ML20211C489

Summary of 970717,0805,0905 & 0911 Telcons Between Westinghouse & Staff Members Re Severe Accident & PRA Issues.List of Participants,Topics for Discussion & Disposition for Topics Encl

ML20211C489
Person / Time
Site:	05200003
Issue date:	09/18/1997
From:	Joseph Sebrosky NRC (Affiliation Not Assigned)
To:	NRC (Affiliation Not Assigned)
References
NUDOCS 9709260224
Download: ML20211C489 (13)

Text

.. . . .

September 18, 1997 APPLICANT: Westinghouse Electric Corporation PROJECT: AP600

SUBJECT:

DOCUMENTATION OF AP600 TELEPHONE CONFERENCE (TELECON) TO DISCUSS SEVERE ACCIDENT AND PROBABIllSTIC RISK ASSESSMENT (PRA) ISSUES On July 17, 1997, August 5, 1997, September 5, 1997, and September 11, 1997, telecons were held between Westinghouse and members of the staff to discuss severe accident and PRA issues. Attachment I contains a list of the telecon participants, and Attachment 2 contains the topics for the discussions.

Attachment 3 contains the disposition for the topics from Attachment 2. The staff requests that Westinghouse enter the items from Attachment 2 into the Open Item Tracking System.

There have been subsequent discussions with Westinghouse concerning the PRA.

Attachment 4 contains informal questions that were asked of Westinghouse and Attachment 5 contains Westinghouse's response. Westinghouse does not have to enter this information into the Open Item Tracking System, because Westinghou-se's response, documented in attachment 5, resolves the issue. In addition, Attachments 6, 7, and 8 contain information concerning the shutdown PRA (see item 4 of Attachment 2 and 3).

A draft of this telecon summary was provided to Westinghouse to allow them the opportunity to comment on the summary prior to issuance, original signed by:

Joseph M. Sebrosky, Project Manager Standardization Project Directorate Division of Reactor Program Management Office of Nuclear Reactor Regulation Docket No.52-003 Atts: As stated cc w/atts: See next page W3 Cn p ogppg

-" >. m e gp g

DISTRIBUTION w/ attachment:

Docket File PDST R/F Stollins/FMiraglia, 0-12 G18 PUBLIC BSheron, 0-12 G18 i JRoe DMatthews TQuay TKenyon

. f BHuffman JSebrosky WDean, 0-5 E23 BPalla, 0-8 H7 ACRS (11)

MSnodderly, 0-8 H7 DScaletti JMoort, 0-15 B18 h [C,3

~

/

MPohida, 0-10 E4 AEl-Bassioni, 0-10 E4 DOCUMENT HAME: A:SVR ISS. SUM n ..iv. . ..n . w. m

• See previous concurrence ne. 6a.. . m in. 6.m c . con woout .n.cnm.m.new. r - con .e .n.com.w.new.m. v . no con I j

OFFICE PM:PDST:DRPM l SCSB:DSSA SPSB:DSSA D:PDST:DRPM NhME JMSebrosky:sgddf RPalla* l {

MPohida* TRQuayilM DATE 09/10/97  !! 09/15/97 09/16/97 09/ y/97 0FFICIAL RECORD COPY 9709260224 PDR 970918 ADOCK 0520 A

l3 g g g {g . , ,l l

Westinghouse Electric Corporation Docket No.52-003 cc: Mr. Nicholas J. Liparulo, Manager Mr. Frank A. Ross Nuclear Safety and Regulatory Analysis U.S. Department of Energy, NE-42 Nuclear and Advanced Technology Division Office of LWR Safety and Technology Westinghouse Electric Corporation 19901 Germantown Road P.O. Box 3!5 Germantown, MD 20874 Pittsburgh, PA 15230 Mr. Russ Bell y fir. B. A. McIntyre Senior Project Manager, Programs 4dvanced Plant Safety & Licensing Nuclear Energy institute Westinghouse Electric Corporation 1776 Eye Street, N.W.

1 Energy Systems Business Unit Suite 300 Box 355 Washington, DC 20006-3706 Pittsburgh, PA 15230 Ms. Lynn Connor

. Ms. Cindy L. Haag Doc-Search Associates Advanced Plant Safety & Licensing Post Office Box 34 Westinghouse Electric Corporation Cabin John, MD 20818 Energy Systems-Business Unit Box 355 Dr. Craig D. Sawyer, Manager Pittsburgh, PA 15230 Advanced Reactor Programs

.' GE Nuclear Energy Mr. M. D. Beaumont 175 Curtner Avenue, MC-754 Nuclear and Advanced Technology Division San Jose, CA 95125 Westinghouse Electric Corporation One Montrose Metro Mr. Robert H. Buchholz 11921 Rockville Pike GE Nuclear Energy Suite 350 175 Curtne" Avenue, MC-781

, Rockville, MD 20852 San Jose, CA 95125

! Mr. Sterling Franks Barton Z. Cowan. Esq.

