ML20133A770
| ML20133A770 | |
| Person / Time | |
|---|---|
| Issue date: | 09/27/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1451, NUDOCS 8510020309 | |
| Download: ML20133A770 (178) | |
Text
ORIQlHAL48s'"d*
UNITED STATES OF AMERICA
(
NUCLEAR REGUL,ATORY COMMISSION
\\
/
l In the matter of:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Class 9 Accidents Subcommittee Meeting Docket No.
i tx_-
l 1
l Location: Washington, D. C.
Date: Friday, September 27, 1985 Pages:
1 - 161
.de f
- 'o \\0:: 3emove "o TL- 'sSX*IM'E m,
AbN I
i o
{T Court Reporters 1625 I St., N.W.
t) j 07 Suite 921 Washington, D.C.
20006 o l (202) 293-3950 0510020309 050927
- 45 PDR
1 1
UNITED STATES OF AMERICA
/ ')
2 NUCLEAR REGULATORY COMMISSION (s,/
3 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 4
5 CLASS 9 ACCIDENTS SUECOMMITTEE MEETING 6
7 Public Meeting 8
9 1717 H Street, N.W.
10 Room 1046 11 Washington, D.C.
12 Friday, September 27, 1985 l
13 I
i t
14 The Su' c ommi t t e e met, pursuant to notice, at 8:30 l
15 a.m.,
William Kerr, Chairman of the Subcommittee, presiding.
16 ACRS MEMBERS PRESENT:
17 W.
Kerr P.
Shewmon i
18 D.
Ward l
19 ACRS CONSULTANTS PRESENT:
l 20 M.
Bender 1.
Catton l
21 M.
Corradint P. Davis 22 ACRS STAFF MEMBER PRESENT:
23 D. Houston 24 25 l
l i
I l
2 1
SPEAKERS:
2 M.
Silberberg J.
Mitchell 3
D.
Ross R.
Meyer 4
O.
Marino M.
Ernst 5
T.
Lee P.
Cybulskis 6
M.
Cunningham B.
Durson 7
2.
Rostoczy 8
9 10 11 12 13 14 15 16 17 to 19 20 21 22 23 24 25
3 i
PR OCEED I NOS
(
2
[8:30 a.m.3 3
MR XERR:
Sin:e all these people struggled to get 4
here, I think we should start on time, and I will, f
5 This is a meeting of the Advisory Committee on 6
Reactor Safeguards, the Subcommittee on Class 9 Accidents.
My 7
name is Kerr.
I am Subcommittee Chairman.
8 Paul Shewmon and Dave Ward are other committee 9
members who made it.
10 As consultants, we have Mr. Bender, Mr
- Catton, 11 Mr. Corradini, and Mr. Davis.
12 Mr. Houston is the cognizant Staff member present.
13 The rules for participation for the meeting were 14 announced as part of the notice for the meeting published in i
15 the Federal Register on September 10th.
A transcript of the 16 meeting is being kept.
17 1 ask that each speaker identify himself or herself 18 and use the microphone, s
t 19 We have received no written comments from members of 20 the public nor requests for time to make oral statements.
21 In a sense, we owe the Stati an apology for this P
22 meeting.
They previously spent two rather filled days l
23 discussing the report, and they perhaps think that that should r
24 have been enough.
I hope, however, that they recognise, that i
/"~\\
l
-/
l 25 those of do who have read the report, that it represents an l
l I
4 l
1 effort to encompass several years and several tens or perhaps I
\\
2 hundreds of millions of dollars of investigation, and as such, V
l 3
at does involve a lot of information that has been developed l
l l
4 over the past several years.
~
l 5
The task of this subcommittee and ACRS, as 1 think 6
as also the task of the Staff, is to try to assimilate the 7
information, and having done that, to try to integrate the 8
various pieces of the puzzle into a coherent pattern, which 9
perhaps w!!! point the way to a better and a safer approach to 10 reactor operation regulation.
11 The Staff, I think, has received from various 12 members of the subcommittee and the consultants several 13 comments which have arisen as a result of our reading of the 14 report and of the previous meetings, and 1 think they will 1 hesitate to use the word 15 give you some idea of but inability, perhaps, to 16
" frustration;" it's overworked 17 comprehend the total picture as well as we should and maybe 18 will give you some guidance as to those things that we would 19 find helpful in our further deliberations.
f 20 1 would hope that the Staff may also identify any i
21 areas, if there are some, in which you would like spectile 22 comments from the ACRS when we write a report.
23 Those are my comments, and 1 would ask if there are l
24 comments by other members of the subcommittee or consultants 9 I
25 (No response.3 l
S 1
MR. KERR:
I see none, and I will turn things over 1
2 at this point to Mr. Silberberg.
3 MR. SILBERBERO:
Thank you, Mr. Chairman.
I'll just 4
he very brief.
(
5 We do appreciate the opportunity to come back and 1
6 discuss 0956 with you.
We do recognise that, from your notes I
7 and comments, from both the subcommittee and the consultants, l
8 that there'are some areas that are worthy of further 9
clariftoation and discussion and hopefully dialogue with us, 10 and we have made special preparations to address those 11 throughout the morning in our various presentations that the 12 Staff will be making.
13 And so our objective is to provide you with as much 14.
Information as we possibly can to support your needs in the 15 preparation of your letter.
16 As far as identifying areas of specific requirements l
17 or specific comments that we would like addressed within that, j
18 we will be thinking about that, and we will try to pass that t
10 on yet this morning before we are done.
20 We are providing you with additional detail and i
21 clarification in the areas that you noted that you needed l
22 more, and the agenda shows five specific items, which !
L 23 L411 eve are important items, and then there is actually a l
24 slith on major uncertaintles/ program for resolution.
\\
25 But if there are any other comments that do come up l
i l
I
l l
6 1
that are not on the agenda, we would be pleased to address
(
2 those, to the extent that we can.
3 Now at this point, the first item on the agenda was
[
4 the issue that had been raised at the previous meeting 5
regarding the intended application of this work, and Dr. Ross
(
6 has expressed a desire to address that issue.
l l
7 MR. ROSS:
I want to discuss and sort of review l
l 8
where we are and where we are going, and I think we ought to or at least I have an opinion -- on 9
also have some dialogue i
j 10 what the role of the ACRS is or ought to be.
l l
11 The Commission adopted the Severe Accident Policy l
(
12 Statement and has some guidelines that we are using as the l
13 objectives for this work and the companion document, l
14 NUREG-1150, and I am referring to the version as published in l
15 the Federal Register on August 8th, back in Paragraph C, 16
" Policy for Operating Reactors "
17 The first point was, operating nuclear power plants 18 require no further regulatory action unless significant new 19 safety information arises.
20 The second point was, in the latter event, a careful 1
l 21 assessment shall be made of the severe accident vulnerability 22 posed by the issue and whether this vulnerability is plant or 23 site-specific or generic.
I 24 The third one is to identify the most cost-effective i
\\'-
25 options for reducing the vulnerability, consistent with
7 1
cost-effectiveness criteria of the Commission's backfit O
2 policy.
the fourth issue goes beyond 3
And the fourth is 4
current regulatory requirements -- generic rulemaking is 5
preferred, and in other cases, you do conventional practice 6
bulletins, orders, generic letters, 5054(f) letters, et 7
cetera.
8 Now where the EDO family is headed this is NRR is working with IDCOR to develop -- or 9
and Research mostly 10 IDCOR proposes and the NRR would dispose of what's called an 11 individual plant methodology, which perhaps has been made 12 available to the subcommittee, 13 This would be a method by which each member of IDCOR 14 would take the IDCOR reference results, and through a lot of 15 engineering work, justify the extrapolability or applicability 16 of the reference work to an individual plant.
IDCOR has 17 recently sent in some information to us, proposing a 16 methodology, and NRR will be reviewing it, and presumably in l
19 due course in a few months, some formula, some algorithm, 20 something called individual plant methodology will be 21 approved.
22 Part of the NHH approval mechanism is having an 23 independent audit of the risk of these reference plants.
The l
24 two ingredients of this independent audit are (a) a collection i
and and this is what 0956 is 25 of analytical procedures l
8 1
(b) is an applications report where we take the methods and
()
2 systematically apply them to five plants, or five or six l
L 3
plants, only four of which are IDCOR reference plants, and 4
also recalculate, to the best we can, the arrival rate of the 5
dominant sequences, so you come up with a risk statement.
1 1
6 NRR, in its review procedures of the individual l
7 plant methodology, would consider it the research input, which 8
is these audits of the IDCOR work.
Part of the work that we 1
l 9
are doing on this audit is a systematic assessment of the 10 role that backfits could play on these reference plants, 11 looking both at the cost-effectiveness of reducing l
12 vulnerability at the front end, reducing the arrival rate of l
13 various sequences as well, and mitigation on the back end.
r 14 We have had many discussions with the subcommittee l
15 here and at Sandia on the Govere Accident Risk Reduction 16 Program.
17 All of this should be available roughly in the first 18 half of 1986.
It comes in in a linear fashion; it doesn't 19 come in all at once.
It comes in plant by plant.
20 80 that is roughly what I call the EDO family 21 program on how we're going to implement the Severe Accident l
22 Policy Statement, what we're going to do, and roughly when I
23 we're going to do it 24 MR. SHEWMON:
Everybody was declared safe enough, 25 unless what happenedt Unless there was some reasonable ides l
F 9
1 to differ with that or reasonable basis for it or what?
2 MR. ROSS:
The direct quote is, "Unless significant 3
new safety information arises."
4 MR. SHEWMON:
Okay.
Now it is not the utility's 5
responsibility to review for this potential deviation or new 6
information.
The Staff is doing it by a comparison with 7
reference plants; is that it?
It is on the Staff to decide 6
whether or not they think there is a risk, a significant risk 9
there' 10 MR. ROSS:
No.
The premise is incorrect.
This was 11 not written in a vacuum.
The people who wrote this new full 12 well that IDCOR was going to systematically look at each and 13 overy plant.
14 Now there is no law that requires them to do it, 15 but they are doing it.
So the burden is on the regulated 16 industry.
17 Furthermore, they are going to do it with a 16 methodology that we are going to review and approve.
What wo 19 are doing is only looking at a selected small set of plants, 20 as a tool to and the Regulatory Staff in deciding whether this 21 extension, generic applicability formula makes sense.
22 But, no, the industry is doing it.
They are just 23 not required to do it.
24 MR. DAVIS:
A related question, Denny.
I understood 25 the IDCOR program was to terminate at the end of this year.
10 1
Has that been changed now?
In which case they would not (O
/
)
2 MR. ROSS:
The IDCOR termination thing is v
3 systematic, in that when the management of IDCOR approves the 4
extension, they approve it a year at a time.
5 Now probably if you looked at the contracts on the I am really speculating here -- you would find that 6
books 7
by the end of the year, the Management Board of IDCOR has to 8
get together to extend it.
9 But, you know, they are on a roll now.
They are 10 doing good work.
It would boggle the mind if right in the 11 middle of this they just went out of business 12 MR. DAVIS:
Okay.
You said that IDCOR was going to
~'}
13 look at each and every plant.
It was my understanding that xs 14 they were just going to develop 15 MR. ROSS:
No, no, no.
I didn't say that.
I said 16 IDCOR is producing an individual plant methodology.
Each 17 plant has to take this and do it on its own.
They produce an 18 extrapolation report, which will be quite thick, and it may 19 cost some plants as much as a million dollars to do it.
But 20 the IDCOR isn't doing it.
Each plant must do it on its own.
21 MR. CORRADIN!:
Could I ask a question about that?
22 1 want to review, because I want to make sure I understand.
23 Olven the individual plant methodology, it is 24 developed by IDCOR, reviewed by the NRC, used by the utilities I
or NRC or NRR N'
25 and plants individually, and then NHR t
)
11 1
comes back and looks at that application and reviews it,
[3
( s) 2 somehow using the 0956 methodology as some sort of comparison?
x 3
MR. ROSS:
Well, you were doing all right until the 4
end, Mike.
5 My guess is, if you just strike the statement about 6
using 0956, because, you see, the front end of the source 7
term methods, 0956, and the i n t e r; r a t e d thing in NUREG-1150 8
will be used by NRR in the first instance to approve the plant 9
methodologies.
There should be no need to go back and look at 10 it again.
11 Now there many be some interesting variations at 12 unique plants.
I don't know.
For example, we are considering
['T 13 and we have represented to OMB that in Fiscal
'87 we would i
1
'wJ 14 do an additional two plants, one D&W and one Combustion.
It 15 well may be that NRR, when they look at this, will decide that 16 one or more unique plants need an additional study of some 17 sort.
We'll just have to cross that bridge when we come to it 18 MR. CORRADINI Just one last question, and that is, 19 not all of the utilities in the country belong to IDCOR.
What or how are you going to be sure that all the 20 is going to 21 utilities are going to use their methodology?
Will it be I don't want to say coercion, but 22 simply by 23 MR. ROSS:
Well, we're not above coereton, Dut I'm 24 not sure if your statement was right about not all uttllties.
!7_ )
N/
25 MR. COHRADINI 1 know three of the Wisconsin
n 12 l
1 utilities are not members of IDCOR.
I 2
MR. ROSS:
I'm not sure that anybody has thought 3
about that.
4 MR. ROSTOCZY:
Yes, we are aware that some of the S
utilities are not members of IDCOR.
Neither have we 6
considered the IDCOR methodology as something that necessarily 7
has to be used.
But what we expect to issue at the end of our 8
work is a requirement that each utility must look at his own 9
plant, examine his own plant for the purpose of finding 10 outliers or any weaknesses, vulnersbilities in the plant.
11 We will provide general guidance of what is this 12 examination supposed to include, and at the very same time, we 13 expect to have proved the IDCOR methodology as one possible 14 way to meet these requirements.
15 If somebody asks, "Is it a Bechtel one,"
i t '. s up to 16 him.
If they want to go somewhere else, that would be 17 available.
18 One other comment might be in connection with 19 the examination being done by the tttilities of the policy 20 statement.
It states that it is the expectation that this 21 examination will be done by the utilities.
22 MR. BENDER:
I wanted to shift gears slightly and 23 ask about the judgments developing out of NUREG-0956.
In the 24 IDCOR report they itsted some things that they thought had 25 evolved as a result of the analysis and cited them as
r 13 1
important, like the time until containment fails.
That's an (O
i 2
illustration.
j 3
I couldn't get the same flavor out of NUREG-0956.
4 There's a lot of statements in there about we know more.
5 There's not a heck of a lot in there about saying well, what 6
we know about it gives us a better way of going about the 7
analysis, and here is why.
8 The commentary is sort of smeared out, and it seems l'
9 to me that you have to get that story consolidated better.
10 MR. ROSS:
Well, I thought I understand your I
11 point.
I thought we had emphasized, even maybe to an l
12 excessive degree, the role of the containment 13 MR. BENDER:
Well, I just picked that as an
(
14 illustration.
But you know, the fact that containment is I
15 important is one thing.
The way in which the analysis 16 indicates that you should consider containment is something 17 different 18 You're going to talk about WASH-1400 in a little 19 while and how it differs from what you're doing right now, but 20 it seems to me this report needs to say more about what 21 changes in the methodology come out of the research program.
22 And it really doesn't say that 23 As a matter of fact, mostly what you find in it are 24 a lot of subjunctive statements that say we don't know enough 1
25 to be able to make good judgments.
And since you are going l
14 1
ahead with this thing and not waiting for the rest of the m'(
2 whatever research program it is to get done, it seems to me 3
the' explanation and rationale for it need to be developed 4
better.
5 MR. ROSS:
I think that's a fair comment.
Why don't-6 we just accept it.
Remember, -- and maybe I didn't make it 7
clear -- we intend to reissue this document in due course 8
after the round of comment, and I think that is a fair-9 comment.
Let me talk a bit --
10 MR. WARD:
Denny, can I ask a couple of questions?
11 The IDCOR methodology which eventually is going to be 12 something that you would approve or the staff would approve as
)
13 an acceptable way to look for outliers does it emphasize O
14 outliers related to the prevention of core melt, or to the 15 mitigation of core melt?
N 16 1 mean, it's sort of a simpleminded way of putting 17 it.
You said it might cost a licensee a million dollars 18 MR. ROSS:
That's on the front end.
That would be 19 front end, the prevention end.
I think it would be mostly in 20 that end, looking for outliers.
I see your question.
21 MR. WARD:
1 mean the outliers you're looking for 22 are in the design of the plant as related to prevention of 23 core melt accidents.
Is that the emphasis?
24 MR. ROSTOCZY:
The emphasis in the IDCOR work is to
{
\\s/
25 take a complete view of the plant and look at both prevention
15 1
and mitigation type of events, so they are going to address I )
2 both of them and provide enough information on how to do a 3
review that both of these can be picked up.
4 The way I picture it and I really haven't seen 5
the details of the methodology; we have seen a general 6
presentation but we haven't seen the details yet.
The way it 7
is being set up it is more like a question and answer type 8
approach where a set of questions are posed to the utility and 9
he~has to answer this questions as to his plant.
As long as s
10 let's say the answer is affirmative, he goes to the next 11 question, but it goes through the entire process and picks up 12 both kinds of potential shortcomings.
13 MR. KERR:
I looked at the same thing Zultan did and
(
V 14 I had the impression that it would be system engineering 15 driven, and there would be a lot of people at the plant in the 16 general category of systems engineering looking at the front 17 end.
I think that's where they're going to spend a lot of 18 their money.
I don't think they're going to spend a lot of 19 their money on analysis, rerunning the IDCOR family of codes.
20 MR. WARD:
So that procedure really won't have all 21 that much relationship to 0956, for example, and the work 22 that's been done.
j 23 MR. ROS?OCZY:
One more thing.
They are trying to 24 minimize the work --
O s/
25 MR. WARD:
Could you answer that question?
l
._~
i i
16 t
i 1
MR. ROSS; I think the answer is -- to the type of 2
work in 0956, I don't think you'll see much of that in the s_-
3 individual plant studies, right.
1 4
4 MR. ROSTOCZY:
IDCOR is trying to set this up in 5
such a way so that it's a minimal effort for the i
l 6
utility.
They don't know yet exactly where they are going to 4
7 end up, and obviously there will be dependence on how close 8
the plant being examined is to a reference plant, and how 9
knowledgeable the individuals who are working on it are.
10 But they are thinking in some general terms of 11 something on the order of maybe a couple of manyears for a 12 plant, so it's a relatively small effort, But they think it 13 will be sufficient to look at individual plants.
i 14 MR. WARD:
One more question, Denny.
In a couple of a
15 committee letters on the subject, we recommended that you 16 should concern yourselves somewhere in this effort with for lack of a better term, 17 developing a way to look for operational outliers as opposed to design 18 what I call is there any activity of that sort 19 outliers.
Do you recall 4
20 going on?
21 MR. ROSS:
I believe the answer generally is no, but l
22 Mal is here somewhere.
Mal Ernst.
23 MR. ERNST:
On the reference plants, we are looking
.4 at the operating history of the plants to the extent we can; 1 O 25 not an extensive look.
We are also looking at plant i
l 1
17 1
procedures and test and maintenance procedures to some limited 2
extent.
So it's not just the design of the plant but it's 3
also the operation.
4 MR. ROSS:
For example, look at the station blackout 4
5 which we analyzed in a certain way.
How will you decide what 6
the arrival rate is and whether or not it is markedly higher 7
or lower for a given individual plant?
Is there an 8
operational outlier that would make that sequence more or less 9
likely?
I don't think we're going to look at that because 10 we're not looking at individual plants.
11 Now whether there is a way in the formula to tell i
12 individual plants to look at it for operational outliers, I
/}
13 don't know.
I think we'll just have to take that and consider 14 it when we review it.
Maybe there is, maybe there isn't.
15 MR. SHEWMON:
When you talk about looking at it, is 16 this looking at paper or looking to see whether it's on the 17 southern tip of a Florida peninsula, or whether they have had 18 many blackouts, more than usual in the last five years?
Or 19 what are we talking about?
20 MR. ROSS:
I think from Mr. Ward's thrust, it was 21 looking at, in this case what I would call the St. Lucie type 22 response; what do you do if you have this blackout, how 23 quickly can you recover, what are your alternate 24 MR. SHEWMON:
It's not experience or location but
\\
25 paper as to whether you think they have the drills down for
18 1
what they do if it came.
C
(
2 MR. ROSS:
That was the thrust as I understood it.
ws 3
And the thing is -- I shouldn't have brought up this example 4
perhaps because I think it's covered by the resolution of 5
A44.
I think that is going to require each plant to look at 6
vulnerabilities to loss of station power.
7 But systmatically looking at operational 8
vulnerabilities, I think we will just have to write a note and 9
see if we can't get that part of the individual plant 10 methodology.
Maybe it can be done, I don't know.
11 MR. WARD:
Yes.
Well, it isn't obvious that the 12 approach that IDCOR is taking, which is, you know, I think
[
13 based on PRA methods or experience in that sort of thing, 14 really has any hope at all of doing the sort of thing we were 15 just talking about.
16 MR. ROSS:
Now, if you will talk about the back end, 17 I think there's general agreement that operator actions t
18 to mitigate a given core degraded siate would be very heavily 19 an output of the work, both ours and theirs.
I think you will 20 see a lot more plant-specific emergency operating procedures 21
.to cope with degraded cores, especially for the BWR family.
22 You'll see a lot of that.
But I don't think it looks at the 23 front end so much.
24 I have one last point and that is what can the ACRS 25 do.
I think there are three separate things that ACHS could
1 19 1
1 do.
In reading some of the comments I can feel sort of a 0
2 sense of frustration on the ACRS's part.
How do you respond
(
3 to all three at once, and I think the answer is that you 4
don't.
5 You have the NUREG-0956 which is a limited purpose 6
document, and that has little or nothing in there about 7
applications.
And then in a few months, six to eight months, 8
you will have an additional NUREG document, NUREG-1150, which 9
is the front end and back end put together for the five 10 reference plants.
I don't 11 Somewhere in that time span there will be 12 know how it will be documented but there will be a report by
/g 13 NRR on its review of IDCOR's individual plant methodology, and
%)
14 you'll have an opportunity to comment on that.
15 I think the last would be the most important one to
'he ACRS in terms of satisfying the itch that you can't quite 16 t
17 scratch yet.
So I guess you have to decide -- if you're going 18 to write a letter on 0956, what are you going to say about 19 applications?
And you may just decide that you can't say 20 anything.
21 If you go beyond this and suppose we have now 22 finished all the work for operating reactors, what happens 23 next?
Well, we have a menu of about ten areas where generic 24 rulemaking is indicated, going beyond closing the books on w/
25 operating plants, Appendix J,
Appendix E,
equipment
20 1
qualification, Part 140 on ENO, operation of engineered safety On
( j 2
features, siting, Reg Guide 1.97 and beyond on accident 3
monitoring and management, prioritization of generic issues 4
and so on.
5 We have a long menu, and you may want to know to 6
what extent can the 0956 methods be used; what is the 7
uncertainty.
If I approve 0956 as an ACRS, am I automatically 8
approving its generic use in rulemaking?
I think the answer 9
is no.
10 Suppose we came in a year from now or two years from 11 now with a revised Appendix J which is risk-based and we are 12 taking into account the reduced source term that we think is i
/'N 13 justifiable on the 0956 methods, I think you would look at the 14 uncertainty and the precision and whatever at that time and 15 not try to decide today what the uncertainty is, what we know 16 and what we don't know as this method is applied for any 17 particular purpose.
I think you're going to have to conside,r 18 the appropriate uncertainty for the use at hand.
19 I have a feeling this is probably the single biggest 20 concern that the subcommittee, and perhaps the full committee, 21 has.
So maybe-this is a good time to ventilate it.
But i
22 anyway,that's all I wanted to say this morning.
23 MR. KERR:
Denny, I got the first two but I somehow 24 missed number three.
A-25 MR. SILBERBERG:
I recall the list, Dr. Kerr, that
21 1
was presented on August 2nd.
2 MR. KERR:
No, you said there were three things that 3
you thought the ACRS might consider -- one was 0956, one was 4
on 1150 --
5-MR. ROSS:
And the third une was NRR review and 6
approval of the individual plant methodology by IDCOR.
I 7
don't know what form it will take.
I don't think it will be a l
8 NUREG, will it, Zultan?
9 MR. ROSTOCZY:
No, we don't expect it to be a 10 NUREG.
What we expect to do is after the information is 11 available so all the reference plant analyses are on the 12 table, then there will be a written recommendation for
/ 'T 13 Commission consideration, so it will be more like a Commission U
14 paper, some attached material and referencing reports like 15 these NUREG's and possibly other NUREG's.
16 We are working with some technical assistance 17 contractors, working with us, and there will be some 18 referenceable report coming out of the evaluation, and those 19 will be used in a similar manner like 0956 20 MR. ROSS:
Okay, that's all I had this morning.
21 MR. KERR:
I think that is helpful 5
22 Let me go back to number 1,
the comment on 0956l4 Y
23 which is I think our most immediate concern.
The reason that i
24 I asked for its use and application is because it's my v
25 impression that we are being asked to say is 0956 an adequate
22 1
or useful report, or something.
i
("'
2 Now I can comment on a rose and I can say it's a g
3 beautiful flower, without worrying about what it's going to be 4
used for.
And if I were a poet I could write rhapsodies about 5
a rose.
I don't know how to comment on this report in that 1
-6 way.
I could say it's a beautiful report, it describes a lot 7
of work that has been done and work that needs to be done.
8 But I also think when you ask, is it acceptable, is it s
9 adequate, is it something, that one has to have something in 10
. mind as to what it's going to be used for or how it's going to 11 be looked to because it's not just a report in isolation; it's 12 a report, as you pointed out, that is going to be part of an
\\
13 analytical procedure, for example, and the analytical 14 procedure is going to be used to analyze plant performance, I 15 assume.
16 And so it is in that sense that I ask, what is going 17 to be used for.
You know, how does it fit into the total 18 regulatory process.
Because unless I have some idea of how 19 that is going to occur, it's hard for me to say yes, it's 20 adequate or more than adequate, or here are the shortcomings.
21 And it's in that sense that I am asking.
I don't
'22 mean that I think you have all those questions answered yet; 23 in fact, I'm pretty sure you don't.
But insofar as one does, 24 it is helpful to me to comment on it if I know in what way it s
25 is going to be applied.
