ML20203N421
| ML20203N421 | |
| Person / Time | |
|---|---|
| Issue date: | 09/26/1986 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1550, NUDOCS 8610090005 | |
| Download: ML20203N421 (341) | |
Text
OR/GWAL M88/Efd 4 UN11ED STATES NUCLEAR REGULATORY COMMISSION IN THE MATTER OF: DOCKET NO:
ADVISCRY COIO1ITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SEVERE ACCIDENTS and SUBCOIGIITTEE ON OCCUPATIONAL AND ENVIRONMENTAL PROTECTION SYSTEMS O . .
LOCATION: WASRINGTON, D. C. PAGES: 1 - 249 DATE: FRIDAY, SEPTEMBE'R 26, 1986
'~;Qp.
'p p rz. (, g%i *~:p.#.*") f^ {} O k 0 s f 3, I; 3 ;; ', }' 1 , , ;! . "a l- -' , ' i
- c. ._. ,
/)
s a /; u, .1 u.....- .
Do Not Remove fmm ACRS OLiice ACE-FEDERAL REPORTERS, INC.
O s OfficialReporters 444 North CapitolStreet
( Washington, D.C. 20001 (202)347-3700 8610090005 860926 PDR ACRS T-1550 PDR NATIONWIDE COyERAGE
({) PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSIONERS' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS FRIDAY, SEPTEMBER 26, 1986 The contents of this stenographic. transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards (ACRS), as reported herein, is an uncorrected record of the discussions recorded at the meeting held on the above date.
No member of the ACRS Staff and no-participant at
[) this meeting accepts any responsibility for errors or inaccuracies of statement or data contained in this transcript.
6 O
i 1
1 'l UNITED STATE OF AMERICA 2 NUCLEAR REGULATOR COMMISSION 3
4 5 Advisory Committee on Reactor Safeguards Subcomittee on Severe Accidents 6 and -
Subcommittee on Occupational and 7 Environmental Protection Systems 8
9 Friday, September 26, 1986 Washington, D. C.
10 Before: Dr. Dade W. Moeller, Member 11 Mr. Jesse C. Ebersole, Member Dr. Carson Mark, Member 12 Dr. Chester P. Siess, Member
- O 13 Dr. Michael L. Corradini, Consultant r 14
', 15 16 17
)
18 19 20 21 22 23
() 24 25 ACE FEDERAL REPORTERS,-INC. Washington, D.C. (202). 347-3700
2
,-s 1 P R O C E E D I N G S:
f 1 w/
2 (8:30 a.m.)
3 DR. MOELL ER: The meeting will now come to 4 order.
5 This is a ' joint meeting of the Advisory 6 Committee on Reactor Safeguards Subcommittees on 7 Occupational and Environmental Radiation Protection Systems 8 and Severe or Class Nine Accidents.
9 I am Dave Moeller, and I will be serving as 10 chairman of this subcommittee meeting today.
11 The other ACRS members with us are Chester 12 Siess, Forrest Remick and Carson Mark.
[l
\_- 13 We also have with us M. Corradini who is 14 serving as a consultant to the subcommittee.
15 The purpose for the meeting is to allow the 16 subcommittees to gather and exchange information with the 17 NRC staff and Public Service Company of New Hampshire.
18 The subcommittee will review results of the 19 Seabrook probabilistic safety assessment and updates to 20 this report which includes the impact of the revised source 21 term to some degree on the strength and leak tightness of 22 the Seabrook containment.
23 Although the subcommittee meeting will dwell 24 heavily on Seabrook, it involves a much wider range of l[ )
25 variations, many of which are generic to the subject of ACE FEDERAL REPORTERS, INC. Washington, D.C. (202).347-3700
C 3
1 nuclear power.
2 These include the NRC safety goals and their 3 implementation into the regulatory process, the NRC source 4 term research effort and the development of the new source 5 term code package.
6 As I say that, though, I want to acknowledge 7 that the approach being taken in the Seabrook Station 8 assessment seeks to divorce the source term analysis from 9 the questions from these, from these particular. questions.
10 Thirdly we will be looking at emergency 11 planning and any changes that might be advisable in a 12 generic sense in light of the lessons learned at chernobyl.
13 The ACRS has had and is having meetings on all 14 of these subj ects and today's meeting is essentially, then, 15 part of this total effort.
16 Our meeting today is primarily a status report 1
17 on the subject, and there probably will be additional )
18 subcommittee meetings before this matter is taken to the l 19 full committee. I 20 We will, though, offer, I hope, a range of 21 comments from subcommittee members today, and I hope that 22 we will be able to interact both with the applicant and NRC 23 staff to at least give you something in the way of our 24 thoughts and responses.
25 Seated on my right is Mr. E. Igne, who is the ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
. 4 1
- -)
l
'1 cognizant ACRS staff member f or the meeting.
2 The rules for participation in today's meeting
! 3 have been announced as part- of the notice that was l 4 published in the Federal Register on September 9, 1986.
I~
5 To assure'that our meeting is properly
- 6 recorded we ask that
- each1 speaker first identify himself or
]
7 herself and speak with sufficient clarity and volume so ,
8 that he or she can be readily heard.
l 9 As of this morning we have received no written ;
}
} 10 comments or requests for time to make oral statements from i
i 11 members of the public.
12 If, however, there is anyone here from the l 13 public who desires to make an oral statement'during the i 14 meeting today, we ask that you check with Mr. Igne during 15 our first break this morning, which should be some time l 16 around 10:15 a.m.
! 17 If you will do that, he will try to schedule .
18 your presentation as part of the agenda later in the 1 19 program. '
20 Written comments also could be received by Mr. :
1 21 Igne at that time, if you decire to have them mado part of 22 the written record.
1
,' 23 We will now proceed with the meeting. But
() .
24 bef ore we do so, let me ask if any of the subcommittee l
25 members or the consultant have any comments or questions or i
' (202)'347-3700
~
ACE FEDERAL REPORTERS, INC. Washington, D.C.
1 5
1 7
gg 1 remarks at this point?
l
'V
- 2 DR. SIESS: None.
i 3 DR. REMICK: No.
l 4 DR. MOELLER: There being none, we will move 5 ahead and call upon-our first discussion leader. This will i
! 6 be a presentation by the NRC staff, i
) 7 It will be led by Mr. Vince Noonan, Director i
!. 8 of PWR Project Directorate Number 5, Division of -PWR i
l 9 Licensing, the Office of Nuclear Reactor Regulations.
1 l 10 Mr. Noonan.
11 MR. BLAKE: Good morning. My name is Vince j 12 Noonan. I am the Director for Project Directorate Number I
j 13 5. I am responsible for licensing for the Seabrook 3
i- 14 Station.
j 15 (Slide.).
16 MR. NOONAN: This morning we are .here with the 17 utility to present to the subcommittee some views on a-i 18 status report on our review of the- technical adequacy of 19 the Seabrook Station, management and emergency planning . .
- 20 study and emergency planning and sensitivity study, 21st of ;
- 21 July, '86.
22 This meeting f rom the viewpoint of the staff I l {
23 particularly is a status report. We have not progressed 1 I
() 24 very far in the study. We are here to tell you the things
- 25 we are doing.
i :
I ACE FEDERAL REPORTERS, _ . _ . , . . . _ _ _ INC.. , _ _ . . , .Washington,
_ _ ' . . _ . . - . - D.C.
_ . . . , _ , _(202) i
_ _ . _ .347_3700_
6 g- 1 We are focusing on the containment structural LJ 2 integrity, containment bypass and the interfacing systems 3 LOCA f or this particular portion.
4 Results we have identified, these are 5 preliminary, we have looked towards identification of 6 significant technical issues, evaluation of major risk 7 contributors in the event -- identification of areas for 8 further risk reduction.
9 I might like to make one point here. Since 10 this is a status report I would like consideration to maybe 11 revisit with the subcommittee at some later date where we 12 might have more definitive results to discuss with the 13 subcommittee prior to going to a full committee meeting in 14 the future. -
15 DR. MOELLER: Of those focus items, the first 16 two are readily clear, I think.
17 Could you expand a spec on the third one, 18 interface systems LOCA.
19 MR. NOONAN: Let me introduce Mr. Rossi and 20 Mr. Long. One of the two of them can answer that.
21 MR. ROSSI: Ernie Rossi. I will answer that.
22 What we are looking at in the case of 23 interf acing systems LOCA would be any breaks or losses of
/ 24 U) integrity in the system that might simultaneously cause 25 damage to the core, at the same time bypass containment.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
i 4 7
t 1 This would be any piping outside the 2 containment or lines outside the con'tainment that would do 3 that simultaneously.
4 DR. MOELLER: Thank you.
5 DR. MARK: Mr. Noonan, have you said, and I 6 missed it, what your schedule for completing your review is 7 at present?
8 MR. NOONAN: Dr. Mark, we have not really 9 established a firm schedule. What we are trying to do 10 right now is achieve some result toward the end of October, 11 mid November timef rame, to sit down with NRC management and 12 tell them where we are at. That 4s our goal at this point
( 13 it time.
l 14 DR. MARK: We would be thinking of a 15 subcommittee meeting, Mr. Chairman, the end of November, 16 maybe?
17 DR. MOELLER: Does that sound reasonable?
18 MR. NOONAN: That would be reasonable. '
19 DR. MOELLER: All right.
20 (Slide.)
21 MR. NOONAN: I would like to preacnt a bricf 22 background on the licensing status f or the Seabrook 23 Station.
() 24 In the saf ety review area we have one fire 25 protection issue that is incomplete. The information has i l
l ACE FEDERAL REPORTERS, INC. Wa shington, D.C. (202) 347-3700
8 1 been submitted to the staff and the staff has reviewed the 2 information.
3 I understand they are preparing their safety 4- evaluation.
5 On the emergency planning issues, the on-site j 6 work is complete.
l
- 7 Regarding the off-site work, as everybody 8 knows that is still in the hearings.
l
- 9 The New Hampshire plant exercise has been i
10 pe rf o rmed. The Massachusetts has yet to be done. There is 11 coordination work yet with FEMA and the regional assistance q 12 committee.
13 We will have final arguments at hearings next
! 14 week starting the 29th to talk about the emergency action 15 levels on the on-site work. And the off-site emergency 16 planning hearings have been postponed and not yet 17 rescheduled.
18 DR. MOELLER: On the off-site emergency 19 planning review, I think again I am familiar with the first 20 couple.
. 21 What about the coordination with FEMA? Has 22 FEMA been able to respond in a timely manner?
23 MR. NOONAN: Let me turn to Mr. Perotti from 24 the Office of Inspection and Enforcement. Don?
25 MR. PEROTTI: What is your full question, sir?
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
9 73 1 DR. MOELLER: We often read in the material C 2 provided to us that PEMA will respond by saying they are 3 quite busy or short on budget and can't keep up with what .
4 NRC proposes as a schedule.
5 Has that been true here?
6 MR. PEROTTI: Let me answer by giving a quick 7 rundown of the scheduling. Perhaps you can draw your 8 conclusion f rom that.
9 The New Hampshire plan was initially submitted 10 in December of last year with subsequent revisions 11 throughout 1986. The last revision of the New Hampshire I
12 state plan was submitted this month, September.
13 A couple weeks ago the FEMA schedule for 14 review and evaluation of that plan was something on the 15 order of October of this year.
16 They recently changed the schedule to 17 accommodate these more recent revisions to the New 18 Hampshire state plan and now anticipate a finding on the 19 New Hampshire plan by around April 1st of 1987.
20 DR. MOELLER: So they are pretty well behind?
3 21 MR. PEROTTI: They are behind, yes. And the 22 staff concludes that off-site planning, in particular, the 23 Massachusetts plan which has not been submitted, will
() 24 probably impact f ull power licensing f oresee brook.
25 DR. MOELLER: Could you comment on this ACE FEDERAL REPORTERS, INC. Washington, D. C.
(202) 347-3700
' 10 4
j 1 regional assistance committee? What is that?
!( 2 MR. PEROTTI: The regional assistance 3 committee is the committee made up of nine federal agencies 1
4 of which the NRC is a participant.
l 5 We have a NRC Region 1 representative on that 6 committee. The committee is chaired by FEMA Region 1.
7 Ar.d they assist and coordinate among the 8 federal agencies and assist the state, in this case, both 9 states, in deve: loping and upgrading the off-site emergency 10 y response plans.
! 11 Our NRC Region 1 frequently meets with these i
12 other member, committee members to discuss all of their 13 efforts.
] 14 DR. MOELLER: Thank you.
} 15 DR. MARK: How far is the nearest point in t
16 Massachusetts f rom the plant?
17 MR. PEROTTI: To the best of my recollection, 18 just looking at the emergency planning zones depicted in 19 that plan, it seems something on the order of six to seven 20 miles.
21 Perhaps somebody else can --
! 22 MR. DERRICKSON: Bill Derrickson.
l 23 Two miles 690 feet. *-
() 24 DR. MARK: Thank you.
25 DR. MOELLER: Roughly.
l
! ACE FEDERAL REPORTERS, INC. . Washington, D.C. (202) 347-3700
a 11 I
i
! 1 Go ahead, Mr. Noonan.
I j 2 MR. NOONAN: There are two safety issues being
! 3 heard at the board next week, one is on qualification and i
i 4 the other on control room design.
l 5 Both reviews have been completed by the staff.
6 Saf ety evaluations have been written and published to all 7 parties. ;
l 8 The staff's testimony has been written and j 9 submitted to the board and they will be heard next week.
10 There is one other item on here I did not put.
i 11 I apologize. The applicant has requested what we call a
- 12 5050 C 05 request and submitted it to staff. Staff has j
13 reviewed that request.
} 14 We have written our affidavits to the board in I
15 support of that request. That issue is now in front of the
] 16 hearing board.
) '
17 We have not yet had any response from the 18 hearing board on that particular issue.
19 Finally what I would like to do is present the i
j 20 latest SALP report given to the utility. As you can see I
21 this is probably one of the better SALP reports given for a i
- 22 utility, almost all 1st and a 2 in the emergency planning ,
23 area.
4
() 24 The utility has been very responsive in
- 25 submitting to the staff's questions. That is reflected in 1
1 ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
12 1 the SALP report.
3 u-)
2 With that I don't have any further 3 discussions. If there is any question, I would be glad to 4 answer it.
5 I would like to turn the meeting over to Mr.
6 William Derrickson from the utility.
1 7 DR. MOELLER: Questions for Mr. Noonan?
8 Carson?
9 DR. MARK: In the report the only item 2 is 10 the emergency preparedness which is said to be improving.
11 Ir. this relevant to the New Hampshire power 12 people or to the state authorities?
13 MR. NOONAN: That particular part of the 14 report was prepared by the office of I & E in the region.
15 Done, can you respond to that?
16 MR. PEROTTI: I am not quite sure how to 17 approach the question. Could you rephrase it or restate 18 it?
19 DR. MARK: The only thing where a question is 20 raised on that list is the emergency preparedness.
21 That a comment concerning the utility people 22 or concerning the local administrations?
23 MR. PEROTTI: That particular item I believe
() 24 refers to the on-site preparedness, because in December of 25 1985, a team of inspectors went to' the site to perform an ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
1%
- 1 appraisal and found in the middle of the first week that 2 training, equipment, procedures and the like had not really 3 been completed to a point to make an evaluation, and the 4 team cut short the appraisal and lef t the site.
5 The team went back then on two different i ~
6 occasions, and I participated in the follow-up appraisals.
! 7 I believe that rating reflects the fact that there were
{ 8 incomplete areas and they were just not prepared at that 9 point in December, 10 DR. MARK: Thank you.
11 DR. MOELLER: How would this rating compare 12 nationwide to other nuclear utilities?
! 13 MR. NOONAN: You mean on the overall?
14 DR. MOELLER: Yes.
4 15 MR. NOONAN: I would say about one of the best i
16 I have coen. It's clearly the best I have seen in my 17 directorate.
18 DR. MOELLER: Thank you. If there are no more d
i 19 questions, then, we will move on and call upon Mr.
1 J
20 Derrickson to begin the public service of New Hampshire 21 preacntation.
I
- 22 MR. DERRICKSON
- Good morning. Thank you for ;
! 4
- 23 taking your time for us. We appreciate your efforts. .
() 24 (Slide.)
25 MR. DERRICKSON: I will be fairly brief. Karl l l
1 ,
I j ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
14 1 Fleming and Jim Moody have the basic burden here to present 2 the technical aspects of .our saf ety assessment.
4 3 I will try to give you a project overview. I 4 was going to talk about the SALP rating but Vince has taken 2
5 care of that f or me.
6 We will tell you where we are with respect to j 7 low power testing and the overall situation in the area 8 with the emergency planning.
9 (Slide.)
10 MR. DERRICKSON: By way of status Seabrook
- 11 Unit 1 is 99.99999 percent complete. We have been waiting l 12 for a license since the summer.
i s 13 We wrote a letter to Harold dent ton in July l
) 14 stating the plant was complete and ready to receive a 1
15 license. It remains that way. Unit 2 is 24.1 percent l 16 complete and on hold, i
17 In the case of Unit I we are continuing to
{
1 18 maintain the state of the plant, security systems and so 19 forth as if we had a license in anticipation that any day 20 we will get a ruling f rom the ASLB on the 5057 C and 21 receive a O power license from the NRC.
7 J
22 DR. MOELLER: What does a O power license
- 23 permit you to do?
s
) 24 MR. DERRICKSON: Our request is to be able to 25 load f uel and do the hot testing with the coolant system at
\
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
15 1 operating temperature and pressure.
2 We have several tests to run, trom tests trom 3 the original hot function tests. This whole ettort trom 4 the day we receive the license to completion of the hot 5 f unctional tests will take about a month or six weeks.
6 So we believe we can gain some time. The 7 hearings f or five percent power are next week in ports 8 tiouth, New Hampshire.
9 And assuming we get f avorable results trom 10 those hearings sometime in the next few months, we would 11 hope to receive a five percent license and so we would 12 continue to move on on with the process, which brings us to O
k/ 13 why we are here today, because the next thing we need is to 14 resolve this emergency planning situation.
15 The saf ety assessment is a key part of that.
16 Hopef ully something can be worked out wherein we can 17 demonstrate by some means that we can adequately protect 18 the health and saf ety of the public and move torward and 19 run Seabrook Unit 1.
20 That is our game plan and that is why we are 21 here.
22 DR. MOELLER: Thank you.
23 ( Sli de . ) l I
() 24 MR. DERRICKSON: The next thing I was going to I 25 show were the SALP ratings. Since Mr. Noonan has shown ,
l l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
16 1 those, I guess I need not repeat them. ;
( (Slide.)
2 t l
3 Mk. DERRICKSON: In terms of the low power 4 testing we have no open items with NRR Region 1 tor the 5 5057 C requests as we understand it.
6 There are three items to be heard in the !
7 hearing next week. Theory equipment qualitication, time 8 durations, control room design review and emergency action 9 lev els.
10 We will have to await the hearing to get the 11 results.of those three contentions.
12 DR. CORRADINI: Low power testing is tive 13 percent?
14 MR. DERRICKSON: That's cor rect.
15 DR. CORRADINI: Okay.
16 DR. MARK: What's the signiticance of the time
'17 duration in your 11rst bullet there on . equipment 18 qualitication?
19 MR. DERRICKSON: It has to do with the 20 longevity of the equipment in an accident condition, 21 inctrumentation and co forth.
22 DR. MARK: How long it can stay at 300 degrees 23 and things like that?
() 24 MR. DERRICKSON: Yes.
25 DR. MARK: Thank you.
ACU FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
i 17 1 DR. MOELL ER: Control room design review, is i 2 that the emergency HVAC system. . !
3 MR. DERRICKSON: Saf ety system.
4 DR. MOELLER: All right, SPDS.
5 MR. DERRICKSON: We have responded to each of 6 the contentions and filed the data required. We don' t 7 believe the items are open any longer f rom our point of
'8 view.
9 That is what the hearing board is going to 10 hear.
11 DR. MOELLER: Thank you.
12 DR. REMICK: What was the substance of the 13 contention on the SPDS? What was the argument?
14 MR. imIDRAND: Dave Maidrand.
15 The control room design review was the 16 original issue. When the hearings were tirst held in '83 17 we had not completed the review so the board ruled the 18 issue was ripe f or litigation in the f uture.
19 Since the issue has been resolved between l 20 ourselves and the statt. The only question the board I
21 allowed in ior the rehearing next week is some issues on i
22 the SPDS, things the statt said we had to add to the stream 23 bef ore we could have it completely acceptable.
() 24 The schedule f or that was agreed upon, I think ;
25 it was the first of June. Merely intervenors claiming this l l
, l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 l
18 4
i 5 1 need be done now or it's not enough.
f 2 DR. REMICK: Bef ore f uel loading?
i I
3 MR. MAIDRAND: Yes.
4 ,
(Slide.)
i 5 MR. DERRICKSON: One last item I nave to 1
6 discuss is emergency planning _in general. Just a recap of i
- 7 what we talked about previously.
! 8 We did hold a graded exercise in New Hampshire ,
l 9 in February, excuse me, in March. We did get results that .
l 10 were, there were punch lists to be done.
i j 11 We sumbitted plans, they were submitted two i
12 weeks ago. We will schedule another graded exercise and l 13 there will be a hearing in New Hampshire. The process is 14 moving along.
]
J 15 The last bullet states our problem, that the
) 16 governor of Massachusetts announced last -Saturday that he .
I 17 will not submit emergency plans for the State ot i 18 Massachusetts to FEMA, thus declined to participate in the 4
l 19 process, which really torces us to seek an alternative. -
20 The safety assessment is a key ingredient ot
]
21 whatever alternatives we can arrive at and have NRC and I
22 FEMA agree to, and the phase to get this safety assessment l
23 reviewed by the ACRS is really step two in the process.
() 24 Step one was f or us to write it and submit it.
l 25 We really need to know the conclusion of the NRC as to our l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
19 1 results so that we can move f orward. We really can't move 2 forward until we know we have some level ot agreement.
3 DR. CORRADINI: ' So you are saying that thin 4 document, or the report which you are going to discuss 5 today, has bearing on the last bullet, but its original 6 intent was just as an emergency planning as a technical 7 backup f or your emergency planning work?
8 MR. DERRICKSON: No, the answer I think is 9 probably no. We started this effort some time ago because 10 we had been getting indications from the State of 11 Massachusetts that in fact what happened last Saturday 12 might happen.
13 DR. CORRADINI: You are hitting exactly what I 14 am curious about. Is this study a direct result of what 15 you expected in bullet 4, or is it just a coincidental l 16 connection? i 17 MR. DERRICKSON: I am not sure whether or 'not f la we would have done it anyway. But given the tact that we 19 have had some ditticulty in Massachusetts prompted us to do f 1
20 this in 1985.
21 It nothing cice, it cet tainly lent u benbu or 22 urgency to the situation.
23 DR. CORRADIN1: Thank you.
24 DR. MARK: Can you go ahead by just agreeing 25 with URC and FEMA that your plans are okay and ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
20
}
t
~
1 Massachusetts can sit up there and suck its thumb all it 1
O 2 wants?
. 3 MR. DERRICKSON: We are not sure how that will l
l 4 work out. There is a decision to be made in the Long 5 Island lighting case for Shoreham. We are watching closely
- 6 what happens there.
7 In fact our plans, the exercise in New
, 8 Hampshire will.come possibly af ter the decision in Long 9 Island.
{
1 10 So we are going to know before we finish with
! 11 New Hampshire whether or not that path is really viable.
- 12 He are trying to learn from everybody. We I
- 13 want to look at every possibic alternative and everything 14 we can do so that we increase our range of options in case I 15 one option doesn't work.
I 16 This saf ety study may broaden the rcnge of
}
i 17 options it in tact you agree as we have demonstrated that i
! 18 the risk f rom the plant is tar, tar lower than the original i
l 19 documents, WASH 1400 and NUREG showed.
i l 20 DR. MARK: It you had an emergency zone ot 2 1
- 21 miles 690 teet you would be in the clear?
4 j 22 MR. DERRICKSON: I think the answer to that i
l 23 question is yes.
( 24 I would like to turn the program over to Jim 25 Moody at this point who is our manager of risk assessment.
1 ACE FEDERAL REPORTERS, INC. Washington, D.C. ( 202) 347-3700
21 1 He will start the --
2 DR. MOELLER : We still have a question.
- 3 DR. REMICK
- I have a couple questions
- 4 relating to the status of Unit 1. They aren't directly 5 pertinent.
6 DR. MOELLER: Go ahead. !
^
7 DR. REMICK: I am interested in the status ot i
j 8 your licensing of personnel tor Unit 1 and, in particular, 9 do you recall what kind of success ratio you had on t
i 10 licensing people f rom the NRC?
l 11 HR. DERRICKSON: Joe, do you have the 1
i 12 specifics on that?
z
- 13 MR. GRILLO
- My name is Joe Grillo.
l 14 DR. MOELLER: Go to a microphone please.
I 4
15 !!R. GRILLO: Operations. department.
l j 16 I believe the number is 38 people licensed 1
17 which corresponds to a 98 percent first time pass ratio.
l 18 Everybody that we put up f or a license has got a technical lj 19 license.
i
! 20 DR. REMICK: How many shifts will you have?
1 i
l 21 MR. GRILLO: Six.
I 1
i 22 DR. RE!1ICK: Are you licensing your SRA's or (
l 23 O's as operators?
a
() 24 MR. GRILLO: Yes we started sending the-senior 25 people on shift through a good college training program 1
__ A}E y DE M REPORTERS, INC. Washington, D.C. (202) 347-3700
. . - .- - . . - . . . . - _ - _ - _ - - - . . . - . . _ = . - - _ _ . .
u 22 l
l l .
I with Memphis State University and they have most all j 1
2 received degrees. There is one person who doesn't have a 3 degree.
4 DR. REMICK: So it is not a requirement but 5 you encourage that, is that correct? ;
6 MR. GRILLO: We encourage the degree, yes.
7 DR'. REMICK: I think you would' agree, the t
8 question of whether they will be licensed, it the STA's 9 will be licensed as SRO's or if you are going to have a ;
10 separate STA?
i 11 MR. GRILLO: We will not have a separate STA, 12 will have STA credentials.
O 13 DR. REMICK: So there will be two on shitt.
I 14 MR. GRILLO: Absolutely.
l 15 DR. REMICK: What is your philosophy on l
16 degrees f or SRO's? Do you encourage it, or is it a )
17 requirement.
18 MR. GRILLO: It's not a requirement. We
]
l 19 certainly will back anybody who would like to get a degree. ;
, 20 We do provide college level training programs I
21 in engincering related cubjecto.
t l
l 22 DR. REMICK: Thank you. l I
23 DR. MOELLER: Go ahead.
( 24 DR. CORRADINI: Are all these degree programs 25 in the Memphis State program 7 ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
23 1 MR. GRILLO: Well, early on we had contracted 2 with Memphis State. We put enough people through it at the 3 time to get very down into our organization.
4 So we haven't had to keep going with the 5 course. Until the NRC had made a decision to allow us to 6 use our STA's, se haven' t started the program. We do -
7 intend to put it into action again, once again.
U I am not sure, that didn ' t answer it. I am 9 not sure we are going to go back with Memphis State, is the 10 answer to your question. It will be another college or 11 Memphis State.
12' DR. REMICK: In fact I think Memphis State is
, O k' 13 of ficially out of that business.
14 MR. GRILLO: Yes.
15 DR. REMICK: The organization is somewhere 16 else.
18 DR. REMICK: Yes.
19 DR. 110ELLER : On your health physics 20 personnel, do you encourage them to seek certification by 21 the American board of health physics?
22 MR. GRILLO: I can't answer that. I am in the 23 Operations Department.
24 DR. MOELLER: I just wanted to see if your
_s 25 philosophy went beyond the operators.
ACE FEDERAL REPORTERS, I!K:. Washington, D.C. (202) 347-3700
24 1 MR. GRILLO: I do know we Edo have a certified 2 health physici.st there.
3 DR. MOELLER: If you have at least one, you 4 are doing very well.
5 Thank you.
.6 I would encourage you to encourage that group i 7 to pursue certification, certainly f or your top people.
8 Go ahead.
9 f1R. DERRICKSON: I would like to turn the 10 podium over to Jim Moody, who will begin the technical 11 presentation.
12 MR. MOODY: Good corning.
13 DR. MOELLER: As you go through this, Mr.
14 Moody, it would help me and perhaps others if you would 15 clearly distinguish the diff erences between, f or example, 16 the risk management emergency planning study and the 17 cmergency planning sensitivity study and the old NUREG 0396 18 or whatever it is and so forth.
I 19 MR. MOODY: Yes, we will do that later on.
5 20 DR. POELLER: Thank you.
21 MR. F.00DY : I am Jim Moody, Manager at 22 Reliability and Saf ety Engineering. I i
23 Sharing with the prepared presentation is Karl 24 Pleming, Pickard, Lowe and Garrick. l f Karl has been involved 25 with Seabrook activities since April of 1982.
hCE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
25 h 1 We also have principal investigators f or the O 2 Seabrook Station, probabilistic saf ety assessment here.. We 3 have support f rcm Yankee Atomic and New Hampshire Yankee 4 operations.
t 5 DR. MOELLER: Excuse me. You say'Karl Fleming i
} 6 has been involved.
l
} 7 To what degree?
8 MR. FLEMING: I have been ' project manager of l 9 the PSA which was started in 1982, continued to '83. I-10 continued as project manager of the more recent study i
11 referred to, risk management emergency planning.
i 12 DR. _ MO ELL ER: Is that five percent of your I
13 time.
14 MR. FLEMING : I would say about 50 percent of
) 15 ny time over the last several years has been on Seabrook.
i 16 DR. MOELLER: Fine, thank you.
I i 17 MR. MOODY: What we will do is- give a briet 18 overview of the Seabrook PSA completed in 1985.
19 As you will see it's a comprehensive eff ort, I
a 20 over 20 man-years ot study by Pickard, Lowe and Garrick the I
i 21 subcontractors and about tivo of the utility cupport. We t
22 Will provide a little moreover view of the update 23 activities. l
! 'i
() 24 The primary tocus is going to be the risk l
25 management emergency planning study and sensitivity study.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
26 .
1 DR. MOELL ER: Are you at liberty to tell us
/
2 roughly what the PSA cost? You know, a ballpark?
7 3 MR. MOODY: About S3 million.
4 DR. MOELLER: Thank you.
4 i
5 MR. MOODY: That is outside costs.
4 6 The PSA update, we are not talk just about a 7 PSA update. The intent was to look at the risk 8 sensitivity, diff erent protective action strategies.- In 1
9 doing so we wanted to take a hard look at the risks at 10 Seabrook and resulted in not only an update but also a 11 re-evaluation of risk, antic action strategies.
12 The conclusions we came up with was that the .
O 13 risk was very low with no immediate protective action.
14 If you look at this low residual risk it's 15 spatially located very close to the reactor.
16 ( Sli de . )
] 17 MR. MOODY: As we get into the rest ot.the-1
- 18 presentation we will provide details on those conclusions.
19 Okay, the Seabrook Station probabilistic l
]
20 safety assessment, f ull-scope level 3 PSA, f ull scope. in 21 that there is tuli treatment of ev ents. I have more on 22 this.
23 Level 3 in that we had a plant-specific f 24 containmentanalysisandsiggconsequencemodel.
25 DR. CORRADINI: Did you consider external ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 3'47-3700
27 4
4 1 events?
l 2 IIR . MOODY: Yes, I will discuss them in a I
3 minute.
- 4 Yankee Atomic Electric Company provided the ,
5 utility project management review f or Public Service of New 6 Hampshi re.
7 Pickard, Lowe and Garrick was 'the main 8 consultant and we had a number of subcontractors. There 9 have been two reviews of the plant model 'and containment 10 model by Brookhaven National Laboratory and Lawrence
- 11 Livermoore.
12 DR. MOELLER : Were those at your request, the '
13 Brookhaven and Lawrence Livermoore?
, 14 MR. E00DY: I don't believe so. We submitted 15 the studies to :the NRC f or their information.
' 16 DR. CORRADINI: Will we get some indication of '
i l
~17 these conclusions from those two reviews here? ;
1 10 MR. MOODY: No.
19 DR. MOELLER: The NRC staff commissioned those 20 reviews?
- 21 MS. DOLITTLE
- Yes, we did.
1 i
22 DR. MOELLER : Would you give us a quick ;
J l
23- summary of the conclusions?
(). 24 MR. B ARAT: Richard Barat, Reliability and 25 , Risk Assessment Branch.
ACE FEDERAL REPORTERS, INC. Washington,. D.C.
.(202) _347-3700
28
~
1 Ubere the Seabrook PRA was submitted 1n '84 C sS. i 2 staff initiated two reviews. One, review of the f ront end !
3 of the PRA with contractor assistance f rom Livermoore and 4 review of the back end, containment perf ormance source term l 5 and consequence analysis with contractor assistance f or ,
6 Brookhaven.
7 The Brookhaven review was completed, and I 8 think I could characterize it as being a f avorable review.
9 That review will torm part of the basis for 10 the Brookhaven, Brookhaven's current work on. reviewing the 11 emergency planning study I think when Trevor Pratt speaks 12 later today he can probably give you more of the details ot 13 that review.
14 The Livermoore review of the tront end was 15 terminated in December of 1984. At that time the review 16 was far trom being complete.
17 DR. MOELL ER: Why was it terminated?
18 MR. BARAT: The Livermoore contractors 19 generated a large volume of questions concerning the 20 details of-the PRA and the plant.
21 At the time it would have required a great
~
22 deal of time on the part of the utility' to respond to those 23 questions.
f 24 The utility at that time did not have a 25 contractual arrangement which would have allowed us to have ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
I 29 1 access to their PRA consultants, and that was the reason 2 for terminating the review.
3 DR. MOELLER: Mike, does that --
4 DR. CORRADINI: So the bottom line is that 5 when we get to Trevor, he will talk about the back end and 6 the front end is still in limbo?
7 MR. B ARAT: That's right.
8 DR. REMICK: I have a trivial question on the 9 previous slide. I notice you referred to this as a 10 probabilistic saf ety assessment but Pickard, Lowe and 11 Garrick are PRA consultants.
12 I am wondering if there is something I am O 13 missing here. ,
14 MR. MOODY: We just PRA and PSA 15 interchangeably.
16 DR. REMICK: Interchangeably, okay.
17 ( Slide . )
18 MR. MOODY: Give you an idea of the scope, 19 comprehensiveness of the study. Af ter thorough evaluation, 20 initiating events, we ended up with 58 distinct initiating 21 . event categories. The next slide discusses these in 22 greater detail.
23 39 plant ~ damage states. The reason reason for
() 24 the large number is it's a level 3 PSA to keep track of 25 whether the melt is early, late, high or low pressure. We ACE FEDERAL REPORTERS, INC. Washington,-D.C. (202) 347-3700
30 8 1 keep track of containment systems such as containment spray O 2 and isolation. This takes care of the dependency question 3 and reduces uncertainty.
4 There are 14 release categories at the end of 5 the containment model, included WASH 1400 type release 6 categories as well as plant-specific.
7 There were 16 event trees in the plant model.
8 Full treatment of dependent events, common cause failures 4 9 were modelled at the system level. External events such as 10 airplane crashes, seismic. Internal hazards such as tires 11 and floods and explicit modeling ot tunctional dependencies 12 was done at the event tree. !
13 The containment model was plant-specific and 14 it was enhanced considerably since -- from~anything done 15 bef or e . I am not aware of any containment analysis equal I
- 16 to what was done in '82 and '83 of the Seabrook 17 containment.
18 Site model is again site specific. Used the 19 PLG CRAC code which allows open puff release.
20 DR. REMICK: Would you briefly explain what 21 multi-putf release treatment is?
22 / MR. MOODY: Keith Woodard.
23 MR. WOODARD: Keith Woodard.
24 When a release is taking place it can occur 25 over a longer period of time. For the longer duration i ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
... . ~. - . ., , . .. . . . _ _ - . . , - . . - - - . . , , . - . . . - - -
31 J
l 1 relea'ses it will break them up into separate pieces or puff 2 segments.
3 DR. REMICK: Is that just for computational 4 convenience?
5 IIR. FD0DARD: No, it more accurately assesses 4
6 the consequences 01 a release.
7' For example, decay is taken into account 8 between the various releases, changes in wind and other 9 Weather conditions are accounted f or.
10 MR. MOODY: Out of the 58 initiating events 38 l
11 were -
Initiating events, six LOCA incidents,14 l l
l 12 different transient, support system failures such as loss
( 13 of of f-sit'e power, loss of primary component cooling water, l
14 service' water, DC. power.
15 Actually these are two earthquakes, a large l
16 LOCA and a transient . type earthquake that is broken up into 17 eight discrete initiating events. i 18 There are nine fires, typically f ail support 19 system such as AC power, --
20 DR. MOELLER: Do any of the fires result from i
21 earthquake?
22 MR.. MOODY: No, sir.
23 DR. CORRADINI: Can I ask a question?
t
() 24 How am I reading this? This is number of 25- ev ents, in.what way? I don't understand.
ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347--3700
32 1 You looked at 38 possible causes and this is 2 how they are broken down 'into categories?
3 MRacMOODY: No. All we have here is six loss 4 of coolant, yes, different categories.
4 5 DR. CORRADINI: You have .38 events, or am I 6 wrong?
7 MR. MOODY: In the common cause area?
8 DR. CORRADINI: Yes.
9 MR. HOODY: Yes.
10 DR. CORRADINI: These are ways of getting to 11 what leads me to a severe-accident?
12 I don't think I understand the graph, r
13 MR. MOODY: Mr. Flemming.
14 MR. FLEMING : What that represents is the 15 number of diff erent initiating events that we quantified 16 using cur event tree models.
17 The initiating events up there, he's giving 18 you breakdown of the coverage of initiating events by major l
19 class.
J 20 DR. CORRADINI: I have you. I 21 11R. FLEMIDG : We are not into results yet. j 22 DR. CORRADINI: Thank you. l l
23 DR. RE!!ICK: Could you' give me an example or l 1
24 f two of truck crashes, what was the initiating event there.
25 MR. MOODY: A truck crash into the i l
l 1
ACE FEDERAL REFORTERS, INC. Washington, D.C.- .(202) 347-3700
33 1 transmission system resulteding in an unrecoverable loss of 2 off-site power, 3 DR. REMICK:- So off-site truck. crash?
4 MR. MOODY: On-site but near the switchyard.
5 DR. REMICK: Switchyard. Is that the only 6 truck crash?
7 MR. MOODY: Yes.
8 DR. REMICK: Not a sabotage tyre thing, 9 accidental?
10 MR. MOODY: Accidental, yes.
13 DR. MOELLER: Are there any railroads around?
12 NR. MOODY: Not on-site.
O. 13 MR. FLEMING : No.
14 DR. CORRADINI: -Are we going to get~into this 15 in more detail so maybe I shouldn't ask now?
16 MR. . MOODY: I am not going to get into 17 . initiating events in more detail.
18 DR. ',ORRADINI: How does one choose it? Bas 19 'this been done enoagh times that you know what bag to go 20 into to pick up the 58?
21 Do you'know what I am asking?
.22 MR. MOODY: Yes.
23 DR. CORRADINI: Is it simply a matter of
() 24 - judgment, you have done this enough times that you -know 25 which-ones to consider.
ACE FEDERAL REPORTERS, INC. Washington,-D.C.. (202) 347-3700
34 1 MR. FLEMING: The 58 initiating events is the 2 result. of a process that is documented in the PRA of 3 considering every conceivable.way in which a plant 4 transient could occur.
5 For example, by doing a f acility tree analysis 6 on the plant heat bound.
"7 Ke did tailure modes and eff ects analysis et 9
8 support systems and looked at virtually all the initiating 9 events considered in other PRA's. That led to a list or 10 several hundred potential initiating events whicn were 11 screened and grouped.
12 DR. CORRADINI: Okay.
I
\w MR. FLEMING : And this is the result.
14 DR. FOELLER: Go ahead.
15 MR. MOODY: Four floods, one an external riood 16 that failed service water. The others were internal that 17 atfected AC power.
18 We had airplane crashes into structures and we 19 had turbine and tornado missiles'into structures.
20 (Slide.) l l
21 MR. MOODY: Conclusions from the Seabrook PSA )
22 in 1983 with regard to early health risk, the NRC safety -
23 goal was met with large margins and we found~the l 24 interf acing LOCA was dominating results as it was in WASH 25 1400 and other PWR PRA's. Latent health-risks, NRC safety i
ACE FEDERAL REPORTERS, INC. Washington, D.C.- (202) 347-3700
35
-.,, 1 goals were met with very large margins.
\-)
2 Support system failures that were dominating 3 the results, support system f ailures were not only 4 dominating core melt f requency, but they were also f ailing 5 the containment heat removal system resulting in a long 6 delay in over-pressurization of containment.
7 Core melt was between two and three times 10
- 8 to the minus 4.
9 Primary containment was very strong. Early 10 f ailure was unlikely and was dominated though but 11 interfacing LOCA's.
12 DR. CORRADINI: If one were to get rid of
-' 13 interfacing LOCA's as dominant, what were the next 14 couple --
15 MR. MOODY: Earthquakes.
16 DR. REMICK: Would you repeat what you said 17 the core melt f requencies were?
18 MR. MCODY: Between two and three times ten to 19 the minus four.
20 DR. REMICK: From a public health risk 21 apparently risk is low.
22 How about f rom a corporate financial type 23 thing. The company f eels that is adequate?
( ) 24 MR. MOODY: I am not sure I can answer that' 25 question.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
36 1 DR. REMICK: Have the corporate people thought 2 about that?
3 MR. MOODY: Yes. We are committed to looking 4 at core melt f requency during the next year or so.
5 DR. CORRADINI: If I take away risk and put in 6 just probability of sequence, of the sequence happening 7 besides interfacing LOCA, what was the next set of 8 sequences,-still earthquake?
9 MR. fr0DY: Up here?
10 DR. CORRADINI: Yes.
11 MR. MOODY: Yes,' I thought that was the 12 question you asked.
1")
k_/ 13 DR. CORRADINI: You answered I thought in 14 terms of risk.
15 I am interested not in terms of risk but in 16 terms of probability of occurrence. Forget about the 17 consequence of it.
18 Is it still earthquakes af ter interf acing 19 LOCA? I wouldn't think so.
20 MR. MOODY: Station blackout.
21 DR. CORRADINI: Yes.
22 f1R. FOODY: You are referring down here, 23 aren't you?
24 DR. CORRADINI: Yes, noted interface and LOCA
(_/) l 25 dominates though. .I am putting the two together.
ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
37 1
1 MR. MOODY: The station blackouts result in 2 loss of containment heat removal and delayed latent 3 overpressurization --
4 DR. CORRADINI: I may be getting ahead of 5 myself. You can hold me back.
]
6 Did you find anything diff erent trom for 7 example the Zion probabilistic safety study in terms of 8 rank ordering of frequency of events?
9 MR. MOODY:- Karl?'
i 10 NR. FLEMING: I will get into that in detail 11 later. Looking at all the Westinghouse PRA's we have had 12 done, and we have done partial or complete on seven or
! 13 eight, I don't know the exact number, it's difficult to 14 find any similarity between any of the sets of results.
15 So there isn't auch similarity at all between 16 Seabrook results and sign results just as there wasn't a 17 whole lot of similarity between Indian Point 2 and Indian 18 Point 3.
1 4 19 MR. MOODY: I am going to' turn it over to Farl 20 in a minute, . anyway.
21 The only other point I . wanted to make was that 22 long-term overpressurizations f or Seabrook were measured in 4
23 days.rather than hours, again, due to the containment.
[j 24 Any other questions?
25 DR. REMICK: A question on the, why is ACE FEDERAL REPORTERS, INC. Washington, D. C. . (202) 347-3700
38 i
j 1 containment considered very strong? Was this intentional 7-)
V 2 design or is this af ter-the-fact analysis, realizing i 3 perhaps it's stronger than we realized?
4 I am not quite sure.
i
! 5 MR. MOODY: We real licensed it as a result of i
6 the Seabrook PSA. As f ar . as ,the intent of the designers, I l 7' guess I would have to ask --
. 8 MR. DERRICKSON: I think we have a slide on
, 9 that.
10 MR. TOLAND: Rich Toland, United Engineers.
11 DR. SIESS: Would you use. a nicrophode, 12 please?
13 MR. TOLAND: United Engineers =in Philadelphia, 14 Engineers and constructors of Seabrook Station.~ including 1
10 15 containment.
z_
16 The question was asked ~ot me yesterday and I
! 17 gave some thought to it. In ettect this containment was ,
J l 16 designed tollowing the basic design. rules applied to other 19 containments we have designed, really predicated within - the -
20' ASME pressure vessel code.
j j 21 The reason it turns out stronger than say 22 prior generations or others.we have-designed similar to 23 this par ticular one, like Indian Point, f or instance, is
( 24 it's principally located .within the description of a load i i 25 set itself that is design basis f or the containment.
~ ~
l ACE FEDERAL REPORTERS,. INC. Washington, D.C. (202) 347-3700
39 1 Design accident pressure is 52 psi gauge, 2 Zion, Indian Point, those are about 46, 47 psi gauge, large 3 dry containment PWR's.
