ML20245H825

Responds to Request for Addl Info Re Consideration of Severe Accident Mitigation Design Alternatives.Tables Listing Current Estimated Core Damage Frequency Per Reactor Yr & Dominant Population Dose Sequences Encl

ML20245H825
Person / Time
Site:	Limerick
Issue date:	06/23/1989
From:	Hunger G PECO ENERGY CO., (FORMERLY PHILADELPHIA ELECTRIC
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 8906300111
Download: ML20245H825 (37)

Text

{{#Wiki_filter:,,. P ] b 'l it. k 1 ] , PHILADELPHIA ELECTRIC COMPANY l NUCLEAR GROUP HEADQUARTERS 1 l .l 955-65 CHESTERBROOK BLVD. 1 'l WAYNE. PA 19087-5691 (215) 640 6000 June 23, 1989 d Docket Nos. 50-352 j 50-353 I License.Nos..NPF-39 -NPF-83 ] U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, DC 20555

SUBJECT:

/ Limerick. Generating Station, Units 1 and 2 Response to Request.for. Additional Information-Regarding Consideration of Severe Accident Mitigation Design Alternatives Gentlemen: NRC letter dated May 23, 1989, requested Philadelphia Electric Company (PECo) to provide additional information concerning . severe' accident mitigation design alternatives (SAMDAs)'lfor the. Limerick Generating Station (LGS). The issue of SAMDAs is being litigated before an' Atomic Safety and Licensing Board as a result'of 1 l the decision of the United States Court of Appeals for.the Third Circuit remanding this matter ~to the NRC for further consideration. The additional information was requested in order to allow preparation of an NRC staff position with respect to this issue. The specific NRC questions and our responses are provided in the attachment to this letter. i With respect to the information provided in the attachment, it should be recognized the importance of utilizing the most up-to-date information as to plant design and analysis methods when modeling the facility and the phenomenology associated with severe = accidents when examining SAMDAs and the question of whether they are f cost-beneficial. If, for example, the base case off-site risk from \\ severe-accidents is over-estimated, the benefits of any mitigation I design alternative which would reduce that risk would likely also be over-estimated. Similarly, if the most up-to-date information concerning the dominant accident sequences and associated radioactivity releases were not utilized, the mitigation measures 8906300111 890623 PDR ADOCK 05000352 /f g/I P PDC 1 i

\\

if

.Doguannt Cohtrol D3sk June:23, 1989 Page 2 L being. examined might appear to be cost-beneficial, but in fact would not be since the mitigation design alternatives would not_be based on potential actual sequences. The evaluation of EAMDAs conducted as part of preparation of the attached responses should be considerd as a. screening process only. Should any SAMDA appear to be close to-cost-beneficial as a result of this initial screening, this mitigation design alternative would be required to be optimized so as to maximize its benefit and, at the same time, minimize its cost. Moreover, a-detailed examination of the associated dominant accident sequences being mitigated and phenomenology must be conducted to validate the result. 1 Please note also that there is a significant scope and regulatory impact uncertainty factor associated with the design alternatives discussed in the attachment, particularly given the short response time. There is little, and in some cases,-no actual p' design, licensing, or installation experience with most of these ~ design alternatives. Should detailed design, licensing, and ultimately, construction efforts proceed,. additional complexities and problems:would.most likely arise~that would further increase the final. installed costs. Therefore, we consider that the likelihood of the estimated costs given in the attachment being overstated is extremely small. If you should have any question, or require additional information, please contact us. Very truly yours, fr G. A. Hunger, Jr. Director Licensing Section Nuclear Support Division cc: W. T. Russell, Administrator, Region I, USNRC T. J. Kenny, USNRC Senior Resident Inspector, LGS l l i - _b

w ,k A Av ~ . Attachment r / OUESTION 1 10n.the basis of PRA-results to-date, identify those accident

sequences that are expected to dominate the overall mean frequency projected for severe core damage and1for the -

significant'off-siterrisks (i.e., projected risk-of-early fatalities ~and persen-rem).. It is suggested that those' sequences that collectively contribute 90% to the overall mean, frequency for' severe. core damage be identified.as dominant and each . described. 'For.these dominant sequences,?present~the projected- ~ imean value.for'each, considering that three categories (i.e., internal. initiations,. fire initiations and earthquake ' initiations) will likely contribute to the overall results.

RESPONSE

Thefcurrent estimate of core damage frequency (CDF) for Limerick . Generating' Station Unit 1;(LGS-1) is given in Table 1-1. The. sequences that dominate the CDF are identified.in Table 1-2. .The sequences expected to dominate the offsite risk (population dose and early fatalities).are. identified in Tables.1-3 and 1-4,. respectively. All values are point estimates.except seismic which are.the means of calculated distributions. rSubsequent to the' initial.developmen't of the' LGS'Probabil'istic Risk Assessment,(Reference.1), in. response to the Commission's 'May 6,.1980 letter,.and the Severe Accident Risk. Assessment. .(Reference 2), developed in accordance with the requirements of' 'the National Environmental Policy Act, Philadelphia Electric -Company's (PECo.).PRA activities have concentrated on the updating and use of the internal initiator portion of the Level 1 PRA in accordance with the Commission's June 7, 1984 letter and PECo's July 23,.1984 response. The core damage frequencies for the internally-initiated sequences given herein are based on the November 1988 update of the LGS-PRA modified to include a Limerick turbine trip frequency of 2.55 scrams / year justified by actual Limerick operating experience (first two operating cycles).. The frequency of other initiators (other transients and LOCAs) remains the same. The current total transient frequency utilized is 6.7/ year. This is conservative and is expected to go down further as additional site-specific. data are accumulated. In order to provide a reasonable basis for evaluating mitigating designs, the externally-initiated sequences have been updated, to the extent possible in the time available, to account for major new information as described in the next three paragraphs. The fire CDF has been updated to reflect the current plant fire protection design (Rev. 11, of the LGS Fire Protection Evaluation Report - Reference 3), the latest plant logic models of the November 1988 update of the PRA, and the initiator frequency and 1-1

s W .3 suppression probability from the Sandia Fire Risk Scoping Study l (Reference 4). Even after this updating, the results remain conservative. Areas of conservatism include: the modeling and assumptions on the extent of damage given failure to suppress a fire i.e., it is assumed that all unprotected shutdown methods in a zone fail if any fire in the zone is not suppressed in 10 minutes; the modeling of fire suppression, mainly based on manual detection and suppression data (Reference 4); and in the determination of initiator frequency, which took no credit for cables at LGS upgraded in accordance with IEEE 383. The seismic CDF has been updated to include revised fragilities based on actual LGS equipment seismic qualification data for a number of components (electrical equipment, SLC test tank, N2 accumulators and RHR heat exchangers) versus the generic or surrogate plant data used originally in SARA where plant specific data were not then available, a more recent assessment of ceramic insulator fragility and analysis of recoverable electrical system failures (i.e., circuit breaker trips). The flooding CDF has been revised _to reflect the results of the detailed flooding protection analyses recently completed, the updated logic models of the November 1988 PRA update and the occurrence of. spurious fire suppression initiation summarized in the Sandia Fire Risk Scoping Study (Reference 4). The relative risk rankings of sequences given in Tables 1-3 and 1-4 were arrived at considering the accident class, as defined in SARA and given in Table 1-5, and the associated conditional risk for that class as calculated in SARA. References For Question 1 Resnonse 1. "Probabilistic Risk Assessment, Limerick Generating Station", Philadelphia Electric Company, September 1982. 2. " Severe Accident Risk Assessment, Limerick Generating Station", Philadelphia Electric Company, April 1983. 3. " Fire Protection Evaluation Report, Limerick Generating Station Units 1 and 2", Philadelphia Electric Company, Rev. j 11, February 1989. j 4. Lambright, J. A., et al., " Fire Risk Scoping Study: Investigation of Nuclear Power Plant Fire Risk, Including Previously Unaddressed Issues", Sandia National Laboratories, NUREG/CR-5088, January 1989. 1-2