23 1
Let me take for example one of the things that I
(
2 have found, and I think it is part of the list of 10, is that 3
it ought to be considered as a substitute -- that's not quite 4
the wording for it -- for those problems that make use of the 5
TID 14-844 source.
I am not sure I know what that statement i
6 means.
I mean, the 14-844 source, in effect, points out very 7
specific numbers of isotopes that are in containment.
8 Does that statement mean'that there now exists a 9
recipe, which I can find somewhere in 0956 or in BMI-2104, and 10 I can run through that and that gives me those isotopes?
Is 11 that what the statement means?
Or does it mean somebody ought 12 to take this report and develop such a recipe?
~
13 MR. ROSS:
The latter.
And we fully intend to.
14 14-844 and the footnote to Part 100 need changing.
15 They are not right.
They need changing, and in time we are 16 going to do it.
17 MR. KERR:
But at this point you don't know how one 18 would use 0956 to do it.
It's just that you are pretty 19 certain that one could.
20 MR. ROSS:
Well, the thing is you would use it in 21
- part, I have a list here of ten items that I just referred 22 to.
The title of this list is " Regulatory Areas Targetted for 23 Source Term Related Changes."
Part 100 is one of those.
If f
24 we are going to change the regulation, we need a technical S
x-25 basis.
24 1
What I think will happen is NUREG 1150 will be the (G -)
2 technical basis for such a change, and then we would refer to 3
it in the supplemental information, the notice of proposed 4
rulemaking, and we will say readers who want to understand the 5
basis for changing it, NUREG-1150 exists.
NUREG-1150 uses as 6
an analytical procedure the NUREG-0956 methods.
7 It also has more. It has applied then to a wide 8
-variety of plants, the three different kinds of boilers and so 9
on.
10 MR. KERR:
Well, I was asking what I hoped was a 11 more restricted question.
I was thinking of the 14-844 not in 12 terms of siting because I think I would tend to put siting on 13 a back burner for a while because I don't think it's a dead
%.J' 14 issue but I don't think it's a real important issue at this 15 point; I was thinking, for example, of environmental 16 qualification of equipment inside containment.
17 Right now you have a very specific recipe which says 18 you put this much out and you put-this much something else, 19 and then you environmentally qualify ~it.
I thought that the
~20 statement in 0956 meant we believe that you ought to use the 21 BMI or whatever codes to do that, to produce numbers of curies 22 or whatever.
But I just saw the statement that it ought to be 23 considered, and I don't knog what ' hat means.
t i
I. don't think 24 MR. ROSS:
Well, 'what it3means is
(
i 25 it means anything serious.
You mentioned equipment inside
25 1
containment, and certainly that's an answer. For example, when
(~x
)
2 we do the ice condenser, when we do Sequoyah, we will have i
3 done an accounting of where we would expect the fission 4
products to be.
There are filters associated with that plant 5
that should filter anything, leakage -- you know, there is an 6
annulus and it collects stuff with charcoal 7
Is there any elemental or organic iodine coming out 8
to be filtered?
If so, you want one kind of charcoal.
If 9
not, maybe you don't want any charcoal at all.
10 On the other hand, maybe the load on the HEPA 11 cartridges are much higher.
You need to do a systematic 12 reappraisal of the location of these fissior products. They
'~h 13 affect filter performance, EFS performance, they affect d
14 qualification'of pumps and seals.
That's why I mentioned the 15 EQ rule has to be looked at, the source term for equipment 16 qualification certainly, but before we change it, we have to 17 do some more work in addition to 0956.
18 What we think you should be looking at 0956 is has 19 the Staff presented the best available technology for 20 analyzing how hot corium interacts with concrete?
If aerosol 21 settled out of a reactor building, have they used the best 22 available technology or have they made any fundamental errors 23 in how the core melt progresses?
24 They have comments on steam explosion.
Does that 4
25 represent today's science?
I think you should look at the
26 1
whole report as a science report and not as a regulatory
(,/
2 report, not as a licensing report.
We have deliberately tried 3
to sever.
If we wanted to change the regulations, we would 4
have put a new chapter in this report that said we want to 5
change Part 100 and here's how we are going to change it.
We 6
are not ready for that.
7 MR. KERR:
Well, thinking of the Battelle suite of 8
codes as science -
_I mean those are tools.
Now, the science, s't u d y 9
the fundamental information has come, apparently, from a 10 of the way in which the parts of molten core behave and in 11 which fission products are released.
Given that fundamental 12 information, one has tried to put together some models to try
[
13 to describe processes which have not been looked at 14 experimentally.
I mean the whole program hasn't been.
15 So it seems to me 0956 is partly science but it is 16 also -- and I don't know which is the larger part -- it is 17 also a description of a set of tools which have been developed 18 for modeling.
.19 MR. ROSS:
Well, wait a minute now. I just can't 20 agree with you there.
If you look at the computer code TRAF 21 MELT, its fundamental equations are physics in there, and 22 there is some experimental information.
They were put 23 together for convenience in something called a computer code, 24 tut that doesn't mean it's not a scientific device.
25 MR. KERR:
I'm talking about the whole suite of
27 1
codes.
There are certain parts of those codes, certainly, (A) 2 that depend fairly heavily on good information.
There are 3
other parts that are very uncertain, as the report makes 4
clear.
I mean the report doesn't necessarily represent 5
things, I think, and overall it seems to me that this suite of 6
codes is a tool that is meant to describe the behavior of a 7
series of very complex events.
8 MR. ROSS:
Yes.
9 MR. KERR:
So to me, it is more than just science.
10 It is science plus some suggested tools.
I mean I don't want 11 to argue the point.
I'm telling you what my viewpoint is.
If 12 you have a different one --
13 MR. ROSS:
Okay.
If you accept that hypothesis, 14 then what?
15 MR. KERR:
So now I am commenting on something that 16 is supposedly to be used for something or other, not just bare 17 bones science, and what I am asking is; For what is it to be 18 used?
Well, the answer is here are ten things.
No quarrel 19 with that. But even within the report, it is not clear to me, 20 for example, as to how I use this set to get the environment 21 inside containment.
22 From what you have said, I gather that from your 23 point.of view you would say one does this now on a 24 plant-specific basis.
If you have Sequoyah, for example, 25 which is an ice condenser, you get the equipment that is
28 1
appropriate to Sequoyah containment, and in order to
(
2 environmentally qualify the equipment in Sequoyah's 3
containment, you run the suite of codes and you get the 4
isotopical content in containment and that's what you use.
5 MR. ROSS:
I'm not quite sure it would work that 6
way. Let me explain it.
We have a branch in our Engineering 7
Division, our Electrical Branch, that is responsible for Reg 8
Guide 1.89 on qualification of electrical equipment.
They 9
have an accident source term that went with that.
10 What we plan to do is, once we get the five or six 11 plants studied, we would tell this branch to take a look at 12 the guidance that you issued for equipment qualification and 13 its source term, take a look at this new understanding of what Os 14 the radioactivity is likely to be in the spectrum of p l.i n t s,
15 and make recommendations whether or not the Commission must 16 change its avgulations with respect to equipment qualification 17 based on the fact that this whole set may have been placed in 18 a new or different light, 19 If it needs to be changed, then we may well come up 20 with a simple recipe, a simple source term that is applicable 21 to all plants.
We haven't ruled that out.
I don't know how 22 it will come out.
We haven't started that work yet and we 23 can't start it until we get this done.
24 MR. KERR:
Do you see why I have a little difficulty 25 with what you said in interpreting what that statement
l 29 j
1 means? You are viewing this in the context in which you have a 2
lot more information and a lot better idea of where things are 1
i 3
going than I do from reading the report.
4 MR. ROSS:
Okay.
Well, we had guidance about a year 5
or two ago from Mr. Dircks.
He made it very clear that he 6
wanted this to be -- I still think it is -- a scientific 7
report.
He did not want regulatory changes to creep into this 8
report.
Now, I think we could provide what you want, but I 9
don't want to put it in this report.
l 10 What additional stuff do you need from us along with 11 what I have just said?
Do you need more dialogue for written 12 purposes, or do you have a suggestion of what more we could
~
i 13 give you?
14 MR. KERR:
We'll, I tried to tell you by giving you 15 an example of the sort of thing that embraces the question I l
l 16 asked when I said how is it going to be used.
The response is l
l 17 it is going to be used for, for example, this list of ten 18 things. So I asked about TID 14-844, and I gather what you are 19 saying is one is going to reexamine a number of issues and we 20 are going to make use of the suite of codes in some fashion 21 which is yet unclear to decide whether to change the existing 22 source terms or not.
23 MR. ROSS:
Zoltan, the NRR had a paper of eight or 24 ten pages along this line.
What is the status of that paper?
\\-
25 MR. KERR:
Well, isn't that what we did a
(
--,,,_,,r-,,., ~, - - _ - - - -.
30 1
presentation on on the 2nd of August?
2 MR. ROSTOCZY:
Yes.
You have received the 3
presentation, which is basically those ten items.
We are 4
maybe a half-step further ahead.
We have given some thoughts 5
to how this might be used, and we intend to discuss it with 6
the industry next month on how it might be used.
7 The present thinking is along the lines that, as you 8
mentioned, the different type of applications require a 9
different form of source terms.
For the evaluation of 10 individual plants in terms of severe accidents where we had 11 calculations available, at least for the reference plant, the 12 use of that detailed calculational approach like described in
/~h 13 0956 is an appropriate way.
So that is one form how it can be 14 used.
15 For other purposes, like for equipment 16 qualification, you need the source term that is relatively 17 simple that an equipment manufacturer can use years ahead of 18 the application of the equipment. Furthermore, his goal is to
.19 qualify his equipment and design his equipment in such a way 20 that he can put it into any plant.
He doesn't want to be 21 limited only to a certain type of plant.
So he would need a 22 different form of source term.
23 At present, we are thinking at most of three types 24 of source terms, One of them is a relatively simple one that 25 possibly could be used across the board, or close to it.
The
31 1
second one maybe, which is either a set of tables or some-kind (O) 2 of procedures, relatively simple procedures described.
These 3
would be then available for a group of plants.
So maybe 4
boiling water reactors of certain types with a certain 5
containment would have a relatively simple recipe or tables 6
that one could go to and use.
7 The third one would be the detailed calculation 8
using the suite of codes, either this suite or an equivalent 9
suite of codes.
Obviously and the approach would be 10 probably that this would be optional which one you use, so all 11 three or two or whatever we end up with would be available to 12 the applicant or licensee, and he could choose which one he
["S 13 wants to use, y
14 If he elects to use the detailed codes, that 15 obviously would cost him more.
I am not sure exactly what it 16 takes to apply this to a plant that has never used any of 17 these codes, but I would venture to guess that we are talking 18 in the millions of dollars.
It's more than a million dollars, 19 probably, to apply that.
20 On the other hand, if he uses the simplified one or 21 one of the simplified ones, then he probably will be taking 22 some penalty because every time when you generalize, every 4
23 time who you develop it in such a way so it applies to a group 24 of plants instead of an individual, you have to cover that in 25 some fashion so it includes some bounding type of thinking,
32 1
and probably will be giving some penalty.
(j 2
Now, for his given application, the penalty doesn't 3
bother him, then he probably will use the simple approach and 4
will not go to the detailed.
The way we are planning to 5
approach it is to take those ten items that Denny mentioned on 6
the list, discuss it with the industry, and see for each of 7
those applications what kind of a form would be needed so it 8
can be easily put to practice.
9 If we can agree on those, then we would kind of 10 group those ten items into groups of what types of forms are 11 needed, and then we would go ahead and develop these forms.
12 Now, you are correct in that that this one is not in
/~'\\
13 0956 because 0956 describes only one of these, provides the 14 means for one of these options for t i.e detailed calculation.
15 It is not going to be in 1150 because 1150 is the application 16 only of this detailed methodology.
17 This will have to be a third piece of work, and the 18 third piece of reporting, so another report or something that 19 will derive from the detailed a p p r'o a c h some simplified 20 approaches in some appropriate manner, and it has to be done 21 after the application of the detailed methodology for the 22 reference plan because you must have something in front of you 23 to see how does it work out when you apply it to a given 1
24 reference plant, and the with some thinking about other plants i
b 25 in the same group, you possibly can derive some simplified
33 1
source term that could be used.
T 2
Now, this is just present thinking and present 3
discussion on this.
Exactly how is it going to end up and how 4
is it going to evolve, we don't know for sure, but we hope to 5
learn a lot more in the next few months, and certal,nly by the 6
time the reference plant analysis is finished we should be in 7
good shape to go ahead with this and do it.
8 MR. CORRADINI:
I was listening and writing down a 9
number of things.
I want to make sure I am following here, so 10 I am going to ask some questions, and assuming 1 understand, 11 maybe I can ask a final question.
12 If I were to list three things, basic science, 13 computer tools for a specific sequence analysis, probabilistic 14 accident sequence analysis, containment event trees and 15 quantification, quantification of source term and then 16 consequence analysis, the way I see it is the basic science 17 comes out from somewhere supported from research and doesn't 18 necessarily have to appear in 0956, although parts of it 19 appear or are referenced.
Okay?
20 What I mean by that is that the person who develops 21 the specific correlation and all his work is not reported in 22 0956; it's referenced as part of the development.
23 MR. ROSS:
There are brief summaries, but by and for example, the VANESA manual was just published 24 large 25 this July. If you want to know the science of VANESA, you
34 1
would go to that report, you would not go to 0956.
If that is
- ~N f( )
2 your point, sure.
3 MR. CORRADINI Yes.
The second one, computer tools 4
for a specific-sequence systems analysis, that is basically --
5 the information there is basically found in BMI-2104 for a set 6
of sequences, and in NUREG-0956, which discusses currently the 7
Battelle codes as the suite that is being used to pull 8
together the basic science in some sort of workable 9
engineering calculation.
10 MR. ROSS:
Yes. You may have to reach beyond 2104 if 11 you wanted to get TRAPMELT or something, and MARCH II, but I 12 think you have met that.
[O'}
13 MR. CORRADINI Okay.
And now, the last four, 14 probabilistic accident sequence analysis, that is currently 15 ASEP.
That is the front end, trying to decide what are the 16 dominant accident sequences.
17.
MR. ROSS:
I can ask Mal.
I think it is not quite 18 ASEP.
The front end that we will be using in 1150.
19 MR. ERNST:
Well, it's ASEP expanded in te reference 20 plant analysis.
21 MR. CORRADINI.
Okay. ASEP plus.
Okay.
22 Then in containment event trees, quantification and 23 construction, is the SARP work, okay*
Quantification - of the i
24 source term is actually using two, that is, the computer tools 25 for specific sequence analysis, the Battelle suite of codes, l
i
35 1
and the SARP work somehow put together and coming up with some
(
2 sort of list of what is the source term and what is its 3
probabil.ty given a whole range of the dominant accident 4
sequences.
5 And then the sixth thing is the consequence 6
analysis, which right now is CEAC-2.
7 MR. ROSS:
Right.
8 MR. CORRADINI So, given all that, my question goes 9
something like this.
If I look at that, it looks like the 10 research and the product of the research, which are these 11 reports and a description of that, is basically being driven 12 by a complete PRA.
I mean I sense the whole thing.
It's a 13 big package, and if the package, if I look at it in total, it t
14 looks like a big PRA for five plants.
15 MR. ROSS:
If you will include 1150 in what you just 16 said, yes.
17 MR. CORRADINI-Yes.
18 MR. ROSS:
I don't disagree.
Let me poll the 19 delegation a minute.
20 MR. MEYER:
I think there was one inaccuracy in your 21 statement about the source term, about item No.
4.
There is 22 not a weighted source term.
There is a source term calculated 23 for each event.
Then there is a weighted risk.
24 MR. CORRADINI Right.
But given the sequence, i
N 25 given a calculation, given an event tree which tells you the
36 1
probability to a point, then you get a source term at the end
.f%
(
)
2 of all that.
Right?
3 MR. MEYER:
You can get a source term independent of 4
the probability.
5 MR. CORRADINI Right, but the source term I get is 6
dependent upon the sequence I take.
I don't get a source 7
term; I get many.
8 ME. MEYER:
Correct.
9 MR. CORRADINI Okay.
So my question goes something 10 like this, and that is: Given that that is all like five bit 11 PRAs, the direct use for the ten items that you had listed 12 previously, I sense what you are saying is now that you have 13 done all this work and you h' ave got this big set of tools, you 14 can address all of these ten items because somehow by doing 1 mean you have 15 the big thing, you have enveloped the need 16 got now all the tools you need to do the big thing, the big 17 five PRAs, and all I need is subsets of that to do all the 18 applications.
19 Is that basically what you are saying?
20 MR. ROSS:
I think so.
Let me q'i v e an illustration 21 on Appendix E.
There has been a lot of notoriety about 22 emergency planning and source term.
I think when we do 1:50, 23 we will have the basis for saying for these six plants what is 24 the revised off-site source term as a comparison with what.is 25 on the books today as the basis for emergency planning.
37 1
Has-there been a lot of talk about shrinking the (p) 2 planning zones or has there been a more graded response than 3
we now have?
I think there are six reference plants in 1150.
4 The analyses are all the way out conducted from the front end 5
through the back end, and getting through the CRAC Code would 6
be the technical basis for either doing nothing or for 7
changing the regulation in some way.
8 Now, if that was your question, then that's a simple 9
answer.
10 MR. ERNST:
I think I would like to add to that and 11 try and build a bridge here.
Zoltan and I haven't talked too 12 much about the detailed contents of 1150.
We are presently
/~ \\
13 coming up with an outline that is maybe 15 or 20 pages long 14 and will be ready for discussion.
15 I would anticipate having a chapter, or, well, three 16 basic chapters in 1150. One is the display of the risk 17 information on the six reference plants, the insights, drivers 18 and things of that nature, another basic chapter on 19 uncertainties, treatment of uncertainties and insights one can 20 get, and the more or less an integrating chapter, which is 21 applications, which would not go so far as what Zoltan was 22 saying, more than likely, of picking specific numbers and 23 saying this should be the regulatory usage, but in drawing the 24 insights that we get from the risk chapter and the uncertainty 25 chapter and melding them into some judgments of better ways to
g-38 l
1 apply and the uncertainties one has to consider and the r~m
(,)
2 various types of applicatirn.
That kind of a bridge I 3
would expect to be in 1150 4
MR. CORRADINI-So let me just say it again because 5
you are asking us to comment on 0956 and I am trying to figure 6
out where it fits in.
So if the whole group of things implies 7
five big PRAs, then 0956 is one part of the group, and t h e' 8
claim is that to go after those ten items that may involve 9
regulatory changes, you would feel comfortable that all of 10 this information gathered, and part of it being published in 1
11 0956, is useful because any of those ten items will just be I
12 using a subset of the tools developed and of the science
(~')
13 developed, either a subset or some large group of the whole.
V 14 MR. ROSS:
If I understood you, it's the other way 15 around.
A subset of NUREG-1150 is the 0956 methods.
16 MR. CORRADINI Right.
But that whole thing, when 17 looking at these ten items which were listed in the August 10 18 meeting, may not use all of what you have developed.
It may 19 use a subset.
20 MR. ROSS:
Yes.
I think Zoltan explained l
21 that..There may be some collapsing or consolidation in areas.
l l
l 22 MR. CORRADINI.
Then given that, the only thing I 23 can think of that we can comment on in 0956 is whether somehow l
24 looking at the ten items, and at this point I'm not sure if 1
(
l
\\
l 25 could, if it is adequate or not, or looking at the big picture j
l l
l
39 1
of the PRAs and saying that your original objective or one of L
()
2 the original objectives was to come up with five large PRAs 3
for these plants as some sort of basis. That would be the only 4
way to comment 5
MR. ROSS:
Of course, the ACRS can do what it 6
wants.
The more of that you do, I think the worse off you 7
are because we are not tendering 1150 to you or anyone else.
8 We haven't done it 9
MR. KERR:
I'm sorry.
The more of what we do?
10 MR. ROSS:
The more you focus on the regulatory 11 applications, the more you are getting distracted from the 12 message we were trying to give.
However, if the ACRS 13 subcommittee or full committee decides to pass on the document i
14 and say you cannot comment until you see the regulatory i
15 application, that's okay.
I don't know what the Commission 16 would think.
You would have to ask them.
But clearly, that's 17 an option, just to pass.
But you can't tell your regulatory 18 body that you ought to advise the Commission on how to 19 regulate.
You need the regulatory closure.
l 20 I can understand that, and if you want to do it, 21 that's okay with us.
l 22 MR. KERR:
Well, Denny, there is an alternative, and 23 it's the one which 1 think I'm trying to take, and that is to 24 ignore regulation and ask yourself what you can tell about the
\\
25 risk of a plant.
You can talk about that independently of I
40 1
regulation. And this, I think, is one of our principal
(
2 interests: what have we learned about how to calculate the 3
risk of a plant?
4 MR. ROSS:
Well, you won't find that in this 5'
document, and we have no intention of putting it in there.
We 6
have a very limited risk statement on Surry, but I even hate 7
touse the word.
8 MR. KERR:
But surely you must mean that,this 9
document fits into a methodology that will calculate risk.
10 MR. ROSS:
You bet.
11 MR. KERR:
And it is that fit that we are trying to 12 understand.
13 MR. ROSS:
Well, we gave you, you know, 20 or so 14 pages in Chapter 6 of how we thought it fit.
15 MR. KERR:
I'm not trying to be critical of the l
16 report, although it does try to encompass a lot of things, as 17 I think you will recognize.
I am saying it is that sort of 18 thing, not how the regulators are going to ultimately apply 19 it.
I can't predict that at all But I do think we need to 20 understand how this information is to be used to characterize 21 the behavior of plants in abnormal situations.
i 22 MR. ROSS:
That I agree with. I see what you mean, 23 and I agree now.
I hope that we provided that in Chapter 6,
24 but if we didn't, you may want to pursue your questioning to k
l 25 see what more could we have done or could we still do.
L-----
41 1
According to the clock, I'm supposed to hand this to 2
Mel now.
3 MR. BENDER:
I have one more question I would like 4
to raise.
In looking through the report, it has a number of 5
places where it says we are still developing more 6
information.
You could deal with those matters as it would be 7
nice to have that information, or you could deal with it as 8
without that information you are limiting what you can do with 9
what is in the report.
10 My own perception is that additional research is not 11 going to add a lot to the usability of the information in the 12 report.
It may make you more comfortable, but I am a little
/
13 puzzled about how to. react to the anticipated further 14 research.
Does it put a limitation on the usability of the 15 information that you want to convey in here?
16 MR. ROSS:
I thirk nort ni a soft, tan +at1"e yes 17 It is going to limit it in the following sense, As you know, 18 the QUEST study did a first step, albeit crude, uncertainty 19 quantification.
We have a follow-on study at Brookhaven which is that 20 I think will be done by the end of calendar 1986 or on into 1987 that will do a better job on 21 about right?
l 22 uncertainty, 23 I thing the uncertainties will still be appreciably 24 high, and when we start trying to do such things as do we have
\\
25 a basis to modify Appendix J or not, the uncertainty still may r
I
42 1
be too large as to do anything very meaningful r%
)
2 1 think in terms of what I refer to as closing the q
l 3
books on severe accidents, that is, deciding once and for all l
4 do we need anything else or not, these methods, even with 5
their uncertainties, are perfectly acceptable, and I think the 6
Sandia work will.show that.
7 So I think the research is intended --
8 MR. KERR:
Excuse me.
When you refer to the Sandia 9
work, you mean 10 MR. ROSS:
The SARRP, or the severe accient risk 11 reduction program.
I think, even taking the broad 12 uncertainties into account, we will be able to make regulatory
[ \\
13 decisions.
I think that the research is needed to reduce the 14 uncertainties for some aspects of generic rulemaking.
I think 15 the quantification will probably still show many or several Ib uaders of magnitude. It will not preclude us from moving 17 vigorously in some of this list of ten areas.
18 MR. BENDER:
hell, are you looking for comments on 19 whether further research is necessary, in view of that, or are 20 you just looking for comments that say assuming that the 21 research doesn't change, what are the implied conclusions of 22 this report, and here is how you might use it?
23 MR. ROSS:
I think that would be certainly a good i
24 point I think when you look at the various things that are i O j
25 analyzed in here, like pool scrubbing and core-concrete and
43 1
aerosol deposition, the committee's views on whether this is
()
2 just witchcraft or solid science or some more in-between or 3
whether.it is something that could be.made better by pursuing 4
the identified research program is certainly in order, that's 5
the kind of thing that would be very useful to us.
6 MR. KERR:
Other questions?
7
[No response.3 8
Thank you, gentlemen.
9 The agenda, surprisingly enough, shows a break at 10 this point, and I don't want to avoid the agenda, so let's 11 have a ten-minute break.
12 Oh, I'm misreading the agenda, so let's go ahead.
13 Ms. Mitchell
' 'N 14 MS. MITCHELL:
The agenda item is called a 15 Comparison of NUREG-0956 versus WASH-1400.
That is a huge 16 topic and could take several different forms.
'I was not 17 absolutely sure what you had in mind, so I brought what I 18 thought was the interesting part of it that perhaps was not in 19 the presentations that we made to you either of the two days 20 that we have been here.
That is sort of what are the major 21 advances in the source term technology that is described in 22 NUREG-0956 versus what it was in WASH-1400 days.
23 There are a list of major advances on your first 24 page of your handout, and I thought I would try to go through n/
w-25 them. And then I brought back something which you have seen s
44 1
before, which is what does the comparison mean if you look at A()
2 a sequence as it was binned into a release cateDory in 3
WASH-1400 versus how do we see that single sequence today in 4
the BMI-2104 analyses.
5 I think that at least that is what I interpreted 6
your question to be.
7 MR. KERR:
Let me see.
I think you interpreted it 8
correctly.
If, for example, you are going to take fission 9
products chemistry and say here is an example of what we are 10 doing in WASH-1400 and here is how the new information on 11 fission products changes some.
12 MS. MITCHELL:
Yes.
Well, not piecemeal The
(
13 information, say, in a vacuum, what did we learn, we changed U) 14 the iodine chemistry from the elemental and melkeular iodine 15 that was assumed in WASH-1400 and made cesium iodide.
There 16 is no one --
17 MR. KERR:
'But can you~new pick out something that 18 is different as a result of that and how different it is?
19 MS. MITCHELL:
No, I don't think so.
Not the one 20 item.