U.S. Department of Energy Eckert Seamans Cherin & Mellott NE-50 600 Grant Street 42nd Floor 19901 Germantown Road Pittsburgh, PA 15219 Germantown, MD 20874 Mr. Ed Rodwell, Manager Mr. S. M. Modro PWR Design Certification Nuclear Systems Analysis Technologies Electric Power Research Institute Lockheed Idaho Technologies Company 3412 Hillview Avenue Post Office Box 1625 Palo Alto, CA 94303 Idaho Falls, ID 83415 Mr. Charles Thompson, Nuclear Engineer AP600 Certification NE-50 19901 Germantown Road Germantown, MD 20874

AP600 PHONE CALL PARTICIPANTS CONCERNING LEVEL 2 AND PRA SEVERE ACCIDENT ISSUES JULY 17, 1997 BABE ORGANIZATION CINDY HAAG WESTINGHOUSE BOB LUTZ WESTINGHOUSE JIM SCOBEL WESTINGHOUSE SELIM SANCAKTAR WESTINGHOUSE BOB PALLA NRR/DSSA/SCSB HIKE SNODDERLY NRR/DSSA/SCSB JOE SEBROSKY NRR/DRPM/PDST AP600 PHONE CALL PARTICIPANTS CONCERNING EQUIPMENT SURVIVABILITY AUGUST 5, 1997 HABE ORGANIZATION

' CINDY HAAG WESTINGHOUSE DICK HILLER WESTINGHOUSE JIM SCOBEL WESTINGHOUSE DAN MCDERMOTT WESTINGHOUSE a

BOB PALLA NRR/DSSA/SCSB MIKE SNODDERLY NRR/DSSA/SCSB DINO SCALETTI NRR/DRPM/PDST J0E SEBROSKY NRR/DRPM/PDST AP600 PHONE CALL PARTICIPANTS CONCERNING SANDAS AUGUST 5, 1997 M6BE ORGANIZATION CINDY HAAG WESTINGHOUSE TERRY SCHULZ WESTINGHOUSE JIM SCOBEL WESTINGHOUSE B0B PALLA NRR/DSSA/SCSB DINO SCALETTI NRR/DRPM/PDST JJE SEBROSKY NRR/DRPN/PDST AP600 PHONE CALL PARTICIPANTS CONCERNING SHUTDOWN PRA SEPTEMBER 5,.1997 HABE ORGANIZATION SELIM SANCAKTAR WESTINGHOUSE ISAAC WALLACE WESTINGHOUSE SUE FANTO WESTINGHOUSE JIM FREELAND WESTINGHOUSE BOB PALLA NRR/DSSA/SCSB MARIE P0HIDA NRR/DSSA/SPSB J0E SEBROSKY NRR/DRPM/PDST Attachment !

- - ~ - - - -

t.

AP600 PHONE CALL PARTICIPANTS CONCERNING SWTDOW PRA SEPTEMBER 11, 1997 Natif, ORGANIZATION SELIM SANCAKTAR WESTINGHOUSE ISAAC WALLACE WESTINGHOUSE C1NDY HAAG WESTINGHOUSE JIM FREELAND WESTINGHOUSE MARIE P0HIDA NRR/DSSA/SPSB JOE SEBROSKY NRR/DRPN/PDST 1

a b

4 k

<v-e-T'-rr r-e-m-* . p-- .

y.-p ,y.y y3.,yy .9 -,s,w--4g 7.--. y w g ..,.7.,. ,-w.m.. , 9_- . ,, ,.-r-'s * --n-- '- @

. I

Discussion Topics for Meeting /Telecon with Westinghouse

1. PRA Insights: (1) the completeness of insights, (2) the level to which related system features are described, and (3) the need to tie the -