4 Seismic was.25 G, ground acceleration, fixed 5 base configuration so it goes in as relatively very. strong 6 seismic response.
7 It results very directly in a higher 8 percentage of steel in the construction of the containment, 9 both together.
10 In comparison between Indian Point and 11 Seabrook you would tind on the average there is about 20 12 percent more principal steel in it. i O 13 That carries directly over f rom capacity ot l
j 14 Indian Point to see brook with the same basic types of j 15 increasing capacity. ,
16 DR. SIESS: I can't see ofthand why seismic l
l 17 loadings incroaces the amount.of WPPC. i l
18 l MR. TOLAND: It does not. It's largely l l 19 predicated by the Zion -- itself. So relative to 46 at l
20 Indian pointed you have about a 12, 13 percent increase.
l 21 Seismic steel however does f unction to carry load.
22 It does have a contributory ef f ect in that l
23 regard. When you look at that, yes, it will.
I 24 DR. SIESS: By seismic steel you mean what.
{
25 MR. TOL AND : Large diagonal bars.
ACE FEDERAL REPORTERS, .INC. Washington, D.C. (202) 347-3700
40 1 DR. SIESS: They don't go' up f ull height, do
- 2 they?
3 MR. TOLAND: Number 18 is spaced 11 inches on i
4 center measured across to 110 feet which is only a short i 5 distance below the spring line.
1 I
6 Then they are carried above the spring line to 7 60 degrees a pecks from the spring line.
8 DR. SIESS: That explains why.
9 MR. TOLAND: Liner'also is half inch in the 10 dome and three-eighths inch in the cylinders. The liner 1-11 thickness in the dome is larger,-greater thickness.
12 DR. SIESS: Can I find a sketch' somewhere that l 13 'shows the configuration of the seismic steel?
14 MR. TOL AND : Yes, I have it. I will be glad 15 to talk with you.
't j 16 DR. SIESS: Thank you.
- 17 DR. MOELLER
- In reading some- of the i
- 18 literature on Seabrook or perhaps it was another plant,-but J
19 I recall intervenors saying, well, there are voids or holes 20 in the concrete. What assurance do I have- that there are
, 21 flow voids in Seabrook?
22 MR. TOLAND: I think-you have to look at that 1 23 in two regards. The first is the construction itself. The !
l
- ( 24 outer surf ace is obviously exposed. The outer surface is l l
25 one on which you take Great Plains to assure that you do
]
i
! 1
^
1
- . ACE FEDERAL REPORTERS, INC. Washington, D.C. (202)_-_- 3 47 -3 7 0 0
41 f~ l not have voids.
G 2 The_ question as to whether or not you can have 3 a void behind the liner becomes somewhat more conjectural.
4 Someone said when they banged on the liner they heard avoid 5 behind it. What does that mean?
6 I don't know if it just means the concrete has 7 shrunk away f rom that liner and you have five mills gap 8 there.
9 Is that a consequence? Absolutely not.
10 Then you look at the behavior of the concrete 11 structure itself under these loads. What does a void do to 12 you? Very little.
A ' couple of . questions.
13 DR. SIESS: Can I 14 assume that in. the structural integrity tests the concrete 15 cracked.
16 MR. TOLAND: That's correct, cracked 17 partially. I would have to explain it to you.
18 DR. SIESS: What would be the average spacing 19 of the vertical cracks.
20 HR. TOL AND: They were spaced to my 21 recollection, I can check, was probably between 14 and 18 22 inches.
23 DR. SIESS: 14 and 18.
() 24 MR. TCL AND : I would say probably 15, 25 somewhere in there.
j ACE FEDERAL REPORTERS, INC. Washington, D.C.
(202{ 347-3700
42 l
I 1 DR. SIESS: Is that the kind of figure you use 2 to get variation in steel stress at ultimate,17.7 per 3 square inch. ,
~4 MR. TOLAND: .I would redirect that-to Mr.
1 5 westerly calls I did not calculate all of it myself; i 6 DR. SIESS: That is an awful lot of stressed 7 developed in nine inches.
8 IIR . PISLEY: Don Wesley.
9 We did the f ragility calculations f or the i
- 10 containment overpressurization.
11 'In answer to your question we looked at 12 imrerical formulas available in the'11terature and came up
'll 13 with vertical crack spacing of~between ten and 30 inches i
i 14 with no effect of the vertical I beams which support the
') 15 liner. .
l 16 These vertical I beams are pretty substantial.
- 17 They are spaced at 20 inches. We use a median valueoot 20
! 18 inches for our vertical spacing to come up with an average l 19 rebar strain.
I 20 DR. SIESS: But you have got a tremendous 21 ditference in strain between the crack location and the i
22 between-crack location. That is only ten-inchs..
23 MR. WESLEY: There is a tairly significant 24 ditterence --
l 25 DR. SIESS: Yes. You have got a strain )
ACE ' FEDERAL REPORTERS, INC; Washington, D.C. (202) 3'7-3700 4
43
- 1 difference from ten percent of the crack down to.I think 2 two percent of the intercrack locations, to get eight 3 per cent strain difference in ten inches is a gosh awf ul --
4 IIR. IISLEY: Those numbers don't gel with my 5 memory, Dr. Seiss. I think we used a strain at ultimate 1
6 strength of about 7 percent.
7 DR. SIESS: It says ten here. Averaged 4.7.
8 MR. WESL EY: Average of 4.7. So, yes, there 9 is a significant gradation in the strain.
10 DR. SIESS: I think the bond stress it would 11 take to develop that in ten-inchs is probably extreme. I 12 think it's a conservative assumption. I am not going to 13 argue with it.
14 DR. MO ELLER: Thank you. That is helpful.
15 MR. MOODY: Karl Fleming will talk about the 16 PSA methodology results.
17 MR. FLEMING : Jim Moody just touched on some 18 of the highlights of the methodology we utilized in the 19 Seabrook PRA and gave you some of the bottom line 20 conclusions reached in terms 01 the risk levels.
21 I would like to take the discussion down about
~
22 one layer et detail greater and talk a little bit more 23 chout the methodology so we can understand the results a
() 24 little bit better and give you some more of the numerical 25 results we obtained..
i r
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
44 1 One of the reasons f or going into the 7-V) 2 methodology question is that at about the time the Seabrook 3 PnA was being planned and initiated, there were some very 4 e:: tensive reviews being conducted by the staf f and 5 contractors on the PRA's that PLG perf ormed on Zion and 6 Indian Point.
7 Through those review there was a lot of 6 insights we got from directions to take in terms of 9 enhancements to methodology.
10 At the time we were planning the Seabrook PRA 1
11 it was an objective to make sure that when the study was 12 completed that the methodology wouldn't become obsolete
)
%/ 13 very rapidly. For the main purpose of using this as a risk 14 management tool, which they intend to do.
15 Also because of the fact that the NUREG 0396 16 document which is' a very important basis f or the current.
17 ten mile emergency planning zone did use information trom 18 the reactor saf ety study to characterize plant risk levels 19 so we can understand what our. current picture on Seabrook a 20 little bit better, it is worthwhile to examine what has 21 happened to methodology since the 'carly 1970 's.
22 ( Sli de . )
23 MR. FLEMING: I think some of the most O
() 24 important enhancements or differences between the 25 methodology used in the Seabrook PRA and that which had ACE FEDERAL REPORTERS,. INC. Washington, D.C. (202) 347-3700
45 l
1 been used in the reactor safety study I have tried to l
) '2 summarize on this slide.
3 4
i 3 As far as the plant model is concerned it's 4 probably the area that has the greatest difference. We 5 have adopted what we call a modular riseed event tree scenario model to characterize the accident sequences that
~
6 7 we analyzed.
8 As opposed to using more abbreviated event 9 trees that then dictate the use of large length fault trees 10 to construct scenarios.
~
11 That is one of the major differences that 12 needs to be understood to understand what, how to examine
,e~
13 the results.
14 We also, there has been a lot of-work in 15 trying to enhance the treatment of dependent events. One 16 of the big lessons or insights we got f rom the Zion-Indian 17 points reviews was that it seemed to be. pointing more in 18 the direction of complete treatment of dependent events.
t 19 This is also based on our review of operating
]
- 20 experience that shows _when we have incidents or accidents i 21 it's not the unf ortuitous concurrence of many -independent 4
22 events that tends to cause these, it tends to be dependent 23 events that link together many ditf erent systems and 24' structures and-so forth.
) 25 So we took our best shot at trying to enhance 4
_ ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
_ _ _ . ~ -
46 1 treatment of dependent events in Seabrook PRA. I will talk 2 core about that in a second.
3 Also, at PLG we have worked hard at 12 .
4 incorporating external events into the PRA's. With the 5 'Seabrook PRA we started moving in the direction of using a 6 single integrated plant model to analyze all the internal 7 and external events.
6 Prior to this for example seismic events had 9 been analyzed with a simplified plant model in comparison 10 to the models used for, say, LOCA's and transients.
l 11 What we have done at Seabrook is constructed a 12 single integrated model and all events in the study were 13 analyzed in that model. I will describe that in more 14 detail too.
15 'Because we had the opportunity to do 16 plant-specitic PRA's on -plants with operating experience, 17 we have developed a very good data base and are able to 18 quantity the variability in component tailure rates trom 19 plant to plant.
20 We used an enhanced data base for the 21 quantification of the Seabrook PRA.
22 In the contain. Analysis area we noted some, 23 there have been some d1111culties in previous PRA's and the
) 24 inappropriate enter mixing of two different types of 25 probability quantities.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
47
- 1 The one type, relative f requency type of 2 probability which we often associate with the plant model 3 had been mixed together in the containment analysis with 4 probabilities that represents the expert judgment of the 5 analyst on how phenomena would occur inside the containment 6 event trees.
7 So what we did in Seabrook is we sorted out 8 those two ditferent types probabilities and just to keep 9 ourselves straight on that, we even use two diff erent 10 words.
11 Ue use "t requency" to cone note the relative 12 or f requency connotation of probability.
(*
\- 13 then we use " probability" we are only talking 14 about the subjective interpretation of probability. That 15 has an impact on how you 'do the containr.ent of entry l 16 quantitication.
1 17 We also took a realistic look at containment 18 f ailure modes, as well as an assessment ot the containment 19 pressure capacity which we believe is much more realistic 20 than previous ~ containment analyses ot that type. We will 21 talk more about'that later.
22 As has already been mentioned, in the site 23 r.odel area we utilized an enhanced torm ot the consequence j ) 24 model originally developed in the reactor satety study.
25 Enhancements are very JDportant for studying i
ACE FEDERAL REPORTERS, IN C. Washington, D.C. (202) 347-3700
48 1 eDergency planning because the enhanced CRACIT model, as 2 Keith explained earlier, allows for treatment of time 3 dependent releases.
4 Ue have scenarios at Seabrook where releases 5 are spread out over long periods of time.
i -
6 The assumption that that all.comes out in one 7 puti, which would have to be made using the old CRAC code I
o for example, would get an unrealistic characterization of i 9 consequences.
10 Similarly there are other features of this 11' CRACIT model that enable us to model the approximate actual 12 path of evacuation as opposed to just assuming as in
- v 13 previous mold else that all people move radially outward to 14 some fixed distance and then across a particular path that 15 is kind of arbitrarily selected.
16 Ue actually moved, simulate movement ot people 17 f rom sector to sector through our model to approximate the 18 actual codes and speeds of the-evacuation plan.
19 We also in both containment and site analysis, 20 we have tried to quantity uncertanties. We do n ' t try t.o 21 hang our hat on one source term or set .ot assumptions tor
- 22. modelling our consequence code.
23 We come up with a discrete set 01 alternative 24 assumptions that seem to represent the range 01 possible 25 values.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
49
. s l' We actually then perf orm multiple analyses f or 2 different sets or assumptions.
3 In this way we attempt to quantity our 4 uncertanties. And that was done in both the contain. And 5 site analyses.
6 Probably the final thing that is, tends to 7 contuse people but once they understand it I think can get U Dore insights out of the-results, is that we package our 9 risk results and present results using a matrix procedure 10 that enables us to sort 01 slice through the risk pie in 11 many diff erent ways and get ditterent . perspectives on what 12 the contributors are.
f' 13 Finally we end up presenting individual 14 accident sequences that trace through the event trees. It
, 13 15 you want to we can actually convert the results to cut 16 sets, which is a f ormat some of the other tault tree 1
17 approaches have adopted.
18 So that knowing some of these enhancements I 19 think enhances the ability to tind out what the results are 20 telling us.
21 DR. MOELLER: Excuse me. You mention the CRAC 22 code.
23 MR. FLEMING: Yes.
24 DR. MOELLER:
( ) One of . the questions people have 25 questioned is the ability to evacuate people on the beach I
! ACE FEDERAL ~ REPORTERS, INC. Washington, D.C. (202) 347-3700
50 1 guess east of the p3 ant.
7-$
V 2 How is that handled in your analysis?
3 MR. FLEMING: I will give you a general 4 response. If I don' t get all the details right, maybe 5 Keith could help me out here.
t 6 There is emergency planning studies done by a 7 separate consultant that are used in support of the regular 6 n energe'cy plan. In those studies they calculate evacuation 4
9 speeds f or various scenarios and we utilize that 10 information directly in our model.
11 We move people -- we assess warning time, 12 release times and times for the various puf ts of the 13 c. release.
14 We allow the people to move as dictated by the 15 separate trattic studies-through evacuation speeds. We 16 make some assumptions about people who may not be evacuated 17 in time.
i 16 In the nore recent studies we have done 19 sensitivity studies where we have basically taken all the j 20 evacuation out and done a wide range of cases.
21 So that is generally how we do it.
22 DR. MOELLER: Is the beach within a one mile i
23 distance, or is it.more than that or less? i i
24 ER. EOODY: Two miles.
25 MR. FLEMI!E : As you draw the two mile radius !
1
. ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
i 51 l
l
. - 1 there are small portions of the beach that get picked up l
2 within the two mile radius.
- 3 DR. MOELLER
- Within two miles.
l 4 MR. MOODY: Most of the beach areas as it goes 5 up and down the coast would actually be outside two miles.
6 There is one little piece that is in it.
7 DR. MOELL ER: That is helpf ul. Thank you.
f 8 (Slide.)
9 MR. FLEMING : I mentioned the treatment of 10 dependent e. vents. I want to go'into a little more det' ail 1 11 there.
12 We attacked these tron several ditterent ways, 13 and Mr. Moody touched on one in treatment or initiating 14 events.
15 We have basically not restricted ourselves to 16 looking at any particular kind of event ex'cepted tor, you 17 know, the only requirement is that we are trying to pick up
'16 any event we think could contribute to risk. !
19 We haven't made any assumptions to contine it 20 to say internal events, whatever that means.
21 One can have a general breakdown into three 22 broad categories. Internal events people talk about, which 23 include generally transient events and LOCA's.
()
l 24 We also have internal plant hazards which are 25 really inside the plant but many people otten call them j i
\
'
- 9 ' *
- 52
- 1 external events as well. Things like fires, floods, 2 turbine missiles, truck crashes and so forth.
3 Then there are events actually outside the plant, 4 seismic events, external tloods, air crash crash and so 5 forth. Those are the general areas.
6 As I mentioned we went through a very thorough 7 qualitative analysis prior to selecting the 58 events Jim 6 told you about.
9 As far as the modelling of tunctional 10 dependencies these are dependencies that stem trom the way 11 the plant is designed. It's the design basis and 12 interf aces between systems.
O 13 Ue take an approach of explicitly modelling 14 these in various ways.
15 The first way we utilize is the dependency 16 matrix which is a concept that we borrowed from the I rep 17 methodology but we have expanded on that in' terms of the 18 level ot -detail we go into in describing inter-system 19 dependencies.
20 That type matrix or diagram. is usef ul _ to.
21 interact with the plant operators and design people to make 22 sure we bave made an accurate assumption about how the 23 plant is actually built.
( 24 We then model many of these tunctional 25 dependencies directly in our event trees because we have ACE FEDERAL REFORTERS, INC. Washington, D.C. (202). 347-3700
53 1 adopted a technique of using the event tree technique to 2 take care of the tact that service water train A f ails, you 3 are going to lose primary component cooling water and so 4 forth rather than depend on a fault tree linking procedure 5 to pick those things up.
6 We adopted that approach because we have found 7 it easier to get the plant operators to review these and C understand them than to have them look at huge link tault 9 trees.
10 Howev er , it's very clear to us that one can
- 11 model these things with either event trees or fault trees.
14 12 If it is done properly, either approach should be adequate N- 13 to handle it.
14 Since this is a level 3 PRA we have to worry 15 about a containment and a site analysis. One thing that is 16 a little bit unique about our approach is that we have put 17 containment systems in the plant event trees.
18 He do that because the containment systems are l 19 dependent, inter-dependent on the same support systems, f or 20 example, ~ that the core cooling systems are.
21 So we found the previous practice of putting l
22 containment systems in the containment event tree' kind ot 23 very ditticult to really treat the dependencies that go
() 24 back and torth.
25 So we put the containment systems in the plant ACE FEDERAL REPORTERS, INC. Washington,-D.C. (202) 347-3700
l 54
, s I model, t )
y 2 Also by doing that we can rcviews the 3 uncertainty in the quantitication of the containment event 4 tree.
5 So we make the trade-ott of putting the 6 complexity in the plant model and having sof tware available 7 that will handle that problem and we then dedicate the 8 containtent event tree to look at phenomenological issues.
9 DR. CORRADINI: You are then going to have 10 some feedback back from the processes that occur after core 11 melt back to your containment sy stem s, right?
12 That is, I may have a physical event that l .
is' 13 occurs alter core melt in the containment that will disable 14 the containment sy st em. So you have to go backwards.
15 IIR . FLEMING: In general that has to be 16 considered. If you are are generating a new environmental 17 stress on systems assumed to be working you would have to 18 pick them up again as consequential ettects.
19 In general we find there are only very l
l 20 selected cases where that comes aDout in practice. But you '
21 arc right. That is a good point. :
l l
22 DR. CORRADINI: I guess what I am curious 23 about is your method seems more-reasonable than some others 24 that I have seen in that you put equipment reliability with f
25 equipment reliability.
ACE FEDERAL REFORTERS, INC. Washington, D.C. (202) 347-3700 Ihm
55 1 But I am curious, how do you do the feedback?
2 What if I had a hydrogen combustion event that would 3 somehow disable --
4 MR. EL EMING : If we had a hydrogen combustion f
5 event that was occurring when there may be some systems 6 operating inside containment we would have to put those 7 systems back in and re-ask the question of the 8 environmental damage to the system.
9 We would have asked the system reliability 10 questions previously and then we would put that system back 11 in and re-ask about the specific issue, did the hydrogen ,
s 12 burn cause fire to that system.
13 At the Seabrook Station that particular issue 14 didn't arise. By the time we get into hydrogen burns there 15 really are no active systems working in the containment we 16 take credit tor.
17 MR. TORRI: Fred Torri.
18 I might add to that a little bit. This issue 19 came up principally in connection with tailure of bank 20 coolerp in the large dry PWR's Zion Indian Point study.
21 Seabrook deals exclusively a separate and dedicated spray 22 system inside containment with no active hardware inside 23 the' containment.
() 24 Therefore, for that reason this issue was not 25 really a sign 111 cant issue in the Seabrook study.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
56 1 DR. KERR: In terms of containment, one of the 2 fission product control systems of course is the spray 3 system.
4 Now, if you read LER's at operating PWR's you 5 note many a time when they have been valved out and in 6 ract, utilities have been fined f or inoperable containment 7 spray systems.
8 Are they doing anything ct Seabrook to make 9 sure containment spray is available and how did you treat 10 it?
l 11 MR. FLEMING : We treat it as all systems, we 12 make a very caref ul search to find these events in the data 13 base. Then we assess whether these events are applicable 4
14 to Seabrook.
15 That in some cases will result in screening 4
16 out some events that don't apply because of their 17 procedures or because of interlocks or tech specs or ,
18 wha tev er .
19 Then we will quantify probability that things 20 may be inadvertently valved out due to human errors and
, 21 test and maintenance. We consider that for any systems as l
22 well as containment spray system.
23 I will get to another comment on that.in a
( 24 second, but in getting back to the dependent event i 25 treatment we also added a spatial interactions test, at the l
l ACE FEDERAL-REPORTERS, INC. Washington, D.C. (202) 347-3700
57 1
- 15 1 time Seabrook tests were being perf ormed there was an awf ul O -2 lot or ' discussion about systems interaction.
[
3 DR. MOELLER: Right.
4 MR. FLEMING : There were questions raised i i
5 whether working, doing PRA's trom -plant drawings and P and 6 I diagrams you might not be able to = tind. the, events and so 7 f orth we were aware of the special studies on systems l 8 interactions at Indian Point. j i 9 So we retlected on that--and decided to add <
10 another task to check the completeness ot our treatment ot 1 11 these types _ events. We added a thing called spatial I
- 12 interactions task.
)
}
13 That task is a, standard part of our PRA l 1 ,
I 14 methodology f or all the PRA's we 'have -started since 15 Seabrook.
I 16 What that consists of is first. developing a f
17- location dependence into thEt model by cod 11ying where all -f 18 the equirment is that is important'to_ save in-the model in- ,
- 19 our plant and performing a plant walkdown to brain storm j 20 with a team of people including a-structural expert, t
21 external event expert and: plant modelling expert and 22 someone who knows the plant very well and identity what 23 might take place due to any kind .of . physical event like a 24 beismic 2 over 1 situation or tir'e or flood of a.particular 25 type in a particular ' location, and several -- going through {
l ACE FEDERAL REPORTERS, INC. Washington, D.C.- (202) 347-3700 'j
58
-s 1 the process f or Seabrook we spent about a week with the 2 team going through the plant location by location.
3 We came up with several hundred scenarios that 4 were, looked potentially interesting. .Then we screened 5 those in a sort of simplitied risk model with .very 6 conservative assumptions about f requency or occurrence and 7 Im pa ct.
8 In ettect we .were in that screening process, 9 we were doing little simplistic external events analysis at
'10 every location in the plant, I am talking on the order of 11 120 locations that were already pretty well identitied in 12 the iire bazards study that was done for appendix R.
13' Out of that process came about 15 or 20 14 scenarios that tended to be mostly tire and flood scenarios 15 because they seem to be the most interesting ones with some 16 missile problems as well.
17 Then more detailed external events type.
18 analyses-were done on these particular events.
, 19 What we added to this, by adding this task we 20 basically provided a more systematic way to identify what 4 21 detailed events we wanted to take a look at.
22 Prcviously this had been done pretty 23 subjectively by just looking at drawings and taking a more 24 limited plant walkdown and subjectively assessing what were
- 25 the most important things to look at.
4 ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
. - - - . . _ ~ - . _ . . . - - - _. _ _ _ _ - - _ _ - _ _ _ . - _ -
4
'59 i
t
'l .In this case we actually did a systeDatic look 2 at' all the locations' and we learned some things about the
~
I 3 plant that weren't really appreciated.
4 We tound areas in the HVAC system with fires I
5 ~ could be postulated to occur in certain rooms but the EVAC i 6 system will actually get' smoke propagated to ditterent 7' areas.
8 Further evaluation of those scenarios indicated
! 9 they weren't very high risk contributors.
1
- 10 DR. MO ELLER
- What does F T stand f or in the' 11 tirst line of the third bullet?
i 12 PR. FLEMING: Location dependent tault tree l
O 13 model.
i .
i 14 DR. EOELLER: Thank you.
]
< 15 MR. MOODY: We 'used the -- the final bullet, 16 system level common ;cause f ailure analysis, this is an area 17 with probably the greatest variability among PRA's, -even
)
l 18 those published in the last few years.
19 Some have just omitted this altogether. We do
~
20 a very systematic approach of looking at other kinds of 21 dependent events largely caused by thingo like design error
- 22 - and human errors that can actually ~def eat the redundancy at i
23 the system level in ways other than those already i
() 24 identified previously.
1
- 25 Talking about undersized valve operators, !
i o l
' ACE' FEDERAL RE pRTERS, INC. _ Washington, D.C.- 1202) 347-3700
60
~
1 leaving redundant valves in the wrong position, al1 kinds 2 of events like that.
3 We actually modeled those at the systems 4 analysis level f or all redundant active components, and we 5 have been evolving some criteria on when to apply these.
6 models.
7 I' have to cay I have ncycr ccen other PRA' c do i
8 as complete a job as I think we have done in this area.
9 One aspect of this is to review the data to I
] 10 identity instances where such events have occurred.
11' It's surprising that in our review ot-the data i
i 12 base we have found on the order of. several hundred of these 116 13 events, rost of which had no consequences ot any 14 significance because they didn't happen in conjunction with 15 an initiating event for example.
16 In some cases that -- they do have that 17 characteristic as well.
18 But we go through and screen these events 19 event by event, and identify which are applicable to i 20 Seabrook station or whatever design we are analyzing. We 1
21 come up with design specific data .f actors as one component 22 of this process. That is applied at the . systems analysis 23 -
level. l
() 24 PMen you look at the results of Seabrook at 25 .the system analysis level these events are v'e ry important. l l
ACE l'EDERAL REPORTERS, IN C. Washington, D.C. (202) 347-3700 l.
61
'l Nearly all the systems have common cause tailures as a very 2 important contributor to their unavailability. ;
3 DR. MARK: You mentioned in this connection 4 design errors.
5 MR. FLEMING : Yes.
6 DR. MARK: Give me an example where you go 7 about having a design error, that is, they t orgot to put 8 the containment on, or what?
9 MR. FLEMING : A good example, this has 10 happened in a couple diff erent places is that you might 11 have some plotter operated valves that have a pressure drop 12 across them and they don' t put a big enough valve operator r
I) k- 13 on to open it. That has happened a number of times.
]
14 Ele can do deterministic . investigations f or 15 Seabrook or any other plant to convince ourselves that 16 maybe that doesn' t apply in that case. But we also know it 17 haprens with some treguency. It has happened on a number 18 01 occasions.
19 It we can't convince ourselves that' they' have 20 not ruled that out <or done some special tests f or that, 21 then we will through our generic data base butt'a w
22 probability in there f or the possibility that there might
- 23 be a design error in say a motor operated valve. ,
( 24 DR. MARK: But it would seem to me you would
- 25 go and look at th* valve and decide that it was. properly ACE FEDERAL REFORTERS, INC. Washington, D.C. (202) 347-3700
62 1 sized, then you wouldn't have to do the arithmetic..
2 MR. FLEMING: We do that. That is the ideal 3 situation. We always try to do that and only apply the 4 data that is applicable.
5 But in many cases it's di111 cult to dismiss 6 the possibility that some every. these events might occur.
7 DR. CORRADIHl: Can we stick with that oxample 8 since I am still not clear how you do that. Let's stick 9 with that undersized valve operator.
10 I don't know how many of these you have in the 11 plant, must be at least hundreds.
1 12 You then have sorae base data you use to come s/ 13 up with your probability.
14 MR. ECODY: The same data --
15 DR. COFRADINI: Or do you go out and do some-16 sampling on the actual plant?
17 MR. FLEMING : In the case of the common cause 18 data we basically look at all the data available to us in l
19 the torm ot licensing event reports or in-house searches i
20 for plant-specific data if we happen to be doing a PRA on a 21 plant with operating experience, and we classified events 22 as to whether they are tailures or not.
23 Or those that are tailures, some small
()
r 24 fraction of those are common cause events. We identity 25 them by noting that there's two or more similar components ACE FEDERAL REPORTERS, INC.
.-- . Washington,,D.C.
(202) 347-3700 -,
i -
'63 1
1 that are failed at the same time.
2 In addition to ~ quantifying f requency: or tailure 3 rate f or tne component we, using the same. inf ormation, 4 quantify a f raction of those f ailures that are common
. 5 cause.
1 6 That traction ranges f rom 1 to 10 percent 7 depending upon the component.
6 DR. CORRADINI: You come up with those ' numbers 9 from --
)'
10 MR. MOODY: From data. The same data we used i
11 to quant 11y the f ailure rates.
- 12- Any time we have looked in the case of I dent
, 13 Ca11tornia redundant active components like pumps, valves, i
!- 14 breakers and so forth, any time we have looked at these 15 data we tind a certain f raction that have been common i
j 16 cause. ,
17 We take them on an' event by event basis and.
i 18 see which events really. apply to that particular design, 19 depending on the design features or special studies that-20 have been done or design features that have been put in to 21 eliminate such causes and so for the.
2 22 DR. CORRADINI: Is that what you - -
l 23 MR. FLEMING: The residue which is lef t goes 1
() 24 into a beta factor. Those get applied to every redundant l 25 active component in our plant model.
[ ' ACE FEDERAL REPORTERS, INC. Washington, D.C. (202)' 347-3700
64 1 DR. CORRADINI: So that is what you mean by 2 design specific beta. factors.
3 MR. FLEMING: 'Yes.
4 DR. CORRADINI: You move out those not design
~
5 specitic to the plant.
6 HR. FLEMING:- That are not~really applicable 7 on tne walkoown.
17 6 MR. CORRADINI: Did that result in any spatial 9 interf erences that warranted changes to the plant?
10 ER. EL EMING : I don' t believe so. The tires 11 that popted up out of the spatial interactions task were-
- 12 analyzed in detail, and collectively they made about a ten 13 percent contribution to core melt i.equency which I will 14 show you in a tew Dinutes.
15 We didn't really tind anything that would j 16 be -- would have indicated some problems with appendix' R or 17 anything like that. The plant was very well physically 18 separated.
19 The only areas where we got into, where you 20 could get into problems without crossing barriers would be
' I 21 places like the control room, cable spreading room, which i 22 10 pretty cosmon to all plants.
23 DE. EEEICK: This was not limited just to 24 tires, the walkdown, is that correct?
(A/ i 25 MR. FLEMING: It was very broad, very. general.
D l
. ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 l
.- - - . . = - . .. .- ..
65
- g- 1 That is really the process by which this odd ball truck
(_g) 2 crash thing came up.
3 It wasn't the actual team, as I-recall the way 4 it went it wasn't the team itself.
] 5 But while the team was walking away from'the 6 plant after one of the walkdowns one of the operations 7 statt said, gee, one thing I have always wondered about is 8 we have these S F 6 transmission lines coming into the i
s 9 plant. Maybe there is a possibility ot - a truck cash. So 10 we looked at that, 11 It didn't turn out to be 01 major importance 12 but it warranted quantification betore we cane to the l 13 conclusion.
14 DR. REMICK : ~But it didn' t' warrant a truck i
15 barrier around the switchyard?
I 16 MR. FLEMING : No.
- 17 Now, having---
18 (Slide.)
l 19' MR. FLEMIDG : We end up with very large event l !
j 20 trees. Rather than putting- all systems and tunctional 21 dependencies in one grand event tree and maybe have to 4
i 22 Wrestle with problems of 11tting it into the computer we 23 have adopted a modularized event tree modelling technique I
() 24 where 11 rat we group the initiating events on the lett two 25 tigure out which event tree structures we can analyze the ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347-3700
66 1 groups of initiating events by.
2 That grouping for Seabrook is indicated here 3 on the left-hand side. And the box indicates that all 4 these initiating events within a box were analyzed with the 5 same model, sorte set ot event trees.
I -
6 Then separate quantitications are made with 7 ditferent numbers to represent the unique singular impact i
6 of that initiator on the plant.
9 For example it you have a loss of oli-site 10 power, when we go into the auxiliary system event tree we 11 can't have any electric power coming into the i
12 saf ety-related busses trom the switchyard.
13 So the unavailability of that nonessential 14 power-is replaced by a value et 1 tor that particular case 15 as an example.
i 16 We 11rst run our scenarios through an event 17 tree ttat has only what we call the auxiliary or support 18 systems.
19 The way we define these systems is that the 20 f ront line systcms are the ones that actually provide your
,- 21 basic saf ety tunctions. Then all other systems that just 22 serve the f ront line systcms are called support or 23 auxiliary systems.
24 They include things 111ke electric power, l 25 component cooling water, service % 3ter and so f orth, i
l 1
ACE FEDI:RAL IIEFORTERS, INC. Washington, D.C. (202) 347-3700
67.
1 The event tree is quantified separately f or O 2 each diff erent initiating event whenever identified. That 3
~
it put into groups to represent the way in which the-4 initiating events act on the auxiliary systems.
5 Ue then have our results ascessed with 6 treguencies and what we call impact vectors which are 7 different states that you can terminate in the auxiliary 8 system model end.
9 Imract vectors simply reflect what kind of 10 downstrebn effects on the front line systems can be 11 produced by their dependencies on the auxiliary systems.
12 Then we go through two stages of event trees 13 that represent the f ront line systems.
14 The first stage is the short-term early 15 temporal response to the initiating events. i I
16 The second stage, thingc like questions about '
17 reactor coolant inventory control, reactor trip, turbine 1 18 trip, things like that.
19 The long-term tunctions are monitored in i
20 reparote (vent trees called long-term system response 21 trees.
22 In those event trees we handle the long-term 23 core cooling and containment integrity questions and 24 containment ' isolation Laths and so forth.
25 So f or a given accident sequence in a plant ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
i 68 t ,
i 1
I codel, we are not in the containment area ret, actually l1 l 2 works its way trom the initiating event through three i
3 stages of plant event trees.
4 Then we terminate the sequences in what we 5 call plant states. Plant states include successtui states ;
6 where core cooling and there is no degraded conditions l
4
, 7 existing. Then a variety of what we call plant damage-j 8 states.
i 9 In the way we did this analysis they were all a
- 10 core melt states with different specifications to perform i
11 the containment analysis. I will get into some of the l
j 12 details later.
13 The pressure inside reactor vessel at the time 14 01 nelt through to know it there is going to be dispersive 1 15 pelt through or not, whether there will be water in the 16 cavity or not, what are the conditions for rebed cooling, ,
l 17 whether they are established or not, some of wh.*.ch depend j l
! 10 on the response of the active plant systems.
19 lie airo track leak paths that may have opened l
20 up in the plant model, whether we have containment I
{ 21 isolation system tailures or an open steam generator tube 22 situation or a purge of the containment which is in l 1
23 progress and so forth.
j f 24 All those are handled in the plant damage l 25 states.
ACE FEEERAL REPORTERS INC. .
Washington, D.C. (202) 347-3700
69 l
1 So by using this technique we break up the 2 overall plant analysis problem into a set of individual 3 event trees wnich are quantitled separately and linked j 4 together using special sof tware that we have.
5 I would like to make a comment. Right now at 4
6 New Hampshire Yankee, the New Hampshire Yankee risk team 7 under Jim Moody's direction is going through the process 01 6 regenerating all the Seabrook results on their computer 9 which is an important step towards their using the tool on 10 a day to day basis and not being so dependent on the 11 consultant.
- 12 (Slide.)
k' 13 MR. FLEMIDG: Having characterized the 14 methodology in a-little bit better detail let's get into 15 some of the can he features of the risk contributors Jim 16 alluded to. This gives you a little more detail.
17 h'e find it convenient to group accident 18 sequences into three broad groups for purposes 0t getting i
19 insights as to what is driving the risk protile.
l
! 20 Group 1 are the scenarios that have a 1
21 potential ict earJy health ettects, involve early gross 22 rupture of containment or large containment bypass.
t 23 Group 2 are an intermediate range et j
() 24 consequences that has significant potential tor latent 25 health etiects and probably little or no potential f or i
i ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 ;
70 '
i 1 early health ef fects depending on witat kind of evacuation 2 strategy might be assumed.
i 3 These typically involved things like degraded j 4 containments, long-term overpressurization, or small 1
5 leakages f rom containment that don't create a big enough 6 source term for immediate lite threatening doses.
7 The third group, given a core melt, is the 6 type et group, the most favorable. That is-where 9 containment r(mains intact which is -- which requires 10 maintaining the conditions for debris bed cooling and also 11 keeping containment heat removal systems in operation as 12 one et the conditions necessary for that.
13 Looking in terms of the, of tnose three 14 groups, as Jim mentioned we had core melt trequency, mean
] 15 of the uncertainty distribution we quantitled, 2.3 times i
16 ten to the minus 4 per reactor year.
17 That was roughly distributed in terms of 26 10 percent in the intact category, 73 percent in the degracied 19 or group two latent health category and one percent in the
{
20 early health offects category.
21 These ratios are typical. These gross numbers i
22 are kind of typical in what we see in PWR's in large dry 23 containments.
l 24 DR. CORRADINI: CouJd I ask a detailed 25 question now?
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
71 l
1 MR. FLEMING: Please.
O.
2 DR. CORRADINI: When Dr. Seiss had asked 3 bef ore about containment, the containment strength f or 4 Seabrook, the ultimate capacity used in the calculations is i
5 what?
6 MR. EOODY: I will get into that in a second.
I 7 Ke came up with a tragility curve, ditterent tragility 8 curve f or dry sequences. that. had higher temperatures than i
9 Wet sequences that had lower temperatures.
i 10 And the medium pressure capacities tor the dry 1 11 sequences were in the vicinity or 190 psi, which means at 12 190 psi we assumed a 50/50 chance or containment f ailure i O 13 versus attack.
i l 14 For the wet sequences where lower temperatures 15 were calculated inside containment they were in the
_;19
, 16 vicinity ot 230 absolute psi for the wet sequences. That 3
i i 17 was the median. We had a curve which I will show you later I 18 today.
l q 19 DR. REMICK: Just an observation on that ;
20 slide. I will suggest on the next to last column it you a
21 called that core melt group trequency, by looking at the
! 22 slide and seeing it was risk and also it had containment l l
23 response and such words as early containment tailure, my
() 24 iirst reaction was when you said group trequency this was .
1 25 1requency of the early containment ta11ure. j i l l
l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
72 1 Really, this does not include any probability a
O 2 or tailure et cont ainment ~. It's a core melt frequency, is
~
3 that correct?
4 MR. FLEMING : That's right. The absolute 5 frequency anC events per reactor year is 2.4 to the 10 6 minus six f or the early group,1.7 times minus 4 for this j 7 group and 6 times ten to the minus 5.
l S DR. REMICK:. It is misleading if you look at 9 the chart without an explanation.
i 10 DR. CORRADINI: So this is core melt?
4 11 MR. FLEMING: Percentage of the core melt i
- 12 frequency. Simply the ratio of these absolute f requencies 5
13 tc the total.
r 14 DR. REMICK: You say total release f requency, l
15 I am thinking release f rom containment.
i 16 MR. FLEMING : Okay.
t 17 DR. REMICK: That isn' t what you mean, I don' t 18 believe.
19 MR. TORRI: No. .That is a good point.
j 20 MR. FLEMING : Good point.
21 I wanted to point out the reason we had as i
i 22 much ac 73 percent of the core melt f requency in the second 23 group stem from the f act that support system tailures which 24 serve both the core cooling and containment systems were
- 25 involved in those scenarios.
b
+
ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347 -3700
73 g- 1 Or stated another way, the most likely
()) 2 scenarios we identified that had involved core melting 3 involved a support system problems, like in-service water, 4- component cooling water or electric power that, it
, 5 occurred, would degrade not only your core cooling systems 6 but also containment heat removal. That is why that number 7 tends to be high.
1 0 These results did not, while they took credit 9 f or operator actions up to the point of core melting and 10 possible actions to avert such a core melt, we did not i 11 consider in the Seabrook PRA actions that might take place 12 over long periods ot time in some instances between the 13 time of core melt and containment tallure, actions that 14 could possibly be taken to save the containment.
15 Ue did not conclude those in the original I
16 PRA's.
17 In the updated results we took a limited icok I
18 at containment recovery actions for one group of actual 19 sequences, but only a partial look.
20 DR. REMICK: In arriving at the friend cease 21 did you distinguish in any way between initiation of core 22 damage and core on the floor?
23 Or did you assume once you initiated damage to
/'T 24 the core there was no difference?
(/
? 25 MR. FLEMING: We didn't really distinguish.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
74 i
1 In a previous PRA which we had done on Midland we 2 entertained the possibility of having separate damage 3 states f or severe core damage short of melt and the core on 4 the floor type et thing.
5 In the analysis done there, because the time 6 interval between the on-site of the initiating event and
, 7 time 01 core uncover to the subsequent time as we 4
8 calculated using the tools available to getting the core 9 out of the vessel, the separate increment of time was 10 rather small.
11 It was telt that problem list particularly it 12 wouldn't have a big impact on the results.
i 13 That was a judgment we made at the time. I am 14 not sure whether it is supported by the most current tools 15 as f ar as how long it takes to melt the core- down.
16 But the assumption we made here was there was 17 no distinction, all core melts were assumed to go out into 18 the containment.
19 DR. MOELLER: In terms of early health eff ects 20 do you assume some threshold dose?
21 MR. FLEMING: Yes. There are some models 1 22 assured f or the probability of f atality as a f unction ot
- l. 23 dose received which does have a threshold.
24 DR. MOELLER: Okay.
25 ER. FL EMING : Keith, you can elaborate if you ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
75 1 want. There is a threshold assumed.
2 DR. MOELLER: Could you just elaborate a bit.
3 MR. KCODARD: Yes. We used the standard 4 assumptions that.are in.the CRACIT, HASH 1400. I believe 5 the threshold f or early f atality was somewhere around 240 6 rem.
7 Then there is a probabilistic curve up to 6 about 600 rem, as I recall.
9 DR. MOELLER: Now, in terms of, I guess, 20 10 latent ettects or in terms oi, maybe they don't even apply 11 NRC regulations, what I will driving' at, you know, the 12 current regulations pertaining to whole body dose limits or 13 guideline, design guidelines, versus thyroid are totally 14 out ot proportion if you use the recommendations of th e 15 National Commission on Radiological Protection for routine 16 exposures as reported in their publication 26 or reactor i
17 nuclear pcwer plant accidents as published in their j 18 publication 40. I l
l 19 Did that come into your calculations anywhere, l
l 20 uhere the thyroid may be controlling versus the whole body l
21 doEe7 I
22 MR. hD0DARD: Not in the early eff ects.
23 DR. MOELLER: Well, certainly not.
() 24 NR. h00DARD: Addressing latent --
25 DR. MOELLER: Yes.
ACE FEDI:RAL BEPORTERS, INC. Washington, D.C. (202) 347-3700
76 3
4 1 MR. WOODARD: Basically the latent doses are l
O 2 received over a long period based on relocation assumptions 3 and other factors in the model.
4 As f ar. as the thyroid goes, I believe we assume 5 that five percent of the thyroid cancers are tata11 ties.
6 It's either five or ten percent. That is the way it's, ;
! 7 that is the way the thyroid is treated in the latent 8 effects calculation.
9 DR. NOELLER: Okay.
10 f tR . ILEMING: The final point I wanted 1 to make 11 on the slide is talk a little about the contributors 12 alluded to earlier.
13 What we have in this column are really groups 14 of accident sequences grouped by initiating event. When we T
15 say teismic 24 Imrcent, there is a large number of i 16 individual accident sequences there, but' they all have a i
17 seinmic initiator to them.
s 18 SimilarJy, loss ot oil-site power, there are 19 many scenarios depending on exactly which system response i 20 we are talking about.
21 This was total percentage due to that whole i
22 class of sequences.
l 23 The single, as Jim Moody indicated, the most 7
24 1requent single event involving core melt was station 1
- 25 blackout involving loss of cit-site power, tailure 01 both l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
(
77 1 diesels and nonrecovery of either bef ore a lostulated pump-l 2 seal LOCA was assumed to result in a core melting 3
3 condition.
4 A large fraction that show up involve an 5 assumed pump seal LOCA. At.the time it was done the 6 current thinking at that point'in time, cooling at roughly 7 20 gallons per minute per pump or roughly ten to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> I _ :
8 at which time they would degrade entirely and tiow rates 9 would go up to a maximum of 300 gallons per minute per j
, 10 pump.
- 11 That was assumed in the-original PRA. W'e have 12 not incorporated the results of the new French tests that 13 possibly indicate the Seabrook model was conservative.
4 14 DR. CORRADINI: What do you mean by i
) 15 conservative in this case 7 i
f, 16 MR. FLEMING : We haven't really completed our
- 17 evaluation of, and interpretation of that data.
16 But there is some case f or an interpretation that 19 under certain conditions, the pump seal LOCA would not be l 20 permanent. Even -it .the plant were shut down and you l 21 pressurize as we assumed at, we assumed at Seabrook the f 22 pump neals continue to leak.
i 23 It the systems needed to restore electric
() 24 power are not recovered: beiore the seals actually went 25 tully degraded, we assume, like I say, that the leak would ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
i 78 1 be permanent.
]I 2 There is some indication trom the tests that 3 if one shuts down and pressurizes the leak will be reduced 4 to a very small level. So that is what I meant by
)
5 conservatism.
] 6 The other comment here is that interfacing i 7 system LOCA was analyzed using the same techniques,
)
- 6 assumitions and conservatisms that have accompanied that
- j. 9 event ever since its initial ident111 cation in the reactor i
10 sat ety study.
11 Namely, it has been assumed that once you 12 postulate rupture of the valves that normally separate the t
j 13 low pressure piping f rom the reactor coolant system that a l' -
14 pipe break would occur, all the coolant would be washed out i
15 that break.
