ML20198L256

Submits Response to NRC Comments on PRA Attachment 54B

ML20198L256
Person / Time
Site:	05200003
Issue date:	10/01/1997
From:	Haag C WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:	Joseph Sebrosky NRC
Shared Package
ML20198L239	List: ML20198L234 ML20198L256
References
NUDOCS 9801150209
Download: ML20198L256 (19)

Text

. _ _ _ . _ . .- .- . - - - - - .. .

s *

.Frome _Cindy Haag cheagelewesmail.com>

Tot. WND2.WNP5 (jms 3)

Dates 10/1/97 9:32am Subjects Response to NRC Comments on PRA Attachment 54B Joe - Please forward this to Marie. Thanks.

Selow is a summary of W resolution path for Marie 8s taxed comments. We can discuss them further during the 1:00 telecon today.

Cindy According to the faxed comments frow Marie Pohida on 9/20/97 on the sensitivity studies reflecting unavailability of the IRWST injection path through the RNS (valve RNS V023), cases 3 and 5 of attachment $4B reflected only failure of this path'due to hardware failure of the MOV (i.e., setting basic event RN23 MOD 58 to 1.0) .

Westd.nghouse agrees with Marie that, in order to make this path totally unavailable, the operctor actions for opening this valve should also be set to 1.0. Therefore, sensitivity cases 3 and 5 have been rerun to include operator actions RMN.MANO5 and RMN. MAN 05C tot to 1.0. The prelimiaary res alts (still to be calenoted) for these cases change as follows:

. Case 3 CDF changes from 1.2E 06 to 2.0E 06

. Case 5 CDF changes from 6.2E 06 to 7.8E.06.

In response to Marie's other comment on the description for basic event IWX MV-001, I have changed its description in all the core damage output files to reflect common cause failure of the IRWST squib valves. (See attachment $4A for description of how squib valves were represented in the creet for 54A &

$48).

Regards, Isaac & cindy CCs. udi . inte rnet 3 ( 'HAAGCLeigor . wesmail . com" , "WALLACITo. . .

A 3 Attachment 3'

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

10,/40/97. 7110 09:48. FAI 412 374 5535 Ap600 2002 s

h

'( FAX TO JOE SEBROSKY (NRC)

/h pages total October 30,1997 Subject MARXUP.OP AP60'l PRA ATTACHMENT Sea AND CHAPTER $9 Joe.

As we agreed during a shutdown PR A telecon on October 20.1997. Wutinghouw would pro jde a mark up of Rev.

. I of draft AP600 PR A Atl chment $45 (submitted by Westinghouw lener DCP/NRC1067, dated 10/8/97) and PRA

• Chapter 59 to addms Marie Pohidai shutdown PRA documentation inves. The NRC documentation issues Mre- i summmisad in an informal correspondence (email) recoved from you on 10/16/97.

Alsached to this fan we pagn with mukups to Attachment $48 and to Chapter $9.

Watinghouse believes the markups twolve reportmg luuss 2. 3. 4,7. 8, and 9 nf the 10/1697 NRC mail.

Shutdown Evaluation Report section 6.1 will be revised (reporting issue 1) as part of the reporti t nt revi6 ion. Note that reporting issue 5 (AD$ fault tree pictures) were submitted to NRC via Westinghouse lettw DCP/NRCl 10$. dated IW24/97, During the 10!?Q97 telecon we discussed reporbng issue number 6. pertaining to common csun failure estimates, and there was no Arther Westinghour stion on this luue.

Piene provide a copy of the attachment to Marie Pohida. The markups will be included in Attachment $4B and Chapen 39 which will be ir cluded in Revision 11 of the AP600 PRA.

If you have quotjons on the attached information. please call me.

A copy of thn fat will be placed in the Westinghouse informal correspondence.

Regards, bldV '

Cymhie Has Wec. shou Advanced Plant Safety & Licensing cc: E mformal correspondence

.I i

Attachment'4 L . . . _ _ . _ _ __ _ _ _ _ . _ . _ . . . . _ - . . _ _ _ . _ _ _ . _ . .

. t o<30/s?, TW 09:46 l'Al 412 394 5535 Amo L

A1TACHMENT 548 t

( SURGE LINE ASSESSMENT FLOODING EFFECT ON LOW POWER AND SHUTIV -

548.1 ELh* = "'_^!_ =, L.: :_ v S-4w cr f sk.fy M.n 1 Akg (p O w,y L. a Fs.c An examination of the surge line flooding concerns was conducted on the APtW design to determine the capability of the IRWST gravity ispection funcuon during reduced invento conditons l' normal residual heat removal (RNS) cooling is lost.

The results of that examination indicated that, if RNS cooling is lost during reduced inve conditions with the reactor coolar ystem ~ pen, a vont path through the ADS 4th stage is required to preclude the oceanence of surg. line flooding and thereby not affect gravity injection.