r -- +. .4 - , TABLE 1-1 ~ CURRENT ESTIMATED CORE DAMAGE: FREQUENCY (Per Reactor Year) Internal Initiators-5.9E-06 ) Transients.. (2.lE-06) l Loss of Offsite Power (2.3E-06) ATWS (1.2E-06) LOCA '(2.7E-07) Seismic 3.4E-06 ' Internal Fires 4.2E-06 Others. 0.2E-06 l (Internal Floods and other Special Initiators) Total Estimated'CDF 1.37E-05 4 1-3

y g.s ..s q 1 g L a l.. 1 i-L ~ TABLE l-2 DOMINANT CORE DAMAGE SEQUENCES ^ (See Notes) ) l 1.- 1.90E-006 13.9% F44QUV-1 Fire ~incFire Zone:44 (F44) with core. damage resulting-from combination of fire-induced and random failures' leading to failure' of.high pressure (QU) and low pressure injection -(V). The frequency of this sequence is conservative. 2. 1.80E-006 13.2% TSESUX 'l Seismically-induced loss (TSES) of offsite. power-followed by, seismic and random failures of high pressure injection-(U) and depressurization (X). .3.- 8.60E-007 6.3%. TSRB -1S Seismic (TS)1 failure of. reactor building (RB) resulting; in ' failure of all injection. 4. 8.20E-007 6.0%.F2QUV 1' i Fire in' Zone 2 (F2) with' core' damage'resulting from the .f combination of fire-induced and random failures leading toL i failure of'high' pressure (QU) and low pressure-injection ' (V). The frequency of this sequence is: conservative.. 5. 17.30E-007' 5.3% TE50SP2DG2RmC 1-Loss--of offsite power-followed by failure of all onsite power.(TES) and failure to recover.offsite:(OSP2) or -onsite (DG2)~ power in 2-hours and failure to initiate alternate room cooling (RmC) in 2 hours.- 6. 6.70E-007 4.9% TCVQUV 1 Loss of condenser vacuum (TCV) followed by failure of high pressure (QU) and low pressure injection (V). 7. 5.10E-007 3.7% F45QUV 1 Fire in Fire Zone 45-(F45) with core damage resulting from combination of fire-induced and random failures leading to failure of high pressure (QU) and low pressure injection j (V). The frequency of this sequence is conservative. i 8. 4.90E-007 3.6% TE50SP2DG20SPSDG50SP10DGlO 1 Loss of offsite power followed by failure of all onsite power (TE5) and failure to recover either in 10 hours. 1 l 9. 4.80E-007 3.5% TSRPV 3/S Seismically-induced failure of the reactor pressure vessel supports (RPV). 10. 3.80E-007 2.8% TEBCC 1 Loss of offsite power (TE) and common cause failure j of all batteries (BCC). t 1-4 l i _m_______________._ J

mx 4 > i.e .L 4 s 1 .i 6 TABLE 1-2' Continued DOMINANT CORE DAMAGE SEQUENCES L .(see notes) -11. 3.30E-007. 2.4% TCVQUX 1-Loss.of condenser vacuum (TCV) followed by loss of high' pressure injection (QU) and failure to.depressurize the reactor 1(X). L 12. 3.20E-007 .2.3% 'F47QUV. 1 Fire in~ Fire' Zone 47'(F47).with core damage.resulting from-combination-of. fire-induced and random failures leading.to. loss of high pressure (QU) - and low pressure ' injection (V). The frequency.of this initiator is conservative. 13. 3.10E-007..2.3% TMQUV 1 3: Isolation transient (TM).followed by loss of high pressure (QU) and low pressure injection (V). i 14-2.0DE OO7

1. 5 %. TCP2LRU' 4

Loss'of: condenser vacuum ATWS-(TCP2), SBLC. works, operator successfully' lowers level (LH) but fails to control low pressure injection after depressurizationL(U'). 15. 1.80E-007 1.3% TTQUV 1 Turbine trip (TT) event followed by failure of high pressure (QU)fand low pressure. injection (V). 16. 1.80E-007 1.3% F2QWFWECC 2 Fire in Fire Zone 2 '(F2) followed by fire-induced and random failure of all heat removal (WFW).' Containment vented successfully but injection fails (ECC). 5 The-frequency of this initiator.is conservative.- 17. - 1.70E-007 1.2% TElUHURX 1 -Loss of offsite power (TE1) followed by failure of HPCI (UH), RCIC (UR), and depressurization (X). 18. 1.60E-007 1.2% TSESCMC2 3/4 Seismically-induced loss of offsite power (TSES) followed-by either random or seismic failure to insert control rods (CM) and failure of SBLC (C2). '19. '1.50E-007 1.1% TMQUX 1 Isolation transient (TM) followed by loss of high pressure injection (QU) and failure to depressurize the reactor (X). 1-5

$j m a ,+ "i 4 M TABLE 11-2' Continued DOMINANT CORE DAMAGE SEQUENCES + .(see notes). 2 0 '. 1.20E-007~ 0.9%.TMP2LHU' 4 ~ Isolation-transient ATWS-(TMP2), SBLC works, operator. successfully'lowersLlevel-(LH)..butifails to control low... pressure injectionsafter'depressurization (U').. 21'

1;20E-007?