We made changes in the core melt progression in the 21 hydrogen generation in the time of containment failure. All 22 together, the entire package of codes was used.together, and 23 to isolate 24 MR. KERR:
Although you know that the fission r
w-25 product chemistry is different, it is impossible to pick out
4G 1
what effect that had on the new results.
O j
g 2
MS. MITCHELL:
To give you a quantitative number, we l
1 3
don't have a quantitative number because it has not been 4
determined without changes in retention in the RCS.
5 MR. KERR:
Well, it is a little difficult to know 6
whether an improvement has been effected or has occurred if it 7
can't be identified, and one can certainly say as a general 8
principle that if you use better information, your final 9
result ought to be better.
It would be comforting if one 10 could point to places in which it is obviously different.
And 11 I don't think this is easy. Don't misunderstand me.
It is 12 just that it would be nice if one could do it, and then you 13 would know here is what happened when you changed.the fission 14 product chemistry.
15 The risk is obviously less or greater or different 16 17 MR. SHEWMON:
There are a group of tables here in 18 the second page of your handout which show numbers which 19 presumably she believes in or got someplace.
20 MS. MITCHELL:
Yes, but they are not the result of a 21 single change.
That is the change from elemental molecular 22 iodine to cesium iodide.
23 MR. SHEWMON:
If professors just tried to tell 24 stories and give generalizations to their students, that is 25 not your way of doing things, so --
4 46 1
MR. WARD:
Yes, but I think it is contrary to the
.')
2 characterization of 0956 as a scientific report.
I mean I was 3
not as uncomfortable with that characterization as Bill was 4
earlier because it is ?.rying to provide you a better 5
understanding of this sort of thing, but if you haven't been 6
curious in developing thatt understanding about the sort of 7
question Bill just asked, I'am bothered again now.
8 MS. MITCHELL:
Oh, indeed.
If you look at the QUEST 9
study, which was for two specific plants comprising three 10 sequences in total, they went through and looked at changes in 11 modeling one at a time.
That is a presentation that I think 12 we have had before but is certainly much more extensive than
/"'
13 a half-hour presentation.
v MR. SHEWMON:
Someday maybe you will get to the 14 8
15 second page here, and if you do, I see here release fractions, 16 iodine with water and without water differing by darned 17 little, yet what started this all, in a sense, was the fact 18
,that TMI-2 was more than your factor of 2 different from 19 WASH-1400, and so how you argued yourself out of THI-2's 20 performance to arrive at the numbers you have here is 21 something I would look forward to.
22 MS. MTTCHELL:
This is a containment failure case.
23 That particular one that you point out as the V sequence 24 bypasses the containment.
( )
N-25 MR. SHEWMON:
On a concrete containment?
e,h' 47 1
MS. MITCHELL:
The V sequence bypasses it.
In TMI-2
'(%d) 2 it released water to an auxiliary building. That's just not a 3
containment failure case in the same sense that the V sequence 4
5 MR. SHEWMON:
Bypass the containment?
Maybe you can 6
help me and say what you mean by bypass.
By containm'ents 7
there, there was a pipe that went out and that's what you mean 8
by a bypass?
9 MS. MITCHELL:
Yes, right.
The V sequence is the 10 interfacing system LOCA.
11 MR. SHEWMON:
And where did the water stay?
The 12 main point of the water is that the cesium iodine is soluble
(s, 13 in water, I thought.
Nj 14 MS MITCHELL:
With water and without water refers 15 to the situation in the aux building at Surry.
Specifically 16 where does the interfacing system LOCA happen in Surry?
Maybe 17 at the first elbow following the isolation valves that failed 18 in order to give you the V sequence, and whether or not the 19 refueling water storage tank drains back out through the other 20 end of this pipe failure and puts a layer of water -- I think 21 it was 3 feet high -- over the release location for all of the 22 fission products that were coming out of the core which melted 23 inside.
l 24 Now, at Surry the building into which this LOCA O
25 would take. place is a building that has a very large, very i
s-e..,,
I i
48
[
1 flimsy door that would not last through the blowdcwn from the f
2 primary system for the failure, so that in essence, in Surry 3
the building is open to the atmosphere from there on.
The l
4 question is will the location be under water or will it not?
I 5
MR. SHEWMON:
Fine.
Thank you.
l 6
MR. WARD:
Okay. And if it is under water, the 7
iodine release is about a factor of 5 less than -- and I guess 8-that was Paul's problem. That doesn't seem as spectacular as 9
some of the early --
10 MR. SHEWMON:
I would wonder about the robustness of l
11 that number, and it must come from --
12 MR. SILBERBERO:
It's related to scrubbing.
/N 13 MS. MITCHELL:
It's related to pool scrubbing from a
\\uJ 14 pool of water that is hot because you have blown down into 15 that pool of water.
16 MR. SHEWMON:
So the failure is assumed because 17 that's the most conservative thing, not the most realistic; is
{
18 that right?
19 MS. MITCHELL:
It's a check valve failure which l
i l
20 allows a blowdown into the low pressure piping, and if it's 21 assumed to fail at an elbow, would you push the whole rest of 22 the pipe away?
23
.MR.
KERR:
But now this factor of 5 would almost 24 exist, probably, if you talked about elemental iodine rather 25 than cesium iodide because elemental iodine itself is fairly 1
_ _ _ _ - - - _. - -,, _ _ _. ~. _ _ - _ - - _. - _ _ -.. -. _ _ _. _.. _ -.. -, _ _ - _. _ - _, -.
I 49 i
i soluble in water, so it is unlikely that the iodine chemistry 2
change had much influence on that number.
(,
t 3
MS. MITCHELL:
That's true.
4 MR. SILBERBERG:
You see, that's the point we are
(
5 trying to make. It will vary.
In other words, the importance l
6 of cesium iodide will vary with the sequence.
7 MR. KERR:
Mel, I am with you all the way, but what l
8 I was trying to explain is what I would have liked to see as 9
an evidence that, yes, the iodine chemistry change had a 10 significant impact on risk, and here is how we know.
l 11 MS. MITCHELL:
You see, common to both of those 12 analyses is the assumption of cesium iodide and its aerosol
/"N 13 deposition. processes from the core all the way out to the 14 break, so that its transport within the RCS is as cesium I
l 15 iodide, and that is common to both.
The difference between i
16 them is only water above it in one case and not above it in 17 the other case.
18 MR. KERR:
My point was, had WASH-1400 done those 19 two cases, with and without water, it probably would have f
l 20 gotten about this factor of 5 difference in release.
Even i
21 assuming --
22 MS. MITCHELL:
No, I don't think so because 23 WASH-1400 gave credit for the MARK I suppression pool, the DF 24 of 100 for cold pools, and a DF of unity for hot pools. This 25 would be a hot pool; therefore, I presume, though I can't go L
50 1
back and say what they might have done, that they would have A()
2 had a DF of 1 3
MR. KERR:
But if they had been realistic about 4
iodine's solubility in water, and it is certainly a good bit, 5
and that was done even before people started talking about 6
cesium iodide, it is certainly unrealistic to use a DF of 1 7
when you bubble elemental iodine in water.
8 MS. MITCHELL:
They did.
9 MR. SHEWMON:
He didn't say whether they did or 10 not.
It say whether it was realistic or not.
11 MS. MITCHELL:
Oh, okay.
12 MR. KERR:
But go ahead.
We are interrupting too
[~
13 much.
iL 14 MR. SILBERBERG:
Let me say that to the extent that 15 we can, as Joclyn goes through these advances, perhaps provide 16 at least a partial insight as to where we can and might impact 17 and try to answer your questions, Dr. Kerr, we will try to do 18 that 19 MR. CATTON:
You have listed on these major advances 20 retention in reactor coolant system and core melt progression.
22 MS. MITCHELL:
Yes.
22 MR. CATTON:
I recollect that Mel one time, when he 23 was asked what was the most important variable, he said 24 temperature, temperature, temperature.
Now, when you go x-25 through these, I would like you to tell me how you have made
51 1
1 advances in estimating the temperature because, to my 2
knowledge, it is still an input into MARCH, and as long as q
f.
-3 that is the case --
4 MS. MITCHELL:
The temperature is an input to MARCH'?
5 MR. CATTON:
Yes.
You don't calculate the 6
temperature.
7 MS. MITCHELL:
Do you imply that because you have 8
chosen types of parameters that the temperature must fall out 9
as a given?
10 MR. CATTON:
Well, if you don't couple the core melt 11 somehow to its boundaries as something you know, then you are 12 not calculating the temperature in any realistic way.
- Anyway, 13 it's just a comment. And as far as retention in the reactor se 14 coolant system, until you somehow have built in this scenario 15 that was initiated by the Dutch in 1981 and is now pursued 16 heavily by EPRI, you haven't really made much of-an advance, 17 as far as I '- m concerned, in retention in the reactor coolant 18 system.
19 MR. SILEERBERO:
And that is listed as an open item 20 in Chapter 7,
and we agree that it's an area that --
21 MR. CATTON:
It was listed here as a major advance.
22 MS. MITCHELL:
Ah.
But we have stated that some 23 of the areas that we have called major advances are still word 24 for word in our' list of major uncertainties, that we have done
- O->
25 things which are better than what was in WASH-1400, but we m
62 1
haven't gone the whole way yet, and that core-concrete
/
2 interactions are listed as an advance.
3 They are also listed as an uncertainty. Containment 4
pressure loads is listed as an advance, and you will find it 5
in-the list of uncertainties because we have made progress but 6
not answered all the questions yet.
7 MR. SHEWMON:
Ivan, what the Dutch were concerned
~8.
about was hot gas rupture of piping above the pressure vessel 9
MR. CATTON:
That's right.
Actually, it was the 10 recirculation between the heated core and the rest of the 11 primary system.
12 MR. SILBERBERG:
I suggest that'we go on item by 13 item, and then as things come up, we can address them.
14 MS. MITCHELL:
Okay.
The fission product chemistry 15 advances have been both in the in-vessel and the ex-vessel 16 areas.
One of the major ones that we talk about most in the 17 in-vessel is the consideration of cesium iodide rather than 18 elemental.
The justification for this is basically a 19 thermodynamic consideration and the fact that there is a great 20 excess of cesium to drive the reaction in that direction.
21 There are caveats that go along with this advance 22 that elemental iodine or molecular iodine may very well be 23 formed in the containment during hydrogen burn -- some of that and also from reactions either in 24 has been seen recently O
25 the vessel or elsewhere with the boron; for instance, from the l
i 53 l
i boron carbide control rods in BWRs or from the boron control (O) 2 shim in PWRs.
3 And a question that is still unanswered is the l
4 processes by which you would get iodine reacting with some 5
organic compounds to form volatile organic compounds.
6 Another in-vessel process that has been taken into 7
account versus fission product chemistry is the tellurium 8
reaction with unoxidized sircalloy cladding.
This is a 9
specific example that can be either good news or bad news, 10 depending on the sequence in the plant.
11 The good news is that it doesn't come out in-vessel, 12 and the bad news is that it can be retained to end up on the
N 13 floor of the containment and come out late in the accident 14 sequence.
So I think that that is a change which is an 15 improvement, but you can't say as a generic statement whether 16 it is something that went to make larger source terms or 17 smaller source terms.
18 MR. CORRADINI.
Can I ask a question?
You brought 19 up that one, and I was going to ask, if I go back to your list 20 of comparisons with Surry -- I am looking ahead, and maybe 1 and Peach Bottom, is that the main reason in 21 shouldn't be 22 the Peach Bottom TCC sequence why I go from 1 in 100 to 48 23 percent release of the tellurium in the TCC sequences 24 specifically because of the holdup of the tellurium with the
,s 25 strcalloy?
54 1
MS. MITCHELL:
Yes.
You hold up the tellurium with
("%
k, )
2 the zirc.
You have a substantial amount of ziro in the 3
channel boxes so that you retain a substantial amount.
4 MR. CORRADINI All right.
Now can I ask kind of an 5
unrelated question?
6 MS. MITCHELL:
Sure.
7 MR. CORRADINI:
It is of interest to me.
8 Now, given this change from WASH-1400 to the current 9
set of knowledge and how then it fits into the systems
~
10 analysis using the Battelle codes, where does IDCOR's 11 calculation -- do they disagree with this change in physics or 12 do they agree with i t?
13 MS. MITCHELL:
I think they agreed.
I think that 14 they have agreed to take this effect into account.
Whether 15 they have adopted the specific method or model used in the 16 code, I am not sure, but the retention by unoxidized zircalloy 17 they will take into account.
18 George Marino may have a comment more on that 19 MR. MARINO:
I would like to clarify.
They did take 20 that into account from our meetings and all the experimental 21 evidence we had that showed that tellurium can be retained by 22 unoxidized zirealloy.
However, the later stage of the 23 accident, the way they modeled the core-concrete interaction 24
.and the heat exchange between the core melt, material and the 25 containment freezes the core melt such that the release of
.,,,.m.my m.
,c_.. _., _.
,_.___-,.-_7.
,.__.,,,r_,,,
55 1
tellurium at that stage of the accident is still nil for O)
(
2 them.
So they see no effect of putting in the tellurium w/
3 retention model because their later model negates its effect.
4 MS. MITCHELL:
I think also their models for 5
containment failure mode of containment failures also tend to 6
make several of these effects extremely small.
7 MR. CORRADINI I know I'm not supposed to think 8
ahead, but I am thinking ahead because what Denny had said 9
previously was it is the IDCOR methodology and the variance of 10 the IDCOR methodology that is going to be used by the utility 11 people to decide if they are better or worse, and if there is 12 still, as I sense, not unresolved but, it seems to me, an
[
13 unresolved set of calculations down here, then somewhere along N
14 the line there is going to have to be some resolution of that.
15 MS. MITCHELL:
You bet.
16 MR. SILBERBERO:
In fact, there is a p?ccess that 17 has been taking place that Dr. Rostoczy has been chairing 18 along with people from Research is do just that, and when that 19 process is over and if, indeed, there are important 20 differences, somehow I assume that the statement on 21 accep'tability of the IDCOR generic applicability may, in fact, 22 have to include an exception like that 23 MR. CORRADINI-All right 24 MR. BENDER:
In trying to understand these A
25 sequences, it is necessary to have some understanding of what u
56 1
i s' implied by them.
The V sequence, I think, implies that
)
2 overpressure of containment will not occur.
g 3
MS. MITCHELL:
That is true.
It totally bypasses f
4 the containment.
The containment is absolutely intact.
l 5
MR. BENDER:
So the release mechanism for 6
radionuclides in that case cannot involve a bursting of 7
containment, so one might want to argue that for that kind of I
8 thing, retention processes are dealt with differently than 9
they would be in the case of the TMLB' event.
10 Now, I don't know whether you are able to discuss 11 that here, but somewhere along the way, to make a reasonable 12 comparison between WASH-1400 and the current methodology, we 13 need to get some of that kTnd of understanding out in the 14 open.
I think it's too complex to talk about here but you 15 have to do it somehow.
16 MS. MITCHELL:
The mechanisms are not different.
17 They may result in different answers at the end.
The TMLB' 18 sequence for in-vessel considerations is, as we model it 19 today, a high pressure sequence.
We don't take into account 20 the possibility of rupturing either a pipe inside containment 21 or the steam generator tubes.
We model as a high pressure 22 sequence with a-long residence time for fission products 23 within the reactor coolant system, which results in a very 24 high retention of those fission products which are released 25 in-vessel by agglomeration and settling and all those other
_. -.. - _ - ~
57 1
processes.
2 For the V sequence, it is not a high pressure 3
sequence, and so the residence time is shorter; but you do 4
have a long run of pipe outside of the RCS down into the s
5 safeguards building, but you do not have the following vessel i
6 failure and the core on the floor.
You don't have an 7
overpressure possibility in the containment.
i 8
MR. BENDER:
Well, I am trying to think, as everyone 9
inevitably does, about the future application, and so you t
i 10 think in terms of, well, how is it going to influence 11 emergency planning, how might.it influence a release of the I
12 radionuclides?
If they do get out of containment, where is
"')
13 the fallout going to be?
And maybe it has
.t o do with what
[V I
14 really comes out.
i 1
15 The high pressure system when it bursts containment 16 carries a lot with it, and the low pressure system has no 17 driving force behind it, and even if it gets out, it won't go i
-18 anywhere.
19 Now, there is a certain amount of that understanding 20 that has to be developed, and I'm not quite prepared now to be 21 able to assess it without somebody telling me more about what 22 assumptions go with the time when whatever is in there departs 23 from containment.
While it is in containment, I don't care in 24 terms of public threat.
I may care for some other reason, but 25_
I really don't care in terms of the public threat until it 4
I
58 1
gets out.
So the circumstances when it happens are just 2
crucial to the whole issue.
3 MS, MITCHELL:
I totally agree.
4 MR. KERR:
Take that as a statement and go ahead.
5 MR. BENDER:
That's all it is.
6 MS. MITCHELL:
The ex-vessel changes in fission 7
product chemistry are basically the incorporation into the I think it's 125 different chemical species, 8
VANESA Code, a 9
and to take into account the chemistry changes that are 10 occurring in the melt on the floor of the containment due to 11 different types of gases which may be bubbling through, 12 sparging through the melt at different times.
13 MR. CORRADINI' Can I ask a question there?
I'm
\\
14 sorry.
It's kind of a special point.
But I didn't 15 understand something when VANESA was first developed, and I 16 thought I might as well ask it.
CORCON already has the 17 capability of following fission products.
It just never was 18 utilized.
Why was VANESA developed in difference to it?
19 MS. MITCHELL:
I don't necessarily buy your l
20 statement that CORCON follows l
21 MR. CORRADINI-The chemistry package in CORCON is a i
22 general chemistry package.
All that was required was to put 23 in the species of the fission products to follow them and they 24 would be automatically released.
25 MR. KERR:
Mike, is this just a question of
.-. _ _ _ _ - - -. _ _. ~ _. _., _ _
l l
59 l
1 curiosity or does it have something to do with 0956?
2 MR. CORRADINI:
It is a specific question so maybe I 3
should just drop it.
4 MR. KERR:
Why don't you pursue it later on?
I 5
don't mean it's not an important question, but
~
6 MR. CATTON:
We are going to get to core-concrete 7
interactions later.
I think we both have questions with 8
respect to that.
9 MS. MITCHELL:
Okay.
10 Well, we are still in the short top part of the list 11 in the reactor coolant system.
The retention in the reactor 12 coolant system was not accounted for -- the potential for it
/
13 as a function of the sequence was not accounted for in v
14 WASH-1400.
It was recognized in the work that was done for 15 that document that the potential existed for retention, 16 especially of the volatile fission products, and it was felt 17 that because of revaportsation, which is listed in the areas 18 of uncertainty, that the retention would take place but that 19 it would be only a time delay before it was released later, 20 and so they didn't take it into account.
21 MR. KERR:
Ms. Mitchell, I am going to suggest that 22 you assume that most of us are sort of familiar with the 23 significance of these statements and rather suggest that, l
24 insofar as you can, you point out the changes in risk or
(
\\'
25 something that occurred because of these changes.
i 60 1
MS. MITCHELL.
Okay.
(
2 MR, KERR:
Either qualitative or quantitative, 3
MS. MITCHELL:
For the retention in the reactor 4
coolant system, it is important for those sequences that !
l 5
mentioned that have a long retention time, a long residence 6
time in the RCS.
For instance, the retention for cesium 7
iodide or cesium hydroxide for TMLB' for Surry is about 85 8
percent, where for the AB sequence it is about 6 percent, or f
9 n e g l'i g i b l e.
i 10 MR. KERR:
Okay.
Now, in WASH-1400 instead of being 11 85 percent, what was it.
12 MS. MITCHELL:
Nothing.
There was no retention.
13 For some sequences you will get retention; for other i
14 sequences, you will not.
You will not have high retention.
e 15 The mechanisms are agglomeration and deposition, evaporation, 16 condensation, the processes which are taken into account by 17 which you end up with a net retention.
18 MR. CATTON:
With respect to retention in the I
19 reactor coolant system, I have heard arguments that the 20 deposition takes place, then it migrates through the steam 21 generator tubes and then it's out again.
How do you address 22 that?
23 MS. MITCHELL:
The specific question about whether 24 or not it will end up in the steam generator tubes.
25 MR. CATTON:
If it gets into the steam generator 4
l 61 l
l 1
tubes, it's obviously going to get back out again, and the (G) 2 path is a more serious one.
i 3
MS, MITCHELL:
If'the steam generator tubes fail.
I 4
MR, CATTON:
Well, if it gets there, they will.
(
5 MS, MITCHELL:
That is being evaluated.
6 MR. CATTON:
Well, what they feel happens and 1 I
7 sent the three papers in.
~
8 MR. KERR:
Ivan, you made the point and they agreed I
9 that it has to be investigated.
10 MS. MITCHELL:
It is being investigated.
11 MR. KERR:
I don't believe we are going to settle 12 the issue, 13 MR. CATTON:
Well, the only reason 1 am raising the 14 point is I don't think they can say anything about the 15 differences between WASH-1400 and what they have here until 16 that question is addressed.
17 MR. KERR:
You have made your point.
I think they 18 understand.
19 MS. MITCHELL:
Okay.
20 The core melt progression also includes the 21 consideration of the generation of hydrogen in vessel The 22 code suite is not mechanistic in this area, but since the time 23 of WASH-1400, there have been some experiments for 24 instance, the PBF in-pile test or the KFJ out-of-pile test 25 that are used to let me use the phrase " guide input" and to
I 62 1
choose model options, where the code, as I say, is not
/N()
2 mechanistic.
3 Some of the things that have been looked at are the 4
liquefaction with sircalloy rather than true melting of the 5
UO-2, the slumping behavior and the behavior of the control t
6 rods.
There are some changes that are made in.the code suite, 7
the source term code package relative to the treatments of l
8 silver, indium, cadmium, control rods versus what is done in 9
the BMI 2104.
10 MR. KERR:
Would you say that the result of the 1
11 changes in core melt progression would indicate a higher or 12 lower risk compared to the WASH-1400 prediction?
[V)
13 MS. MITCHELL:
I think that is a plant-specific
/
14 question.
The changes 15 MR. KERR:
Well, let's take Surry and be plant i
16 specific.
17 MS. MITCHELL:
I think for Surry that it probably i
18 has lengthened the sequence that the overpressure of a 19 combustion of hydrogen is delayed versus what they thought.
20 But I'm not sure that you can make that same statement about i
21 Peach Bottom.
22 MR. KERR:
So in Surry it leads to a decrease in 23 risk?
24 MS. MITCHELL:
Yes.
("
25 MR. XERR:
I'm not trying to make you generalize.
63 1
MS. MITCHELL:
Okay.
I would be cautious in that 2
specific question, I would be very cautious about other
( f 3
plants.
4 MR. KERR:
I would, too.
5 MS. MITCHELL:
Okay.
The aerosol behavior in 6
containment also includes ice beds and suppression pool 7
effects.
The retention in the containment, in the RSS was 8
empirically based, while NAUA has the mechanistic treatment of 9
the same types of phenomena that give you deposition in 10 TRAPMELT, that is, agglomeration and deposition and all those 11 other wonderful things.
12 The RSS was arbitrary in their' pool DF.
We already 13 mentioned that it was a factor of 100 for cold' pools and unity 14 for hot pools, and of course, they did not treat an ice 15 condenser system, so I don't know what they would have done 16 with the ice.
17 The SPARC and ice DF are mechanistically based.
18 There are some experiments which are available now to compare 19 with the SPARC calculations, and there are in the future going 20 to be some experiments to qualify ice DF.
21 The fission product release from the fuel was at the 22 time of the RSS~ empirically based, and it still is, but it is 23 based on improved data both from Oak Ridge and from KFK.
24 MR. SHEWMO!4 :
Have you gotten into whether plants 25 with silver cadmium control rod materials would be appreciably i
I s
l
64 l
I 1
different in aerosols containment than those with hafniums, or (j
2 coron carbide.
In the one case you would have a good deal l
3 more volume of aerosols than the other.
I 4
MS. MITCHELL:
Yes.
The BMI-2104 assumed that there i
(
5 was a significant nonradioactive silver aerosol.
As I 1
6 mentioned, that has been changed based on some new information 7
for the source term code package.
8 MR. SHEWMON:
My question was does it make a 9
difference whether they have silver or cadmium, or have you 10 tried to treat that?
l 11 MR. MEYER:
We haven't treated hafnium, but the 12 hafnium in the model, I am sure, would hat e essentially zero 13 aerosol.
v 14 Mk. SHEWMON:
That's my point 15 MR. MEYER:
The difference is probably not as large 16 as you might think at first.
The principal effect of the 17 large mass of silver in the aerosol is to enhance the 18 agglomeration and settling, so.it principally leads to greater 19 retention in the primary system.
If you look at the numbers i
I can't recall the numbers, but 20 and take that out of there 21 my recollection is that it would make a percentage change 22 rather than an order of magnitude change in the retention by 23 taking it out.
l 24 MR. SHEWMON:
Okay.
?
(~%
25 MR. CATTON:
With respect to the aerosol Lehavior,
65 1
has the question been resolved about why the German tests
(
2 showed no aerosols?
3 MS. MITCHELL:
We are down in the next to the last 4
item in core-concrete?
5 MR. SILBERBERO:
That will treated in that subject.
6 MR. CATTON:
But aerosol behavior in containment?
7
'MR.
SILBERBERO:
We
- p. refer to treat it there, but it 8
doesn't matter.
9 MS. MITCHELL:
I look at it as a core-concrete 10 problem, not an aerosol.
This is where one has some in the 11 containment, what does happen to it.
12 We are down on fission product release from the
/~h 13 fuel.
I already mentioned the tellurium holdup is accounted 14 for in the fission product release from the fuel based on 15 experimental data.
I do want to point out that this is not 16 for terminated accidents.
We look upon this as one of the I think this goes to one of your questions that are 17 areas 18 received, what about terminated accidents.
19 This is one particular place where one would have to 20 be cautious because the empirical correlations on which CORSOR 21 is based are melted fuel, and there have been comparisons for, 22 say, the PBF tests for whtoh not all the fuel is melted with 23 the CORSOR or NUREG-0772 predictions, which is where CORSOR 24 came from, that show that there are not close correlations and n
\\/
25 you better be careful if you are going to terminate an l
66 1
accident at a place.
You may do just fine when you get it all 2
done at the end, but if you are going to stop it in the 3
middle, you ought to be cautious.