Therefore, to address surge line flooding. Westinghouse has included the requirernent for 1 4 A!TS stage 4 valves to be opened, in the success criteria for events dunng toduced in conditions. This success criterion for 1 out of 4 ADS stage 4 valves to be opened is determined by analyses documented in Section 4.8 of the AP600 Shutdown Evaluation Report, WCAP (Reference 5481). Technical Specification 3.4.14 allows for 2 of the 4 ADS stage 4 valves to be out of service during reduced inventory conditions. This technical specification provisio, is to allow for unscheduled re 4,["Wk., ._ _ _ __ - , pairs of as many as two ADS stage 4

..-____.._.....,.,_..--__..__,.w_e..._...,._a...-

. .. .n. _

wry " *riz;; = t: t:^t::: .=A -! " "?lL 2^'! M ?:2= ". J L __ ..

i. L., .h All shutdown drained cai.es are aNected by the surge line flooding issue. These

{ cases are as follows:

IEV RCSOD RCS overdraining when entering mid loop condition l IEV446MiD LOSPD Loss of offsite power during mid loap operation IEV RNSD Loss of RNS during mid loop operation IEV CCWD Loss of CCS/SWS during mid loop operation IEV LOCA24D LOCA through RNS V024 during mid loop operation The method of evaluating the effect of this success criterion on the shutdown PRA involves tit

'following process:

a.

Include ADS stage 4 in the shutdown event trees for drained conditions. The revised event trees are provided in Figures 5481 through 548 5.

4 b.

Construct two ADS fault trees: ADASD for casts with lese-of offsite power available. and ADASDP for the station blackout case. The su. cess criteria summarv tables for th trees are provided in Tables 5481 and 548 2.

1 i c.

Construct ADS support system fault teses for the fault trees developed in item b.

i d.

Construct initiating event sequences for the reduced inventory cases to include ADS success l

or failure, as appropriate. These sequences used the fault trees for the IRWST, RNS, CCS, and SWS that are reported in the shutdown PRA, Revision 6, without modification, k

m.< i .r o,.n m.ma I

E. 10/30/97, TIRI 09:49 FAIl413-374 5635- -AP600 0004

~

l i

b-t to PILA C" $d[ Attach; $4Bf .,

( .A.

= The bueline and focuwd shutdo,rn PRA as reported in Revision 6 is requentified to include the effect of the additional ADS success criteria for surge line flooding. The results of this evaluation are used to derive the shutoown PRA insights, as opposed to the results reported in shutdown PRA Revision 6.

3 Note for the first design change J 5:wed abovS, pertaining to the IRWST i Section valves, the four equib valves have new dependencies on I&C s3 stems and electre power sy,' rems for actuating signals.~ The protection and .  ;

' safety monhoring system along w;th Class IE de power provide the safety related signal to these valves, with the '

' diverse actuation system and the non Class IE de poww providing the nonsafety related signal. Each squib

. valve seceives a safety related s:tuatson signal from a different division. providing a robust redundancy configuration. Smce the check velvas, which did not need a signal to open, have been chenpd to squib valves.

. the IAC and power dependencies prnent a new challenge to the IRWST actuation model. However, as

. discoued below, the IAC and p:*er dependencies are not included 6 the sensitivity study models, e

Because of the robust redundancy of the safety-related signals, and the fact that the nonsafety selated i

systems provide a backup to the safety.related signals, the !&C and power depenuancies will be dominated by common cause failurcs, with random failures being negligible. The non-drained condition shutdown casa already include thow I&C and power common cauw failures in the event sequenen that include failure of the ADS, and in fact, cutset reduction rules will prevent those common cause events from occurring in the sequences involving failure of IRWST, since IRWST follows sucensful ADS y' - operation. Therefore, there is no impact on the design change sensitivity study for the non drained cases due to the new dependency on I&C and power systems.

For the drained condition' shutdown cases modeled in shutdown PRA revision 6. the design for the IRWST required the two IRWST motor operated valva upstream of the check valves to be closed. To inject watw into the RCS upon loss of core cooling, one of these motor. operated valves had to open, These two motor-operated valves received safety related 6;gnals and power from separate divisions. The design has since changed (as noted above) to include four squib valves with a success criteria of one-out.of four to open, and the motor operated valves are no longer required to be closed during drained conditions. Since each of the four squib valves receives a separate safety related IAC signal and a separate safety islated power source, the redundancy scheme has improved. Therefore, the Jesign change impact sensitivity study for the drained cases is conservative with respect w the associated system dependencies.

10/30/07 TH1' 09:49 FAI 412 374 5535 AP600 0005 e.

Add success of the operator act>on to stop RCS drain down in applicable sequences for the RCS overdrain initiating event (RCSOD); this is, represented by basic event RJD6 MAN 04 SUCC.

f. Group the non-drained initiating event sequences from shutdown PRA, Revision 6, with the newly constructed initiating event sequences for the drained cases, discussed in items d and e above. Quantify this group of sequence to assess the effect of the changes in the drained sequences on the braline and focused sh wn core damage frequencies. g

,o o Q wn fr%dh IMs as v94 w, bps g

• ---6'- u h, r: d i:_ t7: : S Mx rd .':xxd ^d::: xn t:,;: -

^i

.