0.9% TMSQUV. 1 Manual shutdown._ (TMS) o followed by failure 'of' high pressure.-(QV)_~and-low pressure injection-(V).- H2 2.' 1.20E-007 0.9% TSRBCM. 1S-Seismic (TS) ' failure of Reactor: Building (RB)1results in failure of'all" injection and failure to'scrami(CM). 23.: 1.20E-007' O.9% TTPPU' 4 Turbine trip ATWS ;(TTP)' with"a stuck open relief ? valve > (P)f followed by failure: of operator to control low pressure. injection after depressurization-(U'). 24. '1.00E-00'7- 'O.7% '.VR1 3/S- . Random reactor vessel' failure. q The'above sequences add up to'approximately 82% of'.the Total CDF. An additional 18 sequences bring.the' total to-90%.- Each of.these additional sequences contribute.less than.1% and:do not add any: additional new functional failures not included'in the top;24l . sequences.. -The information provided for each sequence is:- its rank by.CDF, 3 .thefannual sequence frequency, the percent contribution to the total, the failure' event making.up'the sequence end the accident class. The accident classes are as defined in SARA.with Arabic numerals: replacing Roman. numerals.. See Table 1-5. i .i i 1-6 a

~l e i.i !. O -a J 1 i s TABLE 1-3 'DOMINANTiPOPUIATION DOSE-SEQUENCES 1 . Accident %' Contribution to Total-Rank Sequence Class Poculation Dose 4 . 1c F44QUV 1 10.6- ' 2-TSRB 1S '10.1-3 .TSESUX 1 10.1- .{ 4-

TSRPVs 3/S 8.2 51

~F2QUV-1 4.6 - 6:

TE50SP2DG2R C 1

4.1 ' m 7 TCVQUV. 1 '3.7 j 8J .TCP2 LHU ' ' 4 3.3 _ l SF F45QUV 1

2.8 10' iTE50SP2DG20SP5DG50SP10DG10 1

2.7' 1 11 TEBCC' 1 2.1- ' ' 12 TMP2LHU 4- .1.9 - 13 "TTPPU' '1-1.8-4 1.9 14 ' ' TCVQUX ' 15

F47QUV 1

1.8' ' 16'

TMQUV.

1 1.72 17 .TSRBCMa 1S-124 '18 : TCP2U' 4~ 1.1 . 19 ,TSESCMC2 3/4 1.1 20 TTQUV 1 '1.0 ' 21 '. "F2QWFWECC 2 1.0~

These sequences contribute about 80% of'the estimated population-dose..The next 28 sequences'would bring the total to

~ 'approximately 90%..Each'of these would add less than 1% of1the population dose.- The only additionalLfunctional failures -occurring 11n,these additional sequences are random reactor vessel 1 failure and failure of-pressure 1 suppression following a large LOCA. The sequence.de'finitions are given in Table 1-2 except for the' 4 following: _ TCP2U' - Loss of Condenser vacuum ATWS (TCP2), SLBC works and operator fails tx) control low pressure injection after depressurization (U'). i i 1-7

i -I.. l TABLE 1-4 l

l DOMINANT EARLY FATALITY SEQUENCES

' Risk Accident % Contribution to Total Rank Sequence class Early Fatality Risk 1 TSRPV 3/S 66 2 TCP2LHU' 4 12 3 TMP2LHU' 4-7 4 TTPPU'~ 4 7 These sequences contribute over 90%.of the early fatality risk. The. sequences.are defined in Table'l-2. e i 1-8 i

's '

• .2 y.

, TABLE 1-5 ' ACCIDENT CIASSES 1 CIASS DESCRIPTION EXAMPLE j u 1 1; - ( or.: I) Transients or IDCA's involving. loss TCVQUV: of coolant makeup to the core.- Core melts in~an intact 3 containment. ] i ' 2-(or II) ' Transient or LOCA's involving' loss F2QWFWECC .of long, term heat removal. Long- .' term core melts in a failedior'open i containment.- 13 (or. III) Transients with failure to scram TCP2LHV with failure ~of'all injection. .-Rapid'. core melt in an intact containment. .f ~ .4 (or IV) Transient with failure to scram and TCP2LHU' failure to shutdown. Rapid core' melt.in a failed or open-containment.

S-Core melt due to' reactor pressure

. VR1'- vessel failure with early. containment failure.

4. g

.1S Earthquake initiated transient with TSRB failure'of all injection. Core melts into an open containment l l i 1 1-9 a- _. _. _ ~ n

1 QUESTION 2' For the internal and fire initiated sequences, assess the potential severe accident design mitigation alternative (s), that (if put in place or installed) have a reasonable chance of reducing the. projected severe core damage frequency and off-site risks and (1) which may result in a substantial increase in the overall protection of the public health and safety, and (2).which are. justified by the attendant direct and indirect costs j associated with putting the alternative into place. As noted, this assessment should be limited only to those internal and fire initiated sequences (exclude those sequences initiated by earthquakes over any portion of the earthquake hazard spectrum). Regarding this exclusion, it is the staff's opinion that the incremental severe accident risks due to the nuclear plant relative to all other risks that could potentially be presented by severe earthquakes (up to those large enough to cause the severe core damage accident) would be negligibly small, (i.e., so small that the projected risk reduction benefits attendant to seismic related plant improvements would represent a very remote 1 and speculative projection given the uncertain, competing risks presented to the public off-site from the severe earthquake itself). ~

RESPONSE

ScoDe For the purpose of this evaluation, the range of Severe Accident Mitigating Design Alternatives (SAMDAs) identified in the basis of the LEA contention as defined by the Atomic Safety & Licensing e Appeal Board (ALAB-819, dated October 22, 1985) were initially considered. The SAMDAs identified by R&D Associates (Reference 1), were then considered. The seven SAMDAs listed in Table 2-1 were then further evaluated as representative of the classes of SAMDAs applicable to Limerick. Each is discussed below after a general discussion of the approach to the evaluation. Evaluation ADDroach The design for each of the SAMDAs developed in Reference 1 was reviewed and a revised design basis developed by adding or eliminating features which were considered either needed or not needed to achieve the desired mitigation objectives. The basic design requirements were then translated into design concepts for cost estimating purposes. The cost estimates include both initial and annual costs as appropriate in such categories as engineering, materials, construction, replacement power, regulatory, health physics support, training, maintenance, and QA. It was assumed in estimating the costs and benefits that: j 1 o New equipment is non-safety related unless failure of 2-1

ko', he ! 3p > [ 5 Lp,h1' ^ o

nWp

] ^~ ic the' equipment.could have'an-adverse impact on other. S' safety-related equipment.c j x + .o . Structures,,. systems andjcomponents added'by the. modification and,in the' reactor.enclosurecand control; l structure.will meet LGS Seismic Category IIA' criteria. As described in the1 Limerick FSAR,'thoseLcomponents j listed as> Seismic' Category. IIA are either designed:to-J Seismic' Category:I criteriator.are reviewed to(identify 1 mg ' those whose ' failure could' result in J oss of; required J function of Seismic Category I structures,# equipment,- H or. systems required after an SSE.L Components 3 identified:by.this review ars= considered safety-impacted items and are either analytically: checked to-4 confirm their: integrity.against collapse whenLaubjected; to seismic' loading from the'SSE'or are-separatedLfrom . Seismic Category 'I equipment 5 by a barrier.:, Structures, systems, and components not~ located in. safety-related area, whose-sole function.is-mitigation of severe. accidents will be designed'and: constructed to Seismic ,, y Cate' gory II (non-seismic. category I). criteria..Such structures, systems:and components will comply.with, high-quality industrial coder and standards, e.g.,;the. S' Uniform Building Code. o The designs should not compromise or invalidate.the-existing-design basis of'the plant. ~ ' Costs were estimated for two units and then divided by: 2 to. obtain=a per. unit cost. 'The prcsant worth of the annual costs. was calculated.usingza:40 year plant: life-and:a' discount' rate of 10%. Allicosts are in:1989' dollars. It should be noted that there is a significant scope and regulatory impact uncertainty fa'ctor in the' design concepts, which were developed over a'short period of time for this~ report. There is little or, in some cases, no actual design, licensing or: ' installation experience with these concepts.. Should' detailed-design, licensing and construction proceed,-it.is therefore - i likely that additional complexities.and problems would arise to l further increase the final installed costs. In any case, it is-M very-unlikely that the estimated costs provided herein have been significantly' overestimated. The benefit associated with each SAMDA was quantitatively i assessed-in terms of the estimated man-rems /per. year averted as a result of its installation. The' basis of.this assessment were the internal, fire, and flood core damage frequencies summarized 1 in:the response to-Question 1 and the containment analysis,, source' term analysis and consequence' analysis of the Limerick Severe Accident Risk Assessment (SARA). The conditional population' dose out to 50 miles, given an accident of the various internal, fire'and flood accident classes, is given in Table 2-2 along with the total accident class frequency. The classes are 2-2