4 MR. CATTON:
How important is temperature to fission 5
product release from fuel?
6 MS. MITCHELL:
The mechanism in CORSOR is a 7
temperature correlation.
We point out that this is not true.
8 It should be 9
MR. CATTON:
Have you done sensitivity studies and 10 tried to estimate things like the heat transfer coefficient to 11 see if they are reasonable?
12 MR. SILBERBERO:
Variations of heat temperature
(N 13 coefficient within the core?
14 MR. CATTON:
Within the core and its boundaries.
15 And I know the code doesn't have one, but you could at least 16
. estimate it from the temperature that you assume.
17 MS. MITCHELL:
Is this a CORSOR problem or is this a 18 MARCH problem in calculating the temperature which is fed into 19 CORSOR?
CORSOR has only a temperature dependence.
You have 20 to have the right temperature.
21 MR. CATTON:
But you have guessed the temperature, 22 and what I am trying to establish is whether or not it's a 23 reasonable one.
24 MR. SILBERBERO:
Well, we could go right to the C-)
25 horse.
Pete?
. _ _ _ _ _ _ _. _ _ _.~ ~
67 1
MR. KERR:
Do you understand the question?
2 MR. CYBULSKIS:
I'm not sure I want to answer, being i
3 addressed in such a manner, but I'll try anyway.
I l
4 With regard to the point that Ivan is raising, !
l 5
think we do input a core melting temperature into 6
MARCH. That's input. That's an assumption that tends to limit t
7 the maximum core temperatures that we predict.
It does not l
8 totally limit it.
You can exceed the melting temperature.
In 9
a typical calculation, you will not.
10 As far as the heat transfer to the boundaries of the 11 core, MARCH is coupled to the boundaries both by radiation and 12 convection.
13 MR. CATTON:
Excuse me.
If you fix the temperature 14 in any way, then the heat transfer is also affected.
Have you l
15 tried to estimate what the effect of the heat transfer i
16 coefficient at that fixed temperature is, or do you care?
17 MR. CYBULSKIS:
Perhaps the best way I can answer 18 the question is in earlier versions of MARCH, the core was 19 basically decoupled from the boundaries.
In the current 20 versions, it is coupled to the boundaries.
We do not see a 21 major effect on core melt progression or timing of events.
We i
22 do see some of the structures, particularly the structures l
[
23 immediately adjacent to the core, getting very hot, in some I
24 cases approaching the melting point in a number of the t
O
. \\)
l N
25 sequences.
i
~
l
i 68 1
MR. KERR:
We will have no more orthoganol
(
2 discussion in this meeting.
3
[ Laughter.3 4
MS. MITCHELL:
Okay.
5 Core-concrete interactions.
This, as I mentioned 6
before, appears both on the list of major advances and the 7
list of major uncertainties, and we went out of our way to use 8
exactly the same words to make sure that everybody recognised 9
that they were the same area.
The Reactor Safety Study had a 10 very simpitfied view of core on the floor processes.
Now we l
11 have a more mechanistic view in CORCON and VANESA, which are i
12 based on some large-scale tests, f
13 1 think there are certain insights that have come
\\
! should not use that word.
I 14 out of that that the inert 15 should use nonradioactive.
The nonradioactive materials are 16 important, and in most of the cases are calculated to be 17 copious.
18 You need to take account of the heat and mass 19 transfer, and you need to take account of chemical reactions 20 and chemical changes in the melt and in the gas phase.
There this was 21 is a lot of way to go yet.
People have brought up 22 brought up just earlier, the seeming contradiction between 23 Sandia tests where large amounts of aerosols were viewed, I
24 and the beta tests at KrK, where the amount of aerosol was i
k 25 perceived to be much less, and whether or not that is a i
i
. ~ - - - - - -. -_-..--. -
4 69 l
i 4
1 contradiction-1
(
2 MR. CORRADIN!"
There are three experimental 1
3 differences that ! am aware of.
One is that.
Two is that the 4
erosion rate in the dowcward direction is significant in beta, 1
e i
5 and I have not seen any published data from Sandia that shows I
6 the downward erosion rate being any different from steady i
i 7
state tests to transient tests to confirm what the beta
]
l 8
tests have shown.
I I
9 And the third thing is that at high gas velocities, i
I i
j 10 you essentially get an emulsification of the two layers.
i i
l 11 There is no layered pool any more.
All three of which could j
12 have a very strong impact on source term release.
t i
j 13 MS, MITCHELL:
Yes. There also are other i
14 differences, in that the temperatures are very different in i
15 tests, and that also has i
f I
i 16 MR. CORRADIN1 The downward erosion is coupled. If 4
o 3
}
17 I erode that fast, I cool off the pool because I am just i
J 18 basically diluting the mass with cold ablated material.
1 I
19 MS. MITCHELL:
We brought our local expert,' Fred i
F 20 Burson, to talk about i
l 21 MR. BURSON:
I would like to suppress, to some i
)
22 extent, if possible, this I won't call it myth, but i
23 whatever it is, misperception about the disparity between l
24 aerosol generation in the beta tests and Sandia.
I talked at i
4 25' great length to the people out there yesterday on the subject,
{
l
t I
L 70 1
and I think the first thing to say is that the disparities 2
that are seen 'are not inconsitent with what one would expect.
j 3
The German concretes are highly basaltic, and one l
i l
l 4
does not expect to see the aerosol production from basaltic
\\
I 5
concretes that you do from highly carbonaceious concretes.
1 1
6 That is the first thing.
7 The test that is being compared, as we refer to as 8
contradictory, was one of the coil tests, which was done with 9
completely, well, what we used to call CRBR concrete, totally i
10 11mestone concrete, plus the fact that the steel and by the 11 way, it's a pretty large test, about 200 kilograms of 12 stainless steel.
The steel was indeed stainless steel and 13 contains a fair amount of magnesium.
14 The German concrete is a thermitic seal 15 MR. SHCVMON:
You mean it's stainless?
16 MR. BURSON:
There is some in it.
i 17 MR. SHEWMON:
18 MR. BURSON:
I beg your pardon.
- Now, 19 that is a very large contributor for producing aerosols. So in 20 the cases where the concretes at Sandia had smaller amounts of 21 basaltic concrete, they do not see this high amount of 22 aerosols.
They see them, but it was only in those coil tests 23 where the concrete was almost entirely carbonaceous, and that 24 is the source of this great concern.
1 l
25 Another point that ought to be made is the fact that
71 1
there were no quantitative aerosol tests made at the O
2 facility.
I am going to make an interesting comparison of the (V
3 test that was done two weeks ago, which I think is going to be 4
significant.
5 This is a visual observation, and speaking 6
authoritatively, one really cannot trust visual observations 7
too far.
Even if you have 100 percent good eyeballs, it's a 8
very, very nonlinear device, and it's not a very good 9
instrument for making quantitative measurements.
10 In the beta test, the melt is observed from about 2 11 or 2.5 meters above the melt, and it is bottom lit, and it 12 also attenuates or reduces the sensitivities of visual
(h 13 observation of aerosols because of the small angle scattering (v/
14 of the light which is coming up towards the camera.
15 So quantitatively there might have been anywhere 16 from hundreds or tenths of a milligram per cubic meter, I mean 17 of a gram per cubic meter, up to maybe a gram, and still have 18 seen the same observations.
19 In other words, when you get up to 5 or 7 grams per 20 cubic meter, you get into the area where you are beginning to 21 get opague.
22 Now, I want to refer to the test that was done two 23 weeks ago.
As most of you know, we have arranged with the XFK 24 people to do some tests with American concrete.
The first
(
)
N/
25 test was done two weeks ago tomorrow, as a matter of fact, the
72 1
14th, and it was good and it was bad -- good news and bad 2
news.
g l
'%s I will give you that first 3
The bad news was 4
that the oscillations of the load in the crucible were 5
sufficiently great'that it created some disturbances.in the 6
power supply and they blew a converter for about a minute and 7
a half of the test, so the test was run long.enough to get 8
some steady state observations but not nearly as long as we 9
would like to have had it go.
l 10 Now, the i nteresting thing is the follcwing. This
(
11 very dense cloud of aerosol was observed about 5 seconds after l
12 the pour began.
It took about 20 to 25 seconds for the
(}
13 pour.
The surface of the melt was not visible for about 12 to
\\s J 14 15 seconds.
By the time it had cleared up, the last five l
15 seconds of the pour were still visible.
In other words, it 16 was still pouring and this dense cloud of aerosol had, cleared l
17 up, and it didn't look any different than any of the previous
(
i 18 German tests had looked at that time, l
l l
19 So this concrete was a limestone, common sand 20 concrete, which is common to our reactors.
There will be l
21 other tests in November, which is the high limestone concrete.
22 and we do expect to see aerosols there, maybe five to ten l
23 times greater, i
24 In the test that was run, the gas generation rates
\\
25 were, as approximately predicted, about 2.5 to 3 times what l
l I
- +
m m m - -
m m
m.
73 1
has been seen in the silicious concretes of the Germans.
So
()
oh, I should make one observation.
It is at 2
~at this time v
3 least speculated, not sure yet, that the transient aerosol may 4
be due to a parting grease that is on the mold, and that' cloud 5
has been seen with the German tests, also.
It seemed visually l
6 to be not quite as dense, but on the other hand, in the common 7
sand concrete that was run two weeks ago, the gas rates, after e
all, were two to three times as high.
9 So I would caution pursuing this discrepancy between 10 Sandia and the German tests any further.
11 In answer to your question, Mike, earlier about 12 VANESA, the chemistry model in CORCON does not do anything
/~N 13 about release of the fission products from the melt.
VANESA (x
14 is a bubbling model which has to do with the sparging of 15 fission products out of the pool, and there are two 16 processes.
One is vaporisation.
In other words, they go with the gases from the fission products become 17 the bubbles 18 dissolved as the dissolved fission products become vaporized 19 and get into the bubbles and then carried out.
20 MR. KERR:
Brad, I ruled that discussion for a l
21 caucus, not related, l
1 22 MR. BURSON:
Okay.
I hope that answers some of the j
1 23 questions.
24 MR, KERR:
How, you have said that in your view
/
25 there is no significant difference between the German and the
t' 74 1
Sandia tests.
.O 2
MR. BURSON:
Not to the extent that it has been
( j 3
advertised, that's right.
O MR. KERR:
No, the word I used was "no" significant 5
difference.
Is that fair?
6 MR. BURSON:
I think that's fair, yes.
7 MR. CATTON:
Well, it looks like it is still open.
8 MR. KERR:
I am trying to get what Burson said.
My 9
interpretation of what he said was that there is no i.
i 10 significant difference.
In Catton's view, the question is 11 still open.
12 MR. CATTON:
That's correct.
13 MR. BURSOll:
There is a lot of analysis yet to be
("%
V 14 done in those tests.
3 15 MR. CATTON:
That is why it's still open.
I think 16 there are>still some questions in my mind, also, about 17 VANESA.
VANESA is a model based on small bubbles, much as s
18 chemical engineers frequently do.
When you look at the 19 process, it is'a very complex, two-phase slow process with 20 slugging and all sorts of things going on, and I'm just not 2i sure that VANESA is the proper code or is something that can 22 be used without a great deal of uncertainty, l
23 MS. MITCHELL:
Oh, I think you are right.
As I 24 said, we list it as an area of uncertainty.
( ~)
k%s/
25.
MR. CATTON:
It's easy to be off factors of 10 on
.-4-----
75 1
surface area in problems like that.
2 MR. KERR:
Where are we now, core-concrete 3
interactions?
4 MS. MITCHELL:
Well, I had finished with 5
core-concrete.
6 MR. KERR:
All right. Let's go to containment.
7 MS. MITCHELL:
Okay.
Pressure loads.
There have 8
been improvements in several areas in containment pressure 9
loads since the BMI-2104 calculations were done.
The INTER 10 subroutine that had been in MARCH, which was used for 11 calculating the ex-vessel gas generation rate for determining 12 the pressure loads, has been replaced with CORCON MOD 2 to get
(N 13 better calculations of the gas loading.
U 14 There also have been a better understanding of the 15 flammability limits for hydrogen-air-steam combination and 16 understanding about the combustion of hydrogen.
In that 17 respect we have mixing effects and geometry effects that can 18 be taken into account.
Also, advances have been made in the 19 ability to calculate the steam spike.
20 MR. KERR:
Mr. Shewmon.
Excuse me.
21 MR. SHEWMON:
No, finish.
I want to go back to 22 core-concrete.
23 MS. MITCHELL:
Go ahead.
24 MR. SHEWMON:
If one could get a few garden hoses or O) i
25 supply of water under the pressure vessel, it would go a long
76 1
ways towards ameliorating the reaction down there.
Could you
)
2 tell me what the status is currently about either a water 3
balance in the containment, about how much water would be down 4
there, whether the utilities would be encouraged or 5
discouraged from trying to get more water down there or given 6
credit for it or what the status of that is?
7 MS. MITCHELL:
As a prevention or mitigation system, 8
you are probably right.
One of the curves I showed when we 9
were talking about Chapter 4 was the TML calculation for 10 Sequoyah in which there is predicted to be a significant water 11 pool above the core on the floor.
12 Whether the core always becomes coolable is a 13 question that will be addressed by some experiments in the 14 ACRR under recoolability. But regardless of that, you do have 15 the scrubbing effect even if the core continues to chew at the 16 concrete because it's not coolable.
17 MR. SHEWMON:
If the core got through the water in 18 the first place, or if it gets thick enough, yes.
But the 19 answer to my question was not whether it was a panacea but 20 whether it was being seriously considered, or whether the 21 utilities would be given any credit for it by the Staff if 22 they did try to go that route.
23 As I understand, GESSAR II at least talks about --
24 MS. MITCHELL:
The UPPS system?
I'm forgetting.
I
-~M (V
25 think the UPPS system was basically to stop the core melt in
77 1
the first place and to be able to get water inte the core and i
2 not 3
MR. SHEWMON:
Well, what is the answer to my 4
question, then?
5 MS. MITCHELL:
Well, we take the plant as we find it 6
today.
In 1150 perhaps that would be something that we could 7
suggest as one of the things to look at as ways to get water 8
9 MR, KERR:
The answer to his question is you have 10 not looked at that?
11 MS. MITCHELL:
No.
For whatever sequence -- it is 12 a sequence-dependent and plant-dependent.
If it is perceived,
/ \\
13 as in TML for Sequoyah, that there will be water in the 14 cavity is taken into account.
15 MR. SHEWMON:
There are six or eight lines for 16 getting water in the neighborhood of the core now, and so my 17 interest is in the degree to which those and their use are 18 ever considered or whether it's so much more attractive to run 19 big tests in which you lump a lot of stulf into concrete that 20 we get carried away with that model only.
21 MR. KERR:
That sounds like a biased statement.
22 MR. CYBULSKIS:
If I may interject a word of 23 clarification, in the work that is being done in support of 24 1150, that is one of the key questions that we are trying to 25 address, whether there is or is not water in the cavity at the 1
l
78 1
time of reactor vessel melt-through.
2 MR. SHEWMON:
And what might happen if more water 3
was brought in?
4 MR. CYBULSKIS:
Principally, we are trying to take 5
into account analytically the water that would be expected to 6
be there given the sequence.
7 MR. SHEWMON:
Thank you.
8 MR. WARD:
But in your judgment, is there important 9
potential for improving the situation at certain plants by a 10 provision that water be added, or maybe you don't have an 11 opinion?
12 MS. MITCHELL:
If you can get the water -- excuse 13 me.
14 MR. LEE:
Tim Lee from Containment Systems Research 15 Branch.
16 I would like to bring up one test conducted at 17 Sandia to your attention, which may give you some idea of the 18 effect of the water addition to the core-concrete 19 interaction.
The Swiss test conducted at the Sandia added the 20 water 2 to 10 minutes after the pour or the melt on top of the 21 concrete, and the results showed that it doesn't significantly 22 slow down the concrete ablation.
23 MR. SHEWMON:
I am not surprised.
Thank you.
It 24 did?
It did not
\\
J 25 MR. CORRADINI Can I ask a question about this
i 79 1
because it's kind of important and I don't want to leave this.
4 2
MR. KERR:
Go ahead.
3 MR. CORRADINI And that is, it depends, water in 4
the cavity, whether it is helpful or not, depends upon the 5
initial conditions before that If I am asking the question: I 6
have water in the cavity and then the fuel comes down, that 7
may be a totally different effect than if I have a dry cavity 8
and then the melt comes down and then I have water addition.
9 MS. MITCHELL:
That's right.
10 MR. CORRADINI:
And I think what Professor Shewmon 11 was asking before is the effect of water' in the cavity, and 3
12 the only experiments that I know of -- and I may be wrong 13 about this, so he can correct me -- but the only experiments I 14 know of that address that, although at a very small scale, 15 were the Argonne experiments, the 1/30th scale of Zion, where 16 they did small mass pours, scale pours of melt i nto a
' 17 simulated cavity for Zion with and without water present, and 18 what they found was the amount of atmospheric heating and 19 sweepout were significantly affected by presence of water in 20 the cavity.
21 So I think water in the cavity, or making sure there 22 is ~ water i n the cavity is a very important factor before 23 vessel breach and maybe even after vessel breach, regardless 24 of the Swiss studies.
- 25 MR. LEE:
Excuse me.
I think the test you are
,,n r
-e,-
m
--.,-n,r---
.,;--.-...,,vn-_,n,,n.,-,,
..r-.-
,---.w.-,,,n,.,,,-.~.,,n.,.-
v..,-..
80 1
talking about is a high pressure ejection test, but the Swiss e-( w) no gravity.
2 test is just a 3
MR. CORRADINI Right. But I'm saying the cavity is 4
there and it could occur under a number of situations, and I 5
think water in the cavity is a very important thing to have, 6
both under quiescent purposes, where I've got a pool and I've 7
got water on top, as well as when I've got an ejection of 8
material into a cavity.
So high pressure or low pressure. And 9
the only experiments I'm familiar with that address the whole 10 range were the 1/30th tests at Argonne.
11 I may be wrong about that but I think that's true.
12 MR. KERR:
Mr. Marino.
O-13 MR. MARINO:
I would just like to add a comment to 14 the discussion.
IDCOR has been doing some uncertainty 15 analyses for their entire process, and in some of the cases 16 where they would get a large variation or a large uncertainty 17 in their results, they have then looked at what effect 18 operation action would have on that tremendous range of 19 uncertainty and have always.come to the conclusion that if you 20 can get water into the cavity vis-a-vis some operator action, 21 taking some water somewhere in the system and dumping it on 22 the cavity or bringing fire trucks up and hosing down the l
23 thing, the uncertainty in the source term just diminishes 24 dramatically.
\\/
25 So it is very important, if you can get water on
81 1
there -- but we cannot take credit for things that they have
/~~N
(
)
2 no operating procedures, so you would get a wide range of 3
uncertainty.
4 MR. SHEWMON:
Well, if the water was there before --
5 you know, you've got a lot of water in this plant, which is 6
likely to go to the lowest point.
That is part of it.
The 7
other, I agree.
Rattling around in my head yet is something 8
that Ivan or some group put out a while back, that if you
.9 could get a rebed down there and have water under it, it 10 doesn't take a lot of water to do a lot of good.
11 MS. MITCHELL:
This is known as a este catcher.
12 MR. SHEWMON:
That's one model, yes.
(g 13 MR. KERR:
Gentlemen, I'm going to suggest that all b
14 of these ideas are so important that we need to have a session 15 for designing this system later on; but today I want to try to 16 pursue 0956 and related issues.
17 MS.-MITCHELL:
Okay.
I only brought along these 18 comparisons with WASH-1400 -- you have already seen them --
19 in case there were any specific questions on them.
I didn't 20 intend to go over them again, but just to refresh your memory 21 on what you have seen before.
22 I think we have already talked about the V sequence, 23 and the ATWS.
That completes the presentation unless there 24 are specific questions.
/~N 4
A-25 MR.
KERR:
Okay.
A ten-minute break.
82 1
[ Recess.]
O 2
MR. KERR:
A 1.1 right, Where were you?
( j 3
MS. MITCHELL:
I was going to talk about (a),
(b) 4 and-(c), and Ralph was going to talk about (d) and (e) 5 MR.
KERR:
Sounds reasonable.
Why don't you go 6
ahead?
7 MS. MITCHELL:
Okay.
(a) is the methodology versus 8
the suite of codes.
I'm not exactly sure what you had in 9
mind, and I was going to ask you to rephrase the question.
10 MR. KERR:
What happened to that chicken who just 11 ducked out?
That was Mike's question.
So if you don't mind 12 going to (b) before (a) -- do you mind?
("'T 13 MS. MITCHELL:
No.
The core melt interaction 4
U 14 predictions,.as I read the'information that came a couple of 15 days ago, basically what you're asking about in this case is 16 steam explosions and direct heating.
And the question is:
17 how were they handled vis a vis 0956.
18 MR. KERR:
Let me ask a very over-simplified 19 question, maybe even simplified enough to be stupid, but on 20 the basis of what you have done up to now. would you recommend 21 that regulatory worry about in-vessel steam explosions as a 22 hazard, or forget about them?
23 MS. MITCHELL:
I am not the expert to answer that 24 question.
George Marino feels that he wishes to answer.
\\--
25 MR. MARINO:
I didn't mean to say I could answer the
i 83 1
question.
I was prompting you to say something.
But I will
,O( )
2 attempt it.
I think with the evidence we have today, that 3
regulatory implemention should not worry about steam 4
explosions at the moment.
There are some experiments planned 5
for the future -- I wish Bob Curtis was here to explain them that could shed some light on it.
6 better 7
But I think right now the evidence and the expert 8
opinion seems to be, don't worry about it as far as the 9
probability goes.
10 MR. KERR:
I don't find that in 0956.
I tried to 11 find i t '.
12 MS. MITCHELL:
No, you won't find it in 0956.
It
['N 13 comes out of the containment event tree.
As a matter of fact, 14 for the Surry containment event trees which are given in 0956, 15 in the -cent ral case they are -- the probability is zero.
16 MR. KERR:
Well excuse me, I don't see how you could 17 possibly get this out of an event tree because whether or not 18 you have steam explosions can be put into an event tree, if 19 you know whether or not you have steam explosions.
i 20 But I fail to see how the event tree can tell you 21 whether you're going to have a steam explosions or not.
22 MS. MITCHELL:
The event tree will tell you what the 23 consensus of expert opinion as viewed by the people who put it 24 together.
25 MR. KERR:
It will if you have told the event tree
84 1
to say that.
But the event tree won't produce that.
I'm j
2 trying to distinguish between science and a tool Now tSe 3
science, it seems to me, has to give you some information on 4
which you base your judgment about the risk associated with 5
in-vessel steam explosions, and since 0956 is dealing to a i
6 considerable extent with the status of the science, it seems 7
to me, since that's a fairly crucial issue on early 8
containment failure, that if you can get rid of that you have 9
gotten rid of one of the early containment failure modes that 10 at least in the past had been considered important.
11 I think 0956 ought to be as unequivocal as one can 12 be about that question.
Why not?
13 MS. MITCHELL:
Well, the containment behavior that
(~s 14 was assumed, the preliminary containment behavior for Surry, 15 assumed the probability was essentially negligible; it was 16 sero.
17 MR. KERR:
I'm-not talking about Surry; I'm_ talking 18 about in-vessel steam explosions generally.
And it could 19 happen in places other than Surry.
And it seems to me if 20 there's enough information so that you have some confidence in 21 making a statement about that, it ought to be in 0956.
22 MS. MITCHELL:
Mel is shaking his head yes.
I would 23 say --
24 MR.
KERR:
And I don't know how much confidence you j
f)
\\_)
25 have; that's the reason I asked the question.
l i
l' 85 1
MS. MITCHELL:
The 0956 is-the fission product r" \\
?
2 release and transport set of codes.
3 MR. KERR:
I'm sorry, 0956 discusses a lot of stuff l
t I
4 other than that.
One of the things it discusses is how much l
l 5
stuff gets outside containment.
And one of the very important l
6 early containment failure mechanisms could be that, if it is a l
I i
7 valid process.
8 MR. SHEWMON:
Fardon me.
Is that also the process 9
that requires that -- the head take-off issue, which I think l
10 is' nonsense.
l
(
11 MR. CATTON:
So do a lot of people.
i i
l 12 MR. KERR:
I'm saying that if the consensus among 13 the staff is it's nonsense, it seems to me they ought to say 14 it.
15 MR. SHEWMON:
Well, if they can justify it they 16 ought to stick their~ neck out.
i
(
17 MR. SILBERBERG:.There is a' simple solution to l
l 18 this.
The fact that there is a conscious judgment based on l
19 expert opinion and information and discussion on in-vessel l
20
' steam explosion varied within the context of doing the I
21 containment event tree work in 0956 -- one can extract that 22 body of opinion or that data back into 0956.
l
~
MR. KERR:
I sure had difficulty because I was 23 24 unable to extract it from my reading of 0956.
4 25 MR. SILBERBERG:
No, I'm saying that's something we
k 86 1
could do.
That is a way of doing it.
2 MR. CATTON:
Mel, why don't you just simply 1
3 reference in the report that steam explosion needs research.
4 MS. MITCHELL:
It is referenced.
It is referenced 5
in Appendix B.
6 MR. CATTON:
Because their opinion was that it was I
7 negligible.
8 MR. SILBERBERG:
Right, and that's probcbly what we 9
will do.
We will refer to that body of opinion.
10 MS. MITCHELL:
There is some disagreement with that 11 conclusion.
i 12 MR. CATTON:
But-you mentioned experts and
["'}
13 conclusions.
Supposedly, the NRC Staff brought together all
%/
14 of the experts in the field who have reached that conclusion.
15 MR. SILBERBERG.
I don't think it's an 16 insurmountable problem.
4 17 MR. KERR:
I interrupted you when you were talking 18 about core melt interaction.
19 MS. MITCHELL:
That was one of the two areas that I 20 gathered were areas of interest.
The other one vas direct 21 heating.
I did want to say, before we leave steam explosion, 22 that in 0956, in Appendix D, we talked about the fact. that 23 separate bin is not considered to be required for the steam 24 explosion sequences because of some work that was done by Dana
\\-
25 Powers indicating that the oxidation of ruthenium -- it is an i
_. -~.
87 1
early containment fai7ure mode but not one that has a bin 7~()
2 assignment necessary far it.