16 One would end up with a core melt, containment 17 bypass with no hope of operator recovery.
ii 18 Ele will be talking in our presentation af ter j 19 the break on what we have been doing in our more recent 20 update study to provide more realistic treatment of that l
1
- 21 event.
1 i
l 22 DR. REMICK: Just a rersonal observation on 1 23 that core melt ircquency. I guess I am personally
( 24 surprised to tind it as high as it is.
- 25 It I understand the statement made by the l
1 ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
Y 79 4
l I applicant earlier, you -have a tollow-on program where you LC.
2 are going to be looking more in detail at this to see the l 3 Lossibilities of reducing that, is that correct?
4 ER. MOODY: Yes.
5 DR. REMICK: To me that represents a l 6 ' substantial corporate risk, apparently not a public. risk we 7 are hearing.
j 8 For example, I don't know it there is loss of a
9- ott-site power it one could minimize what the possibilities 10 are or minimizing that risk.
11 MR. FLEFING : While I don't disagree with you 12 I would clarify that these results are quoted as mean ;
l \~' 13 values.
1 14 DR. REEICK: Yes, I realize.
{
15 MR. FLEMING : The comparable results f rom a
j 16 reactor saf ety study mean values for PWR's is about 1 times i
17 ten to the minus 4.
l 1
10 Ey own personal opinion is that in view of the !
2 i 19 . treatment of dependent events and common cause events in- t i i
! 20 the study, I am wondering whether those results are 21 markedly ditterent at all.
i 22 DR. REMICK: No, I agree.
t i
23 MR. FLEMING : On~the other hand --
i
() 24 DR. REMICK: It's a question oi whether'that 25 is, 11 I was the CEO, it I would want to take that type of i
i
?
ACE FEDERAL .2FORTERS, INC. Washington, D.C. (202) 347-3700
80 s 1 risk long-term, b 2 MR. FLEMING: That's right.
3 DR. REMICK: That is why I say a personal 4 observation for what its worth.
5 MR. FLEMING: Yes.
6 Well, that concludes this segment.
7 DR. MOELL ER: We have a couple more questions.
8 Carson?
9 MR. FLEMING: Yes?
10 DR. MARK: I don't see station blackout on 11 that column there.
12 What is the ratio between loss of off-site b
L/ 13 power and station blackout?
14 MR. EL EMING : Okay, some fraction of these 15 loss of off-site power events, a large fraction would be 16 station blackout. But it wouldn't be a hundred percent. I 17 don't know the exact number. But it would be large. l 18 DR. MARK: I wonder why it isn't --
19 MR. FLEMING : There are other events involving 20 seismic initiators, fires and others that also involve a )
21 consequent station blackout.
22 So when we say station blackout --
23 DR. MARK: I was thinking of the failure to O)
(NJ 24 start of the diesels. That can't be a very large number, 4
25 surely. Isn't it maybe two percent or one pores?
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
81 1 MR. FLEMING : Failure of the diesels is a few 2 percent for failure to start. Then the failure rate is, 4
3 yes, it's a significant failure rate during operation.
4 If one looks at the failure rate to start and 5 failure to run for, say, six hours of operation, the total 6 failure probability of one of these diesel generator trains 7 is approaching 10 to the minus 1 or ten percent.
8 So we have two diesels. Even without common 9 cause treatment we have f ailure of 10 to the minus 2 given 10 loss of off-site power.
11 Then recovery actions are presented that 12 reduce frequency for overall station blackout. The overall i
/"T ll 13 station blackout f requency f or that top sequence is only in 14 the vicinity of 3 times 10 to the minus 5 for that one 15 sequence, loss of off-site, failure of both diesels, 16 failure to recover off and on-site power.
17 That is the single most f requent sequence we 18 identified that goes to core melt. ,
19 DR. MARK: That seems high.
20 MR. FLEMING: It's also important to note that 21 many of these seismic events and some of the fires and some 22 of the others involve a consequent station blackout 23 because, f or example, if you have a loss of off-site power
()
( 24 and service water pumps fail, then cooling is lost to the 25 diesels.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
82 1 DR. MARK: Yes.
~
2 MR. FLEMING: There are all kinds of scenarios 3 that get you to a station blackout. Only some are in the 4 loss of off-site power initiating category.
5 DR. MOELL ER: You mentioned two diesels.
6 Weren't two diesels installed for Unit 27 7 MR. FLEMING : Yes.
8 DR. MOELL ER: They are sitting there idle?
~
9 MR. FLEMING: That's correct. The analysis 10 that was done in 1982 and 1983, on Seabrook was a two unit 11 analysis. We assumed Unit 1 and 2 would bot'h be built and 12 run separately.
/~
(~s) 13 DR. MOELLER: Right. Would it help, or how 14 much effort would be required to make the unit 2's diesels 15 also available? -Would this do anything?
I 16 MR. FLEMING : I believe there is a possibility '
17 to back feed. l 18 MR. RAND: We have looked at that. Use of 19 unit two busses, separate skid mounted diesels and !
20 alternatives to other off-site power sources powered 4
21 from -- .
22 DR. MOELLER: All of that is being considered ,
23 or has been reviewed. Thank you.
() 24 Any other questions? Thank you.
25 We are at the time for a break. Let's take 15 ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
83 I minutes.
2 (Recess,)
3 DR. MOELLER: The meeting will resume and we J 2 4 will move in then to the review of the Seabrook Station PSA 1
5 update and the speaker then, we go back to James Moody.
6 MR. MAIDRAND: If I may, Mr. Derrickson had a 7 prior appointment but will be back right after lunch.
8 DR. MOELLER: Thank you. Let me say this has 9 been very inf ormative up to this point. We appreciate it.
10 DR. MARK: Do you expect it to fall apart now?
11 (Laughter.)
12 DR. MOELL ER: No, I think it will continue on 13 a good approach the rest of the day.
14 (Slide.)
15 MR. MOODY: Before I get started I would like 16 to correct a number of Fred Torri. Medium is 213 psi, not 17 230 --
18 DR. REMICK: Still large.
19 MR. MOODY: Yes. I am going to provide an 20 overview of PSA update activities. This is not just 21 updating the PSA, it's actually been the use of the PSA 22 that has created the update.
23 Last year an effort started with regard to
/ 24 enhancing the technical specifications.
25 One part of that effort included using the ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
84 1 PSA. We actually looked at the sensitivity of core melt-2 f requency from support systems when we varied the 3 conditions of operations and surveillance.
4- The most primary focus here of course is on 5 what we refer to as RMEP's, risk management emergency 6 planning study and sensitivity study' using WASH 1400.
7 We have also looked at seismic capacity. As 8 'you noted earlier, seismic was contributing to early health 9 effects.
1 .
! 10 What that entailed was looking at the most i
l 11 recent detailed qualification reports for some of the 12 important equipment. He found higher capacities than what-13 were determined originally.
14 The quantitative effects of . the seismic 15 capacity update is not included in the risk management and 16 emergency planning studies.
17 However, that is part of an ongoing 18 comprehensive effort to update the PSA next year,. updating i 19 and computer rising the study. And the study will be used 20 as input and dec~ision-making on plant changes and will 21 maintain. current.
4 22 Reliability and engineering group also has 23 responsibility for monitoring not just safety system
() 24 unavailabilities, but plant availability and reliability, ,
, 25 like plant trips, outages.
ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347-3700
85 l
f-1 (-Sli de . )
o 2 MR. MOODY: Obj ectives. RMEP's study, wanted 3 to reexamine the technical basis for the 10 mile EPZ, NUREG 4 0396 and part was WASH 1400.
5 We wanted to develop enhanced methodology f or, 6 PRA methodology for establishing a plant and site specific 7 EPZ.
8 This methodology included NUREG 0396 9 comparisons as well as comparisons to safety goals, WASH 10 1400 on early fatality curve. We looked at the spatial 11 distribution of risk around the reactor.
12 So we applied this methodology to the Seabrook 13 Station, we updated the PSA and then examined the risk 14 impact or risk benefits of the different protective action 15 strategies.
16 We included uncertanties and sensitivity 17 studies, namely on the source terms. And we-had.a peer 18 review group that was established to provide an independent 19 review of the results of the conclusions.
20 -(Slide.)
21 MR. MOODY: Principal investigators are Karl 22 Pleming, you have met. Fred Torri provided the expertise 23 in the containment and source term area with support from.
Bob Lutz from Westinghouse and Bob Henry.
-( ) 24 Ken Deremer was 25 involved and Keith Woodard.
ACE FEDERAL REPORTERS, INC. Washington,-D.C. (202) 347-3700
86 1 (Slide.)
0- 2 MR. MOODY: These results are going to be 3 shown later by Karl, more quantitative. We found with no 4 immediate protective actions early health risk was less 5 than WASH 14000, met safety goals with wide margins and 6 confined to an area very close to the site as stated 7 earlier.
8 From that there was small risk reduction by 9 any evacuation at all.
10 The NUREG 0396 criteria, dose versus distance, 11 was met at one mile or less.
12 DR. CORRADINI: Can I ask a question there?
13 Since I am not f amiliar with all the details of 0396.
14 MR. MOODY: Sure.
15 DR. CORRADINI: You are going to give a plot 16 somewhere in here.
17 MR. MOODY: Yes.
18 DR. CORRADINI: Of dose versus distance ,
19 showing that the criteria ~is what, that it has to be --
20 MR. FLEMING: The criteria is the values of 21 the NUREG 0396 curves at ten miles. ;
3 l
22 DR. CORRADINI: I see. You are showing those
)
23 values occur at this much closer in distance?
A
() 24 MR. FLEMING : That's right, or less. .
25 DR. CORRADINI: ,W ith the same assumptions as l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 l
87 1
1 0396 in terms of applying those in other words, using the 2 same set of assumptions that are used in that document?
3 MR. FLEMING: Except for-the assumptions about 4 the accident sequences and other things that are 5 plant-specific.
6 DR. MARK: Is Seabrook pretty much on the 7 waterfront?
8 MR. FLEMING: Yes.
9 DR. MARK: So that.there is a half space in 10 which you don't have any exposure except canoes, things 11 like that?
i 12 MR. FLEMING: That's right.
13 MR. MAIDRAND: No, actually the beach is to 14 .the east. Once you get past the peach you are. into the
~
15 ocean.
16 DR. MARK: A beach is between the plant and 17 water?
3 18 MR. FLEMING: The plant is roughly two miles 19 from the beach.
20 12. MAIDRAND: East from the plant is about a 21 two mile what I call tidal marsh. You come to the cause 22 way, bridge and beach area.
23 DR. CORRADINI: Do we have a picture of that.
() 24 MR. MOODY: Do you have something, Keith?
25 MR. WOODARD: Yes. It's not going to be real ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
88 1 good.
2 (Slide.)
3 MR. WOODARD: .It's not going to help real 1
4 well. This is the coast line right here. And the plant is 5 located about two miles inland from the coast line.
6 This is the beach here. This is marsh in 7 here.
8 DR. CORRADINI: The discharge canals are still 9 out there in the ocean, right?
10 MR. FLEMING: Another mile out.
11 MR. TORRI: There is a thin strip of beach 12 that goes up to the east of the plant about one and 1
v 13 three-quarters of a mile from the plant.
14 (Slide.)
15 MR. MOODY: We found these results and margins 16 came f roin two areas. First, containment effectiveness was 17 very strong. Enhanced V-sequence model Charles will talk 18 about in a few minutes.
19 In combination this provides f or very low 20 frequency of early containment failure.
21 And the source terms we looked at, best 22 estimate source term, conservative source term. Based on 23 these results we decided to do a sensitivity study. We
()
24 looked at the WASH 14000 methodology source terms and the 25 results for that.
ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347-3700
89 1 (Slide.)
2 MR. MOODY: They are summarized here. Karl 3 will have terms that show the results with comparisons.
4 We found using the WASH 14000 methodology 5 source term that the WASH 14000 early fatality risk curve 6 was approximately met with a one mile evacuation. The NRC 7 risk safety goal was met with no immediate protective 8 actions.
9 When we look at the conditional frequency of 10 exceeding whole body doses versus distance, we were meeting 11 the NUREG 0396 frequencies at ten miles, fcr Seabrook we 12' were meeting them at one mile.
A l'3 DR. CORRADINI: This is with the enhanced 14 study and those three factors.
15 MR. MOODY: Yes.
16 DR. CORRADINI: Maybe I shouldn't ask if you 17 get into it later, but I am curious about the breakdown of 18 the three things you mentioned in the previous flight, 19 containment effectiveness as a contributor and enhanced 20 V-sequence and source terms.
21 MR. MOODY: If you look at safety goal 22 calculations later on, which Karl will show, using WASH 23 14000 methodology source terms the safety goal increases
() 24 about two orders of magnitude.
25 If you look at the frequency of early ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) .347-3700
90 l
I i'
1 containment f ailure at Seabrook with the enhanced 2 'V-sequence model you see a couple orders of magnitude in 3 f requency of early containment f ailure.
. 4 MR. FLEMING : If I could clarify, these 5 results here f rom the sensitivity study only the 6 effectiveness of two of those factors, namely the enhanced
- 7 V-sequence model and the effectiveness of the containment.
8 DR. CORRADINI: Yes.
! 9 MR. FLEMING : In the previous results, RMEP I 10 study, all three factors are included.
i
{ 11 (Slide.)
j j 12 MR. MOODY: As I mentioned, we established a
! 13 peer review group, independently review the studies, 14 verified results. Rober t Budnitz, Chairman,- Future 15 Resources Associates; David Aldrich, Science Applications 16 Incorporated; Joseph Hendrie; Norman Rasmussen; Robert 17 Ritzman; William Stratton and Richard Wilson.
f l 18 DR. MO ELLER: ' Excuse me. On page' ll5.and-ll6, l 19 on 115 you point out that you set up a multi-stage t
l 20 technical review for the probabilistic safety assessment.
21 And you list Professor- Rasmussen as being a member of that 22 team.
23 MR. MOODY: Yes.
() 24 DR. MOELLER: .Then you have him reviewing the '
i 25 same report, if I am reading this.
I ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
91 4
1 MR. MOODY: No.
2 DR. MO ELLER: How can he review his own work?
3 MR. MOODY: 'There were two reports being 4 referred to. The first was the original Seabrook Station 5 probabilistic safety assessment completed in 1983. He was 6 part of the peer review on that.
7 Then on the risk management emergency planning I 8 study, sensitivity study, he was also a member of that.
9 DR. MOELLER: So it is two different studies.
10 Thank you.
11 MR. MOODY: Yes. The conclusions of the peer 12 review group are documented in appendix E of the study. I O
O 13 will summarize it briefly.
14 (Slide.)
15 MR. MOODY: They concurred with the ' principal 16 study findings. Namely, overall off-site risks are very 17 small, early health risk is lower-than thought to exist 18 when the general EPZ was established.
19 Confined areas very close to the site. They 20 felt conclusions were robust even in light of the
, '21 uncertanties.
22 In fact they believe the best estimate 23 probably overstates the actual ~ consequences.
1
() 24 They recognize the eff ectiveness of ,the' Seabrook 25 containment as being a major factor, as well as some h
ACE FEDERAL REPORTERS, INC. .
Washington, D.C.
'(202).347-3700
92 t
1 others.
2 (Slide.) i 3 DR. CORRADINI: Is it essentially the key i
1 4 factor?
]
5 MR. MOODY: Containment?
6 DR. CORRADINI: Yes.
4 l
. 7 MR. MOODY: I' believe it is, yes.
8 DR. CORRADINI: So if you throw all things
- 9 aside, the strong containment dominates the conclusions?
10 'MR.' MOODY: It's a combination of things.
i 11' DR. CORRADINI: Let me put it another way.-
12 If I have the strong- containment and I 'can v 13 assure a small probability of bypass, is that essentially 14 the whole picture right-there?
15 MR. MOODY: Yes, bypass includes containment 16 isolation.
17 DR. MARK: You have. probably said' this, and I 1 18 should know. What did you do about source terms?
19 MR. MOODY: Would you like to address that?
20 DR. . MARK: Are you using the TID 844 from so' i
21 many years B C, or what?
22 MR. MOODY: Are you talking about the 23 sensitivity study? In the WASH 14000 methodology i
( 24 sensitivity study we took source terms using the MARCH and 25 CORRAL codes for Seabrook.
4 a
ACE FEDERAL REPORTERS, INC. Washington, D.C. .(202) 347-3700
93 i
).
- 1 MR. FLEMING: .If'I might' clarify, we address
< (
2 those in the thick book, the --
$ 3 (Inaudible . )'-
4 We took our best shot at a -- using a -
I 5 combination of tools using a MAAP code developed in the i
- 6. IDCOR program and other calculations we had done in 7 support -- and conservative set of source terms arrived at -
l 8 indirectly from our. CORRAL results.
9 And also considering NUREG 0956.
)
10 DR. MARK: I was wondering if you had adjusted
, 11 to 09356 at least.
12 MR. FLEMING : We did a set of calculations 13 with 0956 source terms. The'n.we weighted those source i 14 terms probabilisticly to represent uncertanties. We put-1 15 most of the weight.on the best estimate and smaller weight
) 16 on the conservative source terms..
i 17 That was what we-did in the original: emergency 18 planning study, 0432.
19 Then in the sensitivity study we went back and
}
20 .lef t everything else the same but changed the source terms i 21 to what CORRAL would calculate,-the WASH 14000 methodology 1 22 for the Seabrook configurations and used those source terms-i l 23 to' demonstrate that the- containment eff ectiveness in the
~
24 reduced significance of the V-sequence in and of itself-l 25 seemed to lower the risk levels that justified shorting --
i ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
94 1 DR. MARK: You regard that as current besk O- 2 estimate source?
, 3 MR. FLEMING: No we regard the WASH 14000 4 source term methodology 'as a very conservative set of 5 source terms.
6 DR. MARK: The numbers we just saw for early 7 fatalities are --
8 MR. PLEMING: Conservative.
9 DR. MARK: Using that 1400 approach to the 10 source?
11 MR. FL EMING : That's correct.
12 DR. MARK: Since that is probably due for a 13 change and to some extent at least some items reduced, 14 these numbers might be reconsidered.
15 MR. FLEMING: That's right. In fact, we 16 regard the risk inf ormation in PLG L 432 as our very best 17 statement today of what the risk levels at Seabrook are and 18 what the risk benefits of evacuation truly are.
19 Sensitivity study was done in the spirit of the sensitivity 20 study.
21 This is what would get if you assume, type of 22 a study, as opposed to what we - really think the results -
23 will look like.
() 24 Our best statement of risk levels is in the 25 risk management emergency planning study.
I ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
)
95 4 1 DR. MARK: Thank you.
i 2 (Slide.)
3 MR. MOODY: I might add, the peer review-group 5
4 conclusions I presented there were based on the RMEP study, 5 not the sensitivity study.
6 DR. MOELLER: As Dr. Corradini has pointed 1
7 out, one of the basic factors in your reduced estimates-of.
8 population dose is the Seabrook containment.
9 MR. MOODY: Yes.
10 DR. MOELLER: Which members were experts on 11 contain; containment structures? Py presumption being if 12 this is a key factor you certainly want top people on the 13 review group.
14 MR. FLEMING: If I might comment, the 15 containment analysis was perf ormed as part of the original
! 16 PRA's.
17 In a risk management planning study we simply 18 utilize the same results without any modification or update 19 as far as containment pressure capacity and so forth.
? 20 By the time we had done the risk management-21 emergency planning study there had already been an NRC 22 staff sponsored review by Brookhaven'on the containment 23 analysis.
() 24 So when it came time to assemble the peer 1 25 review group f or the emergency planning study we only, felt
! ACE FEDERAL REPORTERS, INC. Washington, D.C.-
(202) 347-3700 --
4
- 96 1 we only had to address those areas that were changed in the
_O 2 PRA. ;
i i 3 That is why we did not choose a containment, I 4 structural expert.
i -5 DR. CORRADINI: Can I follow the question,.
4 6 your answer up, though, a little bit?
i 4 7 Are we going to hear-later today from i
8 Brookhaven their conclusions about structural analysis in 9 relation to their review of it? ;
10 I saw Charlie here from Brookhaven.
1 11 MR. NOONAN: Yes, we will this af ternoon have 12 a presentation, both Trevor Pratt and Charlie. i f
13 DR. MOELLER: All right, the answer. helps me a 4 <
l 14 lot.
. 15 The next point on the- peer review, .their 16 letter as you say. is in the appendix where Dr. Budnitz 17 summarized the group's findings. ;
18 But it does have qualifications. They say in i
19 terms of the time they have available, they even say they 20 weren't able'to_go-into it in great depth.. l
) 1 I
21 According'to what I read-they. spent two days, 22' . October 30 and 31st or something up at the site in a formal t
1 23 meeting.
() 24 MR. MOODY: 'Yes.
1 25 -DR. MOELLER:. They implied they had been
.i I
Washington, D.C. (202) 347-3700
_2 _ _ ?c= r=oza^'ia=>oa*=as
. . 1ac._. _ _
97 1 provided the report, I think it goes without saying, 2 provided ahead of time.
I 3 -
How much time beforehand did they have to look 4 at it?
5 MR. FLEMING : A month?
4 A month.
6 MR. MOODY:
7 DR. MOELLER: That helps me there. Then they.
8 f ormally met f or two days and argued and discussed.
i 9 MR. MOODY: Yes.
10 DR. MOELLER: Then a month or so later they 4
.11 issued their report.
12 MR. MOODY: Yes. Detailed presentations.were j
I
[~)
\s- 13 given to them and there was significant interaction, plant j 14 walkdown.
i 15 DR. MOELL ER: All right, i
16 MR. MOODY: Soon af ter - they .got the reports j
! l
, 17 they requested major portions of the original PSA including i d
18 structural and containment analysis.
i 19 DR. MOELLER: That helps me considerably. One q 20 thing I would have recommended, it says William Stratton, l 1
21 consultant, he:is at Los Alamos, he has been chairman of i U
22 .the ACRS. The gentleman has tremendous background. l
-i l
23 A little biographical sketch of each of;these i
. -( ) 24 people to me would have enhanced the report.
l 25 DR. MARK: Do you really need one on' Joseph j l
ACE FEDERAL REPORTERS, INC. - Washington,{D.C.
i (202)]47-3700 __J
98 l 1 Hendrie?
2 DR. MOELLER: Well, you and I don' t.
)
3 DR. MARK: 'Sure.
4 DR. MOELLER: Former Chairman of the NRC. '
5 There was -one member I didn't know that much 6 about. I knew all the-others.
7 But I would.give a one paragraph bio sketch on 8 each one.
9 MR. MOODY: Yes.
10 DR. CORRADINI: -I am still curious. The 1a 11 review committees can be different.
12 What was their specific charge? Was theirs T
U 13 not to look into detail, simply to look into overall 14 assumptions, approach and conclusions?
15 MR. MOODY: Yes. Assumptions -- th'ey were not i
16 asked to do detailed calculations.
17 However, their conclusions were based on1the 18 fact that there was quite a bit of margin there. They felt i
19 the sensitivity and results, the margins couldn' t- be i
20 affected that much with minor details of frequency and 21 things like that.
22 I will turn it over to Karl Fleming _again.
23 DR. MOELLER: 'Back on the charge of the peer
() 24 review group, 'it was clearly to let the chips fall where 25 lthey may, I mean do an independent analysis and tell us ACE ~PEDERAL REPORTERS,.INC. . Washington, D.C.- (202) 347-3700.
. _ _ _ _ _ , _ __ __ _ . . _ . . _ _ - , __ . ____._ --, . - ~ _ . ,
' 99 1 your conclusion?
2 MR. MOODY: Yes.
3 'DR. MOELLER: Okay.
4 MR. TORRI: To whatever depth they. felt-5 necessary to go.
i 6 DR. MOELLER: I think that is important.
7 Thank you.
l 8 MR. FLEMING : What I would like to do now is j 9 to go back over some of the methodology questions and talk
- 10 about what aspects of.the methodology were in fact' changed 11 in the more recent update to evaluate the emergency
. 6 12 planning options.
I 13 (Slide.)
14 MR. FLEMING: Sort of an overview, if you have 15 the slides you can read this. I don't expect you to read 16 the, this lineup here.-
17 As a level three risk model this has three' l
18 major elements, plant model, containment model and site 19 model, f 20 Inside each of the major boxes are some of the l
21 major elements of the models that have to be done just to l 1
22 have a level three model in place.
23 I kind every indicated with a cross hatched 1
h 24 code there the particular areas that were updated and 25 affected with the update that was done.
i ACE FEDERAL REPORTERS, INC. Washington, D.C. _( 202) _347-3700
100 1 I think that in an overall sense we basically V
2 utilized the same model used in the original PRA. There 3 were only some specific things that were actually changed.
4 Some were in the plant sequence model and the chief update 5 there was a complete relook at the interf acing system LOCA 6 scenarios.
7 I will describe a little in detail what we did 8 dif f erently there.
9 In the containment area as I mentioned, we 10 went with the same basic containment event tree, event tree 11 quantification, pressure, ultimate capacity analysis and so 12 forth.
kJ 13 But we took a fresh look at source terms. I 14 will tell you a little about what we did in source terms in 15 a few minutes.
16 As far as the site model is concerned, again 17 we used the same basic model, same basic population data 18 used in the original study with some very minor updates to 19 reflect anything we learned since the original PRA was 20 completed.
21 The major change here was that we assumed 22 different evacuation str~ategies. There really is a lot of 23 conservatisms built into this treatment in the following l ) 24 way.
ws 25 As I mentioned earlier we had some information gbbm ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
101 1 about the evacuation speeds for an assumed ten mile 2 evacuation. When we went in to model shortened evacuation 3 distances we did not go back-and take a look at how much ;
4 faster that evacuation might go for the inward part of the 5 population if the highways would be congested with less 6 evacuees.
7 So we didn't actually take credit for anything 8 like that. We used the same ~ speeds, but only confined our 9 evacuation to limited zones.
10 DR. MOELLER: Excuse me, though. That 11 statement, which I read in the report, assumes that when 12 you order an evacuation within the first mile that the l
13 people in the second, third, fourth, et cetera miles will 14 not move.
15 MR. FLEMING: That's right.
16 DR. MOELLER: You have no control over that.
17 MR. FLEMING: That's right.
18 DR. MOELLER: Unless you quietly tell the 19 people.
20 MR. FLEMING: Right. It's, it should be 21 viewed as a sensitivity to simulate what would happen in 22 the extreme case that the only, that only people up to a 23 certain distance evacuate and everybody beyond that had j ) 24 normal activities.
25 In actuality if you were planning for an i
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
l l 102 l rs 1 evacuation, you would expect to see movement of people on l (_) 2 an ad hoc basis.
3 DR. MOELLER: Right.
4 MR. FLEMING: We didn't try to model the ad 5 hoc effects. We simply fixed the distance over which 6 people were assumed to be evacuateing and varied the 7 assumptions of how f ar the distance would be.
8 DR. CORRADINI: ,Somewhere in your model you 9 feedback if I tell the one mile people to move and all ten 10 miles move, that is going to cause a great degradation in
- 11. your evacuation?
12 MR. FLEMING: That's right. That is why I say 13 we didn't-take credit for the speed up. In other words, if 14 there is a one mile evacuation that first mile moved at the 15 same speed as we had assumed to go in the ten mile 16 evacuation.
- 17 DR. MOELLER
- Oh,-okay. All right.
l
! 18 MR. FLEMING : That is the way we treated it.
19 The big difference in 'the site area was that 20 we ran a number of cases. We ran a very, very' large number j i
21 of CRACIT model cases to simulate different source term 22 assumptions, different sensitivities and so forth.
23 That was exercised many times for many
() 24 evacuation strategies.
25 That is basically an overview of.some of the l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 l
-103
- 1 general areas where we changed the methodology.
2 (Slide.)
3 MR. FLEMING : A little bit greater detail on 4 this, more detailed changes.
5 In the plant model area, we have enhanced 6 treatment of the interfacing LOCA. We did modify the 7 seismic analysis a little. bit. We didn't change anything 4
8 on the seismic fragilities, even th'ough'a fragility update
- 9 had been done.
10 But we did change some assumptions about the 11 behavior of air operated valves in the containment purge 2 12 system.
13 Back in the original PRA we.found that'the 14 next contributors below the interfacing system LOCA were 15 scenarios in which a seismic event occurred and there was 16 an assumed open purge condition taking place, the purge 17 line was open, which we calcuiated probability for its 18 existence.
19 Then due to seismic induced failure to plant 1
20 equipment that were modelled it was determined, there were 21 scenarios in which -- there was flow actuation signal 22 delivered to the containment isolation . valves and 23 containment purge lines were left open.
a
() 24 Some of the scenarios involved loss of 25 off-site power condition, super imposed on plant equipment l
ACE FEDERAL'_ REPORTERS, --.INC.
Washington, D.C.
. _ _ _ ., ~ . . . _-
(202) 347-3700
104 1 failure.
C<
2 In the original PRA we didn't take credit for 3 the fact that if you had a loss of off-site power the 4 instrument air system that holds these air operated valves 5 open is powered by nonessential power.
6 If you had a loss of off-site power that power 7 would stop, air compressors would stop and the way the 8 valves were designed the air would bleed out and valves 9 eventually shut.
10 We just assumed the valves would stay open.
11 Conversely if off-site is not loss compressors 12 would stay running unless the operators take action to shut L- 13 it off. The valves would stay open. So we corrected that 14 model for the enhanced seismic analysis.
15 We also, just f rom a standpoint of looking 16 towards risk management, what steps might be taken, we 17 started to consider the possibility of containment ;
18 recovery. I 19 So for loss of off-site power induced station 1
l 20 blackout events we looked at what might be done in the very l 21 long interval of time being-estimated between the timing of i 1
22 core melt and gradual overpressurization time of l l
23 containment which is several days. l l
() 24 We felt in that timef rame all sorts of things 25 could be considered, like bringing on mobile power l
l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
105 I 1 supplies, diesel generators or gas turbine units on the 2 back of trucks, using fire water pumps and so forth.
'l 3 So we' considered' containment recovery. We i
4 found for a given set of sequences there was a pretty high l- 5 likelihood that over the long period of time you could in l 6 fact recover the containment if not the core for those 7 scenarios. So we considered that.
1 8 However, that consideration has essentially no
! 9 impact on the emergency planning results that I will get I'
10 to.
i j 11 It was just done for the purposes of general
- 12 risk management, things to consider to actually manage the i
13 risk or lower the risk. ;
14 And in ~ support of some parallel activity we 15 had done last. year, to support some tech spec work on l ~
16 Seabrook, we used the model to do'a. number of sensitivity
~
17 studies on what would be the effect of varying allowed out l'
4 18 of service times and surveillance intervals and so forth. .;
}
l 19 In support of that effort we had a chance to
- l
! 20 go completely over our systems analysis work. There were a -l I' l j 21 couple areas where we decided to include some common cause l l
i 22 .models, you know, common cause possibilities that we had l I
q 23 judged earlier was not important.
24 .That led to a slight increase incore melt .
I 25 f requency, something like 2.3 times 10 to the minus 4 to I
I
- ACE. FEDERAL REPORTERS,.INC. Washington, D.C. (202) 347-3700
-+s 106
- 1 2.7 times 10 to the minus 4 af ter that enhanced treatment.
2 Again, since those sequences don't contribute 3 to early health risk it had very little significance to the 4 emergency planning results we will get on .the source term 5 area, I kind of summarized this in response to an earlier 6 question but basically back in the original PRA we had 7 already taken a shot in the 1982-83 timeframe of trying to 8 take advantage of the then existing IDCOR and NRC sponsored 9 research results that indicated source terms were 10 conservatively calculated using the WASH 14000 methodology.
11 We had actually quantified uncertanties in the 12 source term by having actually four discrete source term 13 assumptions for each release category and each of those 14 source terms was probabilistic1y weighted and the upper 15 bound source term used in the original PRA was in fact the 16 ones calculated by CORRAL.
17 Only a 2 percent probability weight was assigned 18 to our belief that those represented correct source terms.
19 And then some analyses were done to identify 20 lower source terms at three lower levels with different 21 probabilities to indicate where we thought the source terms 22 really were.
23 So we had a lot of source term information 8() 24 from the original PRA to use as a basis for the updated 25 study.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
107 1 We also had the benefit of the results that were 2 already generated in the IDCOR program for Zion. We 3 thought based on our having participated and performed 4 PRA's for both of the plants, we felt there was a good case 5 that could be made for using some of those nource terms 6 that had been calculated for Zion.
7 So we performed a detail, very detailed 8 comparison which Westinghouse helped us with to actually S look point by point at all the different features between 10 Seabrook and Zion that could have an impact on source 11 te rms.
12 As a result of that we decided to select some of 13 the source terms calculated f or Zion.
14 But there was one source term for the 15 interfacing system LOCA that because of different plant 16 configuration the Zion model wasn't appropriate for. This 17 is a new type where the bypass is created at an RHR pump :
1 18 seal.
19 In the case of Seabrook that would occur in 20 the -- a very tall flarow RHR pump vault which, if those 21 seals would fail, could very quickly become submerged and 22 flooded, whereas in the Zion plant those RHR pu p seals are 23 located in the auxiliary building and configuration is l 24 different in terms of flooding potential.
25 We decided to perform some Seabrook-specific
< l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 1
108 i
j~ 1 analyses using the MAAP technology developed in the IDCOR '
() 2 program and Westinghouse performed those calculations for
{
3 us.
4 So that got us on the road for coming up with i
And we reassessed the
^
~5 some best estimate source terms.
6 uncertanties originally assessed in the PRA, took into 7 account NUREG 0956.-
8 We took into account the issues that had been 9 discussed in the deliberations between'IDCOR and the NRC i
10 and various researchers on what the key. uncertanties were 11 for different physical processes that were being explored.
p 12 We ended up, then, with a set of conservative 13 and best estimate source terms. The best estimate,7you 14 know, the best way to view them is their IDCOR like source 15 terms for the best estimate and then some conservative 16 source terms for the other set.
17 Then both of those were run through=the 18 overall risk model and weighted to represent our 19 uncertanties.
l 20 As I mentioned earlier we also went back and 21 did numerous sensitivity studies which I will get into.
22 DR. MOELLER: Excuse me. This slide mentions-23 that you develop some Seabrook results with a MAAP code.
1
')
( 24 At a subcommittee meeting here on Wednesday, the staff, I 25 believe, indicated, the NRC staff, they do not, that they i
ACE FEDERAL REPORTERS, INC. __ W a .shington,
~ , . , . _ , _ _ D.C..
(202) 347-3700
1 i 109 l l
- i 1 do not curredtly have access to the MAAP code and it really
. .O
- 2 has not been evaluated by them. That could create some 1
. 3 problems or questions.
4 MR. FL EMING : Well, there are two,-there were 5 two kinds of sensitivity studies that we did tha't may 6 obviate the need to consider that. And that is that in the 7 RMEP study, we' did some sensitivities in there where we 8 basically assumed that all the weight goes on our 9 conservative source terms.
10 Those conservative source terms, they envelope.
11 the most conservative ones we assessed at that time and the 12 NUREG 0956 source terms.
IO 13 So we have a case in there where we have done i 14 all these calculations with~ simply 0956 source terms and 15 just forget temporarily about the IDCOR source terms.
~
16 So that information is available.
17 Then in the sensitivity -study report, the P L 18 G 0465, I believe that is the correct number, all those 19 results are generated just with the CORRAL source terms.
20 So whether there'is a need to rerun MAAP, I 21 don't know.
22 DR. CORRADINI: You have now named three 23 different ones. I get conf used yith just one.
() 24 0956, when you talk about that what exactly do 25 you mean?
I ACE FEDERAL REPORTERS, INC. Washington,'D.C. (202)~ 347-3700
110 1 Do you mean you ran the source term code 2 package or took 2105 results and 0956 results and used them 3 and assigned them to bins of sequences?
4 MR. FL EMING : The latter one.
5 DR.'CORRADINI: Okay.
6 MR. FLEMING: The first thing you have to do
[ 7 is try to match up the scenarios.
l 8 DR. CORRADINI: Right.
9 MR. FLEMING: From the two studies. When you 10 do that you find there are some see brooks in their that 11 simply no other study had addressed for example.
12 So we don't always have a comparable source 13 term for all the accident sequences.
14 DR. CORRADINI: 'Right.,
9 15 MR. FLEMING: We took our best shot at looking
^
> 16 at all the source terms we calculated and then running them 17 through the consequence model.
18 DR. CORRADINI: I j ust want to say it outloud .
l 19 to make sure I understand. l 20 For example in 0956, really not there, that 21 just describes the supposed technology. In the BMI 2104 i
22 there were 24 discrete: deterministic calculations done.
l 23 Those are what you used?
24 MR. FLEMING: Yes. \
25 DR. CORRADINI: You took the 24 and said all.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 3 4~ ,700
111
- 1 right, I have a myriad of them coming out of your 2 containment event tree and based on the coupling of the 3 physics that came out of your event tree and these 24, you 4 assigned --
i 5 MR. MOODY:- Sequences-6 DR. CORRADINI: To source terms?
7 MR. FLEMING: Yes.
8 MR. TORRI: W'e only used source terms 9 calculated for Zion and Surrey.
10, DR. CORRADINI:- From the BMI 2104. Out of the e
11 24 what does that reduce to? Less than ten?~
12 MR. TORRI: Yes, it's -- I would think it 13 would reduce to an actual set of source terms used out of 14 the source term code package calculations which amounted to 15 maybe three or four out of the seven source terms we had in 16 this study.
17 DR. CORRADINI: So you had essentially a-18 grouping of four to choose from.
19 MR. TORRI: Yes.
20 DR. CORRADINI: Some early, some late.
21 MR. TORRI: Yes.
22 DR. CORRADINI: Did you modify them or stick 23 with them straightaway.
() 24 MR. TORRI: We stuck with' those.
25 DR. CORRADINI: All right.
ACE FEDERAL' REPORTERS, INC. Washington, JD.C. (202) 347-3700
112 1 MR. TORRI: For the missing ones we used the O 2 best information we had from the original PRA source term 3 calculations which were based on march and CORRAL.
4 DR. CORRADINI: I guess that is not too bad 5 under the consideration that the containment is so' strong 6 all your early source terms that you had available to you 7 you never used.
8 Is that a fair way of saying.it?
9 My worry is the 0956 calculations don't have 10 Sirect heat in them for example, but because of the 11 strength of the Seabrook containment, you don't see an 12 intersection of a lack of knowledge in one area, and, 13 theref ore, a source term which isn't possibly 14 representative?
15 Do you see what I am asking?
16 MR. TORRI: That's correct. That's correct.
17 DR. CORRADINI: But you check for those like 18 that?
1 19 MR. TORRI: Yes.
20 DR. CORRADINI: Let me ask my question finally 21 in this way. , ;
22 MR. TORRI: Those source terms were 23 represented, early containment --
b)
(/
24 DR. CORRADINI: But never used.
25 MR. TORRI: Well they were used at whatever -
I ACE FEDERAL REPORTERS,-INC.
W'ashington, D.C. (202) 347-3700
113 l
1 f requency the study indicated they occurred. There still 2 is early containment failure in all-of these calculations.
3 DR. CORRADINI: Yes, but let me ask you, way 4 back here in one of the vugraphs I got the impression it 5 was so small as to .be negligible, early containment-6 failure.
t 7 Am I misreading that somehow?
l 3 8 MR. FLEMING: We had three release categories 9 in our current study that involve early containment f ailure 10 every one study or another.
11 They go by code words S 1, like the steam 12 explosion category; there is S 6, which involves a very j O- 13 large open purge in containment. '
i 14 DR. CORRADINI: Here is what is confusing me, 15 it's probably my confusion.
I i 16 In your summary table you put up previously l
17 where you had groups 1-3, nowhere did I see under early 18 containment f ailures the things you just mentioned. )
l 19 MR. MOODY: They are there.
l 20 DR. CORRADINI: But they are so small as to be )
I 21 zero.
22 MR. MOODY: The frequency is small. The 23 sequences are there. I am giving you more information on 24 that coming up.
I 25 DR. CORR 7.DINI: Fine.
i
} ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
114
- 1 (Slide.)
2 MR. MOODY: As far as looking at emergency 3 planning we reviewed carefully 0396. We looked at the 4 relationships between the PRA results that-were utilized 5~ there which were basically reactor safety study results, 6 and the criteria that were used to evaluate where to set 7 the EPZ and what we f elt was that having a plant-specific 8 and enhanced methodology PRA model for Seabrook station 9 that we had enough resources to actually revisit the 0396 10 technique of selecting the EPZ and actually enhance the 11 methodology itself.
12 That is what we are trying to do here.
The way we did that was th'at we went back and,
~
13 4
14 you know, re-did the 0396 type curves as one element of our 15 methodology. We basically adopted the 0396 technique.
16 Then we tried to extend it a little by looking 17 at other kinds of risk factors or ways to look at risk 18 results so you could see the relationship between risk 19 levels and the evacuation strategy you want to propose.
10 20 We did that first by base' lining the process, 21 by doing a set of calculations assuming no evacuation, 22 normal activities persist for the entire duration of the 23 accident sequence. ,
() 24 What that does basically is it sets an upper 25 bound on the total potential one has to avert risk through l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 f
115 i
l 1- evacuation.
i' 2 It's not the actual risk you avert because 1
3 evacuation isn't always a hundred percent effective. But
- '4 it is base like I the total potential that you have to
! 5 start with. So we did that first.
! 6 Then.before we got into different evacuation l ,7 strategies we looked at the no evacuation case, how is that .
8 risk, in other wo'rds, the accident frequency weighted into a;
l i 9 the number of health. effects that would occur, how'is that ,
i 10 risk distributed around the site. -
7
-11 We quantified a spatial distribution of risk h 12 around the site which is in the -muu? study. The conclusion 13 -was as Jim mentioned, the spatial distribution is very 14 close to the site. It didn't extend to a large distance.
15 That gives you an-idea of what you have to 1 .
j 16 work with in terms of the total gain you could possibly get
- 17 through evacuation.
l 18 DR. MOELLER: Did you-try.like,Dof course, you
- 19 didn't use the last two of your action strategies, or I a
{ 20 gather since you found the others gave you good results.
21 Did you try the one mile sheltering? 1
{
22 MR. FL EMING : More recently in response to l 23 some questions we have been getting from the staff in
() 24 support of their review, we have begun to look at a couple 25 cases where we do only shelters.
I ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
.- - . - - . . - , - - - _ - . ~ , - . - - . . - , . - - . . - - . - , . . , - . , _ . . , - . - - . . . - . - , ... .
116 1 We have' results out to two miles we are just 2 checking right now.
3 DR. MOELLER: Do the homes in that area have 4 basements?
5 MR. FLEMING: Some of them do.
6 DR. MARK: What fraction of the people you 7 have to think about are on the beach Saturday, July 8 afternoon?
9 MR. FLEMING: In our characterization of the 10 population distribution we modeled the probability that the 11 accident might occur in that fraction of time when there i l'2 might be a large population on the beach and actually treat 13 that effect.
14 Then through the Monte Carlo simulation 15 process the model goes through, selects at random different 16 start times.
17 In some cases we find scenarios where the 18 accident is occurring with the beach population, with due 19 regard to the probability of that occurring. That answers 20 it?
21 DR. MARK: Yes. But they don't have 22 basements, Dave.
23 MR. . FLEMING : That's right.
( 24 DR. MOELLER: Yes.
25 MR. FL EMING : Then what we did was in the ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
117 1 . study we looked at actually four different cases of 2 evacuation. Evacuation out to one, two and ten miles.
3 W'e published the results with a. combination of 4 evacuation and sheltering, evacuation out to two miles.and
{ 5 sheltering of everybody out to ten miles beyond that.
! i 6 Those re'sults are indicated 'in the report'and we will get 7 into them in greater detail.
} 8 The other thing we want to do, we recognize l.
- 9 any kind of risk calculations like this have large j
i 10 uncertanties.
}
11 We .can't really hang our hats on any 12 particular set of assumptions and data. We have to look at 13 a range of possible assumptions, each with some probability 14 of being correct.
4 15 So we did a very caref ul job by having 'two
~
- 16 sets of source term assumptions. , We also had two sets of-
- 17 consequence modeling assumptions.
[ 18 A lot of the data 'that have to be used in the ,
19 CRACIT consequence model, there are some input data-daat I
20 have uncertanties.
i
- 21 We actually;did two sets of-assumptions'in f 22 ' terms of for example how many health effects would be
- 23 produced by how many man rem of dose and so forth. That a
() 24 characterized the spectrum of opinion that exists in the 25 literature about those- types of factors.
i 1 -ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
118'
]
l 1 We also in our plant-model are very caref ul to 2 quantify the uncertainty in the estimation.of the accident 3 frequencies. !
4 On top of the uncertainty analysis, which is-in a way a kind of . sensitivity study, we did a more' 5
i
! 6 classical type of . sensitivity . study showing what the
~
! 7 effects of the results would be if you use extreme source j 8 term assumptions or use means versus mediums and so forth.
9 Those are published in the report.
! 10 Finally we tried to pull together all the i
I j 11 possible criteria we.could envision might be used to judge l l
} 12 acceptability of risk results< including safety goals and
- 13 other factors.