Note, the shutdown PRA, Revision 6, reflects the conservaeive modeling of RNS taes in sequences associated with the loss of effsite power (LOSP) initiatlag event. Dis conservatism consists of using an OR gate instead of an AND at the second level of the RNS fauk tree files RNP2. RNT2, RNP2D and RNT2D. By correcting of these fault tree nies, the baseline core damage hequency changes from 5.5E 08 to 4.7E-08 events per year; the dominant core damage sequences with single RNS train failure from LOSP events are eliminased by these corrections.

The corrected RNS fault tree files are used in this evaluation to resolve the surge line issue.

54B 1.1 Baseline PRA Shutdown Core Damage Frequency seeks .

By incorporating the changes discussed above into the shutdown PRA model, the basr9ne cove damage frequency (CDF) changes from 4.7E 08 to 6.7E.06 events per year; an increase of 49 percent. However, a CDF of 6.7E.06 intill very low. De sequence results for this baseline case are shown in Table 54B 3, and the top 200 cutsets from the quantification output file are shown

( l In Table 54B 4. De associated corrponent importance file is shown in Table 54B 5.

545.1.2 Focused PRA Shutdown Coes Damage F:wquencyp "E",Y Resuks By incorporating the changes discussed above into the model from the shutdown PRA, Revision 6, the focused PRA sensitivity study core damage frequency changes from 4.lE-07 to 5.lE-07 l cvents per year; an increase of 24 percent. However, a CDF of 5.lE 07 is still very low. De sequence results for this focused case are shown in Table 54B-6, and the top 200 cutsets from this quantification output file are shown in Table 548 7. De associated component importance file is shown in Table 548 8.

54B.2 Design Change Effect on Surge Liney-:: "*.;., M:5 se e.

Dis section documegtg,tp effect of design changes on tne baseline and focused PRA results from the surge line :xn..i .;i A addressed in the previous section (54B.1) of this attachment. De cutset output files for the results in subsections 54B.l.1 and 54B.1.2 form the starting point of these additional sensitivity studies.

Dese particular design changes, made subsequent to preparing the design reported in the shutdown PRA, Revision 6, are as follows:

1

1. IRWST injection check valves maintaining the RCS pressuit boundary changed to squib valves; (i.e., check valves 123A/B and 125A/B are changed to squib valves 123A/B and

, 125A/B, respectively) ~4t*,c vgb Aco wva. to isclair. d a T h h d ep4 L' b pnnQs. _ AYT hQt6c" 4tn cM gm. Mew mehr .r.pvpld vdets suA46 acc.

i j m t er oma A=du== sea K4

% b. ce b b M3 d., Agdd+rd Jhm5 +

2 l

10/30/97 Tif0 09:30 l'AX 412 374 5838 AP600 ~ @006 2.

RWST recirculation paths changed from 10 inch ad 4 inch lines to 6 inch lines in al 3.

(

IRWST motor operated valva liga & B and ckek valves 120A & B maintaining the IRWST water level changed to squib valves '

4.

CCS valves to the RNS heat exchanger changed from manual to air operated 5.

The capacity of the service water basin reduced;its duration changed from 24 houn to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> 6.

Service water system valves V037A & B changed frorn air-operneed to motor operated

-7. The number of PRHR heat exchangers changed from two to one 8.

RNS check valves V015A & B on the DVIlines changed to stop check valves.

Dis evaluation is an extension of the studies documented in Attachment 54A. The same ba event changes, shown in Attachment 54A, are made in these studies for the base cases. Additional basic event changes related to the 4th stage ADS components and human error probabilities (HEPs) are made to perform sensitivity studies in addition to those in Attachment $4A.

In these studies. $e effect of the design changes are evaluased together by performing the steps in Section 54A.9, l'ut on the surge line related cutset files in other words, the sensitivity studies, that assess the effect of each design change (as documeraad in Sections 54A.1 through 54A.8) am not W ap LweL h"ina this ~s . saurcise.

t

'!he sensitivity studass for this evaluation, consisting of the following cases, are documented in the subewtions that follow:

• Case 1 - Effect of daign changes on the baselir.c surge line core damage frequency Case 2 -

Effect of design changes on t!*.e focused Pita sensitivity study surge line core damage frequency Case 3 - Effect on case 1 core damage frequency with single IRWST train and 2 ADS 4th stage valves operation Case 4 - Effect on cue I core damage frequency when all HEPs are set to 0.5 Case 5 - Effect on case 3 core damage frequency when all HEPs are set to 0.5.

548.2.1 Case i Sensitivity Study Effect os Beneline (Searge Line) Results The Case 1 sensitivity study estimates the effect of the. design changes on the baseline shutdown core damage frequency of 6.7E.03, discussed above in subsection 548.1.1.

This endmation is conducted by making changes to the basic events in the quantification output file from the surge line Level I baseline case. The following steps were done:

a) Revise basic event probabilities associateJ with the new IRWST configuration and unscheduled maintenance unavailability disonsed in subsection 54A.1; IWX MV OO1 is set to 2.6E-05. IWX CV AO set to 3.0E 05. h0VMOD05S set to 3.0E-03, and IRWMOD06S is set to 0.

