ML20132B635
| ML20132B635 | |
| Person / Time | |
|---|---|
| Issue date: | 09/23/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1448, NUDOCS 8509260254 | |
| Download: ML20132B635 (416) | |
Text
""*""
ORIGINAL UNITED STATES OF AMERICA O NUCLEAR REGULATORY COMMISSION In the matter of:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Combined Mocting of the Subcommittees on Structural Engincoring, Seismic Donign of Piping, and P.ctal Components Docket No.
O Location: Wanhington, D. C.
Date Monday, September 23, 1985 Pages: 1 - 327
)
- cmaametapsmce
- Cobif Reporters
( 1625 I St., N.W.
Suito 921 Washington, D.C. 20006 l}y92g4 050923 (202) 293-3950 T-1440 pon
l l
1 I 1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION 3
4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 5 .
COMBINED MEETING OF THE SUBCOMMITTEES ON STRUCTURAL 6
ENGINEERING, SEISMIC DESIGN OF PIPING, AND 1
METAL COMPONElffS e
9 Nuclear Regulatory Commission to Room 1046 1717 H Street, N.W.
11 Washington, D. C.
12 The subcommittees convened, pursuant to 13 notice, at 8:30 a.m., Chester Siess, Chairman of the Structural Engineering and Seismic Pipe Subcommittees, is presiding.
16 17 ACRS MEMBERS PRESP.MT1 14 C. SIESS, Chairman P. SHEWMON 19 J. EBERSOLE C. MARE 20 21 ACEB CDMSULTANTS PREBEMT1 22 E. RODABAUGil M. BEM)ER 23 24 COGNIRAMT STAFF MEMBER 1 25 A. IGNE O
= = . - . - .--.-... _ -- - - .--- -
i . i 2 [
$ l 1 i l
i I '
STAFF AND PRESENTERS PRESENT2 t I' ;
2 J. RICHARDSON l l'
1 3 D. GUIY I
I R. BUDNITE 4 J. COSTELLO 6 J. O'BRIEN 5 R. KENNEALLY j G. ARNDT :
6 H. GRAVES l P. COTA t 7 ,
t 8 i
\ i 9 i i
i 10 l l
11 l l
t tg ,
13 14 15 16 i .
i 17 i
18 I
19 20 i
j 21 ,
2 l I 22 i l 2 ,
! I
, 24 i 2h )
l l
i 1 l 4
3 i PROCEEDINGS 2 MR. SIESS: The meeting will come to order.
3 This is an open meeting of three ACRS 4 subcommittees to review the status of the research that is 5 being carried on in the mechanical structural engineering 6 branch of the NRC, and I believe that these three
! 7 subcommittees cover all of the research. It is the 8 Structural Engineering Subcommittee, the Subcommittee on 9 The Seismic Design of Piping and the Metal Components 10 Subcommittee.
1: I am Chester Siesa. I am Chairman of the 12 Structural Engineering and Seismic Pipe Subcommittees, and i3 on my left is Paul Shewmon, who is Chairman of the Metal u components subcommittee, is We have two other ACRS members present, Mr.
! 16 Ebersole and Mr. Mark, and then we have two consultants, 17 Nr. Rodabaugh and Mr. Bander.
is We are going to cover all of the -- I think it l 19 will be all of the research in the Mechanical Structural 20 Engineering Branch, or a pretty substantial part of it. It 21 will be divided up into several areas, the seismic margins 22 type work, work on containment integrity and on piping 23 reliability.
24 The meeting was planned to go two days with the 25 last portion being pretty much in the area of the Metal
- C
_ _ _ _ _ . . _ _ . -, .__ _ _ _ _ . _ . . _ _ _ . . _ _ _ ~ . _ . . . . -. _.,
4
() i Components Subcommittee. Because some people that we 2 expected to be bare are not here, I don't know what the 3 schedule will look like. We have a schedule and we will )
4 tell you better by 4 or 5 o' clock this afternoon what it 5 will be for tomorrow. I suspect that we may go on to 6 tomorrow, but I am not sure.
7 I am not sure who is here.
8 Jim, who have got from the Mechanical Structural 9 Engineering Branch here?
- io MR. RICHARDSON: As on the agenda, Mr. Guzy has not show up yet, and he happens to be the first one on.
If l ii 12 he does not show up, I will give the overview and Bob d
i3 Budnitz will speak for the expert panel. Dr. Costello is
! u here, and Dr. O'Brien I am sure will be here, and Mr.
i 15 Kenneally and Mr. Arndt and Mr. Graves are here. Mr.
l 16 Kassawara from EPRI has been sent to Mexico City as a i
i i7 member of a team that is going in for an early look at the is damage and the lessons to be learned out of the Mexico City i, disaster.
) 20 And apparently EPRI, it was so bactic on Friday on i
! 21 the West Coast in assembling this team, I never did get to speak to Mr. Kassaware or to Dr. Wall. So I am assuming, 22 23 since there is nob,w; 53re from EPRI that EPRI probaoly 24 will not be reg resented, 25 We will try our best to fill in and describe for lO 4
I
- u. - _ _ - _ _ - _ _ - - - - - _ - - _ - _ _ _ _ - - - - _ _ . - _ - _ - _ _
5
() i you what EPRI is doing from our perspective, and of course that is all you are going to get is an NRC perspective, 2
a wnAch will be incomplete and imperfect, but we will tell 4 you what we are doing with them in conjunction with, and we 5 have a pretty good idea of their programs.
6 MR. SIESS: Well, that will be better than 7 nothing, but I an extremely disappointed that EPRI does not a have a representative bere.
9 MR. RICHARDSON: I am, too.
to MR. SIESS: Either a project manager or somebody 11 more technically knowledgeable about the work. I had hoped 12 that we could hear enough about what they were doing to be i3 able to see bow it relates to what the NRC is doing. I l
u think that is extremely important because we are seeing a is decrease in budget and we are seeing a need for greater 16 dependence on industry-sponsored research, and it was my 17 hope that a major focus of this meeting to see what you were doing and the industry was doing, how they related and og how the cooperation was developing.
20 But I think it will help if you tell us what you 2i know about what they are doing and the extent to which you 22 work and their work is being coordinated.
23 Is there somebody here from NRR7 24 MR. COTA: Yes.
25 MR. SIESS: And you represent which group?
(~l
\>
i l
6
() i NR. COTA:
MR. SIESS:
The Standards Coordination Branch.
Is there anybody here from the 2
3 Containment Performance Working Group?
4 (No response.)
5 Am I correct that the user la the Containment 6
Performance Working Group?
7 MR. RICHARDSON: Yes.
8 MR. SIESS: And they are not here.
9 MR. RICHARDSON: They were invited.
io MR. SIESS: Well, maybe they will show up.
n To continue with the introductory remarks ---
12 MR. SHEWMON: Could you tell me --- ,
1 13 MR. SIESS: Let me get rid of one paragraph.
i4 MR. SHEWMON: Okay.
MR. SIESS: And finish up the boilerplate. A 15 16 transcript of the meeting is being kept. So it is i7 requested that each speaker identify himself or herself is when they first speak and use a microphone.
19 We haVe received no written statements from 20 members of the public and no request for time to make oral 21 statements from the members of the public.
i 22 Now Paul.
23 MR. SHEWMON: In one sentence is the Containment Working Group? It sounds like a fine idea, but I wasn't 24 25 awareof it.
O
i 1
7 f
1 MR. SIESS: There are two groups dealing with 2 containment in connection with the severe acciuent researca ,
1 3 and the severe accident policy statement, one called the 4 containment loads working group, which is concerned with ,
5 what happens inside the containment from the molten core, i 6 et cetera, and then there is a Containment performance 7 working group, which has been concerned chiefly in the past
, s year or so with the possibility of leakage from containment 9 under high pressures and high temperatures as well as io containment ultimate capacity.
11 MR. SHEWMON: The owners group NRC?
12 MR. SIESS: No, not owners group. .Did I say 13 owners group?
14 MR. SHEWMON: No.
15 MR. SIESS: These are NRC working groups.
16 MR. SHEWMON: Okay.
17 MR. SIESS: The containment performance working is group is a mixture of NRR research people and it is part of 19 the severe accident program.
20 I have oeen keeping track of that group. NUREG qi 1047, I believe the draft is out for comment, which is the 22 report on that. There is also a report out on the 23 containment loads working group that the Class 9 Accidents 24 Committee has reviewed.
25 Are there any questions from members of the 0
- _ ~
8
() i subcommittee or the consultants or any comments before we 2 get started?
3 (No response.)
4 Our agenda is broken down, as I said, with the 5 seismic program first and then the piping research program, 6
other than the leak before break part. That is mostly --
7 well, we will worry about bow that divides up. Then the a
containment integrity and then the leak before break, and 9 the leak before break is more specialized I believe into io the balance of plant part, and we will let you explain that n when we get to it.
12 Jim Richardson -is Chief of the Mechanical
) 33 Structural Engineering Branch, and we will turn it over to u you, Jim.
15 MR. RICHARDSON: I might ask you pleasure on the 16 balance of today and tomorrow. Would you prefer that we l i7 speak from the podium with viewgraphs, or would just rather l
i la we sat around the teole and talked from the handouts?
j9 MR. SIESS: Are there enougn handouts for the 20 P'0Ple in the audience?
21 Does everypooy in the audience nave a copy of tne 22 nandouts? We nave copies for tnem.
23 (Copies were distriDutea to tnose not naving the 24 nancouts.)
25 And in that case I think we signt just as well sit
- O 1
I 9
I around the tanle and work from the handouts. They don't 2 have page numbers on them, but if you take them in order, a we will ---
I 4 MR. RICHARDSON: Yes, they will take them in the ;
J 5 general order. What you have before you ---
6 MR. SIESS: Does.anybody object to that?
- 7 (No response.)
e I real handouts a little better than the screen.
i 9 MR. RICHARDSON: What you have before you is a io package of all of the viewgraphs that the staff will be n using. What is missing from that package is what Bob 12 Budnitz will present, and I believe that Al Igne has copies i3 of his handouts, and of course the EPRI handouts.
14 MR. SIESS: Well, if anybody feels more is comfortable at the podium with slides, they can do that. i 16 Jim, obviously this review has some relationship 17 to our annual reports to the Commission on the research is program and budget and to the Congress on the research 19 program and budget. But the committee is-reconsidering the 20 scope of its reports. Frankly, we are trying to cut back I
21 on the scope, and there is some feeling that our 22 involvement with the budget down'at the line item level 23 Practically is not too fruitful.
24 So I would suggest that you give us much less
- 25 emphasis on the budget than maybe we have gotten in the O
10
() i past when we were making these reviews, but don't ignore it completely. You have a- couple of pages and it would be 2
3 nice to mention it just to get the scope of this, but then 4 not too much detail after that.
5 MR. RICHARDSON: We have structured it somewhat 6 along that line in that each of the individual 7 presentations you will see the budget, but in some areas we a
have combined individual fins when it made sense to bring 9 them together because they had common objectives and scope io that were very similar that it made more sense to bring ,
ii them together into a package that gave you a better feeling 12 for integration.
i3 I will start off with the overall budget.
( i, In 1985 in the seismic area we are spending about 33
$5.2 million and about $1.4 million of that will represent 16 Programs that are terminating in FY-1985.
i7 In the containment area on the nest page in 1985 j
la we will be spending about $3.9 million.
i9 In '86 our seismic budget will increase to 20 approximately $6.3 million.
21 And in 1987 to $7.7 million.
22 The reasons for the increase are primarily 23 centered around increased ~ emphasis on testing. We are 24 going from an analytical phase in the seismic program to 25 testing in the validation area in validating our methods O
l
11 i and improving our fragility and data base. Those account f) 1 2 for most of the increases in 1986 and 1987.
3 MR. SIESS: Now that is component fragilities 4 primarily?
5 MR. RICBARDSON: Yes, component fragilities.
6 There is an increase in the seismic category one structures 7 Program, which is also a fragilities like program, and it 8 is increasing through 1987.
9 It is our stated goal that the seismic program io will be for all intents and purposes ended in 1988 We '
ii would like to have the seismic issue pretty much resolved, 12 altnough I am quite certain that beyond 1988 there will be i3 seismic questions remaining to be answered. But we would 14 bope that the bulk of those questions would be put to rest 3
15 and that our seismic budget beyond 1988 would be rather i
16 small.
17 MR. SIESS: Jim, just for your information, I just is received a copy of a letter report on the category one 19 structures program from the new project manager down there.
20 MR. RICHARDSON: Yes.
2i MR. SIESS: I forget his name.
22 MR. RICHARDSON: From Joel Bennett?
23 MR. SIESS: Yes, that's it. I just go that last 24 week, and it was very helpful and a good report.
25 MR. RICHARDSON: Good. I assume that all of you ,
l l
_, _ _ - . ___..____. _ _ _ _ _ _ ~_ _ _ _ . _ , _ _ - , . . _ _
d 12
() i have received copies of our seismic safety research program 2 Pl an which documents and tries to show the overall seismic ,
3 program and how it is integrated with the programs within 1 4 the NRC and within industry as well as foreign elements 5 involved in seismic research.
t 6
Another report-that nas also Deen recently 7 published, which I believe you should have all received, is a
the first report, the first substantial report from the ;
9 expert panel on seismic margins, which was published I io believe in July. Those two should have been in your hands n for the last few weeks anyway. ,
12 MR. SHEWMON: Do you have NUREG numbers on either 33 of those?
u MR. RICHARDSON: Yes. The Seismic Safety Plan, Roger, is NUREG7 l 15 f 16 MR. EENNEALLY: NUREG 1147.
17 Mk.' RICHARDSON: And the seismic margins report, is which is entitled An Approach to the Quantification of i9 Seismic Design Margins, I believe, and, Dan, that NUREG l 20 number?
21 MR. GUEY: It is NUREG CR-4334.
22 MR. SHEWMON: I think I have seen that, but I 23 don't remember seeing the other one.
24 MR. SIESS: The other one was a lot earlier.
25 MR. RICEARDSON: Yes.
O I -- . , . - _ . . . . , . . - - -m,,, ,
7, . . . . - - . _ _ . . . . . - . _ , , , _ . _,
13 h) 1 MR. SIESS: It was quite some time back.
2 MR. RICHARDSON: June I believe was the publishing 3 date of that report.
4 MR. SHEWMON: Okay.
5 MR. MARK: What do you expect of the program on 6 non-linear behavior of concrete? It has 'just about to start," and you say the whole thing will end in '88, and a that is a pretty hairy sounding study. So it can't 9 possibly Come to any Conclusions in that kind of time.
j 10 MR. SIESS: What page are you looking at Carson?
11 MR. MARK: I am looking half way down the list of 12 numbers on the first page.
) i3 MR. RICHARDSON: I would ask that we defer that.
, 14 We will be talking specifically about that program. It is 4
15 not as ambitious as the title would probably imply.
16 MR. SIESS: The emphasis there is on the 17 reliability analysis and not on linear. This is the 1
l 18 Brookhaven work.
i 19 MR. MARK: That helps me.
20 MR. RICHARDSON: And, as you will see when we get 21 to that slide, that program is not completely defined. We 22 have where we think we are going in it, but we will have -
23 not resolved it completely with the user office.
24 MR. EBERSOLE: Chet, may I ask a question? I 25 would like to get some feel for the perspective with which ,
O
14 i you approach this complicated problem. There is an action 2 Plan, TAP-45, which looks at the hypothesis, you know, that 3
we might have to have a dedicated shutdown heat removal 4 system and some improvement to it.
5 In the long run in an earthquake we have got to 6
get the reactor shut down, and I will dismiss the thesis i 7
that we are going to fail the scram system, and we have got a
to get the heat out afterward, and that is the end point of 9
whatever we do in the seismic field, we have got to get the 10 heat out.
ii In doing all this work, which involves thousands l 12 of components and pieces of equipment, do we ever look 13 sideways and say if might be more practical to diminish the i, target size to a shutdown heat removal system with is conservative competence and give this other stuff the soft 16 touch as the more practical approach to the whole thing?
- 7 MR. RICHARDSON
- I think you will see, and we are is pretty enthused about what is coming out of our seismic i, margin studies, and in fact that study ~by examining more 20 than a dozen PRA's has been able'to start to cull systems 21 that really aren't involved in the safe shutdown and decay 22 heat removal systems, and in fact it has allowed us then to 23 in a rather straightforward and much simpler way arrive at 24 estimations of margins by doing exactly that, culling 25 unnecessary systems out of the --
O l
- 15 i MR. EBERSOLE: Even though they may be classified.
i 2 MR. RICHARDSON: Oh, yes.
! 3 MR. SIESS: The point is that the seismic margin i 4 study _ tended to get away from the initial question, which i
5 was the seismic margins of components and structures, and j l
6 to look into the seismic margins of the plant and findings 7 those systems that you needed to protect the plant rather a than the margin in a particular component.
9 When we hear from Bob Budnitz, I think you will to get a picture on that, and it is very important. First n they broadened the look narrowed the focus, i 12 MR. BENDER: Well, the focus is on the existing 33 installations as opposed to something else. <
14 MR. RICHARDSON: Yes. The objective is to look is at, examine and estimate margins of operating plants, 16 Although there may be spinoff on improvements in design 17 Criteria that would be applied to new plants, but that is certainly is not our objective.
19 MR. EBERSOLE: You know a case in point is the 20 relay chatter thing, and why do we ha.ve to tolerate it?
21 MR. SIESS: Well, that is addressed by the seismic 22 margin panel. I think that panel brought a very important 23 perspective because prior to that our attention had been 24 just seismic margin of everything and just looking at each 25 piece whether it turned out to be that important or not.
l0 l
16 i And in fact to some estent that is the way the regulations 2 are. Anything that is safety related or important to 3 safety has to be seismically qualified.
4 And there has been some focus here that if you are 5
really worried about the survival of a plant, everything is 6 not equally important.
7 Okay, Jim, go ahead.
8 MR. RIC8ARDSON: Moving'on from the budget and 9
into an overview of the two elements of the program that to you are going to hear today, namely, our seismic research j ii program and our containment research program, and I might 12 say that the piping is really a subset of our seismic
, 13 Program because our piping programs are seismically related i4 programs. So they fall for budgetary reasons within the is contest of seismic research.
We are driven in our seismic research primarily by 16 l
17 the fact that new seismological information seems to be is indicating that for some operating plants the design basis i, earthquake may not have been high enough.
- o The ongoing seismicicy studies being conducted by 21 the NRC and by industry are indicating that there may be 22 for some plants a real problem that the Commission is going to face in the somewhat near future of having to address 23 t 24 the problem of increased design basis earthquakes on 25 Operating plants.
O l
l
F
' 17 i MR. SHEWMON: Is the art of prediction in that
}
2 area getting to where you can say we are concerned about 3 that plant oc the possible earthquake at this point but not i q
4 for that one east of the Rockies?
I 5 MR. RICHARDSON: The answer to that is yes, but 6 whether the art has been honed sufficiently enough for i
7 people to have confidence in that art form, I don't know.
8 I know that this eastern seismicity study will be producing :
9 site specific seismic hasards curves with large uncertainty io bands associated with chose hazard curves. And that is 11 about the extent of my expertise in that area.
12 Whenever I look at that and hear those folks talk, i3 I am convinced that it is truly a black art form of which I
() i4 have little knowledge.
~
15 MR. EBERSOLE: I agree with you fully.
I 16 MR. RICHARDSON: But, nevertheless, the results of j 17 their studies are handed to us on a platter and we had is better do something them.
19 MR. SIESS: Jim, I think you have described the 20 issue from a strict regulatory point of view, that is the 21 concern that the design basis earthquakes may go up because 22 of new knowledge or new interpretation. But from the point 23 of view on this side of the table, I think there is another 24 aspect to the issue.
25 One reason you are doing seismic margin research uO l
l
i 18 i is because the ACRS asked for it.
2 MR. RICHARDSON: Exactly.
3 MR. SIESS: And one reason the ACRE asked for it i 4 was the sort of thing you mentioned. The other reason,
! 5 however, is not as easy to define in regulatory terms, but 6
it is Simply that the safe shutdown earthquake seems to be ,
7 a probability in the neighborhood of ten to the minus three !
g and ten to the minus four per year, and there was concern j r
9 that larger earthquakes are possible clearly. They are 10 Probable at some lower probability than ten to the minus ii three and ten to the minus four. And there was concern 12 that if we have a larger than design basis earthquake, we '
13 want to be sure we are not on the edge of a cliff and are i4 going to have a catastrophic failure.
15 Now I have to separate that from a regulatory i
16 concern because within the present regulation that can't 17 be, but under the severe accident policy presumably that is can be looked at.
4 19 MR. RICHARDSON: I think the procedures that we 20 are developing and the approach that we are developing can 2i address both of those questions, although the approach that 22 we are developing will not be able nor does it have the i
23 capability of determining the absolute margin. And, as you 24 will hear later, we have rather taken an approach of 25 specifying an analytical earthquake that we would examine O
... _ _ . ._ ---- . - _ _ _ - . - - _ _ . . - _ - - . - _ . _ - _ - _ - = . - .
19 i
i and determine whether the plant is capable of sustaining 2 that earthquake. l 3 NR. SIESS: What will come out of course, and I 4 .have heard some of this before, from the regulatory point 5 of view, if a plant was designed for .2 and some new l 6 interpretation of the data says it should be .22, this in 7 the past has required a complete review and reanalysis of a a plant.
9 I think one thing that is coming out of the
- io seismic margin studies is that the difference between .2 11 and .22 is lost in the noise and we shouldn't be all that 12 concerned about it.
l
~
13 MR. RICHARDSON: But we may be able to demonstrate 14 for that plant that it is fully capable of' sustaining .3 G.
15 NR. SIESS: Yes.
16 MR. RICHARDSON: And then not to worry then about 17 .22 G.
l 18 NR. SIESS: That might be the high confidence low i 19 probability value which is something else that may be new i
20 to people and John will talk about that.
! 21 MR. MARK: On this general area there is an agreed I
I 22 thought of the following sort? We should worry about a i
l 23 Plant being able to stand the ten to the minus four from 24 the new hasards curves, or the ten to the minus five, and 25 we would really worry if it is only capable of taking the l 1
i O
20
() i ten to the minus three earthquake, because there is a 2 bigger earthquake, as Chet said, with some lower I 3 Probability and you can't follow that curve all the way 4 through and make any sense. There should be a cutoff of
- 5 some kind, it seems to me.
6 MR. RICHARDSON: That is right. .This program is i,
7 not making that determination. We are developing the a
procedures and approach to determine what the capacity of a 9 plant is or the probability of a plant sustaining an <
io earthquake at a given level. I think it is up to the ii regulatory staff to establish for particular plants what
! 12 earthquake level that plant ought to reasonably sustain. ;
, i3 MR. SIESS: Well, Carson, in seismic PRA's i4 theoretically at least you can have a hazard curve that 15 goes out to ten to the minus eighth, a 90 Richter j earthquake, you know. And in the seismic PRA they do use a 16 l i7 probable curve. What they have usually done is have a is cutoff. They have said that some earthquakes that it
- i9 probably is not physically possible. Indian Point did this l 20 and Elon did this.
I Now there have been some studies made I think by 21 l
l 22 Livermore, sensitivity studies of the effect of putting this cutoff on the maximum earthquake, and I think that is 23 24 the sort of thing you are talking about to some limit, but 25 that is in PRA's. These guys are not really -- they are ,
O
21
) i working at the deterministic end of the response rather 2 than what the hazard is.
3 What we cannot forget is that the uncertainties in
, 4 the hazard have got to be many, many times greater than the 1 5 uncertainties in the response.
Chet, may I ask another question.
6 MR. EBERSOLE:
7 Is there any Clear interlinking of the logic of leak before a break in the seismic hazard? I have always been bothered 9 by the thesis of leak before break ever since, you know, I 10 learned my father put cornmeal in his Model T to plug the 11 cracks in it.
12 (Laughter.)
13 And I can't help but think that sort of phenomena 14 exists even in reactor systems, especially those that load is it with boron. So these things, can they sit around and 16 rest for years and then suddenly be approached by a seismic 17 event and all hell break loose?
i l 18 MR. RICHARDSON: You are raising questions that I i9 think can be partially answered. Certainly the leak before 20 break concept is coming more and more into our thinking, l
21 especially in the area of reactor coolant system piping.
22 MR. BBERSOLE: Well, in your str:3s studies do you i
23 contemplate cracked pipes that aren't leaking in the 24 seismic?
25 MR. RICHARDSON: Yes, and you will hear more about 1
0 l
l
22
() i that, which resulted in a modification to Design Criteria 4 ---
2 l
3 MR. SIESS: That is in the load combination.
4 MR. RICHARDSON: --- which is in the load 4
5 combinations work that you will be hearing from Dr. O'Brien 6 later on. But you are also touching on the subject that we 7 are just starting to address in the Office of Research, and a
that is the effects of aging and its potential and 9 implications for life extension of operating plants. So io these are areas that we are just now starting to address.
11 But we are certain conscious of the aging effects 12 throughout the plant on component fragilities, behavior of i3 pipes and to date we are basing a lot of our knowledge on
) u Past experience.
is A lot of the data that has come out of past I
16 earthquakes, the San Fernando, the Chilean and now the i7 Mexican earthquake is giving us a lot of data on some of
, is this older equipment and how it behaves in large i9 earthquakes.
20 MR. BENDER: There is a certain amount of arm 21 waving that goes on in this question of plant aging and 22 what it means to seismic fragility. Aside from stress 23 corrosion, what else is out there?
l 24 MR. RICHARDSON: In what little I have been able 25 to glean from what has come out of aging and the experience O
L
23 l 1 data base is showing that a lot of this mechanical 2 equipment and a lot of the electrical equipment that has a a lot of age to it does very well in earthquakes, and that 4 the aging process, other than its effects on materials, 5 such as corrosion and erosion type effects, that mechanical 6 equipment and electrical equipment stand up pretty well 7 over a long period of time. That is one or the pleasant s findings that are coming out of the experience data base is 9 that this old equipment is behaving very well in large io earthquakes.
11 MR. BENDER: I guess I am really trying to get a 12 little bit more physical understanding of what it is we are i3 going.to be looking for. Rotating machinery we know that u the shaft vibration may get you in trouble.
! 15 MR. RICHARDSON: Yes.
i6 MR. BENDER: But it is not a function of age. It 17 is just the bearings are wearing and they shouldn't be
- is wearing that much. There is something else out there that ig you might be concerned about. Both might be shaking loose l 20 or something like'that. '
21 MR. SHEWMON: Are you gentlemen assuming that you l 22 cre talking only about metal components?
23 MR. BENDER:_ I guess I am trying to get a little 24 bit better understanding of when we talk about seismic 25 margin what are we looking for?
O
24
() i NR. SIESS: Yes, but you got off into aging.
Are we looking for just the 2 MR. BENDER:
3 structural response or are we looking for the behavior of 4 certain. pieces of hardware?
5 MR. SIESS: No, functional.
i 6 MR. BENDER: And if so, what are the hardware 7 questions we are trying to address? That is the point a
which I was trying to get to.
9 MR. SIESS: Gentlemen, I think what we ought to do io is let the staff go ahead and discuss discuss some of these it programs, and then when we finish, at a certain stage we 12 will go back and talk about the implications, because we e
i3 are asking questions that I think will be partially a answered in what is presented and maybe more will arise. I i3 think this has been a helpful discussion. But at the end 16 of each of these, although it is not on the agenda, I am I going to take some time to open it up to see what people i7 la think of the program as it has been described.
i9 Jim, one thing I would like for you to single out, 20 and maybe you have already got it, but a number of the
{ 21 Programa that we have here have clear-cut regulatory 22 implications and at least proposed regulatory 23 i"Plementations.
24 MR. RICHARDSON: Yes.
25 MR. SIESS: And as you well know, the NRC research (2) l 1
' 25
) i program has been strongly criticized as frequently not 2 leading to changes in the regulations. And I think here at 3 least some of these programs have some fairly clear-cut 4 changes in regulations proposed or already adopted, and I l 5 would like to have you emphasize those or summarize them. ,
l 6 MR. RICHARDSON: Yes, I think you are right. We 7 baFe had that stone Cast toward us and I think it has been s scmewhat unfair in that I think we have had and will 9 Continue to have an impact on the regulatory impacts.
~
io In fact, on each of the programs that will be
! 11 presented today, one of the key elements is the regulatory i
12 use, and that will be emphasized by each speaker on each of i3 the elements.
14 MR. SIESS: I do think it is important.
15 MR. RICHARDSON: Yes, and that will be emphasised 16 throughout the day.
17 If I can go on with.the seismic program.
I is Basically our objective is to validate the methods that we i
19 have developed over the past. years. We are now out of the 20 Ph ase of developing the theoretical methods and we are now 2- in a position to start to validate some of these methods.
22 Our aim is to validate the more complex methods 23 that can be used to benchmark some simpler methods that 24 would be used in a day-to-day operation. We will be 25 conducting experiments on large integrated facilities m
26 l
() i throughout the world.
facilities, one in Germany, the HDR facility that in early We have centered on two primary 2
3 1986 will be vibrated with an eccentric shaker and the 4
whole building vibrated, and we will be monitoring that
! 5 very closely and using that to validate several codes.
6 MR. SHEWMON: Do these things always go up to 7
below yield at any point in any component, or is it to get a
characteristic frequencies or damping values?
9 MR. RICHARDSON: All of the above, plus we hope to io get into the inelastic region in the components.
11 Fortunately, this is a decommissioned reactor and the i
12 ambition is to within the parameters of safety to do as i3 much damage as we reasonably can.
34 MR. SIESS: Because there is something next door.
MR. RICHARDSON: Yes, there.is a reactor right 15 a next door that makes the safety authorities very, very i7 nervous. We are also negotiating with the Japanese to
} is participate on a particular test on the Tadotsu shaker table, which is the world's largest shaker table where they l 39 i 20 are in the midst of an experimental program of testing 21 large-scale PWR and BWR components and systems.
22 We have identified one particular system that we 23 are interested in and we are negotiating with the Japanese 24 to participate in that test. Those negotiations are still 25 going on and at this time I cannot give you a definite yes O
. - _ _.- - _-.- .- ,., , _ ,___7_ - . _ _ _ .-,-.__--.y .,__-m___,_,.,._.___,_,---.,_____-____,,-.__.-.._.w..-m_._,- _ . . , , , _ . _ _ . _ _ - . ,
27 k) i or no as to whether we are going to participate.
You said large, scale. What is meant by 2 MR. MARK:
3 that, one/ tenth linear or something?
4 MR. RICHARDSON: One-half to one-third.
5 MR. EBERSOLE: Can you tell me what that one 1
6 system was?
7 MR. RICHARDSON: It was a PWR loop with a reactor ,
a vessel, steam generator pump, cold leg, hot leg and 9 Crossover leg.
10 MR. EBERSOLE: It wasn't an aux feedwater system?
i 11 MA. RICBARDSON: No. All of the systems that they l 12 are testing are primary loop systems, both BWR and PWR, l
33 plus containment vessels. They tested a very large steel
- ) 14 containrent, and they are going to test a smaller concrete i
l 15 vessel.
l 16 MR. EBERSOLE: That is interesting because that is 17 not where the heat has to go. .
4
, is MR. RICHARDSON: Well, you must understand, the i 19 overall objectives of the Japanese program is not a ;
- 20 reseerch program. Theirs is what they classify as proving l 21 tests, and they are not going beyond the design levels, and i
! 22 they are primarily what I would call public relations type
! 23 tests to demonstrate to their public that these reactors '
l 24 are fully capable of sustaining design basis earthquakes.
l 25 MR. BENDER: Your purpose in these tests though is c l
1
28 l
i to validate some computational methods?