- 14 ( Slide .-)
i j 15 MR. FLEMING: So that was the methodology we
! 16- used in our best statement of risk this day;from the risk 17 l l management emergency planning study.
11 18 The sensitivity study, :like'.I said it's simply.
19 a sensitivity study, does not really reflect what'we think
! . i
! 20 is the best characterization of risk. 1 l 1 4
21 It's a censitivity ctudy cbtaincd by using 22 WASH 1400 source term methodology holding all other j 23 assumptions fixed at best estimate levels.-
24 We wanted to see the singular effect of source 25 terms and not get, complicate the picture.by varying many l 1
ACE FEDERAL REPORTERS, INC. Washington,' D.C.. (202) 347-3700 l
119
-1 parameters at the same time.
2 We simply fixed .everything, accident 3 frequencies at median values, assumption that the WASH 1400 4 source terms were the only ones to -use.
5 However, I want to point out they were 6 developed using the tools of the computer code, the CORRAL 7 code, but were done for the Seabrook-specific sequences and 8 configuration we defined in our PRA.
9 That included some source terms that were not, 10 you won't find in the reactor safety study because of the 11 way in which they bin their accident sequences, the way in 12 which we bin ours.
13 (Slide.)
14 MR. MOODY: We have already alluded to some of 15 the comparisons we make or criteria, if you want to call 16 them that.
17 One thing we have done is we have used the 18 values, you know, by select the EPZ at ten miles is- the 19 implicit statement that-the values at ten miles are 20 acceptable in some way of speaking.
21 So we need those valuee for the 1 5, 50 and 22 200 rem whole body gamma doses.
23 I will show what the results look like in a O
b 24 second.
25 We have also seen the WASH 1400 risk results ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700 _,
120 '
1 l' to be another statement of acceptability that .has been made O
- i. 2 f rom time to time about risk levels because of certain comparisons that have been made.
~
j 3 4 We made. that in two different ways. One, we .
5 looked at the risk curves in terms of mean values of the 6 uncertainty distributions versus median values.that had I 7 been published in the reactor safety study.
i 8 We even looked at different ways in' which you i
j 9 could come up with mean values from the reactor safety 10 study and made those kinds of comparisons.
11 Then we looked at the NRC individual and 12 societal risk safety goals. We' also can show you those i
13 results in terms of mean values or-median. values and what l.
14 to use has been changing as you know with time. -[
15 The best set of 'results we have f or safety ;
f 16 goals are presented in terms of mean values.
I 17 Another type of criteria is to see, what kind .
{
t .
j 18 of incremental payoff do you get by stretching the
- 19 evacuation out to progressive distances. That is sort of a j 20 diminishing return criteria we utilized..
j .
l 21 In summarizing the methodology, by looking_at i I
i 22 it from these ways you are getting all the benefits one has
) 23 in having a f ull plant-specific PRA to provide input 'into i
l () 24 this process.
25
~
You don't have limitations that.may be I l l ACE FEDERAL REPORTERS, INC. Washington,'D.C. (202) 347-3700 j
121
,3- 1 associated with just one set of criteria.
G (Slide.)
2 3 MR. FLEMING : Getting into some of the 4~ specific results which Jim had described in terms of words 5 earlier, these are results f or the early f atality risk 6 curves.
7 The solid line is just reproduced from the 8 reactor safety study f or early f atality risk.
9 We have a number of curves on there for 10 Seabrook station. The dashed lines come from our 11 sensitivity study that we generated assuming WASH 1400 12 source terms.
(~/3
'- 13 Theme one can see with no immediate protective 14 actions the calculated risk levels f all outside the WASH 15 1400 results.
16 For a one mile evacuation they cross. That is 17 what Jim meant when he said the WASH 1400 was approximately 18 met. It exceeds WASH 1400 in the high frequency low 19 consequence area then the high end tail comes inside that.
20 If you take this as a strict line that you 21 have to be inside for the whole envelope, you would say one 22 mile maybe isn't quite enough.
23 But two miles clearly gets one several orders r~%
! ) 24 of magnitude below this curve all the way across the range
~a 25 we calculate down to 10 to the minus 9 frequency.
ACE FEDERAL REPORTERS, INC. Washington, D. C. (202) 347-3700
122 fs 1 We do the same calculations taking~ advantage U 2 of the source term technology in addition to the can 3 takenment, we have results off scale. We have them 4 presented in tabular form in our report. But they can't 5 even appear on the scale 10 to the minus 9.
12 6 It's been generally assumed in PRA's that risk 7 estimates below ten to the minus 9 frequency are beyond the 8 ability of current tools to calculate so we simply truncate 9 our curves above that level.
10 DR. MOELLER: We have a question.
11 DR. CORRADINI: You said the three dashed 12 lines are using WASH 1400 source terms?
O s' 13 MR. FLEMING: Yes. The missing line that you 14 would see is with the source term enhancements we took 15 credit f or on the first study.
16 DR. CORRADINI: Which was 0956 based?
17 MR. FLEMING : Which included them as a 18 conservative -- we did two sets of source terms.
19 DR. CORRADINI: I understand that.
20 MR. FLEMING: Right.
21 DR. CORRADINI: I am' trying to get a feeling 22 for if I use the 0956 methodology developed source terms, 23 how you link them, do I show up here at all or are they
\ j) 24 down in the corner?
25 MR. FLEMING : I am not sure exactly where they ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
123 1 would appear. These results show the problem list 73 V
2 particular weights on the best estimate. There was a ten 3 percent weight assigned to those curves.
l 4 MR. TORRI: Twenty.
5 MR. FLEMING : Yes, 20 percent.
I 6 I don't know exactly where those results -- if 7 we went back and used 0956 type source terms these results 8 may start to appear on the bottom part of the curve. May 9 not appear.
10 . DR. MOELLER: Say that once more because I had 11 the same problem.
12 The dashed lines are the methodology using the 13 WASH 1400 source term.
14 MR. FLEMING: That's correct.
15 DR. MOELLER: What would have been off scale?
k 16 MR. ELEMING: Using RMEP source terms.
j 17 DR. MOELLER: Okay.
, 18 MR. FLEMING: Which include mix of best 19 estimate and conservative probabilistic --
1 J
20 DR. CORRADINI: Let's go a little further.
21 What you call RMEP's source term is a weighted value where i
22 some of it is from essentially the BMI 2104 calculations, 23 some from CORRAL calculations and some from MAAP
() 24 calculations?
25 MR. FLEMING: That's right.
ACE FEDERAL REPORTERS, INC. Washington, D.C. .(202) 347-3700
j 124-1 1 DR. CORRADINI: You weighted those based on 2 some subj ective judgment --
- 3. MR. FLEMING: That's right. Generally 80 -
. 4 percent on the source terms, 20 percent weight on the i 5 conservative source terms.
6 But also we showed you what you would get-if 1 7 you put one hundred percent weight on the-conservative i
! 8 source terms.
l 9 DR. CORRADINI: That is.not shown-here?
10 MR. FLEMING : No, it's shown in other-safety I
I 11 goal formats. t i 12 DR. MARK: If the source term.for which you 13 have drawn these curves, is it mor.e conservative than your 14 conservative?
j .
] 15 MR. FLEMING : Yes. Our conservative source l' 16 terms are considerably less, in most cases they are 1
l 17 considerably less than the conservative source. terms we :
1- -i j 18 used on RMEP's. I
{ . .
19 There is one exception which has to do with l
1
} 20 the way in which the RHR pump seal source term was' treated.
I i 21 In the, in thene reaulta down off scalc, the i
i f 22 conservative source terms were developed assuming dry 23 conditions, for some reason even though it's a pump seal
() 24 leak we come up with a conservative source term where we
} 25 didn't take effect, take credit for the scrubbing of the i
i I
' ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
125 1 water.
O 2 DR. MARK: Are these things we'can't see 3 because they are off the map assuming two mile evacuation 4
l 4 or one mile-or none?
i 5 MR. FLEMING: The ones that don't appear have 6 no evacuation. Without any evacuation, our best state of ,
i
~7 knowledge, there is no risk.
8 MR. MOODY: In the RMEP study conservative 9 source term is weighted at 10 percent, best' estimate at 90 10 percent, is that correct? -
j 11 MR. TORRI: That's correct.
I 12 DR. CORRADINI: I don't want to ask too many
} 13 questions here if we get more into detail later.
i
} 14 Is that going to be the format or is this our l 15 last crack at asking the questions that make up the curves?
! 16 MR. FLEMING: If you want to ask a question on
! 17 this kind'of curve, I think that this is a good time to ask 4
! 18 it.
I 19 DR. CORRADINI: My_' thinking goes something I l 20 like this. If early containment, early health effects is i
1 71 only going to be effected by early containment failure, not. '
j
} 22 affected by phenomena that is too low because of strength I i .
l j 23 of contain, then it's totally dominated by containment j 24 bypass.
1 1
1 25 Then are those dashed line curves taking into 1
[
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
2 126 1 account the scrubbing you would get from a submerged event 2 V or V dry --
3 MR. FLEMING : Yes . Well, there are two types 4 at diff erent f requency levels. There is an event S, with 5 an RHR pump seal that is submerged, because the ground rule
- 6 in the sensitivity study was to match as closely as 7 possible the WASH 1400 source term methodology.
8 hten we saw the WASH 1400 take credit f or.
9 scrubbing like BWR suppr,ession pool effects we decided in 10 that case to take credit-for that particular case.
13 11 We have another V sequence that assumes a pipe i
12 break. Those are all assumed to be dry.
13 DR. CORRADINI: The split between those is 14 essentially based on the structural analysis as to where 15 it's going to break?
16 MR. FLEMING: That's right.
I 17 DR. CORRADINI: What was that split? !
~
18 MR. FLEMING: I will show those results in a I 19 few minutes.
20 Roughly, the dry pipe break type came out 21 about an order of magnitude lower in accident frequency I
- 22 than the pump seal type. l 23 DR. CORRADINI: Okay. Then one last question. j 24 That is, for station blackout when I am at high pressure, l 25 one of the possible places that I can have a failure is not ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 j
127-1 at the pump seals, but the steam generator tubes, so I get 2 a complete by as of containment via that route. Was that 3 considered?
4 MR. FLEMING: It was, but not found to be'a i
2 5 credible pathway for that sequence.
6 DR. CORRADINI: Based on what calculation?
7 MR. MOODY: Based on the thermal' hydraulic 8 calculations we had performed and timing of the vessel melt 9 through that was calculated.
10 Do you want to elaborate a little bit?
11 MR. TORRI: I missed the question.
12 MR. FLEMING: He is asking whether we 13 considered heating up of steam generator tubes during a 14 station blackout event as containment bypass possibility.
15 My response was we considered that but not found to be a 16 credible release path.
17 MR. TORRI: That was considered in the 18 original PRA.
I 19 MR. FLEMING: Yes.
20 MR. TORRI: As a potential effective station i 21 blackout sequence.
I 22 From that time the consensus is that the tubes !
l 23 would not fall prior to vessel melt.
24 DR. CORRADINI: My reason for asking the 25 question is that if I have a loop seal, if my. loop seal is l l
1 l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
128
~_s 1 open and I am going to have enough flow through it to heat (v) 2 up the loop seals to fail or R C pump fail, then the second 3 place or another place where I have a weak link appears to 4 be in the steam generator tubes.
5 MR .' FLEMING : The R C pump fail doesn't fail 6 because of a heat up but because of a flow reversal of 7 ambient primary coolant temperature through the seal where 8 seal material can only withstand a few hundreds degrees F.
9 That is the assumption that was made.
10 The pump seal kOCA has already run its course 11 bef ore you uncover the core. It's not failed due to 12 thermal effects.
O
(~) 13 DR. CORRADINI: I see.
14 MR. FLEMING: It's failed because of support l 15 systems that normally keep the seals at 200 degrees F, 16 fail, then normal ambient primary coolant reverses flow and 17 exposes it,to 6, 700 degrees. That is what fails the 18 seals.
19 DR. COR7ADINI: You said something else, this 20 is my last question on .his point. You said that the time 21 between a heat up of the primary system at loss of off-sito 22 power at station blackout and time essentially f ailing the 23 vessel is short enough that you are not concerned with heat
/~x
/
24 up.
(s) 25 MR. FLEMING: Yes.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
129 1 DR. CORRADINI: Of'the steam generator tubes 2 and failing that way.
3 What is that time differential? I start off 4 on this path of off-site power and high pressure meltdown.
5 How long between then and vessel breach?
6 MR. TORRI: Specified in the starting 7 condition?
8 DR. CORRADINI: My. logic is_something like 9 this. I have a high pressure station blackout at the 4
10 fence, and one of the possible paths is bypass of 11 containment via steam generation tube rupture. That takes 12 time to get .the circulation pattern going and heating up I
L- 13 the tubes. ,
1 14 MR. TORRI: Right.
15 DR. CORRADINI: You said that time is not 16 available because I soon thereafter failed the vessel and 17 go all the way through and depressurize.
18 MR. TORRI: Right.
19 DR. CORRADINI: What did you calculate or 20 estimate that time to be.
21 MR. TORRI: From the beginning of core 22 uncovery to vessel melt?
23 DR. CORRADINI: Yes.
() 24 MR. TORRI: Less than one hour. This is based -
25 on the assessment at the time --
ACE PEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
130 v
4 1 DR. CORRADINI:' I think that is something you O
~
2 may want to look at only because I am trying to remember l 3 the calculations, estimates we have seen: that with :-natural i 4 circulation .you may. tend ' to delay heat up of the core to f 5 melt and, at the same time, heat up the whole primary 6 circuit, which means your weak. vulnerability is the steam j
I 7 generator . tubes and I bypass, you bypass containment via I
8 that thought.
9 MR. FLEMING: .If that turns out-to be through 10 there would be another effect, the additional recovery we 11 had not considered would have to 'be considered.
12 DR. CORRADINI: Right.
13 MR. FLEMING: Right.
! 14 DR. CORRADINI: If I follow your picture you j 15 are saying all the things you traditionally think about I 16 can't be here.
!14
- 17 The first thing that comes to my mind is tube
} 18 rupture because you are at high temperature heating the f 19 whole circuit. That tends to delay melt and heats the rest i 20 of the circuit up.
- 21 MR. FL EMING
- In a second I will chow what wc ~
! 22 have done using that same logic. How we have gone back and I 23 looked at the big issue of containment bypasses, what . seems l 24 to be the picture here.
I 25 DR. CORRADINI: Okay.
l 4
'(202) 347-3700 ACE FEDERAL REPORTERS, INC. Washington, D.C.
131 1 MR. MAIDRAND: I don't know how you are (s.J-)
2 getting steam generator up to the temperatures you are 3 concerned with degradation. Maybe implies understanding.
4 DR. CORRADINI: The calculations I have seen, 5 again, I don't know what to believe many times so I just, I 6 am looking at a range of things is that if I am in a 7 station blackout situation I have natural circulation in 8 the core which tends to keep the temperature of the core 9 low for longer periods of time.
10 To do that I distribute the heat through the rest 11 of the primary system which means I am heating up 12 components of the system n. ore uniformly.
[k/~)
13 MR. MAIDRAND: The safety valves are relieving 14 at certain pressures.
15 DR. CORRADINI: Saf ety valves aren' t going to 16 do anything f or you at this point. You are holding 17 pressure just by your PORV.
18 You are basically heating the whole system up 19 via decay heat. You are distributing the heat around the 20 whole system and the weak link could be the tubes.
21 DR. MOELLER: Well, go ahead.
22 MR. FLEMING : Okay, I want to go into some of 23 the other kinds of results now.
() 24 (Slide.)
25 MR. FLEMING: Another kind of result we looked ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
132 1 at was we looked at the total risk integrated out over the O 2 population which turns out to be like the area under the
! 3 risk curve that I showed in the previous slide. Get you 4 the total expected number of consequences per unit time.
5 And looked,at that, plot that as a function of evacuation 6 distance assumed. ,
7 This isn't to be confused with spatial 8 distribution of risk around the plant when I say no l 9 cvacuation. There are curves like that in there and one j
i 10 tends to conf use the two.
I 11 There is total aggregate risk integrated over 12 the whole population as a function of evacuation distance.
13 Here you can see that results fall-off, you 14 know, results come down .rather rapidly in the first mile to 15 two miles.
16 Then the additional risk reduction by further 4
17 evacuation tends to level off such that there is very, you 18 know, essentially no additional benefit in terms of total 1
- 19 aggregate risk reduction when we move the evacuation zone 20 from say two miles to ten miles.
21 Further, if we consider just sheltering the 22 population between two and ten miles, evacuate out to two 23 and shelter to ten we almost get as much. The total ;
G 24 potential for f'urther reduction we obtain f rom evacuation,
/j 25 we get almost that amount in a visual sense just by ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
133 1 sheltering.
2 DR. MARK: What does risk consist of ? Early 1
3 f atality or delayed?
l 4 MR. FLEMING : In this chart it is only the 5 early f atalities.
6 We also looked at latent cancer fatalities.
7 Results are presented in-the report. We found generally 8 that is not sensitive at all to the evacuation assumptions.
9 That comes f rom the way the latent cancer 10 fatalities are estimated, a large fraction of the 11 contribution comes from dose contributions that are 12 delivered af ter the event is over, af ter the evacuat' ion has 13 taken place and so forth.
- 14 For that reason you don't get a tremendous 15 change in the risk levels.
4 16 In fact we could hardly see any difference at 17 all on the latent evacuation risk. The differences were in 18 the early health risk area.
19 DR. REMICK: Were'they conducted 10 miles 20 within boundary, or 50 miles?
21 MR. FLEMING : We did a number assuming 50 all 22 the way up to ten and found we could still meet the safety 23 goal average over 10 miles with a f airly wide margin. Not
() 24 quite as big a margin as if we average over 50.
25 DR. REMICK: Am I correct in deducing there ACE FEDERAL REPORTERS, INC. W'ashington, D.C. (202) 347-3700
i 134
,, 1 must be about 10,000 people within a mile of the site
(_/
2 boundary? The way you have the safety goal plotted there 3 you are --
4 MR. FLEMING: Oh. It's about 2,000.
5 MR. MOODY: 40,000.
6 DR. REMICK: .If I look at the curve it looks 7 like, excuse me.
8 MR. FL EMING : This is 40,000 times 5 times 10 9 to the minus 7 I believe.
10 The safety goal for individual risk for_early 11 health risks is --
4 12 This is just a line indicating where we would 13 start, if we maneuvered this line it would indicate we 14 violated the safety goal.
1 15 ,
DR. REMICK: I was interpreting your plot to 16 be 5 times 10 to the minus 3. You are basically saying i
17 it's 2 times 10 to the minus 3.
18 MR. FLEMING: I may not have been accurately 19 plotted this.
20 (Slide.)
21 MR. MOODY This is just a cafety goal 22 calculation itself. This is average individual risk within l
23 one mile of the site boundary which at Seabrook is roughly, j
() 24 site boundary is roughly half a mile f rom containment so
- 25 average risk is about one and a half miles from the site l 1
l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
4 135
't 4
lO 1
2 boundary.
This shows the way in which the saf ety- goal-f j 3 was developed, background risk from accidental death
. 4 statistics, ' five fatalities per 10,000 population per year, 5 one tenth of one-percent becomes the safety goal.
6 In our.best statement of what the risk levels 7 are at Seabrook Station today with no'immediate protective
]
) 8 actions, no evacuation, no shelters, we-would meet the 9 safety goal with very large margins.
}
i 10 When we back.off the to WASH 1400' source term' 11 methodology with no, protective actions we come closer to 12 the saf ety goal but still meet it. We are, you know, .about
}
i
- 13' a factor of 1.5 away from it at that point. !
14 With a one mile evacuation we again achieve a l 15 pretty sizeable margin of a factor of around 10.
]
j 16 For two mile evacuation results are off scale.
- 17 Drop down in the ten to the minus ten range.
I 18 DR. REMICK: I have.a question. It's a little !
j 19 hard to state.
20 But you are making the proper comparison here l
21 on public risk for the commission's' safety goal.but you-
~
1 22 probably know the Commission at one time was cons 1dering a l
i
~
23 core melt frequency.
24 Then there was discussion about whether it' 25 should be core perf ormance guideline.
i I
) ACE FEDERAL REPORTERS, INC. Washington, D.C. . (202) 347-3700
136 7s 1 In the published safety goal the Commission V 2 said to the staff in my terminology, maybe you ought to 3 consider a surrogate safety goal which would look at no 4 major release, environmental release greater than ten to 5 the minus six per year.
6 Have you looked to see if you meet the 10 to 7 the minus 6 for major release?
8 MR. FL EMING : Yes. We were right on that j 9 boundary, it was in the vicinity of 2 times 10 to the minus i
10 6, but in the updated we are below that,10 to the minus 7, 1 1
11 four times 10 to the minus 7. 1 l
12 Frequency of large scale releases, which is !
/~T I km/
13 equivalent to the group one. category I showed earlier.
, 14 Now, question was brought up earlier, what is 15 more important, you know, the containment effectiveness or 16 source terms or what?
17 This result has source terms in it, updated 18 source terms. These were the old source terms.
19 What we are saying here, I don't know what the 20 curve would like look. I guess I don't know what the curve 21 would look like if we did WASH 1400 with no evacuation, 22 what that would look like.
23 But I guess by looking at the results in A
/) 24 various ways I think we came to the conclusion the v
25 effectiveness of the containment is probably the greatest ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
137 i t
- 1 effect that we have in getting these f avorable results.
2 But the source terms have comparable order .of magnitude 3 effect I believe because you can see the source terms _alone 4 will change the results by two orders of magnitude.
5 'DR. MOELLER: I thought in your report i
j 6 somewhere it implied or, and I agree with what you are 1
1 7 saying, but the report I thought or one of them implied the i
8 source term didn't have all that much impact.
i 9 MR. FLEMING: Yes, I don't think that is l
j 10 correct. I would think the source ~ term --
11 DR. MOELLER: Sure, j .
12 MR. MOODY: Has comparable impact.
I j- 13 DR. MOELLER: Can you show me later when you k -
j 14 are showing these risk curves how you get the last two bar i
j 15 graphs?
i 16 MR. FLEMING:- I can describe it right now. We 17 estimate the total risk to. the population within one mile i
i 18 of site boundary. That is an accident frequency weighted
.l 19 by the number of health effects that would occur.
20 Then we simply, so that is the total risk in 4
21 terms of number of fatalities per year. We divide that by t
I 22 the population that lives in that zone to get an average i
j 23 individual risk. l 4 I i
() 24 DR. MOELLER: All right.
l j 25 MR. MOODY: That is basically how we do it.
I I ACE FEDE'AL R REPORTERS, INC. Washington, D.C. (202) 347-3700 j
138 1 DR. MOELLER: Okay.
2 DR. REMICK: Am I correct? I understand you 3 are one mile f rom the site boundary f or what you just did 4 but when we talk about evacuation that is one mile from the
' s s
5 reactor, not the site boundary?
16 .
6 MR. FLEMING: Yes. It's unfortunate that we 7 had that confusion. It stems from the fact that we 8 originally set up our Seabrook site model with increments 9 of spacing, grid spacing based on distance from 10 containment. '
11 It became very inconvenient and costly to 12 consider rematching the spacing. When you huve to be 13 careful when you see the miles in the reports.
14 I think we have accurately stated what we meant.
15 It was a point of some confusion.
16 DR. MOELLER: To straighten this out, when you 17 say you are evacuateing one mile, what is that distance?
18 MR. FLEMING : From containment. If we 19 evacuate one mile from containment we are still leaving a
~20 half mile of people into the safety goal calculation that !
21 don't get evacuated. j i
22 If we go two miles -- see what I am getting I l
23 at? The safety goal goes out from the site boundary. l
() 24 DR. MOELLER: Okay. So,your exclusion zont is 25 roughly a half mile in radius, is that what you are telling !
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 l
139 l l
l 1 me?
O- 2 MR. FLEMING : Yes.
3 MR. MOODY: 30,000 feet.
4 -
MR. FLEMING: 30,000 feet.
i 5 So all the evacuation distances in summary are 6 ' measured in distance f rom containment. The only time 7 distance from the site boundary is in the safety 8 calculation.
9 DR. MOELLER: Where you talk about evaluation 10 f rom one mile f rom containment, again,_ you own half that 11 distance.
12 MR. FLEMING: That's correct.
13 DR. MOELLER: Or more. And you have total 14 control of that, anyway.
15 MR. FLEMING : That's correct.
16 DR. MOELLER: All.right.
17 MR. FLEMING : That's correct.'
18 (Slide.)
19 MR. MOODY: Next we get into the NUREG 0396 20 comparisons. These have come up several times today.
21 What we have here is two'different, three 22 different sets of results. Solid line represents the 23 curves right out of NUREG 0396. The document used to set
() 24 the, justify ten mile evacuation zone for all -plants.
25 These are curves for 200 rem and 50. This is ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
140
- 1 like a double risk, doublely conditional risk curve. It's 2 a frequency of exceeding dose at or beyond a given 3 distance.
4 We come up with dose versus distance curves.
5 The dashed lines again represent the WASH 1400 6 methodology source terms foresee brook and the dotted lines 7 down here represent the good source term or updated source 8 term Seabrook results.
9 As far as the WASH 1400 methodology source 10 terms results are concerned, for boat 50 and 200 rem we 11 start at levels below where the NUREG 0396 results stop at 12 ten miles.
13 So the curves even begin to appear at a lower 14 distance than they end at ten miles.
15 As you can see for the.200 rem curve, the key 16 cu rv e, I think, f or judging the ten miles _ in NUREG 0396, 17 essentially goes to 0 between one and two milec. Thic is 18 two miles, one mile. Between one and two miles it's 19 dropping off very rapidly.
20 DR. MARK: How does that glitch on the 50 rem 21 curve four to six miles, that must be an artif act of the 22 computer or what?
23 MR. FLEMING: I am not sure what you refer to.
( 24 DR. CORRADINI: The glitch, the step at 50 25 rem. That curve.
ACE FEDERAL REPORTERS, INC. Was}nington,D.C. ( 202) 347-3700 -
141 1 MR. FLEMING: Oh, that. That is what these 2 risk curves look like. They represent the --
l 3 Keith has to explain this, I think.
l 4 MR. NOODARD: Yes, if you were to look inside 5 at what made up that outer curve you would see a number of l 6 individual cumulative distributions from each release 7 category.
8 Then they are f requency weighted so they move up 9 and down.
10 That happens to be one that has what we would i
11 call a larger tail with a lower frequency. So the one that 12 bound it --
13 DR. MARK: -- -like a solid curve which goes l
14 vertically down at some point. This thing tells me that I, 15 at 50 rems, just'as much as six miles as four. And I' don't 16 believe it. ,
17 MR. FLEMING : It stems f rom the f act that we 18 had a discrete set of accident categories.. There.is one 19 category that has a curve that does fall off sharply, 20 another starts over here somewhere, this was. the sum of ' the . .;
I 21 curves, and causes a trail to truncate over here..
22 DR. CORRADINI: That leads me to my question. l 23 I am asking, given~ core melt, how do I add these . things up h 24 to get probability?
~
25 And you'say.part of it is probacility o'f the I
AcEFEDERy_ REPORTERS, INC. - Waspington,pC. }202 { 347-3700, -j
142 e
1 release category and the other part, then you have to add-2 all those guys up. Part of it is the consequence 17 3 statistics in ' terms of meteorology, population, evacuation.
4 MR. FLEMING : That's right. Basically the way 5 you get this is you basically sum up vertically over each i., 6 of the release categories weighted by the probabilities and 7 then this happened to be normalized'against core melt 8 frequency. ,
9 So this, you start out by having the-10 frequencies of actual sequences, each has a separate . curve.
11 You' sum them, directly vertically. Then at the end you 12 divide the total by core melt frequency so you get the 13 normalization, given core melt.~
14 That is basically how they .are arrived at.
15 DR. MOELLER: Now, if the RMEP's gives you 50 16 rem, you know, your curve down there, how do you now relate 17 that to no need to evacuate beyond --
l 18 MR. FLEMING :_ The way we interpret 0396, if I 19 take the values at ten miles, I make a vertical slice up_
20 this way, the f requency values f or what they. considered-21 acceptable,for 50 rem was up here somewhere.
]
22 DR. MOELLER: You are way below that.
23 MR. FLEMING: Yes, we start. out down here.
() 24 DR. MOELLER: Okay.
25 MR. FLEMING: We don't need any evacuation ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
t 143 I
1 zone to meet 50 rem, using that definition.
- O 2 Same for 200 rem. It does gown to about, down
- 3 to about this point. It's pretty close. This starts out 4 just a little below where 0396 ends at ten miles and then 5 drops off rapidly.
l
- 6 One thing that, you know, it's been sort of an 7 anecdote in discussions about 0396 is the so-called knee in 8 the curve. Somehow they took the curve and came up-with 9 ten miles.
I
- 10 A locality of people have tried to come up t
11 with scientific definitions of what the knee in the curve 4
12 meant but the curve actually goes-out beyond ten miles 13 bef ore it actually goes' vertical. If you want to take a 14 vertical criteria they would have taken 15 miles.
t 15 Somewhere you have to wrestle with how you 16 come up with ten miles. We said we will just go vertically .
17 ten miles and that point represents a frequency scale we 18 use as a criteria.
19 Seems to be the most^ logical way to use the 20 results.
21 DR. CORRADINI: If one chooses another~ measure 22 one could get different conclusions then is your point.
- 23 MR. FLEMING
- 'Sure. -Absolutely.
, - () 24 (Slide.)
25 MR. FL EMING : Now the similar curves for one i
ACE FEDERAL. REPORTERS, INC. Washington,'D.C. .( 202) -347-3700
144 )
1 and five rem doses are indicated in this slide. We have 2 the benefit of being able to see all the curves.
3 On the previous slides we only saw the 50 rem 4 results for the good source term, best estimate source term 5 study. Now we see all three sets of curves, three pair.
6 Each pair has a one rem and five rem in-it.
7 The dash lines are the source term methodology for 8 Seabrook. We start below where these curves end up at 9 something like 50 or 60 miles.
10 We are getting lower risk of exceeding one and 11 five rem at 'one mile than achieved at 50 or 60 miles in 12 0396.
13 These dotted curves represent the best 14 estimate, take advantage of the best estimate source term 15 information.
16 Those results are considerably lower. That is 17 the results for one and five rem.
18 From looking at both WASH 1400 safety goals and 19 NUREG 0396 it was basically shown all those criteria can be 20 met with a one mile assumed evacuation.
21 ( Slide . )
22 MR. FLEMING: This has come up before. The l 23 extent to which that last conclusion can be made was, while 1
l
() 24 not wholly $urprising to us, we didn't realize how powerf ul l 25 of an effect the containment analysis was having. So we l
l l
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
145 1 went back and took a look at WASH 1400.
O- 2 If you look at the WASH 1400 release 3 categories f or PWR's. You get core melt f requency of 4 something like 1 times 10 to the minus 4 based on mean 5 values.
6 If you take categories.PWR 1, 2 and 3, those-7 were .respectively f or the steam explosion, 8 overpressurization and V-sequence and hydrogen burn, all 9 involve very early containment f ailures and large 10 contributions to the risk curve, early fatality risk 11 curves.
12 If we use those as a measure of what the O 13 release f raction is in WASH 1400 we- would calculate for 14 PWR's 34 percent of the WASH 1400 results were large scale l 15 releases. h -
18 l 16 The Seabrook results were identified as being )
17 1 percent. Further reduced to one tenth of one percent in 18 the updated study. l l
19 That kind of gives you, that is where the 20 reduction is taking place when we don't take credit for -
21 advance source term technology.
22 Why is that low? I will show you what we have 23 done to sort of go back into our results.
24 (Slide.)
25 MR. FLEMING : On this chart we have indicated i
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
146 1 all the results we have analyzed that have potential f or 2 containment bypass or early containment f ailure. I have 4
3 sorted them into some groups here. ,
a
) 4 We have initiating events that open up a
! 5 containment bypass condition. Those. include what we call l
6 interfacing LOCA's and steam generator tube rupture which I 7 was another example mentioned earlier.
i .
[ 8 Those are ways in which the initiating event 9 itself can actually open up a pathway for bypassing .
10 containment.
1 11 We also have some external events that can l
3 12 f ail containment f rom - the outside in in the form of' missile-
. 13 penetration due to assumed aircraft crash or turbine j 14 missile, for example.
i l 15 We did include the analysis of those in the 16 Seabrook analysis.
i 17 Finally we have all the other more :
] 18 conventional containment failure scenarios that'are -
i '
19 normally considered.
20 Those-in which containment is assumed to fail i 21 either due to an isolation f ailure or by the degraded core- ;
22 phenomena acting on' the containment and have 'it lose its i 23 structural integrity ~ early on in the scenario.
l
() 24 That is what was happening in the' hydrogen 4
.25 burn scenarios in PWR 3, in the steam explosion scenarios i.
, ACE FEDERAL REPORTERS,.INC.- Washington,_D.C. (202) .347-3700
. _ . _ . _ . . _ _ . _ _ , . _ _ _ _ _ _ _ _ . _ . _ . _ . . . _ _ . _ . . , _ . _ _ , _ - - . , _ . . _ . ~ , - , . . _ _ _ . , _ , _
i t 1 47 1
1 in PWR 1.
l 1
2 So we have all these in Seabrook. I guess 3 the, you know, question is, you know, why is our frequency
!~ 4 of those scenarios so low?
5 Well, if we take all those scenarios and just 6 group them together, what are the results, what do they i
7 look like now?
- 8 (Slide.)
1 9 MR. FLEMING: Results look like now for all 10 these early containment failure scenarios, they end up in _
11 one of three release categories we call S 1 which is like 1
12 PWR one in WASH 1400, S 6, typified - by open containment
! 13 purge, eight inch containment purge opened in the f ailed -
14 close as one typical example.
i
, 15 Then S 7, our new 1GIR pump bypass where ' the by i 16 as occurs at the RHR pump seal in the containment vault.
1 l 17 If I put all these scenarios together, ~ I -get j 18 basically the contribution to the early release f requency 19 which I said was one tenth of one percent of core melt or-20 something like in the middle 10 to'the minus 7 frequency.
~
21 We have the most likely scenarios 'being 22 earthquakes that do not cause afloss of off-site power, 23 and, therefore, the compressors stay on, . but the earthquake
() 24 is assumed to fail the solid state protection system as one 25 of the most important sequences.
ACE FEDERAL REPORTERS, INC.
Washington, D.C.-
'(202) 347-3700
148 1 There is no actuation signals to the core
() 2 cooling and containment isolation systems and we get into a 3 problem with the most likely way to get into the, this 4 category.
5 We still have interfacing LOCA. We have 6 preassessed these but they have moved into second place in 7 terms of risk contribution, now at 13 percent.
8 There is really two general types of scenarios -
9 in this category. Tnere is the pipe break type where we 10 don't have any rotential for scrubbing.
11 Then we have the RHR pump vault type bypass 12 where there is a high potential for scrubbing effects.
13 Both of those make up this 13 percent with the 14 split in frequency about an order of magnitude between the 15 two of them.
-16 We also have turbine missiles impact the 17 containment building with an assumed failure.
18 We had the reactor vessel steam explosions that 19 are still in the study but the probabilities are very low.
20 That is why the total frequency is low.
21 We also have aircraf t crash impacting containment 22 and this refers to aircraft crashes that exceed those that 4
23 were shown to be defended against in the original design of
() 24 the Seabrook containment.
25 It was'shown in the FSAR that the containment l
l ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347-3700
149 1 could withstand an FB-111 impact of a certain weight, 2 roughly 80,000 flying at subsonic speeds.
]
3 We took credit for that in a sense that we did 19 4 not assume that such impacts would f ail the containment.
i 5 We calculated probability of having an impact 4
j 6 more severe than that due to a heavier aircraf t or higher 7 speed aircraft.
8 Those were assumed to penetrate the 9 containment building without any credit for the ability.of 10 the containment to stop the aircraf t, essentially a 11 simplistic f ragility analysis, assume that once you go 12 beyond design basis it always fails.
13 Then we assumed the aircraft would fail a i 14 steam generator or pressurizer, knock it over and basically 15 cause a LOCA. And a very, very large hole in the 16 containment on the way in.
i 17 We also have, not shown on the list, making an 18 insignificant contribution were steam tube rupture i
19 scenarios that ended with open containment purge and core 20 melt.
21 The reason those don't show up is first we 22 found it highly unlikely such a rupture would even have a.
23 core melt and if it did there would be a high chance the
() 24 operator could take steps to isolate the break by 25 depressurizeing and bottleing up the failed stem generator.
I I
ACE FEDERAL REPORTERS, INC.- Washington, D.C. (202) 347-3700-
4' 150 ,
I i
1 DR. REMICK: On those percent figures is that O 2 percent contribution to the f requency of the occurrence?
3 MR. FLEMING: Yes..
4 DR. REMICK: It is, not risk?
I 5 MR. FLEMING: Not risk.
6 DR. REMICK: Okay.
7 DR. MARK: Could you say again maybe what you
- 8 did say about the earthquake? .
]
9 Is this an SSE or all earthquakes up to San 10 Francisco level or what?
i 11 MR. FLEMING :. We did an earthquake,.in our
} 12 standard approach to earthquake treatment or seismic events 13~ we considered a spectrum of earthquakes that ranged from
. 14 below the SSE level to well above it, and we take it-high j 15 ' enough to capture the situation..where everything is likely 16 to fail in the plant. That is the spectrum of i
17 considerations.
i 18 Then we estimate the f requency of the 19 earthquake as a function of, in this case, . peak horizontal
~
20 ground acceleration and we estimate seismic f ragilities of 21 ' structures and components, and incorporate 'those into our-22 model.
23 The particular scenario.that comes up,-the
()
~
24 kind of scenario that comes up here are' earthquakes much
- 25 more severe than the design basis earthquake or'SSE that i
i i
i
-ACE FEDERAL REPORTERS, INC.
Washington, D.C. (202) 347-3700 _ _ . _ . . , _
151 1 f ail the solid state protection system and which is the l 2 nerve center for sending actuation signals to all the j- 3 safety says systems in the plant. .
f 4 DR.' MARK: I vaguely understood that. You are 5 talking about earthquakes f rom which you would surely have 6 lost off-site power.
f 1 7 MR. FLEMING : We calculate a , probability that t
8 you will or will not lose off-site power and those that-
}
! 9 have a loss of off-site power we assume that the air will-
{
10 be lost to those containment isolation valves if they are 11 open and that we take credit for the high chance that they 12 will close. -So those scenarios kind of go away.
i 13 The scenarios remaining are the .ones where Lthe unf ortuitous combination of events of 'the earthquake, no .
14 15 loss of off-site power but damage inside the plant that:
16 fails the S S P S.
I 17 Since this updated result was . completed :irr 18 parallel with this we had reassessed the seismic I
- 19. f ragilities of key components. S M.A who'had originally
. 20 done the seismic f ragility work took into account. new 1
21 information made available to them since new test reports I 22 and so forth. And we were responding to some comments 23 Lawrence Livermoore made that .they thought our f ragilities-() 24 were overly conservative.
- 25 Those capacities were reassessed.. And.they ACE' FEDERAL REPORTERS,-INC. Washington,~D.C. (202) 347-3700
152 1 have not been incorporated flew these results. These
-)
2 results are based on the old Seabrook seismic fragilities.
3 DR. MARK: It seems to me your solid state 4 components are not terribly fragile for shake, are they?
5 MR. FLEMING : It's things like cabinets, not 6 the components.
7 MR. MOODY: 120 volt AC power that is required 8 to operate the solid state --
9 DR. MARK: Okay, that is what gave way.
10 MR. MOODY: Yes.
'll MR. FLEMING: Yes, 12 DR. MARK: Is it pretty well tied down?
13 MR. FLEMING: We found it's tied down better 14 than we originally assumed. - We anticipate when the results 15 are incorporated in a future update that probably won't be 16 completed until ' 87 --
17 MR. MARK: There was a time not many years ago 18 when Brookhaven was opening all the breakers in the whole 19 plant with a small earthquake. You don' t do that here.
20 MR. MOODY: That has been assessed for some of 21 the key components during this update. But in parallel we 20 22 are perf orming a detailed -relay chatter review in-house.
23 DR. MARK: Okay.
() 24 MR. MOODY: We are going down that path also.
25 DR. MARK: That isn't finished yet?
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
.153 1 MR. MOODY: No, the electrical part is 2 finished but not checked.
i 3 Systems people _ are getting involved and we-4 will _ probably go back to S M A if there are relay chatters
. 5 of significance.
6 MR. FLEMING: In response . to a question ,
7 received this week f rom staff in support of their review,.
8 we are providing staff with some roughest-mats we are 9 making of the combined effects of relay chatter and more 1
10 favorable fragilities.
11 A more detailed look at that will be made
! 12 later.
13 DR. MARK: Thank.you.
i l 14 MR. FLEMING : Of course, a very big ~effect has 1
- 15 been made by the new treatment of the interf acing system i
16 LOCA because this 13 ' percent contribution used to actually 17 be a much greater contribution in this earthquake, the 18 ranking switched around. That stems f rom a net reduction 19 in the contribution to the interf acing system LOCA.
20 (' Slide. )
21 MR. FLEMING : What we noted in the Seabrook 22 PRA, retrospective look at it, in all the analyses that 23 have ever been done that we are aware of, it was pretty
() 24 -clear to us these events had not been analyzed with the 25 same degree'of detail and realism as events like LOCA's and ACE FEDERAL REPORTERS,'INC. Washington, D.C. (202)-347-3700 I
154
- 1 transient events typically treated in PRA's today.
v 2 There were insights provided in the IDCOR 3 program, determineness particulars calculations done at 4 Zion for example to show that there is a good chance the 5 RHR piping might be able to withstand the kind of pressure 6 transient you would see if you postulated rupture of either 7 check valves or motor operated valves depending on which 8 side of the system you are talking about, normally isolate 9 the low pressure RHR piping from the normal reactor coolant 10 system pressure.
11 And several other- things which had been 12 basically, neglected could be conservatively treated in the 13 Seabrook and all the other PRA's.
14 W'e thought bef ore any definitive statements 15 were made about leading health risk that the leading 16 contributor ought to be re-examined and subjected to the 17 same treatment we had given say.high core melt 4
18 contributors.
19 There was not as much motivation in the original 20 PRA because the results with the conservatisms included 21 were indicating a favorable result with are expect to the 22 safety goal.
23 But the enhanced treatment we considered for
() 24 the interfacing system LOCA led to the following set of 25 enhancements if you want to call them that, or changes, ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
1 155
[.
i 1 some of which have an unf avorable overall impact, some of j -2 which have a f avorable impact.
3 To go down the list we considered a more-4 complete set of valve failure modes. _The traditional
- 5 sequence type analyses what is assumed ~is that you have a 6 -rupture of say the inboard check valves that normally cease
~
7 the 1arge portion of'the-pressure drop between the reactor 8 coolant system and RHR system.
9 Then at some point in time between that 10 rupture and the next time they decide to do a test on the i
11 valves a second rupture occurs, sort of'a random coincident '
c
! 12 fa'ilure of two-series check valves.
1 13 We added to the model the chance that the 14 outboard check valve would f ail at the instant it becomes
,i 15 loaded af ter the first one- becomes available...That 16 enhanced the treatment of that failure mode.
- 17 We also considered the fact that there is.some i
j 18 possibility that the-inboard valve might: leak and the i'
19 pressurized space, the inner space might be. pressurized and i-
! 20 the loaded valve might be the one on the outside. So' we i
j 21 considered you ~could have the valves occur in either 22 sequence.
- 23 We treated that probably more conservative,
() 24 that aspect more conservatively than we did earlier. -
l 25 Also in the old V sequence analysis or-4-
3 ACE FEDERAL REPORTERS, INC. Washington, . D.C. _.(202) ~ 3 47-3700
156 ;
- 1 Seabrook analysis, frequency of the valve ruptures was 2 estimated on the basis of a search for very severe check'
- 3 valve ruptures af ternoon not finding any in a large . ,
4 . component population of evidence examined. l 5 It was one of the zero estimates where at best r 6 you can only set an upper bound on what the true frequency 7 is.
8 We decided that that lef t us a little bit cold i
, 9 on a method of quantification, so we decided.to.take a 10 better approach of actually trying ' to look at .the' frequency i 1 11 of valve ruptures as a f unction of the size of the . leak.
12 We did have some data on valve leakage events..
13 No major ruptures were found but we found check . valves .that
- 14 leaked in the vicinity of 200 gallons a minute with a few 15 thousand psi pressure drop across them.
16 So we considered that, i
l 17 We also took more attent' ion at modelling the:
l 18 very important response. RHR relief valves,.and in 1
i 19 particular there is one set of relief .v'alves inside
] 20 containment set at about 450 psi.
1 21 They have a capacity to- relieve at normal 1 1
22 reactor coolant system pressure of something like 2,000 23 gallons a minute, a thousand gallons a minute each or 900.a
() 24 minute each,1800 gallons a . opinion f or the .two valves.
25 If one.has a check valve leakage that has a ACE FEDERAL.' REPORTERS, INC. . Washington, D.C.
- ( 202) ~ .3 47-37 00
157
- 1 lower flow rate than that, the RHR piping will never_ see a 2 pressure increase much in excess of a set point on .the 3 relief valve.