4 insights to specific ITAAC, COL action items, SSAR sections (rather than PRA chapters that may disappear). Examples of weaknesses:

writeup for RPV insulation design and referenced PRA section(does not1) insights mention key features of the system /, redundant / diverse cavity flood lines and valves, ball in cage check valves, buoyant dampors, flow areas and gaps / clearances) or provisions to assure that the system / components will function as designed (e.g., ITAAC, SSAR figures, tech specs), and only refers to Chapter 39, which may not become part of DCD, (2) insights writeup doesn't discuss relationship between PCCS and containment

integrity and identify mechanisms / commitments for assuring that drain plugging will not occur, (3) items that will be dispositioned to severe accident management should refer to COL action item rather than the SAMG framework, i 2. Cost beneficial SAMDAs

additional information is needed regarding the system description, costs, and risk reduction estimates for two SAMDAs:

(1) diverse IRWST valves, and (2) self-actuating valves. The first of these appears cost-beneficial when averted on-site costs are considered.

Justification for not implementino this SAMDA is needed.

3. Requantification of Level 2 sensitivity and importance analyses for internal events at power assuming that events with diffusion flames near the containment shell will result in containment failure.

4. Justification for basing Level 2 portion of shutdown PRA on earlier (Revision 3) event trees on which the staff had many questions. These questions were resolved by going to a significantly modified approach for the at-power PRA, but are still relevant to the shutdown PRA.

5. Capability / pathway / procedures for venting as an accident management

, strategy needs to be addressed. Although venting is not expected to be necessary in most sequences, it may be needed in the event of reactor l vessel failure (since deterministic calculations indicate that early containment failure from steam explosion is not likely).

6. The accident management COL action item, and what specific issues are flagged for inclusion (e.g., should include use of post-72h SSCs).

l Attachment 2 i

4

... .-r - - , - ,--,-t-- - - - ~ , - - , - - - . . -

e n . ,. - - - = ,

-,m- -- - , , , - ,> ,ww- ~v r

I l

7. Appendix D of the PRA contains Westinghouse's equipment survivability assessment. The staff would like to talk about this assessment in general and at this time has four specific concerns which are as follows:

a) The staff is having difficulty' correlating the time frames given in the assessment to the environmental conditions that are expected during the time frames.

b) It appears that Westinghouse is qualifying entire systems instead of specific components. The staff feels that another level of detail is needed to assure proper procurement of specific components. Westing-house shm!1d refer to Section 19.11.4.4 of the ABB-CE System 80+

design control document as an example of the level of detail the staff has found acceptable in the past.

c) The staff believes that the Global hydrogen assessment maybe a) pro-priate, however, Westinghouse does not address local hydrogen surn-ing.

d) The staff would like to further discuss where certain systems /

components are addressed in Appendix D for example: cavity flooding system valves, the equipment used to address TMI issue (NVREG-0737)

Item II.F.2, post accident sampling system SSAR 9.3.3.1.2.2 states this function is performed by the primary sa(mpling system for the AP600), containment vent, and containment sprays.

Disposition of Items from Attachment 2

1. Westinghouse will consider revising the insights from the level 2 and 3 PRA to address the staff's concern.

2. Westinghouse will provide the staff with further information on the diverse IRWST valves SAMDA.

3. Westinghouse will provide a sensitivity and importance analysis to address the staff's concern. Subsequently provided by Westinghouse in an August 29, 1997, letter (NSD-NRC-97-5304).

4.

• This item was discussed during the conversation on September 5, 1997, and September 11, 1997. Although these conversations focussed on level 1 PRA, the level 2 PRA was also discussed. Concerning ths level 2 PRA Westinghouse agreed to evaluate requantyfing the level 2 analysis, or to provide justification for not redoing the analysis.

Concerning the level 1 PRA, the staff discussed an August 21, 1997, letter (HSD-NRC-97-5285) pertaining to the AP600 shutdown PRA and surge line flooding (RAI 720.303). The staff was concerned that Westinghouse response was based on an old revision to the PRA that did not model the squib valves in the IRWST injection line. In addition, the staff was concerned that the results of the letter reflected a success criteria change in the shutdown PRA. Westinghouse informed the staff that they would not requantify the shutdown PRA. Attachment 6 contains the actions that the staff expected Westinghouse to take to resolve the issue. The information in Attachment 6 was faxed to Westinghouse. Please note for item 5 of Attachment 6 the beginning of the sentence should read as follows: "For cases 1 and 4 above...".

In addition, Attachment 7 contains information that was faxed to Westing-house prior to the telecon on September 5, 1997. Attachment 8 contains

' Westinghouse's response to the questions. Attachment 7 and 8 are included for information only, because Attachment 6 contains the Westing-house action items.