2 MR. RICHARDSON: Yes, and the tests that we are
! 3 interested in would extend that test beyond the design 4 basis up to the limits of the table, and we would induce
]
5 large inelastic deformations in the pipe. That is the crux i 6 of the negotiations and those have not been concluded yet, f 7 MR. SIESS: They want to show it won't fail, and f i
8 you would like to carry it to failure.
9 MR. RICHARDSON: Yes, although we have gotten over ;
i 10 that hurdle and the Japanese now agree that that would be a ;
i n good idea to do that. The problem now is one of finances.
12 As I mentioned before, we are developing ;
] i3 procedures to estimate the design margins and we are i4 concentrating in the piping area toward a goal of i 13 eliminating snubbers and pipe supports and going to more (
16 flexible systems where possible. i i
! i7 You will be hearing more about our piping program -
i la later this afternoon I believe. ;
1
! i, We are currently assembling a large fragility data I
20 base, and based on the results of the data that exists, we l
l 21 will define what specific fragility tests we must conduct !
22 ourselves. We anticipate that those tests will start later i
i 23 this year.
l 24 And in the margins area we want to be able to I I
25 demonstrate the margins procedures with at least one, if l
. O l l
i e
-.www,w--w,. g --~ ~ --,ww,-,--- -,- w-e, .m,r-
29
) i not two, trial plant reviews on a PWR and a BWR.
2 We are closely cooperating with the Electric Power 3 Research Institute, the industry in general, principally 4 through the SQUG effort, and we have international 5 cooperation in Germany and Japan, as I have mentioned.
6 The next two viewgraphs go into quite a bit more 7 detail than what I have stated, and rather than my just a reading these viewgraphs, these were viewgraphs that were 9 incidentally prepared for an OMB briefing last week for our to '87 budget, and I thought it might be helpful for you to u see them. It just goes into a little more detail on the 12 basic issues, objectives, what the elements of the program i3 are and how they are integrated with other programs and in u general how the programs will be used.
15 You will see much more detail on this on a project 16 by project description later today. So I think it would be u somewhat redundant in my reading these viewgraphs.
is The next viewgraph shows a schematic of how we 19 have integrated our seismic program with the earth sciences 20 Program and how some of the related projects, both foreign 2i and domestic, on the left side and the regulatory l 22 applications on the right side.
I i 23 MR. SIESS: Jim, you have got a lot of other
! 24 groups working on this, German, Japan and industry and so l
25 forth.
O l
l
1 l 30 l i NR. RICHARDSON: Yes. l 2 MR. SIESS: In terms of those other activities l 3
that you have a piece of or detailed knowledge of, what i
4 proportion would you say the NRC research is of the total
. 5 research that you have access to?
I 6 I would guess, and this is a pure guess because I 7 am have made some estimates on the value of our '
! a experiments, but all of the other research probably we have i
- 9 a leverage of about four to one. About four times our 10 budget in total is being spent in this area.
l 11 MR. RICHARDSON: I think of all of your learning i
12 you are paying for about 20 to 25 percent of it.
i I
33 MR. RICHARDSON: Yes. But that does not include 34 access to some very expensive integrated test facilities j is and the capitalization of those to give us much more 16 leverage, somewhere in the area of about three to four l
! i7 cents on the dollar. There we are getting quite a bit of ,
1 is leverage and that would not be included in this four to one
- i, leverage that I just spoke of.
I The next viewgraph is just a matrix to show all of !
i 20 i
21 the other programs that ---
22 MR. SIESS: Just to be sure we are on the same one j
, 23 each time, why don't you just hold it up.
! 24 MR. RICHARDSON: It is a program slide called 25 seismic integration. It is just a matrix to show all of I
O i
31 i our contractors, coupled with other sources that are 2 involved in research that relates to that project and the 1
3 relationship, whether it is a joint program or in fact 4 complementary, and that is just merely coordinating to make 5 sure we are not duplicating each other. ,
6 But we do have quite a bit of interaction with 7 EPRI, foreign agencies in Japan, Taiwan, France and the a United Kingdom and industry sources such as the SQUG effort 9 as well as the large industry effort over the past years in
! to equipment qualification testing which in many cases turned i
11 out to be good fragility testing. They went well beyond 12 the design limits and in many cases failed the components, 13 and we are getting quite a bit of fragility data that we in
! 14 fact were pleasantly surprised was there and we have gotten is access to it.
MR. SHEWMON: Sometimes these failures, at least I f 16 l 17 remember one diesel, or something like this, an oil line is failed, and that was what limited the fragility and f 19 presumably somebody could brace the oil line and that would 20 then change it to a different point. Now when you say l
i 21 fragility data of the equipment -- okay, go ahead. l 22 MR. RICHARDSON: Well, yes. In fact, you have 23 brought out one of the points that in our seismic margin 24 study we discovered, and that was that the fragility of the 25 diesel was not so much the diesel itself, but the
32 i peripheral components in support of that diesel were the 2 more fragile part of that system.
l 3
But I think the question you are getting after is ;
4 if we were to fix than, then what is the fragility. And in 5
some cases we may have data on the next layer and in other i 6 cases we don't. We have merely what the weakest link is, i 7 I might say by fragility I am talking about a
functional fragility and not necessarily structural 9 fragility, but when does the component lose its ability to
! to perform its safety function. ,
n MR. SHEWMON: Yes.
12 MR. BENDER: Jim, it seems to me this is the right i3 place to ask about this question. There are a lot of items
, i4 listed on this table and it is hard to get a sense of Proportion about which ones are most important and which
~
is
- 16 ones involve the most expenditure of effort and which ones 37 are the ones that have the least information about them.
l 18 Now all three of those things, it seems to me, have to be i, dealt with at some time or another.
20 Have you at any time every tried to taho the whole 21 seismic research business and not just NRC's part 6f_it and
( 22 Put it all together so you can see what the whole picture I 23 is likely to be?
i 24 MR. RICHARDSON: We have tried to do that in our 25 seismic plan that we have developed and that I referred to O
33 h) i earlier, which was to try to give the reader a sense of 2 what the whole picture looks like. I am not sure we have l 3 been totally successful in that, and I am not certain that l
}
l 4 we understand the whole picture a hundred percent, but I 5 think we have got a pretty good idea of what is going on
! 6 and where the real weaknesses are.
7 I hope that by the end of today ---
8 MR. BENDER: It is some other plant?
9 MR. RICHARDSON: It is another plant. It is to called the seismic safety research program plant and Dr.
11 Costello is holding up a copy of it.
12 Al, I don't know. Did the consultants get a copy 13 of that?
O, 14 MR. IGNE: I don't think so.
I 15 MR. BENDER: I don't know that some of us need 16 it. The fact that I don't have it doesn't necessarily 17 bother me any. It is just a matter of knowing whether it is exists. Thank you.
19 MR. EBERSOLE: May I ask this question. In the 20 instrument world there is the matter of the effective 21 instrument failures on safety systems, and in the 22 electrical world there is GDC-17 about the design of non-23 safety apparatus and the electrical high impacts on the
- 24 safety. Where is there in here a look taken at non-seismic 25 equipment and its operational effects on the safety 0
34 i equipment?
2 MR. RICHARDSON: In our seismic design margins 3 Program we really don't differentiate between seismic and 4 non-seismic.
5 MR. EBERSOLE: On, you looked at the whole thing?
6 MR. RICHARDSON: Ne looked at the whole plant.
7 MR. EBERSOLE: Okay. I didn't know that.
s MR. RICHARDSON: And the interactions between
, 9 safety and non-safety systems is accounted for or examined, io and it is examined primarily by looking at the PRA's that it have.been performed to date and that has guided ---
12 MR. EBERSOLE: Lots of PRA's don't look at this topic. They look at the safety systems. They wouldn't 33 i4 find a domestic water tank on the roof.
MR. RICHARDSON: Oh, I think they.do. I think Dr.
15 i Budnitz will able to address that in more detail, but I 16 i7 think I would disagree with you. I think you will find
?
, is that the seismic PRA's that have been performed to date, l
i9 which there have been 12 to 15 that have been performed, do 20 look at non-safe category one safety tanks.
21 MR. MARK: They look at those tanks, but usually 22 if it is non-seismic qualified they say it fails.
l l
24 MR. RICHARDSON: No. The PRA is an estimate of 25 the fragility of those tanks.
O l
l
. . - , , ,...-.,-,,,.y,
35 i MR. MARK: The PRA doesn't really go at things and
)
2 get real numbers on fragility. It puts the numbers in.
3 MR. SIESS: Well, they do the best they can.
4 MR. RICHARDSON: That is one reason that you find 5 that, in my judgment, many of the seismic PRA's are, 6 conservative in giving you probably lower failure 7 probabilities than what reality is because of some of these a conservative assumptions, and an example is the crib house 9 failure at the Zion plant when it was discovered that the io crib house roof may in fact fail, and when it fails they
- 11 assume that it wiped out all six pumps housed in that, when 12 in fact that would not be the case.
I 33 MR. MARKr. Yes, that is conservative all right.
14 MR. SIESS: But, Carson, they do try to quantify.
is For example, an earthquake will cause a loss of.offsite 16 Power because of the failure of ceramic insulators, and 17 that isn't an arbitrary assumption. They will have data on
- is what will fail ---
i 19 MR. MARK: That I understand, and I understand 20 that they can tell you if the failure of this important.
21 MR. RICHARDSON: Yes.
, 22 MR. SIESS: And you'can estimate the probability.
23 MR. RICHARDSON: You will find in the description 24 of the seismic margins procedure that has been developed l 25 that in fact tanks have been identified as one of the
- O l l
- l l
l 1
36 i
() i potential weak links, and the person or persons conducting a review is directed to look at the support of those tanks 2
3 and perhaps do a simple analysis of the support of those 4 tanks to make sure that it is adequate for the design level 3 that is being looked at. i 6 MR. MARK: Thank you.
7 MR. SIESS: Jim, the question Mike Bender asked a
really can be answered by the kind of study that was done 9 by your s3.a_;c margins panel looking at PRA's. It can to tell you what is important, it can tell you where the ii biggest uncertainties are, and that is'certainly what they 12 have done. It doesn't tell you where you are spending your i
money the most.
()
33 i4 I think one question is, let's say, that we look
{ i3 at the seismic margin, the expert panel, and believe what i
d 16 they have come with. It tells us a lot about what we need 17 to do work on.
is MR. RICHARDSON: Yes.
19 MR. SIESS: It at least suggests that there.are
, 20 some areas where learning more isn't important and some 1
1 21 areas where learning more may be impossible.
l l 22 MR. RICHARDSON: Yes. And in fact we have 23 specifically asked the panel to make recommendations to us 24 on what research needs to be done and where the holes are, 25 and that coupled with the overall insights that are being O
t 37 l
f) i gleaned from this' study I think will help us direct and 2 perhaps redirect where appropriate our resources.
3 MR. SIESS: I was going to say that another i 4 question you can ask them is what research are you doing 5 that you really don't need to be doing.
6 MR. RICHARDSON: Yes.
7 MR. SIESS: Now, again, you see, some of your
- a research is driven ~ strictly by safety requirements like the 9 PRA which really doesn't, you know, the expert panel looking at PRA's and not really relating it to design bases io 11 or regulations. It is looking at actual behavior on a 12 probabilistic basis.
i3 Now some of your research is driven by needs from 14 NRR from a strictly regulatory point of view. And in case is you didn't get it, I am making a distinction between 16 regulation and safety.
17 MR. RICHARDSON: Yes, we recognize that I is distinction also.
19 MR. SIESS: I mean they might tell you that you 20 are doing some things because somebody wants them answered, 21 but they really don't need to be answered. It is g not 4
22 important.
23 MR. RICHARDSON: And we are trying our best to l
24 identify those areas and interact'with the staff and where 1
25 appropriate disagree with them and get the issue resolved O
__.-___...m__,, . . _ _ _ . , _ . __
% _ _ ,_ _ ms,__ __ , ,. . .s..
38 i before we start spending money on it.
2 MR. SIESS: Now you do have NRR liaison on the 3 expert panel study.
4 MR. RICHARDSON: Oh, yes. Jim Knight and I are co-5 chairing the working group and there is substantial support 6 and participation by NRR on that working group.
7 If I can now turn very briefly to our containment a program, which is the other leg of our research that we 9 will be talking about today and tomorrow.
10 MR. SIESS: Jim, the way the discussion has been ii going, would we be better off going into the details on the 12 seismic margins?
MR. RICHARDSON: Yes, and we can return then to a 13 i4 brief overview of the containment integrity, and I think 15 that would make ---
16 MR. SIESS: We can separate those completely then.
17 MR. RICHARDSON: That would make good sense.
, 18 MR. SIESS: It will louse up the sequence on the t 39 viewgraphs, but we can probably ---
20 MR. RICHARDSON: I think we very easily return to 21 a very quick overview.
i 22 MR. SIESS: Okay. I think we ought to take one 23 subject completely and then try to tie it up and go on with 24 the next one.
l MR. RICHARDSON: Okay. With that in mind then, I 25 O
i 39
()
I i would like to bring on Mr. Dan Guzy, who is the Program 2 Manager for our seismic margins research program, and with ,
1 him will be Dr. Budnitz, who is chairing our expert panel 3 l l
4 in support of this program.
- 5 Dan I think wants to use the podium with 6 viewgraphs.
7 MR. BENDER: Can I ask again when the right time a is going to be to try to understand what hardware we have 9 concern about?
I 10 MR. SIESS: During and after the detailed 11 presentations. If you don't get it from that, then
- 12 interrupt.
i3 MR. GUEY: I am going to lead off with some of the 14 budgetary and planning discussion of the program as a is whole, and then as far as the approach and the techniques 16 that are involved with the seismic margin review, Dr.
17 Budnitz will come in later on.
l is (Slide.)
ig Essentially I will make a couple of corrections to
{
20 the slides as I go through. One of them is that cur '87 l
21 budget is 900 K, which is consistent with what Jim gave j 22 earlier.
23 Right now this margin study is all under one fin.
24 The fin is through Livermore, but the work as a whole as 25 been primarily done through the expert panel, which Dr.
I
' ).
i
. ~.__ - -_-. _
40 i Budnitz chairs. And the products and the development to 2
date has been mainly an expert panel product.
3 (Slide.)
4 I think you discussed the issues well enough. I 5 just want to talk a little about some of the smaller 6 objectives or some of the sub-objectives of the program. l l
7 The objective as we have laid out now is to 8 develop a means to evaluate the margins in nuclear power 9 . plants, or to be more specific, a means of assessing in whether the margin is adequate and to the degree that it n gives us confidence in that measure.
12 The first step in doing this was to assess the 13 existing margins information, which is to look at the PRA's a that have been published to date and reviewed by the NRC 15 and the panel, to look at the recent experience data, and 16 in thi8 area the Panel has used the SQUG data to a large i7 extent, and also tests that have evolved in the last few is years, such as the ANCO pipe test and other new physical i9 tests that have been performed.
., 20 This phase has been completed. Having reviewed 21 this, the next step was to develop margins review l 22 Procedures. In the first step in doing this, an overall 23 approach was formulated by the panel and has been 24 published. It has been published in this NUREG we just 25 mentioned, CR-4334.
O
41 1 The panel now is in the process of fleshing out 2 what was given in that report as far as an outline of the-4 3 margins review.
4 MR. SIESS: Now margins review means for ,
5 individual plants?
6 MR. GUEY: For an individual plant and for 7 specific earthquake levels. That is the way it has been a broken down.
9 Another objective of this is to identify those io areas where more information is needed such as where shall ;
11 we do more to improve our reviews, and the panel has !
12 identified a number of areas such as relay chatter, 13 fragilities in general, design and construction errors and u other areas where the panel feels that it would be is V4rthwhile to spend a little more time and improve what we 16 know now about power plants.
17 MR. SHEWMON: What units do you usually express is margins in?
19 MR. GUZY: The G levels, if that is what you are
. 20 talking about, with this being tied back to peak ground 21 acceleration.
22 MR. SIESS: Don, your identify information needs.
23 Now the expert panel really covered the issue of seismic 24 margins, which is the entire issue in terms of seismic
- 25 design of nuclear plants. Can you relate their information
- 0 I
l
. i 42
() i needs to those portions of the research program that deal with seismic responses?
2 3 MR. GUEY: Well, one of the areas, as I mentioned, 4 was fragilities, and we have a program in the fragilities 5 area. Relay chatter, it is not under our branch, but there j i
6 is a project under DARO looking at relay chatter and its 7 effect. Design and construction error is a hard one.
8 There is some work being done again through DARO in this e area. I think it is something that we will have to try and i
get our hands around later on in the out years hopefully..
io ,
11 MR. SIESS: Are there any things that you are 12 doing that you cannot relate to an information need that 33 came out of that report? Do they indicate soil structure 34 interaction as being an information need?
15 MR. RICHARDSON: I would say primarily the area i 16 that we are doing that was not in that information area is i7 in our piping research.
is MR. GUEY: Bob wants to say something.
{
i9 MR. BUDNITE: I have a comment about that. The 20 idea of trying to ascertain the margin of a specific plant 21 is a good one, but it is different from the idea of trying 22 to understand the actual behavior of a plant in a given 23 earthquake. And because of that, the intellectual needs, l 24 that is the needs for various kinds of knowledges is quite l
l 25 different.
O L
,- ~ ,, - . - - . . - . . - - - . . , . - - . . , . - . . . . - - - .- _ - - - - , . - - . ,
i .
43 l
() i 2
It would be a mistake to think that because you knew that the margin of a plant was such and such, that is 3 that you had high confidence that it would survive .35 G, 4 let's say, and that all that you wanted to know about that 5 plant in earthquakes was understood. That is just not so.
6 MR. SIESS: Then maybe we ought to have another 7 panel of the same general level of Competence to look at a the other aspect to try to give you a tie to your research 9 framework.
10 MR. RICHARDSON: We were hoping that maybe the 11 ACRS was that panel.
! 12 MR. SIESS: Don't count on it.
i3 (Laughter.)
f 14 MR. GUEY: Maybe we should also mention there is a l is panel, which I guess many of the members of this panel is a -
16 subset of, which is the validation panel. They are also '
17 looking at PRA's and the shortcomings and the types of is research that would help us improve PRA's, the seismic 1
19 evaluations in general.
20 MR. SIESS: The thing that bothers me is that the i
21 research you need to improve PRA's may not really be the 22 research you need to improve safety.
23 MR. RICHARDSON: And I think that our piping 24 research is a good case in point. It has not been 25 identified in the seismic margin panel as an information need for seismic margins, and it has not been identified 0
l_ _ - - - - _ _ _ - _ _ _ _ _ _ _ _ .
44
() 2 i by the validation panel, but it is information needed to improve safety of piping systems by reducing the number of 3
anubbers that are in piping systems, to improve the ASME 4 code by redefining the failure mechanism of piping, and I 5
think that is a good example of that type of research.
6 MR. SIESS: There are a lot of things we don't 7
know about the seismic behavior or the seismic response of a a nuclear plant. But the question is how many of the 9 things we don't do we need to know in order to be able to io decide whether a plant is safe enough now or to be able to
, ii design a plant that will be safe enough. We can't know 12 everything about everything.
13 MR. BUDNITE: There is another point about that, a too, and that is that although PRA is a wonderful tool for is guiding what you need to know, you certainly can't rely on a it entirely in a research program.
i7 I remember, and it would have been about 1979 when is the Lewis report came out and Saul Levine was still the 39 Director, and people in the research program began to i 20 realize how important PRA was going to be.
21 Chuck Serpan came running into Saul's office, and 22 I was Saul's Deputy at that time, and said but PRA can't i
23 help me plan my program, and Chuck is involved with the 24 Primary vessel integrity, et. cetera, and of course he is 25 completely right.
O
45 i No PRA has ever found that the primary vessel k
2 integrity contributes to risk, and we are thankful that it 3 doesn't. But there isn't anybody I know that wants to 4 dispense with that program because it doesn't contribute.
- 5 So you have to be very careful --- '
i 6 MR. SIESS: Why?
l 7 MR. BUDNITz: Because -- well, you know why, too --
! a because we need the program to assure that it continues to 9 be so in. ten words or less.
10 Now that is just to be sure to re-emphasize, and 11 it is true here, too, we are talking about piping, PRA's 12 are useful for one important part of our understanding, but i3 they cannot guide the whole program. It is just there is i4 more to it.
- 15 MR. SIESS
- A good point.
16 MR. GUzY: Okay. A little more about where we are 17 going and where we are. In '85, this year, essentially the is first major milestone was the development of a seismic 19 . design margins program plan. The word " experiment" should 20 be scratched in this. It is a typo I guess.
- 21 This one itself was a major accomplishment to get 22 the NRC working group to agree on a direction and to 23 ess'ntially e have the panel be focused also in a direction 24 as far as what we are going to lead to in our seismic 25 margins work.
46 i (Slide.)
2 The next step, once this plan was laid out, was 3
essentially to do the assessment of existing data and to 4 develop an approach to doing seismic margin review. One of 5
the significant things in here *. hat the panel with the 6 working group's approval said, and as Dr. Siess had ;
7 mentioned, that the way they get their hands round the
- a problem was to focus on specific plant reviews for specific 9 target levels, and also the development of what we call the
- io high confidence of low probability of failure approach, or
- in the HCLPF approach as it has been called, as opposed to 12 looking at medium failure values.
So the report that has been published outlines the 4
13 i4 lessons learned, you might say, from the PRA's and from the 15 Physic.a1 data. It discusses the philosophy behind this 16 BCLPF approach and also the screening approach where you i7 use your systems information to focus on systems that are 18 really important to plant core melt, and to develop this, i i
i9 the rough outline with the review procedures would be.
NR. SIESS: How do you pronounce that HCLPF7 l 20 MR. GU5Y: HCLPF. It is either that or saying 21
! 22 high confidence of low probability of failure.
23 MR. SIESS: You should always put some vowels in a 24 acronym. It helps.
I 4
25 MR. GUEY: Well, that is the expert panel's words.
l O
i
47 Okay, the activities right now is still the expert
) 1 2 panel is doing the majority of the thinking, you might say, I 3 and the work that is in progress right now is these more 4 detailed guidelines for more step-by-step recommendations 5 for how to do a margins review.
6 I will just say this is a little misleading. The 7 approach now is to develop trial guidelines. So guidelines a that essentially will be used for a trial plant reviews in 9 a research Context, and then to finalise them once those io reviews have been finished.
11 MR. RICHARDSON: I might add at this point, the 12 success of conducting those trial reviews is totally l i3 depending upon our acquiring a cooperative utility because
) u we have no licensing basis to force a utility to engage in is a margins study.
16 So we are in the process of trying to get a utility to voluntarily subject their plant to this study. !
17 18 MR. SIESS: If they already have a seismic PRA, I
19 there is no need for this; is that correct?
20 MR. RICHARDSON: We really don't want to do that 21 because we would like to have a plant that does not have a 22 PRA. Ideally we would like to have a plant that does not l
23 currently have a PRA, but will have one in the near future !
24 that we can benchmark our results with to assess the 25 accuracy.
O V
. =. .-
l 4
48
() i MR. SIESS: These reviews are essentially a
! 2 substitute for a seismic PRA? i
! 3 MR. RICHARDSON: No, no, they are not a
! 4 substitute. The results will not be risk numbers.
i MR. SIESS: No, but they will tell you whether you l f 5 6 should worry about it. j i
i.
l 7 MR. RICHARDSON: Perhaps, but I think it would be i e a mistake.to advertise this as a simplified risk approach 9 or that risk numbers will result. There are going to be j io margin numbers.
11 MR. SIESS: Well, I don't get that much comfort 12 out of seismic risk numbers because I think the I uncertainties in a seismic hazard dominate the thing so i3 j
i4 auch that it doesn't help.
l Mike.
). is j 16 MR. BENDER: Well, the SEP's might be a framework j i7 to work in.
18 MR. GUEY: There are a lot of similarities between i, the seismic margin reviews and what is being recommended in ,
20 the SEP's. You might say it is a new approach.
! 21 MR. BENDER: So if you are looking at something to ,
j 22 try to approach it, it seems to me that is the better place
} 23 to look.
l 24 MR. SIESS: But if you had done a seismic PRA, there would be no need for this sort of thing. Am I j 25 lO i
i I
I l
49 i correct?
2 MR. RICHARDSON: I suppose, yes. If a PRA were
, I 3 completed, there would be no need to do this study.
4 MR. GUIY: I think that is true. -There is a i 5 slight difference between a PRA and this, and I think the
- 6 next report that comes from the panel will explain this 7 very well.
8 As is being outlined now, the approach of the j 9 margin review, we will spend more time looking at to fragilities. We will essentially do a better job in 11 looking at fragilities. Of course, it doesn't look at 12 hasard and it is a more focused look at the failure of l
i3 components and systems than say in a PRA.
- 14 MR. RICHARDSON
- And I think in that sense it is a is better look than a PRA, although you are using the PRA 16 approach to focus is on a few select systems. Once you 17 have selected those systems, then this approach gets into i is more detail than a PRA does.
l 19 MR. GUEY: Okay, back to scope accomplishments.
I 20 This next panel report has been -- I guess there was a 21 meeting last week with the panel to finalize their comments 22 on it. There will be a version that will now be circulated 23 within a matter of weeks to the NRC working group, and the 24 NRC working group will provide its cosmients to the panel at 25 the end of October. And hopefully after that it will be
50 i finalized as a report. It probably won't be a NUREG. This !
2 is probably incorrect.
! 3 It will be finalized at least in this stage to be 1 4 a reference for these trial plant reviews or review. I say l 5 review or reviews because the working group now is tending i 6 to recommend that instead of doing two plant reviews or two l 7 PWR plant reviews at this stage, that they would prefer to j g look at say two levels of seismic review, two target 9 levels.
1 10 And being what the budget is, we assume that we it would be best off doing one plant with two levels of j i2 review, plus plant selection is still a question.
33 I should mention that there are -- between the f i4 panel and the working group, they have been narrowed down i
1 15 to approximately five plants which we are exploring -- five l 16 P l ants that were good candidates for this type of review i7 based on their vintage, whether they have a PRA done, which !
is would more or less exclude them, and ---
ig MR. SIESS: What do you mean by target levels, two 20 target levels?
l 21 MR. GUEY: Target levels, the way I guess the i
l 22 approach is now is that you would essentially review a l
1 23 Plant for a specific target level and determine --- .
l 24 MR. SIESS: A G 1evel?
25 MR. GUEY: G 1evel.
O
51 f'
i MR. SIESS: Something 10 percent over its design
(_ .
2 basist is that a level? Is that what you mean, and another 3 one for 50 percent over, a double?
4 MR. GUEY: Right, you could talk in those terms.
- 5. Basically the levels now -- it was convenient to pick i 6 levels that were say .3 G, you know,- SSE to .3 G and then 7 say the neXt level would be .3 to .5 G, a specific value in a those ranges 9 MR. SIESS: Oh, using their levels, the panel's 10 levels?
11 MR. GUEY: Yes.
12 MR. SIESS: Because I can visualize two kinds of i3 regulatory cases. One is the kind we have seen with the i
) 14 hazard curves where somebody has gone up from .2 to .22.
l l 15 The other I can visualize is as a policy matter Charleston 16 gets cut loose from Charleston and starts drif ting up and 17 down the Atlantic reaboard, and we have got .2 going to .4 i.
! 18 and .2 going to .5.
l 19 MR. RICHARDSON: That is precisely why we like to 20 do two different levels at one plant to see what the 21 Problems are at a high level and an intermediate level.
l 22 MR. SIESS: The other kind of level is the ACRS's 23 larger but less probable earthquake. Now those are our l 24 words. They escalated from the first case, you know, of 25 the slightly increased one, and I don't know what it i
(0 l
l
i 52 i means. But.certainly if I have got a plant that was 2 designed for .2 and you can show me that there is a high 3 probability that it will survive .4, then I have moved the t
4 cliff far enough out that maybe I am comfortable at something larger but less probable. I don't know.
5 l 6
What you are thinking about is these that you !
7 might be able to justify or might have to use from a a
regulatory point of view.
! 9 NR. RICHARDSON: Although we do want to hopefully 1
io in this candidate review plan pick a fairly high level to ,
ii start addressing the ACRS type question and a lower level
~
12 that might address the typical regulatory question.
i3 Wnetner we can get a utiluy w an fact suosit a plant to 1
i4 that high a level of review, I don't know.
15 MR. SIESS: For a lot the SEP plants any earthquake is a high level because they weren't designed l 16 17 for them. For Big Rock of course .2 is what, four or five l 18 times what it was designed for.
19 MR. GUEY: The problem with the two levels of f
20 review is the higher you get, the more detail and the more l
l 21 expensive the review is.
, 22 The trial plant reviews should begin next year 23 once the plant is selected, and this will be primarily done i
24 by an organization to be picked, probably a national lab 25 with other consultants following the procedures that.will O
2 53 J
be laid out in these trail guidelines.
) i 2 And the panel will in a way be out of the 3 production mode and fall into more of a review mode. The l
4 panel will serve as a peer review group to monitor these 5 trial plant reviews as they go along.
6 MR. BENDER: Before you get off that point, it 7 seems to me in trying to deal with the translation of a a seismic design at one level to a seismic-design at another :
} 9 level that you need to know a few things that sort of ,
10 characterize the problem.
l 11 First, there is some threshold that you have to
- 12 consider that varies, depending upon what the initial i3 design was. When people started out with a desig.1 _.sel of ,
O 14 .1 G and assumed that the uniform building code took care is of everything, you don't really know whether you had any 16 margin at all when you got beyond .1 G. When you start 7
\
! 17 with something that is at .2 or a little above that, and f is then you double it, there is a different threshold that may 19 cause you concern.
i 20 Now if what we are really worrying about is 1
f 21 whether what is going to happen if we translate the 22 Charleston earthquake from Charleston to the entire eastern 23 seaboard, it seems to me that you need to be thinking some 24
~
about how that threshold is influenced by that change. Are 25 we getting that kind of flavor into this study? l O ,
l
54
.i MR. GU5Y: It is a flavor that older plants with different design criteria will have to be looked at I guess l
) 2 3 in more detail than newer plants. Essentially there are l 4 some caveats about the use of some of the streamlined 5
recommendations I guess that the panel is coming up with as l 6 far as doing these reviews, and if you have a very old J
l, 7
plant, then you will less of these convenient tables and I a screening devices than you would have with a newer plant.
9 But there is the systems approach which would 10 apply I guess to all plants and it would be essentially ;
ii focusing on those systems that a systems analyst says are i
) 12 important to preventing core melt.
i3 MR. RICHARDSON: I think the procedure is there to i
! i4 consider all vintage of plants. It would depend then on ,
! 15 that vintage of plants whether it falls through the screen l
or not and whether it filters through and nothing need more l 16 j u be done, or an older vintage plant may well not screen out f is and you would have to then do an analysis or a closer j i, examination of particular components within that older 4 20 plant because we can't a priori screen it out because of 21 the vintage of the plant.
! 22 MR. BENDER: Bob, do you have any comment on that 1
23 point?
l
! 24 MR. BUDNITE: No.
25 MR. GUXY In FY-87 our plans now are to have lO i
i l
1
, - - - . - , - .- n - . , - . . , , _ - . - . . , , , - ~ , , , - - - - . - _ - . , , , - - , _ - _ - . , _ - . , , - - - _ , - , , , - - -
55 7hj i completed or nearly completed the trial plant reviews and 2 to use that information as -- the lessons learned from that 3 will be used to improve the guidelines, the trial 4 guidelines as written now.