3 MR. KERR:
Okay, I saw that and I didn't realize 4
that that was a reference te the fact that in-vessel steam 5
explosions were considered to be very low probability.
6 MS. MITCHELL:
No.
I'm talking about what does it 7
do to the fission product chemistry.
8 MR. KERR:
But I also guess I was a little amused, 9
you give a reference a private communication from one of your 10 contractors.
11 MS. MITCHELL:
Well, by the time NUREG-0956 is 12 published final, there will be a report which will be
(g 13 citeable.
It is a report that was written for the Grest 14 people and is now not available to cite.
That's a problem of 15 trying to put something in --
i I
1C MR. KERR:
The only way it could see the light of 17 day would be for some newspaper reporter to use the Freedom'of i
i 18 Information Act, and then it would b e' available.
19 MS. MITCHELL:
No.
20 MR. CORRADINI' Has it just not gone through Sandia i
21 approval?
Is that the problem?
22 MS. MITCHELL:
No, it was written for another venue; 23 it was written for OECD.
And those particular reports are 24 restricted, so if you called it on an FOIA one wouldn't get it
[A\\
\\-)
25 anyway.
t y
_._,....m
,.,_,_,..__,.,m,
88 i
i 1
MR. CORRADINI:
But if Dana just simply issued a m
2 Sandia report with the same words in it --
3 MS. MITCHELL:
I was going to say it will be-out 4
because as I understand it, it will be in the IAEA in late 5
October, early November.
So by the time 0956 comes out final, 6
there will be a reference which will be other than what is 7
there.
8 MR. KERR:
Okay, go ahead.
9 MS. MITCHELL:
Direct heating is the phenomenon of 10 high pressure injection transferring chemical energy to 11 the containment and causing early containment failure at 12 vessel failure, concurrent with vessel failure.
\\
13 This is handled also in the containment event trees 14 by having a split fraction between sequences that have early 15 containment failure due to that phenomenon and those that 16 don't 17 From the point of view of fission prod.uct chemistry, 18 one cannot get this phenomenon without having chemistry 19 changes.
So that the source term code package now that is 20 being run at BCL now has, let me call it, a tracer such that 21 by hand outside of the suite, one can take account of the 22 chemistry changes that would occur --
23 MR. KERR:
Now, when you do this branching between 24 these two possibilities, the probability is gotten by a
- O 25 feeling of expert opinion?
89 i
MS. MITCHELL:
Yes.
There was a problem run through
(
2 the containment loads working gr.oup that is just very briefly 3
talked about in Appendix C, and although I haven't looked at 4
it myself, presumably more thoroughly.in their report, that 5
talks about the loading that may come and how much of a core i
6 you would need.
7 MR. CATTON:
My recollection is that that particular 8
problem when it was run through that group was highly 9
controversial.
99 percent of them didn't believe it and they 10 did the exercise only because they were told.
I would hope 11 you wouldn't be referring to it in any strong way.
12 MR. KERR:
No.
In fact, they're using numbers from 13 that; that probability number.
14 MR. CATTON:
Well, that's a very small probability 15 number that you associate with it.
I would hope that it is 16 next to zero.
17 There is.another thing, though.
In order to have, 18 the direct heating problem really manifest itself you need to 19' have the high pressure to drive it.
20 MS. MITCHELL:
You bet.
21 MR. CATTON:
And this gets back to the question of 22 recirculation, and some of us feel that you're putting that in 23 the wrong bin.
i 24 MR. KERR.
Ivan, I wish you.had mentioned that
\\
25 direct heating earlier in that high pressure problem.
1 l
l 90 l
1
[ Laughter.]
i f
L 2
Go ahead.
3 MS. MITCHELL:
That's all I wanted to say about 4
direct heating and steam explosions, unless there are any 5
other questions.
6 MR. KERR:
Mike, you had comments -- and I don't 7
know whether you want to elaborate on them -- about core melt 8
interaction predictions.
We are on (b) at this point.
9 MR. CORRADINI:
Well, I asked it previously.
I am 10 most concerned with somebody who is trying to analyze the 11 Argonne experiments because they're the only closed-group 12 experiments, although small_ scale, which looked at the effect
[~
13 of trying to get a high pressure melt ejection and low 14 pressure melt ejection and the effect of water in the cavity 15 and structure.
16 And granted, when we went through with the 17 Containment Loads Working Group there was a great deal of 18 argument on whether they were good experiments or bad 19 experiments.
My impression is they're the only complete 20 published group of experiments that we have.
And it seems to 21 me that the methods, whatever they are in this, ought to at 22 least look at the experiments to try to determine whether or 23 not they can successfully predict what was seen in those 24 tests.
\\-
25 So I said it earlier, but --
91 1
MR. KERR:
I thought ycu also had some questions and
?
2' maybe they ought to be raised in another context, about the 3
temperature predictions of the melt.
4 MR. CORRADINI:
Okay.
My questions there is, I had 5
asked it
- 1. guess at the August 1st meeting and I got a little 6
confused with the answer.
And that was, the core melt 7
interaction with concrete calculations from BMI-2104 were such 8
that -- and then the auxiliary calculations in the QUEST study 9
-- were such that it looks, even with the two fixes that the 10 QUEST study made, which one was the interfacial heat transfer 11
. coefficient and one was the averaging of viscosity, you still 12 got.a temperature excursion when you did the calculations.
13 That is, the temperature went up instead of going down.
And v
14 that should have had a very big effect in the VANESA 15 calculations for the source term release.
16 And that still strikes me as an unphysical 17 result.
And at the August 1st meeting I was left with the 18 impression that yes, they knew it was unphysical but they 19 didn't think it was a big effect.
20 But I'm still left with a problem there because I 21 think any temperature increase -- do you know the plots I'm 22 thinking of?
You got back to Chapter 7 of the QUEST study and 23 then Chapter 5 of the 24 MS. MITCHELL:
Which volume are we talking about?
l 25 TMLB'?
i l
92 1
MR. CORRADINI TMLB' lq y
2 MR. MEYER:
These are CORCON Mod 1 results?
\\_/
3 MR. CORRADINI Yes.
And then in the August 1st 4
meeting I brought all this up and then Dana said -- I remember 5
he specifically said, well, the CORCON Mod 2 calculations are 6
very similar, which struck me as odd.
But all right, given 7
that, then you get a temperature excursion, which then ifI 8
take that temperature profile and stick it into VANESA, that 9
drives the whole process which means I'm going to get a source 10 term which seems to me to be unphysically high.
Does that 11 make sense, what I'm asking now?
12 MR. MEYER:
Yes.
There were several cases where 13 that happened in calculations done at Battelle either on 14 BMI-2104 or on the similar Limerick calculations.
15 MR. CORRADINI:
Yes.
But the traceback -- and again 16 I'm getting into details, but I think it's kind of important 17
-- the traceback was that BM1-2104 used CORCON Mod 1 18 unaltered.
QUEST came through, saw that you get an enormous 19 temperature excursion, made two model alterations; one on 20 viscosity, one on interfacial heat transfer coefficient l
21 between pool layers.
But you still get a temperature 22 excursion.
Not as big but it still goes up instead of down, 23 which is unphysically -- which seems unphysical to me.
24 And in either case, that would then impact the O
/
25 VANESA calculation because that totally is driven by
93 1
temperature.
(O) 2 MS. MITCHELL:
No.
The VANESA calculation i t, not 3
totally driven by temperature.
4 MR. CORRADINI:
That's the answer Dana gave me but I 5
did a little calculation and I sent it off to Professor Kerr 6
where if you look at just a -- you just look at a 7
normalization of what I vary -- and I remember Dana's answer 8
to me was it's velocity, it's kinetics and it's -- it's 9
surface area, he said, and the kinetics associated with 10 surface area in VANESA, temperature and velocity.
11 But I have no feel for the kinetics in VANESA since 12 there's no documentation, and without that I don't know what
/
13 mass transfer coefficients or surface area models he's using.
14 So given that, if you look.at temperature and 15 velocity, I did a little calculation which shows the 16 temperature far and away is the dominant effect in terms of 17 source term release because it determines the vapor pressur,e 18 for the release of the materials.
19 MS. MITCHELL:
If you have a temperature excursion, 20 an unreasonable temperature excursion, I think that's probably' 21 true.
The releases just go up.
If you cross through the knee 22 of the curve, then just a few hundred degrees centigrade can 1
23 make a big difference in the release.
i 24 MR. CORRADINI So my question goes something like
~g
\\ss 25 this:
Is this Battelle suite of codes, which I guess CORCON 1
1 l
94 1
Mod 2 now is part of --
m k,,)
2-MS. MITCHELL:
CORCON Mod 2.
3 MR. CORRADINI Do they still see, under 4
circumstances, what I would call a temperature excursion, but 5
an increase in temperature with time, whereas in all the 6
experiments and in all the analysis of experiments you see a 7
characteristic decrease in temperature with time.
And that 8
difference of just a few hundred degrees could make a 9
significant difference in source term.
10 MS. MITCHELL:
One of the things that bothers me is 11 I'm not so sure that the teuts really had the zire oxidation 12 in that can be a source of heat, in addition to decay heat or
[V~')
13 power.
You are basically inputting --
14 MR. CORRADINI Okay, but let me take you back to 15 one set of calculations I have seen the NRC publish, and that 16 is the CS&I standard problems, which Dr. S'peiss went to Europe 17 and presented with Theofanous and Pratt doing the 18 calculations, and in those he had a large amount of zire 19 oxidation and still CORCON, in those calculations, showed a 20 characteristic very rapid falloff in temperature.
Which is 21 different than what I had seen in the BMI-2104 calculations 22 and the QUEST calculations, which it hung up and in some 23 certain instances with certain combinations of materials not i
24 only hung up but went up further than the initial temperature.
I 25 And that can make a big difference in the source
[.
~
~
95 1
t e' rm release.
p) 2.
MR. KERR:
The problem you can't answer the
(
3 question now, but do you understand the question he's raising?
4 MS. MITCHELL:
Yes.
We will have to give you a l
5 promissory note on it.
6 MR. KERR:
And I'd think you would want to 7
investigate it.
8 MR. CORRADINI' Yes, and I don't expect you to come 9
up with the answer right now.
The one linking thing there 10 that's important -- and I bring you back to the experiments 11 because I think it's very important -- is that when I have the 12 chance of getting these high temperatures, simultaneous with 13 that in the beta experiments I have high gas release.
The 14 high gas release causes an emulsification of the layers, and 15 the layered geometry which is assumed in CORCON is not seen in 16 the experiments.
Do you follow what I'm saying?
It's 17 essentially emulsified.
18 MS. MITCHELL:
Yes.
19 MR. CORRADINI And the fact that I've got a 20 homogeneous mixture now instead of a layered structure should 21 also change the chemistry in the source term release 22 tremendously, it seems to me.
1 23 MR. MEYER:
Well, the layered nature of VANESA is 24 assumed not to be important.
I think you could have assumed 25 it was homogeneous and done the chemistry the same way.
=-
96 1
MR, CORRADINI-Okay. Because now I take an oxide,
)
2 which couldn't somehow affect the oxidation or reduction of
- g 3
fission products of the metals, and now I have them 4
4 homogeneously mixed, and it seems to me.that would affect the 5
chemistry in some way.
6 MR. MEYER:
The way that VANESA is written is 7
essentially the same as if it were homogeneous.
I'm not sure j
8 that the same --
9 MR. CORRADINI:
That's an interesting point that I 4
10 didn't know.
Are you saying it makes no distinction of 11 layers?
J l
12 MR. MEYER:
In VANESA -- it might make a difference
("N 13 in CORCON, but my understanding is it does not make a i
14 difference in VANESA.
15 MR. CATTON:
It might make a difference in the real 16 world.
17 MR. CORRADINI:
Yes.
18 MR. MEYER:
VANESA, whether you pictured the melt as 19 layered or homogeneous, it operates. essentially as if it were 20 homogeneous.
21 MR. CORRADINI.
Okay.
t 22 MR. MEYER:
Now, with regard to CORCON MOD 2,
I 23 believe that the hangup of the temperatures is still being 24 seen in some calculations, and those are being looked into, O
25 but I don't believe the second peak is seen in VANESA.
I i
_. _. - ~,. -,,,
97 1
think that was an artifact of the MOD 1 code, which was
)
2-understood, and that problem is eliminating the MOD 1 code 3
from the suite of codes.
4 By the way, the VANESA manual is in the public 5
document room downstairs and has been there since late July.
6 MR. BURSON:
All 400 pages of the VANESA document 7
are on my desk, and I will be happy to send it to you.
8 MR. CORRADINI Okay, because in August I remember 9
we asked about it.
10 MR. BURSON:
It is available and you can have it.
11 MR. CATTON:
I would be delighted to see that also.
12 MS. MITCHELL:
It's downstairs. At lunch hour you
^
13 can look at it.
14 MR. CORRADINI-I will go.
15 MR. SHEWMON:
Fred says it's not cleared yet, and 16 you say it has been downstEirs since July.
17 MR. MEYER:
A draft was put in the Public Document 18 Room in late July even though they weren't finished reviewing 19 it and getting it out the door as a formal report.
20 MR. CATTON:
I have one more comment on all of 21 this.
The CORCON package that KFK, I guess, has been modified t'emperatures and 22 in order that he can better match or measure 23 so forth, and the penetration rates.
It seems to me that the 24 heat transfer is even higher and that this mess is going to be O
k-25 even colder, which drives you further away from having these
98 1
high releases.
I would think that those modifications ought em
(
)
2 to be made. straight away and that you ought to run a check 3
case to see what it's doing to you.
4 And what it does even further is it makes the VANESA 5
code even less of the tool that you could use.
i 6
MR. BURSON:
I guess I had better respond to that in 7
this way.
If we were sure what that modification ought to be, 8
we would have done it already.
Both the VANESA code and the 9
CORCON code have modifications in them to compensate for the 10 disparity between axial and radial ablation.
The one in the 11 CORCON code, which is sort of an analog to transition boiling 1
12
-- Randy Cole feels that it does the job in terms of
(
13
_e x p l a i.t i n g the beta cases but he is not at all sure about the
\\
14 mechanisms. involved.
It may be getting the right answer for 15 the wrong reason.
So to put that modification in would be 16 premature.
17 Michael Ryman, who is the author of the VANESA code, 18 has another fix in that code.
They both give good agreements, 19 and by good, I mean like by 10 or 15 or 20 percent in the 20 ablation rates or gas production rates, which is very, very 21 good in point of reality.
But it would be premature to put a 22 correction sure if you're not sure it's the right correction.
l 23 So it's a temporary fix.
24 MR. SILBERBERG; I think your point is well taken.
A k _,)
25 MR. CATTON:
I certainly would agree with that, but J
l i
s i
l l
1
i 99 1
on the other hand, the only data available ought to be used.
ry (a) 2 MR. KERR:
Does that complete our discussion of Item 3
B, as far as you are concerned?
4 MS. MITCHELL:
Yes.
l 5
MR. KERR:
Would you now go to Item A, and Mike, I 6
could, I think explain what you had in mind.
I believe it was 7
you who asked a question on Item A, but I thought it would be 8
better to have you explain it.
9 MR. CORRADINI Part of that question, actually, I 10 was asking this morning about trying to understand the l
11 methodology here because the methodology that I asked Denny 12 about before was basically you have six pieces.
You have got i
l (s 13 the basic science somewhere.
You have got the computer tools v
8 G
14 that basically do the systems analysis for a given sequence.
15 You have'got the PRA sequence analysis, and then you have got 16 the containment event tree, which is now in the form of a 17 computer tool rather than just simply a hand drawing of it, 18 and then you have got the consequence analysis, which is CRAC 19 2,
all of which feed into quantification of the source term i
20 and its effect.
l 21 My point was that in 0956, those six things and how 22 they fit together are really independent of the Battelle suite 23 of codes.
You don't necessarily need to use the Battelle 24 codes to do parts of these things.
If one set to it, one 0%
ss 25 could do it all by hand, so the methodology could be set up
100 1
independent of what you use.
i
,-sg
( )
2 For example -- I will just take this as an example 3
-- you may cast this all aside in six months and use MELCOR 4-for 50 percent of all of this, or you may go back and NRR may i
5 decide to heck with all this, we are going to do hand 1
l 6
calculations.
But the methodology should be fixed. That was t
I 7
my point, Does that make sense?
8 MS. MITCHELL:
Is that --
9 MR. CORRADINI.
That's the thrust of it.
It seemed 10 like from 0956 I got more of a taste what the codes were i
11 rather than the methodology.
12 MS. MITCHELL:
0956, as you point out, is only 13 addressing part of it.
It doesn't even address -- except to 14 state that one needs to have sequences defined and their 15 probabilities determined.
16 MR. CORRADINI:
Okay.
But the one thing that 0956 17 does do, at least does advertise, is that it presents a 18 methodology.
19 MS. MITCHELL:
No.
No.
We don't have that word.
20 We are talking about the analytical procedure for source term 21 calculations, so we did not do that.
l 22 MR. WARD:
Does that fix your problem?
l 23 MR. CORRADINI' I guess.
24 MR. SILBERBERG:
We have expunged the word t
25
" methodology" from the test.
3
,,c
101 1
MR. WARD:
What kind of methodology do you mean?
)
2 MR. KERR:
It means analytical procedure.
3
[ Laughter.]
-4 MR. CORRADINI; The only thing I'm concerned about 5
is I don't want to -- well, it seems to me that one chapter.in
~
j 6
0956 wants to spell out, independent of what tool you use, how 7
one would go about it.
I mean if all somebody gave is a l
8 person a pencil and a paper in a room, maybe not even give him 9
a calculator, how you would go about thinking through the 10 process.
11 MS. MITCHELL:
You are asking about how would a 12 person who wanted to come up with the risk of Plant X,
how I
[)
13 would that person start out doing the problem?
N~sl l
14 MR. CORRADINI
- Right, 15 MR. SILBERBERG:
Without these codes.
16 MR. CORRADINI Yes.
You don't need a code to do 17 it.
18 MS. MITCHELL:
1150 should say this is 19 MR. CORRADINI Okay.
So you think this isn't the 20 place to write that down?
21 MS. MITCHELL:
This is only one piece of it.
This 22 is not to say --
23 MR. CORRADINI.
I feel for the people who write 1150 24 because they are really getting the buck.
k.
I 25-MS. MITCHELL:
Well, they are here listening to what j
i i
i 102 i
l' we tell them they are supposed to do.
2 MR. BENDER:
Well, just setting aside that point, 3
the IDCOR people have developed some procedures, and maybe 4
some comparative presentation of how what you have got here 5
compares with that would hrlp shed some light on this.
I 6
don't want you to do it hera, but I think it would help.
7 MS. MITCHELL:
Are you talking about just strictly 8
the source term part of it or are you talking about how they 9
define the dominant sequences, containment failure modes that 10 11 MR. BENDER:
Well, I would do the whole thing if I 12 could, but under the constraints which you have laid out here,
[
13
.I would say take the package which you think this represents
\\
14 and fit it into those set of procedures so we can see how-15 somebody else might apply the information.
16 MS. MITCHELL:
They will certainly look at that as 17 they audit the procedures that are being developed.
Are the 18 IDCOR models or procedures correct?
19 MR. SILBERBERO:
That is qualitatively available in 20 terms of major differences in modeling that have arisen.
By 21 reading the test documents from the IDCOR work and reading the 22 backup to 0956 --
l 23 MR. KERR:
Excuse me.
What is meant by the backup?
I 24 MR. SILBERBERG:
That is BMI 2104, and there has t) 25 been a discussion and dialogue on the different -- on l
l l
103 I
l 1
qualitatively, on major modeling differences, I
2 The next step is there was a MARCH / MAP comparison I can't recall 3
that was done by Battelle, and I am not sure that would indicate how they would compare 4
all the details if I ran this code and this code, how would they 5
6 compare? -- but without actually, as Mike says, stringing out 7
the assumptions in each of them.
8 So the thing that you are referring to is largely 9
not available except in a qualitative sense, but certainly it 10 is a useful thing to do.
11 MR. BENDER:
I don't know that I am a strong 12 proponent of it, but a great deal depends on when 1150 is
/i 13 coming out.
If it's coming out in a few months, I would wait I
(J 14 for it.
If it's going-to be forever to come out, then I would 15 sit down and I would say, well, what are these other people 16 out there doing that I'm going to use as the basis for 17 judgment, just to see where you fit things together.
18 MR. MEYER:
Mike, the whole exchange that we have 19 had with the IDCOR people really has been directed at exactly 20 that.
21 MR. BENDER:
That's why I am encouraging you to do 22 it this way, so that other people besides you guys in your l
23 little narrow sphere are able to see what's going on.
I don't 24 care about me because I don't really care about the whole
,l O N-25 analysis,.but the people in the world as a whole might find j
-- -~ ~. -.... - -.. -.. _. - - -
104 1.
the presentation clarified.
l l
(
2 MS. MITCHELL:
There are minutes put out of every l
3 meeting that has been held with IDCOR, not only the summary of 4
someone designated to do the summary, but also all of the 5
Vu-graphs that were used in'the presentation are available in i
i 6
the Public Document Room, and they were all open meetings.
l l
7 MR. SILBERBERO:
And the dramatic differences in the i
8 analyses are highlighted in those discussions, in particular.
i i
9 MR. BENDER:
I know it's all out there i
10 somewhere. Nobody has the time or the knowledge to go find it 11-item by item.
Just put it together in some way.
l l
12 MR. KERR:
Did you ever try to reconstruct a meeting 13 from the slides and the minutes?
I have, and I can't.
I mean
(
14 they are better than nothing, but.
And what Mike is 15 suggesting would be a r.ther large undertaking. It would be 16 helpful to have such a document, but it wouldn't be easy to 17 do.
18 MS. MITCHELL:
Yes. Well, right now I believe that 19 the sequences are defined very differently between IDCOR and 20 the BMI 2104.
In 0956 there are only three out of the vast l
21 number of sequences that were looked at that we chose to make 22 some limited comparison in Chapter 5.
It was because I felt 1
23 that the others were so different, they are not the same 1
l 24 sequences, so you are trying to compare not analytical
(~
~ procedures, not what would MAP give you versus what would
\\_
25 l
i
_ ~ _ _.. _... _.,,
105 1
MARCH give you, but the definition of how the sequence goes.
/
j (N) 2 MR. SILBERBERG:
But I think his point is that even j
3 if just what is in the assumptions and models and science or 4
engineering is in the code, just by themselves it would be 5
useful How they manifest themselves and the result, you are 6
-right, the plants that exist right now that they have studied 7
don't quite match up with sequences.
8 MS. MITCHELL:.The sequences d o n '. t match up.
9 MR. BENDER:
The logic of the computation, that's 10 really what people want.
11 MR. KERR:
The value of such an exercise, it seems 12 to me, would force somebody to have an overall understanding
[~'i 13 of the process.
I get the impression even from 0956 that N.Y 14 there is
- a. lot of information out there but that it is 15 fragmented. I mean look at the chapters.
And there are people 16 who undertstand this part and some others who understand this 17 part and some others who understand that part, but there 18 probably are not very many people who understand the whole 19 process in detail And somebody needs to-20 MS. MITCHELL:
I agree with you, there really aren't 21 experts in everything.
We did point out that it crosses so 22 many disciplines.
23 MR. KERR:
I'm not asking for experts in 24 everything.
I'm asking for a few technically qualified people n'
25 who have a general understanding of the total process.
They
106 1
don't'have to be experts.
And that, I think, is important 2
because this eventually, if it is to be used,_has to be used 3
as a total process, not in little pieces.
4 MS. MITCHELL:
Yes.
All of the people who represent 5
-- I think they need to be experts in it, and if it takes four 6
people walking in locked step to every meeting, the people 7
from the front end who will define how does the sequence 8
begin, and the people who are doing the containment event tree 9
who need to listen to everything and take advantage of when is 10 the hydrogen coming out, how much hydrogen do you get at'the 11 time of vessel failure to influence the containment matrix, 12 they all need to show up together to make sure that the
('s 13 problem that is being solved is indeed the one that will lead-14 you to the right best answer for the reason that you are doing 15 the calculation.
16 And to the extent that nobody is an expert in 17 everything, there needs to be more than one person.
18 MR. KERR:
Okay.
Are we now at (c)?
19 MS. MITCHELL:
The containment performance 20 predictions.
21 In the material that came is the recognition that 22 the behavior of the containment in response to the challenges 23 is a very important factor, we say, for most sequences.
For a 24 lot of sequences, it is the largest single factor influencing 25 the source term, and that certainly is true.
107 1
Right now there is good news and bad news with
/O
)
2 respect to the containment performance.
There are
's /
3 improvements compared with what was done for the reactor 4
safety study in making more inclusive event trees and in 5
bringing to bear more information than was available at the 6
time of the RSS.
7 In your package there are two pages, one with the 8
good news on it and one with the bad news on it.
The good 9
news basically is the fact that we had two working groups --
10 containment loads working group and the containment 11 performance working group that tried to look at leak before 12 break-behavior, ultimate pressure capability of the
}
13 containment'.
s/
14 There are also the QUEST and IDCOR studies, things 15 like the SASA program, and the large list of sources of 16 information that were brought to bear on the containment 17 behavior problem as listed in Appendix B of 0956, and I think 18 there are probably 25 sources of information.
19 The different kinds of. questions asked fall into 20 basically four categories: phenomenological events, how much 21 hydrogen is generated by the time of vessel failure, how much 22 comes out during the core-concrete interaction, does the RCS 23 fail at high pressure or low pressure, and if it fails at high 24 pressure, is it going to fail through the bottom at high 25 pressure or the steam generator tubes or the hot leg piping or
l l
i 108 1
some of the other locations?
(,/
2 Is the equipment going to survive in this 3
environment that it is being asked to be worked in, and what l
4 are the failure modes?
Will it leak before break?
Will it 5
open catastrophic -- sir?
6 MR. BENDER:
Well, I better let you finish.
7 MR. KERR:
That's a good idea.
Why don't you.
That 8
was just an expression of extreme interest that you saw.
9 MS. MITCHELL:
The bad news about containment 10 performance predictions right now is that they.are 11 subjective.
They are appeals.to expert opinion to take 12 information from all of the sources and try to distill out of
[\\ }
13 it some nuggets of wisdom, and they require extensions of all J
14 of this information.