5. Westinghouse will provide a discussion that addresses the requirements of 10 CFR 50.34(f)(3)(iv). Westinghouse will also look at how the issue was resolved in the General Electric final safety evaluation report (Sec-tion 20.5.44) for guidance.

6. The NRC will look at the revision to the severe accident management guidance that Westinghouse provided in Revision 1 to WCAP-13914. In SSAR Revision 15, Westinghouse provided a change to Section 13.3.1 to add how tha post-72 hour actions are to be treated.

Attachment 3

• 1 1

7. The staff requested Westinghouse to read the requests for additional i information that were sent to Westinghouse on the subject (RAls 470.34, 470.35, 470.36, and 470.37 .

RAI 470.36 was not addresse)d in Appendix D to theThe PRA.The staff was also co staff agreed to review the Appendix and to reconvene the conversation with Westing-house concerning the issue at a later date.

the staff issued RAls to Westinghouse on the s(Subsequent ubject in a letter to thedated telecon September 11, 1997. Therefore the phone call was deferred and this item is superseded by the September 11, 193/, RAls.)

3

Informal questions sent to Westinghouse concerning the PRA The sensitivity case for containment isolation (50.6.2.2) a> pears to represent the effect of increasing the isolation failure probability )y about a factor of 5 for all accident classes except '6*. For '6" it looks like the failure probability is increased by about a factor of 60. (I am basing this on the split fractions shown in figures 43-3 through -11). Can you confirm the followingt

1. Why the-15 split fractions for accident classes lA and 3D/10 are an order of magnitude higher than for the other accident classes that are also quantified based on the CIC fault tree.

2. Why the IS split fraction for accident class 3C (based on OTH-CNB+CID) is not greater than the split fraction for accident class 3A (based on CID).

3. Why accident class 6 has been treated differently in the sensitivity study.

l 4. Why the IS split fractions are so much lower in the focussed PRA than in the baseline PRA for accident classes IA (a factor of 15 lower) and 3D (a factorof23 lower).

Attachment 4 i

~

. o From: 'Scabel, James H? <tcobeijhewestinghouse.com>

To: Robert Pelle[SMTP:RI.739erc.aov)

Dane. Wed,10 Sep 1997 09.50 $6 0400 To: Bob Palla

Subject:

. Desponses to your questions in your e mail 9/10/1997 B(L, Here are the toeponses so your queenons related to containment !aoleuon eensitivity mode (15 node)in the AP600 PRA containment event ires.

1. .

In the accident classes l A and 3D, the Boolean multiplication of accident sequence CDF cutases and the 12 failes eussets share more IdtC related failures than d.ose cutsets that are in the low premure RCS accident sequences. For that reason, the I&C fallwes that dominais CDF for 1 A or 3D also fail the containment isolados signal. I briefly looked at the top cutases for (l A and CIC (slwe), and also (3A and CIC -

fadwe) cussets and observed that top cutsets contain falwes like common cause software; loss of ESF oebiness, etc., which support the above statement.

2.

he C1D fault tree used in 3A has a fallwe probability of 1,64 E 03, whomas the XCID ( CID +

0'IRCNB) fault tree used for SC has a failure probability of 2.64 B 03, as expected. When the Boolean amitiplicesion is performed on the product of accident class CDP cutness and the IS mode syssem failwe cutests, the 18 aode split fracdons for 3A and 3C two out to be 2.62 E43 and 2.65 E 03, respectively, he differense is due so the following:

Since the CDP cueness for 3C only involve RV ru,iews, they are IF; dat from the l$ node cuensis; thus the IS split traeuon is the same ( or about the same) as the XCID fault tree result. On the other hand, thac are some common actuanon or component failwes between the cutsets of 3A and CfD to make their product to come out higher than a scalar probability multiplication; thus the split fraction for 3A 15 node goes up somewhat froen she fauk iree value of 1.64 E-03 to the calculated value of 2.62 B 03.

3. _

he 15 mphs fractions accident elatees IA and 3D have higher failure probabihties (at the order of 0.01) than the rest of the accident classes (at the order of 0.001). he intention of the IS sensiuvity case was to raise the is split fIractions of thoac accident classes with 10 2 IS failwe to 0.1 and raise the remaining spht fractions to 0.01 in the Accident class 6 was inadvertently included in the first group, despise its IS split firection oeiginally belag ceds of 0.001, I have set the accident class 6 IS node spht nfracte to 0.01 in the same sensitivity analysis and observed that the resuk is the same as the one in the subitad case with is = 0.1 within two significant figwes.