5 Now the guidelines are primarily written, at least 6 in the systems area you might say, for PWR's because that 7
is where the vast number of PRA's have been, PWR's, and a there is only one BWR in the data base now.
9 However, in the next I guess year or so there will to be more BWR's being reviewed. Then the systems part of the 11 guidelines will be updated to complete the systems aspects, 12 the system screen aspects of the BWR systems.
i3 So I guess the point should be made that the
(,)
, u fragilities -- I think the panel feels that fragilities is is pretty good for both types of plants, except for maybe some 16 BWR specific components like the containment.
17 NR. EBERSOLE: May I ask you a question. Since
- is you broke down into PWR's and BWR's, you recall that 19 subsequent to TMI-2 it was found that I think it was five 20 or ten plants didn't even have seismic or other safety 2
- qualifications on the aux feedwater systems, which are the 22 fundamental systems to get the heat out. There is no 23 other, and you have got to get it out through the secondary 24 side. Are you focusing on that particular aspect of a PWR 25 in view of its past somewhat miserable history?
1
56 i MR. GUZY: Well, it would be focused on to the 2 extent that a plant walk-through, which is more than just 3
our walk-down or walk-through, would review the design 4 document or design criteria. If the aux feed system wasn't 5 designed as a modern plant, you might say, then that would i
6 be caught and would be reviewed. So it would be something l
l 7 that would be caught in the process.
8 MR. EBERSOLE: Would its auxiliary systems that
, support it be caught as well?
f to MR. GU5Y: Yes.
1 i MR. EBERSOLE: And on the boiler, are you going 12 inside at the shutdown system, the scram system?
i3 MR. GU5Y: It looks at all the systems.
i4 As I said before, the systems aspects is what we is are getting out of the PRA's or the BWR's, and I guess our 16 planning right now would be that once the trial guidelines i7 incorporate the screening, the system screen aspects of the is BWR's, then we would go ahead and do another trial margin i, review this time on a BWR.
20 MR. BENDER: Before you get off that point, there 2 is a time relationship here that has to be considered. Did 22 the PRA's tell you that.
23 MR. GU5Y: Time as far as ---
24 MR. BENDER: The earthquake affects different 25 things at different times and for certain durations, and I O
e i
i 57 l
j i am wondering how that gets cranked in?
MR. GUEY: That is cranked basically the same way l
2 j 3 it is cranked in in the design process and in PRA's. In
. 4 other words, there are certain assumptions made about the j 5 earthquakes that do damage. I don't quite understand where I
6 you are coming from. Are you talking about a duration 7 longer than you would normally experience?
8 NR. BENDER: No. I am saying that there is a f
9 certain shaking that goes on for a short time. The relays, l
)
! 10 for example, have to work for a certain period of time and 11 then you forget about them. So the increment of time j j 12 associated with relays is one thing. The aux feedwater I i3 systems may come in at a different time and for a different t4 period of duration. And so the issue of fragility,
! is reliability and resistance to seismic events has a variable i
i 16 time aspect, and I think it is important to understand i'
17 those time relationships.
i
- is MR. EBERSOLE
- Mike, you are saying you could i, identify a delay function of some sort that says during the i
} 20 interval of shaking it would damage those functional 21 requirements anyway, but after it was over you would have 22 to work.
i i 23 MR. BENDER: That is the kind of thing you have to i
! 24 think about. l i
25 MR. EBERSOLE: Provided you put in a master delay 0 '
i
- _ _ _ ~ , . , _ _ , _ . _ - _ . . _ . _ . _ , . _ _ , _ . _ _ _ _ _ _ , _ _ _ _ _ _ _ _ _ . . , - . . _ , . _ _ . _ , _ . . _ _ _ _ _ _ . . _ _ _ . _ , _ . _ . . _ _ . . _ . _ . . , _ _ _ _ . . . - _
58 l
1 of some sort that says don't answer the calls in this l
- 2 interval.
- 3 MR. GU5Y
- I think this attempts to be addressed !
in PRA's and this attempts to be addressed here. Now there 1
4 ,
l 3 are areas where we have noticed shortcomings, such as relay '
6 Chatter and ---
! 7 MR. EBERSOLE: Well, that would be a way of 4
a putting relay chatter out of the picture by saying let it
} 9 chatter, I asi not going to use it.
1 I think it could be done in PRA's,
- 10 MR. BENDER:
l ti but I don't think it is.
l MR. GURY: There are some instances I know, such 12 k i3 as the assumption that operators can do certain things and 34 where it is attempted to be built in. This is an area --
15 at least the relay chatter is an area that is identified as i
16 one needing more study.
j i7 I guess the completion of the trial plant reviews f is for a BWR essentially, and whatever revisions comes to the procedures, will essentially end the work of the panel and f i, 20 the contractors in this stage.
21 At this point we hope that we will have a process i 22 or a procedure that is ready to turn over to licensing. Of 23 course, licensing will be reviewing as the working group is ,
i i 24 still following and monitoring all this work, and we are i 25 having their input as far as the research and the form and
!O
59 4
i content of these reviews.
- 2 There is another item which I have added under i
i 3 this topic, but it is something that we haven't really i
1 4 started to address yet. It is hoped that out of these i
l 5 margin studies we could get a start on a procedure or a j 6 Process for doing post-OBE evaluations. We haven't really 7 laid out anything formally on this yet, but we know that s
there will be a need somewhere along the line to come up 9 with some criteria for assessing whether when an OBE hits a 10 plant, you know, what should you do. If you have to shut 11 it down, what kind of assurances do you need to get the I 12 thing going again?
I l . 13 NR. SIESS: Oh, come on. You know you are going j Davis-Besse has been shut down for three
, 14 to shut it down.
I
! 15 months for an incident that didn't do anything.
!j 16 MR. GUEY: Well, maybe they should just start it t
! 17 up again.
18 MR. SIESS: And the last value impact statement I
! 19 saw on the need to look at post-OBE evaluations u de some 1
l 20 estimate as to how much faster you might be able to start i
21 it up if you had the procedures all worked out in advance, l 22 and that was a pretty cloudy crystal ball.
23 MR. RICHARDSON: I don't envision a cook book l 24 Procedure, but I think this program will give us some i 2s insights on what areas we ought to be looking at. i C
4 j 60 i
1 i MR. SIESS: Oh, just the expert margin study is 2 going to tell you what to look at when it happens, but I
_3 have a very dim view of trying to write a manual in advance 4 of what you are going to do following an OBE. I think it 5
could be.done, and I think it will be completely ignored 6
when the OBE occurs and the plant will be shut down and,
! 7 you know, everybody will look at everything for God knows i
- a how long. ,
i l 9 MR. SHEWMON: In a somewhat more positive vein ---
! to (Laughter.)
i
! ii --- you visualizing something like the walk-12 throughs that one goes through now?
j 33 MR. RICHARDSON: We think that this margin study i4 will givs us some insights on where you would start to look i3 at a plant that experienced an OBE.
i 16 MR. SIESS: They already have.
l i7 MR. RICHARDSON: Sure, and it is not much more l
18 than that. This is a very low-level effort.
j i, MR. SIESS: It is just fall-out from what you have .
20 got.
j 21 MR. RICHARDSON: Yes. That is what it is. It is 22 gathering together those insights and fall-outs from this 23 program and putting it into some logic that you might start 2.s looking at for a plant that experienced an OBE. No, I do 25 not want to advertise this as a detailed inspection manual
]
!O
61 i for post-OBE inspection.
)
2 (Slide.)
3 MR. GUEY: As my final slide, we were asked to put 4 something in regulatory use, and I think you discussed the 5 hazard and the Charleston earthquake issue, and this is one 6 of the objectives to be able to satisfy a hypothetical case 7 where a larger earthquake had to be reviewed.
8 The other item is that I think on plant specific 9 and issue specific seismic items these guidelines will be to useful for looking at the need to upgrade seismic 11 procedures. And also I think that perhaps the whole 2
12 procedures or part of the procedures can be used to find --
i3 and I used the word " weak links" which the panel does not i4 like to use any more, but it would be a way of giving is assurance that there is nothing bad that you haven't caught 16 out there.
i 17 So I think there will be plant specific items and l
i is specific issues so the margins review can be used with 19 other than just the increased hazard at the site.
20 MR. SIESS: Don, if you had known then what you 21 know now, and if you had been in a position to make the 22 decision, would you have ordered the five plants shut down?
23 MR. GUEY: Oh, no. I think most of NRR would say 24 that.
25 MR. SIESS: Good.
O
62 i MR. GUEY: Okay, that is all I have. I guess now 2 I can turn it over to Bob Budnitz.
3 MR. SIESS: First, are there any questions for Don
! 4 Guzy?
5 (No response.)
6 Then before we get to Budnitz, I am going to l 7 declare a break.
8 MR. RICHARDSON: Dan, I would ask you to prepare 9 yourself after Bob Budnitz to give us your insights on what .
10 the EPRI margin study is since Bob Kassawara is not here. ;
i 11 I think you are probably in the best position to give 12 insights on what their program is.
MR. GUEY: Okay. When I heard it, Bob heard it.
i3 i4 So can both field questions together.
15 MR. RICHARDSON: Bob Kassawara presented to the i 16 Panel last week what their program is.
17 MR. BUDNITE: And he is Mexico.
is MR. RICHARDSON: Yes, he is in Mexico.
39 (Recess taken.)
! 20 MR. SIESS: The meeting will reconvene.
21 Are you going to use the lectern, Bob?
22 MR._BUDNITE: Yes, I am going to use this.
23 MR. SIESS: Do you have Dr. Budnitz' handouts as L i 24 separate sheet?
25 (Copies were distributed.)
O
63
) 1 (Slide.)
2 Well, I had about 10 key points to make, and all 3 11 of them have already been made.
4 (Laughter.) !
5 So what I am going to try to do is fill in some of l
6 the areas where Dan Guzy's presentation and your questions :
7 was incomplete just because there wasn't time or an issue a
was raised that wasn't explored enough.
4 9 MR. SIESS: Or we didn't understand the answer. ;
10 MR. BUDNITZ: I think the most important thing for 11 you to understand is what the margin reviews that we are 12 talking about doing can do and what'they can't do.
i3 And in order to explain that well, I am going to u try to start out with a discussion of ideally what you 15 would like to know and then what we are going to review for 16' and describe the differences.
17 Now ideally what the regulatory staff says they is want to know is what is said here. Th'e regulatory need
- 9 that was expressed to the panel when we started was in 20 Plain English how large an earthquake can each plant 21 actually withstand. And everybody understands why we would 22 like to know that. I would sure like to know that plant 23 "x" can withstand a certain size earthquake, and then the 24 next larger size however parameterized is where it starts l
25 running into a compromise of its safety, and compromised
/"'
l
A 64
() i defined as it would lead to a core melt. We have used the 2 end point of core melt in a PRA sense as our figure of 3 merit.
4 Now, unfortunately, we haven't been able to devise 5 a seismic margins review approach that can answer that 6 question, and that is how large an earthquake can each 7 plant actually withstand.
g We do have in the engineering community a way to 9 answer that question, and it is the PRA. Unfortunately, l
io the PRA answers that question with important uncertainties, it large enough uncertainties so that it is very hard to get a 12 handle on precisely what the answer is.
i3 But I am going to put up, and you have this in your handout, but I am going to put up what the orderly i4 is methodology is that exists and is being practiced that 16 answers that question.
17 (Slide.)
is I think it is very important for you to understand 39 that this in my view, this PRA methodology, and that is 20 what I have here, can in principle answer the question.
21 MR. SIESS: But the uncertainties aren't due to 22 the PRA.
23 MR. BUDNITE: No. The uncertainties are not due 24 to the PRA methodology, and the methodology can taswer the 25 question. The uncertainties are due to the fact that we O
65 i don't know enough about a whole lot.of things to answer the
, 2 question.
l 3 It is also important for you to realize that the 4 PRA has a risk end point having to do with probabilities.
5 However, even if I divorce the probability part of the PRA, 6 I am still left with uncertainties on the other side. You 7 see, I don't have to know about the probability of an a earthquake of a given size to. answer the question I put up, 9 which is this question over here.
io (Slide.)
11 There are no probabilities here of whether .4 G is 12 a 10,000-year earthquake or a 40,000-year earthquake or a i3 6,000-year earthquake.
- ) u MR. EBERSOLE
- I just want to ask. You say l 15 determine capacity or fragility for each component. What
.i 16 is your confidence level that you know what the components 17 are considering auxiliary support dependencies and is dependencies on dependencies? I mean it is a very 19 complicated ---
20 MR. BUDNITE: Yes, sir. That is a crucial 21 question. The point I was making here, and I just want to 22 reiterate the point, the point I was making here is that we 23 have a methodology that in principle can answer the 24 question I posed.
25 MR. SIESS: If you knew everything.
I
i 66
() i MR. EBERSOLE:
MR. BUDNITE:
If you knew it.
All right. Now, furthermore, that 2
3 methodology, although it is a PRA methodology, is in 4 principle divorced from your knowledge of the recurrence 5 interval of earthquakes because the question that I posed 6 here doesn't speak to the recurrence intervals of 7 earthquakes.
g So even though everybody says ah, but the PRA 9 bottom line is very uncertain on account of I don't know 1
10 what the recurrence is, even though that is true, that it uncertainty is not present in the answer to this question 12 using this methodology. I just want to be sure you 13 understand that point. It is a crucial point.
ja MR. SIESS: Now, Bob, it helps me a great deal if 15 I think of the PRA,'when you say PRA here, as a l
16 probabilistic reliability assessment and not a ,
i7 probabilistic risk assessment.
is MR. BUDNITZ: Well, it is the PRA methodology, but i
39 only taken to the stage ---
20 MR. SIESS: It is a reliability methodology.
21 MR. BUDNITE: Well, in the context of seismic 22 response of plants it is in fact -- if you knew everything, 23 it is a seismic capacity methodology.
24 MR. SIESS: It is a probabilistic reliability 25 study.
f 1
67 l l
MR. BUDNITE: All right. It is a probabilistic f) i 2 way of understanding the capacity of the whole plant. Now 3 there 'is a big if, and then I will get to answering Jesse's 4 question. And, by the way, that is not the only thing, but 5 it is one of them, 6 I think you need to understand my conviction, and 7 now I think I speak for the whole panel, that if you could a do a seismic PRA without -- and never mind the hazard, but 9 the rest of it, you see, that you could, using that, get a 10 handle on the seismic capacity of the whole plant and of 11 its functions and systems and components if you could do 12 it.
13 Now you have major uncertainties all the way down 14 here, and I will answer Jesse's question as one of a set of is questions about the uncertainties in each one of these 16 things, and I am going to try to explain to you why we 17 think that the seismic margins review method we have is proposed overcomes some of the uncertainties that are ;
19 present in the PRA and gives us a good handle on.what we 20 want to know, which is the seismic margin of the plant with 21 high confidence.
22 The crucial-thing, and I think it is very 23 important for you to realize it, the crucial thing is the 24 judgment of the fragility engineers on our panel that you l 25 can arrive at a PGA level, and a peak ground acceleration '
O
68 l
() i is our figure of merit for measuring motion, although it is an imperfect one, but that you can find a PGA level at 2
3 which you can say that a particular component has a high ,
l 4 confidence of low probability of failure.
5 And that having found that, the engineers on our ,
6 Panel who work in this area feel that the determination of 7 that level is more robust than the determination of the a
mean or the median fragilities which go into the overall 9 risk stuff of. seismic PRA.
l io See, you need the whole fragility curve to fold in 11 with the hazard curve to come out with a risk number or the 12 core melt probability. And much of the uncertainty in the j 33 core melt probability and in the overall offsite 34 consequences and risk comes from the fact that you have to 15 have the whole hazard curve and you have to fold that whole 16 curve in with the fragility curve and the hazard curve, and 37 so much of the uncertainty comes from that, is Now what we have determined is that we believe you 19 Can find a point for a Component at which the engineering 20 community can state with confidence that there is a l 21 confidence that at that point at least there is a low 22 Probability of failure. j 23 And having found that point, you can use the --
24 and we call that the HCLPF point, the point where there is 25 high confidence and low probability of failure -- you can O
69 k) i use the HCLPF points for.all the components, if you can determine them all, and all is a bid word, so I am going to 2 !
i 3 have to modify that in a minute, to work out what the HCLPF 4 point is for the plant using systems combinations.
5 And that that approach is fundamentally different 6 from the PRA which needs the whole fragility curve, the 7 whole hazard curve and all that convolution to get its a answer about risk.
, 9 Now let me then go on to answer the question 10 . here. To determine the capacity or the fragility for each i
11 component is a neat trick. And if you talk to the people 12 that do this for a living, Kennedy, for example, who has i3 done more of these than anybody else in his team, and there u are others who are practicing this art as well, they will is tell you, and everybody in the community knows, that the I i6 uncertainties with which No. 2 can be done are very large, ;
17 large uncertainties, and you really can't pin it down very is well.
19 There is also the problem that a PGA as a one j 20 Parameter figure of merit measure of earthquake size is a ,
21 very inexact thing. That is, I can draw for you a hundred 22 different records that all have the same PGA, it might be i 23 .3 G, that might look very different and no one claims that 24 components will behave the same in the fragility sense, and 25 we are not talking about functional fragility, for all O ,
70 i those hundred records even if it is all at say .3 G. So 2 that is another uncertainty.
3 MR. SIESS: Now wait a minute. The respectful l 4 content of the earthquake can vary a lot for a given PGA.
5 MR. BUDNITZ: But even with the same spectral 6 Content, and of Course you are right there, too.
7 MR. SIESS: And how you have got to go with your I
e soil structure interaction and floor response spectra to 9 get up to the component.
10 MR. BUDNITZ: The response and all that stuff, ti right.
12 MR. SIES3: So all of those uncertainties are 13 built in.
14 MR. BUDNITZ: Y e s , ..; ,i..
15 MR. SIESS: The only uncertainty that you have 16 left out is the probability of that PGA at that site.
j 17 MR. BUDNITZ: But now I am going to come to the t
is difference, and you have just put your finger right on a i, key reason why the seismic margins review approach we have 20 Put together we think has more chance of getting a 21 unanimity in the community, that is of people agreeing'that 22 at least there I have got confidence, than the the PRA has 23 for the whole fragility curve. So let's keep going.
24 Here we are determining local motion and 25 response. That is a neat trick. And then you have got to O
71 1 put two and three together and determine which components 2 fail. This is the PRA. This is logic and the methodology 3 surely works. It is just a question that we all know, and 4 you don't have really enough information to do this with 5 any accuracy.
6 And then you analyze the system's dependencies and 7 you say well, gee, which things that failed then will lead 8 to a core melt, and you go down the thing.
.9 MR. SIESS: Do you have any feel for the relative 10 uncertainties in 2 and 3 compared to l? ,
11 MR. BUDNITI: Two and 3 are probably dominant
- 12 compared to 1.
13 MR. SIESS: Oh, come on.
14 MR. BUDNITE: I am not talking about the is Probability of a .3 G earthquake. I am talking about ---
16 MR. SIESS: Well, I was.
17 MR. BUDNITI: Oh, no, no. I was talking about how 18 well saying something is a .3 G PGA earthquake, how well ,
19 that parameter characterizes the assemble of .3 G 20 earthquakes.
21 MR. SIESS: Okay.
22 MR. BUDNITE: This has to do with finding a 23 Parameter to characterize your earthquake and not the 24 Probability of it. Okay?
25 MR. SIESS: Okay, I see what you mean.
I l
l
72 i MR. BUDNITE: Have you got a blackboard?
{)
2 (Mr. Budnitz proceeds to draw on the blackboard.)
3 Now I want to now show you why we arrived at that
, 4 conclusion. I don't have a viewgraph about it, but it is 5 very simple to understand, and then you will come to my
, 6 next viewgraph.
7 If this is then PGA in some increasing axis, and a
this at arbitrary units, G's, and we will take a single 9 component. But I don't mean a capacitor in a circuit. I an io talking about a pump or a valve, something like that.
it If I shook that component with this assemble of 12 100 .3 G earthquakes, let's say, and .3 G was the point at i3 which the thing was getting into trouble, and of course
- a that is low for a component like this.
i 15 MR. EBERSOLE: Don't you have to specify a a frequency you are talking about?
17 MR. BUDNITE: Yes, there are a whole lot of things you have to specify. That is the point. There are a whole
) is 39 lot of .3 G earthquakes, and this is the probability of i 20 failure.
i 21 What I am going to find is a density curve that 22 looks like this. Some of the .3 G earthquakes and, you 23 know, people understand that, there are a whole lot of 24 things going on. Some .22 G earthquakes are' going to fail 25 the thing even though .3 it is supposed to and some .35's C)
l 1
73 I
) i aren't on account of the frequency and the duration and the 2 ground motion response and all that stuff.
3 MR. EBERSOLE: It gets very messy.
4 MR. BUDNITI: All right. Now even if we knew the 5 thing very well, there would be this spectrum just because 6 .of what I have done here in parametizing my earthquake with 7 a single parameter doesn't work. It Can't Capture the a whole complexity of an earthquake in one parameter.
9 Now, unfortunately, we don't know as much as we 10 would like to know even about that. So even though this 11 might be my best judgment, there is a family of curves like 12 this that together tell me what we do know. You might i3 think, gee, this is the most probable one, and this one u only has a 10 percent chance of being right and this one maybe has another 10 percent chance of being right.
~
is 4
16 And in the classic way that Bob Kennedy started 17 parametizing these things a long time ago, he started is drawing three curves which were always in his pencil drawn i9 in a cumulative way so that you draw the curve like this 20 and this middle curve looked like this, you know, this is 21 100 percent and that is zero percent.
22 Then he would draw one curve over here and one 1 23 curve over here and this would be the 50 percent curve and i
24 this would be the five percent and that is the 95 percent.
25 MR. SIESS: That is probabilities ---
1
- - m - -, y-n, -- --
e e -
, .q,m,.-4,3,- 7..--_-,y nw ,,-g,--4 - - , - , , . , ,,,n,,,
74
() i MR. BUDNIT1: Yes, and this le going from zero to If this is the component, this medium curve 2 100 percent.
3 that goes from zero to 100 percent is just this curve here 4 that I just did the integral.
5 Now I have in my mind divorced myself from.those i I
6 Curves to go to these because I think it shows better to me 7 what is going on with our HCLPF thing. Now what-we have a
done is we said well, there is this family of curves and 9 you get a different analyst and he will draw another curve,
, io you know, and it might be steeper or shallower or whatever, ii which then shapes over here.
12 We have concluded that there is a point somewhere 1
33 here where the assemble of engineers that we are. dealing i a with, that is the practitioners in this business, can all I is agree that it kind of comes down. Maybe I didn't draw the 16 axis right, and I want to do this right. Maybe the zero is i7 down here, you know, and these curves are all up here j
is somewhere and there are certainly components like that.
39 We have identified a point that somewhere down 20 here, and it is not precisely identified mathematically,
, 21 although I can write a mathematical definition if I put log
! 22 normals on these curves ---
l 23 (Laughter.)
24 --- but I mean that is nonsense.
25 There is a point here that we call the HCLPF O
l .
I 75 i point.
i 2 MR. SIESS: And it is not a mean and it is'not a 3 median.
4 MR. BUDNITI: It is a judgment. But it is a point 5 at which we think, and now I am talking about my fragility 6 colleagues because that is not my expertise, where we think 7 the community can agree that for that component there is a I
a high confidence at that point of low probability of 9 failure.
10 Now it is not up here with these curves, you~see.
l 11 I mean up here you would say, gee, maybe half the people j 12 don't think they have high confidence, but it is also not I
i3 way down here, you see. It is wherever the community kind l
14 of settles.
15 MR. BENDER: I am going to come back to the point 16 I made a few times here and have gotten Zero answers so far 17 on it. To have any real appreciation of whether the is fragility experts are any good or not I have got to have a i
19 better understanding of what the hardware is. I have yet l
20 to see the hardware list, and I don't have any comfort ---
l 21 MR. BUDNITE: I will answer your question in a 1
minute.
22
! 23 MR. BENDER: Well, everybody has been answering it 24 in a minute, and it becomes hours after a while.
25 (Laughter.)
l0
. . . _ _ _ _ - . _ _ _ _ . . - , _ . g._, _ , _ _ , __ _ . , . _ _ , , . _ _ _ _ . _ _ _ _ _ . - - _ _ _ _ _ _ _ _ _ _ _ _ .m.._ ___. . _ . . _ . _ _ _ _ _ . . _ _ _ , , _ , _ _ . _ _ _ _ _ _ _ _ _
76
() i .MR. BUDNITI: But you see the reason why the engineering community that deals in this game is willing 2
3 and able to say something about this but has a greater
, 4 diversity of opinion about what the median is is because we 5 have a heck of a lot of experience data out there from 6 earthquakes that have happened and from testing that has '
7 been done in which we know that certain equipment, large 8
numbers of certain equipment have ridden out earthquakes of i
9 a certain size without functional failure.
10 Just to give-an example, I saw slides at the SMIRT 11 meeting in Brussels that Peter Yanoff took in Chile of I 12 guess a hundred motor control centers and all did great at
) i3 earthquakes, I don't know, at half a G or more, and they a continued to function afterwards.
15 And because if you have a hundred and they all 16 work at a level, whatever it was, you have a feeling where i7 this lower bound is. It is flat in here, but you don't is know what the shape of the curve is because in order to i9 know what the shape of the curve is you would have to have 20 a whole lot more experience data than we have got and you would have to be able to analyze it'better than we can.
f 21 l 22 So I as trying to explain why we think the 23 engineering people, the fragilities people think that they l 24 can find a lower bound -- the lower bound is a mathematical 25 term that I don't like -- but an area where they can state O
77
) I with high confidence that there is a low probability of 2 failure based on experience data, test data, analysis and 3 so on.
4 And then work with that through a methodology not 5 too different from this to come up with a HCLPF point for 6 the plant based on you have to have this component and that 7
tank and that function and that system altogether failing 8 in order to get a core melt accident sequence.
9 MR. SIESS: Now, Bob, you have mentioned it, but I 10 think it needs to be emphasized, that the best estimate 11 point, or whatever is used with the PRA, is somewhere ---
12 MR. BUDNITI: It is in here somewhere.
13 MR. SIESS: --- somewhere over in there, and that 14 difference is a multiple of ---
)
i 15 MR. BUDNITZ: A factor of two maybe or more.
16 MR. SIESS: More.
17 MR. BUDNITE: Well, it varies. In fact, I want to
- i j is make that point quantitatively.
19 MR. SIESS: We are so conditioned to think in best 20 estimate points ---
21 MR. BUDNITE: I want to make that point 22 quantitatively, and then I will answer Mike's question.
23 Quantitatively you have a hazard curve, which is, you know, 24 on another, and this is PGA. So it has got log scales on
! 25 this in the black thing that I am drawing and it is coming 1
!O
78
() i down like this and it is very uncertain what it is. And you have got to convolute that with these curves, and what 2
3 dominates that in terms of your answer has a lot to do with 4 drawing the shape of this median curve, this curve here.
5 MR. SIESS: You ought to draw that hazard curve 6 down on that bottom one.
7 MR. BUDNITE: All right. And that is why the a
answers on core melt probability, you know, are so 9 uncertain. But, darn it all, the knowledge of what this io point is or its approximation has more engineering validity it for the reasons I said, having to do with test data and 12 real experience data on earthquakes that have happened and 33 analysis and then that stuff. I mean that is just the a fundamental reason why the high confidence point has been is selected for this margin.
16 MR. EBERSOLE: Bob, let me comment on the real i7 experience data. Probably the reason that that equipment is worked is that the company learned through bitter ;
39 experience that some tough teamster was going to throw that 20 stuff off a truck and it was going to hit the ground and it 21 had to survive and couldn't be shipped back.
22 MR. BUDNITE: I of course understand that 23 completely, but of course if the stuff in the nuclear plant 24 .
was handled at least as well as that stuff in Chile, you 25 know ---
O
79 i MR. EBERSOLE: And was built the same way.
2 MR. BUDNITI: Yes, and came off the same assembly ;
3 line of General Electric and Westinghouse or whatever.
4 MR. EBERSOLE: Some genius driven by an MBA would 5 say oh, I am going to pack this in 14 feet of rubber and it
~
6 ain't ever going to be shaken.
7 MR. SIESS: And you don't think they are going to 8 Change the design just because of that?
9 MR. EBERSOLE: If shipping can be made to io compensate for fragility, 11 MR. BUDNITI: There are questions ab'out that and l
12 the applicability therefore -- and the data in Chile is an j .
i3 example. You know, that wasn't even designed for seismic
, 14 and it was an oil-fired plant and had nothing to do with, ,
is you know ---
16 MR. EBERSOLE: That is what I say, it was designed ,
17 to survive brutal treatment.
is MR. BUDNITE: Right. I have got to tell an 19 anecdote about this because it happened to me two eeeks 20 ago.
21 I just sent a son to college, and he didn't take 22 along our PC because his roommate had one. Well, his 23 roommates didn't work. So he' called up and he said send me 24 the PC. It is an Apple. I need it, word processing and 25 stuff.
1
, ,-.-e , . , , . - - - - - - - . - , - - - - - . . - , - , -- .
80 i So I sent him the PC by UPS. I called up UPS and 2 they said well, you know, you are going to have to pack it 3
so that it can survive a three-foot drop.
4 (Laughter.)
5 United Parcel Service. So I packed it so it could 6
survive a three-foot drop. I shook it, you know ---
7 MR. SIESS: To an essentially unyielding surface.
! 8 MR. BUDNIT5: Yes.
9 (Laughter.)
10 So I Put it inside a box and I had packing and I n put it inside another box just like you are suggesting, and 12 I was appreciating that it is a $2,000 machine and I 13 insured it anyway, and so I got it in there, and I brought u it over to the UPS thing, and the guy shook it and he says, 15 well, we are going to put fragile all over its six-sides so 16 it will never be dropped.
17 Laughter.)
! is But of course that is their packing criteria. So 19 I Paid him the $11 and he put it on the shelf, and just 20 then another customer came in and be put his box on top of 21 mine and the both of them fell.
22 (Laughter.)
23 MR. EBERSOLE: Was his box on top of yours?
24 MR. BUDNITE: Yes. But his box had clothes in it 25 and mine had the PC in it. So what was I going to do. 'So ,
l0
81 l
() 1 2
anyway, it went off to college.
called me and said I got it and I plugged it in and it About a week later my son l 3 works.
4 (Laughter.
5 MR. EBERSOLE: So it didn't have anything to do 6 with seismic criteria.
7 MR. BUDNITI: No, no. But the fragile didn't have a anything to do with anything, right? It had to survive the 9 three-foot drop, and it did survive it.
io (Laughter.)
11 That anecdote has nothing to do with this, but I i
- 12 thought I would share it with you.
i3 Maybe it does apply.
()
14 'Now having explained why we got to using a BCLPF, is or high confidence of low probability of failure criterion,
- 16 because the engineering community can agree and settle on 17 it, and now I will come to Mike's question, which is the i is second key thing.
19 There are three things, and the next one is the 20 systems, and I will come to yours next.
21 You see, what we believe is that we can determine 22 the HCLPF. point for any component or structure in the 23 plant, and the fragility people say that, with varying i
24 levels of uncertainty, but for any component in the plant, 25 with a couple of exceptions.