(

new I or DreA Aniweem Set 3

'* AP600 @ 007

-.10/30/97:

. , THU 09 50 FAI 412 374_5535 d.b S45y7' Conchas6en

'[ The design changes were evaluated to determine their effect on the PRA mults and insights,W

\ *

, t :-g ' r "9; nfj. The previous subsections provide details on the five quantitative sensitivity studies that basically assess the effects of these design changes on the shutdown (surge 1 lue) baseline and focused PRA core -t- ; hequencies. The resuhs of this evaluation are as follows:

a) When se design changes are incorporased together, the baseline (surge line) shutdown PRA core damage frequency changes from 6.7E-08 to 7.0E 06 events per year; an increase of only 4 percent.

b) When the design changes are incorporated together, the focused (surge line) shutdown PRA core damage frequency changes from 5.1E47 to 4.8E-07 events per year; a decrease of 6 percent.

c) Although some top cutsets are reordered, the top 50 to 100 cutsets are more or less the same.

This can be seen when the cutsets for the baseline and focused shutdown studies in Section 54B.1 are compared with the results from Cases I and 2 in Section 545.2.

d) Although basic events impottance are changed and reordered in some cases, the most important events are relatively the same, as shown by their "importance % decrease" values,,

in the related component importance tables, e) When the HEPs are set to 0.5 in the surge line baseline case, the core damage frequency becomes 2.4E 06 events per year, which is about 34 times greater than the base case core i

damage frequency of 7.0E.Og events per year. This inusase in the base case core damage frequency is somewhat signincant even though a core damage frequency of 2.4E 06 is still quite low. The result indicates that the operesor accons are important in maintaining a very low core damage frequency for intsmal events at shutdown.

f) Although the functions of the IRWST and the RNS and i:s surport systems (CCS and SWS) ao still importsat to maintaining plant safety during shutdown, the function of ADS becomes l more important than reflected in Revision 6 of the shutdown PRA, since the ADS 4th stage is now required to preclude the effects of surge line flooding and thereby maintain plant safety. The importance of these functions are due mainly because of their required operability during reduced inventory conditions, where imtiating events dominate the plant shutdown risk.

548.4 References

$481 "AP600 Shutdown Evaluation Report." WCAP.14837. Revision 0. March 1997.

d i M o? .5 4 y M g4 m M A N M t.4 4g ap Y[ N b n u,,W s n.

s2f be f ced k Asst. . L an sto &

,so,-, g a &w}sucnzs

• c is

-t- ~tJ v #ss c, . w. p gg

AP600 @ 008 ;

10/30/97 THtt 09:51 FA1418 374 5535

( I l

• A reduced reliability in the abiury to isolate the containment does not signincantly change the large release frequency.

I

• here are no operator actions that could be signincandy improved that would result in I a sipificant reduction la the large release frequency. .

l

• A rechiced reliabiury in the passive containement cooling syseem does nor signincantly I change the large release frequency.

I e Common cause fsDures dominees the basic event @w. This shows that single 1 independent f=0ures do not have a large impact on the large release frequency for

. l- AP600 and reflects the F-M ^ y and divnsity of protection against large releases.

l

• h potential for a release of radioactive maestials to the environment is very small.

I This is largely des to the very small core damage frequency and very smau release i frequency. The contaiarnest des'gn provides enhanced deposition of core insterials that l- could be released in a severe accident, and the passive containment cooling system I rninimizes tbs energy available to expel such meserials froen the contamment.

i-l The issults of the at power analyses show the AP600 design includes redundancy and .

I diverney not found in current plants. De safety related pasalve syssetna do not requare ac l power or operator actions to actuate, and the plant design is robust in the prevenson anil l mitigation of the consequences of an accident. "!be AP600 core damage 7.wy and large l release frequency ase rouch lower than has been seen in current generation plants, despite the

('

i many conservatisms buik into the PILA inodels % assurned dose to the enviranawar gives i a sevee accident and a large release is won within the goals set for that analysis.

59J Core Daninge and Severt Release Frequency freen Events at Slustdown f

^

59J,1 Susunary of Shutdown Level 1 Raoubts 6.'1 The low-power and shutdowa ===wnt calculated a core damage frequency of ScSE 08 events per year. The top accident sequences contributeXpersent of the level 1

shutdown core damage . These doadnant sequences rt(ult from

a to

• Fauure of nonnal residual beat removal system doe to a loss of component cooling or service water system initiating event during drained condition, which concibutes Percent of k shutdown cas darnage fraquency q, cy u,,f , ydw Revision: g T@ 59-75 mTe September 30,1996 On*l&W8

10/30/97 THU 09:51 FAI 412 374 5535 AP600 @ 009 s9. FRA Raudes and Jasights M -

( -

1-x% ss of offeine power (LOOP) inidating event * *s drained cc:xhtion, with failure of recovery withm 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, which contri s 13.6 percent of the shutdown cose e eiwi

• less ormal residual beat rerao system inidatmg event during drained condition, which c 'butes 10.4 penent the shutdown core damage frequency

• Loss of offsite wer ting event during drained condition, with success of grid  ;

recovery witlua , which contributes 5.4 percent of the shutdown core damage frequency

• Loss-of c t weiden initiadsg event due to inadvettet opening of RNS V024 dunng cold shutdown ondations, which contributes 5.0 percent of the shutdown

]

4 cm e faquency D C,

• eactor coolant system ov ' g event during drainage to mid. loop, which hQ . contributes 3A percent of the shutdo core damage frequency ND .

p, The descriptions of the dominant sequences are provided in the following paragraphs.

h h a

\

Loss of Composest Coohng or Service Water System Initiating Evesd during Drained Coodition Tw La % . u s.