4 So the assumption that you go directly from 5 check valve ruptures to piping f ailure doesn't really take 6 due account of the behavior of the relief valves. So we 7 modeled them, gave them a chance to work or f ail uaing data 4
8 on the reliability of the relief valves.
,9 We also looked at the piping in the RHR 10 sy stem. In-the Zion IDCOR study they did deterministic 11 calculations that showed the stresses in the stainless i
12 steel piping. in the RHE system designed to 600 psi would 13 not even approach a conservative value for yield stress.
14 Therefore, deterministically assumed in that work that it 15 would remain intact.
16 We wanted to treat it in a more probabilistic ,
I i
17 way. We came up with a f ragility curve f or the piping.
18 That curve was arrived at by calculating the pressures at 19 which one would obtain yield stresses in the-largest piping 20 and ultimate stresses in the largest piping based on 21 conservative values that had been selected for yield and i 22 ultimate stress of those materials. l l
23 Then despite the fact that there is no
() 24 evidence that materials fail at yield stresses, not to l 1
. 25 mention conservative stresses, we assumed a one percent- !
i ACE FEDERAL. REPORTERS, INC. Washington, D.C. (202) 347-3700
- _ . . . ~ . _ . - . _ . . - . _ . - _ . . . . .. _ . - . ..
158 1 probability of f ail'ure at yield and 99 percent - probability l
O 2
~
of failure at ultimate.
4 3 We used those two points to construct a 4 f ragility - curve which we use for the RHR pipe.
5 Fe also recognized that there,may be other- >
6 kinds of failure modes that are not the intrinsic i
7 properties of the material but rather. construction defects 8 that may go on-undetected and.not be revealed at the time 9 the system is cold hydro and so forth.
10 We designed, assigned a probability based on i 11 our subjective judgment that there might be some hidden 12 flaw in the piping which would reveal itself once 13 pressurized.
14 The. combination came up with a f ragility curve 15 essentially which we estimated if f ully pressurized,' that
, 16 there would be roughly a one percent chance.the piping i 17 would break. That was the f ragility model we used in the.
j-18 analysis, i
4 19 We.also had a.t'reatment of the RHR pump seal-
~
20 leakag'e. We didn't know exactly how large or small the 21 pump seal leak-would be, so we assigned'a range of values 22 ranging from very small leakages up to the upper bound leak 23 rate you could get if you assumed the 'whole seal 0-ring 24 material and so f orth blue out cf- the pump.
i 25 Then we looked at a spectrum, we;did ACE FEDERAL REPORTERS, INC..
Washington, D.C. J 202): 347-3700
. _ - , . - . . . . _ - . - . - . - , , - , . . . ~ . . . ,
159 1 sensitivity studies to convince ourselves the results O 2 weren't really sensitive on where we set that RHR pump seal 3 leak rate.
4 We also brought in operator actions on one 5 type of interfacing system LOCA involving check valves. If
~
6 the operator can diagnose correctly the accident sequence, 7 he can actually isolate the leak using the installed motor 8 operated valves in the system, but there are other i
9 scenarios he can't based on the valve arrangement they 1
10 have.
11 But operator actions could be taken to either j 12 isolate the leak and terminate the process or even if the 13 leak persists, to maintain adequate core cooling. Some 14 consideration was given to that,-with due regard to the i
15 environmental effects that would be associated with RHR 16 pump vault flooding.
17- RHR pump vaults include containment building 18 spray sums, RHR and high pressure injection pumps but not 19 the charging pumps.
2 20 We did identify, you ' know, there is high
- 21 potential that depending on what kind of environment you 22 had in the RHR pump vaults that you would in fact fail l
i 23 those pumps. We quantified that.in the study.
() 24 Still even with failure inside the vaults 25 there are ways using charging pumps that one can prevent ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
160 1 core melt which we took credit f or.
Os 2 The last thing we did is thermal hydraulic end 3 source term analyses which Westinghouse perf ormed f or us 4 using these MAAP computer codes-include including the 5 thermal hydraulic model they have on the f ront end.
6 We modeled a couple scenarios, one assuming wet, 7 one dry RHR pump vault conditions. W'e used and assumed 8 maximum pump seal leak rate for that condition and 9 quantified uncertain.
10 So those are all of the things we did.
11 DR. REMICK: How do you model in, in the case 12 let's say that the operator either misdianoses the 13 situation or properly diagnoses the situation but 14 inadvertently presses the wrong button and isolates the 15- wrong system.
16 I am thinking of the Davis Bessie, the steam 17 feed rupture control system where he punched the wrong 18 button and isolated the aux f eedwater system by mistake.
19 How do you factor that in? You mentioned --
20 MR. FL EMING : We consider that in quantifying 21 the probability that he does not properly diagnose the 22 event.
23 Then, after we resolve which scenario we-are 24 talking about, does he or doesn't he, if he does correctly (f
25 diagnose it we then consider the possibility he may i
ACE FEDERAL REPORTERS, INC.
Washington, D.C.
(202) 347-3700
161 j 1 improperly f ollow out the procedure.
- 2 If he doesn't diagnose it properly, we assume the 3- operator doesn't officer any difficulty at all.
4 What is more difficult is he takes action to i 5 make things worse. We looked at scenarios like that by 6 going over the sequences in the study on the simulator with 7 the operations group.
8 And we chased down various scenarios and 9 convinced ourselves that the scenarios interested in, that 3
10 the errors of~ commission were not of real concern because j 11 most of the ones we postulated tracked down in the 12 simulator on the original PRA, there was flow real
, 13 consequence of misdiagnosis.
l 14 In this case there is a potential for 15 misdiagnosis between a LOCA inside containment and one 16 outside containment.
17 We flagged that potential. It~is now being 18 addressed in the training program. In; fact, . plans are made 4 19 to actually add a module into the core so that the 20 operators can -- because you really would have a LOCA at 21 both places if you had one outside containment because the i
22 RHR relief valves vent into the pressure relief tank inside ,
23 the containment. l J) 24 Eventually, if not mitigated, the ruptured l
25 disk or whatever it is on the tank would open and you would 4
l I
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 !
162
. 1 . have conditions of a LOCA inside the containment. Then
! ( ).
j 2 possibly seals might' give way later on- in the scenario.
! 3 .So there is_ potential for misdiagnosis but it
- 4 is being addressed as part.of the training program.
1 ..
5 The net impact, one way to look at the net 6 impact of these, all these improvem'ents is, I tried to !
i 7 capture that in sort of three basic numbers.
8 (Slide.)
) 10 analysis we have quantified f requency of I initiating valve 1
11 ruptures which at a minimum will give you a low because of 12 the presence of the RHR relief valves.
l 13 In the original PRA that was assessed at 1.8 4
14 times 10 to the minus 6. At the updated analysis it was 15 assessed at a higher frequency.
3 1 16 That reflects the fact we backed off on the 1
a 17 definition to include events th'at .would have leakages, leak 18 rates greater than 150 gallons a minute. That is the
- 19 capacity of one charging pump.
20 So that is where we started our analysis. So 21 we were more complete in our coverage of initiators, but s
22 because the original PRA only considered the mass of 23 ruptures as initiators.
l 4
f) 24 Our' frequency indicates as you would expect,
- 25 higher, even though we used a different approach to l 1
1 I l ACE FEDERAL REPORTERS, INC. -Washington, D.C. (202) 347-3700 l
163' f 1 quantify it.
1 2 But then in the original PRA this goes on to 3 be a valve rupture with LOCA and containment bypass and a 4 melt, the same frequency because we assume those factors 5 were 1.
3 6 Once you ever the initial things go, the melt 7 and containment bypass was inevitable.
8 In the updated analysis th . is not quite the 9 case. We assumed roughly 5 percent of the valve ruptures 10 with LOCA's would have a valve rupture in a LOCA in a i
11 containment bypass condition.
12 What that means is- that that 5 percent roughly i
13 represents the probability that the leak size will be less 14 than the capacity of the RHR relief valves.
15 If that occurs, all you will have is the LOCA 16 inside containment with the opportunity to mitigate that 17 with full coverage of ECCS equipment in the plant.
18 But we drop down to this f requency for a valve 19 rupture, LOCA, containment. bypass and melted. It goes to a 20 lower frequency because now we are' talking about more 21 severe valve ruptures to challenge the capacity of the RHR l 1
22 relief valves.
23 Only then are we then trying to challenge the l
() 24 integrity of the RHR piping.
25 There is some chance the piping won't break. If l 1
l l
l ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700 ;
164 l 1 it doesn't break the operator has a chance to save the 2 plant and save the melt.
3 That is why we have basically gone down to one 1
4 and a half orders of magnitude, about, between where we 5 started out and where we ended up in terms of containment 6 bypass frequency.
7 So that is 'the net eff ect of those 8 enhancements.
9 At this point I am going to stop and ask for 10 any questions you might have because what I am going to <do 11 is, next, I am going to turn the~1ast three slides in the i
12 precentation over to Fred Torri, our. containment expert.
13 He really should be presenting that.
4 14 If you have questions on the V-sequence, f or a
4 15 example, now would be a good time.
16 DR. MOELLER: I hear none.
17 MR. FLEMING: Is there a question back there?
18 DR. MOELLER: No, I guess not. 'Let's go 19 ahead, then.
20 MR. FLEMING: Thank you.
21 MR. TORRI: I am Fred Torri. - I am from I
22 Pickard, Lowe and Garrick.
23 The presentation I am going to give you on the
() 24 containment failure analysis is intended as a brief 25 overview of the containment analysis work. It.was done i
i ACE FEDERAL REPORTERS, INC.
Washington,.D.C.
(202) 347-3700 . , . ~ ~-
165 1 concentrating on the analysis. that were done bef ore the
,.O 2 containment failure.
I i 3 It is not intended to be a detailed, 4 addressing a detailed, in detail all the facts and analyses 5 we made.
~
l 6 The objective 2of the containment failure analysis 7 was to determine.the expected location, pressure and leak 8 area at which failure of the containment, of the Seabrook 9 containment would occur.
- 10 I have to stress again, expected. We defined 11 that as a 50 percent probability of f ailure.
, 12 And then, since we are addressing expected l
1 13 behavior, or most likely behavior, we also are interested i 14 in determining what are the uncertanties about the pressure 15 and leak area that we determined from this analysis.
16 We did not try to answer the question, where
~
17 is a lower bound containment failure pressure, as has been. l l
1 18 done in the past when containment failure was addressed.
19 We felt that approach was inconsistent with 20 the spirit of PRA, which is attempting to find the expected i
i 21 behavior of the plant, and not a lower bound or 22 intentionally conservative behavior of the plant.
I _
23 As a result of the early examinations of the 24 behavior of this containment under pressure, it became i
l l 25 obvious to us that an assessment which would define 1
t 1
Acsy sosa g a m a3sas, m C. -Washington, D.C. ]202) 347p700-_,
166 1
1 containment failure as some strain limit or some stress 2 state such as yield would be very conservative.
3 The reason, we believe, for this assessment 4 wasthecharacteristicbehaviorassociajdd'withafully 5 rebar containment, as opposed to a post tension containment 6 where the load carrying members, the tendons, are designed 7 to slip relative to the structure. And that is done 8 intentionally so the tensioning mechanism can function.
9 This is not the case in the rebar containment. l 10 The see bar' con the Seabrook containment as seven full 11 layers of rebar spaced across the concrete f all which 12 ranges from three and a half to four feet thickness.
13 Our first objective was to define containment 14 f allure types.
15 (Slide.)'
4 16 MR. TORRI: We decided that this was necessary 17 because in terms of calculating consequences, it.is 18 important to distinguish whether you have a small 19 increasing leak rate early on in an accident sequence or 20 whether you, whether you wait until' you have a gross 21 containment f ailure mode where then the release occurs very 22 quickly.
23 We defined three containment failure types. j 24 Containment f ailure type A is defined as a small leak, in 25 the range ~of .02 square inches.to six square inches
)
. ACE FEDERAL REPORTERS,.'INC. Washington, D.C. -(202) 347-3700-
167 1 equivalent. orifice.
73 U
2 For these, lower bound, .02 square inches as 3 defined by the design basis leak rate of the containment of 4 .1 percent per day.
5 The upper bound is defined as that leak where 6 if it existed pressurization of the containment would no 7 longer continue, but the leak rate would be just enough to 8 level, cold pressure level in the containment.
9 Therefore, within this range of leak, 10 pressurization of the containment will continue and will 11 lead eventually to either a type B or C failure mode. But 12 the release characteristics are such that the release l}-
L 13 increases from the design basis leak rate to some larger 14 leak rate, well bef ore the type B or type C f ailure occurs.
15 The type B containment failure ~is defined as that 16 ranging from six square inches to 60 square inches. For.
17 this leak rate the blowdown time, or the time over which 18 the release from the containment occurs, exceeds one hour, 19 and, therefore, is modelled in our consequence analysis as 20 a multi-puff release.
21 The one hour ctfi er! . stems from the fact that I 22 weather data is only available in one hour increments and, 23 therefore, our consequence calculations are staged in one I
() 24 hour increments.
25 A type C containment,_ this is approximately a ACE FEDERAL REPORTERS, INC. Washington, D.C. ~(202) 347-3700
168 1 'little less than half a square foot hole or orifice.
- _(
. 2 Type C containment f ailure is a gross
- 3 containment failure in excess of 60 square inches and we 4 assume instantaneous release of all the activities within l 5 the containment in.that case.
i 6 In other words,.those failures are modeled as
. 7 single puff releases.
8 The next step in the analysis then was to.look
- 9 at the shell behavior of. the containment and to . quantify 10 the pressure at which the - containment shell, in ' absence of '
11 any of- the penetrations, would f ail.
12 And 'we defined, or calculated, this was done i
13 by SMA, Don Wesley in the back of ' the room was in charge of l . .
14 that work at SMA, and SMA calculated deflections that would
- 15 occur in the containment wall if f ailure of the she11 'would ,
16 occur according to the-definition of a failure Dr. Seiss 17 addressed earlier and Don Wesley had already responded to.. ,
18 Those deflections were large. They were measured i
19 in feet. The deflections f rom the original position of the l
. 20 containment wall to the failed position of the. containment 21 wall.
l 22 And failure pressures were high. The data-is
- 23 in the S S P S A. There was no containment failure mode of
- 1( ) 24 a shell, no
- containment shell failure mode below 230' psia .
25 the wall hoop failure mode was the~ dominant she11'failureL i
! H t'
- ACE' FEDERAL REPORTERS, INC. Washington, D.C.
(202) 347-3700
. . . _ _ . . _ _ _ , _ ~ _ . _ _ , _ . . _ _ , .
l 169 r.
l- mode.
2 The next step, then, was to recogni-e'that 3 large deflections at failure had.to~be recognized and 4 addressed.
5 (Slide.)
6 MR. TORRI: At those deflections it is very 7 well 'possible that local penetrations will fail at a lower 8 pressure and lower deflection simply. if for no other reason 9 that interf erences between penetration piping and adj acent 10 structures in the outside of the containment are beginning 11 to occur.
12 We defined what we felt was quite a i
13 comprehensive set of local containment failure modes which
, 14 addressed fluid systems-penetrations.
15 Those. are the component penetrations and so 16 forth, feedwater and steam lines, fall transfer tube, !
j 17 electrical penetrations.
i 18 lie considered the purge line penetrations, the 19 containment penetrations by which the air is exchanged with
! 20 small valves used during operation containment purge,11arge 21- 36 inch set of linea used for ref ueling air exchange during 22 refueling.
23 We considered equipment hatch, personnel. lock i
We also
.(
5
) 24 and other penetrations.Seabrook containment has..
25 considered failures.as a result of liner tearings and weld 3
[ ACE FEDERAL REPORTERS, INC. Washington, D.C.
(202)' 347-3700
170 1 imperfections in the liner.
O 2 For each of these local containment failure 3 modes we determined the pressure and deflection at which 4 failure in these penetrations would occur.
5 And based on-the design and based on the 6 failure mode that was determined we assessed the 7 uncertanties in the' calculated pressure at which each of 8 these penetrations would f ail.
9 The next step, we were interested in two 10 things. Are there contain. Failure modes which would 11 occur before the shell f ailure would occur, which were very 12 high.
~
% 13 Secondly, if these occur earlier, are they of 14 a type A, B or C leakage characteristic because these are 15 the ones that can give you smaller than' category C 16 containment failure modes, any containment' failure mode as 17 a result of a shell failure was assumed automatically to be 18 a gross containment failure.
i 11 9 We did not attempt to consider any smaller 20 failure modes.
i 21 We then considered the f ailure for each -local 22 failure mode and uncertanties and pressures in the failure 23 mode determined an overall composit probability
() 24 distribution for the pressure at which containment failure 25 would- occur.
. ACE FEDERAL REPORTERS, INC. Washington, D.C. (202)' 347-3700
171 1 (Slide.)
i O 2 MR. MOODY: Pressure is given in psia and the 3 analysis considered two distinct types of conditione in the s
! 4 . containment.
f 5 If-during the accident sequence the refueling a
6 water storage tank was injected 'into containment, there was 7 plenty of water covering the debris.
8 The containment conditions were saturated 1
i 9 steam and the temperatures were governed ~ according ato 4
10 saturation temperatures, those are called wet containment-i 11 conditions.
I 12 If the refueling water storage tank is not 13 inject'ed, only the primary coolant system and the ,
- 14 accumulators are discharged into the containment.
15 If boil off of'all that water would occur 4
16' prior to containment failure and the containment atmosphere i
17 would transition to a super heated steam condition as 18 opposed to saturated steam condition. Super heated steam, i
19 gas inventory, so to speak, condition.
20 For those dry conditions, higher temperatures 21 in the containment atmosphere at f ailure are predicted and - -
22 the effect of these higher temperatures on the strength of 23 materials that are carrying the load prior to failure was l(f 24 considered.
i-l 25 That-resulted in the derivation of two:
l l
{ ACE-FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
172
~
1 different containment f ailure probability distributions.
O 2 The containment failure. probability 3 distribution for dry conditions is shown on the dotted 4 line.
5 You see that the 50 percentile, two, three, 6 four, five, 50 percentile confidence level f or dry 7 condition containment failure is at approximately 190 psi A
~
8 the one percentile confidence level, the 99 percentile 9 confidence level, I should say, the pressure at which we 10 are 99 percent confident that for dry conditions 11 containment will not fail, is given by a pressure of 12 approximately 150 psia.
13 So what in previous containment pressure capacity 14 analysis was considered to be the point of f ailure 15 according to our analysis for the Seabrook containment, 16 would correspond to roughly a one percent f ailure 17 probability. ,
18 For wet conditions the solid curve shows the 19 containment failure pressure probability distribution. It 20 shows that the median failure pressure is approximately 215 21 psi, a number mentioned before today. The one percent
~
22 failure probability for wet condition is around 163 or-2 23 psia.
() 24 Contributions to these failure pressure 25 distributions by type B and type C are shown for the wet i ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
173 1 containment f ailure pressure distribution by the dotted O-2 line for the type B, and the dashed line for type C.
3 If you take the ratio of type C containment 4 failure probability to total containment failure 5 probability, that is shown in this solid curve here with 6 the scale on the right-hand side, which reads failure 7 fraction for type C failure.
8 An interesting conclusion that we reached was 9 that if failure did occur both in the low probability and 10 high confidence range at the lower pressure ranges, in 11 other words, if the containment f ails at a pressure below 6 -
12 about 19,0 psia or 186 psia, we were quite confident that l 13 that type or f ailure was a type B f ailure with a slow 14 release and only had a probability, conditional probability l l l
- 15 in the range of 2 to 3 percent, that a low pressure failure l l
l 16 would be a gross containment f ailure mode. I 17 l Whereas, if the containment f ailed at higher i i
l 18 pressures you were getting closer to where the shell {
l 19 f ailure ~ modes were dominating.
20 And we reached a plateau where conditional 1
21 probability of f ailure was in the range of 40 percent for i
22 the type C containment failure mode. !
23 These results were used directly in the 24 containment response analysis in that the distribution of 25 containment f ailure according to the type of accident ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
- 4. s m ma4- ~ J a- e 4 4 +4r, 17 4 4
1 sequence was used to determine, in determining containment
-O 2 failure probabilities:in the.co.ntainment response analysis 3 and in the different release categories that Karl Fleming 4 discussed' earlier.
5 Are there any questions on the presentation that 6 I gave on containment f ailure analysis?
7 DR. MOELL ER :- Any' questions?
8 Dr. Seiss is our main man, and he has no 9 questions.
10 MR. TORRI: Thank you.
11 DR. MARK: Is there way of treating-the 12 containment unique to the Seabrook study, or.is it how they 13 are usually done?
14 MR. TORRI: According to our knowledge, it'is 15 unique in the sense that we don't know any other study that 16 has attempted to distinguish this kind of detail in.
17 understanding how the containment fails..
18 I think the other studies,.for whatever reasons, l
19 have tended to address the lower bound type containment 20 failure-conditions,. the condition which says-at this 21 pressure I am very, very confident ~that the containment l
22 will not fail and I am not willing to go beyond,,and l
23 defined that as failure.
)
O 4 oa a^ax= ** it eet te 1so ve==a 1e -
25 assumed the containment is'n't there at all. I l
l l
. ACE FEDERAL REPORTERS, . IN C. Washington,'D.C. (202) 347-3700 l
175 i V~ l
,s 1 - MR. TORRI: h'hatever the result. There'was a
~
2 whole range of results.
3 DR. SIESS: I don't think anybody else has 4 looked at penetrations in quite the same detail we have 5 seen here.
6 DR. CORRADINI: The Zion results were 50/50 7 probability at 150 pounds, wasn't it?
8 The Zion curve, if I were to translate it back 9 to something that I am f amiliar with f or the Zion safety 10 study, they generated the same curve. At 50, 50 percent 11 probability it was 150 pounds, was that correct?
12 MR. TORRI: I would like to refer the question O
V 13 to Dick Toland.
14 MR. TOLAND: This is on recollection but Zion 15 also took a lower bound approach. The 150 that you are 16 referring to is psi absolute. Psi gauge was 135 which was 17 correspond to the value he gave. - I think you are talking 18 in terms of gauge pressure?
19 MR. TORRI: No, on the graph I was talking in ,
20 absolute.
21 MR. TOLAND: 150 corresponds to that then. So 22 the 150 is correct.
23 DR. SIESS: Wasn' t the Zion' figure based j ust O
(__)
24 on shell failure?
25 DR. CORRADINI: Yes.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
176 1 MR. TOL AND: But as a percent strain, put a 2 certain strain on it and you assume it fails.
3 DR. SIESS: Did not consider any of the 4 hatches?
5 MR. TOLAND: No, they did look at hatches.
6 DR. SIESS: That curve didn't consider 7 hatches.
8 MR. TOLAND: They-looked at hatches, did look 9 at a few penetrations, did look at some temperature 10 effects.
i 11 MR. TORRI: We have perf ormed prior 12 containment f ailure containment response analysis on other 13 containment types that were not of the rebar type, and for 14 those also adopted a strain limit or definition of f ailure.
15 DR. CORRADINI: One last question now. I 16 guess I thought I understood it. Now I am not so clear.
17 Wh'en you say dry, you have some sort of time
. 18 that you are assuming you hold the containment at high 19 temperature to reduce its strain.
4 20 What is that time?
21 MR. TORRI: What happens is'that you l 22 transition some, oh, a fraction of a day, several hours 23 ,
af ter vessel breach. You have boiled off all the water in j
() 24 the containment. Up to that point the conditions are
? 25 saturated.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
177 1 Then you are transitioning through a-
.( ) 2 containment heat up period where the decay heat is. taken up 3 by heating up the containment atmosphere but not adding 4 more steam to it.
5 That is a slower pressure rise, but associated 6 with that slower pressure rise, is a -- heating up of the 7 structural members of the containment wall is what 8 determines reduction in the strength of the materials.
9 DR. CORRADINI: You do it separately.
10 MR. TORRI: Yes.
11 DR. CORRADINI: You do the containment loading 12 and containment response.
13 MR. TORRI: Right.
7 14 DR. CORRALINI: So if containment response 15 buried in here is several hours of heating to define as 16 dry?
17 MR. TORRI: Yes.
18 DR. SIESS: What materials were involved?
19 MR. TORRI: Liner and first row of rebar were 20 exposed, when you got to the second row you'were far enough 21 in that there was not very much of a temperature increase.
22 DR. SIESS: And the penetration material.
23 MR. TORRI: I would have to refer back to Don
() 24 Wesley, if I can, to answer that part of the question.
25 MR. WESL EY: I don't have a very good answer l
4 ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
17 8 1 'for you, Dr. Seiss. What we did in our fragility 7--
(
2 evaluation is we based our temperature condition on the 3 design base system.
4 We then gave in essence a knock down factor 5 for increases in temperature, as Fred said, of the liner, 6 and the first several rows of.rebar.
7 Some things we did not attempt to address were 8 temperature effects on' the elastomer seals.
9 DR. SIESS: What about-the steel in the 10 penetrations?
11 MR. WESLEY: The steel-in the penetrations, 12 again, we have a knock down factor f or it.
13 DR. SIESS: All right. Thank you.
- 14 MR. TORRI
- The behavior of the seals was 4
15 considered in the context of their design in terms of 16 the -- particularly on the equipment hatch which is 17 designed --
18 DR. SIESS: How much temperature rise did you 19 figure, I saw somewhere 700 degrees F.
20 MR. MOODY: That was the maximum we found.
21 DR. SIESS: That's right.
22 MR. TORRI: This is again in the atmosphere.
23 The conditions typically associated with
() 24 saturated steam at containment failures were at 400 degrees 25 or below.
ACE FEDERAL REPORTERS, INC. Washington, D.C. (202) 347-3700
.. 17 9 1
f 1- The last vugraph summary is' going to be l.Q.
- 2 presented by Jim Moody. ,
- 3 MR. MOODY
- That concludes our prepared 4 presentation. All we had was a summary slide.
! 5 (Slide.)
i i
6- MR. MOODY: Which has been said about six.
i l 7 times I think.
8 DR. MOELLER: .Okay.
4 i 9 Are there any more questions ' f or any of the
- 10 people who have made presentations this morning?
i j 11 .I proceed assume all of you will be later this 12 afternoon in case something comes up.
i 13 MR. MOODY: Yes.
i 14 MR. PLEMING: Thank you, very much.
15 MR. FLEMING: Yes.
1 16 DR. MOELLER: Thank you'very much. We ;
i l 17 appreciate it.
j 18 We will now take one- hour break for ' lunch and t
1 19 resume at 1:50.
20 (Whereupon, the meeting-was recessed at 12:50 21 p.m., to resume at 1:50 p.m.)
{
1 22 i
^
23 O 24
! 25 l
3 . .. ,, - _ . _ ._ .. _ .
.-..____...,,b*
?.', ,, .*, .* . . . . . . . ._,__
28135.0 180 i
1 AFTERNOON SESSION ( 12:50 p.m.)-
, 2 MR. MOELLER: The meeting will resume. .The next ;
i
- j. 3 item on' the agenda is a presentation-by the NRC Staff,
! 4 which'will include an overview of the scope and focus of l
5 their review of the Seabrook work. This will be led tar 4
6 S teve. _ Long .
- 7 (Slide.)
Il '
j 8 401..LONG: My name is Steve Long, I'm the I
9 technical review project manager for the Seabrook emergency r
]
q 10 preparedness sensitivity study. I am going to lead off
- 11 this morning, or this afternoon, a presentation of the 12 Staff's focus for the review. We are just midway into the
( 13 review. There are no conclusions.at this point. We are 14 still trying to develop the total scope of the review.
- 15 We would like to have both your thoughts on' the 4-l 16 study and your thoughts on our scope. I'll start this off i
t 17 with sort of a picture of why we are'looking at the things
- 18 we are looking at. Then we'll have some additional people 19 from I&E to compare the document that Pickard, Lowe and-i 20 Gerrick has produced on NUREG-0396, which was the basis for 21 some of our emergency planning preparations. Then we'll 22 have a presentation of some of the personnel on Seabrook 23 who were contracted with us for help with the study. The 24 emphasis will be pretty much the methods that they are 25 using and.more detail on the area that they are focusing on.
- (7) _
i ,
~
- ACE-FEDER.AL REPORTERS, INC.
l 202 347-3700 Nationwide Coverage 804 336 6646
1 4
28135.0 181
.g-9RT ,
d .
1 The purpose of our review is, at this point, to 2 check the robustness of the conclusions the company has 3 reached by identifying and doing an audit-type review or a 4 little bit of independent checking on the most sensitive
, 5 aspects of the study.
l 6 (Slide.)
7 So that we are looking at-the conclusions of the 8 study for robustness. Those three conclusions are the
} 9 individual risk of early fatality at Seabrook, that is, b
i 10 being within the safety goal of the Commission; also, .that !
}
i 11 a one-mile evacuation at Seabrook provides a similar risk i
l 12 level with respect to early fatalities, so what the 1
- 13 WASH-1400 results'showed with a 25-mile evacuation; and the i 14 third one is that the probabilities of specific 15 radiological exposure levels at one mile from Seabrook are
- 16 less than the corresponding probabilities shown at 10 miles f
l 17 in NUREG-0396.
- 18 It is important here to recognize that there's a l 19 difference in how you would look at some of these 20 conclusions when you get into the'PRA.
21 (Slide.)
22 I see my slides did become scrambled. As a
~
23 matter of fact, one of my slides has become lost.
. 24 MR. MOELLER: While you are doing that, what 25 role will PEMA play in any decisions you reach?
ACE-FEDERAL REPORTERS, INC.
- 202-347-3'(o ' Nationwide Coverage 80tk336-6646
28135.0 182 9RT 0
1 MR. LONG: If I can:get Dave Matthews to speak 2 to that; Dave is from I&E and he's'really the-FEMA liaison.
3 MR. MATTHEWS: .This is Dave Matthews. I'm chief 4 of the emergency-preparedness branch in I&E.
5 At the present time, FEMA is not playing a role.
! 6 in this~ activity, insofar as we are just reviewing the i
j 7 technical presentations of Public Service of New Hampshire l
.! 8 against the primary technical considerations, which form
! 9 the basis for the rule that was published in 1980. At such I
l 10 time that a request might be made that would involve
- 11 determinations that would go beyond the primary technical-12 determinations, we would, of course, then be working with
' ( 13 FEMA with regard to what the implications-of any such i i
j 14 decision such as that would be.
1 15 MR. MOELLER: So, at some point FEMA will be
}
- 16 involved?
I
! 17 MR. MATTHEWS: Yes. But in the determination I
! 18 with regard to the technical. aspects of such proposal,' FEMA i 19 has stated in more than one instance than this one they did l 20 state it to me, that as a technical agency we look to you
- 21. to make a determination as to whether there's technical <
22 validity to such a request; with regard to what the l 23 implications in other arenas might be, and the efficacy of i
l 24 such an approach, then we'd be prepared to give you our l 25 comments.
ACE-FEDERAL REPORTERS, INC.
202-347 37(x) Ntionwide Coverage 801336 & l6
28135.0 183 RT J
l MR. REMICK: I have a related question. If this 2 request came in, it would be a request for an exemption of 1 3 a Commission regulation; is that. correct?
I 4 MR. MATTHEWS: Yes. Request for exemption, ,
5 would certainly be exemption from a Commission regulation.
6 That's what I understand; yes.
q 7 MR. REMICK: Do you know where in Part 50 that j
j 8 would be?
I j 9 MR. MATTHEWS: Yes. It's in several places but >
10 most pointedly it is in 50.47, and in Appendix E. Okay?
11 MR. REMICK: Appendix E does provide for it?
a 12 MR. MATTHEWS: Yes. It specifically dictates 1 -
13 the requirement associated with the extent of planning that ,
I 14 is required and establishes a' criteria that it~should be i
15 about 10 miles in radius for the plume exposure pathway and 16 about 50 miles for the ingestion pathway.
2 17 MR. REMICK: "About"?
i ,
18 MR. MATTHEWS: About. That's a criteria applied i 19 with considerations of topography, jurisdictional 20 boundaries, things specific to.the site, by the Staff, that 21 would dictate to be minor variations in other than a 10-mile i 22 radius circle.
)
23 MR. REMIC K: Does Appendix E itself address 24 possible exceptions? I can understand the "about" judgment 25 call; one or two miles is not "about" 10 miles, I don ' t
- ACE-FEDERAL REPORTERS, INC.
M-347-370) Nationwide Coverage 800'336-6M6
- - ~- .
, 28135.0 184 f-qRT
(_)
i 1 think.
2 MR. MATTHEWS: So that's right. In previous
! 3 cases "about" has been interpreted to the extent it's 4 necessary by hearing boards to mean plus or minus two miles.
5 Those'are the kind of variations that are usually seen. We l
, 6 have very few circles.
j 7 MR. REMICK: So presumably.it would be 50.47 --
[ 8 MR. MATTHEWS: That's the requirement. If it' 9 was an exemption it would have to meet the exemption 10 criteria in 50.12(a), which.was recently published in a 11 different form, which causes the quote to be made, "no !
! 12 undue risk to the public health and safety, common defense
,' 13 and security," and then there are six categories of special 14 circumstances listed by the Commission, one of which needs f 15 to be. met.
l 16 MR. REMICK: So if this came in as-a request, 17 certainly would it be that criteria that.would --
18 MR. MATTHEWS: Yes.
19 MR. MOELLER: The NUREG-0654 says that it need
]l j 20 not be a circle. Do the NRC regulations say that, also?
i 21 MR. MATTHEWS: Yes. !
l 22 MR. MOELLER: Okay. Go ahead. ;
{ 23 MR. ~LONG: There are some things we are focusing j i
24 on in the review of the conclusions in the study. First of '
25 all, we are looking at the WASH-1400 source term, " sensitivity i
ACE-FEDERAL REPORTERS, INC.
202 347-37a) Nationwide Coverage 80fk3W6M6 i
t 28135.0 185 RT 1 study" most-heavily. There was a policy conclusion.that we 1 -
2 should look there first, both from the sake of robustness,
'3 'and also because of the lack of a generic acceptance right 4 now of a different source term.
S- The focus for the criteria are on the early 6 fatalities, as opposed-to some of the other health effects, 7 and on the whole body doses when we are talking about 8 distributions of doses with distance.
9 (Slide.)
10 MR. MOELLER: Now, in terms of your doses, I
, 11 keep reading in NRC material that you may consider doses to 12 organs other'than the thyroid?
O 13 MR. LONG: In tetms of the review or in terms of 4
14 the regulations?
15 MR. MOELLER: I gather both.
i 16 MR. LONG: What I'm saying here is we have ,
17 regulations for doses to the thyroid, and I guess really
- 18 the doses deal with a critical _ organ which, depending on 19 what sort of radionuclide you are talking about and the -
{ 20 transport mechanism, it can be a lot of different organs in l 21 the body. That becomes a very complicated sort of analysis i
, 22 to determine that. You would have to go-all the way i 23 through to the back end, and then check.
24 In terms of looking at the results of this study, 25 we are trying to look at the back end as,it has been 1
- ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 804336-6M6 -
_ . . ~ . _
~
28135.0 l86 j (-qRT 1
1 presented and figure out what is likely to be most 2 sensitive, and then go look at those pieces first. It is 4
3- not saying that we have conclusively tracked down every i
4 detail, it's just at this point saying: This is how we are l 5 focusing our review on the trail pr'esented. ,
6 MR. MOELLER: Do you intend to use the effective 7 dose equivalent in setting the priorities, or what is 8 controlling?
9 MR. LONG: I'm not sure I understand the i
10 question in the context of what-I just said. I think we 11 are missing each other here.
12 MR. MOELLER: Okay. If you are looking at the 13 thyroid, the dose to the thyroid and the dose to the lung 14 ,
and the dose to the whole body, there is an established 15 mathematical formula for deciding how many rem to the 16 thyroid is equivalent to 1 rem to the whole body.
17 MR. LONG: I think I see what you are saying.
18 MR. MOELLER: This you call the effective dose 19 equivalent. For the thyroid it takes about 30 rem to the 20 thyroid to be equivalent to I rem to the whole body. ,
21 MR. LONG: I don't believe that was the way it 22 was done in the original CRAC calculations.
23 MR. MOELLER: Probably not.
24 MR. LONG: We are at this point just reviewing
- 25 what was presented to us by the company that was based on ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coserage 804 336-6646
l 28135.0 187 i
) I calculations with those codes and trying to back off from 2 that what looks like the most' sensitive contributors to 3 those doses.
l 4 MR. MOELLER: To digress for a moment, what is 5 the date cf the CRAC code which you people were using? You 6 know, roughly when was it? As of what date?
4 7 MR. WOODWARD: Keith Woodward, Pickard, Lowe &
8 Gerrick. It was first written, I believe, in about 1980.
9 It's an extensively modified version of CRAC, which .
10 includes variable trajectory, wind directions and 11 population evacuation.
12 MR. MOELLER: I guess what I'm driving at, it
( 13 does not, apparently, incorporate the newer health effects 14 model study that was done?
I 15 MR. WOODWARD: That's correct. The-one change 1
16 we did make was to go to a linear relationship for the dose 1
l 17 effects.
j 18 MR. MOELLER: Thank you.
19 Carson?
20 MR. MARK: The old way of treating this, I think ;
21 I'm right, CRAC, CRAC is an updated version of handling the i 22 same problem. In the old one it was only possible for 4
, 23 people to move on radii, I think. They insisted on going 24 downwind so that the cloud would keep up with them. j
/ 25 Does CRAC allow people to get the heck out of 1
1 ACE-FEDERAL REPORTERS, INC.
202-347-37tn Nationwide Coverage 8 m 336 # 46
28135.0 188
.RT 1 the way of the cloud?
d 2 MR. WOODWARD: Yes. CRAC actually models the 3 trajectory of the plume, which can change with time, and 4 the trajectories of people in their evacuation routes; so 5 if they are evacuating cross-plume, they'll only be in the 6 plume a very short time.
7 MR. MARK: That seems like an improvement.
8 MR. MOELLER: Go ahead, Steve.
9 MR. LONG: The point that I'm trying to get to 10 here is that as we look at the conclusions of the study, 11 there's really two areas that -- where we want to look for 12 risk contributions. One has to do with the probability of 5 .
i - 13 early fatalities, given evacuation. Here there appear to 14 be contributors from several different release categories, 15 at least the early containment failure and some of the 16 bypass sequences. And these release categories have 17 contributions fr.om several different initiating events and 18 sequences in the front end PRA.
19 Originally, in 1983, event V was dominant, so we 20 ,
also are going to want to focus on why it is no longer 21 dominant.
22 When you move to'the dose versus distance curves, 23 there is an assumption of no evacuation in the results that l
24 are presented in the study, although there are additional l
i
(^)
'x_ '
25 ;
t results available that show the effects of evacuation.
1 ACE-FEDERAL REPORTERS, INC.
202 347-3 Nil Nationwide Cmerage N O-33MM6
l i
28135.0 189 4
RT i
1 These are dominated by a. single release category that has j 2 seismic initiators. That's essentially a seismic-induced
- 3 station. blackout with a failure to isolate containment.
4 I should point out here that the names of the l 5 release categories are not too instructive in determining i
6 where the risk is coming from. This release category is ,
) 7 S-2, and it's called an early containment failure, but in l
8 fact, most of what has been put into this, 93 percent of it, 9 is a seismically induced set of failures which leave the 10 containment effectively open and the timing of the release 11 is determined by the timing of the core melt as opposed to f
- 12 the timing of containment failure.
l 13- MR. MARK: When you look at early fatalities --
4 14 and I presume you also look at delayed effects, health 15 effects, have you caught up yet, with BEIR-3 or are you
.s 16 still back at the WASH-1400 era?
4 17 MR. LONG: Again, one of the things I was i
18 presenting earlier was that right now we are not trying to l
19 scope the study review by looking at that distinction. I 20 believe you have gone to a threshold model for -- did I say 21 latent effects, by the way? Or related effects?
22 MR. MARK: . Dose /effect relationships have been 23 changed, have been modified by the National Academy, by the 24 NCRP, by the ICPR, on several points since WASH-1400.
25 Are you using a recent version of such proposed ACE-FEDERAL REPORTERS, INC.
202-147 3700 Nationwide Coverage 800-3364M6
28135.0 190 1 correlations?
l 2 MR. LONG: Again, from the Staff review l 3 perspective, an> are looking at what the Applicant has 4 presented for the early fatalities and using that as a 5 gauge. We are not looking at the'early health' effects or- !
6 the latent effects right now, late cancers or thyroid I 7 nodules.
8 To the extent that the company's work has'been i
! 9 updated -- I think they said they made a couple of j 10 corrections, changes to the dose / response curves. I'm not 11 sure if they are completely consistent.with BEIR-3 or not.
4 12 That would be something we could look at at a later time, f 13 but it's a detail at the moment for our focus.
1 I
, 14 MR. MARK: Thank you.
l 15 MR. MOELLER: To offer a comment, I would have -
16 said it more strongly than Dr. Mark. I would think, in
- 17 your review, that's one of the things you absolutely would j 18 be looking for, is to be sure that the most recent, 19 up-to-date, so far as the most accurate models and 20 quantitative relationships are used in every step of the 1
j 21 analysis.
22 MR. LONG: I should probably wait until the 23 people from Brookhaven get up here to talk about the actual l
i 1 24 sensitivity of moving f rom the WASH-1400 to the BEIR-3 j 1 .
j 25 methodology. Right now, though, we are looking for things i
1
- 1 l ACE-FEDERAL REPORTERS, INC.
20L3474700 Nationwide Cmerage 8 m 3 h 6646
28135.0 191 RT f
i
. 1 that look like they will give us an order of magnitude, or j 2 in some cases we are really looking for a 2 or 3 3 order-of-magnitude change in the conclusions, to see if the 4 conclusions are really challenged by the changes we would 5 make.
6 There is a very large list of things that we 7 could look at to see if the conclusions are really 8 precisely stated, and the review effort at this point has 9 not been expanded to include those things.
t 10 (Slide.)
i 11 So, to sort of summarize how we are~ approaching 12 it, the factors that are shown by the current study to i
O 13 really be important for risk mitigation are things that 14 we'd like to focus on here. And I'll get to that list in a 15 little more detail in just a moment.
16 Also, areas where there has been a significant 17 risk reduction between the conclusions in the original PSA 18 in 1983 and the conclusions that have been presented in the i 19 RMPs and the sensitivity study to that.
I 20 (Slide.)
- 21 Two of the plant features that really attract 22 attention are, first of all, the containment structure, 23 where there has been a statement about the actual failure 24 or bypass of the containment structure is only about a tenth j 25 of a percent of the core melt frequency and it's this tenth 1
! ace-FEDERAL REPORTERS, INC. i
{ 202 347-3700 Nationwide Coserage NG336-6646
28135.0 192 RT 1 of a percent that's really contributing the risk to the 2 population.
3 We really need to be sure that the containment 4 structural integrity is great enough to really justify that 5 kind of conclusion. And, also, that the system's bypass 6 potential for that potential, either with interfacing 7 systems LOCA or bypass of the atmosphere inside the 8 containment, out of the containment, are really as unlikely 9 as modeled.
10 MR. MOELLER: Is it not the Staff that 11 commissioned the Brookhaven review of the containment 12 calculations?
\- 13 MR. LONG: That's right.
14 MR. MOELLER: Are you beginning to get some 15 answers to your questions?
16 MR. LONG: We are beginning to get some answers.
17 I would like to let -- for the containment -- Charlie 18 3 Hofmayer tell you just how far we've gotten.
l 19 l MR. MOELLER: Okay.
l 20 ! MR. LONG: By the way, the containment enclosure, f
21 although it would contribute to protection of the public in 22 minor release accidents, really looks like it won't give i
23 l you much credit when you are talking about large 24 deformations of the containment. So in our initial review
(~x., 25 here we are ignoring the containment enclosure as having e i uJ ,
ACE-FEDERAL REPORTERS, INC.
, mm.m, m- m-, ->_ j
28135.0 193 RT 1 much effect.
2 The other factor that is attracting our 3 attention is the nature of the RilR vault. As you heard the 4 company this morning say, it's a very deep, narrow vault l
5 and they've taken credit in a couple of ways for it becoming l 6 flooded and reducing the source term. So we are very 7 interested in the piping configuration in and out of the 8 vault, the strengths of the piping and so on, to see if 9 this feature is of significant -- is of as significant a 10 benefit as claimed.
11 (Slide.)
12 Coincident with the structures are the k/
13 assumptions in the modeling. As I said, the interfacing 14 systems LOCA is very important in the original study in 15 1983. It is still important from the standpoint of the 16 risk of early fatality but it's way down in the noise from 17 the standpoint of exceeding a certain dose at a certain 18 distance.
19 The completeness of the interf acing systems LOCA 20 scenarios is something that we have been looking at.
21 Something that Dr. Corradini brought up this 22 morning of the potential degradation of the steam generator i 23 , tubes is something we identified. The potential for hot l
24 fluids going through a penetration in some sort of blowdown 25 outside containment, perhaps failing a penetration, is ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Cos erage 8(n33MM6
i ,
28135.0 194 RT 1 another area.