1 i
82 i The exceptions are that they are still not -- and 2 nobody really knowing what to do with relays and circuit j 3
breakers and the like, because even though we know that 4 they chatter, nobody quite knows what function means on 5
this PGA axis, and there is a whole open question there 6
which is now being addressed with research programs that 7 aren't going to have answers for a year or two at least.
8 And the other thing has to do with design and construction 9 errors.
10 You never really know when you look at a sheer ii wall that has been built 17 years ago and you have design 12 drawings, you never really can know whether the thing was constructed with the stuff left out of it. We don't pretend i3 i4 to address those things.
15 But, except for that, we believe that this 16 methodology not only can address every structure, system 17 and Component in the plant, but our report, 4334, that you is have has a list by class of every one of those things and i9 it gives you guidance on what to do to determine the HCLPF 20 Point for that. You know, it might be a motor operated 21 valve with a cantilever or it might be a vertical tank with 22 an aspect ratio of four to one that may or may not be 23 anchored properly or whatever.
MR. EBERSOLE: Let me ask you a question. When 24 25 the industry, being what it is, can't even find the scram 1 O
83 j i
breakers at Salem, why do you think you are going to find
) I 2 everything?
3 MR. BUDNITI: Oh, we don't think we can find 4 everything.
5 MR. BENDER: Well, I think you have been trying to 6 answer my question, and I think the answer is that people 7 are going to go down and make a list and make a judgment t
8 about it and they are going to say that you can rely on 9 that judgment.
j 10 I would like to rely on that judgment because that ti is the only thing we have got.
12 (Laughter.)
t 13 But there is no real numbers game associated with 14 this and there is not very much loading information you can is work from. But it seems to me that if we had some set of !
t 16 comparative data where the hardware that survived the 17 Chilean earthquake could compare with the things that we is need to know about for the PRA or anything else you want to 19 call the risk study, then you might develop some comfort 20 that that threshold that you are talking about exists.
21 MR. BUDNITE: The word " threshold" is not really 22 what ---
23 MR. BENDER: It is just the threshold. It says 24 that if you stay below this there is a very strong 25 Probability that nothing will fail in this kind of shaking i
84 i environment.
! 2 MR. SIESS: Mike, the Chile studies. we were shown l 3 an inverter that was identical with one that went into J
! 4 McGuire I believe, including the drop pan on top of it.
3 MR. BUDNITE: It came off of the same assembly l l l 6 line at Westinghouse, as I remember, the same model number.
7 MR. BENDER: I am not trying to fight with Bob ,
i about his criterion. I think he is right.
s ,
1
) 9 MR. SIESS: But they are trying to relate 1
io equipment.
) in MR. BENDER: We have more than just a few examples 12 on a systematic list that you can compare.
l MR. BUDNITE: Absolutely. Let Jim say something, i3 1
l i4 and then I am going to come to the systems side to tell you how to choose what to look at, and that is the third leg of i is ;
16 this triangle.
37 Jim.
1 j is MR. RICHARDSON: I think to an extent, and maybe i9 not to the total extent, but our component fragilities l
20 Program is addressing exactly that issue, and John O'Brien I i
l 21 this afternoon will be describing that program where the l
- 22 first thing we did was develop, one, an importance list on l 23 what components do we really need to know the fragility
- 1
! 24 and, secondly, how much of that hard fragility data, that
! 25 is actual test data, can we get our hands on, and then what 1
85 h) i additional fragility test do we need to perform to complete 2 that data base-that will do just that, and then go back and !
- 3 evaluate these judgments that have been made on PRA's and
! 4 margin studies to see numerically and based on actual test 5 data how good their judgments are.
6 MR. BENDER: Just to respond for a minute. I -
7 think if you get all that information you will be a lot a better off. .I don't know whether you have it right now. ;
l 9 MR. RICHARDSON: We have some of it.
10 MR. BENDER: I think there are two kinds of things i
j 11 that are needed. One is the systems that are important.
1 12 MR. BUDNITE: And I am coming to that in the next 13 slide. !
u MR. BENDER: And the'other is the systematic list, is and when I talk about a systematic list I am thinking about 16 more than just a system. I am talking about where it is 1'
17 and what kind of condition it is installed under, what kind l 18 of instructions go with the lastallation and how much !
l 19 quality control there is over the installation.
! r 20 MR. RICHARDSON: Maintenance, too.
i 21 MR. BENDER: Maintenance is an additional item, 22 but I will set that aside for a minute. But no matter what 23 the list is, if-you are going to do this thing, the 24 criteria that go with it have to be laid out, and I have
! 25 yet to see, because I haven't been looking very much, i
_.._,,_1__ , _ , . _ _ _ , , , _ _ , , , , __
86
, 1 that systematic discussion and presentation as the basis.
i 2 (Slide.)
3 MR. BUDNITE: Now I am going to come to the 4 systems discussion to show you what we h' ave done, even
- 5 though we admit that it is only an approximation or an I
! 6 incomplete first pass, about that.
i Now in the systems discussion I am going to l 7 i
l 8 concentrate on pressurized water reactcrs, PWR's. Dan 9 explained why, but I will just repeat it.
10 We have in the literature a little more than a ti dozen PRA's complete, seismic PRA's on PWR's. Only six or 4
12 seven of those have been published and reviewed and all.
i3 There are a comparable number, about another half a dozen, IO i
u that members of the pane either knew about or worked on or had studied and which are all going to be published in the !
is
! 16 next 3, 6, 9 or 12 months.
17 So we are relying on about a dozen seismic PRA's is that were done in'every case in conjunction with a regular i, internal initiator PRA. So that we have the full, l ;
20 including earthquakes, PRA. Now some of them are old and l
21 some of them are recent and some of them have a'better j 22 methodology than some of the others because the methodology of seismic PRA has been improving.
f 23 i
24 But we have tried to incorporate all the lessons 25 from these dozen into a method for selecting which systems O
,, yr,- - ,.---- - +,w..y-----,., - - - - m-.-r--,.--- , - . - - . , -
87 i really in a pressurized water reactor for earthquakes.
2 MR. EBERSOLE: Bob, let me try a shot in the dark.
3 MR. BUDNITZ: Sure.
4 MR. EBERSOLE: Were there in those PRA's set forth
! 5 ambient environmental conditions for auxiliary feedwater 6 Pumps, like control of temperatures and, you know, the 7 environment?
8 MR. BUDNITI: The auxiliary feedwater pumps were 9 examined --- ,
j 10 MR. EBERSOLE: I didn't say that. I said the 11 ambient and supportive temperature control requirements for i
- 12 the environment in which they sit. I am trying to find a i3 coupled, you know, a dependency which tends to be ---
14 MR. BUDNITE: I don't understand your point.
l 15 MR. EBERSOLE: What is that?
16 MR. BUDNITZ: I don't understand what you mean by 17 the ambient ---
- is MR. EBERSOLE: Okay. Suppose the aux feedwater 19 pump is a steam driven pump, but you find it is 20 unfortunately cooled by an electric fan.
21 MR. BUDNITZ: Yes, that is included. As a general 22 example, instrument air, service water, JJ, wil the support 23 systems are included.
i 24 Now it is fair to say that the most recent PRA's 25 have done a better job than the first three or four, but 1
i
5 88
() i the methodology is capable and has been effectively applied to look at all of the support systems that support and all 2
3 the components that support those systems.
4 MR. EBERSOLE: All right.
5 MR. BUDNITZ: Now our screen methodology insists 6 that one.go to the level of looking not only at the support 7 systems, but at the support systems that support the a support systems. So that, for example, you might have an 9 instrument air system that is supported by batteries, as io well as the aux feed pump supported by batteries through ii instrumentation or something. You have to do that.
12 The PRA methodology is not only capable of doing i3 that, but in the last two or three has gone to that level, i4 although it is fair to say, and I just have to repeat, that i
i3 all these things are approximations. That is, some guy has ;
i6 got to make a judgment about what matters and what support 17 things do and what support things don't really contribute is to the failure, and failure meaning functional failure, j i, because the infinite list is infinite. It is too large.
20 So there are systems judgments supplied, but there 21 has been an attempt, an intellectually honest attempt to 22 try to do the best you can and all that stuff.
- 23 Now here is the important lesson, and I.have got 24 to put it up.
i l
25 (Slide.)
O l
89 You have it in the handout, but this is a crucial
) 1 2 lesson that we have learned. Just look at the top half of ,
3 this. The bottom half is that thing I drew on the board.
4 What we have found is that in the dozen PRA's, 5 seismic PRA's to date that core melt is dominated by 4 6 failures in a very few functions. Failures in other 7 functions don't contribute.
, a In fact, there are three, but I left one of them
- 9 off. The one I left off is, and I really should have typed 4 10 it in because it is not unimportant. It is maintaining hot 11 shutdown, and we haven't differentiated from between hot ,
12 and cold shutdown because the PRA's don't, and one can 1
i3 dispute whether we should. But, anyway, we will talk about 14 hot shutdown because our criterion is if you get to hot is shutdown safely, then we are there. Now that is.not quite 16 right. You do ultimately have to have RHR to get to cold 17 shutdown, but we haven't made that thing.
18 MR. EBERSOLE: Before you leave that, ECCS, what i
19 is ECCS?
i I
20 MR. BUDNITE: I am going to describe it.
22 MR..BUDNITE: Let me just say it. There are two i
23 things that have dominated the core melt probability in all i 24 these PRA's that we have on our shelf. Now we are not sure 25 that they are going to dominate all 70 pressurized water b
90 reactors, but we are making a leap here, and that leap is f i 2 something that we want the systems community to review, and j I
3 we are hoping to get review from them as they chew on 4 this. But that leap is that if we examine those systems 5 that support this function and those systems that support l
6 this function and the components and systems that support 7
those, and if we can determine what the margin is for s
those, that is what we are calling the margin for this 9 plant.
10 Now I will describe what systems and functions it are. ECCS is a jargon for a whole lot of things.
12 MR. EBERSOLE: Oh, no, and let me --- ,
13 MR. BUDNITZ: --- which includes high-pressure i4 injection and ---
15 MR. SIESS: Not seismic PRA.
16 MR. EBERSOLE: Bob, if you go to the FSAR's you 17 will find that ECCS is a fairly small set of systems is intended to mitigate pipe break.
i9 MR. BUDNITZ: Oh, we are not talking about that.
20 MR. EBERSOLE: Well, we have got to have some kind 21 of commonly understood language, and ECCS is not it.
22 MR. BUDNITZ: What we are talking about is those 23 systems that are necessary i.r t inventory maintenance ---
24 MR. EBERSOLE+ Mb all circumstances.
25 MR. BUDNITZ: Whac; l
I
91
() 1 2
MR. EBERSOLE:
just for pipe breaks.
Under all circumstances and not 3 MR. BUDNITE: No, no, no. Under transients, under 4 small LOCAs.
5 MR. EBERSOLE: ECCS is just for pipe breaks. Go l 6 back and look at the FSAR's.
7 MR. SIESS: Oka'y. Let's scratch ECCS.
8 MR. SHEWMON: The engineering community defines it 9 differently. So you are going to have problems if you try 10 .to redefine the word.
11 MR. EBERSOLE: You have got to say what ECCS is.
12 MR. SIESS: These are not the words they use in i3 the report, if I recall.
()
14 MR. EBERSOLE: Right.
1
~
15 MR. BUDNITE: I typed the slide up, and I 16 apologize for ---
17 MR. EBERSOLE: Well, ECCS is a misnomer. I invite is you to look at any FSAR. #
19 MR. SIESS: Jesse, let's accept that. Scratch 20 ECCS. They use different words-in the report because I i
21 remember it.
22 MR. BUDNITZ: Okay. Now the function that we are 23 talking about is maintaining the inventory and ---
24 MR. EBERSOLE: And keeping water on the core.
25 MR. BUDNITE: --- keeping the core covered during C
= .. - - - - - . _. . - - _
l 92 4
i
() i the early phase of the accident, and the early phase mean, as opposed to what in PRA systems jargon is what called the 2
3 late phase, which has to do with maintaining it long after 4 all the transients have happened, and that late phase has 5 to do with RHR and that sort of stuff.
6 Now this may surprise those of you who have 7
worried about RHR-all your life that you have done such a a good job. In fact, there is none of these 12 or so PRA's that we have.on our shelf that that find that RHR
~
9 io contributes during a seismic initiated accident. That is 11 great, by the way, and that is what we found. This is the 12 stuff that contributes to those things.
i3 So we decided that in screening plants to pick the
() 14 systems to do the fragilities analysis on that we are not is going to look at RHR, but we are going to look at the 16 systems and components and subsystems that support these i7 two functions. And that is a judgment that the margin of is these plants will be dominated by whatever supports these i9 are and not by RHR. RHR turns out to just be held for 20 stout.
2i MR. BENDER: Bob, I think there is a big class of 22 uncertainties that can be dealt with that way.
23 MR. BUDNITZ: Yes, there are.
24 MR. BENDER: The fact that we don't find the RHR t 25 to be a dominant system has to do mainly.with the way in i
O l
l t __
i
! 93 which we see the earthquake influencing the overall f 1 2 operation of the plant.
3 MR. BUDNITE: That is fair.
!, 4 MR. BENDER: But if, for example, we were to take 5 the case that happened at Mexico City in which we had a 6 very severe second shock occurring at an incremental time 7 later when that plant was operating, I don't know what a actually you would get.
9 MR. BUDNITE: The plant wouldn't be operating.
~10 MR. BENDER: You would be taking away the i
11 afterheat.
12 MR. BUDNITZ: And you want RNR to do that.
13 MR. BENDER: And what its survival capacity is 14 under those circumstances has to be thought about.
l 15 MR. SIESS: Was it lower the second time than it L
! 16 was the first?
17 MR. BENDER: It was lower, but not a hell of a lot is lower.
19 MR. SIESS: No, I mean do you think the RHR could 20 survive the first shock but not the second? Is that what 21 you are saying? ,
22 MR. BENDER: It would shut down the first time.
i 23 Usually it is not in business. So it doesn't see the 24 thrust of the first loading.
25 MR. SIESS: So you are talking about the O
94 difference between the equipment that is not running and 4
(I i 2 the equipment that is?
3 MR. BENDER: That is an important consideration.
4 That is why I said that a systematic thought process is
- 5 needed. .
6 MR. BUDNITZ: That is fair.
7 MR. BENDER: I might come back to the same answer a you have got, and I am only exploring it because I think I i 9 understand it what you are talking about with io uncertainties.
ii MR. BUDNITZ: Now I have to say some things to 12 modify what I said so you will understand the level of i3 inquiry here.
a When the PRA's found that RHR was, as I said, held i3 for stout, that is it was stronger, sufficiently stronger, 16 and it was these other things that were the dominant 17 contributors at lower PGA levels to seismic initiated is accident risk.
39 There are vulnerabilities to support systems. You 20 know, for example, you might lose component cooling water, 21 and that might happen at a lower level. So then you say ah 22 ha, but RHR went out on account of that. But that same 23 component cooling water, the service water system design, 24 for example, supports these things, too. And so these 25 turnout to be for other reasons the things that dominate O
95
() 1 and not the RHR. There are some of those underlying dependencies of several of these things like the service 2
3 water systems, but it it is these functions.
4 What we found was that any plant where these 5 functions survived, survived in the PRA for an earthquake.
6 MR. EBERSOLE: Would you call the aux feedwater an 7 ECCS system?
8 .MR. BUDNITZ: No, we are not defining it that way.
9 MR. EBERSOLE: Oh, you aren't.
10 MR. BUDNITZ: No.
11 MR. EBERSOLE: Then how do you get the heat out 12 early on of a PWR?
13 MR. BUDNITZ: If you lose it, there are other ways 14 around.
15 MR. EBERSOLE: What is that, bleed and feed?
16 MR. BUDNITZ: Yes.
17 MR. EBERSOLE: In some plants.
18 MR. BUDNITZ: In soma plants. Well, there was 19 none of the 12 that were looked at at which you couldn't do 20 it, and that is at which if you lost it you couldn't do the 21 other. But aux feed was not defined as a piece of this 22 function.
23 MR. BENDER: Let me add one more piece of fuel to 24 the fire because one of the things that is of more concern 25 in these kinds of events is the shaking scares the hell out O
96 A
i of the operator.
2 MR. BUDNITZ: Yes, and that is not here.
3 MR. BENDER: And the importance of his response 4 has to be in that picture.
l 5 MR. BUDNITZ: That is not here. Like design and 6
construction areas, it is an area that we haven't addressed 7 properly and which we are not going to be able to until j s more research is done. I agree.
9 MR. EBERSOLE: I know now you weren't looking at jo Palo Verde.
1 11 MR. SHEWMON: I had a question somewhat more 12 generally, but the same thing that Mike brought up, and '
l
, j3 that was to what extent what happens in the plant is
! ) u independent of human activity? You brought up maintenance 15 and said that was a factor that came in.
16 MR. BUDNITZ: In the fragilities determination.
i7 MR. SHEWMON: Most of these operator actions are is not necessary or how soon does it come in?
19 MR. BUDNITZ: Well, let me describe how the 1
i 20 margins review is done, and then you will have the answer.
21 The way the margins review is done in our plan, l 22 which we are going to try on a trial plant, you know, l
f 23 coming up, is the first thing that happens is the systems 24 analyst sits down with the plant drawings and the 25 operations and walks around and determines which systems O
97
) i and components and subsystems as part of all that stuff 2 support.these two functions. That is the first thing he I
3 does. And by that, that includes tanks and it includes,- l 4 you know, fuel lines to the diesels, and that sort of 5 thing.
I 6 Oh, by the way, I misspoke when I said there were 7 only these two. There is the electrical system, and that i a is the diesels.
9 MR. EBERSOLE: Well, it is just a support system.
10 MR. BUDNITZ: Yes, absolutely. So the fuel line 11 from the day tank to the diesel is in here because it is 12 part of the support system that supports these things. So i3 that is all there.
14 Now he sits down and makes that whole list. That is is the first thing that happens. He gives that to the 16 fragilities analysts. The fragilities analysts then go and 17 -walk through the plant and make a first pass determination l 18 as to which components look like they are so strong that 19 they are well above the level you are reviewing at.
20 Suppose it was a plant with a .15 G earthquake and 2p this review is being done at .3 G, just to pick an 22 example. They make a walk-through and it is basically a 23 triage approach in which he says gee, these components are 24 really -- I can tell by my walk-through or by my knowledge 25 or by my analysis or whatever else, that there is a whole l
-. . .. . . -- =-.-
98
() i 2
class of components there, you know, very much stronger than that, there is a class that I know are suspect, and
} 3 there may even be a whole class that I know are weaker than 4 that.
5 MR. EBERSOLE: Let me pick one particular one.
i 6 How about a B&W plant which has very little water.
7 MR. SHEWMON: Why don't you let him finish.
8 MR. EBERSOLE: But I want to get him while he ir i
9 on the topic here. A Babcock and Wilcox plant uds goc 10 almost no water au the secondary. It is gone real quick.
11 You don't account for aux feedwater. So it is try. So you 12 say well, I jump to bleed and feed. And so I open the .
i3 bleed and feed valve and I oversteam into the containment.
( As you know, the plants built today depend on 115 u
is or 230 watt solenoids to hold those valves open. They are i
16 not environmentally qualified for the very steam 17 environment-in which~they are going to produce. How did is you keep them open?
i9 MR. BUDNITI: I didn't hear the rest of it. j 20 MR. EBERSOLE: How do you keep these valves open 21 that are not environmentally qualified for the environment 22 they produce?
23 ~ MR. SIESS: That is not an earthquake, Jesse.
24 MR. EBERSOLE: Oh, it is under his circumstances. j i
l 25 He has caused all this to be needed by the earthquake.
O 1
n - - - - - - - - _ . . _ . .
99 MR. SIESS: What did the earthquake do?
f) 1 2 MR. EBERSOLE: It lost his aux feedwater. He just 3 said it.
4 MR. BUDNITZ: No, didn't say it. I said that ---
6 5 MR. EBERSOLE: You said you didn't count it.
6 MR. BUDNITZ: You don't need it.
7 MR. EBERSOLE: Well, okay. So in that context it 8 is gone. So you are relying on -- and I am just coming 9 down to a particular part of the design, which is the to unqualified, non-environmentally competent bleed and feed 11 solenoid valves.
12 MR. BUDNITZ: I can't talk about the environmental i3 qualifications. But in terms of earthquake that valving is 14 on our list.
15 MR. EBERSOLE: It is not qualified to resist the 16 very environment it produces.
17 MR. SHEWMON: Now you have also postulated a LOCA; is is that it? !
19 MR. EBERSOLE: No, I haven't postulated a LOCA. I 20 am just getting the heat out.
21 MR. SHEWMON: How did you get this environment?
22 MR. EBERSOLE: I opened the PORV's,'as he said he 23 was going to do.
24 MR. BUDNITZ: I can't speak to environmental 25 qualifications, Jesse.
t
4 e
100 1
i MR. EBERSOLE: Well, if you don't, you aren't with
{)
2 it.
3 MR. SIESS: No, this is seismic, Jesse.
4 MR. EBERSOLE: Look, if the seismic produces an 5
environmental condition that can't be survived, it is a 6 spinoff for the seismic condition.
7 MR. BUDNITZ: He is absolute right.
8 MR. SHEWMON: Now the PORV exit doesn't go into 9 the containment. It goes into the tank, doesn't it?
10 MR. EBERSOLE: Oh, it does.
t 11 MR. BUDNITZ: No, it goes into the tank.
12 MR. EBERSOLE: And then it comes out of the tank i3 very quickly when you keep it open, a MR. BENDER: All Jesse is doing is amplifying the is Point that I made before.
16 MR. BUDNITZ: Oh, I understand your point.
17 MR. BENDER: You have to have a systematic 4
is understanding of the conditions in order to be able to go 19 further.
20 MR. BUDNITZ: Now the valves in terms of their
. 21 seismic behavior are in this ---
i 22 MR. EBERSOLE: They are not seismically competent 23 to open in the first place. Right now they are not
, 74 provided with electric power which is competent to l
l 25 withstand a seismic event.
101 i MR. BENDER: That is not the right way to play the
)
2 game. I think you really ought to think of it in terms of 3 no matter of whether it is seismically qualified or not.
4 MR. EBERSOLE: It might work.
5 MR. BENDER: What is the expectation that it will 6' work?
f 7 MR. EBERSOLE: I agree.
8 MR. BENDER: That is the thing that you ought get.
9 MR. BUDNITE: Now whether it is seismically to qualified by the NRC, and I don't mean that pejoratively, 11 but what I mean is that may not be relevant to whether it 12 works ---
13 MR. EBERSOLE: And it might environmentally i4 survive.
I I 15 MR. BUDNITE: The seismic performance of that is 16 reviewed in this.- So that is important for you to 17 understand. The seismic performance is reviewed in this.
i 18 Now whether it is environmentally qualified.by the steam 19 environment I must say isn't part of this.
20 MR. SIESS: You are assuming that, except for 21 design and construction errors, that the plant has been 22 designed according to current regulations.
73 MR. BUDNITE: Yes, that is right.
24 MR. SIESS: And that the equipment is designed at f
25 least to take the SSE and that the equipment has been 1
-= - _- .. . . .-
102 i designed to take the environment that is now specified.
2 MR. EBERSOLE: Of course, this hasn't been so l 3 designed.
4 MR. SIESS: I know, but they can't help that.
5 They are looking here at the seismic capability of a plant, l 6 and if a plant has other weaknesses that won't let it 7 operate, then that they can't look at.
8 MR. BUDNITE: What I was going through was to show
.: 9 you how the human factors part came in.
- 10 MR. BBERSOLE
- But just in a summary though, if I ii the seismic event produces ~an unacceptable environmental 12 condition, you have had it, haven't you?
g MR. BUDNITZ: Yes, sir, absolutely.
i4 MR. SIESS: Jesse, it is the same thing. If the
- 15 operators have not been trained to take one of these 16 alternate paths ---
37 MR. EBERSOLE:
It is the same thing.
is MR. SIESS: You can only go so far with seismic.
, i9 MR. BUDNITE: If that is true because it was 20 caused by something else than an earthquake, you have still 21 had it. So we can't cope in this program with those i
22 things. I am not saying that it is unfair, but you are 23 right. I mean you have had it.
24 What I was going through is the logic of our 25 review to show you -- I was answering the question about O
\ __- .. -_ - . . . .___ _--.- -_ ._ . - - . - . - . . . _. . - _ .
103 i human operator intervention.
2 So I said the first thing that happens is we go 3 through and we make the list of all the systems, subsystems 4 and components, and those valves and so on are in that I 5 list. And the fragilities guys go through and they,make i
6 this pass to decide what really looks like it is stout 7 enough that you don't have to worry it and what stuff is ;
]
e sort of still worthy of more study.
1
{
9 Then having done the walk-through, and the walk-10 through is more than just a walk through on your feet, but 2
it it has to do with studying drawings and the like, then the ,
i 12 seismic fragilities experts and the systems people sit down 13 together and look at which things survived this pass as 14 still being weak enough to be of concern, and weak enough l 15 meaning that the HCLPF level is not surely, you know, real
! 16 stout, and sit down and go through and decide which 17 accident sequences the systems guys can construct fron is event trees and fault trees that involve the components j
- 19 that are left.
I j 20 And what you are going to find is some of the l 21 components that look like they might have been weak down l
l 22 here don't contribute to any accident sequences because the 23 stuff that they were in sequence with was strong. So you l
24 are left only with a subset of those that are in the 25 combinations that they survived as accident sequence C
i l
104 i combinations, and then those are done in greater detail by 2 the fragilities people. They go back and they do more 3 refined analysis and they try to figure out which ones of 4 those really have HCLPFs above an which ones don't at the 5 level you are reviewing at. So it is iterative process.
6 And what you end up with is a list of components, 7 and they are at the component or system level, you know, a and a component might be a pump or it might be a battery 9 rack or something, or it might be a structure or it might 10 be a well that supports something hanging on it, which have is the dual properties that they fit into accident sequences 12 in the PRA sense that the seismic event affects, and that 13 their HCLPF value is below or near this level.
i4 And from that combination of systems and seismic is analyses you can determine what we call the HCLPF value for i
16 the Pl ant. That is, there is some accident sequence whose i7 BCLPF is lower than the others, and that is the point and is that is the margin.
i9 MR. SHEWMON: You are talking about operator 20 action and if we keep listening long enough you will get to l 21 it.
22 MR. BUDNITZ: Yes. Now in the systems analysis 23 that examines which of these components fits into what l 24 accident sequences, the normal PRA operator error approach 25 is used to see whether or not operators are involved and, 1O l
- 1 105 1 1
l 7 i if so, what their error rate is.
4 2 As it. turns out, most of the seismic sequences 3 that have dominated risk in the PRA's we have on our shelf 4 have not had much operator action in them. That is as it 5 has turned out. A few of them have, but mostly they i 6 haven't.
7 And if that is true of the plant you are looking 8 at, then there isn't any operator in the picture. But if i
9 it is, then the operator is in here and you have to assess to whether or not that. operator contribution is or is not a
- 11 contributory part of that accident sequence. So it is some 12 combination of PRA type seismic and fragilities analysis.
i3 MR. SIESS: But you don't change the operator
< ( 14 frequency because of the seismic?
15 MR. BUDNITZ: No, we do not, but we are assuming i 16 that over the next'few years enough research will be dona, 17 anc we don't know that it will be successful, so that we is will have better error rates during and after earthquakes l 19 than we now have.
20 Right now the standard assumption has been that it
- 21 is the same, but you can make more pessimistic assumptions 22 and it changes the numbers.
23 MR. BENDER: I think you are right in the approach 24 you are taking. I don't see any other way to do it. So i 25 you have to do it, from my viewpoint, in a way like this.
l
- 0 l l l
L. _ __- _ .______ ._, __ _ _ _ _ __ _.____ ._. - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
- _ .. . .= - . _.._. _. - - - . - _ . -
106 i
l i MR. BUDNITZ: I don't want to gloss over these
[}
2 Points.
3 MR. BENDER: But the things that bother me most 4 have to do with the comprehensiveness of knowledge that is 5 associated with it.
MR. BUDNITZ: Yes, and that is a fair point. It 6
7 is'a real problem.
8 MR. BENDER: And at the moment I have some !
l p 9 discomfort with that' thing, unless you can follow the tact io Jesse is talking about. If you can narrow it down to a few it things and know that that increment of components is small
- 12 enough then to look at it, you may be okay. But right now l i3 I think I would have to say that logic is missing.
) 14 MR. BUDNITZ: What do you mean by missing?
15 MR. SIESS: Mike, do you think they are any worse 16 off in this approach than they would be in trying to do a l
l 37 full-scope seismic PRA?
is MR. BENDER: Well, I think if'you do a full-scope
! 39 seismic PRA, it will be this thing. That is the only thing 1
20 they could ever do. And all I am saying is that ---
21 MR. SIESS: I don't.get you. There have been full-22 scale seismic PRA's.
23 MR. BENDER: Well, if you have to do something 24 that invol'res the whole spectrum of events, the logic get 25 so complicated that you eventually just wind up saying O
l 107 k) I well, here are the things that I am going ---
2 MR. SIESS: Well, I am trying to understand your 3 concern. I thought your concern ---
4 MR. BUDNITE: I don't understand it.
5 MR. SIESS: --- was that this screening they do i
l 6 July not be complete. ,
7 MR. B2NDER: My concern is that without having a a better picture of what the screen consists of that it is 9 very hard to know whether the scale of it is such that it to is sufficiently uncomplicated to be able to trust it.
i 1 11 MR. BUDNITI: I can answer that. I think, without 12 giving a satisfactory answer about how it is going to work i3 until we do the trial, but I can answer it in terms of our 14 intent and what we wrote down.
15 MR. BENDER: Bob, before you do this, I think what 16 Mike is saying here is compatible with what I would call 17 life boat logic. I don't know what is going to sink the
- 18 Titanic, but I know I am going to have a rubber boat on the 19 deck which is going to float after it sinks.
20 MR. BUDNITZ: Well, let me answer the question sa 21 follows. Our intent is to analyze all, and I really mean 22 all the systems, subsystems and components and support i 23 systems as far as the PRA approach can identify them
- 24 through the fault trees right down to the basic events that 25 support these functions.
- _ _ _ . _ .~-._ _ ._ -_ _. _ _ -- - _ _. ,__ _ ._ _ _ _ -__ _ _._._ _ _ _______.___ _ _
108
( ) Now you say, gee, but that is too much. In fact, 2 it turns out that if you limit yourself to these functions, 3 you have eliminated 80 or 90 percent of what you would have 4
had to otherwise review and we are left with only 10 or 20 5 percent of what you would have had to otherwise review.
6 Now it turns out that much of the 10 or 20 percent !
i 7 that you are reviewing is the support systems, you know, a
because every support system is involved, and it is a much 9
smaller number of what I call the front lines, the things
- io that support systems support.
in Therefore, we believe you have a chance of doing 12 the thorough review that a seismic PRA hasn't been able to '
i3 do because there is too much there. And, furthermore, and O
i i4 this is a crucial point, we believe that using this is screening approach, you start out by throwing out of your l 16 review those things that, you know, hell, that is a 2.0 G 17 HCLPF, and it is probably a 6 G median, and there are 1
is things like that that people know about. I am talking ;
i 19 about, for example, small, compact valves, and you don't i
20 worry about them any more.
2 Having done that preliminary screening, you are 4
22 left with a very much shorter list than the poor seismic ;
23 PRA analyst has to do when he has got to find the median 24 fragility, and I won't say for everything, but for a very 25 large number. Although it is fair to say that, you know, i O
109
) i in the PRA's they have only ended up with two or three 2 dozen of these because of other ---
3 MR. EBERSOLE: Even so, you are still using 4 Titanic logic that you are going to keep it afloat, all 5 these complex systems, and there is no integral capacity 6 parked to one side that is going to save it because it is 7 particularly seismically competent.