'd %m 92?loss of decay heat removal initiated by failure of the normal residual be ys removal syst*m as a resuk of failure of the component cooling water or semce water system 23 during mid loop / vessel flange operation, which has an estimated duration of 120 hours0.00139 days <br />0.0333 hours <br />1.984127e-4 weeks <br />4.566e-5 months <br />. Cat damage occun if automatic and manual actuation of the in-containment refueling waser i storage tank injection valms sad manual actuation of the normaliesidual beat removal system I pump suction valve fait ce ifA%5%4 M %.

The major contributon to core damage faquency due to loss of cornponent cooling water systerafservice water system during drained condition are:

l

• C&m cause failure of the in-containment refueling water storage tank innection valves and nonnal residual heat removal system pump suction valve Common cause failure of the strainers in the in-containment refueling water storage tank 03..cm W-gA n5sy 9syk As Sep r 30,1996 m w w + se * ,t.: w as m 59 76 asw YO

• 110/30/97 THtt 09:52 FAX 412 374 5535 AP600 @ 010 Insart to PRA Che_= 39, sme4n 39.5:

(. ..

Failure of normal residual heat removal synom due to a loss of component cooling or service water system initieting event during drained condition, with failure of fourth Hage ADS, which contributes 22 9 percent of the shutdown core damage frequency Loss of normal residual heat removal system inicating event during drained condicon, with failure of IRWST injection, which contnbutes 8.6 percent of the shutdown core damage frequency Loss of normal residual heat removal system laitiating evcnt during drained condition, with failure of founh stage ADS, which coninbutes 4.4 percent of the shutdown cors damage frequency Loss of coolant accident initiating event due to inadvenent opemns of RNS V024 during safe / cold shutdown conditions, which contributes 4.0 percent of the shutdown core damage frequency

=

Loss of offsite power initiating event during drained condition, widi failure of grid recovery within I hour, which contributes 3.0 percent of the shutdown core damage frequency e

- Reactor coolant system overdraining event during drainage to mid loop, which contnbutes 2.8 percent of the shutdown core damage frequency.

E f

F.

l 10/39/97 THU 09:52 FAI 413 394 5535 AP600 @ 015

, . 89. PRA Rasalm and Insights l

(t h Esse of Oedte Power Ishissing i rest during Draised Coedities (with faRore of grid recovery withis I bour)

' Itis segosace is inidated by loa of offshe poww dwing add-looprassel-flange operadon, which has an h desdon of 120 bours, la this sequenos, the normal midual heat i

removal system fails to restart automatically following the instiaang event, and the grid is not recovered withis i bour. Core esmage occurs if automatic and manual actuados of the in.

I_ containment afueling water storage tank injection valves and manual actuation of the normal I residual beat removal systers peep section valve fau.

The nasjer contributors w core damage frequency given loss of offsite power (wnhout grid secovery) dunng draised coo &tico are:

Softwas common cause fades of preesctics and safety anonitonas systern5 tant control

, system inseumentation and control logic cards 3 Fanee of a nonnal residual beat amoval system pump to essert er ran

. Fauwe of a diesel generator to start and nm

. i FaDure of mais circuit breaker 100 (or 200) e opea  !

. Faume to scover ac power within 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />

, I *  !

Common cause failure of abs la<ootainensat mfooling water storage tank injectica l valves and normal.sesidual beat uneoval system pump sucdon valve

{

Common cause faDum of the strainers in the in costainment refueling water storage tank

@ Less of Normal Residual Heat Roanosal System Initiating Event during Dralmed coednes  !

% % a.mM sy a C. _ _.. S . loss of decay best removal initiated by fadme of the normal residual heat removal systern &uing drained conditions. 'Ibe loss of decay best wooval occurs following failure of the sonnel residual beat removal system due to nonnel meidual heat removal system hardware fanks during add-loophessel flange operation. Coco damage occurs if autoamne and 1

mooni acmation of du in<catalasset mAwling wame sierage tank injecdos valva and manual assustion of the normal resideal beat rossoval synem pump suction motorW

valve faa er 3 if *W5 54m[ vh ML 1

a ll T6 Revident 8 Seyesember 30,1996 59 77 se* 8Wa**

10/39/97 THU 09:53 FAI 418 374 5535 'AP600 @018

. - l

( .

The major contributors to cost damage freqeancy due to loss of the morenal assidual beat removal system during drained condstion are:

t~- = -

, _ 16 ,s ,y  ; _ . u _. 7 - g _ _ 7 _ ; 7_ P) $ bak I * * -

Comenon cause fadure of the in --

refueling water storage tank injecdos valves and resi best removal system ousnp sucten valves e 6,- m. ., ma ' t ,,4 ,

4 waW

• A 4. % s s* 9 9+

Cm. cease failure of the strainers in the in containment isfueling waser storage tank of Ottsite Power Initiating Evest during Ceedities (wieb seosses of grid ery within 1 beer)

'his is initiated 'oy loss of offsite during raid loop /vssael flange operanca.