.'c i i 1 D ' defined'in Table-1-5.' The source: terms and resulting population dose:are believed to be" conservative as'they.are. based on source i term technology.of the 1981-1983' time frame.l An'. adjustment was ) d made to the SARA results to account for the benefit of the existing plants capability-to. spray or inject water:into the j drywell after a-core melt. The original'PRA/ SARA did'not' include i ithis. 'The~ averted dose wasithen-assessed;by examining the . effectiveness of each SAMDA on each accident class. The benefit'of the estimated' reduction'in population dose was estimated'using'$1000 per man-rem (References 2,,3 and 4) and the

present worth at 10% for 40' years.. The $1000 figurelis used as a:-

surrogate to represent all the offsite effects. Details of the-assessment of~each SAMDA are provided on pages 2-8ff. L I ( ) e f 2-3

a "= ,L ) N '? [ Summary of" Cost' Benefit-Results' i LThe: costs and. benefits of'the mitigation systems are" summarized' -in, Table:2-3. The table provides the1 following: Benefit: LThe' estimated risk reduction,in dollars;perfyear ]l calculated from-the estimated man-rem per year. averted.by the-mitigation deviceftines $1000?per-man-rem.' 1] Total, Benefit: The present: worth'in dollars of the1 yearly benefit?

assuming aL40 year plant life'and a'10% discount-

. rate. 3 Total' j Cost:~ The. total' cost of the mitigation device. including-construction. costs and the present worth of annualEoperating costs;over a 40 year plant life. Benefit / Cost-Ratio:- The. ratio of the total benefits to total costs. JL value areater than 1.0.would indicate a cost : beneficial. mitigation' device.

Cost / Man-rem

< Averted: The cost per man-rem averted. A cost;less than $1000/ man-rem would indicate a cost beneficialJ mitigation' system. The.results presented 'in Table 2-3'show that none of.the mitigation 1 systems examined are cost beneficial. Iri fact, the -results' indicate that no mitigation system is within an orderLof ~ magnitude '(factor of 10) of being cost beneficial. References for Question 2 Resogngg 1.

Dooley, J.L.,

et,al., " Mitigation Systems for Mark II Reactors", RDA-TR-127303-001 (Preliminary), May 1984. 2. Haaberlin,-S.W., et al., "A Handbook for Value Impact Assessment",.NUREG/CR-3568, December 1983. 3. Kastenburg, W.E., et al., "Value/ Impact Analysis for Evaluating Alternative Mitigating Systems", NUREG/CR-4243, I January, 1988. 4.

Stello, V.,

Jr., to the NRC Commissioners, " Mark I Containment Performance Improvement Program", SECY-89-017, i January 23, 1989. N 2-4 _ _ _ = _ _ - _ _ _ _ _ - - _

4 j 3 ( ~ 1 x q t TABLE 2-l' ' SEVERE ACCIDENTLMITIGATING DESIGN 3 ALTERNATIVES EVALUATED ' o' POOL HEAT REMOVAL SYSTEM .A separate.' independent' dedicated system'for. transferring heat from the suppression pool to1the-spray pond, utilizing'a: diesel driven 3,200'gpm pump and k -heat: exchanger without dependence on the Station's-present AC ' electrical ' power: or other: systems. - The: diesel :is-cooled with water' tapped off the spray pond suction line. o DRYWELL' SPRAY A new-dedicated' system for heat and" fission. product-removal.using the Pool Heat Removal System described: above..to inject water'into the drywell. o CORE. DEBRIS CONTROL (" CORE CATCHERS") Two techniques, either a basemat rubble bed, or.using a dry crucible approach, to contain the. debris;in'a known-stable condition in the containment. o ANTICIPATED TRANSIENT WITHOUT SCRAM'(ATWS). VENT

A large.wetwell vent'line to an. elevated release point.

j; to remove heat added to the pool in an ATWS event. p o FILTERED VENT Drywell'and Wetwell vents to a large filter (two types - gravel or enhanced water pool)ito remove heat'and fiss, ion products. 'o LARGE H2 RECOMBINER Independ'ently powered recombiners to remove H2 from the containment in'the long-term after a severe accident, i o IARGE CONTAINMENT VACUUM BREAKER To restore containment pressure to atmospheric level through'20" valves in certain severe accident cases where a vacuum has been produced. e l: 2-5 1 1 _ _ A

E z-I . TABLE 2-2' LIMERICK RISK (POPULATION DOSE) PROFILE _BY CLASS 1 CONDITIONAL'50 MILE CLASS FREQUENCY POPULATION DOSE RISK (per year) (Man-Rem) (man-Rem /Yr) j 1 8.6E-6 5.4E+6 48 i i 2 1.7E-7 9.3E+6 ~2 i 1 3 . 2.7E-7 5.4E+6 l'.. -i '4 1.lE-6 2.7E+7-28 i S 1.0E-8 4.6E+7 0 i i i l 1 l 2-6 I

1.- [ L TABLE 2 l COST / BENEFIT COMPARISON COST / TOTAL TOTAL BENEFIT /- MAN-REM MITIGATING SYSTEM BENEFIT BENEFIT COST COST RATIO AVERTED Dedicated Suppression Fool Cooling $6,000/Yr $57K(1) $25,600K .002 $449,000 Enhanced Drywell $54,000/Yr $516K $46,500K(2) 011 $ 90,100 Sprays $27,500K(3).019 $ 52,300-Rubble Bed Core $13,000/Yr $124K $38,400K .003 $310,000 Retention-Dry Crucible Core $57,000/Yr $545K $119,000K .005 $218,000 't Retention ATWS Vent $27,000/Yr $258K $ 3,900K .066 $ 15,100 Filtered Vent $24,000/Yr $229K $11,300K .020 $ 49,300' -(Gravel Bed) Filtered. Vent $24,000/Yr $229K $ 5,700K .040 $ 24,900 .(MVSS) Large' Hydrogen 0/Yr 0 $ 5,200 .0 Recombiner Large Vacuum Breakers 0/Yr 0 0 .0 l 1 K denotes that the item is in thousands of dollars l l 2 New drywell spray nozzle distribution header j l 3 Use of existing drywell spray header ] 2-7 l .l