8 MR. SIESS: It is existing plants as they are 9 built, Jesse.
10 MR. EBERSOLE: Yes, right. So he has to do that.
11 MR. SIESS: They don't have a bunkered, 12 seismically designed dedicated heat removal system.
_ 13 MR. BUDNITZ: Now let me tell you what I think is
, k- / 14 the biggest vulnerability of this as a potential, and then is we are going to find out through the trials whether it 16 really is an actual or only not a worrisome vulnerability 4
17 to this approach.
- is It could be that in the course of doing these i, margin reviews that you find out that the systems that are
- 20 involved have -- the systems and functions and components 21 that have the HCLPF values in the area of interest that 22 aren't real strong combine in a system sense to make 23 accident sequences that are very complicated accident 24 sequences, and not the kind of simple cut and dried ones 25 that the PRA analysts have traditionally used in their O l
l 110
() i 2
internal initiator events.
It could be the accident sequences that remain are 3 so complicated that we really don't know how to analyze 4 them. As a systems analyst that is what I know more 5 about. I am not sure how that is going to come out.
6 MR. EBERSOLE: You are talking about something i
7 that, for instance, would disable such a root system as the a service water system?
l 9 MR. BUDNITZ: Yes, sir. In fact, that was what io dominated the risk at Zion, the seismic risk.
l 1 Now I want to reiterate though that we are not 12 looking at reactor safety in the whole. For example, if we find that the large accumulators for the large LOCA ECCS
()
33 i, are stout and that the support systems are stout, I am not is going to back up and ask whether or not if they are called 16 they are going to cool the cote. I mean, you know, the i7 research program spent a half a billion dollars asking that is question, if it is isn't solved, there is nothing we can do i
i, about it.
20 You can imagine that they might now cool the core t
21 during an earthquake, you know, and there are a whole bunch 22 of things. Maybe the operator got confused and he did 23 something he shouldn't have done. I mean there are a whole 24 bunch of things there that we haven't gotten into yet, but 25 we think that this pass should give us a lot more insight
() :
a 111 I than we have ever had before.
I ) As a matter of fact, the 2
MR. EBERSOLE:
3 accumulators have nitrogen in them which can plug the 4
iiiverted U tubes and stop cooling the core.
MR. BUDNITZ:
So I just want to wrap it up with a 5
6 comment about the future and then I will be done.
7 (Slide.)
We decided at the last meeting, the panel, and Dan a
9 talked about this earlier, we decided that we would like to 10 have two trials, one at a review level that is not too much I mean just to give an example, take 11 greater than the SSE.
We would 12 an eastern plant with an SSE about .1 or .15 G.
13 sure like to have a review at let's say .2 or .3 G for a plant like that. We would also like to have another trial
%/" 14
.4 or .5 G, which is such a is of this methodology at about 16 large earthquake and a whole lot more components are in 17 this picture than the other one.
The reason why we want both.is because we suspect is l 19 that the first one I mentioned, the .25 or .3 G review for I 20 a plant whose SSE is .1 or .15 will be used often and will 2r give great benefit to plants that do it because we expect And, 22 that a lot of plants are going to be great there.
23 therefore, we want to have a trial there.
t I But we also want to have one at a higher level 24 f where we suspect there may be problems and issues there 25 lw_,'I U. -
I i
112 ;
l l i because everything is so much more complex because we don't 2 think it is fair not to do it. t lI
- Now we originally said, gee, let's do plant "X" at 3
i 3
! 4 this first one and let's do plant "Y" at the other one, and i, ,
! 5 Dan has mentioned our feeling that maybe we ought to do the I 6 same plant first at the one and then at the other. You can j 7 do them in the opposite order. If you start out with a .4 g G earthquake and then come back to the .2 you have learned 9 the lessons that all right you have got to do it in this ,
{
j io order in order to understand how the review would take l
) 11 place.
i 12 But we think that if we do two levels like that we
! 13 will get different insights even on the same plant which a will tell us whether an independent team that is going in is and going to do let's say a .3 review for a .15 plant can i
16 or can't approach the probleia in ignorance and use the
}
- 17 methodology to see if it works.
is I would urge you to think about whether that makes i, sense and give us some. advice about that, because that is l
20 one of the crucial kind of logistical decisions we are
! 21 making right now as we select a plant for review. And, by 22 the way, the first plant is going to be a PWR.
23 The second comment I want to make has to do with l
24 boilers and then I will shut up.
25 MR. SIESS: Let me say one thing. It is
~ .
113
() 1 refreshing to hear somebody talk about how the results are going to be used.
2 3 (Laughter.)
4 I said it is refreshing to hear somebody talk I
5 about how the results are going to be used and to .
6 demonstrate that quite a bit of thought has been given to 7 how the results are going to be used.
8 MR. BUDNITZ: We have designed it for use and not 9 just for fun.
10 Now the other thing has to do with boilers. There it are about well, more or less a half a dozen BWR's whose 12 PRA's have either been completed or are in process now, 13 seismic PRA's. But it will be another 3, 6, 9 or 12 months 0,
14 before we have enough of them on our shelf and have is reviewed them and understood them so that we can come to 16 some systems conclusions about which syctems are important 17 for them like the conclusions we came to for the la pressurized water reactors. And until we have done that, 19 if you wanted to do a margin review on a BWR, you have got 20 to everything at least for the time being. That is all we 2i are willing to say.
2 2 ,,
Now I have my own views, and I know that Paul
,3 Amigo on the panel has his own views and Garth Cummings has 24 his own views. We all have our own views about how the i 25 boilers are going to come out because I have been staring i
i
114
() i at these things for some time. But we haven't had the systematic review of enough of them to be ab3e to say 2
3 something generic about them yet. And when that happens, 4 we will have a list for boilers and then we will want to do 5 a trial one of them.
6 Mk. SIESS: What do you see as the significant ,
7 differences between the boiler and the PWR, how the control l 8 rods go in?
9 MR. BUDNITE: No, no, no, it is not control rod io questions. It is the support systems, and I don't mean the it support system like water and so on, but it is the 12 injection system. For example, the boilers have -- and, i3 you know, the RCIC system is not the world's best system.
i4 MR. SIESS: Okay) 15 MR. BUDNITE: And that is just an example. And 16 then there is the whole question about depressurization.
i7 The whole ADS'or manual depressurization function and its is role in these sequences, and in any of these sequences, i, requires some very careful thought and we just have to do 20 that.
21 MR. EBERSOLE: Well, when you do the boilers, are ,
22 you going to invoke containment venting, because if you 23 don't, you are going to immediately find that is where they 24 are' going to die.
25 MR. BUDNITE: Yes, that is fair. That is a fair O
l 115
)
i comment. It is not where, but it is a vulnerability.
j 2 MR. SHEWMON: My question had to do with why there ,
i 3 have been more PWRs done than BWRs?
4 MR. BUDNITI: Well, I think I can answer that in 5 very simple terms. The NRC ordered, or at least twisted l 6 arms for three high-population sites, and they were sion, 4
2 7 Indian Point and Limerick, and the Eion and Indian Point a study has been out for, I don't know, three or four years, 9 and Limerick is just now out. And when the Zion and Indian 4
10 Point studies were completed by Pickard, Lowe and Garrick,
- 11 John Garrick to his infinite credit or whatever succeeded i 12 in selling his services to a whole lot of PWR people to go i i3 and do theirs with the methodology that he had had.
14 For example, he want and sold it to Seabrook and
- 15 Seabrook has now done one. And that I think was kind of an i
16 historical thing. It wasn't a bias per se. On the other I 17 hand, you know, the NUS crowd that did Limerick has now l
18 gone and done a comparable job at Shoreham, and so we are l
i 19 going to have that one. Actually I have some information j 20 about that, but it hasn't been reviewed. So it is just a
- l 2: question of history. There wasn't any bias the way there 22 has been in the research program's 10-year history.
f 23 MR. SIESS: Does the same ratio hold on internal 24 event PRA's, or is that just one seismic that we have seen 25 all the PWR's and not the BWR's?
O I
i 116 i
() i event PRA's, too.
MR. BUDNITI: I think that is true of the internal Well, let me think. '
2 3
MR. SIESS: Some RSSMAPS has ---
4 MR. BUDNITE: Well, obviously if you think about 1
5 BWR's and the regular PRA's, I mean Peach Bottom of course t
6 in the beginning and then there was Grand Gulf and 7
Limerick, and Susquehanna, which isn't out, and shoreham J 8 and Kuosheng in China ---
i
) 9 MR. EBERSOLE: Are you going to do a spectrum of f
io boilers like Nine Mile and ---
i n MR. BUDNITE: We are going to study what PRA's i
have been done, and that is not a spectrum. That is, let's i2
! i3 see, there are going to be Mark II's.
34 MR. SIESS: Well, you have got a Mark I, don't I
i 15 you? Isn't Millstone I a Mark I?
16 MR. BUDNITE: Yes, but Millstone I is not a good i
1 i7 one.
is MR. EBERSOLE: The Limerick boiler doesn't have
! i, one yet, does it?
20 MR. BUDNITE: Yes, i
21 MR. EBERSOLE: Limerick has a seismic?
- 22 MR. BUDNITI
- Yes, sir, and that of course a Mark
\
23 II, and Shoreham does, and that is another Mark II.
24 Kuosheng is of course a Mark III.
' MR. EBERSOLE: Limerick has a containment venting ,
25 i
lO -
i I l
! i l
l-.. . - _ _ _ _ _ _ . _ . . _ . . . _ _ . _ . _ . . _ . _ , _ . , _ _ _ , , , . _ _ . _ . . _ . _ . _ _ . = . . - _ . . . _ . _ _ , , , _ _ _ . _ . - . . . . . . _ . . . , _ . _ _ _ _ . _ , . , _ , . . , . , - ~ _ ,
117 i
() 1 2
system and none of the others do at the moment.
MR. BUDNITZ:' The bag we are going to catch is the 3 bag is the bag that has been done. We are not going to do 4 anything else on that.
5 So now I am done.
6 MR. SIESS: You are done, you said?
7 MR. BUDNITZ: Yes.
8 MR. SIESS: Are there any more questions of Bob?
9 We have been hitting him pretty hard.
10 (No response.)
u I would comment to you trying to read their report 12 because there is an awful lot more in it than Bob has been i3 able to cover in his abbreviated presentation, and his u presentation dealt more with the background and the is philosophy and the approach than the details.
16 MR. BUDNITZ: Can I make a pitch? I want to make l
17 a pitch.
is I think that this approach, and now I am speaking 19 for the panel -- by the way, the panel is proud of it and 20 that is because we did a lot more than we thought in a 21 shorter time than we had originally planned -- I think that 22 this approach has the potential for addressing a very large 23 fraction of the seismic related issues that are coming up 24 last year and this year and next year than any of the other 25 seismic things now going on partially because the other l
l l
118
, i things Jim is talking about are still two or three years 2 away before we have got all the information necessary.
3 Whereas this, you know, has a chance of actually doing 4 something in the next year or two, and it builds on whole 5
lot of PRA experience that is behind us. .
6 MR. SIESS: I think there is an important point 7
that has been brought out at least to me. I think there a
was a tendency again again at least for me to think of this 9 methodology as being a sort of an abbreviated short-cut to . substitute for a full-scale seismic PRA.
11 MR. BUDNITZ: That is of course completely 12 erroneous. !
13 MR. SIESS: And I have gotten a different 14 impression. I think it is something that I believe now has 15 the potential to be more effective in finding seismic 16 outliers of evaluating seismic margins than doing a full-17 scale seismic PRA.
18 MR. BUDNITZ: Yes, I agree.
i, MR. SIESS: And if that is true, I think that is a 20 very important. thing in the way we look at it as well as 21 the way we use it.
22 MR. RICHARDSON: I would say the only danger in 23 that is that if you have a plant that is an outlier that 24 has not been picked up in the PRA's that have been done to 25 date and its systems or behavior is enough different that O
119 1 you might use the wrong screening criteria, then that is a 2 danger that I think you have to keep in mind in reviewing 3 some of these older plants.
4 MR. SIESS: Well, if you go back and look at SQUG, 5 for example, it says look at component supports. Now SQUG 6 doesn't say look at component supports just in these 7 systems. It says look at component supports. So if we do a that sort of thing as well as this, you know, we will cover 9 some of those outliers.
10 MR. BUDNITE: Just to throw out one other thought, 11 just around the table one of you, and maybe it was Chet, 12 said well what about Millstone I, it would be an example.
i3 We don't think, the panel does not think that this !
14 methodology should be used for the small handful of the is really older plants. And part of the reason, and maybe the 16 main reason for that is that the systems insights that we 17 have gained, and I am only talking about the older PWR's is for the time being, and of course Millstone I is a boiler, 19 but say a real old PWR.
20 MR. SIESS: San Onofre.
21 MR. BUDNITE: Yankee for example, Yankee Row. The 22 reason for that is that we don't think that the systems 23 insights that we have gained are necessarily applicable to 24 that handful of early designs without some more careful 25 looking than we'have done. And so we would not recommend l
120 4
() i it to be just easily transferred to one of them. On the other hand, do you want to transfer it to one of the new, 2
3 large four-loop Westinghouse plants, and we think that 4 would be applicable.
5 MR. SIESS: Fortunately, some of those older 6 designs were simply enough that you could look at them in 7 other ways.
e MR. BUDNITZ: That may be.
9 MR. RICHARDSON: If you would like, Dan Guzy and io Bob can help him give you some of their impressions of what 11 the EPRI program looks like. Bob Kassawara presented to 12 them last week an overview of their program and they can 33 give back to you what their impressions were.
14 MR. SIESS: Pine. This is the time to hear it.
i3 MR. GUZY: Okay. Well, basically their approach 16 is very similar to ours. In fact, they intend to use some 17 of the information in this NUREG we just mentioned, and we is have also given them earlier drafts of our guidelines.
i9 They plan to use those in their review.
20 They are starting later than we are, and that 21 gives them an advantage to be able to look at this stuff 22 before they begin. But they are in some sense maybe a 23 little bit ahead of us because they have got a plant picked 24 out. The general configuration of people will be SMA will 25 be doing the fragilities work, PL&G will be on the systems O
121 I work, and they have Woodward Clyde as a consultant doing
) '
2 the soil structure reaction guidelines type work.
3 But the intent of these three contractors will be 4 to develop guidelines, generic guidelines that can be used 5 and plant margins that can be used and that can be used by 6 a utility. Their target review team would be comprised of ;
7 utilities. So in some sense they have to be a little more, 8 or at least in their thinking they have to be a little more 9 prescriptive than say what we are developing here.
10 The plant they have picked, which I believe is 11 public, is Catawba and Duke will be the review team that 12 uses these guidelines in their trial plant review. As far 1
1 13 as was presented to us last week at the panel meeting, and !
l() 14 there are a lot more similarities than differences at this i
j is point anyway in the two approaches, the only thing that I 16 noticed that was different last week was that I think there 17 are some slight differences in the systems approach, and is maybe Bob could explain this better. !
19 They have a success oriented or success tree type ,
20 of approach where they pick a successful chain of systems l l
21 that they can show has margins or they.have confidence in 22- performing, and they will shut the plant down to prevent ;
23 core melt.
I 24 But other than that, the SMA is heavily involved 25 in all the fragilities work that has been done so far and 0
-_x__ _ , _ - , - - _ . . ... - - . .
! 122 i
i there are a lot of similarities with the ---
2 MR. SIESS: SMA is doing the work for EPRI?
3 MR. GUEY: Yes.
4 MR. SIESS: Is there anybody else involved in it?
5 MR. GUZY: SNA and then PL&G is doing the systems 6 work. Then Woodward and Clyde, it is my understanding, 7
will be doing some soils type work. I think they are going a
to address soil structure interaction guidelines and I
, 9 think they are going to address liquefaction. l l 10 MR. SIESS: How long have they been working on it?
4 11 MR. GUZY: They have juac utarted. I am not sure 12 when it was awarded. I think it was awarded --- .
1 13 MR. BUDNITE: September 1 about.
i4 MR. GUIY: I think their schedule now, which I 15 believe is optimistic, will be.the guidelines themselves 16 will be finished somewhere around February, and then Duke 17 Power will begin their review then and finish in October- ,
is next year. Now I think there is some skepticism as to 19 whether they can do that.
20 MR. RICHARDSON: But they have a tremendous 21 advantage in that essentially they are taking off from what l 22 we have done. Another point, SMA's involvement, of course 23 Bob Kennedy has been one of the principle contributors to 24 our panel. He has now resigned from our panel and will no l 25 longer participate in our panel to avoid the obvious or at O .
i 123 f) i least apparent conflict of interest.
2 We intend to and will continue to work closely f 3 with EPRI in coordinating these efforts to make sure that 4 there isn't unnecessary duplication, although I think there 5 is some desirability for some duplication. But we are 6 going ta be in very close communication and working closely 7 with them.
8 We are going in very close communication and 9 working Closely with them, but we have made a conscious to decision not to pursue Catawba as our candidate review 11 plant so that we might remain independent of their program.
12 MR. GUZY: Well, originally we said to exclude it, 13 but I think there were some people on the panel, the expert 14 panel that would recommend it. So I think we are going to 15 consider that as one of the candidate plants now, at least l
16 for this next cycle of selecting a plant.
17 MR. RICHARDSON: I will be bet against it though.
19 MR. SIESS: Carson has got a question.
19 MR. MARK: One of the things that I found very j 20 encouraging about what we heard from Bob Budnitz was having
( 2i separated the hazard curve from a real definable question 22 of how big an earthquake will.this machine survive, 23 something in words almost like that. Is that the same 24 _ question that EPRI is going to address, or are they going 25 to have the earthquakes on their minds, too?
O 1
. l 124
() i MR. GUZY: I think their objectives are an earthquake such as would come out of these hazards 2
3 studies. But I think as far as their trial plant review, I 4 believe they said that .3 G would be their target 5 earthquake they would review for.
6 MR. MARK: They will put in a parametric 7 earthquake and work on that.
8 MR. GUZY: Yes, probably a .3 reg. guide type ---
9 MR. SIESS: Well, for Catawba they could take to Charleston, for example.
11 (Laughter.)
i2 MR. MARK: For their second round.
i3 MR. BUDNITZ: I had two points to amplify what Dan
() 14 said about the different between EPRI's and ours, the one is that was just mentioned.
16 They are not going to review at what I said the i7 higher level, .4 or .45 or .5 G level that our panel is recommended was an'important thing to try. They are i9 starting at the lower level they said, and I understand 20 that. They are going to have a hard time getting the 2p utilities to volunteer to cooperate and find out that they 22 don't have margin at the .45 with the sort of high 23 confidence we are talking about.
24 You know, when we say a high confidence at a 25 level, we generally think that the actual influence is l
l 125 i
() I 2
considerably higher than that.
just sort of where everybody can say.
We don't know, but that is And so I understand l t
3 why the utility crowd didn't want to do that. l 4 MR. SIESS: The best estimates.
! 5 MR. BUDNITE: Yes. Now the second difference is, 6 and he mentioned it, too, but there is a more fundamental 7 difference than that.
s They are going to do the whole system thing. They 9 are not going to limit themselves to these systems they I 10 say. Now I also believe they are going to end up quickly 11 screening out some things. So they are going to focus down i
12 quickly, but they are not going to start a priori. We want l 13 to start with just the functions I said and the systems
()
14 that support them.
4 is I think that is beneficial. They are going to see 16 whether or not what they end up with is what we have 17 supposed was so from the dozen in our literature. Now, by is the way, I think they are going to end up there. Catawba 19 was very similar. Its systems are just like Zion and l 20 Indian Point, you know, I mean it is not quite.
21 MR. BENDER: It has got an important difference.
! 22 It has a ice condenser in it.
23 MR. BUDNITZ: Sure, I understand that, but that is 24 another kind of a difference. I was thinking about in 25 terms of the NSSS.
i C
e ..,--,-:.-- .. - - - - - . - - , _ - - - - - - - . , - - . . - - . . - - - - .
126 i And because of that, and also because it is PL&G 2- doing the work, we are going to get a more independent 3 systems analysis from them that will be of great value.
4 Then we are going to get a more independent fragilities 5 analysis. The fragilities analysis isn't going to be quite 6 so independent. You know, it is Bob Kennedy and his 7 Colleagues. But the systems analysis is going to be very 1 8 different from anything that we would do in our review, and 9 I think that will help us get some additional insights.
10 MR. SIESS: I may be naive, but it seems to me n that it would be some sort of kick if they did their review 12 on Catawba, if you did one on Catawba and then somebody did 33 a full-scale seismic PRA on Catawba. Now all three may get i4 the same answers and you might get three different answers, is and I am not sure what it would mean, but just offhand it 16 sounds like a good idea.
17 MR. BUDNITE: The advantages of that are obvious la and are counterbalanced by the logistical problems that we i9 are wrestling with about trying to do a review in the same 20 Plant that they are trying to review and getting in there 21 and separately from them or with them and having something 22 that we are ready to publish and they are not ready yet and 23 vice versa. There is a whole set of questions which are 24 counterbalancing and they are not technical questions, but 25 they are important to trying to get the thing done.
lO l
i 127 !
) i MR. SHEWMON: There is also the question of is it i 2 really worth trying to do one thing three times and never 3 do anything else, or not do anything else or what you lose.
4 MR. BUDNITZ: Yes. l 5 MR. SIESS: Well, it is not one thing three 6 times. It is three different things with the same answer, 7 and I don't know if they all agreed how much confidence I 8 would have with any of them then.
5 9 MR. BUDNITZ: By the way, the ice condenser to question is of course relevant, but most of what we are li looking at is the NSSS. It is inside of that.
12 MR. BENDER: I guess I wouldn't be quite as i3 comfor' table in with that kind of answer as you are, but any 14 one you work on will tell you a lot.
is MR. BUDNITZ: By the way, in our report we 16 specifically excluded the ice condenser plants from 17 screening, that is the containment part of them for the is time being because no one has ever done a seismic PRA in an 19 ice condenser. So, therefore, although Kennedy is willing 20 to say and I think John Reed that the building is probably 21 okay, no one has a good. feeling yet with fragilities in 22 those ice beds. I mean that is an open question. So it is 23 worth asking.
i 24 MR. BENDER: I didn't intend to downgrade the idea 25 of tackling the Catawba plant at all. I think it is as (0
128 i good from the standpoint of learning it and you may learn 2 more from it.
3 MR. BUDNITZ: So with the balance between doing ;
1 4 the plant three times, as you said, with the benefits s
thereof and the administrative difficulties, I am learning 6 towards trying to do something else besides Catawba. That 7
is where I am leaning.
8 MR. BENDER: The one point, Bob, where I think you 9 ought to try to get with EPRI is in being sure that both of l io you have the understanding of what circumstances have to be ii dealt'with, and that is, the earthquake is one thing, but 12 the systems environment that goes with it has to be 13 considered.
u If, for example, in the ice condenser you are is going to assume that the ice is going to come down in the i6 earthquake and you are not going to have the opportunity to i7 condense the steam that is being released at the same time, is then the working environment is going to be ---
- i9 MR. BUDNITZ
- - That is fair.
2c MR. BENDER: That is the kind of thing you have to 21 think about.
22 MR. BUDNITZ: That is fair. By the way, I ought 23 to say that Thursday last when the EPRI people presented 24 their plans to our panel, we can to a tentative agreement, i 25 informal and just talking, that as we got into our separate O
e l 129
() I reviews, we would be careful to at least share with each other what our assumptions were so that we didn't make 2
3 arbitrarily different assumptions that would make the 4 analyses not comparable just out of foolishness.
5 So we are going to be careful to do some of that, 6 but we want to be sure that they are independent enough so 7 that there is not, you know, an intellectual lock. It is a 4
8 complicated question.
9 MR. SIESS: Bob, we don't assume that the to earthquake causes a LOCA, do we?
11 MR. BUDNITI: Oh, yes. You are talking about a 12 large double ended LOCA? That is an element of the
___ 13 analysis, but it turns out it is stout, that is you don't
_ u get them.
15 MR. SIESS: Is this one of the assumptions in i 16 this?
4 .
I 17 MR. BUDNITZ: Well, pump seal LOCAs, for example, is are an important aspect.
19 MR. SIESS: Okay. But I am talking about large 20 break LOCAs.
21 MR. BUDNITE: Let me remind everybody that in the g 22 seismic PRAs that have been done to date that issue has 23 been examined for each one, and in each one it was found
! 24 not to contribute, that is, it was stout.
25 MR. BENDER: But that is. based on a given level of l0
130
() i 2
earthquake related to the design basis.
MR. BUDNITZ: No.
3 MR. BENDER: I guess I am not all that close to a J
4 lot of seismic analyses, but my belief is that if you put 5 enough seismic loading on the piping, you can lay it on the 6 floor.
7 MR. BUDNITz: I will follow what Chet said. At a Zion and Indian Point they took the seismic loading up to I l
9 think 1.5 G or somethinc. and concluded with a dynamic io analysis that it was still okay, and that contributors were 11 other than that, as I remember.
i 12 I am just as sure as I am sitting here that they i3 did that.
i
() i4 MR. SIESS: I don't believe it was 1.5. I think 33 they cut off at around 7/10th.
16 MR. BUDNITZ: No, Chet. As I remember ---
17 MR. SIESS: I remember the cutoff curves.
is MR. BUDNITI: Well, you remember the containment 39 was found to fail at 1.1, and then they sharpened their 20 Pencil because they had made some conservative assumptions 21 and they found out it was 1.7 instead. So they did carry l
22 things up to that level.
23 MR. SIESS: They put in a ceiling, and they tried 24 to stop it at some earthquake level they didn't think was 25 reasonable.
O
131 k 1 MR. BUDNITZ: But they carried the containment up 2 to 1.7, as I remember on Unit 2, and I remember the piping a being done -- the dynamic analysis was conservative, but 4 they did the large LOCA piping up to 1.5, as I remember.
5 MR. SIESS: Well, that is not PGA.
6 MR. BUDNITZ: It was ground motion PGA, and then 7 they had to go through the structure in response.
8 MR. SIESS: Don, do you have any more to say?
9 MR. GUZY: No.
10 MR. SIESS: Any more questions.
l 11 (No responee.)
12 So the EPRI stuff tends to parallel the i3 implementation of the expert panel.
()
14 Now we have got one more thing scheduled for this is morning,'and that is Jim Costello. How long that take you, 16 Jin?
17 MR. COSTELLO: Well, probably if I ---
l 18 MR. SIESS: With no questions how long would it 19 take?
20 MR. COSTELLO: Oh, ten minutes.
21 MR. SIESS: Okay. Let's go ahead and do.that 22 before we break for lunch.
23 MR. MARK: Without questions?
24 (Laughter.)
25 MR. SIESS: No.
132
() i MR. COSTELLO: Well, in the interest of expediency and bearing in mind the fact that everyone has copies of 2
3 the viewgraphs, and we should be in the package right after 4
the material that Dan and Dr. Budnitz spoke to. l l
5 MR. SIESS: Can you just cite a viewgraph title?
6 MR. COSTELLO: It is entitled validation of 7 seismic response predictions.
8 MR. SIESS: Okay, go ahead.
9 10 11 12 4
15 16 17 18 19 20 21 22 23 24 25 l
O
. . I 133 7m MR. COSTELLO: This activity is one for which we 4
( j 2
continue to modify the name at every presentation. !
(Laughter) 3 MR. COSTELLO: The only thing that is constant in tihere is, the word " validation" comes in. The difficulty 5
is in finding a sufficiently short title that describes the activity that is involved in this program.but also doesn't 7
imply that there is more activity in here than is going on.
8 9
In the perspective given by Dan and Dr. Budnitz's presentation as well as the questions that they elicited, I 10 jy could summarize where this activity is focused.
g There are the three elements that are involved in either assessing risks or estimating margins from large y, earthquake events. The one is, of course, the hazard curve. The second one is getting from some representation 4 15
-of the motion to how the equipment and structures and j7 components will behave. The third part is the fragility.
18 Dr. O'Brien later on this afternoon will talk l about our efforts on fragility. We are not making any 39 explicit efforts in this branch of the research program on 20 g hazard curves although you will hear that, I am sure, from Dr. Murphy if he comes to talk to you.
22 We are f cusing on the part' in the middle. That 23 is, given s me representation of large earthquake motion, 24 how well can existing calculational methods which are used 25
( .
V i
i . . - .
e 134 l i project what's likely to happen in the vicinity of failure.
2 We are also not neglecting the question of how 3 g od would these calculational projections really have to 4 be in the light of uncertainties associated with the hazard 5
curve and uncertainties associated with fragility 6
estimates.
7 So we are proceeding more or less on a two-fold 8
path. Of course, the effort on how good they have to be is 9 mainly an analytical e f f s, t and a low-level effort, and we
'10 believe we can get by on that subject as long as we just 11 don't forget, as long as we continue to pursue that.
12.
Yes, sir?
13 MR. MARK: On this first viewgraph you mention 34 three contractors, Argonne, Brookhaven, and NBS.
15 MR. COSTELLO: Yes, sir.
I g MR. MARK: And three cooperating organizations.
37 MR. COSTELLO: Yes.
18 MR. MARK: Are funds transferred from the j9 contractor to the cooperating ~ organizations, or are they 20 contributing advice on their own?
21 MR. COSTELLO: As far as transferring funds, I can
~~
22 23 MR. MARK: I don't need to know the numbers.
24 MR. COSTELLO: On the matter of a direct transfer 25 f funds from the NRC to defray costs of an experimental
, w
- .. ._ . _ _ - _ = _ -
- - . _ _ - _ . - - . . .. - . . - ~ , -
135 3
acti vi ty , for example, the activity at KFK, we at the NRC transferred those funds directly. In fact, our ante 2
3 inv Ived putting up money to have a large shaker built, to 1
, contribute to the cost of that shaker t
I assume that.if we are going to have similar 5
g activities in Japan, the transfer of funds overseas will be I
by the NRC.
7 i
~
- " "Y # " "' " " ** "" '
8 9
MR. COSTELLO: .Yes,-sir.
4 I
MR. SIESS: Jim, you touched on something and I 10 yy guess I wasn't quite clear how you ended up.
12 There are clearly'large uncertainties in the t
+
hazard. There are large uncertainties in the fragilities g ~ of a component subjected to a known shaking. And what you c
are Concerned with here is how much a structure or l component shakes for a g'iven ground-motion shaking. You ;
fi 37 are trying to get'from %nat happens out there to what.
I 18 happens in here; right?-
t MR. COSTELLO:- Yes, sir.
l; 39 MR. SIESS: And you said~nothing'has been done-to l 20 4
l 21 address how well you need to.get from one place to the i
g .other. Have you considered that there may be sufficient unce a n y in the hazards and sufficient uncertainty.in 23
- 24 the fragilities that eliminating.all'the uncertainty'in between would still not be much of an improvement?
25
- n v
I
, , , . . ~ . . . , . . , . _ . _ , . . _ , . . . _ . _ . . . , _ . _ _ _ , _ _ _ . _ _ . _ . , _ , _ _ . , . _ _ _ _ . _ _ , . . , _ . _ _ . _ _ , _ . _ , ,, , . . , . _ . , _ .
136 MR. COSTELLO: That is in
)
1 2 MR. SIESS: In risk.