There are only six problems done in the 15 containment loads working group, and that certainly doesn't 16 answer all the pressure loading questions for all sequences 17 in all plants, and there is engineering judgment on what va.lue 18 shall I have for the split fraction of the alpha failure mode, 19 to pick one that we discussed already.this morning.
20 What is done for NUREG-0956 and what will be done 21 for SARP and.1150 is basically to look at three views of the l
22 containment behavior, one that is optimistic and takes a 23 subset of the entire body of information that will lead to the 24 conclusion that the containment will behave in a manner so as
~s 25 to produce reduced source terms, and a central estimate and a
'~
6 109 1
more pessimistic estimate which will lead to larger source 2
terms.
3 MR. KERR:
Let me see if I understand those 4
estimates.
Is each individual asked to make a central or a 5
pessimistic and an optimistic?
6 MS. MITCHELL:
No.
7 MR. KERR:
Or d;es one sort them out once you get 8
them and then you decide which is central and which is 9
pessimistic and which is optimistic?
10 MS. MITCHELL:
Yes.
As I understand it, they took i
11 the body of information represented by the list of references 12 and tried to sort out, the people who felt'there would be a 13 large amount of hydrogen generated in vessel such that you 14 would get a high pressure in the containment early on, and 15 anything that led to early containment failure was l o o k e 'd at 16 as being a pessimistic assumption, and anything which allowed 17 containment to survive forever was looked at as an optimistic 18 estimate.
19 The central estimate, they point out, is something 20 that is much more difficult to come by; that the experts tend 21 to agree that it certainly couldn't be any worse than this or 22 is not likely to be any better than that 23 MR. SHEWMON:
When I look at the back of 0956 here, 24 there is talk about the loads which would lead to general 25 membrane strain for at least the metal containments.
It seems
110 1
to me experts agree on the calculation reasonably well Is
)
2 that such a limited fraction of situations that it doesn't 3
enter much?
Are there usually other boundaries that fail in l
4 the concrete that you are talking about, or wl: e r e is this wide 5
range that you are referring to?
6 MS. MITCHELL:
I think it is in Appendix B.
7 MR. SHEWMON:
What kind of containments?
8 MS. MITCHELL:
The one that is in here is Surry.
9 MR. SHEWMON:
So the only containment you are 10 talking about is Surry?
11 MR. KERR:
Paul, she is talking about phenomena that 12 might lead to failure as well For example, hydrogen
(~N 13 production. How much do you get that might tell you whether N.
it isn't just the failure mechanism; 14 you get early or late 15 it's also the loads, I think.
16 MS. MITCHELL:
Loads, right.
If you have a 17 containment like' Sequoyah that fails at -- I don't remember 18 the number, but like 60 psi, how likely is it if you get a 19 large amount of hydrogen generated in vessel that when it 20 burns, it will fail the containment, so that the pessimist --
21 MR. SHEWMON:
Fine.
I understand.
Thank you.
22 MR. BENDER:
I could shed just a slight amount of 23 light on the question that you are asking.
That is, it's only 24 the low pressure metal containments where this is likely to
\\
25 be an issue.
i l
i l
I m.
111 l
1 MS. MITCHELL:
Early containment failure?
)
2 MR. BENDER:
The metal strain question that Paul is 3
asking.
4 MS. MITCHELL:
Okay.
5 There is a limited sensitivity study in NUREG-0956, 6
and this is for.Surry, to the prob' abilities that you 'would j
7 associate'with answer the specific questions where you had to I
l 8
have recourse to extensions of information or engineering 9
judgment that would say this phenomenon is unlikely or is 10 remotely possible.
To get a quantification of the tree and to 11 produce a containment matrix, you have to assign a probability 12 to that. Is it remotely possible,
.01?
Is it 1?
Is it l
(~'
13
.001?
And there is a limited sensitivity study.
Y,/
14 They found for Surry, as reported in here, that 15 there was less sensitivity to that probability-assignment than 16 to the changes in phenomena that would represent the 17 optimistic or pessimistic walk-throughs, as they call them.
18 So that the people who felt there would be a large amount of 19 hydrogen generated as more important than whether or not 20 something is remotely possible or unlikely.
I 21 MR. KERR:
Since you referred to Appendix B and 22 Surry, I note on page B-15 the statement under Surry that 23 in-vessel steam explosions could also be important in the 24 pessimistic walk-through if test data and analyses were to x_/
25 show that their influence is greater than that attributed.
i l
\\
i
112 1
What is that sentence supposed to tell me?
A 8
2 MS. MITCHELL:
In the pessimistic walk-through for (d
3 Surry, in-vessel steam explosion that leads to early 4
containment failure, the alpha failure mode, is listed at its 5
RSS value of
.01.
There are some --
6 MR. KERR:
No, I'm trying to find out what I am 7
supposed to gather from this sentence.
What is it supposed to 8
tell me?
9 MS. MITCHELL:
If the probability of in-vessel steam 10 explosion leading to early containment iallJre is higher than 11
.01, it would be an important contributt.r right now in the 12 pessimistic walk-through.
13 MR. KERR:
But I don't understand what the message N_/
14 is.
Is it in there to say that there is some likelihood or 15 reasonable likelihood that it may be bigger and hence we ought 16 to be worrying about it, or is it there just as -- I mean I am 17 trying to understand what message I am supposed to get from 18 this sentence that has to do with reactor safety.
19 MS. MITCHELL:
Experts don't agree on the 20 probability of steam explosions.
21 MR. KERR:
Well, I thought I had heard earlier-that 1
l 22 your experts, at least, had given it such a low probability 23 that the current feeling was that it's negligible.
24 MS. MITCHELL:
It's carried that way in the central
,A,
\\~s/
25 estimate.
That is exactly why it is zero in the central
)
113 i
estimate.
But to the extent that there is a residual f'%
( )
2 uncertainty --
3 MR. KERR:
There is always residual uncertainty in 4
any physical phenomenon that is not trivially simple.
I'm 5
trying to get at what it is the Staff thinks about this issue, 6
and that sentence, I don't know what it's telling me.
7 MS. MITCHELL:
Well, I tried to tell you what I 8
thought I was putting in there when I put it in.
Expert's 9
disagree, and there still is an uncertainty on this issue of 10 probability.
11 MR. KERR:
Well, experts are always going to 12 disagree on this issue.
Does that mean that the Staff will
/~}
13 never take a position?
14 MS. MITCHELL:
That I can't answer.
15 MR. CATTON:
It seems to me that the steam explosion 16 review group concluded something.
The experts,.although they 17 didn't agree, at least came to a conclusion that the steam 18 explosion probability was very low.
I don't know how you can 19 put that sentence in the report in the face of that report.
20 MS. MITCHELL:
I will take under advisement your 21 comment.
22 MR. BENDER:
May I ask my question now?
23 MR. KERR:
Is it relevant and concise?
24 MR. BENDER:
Whether it's relevant and concise is a D
25 matter of opinion.
's
l l~
114 1
MR. KERR:
Okay.
Go ahead.
.(O) 2 MR. BENDER:
I am not able to come to a judgment l
3 from what I can see in this report about how one deals with 1
4 the modes of containment failure for the purpose of this 5
analysis.
There are several modes.
One is the abrupt 6
bursting.
I don't at the moment know what things or what 7
acci, dents go with a abrupt bursting.
Is it only ATWS?
Is it 8
ATWS for certain things and some other accident for others? Do 9
I have to envision this by knowing all of the sequences that 10 have to be considered?
11 The second part of the question is does this 12 methodology or computational procedure allow some kind of
}
13 treatment of leaking containments that will limit the capping v
14 of the pressure or cap the pressure, and if so, how is that 15 covered in this discussion?
16 MS. MITCHELL:
You input the containment failure l
17 criterion, whether it is pressure or. time.
18 MR. SHEWMON:
But his question is: If it fails, then 19 what flexibility is there about whether it breaks into four 20 pieces and is spread four miles over the countryside or it's 21 a concrete one and it just sort of develops a lot of cracks 22 that pick up a lot of aerosols on the way out anyway.
j 23 MR. BENDER:
Is that the analyst's judgment or 24 somebody else's judgment?
)
N/
25 MS. MITCHELL:
Those kinds of issues are input l
~,, - -
115 1
quantities in MARCH.
You get them from evaluating the list of
(
2 information from CLWG or containment performance to decide 3
whether or not in this particular sequence, which is -- let me 4
pick one -- is a long, slow overpressure --
5 MR. BENDER:
I take it you are telling me that the 6
analyst has to decide.
But when you get ready to write 1150, 7
or somebody does, can any analyst decide or is there something 8
out there that will be the judgment that is going to be used 9
to categorize each type of containment?
10 MS, MITCHELL:
There is a containment event tree for 11 each type of containment that asks a very large number of 12 questions relative to the accident sequence and parameters 13 that describe the containment itself and its pressure or 14 something: is it likely to fail, what is its ultimate pressure 15 capability?
16 The mode of failure is, as I understand what is done 17 now in the containment event trees, it's to say if you have s 18 rapid pressurization, that you will not leak before break.
So 19 that if you rise quickly to the ultimate failure pressure, it 20 will be a catastrophic failure.
However, there is a split 21 fraction for slow overpressurizations to say for some, in this 22 case small, fraction, that it will leak before break.
23 This kind of thing is based on the containment test 24 that was supposed to leak before break and didn't. There will
\\s 25 be a concrete containment test in'the future, which I think is 1
116 1
a very important test, which will answer the question of
}
2 whether or not a concrete containment always fails with a very g
3 slow, small leak.
Whatever leak you want to put in there can 4
be put in.
5 There is a parametric study in here, a couple of 6
them, of isolation failures and high, medium and low ' leakage 7
rates, which are a function of pressure, so that the leakage 8
area is a function of time as the accident progresses.
9 MR. BENDER:
I read what you are saying as the 10 analysts will decide.
11 MS. MITCHELL:
You bet.
12 MR. BENDER:
And somebody will give the analysts 13 some criteria, but exactly what they are yet, we don't know.
tN_
j 14 MR. WARD:
Well, there is one criterion built in, 1
15 and that is the pressurization rate, you said.
16 MS. MITCHELL:
It is not built into the MARCH 17 calculation.
It is built into whether or not there is a split 18 fraction given for early overpressurisation, rapid.
{
19 overpressurization for leak before break behavior.
In what is 20 done so far, there is no leak before break split fraction for 21 rapid overpressurization, but if you want to determine the 22 risk of Plant Z,
then you have to go through and say these are 23 the kinds of accident sequences that have to be considered, 24 and these are the modes of containment failures that can come
(\\s 25 with those accidents. Is it going to be late?
Is it going to
117 1
be early?
)
2 And for cases where you feel that it is important to
'%/
3 describe that kind of a situation, then you tell the source 4
term person to go out and execute this set of codes for a leak 5
before break behavior for a TMLB' delta late overpressure 6
failure.
And then the analyst will go out and run a 7
calculation that will provide you with a source term that 8
describes the bin of late overpressurization leak before break 9
behavior for TMLB', and the containment event tree person will 10 tell you what is the probability of that particular bin, 11 adding up all the sequences that go into that bin and their 12 respective split fractions from their respective evaluation of 13 the tree.
(w 14 Did I muddy up the waters?
15 MR. BENDER:
You just gave us a lot in a short time.
16 MS. MITCHELL:
I did want to point out that what was 17 done in NUREG-0956 had very limited review in the containment 18 performance area and was referred to as a preliminary 19 evaluation. These trees that are referred to here and some 20 detail given on the Surry one are being peer reviewed for 21 1150, and here is another promissory note for 1150.
22 MR. KERR:
What is the relationship between the 23 activities and conclusions of the containment performance 24 working group and the containment loads working group, if they b\\- J 25 have anything to do with each other, and the r.cipe suggested 9
i
{
118 1
in here, or is there any recipe?
Or are they the same thing?
2 MS. MITCHELL:
The recipe that is in here for 3
determining the cnotainment matrix is to go take all the 4
information, construct yourself a tree by asking the i
i
~
'S appropriate questions, and pick quantification for the 6
branches on the tree.
That is 't h e recipe.
The information 7
source that would allow you to provide a central pessimistic 8
and optimistic walk-thruagh are garnered from such diverse i
9
-sources as the CLWG and the QUEST study.
i 10 MR. KERR:
I don't know whether the answer is yes or 11 no.
12 MS. MITCHELL:
They are completely different i
(~N 13 animals.
One is the source of i n f o r m a t i o'n, and the other is a l
14 recipe for using that source of information and others, i
15 MR. KERR:
But if you have to use a central, a I
16
. pessimistic, an optimistic, then what do you do when you get 17 throught 18 Do you give this to somebody and say, "Do you feel 19 like a pessimist today or an optimist or a centralist?"
20
[ Laughter.)
21 MS. MITCHELL:
In 1150, we have told you that they 22 intend to take account of uncertainties, and the uncertainties 23 in the phenomena, the uncertainties that arise from whether o
e l
24 you define "unlikely" as a conditional probability of
.01 or 25 some other number, they contribute to the uncertainties, and W
t
.. _ _,. _. -. _... _,. ~
119 l
1 there is a group that is working on that and bringing that b)
(
2 into the results in 1150.
3 It is only one source of the uncertainty.
4 MR. KERR:
Mr. Corradini?
5 MR. CORRADINI.
I'll let him go first.
6 MR. CATTON:
What group is doing this?
7 MS. MITCHELL:
The uncertainties?
8 MR. CATTON:
Well, with respect to the containment.
9 MS. MITCHELL:
Sandia.
10 MR. CATTON:
Well, my recollection of that working 11 group is that the people from Sandia were the ones that were 12 out on the pessimistic limb, as contrasted with the remainder 13 of'the group.
14 How are you going to avoid bringing that factor back 15 in and being more heavily weighted?
I would think you would 16 give it at least to a third party to bring that together.
17 MS. MITCHELL:
We have.
18 MR. CATTON:
Who is the third party?
19 MR, CORRADINI:
That's where I come in.
That's what I
20 I wanted to bring up.
There is a group of five of us who are 21 reviewing the SARRP event trees -- myself, Richard Denning, 22 Trevor Pratt, Steve Hodge, and somebody else who I've 23 forgotten.
24 MR. CATTON:
So you are the one who is going to give O
25 us that information?
120 1
MR. CORRADINI WeII, I wanted to explain, because 2
if I were Professor Kerr, I don't think I would understand
(
3 what you had said completely.
1 4
As I understand it now -- 'since I'm supposed to be 5
reviewing it, I think I should understand it -- the Sandia 6
people looked at the Containment Loads Working Group and the 7
Containment Performance Group reports and attended the 8
meetings, and from listening and reading, subjectively picked 9
probabilities to put into the containment event trees.
10 MS. MITCHELL:
There are other sources of 11 information besides just CLWG and CPWG, okay.
12 MR. CORRADINI.
Right.
But I guess what he'had 13 asked before was, how did the two relate?
And how they relate 14 is, they essentially read and listened and then picked numbers 15 for the branch point probabilities in the event trees.
That 16 was your original question.
i 17 MS. MITCHELL:
Well, did I not say that in five 18 hundred words, instead of fifty words?
19-MR. CORRADINI-Yes.
That's what I was trying to 20 get into a few words.
21 MS. MITCHELL:
Yes.
22 MR, SHEWMON:
The Sandia people are the ones that 23 are trying to keep alive the lid going through the roof?
24 MR. CORRADINI You are asking me, and I'll tell you O
-- I'll give you some intermediate background.
25 4
.ne,-.,
-..n-n,--,..
.--,...-_,--w,-n n.,e.,.
.,,--.-,.,.---e,,
e.-.
.n
121 1
Of the five of us that are reviewing this, the
)
2 opinions range as to how to characterize the probabilities 3
assigned.
On one hand -- I won't name names; that's not fair 4
-- but on one hand, one extreme is that the probabilities 5
assigned for the five event trees are reasonable 6
representations of some optimist, of some pessimist, and of 7
some central guy, okay?
8 On the other extreme i n our group, some people feel 9
that because of the nature of the method of picking the that is, requiring that somebody has got to 10 probabilities 11 be an optimist, somebody has got to be a pessimist, and 12 somebody has got to be a central guy -- that you get a 13 biasing, and you don't integrate the true probability in s /
14 there.
You basically lump them in three groups, and by this 15 lumping process, you bias it.
Some people feel that biasing 16 is inherently pessimistic, that if you look at it, the 17 pessimistic people are very pessimistic, and the central 18 people are kind of pessimistic, and the optimists are just 19 there'along for the ride, okay.
20 So there i s argument among all of us as to how to 21 look at it.
Does that give you sort of a characterization of 22 it?
23 MR, SHEWMON:
No.
The answer to my question may 24 well be yes, t
25 MR. CORRADINI That it is biased, yes, If I were
--w
--9 v.
r
-_,-,_m.,--_--,---r,.--.,y,
---,,,y
,-.y-y-,,..,,,,5
'*---'*"~v*v'W-
3,,.,.-,wecy,,-T-'
- ""T"""M7W**""'""W'P'W"'-V*F
--8'*"WW y.
F
122 1
to put it another way, I think a fair thing to say is, if you
?
(,/
2 sat on the committee, you may judge that having those people 3
do it biased it to the pessimistic side, guessing what you 4
might think.
5 MR. KERR:
Has anybody thought about -- since this 6
is not a bad procedure -- simply taking the whole process of 7
saying what is the risk and getting a group of pessimists, 8
optimists, and centralists together and saying, you know, what 9
is the probability of ten people being killed in an accident?
10 Then you wouldn't have to go through all these other 11 calculations, because we agree that containment performance, 12 particularly if it's good, is perhaps the most important item n
/
T 13 in the whole process of predicting source term.
And since we D
14 are now us'ing expert cpinion to determine that, it might make 15 sense to just use expert opinion to do the whole process.
16 MR. CATTON:
It would certainly be cheaper 17 MR. MEYER:
I think hindsight is real good, but i
18 remember that we got into this source term business believing i
19 that if we just did the physics and chemistry of release and 20 transport and deposition right, that we would find magnificent 21 reductions in the releases from the plants, and that is where 22 we have been focusing our work.
And we have found that other 23 things are very important.
24 MR. KERR:
Well, it is not new, however, to have fs 25 people tell you that the containment ought to be looked at in
123 1
detail, I think.
[
)
2 MS. MITCHELL:
I think we have touched on everything
's.d
\\
3 in one way or another that I wanted to say about this, unless 4
there are other questions.
5 MR. KERR:
Okay.
Verification of modeling, the 6
latest data.
I don't know what that refers to exactly, except 7
there were some comments that the models in many cases were 8
not based on very much comparison with data.
And you can say, 9
"Probably, yes, we think they are," or 10 MR. MEYER:
Well, what I was going to say is, I
11 thought that was your question, because I find in your letter 12 the statement that seemed similar to that agenda item.
g 13 MR. KERR:
No.
I wrote something that I thought to
(Y N
14 be a representation of the discussion.
15 MR. MEYER:
The statement is that "there should be 16 more comparison of the code calculations with existing j
17 experimental data," and it goes on to say, for example, "It 18 would be helpful to make a comparison with what is known about 19 the core damage."
20 MR. KERR:
Okay.
I think it was a good question.
I 21 would have liked to have asked it.
22 MR. MEYER:
We agree with that statement.
23 MR. KERR:
Good.
24 MR. MEYER:
And, in fact, we do plan exactly such O
N>
25 comparisons in the near future and have, in a qualitative
. -. _ _ _, _.. _...~ ___ _ ___
124
)
l 1
sense, made such comparisons already.
2 I think the thrust of this item is on making 3
comparisons, validating models against the latest data, recent 4
data, which, in fact, were not available at the time the 5
models were generated, because the models that we are talking 6
about were fully developed in the 1983 '84 time period, and so c
7 they are, at this point, a little bit out of date, and there 8
are more recent data to compare ~them with.
9 We have a program at Brookhaven that is going on 10 right now.
It is just starting in this activity, but it is 11 funded and in existence, to take the exact reference version 12 of the source term code package, which is being used in the
/ '\\
13 NUREG-1150 risk rebaselining study, and to compare predictions 14 with that reference code with whatever information'we can get 15 from TMI-2 from some of the recent PBF experiments.
The LOFT 16 tests probably would have some information of relevance to 17 these codes and the core concrete tests that we talked abou,t 18 earlier at both Sandia and Karlsruhe and others.
19 In addition, the source term code package reference
~
20 code will be compared piecemeal, to the extent practical, with 21 the more mechanistic codes that we have discussed on other 22 occasions, although I don't think that is your interest at the 23 moment.
24 We have, to the extent possible, compared in an 25 approximate way the codes in the Battelle suite with data that
~*
125 1
are fairly recent and with the TMI-2 observations, although
/9
( /
2 those were less direct.
In fact, there has been a lot of 3
discussion about making predictions for the TMI-2, and frankly 4
our early conclusion on that subject was that it wouldn't be 5
of much value to validating the code, because the TMI-2 6
accident didn't go very far down the line of the code suite 7
and offer the possibility of comparing to codes like CORCON 8
and VANESA or some of the others.
9 MR. CORRADINI:
Could I ask a question about that, 10 because that kind of intrigues me in the sense that I was 11 under the impression --
12 MR. MEYER:
We have changed our mind, and we are 13 going to do it.
14 MR. CORRADINI:
Oh, okay; you changed your mind.
15 MR. MEYER:
We do see the value in that.
We did, in 16 fact, do a comparison with the MARCH-1 code.
There have been 17 some comparisons between MARCH-1 and MARCH-2, so that we can 18 make an inference right now of what MARCH-2 would give in 19
- terms, for example, of the metal / water reaction at TMI-2, and 20 that estimate is 35 to 40 percent, which I think is a 21 reasonable estimate, and we will do that calculation with the 22 version of MARCH that is in the code package, which in terms 23 of the core modeling is basically MARCH-2.
1 24 MR. CORRADINI Okay.
The reason I asked it, O
25 though, is it's more than just the meltdown, because from i
1 4
126' 1
presentations I have heard from EG&G people, you had some sort
,/
2 of pouring event of the order of 20 tons into the lower 3
plenum, so you can look at things like whatever you want to 4
call -- debris bed models or the whole idea of what is 5
occurring and at'least the envelope of what occurred at TMI, 6
because evidently, at least from what I understand from 7
presentations, you had a large amount of mass which was very 8
hot and stayed hot for very long sitting on the lower head and 9
nothing failed.
10-So a lot of those sorts of things would be of 11 interest to compare to.
12 MR. MEYER:
Yes, I agree.
Also, we know that those 13 are relatively recent observations, and so it really hasn't k
14 been possible before now to do a very complete calculation.
15 MR. CORRADINI Well, the pressure traces had always 16 been there, but I guess they're getting smarter in trying to 17 analyse them.
18 MR. SILBERBERO:
But we feel that the latter part of 19 that scenario, the remelt and stuff coming down, bringing 20 large quantities to the bottom of the vessel, is better l
I 21 addressed in, let's say, MELPROG.
So in other words, we are by another method, by 22 going to attack it from another 23 looking at it specifically.
)
1 24 Now to the extent that one can get information from 25 that and apply it back and put it in perspective with MARCH,
127 1
certainly that would be the kind of thing that he described.
(Av) 2 MR. CORRADINI The only reason I'm asking that is 3
.because from the presentation.that I heard from EG&G, I heard 4
also a -- not simultaneous, but a series of presentations by 5
the IDCOR people, who are analyzing TMI, but they are 6
essentially using' MAP.
And it seems to me that at that level, 7
you would like to use the codes that you're doing source term, 8
like the Battelle suite and MELCOR.
9 MR. SILBERBERG:
Right.
10 MR. MEYER:
Yes, we are going to do that, and it is 11 scheduled fairly soon.
-12 We had a rather lengthy discussion on CORCON
(~
13 vis-a-vis the KFK experiments and the Sandia experiments, and 14 I guess I would just say that we have watched those 15 experiments and will do careful comparisons with them as soon 16 as we get set up at Brookhaven to carry this work out, and 17 basically thought that the experiments were in qualitative 18 agreement with the codes that we were using and didn't see a 19 basis for abandoning our code suites.
20 MR. KERR:
What does " qualitative agreement" mean?
21 MR. MEYER:
What that means to me is that I relied 22 on people like here like Randy Cole, who are actually running 23 the codes that we are utilizing, and rurning them for the 24 experiments that were being performed or in possession of the 25 results from those experiments, and telling me that they
128 1
didn't see any obvious discrepancy.
)
2 MR, KERR:
Now I understand what you mean by 3
" qualitative agreement."
4 MR. MEYER:
There is one case where we have been 5
making comparisons in an approximate sense with the data, and 6
I keep waffling on that, because we haven't yet been able to 7
make good, crisp, quantitative checks that we will do in the 8
next few months.
9 But in the case of the CORSOR code, which predicts 10 the release rate from the fuel in the vessel, that correlation 11 was based on a series of tests done at Oak Ridge out.of pile 12 using high-burn-up fuel rods from commercial reactors.
And O
13 after that model was developed and while we were using it,
.s-14 several of the PBF tests were completed, and results were 15 compared with the CORSOR model 16 There were some discrepancies, but not enough to 17 make us throw out the CORSOR model in its entire, except for 18 one piece of it which we have changed as a result of 19 comparisons with data that came out since the models were 20 developed, and that is in the area of the silver / cadmium 21 aerosol algorithm in the CORSOR model 22 The tests that were done since that model was i
23 developed show that our original algorithm was not realistic, it will 24 and we did change it recently, and it will be 25 produce much less silver and cadmium in the source term code t--
l 129 1
package that is being used in the risk rebaselining study, b
( )
2 Do you have any further questions on the subject?
3 MR. KERR:
Well, it really wasn't so much a question 4
as it was a suggestion and a query as to what you were doing, 5
and you have, I think, told us what you are doing, and I would 6
encourage -- and I don't think I altogether have to encourage 7
you -- to do whatever you can about trying to compare your 1
8 codes with whatever data exists.
9 MR. MEYER:
Are we ready to go on to the next item?
10 MR. KERR:
Any further questions on Item (d)?
None?
11 Go ahead.
l 12 MR. MEYER:
The next item has to do with the life l
(N 13 expectancy for the suite of codes, and this question comes up i
v)
\\
14 in the context of a code called MELCOR.
MELCOR is a code that i
15 is similar in scope to the source term code package and has 16 been under development as a PRA risk assessment code as a 17 replacement for the earlier MARCH-based risk assessment 18
- methods, i
19 MR. MARINO:
Now what runs very fast is the fission 20 product deposition.