2 "We have asked the' company for additional
~
3 information on the probabilities of additional plumbing 4 configurations contributing to the probability of LOCA 5 outside containment and we also started checking into the 6 check valve failure data and the conclusions they have 7 drawn from that.
8 In particular, they have reduced the probability 9 of having this type of failure substantially by looking at 10 the probability of the ch'eck valves.failing. They have 11 also reduced consequences by inducing some scrubbing terms 3
12 in their effluents, given an-RHR vault LOCA.
4 13 As you heard them say, the most probable 14 location of failure, they believe, is the RHR pump seals.
15 They feel the vault would flood fairly quickly once that i
16 occurred, and by the time of a significant release of s
i 17 fission products in the core it would be a scrubbed release, 18 bubbling up through the vault. We want to check that the t
4 19 credit taken for the scrubbing ~is a ~ reasonable level of i
20 j credit for the isotopes that are scrubbed.
21 The company is also taking credit for operator 22 recovery actions, in event V, and we have considered the 23 potential of misleading the operator, especially in the 24 scenario they mentioned this morning where the RHR valve f 25 relief inside containment brings the operator out of his
?
i ACE-FEDERAL REPORTERS, INC.
202-347-370): Nationwide roserage 8% 3% MA6
28135.0 195 RT 1 initial emergency operating procedures into the LOCA 2 procedure and doesn't draw his attention back to the fact 3 that the LOCA may also be occurring in the RHR vault.
4 We have the procedures and we will be looking at 5 the cues available to the operator, if such an event occurs.
6 There are also recovery credits taken for the 7 station blackout station sequences. We have been looking 1
8 at these areas as well, because there's a lot of time 9- claimed in the strength of the containment for recovery to 10 be available, recovery to be able to occur.
11 In addition the seismic initiators are, as I 12 s a id ,' very significant contributors to the-final risk I
13 values and the seismic fragilities of the really 14 risk-sensitive components are an area where we intend to do 15 a little bit more looking.
16 There has been another area. raised that we are --
17 we have some information on but we are not really' clear on 18 how to deal with it yet and that has to do with the risk of 19 off-normal operations, the risk from nonpower operations; 20 that is, risk during cold shutdown when the containment may 21 either be only partial.ly isolated or the isolation may be 22 violated completely. l 23 I think that covers the things I wanted to get !
24 across'for why we are focusing in the areas you are about I 25 to hear about in more detail, l
a l
ACE-FEDERAL REPORTERS, INC. '
202-347-3700 Nationwide Coserage 800-33M646
28135.0 196 RT 1 If there are any questions as to that aspect of 2 it I would like to address them now. If there are 3 questions on the details of the review I would like to 4 leave them for the people from Brookhaven to speak after 5 I&E.
i 6 MR. MOELLER: Questions?
7 There apparently are none, so let's move ahead.
8 I would like to turn this over, then, to Dave
' Matthews, who will talk about a comparison of the company's 9
10 studies with NUREG-0396.
11 MR. MATTHEWS: Good afternoon. As I' mentioned 12 earlier, I'm Dave Matthews from the office of inspection 13 and enforcement.
l 14 I would like to start off by expanding just l 15 slightly on my response to Mr. Remick a little bit earlier.
16 I may have spoken a little too quickly insofar as 50.12(a) 17 being the only mechanism by which some change or relief to 18 the regulations might be granted. There's another, 19 mechanism in Part 2 for the utility to make a case 20 associated with special circumstances existing that.might 21 warrant not applying a regulation in a specific instance.
22 That type of petition could be made, as I understand-it, 23 through the hearing process and'could eventually evolve to 24 getting to the Commission for' consideration.
r~ 25 MR. REMICK: Would you happen to know the
(_g/
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80 4 336-6646
28135.0 197 RT 1 section of Part 2?
2 MR. MATTHEWS: No. Bob Perlis from the office 3 of the general counsel may know this.
4 MR. PERLIS: Section 2.758.
5 MR. REMICK: Thank you.
6 MR. MATTHEWS: For purposes of review --
7 and it may be that for many of you -- in terms of the l 8 technical basis for portions of the emergency planning 1
i 9 regulation, particularly with regard to the size of the 10 emergency planning zone, it was determined based on 11 considerations in'a document described as NUREG-0396, which 12 was a joint FEMA -- excuse me,.NRC/ EPA document, I think, i
O 13 in the 1978 time frame. '
14 That rationale that exists in NUREG-0396 was 15 also more succinctly stated in NUREG-0654, which is.another 16 document familiar to those in the emergency preparedness 17 field, which is the joint FEMA and NRC criteria for review 18 of state and local and utility emergency plans. It, 19 NUREG-0654, contains guidance to state and local 1
20 governments and utilities with regard to-the scope and 21 nature of accidents that should be planned for.
l 22 The rationale with regard to the size of the 23 emergency planning zone -- and I'm now talking the plume 24 exposure emergency planning zone -- was stated in 1
3 25 NUREG-0654, and I have tried to make it a more succinct 1
~s/
l 1
l ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6646 1
_ ~ _ . _. _ _. _ __ . _ _ ___ _ __
28135.0 198 j RT L 1 statement here.
2 (Slide.)
3 NUREG-0654 wording is a little more expansive,
- 4 but basically it says that the projected doses from the 5 traditional design-basis accidents would not exceed 6 protective action guide levels -- these are the EPA 4 7 protection guide levels -- outside a 10-mile zone; the 8 projected doses from most core melt sequences would not
~
, 9 exceed protective action guide levels outside the zone; and i
10 for the worst core melt sequences, immediate
- 11 life-threatening doses would generally occur -- not occur, 12 excuse me, outside the zone.
1.3 Furthermore, it indicated that one of the V
14 considerations in establishing an EPZ of a specific size 15 was that detailed planning within 10 miles would provide a l 16 substantial base for expansion of response efforts in the
- 17 event that this proved.necessary. And that basically says 18 that if you were to get an accident.more severe than formed 19 some of the foundation for a 10-mile EPZ, which we don't 20 exclude'the possibility of, that the response efforts that 21 would be planned for in the 10-mile EPZ could be expanded 22 to encompass additional protective actions outside this 23 zone.
24 The second and third-of these criteria were l i 25 displayed graphically in NUREG-0396.
, ACE-FEDERAL REPORTERS, INC.-
202-347-3700 _ NationwideCoverage 80 4 336-6646
28135.0 199 1 (Slide.)
2 In this form. This data, in what's known in the 3 industry and among the emergency planning personnel, is.
4 referred to as figure 1-11, and it displays whole-body i
S doses versus distance, given a core melt. And it 6 graphically illustrates, again, the second and third 7 factors associated with the criteria NUREG-0396 and 0654.
8 This data extracted to 0396 is based upon WASH-1400 3
9 categories PWR-1 and PWR-7. Those sequences that did 10 result from a core melt.
11 Again, this particular data is derived from the 12 Surry plant, which is one of the plants done in WASH-1400.
13 MR. CORRADINI: Could I ask a question? I asked 14 it of the Seabrook people and I'll ask it here.on that 15 figure, if you can put it back up, the one you just had.
16 That's, given a core melt and buried'in here, in the 17 Seabrook calculations was both the release category 18 probabilities for the source term as.well as the 19 meteorological statistics. Is that also what's buried on 20 the ordinate here?
21 If I look here and go from, normalize from l'to 22 10 to the minus 30, is.it the release categories and the 23 probabilities of those release categories. coupled with the 24 statistics of the meteorology that give me those 25 probabilities?
O ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage . 80 4 336-6646
28135.0 200 1 RT I
1 MR. MATTHEWS: Yes. Yes. Presuming you had a
. 2 core melt accident.
j 3 MR. CORRADINI: Right. So this curve is for 4 Surry at Surry meteorological conditions or is it averaged
- 5 over 100 sites?
1 6 MR. MATTHEWS: That's a good question. I can't 7 answer. The person most knowledgeable to answer it isn't 8 here today. My recollection is that it was representative 9 of average meteorolocy.
10 MR. CORRADINI: But with the Surry release 11 categories?
12 MR. MATTHEWS: Yes. ,
13 MR. CORRADINI: And those average numbers?
1
- 14 MR. MATTHEWS: Yes.
15 MR. REMICK: Averaged meteorology or averaged 16 weighted meteorology for the Surry site?
17 MR. MATTHEWS: My recollection is it was not 18 weighted for the Surry site. It was.some. aggregate 19 meteorology representative for the United States.
20 MR. REMICK: I see,. assuming some average 21 meteorology, not taking a variety of meteorological 22 conditions in the accidents and weighting with the 23 probability.of the meteorology of the Surry site? You 24 don't think it was the latter?
1 25 MR. .MATTHEWS: Can't be certain on that.
L ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
28135.0 201 RT 1 Yes, Dr. Seiss?
2 MR. SEISS: I may be asking the same question 3 Corradini asked because I didn't understand the question --
4 or the answer -- or both.
5 These are conditional probabilities of doses at 6 those levels at those distances, given a core melt?
7 MR. MATTHEWS: That's right.
8 MR. SEISS: It's not given a core melt that 9 somehow represents all the possible core melts with their 10 probabilities and the kinds of releases that go with them?
11 MR. MATTHEWS: And Monte Carlo'd with the 12 meteorology. That's my understanding.
i
~
13 MR. SEISS: And if I have a plant that's 14 different from Surry and one type of core melt is much more 15 probable, could I still use this?
16 p MR. MATTHEWS: Now I don't understand your 17 question.
18 MR. SEISS: All types of core melt scenarios are 19 not equally probable.
20 MR. MATTHEWS: If you used a plant different 21 from Surry and did a similar analysis you would come up 22 with potentially a different-looking set of curves.
23 MR. SEISS: If I did a similar analysis?
24 MR. MOELLER: See, Carl Fleming wants to speak --
~'
25 MR. SEISS: This was in a NUREG that governed
\_/
ACE-FEDERAL REPORTERS, INC.
202447-3700 Nationwide Coserage 80}336-6M6
i 28135.0 202 RT 1 all the plants, isn't it?
2 MR. MATTHEWS: That's exactly right.
3 MR. SEISS: So Surry conditions were used for 4 everybody?
5 MR. MATTHEWS: That's exactly right.
6 MR. SEISS: Maybe you can explain it better.
7 MR. FLEMING: I was just going to reply to your 8 -question in the affirmative, in that if you went to' a 9 different plant that had a different relative frequency of' 10 different types of releases, that difference would be 11 reflected in these curves. However, what you lose, what's 12 missing is that the results have different core melt O 13 frequencies. This is all formalized against core melt, so-14 if there's drastically different core melt frequencies, you 15 are not getting that difference.
16 MR. SEISS: But a different mix?
17 MR. FLEMING: If you have a different mix, you .
18 will get that.
19 MR. REMICK: Isn't that what our whole 20 discussion is about today, if you take 10-mile arrived at 21 by this method, there are other-ways you might approach the 22 same problem? That's basically what we are. talking about.
23 MR. MATTHEWS: That's right.
24 Just in summary with regard to those curves.
25 (Slide.)
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage - 8043 % %86 3
28135.0 203 RT 1 Narratively what they tell us is there was about 2 a 30 percent chance'of exceeding a PAG dose, if you were an 3 individual at 10 miles.
4 MR. CORRADINI: What's the PAG dose? 5 rem?
'S MR. MATTHEWS: The lower level PAG dose is 1 rem 6 whole body and the upper level is 5.
7 MR. CORRADINI: Wow.- Okay.
8 MR. MATTHEWS: And the range is one that EPA 9 dictates as 1 being the onset of a consideration should be 10 given to protective actions and 5 being one that would 11 indicate, if you had a projected dose of.that value, you 12 certainly ought to take the protective _ actions.
13 MR. SEISS: Which one does the 30 percent refer 14 to, the 1 or the 5?
15 MR. MATTHEWS: I think you'll see if you look at 16 curve, it's probably so similar that it's. indistinguishable.
17 (Slide.)
18 MR. SEISS: It would be "about," then.
19 MR. MATTHEWS: Yes.
20 MR. SElSS: Okay.
21 MR. MATTHEWS: With regard to criteria for 4
22 beginning to review the technical aspects of the public 23 service PSA update, since the WASH-1400-risk portrayal of 24' the Surry plant was probsoly the major factor in the-25 selection of the size of the plume exposure pathway, EPZ, ACE-FEDERAL 202-347-3700 REPORTERS) INC.
Nationwide Coverage 800 336-6646
l l
i 4
28135.0 204 O
1 - . (E)RT I the preliminary evaluation of this sensitivity study at the 2 moment is directed towards an assessment of the probability I 3 of off-site doses at given distances, gi~ven a core melt at 4 Seabrook.
5 (Slide.)
6' Consequently, we are probably-looking for a ful1 7 range of doses versus distance, analogous to those shown in 8 the figure 1-11 for the Seabrook plant. This. requires an 9 evaluation of Seabrook-specific information relating to the 10 probabilities of severe accident sequences, and as Steve
- 11 earlier said containment-failure medes and' probabilities, 12 the release characteristics and the dose calculations.
13 When we~are able to confirm the validity of that
~
14 assessment we'll be able to make a judgment on the relative 15 core melt risks between the'Seabrook plant and this figure 16 1-11, which is representative of, again, the primary 17 technical basis for the EPZ size.
18 That was all I had in terms of prepared remarks.
l 19 MR.-MOELLER: Any questions for.Mr. Matthews?
20 Then do you immediately go into'the next topic?
21 MR. PRATT: What I would like to do is give-you 22 , a orief status of the BNL review of the Seabrook station 1
23 planning sensitivity study.
24 The material, as Steve said, will be presented 25 basically-in three parts.
C, ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwkie Coverage 80 4 336-6646
28135.0 205 RT 1 (Slide.)
2 What I would first like to do is go over with 3 you the risk perspective.
4 (Slide.)
5 In doing this, what I would like to do is 6 concentrate on the curves that you have all been asking 7 questions about already, the dose versus distance curves, 8 and try to go through rather slowly where the numbers are 9 coming from; how do we get these calculations.
10 When I put the presentation together I wasn't 11 sure just what would be presented by other people and how 12 much detail there would be. What I tend to do when I first 13 have to review one of these pieces of work is understand 14 where things are coming from and where all of the building 15 blocks are. So I will.be going through that fairly: slowly.
16 I hope'that I won't be repeating too much what has already 17 been said, but I think it is very.important that we get a 18 very clear understanding of where these numbers are coming 19 from and what they imply so that we can then focus our 20 review.
21 MR. MOELLER: How long have you been on this and 22 how many people are involved?
23 MR. PRATT: This particular study, for about-24 three weeks or so. It's about a half a dozen individuals 25 involved.at Brookhaven. j
! ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nuionwide Coverage 800-336-6646
28135.0 206 RT 1 MR. MOELLER: It's just barely started?
2 MR. PRATT: It's barely starting'and we are 3 hoping to -make some preliminary conclusions next month. So 4 we are moving on a very tight schedule and that's really 5 why I want to spend some time with you about why we-are 6 focusing where we are. We cannot, obviously, cover ,
7 everything in that sort of time frame, and the aim is to 8- really focus our attention.
9 Steve has already given you a good introduction 10 to this and focused us in those areas that we think are 11 important. Obviously containment structural integrity --
12 it's what you have been hearing all day -- is very crucial
('
13 to the Seabrook study, and Charlie Hofmayer from Brookhaven 14 will be looking at that in detail in his division at 15 Brookhaven. He'll give you a brief status of where we are 16 there.
17 MR. MOELLER: That part, we were led to believe, 18 the containment structural study,.had been going on for 19 quite some time. Was I miss --
20 MR. PRATT: We'll cover that in'the next Vugraph.
21 I think we need to clear up what we are doing there.
22 MR. MOELLER: When you said three weeks, you J 23 meant the overall review?
24 MR. PRATT: This very latest review, the review 25 of the most previous submittal. Let me put up the next ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33M486
28135.0 207 RT 1 Vugraph.
2 MR. NOONAN: Dr. Moeller, you know this 3 particular study we received was submitted on the dockets 4 on the 21st of July. What we have done is contract with i
5 Brookhaven, approximately three weeks ago, to do a certain 6 scope of work for us for this test.
7 Some of the structural stuff there that was 8 talked about was previous Brookhaven work done -- done a 4 9 number of years ago. Dr. Pratt will go through that a 10 little bit.
11 MR. MOELLER: That's helpful. I wanted to 12 understand. Thank you.
13 MR. PRATT: Let me to go this Vugraph.
14 (Slide.)
15 I.think that will put the whole thing in 16 perspective.
17 The original PRA was submitted in 1983 and 18 basically two reviews occurred by different contractors.
19 The front end -- and I mean by that the assessment of the 20 core damage frequency -- was done at Lawrence Livermore.
21 The back end review was done at Brookhaven. At that time it 22 was a relatively limited review. It'was about eight or l
23 nine staff months of effort.
24 As part of our evaluation, here, we certainly 25 did not reevaluate the structural compatibility of the Vgs ACE-FEDERAL . REPORTERS, INC.
202-347-3700 Nationwide Coverage 804336-6M6
]
l O l 28135.0 208 1 1 Seabrook containment building. There was no 2 structural-mechanical people involved in that assessment.
3 We were PRA people, looking at the back end to see if there 4 were things that were important to risk.
5 So I would certainly not characterize this 6 document in any way as doing a review, an in-depth review 7 of the structural capability of the building from a 8 structural-mechanical point of view. That will be 9 performed as part of the latest assessment by Charlie 10 Hofmayer's division. Okay?
4 11 That's basically where we are. Then we had this 12 -additional submittal f rom Seabrook, which was a' risk 13 management and emergency planning study. You have heard i
14 the details of that this morning and how that differed from 15 the latest submittal.
16 This is the focus specifically of what we are 17 doing at this time. We will.be using, as much as possible, 18 the information that was generated earlier in our reviews 19 of the front end and back end of the original PRA. .
20 The objectives are to focus in on this review 21 and, as you have heard from the previous presentation, a 22 very important element of that study was the comparison of' 23 the dose versus distance curves with.these in 396. What I 24 intend to do is walk through how those curves were 25 generated in great detail and try to focus in on those o(~s ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80433M686
a 28135.0 209 SRT.
I things that contributed to it and, therefore, why we have 2 to focus our review.
4 3 Basically the codes were based in large part on 4 the PRA, the original PRA, with modifications and updates f
] 5 in PLG-0432, and the important thing is they went back to 6 WASH-1400~ source terms to calculate the risk.
7 What I'm going to be attempting to do --
8 (Slide.)
9 You have seen this so many times you must all be 10 sick of it -- is try to explain where this curve is coming 11 from in comparison to this curve in WASH-1400, and how does 12 one construct this type of graphic.
13 (Slide.)
14 This gives you, in words, the process that was 15 gone through. You virtually calculate the dose versus 16 distance curves for each of.the release categories-that you 17 identified. You then take the probability of exceeding 18 this dose, multiply by the release frequency appropriate to 19 it, and sum it to give you the probability and divide the 20 whole thing by the probability of the total . core meltdown .
21 frequency.
22 "R. CORRADINT: Can I ask a question? Given the 23 probability of core melt is 1 normalized here, the-24 surumation of all the release categor'y frequencies also has
. 25 to add up to l? Using WASH-1400 release probability, ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 2 336 6646
, _ . _ _ .- , . _ __ .-.-_,,.s. _ _ , . - , ,_ ..
f' 28135.0 210
. 1 because there it became the probability of 1.
2 MR. PRATT: Using everything. There may be 3 release categories that there's no containment failure; <
4 therefore there's no risk.
5 MR. CORRADINI: I see. All right.
6 MR. PRATT: In other words, what you are really 7 seeing here is, you.are seeing on this ordinate the
, 8 strength of the containment building. This is basically 9 saying that in NASH-1400, if I assume that the only way I i
] 10 was. going to get a probability of getting 200 rem dose was j 11 by very large early containment failure, then, in WASH-1400 1
12 days,' I had 1 chance in 10 of this event occurring. Okay?
O 13 So, it gives you a measure of the strength of 14 the containment. It's rather-like.a safety belt. If you.
i 15 have a measure of your safety-containment, .l', given a core 16 melt frequency (indicating).
i 17 Now, if you look here in the analysis done .in 18 WASH-1400 it represents 5 times 10 to the minus 5. What 19 this point here represents in Seabrook-is 2.23 times 10 to 20 the minus 5 -- 4. I'm sorry. Times 10 to the minus 4.
21 MR. MOELLER: Go ahead.
22 MR. TORRI: That's the median value of the mix 23 between PWR and BWR?- The mix is 89.98 times 10 to the 24 minus 5 or 10 to the minus 4, so it's not quite as i 25 disparate.
ACE-FEDERAL REPORTERS, INC.
202-347-3XU Nationwide Coverage 804336-6M6
- - , _ , . - . . , _ ~ , -
l 28135.0 211
-1 MR. PRATT: This is the reactor safety study, 1
2 which was a point made half an hour ago. I don't want to 3 get into differences of 10 here but people should be aware i
4 of that. Okay?
5 You have seen this before. Let me summarize it 6 again.
7 (Slide.)
8 What we are talking about is~seven release 9 categories that were earlier identified: Early containment 10 failure, which was categorized much like the WASH-1400 PWR 11 release category 2, as I recall, the way it was done at 12 Seabrook.
13 Then we have an S-2, which is an early increased 14 containment leakage. This represents something like the 15 type A leakage that Dr. Torri identified earlier. Here we 16 are talking about increasing the leakage'from about .1 17 percent a day to about 40 percent a. day leakage.
18 Late overpressure failure is very late in this 19 case.
20 Base mat melt-through, this was really'not 21 quantified as being very low, I believe, in. consequences.
22 Containment intact -- again, relatively low.
23 This is design-basis-leakage.
24 Containment not isolated, this is more like the
, fg 25 C-type release category, and includes in this particular l (m./ l ACE-FEDERAL REPORTERS, INC. l 202-347-3700 Nationwide Coscrage 800-336-6646 !
28135.0 212
,-7RT i
%-)
1 case, it was analyzed as a purge valve open, I believe, an 2 8 square inch diameter opening.
3 You'll see the differences in the conditional 4 consequences for this particular leakage versus that one in 5 a minute.
6 And then this is the containment bypass, the 7 V-sequence. Here we are talking about about the V-sequence 8 submerged, so there's particular scrubbing of that 9 particular release category.
10 (Slide.)
11 These are all taken from the document that we 12 are reviewing. These give you a summary of the release
'l 13 categories. And, as I said before, this is the severe one.
14 This is basically what you would expect if you looked at PWR 15 release category 2 in WASH-1400, and you can see how the 16 others change.
17 What is very significant, if you look at the 18 release category S-2 and the release category S-6, if you 19 recall, these were the leakage characteristic failures of 20 the containment building. The S-2 is a low leakage. The S-6 21 lisamoresignificant leakage.
22 It gives you the time-dependent releases. These 23 times are important for this particular calculation because 24 we were -- the critical time that goes into this is 24 25 hours. Okay? So we are integrating a dose to the
{d}
u ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 800-336 % 46
28135.0 213 BRT (vT 1
1 individuals without any actions being taken for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
2 So, for some of these leakage calculations, 3 indeed, obviously if it's leaking beyond that time, it is 4 not --
5 MR. CORRADINI: So this-is WASH-1400 series 6 values?
- 7 MR. PRATT: This is straight from WASH-1400.
8 It's.the type of thing we were doing at Brookhaven for many 9 years. We were asked to put ourselves in the 1970 time j 10 frame and reanalyze the source terms assuming the 11 plant-specific data of the plant we were reviewing. So we
- 12 did it for Indian Point, for Zion, for Limerick and so on.
15 We took basically the characteristic releases for WASH-1400 14 and applied them to the CORAL code and MARCH code and 15 calculated releases and this is exactly what they did here.
16 MR. CORRADINI: I don't think I catch it. I 17 thought.it was simpler than that. You said you went to 18 WASH-1400 and picked out the release categories.
4 19 MR. PRATT: In this one particular case. This 20 is a large early release that represents something like an 21 interfacing system -- whatever. That is a PWR release 22 category 2.
23 If, for example, I have another release such as 24 this one, where I open up a small hole and I go from .1 25 percent per d.ay leakage to 40 percent per day leakage, what
.0 ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80 4 336-6646
j 28135.0 214 SRT s-)
1 I must then do is I must go back and do my analysis, 2 wherein I run the MARCH code to give me the thermohydraulic 3 conditions in the containment building, calculate the 4 leakage out. I then input to the CORAL code that 5 thermohydraulic data and input the radionuclide release
]
6 groups as taken from WASH-1400. There are tables as to 7 when they come out. I run that calculation and get this 8 answer.
9 Correct me if that's not the process that you 10 did.
11 MR. TORRI: It's actually a PWR-1, which we 12 calculated using MARCH and CORAL for Seabrook, but because
' O 13 the containment isn't there, you get the same answer.
14 MR. PRATT: Very similar.
15 MR. CORRADINI: I missed that._ Where do you 16 start? You start with the release ratchet out of the-17 vessel and redo the containment thermohydraulics to get the 18 release out of the containment?
19 MR. PRATT: If you look in WASH-1400 there'are 20 tables that give you the release from various stages of the 21 accident sequence, the gas release, the melt release, and 22 the vaporization release.- And_they give you prescribed 23 times during which those are released to tho containment. .
- 24. You then must go and do your thermohydraulic 25 analysis on the containment, open up the hole that you want ACE-FEDERAL REPORTERS, INC.
202-347-3700 - Nationwide Coverage 80433M646 '
28135.0 215 RT
~
1 in your containment, and calculate the release, using MARCH 2 and CORAL. But the emphasis there is to use CORAL for your 4
3 deposition of the aerosol particles rather than something I
4 like NARO which gives you very much increase aerosol 5 conglomeration and settling.
6 The reason you are getting such high releases 7 here at such late distances is because you are using CORAL, 8 which has a much slower deposition velocity than if you 9 used-the other. So you would not-get characteristics of 10 this nature using the new source term methods. is that 11 clear now, Mike?
12 MR. CORRADINI: Yes.
O V 13 MR. PRATT: So, having those releases, then what 14 you do is do calculations of this nature. Again, this is 15 taken straight from the submittal. What-this shows on a 16 conditional basis,-assuming you have this rather bad 17 release, this S-1 release -- and we are calling it a W here 1
18 because it.is supposed to be WASH-1400 rather than the 4
19 original S-1 release, which is what the Applicant believes 20 to be the true value. This may be a little difficult to
'21
- read. -
22 Principally this is a 200 rem dose here coming 23 down. This is conditional upon the accident occurring. So 24 we have 1 here, and this shows the 200 rem dose against.
I 25 distance. This is -- the reason why, if you go back-to ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6 4 6
~
28135.0 216 RT 1 something like WASH-1400.
2 (Slide.).
3 You'll see curves of this nature that curve out 4 and' drop down at the.15-mile range. What that is coming 5 from is all of those accident sequences, if you like, that 6 in WASH-1400, that were calculated to arrive in release 7 categories like this one. Okay?
8 So it's really coming from large early' releases 9 in~which the characteristic curve looks like this-and comes 10 down at about the 15-mile mark. This is important.
11 Really what you are doing is you are taking each 12 of these calculations, multiplying it by the' probability, 13 adding the whole thing-together and then dividing by the 14 total core melt frequency. Okay? So that's why you get 15 increase in the curve that you asked about later where the 16' relative frequency of these various events trip in.
17 For example, thic is the S-6.
18 (Slide.)
19 As you recall, the other one was coming in in 20 this sort of range. This one, which is a leakage, a rather ;
21 large leakage, comes in at around about just-below the 22 mile mark.
23 MR. CORRADINI: Could I ask a question, then?
24 The procedure you just stated, then, normalize it based on 25 the average core melt frequency. -What if I take each ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-33MM6
28135.0 217 RT 1 individual release category and divide by the probability 2 of getting to core melt with that sequence? Will I get a 3 different number?
4 MR. PRATT: I don't understand.
5 MR. CORRADINI: Let me repeat. You said you 6 take the release categories frequency, the probability --
7 MR. PRATT: This is conditional upon this 8 occurring. Let me just go a couple of Vugraphs ahead.
9 MR. CORRADINI: I'm sorry. I'll just wait until 10 you are there.
11 MR. PRATT: This is the frequency that was 12 calculated for each of these releases.
i
'~' 13 (Slide.)
14 Okay? So, I would take the conditional value 15 here for the S-6. Nobody can see a thing, there. -- for 16 the S-6. All right? Multiply each of the -- that curve, 17 by the frequency of the S-6, which is 3.2 times 10 to the 18 minus 7.
19 (Slide.)
20 , Okay? I would then add that contribution to 21 each of the other ones, for each of these values, whatever i
22 the curve might be. Then I divide that by the whole tning 23 added up, which is about a 2.3 times 10 to the minus 4.
24 MR. CORRADINI: Okay. I'm with you. But then c' 25 you would normalize it to give it a core melt? You'd mj ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
t 28135.0 218 RT 1 divide by the average core melt frequency?
2 MR. PRATT: 2.3 times 10 to the minus 4; right.
3 MR. CORRADINI: It's a minor point, but I'm a 4 little confused about it. Why don't I simply ~just take out 5 the core melt frequency for each individual sequenco at the 6 beginning and just have the probability for containment 7 failure by the sequence instead of dividing by the average 8 at the end? I could possibily get two different numbers.
9 MR. PRATT: I guess I'm really not following you,
- ~
4 10 Mike, at all.
J 11 MR. CORRADINI: I'll wait and ask --
12 MR. PRATT: You may not get much more, so if you 13 want to keep me going I'm here for you to try to understand 14 it.
j 15 MR. CORRADINI: Let's take two examples, S-1 and 16 S-2, all right? This is the probability of the frequency e
l 17 is some number, 5.2 times 10 to the minus'9.
1 18 MR. PRATT: Right.
19 MR. CORRADINI: Then I take that times the 20 ' probability of exceeding a given dose. j 1
21 MR. PRATT: Yes. j 22 MR. CORRADINI: Sum it up with S-2, get a total, 23 and divide it by the total of the core melt frequency.
24 MR. PRATT: Yes.
25 MR. CORRADINI: Wouldn't I get a different 1
l i
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80433 4 646
28135.0 219
~BRT 1 number if I started off by taking the S-1, dividing by the
- 2 core melt-frequency for that point -- okay? So all I get 3 at this point is the probability of containment failure 4 given a core melt for S-l?
5 MR. PRATT: Yes.
6 MR. CORRADINI: And then multiply that by the 7 probability of exceeding the dose plus S-2?
, 8 MR. PRATT: Well, let me do it.
4 9 MR. CORRADINI: It's not clear to me the numbers 10 are going to be exactly the same.
11 MR. PRATT: I think they have to be. Maybe I 12 can write this on here.
13 (Slide.)
14 This is for Seabrook, if I do it down this side.
15 This number is 2.3 times 10 to the minus 5; all right?
16 This is 2.3 times 10 t.o the minus 6 -- minus.7 --
17 minus 8 -- 4, I'm sorry, 4, 5, 6, 7. (Writing.)
18 So we have 2.3 times 10 to the minus 7, in terms 19 of an absolute frequency scale down the side here.
20 So that the way you can find out.what is 21 contributing here to this calculation is you look at the 22 release characteristics that you've got and the frequencies
- 23 that you have here. Okay? I can see that these releases,-
24 which are the only ones that have any sort of decent t
25 probability, do not have the probability of exceeding 200 ACE-FEDERAL REPORTERS, INC. ,
202 347-3700 ' Nationwide Coverage 80 4 336-6646
28135.0 220 RT i
1 l rem.
2 So this is something times nothing, which gives 3 me nothing.
4 MR. CORRADINI: Okay.
5 MR. PRATT: So where this is coming from -- it's 6 all really coming from the S-2 release which has a 7 frequency of 2 times 10 to the minus 5.
8 (Slide.)
9 That's conditioned -- this is what it.looks like 10 on a conditional basis.
11 MR. CORRADINI: I see what you are saying.
12 MR. PRATT: No. That's S-6.
13 MR. CORRADINI: That's okay. Why don't you go 14 on.
15 MR. PRATT: No. This is really all I'm going 16 through. The S-2 looks like this. This is the one that 17 drops off at the two miles. Okay?
18 (Slide.)
19 Basically if you look at this curve what you are 20 really seeing here in this area is all coming from the S-2 4
21 release. Okay? Because you've got a 2 times 10 to the ,
22 minus 5 frequency. And those are the characteristics. But i
23 if you were to plot, of course, additional cycles on your 24 paper down in here, you'd see the other ones sitting in at i
- 25 much lower probabilities. .Okay? So that's an important.
ACE-FEDERAL REPORTERS, INC.
202-347-370) Nationwide Coverage 804336M46
28135.0 221 RT 1 point. We are not seeing these things disappear; they are 2 just off the graph and very low in probability.
3 I think this is a very important point because 4 what we are really saying is, when we look at something 5 like this, in order for this representation ~to be correct, 6 then those types of accidents that would result in large 7 dose to large distances have to be lower than this 8 particular value, which is 2.3 times 10 to the minus 7.
9 That's an important point, something that really was the l
l 10 purpose of my walking through this so that you all grasp 11 what we are saying. We are actually going to have to say
~
12 that when we get to the-bottom line on all of this, that b
k' 13 that is a reasonable representation.
14 So the main point here is that release category 15 S-2, in terms of the presentation that I showed you for the 16 dose versus distance curve, is dominated by the S-2 release 17 category.
18 Again, this has been mentioned by Steve. This 19 is dominated by seismically initiated events which result 20 in this type A-type leakage failure. Okay? It's a small 21 leakage rather than the larger leakag'e.
22 If it were the larger leakage then we would not 23 have gotten the same characteristics. The characteristics 24 would have gone further to further distances.
l 25 So, it's obviously important for us to know how ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8 @ 336-6616
28135.0 222 RT 1 the containment building performed under these 2 circumstances. We need to know whether'it's a small
-3 leakage or large leakage because it would significantly 4 affect the characteristics of the calculated off-site dose.
5 (Slide.)
6 This I have already said. Basically what the 7 curve is telling us is that all accident sequences that 8 might result in release categories S-1 and S-6 are low,. low 9 in frequency. That's the main strength of the Applicant's 10 statements that these events are, indeed, low. Okay?
11 So, in order for us to look at this thing we 12 have to come up with the basic conclusion that induced 13 early containment failure is very low in frequency. That 14 bypass of the containment function, either.as a result of
, 15 loss of containment isolation or as a result of interfacing 16 system LOCAs, are also low in frequency, around about 10 to 17 the minus 7 or lower.
18 MR. REMICK: .I have a question. As I hear you 1
19 and the previous Staff speakers, I'm having trouble sorting 20 out when you are telling us that this is what is in the 21 Applicant's study and when you are' saying we . endorse it. ;
l 22 Are you not endorsing any part of it? !
23 MR. PRATT: At this point we are not endorsing _
-I '
24 anything.
25 MR. REMICK: At this point -- okay.
ACE-FEDERAL REPORTERS, INC.
202 347-3700 Nationwide Cmcrage 800-33M646
28135.0 223 RT 1 .MR. PRATT: The purpose of the exercise here was 2 to try to let you know where the important aspects are and 3 why we are focusing, in the couple of months we have to do 4 it, on these areas. That really is a nice lead-in to the 5 next Vugraph. Because we are obviously going to have to 6 look very closely at the structural integrity of the 7 containment building and whether we see high failure points 8 at the containment building.
9 (Slide.)
10 We have to look at isolation --
11 MR. SEISS: Excuse me, containment structural 12 integrity, you said? Is that S-1 or S-6?
O 13 MR. PRATT: Pardon?.
14 MR. SEISS: You pointed out previously that 15 whether S-1 or S-6 can happen is rather important. '
16 MR. PRATT: Yes, sir.
17 MR. SEISS: Now you are' listing things to look 18 at. Containment structural integrity relates to which of 19 the source ter'm categories? S-l?
20 MR. PRATT: It relates to S-1, because if it 21 failed catastrophically close to the time of vessel failure 22 you would get an S-1-type release.
23 MR. SEISS: Steam explosion or direct heating?
24 MR. PRATT: Steam explosion, direct heating, et 25
- -) cetera.
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800 33MM6
28135.0 224
,_yRT b
1 Now, in the-Applicant's analysis, that is a very 2 low probability event so that where most of the frequency 3 for the S-1.is coming from are bypass sequences, currently.
4 MR. SEISS: The data for structural integrity is 5 not associated with S-6, is it? I 6 MR. PRATT: It would also apply that#-- what 7 would be in there is circumstances where you might open up 8 a small leakage or large leakage. S-2 is a small leakage, 9 S-6 is a large leakage, and you saw the difference in 10 consequences if we have a.large leakage versus a small. So, 11 again, it's important to know the leakage characteristics --
12 Ms. SEISS: That's the difference between A and
~
t hs s 13 B leaks?
14 MR. PRATT: A and B; yes. By the time you get-15 to C you are. pretty well --
16 MR. SEISS: What's the difference between B and 17 C?
18 MR. PRATT: In terms of hole size? I think in 19 one case the difference between going to a C, would be that 20 you would basically be evacuating most of the -- you would 21 be blowing down the-building, in other words.
22 MR. SEISS: Can you relate the S-2, -3, S-5, for 23 example, to B versus C?
24 MR. PRATT: You might equate the S-1 to a 1
7- 25- situation where you had a catastrophic f ailure close to
(_5/ !
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80N36-646
28135.0 225 RT f
1 vessel failure.
2 MR. SEISS: S-1 is a C. Is S-2 an A or a B?
3 MR. PRATT: S-2 is an A. S-6 is a B.
4 MR. SEISS: And S-3 is a C? ,If you used Greek 5 letters, mayb'e it would help.
6 MR. PRATT: What we are'saying here as'well~--
7 and this is an important' point. I don't want to overkill 8 -this Vugraph.
9 (Slide.)
10 Because of the way we~are doing this analysis 11 here, because we are not doing anything with the people for.
12 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> -- we are not evacuating, we are just leaving them L
13 there and letting them get whatever comes out -- the size 14 of the leakage is important. You might look at, from an 15 overall risk perspective if you were evacuating people, the--
16 leakage area, whether it's small or large, may have not 17 been that big an impact, if it's occurring late. Because
. 18 we are integrating the. dose, here, over.24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />,-the fact 19 that this is a small leakage rather than a little bit:-
~
20 larger leakage, you are talking about the difference 21 between 1.5 inches and 8 square inches;-this is a 22 significant difference. This tails off at.about the two 23 miles. If it was a larger leakage it wouldn't. It would
- 24 be coming down a little bit 'further out. 'So it is 25- important here that we be aware of that.
i i ACE-FEDERAL REPORTERS, INC.
202 -347-3700 Nationwide Coverage 800-336-6M6
. . , - . - , . . __ . , , . . . . . , ._.m.- . . . . .m.. - - . -
28135.0 226
.BRT 1 MR. CORRADINI: Then you brought in something 2 else that I missed,-if that's the case. Early on where the-3 Seabrook presentation this morning pointed out, there were 4 three things that were important: one being the 5 containment; two being the source term; three being -- I 6 guess it was bypass. You are also saying it is what the 7 definition of containment failure is and leakage rate 8 that's important? So that when I have the possibility.of 9 failure, that failure is a type B failure, not a type C.
10 If it were a type C, I would see a characteristically 11 different curve.
12 MR. PRATT: This is A going to C. This is the-13 small one, A.
14 'MR. CORRADINI: I understand that. I am picking 15 numbers making a point. It's not just simply whether I
~
16 fail or not, it's also the degree of failure that's also 17 quite important, because of the timing of the emergency 18 planning is that you have to -- the assumptions are that 19 you have to hold everybody there for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />?
20 MR. PRATT: Recall what we are doing here. This 21 is a way of determining the times of distances that you 22 would have to worry about~, given an accident. Nobody is 23 suggesting that we would leave people around for a day, 24 given one of these accidents. We are saying, if we did, 25 these are the times of radii that we would expect problems ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336 4 46
,,-n - ,. ., . . - ,
28135.0 227 RT 1 to-occur in; therefore these are the times of distances you 2 would have to, perhaps, plan for. Okay?
3 MR. LONG: If I could add something. When you 4 look at the timed duration of the release, even in the case-5 where you don't have evacuation, the three-puff model they 6 are using for the dose calculational allows a duration, 7 long duration release to be spread over a wider area and ,
8 thus you have less dose at a more distant location than it 9 would calculate out if you'kept the plume in the same 10 direction for the entire release. I believe the 11 calculations that we've done at Brookhaven to benchmark the 12 Applicant's results show the import of the three-puff as O'- 13 opposed to a one-puff' release. That has, I think, an 14 important relationship to why the curve drops..off in the 15 S-2 category as opposed to some of the others coming out 16 all in one puff, early or late.
17 MR. REMICK: Does the essumption that people.do 18 not evacuate for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> hold for people on the beach, 19 also?
20 MR. PRATT: Yes. This calculation is done' 21 assuming no normal activity for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and then you. move 22 them out of the problem.
23 Again, it's not a suggestion that that's what we 24 would in.any way do. It's to get a measure, if nothing ~is 25 done, what might happen.
J ACE-FEDERAL REPORTERS,-INC.
202-347-3700 Nationwide Coverage 800-33MM6
28135.0 228 RT 1 MR. REMICK: Do you happen to know if that's 2 where the preponderance of the dose comes from, the people 3 on the beach, on the average?
4 MR. PRATT: I have no idea.
5 MR. WOODNARD: It's really unre. lated in this 6 case, dose versus distance.
~
7 MR. PRATT: Yes. I see.
8 If we had absolute risk values here where we had 9 integrated that over a population, that effect would occur.
10 Just to complete the Vugraph that we were 11 talking about, we would obviously be looking at these two 12 areas which I have mentioned, and also, in terms of p/
s- 13 completeness -- and this has, again, been mentioned by 14 Steve Long, so I need not spend too much time on it -- but -
15 of course ~there are other accident sequences that could 16 lead to early and large fission product . release. 'That's 17 early bypass of the containment building, induced steam 18 generator tube rupture -- these are examples of areas that 19 one obviously must look at in order to convince ourselves 20 that we cannot get into those types of release categories 21 with frequencies much higher than the numbers we are 22 talking about.
! 23 Unless there are other-questions, Charlie 24 Hofmayer can give you a status on the containment
(~g 25 structural integrity review.
U i
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336-6M6
28135.0 229 RT i
1 MR.~MOELLER: Any more questions?
2 Okay. Let's move ahead.
3 MR. TORRI: Mr. Chairman,.may-I make one 4 correction? Fred Torri, Pickard, Lowe and Gerrick.
5 Release category S-6 was really analyzed as entirely 6 separate release, not related to an A, B, or C~ type 7 containment failure. It would, however, fall into the C 8 category,~ size, as an 8-inch-diameter penetration. It is 9 equivalent to a C-size range, but it was analyzed entirely 10 independent of the A, B, C as an 8-inch hole existing at-11 the time at the beginning of the accident. Not caused by 12 overpressurization.
O 13 MR. SEISS: What have I got now?
14 MR. TORRI: What_you have as release category S-3, 15 which is late overpressure failure.
i 16 MR. SEISS: That's a C.
17 MR. TORRI: That can lead either to a B or a.C, 18 depending on the probability distribution which I showed on .
1 19 this bottom, the bottom curve.
20 MR. SEISS: Okay.
! 21 MR. TORRI: However, I might add this, too, when 22 an S-13 release occurred later than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after -- with 23 a greater than 24-hour warning time,.we always assumed it 24 was a C release because it didn't make any difference. The
.25 other thing, of course, where B and C comes into' play is ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 2 33 N
28135.0 230 RT 1 .that each S-2 release leads up to a B or a C failure, 2 eventually, because pressurization continues.
3 So, an S-2 is not just a small hole forever, 4 it's a small hole until the pressure gets big enough to 5 have a type B or C failure.
6 MR. MOELLER: We'll listen then about your 7 review of the containment structural integrity.
8 MR. HOFMAYER: I would like to briefly summarize 9 the status of the review of the containment structural 10 integrity. It's a little more than structural in terms of 11 functional integrity we are looking at.
12 MR. MOELLER: When did you do this? It says you k' 13 had a meeting.
14 s MR. HOFMAYER: Our review started basically in 15 the middle of August with a meeting with the Applicant at 16 Brookhaven wherein they did present the details of the 17 analysis that he conducted concerning the structural 18 integrity of the containment.
19 Primarily they focused on the calculations 20 ! performed by SMA which support the PSA. As a result of l
21 ! that meeting, that basically is when we initiated -- when l
22 l we initiated this review.
l 23 We obtained at that meeting, I would say pretty 24 l much all of the information that exists that they have in (3 25 i their files concerning the containment analysis. We still L. :' \
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cmerage 800-33MM6
28135.0 231 RT 1 have some information on isometric piping drawings we would 2 still like to obtain from them, but I would say they have 3 been very responsive to us . I certainly can't chide them 4 in any way. They have opened up their files and my room is 5 now full of a lot of paper.
6 So we have, I would say, most of the information.
7 So, subsequent to that we didn't obtain all the 8 information at the meeting, we obtained it somewhat later.
9 We have embarked on a review of primarily 10 focusing on the structural analysis performed by the 11 Applicant and we have reviewed -- gone through the SMA 12 overpressurization calculations and have identified certain
's 13 issues, which I'll get to on the next Vugraph, that we 14 really feel need to be looked at a little bit closer.