3 MR. COSTELLO: That's a prospect which has some 4 possibility. I might not say "all" but it might be that 5 there is very'little. That is what we do intend to 6 Pursue. I said, not forget about it.
7 MR. RICHARDSON: But keep in mind, we are not 8 addressing just risks, we are looking at the question of 1
9 margins which is a more deterministic --
to MR. SIESS: Yes, but --
11 MR. COSTELLO: But the question about how to use 12- it, how well you can use it, I think is indeed the focus of 13 our effort at NBS.
e s- 14 MR. SIESS: But really what you are looking at 15 here, when you say " seismic response," you are talking 16 really about getting from ground motions to floor spectra; 17 is that what we are talking about?
13 MR. COSTELLO: Yes, sir 19 MR. SIESS: And, of course, the struc'tural 20 analysts think they can do a great deal there and they are
- 21 deterministic. But when you get to the uncertainties, I'm 22 not sure what you accomplish is all that great. That's the 23 point I am trying to make philosophically.
l 24 MR. COSTELLO: Well, I think we have a little l
~
25 concern with it-ourselves. The question is, how good U)
V
l l
137
, () j- should a calculation really have to be, faced with certain That, indeed, is the focus of our effort at 2 uncertainties.
3 NBS.
4 MR. SIESS: Because usually the calculations 5
involve getting increasingly more complicated, bringing in 6 more variables, using bigger computers and so forth. A lot 7
of times the people that are making those aren't really 8
concerned about the uncertainties. They say, "Give me this 9 information and I'll turn that out."
10 If they improve it from plus or minus 50 percent
- is to plus or minus 10 percent, it really may be negligible.
12 MR. BENDER: Jim, I'm not clear. Is this purely 1
33 an analytical program at this stage?
ja MR. COSTELLO: No.
15 MR. BENDER: What experimental things are you 16 doing?
17 MR. COSTELLO:' The experimental undertakings are
- 18 three, mainly. One which is on going right now is a joint 39 effort with EPRI at a facility they built in Lotung, l
20 Taiwan.
21 MR. COSTELLO: I have a little extra viewgraph 22 which I may slip up here that may help.
23 (Viewgraph) 1
- 1. 24 MR. RICHARDSON: Those are.our three experiments.
25 MR. COSTELLO: This sort of summarizes the
(
I
138 That's the substance of what's going on in
) 1 activity.
2 Taiwan.
3 MR. BENDER: In that list up there in which you 4 are showing some work you are going to approach failures, I 5 understand it. Is that what you want to try to do or not?
6 MR. COSTELLO: Certainly, at HDR. Yes, in the 7 Phase-2 activity at HDR.
8 MR. BENDER: And failure means going up to large 9 displacements?
10 MR. COSTELLO: In elastic response, yes.
11 MR. BENDER: That's not quite the same thing, but
~
U 12 I presume that they may be nearly the same thing.
, 13 And in the Lotung thing, what are you doing?
s, 14 MR. COSTELLO: That's a joint activity with EPRI.
15 I'm sorry Dr. Kassawara isn't here --
16 MR.' BENDER: So am I.
17 MR. COSTELLO: -- to go into that. But EPRI has 18 constructed, in partnership with the Tai Power Company a 19 building which is an approximate quarter-scale model of a 20 PWR containment -- approximate in the sense that they have 21 mass distribution that is similar.
22 The facility has been built near a large 23 horizontal seismic array that is operated jointly under 24 sponsorship of the National Science Foundation and the j 25 Chines ~e equivalent of the National Academy of Sciences.
O{~1
e 139 1
( j That has been going for some time. !
2 What EPRI has added on is this building and --
3 arrays.
- 4 MR. BENDER
- But it's in an area subject to 5
frequent earthquakes.
6 MR. COSTELLO: Absolutely. That's the reason for 7 choosing it. There are a number of earthquakes recorded 8
there every year. The intention on this is to see how well 9 modeling techniques can predict what happens up into the 10 building from down, from input motions, including soil-11 structure interaction effects.
12 MR. BENDER: Have they designed the model, or l 13 whatever the thing is, so that in the kinds of earthquakes i G l k_) ja they see the structure.is very-near its structural limit; i
15 do you have any idea?
16 MR. COSTELLO: I don't believe we will get i7 failure.
18 MR. SIESS: This is for-soil-structure interaction 39 primarily.
20 MR. COSTELLO: I do not expect that we will get 21 failure in the building, and the building does not include 22 any significant components.
23 MR. BENDER: So it's#just a matter of the 24 transmission of the energy to the building.
l 25 MR. COSTELLO: Yes.
i
[
! (_)\
___. ._. _ _ . . . . . __ ___ _ _ .__m. _ . _ - -
i l
I 140 MR. RICHARDSON: But'it will start to perhaps
) 1
! 2 resolve this very sticky question of what is the correct e
I 3 modeling technique to account for soil-structure 4 interaction.
5 MR. BENDER: There are a lot of variables and 6 whether one installation will deal with all the variables, 7 especially if it doesn't see the size earthquakes we are 8 talking about, I think might still be debatable.
9 MR. SIESS: But Mike, it's an attempt to verify -
10 analytical methods and presumably if you could predict what i
! 11 happens there, you'd have a little more confidence in your i i j 12 soil-structure interaction.
13
- MR. BENDER' Every time I've heard this discussed 14 everybody says, "But we don't know the properties of the 4
l 15 materials at the loading condition" -- and I'm talking l 1
! 16 about the soil properties. That's going to be a debate, 1
i 17 still.
i 18 MR. COSTELLO: There will be soil samples taken.
L
- 19 MR.-BENDER
- I'm not talking about this particular 20 thing,- but translating this to something else. The things
21 that are most important.for the bigger events may not be as ;
i 22 important for these smaller events.
MR. SIESS: Well, this isn't necessarily related
[ 23 24 to the bigger events.
i 25 MR. COSTELLO: -But there is an expectation of
( ,
I '
e 141 l l
("\ j getting some pretty good size earthquakes at this gJ 2 facility. The hi story woul'd have i t so.
3 MR. BENDER: I'm not trying to fight with you, I'm 4 just trying to remind you of a few things.
5 MR. SIESS: Simquake is already completed. Have 6
y u gotten all out of that you think you can?
7 MR. COSTELLO: From the Simquake experiments, 8
yes. Now, what EPRI has done also -- although we are not 9 participating in this aspect of it -- has been to reproduce 10 one f those 112 cylindrical shapes that they had at ij Simquake and put it over here for some correlation 12 purposes.
13 MR. SIESS: Now, how does this differ from i4 Simquake, scale?
15 MR. COSTELLO: Scale and the fact that -- the real 16 dif f erence or the important dif f c: ence is that it is real 17 earthquake motion, not generated by explosives.
18 MR. SIESS: Of course, at Simquake they knew the j9 motion.
20 MR. COSTELLO: Measured it, yes.
j 21 MR. SIESS: And they could go from a known soil 22 motion to a known structure motion. And that's all that 23 s 11-structure interaction analyses give you anyway.
24 MR. COSTELLO: Oh, but the big difference was that 25 the motion was imposed by explosives --
O O
e 142 r"'s
() 1 2
MR. SIESS:
MR. COSTELLO:
Yes.
-- and' of rather short duration.
3 MR. SIESS: That's right.
4 MR. COSTELLO: And not at all typical of what you 5 would get from a large earthquake.
6 MR. SIESS: It was intended to be as typical as 7 they can make it with explosions, but you expect to get a 8 better, a more typical input.
9 MR. COSTELLO: Real earthquakes, yes, sir.
10 MR. SIESS: Different soil? -
11 MR. COSTELLO: Terribly different soil, very soft 12 site, which should have the effect of amplifying --
.., 13 MR. BENDER: Tadotsu, is that what it is?
i (s_)'\ 14 MR. COSTELLO: Tadotsu, yes.
15 MR. BENDER: That's actually a PWR configuration 16 that is being shaken?
17 MR. COSTELLO: In Tadotsu there are both PWR and 18 BWR components which are being shaken. We have a pretty 19 good table extracted in Appendix F of the seismic plan.
20 MR. SIESS: Does the ANCO stuff get into this 21 category? It's all finished, but it's part of this?
22 MR. COSTELLO: I beg your pardon, sir?
23 MR. SIESS: The ANCO?
24 MR. COSTELLO: Oh, ANCO's activity --
25 MR. SIESS: That comes under pipe somewhere?
A
= -- - . . - - ._ . . .- -
a I
143 (h i MR. COSTELLO: Yes.
2 MR. SIESS: Okay, we are talking structures. We 3 are talking structure input now.
4 MR. COSTELLO: Yes.
5 MR. SIESS: Now, you've got the CAT-1 structures 6 down there below a line.
, 7 MR. COSTELLO: Yes. The line is to f urther 8
clarify that there is more activity in the research program 9 which could be viewed as validation of seismic j
10 calculational methods. That's why I threw in the two big it ones.
12 MR. SIESS: I see.
i 13 MR. COSTELLO: The difference here is that what we 14 are focusing on in this program-is, in which you go out and 15 try and get agreements with other people to share the costs .
t 16 r benefits of large experiments.
17 As I say, the EPRI one is on-going right now. We 18 have, as part of our participation, contracted -- actually, 1
19 Argonne contracted on a competitive procurement with ANCO 20 to do low-level exitation of the base mat to get some low-l 21 level impedance data.
22 The building is-now poured. They will be going 23 again in the end of October for a week or so to put a
! ;4 shaker up on the building, again at very low level, to get -
I ~
25 4
e
(
144 1 MR. SIESS: What building?
2 MR. COSTELLO: The one quarter model is now 3 poured. ANCO will be over there on a subcontract to 4 Argonne to do Part 2, which is to get --
5 MR. SIESS: And that's a containment-type 6 building?
7 MR. COSTELLO: Containment-type building, yes, 8 sir.
9 MR. SIESS: Now, the CAP one studies have other 10 objectives, ultimate capacity of shear-wall structures.
11 But the one issue that has come out of them pretty strongly 12 is the question of shear-wall stiffness and what it does to
,,, 13 floor spectra, for example, i
(/T
(_ 14 Is there anywhere anything that could be done on a 15 larger' scale to verify that? It looks to me it's a couple 16 years down the line. I know they are working at scale.
17 MR. COSiELLO: Well, we are hoping that one of the 18 asterisks activities up there, the so-called package deal 19 with Miti will get us, if this agreement comes to fruition, 20 data on some pseudo-dynamic large shear-well test the
! 21 Japanese have run in the past, what they call so-called 22 " restoring force" experiments.
23 MR. SIESS: They have done some work.
24 MR. COSTELLO: Oh, yes. We are in pursuit of 25' that. We are i so looking at the possibility -- actually i
7
145
' ~
i Los Alamos is looking at the possibility that there may be 2 some benefit in looking at frequ2ncy shifting on heavy 3 concrete walls that will come from the large shaker
- . 4 experiments at HDR. It is not totally clear that it will 5 be beneficial, however.
6 There is a large wall directly under the shaker.
7 But the motions that that wall sees may not be sufficient 8
for that purpose and, indeed, since there is no 9
instrumentation embedded on the rebar in the wal-1 there may 10 n t be any conclusive outcome. But it's being explored as 11 a possibility.
12 MR. SIESS: I was thinking maybe you could put a 13 shaker on-top of the middle of the diesel generator
(( w) 34 building. It probably couldn't do any more damage than has 15 already been done.
16 I really wonder whether you can shake that shear-j7 wall building or not with one of those things.
ig MR. COSTELLO: Well, again, Los Alamos is now 39 looking at the German safety calculations and trying to 20 come up with some idea of what's expected directly under 21 that shaker. It 's inconclusive as yet, but they'll i
22 probably go after it.
23 MR., EBERSOLE: Don't these big diesels get used 24 in shipping vessels and others subject to the shocks of-25 -
transverse waves and other things and get a practical
\>
i l
e 146 demonstration of their integrity?
) 1 2 MR. COSTELLO: On diesel generators.
3 MR., EBERSOLE: I'm talking about the big diesels, 4 yes.
5 MR. COSTELLO: Well, I think that Dr. O'Brien --
6 on the matter of fragility, Dr. O'Brien will speak to that.
7 MR. SIESS: He's not talking about diesels, just 8 the structures.
9 MR. COSTELLO: Oh, just the building, how you get to there.
11 MR. SIESS: What he's talking about is what 12 shaking the structure can transmit to the components.
- 13 MR., EBERSOLE: Well, that's-analogous to the
(,,) 14 whole, transmitting it to the --
15 MR. SIESS: Yes, but-the question is, soil moves 16 ut here. How much does the structure --
17 MR., EBERSOLE: Yes, okay.
18 MR. SIESS: Then you go the next step --
19 MR., EBERSOLE: It's soil structural, yes.
20 MR. SIESS: Well, it's not just the soil, the 21 structure has to transmit it.
22 MR., EBERSOLE: Back'at the,3,000 horsepower level 23 the argument was always made, they used locomotives and 24 they really get the pounding, so they are good.
25 MR. SIESS: Well, I think somebody said earlier F-q b
l l
1 e , - , e
e 147 diesels have very high --
( i 2 MR. EBERSOLE: Yes. But that sort of changes, 3 Chet, when you jump above 30,000.
4 MR. SIESS: Well --
5 MR., EBERSOLE: They don't build them that big.
~
6 MR. SIESS: I still think it's the auxiliary that 7 are going to fail your diesel.
8 Mr., EBERSOLE: Yes. Oh, I'm sure it is.
9 MR. COSTELLO: I guess in order to give a clear jo picture, we still have to say that the agreements with the ij Japanese are a bit murky. It begins to look now, I think, 12 that we may not come to an agreement on the use of the 13 Tadotsu table on the grounds of cost, as Jim mentioned, s
s ja But on the other hand, as far as the information 15 exchange using existing Japanese data.and on-going Japanese g experiments as benchmark data for calculational methods, it 37 begins to look as if that will go through.
18 Is.that a fair estimate, Jim?
19 MR. SIESS: Something I said earlier may not 20 really have been correct about how accurate you need to be 21 to be able to predict these things.
22 This area is chiefly concerned with finding out 23 how good the methods are, not with improving them.
1 24 MR. COSTELLO: That's correct.
25 MR. SIESS: You might find out that they are good
()
m
e 148 enough within plus or minus_something, and not bother to
) 1 2 improve them.
3 MR. RICHARDSON: That's right exactly.
4 MR. COSTELLO: Yes, absolutely.
5 MR. SIESS: 'Just quantifying it may be enough to 6 fit it in'to the general level of uncertainty and see what 7 it means.
8 MR. RICHARDSON: We would like to have experiments 9 enough that we challenge those codes to see what is the 10 uncertainty band in the calculational methods.
11 MR. SIESS: And it's not just'that.
12 MR. COSTELLO: Is there something systematically
_ 13 .wr ng.
( N
(_,) 14 MR. SIESS: A lot of those calculations are made 15 with the_ usual conservatisms which I'm not sure always 16 intended to be deviations from reality, but they usually-17 end up being deviations'from reality.
18 1 assume in the validation process you will be 19 using as close to real things as_you can, testing just the 20 methodology.
21 MR. COSTELLO: Yes.
22 MR. SIESS: But even the methodology has got 23 conservatisms built _into it.
24 Is that all you wanted to - you didn't go through 25 all of these.
fy
\/'
e 149
- MR. COSTELLO
- No, I didn't think so. Again, if I%))
2 you would like a little bit of the why we are doing it, 3 it's on the second page.
4 MR. SIESS: I think we've covered that pretty 5 well.
6 MR. COSTELLO: And where we are today --
7 MR. SIESS: Yes, why don't you run through those 8
pages briefly.
9 MR. COSTELLO: -- is on the third page'in the 10 write-up, n MR. SIESS: When do you expect to'get results? I 12 think the Taiwan thing only in85. I assume it's there 13 for a longer period of time.
O
'Q i4 MR. COSTELLO: Oh, yes.
15 MR. SIESS: Once you've put it in, there is 16 n thing to do but measure.
37 MR. COSTELLO: Shortly, we are now going to be 18 collaborating with EPRI in '86 on lining up people to make i9 predictions for what's going to happen.
20 MR. SIESS: That's Taiwan, the second bullet in 21
'86.
22 -MR. COSTELLO - Yes, the second bullet in '86.
23 MR. SIESS: Okay.
24 MR. COSTELLO: And '87 and on, it shows up in the 2'S third bullet. That is, we are assuming that nature will be o
i
. = , - - _ _ . _
e 150 P)
(%,' 1 agreeable enough to give us'a nice ensemble of earthquakes 2 over the next few years.
3 EPRI has a commitment with Tai Power to sit it out 4 for five years, waiting for the good earthquakes. So we 5 will go through an interim scheme in which different 6 parties are engaged to predict what the earthquake response 7 will be.
8 Then, in the assessment phase, we will go back and 9 mutually compare on a round-robin basis, to MR. SIESS: What information have the Japanese 11 gotten so far from their real plants? They all instrument 12 it, I'm sure, and at least one site, I thought, got --
, 13 MR. COSTELLO: Well, there is data from two
_ 14 earthquakes at the Fukushima plant, and certain results 15 have been reported in the SMiRT papers and the 7th and 8th 16 floor conference papers, primarily Dr. Tanaka from Tokyo 17 Electric.
is What they seem to have found out, and we had 19 somewhat independently confirmed that using a limited 20 amount of data that they made available to us, was that the 21 convolution-deconvolution procedure does give you a fair 22 estimate of how motion attenuates with depth, and the 23 general trends are correct.
24 .MR. SIESS: What about soil-structure interaction, i
25 locking, torsion?
N iN ), \
151 MR. COSTELLO: That was not available from the
('))
(
i 2 data they made available to us and may not be from their 3 instrumentation.
4 MR. SIESS: Does your information exchange 5 agreement with Miti try to get any of that,'that they might 6 have?
7 MR. COSTELLO: Yes, sir.
8 MR. RICHARDSON: To a point, except -- well, a lot 9 of that data is owned by the utilities and that's harder to 1
in get at. What we are initially after and what Miti has in tentatively agreed are data which they-have direct control 12 over.
13 MR. SIESS: Why would the utilities be reluctant y ja to --
15 MR. RICHARDSON: Oh, I think for the same reason g that U.S. utilities are reluctant to --
37 MR. SIESS: I mean, they came through the thing --
is MR. RICHARDSON: Well, I think there is always the 39 lingering doubt that data will.be misused or some safety 20 question will surface that would be hard for them to 21 address.
22 MR. SIESS: You mean the Japanese have some data 23 they won't even'give to Miti?
24 MR. RICHARDSON: That's --
25 MR. SIESS: I must misunderstand, I thought their l
(~
l N./
i l l r . .
1
152
) 1 system worked differently.
V 2 MR. RICHARDSON: Well, it certainly is different 3 and it's a lot more incestuous than ours. But I think 4 there is some data that is held pretty tightly by the 5 utilities. .
6 MR. COSTELLO: If you would like me to continue on 7 the '85, '86, '87 sequence.
8 MR. SIESS: Sure.
9 MR. COSTELLO: On the HDR activity, I did not 10 mention it in the accomplishment. What we did, we pointed 11 up 500,000 as our share for the construction of the large 12 shaker which will be activated in June, the large coastdown
! e. 13 shaker which will provide the input for the experiments in
/
kss 14 this year. In 1987, we will participate in the small 15 shaker local exitation to the inelastic range of piping 16 loads at HDR.
17 I have not mentioned it explicitly, but of course 18 there will be pre-test calculations, post-test 19 calculations, and benchmarking activities.
20 Finally, in '85 we have looked at two options in 21 Japan, two proposals in Japan. One was the Tadotsu 22 extension of the test of the Tadotsu. facility and 23 modification of a specimen to assure that it could be 24 brought into the inelastic range. That's the one that Jim 25 says is currently f oundering on cost.
(~-
l
e I
l 153
) i 'It's worth only so much to us, the way we see it,
. 2 and it seems that current thinking is it will. cost them too 3 much for their end.
4 That's a clearly forward-looking activity, the 5 test. It would occur and would take place in ~87 after 6
they are finished with this specimen themselves.
7 The information exchange is really more backward 8
looking. There are packages of test. data and experimental 9 programs that they have run for a number of years. The 10 data are in hand. We are exploring the possibility of ti using these data to benchmark response calculations.
12 MR. SIESS: Have=they given you a price on using 13 Tadotsu?
(,) ja MR. RICHARDSON: Oh, yes.
15 MR. COSTELLO: Oh, yes.
d g MR. RICHARDSON: It costs, to run the facility, 17 about.$20,000 a day, $20,000 to $25,000 a day. That is doesn't include depreciation of the facility itself, that's 19 just the operating costs.
20 This set of experiments that we have proposed --
21 MR. SIESS: They can build four automobiles for 22 that, too.
23 (Laughter) 24 MR. RICHARDSON: They are claiming costs of about 25 ' S2 million to run thi s test into the inelastic range. We (s_
e 154
\
have made an offer. Their initial reaction was, they are v; I 2 after thermal-hydrolic codes and that is what is under 3 negotiation right now.
4 MR. SIESS: It seems to me that selling those 5 codes has been one of the most important benefits we have 6 gotten from that old code. What was it, half a billion 7 dollars Bob said? I think he's off by a factor of two.
8 MR. RICHARDSON: Yes. Our office director is 9 contemplating that right now.
to MR. SIESS: Good.
11 MR. COSTELLO: And, of course, on the last page we 12 simply focus on where in the regulatory agenda we suspect
_, 13 that these results will have results in a few years.
_ 14 MR. SIESS: Any other questions?
15 (No response) 16 MR. SIESS: I think this program is really coming 17 along. I don't know how soon we are going to get some 18 results.
19 What's your feeling right now about how well we 20 are able to predict frequencies for flow responses? Some 21 of the early HDR stuff said, you know, we were off on-22 frequency by quite a bit.
23 MR. COSTELLO: I expect that we will find out that 24 at significant shaking levels there is more shifting than 25 we might have originally thought.
7..
U l
I i
4 155 4
4 r s l 3 As to how important that is, is as yet unclear.
.i s 2 MR. SIESS: That gets back into how essential the l
t
} 3 fragilities-are and I'm sure that is looked at.
i 4 Okay, thanks, Jim. !
-I i 5 MR. COSTELLO: Yes, sir. .
l 6 MR.-SIESS: We'll take a lunch break now and be ,
i 7
back at 1:15.
t- 8 (Whereupon, at 12:15 p.m., a luncheon recess was r 1
3 9 'taken until.1:15 p.m. on.the same day.)
(
1 j 10 11 -r r
i
- f 2 1
1 13 1
i 14 i 15 5
.i i 16
'17 i i
- ; 18 i
! 19
, 20
?
l i
-- 22 1
l l 23 i ,
24 I
l 25 L
..,.e._,--. , , - , . , - .-em.,.,--, ,7.m,, yyy., , . .m ,.,,.-e.,,,,,_.-.,,_,..m,.- ,,-.m., . . , _,,,,_,,-w,w;,vyg.-,.m _,%- - -.-
e 156 4
FT
() 1 AFTERNOON SESSION 2 (1:18 p.m.)
3 MR. SIESS: Let's see, we'll now have John --
4 MR. RICHARDSON: John O'Brien.
5 MR. SIESS: -- on component fragilities. That's 6 an awfully negative approach, isn't it?
7 (Laughter) 8 MR. O'BRIEN: I'd like to wish'you all goou 9 afternoon. My name is John O'Brien and I'm here to talk to i
10 you about our component fragility program.
11 (Slide) 12 MR. O'BRIEN: The first slide is an announcement
, ,- 13 that we will continue with Brookhaven and not continue with
/
k_,x) 14 Livermore which this year is funded-to the extent'of 515K.
15 We have tried to keep this a secret because we 16 want to continue to encourage Livermore to try to impress 17 'u s , and I'd be pleased if everybody here keeps our secret 4
18 so they don *t. find out,that'they are out of loot.
19 MR. SIESS: I won't tell Charlie.
20 MR. O'BRIENs Okay. Charlie is the winner, i
21 though.
22 MR. SIESS: Yes, I know.
23 MR. O'BRIEN: Charlie doesn't know he is the 24 winner.
25 MR. SIESS: Yes. He is one of my students, you
(
O
1 1
157
-s know.
) I 2 MR. O'BRIEN: .There is a considerable sum of money 3 involved here. What you don't see on this slide is the 4 515K f or Li vermore, and they actually have been very 5 successful. I'm going to take the liberty of briefly.
6 describing some of their successes this year.
7 The issue is more reliable and realistic component-8 fragility information, and the second sentence-and the key.
9 thing is that we think we have enough data right today to 10 show that component fragility is used in the dozen or so ij seismic PRAs or pessimistic estimates.
12 The source of our data is Wile Laboratories and i 13 some data they have provided for EPRI.
34 If you ask what the money in '86 and '87 is geared j$ to, is to provide a better technical base, information base 16 not based on theory but on test data. That that second 17 sentence there is true that the seismic component is fragilities is indeed pessimistic.
j9 I am going to try and impress on everybody that we ,
20 have had a phobia about earthquakes in the nuclear 21 industry.
! 22 MR. RODABAUGH John, could you name five or-six 23 f the components you are going to be testing?
24 MR. O'BRIEN: Yes. They are all electrical.
25 MR. RODABAUGH: Are'all electrical.
. n m
l l
r
e 158 MR. O'BRIEN: We have looked at it all, but you
) 1 2 will see in short order that we believe it 's -the electrical 3 stuff if you exclude civil structures, the seismic 4 fragility of civil structures, and all you need to worry 5 about is electrical. Mechanical is very rugged for a 6 number of reasons.
7 MR. EBERSOLE: What's the common denominator that 8 gives you all this ruggedness that you anticipate? I don't 9 even think -- is it in the procurement specifications?
10 MR. O'BRIEN: No, it's because mechanical 11 components are designed typically for pressure and 12 temperature anyway, so --
_, 13 MR. EBERSOLE: Well, I'm talking about the
, 14 electric apparatus. What gives you --
^
15 MR. O'BRIEN: Oh, they are not rugged.
16 MR. EBERSOLEs - That's what I mean. But even to 17 the extent that you' find that they will withstand -- since 18 these components are not standardized with any common 19 language other than, you know, individual seismic 20 acceleration requirements, how can you just, on a. common l 21 basis, find that they are resistant to earthquakes beyond 22 the design basis?
23 MR. O'BRIEN: We don't find that they are 24 terrifically resistant to earthquakes. That's why we are 25 focusing our efforts on them particularly with an aim I
[d
e i
159
() 2 i eventually to decide which-characteristics make them more resistant to earthquakes. However, not going to the extent 3 that you make them super-resistant against earthquakes but 4 then fail for environmental reasons.
5 I mean, you've got to balance it, right?
6 MR. EBERSOLE: Right.
{
7 MR. O'BRIEN: You don't want to choose seismic 8
ruggedness at the expense of something else.
i 9 MR. SIESS: John, you said most of your data is jo coming from Wyle?
11 MR. O'BRIEN: No. No, not most of it. I had this 12 sentence here and I had to identify for you the source of 13 that, and it's information prepared by Wyle Labs for EPRI.
! ) ja MR. SIESS: Oh, I see.
15 MR. O'BRIEN: Which is now made available to us.
16 MR. SIESS: How much of the work has been done by l 17 Wyle, qualification testing?
ig MR. O'BRIEN: A lot but not -- it turns out that 39 the utilities have done it themselves; the vendors have 20 done it themselves. I would say that Wyle represents a 21 small fraction of our total data base, maybe less than a 22 quarter.
23 MR. SIESS: The utilities have had their own shake 24 tables?
25 MR. O'BRIEN: des. And it turns out that O
V l
l
160 h 1 utilities like Pacific Gas & Electric and Commonwealth and
%)
2 WPPS -- WPPS has been wonderful in providing meaningful 3 information to us.
4 MR. SIESS: All right, that's interesting. Thank 5 you.
6 MR. MARK: I believe you said that you hope to i
j 7 confirm that the assumptions used were indced pessimistic.
8 MR. O'BRIEN: No, not the assumptions. I hope 9 what I-said was that this second sentence here is born from 10 information that we have received from Wyle Laboratories 11 under a cooperative agreement with EPRI, suggesting that 12 component fragilities are at least twice what is used now 13 in seismic PRAs.-
[']/
, (,, 14 MR. MARK: Well, isn't that to confirm that they 15 are indeed pessimistic, that you hope to do the work?
16 MR. O'BRIEN: Yes.
17 MR. MARK: And having confirmed that --
- 18 MR. O'BRIEN Hold it. We got it from one source, i 19 and I'm willing to quit but not everybody else is willing i
l 20 to quit.
l 21 (Laughter) 22 MR. MARK: Right. Supposing you do confirm --
i 23 MR. O'BRIENs Happy days are here.
24 MR. MARK: Are you then proposing that more 25 realistic values be used in PRAs?
s
m - a -
161
\
.(d i MR. O'BRIEN: Yes.
2 MR. SIESS: It can be a hope, anyway.
3 MR. O'BRIEN: This program is driven by East Coast 4 seismicity. If you ask what's the regulatory impact on 5
this, it's to respond to a Charleston earthquake, a New 6
Brunswick earthquake, the New Madrid earthquake questions.
7 If those are false questions, then this effort doesn't need to exist, we can discontinue it now. If you 8
9 believe that is a problem and the potential for shutting to down reactors on the East Coast is a real potential, then
- j it would be good to have this information in your back i
12 p cket to stop or continue operating reactors, j3 MR. MARK: I believe it's a problem if indeed they
{ 34 are using pessimistic assumptions in getting out the fact j$ that seismic risk dominates the total risk. That's a i
16 disease that it would be nice to get over.
17 MR. O'BRIEN: Yes, I agree.
ig MR. SIESS: John, when they do a seismic i9 qualification on a piece of equipment, they subject it to 20 varying inputs, I know, presumably up to some level beyond 21 that that it has to withstand in the plant.
22 Now, an East Coast plant is designed for around
, 23 two-tenths "G" maximum and we've got some West Coast plants 24 that were designed for .67, .75 ground acceleration.
25 Is there different equipment in those plants than i
J L
e i 162 r
'f %
in the East Coast plants?
{v) 1 2 MR. O*BRIEN: No.
3 MR. SIESS: And that equipment was tested to much 4 higher "G" levels.
5 MR. O'BRIEN: That's'a good case for keeping East 6 Coast plants running, the seismic upgrading.
4
- 7 MR. SIESS
- But none of it was ever carried to 8 failure, in effect.
9 MR. O'BRIEN: Yes.
10 MR. SIESS: I mean, some of it failed in the test 1
11 because something was not attached properly and they went 12 back and fixed it, I'm sure.
+
13 MR. O'BRIEN: We have achieved limited seismic 14 failures. In most cases, we have run to the capacity of 15 the equipment and the test ended --
16 MR. SIESS: Yes.
17 MR. O'BRIEN: -- because we couldn't put more load 18 into it.
19 MR. SIESS: But these weren't qualification tests?
i 20 MR. O'BRIEN: Some were proof tests or 5
21 qualification tests. Others were actually attempts to go 22 as far as you could.
23 MR. SIESS: Okay. But in the qualification tests 24 there was usually no attempt to find the fragility. It was 25 an attempt to qualify it to some predetermined level; l
U
.-e -
, w.., - -v - - + ,--w-- e
-w , - - . .
-- __- . . - _ - - .. . _ _ - ~ - - -- . --
e 163 I
() I 2
right?
MR. O'BRIEN: That is the definition of a i 3 qualification test.
4 MR. SIESS: Yes.
5 MR. O'BRIEN: What we have learned, Chet, is that 6 that a lot of people, after they finished the quality test 7 said, "Let's go all the way."