In this uence, the normal residual beat al system does not restart automaticauy following initiating event, but ths

• is recovered within 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />; however, manual normal iss' her vernoval system re (after grid recovery) fails. Core damage occurs I if automatic manual actuation of onte==nt refueling water storage tank injecting valves and rnan actuation of the residual best removal systern pump sucuan moeurw operstod valve f , '

The major contribu to damage frequency given lou of offsite power (with grid recovery) dunng c

• tion am:

• Software cc -

failure of protecnon and safety monitoring system / plant control system in control logic cards

• FaDwe resi removal sysism pumps to run or to restart I

• cause failors of iKontainment refueling water storage tank injectica vos and normal itsidual system pump suction valve Common cause failure of the in the in-containment refueling water storage tank

]- Lasset Ceelant -'

Accident Initiattag Event due to Inndvertent Opening of RNS V034

, r_., Sase,M stetdown co.dits -

his sequence is a lossof coolant accident inineted by inadveneet opening of RNS-V024 daring bothold shutdown conditions when the reactor coolant system is fuled and pressuriasd (which has an animanad duration of 220 boun). Following the initiating event, the core 4 I _----

Revislem: 3 Sepeamber M,1996 m wecee nw e sspr.ik mass 59 78 YN

. - - , - , . . ~ . - . - .--m . - . - - - - - . . - - . - - - . ' - - --..t -a. - . - - - - - - . - - - - . - . . . . , .

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

. 10/30/,97 THU 09:53 FAI 412 374 5535 AP600 @ 013

,99, PRA Rosetts and lasidias makeup tanks are actuated, and the automatic depressuriantion system actusses. Coro damage

l. occurs if the in<entainment refueling water storage tank injection valves do not open.

De rasjor costributors to core damage fmquency des to a loss of<oolant accident through RNS. V024 during horcold shutdown conditions are:

. gas, -

Inadvertent opening of RNS V024 des to operator enor (an initiating event frequency contributor) l

• Common cause failure of the in<ostainment refooling water storage tank issecoon valves Common cause falhus of the strainers la the inweetamment refueling water storage tank

@ Reacter Caetas, JyMan Cs..-L ' " ; Event emetas Dreimage se Mid Imey

Ws sequence is mitiated by reactor cooient system overdraining dunns drainage to mid-loop

, conditions. Draming to midloop has an estimated dwatice of 39 hours4.513889e-4 days <br />0.0108 hours <br />6.448413e-5 weeks <br />1.48395e-5 months <br />. Following the initiating event, manual isolation of the monnal residual hast removal system fails. Cose damage occurs if sensual actuation of the in containment refooling wanst storage tank injection

l valves and manual actuation of the norunal gesidual heat reseoval systeen pump suction moest.

i operased valve fail.

The major commbutors to core damage frequency due to scact coolant sysenn overdraining initiaand d. iring drainage to mid loop are

Conunon cause faihus of the chemical and volume control system air operated valves to close autornatically upos voceipt of low bot leg level signals and failure of the operator to stop draining (initiating event frequency contnbutors)

Operator fails to isolase the normal assidual beat ismoval systan l

• Operator fails to open the in<nneslammar seAasting water storage tank iq)ectico valves 3 Opentor fhils to open the normaliesidual heat semmoval system pump section valve Coennon canae failure of the in caaa=3===* sfueling weser storage tank injection I valves and nonnal residual heat assoval system pump section valve

= i=^_.. J 1

-r-- -

5tb o ' _ - ./ r" ; J . A ^- ' >

1 1

' i Revisies: 8 4

IN tanw.

3,.3 ,os.aos Sepessaber 30,19M

.,ewwww

+,,,r , + - . --- - * -v--

10 /3,0 / tL7 THU 09:54 FAX 412 374 5535 AP600 Q)014 8P. FRA Raoulm and lesighes 1-4_ l Condadoes ne conclusion drawn thwa the sheodows Level I sandy an as fouows:

ne overall abnedown. c$se dunage frequency is very small.

L 88 .'

• laniasing during reassor coolant sysassa drained wh costribees i approniremely

gnusset of ibe total aboedown core damage fregosocy; loss of decay I best removal capatslity (during drained coednion) des to faDere of the componest cooling water sysesin or service wasst systmo me tbs laisiating events with tbs assalast contribution (M'peroset of the abuedows core damage fregesecy).

A 4

e Coswoom casse faDeres of g f8

• t refselims wiesr storage tank, conspossets comenbane approsimnely M penosat of the total shendown core damage I

q-consnos cause faders _ the io<==*=== sfueling waser storage tank T3 comenbetes apprournais # percent of the total abandown core damage frequency.

I His indwass that insistaines tbs reliabdity of the WA-== refueling weser storage ask valves and strainers is imporust in meistaines tbs cuneet level of leer, i

core damage frequency at abuedown.

1 A

• Human errors are not overly important to shutdown cose damage fregosacy. Deve is no particular dominset contributor. Sensitivity results abow that the shutdown cose damage frequency would reinein very low even with liale credit for opensor accians.