_1 4 i J INDIVIDUAL J 1 SAMDA ASSESMENTS 1 i i l i i 1 f i h i t l 4 ll 2-8 1 i

c. -

f,-. a ..= 4 s j SDedicated Suoeression Pool Coolina ,,; ) System-Description:^ 'This system-is designed to remove heat from Le ' the containment (suppression' pool):during-an~ accident where other. '-meansLof pool cooling have been lost. 'It provides an independent 1 means'of pool coolingfby; circulating suppression pool water, JN .through'a. heat exchanger and returning the waterito.the). suppression. pool.- Cooling water from the: spray pond will-be-circulated throughttheishell-side ~ofla-heat exchangerJand. returned to'the: spray pond. Pump. motive power _is provided;byjan independent diesel--locatedJin a new structure; the-pumps aref 'shaftLdriven from the diesel engine. ' Consistent with Reference 1, ,o the assumed' capacity of each pump.is 3200 gpm, and the heat-2 JexchangerL(approx. 4 000 L f t ). removes 4 5 MWt. The new structure-258:xf40' x 20%high,'will.be located underground. 'Three new power supplies will~be= housed in the-new structure. A diesel engine will be mechanically, connected to-both the' pool and pond pumps. .A' diesel generator:(D/G) willi provide a small source of AC power for operating the, isolation valves at the1 containment penetrations and-at the service water tie-ins, for. operating the HVAC, and for miscellaneous: services. E The third power supply is a battery-backed power supply in1the fl new structure for cranking the diesel sets. The system will'be either manually or automatically. actuated. Seauences Mitiaated: This system will mitigate accident, sequences . where containment failure occurs due to steam overpressurization.- It-will prevent containment-failure and core melt for Class 2

sequences involving loss of containment heat removal (e.g.,'TW).

The heat' removal capacity of the system as designed,_ is insufficient to prevent' pool heatup, containment. overpressure failure and the resulting core melt for the class 4 ATWS sequences..This' system has a low probability of' mitigating. Class 1 and Class 3 sequences since drywell failure from other mechanisms (eg., overtemperature);is not prevented. qualitative Benefit: This system.can be highly effective in-p H preventing containment failure and the resulting core melt for Class.2 sequences. Class 4 ATWS sequences will.not be mitigated. Class 1 and 3 sequences will be successfully mitigated only if drywell overtemperature failure is avoided.. Overtemperature' drywell failure can be prevented if the drywell sprays are -operating (see section on' Enhanced Drywell Sprays). Necative Safety Implications: This system involves extending the containment boundary outside of the secondary containment. A leak or break in the piping carrying radioactive fluids could lead to ~ an uncontained radioactive material release, draining of the suppression pool and loss of containment integrity. -l Quantitative Benefit: The dedicated pool cooling system is I estimated to provide the following risk reduction in man-rem per E year. 2-9 I m........

c, ^ i) a i j I ) Man-rem per year Class Reduction i 1 5 ] 2 1 3 0 4 0 Total 6-6 Man-rem per year at 61000 per man-rem yields $6,000 per year or' { an approximate prese..c 9torth benefit of $57,400. Costs: Initial'. Investment $ 23,117,500 0 & M-(Present Worth): 2,495,000 Total S 25,612,500 i

== Conclusion:== .These benefits do not exceed the estimated costs of $25.6 million and this mitigation: device is'not considered cost-beneficial. f l l l l 2-10 ) 1L __

l Enhanced Drvwell Sorav System (EDSS) )- System

Description:

This system is designed to remove heat from the. containment, provide cooling water to debris in the drywell following vessel failure, prevent high temperatures in the drywell and scrub fission products from the drywell atmosphere and/or limit radionuclides release from core debris / concrete interactions during a severe accident, where other means of I containment heat removal and the existing sprays are inoperable, j The system is designed to circulate 3200 gpm of suppression pool . water through a heat exchanger and to spray this cooled water { into the drywell. The dedicated suppression pool cooling system (DSPCS) (previously described) removes heat by cooling the suppression pool water and discharging the removed heat to the spray pond. The. suppression pool water is discharged through the drywell sprays and is returned to the suppression pool via the downcomers between the drywell and the wetwell. The incorporation of the EDSS requires, in addition to the distribution headers, additional valves and control circuitry from those envisioned for the DSPCS. The spray system will be initiated on very high drywell pressure or very high drywell temperatures; if the DSPCS portion of the system was previously initiated, the flow will be diverted to the EDSS. If, for some reason, the DSPCS is not operating, these same pressure or temperature signals will initiate EDSS operation. The appropriate indications and controls will be provided in the control room. This system is a extension of the dedicated pool cooling system discussed separately in this report. Seavences Mitiaated: This system will mitigate all classes of accident sequences. It will prevent containment failure and core melt for Class 2 sequences involving loss of containment heat removal (e.g., TW). The heat removal capacity of the system as designed is insufficient to prevent pool heatup, containment overpressure failure and the resulting core melt for the Class 4 f ATWS sequences. However, this system will partially mitigate the radionuclides releases by attenuating radionuclides in the drywell atmosphere. It will prevent containment overpressure failure and drywell overtemperature failure for Class 1 and 3 loss of core coolant injection sequences. Hence, there is a high probability of this system mitigating Class 1 and 3 sequences. Qualitative Benefit: This system can be highly effective in preventing containment failure and the resulting core melt for Class 2 sequences. Class 4 ATWS sequences will be only partially mitigated. C1, ass 1 and 3 sequences will be successfully j mitigated, j Neaative Safety Implications: Same as for dedicated suppression pool cooling system. l Quantitative Benefit: The enhanced drywell spray system is estimated to provide the following risk reduction in man-rem per ) 2-11 i

q a. q

year.

1 i Man-rem per year Class Reduction 1 43. 2 1 3 1 '4 9 Total 54 -54 man-rem per year at $1000 per man-rem yields $54,000 per-year or a approximate present worth benefit of $516,000. Costs: The costs shown here for the EDSS also includes the costs associated with the dedicated. suppression pool cooling system into'which the EDSS is integrated. t Option 1 presents the costs assuming new and separate drywell spray headers are required. Option 2 presents.the costs assuming. the. spray' headers and nozzles from one train of-the existing. drywell spray system can be used. Ootion 1 Option 2 Initial Investment $44,016,500 $24,517,000-O & M (present worth)- $ 2,533,000 $ 2,514,000 Total $46,549,500 $27,031,000' conclusion: These benefits do not exceed the estimated costs of $.46.5 million and $27.0 million and this mitigation device is not considered cost-beneficial. p i 9 ) i 2-12 I

U3 1 m-

i3 x

Rubble Bed Core Retention Device l .i . Systenc

Description:

This: system consists of-a 'floodable rubble i bed core retention device located in the lowerJpedestal area of the watwell. It is designed to hold and_ cool the debris, and. l prevent-debris penetration through the basemat.into the soil.