15 MR. MOELLER: Have you done similar studies for 16 other nuclear power plants?
17 MR. HOFMAYER: We have performed analyses for 18 other plants. We also participated in the containment 19 performance working group estimates of containment. But I 20 f would say it's fair to say that a few people have gone to I
21 i the extent this Applicant has, in terms of their analyses.
i 22 So, as far as I know, no one has performed that.
23 MR. MOELLER: Okay.
24 MR. HOFMAYER: As part of the review also we
(~} 25 participated in a site tour on November 8 and 9. This was
<J ACE-FEDERAL REPORTERS, INC.
202-347-37(U Nationwide Coserage 800-33MM6
28135.0 232 BRT 1 in conjunction with Trevor Pratt's groups who hadLthree 2 system engineers up there or valve engineers looking at 3 interfacing systems LOCA and whatnot. We had a piping 4 engineer look at all the containment piping penetrations
- . 5 and a structural engineer walk down the containment. !
6 Unfortunately, I did miss that trip for personal 7 reasons; I would have liked to have seen it myself. But we 8 did walk down the containment;- certainly probably not to 9 the extent the Applicant did, but it did give us a very
. 10 good feel for the relationship with the adjacent structures, 11 what are the hard points in the containment and what do the 12 piping configurations look like.
13 I'm-sure you have walked down many plants. You 14 get an overview, but it's very hard to get 'a total 15 perspective without looking at drawings.
16 Another element of our review, we are embarking t
17 on our own finite element of the containment, which is an 18 axisymmetric finite element analysis. It will give us a 19 baseline to have more confidence on what the Applicant did.
! 20 Not that I'm saying we don't have confidence, but I think 21 it gives us the proper perspective. Also, as part of this 22 analysis we'll be able to focus more on certain shear 23 failure modes which we feel maybe need to be looked at in a 24 little more detail.
i 25 MR. SEISS: To what detail did you model the i
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 8(G336-4%46
28135.0 233 RT 1 penetrations in that analysis?
2 MR. HOFMAYER: In the axisymmetric? It's a pure 3 axisymmetric. What we'll be modeling is the reinforcing 4 steel in the liner and the discontinuity readings. We are 5 not doing a three-dimensional type of analysis.
6 MR. SEISS: Why?
7 MR. HOFMAYER: Excuse me?
8 MR. SEISS: Why are you just doing a smooth 9 shell analysis?
10 MR. HOFMAYER: Part of the reason is, 11 potentially, timing. But one has to look at the specific 12 details of each element. You are basically trying to do
( )
- 13 that nonlinear - you are talking about a very large 14 analysis. You can't do that very simply.
15 MR. SEISS: Is this finite element analysis?
16 MR. HOFMAYER: Yes, it is. We have done this 17 type of analysis for other containments.
18 MR. SEISS: I don' t see what you are going to 19 get out of a finite element analysis of a clean shell.
20 MR. HOFMAYER: You are basically looking at the 21 failure modes, particularly of the shear at the base and i
22 l what effect the cracking has, in terms of being able to 23 ! adequately predict that shear. And the discontinuity at i
24 the spring line.
(N ;
25 We also met with the Applicant early in the week, i
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Cos erage 8 @ 336-6646
l 28135.0 234 BRT O 1 where we discussed a number of issues that we have raised 2 as a result of our review of the SMA cales, and I'll
'3 characterize that on the next review.
4 Our plans are basically to get their' responses
~
5 to those type of questions and have an additional technical 6 meeting with them, probably of several days' ' length, to go 7 over in more detail-the assumptions that they have made in 8 those calculations.
9 We are about a month into the review.
10 (Slide.)
~
11 I just.wouldn't characterize-the containment 12 behavior. This is from the Applicant's report.
13 What we are talking about is a containment that 14 has an initial general yield state _of approximately, as the 15 Applicant. calculated it through SMA, of 157 psig; USEG 16 predicted something like 150 psig in the hoop deflection, t
17 when the containment will reach yield.
18 From that point on, SMA has_' predicted that they 19 will get a mean failure pressure of something in the 20 neighborhood of 216 psig. That will give you a hoop 21 deflection in excess of 3 feet on the hoop, which would 22 almost close the gap between the shield building and the 23 containment. !
24 It also introduces, when you start getting to 25 these very large deformations, you will be contacting ACE-FEDERAL REPORTERS,.INC.
202 347 3700 Nationwide roverage 804336-6M6
~ . . _ . . - -- -
i e
'28135.0 235 RT i
1 adjacent structures and-you are talking about a very large, 2 nonlinear response.
3 One of our concerns and what we are trying to a
4 get a handle on is just how much credit, since this is a l
5 mean failure pressure ---you know, what's really driving 6 the result in terms of gross containment f ailure? ' What 7 pressure, somewhat back here,'does this bec'ome significant?
]
- 8 And can we really, indeed, predict that type of behavior 9 with these types of deformations?
j 10 So, just to characterize it, most people have i
11 been in this range in terms of predicting capability 12 pressures for containments in these type of studies.-
13 MR. SEISS: Some would thin out to 1 percent 14 strain. Some people assume 1 percent strain. That's a lot 15 further out than you have.
16 MR. HOFMAYER: 1 percent strain -- well, I'm not 17 really familiar with who. I know on the prestressed. i 18 containments they do have.
19 MR. SEISS: It would be 8 inches. About.10 l 20 inches.
4 21 MR. HOFMAYER: That would get you to your l
22 pressure of about 170.
23 'MR. SEISS: If Seabrook can assume this thing j 24 would fail at 1 percent strain, would their risk analysis 25 have come out a whole-lot different?
}
'I 4
ACE-FEDERAL REPORTERS, INC. '
202-347-37(X) Nationwide Coverage - 80 4 336 6646
. 4 . . . .
l 28135.0 236 BRT O
1 MR. HOFMAYER: .That's exactly the type of 2 question we would like to pursue with the Applicant.
3 MR'. SEISS: Don't think you'll answer that with l 4 structural analysis. I think you'll answer that with risk 5 analysis. So that's why I was-asking. I wasn't asking you.
6 MR. HOFMAYER: Part of the' problem is, of course, 4
7 they would then have a probability distribution which would 8 maybe force it lower. I don't know. I can't answer.
9 MR. LONG: I would like to add something at this 10 point. The curves that the previous speaker had up there l 11 when he was talking about things being off-scale low, we 12 are really looking for two types of things. One is the i
() 13 subject that would raise some of those contributors 14 off-scale low back up so they are on scale with releases 15 that we are already considering. And this question falls
! 16 into that area. In other words, bring the S-6 or the S-1 17 curve probability up high enough so that term appears 18 important to the sum of the releasing frequencies.
19 MR. SEISS: I can see that for the S-1-type case.
20 Is that the only place where you are interested in where 21 this curve ends?
i
~
22 MR. HOFMAYER: For S-l?
23 MR. SEISS: I could see that very clearly. If 24 you have a steam enclosure,'it either goes or doesn't go; 4 25 you don't know which, and you'll never know exactly which.
ACE-FEDEPAL REPORTERS, INC.
- 202 347-3700 Nationwide Coverage 800 33M646
28135.0 237 RT
! 1 That's fairly clear, but'what I was concerned about or 2 thinking.about is if I. stopped at 1 percent strain and l
3 arbitrarily took that rather than trying to get the 4 ultimate, would the partition between the ' type B and type C
- 5 leaks be markedly different? And to what extent would thate 6 contribute to the risk? Changes in the risk?. That's a lot-7 more complicated, to see how it affects the division I
l 8 between type B and type C, just to see what the ultimate is.
1 4
9 MR. TORRI: Mr. Chairman?
10 MR. MOELLER: Yes.
, 11 MR. TORRI: May I shed some light on'that?
- 12 Torri PLG.
13 If we assume we take off that curve the 1
) 14 percent strain value is souething like 170 psig, that 15 translates into 185 psia. If you recall the distribution 16 that I showed you for dry conditions, the 50 percent 17 failure probability was at 189 psia, only '4 psi higher than 18 what you would get this way.
19 So we really haven't -- because we considered 20 the other failure modes, the local failure modes, we have 21 net taken credit for failure out at 3 feet plus deflections.
22 We found other failure modes that lead to failure earlier, a
23 which is equivalent to you saying, 'let me take 1 percent i 1 ,
24 strain and not worry about it, except that we know, now, l 25 where the leakage was occurring.
l ACE-FEDERAL REPORTERS, INC. l l ' 202-347 3?(o Nationwide Coverage 8043E6M6
28135.0 238 BRT G
1 MR. HOFMAYER: Just one comment on that, though.
2 When you are talking about the dry condition you are 3 talking about the high temperatures, so therefore your 4 yield condition is also lower.
5 MR. TORRI: That's right.
6 MR. HOFMAYER: It dropped down substantially for 7 the ultimate; it would also drop down for the yield. So if 8 you compared it to the wet condition, which would be 9 something in the neighborhood of, what, 200 psia?
10 MR. TORRI: 214 at -- is that figure for wet?
11 MR. HOFMAYER: This is for wet. If you used the 12 dry, this curve would drop also.-
(";
E 13 (Slide.)
14 As a result of our review, there are a number of 15 areas that we feel need to be looked at.
16 Basically we feel that the uncertainty factors 17 assigned to the various pressure capabilities, we need --
18 these are a number of items that we intend to discuss 19 further with the Applicant and pursue in the review that we 20 are doing. Most of them do focus on the issue of high 21 containment strains and large deformation.
22 i The first item of course, though, is the 1
23 uncertainty factors that they have assigned to these 24 various parameters. We want to better understand the basis r~w 25 of those factors.
s' ;
ACE-FEDERAL REPORTERS, INC.
-- x_m_. s_
28135.0 239 7-]RT
'v' 1 An additional concern is the question of the 2 transferring load between penetration sleeves to the wall 3 under high strain conditions. Many of the calculations 4 that we've seen today, it appears that in some of these 5 areas, you know, we are talking about relatively high 6 strain conditions. There are additional loads imposed on 7 the rebars due to the various penetrations, due to the 8 transfer of the loads. We feel that those must be 9 incorporated into the analysis and to study the impact that 10 that might have on the conclusions the Applicant has.
11 Another area is the question of nonuniform crack 12 growth. Basically the liner is assumed to follow the 13 concrete, pretty much that the cracks are all fairly 14 uniform around the containment.
15 We feel that there may be a potential, under 16 these type of conditions where the cracks are not uniform.
17 This may impose higher liner strains or maybe conditions 18 where the liner could be, in localized areas, restrained in 19 a much higher -- higher strains than what they have 20 predicted.
21 There's a question we have on the compatibility 22 of strains on the rebar and liner, in terms of the strength 23 that they have assigned --
24 MR. SEISS: Excuse me, Charlie, how are you rs 25 going to get anything on nonuniform crack growth out of an
(_)
ACE-FEDERAL REPORTERS, INC.
102-347-37(U Nanonwide Coserage 8in336-6646
28135.0 240 RT 1 axisymmetrical finite element model?
2 MR. HOFMAYER: We are not. We are not proposing 3 that that model will answer these questions. I can assure 4 you of that. What I feel is that model will calibrate us 5 with them in terms of the yield capability that they 6 predict and whether we can agree with the fact that the 7 shear failure mode that they have assigned to be very high 8 for this containment is, indeed, the fact.
9 MR. MOELLER: Are these items, now, onos you are 10 expecting the Applicant to provide the additional 11 information to resolve?
12 MR. HOFMAYER: Yes. We have asked them
- i k ' 13 questions on this. We also have asked them to make 14 available to us the people that actually did the 15 calculations at SMA. We have reviewed the calculations; 16 not having the people that did them directly available and 17 the time constraints that we have, we feel it would be more 18 ,
appropriate to discuss these issues with the Applicant plus l
19 { go through their calculations together with them so we can i
20 ; better understand the assumptions that they have made.
21 ! I mean we may or may not agree with them on all 22 of these issues. Once we better understand where they are 23 coming from, then we may assign other factors to these 24 l parameters and see what effect that would have on the risk i
25 (n )
! study.
l l
ACE-FEDERAL REPORTERS, INC.
- >- ~ _ m_ , ,,m3 -
l-28135.0 241 RT-s 1 This would then be some iterative process with 2 Trevor Pratt's group in terms of, as Professor Seiss said, 3 in terms of what sensitivities are brought up. If we lower 4 these numbers and believe that this is a range that we feel 5 may be -- maybe is more appropriate, what impact does-that 6 have on the study?
7 As I mentioned at the discontinuity areas, we 8 can probably address this through some of our studies,'but 9 we also asked the Applicant. The fact that it appeared 10 that in some cases they did separate these ef fects and we 11 are concerned that the shear and the bending and the
] 12 tension are all occurring together; and have they 13 adequately incorporated all three effects into their i
14 calculations?
15 Another question, I guess, that kind of burns at 16 us is: Under these high strain conditions we are 17 interfacing with adjacent structures.- These would then
! 18 create a nonaxisymmetric condition in the bases of the 19
~
containment, and again we are saying, perhaps there's a 20 localized higher strain that we are seeing in the
- 21 containment that is not being taken into account by either 22 of us in terms of the axisymmetric analysis.
23 Puncheon shear at the field transfer building is j 24 something that the Applicant has evaluated and we feel that 25 that.needs to be addressed in more detail.
ACE-FEDERAL REPORTERS, INC. -
202-347-37(U Nationwide Coserage 8%336-ui46
28135.0 242 1 Again,.looking at what effect-that has on the 2 reinforcing steel; in ~ terms of the increased strains in the 3 steel.
4 Material strengths, I don't think we have a big 5 concern here, although we do have a question about the 6 liner strength material. They did use probably about a 50 7 percent increase f rom the specified yield stress to' what 8 they considered the mean yield stress. They did quote 1 9 three test reports that they have. This seemed high to us.
10 We want to explore that matter further.
11 The question of piping penetrations. They did 12 look at some of the key piping penetrations. In our 13 walkdown, I guess we /SAELGSD that they potentially looked 14 like the ones you would want to concentrate on. What we 15 are concerned about, we would like to have more confidence 16 in, is the fact that there are many piping penetrations and, 17 indeed, are the ones we looked at the most critical?
, 18 And the last item is the question of leak area i
19 calculations. They do calculate certain-size openings. I
)
i 20 guess we would like to better understand how they 21 calculated those. We are talking about failure modes 22 occurring, like in the feedwater fluid head: Is it 23 reaeanable, in terms of the leak area sizes that they have U
24 calculated?
1 gs 25 These are the items that, I guess, I would
\-
ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800 346M6
-- _ . , - - . . - . . , _ ,. _ - - . , _ . , . . - , . .. . - . - . ~ - .. - . . . . - . _ . -
3 28135.0 243'
, RT I highlight as the areas that we are focusing our review on.
l 2' If there-are others that you feel we've overlooked or you 3 feel that some of these are unimportant, I would appreciate j 4 .any comments.
5 MR. SEISS: I think-you may find out some of; 6 them aren't important, but I don't,think you can tell now.
! 7 MR. HOFMAYER: That's true. ,
q 8 MR. SEISS: They may not-be important from the 9 risk point of view even if they are important from the 10 structural point of view.
\
11 MR. HOFMAYER: That's another point. If we can j 12 back ~off f rom the pressure we are talking about in terms of i
1 13 the mean and the 5-foot /3-foot' deformation, and it really
- 14 does not become significant until the strains are much
- 15 lower level, then potentially a lot of these questions i
16 can --
l 17 MR. SEISS: The first study that was made two or 18 three years ago, somebody said this was done by the risk i
19 analysts and not by structural engineers. It would be just 20 as bad to have the structural engineers make one without 21 talking to the risk analysts.
, 22 MR. HOFMAYER: That's true.
] 23 MR. SEISS: I hope you guys get.together.
1 24 MR. HOFMAYER: Yes. Trevor's phone number I 25 have memorized.
i ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 80(L33MM6
.g m 28135.0 244 RT 1 MR. REMICK: Every time before they come before 2 the ACRS they get together.
3 MR. HOFMAYER: No, it's more than that. That's 4 all I have to say.
5 MR. MOELLER: Thank you. We'll then move on to 6 hear Mr. Youngblood. Okay.
7 MR. YOUNGBLOOD: I'm Bob Youngblood. I'm 8 speaking for the front end group in Trevor's division.
9 (Slide.)
10 It has been made pretty clear by now that 11 getting stuff out of conta'inment, whether anything will get 12 out of containmont is the important point. Our main effort i ;
'-' 13 in this review is to look at interf acing systems LOCA and a 14 little bit at containment bypass.
15 (Slide.)
16 We have a concurrent effort in support of the 17 generic issue on interf acing systems LOCA that's a much 18 more extensive problem than this and it's actually 19 examining several plants in support of the Staff's I
20 i resolution of that issue.
I 21 l What we've done in the time frame of this i
22 l project is essentially divert effort from that study to 23 ,
this and we'll be looking at the aspects of this analysis 1
24 l that have been highlighted as the enhancements to the event rm 25 l V, old-fashioned analysis. A few of these are just listed
(~, I i
ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Cmerage 80lk3364M6
1 28135.0 245
~BRT
/s b
1 here. They looked more carefully at the more likely break 2 locations, what you might call a lucky break, being a break 3 down at the bottom.of this vault.
4 There are, of course, plant-specific features of 5 Seabrook that make that a particularly favorable place to 1
6 have a break. But the idea of examining what part of the 7 system is going to go is certainly an aspect that has come 8 up in the generic work as well. It is well worth 9 understanding why it won't be the long pipe run that goes; 10 or coming to grips with how unlikely it is that that long 11 run will be the thing to go.
12 They've also, of course, examined which 13 components will or won't be affected by that, and have gone 14 into more detail in analyzing how the operator can go right 15 or wrong in doing this.
16 We hope to actually visit the plant and discuss 17 the actual handling of these scenarios in a little more 18 detail.
19 In the area of containment bypass, we have 20 someone familiar with valve operability looking at these 21 issues. I don't believe we have anything like a concurrent 22 program in containment bypass to piggyback this effort on.
23 But he'll be doing what he can in the time available.
24 I could probably try to tie up more time than 25 this, but between Seabrook and the earlier Staff
(~)
\-)
ACE-FEDERAL REPORTERS, INC.
202-347 3700 Nationwide Coverage 8m33&6M6
l 28135.0 246 f-7RT b
1 presentations, I frankly can't see the point.
2 MR. MOELLER: Okay.
3 Do we have questions for Mr. Youngblood?
4 Okay. Thank you, then.
5 Mr. Pratt, does this then conclude the sequence 6 of the presentations by the Staff?
7 MR. LONG: That concludes the Staff's 8 presentation. I think we have a few concluding comments.
9 MR. MOELLER: Okay. The next item on the agenda 10 is -- would be the concluding remarks by Public Service of 11 New Hampshire. Do you have --
12 MR. MOODY: We just thank you for your time. We 13 await your comments.
14 MR. MOELLER: Then let's move on and do the 15 concluding remarks f rom the NRC Sta f f .
16 MR. NOONAN: Dr. Moeller, we really have no 17 set-up remarks to make at this point in time other than to 18 indicate, as is obvious, we have just embarked on this 19 study. We'll do some further looking. We'll come back 20 with further recommendations to the utility, as to what we 21 think they might do to make this study better. Those 22 recommendations will be documented and we hope to come back 23 and brief the Subcommittee again and tell them what our ,.
24 conclusions are.
25 MR. MOELLER: All right. Do any of the ACE-FEDERAL REPORTERS, INC.
202-347-37(W) Nationwide Coserage MXF3%%4
28135.0 247 RT 1 Subcommittee members have comments?
2 MR. REMICK: I just have something that I might 3 ask the Applicant and Staff. I don't know if they are 4 aware of, but last week in Pittsburgh an interesting paper 5 was given called "A Rational Approach to Emergency Planning 6 Zone Definition," which, from a different perspective, 7 repeats much of what has been stated today.. I found it 8 kind of interesting. The comments here -- this was a 9 reassessment of emergency planning requirements. They say, 10 based on this review or reassessment, it was determined 11 that NUREG-0396 approach, one, is somewhat subjective and 12 very difficult to interpret quantitatively; two, is 13 unwieldy in its application and would be especially so with 14 a large number of source terms; and, three, does not really 15 provide visibility of the importance of accident sequences 16 which dominate exposures.
17 They go on to come to a conclusion in the paper, 18 which I won't endorse either, which comes down to the f act 19 that the suggestion that the 10-mile limit probably should 20 be reasonably reduced to two miles. I just suggest it to 21 you as, I.think, an interesting paper. I'd be happy to 22 give it to the ACRS Staff to make copies and provide. It's 23 a Stone & Webster paper by M.J. Bazon and E.A. Mormon.
24 I think it's worth reading. :
25 MR. MOELLER: Well, let me make some concluding O
ACE-FEDERAL REPORTERS, INC.
202-347-37(M) Nationwide Cmcrage 8@346M6
'28135.0 248 RT 1 remarks, then, or suggestions. If the Subcommittee concurs, 2 then we'll follow this path.
3 It seems to me that, today, it has been 4 extremely beneficial to hear from the Applicant and to gain
- 5 a first cut at what you are doing, what you have done, and i 6 some of the conclusions that you have reached.
g 7 It~ is obvious, though,. equally, that the NRC
{
8 Staff's review is very preliminary, as they had indicated 9 to us. It will be several months before they will be 10 getting -- before they will begin to firm up' their j 11 conclusions.
! 12 We have allocated time at the next full 13 Committee meeting in October for a discussion of this topic.
I 14 What I would suggest is that, at that time we allocate 15 perhaps an hour for the Applicant to come in, Public 16 Service of New Hampshire, and give us, within an hour, a f 17 rundown of what it is that you have been doing. You know, i 18 consolidate what you have done to date, at least you would i 19 begin to present to the full Committee and get them to
} 20 start thinking about some of the ideas that you have in i ;
l I 21 mind. We might give the Staff five or 10 minutes, simply i
22 to say that they are launching their review and they hava a 1 l
23 long way to go. But we could give them that time. And !
24 then at the same time I would take a few minutes at the .
, 25 beginning, just to briefly summarize the Subcommittee i
! ACE-FEDERAL REPORTERS, INC.
mmm, s.e_m_... ,,3 3-j
i i
e i
28135.0 249 RT i
1 meeting of today.
! 2 Does that sound reasonable' for us?
3 MR. REMICK: To me, yes.
] 4 MR. MOELLER: Okay. If that sounds reasonable.
5 Do either -- does the NRC Staf f -- do you have any --
i i 6 MR. NOONAN: .That sounds reasonable to me.
7 MR. MOELLER: Do you agree, for Public Service?
8 All right. If that's agreeable, then, I will conclude the 2
9 meeting, the Subcommittee meeting by pointing out that not i 10 only will we hear, then, the presentation in October before i
11 the full Committee, but it is obvious that we'll have to 12 have one or two more Subcommittee meetings to review this I
13 in depth before we go to the full-Committee with our 14 recommendations. Then at some point'you'll come before the 15 full Committee, at which time, I presume, we would send, 16 formally, some sort of a report to.the Chairman of the i
i 17 Commission.
18 With those remarks,.then, let me tha'nk all of f
19 our speakers, the Public Service of New Hampshire Group and l 20 the gro'up from Pickard, Lowe and Gerrick, and the NRC Staff.
21 With that, I'll declare the meeting adjourned.
I s 22 (Whereupon, at 3:35 p.m., the meeting was i
23 concluded.)
24
() '
i ACE-FEDERAL REPORTERS, INC.
202-347-3700 Nationwide Coverage 800-336 4646
CERTIFICATE OF OFFICIAL REPORTER O
This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY COMMISSION in the matter of:
NAME OF PROCEEDING: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SEVERE ACCIDENTS and -
SUBCOMMITTEE ON OCCUPATIONAL AND ENVIRONMENTAL PROTECTION SYSTEMS
]
DOCKET NO.:
PLACE: WASHINGTON, D. C.
J DATE: FRIDAY, SEPTEMBER 26, 1986 i
were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission.
1 (sigt) .
(TYPED)
CRAIG L. KNOWLES Official Reporter ACE-FEDERAL REPORTERS INC.
Reporter's Affiliation, i
I O
CERTIFICATE OF OFFICIAL REPORTER O
This is to certify that the attached proceedings before 4
the UNITED STATES NUCLEAR REGULATORY COMMISSION in the matter of:
NAME OF PROCEEDING: ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON SEVERE ACCIDENTS
- and
! SUBCOMMITTEE ON OCCUPATIONAL AND
! ENVIRONMENTAL PROTECTION SYSTEMS DOCKET NO.:
PLACE: WASHINGTON, D. C.
O DATE: FRIDAY, SEPTEMBER 26, 1986 were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear l Regulatory Commission.
(sigt) -
(TYPED JOEL BREITNER Official Reporter RhhYhk^Eh k$h
- O
.-. ._ _. - . . - . = _ - - -__ . . . .- . _
h, .
s..
l 0 l l
JOINT EETING 0F TW ACRS SUBC(M1ITTEES ON l OCCUPATIONAL AND EMROMNTAL PROTECTION SYSTEMS ;
AND l
SEVERE (CLASS 9) ACCIDENT 1 l l SEAEPmK STATION I -
SEPTEMBER 26, 1986 3
O r 1
i l
4 VINCENT S' NOTAN, DIRECTOR PWR PROJECT DIPECTORATE No. 5 DIVISION OF PWR LICENSING-A O ,
SLIDE 1
~ _ _ . , . _ . _.
O REVIEW 0F UPDATED PROBABILISTIC SAFFTY ASSESSENT FOR SFABROOK STATION OBJECTIVES
. TO REVIEW TECHNICAL ADE0VACY OF SEABROOK STATION RISK PANAGEENT AND EE RGENCY PLANNING STUDY AND E K RGENCY PLANNING SENSITIVITY STUDY FOCUS _
. CfWTAINENT STRUCTllRAL INTEGRITY
. CONTAINf9fT BYPASS
. IEERFACING SYSTEPS LOCA RESULTS.
O . IDENTIFICATION OF SIGNIFICANT TECHNICAL ISSlES
. EVALUATION OF PAJOR RISK CONTRIBUTORS
, IDENTIFICATION OF APEAS FOR FURTliER RISK PEDUCTION l
O SLIDE 2 1
i CURRENT STATUS OF SEABR0nK REVIEW O , SAFETY REVIEW FIRE PROTECTION INCOPPLETE
, EERCBCY PLANNING REVIB1 ONSITE COMPLETE APPRAISALS EXERCISE OFFSITE INC0lPLETE N. H. EXERCISE PERFORW D PASS. EXEPCISE TO BE PERF0 PED COORDINATION WITH FEPA
~
REGIONAL ASSISTANCE Cat 11TTEE
, HEARINGS EERGENCY PLANNING ONSITE AUGUST,1983 TO BE SUPPLEENTED BY HEARING DURING THE EEK 0F SEPTEMBER 29,1986
_ ENRGENCY ACTION LEVELS OFFSITE NEW HAMPSHIRE - POSTPONED FROM AUGUST, 1986 AND NOT YET RESCHEDULED PASSACHUSETTS - PLANS NOT F0PPALLY SUBMITTED TO FEPA FOR REVIEW SAFETY SEPTEMBER 29, 3986
_ ENVIRONTNTAL OJALIFICATION OF EQUIPENT
_ CONTPOL ROOM DESIGN REVIEW O
SLIDE 3
RESULTS OF LATEST SALP O
PERIOD BIDING PERIOD ENDING RECENT TREND R!NCTIONAL AREA (12/ 31 /84 ) (3 /31 /86 )
A. . CONSTRUCTION 2 1 COPEISTEE B. PREOPERATIONAL 1 1 CONSISTENT TESTING C. FIRE PROTECTION AND HOUSEKEEPING N/A 1 CONSISTENT
. D .' OPERATIONAL READINESS N/A 1 CONSISTENT E. EERGENCY PREPAREDNESS N/A 2 If0 ROVING O F. ASSURANCE OF Cl1ALITY I 1 CONSISTENT G. LICENSING 2 1 C0tSISTENT 4
l l
O SLIDE 4
CEACHERN l' 1
l
~(q h .'L ,
l m
}v% $~
g l September 25, 1986 1
Dr. William Kerr l Advisory Committee on l Reactor Safeguards l Nuclear Regulatory Commission 1717 H Street NW Washington, DC 20555 i
Dear Dr. Kerr:
I am registering with your committee my strong opposition to the reduction in size of the emergency planning zone for the Seabrook Nuclear Power Plant. As your committee discusses this issue, I would O ask that you instead consider expanding the evacuation zone. In light of the 18 mile permanent evacuation following the Chernobyl disaster in the Soviet Union, it is irresponsible and insensitive to the health and safety needs of the people of New Hampshire to reduce the zone at Seabrook.
One of the main lessons of Three Mile Island made clear by the Kemeny Commission and the NRC's own inquiry group was that reliance on engineered safety features alone was not sufficient to assure adequate protection of the public against the nuclear hazards of an accident.
The applicant's request to shrink the Evacuation Planning Zone is a blatant attempt to remove the opposition of the Commonwealth of Massachusetts from the planning process. It amounts to nothing more than a retreat to the pre-Three Mile Island position that the engineered safety features alone are adequate. History shows us otherwise.
Sin . rely, Paul McEachern Democratic Nominee for Governor of New Hampshire PM:jm1 Worth Plaza, Suite #7, Portsmouth, NH 03801 603/433-2288 m 1170 Elm Street, Manchester, NH 03101 603/644-2200
( Kathenne D Amours, Fuof Ageat 154 baden Street, Manchester, NH g eo
l {
O SEABROOK STATION PROBABILISTIC SAFETY ASSESSMENT l
l Presentation to O ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Washington, D.C.
September 26,1986 l
O l -
O Agenda e Project Overview e SALP O Low Power Testing ,
e Emergency Planning O
O
1 l
O l Project Overview e Unit 1 Complete e Unit 2 24.1% Complste l
O i i
1 I
O
O SALP l
l FUNCTIONAL AREA RATING (1/1/85 - 3/31/86)
Construction 1 l Preoperational Testing 1 Fire ~ Protection And Housekeeping 1 Operational Readiness 1 Emergency Preparedness 2 Assurance Of Quality 1 O 'i *"Si"9 1 l
O 1
l l
O !
l l
J Low Power Testing 9 No Fuel Load Open items With NRR Or i Region 1 9 ASLB (Low Power License) l
- Open items !
O Equipment Qualification Time Duration O Detailed Control Room ,
O Design Review- SPDS O Emergency Action Levels
- Hearing Scheduled For 9/29 - 10/3 in Portsmouth, N.H.
Petition 50.57 (c) To Load Fuel
- And Conduct Precriticality Testing Submitted 8/22/86 0
O Emergency Planning I e Graded Exercise Held 3/86 (N.H. Only) e Risk Management And Emergency Planning Study Submitted To NRC
-7/21/86 _
e Revision 2 Of N.H. Plan Submitted 9/86 O e G v. Of M ss. Ann un ed RefusalTo Submit Plans To FEMA- 9/20/86 ,
l O
l
gm -aw -sm,---- - ---.--- -- - ------- ---'a -mm,-. maw-aw6B N m M M f MMeN Ne 4 mms--4* mea M 6 m5P+mb-MM5sh4* 4 hme wah&M E
i I
i 1-1 ll i
f 9
I l
l 1
i I
i l
l 6
i I
i s
h I I k
> ?
t 1
k I
s I
- ;i I
)
i
)
r i
.=
6
?
f i
l I
l >
e i
l l
l l
t t
k 6
h
O O O SEABROOK STATION PROBABILISTIC SAFETY ASSESSMENT by James H. Moody, Jr. - New Hampshire Yankee Karl N. Fleming - Pickard, Lowe and Garrick, Inc.
Presentation to ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Washington, D.C.
September 26,1986 I
O O O MEETING AGENDA APPROXIMATE TOPIC DURATION SPEAKER (min)
SSPSA (1983) OVERVIEW
- HISTORY, SCOPE & CONCLUSIONS 10 JIM MOODY
- METHODOLOGY & RESULTS 10 KARL FLEMING
]
PSA UPDATE
- OVERVIEW & CONCLUSIONS 30 JIM MOODY
- METHODOLOGY & RESULTS 30 KARL FLEMING l
1
\
- SSPSA (1983) OVERVIEW HISTORY, SCOPE & CONCLUSIONS
)
i a
l O ~O O PROBABILISTIC SAFETY ASSESSMENT e FULL-SCOPE LEVEL 3 PSA o PUBLISHED DECEMBER 1983 o PRINCIPAL CONTRIBUTORS
- YAEC - UTILITY PROJECT MANAGEMENT AND REVIEW
- PICKARD, LOWE AND GARRICK, INC. - PRA CONSULTANT
- e NRC REVIEW ~
l
- LAWRENCE LIVERMORE - PLANT MODEL - (RESPONDED i MAY 1986)
! - BROOKHAVEN - CONTAINMENT CAPABILITY NUREG/CR-4540 (FEBRUARY 1986)
i O O O I
(
SSPSA SCOPE AND COVERAGE OF ACCIDENT SEQUENCES e COMPREHENSIVE COVERAGE OF ACCIDENT SEQUENCES
- 58 DISTINCT INITIATING EVENT CATEGORIES
- 39 PLANT DAMAGE STATES (" BINS")
- 14 RELEASE CATEGORIES
- 16 MODULARIZED EVENT TREES e FULL TREATMENT OF DEPENDENT EVENTS
- COMMON CAUSE FAILURES (SYSTEM LEVEL)
- EXTERNAL EVENTS
- INTERNAL PLANT HAZARDS l - EXPLICIT MODELING OF FUNCTIONAL DEPENDENCIES l e PLANT-SPECIFIC AND ENHANCED CONTAINMENT MODEL l - ASSESSMENT OF CONTAINMENT FAILURE MODES
- QUANTIFICATION OF SOURCE TERM UNCERTAINTIES l - ENHANCED METHODOLOGY e SITE-SPECIFIC CONSEQUENCE MODEL
- MULTIPUFF RELEASE TREATMENT
- ACTUAL SITE CHARACTERISTICS l - QUANTIFICATION OF UNCERTAINTY
O O O i
COVERAGE OF INITIATING EVENTS IN SSPSA BY MAJOR GROUP 15 COMMON CAUSE EVENTS (38) en z 10 -
s
= _
O 5 -
8 E E s E ., ;;;;;;; F' b$6s~
3i >
4 LOSS TRAN- SUPPORT EARTH FIRES FLOODS AIRCRAFT TURBINE OF SIENTS SYSTEM QUAKES (9) (4) & TRUCK & TORNADO COOLANT (14) FAULTS (8) CRASHES MISSILES
! (6) (4) (4) (9) t
- O O O CONCLUSIONS (1983)
EARLY HEALTH RISK
- NRC SAFETY GOAL MET WITH LARGE MARGINS
- INTERFACING LOCA DOMINATES
- LATENT HEALTH RISK
! - N RC SAFETY GOAL MET WITH VERY LARGE MARGINS
- SUPPORT SYSTEM FAILURES DOMINATE CORE MELT FREQUENCY LOW CONTAINMENT EFFECTIVENESS j - PRIMARY CONTAINMENT VERY STRONG
- EARLY FAILURE UNLIKELY
- LONG TIME FOR OVERPRESSURE l
1
J %. - . __-.-aw i-...,. _ - . _ _ _ _ _ _ _ _ .
- O O O i
4 l
l l
SSPSA (1983) OVERVIEW METHODOLOGY & RESULTS l
l l
O O O i
4 SSPSA METHODOLOGY ENHANCEMENTS / DIFFERENCES e PLANT ANALYSIS - MODULARIZED EVENT TREE SCENARIO MODEL
- ENHANCED TREATMENT OF DEPENDENT EVENTS
- INTEGRATED ASSESSMENT OF ALL TYPES OF INITIATORS
- ENHANCED TREATMENT OF DATA e CONTAINMENT ANALYSIS - CORRECT PROBABILISTIC INTERPRETATION OF EVENT TREES
- REALISTIC ASSESSMENT OF CONTAINMENT FAILURE MODES
- ASSESSMENT OF SOURCE TERM UNCERTAINTIES S SITE ANALYSIS (CRACIT) - MULTl- PUFF / VARIABLE PLUME TRAJECTORY i - ACTUAL SITE CHARACTERISITICS
- ASSESSMENT OF MODELING UNCERTAINTIES -
)
9 RISK ASSEMBLY & - MATRIX SYNTHESIS OF RISK RESULTS l
- DECOMPOSITION - MATRIX DECOMPOSITION OF RISK CONTRIBUTORS
- INDIVIDUAL SCENARIO RANKINGS
- CAUSE TABLE BREAKDOWN OF SYSTEM LEVEL CONTRIBUTORS
}
I
(
O O O j SSPSA TREATMENT OF DEPENDENT EVENTS S FULL SCOPE COVERAGE OF INITIATORS
" INTERNAL" EVENTS (TRANSIENTS & LOCAS)
- INTERNAL PLANT HAZARDS (FIRES, FLOODS, MISSLES, etc.) ,
" EXTERNAL" EVENTS (SEIM[C, AIRCRAFT CRASH, etc.)
e EXPLICIT MODELING OF FUNCTIONAL DEPENDENCIES
- DETAILED DEPENDENCY MATRICES
- EVENT TREE STRUCTURES & IMPACT VECTORS
- CAREFUL TREATMENT OF PLANT / CONTAINMENT DEPENDENCIES
- 8 ADDITION ~OF " SPATIAL INTERACTION" TASK
- LOCATION DEPENDENT FT MODEL FOR SCREENING ,
- - PLANT WALKDOWN / INSPECTION
- SYSTEMATIC EVALUATION OF ALL POTENTIAL FIRE & FLOOD LOCATIONS e SYSTEM LEVELCOMMON CAUSE FAILURE ANALYSIS
- SYSTEMATIC REVIEW OF OPERATING EXPERIENCE
- DESIGN SPECIFIC BETA FACTORS ,
~
! - PROBABILITIES ASSIGNED TO ALL ACTIVE REDUNDANT COMPONENTS
O O O OVERVIEW OF SSPSA EVENT SEQUENCE MODEL STRUCTURE INITIATING EVENTS AUXILI ARY SYSTEM EARLY SYSTEM RESPONSE LONG TERM SYSTEM RESPONSE TRANSIENT EVENTS OTHER EVENTS EVENT TREE EVENT TREES EVENT TREES RT FSRAC
^
LOCA TT FCRCC >
E.7 L > LLI > LL2 P TLMFW FCRSW E 1.0L PLMFW FCRAC TMLL EXFW FET1
LOPF TMCR St , TO PLANT TMCST STATES LOSP TMPCC l SLOCA l > > SL ?
LIDC MELF LOSW- MPCC AUX LPCC MCR > > TRAN P LTI W E.2 T ACR ,
E.3 T APAB '
E.4 T FLLP l SLBI l 0 > SLI 5 E.5T FLISG LT2 E.7 T FL2SG SLBO E 1.0T FLSW ' {
TMSLB '
SLO ---->
--> SGTR 2 #
l V l >
l
O O O '
SUMMARY
OF PRINCIPAL CONTRIBUTORS TO RISK IN TERMS OF ACCIDENT SEQUENCE GROUPS AND INITIATING EVENTS FROM THE SSPSA Accident Containment Response - Group Group Fraction of Sequence Group Contributing Contribution Frequency Total Release Initiating Events Percent (mean values) Frequency Group I Early Containment Failure 2.4 x 10-6 per Early llealth .01
- Interfacing LOCA 76 Reactor Year or E f fects - Seismic 24 Once in 410,000 Ttf0 Reactor Years Group II Delayed Containment Failure 1.7 x 10-4 per Latent Health .73
- Loss of Offsite Power 40 Reactor Year or Effects, -
Transients 19 Once in 6,000 Fires 15 Reactor Years
- Seismic 15 Others 11 100 Group III Containment Intact fio llealth -
Transients 57 6.0 x 10-5 per .26 E f fects -
SLOCA 29 Reactor Year or Others 14 Once in 17,000
,3 100 Reactor Years Total 2.3 x 10-4 per 1.00 Reactor Year or Once in 4,300 Reactor Years
+-be%.'- 3 w - - - + - _ . _ . _ - -- - _ - _ _ - _ _ m ,,__ _, _e. _s.m -- - ='
- O O O l
l e
1 PSA UPDATE (1986)
! OVERVIEW & CONCLUSIONS l
i i
I
O O O I i
PSA UPDATE ACTIVITIES o RISK BASIS FOR TECHNICAL SPECIFICATIONS e CONTINUAL REASSESSMENT OF PUBLIC HEALTH RISK
- RMEPS
- SENSITIVITY STUDY
- - SEISMIC CAPACITY UPDATE s
I e ESTABLISHED RELIABILITY AND SAFETY ENGINEERING GROUP
- - MAINTAIN CURRENT RISK MODEL '
! - PART OF CHANGE REVIEW PROCESS l - EVALUATE IMPACT OF REGULATORY CHANGES
- PLANT RELIABILITY RESPONSIBILITIES J
l l
l l
O O O RMEPS OBJECTIVES e REEXAMINE TECHNICAL BASIS OF THE 10-MILE EPZ (NUREG-0396) ON A PLANT-SPECIFIC BASIS e DEVELOP AN ENHANCED PRA METHODOLOGY FOR ESTABLISHING A PLANT AND SITE-SPECIFIC EPZ e APPLY THIS METHODOLOGY TO SEABROOK STATION
- UPDATE SSPSA RISK MODEL (1983 - 1985)
- DETERMINE RISK IMPACT OF EMERGENCY PLAN OPTIONS e ADDRESS UNCERTAINTIES AND SENSITIVITIES 4
e PROVIDE DOCUMENTATION AND PEER REVIEW i
i i
i
- - - - - , - - - - _% L-- -_ . . _ . _ _ _ o,- 4-. - -
O O O PRINCIPAL INVESTIGATORS KARL N. FLEMING PLG ALFRED TORRI PLG ROBERT J. LUTZ WESTINGHOUSE ROBERT E. HENRY FAI R. KENNETH DEREMER PLG KEITH WOODARD PLG i
O O O KEY RESULTS e EARLY HEALTH RISK WITH NO IMMEDIATE PROTECTIVE ACTIONS
- LESS THAN WASH-1400 WITH 25-MILE EVACUATION
- MEETS NRC SAFETY GOAL WITH WIDE MARGIN
- CONFINED TO AREA CLOSE Tb THE SITE o VERY SMALL RISK REDUCTION BY ANY EVACUATION e ALL NUREG-0396 DOSE VERSUS DISTANCE CRITERIA
- SATISFIED AT 1 MILE OR LESS e LATENT HEALTH RISK INSENSITIVE TO ASSUMPTIONS l REGARDING EVACUATION i
- O O O I
FAVORABLE RESULTS DUE TO e CONTAINMENT EFFECTIVENESS e ENHANCED V-SEQUENCE MODEL e SOURCE TERMS 4
I i
j O ~O O :
i RESULTS AND CONCLUSIONS OF SENSITMTY
. STUDY i
1 e WASH-1400 EARLY HEALTH RISK APPROXIMATELY. MET WITH 1-MILE EVACUA.T!ON e NRCs PROPOSED INDIVIDUAL RISK SAFETY GOAL MET '
WITH NO IMMEDIATE PROTECTIVE ACTIONS o CONDITIONAL FREQUENCY OF EXCEEDING WHOLE-BODY DOSE VERSUS DISTANCE LOWER FOR ALL CASES SEABROOK
- NUREG-0396 STATION l 10 MILES 1 MILE i -
l 200 REM .03 .02 l 50 REM .12 .03
! 1 REM .30 .06 l
e 1-MILE EPZ JUSTIFIED EVEN ASSUMING WASH-1400 SOURCE TERM METHODOLOGY l
1
O O O 1
PEER REVIEW GROUP o ROBERT BUDNITZ, CHAIRMAN, FUTURE RESOURCES ASSOCIATES, INC.
i e DAVID ALDRICH, SCIENCE APPLICATIONS INCORPORATED i
e JOSEPH HENDRIE, CONSULTANT e NORMAN RASMUSSEN, MASSACHUSETTS INSTITUTE OF TECHNOLOGY e ROBERT RITZMAN, ELECTRIC POWER RESEARCH INSTITUTE l e WILLIAM STRATTON, CONSULTANT e RICHARD WILSON, HARVARD UNIVERSITY .
.i - - - - - - - - - _ _ _ - - - - - - - - - - - - - - - - - - - - - - - - - - - _ - - .