Sorry for the confusion it orested!

4.

In the focused PRA model, the low of offshe power is postulated. Dis causes loss of air and the air operated containment isolanon valves clou due to loss of air, without a need for !&C actuation. hus the IAC fanfares, which dominated base case for both the accident esquence cutsets for I A and 3D and also affected 1$

node failwe, are no longer relevant.

Then, the l$ split fractions for I A and 3D are dominated by other faalutes, just like the otha accident classes; the IS split fractions are similar to those of the other accident classes.

I hope theme explanations respond to your questices.

Regards, Selim Attachment 5 2.

Material that the staff is expecting to receive based on 1

September 11, 1997, phone call l

1. Using the truncated cutsets from the last shutdown PRA quantification, Westinghouse will generate a revised CDF, a revised list of dominant 1 cutsets, and revised importance analyses, assuming the AP600 design changes incorporated in Attachment 54A of the shutdown PRA and the new success criteria (incorporation of the 4th stages ADS valves to preclude surge line flooding).

1

2. Using the truncated cutsets from the last shutdown PRA quantification, Westinghouse will generate a revised " focused PRA" CDF, a revised list of

" focused PRA" dominant cutsets, and " focused PRA" importance analyses, assuming the AP600 design changes incorporated in Attachment 54A of the shutdown PRA and the new success criteria (incorporation of the 4th stage ADS valves to preclude surge line flooding).

3. Westinghouso will provide a justification for not re-quantifying the PRA given the new success criteria.

4. For case 1 above, Westinghouse will provide a sensitivity case assuming minimal compliance with AP600 TS during the entire cold shutdown and refueling period (with RCS level less than 23 feet above the reactor vessel flange). This case will assume: 1-out-of-2 IRWST gravity injec-tion paths are inoperable, gravity injection through the RNS pump suction line is inoperable (RNS valve V-023 is inoperable), and 2-out-of-4 4th stage ADS valves inoperable.

5. For cases 1 and 3 above, Westinghouse will perform a sensitivity study changing all human error arobabilitios to .5. I think that this result will be more meaningful t1an setting all human error probabilities to 1.0.

T

+

Attachment 6

o O

Questions the staff had on August 21, 1997, letter from Westinghouse (NSD-NRC-97-5304)

I have three comments / questions for Cindy Haag about RAI response 720.303 (0ITS Item 3007).

1. The cutsets don't make sense as we discussed last Wednesday.

2. In the revised RAI response, in the fault trees, Westinghouse credits 4 ADS valves being available during reduced inventory conditions. However, TS only require 2 ADS valves to be operable. Westinghouse needs to provide a sensitivity study assuming the COL ap)11 cant follows TS.

Specifically, Westinghouse needs to deterwine 11e CDF assuming only two 4th stage ADS valves are operable and one IRWS1 injection path is avail-able.

3. Why did Westinghouse not update the PRA to reflect this new success criteria considering that 4 event trees have changed?

Attachment 7

Westinghouse response to Attachment 7 Questions i

/ 1. The surge line analysis was a sensitivity study of the shutdown results reported

( in Revision 6 of the PRA, which did not model the IRWST squib valves. The Rev. 6 shutdown model used the MOVs in the IRWST injection lines for drained Cases.

, in the Rev. 6 shutdown model, the ADS was modeled only for non drained conditions and with squib valves. Therefore, in the surge line flooding analysis, ADS squib valves were modeled for the drained cases.

Therefore, the components modeled in the surge line flooding snalysis are consistent with the Rev. 6 model; MOVs are modeled in the IRWST and squib vetves are modeled in the ADS.

The shutdown PRA was not updated since revision 6, but an assessment of the impact of design changes subsequent to Rev. 6 was conducted earter this year.

That asses went included the IRWST design changes such as the squib valves and was transmitted to the NRC as Attachment 54 (letter: NSD NRC 97 5044) dated March 31,1997.

2. The ADS was modeled consistent with the IRWST with regard to Tech Spoos for drained conditions. As was discussed previously with the NRC, the TS allowance for one train of IRWST to be inoperable is not intended for scheduled maintenance. Theretore, both trains of IRWST are expected to be available during drained operations. Similarty, the TS allowance for 2 ADS valves to be inoperable during drained conditions is not meant for scheduled maintenance I unavailability. Therefore, the baseline success enterla should not reflect one available train of IRWST injection and 2 available ADS valves

3. As stated in item one above, the intent of the surge line flooding analysis was not to u(Mate the ohutdown PRA.

Attachment 8