21 MR. KERR:
Other questions on this topic?
22
[No response.3 23 MR. KERR:
I think we have covered that.
24 Thank you, Mr. Meyer.
O N'#
25 MR. MEYER:
The next and last item on the agenda is
130 1
major uncertainties / program for resolution.
2 When we were here in May and I presented what was 3
and is Chapter 3 of the NUREG-0956, at that time we listed and 4
discussed major areas of uncertainty.
There were eight of 5
them.
There is a sheet in your handout that lists those 6
again.
They haven't changed.
7 I am sure you are willing to stipulate that those 8
are indeed major areas of uncertainty, and I probably don't 9
have'to discuss that part of the topic any further.
I think 10 what you're interested in are our programs to resolve these 11 areas of uncertainty, and that we did not present to you 12 because it was in the research chapter that was deferred from 13 that presentation to now.
14 I realize when we were looking at this agenda, that i
15 I was going to be lucky en this one to be last, down near the 16 end of the day.
I knew you had a fairly hard deadline for 17 reconvening;to do some business this afternoon, and so frankly 18 I have not prepared and do not intend to go through what is i
19 written in Chapter 7 of the NUREG, but would rather you ask 1
20 que,stions or indicate w h a't areas of this you are interested in 21 more information on, and between those of us who are here 1
22 today and I think we have most of the areas covered fairly 23 well -- we can-probably give you the information that you
/
24 want.
25 MR. KERR:
I have some questions, but I will defer
~-
6 131 1
to others who may have some.
Are there other questions?
Mr. Bender?
3 MR. BENDER:
I want to ask a more general one, 4
because there are a lot of things in here.
I think it would i
5 be very useful to order them in terms of their importance.
6 Which are the first three which need to be addressed, and what 7
about the next three, in some form like that, so you could get 8
some judgment about which ones are dominating the question'of 9
the usefulness of the whole program.
10 MR. MEYER:
Well, the order of the list that we have 11 in here is sort of the order in which they occur as we track 12 the fission products through the system, rather than the order 13 of importance.
But I suspect that No.
1 is No.
1 Natural 14 circulation in the vessel with the degraded core is probably 15 the'most important of these subjects.
16 Ivan, I am sure, would agree.
17 MR. KERR:
It's hard for me to see how anything is 18
'more important than containment behavior.
19 MR. BENDER:
Well, it seems to me that we ought to 20 first here what they have to say.
21 MR. KERR:
We*I, okay.
22 MR. MEYER:
Well, I was going to say, that runs very 23 close to the containment performance, and it's htrd to tell t
24 which is more important.
I l
25 MR. BENDER:
Well, if you take those two and you put
- = - -.. - ----,.
132 1
those out, and you say that they are the two most important
((%)
2 ones, if I get those sorted out, can I ignore the rest o f-3 them?
If I get those two sorted out, are the rest of them so 4
much lesser in importance that I don't have to worry all too 5
much about them?
6 MS. MITCHELL:
That really makes three.
There are 7
two issues that'are associated with the containment, not only 8
the loads but the mode of failure, so that buys you three:
9 natural circulation and loads and response to the loads.
10 MR. DAVIS:
I have a related comment.
Isn't it true 11 that the significance of these things depends on what sequence 12 you are talking about and what plant?
['N 13 MR. MEYER:
Yes.
14 MR. DAVIS:
It may be that some of them are not 15 significant for some sequences but become dominant for others.
16 MS. MITCHELL:
You bet.
The containment ~ behavior is 17 unimportant for the V sequence.
'.R.
DAVIS:
And that's a little bit disturbing, 18 M
19 because that means they all have very high priority, depending 20 on which plant.
21 MR. SILBERBERG:
In fact, that is a good point,
.22 because that is one conclusion that we would certainly like 23 you to focus on.
24 The second one is that -- and again, depending upon
\\
25 the sequence in the plant, they do not stand in isolation.
133 1
There is a coupling.
Once I deal with if I am talking n()
2 about pressure loads or I am talking about core concrete 3
interactions, what happens previously in vessel and what is 4
delivered to the cavity is significant as an input condition.
5 So they do not stand in isolation.
6 MR. KERR:
Well, let's take the V sequence, since 7
somebody brought it up.
What is the current. qualitative or 8
whatever feeling about the importance of the V sequence as a 9
risk contributor in terms of what has been done and what can i
10 be done?
11 MS. MITCHELL:
You are asking for its probability?
12 MR. KERR:
You can tell me in any terms you want to
[
13 tell me.
What I want to know is, how much -- how it
\\
14 contributes to risk.
Is it a real important contributor, not 15 very important?
16 MS. MITCHElL:
It is very plant-dependent.
I think 17 everybody has pointed out that that is one thing that is 18 extraordinarily difficult to generalize, that it really 19 matters where in the plant the piping goes --
20 MR. KERR:
Well, but I've also gotten an impression 21 that things can be done to make the V sequence much less 22 likely Now if that's the case, it seems to me that that has 23 to be taken into account in determining the resource 24 contribution to that sequence.
I may be completely in error.
25 But what I'm suggesting is, you have said -- and I
134 1
think this is correct -- that when you're analyzing plants, b\\
2 you look at the plants as-is.
t j
3 MS. MITCHELL:
Yes.
4 MR. KERR:
But when you start allocating resources, 5
it seems to me, you know, that you need also to look at 6
whether there is a major contributor that has an easy fix or 7
not.
8 MS. MITCHELL:
Oh, you're right.
Absolutely.
9 MR.
KERR:
Now that's the reason I'm asking, let's 10 take the-V sequence.
Does it have an easy fix?
11 MS. MITCHELL:
I can't tell you, because I don't 12 think that anybody has looked at the V sequence for all of the Os 13 plants.
14 MR, KERR:
But don't you think somebody needs 15 to, if its going to be explored further, if it looks like a 16 major contributor?
17 MS. MITCHELL:
The outlier procedure had jolly well 18 better focus on what do you need to look at for the V sequence 19 to decide whether or not it's an important contributor, and I 20 think that Zoltan -- Zoltan probably left -- I think he would 21 agree that --
22 MR. KERR:
You see, in a way it's a part of 23 containment behavior.
1.1though we've said it isn't, it really 24 is, because it's a fairly important bypass mechanism.
\\
('
\\/
25 MS. MITCHELL:
It's an initiating event.
You don't 1
135 1
have the event until -- it isn't something that's a cause 2
thing, like overpressure failure is caused by the accident 3
happening.
This is the accident which happens that bypasses 4
the containment.
5 MR. KERR:
But the containment being left open is 6
not an accident either, unless you want to define it as an 7
- accident, But it can be, and I expect it is, an important 8
contributor to risk, you know, just accidentally being.open.
9 MS. MITCHELL:
Yes.
That also i s a plant-dependent 10 insight in the fact that it's not so likely that either an 11 inerted or a subatmospheric containment would be inadvertently l
12 left open.
So that is again a plant-specific insight.
- '\\
13 M2. KERR:
Now at some point, is it likely that the
[b 14 Staff will draw some conclusion about which is the best 15 containment?
16
[No response.3 17 It's maybe not important for existing plants, but it 18 seems to me it could be important if we are talking about new 19 plants.
It might even be important for existing plants.
- But, 20 you know, if there is a clearly best containment -- and I do 21 not have an opinion at this point; I'm not trying to sell 22 something -- but it seems to me that if there is a clearly 23 best containment and you can reach that conclusion, it ought 24 to be discussed.
s 25 MS. MITCHELL:
Well, there are standard plant i
?
~..
,,,,_,.n..
_,,.--n..
-,-,.,n,,
I i
136 1
de' signs.for the major vendors which are ~ different.
(s
( )
2 MR. KERR:
I know that.
But I don't think they are 3
different because the manufacturers have looked at this 4
question in detail, do you?
5 MS. MITCHELL:
I think that that perhaps was some 6
aspect of it, the idea that severe accidents are something 7
t h a't you need to c o n's i d e r.
I can't answer.
I am not a 8
vendor's representative.
9 MR. MEYER:
Well, there is a severe accident risk 10 reduction element in the rebaselining program, and i
11 unfortunately we are not the group that's running that, and 12 I'm not sure that I have any comments that you might --
g '%
13 MR. KERR:
Well, I am simply saying again, r
\\h' 14 containment-behavior is very important, we agree.
15 My sidekick says I should listen to you, and I will 4
16 take his suggestion.
17 MR. CUNNINGHAM:
Well, there are two things.
First 18 on your question on the V sequence, I guess several years ago 19 the Staff -- the question of the risk associated with the V 20 sequence.did come up in the Staff, and it was -- just as a 21 matter.of record -- it was addressed as a question of, if it 22-was going to be dealt wi th, -i t was going to be dealt with by 23 reducing the likelihood of the sequence, but it was not dealt 4
24 with from a mitigation standpoint or anything.
It was 25 clearly,. the way to deal with V was to prevent it.
And there
-~- -.,,_,_,-_,- -
-y.,_._..,.,,.,,.,,,w,,mmm%,_.,,..
.,_._...,_,o.m.,.,.-...,__,,.,._,..-,.,...,,,m_,.w,
-y,,-y._
137 1
are modifications that, I believe, that went on in some of the g
2 plants to account for that.
g 3
For the second part of the question, I think that 4
the risk work that will come out will give you an idea of the 5
relative benefit, if you will, of the various containment 6
types.
Whether we actually go out and say, "Well, therefore, 7
everybody should have this," or something is unlikely, but I 8
think that the information will be clear, the relative benefit 9
of the various containment types.
10 MR. BENDER:
Can I go to the other end of the 11 spectrum?
Is there a containment that will resist the ATWS 12 consequences of BWRs?
Is there any assumption that postulates
- ['N 3
that the containment doesn't fail when the ATWS occurs in a 14 BWR?
15 MS. MITCHELL:
The location of the containment 16 failure is very important.
If the fission products go through 17 the suppression pool on their way out, it's a very different 18 situation from failing in a location where for some period of 19 time they wouldn't go through.
That kind of thing ends up 20 being a view --
21 MR. BENDER:
Well, you may be missing the point of l
22 my question, so let me try it again.
l 23 The postulates that have usually gone with the most 24 serious BWR ATWS is that the head blows off the pressure r
25 vessel, and that, in' turn, goes through the containment' t
i
--e v
i--wem r-m
,-m--r-y=-
er*-%-~m---e-%w--->-e----T
- * - - - ^ - -
rm-T --
138 1
structure.
~~
2 Ncw I don't know whether that postulate hangs around 3
still 4
MS. MITCHELL:
I don't believe that's the case.
4 5-MR. SHEWMON:
You keep spreading old stories.
]
6 MR. BENDER:
Well, I'm-trying to get things sorted 7
down to the point of saying, what is the present postulate 8
that we're working with, and what is its limiting case at the 9
moment?
4 10 MS. MITCHELL:
The present postulate on ATWS is that 11 there isn't just one ATWS; there are several ATWS sequences.
12 There are the fail first and melt later, and melt first/ fail
("'N 13 later sequences.
14 MR. BENDER:
I don't care about the coremelt aspects 15 of it.
I am more interested in the loadings that they impose 16 on the containment, because in the end --
17 MR.
XERR:
Mike, that's what she is talking about.
18 MR. BENDER:
Well, I think she is only talking about d
19 the sequence of events.
20 MR. XERR No.
She is talking about core loadings, 21 but at a later point than what you're talking about.
You're 22 talking abou+ an absolutely uncontrolled ATWS where you don't 23 even get pump trip or anything, I think.
And then there are 24 certain members of the ACRS who'say that that will lead to the k~
25 vessel coming apart
=
m
.-~y_..,_,.,m.
139 1
MR. BENDER:
Right now, the game is played that the
(~'N 2
pump trip helps things out.
There is always that possibility 3
that the pump trip won't occur.
4 MR. KERR:
That's true.
5 MR. BENDER:
And in that event, which is like the 6
other kinds of accidents which are excluded from the accident 7
sequence, it still has some possibility.
8 MS. MITCHELL:
Plus also the relief valves fail to 9
operate to keep the pressure below the failure pressure of the 10 vessel?
11 MR. BENDER:
Well, that set of events hasn't been --
12 you've just reduced the probability of it to a level where you
(\\
13
- say, "I
can ignore it."
At the moment, what I'm trying to 14 find out is a little bit about how we're making some of these 15 decisions, what the limiting pressure is going to be, and how 16 you can factor that into the containment loading question.
17 MS. MITCHELL:
The limiting pressure on the vessel?
18 MR. BENDER:
No.
They are combined in this case.
19 If the ATWS event drives the primary cooling system to fail 20 in a certain way, it will impose certain loadings on the 21 containment, and-it's a combination of events that leads to 22 that loading that determines whether containment can survive, 23 whether it will be rapidly loaded or loaded more slowly, and 24 it also' determines whether leakage is tolerable or not.
\\~.
25 I am looking for the spectrum that says, what is the
140 1
worst kind of containment loading that I have to deal with 0) 2 right now?
( j, 3
If the V sequence represents the low leakage, or low 4
leakage but early leakage event, there is the other extreme, 5
which is the worst loading.
6 MS. MITCHELL:
Well, I don't think we are 7
considering a catastrophic vessel failure as an ATWS 8_
sequence.
I know we are not'for Peach. Bottom.
9 MR. BENDER:
I think I'll withdraw the question for 10 now.
11 MR. CORRADINI-Well, let me try to answe'r it for a
-12 PWR.
If you classify the V sequence as a small breach of 13 containment over long times, the other extreme for a large dry 14 PWR would be the direct heating.
I'm at high pressure; I get 15 a melt ejection which gives me a very rapid pressure spike 16 with chemical reaction which sends me over dramatically very 17 large high pressure rises.
18 So I think that's what you're --
19 MR. BENDER:
Well, for the PWR, that is the 20 postulate.
Okay.
21 MR. CORRADINI.
That's the other extreme.
Right.
22 MR. DAVIS:
I had a general question on this whole 23 area of uncertainties.
24 As I read through 0956, I come across several lists k
25 of uncertainties and issues.
One of them, of course, is the
141 1
IDCOR NRC list which is -- what? -- 18 or so problems, and
)
2 then there is the APS list, which is maybe a dozen or so.
3 Then there's --
4 MS. MITCHELL:
It is another 18.
5 MR. DAVIS:
Okay, another 18.
Then'there is a 6
section in the report on the Oak Ridge validation process on 7
page 3-27 and 28, and they have another list of areas where 8
they say the modeling needs to be improved, and also another 9
list which says, "Here are some things that should be looked
'10 at to see if they need to be improved."
That's about 15 or so 11 items.
12 And also sprinkled throughout 0956 in various places
[^N 13 is reference to other problems in the code package.
\\
14-If I line all these things up -- and I haven't done l
15 this, but I get the impression that if you did line all these 16 things up, there wouldn't be any area where someone didn't 17 think there was a problem with it or a major uncertainty.
18 I even thought the ORIGEN calculation was one thing l
19 we could put our hands on as being accurate, but there is even 20 a discussion in 0956 about some problems there.
I 21 Now my confidence in this code package begins to 22 unravel as I read all of these lists of problem areas, and it l
l 2
23 seems to me like someone needs to take a look systematically 24 at these problems and see which ones are important and why and 25 for which sequences to try to determine where we have
142 1
confidence in the calculation and where we need to do more m
)
2 work and how much work we need to do before we can conclude 3
that we have done enough and that every area is being covered 4
adequately.
5 I think that, to me, is a deliciency in the report.
6 There is not a systematic consideration of everything that 7
anyone thinks is'a problem.
8 Now there is some overlap in these lists, but there 9
is also considerable differences in what people think are 10 important.
And I don't know whether they all have the same 11 credibility or whether people have, in fact, gone through and 12 tried to determine which ones are important.
.[)
13 MR. SILBERBERG:
You know, for example, you have D
14 mentioned a number of them, and let's deal with the ones that 15 we have put in the table.
In Table 7.3 is a cross indexing of 16 APS-18 and what we call the Big Eight And you will see that 17 ours pretty well are accommodated by most of the'18.
The same 18_
thing is sort of true of IDCOR.
19 But then there were some open areas that APS listed, 20 which we then discuss and address as, yes, we are doing some 21 work, but the didn't make our Big Eight.
I mean, they were 22 less concerned.
So there was already a conscious separation 23 on the part of the Staff to do that first sorting that said, 24 "Yes, we think, based on what we have seen in BMI-2104 and I,, \\
25 these results, we can make separation."
~-
i i
l
143 1
There is another separation that you could make, m
2 that in fact Tom Kress has been involved in, in a report that 3
hopefully will be published in an OECD CSSI exercise, which 4
says -- that said, " Gee, but depending on the sequence, these 5
things take on more or less different importances."
And l
6 together that was helpful -- I mean, taken together, that was 7
helpful.
8 So to some extent, there has been some of this 9
sorting.
I suspect more could be done, more refinement of 10 that.
11 Do you want to add anything?
l 12 MR. MEYER:
The report really was written almost in l
/'*g 13 a pathological way.
That is, it was a critical review, rather 14 than written as if we were into marketing.
So we tended to 15 try and expose all of the warts and blemishes.
We did, we 16 thought, narrow this down to a small number of areas where we 17 thought the uncertainties were quite.important.
I believe 18 that you will never --
19 MR. KERR:
Excuse me.
Where does that narrowing 20 occur?
21-MR. MEYER:
The eight major areas of uncertainty.
I 22 don't think you will ever reach the point where you will get 23 agreement that we know everything we need to know.
And so I 24 think that's an unreasonable goal, and that the reasonable (Qj
\\s 25 goal is to reserve -- if you think you have reached a plateau
7 144 1
where things are so much better than they used to be that you f( )
2 want to go with that as you continue to strive for 3
improvements, and that's where we are,,and you do get from the 4
authors of NUREG-0956 and from the contractors who supported 5
us in this the recommendation that these methods be used and 6
that you start using them.now, notwithstanding their 7
shortcomings.
I 8
MR. KERR:
Mark?
9 MR. CUNNINGHAM:
If I could add a bit, speaking for 10 Mal Ernst, I think it is the intent of the people who are 11 going to write NUREG-1150, to go back and after the risk 12 estimates have been put together, to go back and investigate
/
13 the sensitivity of that risk to especially the IDCOR issues 14 and at least the more important IDCOR and APS issues 15 quantitatively.
So you will see in 1150, at least as it's 16 envisioned now, a measure of the importance of these things to 17 the risk.
18 MR. DAVIS:
I agree that you have done a good job of 19 outlining the problems that people have brought up.
I think 20 it's very candid and done quite well 21 The other thing that bothers me is, it's not clear 22 to me how much uncertainty you are willing to live with, and I 23 know that problem has been raised before, and I don't want to 24 belabor it here, but it is going to be difficult to tell how 25 much work you need to do, because you don't have a good feel
l 145 1
for what uncertainty is acceptable, and I don't have an answer r"%
)
2 to that either, and i +. ' s going to vary from plant to sequence 3
to how much risk you can stand.
4 MS. MITCHELL:
And issue.
5 MR. CORRADINI:
But can I make a suggestion there, 6
though, to go.on with what Pete was saying?
7 Supposedly you have got the tool, I mean, to do 8
that, and that is, you had said previously, and Pete and I 9
before when we were discussing just privately that you are 10 always going to be subjective about these branch points, 11 because you don't understand the physical process.
So the 12 picking of the branch point probabilities for anyone who
(
~ 1:3 constructs the event tree in a sense is the assignment of
\\
14 uncertainty, because if I knew enough about the physical 15 process, when I came up to a physical process branch point, it 16 would be 0 or 1.
17 Do steam explosions fail the containment?
No.
18 Zero.
And I go on, okay.
Or subsidiary questions.
Do I get 19 core debris bed coolability when I have water on top of a 20 molten core?
Zero or one, if I knew the physical process.
21 So the assigning of the branch point probability is 22 essentially the subjective way of assigning uncertainty.
So 23 one way to look at it is to go through all of this and to see 24 how much of an assignment of uncertainty -- what uncertainty p_
25 you assign by these numbers, these subjective numbers, and
146 1
when it becomes unimportant,
(_j 2
MR. WARD:
I don't really think that's right, Mike.
3 You aren't picking, you know, fractions to go in different 4
directions based only on uncertainty.
There might be a 5
spectrum.
If there is a spectrum of possibilities.
I mean, G
6 it's not a deterministic world.
7 So I don't think that's really a total expression of 8
uncertainty.
9 MR. KERR:
You can't argue with what he thinks.
10
[ Laughter.3 11 That's what he thinks.
12 Continue.
More questions?
(
13
[No response.]
14 Let me, tot of ignorance, try to compare the way in
'which this code package deals with core damage versus coremelt 15 16 and the way in which it was dealt with in WASH-1400.
If I 17 understand WASH-1400, one reaches a certain level of damage, 18 and then it is melt.
19 That's not quite the case in this code package, is 20 it, or is it?
21 MS. MITCHELL:
These are all coremelt.
Everything j
22 that you see in 0956 is coremelt.
23 WASH-1400 had PWR 8 and 9 release categories.
24 MR. KERR:
What does "coremelt" mean in this code N-25 package?
147 1
MR. MEYER:
Why don't we ask Pete Cybulskis to r%
(
)
2 compare the coremelt process, just describe it for BMI-2104 3
and compare it with WASH-1400?
4 MR. KERR:
I don't want a lot of comparison.
I just 5
want to know how it differs, how you identify coremelt in this 6
code. package.
7 MR. WARD:
Are they equating coremelt with loss of 8
assured core cooling or --
9 MR. KERR:
I just want to know how it's identified, 10 as compared with the way it would have been identified in 11 WASH-1400.
12 Do you understand the question?
["~'Y g
13 MR. CYBULSKIS:
I think so.
I'll try to answer it.
14 In WASH-1400, basically we looked at the available 15 safeguards that were available, and somebody came up with 16 probabilities that say that certain safeguards would or would 17 not be available.
Whenever somebody told us that less than 18 the minimum ECC systems were available, we said, that is 19 coremelt.
So in that sense, in today's light, we are fairly 20 conservative.
In today's light, we are starting out at a 21 similar point.
But when somebody says that you will have less 22 than three out of six pumps or whatever, we don't necessarily 23 assume that it is coremelt.
We run the conditions through the 24 calculation and determine whether, indeed, it leads to
\\w -
25 ultimate core melting or not.
d$
148 1
And.in some cases, sequences fall by the wayside C\\
2 where we decided, gee, that wasn't a coremelt after all.
(~J 3
MR. KERR:
Ultimate core melting means that the core 4
temperature reaches some point; is that true?
5 MR, CYBULSKIS:
Ultimate core melting means that 6
essentially all the core will melt.
Just some of the core 7
reaching the melting point and then being cooled off is not 8
defined as a coremelt.
Coremelt would be an unrecoverable 9
coremelt, if I can use that expression.
10 MR. KERR:
But when is an unrecoverable coremelt --
11 how does one identify that?
12 MR. CYBULSKIS:
There is a certain element of dN 13 judgment involved, but I think the clearest distinction would
(
14 be if you ever get to the point where the core slumps into the 15 bottom head, you will not make arguments about recovery.
If 16 you recover the systems or the energy balance is such that you 17 predict fractional core melting, which eventually cools off, 18 you will probably say that that is a recoverable sequence.
19 MR. KERR:
And you can, in your present scheme of 20 things, get to a point at which you reach fractional coremelt 21 and then recover?
You can treat that?
22 MR. CYBULSKIS:
Yes, we can treat that, and we have 23 treated sequences of that sort.
24 MR. BENDER:
TMI-2 would be in that category.
25 MR. CYBULSKIS:
Yes.
That would not be a coremelt
/
149 1
sequence by the normal definition.
A,)
2 MR. WARD:
We l l.,
Bill, I thought your question was 3
more along the lines of, what triggers -- I mean, how detailed 4
are the thermohydraulic calculations?
What triggers the 5
beginning of fuel slumping in a given fuel assembly, let's 6
say?
And in what detail do you look at the non-uniformity 7
across the entire core?
8 MR. KERR:
That probably would have been a g o o'd 9
question, but it's not the one I was asking.
10 MR. WARD:
Well, let's assume that question has been 11 asked, then.
12 MR. KERR:
You want an answer to that one?
]
13 MR. WARD:
Yes.
[d 14 MR. KERR:
Did you understand the question?
15 MR. CYBULSKIS:
I think so.
16 In the typical calculation, we subdivide the core 17 into a number of regions.
Each one of those regions has its 18 own power distribution, which is based on the assumed power 19 distribution at the time of the accident, which is typically 20 an equilibrium situation.
21 When the accident starts, again typically the-core 22 is covered, well-cooled at normal power.
Given the 23 interruption of cooling or failure of the ECC system or 24 whatever, we follow the boiloff of the coolant from the core, O
25 and we track the core temperatures on a node-by-node basis.
/
150 1
I rt the typical calculation, there will be 240 nodes 2
representing the core.
3 As you lose coolant, as the coolant level drops, the 4
core temperatures will start to increase due to lack of 5
cooling, due to metal / water reaction, a combination of 6
effects.
7 The code has an input melting temperature.
If the 8
node gets that high, it takes into account the latent heat 9
effusion, et cetera.
10 In the current set of calculations, as typified by 11 BMI-2104, there is effectively internal relocation of the fuel 12 as successive nodes melt, but no relocation of the fuel out of 13 the core region until the bottom node in a given region is 14 molten.
When the bottom node in that given region is molten, 15 then the fuel is assumed to' move out of the original core 16 region and fall onto the support structures.
And-the code 17 takes the falling out, the successive melting, one node at a 18 time, and conceptually the nodes would slump over a period of 19 time.
Practically, the period of time can be quite short.
20 I.t is a function of the accident sequence, a 21 function of the power distribution and a number of other 22 variables.
23 Then just to finish the story, when we get to some 24 point where the majority of the core is molten, and the 25 majority in the typical analysis is 75 percent, basically what
151 1
we say is, if ycu've gotten that far in the accident, you p-(
2 might as well drop the rest of the core in.
And that is our x
3 core collapse point-4 At that point, we cease to distinguish individual 5
core nodes.
We drop the entire core into the bottom head and 6
take the analysis from there.
7 Does that clarify it?
8 MS. MITCHELL:
I want to point out that there is a 9
sort of a schematic of this on page 3-8 of the core slumping 10 model and the core meltdown model on page 3-7.