O O O PEER REVIEW FINDINGS-i e CONCURRED WITH PRINCIPAL STUDY FINDINGS
- OVERALL OFFSITE RISKS VERY SMALL
- EARLY HEALTH RISK LOWER THAN THOUGHT TO EXIST WHEN GENERAL EPZ ESTABLISHED
- EARLY HEALTH RISK CONFINED TO AREAS VERY CLOSE TO REACTOR e CONCLUSION ROBUST EVEN IN LIGHT OF UNCERTAINTIES e BELIEVE THE "BEST ESTIMATE" PROBABLY OVER-ESTIMATES ACTUAL CONSEQUENCES i
i e SEABROOK CONTAINMENT MAJOR FACTOR i
-=.-.__a._. a - - - _ . . - -_- _.u - - - -.._. m aa + w--
O O O l
SSPSA UPDATE (1986) t i
METHODOLOGY & RESULTS l
i I
i i
o O O BLOCK DIAGRAM STRUCTURE OF SEABROOK RISK MODEL r , , , ,
3
--+ ['//(((///f ivint uourucay +-- ---+
coniAinua=I eviav uoui ca ~
%i / J////
"*" % suecgrganCes sysn us wuuAn exten=AL cO=IA.nua ni Acciot ui SNIERA leon ,
MOOtts 4-MODE Ls
-6 EVlNis MOOfts M F Alt uR$
MODE L 1 sauut AllON MODE L O
%33}/
/ OUAN sf tCAllON/
.co,o ,c IMPACT WODit 4-
,o,oc ,,,gy DE MOGR APHY, .,
naima.
NMI REsPONsf
, umM utitOROLOGY
, , Moorts 0,yg PLANT
! { {
seTEa!s "
- iv'l[u'fi.oI.EEus-7//////////4 l
q PLANT MODEL y ( CONTAINMENT MODEL j ( SITE MODEL y LEGEND ORIGINAL SSPSA MODEL RMEPS UPDATED MODEL
- O ~O O i
UPDATE OF SSPSA RISK MODEL e UPDATED SSPSA PLANT MODEL
- ENHANCED V-SEQUENCE MODEL
- ENHANCED SEISMIC ANALYSIS
- - CONTAINMENT RECOVERY MODEL
- ENHANCED TREATMENT OF COMMON CAUSE FAILURES l
e UPDATED SSPSA SOURCE TERMS i
- EXISTING SSPSA SOURCE TERMS i
- INCORPORATED SOME ZION IDCOR RESULTS
- PERFORMED SEABROOK/ ZION DESIGN COMPARISON
- DEVELOPED SOME SEABROOK RESULTS WITH MAAP
- REASSESSED UNCERTAINTIES
- EXAMINED SENSITIVITIES a'
O O O ENHANCED METHODOLOGY FOR EPZ DETERMINATION e DEVELOP NUREG-0396 RISK OF DOSE VERSUS DISTANCE CURVES BASED ON PLANT / SITE-SPECIFIC RISK MODEL e CHARACTERIZE TOTAL POTENTIAL FOR RISK REDUCTION BY PROTECTIVE ACTION AS RISK WITH NO EVACUATION e QUANTIFY SPATIAL DISTRIBUTION OF NONEVACUATION RISK e CALCULATE ACTUAL RISK REDUCTION FOR PROTECTIVE ACTION STRATEGIES MILE EVACUATION MILE EVACUATION 10-MILE EVACUATION MILE EVACUATION AND SHELTERING OUT TO 10 MILES I e EVALUATE UNCERTAINTIES AND SENSITIVITIES e COMPARE RESULTS WITH ALL AVAILABLE RISK ACCEPTANCE CRITERIA
O O O i i
j EMERGENCY PLANNING SENSITIVITY STUDY i METHODOLOGY J '
o PURPOSE:
DETERMINE IMPORTANCE OF SOURCE TERMS VERSUS PLANT-SPECIFIC FEATURES AND ENHANCED PRA TECHNOLOGY '
l
- APPROACH: RMEPS CALCULATIONS REDONE USING:
I WASH-1400 SOURCE TERM METHODOLOGY i
- BEST ESTIMATE ASSUMPTIONS ON ALL OTHER UNCERTAIN RISK PARAMETERS I
1
l o o o i
RISK ACCEPTANCE CRITERIA UTILIZED i
e NUREG-0396 DOSE VERSUS DISTANCE CURVES FOR 1,5,
!j 50, AND 200-REM WHOLE-BODY DOSES e WASH-1400 RISK CURVES FOR EARLY AND 1
LATENT HEALTH EFFECTS (MEAN & MEDIAN RESULTS)
! e NRC INDIVIDUAL AND SOCIETAL RISK SAFETY GOALS i
l e SPATIAL DISTRIBUTION OF RESIDUAL RISK i
l
O MEDIAN RISK OF :
EARLY HEALTH EFFECTS FOR ,
DIFFERENT EVACUATION DISTANCES l 10-3 i
- L END 10 -
--- SEABROOK STATION PER RMEPS AND WASH-1400 SOURCE g
< TERM METHODOLOGY y (MEDIAN RESULTS)
WASH-1400 d: 10-5 o -
(MEDIAN RESULTS) o
.O i
=
0- ~% NO IMMEDIATE PROTECTIVE N ACTIONS 8
5 10-8 o
T '~ ' N i-uitE N
N
~
s f EVACUATION
!z 10-7 -
N
\
\
8 \ g l e \ ;
' I
\
\
\
ir8 -
\ \
- ~ ---. % [ 2 - ui tE EVACUATION
\ {
RMEPS RESULTS N \
OFF SCALE N
! 30-9 i i \ iiI i 10 0 10 1
10 2 10 3 4 10 S 10 EARLY FATALITIES
l O ~O O
! THE BENEFITS OF RISK REDUCTION BY
- EVACUATION OR SHELTERING ARE:
I e VERY SMALL DUE TO VERY LOW INHERENT PLANT RISK
)
j e FULLY REAllZED BY CLOSE-IN EVACUATION
- NOT NEEDED TO MEET NRC SAFETY GOALS l
i WRC 5Af ETY CO AL FOn a w He LE
- Of Slit souseDARY _
80-3 E
, m .
. ie d - -
l.
3i
, =
1 E 2 ! -
a i l O F ,,4 - -
neoucteose i.e z.uiLE L E A48eG TO anet 5 i i
) g y i I I I e 2 4 S S #
EVACUAT1080 De5TAseCE 188tLES3
j O O O COMPARISON OF INDIVIDUAL RISK .
i 1
WITH BACKGROUND AND SAFETY GOAL VALUES 10'2 3
i 9 BACKGROUND ACCIDENTAL FATALITY RISK 10 (5 FATALITIES PER 10,000 POPULATION PER YEAR) e 10 4 -
j N
ta.
N 10-5 9 SAFETY GOAL (.001 TIMES i
I 8 BACKGROUND RISK)
[ SENSITIVITY STUDY O 10~6 -
WITH NO IMMEDIATE
& PROTECTIVE ACTIONS m
10 WITH 1 MILE
, /,
EVACUATION
$ RMEPS RESULTS z 10 WITH NO IMMEDIATE l
- PROTECTIVE t.CTIONS
] 10-8 i
i i
O O O
) COMPARISON WITH NUREG-0396 - 2OO-REM & 50-REM WHOLE-
- BODY DOSE PLOTS FOR NO IMMEDIATE PROTECTIVE ACTIONS 5
_ . . .....i . .
. .....i . . . . . . . . .
.t - -
[ NUREG-0396 [
. E$ _
SENSITMTY STUDY -
i 4
o E__
o _ -
O p . . . ... . ... . ... .. R M E PS RESULTS
@4 -
(200 REM CURVE OFF -
, ma SCALE) sw n1 i
o us
! jg 0.1 -- --
ou -
z< - -
BZ wm
~
~
h -
.i Da
- o. o
- \ -
og g 1 50 HEM
>r 200
\ -
E!O _ \ REM
' ~
Do Om
\ \g
\
l EC O.01 -
g \ T wG : a -
am _
\ \ -
\
l $m < . I -
! Es -
l \ -
1 os 1 _
$$ l l "
)
i
_\
O.001
[- i e i iiil e i e i i s ial i n 8 ' i e
3 10 100 1.000 DISTANCE (MILES) l
O O O
- COMPARISON WITH NUREG-0396 REM & 5-REM WHOLE-BODY DOSE PLOTS FOR NO IMMEDIATE PROTECTIVE ACTION I i e iiig i ii is_
. i e i iiis l i i i i
~
Oz un w
- - - - - - SENSITIVITY STUDY -
ge - -
i >u RMEPS RESULTS o4 . -
i OU
$W a3 1 REM I
om 5 HEM -
- I cc O.1 -
Rou - -
j - -
l o -
24 -
- .s.z -\
g2 -
- x mm C A . \_._ ___ b_ --- N s
! . . \ -
4 u. u . T oz S. \ g
\
uh
. \
55 \
l
=
no o us **.
\. %*
g \.
\
W E-II-0.01 -*
- I s
\
7 -
I
\
1
$@8 04
\ \
l g _ %.,
., \1 REM -
4 ze - . .
! on %, . g o ., ., 1 REM -
] _ ,
l 5 REM ,
, g l a"i i s 'n : Il il i e !eil i i ' ' ' ' ' '
ioo i.o m l i io l DISTANCE (MILES)
o o o l
l Safety Study Results:
CONTAINMENT EFFECTIVENESS i (Percent of Accident Scenarios) i i
66 % 99 % 99.9 %
i l }
34 % 1% .1 %
i
! WASH-1400 (1975) l SEABROOK STATION (1983) SEABROOK STATION IMPROVED LOCA OUTSIDE CONTAINMENT M00EL (1985)
~ " EARLY i DEGRADED CONTAINMENT q
FAILURE CONTAINMENT OR CONTAINMENTINTACT i
i ____
~
l O O O PRINCIPAL CONTRIBUTORS TO EARLY RELEASE FREQUENCY
! INITIATING EVENTS j WITH CONTAINMENT
! BYPASS
- INTERFACING LOCAs #
! - STEAM GENERATOR l TUBE RUPTURE i ~
)
EXTERNAL EVENTS WITH POTENTIAL l
l CONTAINMENT DAMAGE FOR I
- AIRCRAFT CRASH > EARLY
- TURBINE MISSLE RELEASE
, m LOSS OF CONTAINMENT m 1 STRUCTURAL INTEGRITY ALL OTHER INITIATING 2 EVENTS CONTAINMENT ISOLATION .
FAILURE
I O O O 1
l CONTRIBUTIONS TO EARLY RELEASE j FREQUENCY (S1, S6, S7) i l SCENARIO TYPE PERCENT l
- 1. EARTHQUAKE WITH OPEN PURGE 87
- 2. INTERFACING LOCA 13 i
! 3. TURBINE MISSILE IMPACTS CONTAINMENT BUILDING <1
- 4. REACTOR VESSEL STEAM EXPLOSION <1 l
1 l 5. AIRCRAFT IMPACTS CONTAINMENT BUILDING <1
)
1 TOTAL 100 1
I 1 ___ _
l l O O O I
l l
ENHANCED TREATMENT OF INTERFACING '
) SYSTEMS LOCA l
i e MORE COMPLETE MODELING OF VALVE FAILURE MODES i
e NEW DATA ON CHECK VALVE FAILURES VERSUS LEAK SIZE
! e MORE REALISTIC TREATMENT OF DYNAMIC PRESSURE PULSE e EXPLICIT MODELING OF RHR RELIEF VALVES o QUANTIFICATION OF RHR PIPING FRAGILITIES TO
! OVERPRESSURE
{ e MODELING OF RHR PUMP SEAL LEAKAGE e OPERATOR ACTIONS TO PREVENT MELT CONSIDERED i
e THERMAL HYDRAULIC AND SOURCE TERM FACTORS i
MODELED USING MAAP e UNCERTAINTIES QUANTIFIED 1
1
O O O i
i' i i i
1 i INTERFACING SYSTEMS i
i LOCA KEY RESULTS l
FREQUENCY (PER REACTOR-YEAR)
, UPDATED
! EVENT SSPSA ANALYSIS ,
l l
i VALVE RUPTURES, LOCA 1.8 x 10 -6 7.8 x 10 -6 VALVE RUPTURES, LOCA, -6 1.8 x 10 3.1 x 10-l CONTAINMENT BYPASS VALVE RUPTURES, LOCA, 1.8 x 10 -6 4.1 x 10 -8 CONTAINMENT BYPASS, li MELT i
i
O O O i
- CONTAINMENT FAILURE TYPES i
I A. SMALL LEAK (0.02 SQ. INCHES TO 6 SQ. INCHES) j; PRESSURE RISE CONTINUES 1
i
- B. LOCAL FAILURE (6 SO. INCHES TO 60 SQ. INCHES) i PRESSURE RISE CONTINUES l LEAK RATE INCREASES UNTIL PRESSURE RISE i STOPS I
i C. GROSS FAILURE ( > 60 SQ. INCHES) 4 RAPID CONTAINMENT BLOWDOWN ( < 1 HOUR) i j
l O 'O O i
l i
l 1
l 1
LOCAL CONTAINMENT FAILURE MODES CONSIDERED i
l I
e FLUID SYSTEMS PENETRATION i e HIGH ENERGY PENETRATION i
i e FUELTRANSFER TUBE l 9 ELECTICAL PENETRATION e PURGE LINE PENETRATION e PURGE VALVE SEALS S EQUIPMENT HATCH l 8 PERSONNEL LOCK i
e OTHER PENETRATIONS i e LINER TEARING i
e WELD IMPERFECTIONS 1
)
l
O O O COMPOSITE CONTAINMENT FAILURE PROBABILITY' DISTRIBUTIONS FOR TYPE B (LEAK) FAILURE, TYPE C (GROSS) FAILURE, AND TOTAL FAILURE j .
1 = ;y ::.~c .
-=;m -
- f- ---
. p / TYPtB(L(AK)
/
- F.A.I.LU..REoo c., .
, /
i *
. /
- ? *
~ *
, / %vvre e scaossa
! ~
/ 'w'sSE= cts 1 -
f l -
- s
. i l /
- /
- w : : 2
/
i 3 2 : /
I i 3
! /
. ^
- : I.
G - : 1 2 : I -
g
-[ . I I
i w' :: I
- 1 3
- /
-: I
/ - **
f 4 I i f I 8
g
,1 .
= , ;
, I :
1 1 5 f - ai g i /
i - /
t i
] "g e / I I I I J = u. = = m m = .
ratssunc irsiAi
) -
4 i
O O O l
i i
SUMMARY
e EARLY HEALTH RISK VERY LOW EVEN WITHOUT
- IMMEDIATE PROTECTIVE ACTIONS l
I e BENEFITS OF EVACUATION VERY SMALL & CONFINED TO AREA CLOSE IN TO SITE l
e PEER REVIEW GROUP CONCURS WITH RMEPS &
l SENSITIVITY STUDY CONCLUSIONS i e SEABROOK RISK MANAGEMENT ACTIVITIES ARE l CONTINUING
i . l T
O 10 mil _EEkERGENCYPLANNIllGZONE:. -
UPON4PRINCIPALCONSIDERATIONS -
~~
J i 1) DBA DOSES LESS THAN PAGs EEYOND 10 MILES
! 2) MOSTCOREMELTACCID$NTDOSESLESSTHAN '
I PAGs EEYOND 10 MILES I
- 3) FOR WORST ACCIDENTS -- PROMPT DEATHS ' ;
GENERALLY WITHIN 10 MILES O
- 4) 10 MILES - BASE AREA EXPAND AD-H0C IF fiECESSARY .
a p i
l
\
I i
l l0 '
l i
. . _ _ _ _ , ,..-.._____..___,[,. . , . . _ . . , , . - - . . . - . . _ . . . . , _ . . . - . _ _ _ _ . .._.___.,e _ . . . .__ - . . . . . . - .
'x x
'T , ~
/
z s
C_yb -
~
. S.
Nx
\
gw%smy 35*
m
%o g y, e , -
3 m.sp I
n& 4 c C gg gfj,:-
g }f N y/ A. i~ c
' g - \
A' 8
\j
~, - '+
- ~
. x ~.
w p ( : k ) =
fk
- O
- 9. ,, .% [,rg S.pwhM dN N d
"h ,Y ,$ ,
h(wI" T!{,l"f,L ?.
- 4a$ f
- f O
wzu 4
~~:
6:h@
l
[
I
) , - g.g 4 i s\
y.as x' :- '
ca
%, ~"~~~
ff x.
,Q, ..<s* ;
N"'
_g: qw /
,3, i
g,: ,
..p.-
SEABROOK STATION TWIN PRESSURIZED WATER REACTOR UNITS H
= EACH GENERATING 1,150 MEG AWATTS OF ELECTRICITY ,
- SEABROOK, NEW H AMPSHIRE
! .M @-
z v; Q1 ,
+
- + Q ?& , 7. :. : ,8 b?Q .O N 4
/ g Y~' ,
/+&N%
1-.
'n%g' e<*f w 'y c,_.A i
N 4 e s s x ORg J, 5
? g f ?h y .
s , [ %
\
- .A i-U I ::
a
/
A AA y 3 e
{^ ,
o y
p i
. ,2 4 N . O %. O ,
' *i s !
?
Q g '
3
+ i ,,
f 3l --
- . s % *o a d m erc e i1 D- , Mg \ -
{' (h h 1
-t gg }
O l
0 '
4L 9 - 2 Also Availdle Aary comoone,icoonn2
'Oh M ^ j IL g' %s
/ 'ANrture Cr ' '* *ater storage tank to Ae%eung "9
55 Soray chemica4addmon tank ...
k .,
g(O*j '
- c *aw r+
V4' ncr c
. awes
'A Recovery test tanks (2)
Waste test tarks (2) C.
s Se Peac'or water make up tank l
Z ,
w %;vcyecnysc7"' 59 Boron waste storage tanks (2) ?
i p- . g ' : ab r e's g p., mary dram tank (2)
. ,g : 4 'pA wr 6' Ae sted dra'n tanF (2)
'~~
7A; % ,- .o 3 '*. i s ov. c e ^'e' 62 Aesm waste tank y j^ . . e a -, g-c 3 .ar: o sc ++1ng m- 63 Sona waste processing y>
g 7
(h ,%-
.y, p.
3 3
3'
/r LA U gaa ta"e'es 64 Waste ha;'d degassifier 65 Steam generator blowdown
- ;,c- 4 4 ; ese gere a'1 recovery u n it ..
p we+ y , '.e s ra,e an* 66 Cochng tower (service water) 4 g .c 9- Yanrq- p4 m esso' pur ps
- - y- .. ~ s t , n esu o , . (, a s, e,c e. 6 7 Coobng tower ceramic fm e- 4 - " New' ' 'wr 68 Fans
[ ,; Y s w .: , Va car ~9< > 'me, ' o se'cana 69 Goosing tower swrtcngea-
[r j.;f p 'r9- *a .r c ta 4 Eas= icao og a ea 70 Circu;ating water pumps gf
- wm g . sn' e - Sre"s a s'c r age 71 'ra <ehng screens
,,y 4 e ~ e, p w - a & c /* ' / c ' es soen ae coc %at 72 intake fh.,me and forebay
, fy.^' , 4 Ome4 e, na"ger r 3 Discharge vanes tor back P ~+ . E = Sr L J.
7 Y e ear-
.se 6 **- a seca r v era-4< S >ar' Le casa sneng area 44 Ba hush control 74 Discharge r ser shaft i
, y.,. - . , x g 6.r- ."c' .#
~v a 're 45 B "c :aca ama ms'o' age tanks nina 75 Derrureranzed mater s:
E _s #g' "vs , S'er # vr: 'ae 'ts 46 t e' dcwr aegassaie- storage tanks I
, Key to drawing se. . ja ~ sg. ee a e, y e con n .an* 76 Sea wan If a % 4 e. , -en ,, "e c.r sw. s y c ess ce*asqw e m a8 P"r"a y c'ar1 <entoahon 77 Gas insuiated transmission ";
, aa. ,
u -
-c -
g reaws e suge a r sspaw kne n.-
F B - e: .
c 5'ar w , :.w a w Sec raa', c ' oc ' e ' , 49 Ausn.a v ovoc'rg e,ravst 78 Auchary transformers (4, ;
' e r -w- 79 Step oo trarsformers(3) "
- e .c e m, ~ : .s o ~g"ea'ar a g- s +r 'me s --
er- a, va ces, - s . Jw a'e' "e 22 w r/ess/ewea w a 'e' 5 Le'oewr de rmeranzers 80 Ashary bouer L 1 Aa- o< .a,, en - ce co~ % rece's and f oters 81 Water treatment pit nt
[ E. -
4 e' e tso a 2 3 Nor seq ega'ed enase s'r.or '
Pwary compocent cocong 82 Soner stack
_ r D ese g- - < N a, ,E e< gen , - .oc g se". ce om s opps 83 Condensate storage tank
= w p m u y,t w g a: c, e< ' s a <a Hea e tav /are 52 Pr,rna'y component cochng 84 Primary vent stack mF e. storage cura na - So- ca 25 va " s'eam com y <a <es eroansion tanks 85 Service water pipeng r
{ .
Moo 90oosW -
O O
O m
m PROBABILITY OF EXCEEDING WHOLE BODY DOSES GIVEN A CORE MELT ACCIDENT E o m o r O m a d I i i iiilij i l I i 17 ll 1 1 I fl l 'Ill >
O O
l0 l
Dm 1
cs 2
- OH
- CW
- HO
- a>
or O
gO
_C a rr g
a y
g m>
~
d $
l a
m m m 'o o E $/ 2 m %2 m -
2 g e 5
mm p - m r m - E cl us
~
Z Em 5 ~- mE
- mm X%
- oO om m2
~
CO
- m -<
msF g-o .
2 2
2 O
N O
. 2 m-O i
O O O -
? .
1 CORE MELT ACCIDENTS CALCULATIONS EMERGENCY PLANNING ZONE
- OF ABOUT 10 MILES FOR PLUME EXPO About 30% chance of exceeding PAG doses at 10 miles Less than one chance in 100 of exceeding life threatening do~ses beyond about 10 miles r
O l
NRC STAFF PRESENTATION TO THE ACRS
.. ON THE REVIEW PLAN FOR THE SEABROOK E E RGENCY PLANNING SENSITIVITY STUDY F
SCOPE AND FOCUS OF STAFF REVIEW S. LONG (NRR)
COMPARISON OF PLG-(A56 WIm NUREG-0396 D. MATTiens (IE)
ItTHODOLOGY AND STATUS OF REVIEW:
SOURCE TERMS T. PRATT (BNL)
O RiS o m ySiS - -
CONTAINMENT STRUCTURAL INTEGR!TY C. HOFMAYER (BNL)
CONTAINMENT BYPASS R. YOUNGBLOOD (BNL)
INTERFACING SYSTEM LOCA " "
Y O
PURP SE OF REVIEW O
TO IDENTIFY AND REVIEW TE PORTIONS OF THE STUDY THAT ARE MOST SENSITIVE WITH RESPECT TO Tl!E STUDY'S PRINCIPAL CONCLUSIONS INDIVIDUAL RISK OF EARLY FATALITY AT SEABROOK IS WITHIN SAFETY G0AL 1 MILE EVACUATION AT SEABROOK PROVIDES SIMILAR RISK 0F EARLY FATALITIES TO THE WASH-IIK)0 RESULTS WITH 25 MILE EVACUATION ,
PROBABILITIES OF SPECIFIC RADIOLOGICAL EXPOSURE LEVELS AT 1 MILE FROM SEABROOK ARE LESS THAN THE C0fR 5 PONDING PRGBABILITIES SHOWN AT 10 MILES IN NUREG-0396 l
O l
- . - . - . - . - - -~ .- -__ _.. . .. - - - - - _ - - - - . - - _ . - _ _ _ _ _ _ - _ _ _ _
f
~
8 i
i e i :
t l
! BASES FOR COPPARISONS a
L0 WASH-1400 SOURCE TERM PETHODOLOG i i EARLY FATALITIES WHOLE BODY DOSES ,
I L
I l .
f I
6 L
!O 1
1 i
l 1
I l I i
i, 1
l t
i l
i O
l 4
1 i
7
-wrwy-wy-m g, um * . - - - - - -w-. vm
- w y-wgypg%9y + -y=y-Ty-rer,-w-twqp g .y p ww-ywyw-w--rwym-v-Wg'wr t -w--w'M s Fr~TW'T"T- t'-w 7"Y w -hmT'"'VMMDTT't W*# w 4 W~& "'W ''t*Y- TY'-'MWW74
, MAJOR CONTRIBUTORS TO RISK IO j DIFFERENT CONTRIBUTORS FOR DIFFERENT RISK COPPARISONS:
_. PROBABILITY OF EARLY FATALITIES, GIVEN EVACUAT' ION, APPEARS TO HAVE CONTRIBUTIONS FROM SEVERAL RELEASE CATAGORIES AND EVENT INITIATORS.
EVENT V WAS DOMINANT IN ORIGINAL PSA, DOSE VS DISTANCE CURVES (N0 EVACUATION) ARE DOMINATED BY SINGLE RELEASE CATAGORY AND SEISMIC EVENT INITIATORS.
THESE CURVES WERE NOT PRODUCED IN ORIGINAL PSA, O
t FOCUS OF STAFF REVIEW EFFORT O
~
1 FACTORS SHOWN BY CURRENT STUDY TO BE IlPORTANT FOR RISK MITIGATION .
6e AREAS WHERE SIGNIFICANT RISK REDUCTION OCCURRED BETWEEN UPDATE AND ORIGINAL PSA ;
1
.i
{
l 6
O i
e W
9 l
O
0 0
PLANT DESIGN FEATURES SIGNIFICANT TO REVIEW O ,
CONTAlWENT STRUCTURE RHR VAULT 6
O l
l 1
O
I MODELING FEATURES SIGNIFICANT TO REVIEW
- O j
CONTAIWENT RESPONSE AT HIGH PRESSURE CONTAIWENT BYPASS ASStPPTIONS ,
INTERFACING SYSTEMS LOCAS COMPLETENESS CHECK VALVE FAILURE DATA SOURCE TERM REDUCTION FROM SCRUBBING l
OPERATOR RECOVERY CREDITS EVENT V
- o STATION BLACKDUT i l 1 l 4
6 i
j O
6
-7, _ -, - __ e _,-_,.-_.__r , _ _ . _ , ,_.-,...,m ~.
,t .
O STATUS OF BNL REVIEW 0F SEABROOK STATION EMERGENCY PLANNING SENSITIVITY STUDY DEPARTMENT OF NUCLEAR ENERGY BROOKHAVEN NATIONAL LABORATORY UPTON, NY 11973 O
PRESENTED TO A JOINT MEETING OF THE ACRS SUBCOMMITTEES ON OCCUPATIONAL AND ENVIRONMENTAL PROTECTION SYSTEMS AND SEVERE (CLASS 9) ACCIDENTS SEPTEMBER 26, 1986 ;
l 1
- \
Q BROOKHAVEN NATIONAL LABORATORY l} gj j A5500ATED UNIVERSITIES, INC.(I Lll l l
O PRESENTATION OF MATERIAL RISK PERSPECTIVE W. T. PRATT CONTAINMENT STRUCTURAL INTEGRITY C. H0FMAYER l
REVIEW 0F ACCIDENT SEQUENCE FREQUENCIES R. YOUNGBLOOD 1
O i
i Q BROOKHAVEN NATIONAL LABORATORY l}l}l A5500ATED UNIVERSITIES, INC.(Illl I
~ '
O BACKGROUND SEABROOK STATION PROBABLISTIC SAFETY ASSESSMENT (SSPSA), 1983
" FRONT END" REVIEW: LLNL
- BACK END" REVIEW: BNL (NUREG/CR-4540)
SEABROOK STATION RISK MANAGEMENT AND EMERGENCY PLANNING STUDY, PLG-0432, DECEMBER 1985 SEABROOK STATION EMERGENCY PLANNING SENSITIVITY STUDY PLG-0465, APRIL 1986
] BROOKHAVEN NAT!0NAL LABORATORY l)l)l A5500ATED UNIVERSITIES, INC.(llll
O OBJECTIVES REVIEW 0F THE EMERGENCY PLANNING SENSITIVITY STUDY FOR SEABROOK FOCUS IS COMPARIS0N OF SEABROOK SPECIFIC DOSE VERSUS DISTANCE CURVES WITH GENERIC CURVES FROM NUREG-0396 SEABROOK DOSE CURVES BASED ON:
O SSPSA WITH UPDATES IN PLG-0432 WASH-1400 SOURCE TERMS i
l Q BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(I Ul
. __ ~_- - . .. .
g O '
l BASIS OF DOSE VS DISTANCE CURVES DOSE VS DISTANCE CUP.VES GENERATED FOR EACH RELEASE-CATEGORY PROBABILITY OF EXCEEDING GIVEN DOSE IS MULTIPLIED BY APPROPRIATE RELEASE CATEGORY FREQUENCY ALL PRODUCTS ARE SUMMED TO GIVE TOTAL PR BABILITY O F EXCEEDING GIVEN DOSE TOTAL PROBABILITY OF EXCEEDING GIVEN DOSE NORMALIZED TO TOTAL CORE MELT FREQUENCY I
l
- .l 1
Q BROOKHAVEN NATIONAL LABORATORY l} gj l A5500ATED UNIVERSITIES, INC.(Elll 1
I
-1 - - - -- .
__ . _ . _ . _ ._ _ .. - . . _ . .___. ._ ,- _ . _ = _ . _ . . .
4 O
i TABLE 4-1. SOURCE TERM CATEGORIES Source Analyzed Analyzed Term Containment Failure Mode in the in This Category SSPSA Study .;
i !
S1 Early Containment Failure Yes Yes
^; S2 Early Increased Containment Leakage Yes Yes S3 Late Overpressure Failure Yes Yes
- S4 Basemat Melt-through Yes No*
SS Containment Intact Yes Yes S6 Containment Not Isolated Yes Yes- '
- O s7 coateiameat erPessed (v-se9#eace) "o ves
- Based on the SSPSA res u lts , basemat melt-through sequences were assigned' to category S3 in this study.
(Taken from PLG-0465, April 1986) 4 O BROOKHAVEN NATIONAL LABORATORY l}l3l ASSOCIATED UNIVERSITIES, INC.(I Ell
~
-- ,,,,v ,n,- , , , , ,
-. --..mm - . - - - - . , , s.n~. -- - - - -
.c.- - - - - ..,,,.m.,w
O O O TABLE 4-3. RELEASE CATEGORIES FOR SEABROOK STATION BASED ON WASH-1400 SOURCE TERM METHODOLOGY t Release Release Warning Energy Release Fractions R h se Time Duration Time Release Category I-2 TE BA RU LA (hours) (hours) (hours) (MCA/S) XE 0.I. CS S1W 2.5 0.5 1.0 11.9 0.9 7-3 .7 .5 .3 .06 .02 4-3 S2W-1 4.8 2.0 0.5 0 .03 2.1-4 4.3-3 .023 4.2-3 2.8-3 8.4-4 8.4-5 S2W-2 6.8 4.0 2.5 0 .07 5.0-4 1,3-3 .048 .039 5.5-3 3.4-3 5.2-4 S2W-3 19.8 18.0 15.5 0 .023 1.6-3 2.3-3 .126 .147 .014 .011 1.9-3 TOTAL 4.8 24.0 0.5 0 .123 2.3-3 7.9-3 .20 .19 .022 .01 5 2.5-3 S3W 6.0 24.0 2.0 0 4.7-4 3.3-6 3.2-5 1.7-4 1.5-4 1.9-5 1.2-5 2.0-6 56W-1 1.75 1.0 1.5 0 .15 1.1-3 .10 .11 .02 .014 4.1-3 4.1-4 56W-2 2.75 4.0 2.5 0 .42 2.9-3 .07 .19 .063 .022 .009 .001 56W-3 15.75 18.5 15.5 0 .32 2.2-3 .01 .13 .32 .011 .020 3.8-3 TOTAL 1.75 23.5 1.5 0 .9 6.2-3 .18 - .43 40 .047 .033 5.2-3 S7W 8.5 7.0 2.0 0 .9 7-6 7-4 5-4 3-4 6-5 2-5 4-6 NOTE: Exponential notation is indicated in abbreviated form; i.e. 7-3 = 7 x 10-3, Taken from PLG-0465, April 1986.
BROOKHAVEN NATIONAL LABORATORY l} gjl
' A5500ATED UNIVERSITIES, INC.(IIll
! 1 ll yl gI O l(
Y R
}I C
N O
T I
A S, R E O
I T
B I S
i2 3 4 W A R g "6 6 6 -
1 L E V
g 1 1 1 g2 S L A I g 0 m., .
3 Y R
O G
E )7 N
O I
T A
N N
U D
E T
A T 4 N 0
- A 4 E 0 C R V 5 A 5 E BE H A K
S O A M E U O l1 L N R B
E R UN R R
^ N 1
O(
F E O)
N .
6 V I6
,'1 R T8 U C 9
- )
C A1
. S E El
_ E C IVir L N I A Tp
_ 1 t T CA
( S E x
I T ,
E C
D R6O5 O .
N A
T S
S P40 U
S E -
R TG E AL I
D V I D P
E E m 6
8 SM O Mr o i'1 I f D
O n
.N ke _
1 R a A FO (T
_ =
E R
1 U _
NtN G _
EEi I RRl F 6
6.
6 6
6i
.l -
66 : .
25f 1 x+: 6
- 1 3 4 6
O 6 8 6 I 1 1 1 g g3 * =o e5
- gEi$ E >D _ E ain.
- i -
O 3 '
O O O 3
16 : _
If S.
w 2$
5 7 .g _g R 16 __ __16 w
E z:
5 5 -2 2 16 __ ..
__16 O
w
$$ a w
E5 -3 -3
> 16 _ __16 d
[
a E x266.06 REl1
<z +56.66 REf1
@ 5.66 ret 1 oc 1.06 REli 16 i i 16 16-1 '168 '161 162 DISTANCE (t1ILES)
, FIGURE A-2. DOSE VERSUS DISTANCE CURVE FOR RELEASE CATEGORY S6W FOR NO IMMEDIATE PROTECTIVE ACTION (RUN NUMBER 452)
(Taken from PLG-0465, April 1986). ,
BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl
O O O 18 - if S
=
w id 8 3
-1 -t E 18 __ - 18 N
g ..
m 5 l .
5 -2 l g
-2 g ,- ,e 2
w k i 18 18 g
( ,e 2
cc x 266.86REN
+56.08 REN
@ 5.86 REN m: 1.06 REN 18 i '
19 18-1 'l aa 181 182 DISTANCE (NILES)
FIGURE A-3. DOSE VERSUS DISTANCE CURVE FOR RELEASE CATES 0RY S2W FOR NO IMMEDIATE PROTECTIVE ACTION (RUN NUMBER 453)
(Taken from PLG-0465, April 1986).
BROOKHAVEN NATIONAL LABORATORY l} gy l A5500ATED UNIVERSITIES, INC.(IIll
O O
~
O 16 1 3 ,
o co
-1 -1 E 16 __ - 10 M
^
8 -
= N a ~
16~ --:16 g -
E d
M w
Ei -3 -3
> 18 -- 7 16 e
U I E x266.06 Rett l cx +56.06 Rett
@ 4 5.06 Rett cx 1.00 ret 1 o.
4 .
4 16 ' ' 16 10-1 '186 i
'gt 182 DISTANCE (t1ILES)
.F FIGURE A-4. DOSE VERSUS DISTANCE CUP.'!c FOR RELEASE CATEGORY S7W FOR NO IMMEDIATE PROTECTIVE ACTION (RUN NUMBER 454)
(Taken from PLG-0465, April 1986).
i i
BROOKHAVEN NATIONAL LABORATORY l)l)l A5500ATED UNIVERSITIES, INC.(IIll
O O O 1
18 3
8 m
w id s -1
-1 18 R 18 _
E :
z:
5 U .a -a m ia _-._ _ .te S
w Ei -3 -3
- 18 3 18 ---
x 266.68REf1 a +56.06 REli
@ 5.00 REN
= .t.06 R E ti .
n.
4 4 18 i 18
, 18-1 't a'e 't a l tea DISTANCE (t1ILES)
! FIGURE A-5. DOSE VERSUS DISTANCE CURVE FOR RELEASE CATEGORY S3W FOR NO IMMEDIATE PROTECTIVE ACTION (RUN NUMBER 455)
(Taken from PLG-0465, April 1986).
! BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(IIll
.i i
TABLE 4-2. REVISED C-MATRIX FOR NEW SOURCE TERM CATEGORIES O nt S urce Term Category ge i State S1 S2 S3 SS S6 S7 (frequency) 1F 1.0 (2.0-8) ,
(2.0-8)
IFV 1.0 ;.
(4.6-9) (4.6-9) 1FP 1.0 (1.4-6) (1.4-6)
IFPV 1.0 (2.7-8) (2.7-8)
.2A 3.4-5 1.4-4 1.0-2 0.99 (1.9-6) (6.5-11) (2.7-10) (1.9-8) (1.9-6) 30/70 2.0-6 8.0-5 0.95 0.05.
(3.8-5) (7.6-11) (3.0-9) ( 3.6-5 ) (1.9-6)
O 3r/78 (3.0-7) to
( 3.0-7 )
3FP/7FP 1.0 (1.9-5) (1.9-5) 4A/8A 3.1-6 1.3-4 5.2-3 0.995 (1.1-4) (3.3 10) (1.4-8) (5.5-7) (1.1-4) 7FPV 1.0 (1.2-8) (1.2-8) 80 1.1-6 3.1-5 0.9999
( 1.0-4 ) (1.1-10) (3.2-9) (1.0-4)
Total 5.2-9 2.0-5 1.4-4 1.1-4 3.2-7 3.9-8 Frequency NOTES:
- 1. Exponential notation is indicated in abbreviated form; i .e., 2.0-8 = 2.0 ~x 10-8
- 2. Numbers inside parentneses are unconditional frequencies (events per reactor year) based on mean values. Numbers not inside parentheses are conditional O rre9"e"cies or so"rce ter= cete9ecies 94 ve" t"e 4"dicetee Pie"t e e9e stete.
also based on mean values. Median values of source term categories are. presented in Section 3.
(Taken from PLG-0465, April 1986).
- . _ . w - ., , _ , -
1
. i a i ie iaig i a ie iisig i i i i ie i s.
$$ _ ----- THIS STUDY FOR _
gg ~
SEABROOK STATION OQ c<
. .. ... . ... . . .. .. R M E PS R ESU LTS FO R
~ SEABROOK STATION _
@b -
(200 REM CURVE OFF di w SCALE) a3 om jg 0.1 --
T eu -
z<~ - % -
Ob 5 -
U$
X T-'N -
- m' - \ -
05
>s \ 200 k 50 REM
\
052 \ REM
- u. o 3 0 \
$ 0.01 -
k -
25 i \ \ :
$A o<
- l \ -
g ea -
1 -
h, 05 oc r*
l l
_ \., l
,l s 50 REM
/L 200 REM V '
\
'I ' ' 'I ' '
0.001 1 10 100 1,000 DISTANCE (MILES) 1 FIGURE 2-3. COMPARISON OF SEABROOK STATION RESULTS IN THIS STUDY AND RMEPS WITH NUREG-0396 - 200-REM AND 50-REM WHOLE BODY DOSE PLOTS FOR NO IMMEDIATE PROTECTIVE ACTIONS (Taken from PLG-0465, April 1986) l l
l l
l O BR KHAVEN NATI NAL LABORATORY l}l3l A5500ATED UNIVERSITIES, INC.(1 Ell 1
O COMMENTS ON THE SEABROOK DOSE VS DISTANCE CURVES RELEASE CATEGORY S2 IS ONLY. CONTRIBUTOR TO CONDITIONAL FREQUENCY OF EXCEEDING 200 REM DOSE RELEASE CATEGORY S2 DOMINATED BY SEISMICALY INITIATED EVENTS WHICH RESULT IN EARLY INCREASED CONTAINMENT LEAKAGE ALL ACCIDENTS THAT MIGHT RESULT IN RELEASE CATEGORIES (S1 AND S6) ARE LOW FREQUENCY (< 3x10-7)
O THIS IMPLIES LOW FREQUENCY FOR INDUCED EARLY CONTAINMENT FAILURE BYPASS OF CONTAINMENT FUNCTION LOSS OF CONTAINMENT ISOLATION INTERFACING SYSTEM LOCA Q BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(llll l
l l
l
O FOCUS OF REVIEW CONTAINMENT STRUCTURAL INTEGRITY MAINTAINING CONTAINMENT FUNCTION:
CONTAINMENT ISOLATION FAILURE INTERFACING SYSTEM LOCA C MPLETENESS:
O OTHER ACCIDENT SEQUENCES THAT MIGHT LEAD TO EARLY LARGE FP RELEASE EXAMPLES:
INDUCED STEAM GENERATOR TUBE RUPTURE ACCIDENTS FROM SHUTDOWN WITH CONTAINMENT OPEN l
l Q BROOKHAVEN NATIONAL LABORATORY l} g)l ASSOCIATED UNIVERSITIES, INC.(llll l
i
).
8 t
i
!O 4
i i :
i i I
! l i
- i i !
)
- l i
i CONTAINMENT STRUCTURAL INTEGRITY !
I !
C. H0FMAYER l t
t 1 t j 4 4
l i t
i l i
i i
Q BROOKHAVEN NATIONAL LABORATORY l} gyl ASSOCIATED UNIVERSITIES, INC.IlEll l . - . - _ , _ . .
STATUS BNL REVIEW 0F CONTAINMENT STRUCTURAL INTEGRITY t MEETING AT BNL ON AUGUST 14, 1986 8 OBTAINED:
- DESIGN DRAWINGS AND SPECIFICATIONS
, - CONSTRUCTION AND DESIGN REPORTS
- UE&C DESIGN CALCULATIONS
- SMA OVERPRESSURIZATION CALCULATIONS
- UE&C CONTAINMENT MODEL DATA
- - PENETRATION SLEEVE DRAWINGS O
THER RELEVANT INFORMATION O REVIEWED RELEVANT CONTAlllMENT STRUCTURAL ANALYSES PERFORMED BY APPLICANT'S CONSULTANTS.
8 PARTICIPATED IN SITE TOUR ON SEPTEMBER 8 AllD 9, 1986 8 PERFORMING AXISYMMETRIC FINITE ELEMENT ANALYSIS OF CONTAINMENT.
O MEETING WITH APPLICANT AT NRC ON SEPTEMBER-23, 1986 O FOLLOWUP MEETING TO DISCUSS DETAILS OF STRUCTURAL ANALYSES' l
SCHEDULED FOR MID OCTOBER.
I a
O BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(E Ell
ISSUES RELATED TO CONTAINMENT STRUCTURAL INTEGRITY I
BASIS FOR UNCERTAINTY FACTORS ASSIGNED TO VARIOUS PRESSURE CAPACITIES.
8 MECHANISM FOR TRANSFERRING LOAD FROM PENETRATION SLEEVES TO WALL UNDER HIGH STRAIN CONDITIONS.
8 NONUNIFORM CRACK GROWTH MAY LEAD TO STRAINS ON LINER HIGHER THAN PREDICTED.
COMPATIBILITY OF STRAINS IN REBARS AND LINER PLATE.
& CDNSIDERATION OF COMBINED TENSION, SHEAR AND BENDING EFFECTS AT DISCONTINUITIES. '
8 SGMUSYMMETRIC CONDITIONS MAY LEAD TO HIGHER STRAINS.
g 8 EVALUATION OF PUNCHING SHEAR AT FUEL TRANSFER BUILDING.
8 MATERI AL STRENGTHS USED IN ANALYSIS.
I EXTENT TO UHICH PIPING PENETRATIONS HAVE BEEN ANALYZED.
8 BASIS FOR LEAK AREA CALCULATIONS.
h BROOKHAVEN NAll0NAL LABORATORY l} l}
A5500ATED UNIVERSITIES, INC.(llll
4 l
l
}.
lG i
i REVIEW 0F ACCIDENT SEQUENCE FREQUENCIES t
i l i L t
R. YOUNGBLOOD 1
l i i i l i I
\O
\
- 1 l
i 4
4
/
Q BR00KHAVEii NATIONAL LABORATORY l} gy l A5500ATED UNIVERSITIES, INC.(I(Il i
~
O REVIEW 0F ACCIDENT SEQUENCE FREQUENCIES A FOCUSED LOOK AT:
INTERFACING SYSTEM LOCA CONTAINMENT BYPASS l
O l
Q BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(1(El
I O
INTERFACING SYSTEMS LOCA PSNH SUBMITTAL IS ESPECIALLY-THOROUGH IN KEY AREAS:
PROBABLE BREAK LOCATION PLANT RESPONSE OPERATOR RESPONSE WE WILL BE FOCUSING ON THESE AREAS, AND SPOT-CHECKING SYSTEMS TO TRY TO COME TO GRIPS WITH COMPLETENESS O
BNL IS CONCURRENTLY WORKING ON GENERIC ISSUE 105 ON INTERFACING SYSTEMS LOCA g ,
BROOKHAVEN NATIONAL LABORATORY l} gy l A5500ATED UNIVERSITIES, INC.(llll 4
--- -- - - - -- .-- , , . , - -w,---. - --y ee-,ew, ,
O CONTAINMENT BYPASS ISSUES ISOLATION CAPABILITY OF ALL VALVES COMMUNICATING WITH CONTAINMENT ATMOSPHERE, ESP. PURGE AND VENT VALVES:
SEAL DEGRADATION DUE TO ENVIRONMENT OPERABILITY SURVIVABILITY O
l Q BROOKHAVEN NATIONAL LABORATORY l} gj j A5500ATED UNIVERSITIES, INC.(llli l
~ '
. _ . .