l In the Limerick' plant, the suppression pool water extends into the. lower central pedestal area.. In this concept, the hot core 1

melt debris would be directed through 12-inch diameter holes in j the diaphragm floor _and allowed to drop'into the. lower pedestal l area onto a bed of rubble covered by thoria plates. The inside. L diameter of.th.e pedestal at the bas? mat is approximately 20. feet l and therefore, the volume of the-core material would fill this < area to a depth.of less than 4 feet even allowing for 50 percent L . voids. f. This concept is similar to the' design illustrated. schematically i in Figure 3-13 in Reference 1. A stainless steel cylinder is constructed to act as a heat shield for the concrete walls and; prevent. excessive decomposition. Heat would be removed from the: steel cylinder by surrounding water at the lower elevations and-radiation and convection atLthe higher elevations. Thoria plates' e would.also=be added'and extended up the sides a few-feet, if H necessary. To preclude a steam explosion and minimize ex-vessel hydrogen. generation, the core debris retention system is kept essentially dry until after the hot core debris falls onto the rubble bed. Only after the' material has penetrated into the rubble bed area and been cooled somewhat would water be allowed to percolate up through the bed. <i Seouences Mitiaated: Aside from assuring that the debris will ~) not penetrate into the surrounding soil (a low probability event in any case) this system will provide limited additional-1 mitigation.- This system will not prevent containment failure and the resulting core melt for the class 2 loss of containment heat removal system sequences or for the Class'4 ATWSLsequences. This system maY be successful in preventing containment overpressure 3 failure and overtemperature drywell failure by directing the debris away from the drywell onto the rubble bed in the wetwell i] pedestal and cooling the debris for Classes 1 and 3 loss of core { cooling injection sequences, j 4 Qualitative Benefit: This system has a limited potential for successfully mitigating Class 1 and 3 sequences and essentially no mitigation potential for Classes 2 and 4. l Necative Safety Implication: None found. Quantitative Benefit: The rubble bed is estimated to provide the 2-13

-..* si .e ;. .,8 h' <l' o following risk' reduction in man-rem per year: l ^1 Man-rem per year

{

Class Reduction

j 1.

12 l 2 0 3 1 / 4 0 Total 13 .13 man-rem'per year at $1000/per man-rem yield $13,000 per year,- 'or an approximate present worth' benefit of $124,000. ' Costs: Initial Investment: $37,979,000. O'& Mf(Present Worth) 377.500 Total $38,356,500

== Conclusion:== The benefits.of this system are far below the estimated cost of $38.4 million and this mitigation device.is not considered to be cost. effective. Di O 2-14 t __. __ _ _ _ L

1 I Cooled Dry Crucible Core Retention Device System

Description:

The dry crucible retention device is located below the basemat of the present containment. The truncated l is 6 cone-shaped crucible shown in Figure 3-5 of Reference 1, feet in diameter at the top, 3 feet in diameter at the bottom and about 70 feet long to allow for easy entrance of the molten mass. For this concept, a number of large holes (at least 4 - 12" diameter) will be drilled through the diaphragm floor to direct debris flow to'the pedestal area. These holes will be sealed during normal operation by fusible metal plates. The' pedestal area at the basemat is filled with water. This must be blocked off so the area is dry and the core debris can drop through the holes formed after melting the plates in the diaphragm. slab. Then the hot debris will readily melt through a succession of' thin steel barriers and drop into the lower crucible cone. The cone is waterjacketed and supplied with forced circulation to remove residual heat. The cooling water would be pumped and cooled by a dedicated heat removal system similar to the system described in the dedicated suppression pool { cooling system option. Suppression pool water would be removed i from the core catcher area, pumped through the heat exchanger, l core catcher and then the drywell sprays. I This option would require a 6 to 8 foot diameter hole through the basemat which accommodates the upper section of the core retainer. The material can be broken up and removed out of the access tunnel. The access tunnel will be used for carrying all the required material for fabrication and installation of the core catcher crucible. When installation of the dry crucible and supporting equipment is completed, the tunnel will be used for normal access to the supporting equipment. Unidentified complexities and problems are likely to arise during the licensing, design and implementation of this concept. Since no plant has attempted a similar modification, these unidentified problems are expected to significantly increase the estimated costs. Examples of the uncertainties involved include: impact to the plant during excavation, the effects on the seismic design i resulting from a major change to the containment design, and the I effort required to drill an 8 foot diameter hole through the I containment basemat. 1 Secuences Miticated: Aside from assuring that the debris will l not penetrate into the surrounding soil (a low probability event in any case), the core retention portion of this mitigation 1 system will provide limited additional mitigation. However, the drywell spray portion of this system will provide substantial benefits comparable to the Enhanced Drywell Spray Systam described previously. Qualitative Benefit: Comparable to Enhanced Drywell Spray System 2-15

i," :* (.1 Y. . ~. ?. ; Neaative Safety' Implications: A braak or 1sakuin ths lins: . carrying radioactive fluids:outsideLeontainment could11ead to release.of radionuclides,, draining of;the pool and. loss of. H . containment-integrity.s ,t i i i . Quant tat va Benef ts: The dry crucible with'drywell; spray is estimated to provide.the following risk reduction in man-rem per -{ l year: l Man-rem per' year _ d Class: . Reduction-m 1-45 2-1 3 1 4 10 i Total- < 57 ' 1 .57 Man-rem.per year at $1000.per man-rem. yields $57,000 per year j or>an' approximate present worth benefit of $545,000. J 1 Costs:' .InitialxInvestment: $ 116,817,000 0 & M-(Present Worth) 1;945,000 Total $'118,762,500

== Conclusion:== 1 The benefits of.this: system are'far below the estimated' cost of .$119 millioniandLthis mitigation device is not considered to;be. . cost' effective. h 'l l l 1 2-16

p! 0 0 40 e 'l e + p y, ATWS Clean Steam Vent } ' Eygtgm

Description:

This system consists of.an unfiltered high capacity vent pathway.from the wetwell. airspace to'the; ' atmosphere. This system is designed to' relieve the steam-t Lgenerated during an ATWS (AnticipatedtTransient Without Scram)' 4 ' when reactor coolantfmakeup;is available-and where the~ reactor stabilizes at' an average power level-of.-10% ~ of - full; rated power. Steam is relieved to the: suppression pool via.the! main: steam 4 safety' relief; valves; " clean" steam-is.then vented to the' stack. j from the suppression pool air space ~. The system: consists of piping.from the-Unit 1 and Unit 2 suppression 1 chambers to the north. stack which~is shared by both Units. New piping would be' 1 connected 1to the existing 18-inch purge lines close to'the i containment penetration and upstream of the containment isolation I

valves.