5!

- one action, operwor so recognise en need for toe ser coolant synssa i

depressurization dunng cold shutdown raadiaaae is identined as having a i

significant Malt increase value. His indicasas it is important that the p----!- ; include I

this action and the operators understand and me appropriately troised for k.

Individual compassot faDures me not sign Acast contributors to sbuidown cose damage i

fregesney, and there is no perncular doeninant concibosor. This confinns the at pomer coachinice that single indepsedent compooset faDuns do not beve a large inspect on core damage frequency for AP600 and reflects tbs setbadancy and diversity of l pneestion at abundown as weg.

i 'the bHootniament afueling water storage tank provides a significant beoefk during abandown because it serves as a passive backup to the nennal residual heat atmoval

, sysmuk

. Oe um cas ilure yav 5 4 Ma htD Mc d*f ' "W l

1

<.I i c.4b+ 4 oppros M HF'd '

4f 5 4 a;$(;&,e een damag h l

(

~

+bt g da,w xig p as # 6 a'a f b , A q es w %.At *e,u a +84 ,

Q

" \np

' q A % 4.a-o.

Ir",t. L ',

mw.w.=s iwasse

,g.

e 5940 mL f ====

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

10/40/97 THU 09:55 FAI 418 374 5!35 AP600 @ 015

.- 3 spr. m .m it,a.4 e (s

l Table 59 26 l __

SUMMARY

OF AP600 PRA RESULTS l cosm =~ r. 1 l At. Power Shotdown At Power Shutdown

! Iowmal Evoets 1.75 07 (,,7 M08 1.8E4 1.4E-08 l Inseran) Flood 2.2E.10m g,9g.opn N/A N/A l Internal Firs fSE-Of" 3.55.Of" N/A N/A _ _

I!214E' (1) Siax the (Sternal Ars and internal flooding p====ents were conservedve, hading analyses (and tin at-power med shutdown analyses was not) it is not appropriam to add these resuks. That is, the diffenet analyses are not comparable.

l N/A = not applicable -

J l

l I

i i

k

• i septe r 30,1996 um%. I6

_ _ . , _ 3.

i= ,

e e Discussion Topics Related to Westinghouse Responses to Severe .

Accident Open items (sent to Westinghouse December 10,1997)

Functional Requirements for RPV insulation System General configuration

- conical entrance / transition section (add cone angle)

- cylindrical " riser" section '

- straight horizontal / transition panels, resulting in a single point of contact with the spherical portion of vesselif a panel becomes dislodged inlet at bottom

- sized for negligible pressure drop (add pressure value or flow area)

- inlet assemblies open when cavity filled (add buoyant / passive) pressurization doesn't cause assemblies to restrict flow (say how this is shown)

Steam vent at top

- flow area greater than minimum flow area for circulation loop

- steam dampers open when cavity filled (add buoyant / passive)

- once open stay open until manually closed (indicate by means of offset pivot)

Circulation loop

- minimum flow area for the circulation loop, including annular gaps (7.5ft ?)

- minimum flow area of each individual passage along pathway less than 10in2 to prevent plugging Vertica; insulation panels and associated support members and fasteners

- minimum gap between insulation and cylindrical portion of RPV (5.9in)

- minimum gap between insulation and closest point on octagonal reactor cavity (2in)

- withstand bounding loads without exceeding deflection criteria

> loads: 2.7 psi in outward direction, 5.4 psi in inward direction

> max deflection: 4.0in (to maintain a minimum 2in distance between cylindrical portion of reactor vessel Wnd insulation)

Horizontal and transitional insulation panels and associated support members and fasteners

- minimum gap between insulation and RPV (9.0in)

- withstand bounding loads without er.ceeding deflection criteria

> loads: 2.7psl in outward direction,1 Opsi in inward direction

> max deflection: ???in (such that a minimum ???in distance between spherical portion of reactor vessel and insulation is mainteined) Note: minimum distance should he close to 9ln since deflections permittedin ULPU were ninimal.

Means of determining operability inlet assemblies and steam vent dampers can be placed in the fully open position with no observed binding or interference no observed deterioration of the components or changes that coulu affect the ability of the inlet assemblies or vent damper to change position the circulation loop is unobstructed I?y structures or debris. No blockage or loose materials that could potentially form a clockage are observed in the proximity of the inlet assemblies and steam vent dampers or along the circulation loop. The door between RCDT compartment and reactor cavity cannot be latched, and opens into the compartment immediately below the reactor vessel with a force (or pressure) less than ???.

Attachment 5

.t .;

Discussion Topics Related to Westinghouse Responses to Severe -

Accident Open hems (Sent to Westinghouse December 19,1997)

' Core Concrete interactions (Open item 720.418F, 419F)

1. On p.1 of the markup, the bottom of the ventilation duct is stated to be 2 feet above the  :

cavity floor. This is inconsistent with the ANL CCI analysis and the revised reactor cavity drawings, which indicate that the bottom of the duct is 4 inches above the floor.