8 MR. SIESS: You found that.
9 MR. O'BRIENs That's what I want to rerort to you jo today.
11 MR. SIESS: Yes.
12 MR. RICHARDSON: But there is also the prospect 13 that you will have a component that is so rugged that you
, g_f i4 demonstrate that it's capable of sustaining accelerations 15 far above design where it's no longer of interest what its 16 functional l i mi t is. It's just so far above the design 17 that we really don't card precisely the failure level.
18 MR. O'BRIEN: Yen.
19 MR. SIESS: You know, you get a sort of a feeling 20 when you look at some of these shake tables and what they f 21 can do, that if you can 't bust . it on that shake table, it I'
22 ain't likely that an earthquake can do it.
t 23 (Laughter) 24 MR. SIESS: That's rather crude, but I --
25 MR. RICHARDSON: Yes, that's right.
I
(
s_J i
e 164 1 71 MR. O'BRIEN: Well, to continue. Our objectives
) 1 2 are as indicated on this slide here.
3 (Slide) 4 We have succeeded with cooperation with EPRI. We 5 are more interested in domestic cooperation than foreign
.6 cooperation because they have different electrical designs, 7 different voltages, frequencies and so on.
8 The third bullet here will be completed this year -
9 - as in fact they all will. But this is a Livermore to objective and I just put it there to complete the picture.
11 (Slide) 12 These are future objectives. These objectives are 13 in agreement with what Bob Budnitz was talking about this O' 14 morning for '86 and '87.
15 The key thing is, the first bullet on the
, 16 - objectives is, we intend to depend more on data acquisition 17 than testing. That partly explains why we reduced the 18 dollar figures because it's easier to buy data than create 19 your own data. We are paying two or three cents on the 20 dollar. That means we'll buy a million dollars worth of 21 data from Westinghouse or something like that and pay 22 $20,000 or $30,000 for it.
23 Moreover, we are getting data on equipment 24 actually installed. We are getting data from tests 25 performed in the '70s on equipment.that was installed in a v
1 1
165 I think that's a key j w ide number of nuclear power plants.
technical success, it's a business success. It's not a 2 in that a very significant one success but, nonetheless, 3 money to get the government doesn't need to spend a lot of 4
5 information.
You might not always get the information you want 6
the information as well in the sense that you don't control 7
But it's the way we are as if you had created it yourself.
8 going.
We are going to focus more on data acquisition and 9
testing.
10 ahead of It also explains why Brookhaven ran 11 12 Livermore.
Our scope includes electrical and mechanical 13 electrical components. We are I ,
ja components with emphasis on
(/ components.
focusing on both active and passive electrical i$
g Piping is excluded because that's covered by another j7 program.
It has become, in the course of events, an ig At one time, we thought about exclusively seismic program.
39 It has water hammer-loads.
throwing in pipe rupture loads, 20 an exclusively now become, as a result of NRR requests, concurrent loads are considered.
21 However, seismic program.
22 N w, our accomplishment since this program began 23 people with whom we have purchased data 24 in '84 is a list of We have had people offer us data free as 25 fr m Brookhaven.
.~
h
e 166
) well.
2 The Wyle data I referred to earlier came via our 3 cooperative agreement with EPRI. There is a common data 4 format. Our contractor, Brookhaven, and EPRI contractor 5
ANCO Engineers, have divided the country into two-and we 6
are collecting data and exchanging data.
7 We had a three-day workshop with 22 presentations.
A NUREG should have been out last week, 8
wasn't. But it's due out very soon.
9 10 I think what I'm most proco of is the fourth-bullet there, our prioritization scheme, which was a 11 Livermore, not a Brookhaven accomplishment and I want to 12 j3 show you what we learned.
On that third bullet, is that
( J 14 MR. SIESS: Yes.
15 the workshop that was held the same week as our ACRS 16 meeting?
MR. O'BRIEN: Yes.
U MR. SIESS: Okay. That's twice you guys have done 18 19 that.
20 (Laughter)
MR. RICHARDSON: That wasn't by design, Chet.
21 22 (Laughter)
MR. MARK: On the last sentence of the scope, let 23 24 me suppose' that some component with present assumptions is Then you 25 helping to govern the total systems fragility.
(W, s/
l l
.--p,
I 167
( ) i don't want to stop just when you show that that assumption 2 is exceeded, you would like to know how much.beyond that 3 you could go.
4 MR. O'BRIEN: Beyond you could go.
5 MR. MARK: Now, if it already isn't causing.any 6 trouble, you can stop, I would agree.
7 MR. O'BRIEN: All right. What you point out is 8
very true. What we are trying to demonstrate, greater than 9 presently assumed but not just delta -- a very small delta -
jo - but a large delta so that we can discredit it as a risk it contributor.
1 12 MR. SIESS: That's much greater than you think 13 they are.
14 MR. O'BRIEN: Yes. But we decided now not to go
'i j$ all the way to failure if that's a very expensive -
16 Proposition, which it could be. You don't actually have to 17 achieve failure. Just show that it's not a significant 18 contributor to overall core-melt probability.
i9 MR. EBERSOLE: Yes.
20 MR. SHEWMON: Is it relatively straight-forward to 21 go back and change the -- recalculate the seismic'PRAs from 22 earlier studies? You said it's now believed that 23 pessimistic estimates come through undue conservatism in 24 fragilities.
25 How hard is it to go back and lay that to rest I
U i
i l.
e 168 p%
g ) I with some certainty when you have them?
'J 2 MR. O'BRIEN: I think what it means is, if this 3 curve is still here, one curve expresses load and one that 4 expresses fragility. And what you are saying, if we shift 5 this curve, if the fragility curve is shifted there, they 6 will have to do that integration.
7 But it doesn't strike me as a very difficult 8 calculation if the old calculation is still available in 9 terms of the computer code and all you have to do is change 10 the input and descriptive, the mean value of that code and 11 run it again.
12 MR. SIESS: In a real risk study you have to go
, 13 back and --
14 MR. O'LRIEN: If it already exists and its safe, 15 you load it with new inputs, it shouldn't be a difficult 16 problem.
17 MR. SIESS: But if you were looking at the kind of 18 stuff the expert panel was and just establishing the 19 seismic capacity, that's fairly easy to feed in new 20 fragilities to.
21 John, let me interrupt you a minute before you get 22 further on with what you are doing. You mentioned you are 23 not going to talk about piping because somebody else is 24 doing that.-
25 But what about structures?- It seems to me that
/ * 'l O
e L
I
e 169
() i some of the conservatisms that we have gotten out of seismic PRAs like the pump house at Zion is simply because 2
3 people are unable or unwilling to try to describe or define 4 the fragility of a structure in terms of - you know, we 5
see it on masonry walls, assume that masonry wall comes 6 down, everything comes down.
7 The ceiling at Indian Point is overstressed, 8
therefore it falls down and kills everybody.
9 Is there any attempt anywhere to have somebody try 10 to trace these things out a'little more carefully and ti understand how --
12 MR. D*BRIEN: People are trying to do that.
33 MR. SIESS: They are?
14 MR. O'BPIEN: I'm sort of worried about things i$ like that, ceiling tiles falling down in the control room 16 building, unequal settlement of pump houses, things like 17 that which we do not assume will occur. We just calculate, ig you know, a CPA for a component and we subject it to that j9 and it survives.
l 20 But if, obviously, a block wall next to it falls 21 n a component, that's -- you know --
22 MR. SIESS: It may or may not damage it.
23 (Laughter) 24 MR. SHEWMON: I had the impression that Chet's 25 point was not whether the wall was going to fall on the
- - - = - . - - - _ _ - . - - -
170 pm component, but whether if indeed some of the tiles did fall
( } )
2 out of the roof in the control room that it would make any 3 difference to the operator action.
4 MR. SIESS: They get to a sort of conditional 5 probability. I'm not sure what will happen beyond this, 6 let's assume the worst. You know, the crib-house roof --
7 MR. RICHARDSON: I think the answer to that in the 8 context of our research is, no, we are not conducting any 9 research along those lines. But we would, if in our margin 10 studies it turns out that that is a very sensitive 11 parameter, we may then go in and do further analysis.
12 MR. SIESS: See, at Zion that -- did that turn out
,_ 13 'to be a dominant contributor?
, _ 14 MR. RICHARDSON: Yes, the crib-house was dominant.
i 15 MR. EBERSOLE: Chet, I wonder if they noticed that 16 in such plants as Diablo the emergency procedure manuals 17 were in bookcases with rolling casters, sure to be thrown 18 all over the pace and being "untangleable."
- 19 MR. SIESS
- You know what the answer I got to that 20 was?
! 21 MR. EBERSOLE: What.
22 MR. SIESS: They are not supposed to read those 23 things in emergencies, you are supposed to have that 24 memorized.
25 MR. EBERSOLE: Whose memory is going to run i
k w
1
- -- -g-, + - - - , , , . - y - - , - - , , - , , ,-,e, - - , - . - - w-w- . - - , - _ . -
e i
171
() 2 i through all those steps?
MR. SIESS: What's worse, I have seen them in r 3 -lo se-leaf notebooks sitting on an open shelf.
4 , MR. EBERSOLE: That's a lot of horse. To memorize j
5 all that, you've got to have the written rituals.
f 6 MR. SIESS: I have often worried about chairs on 7
casters in control rooms.
l 8 Gunther?
.l i
9 MR. ARNDT: I would just like to add that a little jo later on I'll give you some updating on some work that we si are doing on establishing design margins and capacities in 12 the structures which we'll input PRA studies for i3 structures. I'll give you a little bit more on that later.
ja MR. SIESS: Well, my problem is that I think some i
15 of these people can't figure out how the structure fails 16 physically, so they make the easiest assumption. And .
l 37 sometimes, you know, if the probability of taking out six 18 pumps was changed from one to one-fifth, it would no longer
( ig be dominant.
I 20 MR. RICHARDSON: That's right. .But we are not 21 doing that on a generic basis, we are trying to look at 22 each structure.
l 23 MR. SIESS: No, the thing would be to encourage i 24 people to not make those kinds of assumptions when they are 25 doing PRAs. We don 't do it in other -- you know, the i
. . _ ~ _ - _ . . _ . _ ~ - - , _ _ _ . - _ _ _ _ _ _ _ . - _ - _ -_. . . _ . . . . _ _ . _
e 4
172 parallel is saying that when a rod reaches 2,200 degrees at
) 1 2 any point along its length, you've got a core melt. We 3 have quit doing that, I think. That was WASH-1400, right?
4 MR. RICHARDSON: Right.
5 MR. SIESS: Okay, excuse the interruption. , Go 1
6 ahead.
7 MR. O'BRIEN: One of our accomplishments this year 8 -- and you don't have this slide and this comes from 9 Livermore -- is our ranked list.
10 MR. SIESS: We don't have it. You've got to move 11 over a little bit to the rights maybe even move it up.
12 (Slide) 13 MR. O'BRIEN: This is car ranked list of key
(. 14 components. That is components which, if they fail, will
' 15 lead to system failure. And these components in general u have low seismic capacity. Engineered safety feature 17 sequences --
18 MR. EBERSOLE: Pardon me a moment. Are you using 19 the broad interpretation of what an engineered safety 20 feature is again, on account of in the FSARs they are 21 limited to about three or four systems designed for LOCAs, 22 period.
23 MR. O'BRIEN: Gee, I could fake an answer, but I 24 don't know the answer.
25 MR. EBERSOLE: Well, I think you better broaden it O
l l
173 I
i for shut-down heat removal system sequences.
rU' MR. SIESS: Well, he's only got it up there, 2
3 Jesse, on sequences. Are there sequences on anything else?
4 MR. EBERSOLE: I'm just saying, you know, that l
5 ESFs are a discreet and limited set'of systems for specific 6 purposes, not including shut-down heat removal in a non-7 accident condition, i 8 MR. SIESS: But if those are the only features 9 that have sequences, he's not wrong.
4 10 MR. EBERSOLE: Oh, there are others that have
- sequences. ,
12 MR. SIESS: Well, that's what I'm asking.
j3 MR. EBERSOLE: Like aux feedwater.
( 34 MR. O'BRIEN: I could tell you where it comes 15 from. It comes from two systems. It comes from the on-16 site power system and it comes from the nuclear 17 instrumentation control systems.
i 18 MR. EBERSOLE: Well, I'm just, you know, objecting 39 to the limited terminology of ESF as used in the standard 20 context.
j 21 MR. SHEWMON: Is the sequence there a set of 22 switches?
23 MR. O'BRIEN: .No. Sequence there is something 24 that delays action. You know, it's a timer.
25 MR. EBERSOLE: Yes.
!O
- . = _ _ .
e 174 I)
(J 1 MR. O'BRIEN: It's something --
2 MR. EBERSOLE: Just a program of the electrical 3 power system so it won't stall out.
4 MR. O'BRIEN: And it's part of -- it comes from 5 two systems, the on-site power system. Look at the key 6 . assumptions, by the way. These could be crippling.
7 MR. EBERSOLE: Yes.
8 MR. O'BRIEN: We, in revising this list, have 9 ignored system redundancy. The reason we did it, this is a 10 generic list-good for BWRs, PWRs, all kinds.
11 MR. EBERSOLE: Why don't you just say load 12 sequences?
,_ 13 MR. O'BRIEN: I just call it sequences.
, 14 MR. EBERSOLE: Yes, sure; that's better.
15 MR. SIESS: On his list he can change anything he 16 wants.
17 MR. O'BRIEN: Unlike what Bob Budnitz said this 18 morning, we assume no pipe rupture. And the reascn we 19 assume it is because we thought it was a good assumption.
20 But if you look through this whole list, you 21 cannot find any mechanical items at all. We looked at 22 pumps, motors, generators. There is nothing mechanical in 23 the whole shebang, they are all electrical stuff.
24 MR. SIESS: Is a breaker mechanical?
25 MR. O'BRIEN: No, I would say --
175
( i MR. SIESS: That's electrical. It's a mechanical 2
device in an electrical system.
3 MR. O'BRIEN: A breaker would be right here, it 4 could be right here, the motor control --
5 MR. EBERSOLE: What's a valve, is it electrical or 6 mechanical?
7 MR. O'BRIEN: For the British, it's electrical and a
for us --
9 MR. EBERSOLE: Well, as a matter of fact, there is jo a demarcation point where the mechanic goes one.way and the ja electrician goes the other, and it makes a mess.
12 MR. SIESS: And never the twain shell meet.
33 MR. O'BRIEN: The key thing is that all these 34 items are from two systems, the on-site electrica7 power 15 distribution system and the nuclear instrumentation control 16 system. That's a research result.
17 This is an important list which would only be ig important for seismic fragilities. You would come up with j9 a different list if you were looking for something else 20 like aging, for_ instance. We know aging has different l 21 imPortance lists.-
22 (Slide) 23 We are starting testing tomorrow at Norco, 24 California and we have a three-bay motor control center 25 loaded with electrical components. I want to take very
("~\
V l
l l
l
\
I 176 1 briefly a few minutes to show you what we are doing.
2 This is sort of like a progress report. We have a 3 five-star Westinghouse motor control --
4 MR. SIESS: Raise it up, please. Whenever we 5 don't have it, raise it.
6 MR. O'BRIEN: We have a five-star three bay 7 Westinghouse electric motor control center. We are 8 investigating anchorage. Everybody believes an anchorage 9 is a key issue in component fragilities. You can take the to same thing anchored different and the component fragilities 11 go up by.an order of magnitude.
12 We are going to excite things off the shake table 13 with relays and circuit breakers on the motor control (R
g ,) 14 center, slide motors off the table.
15 We thought about putting the motor on the table.
16 That becomes a little difficult. In the real situation, 17 the motor will be shaken off, but we didn't do that. And is here is the way we are going to -- we've got four different
- ig configurations. A column is.the same as a bay, and we are 20 going to have it flexible and stiff.
21 That.means you add another brace inside to find 22 out how the stiffness of the cabinet affects component 23 fragility. That's the second parameter that we are 24 varying.
25 MR. SIESS: That top brace is internal?
(v~$
e 177
. ( ) MR. O'BRIEN: This is not internal. This i s --
l \s 2 internal stiffening is something provided by the 3 manufacturer.
4 MR. SIESS: Okay, I'm sorry.
5 MR. O'BRIEN: This is provided extra externally.
6 MR. SIESS: Let me be sure I understand. In a 7
motor control center, are these things.about six feet high 8 that have doors on them and dials on the. front? Okay.
9 MR. EBERSOLE: Do they have doors?
10 MR. O'BRIEN: Frequent failure mode is, the' doors
- ti come loose. You've got a 15-cent lock holding them.
12 MR. EBERSOLE: Well, as a matter of fact what they
- 13 do, some of them are designed'so if you open two doors at
\s,) 34 once and they approach each other, they create short 15 circuits in the plant because they don't look at two-door 16 performance, they look at single-door performance.
17 MR. SIESS: But John, my point is, I look at those 18 things and what seems to be common is to anchor them down l
39 at the bottom to the floor.
20 MR. O'BRIEN: Yes.
i 21 MR. SIESS: Which, from a structural engineering i
22 point of view is sort of the hard way. You are taking the 23 cantilever moment'and not only that, you are getting a 24 fairly low period, fairly low frequency because it's a i 25 cantilever, as opposed to bracing it at the top which takes b
e 178 l
) I a heck of a lot less force to resist the overturning and N/
2 will change the stiffness. It must change the frequency 3 tremendously.
4 Now, is it impractical at these plants to brace 5 these things at the top? It wasn't at Diablo, that's what 6 they did to a lot of them.
7 MR. O'BRIEN: Yes. You can't hang it on to anchor 8 because usually there is nothing up there. I mean, it's 9 easier to just put them on the floor.
10 MR. SIESS: John, some of these rooms you've got 11 enough of these things you could run braces across a whole 12 bunch of them over to the wall.
_, 13 MR. O'BRIEN: Yes, it you have them close to a (N, 14 wall, it will be easier to brace them on the top.
15 MR. BENDER: There are a couple of other things 16 about these mountings that we might want to consider. One<
17 is, I found that a lot o'f people, when they put in the 18 panels, find out that the bolts they intended to set these
.19 things on don't line up with the' holes. So what_they do 20 is, they say, we'll cut the bolts off and plug-weld them.
21 MR, O'BRIEN: You are talking about the bolts on 22 the shelves, right?
23 MR. BENDER: Yes.
24 MR. O'BRIEN: Not on the base. Yes.
25 MR. BENDER: On the shelves. You might want to
./
R s j
l E_________- - - - . .-
e 179 i consider whether you want to look at the plug-weld question 2 and see if you can learn something about whether that is 3 that an alternative. It shows up they are not very good as 4 a --
5 MR. O'BRIEN: It's a little late in the game.
6 MR. BENDER: What?
7 MR. O'BRIEN: It's a little late in the game 8 because the testing begins tomorrow. .
9 MR. BENDER: Well, you know, after you get all jo this testing done, cut a couple of bolts off and pl ug-wel d .--
ii - and see what happens. That's what I had in mind just to 12 close the loop.
_ 13 The other was that at one time I heard.a lot of
(,,) ja people saying, well, look, with all that conduit and stuff 15 that comes down through the top of the panels, that that g represents a substantial amount of stiffening and dampening 17 that isn't given much credit.
18 Are you dealing with that?
19 MR. O'BRIENs No, the only --
20 MR. BENDER: I think that is an argument that says 21 that that particular constraint may-be even more important 22 than the bolts when you get to it. That's what I have 23 heard some people say, including Mr. Stello once. I don't 24 know how authoritative it is.
25 You might want to think about that.
- i 180 I MR. O'BRIEN. Anyway, we will sturly the effect of
( )
2 anchorage and stiffness of the motor control center as it 3 pertains to component fragilities. We emphasize that this 4 is not'a test of the motor control center but a test of 5 electrical components mounted in a motor control center.
6 Real quick, we've got a variety of relays and what 7 we are going to observe during the tests are current 8 fluctuations, chatter of contacts, operation of off-table 9 relays and physical damage to components.
10 There is no standard fragility test, you know. We r 11 are doing one, we are trying to do one.
12 MR. SHEWMON: Tell me what relay chatter is, or 13 what you have as a criteria for_ unacceptable chatter. If .
) 14 it opens once, is that a failure?
15 MR. O'BRIEN: It certainly.is chatter. If it 16 opens once, it's a malfunction.
17 MR. SHEWMON: Yes.
18 MR. O'BRIEN: Yes, I think we will characterize 19 the chatter. We will have a signal that we can monitor, we 20 can tell you how many times it opens.
21 MR. SIESS: You won't know the significance of it 22 until you loo.k at the system.
23 MR. O'BRIEN: Yes.
24 MR. SIESS: Because right now we know that one of 25 the criteria has been, I believe, that it's operable after sq m.)
e 181
() i the shaking, no matter how much it chattered during it.
But the real concern is that the intermittent operation 2
3 will lock in some other function somewhere.
4 M R .- SHEWMON: The cabinet itself will have a 5 frequency of something like a cycle a second?
6 MR. O'BRIEN: We will vary the frequency by 7
varying the stiffness in the anchorage. We believe that we 8
are going to actually have to loosen the bolts to make it 9 less stiff. That's one of the variables.
jo See, we've got braces on the top and we have ij numbers of bolts that are variable in the tests.
12 MR. SHEWMON: But this is in a building.
MR. O'BRIEN: It's on a table, 13 b
T,,/ 34 MR. SHEWMON: Yes, but the one you should be 15 worrying about is in a building.
16 MR. EBERSOLE: May I ask you --
37 MR. SHEWMON: No, let me finish.
18 MR. EBERSOLE: Yes, sure. Go ahead, carry on.
39 MR. SHEWMON: I too, like all the rest of the
~
20 members, pause between sentences.
21 MR. EBERSOLE: Yes.
22 (Laughter) 23 MR. SHEWMON: What I'm trying to get at is, you 24 know, when your teeth chatter, that's a high frequency.
25 When I hear-the word " chatter," I think'of a high frequency n
( i s_J .
l i
e 182 n
I 1 and I wonder if indeed this can be driven at the frequency N.)
2 where " chatter" is a relevant word.
3 And the next thing then is, indeed, if one of 4 these things -- if you ask what the natural frequency of 5 the relay arm itself is, I suspect this is a lot higher 6 than that of the building.
7 So have you gone back to see if indeed that is a 8 credible -- if there is a complete mismatch between what 9 you are going to drive it at and what the arm will chatter 10 at?
11 MR. O'BRIEN: Oh, we are going to drive this at a 12 standard earthquake spectrum, we have agreed on that. That 13 represents a good earthquake.
( \
(,/ 14 MR. SHEWMON: And this is a good earthquake at the 15 third floor-in a plant, the c'ntrol room of a plant?
16 MR. SIESS: You drive it at floor spectra.
17 MR. O'BRIEN: It's a fragility test, so you keep 18 on going up, and up, and up until you don't have any more t
l 19 capacity. It's not a proof test. It goes until you can't 20 cause chatter. And then, to describe chatter, incipient 21 chatter, gross chatter, and then in a gray area where it's 22 just a one-time --
23 MR. SIESS: But chatter has actually been observed l
l 24 in qualification tests, am I correct? It's not a 25 hypothetical thing.
R l t LJ
a 183 MR. RICHARDSON:
' (m)
-s' i A couple points. Number one, 2 this test is not going to be a conclusive test that will 3 lay to rest the fragility of relays.
4 The purpose of this test was to demonstrate that 5 we can do fragility tests and how a fragility test ought to 6 be conducted. And I'm almost positive that we will 7 conclude we need much more testing on relays to learn more 8 about the characteristics of chatter and what drives 9 chatter, and when it does and does not chatter.
10 The other point I wanted to make was that it's my ti understanding that the PRAs, the seismic PRAn, performed to 12 date have shown that the population of relays that are 13 critical to relay chatter are really very few. There is
(-~)
s
, y only about a dozen or so relays in a plant whose chatter 15 really is a deleterious phenomenon that is going to bring g heartache.
37 There are hundreds and perhaps thousands of relays 18 used but only a handful of those relays would cause harm if i9 they were to chatter during the earthquake.
.20 MR. BENDER: Well, these relays that you are-21 testing up here -- you have listed some -- why were they 22 selected?
23 MR. O'BRIEN: Well, first of all, the numbers in 24 parentheses are the numbers because they represent what we 25 find in nuclear power plants.
O
e 184 pTm s MR. BENDER: But I'm just going to Jim's point.
[
) 1 2 Are they selected because they represent relays that are 3 involved in critical functions?
4 MR. O'BRIEN: No, the various designs. Some are 5 solid state, some are electrical and mechanical. They are 6 different designs.
7 MR. BENDER: So this is just to test the hardware 8 without any application in mind.
9 MR. RICHARDSON: Well, I asked that very question 10 and the answer I got was that those handful of relays over 11 the whole population of plants would be quite a variety of 12 relays.
13 So, these were selected to try to represent a wide s, 14 spectrum of different types of relays to sort of scope the 15 Problem.
16 MR. SIESS: I'm trying to understand why there are 17 so few. Any system that involves interlocks, a relay 18 chatter could lock something in; could it not?
19 MR. RICHARDSON: Yes.
20 MR. SIESS: But then your scope is limited in 21 systems, to those two sets of systems.
22 MR. RICHARDSON: Yes.
23 MR. SIESS: So that combination gets you down to r 24 that handful.
l 25 MR. RICHARDSON: That's my understanding. Now, b
I e-I 185 O
f, Oi j I'm admittedly skating on pretty thin ice here and if you 2 test me much more, I'm going to -- l 3 MR. SIESS: But it's obvious it's not every relay 4 because many relays could even change state and not affect 5 what you are talking about.
6 MR. RICHARDSON: Yes.
7 MR. EBERSOLE: Chet?
8 MR. SIESS: Yes, Jesse?
9 MR. EBERSOLE: Will these relays be connected to to their standard loads?
11 MR. O'BRIEN: Yes.
12 MR. EBERSO,LE: In particular, if they are AC
- 13 relays,.will they --
(_ j, MR. O'BRIEN: They will be energi zed.-
15 MR. EBERSOLE: -- will they be connected to moving g armature receivers such as you would have in a big 17 contactor?
18 MR. O'BRIEN: There will be circuit breakers and 39 other electrical equipment in the MCC.
20 MR. EBERSOLE: There is a little bit of a 21 problem. You know, if you open a relay through chatter 1
j 22 before the armature is moved, you are interrupting an l
23 approximately 10 to 15 times over-current with a relay that 24 is not designed for it, and it may-stick.
25 MR. O'BRIEN: I would call this not a separate l
O l
l
186 l I
x
'n i effects test but an integral test, and that we will have
( /
2 three cabinets loaded with lots of electrical things
+
3 including cables to ground things.
4 MR. EBERSOLE: All right.
5 MR. O'BRIEN: We'll have something off the table 6 that we will try to operate during the test and after the 7 test.
8 MR. EBERSOLE: So that the chatter may open the 9 relay before the main cockpit contactor is' closed out.
10 MR. O'BRIEN: And we will have interaction effects II 12 MR. EBERSOLE: Good.
_ _ _ 13 MR. O'BRIEN: It's $150,000 with a subcontract to
~~
14 Livermore.
15 MR. SIESS: $150,000 for the whole thing?
16 MR. O'BRIEN: Two months of testing.
17 MR. SIESS: That's not bad.
18 MR. O'BRIEN: No.
i 19 MR. BENDER: How much is the vendor getting? Do 20 you have to buy these pieces of equipment?
21 John, do you have to' buy this equipment?
22- MR. O'BRIEN: Yes, we.had to buy it. We had a 23 review, a free review, by SOUG and EPRI to certify that 24 this was an acceptable test, something that would help us 25 resolve 846.
l'~1 O
l
e 187
() MR. SIESS:
MR. O'BRIEN:
You can always re-sell it.
And then, my final slide -- that was 2
3 my brief progress report on Livermore activities.
4 (Laughter) 5 (Slide) .
6 MR. O'BRIEN. What we get from this program are 7
recommendations for component fragilities and modes of 8
failure. There aren't going to be any regulatory guides or 9 standard review plans impacted.
jo We may star.t a component fragility data bank and, 11 as I said earlier, the only use is to respond to questions 12 of East Coast seismicity, keep reactors on the East Coast 13 operating if the seismic threat is perceived to increase as ja a result of the Charleston, New Brunswick earthquake 15 issues.
g MR. RICHARDSON: But I think there is another 17 implied regulatory benefit here, and that is that if a 18 component shows substantial ruggedness, the interest in its i9 qualification level, I think, becomes less.
20 And it may help' resolve doubts in people's minds, 21 or questions, on the qualification of a particular piece of 22 equipment if it can be shown that its ruggedness is 23 substantially above the design level.
24 MR. SIESS: It certainly would be nice to know it 25 instead of. sitting around guessing, wouldn't it?
I
188 M
\
d[ 1 2
MR. RICHARDSON:
MR. SIESS:
Yes.
John, Dr. Shewmon suggested that you 3 ought to be able to sell these pieces at a pretty good 4 price since they are now qualified.
5 (Laughter) 6 MR. RICHARDSON: Maybe we can sell it to the Air 7 Force to install in C-5As and make a profit.
8 MR. SIESS: Any other questions for Jim?
1 9 (No response)
I 10 MR. SIESS: Thank you.
11 MR. O'BRIEN: Thank you.
12 MR. SIESS: Who is doing this test for you?
13 MR. O'BRIEN: Wyle Laboratories --
..y g) 14 MR. SIESS: Wyle.
15 MR. O'BRIEN: -- under subcontract to Livermore.
16 MR. SIESS: Out in California.
17 MR. O'BRIEN: Norco, California, which is near San 18 Bernadino.
19 MR. SIESS: All right, CAT-1 structures. Roger?
20 MR. KENNEALLY: I'm Roger Kenneally, the project 21 manager of the Seismic Category-1 Structures Program.
I 22 The viewgraphs in your package follow those of 23 John. The title of the program is "The Seismic Category-1 24 Structures Program," and also identified the contractor as 25 the Los Alamos National Laboratory.
i t
189 Im \ 'j (Slide presentation)
\usi 2 MR. KENNEALLY: The main issue that this 3 particular program is looking at is the ability of the 4 current plants to accomodate earthquakes larger th'an what 5 was used in their initial design.
6 A broad statement of the objective would be to 7
reduce the uncertainties in how these larger earthquakes 8
transmit their loads through the buildings to the systems 9 and components that are needed to operate and shut these 10 facilities down.
ij MR. SIESS: I would rather talk about earthquake 12 motions, rather than earthquake loads.
13 MR. KENNEALLY: Okay.
/"N MR. SIESS: We use generalized loads.
( ,/ 34 15 MR. KENNEALLY: This is an analytical experimental 16 program, and we will be obtaining data on the failure of 17 these different model configurations. But more important 18 is the interaction between the systems and the building 19 themselves.
20 Some more specific objective is to- obtain the 21 experimental data on the sensitivity of the structural 22 behavior through variations in configurations and 23 earthquake motions.
24 By behavior, I'm referring to accelerations, 25 displacements, frequencies,_and the like.
^$
G l
190 n) 4 i We are also trying to obtain changes in floor 2 response spectra that have resulted from these increased 3 earthquake motions.
4 The data from this program can be used to validate 5 computer programs that are used to predict the structural 6 behavior in the elastic and inelastic ranges.
7 MR. SIESS: Roger, on the changes in floor 8 response spectra, if I'm not mistaken, now there is some 9 interest in changes that might result not from increased to earthquake loadings but the reduced stiffnesses they are 11 getting in the shear walls.