j 1 Containment isolation is maintained by two normally-closed,' air-- ) operated, valves in series followed by a rupture disc.- Following-an ATWS, the operator could open these' valves by means of a' key-l locked;~ administratively-controlled switch;11f suppression chamber pressure exceeds approximately 70 psig, the rupture disc will open, allowing.the-excess steam associated with the ATWS to be vented to'the atmosphere via the' north stack. The air-operated valves'are provided with a dedicated power supply and accumulator backup.- ~ The vent lines from Unit 1 and Unit 2 are joined just before entering the stack. In addition to the normally-closed isolation valves'and the rupture disc, each line is provided with a check' valve as a further means of preventing the spread of radioactivity from the Unit undergoing the accident to the.other Unit-. y I Secuences Mitiaated: This system will' mitigate accident sequences where containment failure occurs due to overpressure-zation from slow or moderate steam; production rates. It will prevent containment failure and the resulting core melt for: Class j 4 ATWS sequences (1). 'It will'also prevent containment failure i and core melt for class 2 (e.g.,~TW sequences) characterized by loss of containment heat removal. The system will also prevent overpressure containment failure and provides attenuation of the radionuclides for Class 1 (and 3) sequences-(such as TQUV and station blackout) characterized by loss of coolant' injection to l the core. However, to achieve this benefit drywell failure by other failure modes such as overtemperature and drywell to wetwell pool bypass (e.g., drywell pedestal liner plate failure) must be prevented. L (1) In the absence of containment failure it is assumed that core makeup continues for a sufficient time period to allow i alternative means of reactor shutdown to succeed. 2-17 i -i.Ji

J g',I

q 4-I Qualitative Value

- This system.will be effective in preventing-I J

core melt in Class 2 and 4 sequences and'can be effective in-mitigating ~ class l'and 3 sequences.if drywell'overtemperature failure and drywell,to wetwell pool bypass are' prevented.= Class 4 sequences appear to be more difficult to mitigate than other 1 types of sequences.- This analysis assumes that-the steamLeanfbe 1 L successfully vented at the design-flow rate and that the ATWS l l ' sequences will be mitigated j L . Implications: ' Inadvertent venting.during.an- { E E Neaative' Safety J accident after radionuclides release has= occurred to the containment atmosphere prior.to containment overpressurization H could release noble. gases and a moderately small fraction of the other radionuclides. After vessel failure the release could be - large'because of pool bypass. Quantitative value: The ATWS clean steam vent is estimated to-provide the following risk reduction in man-rem / year. Man-rem per year ' Class Reduction 1 1 2' 1 3 0 4 25 Total 27 ' 27 man-rem per year at $1000/ man-rem yields $27,000/ year or an approximate present. worth benefit of $258,000. Costs: Initial Investment: $3,526,500 0 & M: (Present Worth) $ '353,500 Total $3,880,000

== Conclusion:== The benefits do not exceed the estimated cost of $3.9 million of i the system and this mitigation device is not considered to be cost-beneficial. 1 2-18 i __ _____ ___- _________ L

I ] AL*L !c J Filtered-Vent System .-) System

Description:

This system provides a vent, pathway'from the j 'drywell to a steam condensing and fission product removal device 4 'and from'therefto an elevated release point. The system is designed to. provide the following functions: ;(1)- remova 99% of, a Lthe radionuclides in particulate form and.99% of the. molecular iodine,, (2) accept primary system stored, energy and decay heat for 24. hours without' external cooling,'and (3): process 35_lbm/s

i of steam /non-condensible gases at 70 psig drywell pressure.-

'A1hard pipe vent path is provided from each unit to a common. filtering device. Valving, a rupture disk, and vacuum breakers are located.in each vent path for operational purposes. 'A new. vent stack is located at the filter.to. provide an elevated- . release point.for the filtered stream. Two filter options have been included in this assessment.: The first option is a gravel bed filter (similar to the FILTRA device used at Barseback infSweden) and the;second option is a multi-venturi wet scrubber (similar to the filtering ~ devices used on all other reactors in Sweden). Both devices will meet the: design -performance requirements. Secuences Mitiaated: This device will mitigate sequences where containment failure occurs due to slow steam overpressurization. This system willLprevent overpressure containment failure and mitigate the radionuclides release-for' class 1 and 3 sequences such as. transient initiated and fire initiated sequences'which are characterized by loss of core coolant injection (e.g., TQUV, station blackout).. This device wi11' prevent overpressure containment failure and subsequent core melting.for class 2 -sequences such as transient sequences characterized by loss of ~ containment heat-removal (e.g.,HTW). This device does not have sufficient capacity to relieve the. steam generated by'an ATWS event and hence will not prevent containment failure and core: melt for the' class 4 sequences. This' device is insensitive.to drywell to wetwell pool bypass events (such as'drywell. pedestal-drain line' plate failure). However, drywell' failure from other mechanisms such as overtemperature will compromise the system. Qualitative Benefit: This system can be highly effective in mitigating class 1, 2 and 3 sequences if drywell' failure from overtemperature can be prevented. Necative Safety Imolicationg:. Inadvertent or early opening of the filtered-vent during an accident could release noble gases j and a very small fraction of other radionuclides at a time when the containment is not threatened. j Quantitative Benefit: The filtered vent is estimated to provide the following risk reduction in man-rem per year. i 2-19 =__=:- - - _ -

3. ,y. t i i Man-rem per year Class Reduction 1 23 2 1 3 0 4 0 Total 24 24 Man-rem per year'at $1000/ man-rem yields $24,000/ year or an approximate present worth of $229,000. ' Costs ' Gravel Bed Filter Multi-Venturi-Scrubber System Initial Investment: 10,898,000 5,285,500 0 & M (Present worth) 420,500-406,500 Total $11,318,500 $ 5,692,500

== Conclusion:== The benefits do'not exceed the estimated cost of $11.3 million. for-the gravel bed filter or $5.7 million for the multiventuri scrubber and neither mitigation device is. considered.to be cost - beneficial. j i I i 2-20

s je. \\ va v .x .c q Larce Hydrocen'Recombiners ~ System'

Description:

-'The purpose of this' system is to recombine free hydrogen with oxygen to eliminate.the potential,for uncontrolled combustion..Hydrogensis. generated during a Lpostulated severe accident during the oxidation of metals and 'from'radiolysis of water.- The recombiners are not expected to!be

required prior to venting.