That sort of 11 draws little pictures about that process.
12 MR. BENDER:
I thought you were going to compare it
("'T 13 to WASH-1400.
How close to WASH --
14 MR. CYBULSKIS:
As I recall, that wasn't the 15 question, but I will,.if it is now.
16 MR. WARD:
Well, I think he did.
17 MR. BENDER:
I don't think he did.
18 MR. CYBULSKIS:
As far as the progression of 19 core melting in WASH-1400 versus now, we use very similar 20 models, but we use considerably di'fferent criteria for the 21 relocation of the core into the bottom head.
22 In the current analysis, as I said, we have 23 effectively regionalized relocation of the core into the 24 bottom head.
In WASH-1400, we looked at three different O
'~-
25 models, but the one that basically was used to drive the
152 1
calculation was the coherent slump model, in which we said
(
2 there is no relocation of the fuel out of the core region 3
until the entire core falls, and again, the criteria to l
4 determine that was again 75 to 20 percent molten.
5 So to put it simply, in WASH-1400, nothing fell out 6
of the core until everything fell In the current analysis, 7
we start out somewhat similarly, but we permit the fuel to 8
move out of the core by individual region.
9 MR. BENDER:
How about the radionuclide transport 10 that goes with it?
11 MR. CYBULSKIS:
In the WASH-1400, we used a 12 relatively~ simple recipe for radionuclide release from the f'^g 13 fuel that was tied to fuel melting.
The fractional release 14 of most of the isotopes was directly tied to the fraction of 15 coremelt.
Ten percent of the core was molten, and ten percent 16 of the iodine was released, et cetera.
17 In the current analysis, which is basically CORSOR, 18 we have a model which releases fission products as a function 19 of time at temperature, rather than being strictly tied to 20 timing, and the coefficients, if you will, the release 21 coefficients, vary with' the different isotopes, so that some 22 1sotopes will come out.even before melting, and other isotopes 23 will be released during melting, and some others may not be r
24 released at all r
(
\\
25 MR. MEYER:
We use a core slumping temperature
153 1
now that is not at the melting point of UO2.
Could you N- )
2
. compare that with what was done in WASH-1400?
s 3
MR. CYBULSKIS:
That's a good point, Ralph.
In 4
WASH-1400, we assumed that effective melting point of the core I
5 was that of the uranium dioxide fuel, somewhere in the 5200 6
Fahrenheit range.
7 Subsequent to WASH-1400 and, in particular, based on 8
a lot of the work that was done at Karlsruhe, what peop1'e have 9
observed is that you get liquefaction or dissolution of the 10 uranium oxide fuel by the cladding, which melts at a lower 11 temperature, and that you can have liquid states at 12 temperatures considerably below the melting point of UO2, and
("'N 13 that the melting point is a function of the composition and
\\s,/
14
-the state of oxidation in the cladding.
15 Our codes cannot take into account all this detail.
16 They vary the melting point as a function of time.
We still 17 are limited to a single melting point.
The value that we use 18 has been chosen to be between the lowest point or lowest 19 temperature at which you would expect any liquid phase and the 20 melting point of UO2, and that is the 4130 number that you may 21 have noticed bandied about.
22 MR. SILBERBERG:
And the liquefaction concept is 23 also consistent with the observations --
24 MR. WARD:
One more question.
What in the WASH-1400 l
I 25 calculations -- wasn't the criterion for the start of -- let's
_e.
.,, _..,,,,, -.-.~.
,,7 m.
,+--*-
.e
154 1
say, heat-up of node fuel different.from what you're using
'S 2
now?
3 MR. CYBULSKIS:
No.
4 MR. KERR:
Are there other questions?
5 CNo response.]
6 In reading the discussion of uncertainty analyses 7
that are likely to occur in the future and. I think, also a 8
paper that Mr. Marino distributed, the impression I get is 9
that it is going to be much more what I would call sensitivity 10 analyses.
I saw a reference, for example, to the Latin 11 hypercube sampling technique, which as far as I can tell tells 12 you nothing about the uncertainty in the model It will give O'.
13 you a good bit of information on the uncertainty of results, 14 based on the uncertainty of data.
15 And a lot of the uncertainty analyses, or so-called 16 uncertainty analyses, that I.' v e seen are really that kind of 17 study.
And I don't see that they give one very much 18
' indication of what the uncertainty in the result is in' 19 situations in which there is a great deal of uncertainty in 20 the modeling.
21 I couldn't tell what was going to be done about 22 uncertainty in modeling, except perhaps to appeal to experts.
23 Is that what is anticipated?
24 MR. MARINO:
Yes, that's about exactly right.
You O
~
25 know, uncertainty, defined statistically for these kinds of
155 1-
. systems we are talking about, in my opinion, is virtually
,-m
(
2 impossible.
You need a lot of data points, a lot of studies.
3 You just can't do that.
4 MR. KERR:
I don't disagree with that, but I don't 5
think that a lot of runs of Latin hypercube sampling makes up 6
for that.
7 MR. MARINO:
Well, in Latin hypercube sampling, you 8
do -- after you do a group of experts that say, "Look, I
9 believe that physically that parameter should not go beyond 10 this low range and this high value," and you get a lot of 11 people together and say, " Yeah, we agree."
We run some other 12 detail codes, and you cannot vary that parameter outside these
'13 ranges for that sequence.
14 Now that's an expert opinion or an opinion of people 15 who work in the field.
That's not an uncertainty analysis.
16 But I would say, my confidence is that it's not less than 17 this, and it's not greater than that, and we'll put a 18 lognormal distribution in between, and we'll do that for 30 19 variables, and now we'll use the Latin hypercube sampling 20 method on those 30 variables over those ranges, and you could 21 call that statistical, but it gives you a band.
22 MR. KERR:
I wouldn't call.it statistical, and I 23 don't even quite know what it's good for, because it is, it 24 seems to me, useful and interpretable only if you have v
25 confidence in the model you have, and there is more in a model
156 1
than just parameters.
There is the rela *ionship between (0;
l 2
parameters.
There is the basic structure of the model itself.
l 3
So the mere fact that you get experts to agree that l
4 a particular parameter can vary only over a certain range 5
doesn't tell you whether the model is correct, not at all 6
MR. MARINO:
No.
In fact, you have to rest'rict 7
everything you say by saying, "I
used these-models, and within 8
the uncertainty engineering ranges for those models, this is 9
what we put on as a band of the output parameter," whatever 10 you want.
You must always qualify it with the statement that 11 we don't know if we have all the right models in there or if 12 all the
(~'
13 MR. KERR:
The publications up to now have sort of I\\
14 hidden that caveat, it seems to me, and have used numbers that 15 depend on what I would call sensitivity analysis.
16 MR. MARINO:
The caveat will always be there, and no 17 one will ever know if we -- if we know every physical process 18 that's going on in this system.
That is absolutely 19 impossible.
We will never know that, and we just do our best 20 engineering judgment.
21 Now if there is a phenomenon like recirculation in 22 vessel that may be important, and it's not in the source term 23 code package, we'll have to treat that separately and say, 24 "All right, now we'll do a run, putting in this recirculation
%s) 25 model and seeing what difference it makes in the answer."
157 1
We should also do runs that make -- that change the f%
(
}
2 shape assumed between the two limits of our variables from, 3
say, a box shape or a lognormal and see what effect that has.
4 But when we're all finished, no one will still, just 5
like for the sequence analysis, stand up and say, "We know 6
everything.
We've checked every possibility there is, and 7
this is your uncertainty."
8 MR. KERR:
But you're burning down a straw man.
I 9
didn't pretend that you said that you knew everything.
10 MR. MARINO:
But we never will 11 MR. KERR:
I simply said that the reports I have 12 read seem to sort of indicate that the results of this
[g 13 sensitivity analysis give you a measure of uncertainty.
I 14 think it is unfortunate to represent that.
15 MR. MARINO:
Well, it gives us a measure of 16 something.
It gives me a measure of how confident'I feel that 17 we know a certain parameter within a certain range.
But I 18 can't put a number of it.
I don't want to say 95 and 5.
I 19 want to say, "As an engineer, I believe it is within that 20 band."
21 MR.
KERR:
I don't want you to.
But I also don't 22 want you to run a lot of sensitivity analyses and get numbers, 23 and then have people who maybe don't know what has occurred to 24 treat those as if they are uncertainties.
\\s 25 MR, MARINO:
Well, maybe we should take the word
158 1
" uncertainty" out of the reports that we right and say, "These
/3
( }
2 are engineering judgments of spreads that we believe are 3
possible for this given sequence."
4 I mean, I don't like the word " uncertainty" either, 5
when people tie it to the word " statistical I just can't 6
apply statistics to this thing.
Only in the sense of a 7
sampling.
The Latin hypercube sampling.is s t a t i s t i c a l^.
It's 8
a statistical method.
But, you know, I wouldn't say that that 9
puts the stamp of statistical validity on all this work.
You 10 can't say that, 11 MR. SHEWMON:
Natural circulation is one of the 12 contenders for being an important unknown, and I am not sure I
[N 13 know what that means,
\\
)
As-14 I I-I read what is in the report here, it seems to 15 end up each paragraph with a punchline about what this will do 16 with regard to the source term.
Yet another place where 17 circulation came up had to '9
<aj ' whether or not you can heat 18 pipes above the pressure vessel enough to make them go plastic 19 with the pressure that's there.
20 Is that also considered part of this uncertainty, or 31 does that come under a different code name?
22 MR. MARINO:
No, you are right.
You know, the 2,3 recirculation will affect the heatup of the --
24 MR. MEYER:
Look back in Section 3.
You will get a
\\w /
25 more complete description of the uncertainties.
I'll tell you
159 1
exactly.
(h
(
)
2 MR. MEYER:
337 and 339 --
x-j 3
MR. CATTON:
If you don't couple them, you can't use 4
MERGE and all the other kinds of codes.
You really have to 5
combine the-fission product transport with your calculation or 6
you're not going to get anywhere.
And you don't have that as 7
an indicated plan here.
8 MR. SILBERBERG:
But in fact, that is done.
I mean, i
9 that's part of --
10 MR. CATTON:
Well it may be, you know, but it reads, 11 TRAPMELT, MERGE, RELAP-5 -- those kinds of codes cannot -- and 12 TRAC.
None of them can handle the stratified flow in the
}
13 hotleg that you need to put in there if you're going to handle sJ 14 the transport.
They can't do it.
15 MR. SILBERBERG:
The stratified flow problem in fact 16 has been under discussion with EPRI EPRI has been looking 17 MR. CATTON:
Yes, but EPRI is not using TRAP, not 18 using SKEDAP.
They built their own code to do it.
19 MR. SILBERBERG:
Yes, I understand.
And --
20 MR. CATTON:
And published the results at the Denver 21 Heat Transfer Conference.
22 MR. SILBERBERG:
I think that using TRAC in the 23 context the way the Sandia people and the Los Alamos did, can 24 also give you that information.
)
N' 25 MR. CATTON:
Well, I talked to Dennis Lyles just
160 1
yesterday or the day before.
TRAC'cannot handle that kind of 2
stratified flow.
It cannot.
So I think if you. don't tie 4 3
back to 1(a) somehow, you're not closing'the issue.
4 MR. MARINO:
Ivan, I might be wrong on this but 5
I thought someone was also using the COMEX code to help in a
6 this area.
4 7
MS. MITCHELL:
Yes.
8 MR. MARINO:
Can that handle what you're asking?
9 MR. CATTON:
To a certain extent, the COMEX code 10 can.
11 MR. SILBERBERG:
That's a good point.
In fact, 12 specifically -- and I've forgotten -- that simple stratified
{/
\\
13 flow ~ problem in the hot leg has specifically been assigned to j
s.
14 COMEX and Bill Shaw, using the COMEX code, is.trying to 15 address that piece of it.
Now, is that mentioned in here?
16 No, it's not.
17 MR. MARINO:
That should go in.
That's a good 18 comment.
19 MR. KERR:
Are-there other questions?
20 MR. CATTON:
I don't know why you just don't use the 21-published EPRI work.
22 MR. KERR:
That's not a question.
23 MR. CATTON:
Sorry.
24 MR. SILBERBERG:
We may eventually do so.
In fact, 1
)
25 we're having exchanges with them.
161 1
MR. CATTON:.
I'm aware of that.
2 MR. KERR:
Other questions?
3
[No response.]
4 MR. KERR:
There being none, I want to thank the 5~
staff for participating unde.- rather traumatic circumstances 6
with a hurricane going on and everything.
You are not unused 7
to a free flow of air when dealing with this committee, so 8
maybe. it's not all that unusual 9
I will call a luncheon recess at this point and 10 expect some of you back in about an hour.
11
[Whereupon, at 1.: 1 0 p.m.,
the meeting was recessed 12 for lunch, to reconvene in an unrecorded session later the 13 same day.]
14 15 16 17 18 1
19 1
20 21 22 23 e
24 O
N' 25
Oi U
i 1
CERTIFICATE OF OFFICIAL REPORTER 2
3 4
5 This is to certify that the attached proceedings 6
before the United States NucIear ReguIatory Commission in the 7
< natter of. ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 8
9 Name of Proceeding:
class 9 Accidents Subcommittee Meeting 10 11 Docket No.
O 12 PIace:
Washington, D. C.
L, 13 Date: Friday, September 27, 1985 14 15 were held as herein appears and that this is the original 16 transcript thereof for the file of the United States Nuclear 17 Regulatory Commission.
18 (Signature)
(TypedNameofReprter) [S'uza ge B. G ng 20 21 l
22 23 Ann Riley & Associates, Ltd.
24 4
25
O O'
AtddLO e
l l
1 i
i l
Major Advances in Source Term Technology
- Fission Product Chemistry
- Retention in Reactor Coolant System (RCS)
- Core Melt Progression i
- Aerosol Behavior in Containment i
- Fission Product Release from Fuel 2
- Core-Concrete Interactions
- Containment Pressure Loads i
i l
i
O O
O COMPARISON WITH WASH-1400 SURRY PLANT 3
CONTAINMENT RELEASE RELEASE WARNING ENERGY OF 6
ACCIDENT FAILURE TIME DURATION TIME RELEASE (10 i
SEQUENCE MODE (HR)
(HR)
(HR)
CALORIES /SEC)
I CONTAINMENT NOT APPLICABLE 1.0 2.0 0.8 0. 7.
1 BYPASS LOCA (V)
WITH WATER CONTAINMENT NOT APPLICABLE 1.0 2.0 0.8 p.qt l
BYPASS LOCA (V)
WITHOUT WATER
,s i
j WASH-1400 PWR-2 2.5 O.5 1.0 12
-5
]
RELEASE FRACTIONS j
ACCIDENT NOBLE 1
SEQUENCE GASES IODINE CESIUM TELLURIUM BARIUM RUTHENIUM LANTHANUM l
1 l
V WITH 1.0 8x10-2 8x10-2 2.5x10-2 2.2x10-3 1.3x10-4 7.0x10-5 WATER j
V WITHOUT 1.0 4x10-I 4x10-l 1.2x10-I 1.1x10-2 6.5x10-4 3.5x10-4 WASH-1400 0.9 7x10-I 5x10-I 3x10-l 6x107 2x107 4x10-3 2
2 1
j
O O
O COMPARISON WITH WASH-1400 SURRY PLANT CONTAINMENT RELEASE RELEASE WARNING ENERGY OF 6
ACCIDENT FAILURE TIME DURATION TIME RELEASE (10 SEQUENCE MODE (HR)
(HR)
(HR)
CALORIES /SEC)
STATION OVERPRESSURE 2.5 10.0
.5 el BLACKOUT (TML5'_)
(6)
WASH-1400 PWR-2 2.5 0.5 1.0 12 RELEASE FRACTIONS ACCIDENT NOBLE SEQUENCE GASES IODINE CESIUM TELLURIUM BARIUM RUTHENIUM LANTHANUM
-2 1x10 2 1.3x10-3 1.7x10-4 7x10-2 5.8x10-2
)
-I 0
w-mw
COMPARISON WITH WASH-1400 PEACH BOTTOM PLANT CONTAINMENT RELEASE RELEASE WARNING ENERGY OF 6
ACCIDENT FAILURE TIME DURATION TIME RELEASE (10 SEQUENCE MODE (HR)
(HR)
(HR)
CALORIES /SEC)
0VERPRESSURE 1.5 10.0 RELEASE TO ENVIRONMENT (y')
WASH-1400 BWR-3 30.0 3.0 2.0 1.4 RELEASE FRACTIONS ACCIDENT NOBLE SEQUENCE GASES I0 DINE CESIUM TELLURIUM BARIUM RUTHENIUM LANTHANUM S
TC 1.0 1.6x10-I 1.4x10-I 4.8x10-I 3.9x10-I 1.9x10-3 1.8x10-2 BWR-3 1.0 1x10 3x10-I 1x10-2 2:10-2 2x10-2 3x10-3'
-I I
i i
>m-
-.--e
Source Torm Codo Suite Fission Product Thermal Hydraulic O
Transport Behavior
_q l
ORIGEN MARCH l
i i
Fission Pr,oduct g
g,,,
inventory in Fuel h%r of Reactor Coolant System, =1 g
Molten Core, and u
g]
l Containment l
,l CORSOR l
E i
l l
l
}
Release l
Retained l
l in Fuel I
u from Fuel i
l l
I l
l l
a TRAPMELT 5
MERGE l
m l
'E Detailed Temperature, 3
O l
Reactor Coolant System use Pressure, and Flow in 1
g Transport and e
Retention j
I i
_ _ _ _. _ _ _ _ _ _ _ __ _ _ a l
l l
l l
l-CORCON l
VANESA I
E
-l 2
S Release frorn m'===="a Detailed Core-d El Core-Concrete Melt i
Concrete Temperature l
l and Interactens 3ll l
3 4
!I j
f NAUA, SPARC, ICEDF l
l Containment Transport l
and Retention g
g 1
I L __ _ _ _
O Release of fission products to the environment: Source Term
-....... ~. - -
O CONTAINMENT PERFORMANCE PREDICTIONS IMPROVEMENT COMPARED TO RSS o
MORE.INFORMATION SINCE REACTOR SAFETY STUDY CLWb CPWG RECENT PRA'S SEABROOK SIZEWELL RSSMAP ZION QUEST IDCOR O
SASA o
MORE INCLUSIVE EVENT TREES PHENOMEN0 LOGICAL EVENTS RCS FAILURE MODES EQUIPMENT SURVIVABILITY FAILURE MODES
- O
CONTAINMENT PERFORMANCE PREDICTIONS (CONT'D)
O UNCERTAINTIES IMPORTANT o
SUBJECTIVE EXTENSIONS OF INFORMATION ENGINEERING JUDGMENT o
LIMITED SENSITIVITY STUDY IN NUREG-0956 o.
NUREG-1150 WILL BENEFIT FROM PEER REVIEW 0F TREES o
ADDITIONAL EXPERIMENTS AND EVALUATIONS CONCRETE CONTAINMENT TEST DIRECT HEATING l
PENETRATION TESTS l
l l
' O
g a
O~
s
'='='='=
C's i
ed E
t
!E I
I I
I g
my 4.
3 t
~
m 3
2 2
R b
S d
's 's d 2 5
E t3 g
h 1 3
3 3
3 6.
2 sg sg sg 'sg
""I l
I e
-i 3!
E
=
=
O l3 g1 Ej3
,J.
v 311 ".. '.. '..
il 3
r2
=
3 3
1:
4.
- s. s.
2 3
t
$ 5 E
-~ --
au 3
4
-- I
.h--
s s
jh&
I1 s
n
}
ls
=
O
- 2 l V 2 I2
}
s
=
P14 i
-v--- - -.. - -. - -. - - - -- -. - -..
.--c.
,,y
-.w.,-,
e--
,r,---.-
---w+----.+-
-e_-w..,--r------+---.
Fission Product Thermal Hydraulic i
i i
i O
i p_________________,
l ORIGEN MARCH Fission Product d
l Inventory in Fuel g,
v;,,,,
[gl Molten Core, and
. a Containment l
ml CORSOR l
i l
l l
ul Release l
Retained g
g i
in Fuel l
5 from Fuel i
1i, TRAPMELT i
5 l
MERGE l
I a
1 Detailed Temperatur,e, l
l Reactor Coolant System um.
Pressure, and Flow in Transport and g
l Retention B
I I
i O
' -- - - - -- - - -- - - - -- - -- i VANESA 3
Release from
==
Detailed Core-Core Concrete Melt Concrete Temperature i
3 i
4 l
and Interactions kl l
h I
,l i
+
i
)
NAUA, SPARC, ICEDF l
g l
Containment Transport l
and Retention g
g Release of fission products to the environtnent: Source Term Figure ES.1 8MI-2104 suite of codes as used in the source ters reassessment.
xviii 4
--,--,,----.,e.--,.n--,_-,-
-w,-----,,
.n n-,,,n-,..,.-n-..,
,--.,..,,,--,--e-
0 O
i I
I h
c.
2:
i a
f) 1&
f g
i e
q)
O O
i W
Il Yd
.o 0
-o w o
i g
h h
~
1 6
c i
.Q.
-o s
1 g.--------,
a
,s e
x y
g
(
d i
t I NN O$
I
\\nlk.I
]
L 5,
n...p m.y,.n y
- O l
.---~_._c._.,,,-,-,__,,,_
_-__,_...m.,,,-m..
O O
O Major Areas of Uncertainty
- Natural Circulation in the Vessel
- Core Melt and Hydrogen Generation in-Vessel Releases from Fuel e
Retention and Revaporization in RCS Core-Concrete Interactions Scrubbing by Pools and Ice Compartments Containment Pressure Loads Containment Failure Modes t,
O CALENDAR YEAR 1984 19 5 1988 1987 I
I I
i l
'"*'US'#7.'O'"
1 N
ih.
9 l
2.Coe grossion A
i Generation
- '," s:::'o'.'?s."
M from the Fuel and
- Aerosol Generation d
vap last on of N
Fission Products in the RCS 4
- 5. Fission Product Release A
C and Aerosol Generation L -.a from the Core-Concrete Interaction
- 6. Scrubbing Efficiency M
p of Suppression Pools and Ice Compartments ress eLoads
(
g g
8.
on nment Fellure i
I 1
l Figure 7.1 Important milestones in research programs that are addressing l
major technical areas of uncertainty.
I 1
l 7-2
Table 7.1 Description of milestones shown in Figure 7.1.
Milestone and Description 1.
Natural Circulation in the Reactor Vessel a.
Completed:
COBRA-NC calculat1*on for intact geometry b.
Complete MELPROG calculation for intact geometry c.
Perform preliminary MELPROG calculation for degraded core d.
Complete MELPROG calculation for degraded core e.
Issue report assessing effect of natural circulation 2.
Core Melt Progression and Hydrogen Generation a.
Completed:
Third integral severe fuel damage test in PBF b.
Completed:
Last of 4 integral severe fuel damage tests in PBF c.
-Complete full-length boildown and hydrogen generation test in NRU d.
Release integrated SCDAP fuel damage code package e.
Complete 2 melt progression tests with control rods in ACRR f.
Release integrated MELPROG melt progression code package g.
Begin MELPROG validation experiments V
h.
Complete BWR boildown and hydrogen generation test in NRU i.
Complete last of 6 integral melt progression tests in ACRR j.
Complete MELPROG validation experiments 3.
In-Vessel Fission Product Release from Fuel and Aerosol Generation a.
Operation begun:
ORNL 10-kg core melt facility b.
Completed:
Last of 4 integral tests in PBF c.
Complete initial high pressure fission product release tests at BCL d.
Complete first 2 fission product tests in modified apparatus at ORNL e.
Complete first 2 source term tests in ACRR f.
Complete high pressure fission product release tests at BCL g.
Complete second pair of source term tests in ACRR h.
Complete atmospheric fission product release tests at ORNL V
7-3
Table 7.1 (Continued)
Milestone and Description 4.
Retention and Revaporization of Fission Products in the Reactor Coolant System a.
Completed:
Coupling of the TRAP-MELT and MERGE codes b.
Completed:
Linking of RELAP5/SCDAP/ TRAP-MELT codes c.
Complete chemistry-based revaporization model d.
Complete MARVIKEN tests for TRAP-MELT assessment e.
Complete last of 2 fission product tests in LOFT f.
Complete TRAP-MELT assessment tests at ORNL g.
Complete aerosol resuspension tests at ORNL h.
Complete TRAP-MELT assessment with experimental results 5.
Fission Product Release and Aerosol Generation from the Core-Concrete Interaction a.
Results due from standard problem calculations b.
Complete SWISS tests c.
Complete and analyze SURC tests 6.
Scrubbino Efficiency of Suppression Pools and Ice Comparteents a.
Issue code manual for improved ICEDF code b.
Issue code manual for improved SPARC code c.
Complete comparison with new data for validation of SPARC d.
Complete comparison with new data for validation of ICEDF 7.
Containment Pressure Loads a.
Complete diffusion flame model b.
Complete tests on interaction of molten urania and concrete c.
Complete tests on direct heating of containment atmospheres d.
Develop coarse-mixing model for core-coolant interactions e.
Complete direct-heating analysis 8.
Containment Failure Modes a.
Completed:
Four tests on steel c6ntainment models b.
Complete tests on seals and gaskets c.
Complete tests on electrical penetrations and valves d.
Complete tests on large concrete containment model e.
Initiate seismic tests 7-4
i I
CODE STRATEGY l
l Tier 1: INTEORATED CODES i
Stef Fles COOf PAC 8tAGE 11ee Genersetent e8ELCOR lau.4 Geneseelen8 Tier 2: DETAILED MECHANtSTIC CODES 4
Y SPAAC pASTORASS a
M MM Ea*
woe.,
SCEOF i
l TRApe0ELT RSLAP S SCDAP M.
)
MW M
esageOS eseCTA vtCTOmeA COA.CO.e.s. VAastsA - - - - -
ymen M,
4 COesTAses i
1P l
TRAC ir egetPetoei '
]
4 Cee mesmose Teensport Vanese Coace.ee R.semes Transport in Causamensment Theesned noen
,,,,,, pues i nCs to e..
s e.ces e.
o.as ee C n esa.ne.ie t.ede se a e.ses.
t v
F.;; :::'r. of ACChient Phenomena Hgure 7.2 Two-tier code strategy showing the progression ol' accident phenomena covered by each code.
j