating events.

(Revised Figure 12-10 of LGS-SARA) i

(' CONTENTS (Continued)

Page 12.5.3.3 Comparison with Other Studies 12-13 12.5.3.3.1 LGS PRA 12-13 12.5.3.3.2 The Reactor Safety Study 12-13

. 12.5.4 Latent-Cancer Fatalities--Interpretation and Perspective 12-14 12.5.4.1 Dominant Contributors to Risk 12-14 12.5.4.2 Comparison with Other Studies 12-14 12.5.4.3 Other Measures of the Risk of Latent-Cancer Fatality 12-15 12.5.4.4 Risk of Latent-Cancer Fatality in Perspective 12-15 12.5.5 Whole-Body Population Dose 12-15 12.5.6 Individual Dose Impacts 12-16 12.5.6.1 Bone-Marrow Dose of 200 Rem or More from Early Exposure 12-16 12.5.7 Offsite Costs (Decontamination, Relocation, etc.) 12-16 12.5.8 Individual Risk of Fatality 12-18 12.5.8.1 Individual Risk of Early Fatality 12-18 12.5.8.2 Individual Risk of Cancer Fatality 12-19 12.5.9 Future Trends 12-19 References 12-21 O SUPPLDiENTS 1.0 Impact of plant design changes on the Limerick Generating Station Severe Accident Risk Assessment 2.0 Quantitative information on revised fire analysis for Limerick Generating Station l

() (xi) 11/15/83 l

{

i

CONTENTS (Continued)

Page 12.5.3.3 Comparison with Other Studies 12-13 12.5.3.3.1 LGS PRA 12-13 12.5.3.3.2 The Reactor Safety Study 12-13 12.5.4 Latent-Cancer Fatalities--Interpretation and Perspective 12-14 12.5.4.1 Dominant Contributors to Risk 12-14 12.5.4.2 Comparison with Other Studies 12-14 12.5.4.3 Other Measures of the Risk of Latent-Cancer Fatality 12-15 12.5.4.4 Risk of Latent-Cancer Fatality in Perspective 12-15 12.5.5 Whole-Body Population Dose 12-15 12.5.6 Individual Dose Impacts 12-16 12.5.6.1 Bone-Marrow Dose of 200 Rem or More from Early Exposure 12-16 12.5.7 Offsite Costs (Decontamination, Relocation, etc.) 12-16 12.5.8 Individual Risk of Fatality 12-18 12.5.8.1 Individual Risk of Early Fatality 12-18 12.5.8.2 Individual Risk of Cancer Fatality 12-19 12.5.9 Future Trends 12-19 References 12-21 SUPPLEMENTS 1.0 Impact of plant design changes on the Limerick Generating Station Severe Accident Risk Assessment 2.0 Quantitative information on revised fire analysis for Limerick Generating Station (xi) 11/15/83