After the' containment has been q vented,; oxygen may'be introduced to the containment and.the ~ volume percent oxygen may be < increased with operation.of: the

containment' sprays which would tend to condense the steam in the 1

containment atmosphere.. Hydrogen / oxygen. recombination will then- ~ be required to prevent.the.long-term formation of combustible concentrations, as hydrogen.and' oxygen will continue to be 2 generated due to radiolysis of water and steam inside the- -containment. -Limerick's' primary containment is inerted with: nitrogen. The. existing hydrogen recombiners are designed and operated to control the? containment oxygen' concentration to below 5% to -prevent' hydrogen combustion. The proposed' system is specified to ~ be' designed fors70'psig containment pressure and capable of processing the containment volume within 2-3 weeks. A dedicated ~ power supply is provided but is probably not required since. ~ normal plant power sources should be available over'the long periods of time when the system is'to be used. The' existing Limerick Hydrogen Recombiner System. consists of redundant combiners located outside primary containment in_the-reactor enclosure. The existing hydrogen recombiners can meet the specified capacity requirement for.a severe accident and'the design concept for this system is to employ the existing hydrogen recombiners, upgrading them to withstand theLspecified design conditions and providing a. dedicated power supply. Secuences'Mitiaated: This system does not prevent (early) containment failure or mitigate radionuclides release'for any identified accident sequence. It is viewed as more of a long-term accident recovery system than a short-term mitigation -system. Qualitative Benefit: Reduces the risk of a hydrogen burn if air is reintroduced into the containment following venting to relieve an internal underpressurization condition. Necative Safety Implication: None found. Quantitative Benefit: No PRA to-date has assessed the risk of very late hydrogen combustion resulting from air introduction following venting into a normally'inerted containment. It is i judged that the risk reduction potential of this system is small. i l 2-21 j l

,N _ g.. ..Y : ,~ Costs: Initial-Investment:. $4,819,500 0 & M (present worth) 392.000 . Total' $5,211,500 L conclusions: Since this system is assessed as having a very small benefit and. its costs are high, it is not considered a cost-beneficial system. 2-22

g-@ f Q

g -;.+ V: >Larae Containment' Vacuum Breaker System' System?

Description:

,This1 system provides a large diameter path .from' atmosphere-to: containment'for use when a high degree 1of vacuum occurs in containment. In essence,'it'would consist-of a .large: pipe.with at least two check valvesfin the line. .Secuence Mitiaation: As in Reference 1 the purpose'of this system would be:to avert containment failure due to external- ' overpressure. ;A qualitative assessment;by1thelBoiling Water Reactor owners? Group of the conditionsEthatLwould lead ~to large ' negative pressures concluded that such conditions are not-expected following recovery of normal-containment. heat removal and-termination of venting. Additionally the' reinforced concrete " Mark II' containments such as Limerick'are not expected to fails even'for pressure. differentials exceeding twice the design 1 differential pressure of 5 psid. Therefore the vacuum breaker' - would-not mitigate.any accident sequences. currently. identified.- Qualitative Benefit: None Necative Safety Implications: Any vacuum breaker actuation would-introduce oxygen into the containment and may produce. conditions suitable for hydrogen combustion to occur.. Quantitative Benefit: None g Costs: Not' estimated

== Conclusion:== This ' system was not quantitatively assessed because of: the determination' of no benefit. I 2-23

i 6 w. 4 ~ k L QUESTION 3 Provide the results from (1) and.(2). above. In view of the positive choice by PEco to maintain its PRA in a "living" status since.the PRA became available, you may' elect to use the PRA insights to enumerate and briefly discuss those various alternatives c~onsidered in the interim and/or improvements actually made to the plant design and operational procedures, that would in your judgement, serve the objectives of (2) above and have served to increase the level of public protection through either prevention or mitigation of severe accidents.

RESPONSE

There are several areas where PRA insights have influenced design and procedural' enhancements and increased the level of public protection through either prevention or mitigation of severe accidents. Desian Considerations The Limerick PRA/ Severe Accident Risk Assessment (SARA) influenced several design features that were installed in Unit 1 prior to its licensing: 1. ATWS Alternate 3A fixes including alternate rod insertion, recirculation pump trip, redundant and diverse scram volume instrument sensors, MSIV isolation setpoint change from level 2 (-38") to level 1 (-129"), and standby liquid control system enhancements' including the' addition of a third pump, automatic initiation, injection through the core spray sparger, use of redundant penetrations for injection, and arrangement of equipment for enhanced testability. 2. ADS air supply considerations including the type and location of backup supplies, physical arrangement of piping and valves, use of dual pilot solenoid valves, and the design of safety /non-safety interfaces. 3. MSIV air supply improvements. 4. Fire propagation barriers for reactor enclosure equipment hatches. Other PRA supported design changes implemented subsequent to the NRC review of the Limerick PRA/ SARA are: 1. Improved ADS initiation logic, in response to TMI Action Plan Item II.K.3.18, which uses a timer to bypass the high drywell pressure permissive. 3-1

' (yQ ,S' 2. -AdditionLof manual ADSLinhibit' switches to. improve implementation'of the BWR Owners Group' Emergency 3-l . Procedure Guidelines-(EPGs). Additionally, it'should be noted that even though they tend to reduce. risk and core damage frequency,-the benefit of the existing.drywell spray and,CRD systems'have not been formally quantitatively assessed and included in the PRA at-the~present -l time. A cost / benefit. analysis of installation of a combustion gas: ~ turbine.was performed as a possible design alternative. The conclusion: reached was that installation of a combustion gas turbine for restoring power after a station blackout is'not-cost-effective. .The benefit gained is'small compared to the cost of ' making:theLuodification'and maintaining it over the-life of the plant. Procedural Considerations Improvements in current operational procedures oversthose-in_ j place at.the time of the NRC. review of the Limerick PRA/ SARA, 1 have: reduced. risk. The Transient Response Implementation Plan Procedures, the Limerick-specific emergencyfoperating procedures,. were_found.to'give. clear guidance to the operators to gain l control of: potential accident events. Operator actions of venting' containment and maintaining. injection to the vessel are considered in the updated PIUL.. Limerick has implemented Revision 3 of;the BWR Owners Group EPGs and Secondary Containment! Control and Radioactiv'ity Relcase Control from Revision 4 of'the'BWR Owners Group EPGs. Limerick is scheduled.to implement the remainder.of Revision <4 of the BWR Owners Group EPGs by the end of 1989; The BWR Owners Group review of the applicability of EPG, Revision 4,-to severe accidento concluded that EPG, Revision 4, is a set of effective accident management-procedures capable-of contributing to the prevention and. mitigation of the consequences of core melt. The NRC Safety Evaluation Report Issued September 12, 1988, stated "We~believe that the BWR Emergency Procedure Guidelines (EPG) provide a basis for: a. significant improvement in current emergency operating procedures." i H 'Other operational procedures implemented subsequent to the NRC review of the Limerick PRA/ SARA include. procedures following a loss of offsite power or following.a station. blackout. Actions directed by the station blackout procedure include establishing i alternate HPCI/RCIC. room cooling, reducing reactor pressure to minimize drywell heatup, and isolating' unnecessary DC loads. In the process of performing the work associated with o incorporating the TRIP procedures into the PRA, areas of the L procedures were identified where enhancements were suggested and I made. The following procedural enhancements have been accomplished: l ~~ l 3-2

c

cy y /r* ? ' 3 1. Thefinstruction to inhibit ADS for an'ATWS has been moved to avoid possibly missing the. instruction-at a branch in the procedure. L-2. The_ATWS procedures have been revised'to call for bypassing,the level one MSIV closure signal prior to the required lowering of the reactor water level for- ] turbine trip ATWS with.a stuck open relief valve. i'l 3. LThe. instruction to intentionally deenergize the-reactor-(enclosure when venting the containment with the large s 18":and 24" lines has been eliminated. I 4. 'The containment venting procedure has been' modified so that.with.high~ rates of pressure rise the large-(18"- and 24") vent paths are-opened rapidly. l l l 3-3 l j}}