)

2, Explain how normal leakage is expected to enter the reactor cavity sump, and why core debris would not enter the sump in the same manner, I

3. Since there is a potential for core debris to enter the sump, it is important that this debris  !

remains confined within the sump. The following statements from the markup should be included in the ITAAC for the reactor cavity: "there is no piping buned in the concrete .

beneath the reactor cavity" and "the sump drain lines are not enclosed in either the reactor ,

cavity floor or reactor cavity sump concrete." ,

4. The following items should be clarified in the discussion on the use of MELTSPREAD analyses to assess the effectiveness of the sump curb (p.3 of the markup):

s. how the structures between the reactor cavity and the RCDT rooms (i.e., the steel >

door, the ventilation duct, and the neutron shielding) are treated in the debris transport analyses, and the technical bases for any assumptions regarding the structural failure or melf through of these structures,

b. why the neutron shielding in the vicinity of the doorway would not channel the molten i core debris directly towards the sump, given the angled orientation of the shielding shown in Figure 5.1 of the ANL CCt report, and

c. how the area and volume occupied by the RCDT supports and the sump curb is accounted for in determining the debris spreading area and debris depth, and why this is appropriate (this was called out in the open item but was not clearly addressed in the martrup).

5. Additional design information regarding the door between the reactor cavity and RCDT rooms is needed, given the impor*ance of this flow path for both in-vessel retention of core debris (in which case the 'Joor is assumed to open towards the reactor with minimal force)

F and core concrete interactions (in which case the door is assumed to fail in an outward direction to permit spreading of core debris into the RCDT room).- This information should include the dimensions, orientation, thickness, hinging / latching provisions, and opening force.

L Attachment 6

. 2 Reactor Cavity Flooding System (Open item 720.441F) '

1. The IRWST injection squib valves e o claimed to be diverse from the containment -

recirculation squib valves because they are exposed to and designed to open against ,

different system pressures, As such, the thickness of some of the valve components and the size of the propellant charges are different. Although the difference in the valve design pressure provides some degree of diversity, other failure mechanisms could effect both valves, such as: (1) failure of the valve actuation signal or power supply, (2) maintenance or surveillance errors, particularly if maintenance is performed by the same crew, and (3) failure of the propellant charges due to defects in the chemical composition or environmental / aging effects, particularty if charges from the same supplier and batch are 1

used in both valves. Additional mechanisms or administrative controls to minimize the potential for such common cause failure modes should be identified.

RPV Thermal insulation System (open item 720.423F, 442F, 443F) 1, in the markup of SSAR Section 5.3.5.2, a minimum gap of 2 inches between the RPV and the vertical insulation panels is incorrectly stated to be "under static load conditions associated with containment floodup". This minimum gap should be associated with the

-maximum dynamic pressure load in the direction towards the vessel conditions (12,g5 feet of water) as described in PRA Chapter 3g.

2. A nominal gap between the RPV and the verticalinsulation panels of more than twice the minimum gap (i.e., a nominal gap of at least 4 inches) is prescribed in the revised SSAR section. The dimensions of the riser pipe in the ULPU facility (6 inch diameter) provided a flow area equivalent to a minimum gap of 4.5 inches. The basis for specifying a nominal gap less than simulated in the test facility should be provided.

3. A minimum 6 ft' flow area is specified for the water inlets and the flow path (s) from the loop compartment to the reactor cavity (including the door between the reactor cavity and RCDT compartments). This flow area is substantially less than indicated in AP600 drawings and WEC submittals on other topics, and is believed to be less than the equivalent flow areas simulated in the ULPU tests. For example, the area through the loop nozzle penetrations is 170 ft' based on (previous)

Table B 2 of the PRA 4

the area through the floor between the loop compartment and RCDT room is about 85 ft' based on (revised) Table B-2 of the PRA the flow area through the doorway between the rea:tur cavity and RCDT room is about 25 ft' based on the sketch in the ANL CCI report The basis for specifying a flow area substantially less than shown in the design drawings or simulated in the ULPU tests should be provided.

r--+ ,- ee T w -e w . . - . wwe, -.s --.--~ = e- , -...e .-- ,. ~,- -. - - -, - . ..--tvc--. ...--

• e -S, . .- , m ><o % -rr-r--

o, e .

l 3

4 In order to clearly illustrate the recirculation flow path essential to effective extemal reactor vessel cooling, we request that:

a. the schematic of the reactor vesselinsulation (SSAR Figure 5.3-7) be expanded to include the adjacent RCDT room

b. the expected recirculation flow loop be depicted on the figure

c. . the key junctions along the flow path be labeled S. A minimum flow area through each water inlet assembly of at least 7 in' is specified in 4 order to prevent clogging from debris, a, the basis for the minimum flow area should be provided

b. the minimum flow area should apply to the entire flow path / loop rather than just the water inlets

c. this design provision should be confirmed in the ITAAC l

6. The minimum gap and flow area between the RPV lower head and the conical insulation section, and the associated pressure load, should be specified in the SSAR and confirmed by the ITAAC. (This flow area would be significantly greater than 7.5 ft'since the flow area simulated in the ULPU tests was equivalent to approximately 18 ft'.)

7. The ITAAC inspections of the water inlets and steam outlets should confirm that the closure devices rar?t their design objective of opening passively when the cavity is filled with water.