12 MR. KENNEALLY: That would be an issue of, is 13 indeed the algorithm we are using currently in our design
[ h correct. That is one issue, yes.
t_/
s i4 15 Also, as the higher-earthquake motions are --
16 MR.-SIESS: Well, that's something else, yes.
17 MR. KENNEALLY: That's right.
18 MR. SIESS: But originally, I don't think that 19 change in stiffness was expected.
20 MR. KENNEALLY: That's correct.
21 MR. SIESS: Just for anybody else's information, 22 they are finding under relatively low loadings, they are 23 getting stiffnesses due to cracking of the concrete that 24 are about one-fourth those that you would have if it didn't i 25 crack, and a lot of designs have been made assuming that it l
l /)
(_
1 i
4 w - -- .
e i 191
, i didn't crack.
2 MR. KENNEALLY: That's correct.
3 MR. SIESS: And this can cause some interesting 4 shifts in floor response spectra which at a particular 5 frequency might affect something quite a bit..
6 MR. KENNEALLY: That's true because of this 7
stiffness difference we'are actually saying that the 8
fundamental frequencies of the structures that we are i
9 calculating may not be correct.
10 MR. SIESS: I still find it incredible that. people ij design those things assuming they won't crack.
12 MR. KENNEALLY: The only building, to my 13 knowledge, that is subjected to a crack analysis is the i
j4 containment structure-itself --
15 MR. SIESS: Yes.
16 MR. KENNEALLY: -- and that is because of the over-17 pressurization.
ig MR. SIESS: Yes. But anybody that assumes 39 concrete doesn ' t crack --- I mean, if there is one thing 20 I'm sure of in this life is that concrete will crack.
i
! 21 .MR. KENNEALLY: As we have. mentioned earlier, the l
22 seismic program is highly integrated. This particular 23 aspect of the program will be providing structural behavior 24 information to the validation project that you heard 25 earlier and also the seismic design margins ~ project.
O LJ
e 192
{\.I Q.
1 The interest on floor response spectra will be 2 feeding information to piping reliability and component 3 fragility projects.
4 MR. SIESS: ~You know, if we find out that for a 5 change we have.made an unconservative assumption, we'll 1 6 also need that fragility to justify some of the existing 7 designs.
8 MR .' KENNEALLY: Yes.
9 MR. SIESS: It will be nice if they offset each 10 other.
11 MR. KENNEALLY: To give a little bit of history 12 about where we have been on this program and where we hope j 13 t o g o --
, 14 MR. SIESS: Excuse me again I don't think you 15 have made it quite clear that this is not containments, but 16 this is the things like aux buildings and diesel generator 17 buildings, shear-wall buildings.
18 MR. KENNEALLY: .That is correct, looking at the 19 reinforced Category 1 structures exclusive of the 20 containment building.
21 MR. SIESS: Yes. And that these are structures 22 that fall outside the range of the kinds of buildings that 23 fell down in Mexico, and so forth, where we have a lot of 24 other experience there.
25 MR. KENNEALLY: That is correct.
j' 8 '
l (_,)
i l
1 i .
_. . _- - l
l 193 i
} j MR. COSTELLO: Professor Siess -- Jim Costello --
2 I would like to add on a little bit on the question of the 3 implication of the frequency shift.
4 Clearly, it's possible that one might then infer 5 that if the response spectra had not been widened enough to 6 accomodate expectable frequency shifts, that things might 7 tend towards the unconservative.
8 However, results of the question of the additional 9 damping that will be associated with that kind of frequency in which might tend in itself to compensate for it.
it MR. SIESS: All we can be sure of is, it will 12 change something. Whether it will go up -- it might go 13 down as much as up, depending on where your peak was, I
\m ,/ u tool. Yes, the damping is important. Thank you.
15 MR. KENNEALLY: Prior to this fiscal year, we have 16 done static or seismic' testing on 23 different models of 37 two types of reinforced concrete structures.
18 These were prototypical models of the diesel 19 generator building and an auxiliary building.
20 Two different scales of these models were used.
21 They were constructed with either a one-inch wall'or a 22 three-inch wall.
23 Specifically in the area of the diesel generator 24 building, we had models of one or two floor, and'in the 25 auxiliary building, it was a three-floor model. The models T
- (~'/
s.
l l
l
194 m
%d
) I were so designed that the model of a one-inch wall was in 2 exact scale of a 30-inch wall structure, and a three-inch 3 wall was a one-tenth model of that same structure.
4 And the scaling laws were applied so that the one-5 inch thick wall was a one-third model of the three-inch 6 thick wall. So we were able to extrapolate from our small 7 models prototypical behavior and also used a one-inch model 8 as a scale model of a three-inch structure.
9 Based on that, I will jump down to the last to bullet. We are very confident that we have been able.to 11 scale these results from our testing of the one-inch and 12 three-inch models. However, all of our structures to date
. 13 have been constructed with micro-concrete and there is some s- 14 concern about the behavior of micro-concrete versus other 15 concrete.
16 MR. SIESS: I can see a question as to when it 17 might crack as a difference between micro-concretc or other 18 concrete, or even a one-inch wall, or three-inch wall, or 19 30-inch wall, whether they are made out of the same 20 materials.
21 Are there any other reasons to wonder about the 22 scaleability with the micro-concrete?
23 MR. KENNEALLY: Just the same size crack would be 24 harder and have a different impact on the two structures.
25 MR. SIESS: Okay. And the bond characteristics, I 1
0
-O o
195 l
( j mean, that's another scaling, it isn't the concrete.
%d 2 MR. KENNEALLY: That's right.
3 MR. MARK: In this scaling, in a CO-inch wall 4 there is a term -- large term, possibly -- having to do 5
with just the weight of the structure. In a one-inch wall, 6
that's probably totally --
7 MR. KENNEALLY: Some of the gravitational effects 8
cannot be properly scaled, but a lot of the others are.
9 MR. SIESS: And for some of the dynamic tosts, 10 they put added mass on and it's calculated to give the it response.
12 MR. MARK: Anyway, that's done as well as you can y3 here.
s_f. ja MR. KENNEALLY: Yes.
15 MR. MARK: And these things were shaken and not 16 just pressed.
37 MR. SIESS: Yes. But they have an added mass to is handle the proper frequency response characteristics.
39 MR. KENNEALLY: And in jumping up to the'next to i
- 20 last bullet on that, as has been pointed out earlier by 21 Professor Siess, that we have noticed that stiffnesses of 22 ur model structures have been lower than the computed 23 uncracked crack sectional values by factors of as large as 24 f "F-(
l 25 MR. SIESS: I don't know how the computer got'in O
196 V
(v ') I there. I can calculate the stiffness without a computer.
2 (Laughter) 3 MR. MARK: That was Bob Budnit='s computer.
4 (Laughter) 5 MR. KENNEALLY: During this current fiscal y,e a r ,
6 we have been concentrating on trying to address this 7 stiffness difference. With the aid of the technical review 8 group, a special structure has been designed.
9 This particular structure was designed to have a 10 fixed frequency of approximately 30 hertz. We were going 11 to subject this structure to both seismic and static 12 loadings.
___ 13 Before we went ahead.and built a fairly large
[
(_j 14 structure, we decided it would be best to build smaller 15 .models of this so that we can make sure that our predicted 16 responses were indeed close to what we thought we would be 17 getting. It would give us an opportunity to verify the 3
18 instrumentation locations and make sure in general that we 39 were not going to have any surprises.
20 So we have gone ahead and we have fabricated two 21 one quarter scale models of this structure that we have 22 deemed the TRG structure -- which stands for technical 23 review group. It dces not indicate that the technical 24 review group has indeed designed the structure.
25 (Laughter) 1
(~11 t
x/
I i
i
197
/
i y MR. KENNEALLY: They have suggested it to try and L)m 2 isolate it from other configurations, and we have used that 3 acronym.
4 MR. SIESS: I found that rather strange. It 5 sounds like you are blaming them for the structure, that 6 you really didn't like it, but they recommended it.
7 MR. KENNEALLY: No, we did go around with both 8
ourselves, the contractor, and user offices to make sure 9 that we came up with something that would be acceptable, jo MR. SIESS: It's one of these things, you start in referring to it that way in your conversation, and then it i
12 ends up as the formal designation.
13 MR. KENNEALLY: That's right. We will live with
(-)/
(, ja that until~the tests are completed.
15 Also in this current fiscal year, we are going to g start the fabrication of this_ larger structure. The model 37 is going to be fabricated with prototypical concrete and ig reinforcement rods. Typically, the concrete will be 4,000 i9 psi, 60 ksi reinforcement.
20 MR. SIESS: Now, that's about the size of this 21 half of the table, isn't it?
22 MR. KENNEALLY: Yes. The structure itself -- in 23 fact, let me show you a picture of it.
24 This is basically a rectangular tube with a wall 25 that is placed i n the center of it. The wall that we will.
O i
V i
, v-- __m- - - , - - , . _ . - , -
e 198 p
.I 1 be exciting is 106 inches in height, 90 inches in depth.
V 2 So, it is a very large structure.
3 MR. SIESS: In what direction is the excitation?
4 MR. KENNEALLY: In direction of the wall.
5 MR. SIESS: This way?
6 MR. KENNEALLY: Yes.
7 MR. SIESS: Okay. No end walls, but you have a 8 cross-wall.
j That's right. That's what we are 9 MR. KENNEALLY:
10 studying, is the cross-wall itself.
11 MR. SIESS: Is the cross-wall in the direction of 12 excitation?
- 13 MR. KENNEALLY: Yes.
(,,) 14 MR. SIESS: Where are you going to test this?
15 MR. KENNEALLY: At -- I'll get to that a little 16 bit later. That testing will be coming up in November and 17 we will be keeping the ACRS informed of that definitely.
la As I just got through mentioning, in FY 1986 we i 19 will be testing two of these substantial structures. The 20 first one will be in November and the second one will be in 21 March.
22 The first one will be subjected to a considerable 23 amount of low-level excitation, static and quasi-static --
24 sorry, static and seismic. The second one will be 25 subjected to a quasi-static.
gm d
-199
() i We do believe that these structures can be failed.
MR. MARK: Can you simulate an earthquake signal 2 l 3
with the testing machine?
4 MR. KENNEALLY: Yes, it is. We are currently 5
using a properly time-scaled El Centro.
i j 6 We have also developed some contingency plans 7
based on the outcome of this particular testing.
These 8
plans appear in Appendix C to NUREG-1147. We think we have 9 handled any of the outcomes of this particular testing that ja we might be seeing.
1' 11 If things are going as we are hoping in the
, 12 resolution of the stiffness and solubility issues, we want 13 to initiate a limited number of experiments which will be j
ja to meet the. program objectives and benchmarking analytical 15 m del development.
16 This side-step-that we are currently going through 37 right now on the stiffness difference in scaleability is L
'is not the overall program objective. We are not putting this j
39 money in here to say that we know how to test reinforced 20 concrete structures.
l 21 Also during fiscal-year 1986, Los Alamos will be 22 adding a statistician to the program. The office has come l 23
~
through with an understanding that they would like to have 24 m re statisticians involved in our program planning to make 5
25 sure that we do not get through a series of tests and then
!O l
i i
r
- . . - . -. - - - . - . - - -. . - - ~ - - - -
i' s. *
, l 200 I
i i say, " Gee, if you had only done this, or if we had only a .
! 2 done that, we might have had some better results." !
3 this particular program is one of a pilot grouping
, 4 for meeting this objective.
i 5 MR. SIESS: Lots of luck. .You are likely to end 6 up testing a lot more specimens than you want.
Y 7 MR. KENNEALLY: Well, that is my job, to keep i
8 within the budget constraints. I understand what you are i 9 saying, yes.
l i .
10 MR. SIESS: We tested some bridges in connection i
11 with the ASHO road test many years ago. They had a i
i 12 stati stician designing the road ' test which they could do 1
33 bectiuse they had four one-mile loops, and so forth~and so
- 14 forth.
15 . We had 16 bridges, and that statistician nearly ,
16 went crazy trying to figure-the statistical design of our 17 bridges which were not statistically designed. . We had a 18 four-hour session and it got nowhere until somebody said, 3 19 "You know, what we are really doing is case histories," and
- 20 that solved it. He left the room. He said, "I don't know i
} 21 anything about case histories."
i i
22 It was an extremely successful. test. We got at 1^
-23 least -- for dollars spent, we got a thousand times as much.
24 as they got out of the pavement. You know,fwe can analyze 25 these. things. We are not-just looking for statistical l
l
_ _._ _ _ _ . _ _ - . _ _, _ _ - ~ . . . . . . . .. _ ,- __.___;____,,..u,._.. _. _ . . .. - . - . _ , . _ _ _ . _
201 results, .4e are looking for comparisons with analyses, and
( 3 2 that's a case history.
3 It's worth a try. .They've got to keep the 4 statisticians busy down there.
5 MR. KENNEALLY: Like I said, we have been di,rected 6
to pursue this, and there is some merit to it. But it does 7
have to run a close check, there is no doubt about it. I 8
don't want to spe'nd a considerable amount of money to just 9 indicate that.if I had tried to build floor-concrete models 10 and when I subject them to the same low-level test, I'm i
, n going to get four different values of frequency.
12 MR. SIESS: That 's probably what you 'll get.
13 MR. KENNEALLY: We have already, I understand h) q_ ;4 .that. I don't need a statistician to point that out to me.
15 During fiscal year 1987, we are going to be doing g a lot more work on analytical methods. We do not want to j7 do a lot of testing if we can do a good analytical 18 algorithm and the like. So basically, after the stiffness 39 difference issue has been resolved, we will test only what 20 we feel is necessary to ta): the parameters and make sure 21 that we have confidence in the analysis.
22 Currently, we are trying to do these testings on
~
23 one-inch thick walls. If our results from the testing will 24 allow us to, they are very economical to build and they can 25 be done in the controlled environment of Los Alamos.
i O
(
s
i 202 l
1 Configurations we will be looking at is perhaps 2 additional walls internal to the structure. What we have 3 done so far has been in a very simple box. We will be 4 adding additional walls in there, removing some walls to 5 make them more representative of what might be out in 6 actual plants.
7 There is no way that we will be able to build 8 configurations as they exist. We will try and complicate i 9 them a little bit so that we can get insights on what might to happen and what is critical.
11 MR. SIESS: Well, DG buildings are fairly simple.
12 MR. KENNEALLY: That is'why-we have selected it.
_ 13 MR. SIESS: But again, that program is much in the
,, 14 same category as the thing you mentioned just before, 15 taxing the analysis.
16 MR. KENNEALLY: That's correct.
17 We will also be doing some testing on some two-18 floor models. We would like to have different 19 configurations on the first and second floors. Some of 20 these would probably be done with some thicker walls also 21 to make sure of our scaleability of results and the like.
i 22 MR. SIESS: Will you get some in there that are 23 asymmetric to get some torsion?
24 MR. KENNEALLY: Yes.
25 I'm reluctant to do a lot of testing on the r~
l (m
N.)
\
~ __.
203
(-)
,. ) thicker walls. First of all, the models are very expensive 2 and we had an experience last January where we set the 3 model on the CERL table. We had some power supply 4 problems. We threw 60 hertz worth of noise through.the 5
model, did some damage to it. They felt terrible. They 6
said if it only happened the week before when they had 7
steel cabinets on the table it wouldn't have bothered it.
8 (Laughter) 9 MR. KENNEALLY: That was very close to our 10 resonant range.. We didn't obtain as much out of the 11 experiment as we could have. We did still get some 12 meaningful data. The larger the model, the more expensive 13 to build, transport, and to understand.
s_j i4 MR. SIESS: That goes against a basic rule, but I 15 know why. Usually, larger models are not more expensive to
~
g build. The smaller models got to be watch-making jobs.
j7 But your small models are tricky. You are using hardware ig cloth instead of reinforcing bars.
19 MR. KENNEALLY: We are using welded wire hardware 20 mesh, that's correct, 21 MR. SIESS: That's what makes the difference. If 22 you try to fabricate those out of simulated rebar --
23 MR. KENNEALLY: We are into the watch-makers, 24 that's correct.
25 MR. SIESS: Yes.
m v
e 204 l
)
T(s
) MR. KENNEALLY: The regulatory applications for 2 this particular program will result in changes to 3 regulatory guides, for example, the regulatory guide 4 involving damping and' structures, 1.61; also, regulatory 5 guide on floor response spectra. Is the plus or minus 15
~
6 percent adequate or should other considerations be 7 incorporated into.the regulatory guides?
8 This research will-also be reflected in changes to 9 the Standard Review Plan Sections 3.7 through 3.10.
10 It may be necessary to develop new regulatory.
11 criteria considering non-linear response effects. The 12 basis of some of that will come out of the work from this 13 pr gram.
14 We will also have experimental data and analytical 15 methods that will determine the behavior of the concrete 16 buildings to earthquakes larger than the design.
17 Getting back to the subject that you mentioned 18 earlier on structural fragility, this program will start to 19 contribute to some of that by being able to identify for 20 the structures that we are testing where cracking is going 21 to start in some of these structures.
22 The problem with structural fragility, unlike so 23 much of the other fragility, is to relate the cracking in 24 the building to the physical attachment of supports of 25 piping and the like.- If there happens to be support
/ 1
\ )
i x_-
i
205 ;
attachment in areas where the building is cracking, then it
( i 2
can get to be a problem. If not, the structure has not 3
seen any loss of function, nor has the equipment.
4 MR. SIESS: A couple of questions. This set of 5
experiments where you are going to use the statistician, 6 these are still back in the stiffness issue?
7 MR. KENNEALLY: No.
8 MR. SIESS: No?
9 MR. KENNEALLY: These are more the program jo objective issue when we start changing configurations and it the durations of the input motion, and maybe different 12 types of input motions.
. 13 MR. SIESS: Okay, I see.
,,,) i4 In these studies on the stiffness issue -- and I 15 was trying to remember from the report I read -- are you 16 doing splits on the test to get tensile strength of j7 concrete?
18 MR. KENNEALLY: Yes, we are.
39 MR. SIESS: Okay. I had to get about eight pages 20 beyond where I should have in that report to find out that 21 they were measuring the elasticity and not calculating it.
22 MR. KENNEALLY: Yes.
23 MR. SIESS: Which seemed a little strange at the 24 time, but it was there.
25 MR. MARK: I suppose this is understood well O
V
l l
206 1 .enough. The scaling features of rebar concrete in my
}
2 simple mind need not proceed in parallel . Is there a 4
.3 problem there?
4 MR. KENNEALLY: The problem is, as the concrete i
5 gets smaller in size, aggregate is very small. And t,he 6 rebar itself is very small, to have enough defamation in 7 the rebar to get the simulated bond.
8 MR. MARK: -- concrete cracking, which doesn't 3 9 necessarily go in the same way as bending a rebar, 10 MR. KENNEALLY: No.
11 MR. SIESS: The thing is, they can measure the 12 cracking tendency, the tensile strength. I had raised an 13 issue when we were down there two or three years ago as to
( ,/ 14 whether they were getting shrinkage stresses due to think 15 sections. ,
16 I think they did some tests with moist specimens 17 and found no -- were still getting the effect. .
! 18 MR. KENNEALLY: That's correct. The specimen came 19 out of'the curing chamber. It was tested, and the results 20 were just about the same.
21 MR. SIESS: Now, it's still not quite clear where l
l 22 the cracks are coming from ativery, very low stresses. But j 23 we don't have to understand that to know th'at they are 24 there.
t 25 MR. KENNEALLY: That's correct.
m
I 1
207
( i MR. SIESS: There is not any really serious 2 pr blem about scaling if you can measure the properties. I 3 think the bond question, as to how large the cracks are,-
4 what spacing they will come out with, we have.done a lot of 5
relatively small-scale models and it looks very, very 6
similar to the la_rger scale or full size.
7 But you need to try a couple of scales.
8 MR. MARK: It' Just didn't seem intuitively obvious 9 that this would be the case. But you have looked at it
. 10 enough to know that it does follow pretty well.
11 MR. KENNEALLY: Yes.
12 MR. EBERSOLE: Chet, does this use the same sort 33 of --
( y MR. SIESS: The damping might end up being 15 different because of the cracks in the aggregate ~ interlock,
! g which won't scale. I'm sure you get some aggregate
- 37 interlock across-cr'acks. Cracks won't be as wide with the 4
ig small ones,~but then the aggregate is not as big. But l j9 that's almost going to be in.the noise.
4 20 MR. KENNEALLY: The issue of damping is not going 21 to be an easy one to try and address anyway.
22 MR. SIESS: No. Except it won't come out what we 23 normally assume, probably.
24 MR. KENNEALLY: That's true.
25 MR. SIESS: Any other questions?-
O c.
e 208 TN i%/ ) 1 (No response) 2 MR. KENNEALLY: Thank you.
3 MR. SIESS: I'm looking at the schedule, and we 4 are about ten minutes behind schedule which, for the ACRS, 5 is right on or better. But there was some thought that we 6 might be ahead of schedule because the EPRI people were not 7 here.
8 It doesn't look like we are going to get that far 9 ahead of schedul e , and we will probably stick to the 4
10 agenda, which means we go through'the piping research 11 program today and take containment integrity tomorrow. If 12 that helps anybody that wants to go home or something, I 13 think we can stick to it.
(,
k 14 MR. ARNDT: I will talk about structural load 15 combinations.
16 MR. SIESS: And the viewgraph I've got says 17 reliability analysis and non-linear behavior. Is that part 18 of it or do we skip that?
19 MR. ARNDT: Yes, that is part of it.
20 MR. SIESS: Okay.
l 21 MR. ARNDT: In the. interest of conserving the 22 schedule you are making and the short time that I do have 23 on your schedule, I will do as Roger did and simply 24 describe from the chair here without projecting the l 25 viewgraphs that you already have before you.
1
[3 Y_Y l
e sed."
- 209 (m)
%/
i We have had for the last few years --
2 MR. SHEWMON: Can you help us find the first 3 viewgraph of your presentation?
4 MR. ARNDT: Okay. The first viewgraph follows 5 immediately after what you have had, and it's title, 6
" Reliability Analysis of Non-Linear Behavior of Concrete 7
Structures."
8 MR. SHEWMON: Okay.
9 MR. ARNDT: For the last few years, we have had an 10 on-going research program at Brookhaven entitled, 11 "Probabilistic Load Combinations for Concrete Nuclear 12 Structures."
13 This is really an extension of it. What we are.
ja doing is, we have come to the decision that the research j$
phase over the last few years has come to a good conclusion 16 point with the end of this fiscal year, fiscal 1985, and 37 what we are planning to do in the next phase would perhaps 18 be best characterized by retitling the research and re-19 identifying it, and making a distinct _ break in'the progress 20 in the research that we have been doing.
21 Since this is.for the fiscal years ahead, the 1'
22 . slides that you have before you in this title represent the 23 fiscal '86 and '86 projected work. ,
24 MR. SIESS: Let me interrupt for a moment. The 25 previous work on load combinations, then, is completed?
!O l
l _-
210 1
MR. ARNDT: Almost. What is being done currently
( j 1 2 is a wrap-up summary report by Brookhaven --
3 MR. SIESS: Okay.
4 MR. ARNDT: -- to tie together all of the work 5 that has been done. What it will do, we will have 6 completed at the end of fiscal 1985 development of 7 probabilistic phased load combinations and load factors for 8 concrete containments and for concrete shear-wall 9 structures.
10 MR. SIESS: That will take care of ASME 32 and 11 359?
12 MR. .ARNDT: Yes.- The idea will be that these will
__ 13 be provided to these groups for their use as they see fit.
{(_)h 14 We have not carried it --
15 MR. SIESS: And those are not non-linear.
16 MR. ARNDT: These are linear.
17 MR. SIESS: Yes, okay. So that's finished, and in 18 this now you are going to go a step --
19 MR. ARNDT: The difference here and the shift in 20 phase and the reason for changing-the title are two-fold.
21 One, we are going from the elastic linear behavior 22 of the structures up to including design loads, beyond the 23 design loads into non-linear elastic behavior of the 24 structures.
25 We are also not going-to go as far as we have in m
\
l 211 i
() I the past years looking for the load factors. We are simply going to go as far in this on going research as the 2
, 3 reliability analysis for these structures.
4 The reason being that we are not at this point 5 staring at a whole lot of new designs coming in for ,
6 licensing. And the focus and the need of NRR for the 7 research is looking at structures that are already built 8
and evaluating their behavior under higher loads.
9 MR. SIESS: Now, .this is seismic loads.
10 MR. ARNDT: Principally seismic loads, but not ij limited to. The three loads that are being used are the 12 dead load, live load, and seismic load. But the 13 methodology should be applicable for more than that.
ja MR. SIESS: Now, this would not be containments 15 under severe accident loads.
16 MR. ARNDT: Yes, we will be going into 17 containments.
ig MR. SIESS: At severe accident loads, you look at 39 that?
k 20 MR. ARNDT: Yes.
21 MR. SIESS: Okay.
[ 22 MR. BENDER: Does this mean that the're will be l
l 23 changing loads as a function of time? I think with the j 24 seismic load it is obviously that way. But the other 25 loadings that go with it?
/3 t !
V l
e 212 M
MR. ARNDT: Dead load and live load are really (v) 1 2 statically applied loads.
3 MR. . BENDER: Well, I think a dead load obviously 4 is statically applied. The other load is pressure loading 5 or something like that -- I'm not really sure what you are 6 saying. Is it pressure loading that you are going to add 7 to the other two?
8 MR. ARNDT: I would like to present Nilesh 9 Chokshi. He is with NRR.
10 MR. CHOKSHI: Methodology -- and duration. It can 11 account for good.
12 MR. BENDER: Okay.
_, 13 MR. ARNDT: Now, the viewgraph you have before m, 14 you, I'd like to make one modification. It illustrates 15 scope with.four tasks following i t.
16 That's really a fiscal 1986 scope. The 1987 scope 17 is virtually the same thing, just completing Tasks 1, 2 and
{
18 3. And-Task 4 indicates that there be a report on 19 stochastic equivalent lineari ation technique. That would 20 be provided at the end of fiscal '86. But at the end of 21 fiscal '87, we would have a wrap-up report for this on 22 reliability analysis, not just for non-linear structural 23 behavior.
24 You will note that 1985, it shows zero dollars 25 funded. But that's simply to keep things separate. We
{
- ,r s
, \
i i
- - ,_-.w. - - , _ . - . _ _ - -- . . _ _ -
I 213 want to be able to exercise a better understanding and a
( ) ;
2 good break point between what has been done to date and i
3 what will be done here.
4 There is a little bit of fine-tuning still to be 5 done by us in our communication with the user and the
! 6 contractor in establishing the details and refinements of 7
the needs, and exactly what will be done by the contractor 8
and how it will-be done to satisfy those needs.
9 MR. MARK: Can you easily explain fgr my benefit 10 in more than four words what is meant by " stochastic n equivalent linearization technique?"
12 (Laughter) 4 13 MR. ARNDT: I don't blame you. I view myself as 34 someone who is not an expert in the field that I am looking 15 at, and I hope that that's a benefit to try to keep using 16 plain English.
17 MR. MARK: I got that far.
- 18 Stochastic is simply a varying random load.
39 Probabilistic is --
20 MR. SIESS: He's on the third line.
21 MR. ARNDT: Oh, Task 4.
- 22 MR. SIESS
- No, it's the third line in Task --
l 23 MR. MARK: I was really in Task 4.
1 24 MR. SIESS: Okay, it's the same thing as the third i
25 line in Task 1, yes.
U l
1 . . . , -
214 y
) 1 MR. ARNDT: Okay. What they want to do at
-x__j 2 Brookhaven is represent the non-linear behavior of-the 3 structures with some other approximate manageable 4 technique. This is what they propose would be an approach 5 that would approximate this inelastic behavior of a 6 structure which will crack and deform.
7 MR. SIESS: Well, this is what has been done in 8 other structures.
9 MR. ARNDT: Right.
10 MR. SIESS: NUMARC's approach, you can change the 11 damping and --
12 MR. ARNDT: It's an approximation to what actually f3 13 i I A/ 14 MR. SIESS: The trouble is, I don't think it's an 15 approximation you have never generalized.
16 MR. ARNDT: Pardon?
17 MR. SIESS: I don't think you can ever generalize is it. I think you can get a linear approximation for any 19 given case, but to get one that is perfectly general, 20 you've got to do the general case first.
21 MR. BENDER: What happens, you wind up with some 22 kind of pseudo-stress value that's a lot higher.
23 MR. SIESS: A pseudo-linear.
24 MR. ARNDT: Pseudo, yes. The relationship, I'm 25 not sure at this point.
!r ,\
v I
! l l
l
l l
4 215 l
(Q j j MR. CHOKSHI: Integral stochastic linearization, 2 for example, you are interested in a mean response of a 3 structure, and you match your parameters, stiffness and 4 damping, to obtain an equivalent linear system which -- the 5 mean response, whatever you are interested.
6 And depending on the quantity of interest, you 7
define your equal linear systems which will give -- square 8
values, or the probabilistic response you are interested 9 in.
10 MR. SIESS: You can get an equivalent linear ij system to match any desired outcome of a non-linear 12 system. But it won't be the same equivalent linear system, j3 depending on whether you are looking for velocity or
- s_ ) ja acceleration. It's been done.
j$ MR. CHOKSHI: In this case, it will be mostly 16 extreme statistics we are interested in.
J 37 MR. SIESS: Yes. And then they are going to do it ig probabilistically, which makes it real interesting. And 39 nobody will understand the results except the guys that are 20 doing it.
21 (Laughter) 22 MR. SIESS: There may be three or four other 23 people.
24 MR. RICHARDSON: I particularly, as the manager, 25 have to express some ignorance in this area. I am really N <
i
- ~
- n .., . - , , - , , - - . , - - , , - - , , - . , , , - , , ,, ,, ,, --
e 216 I soliciting your and the committee's reaction to the 2 importance and need of this research.
3 MR. BENDER: What are you going to do with it, do 4 you know?
5 MR. RICHARDSON: Nilesh can answer that bett,er 6 than I can.
7 MR. CHOKSHI: I think the uses we are looking 8 forward to is, for example, the PRA for the latest matters 9 in structural linear is very, very approximate, subject to 10 judgment. Or the use is the multiplication of factors of 11 safety, opened a lot of questions.
12 Here, what we are doing is taking the r
., 13 probabilistic aspects of loading into consideration and
, 14 then use a random vibration technique, and then calculate 15 .in probabilities, rather than making judgments as to the l
16 normal distribution estimating a million factor of safety 17 and assuming some uncertainties. That is the main benefit.
18 This method was used for GESSAR II containment 19 evaluation.
, 20 MR. BENDER: Well, is the answer more meaningful i 21 and if so, why?
22 MR. CHOKSHI: Well, to the more meaningful, EPRI 23 identified the structure as being a critical item, 24 particular structures or particular structural elements.
25 And then you want to examine whether that is based l
217
() i on a simplified assumption or it's really a critical item.
MR. SIESS: The use would be in FRAs, then, 2
3 primarily.
4 MR. CHOKSHI: Among other things. Other things 5' will be a margin. I think this will give you margins much 6
better distribution in a big area.
7 MR. SIESS: Is this for steel or concrete, or 8
both?
9 MR. CHOKSHI: Currently, it's for concrete. As 10 1 ng as you can define limit states, it's not really n limited by --
12 MR. SIESS: Well, you mentioned GESSAR and that's 13 a steel containment.
Og
(,,/ MR. CHOKSHI: Right. We had added -- for GESSAR i4 i$ containment in the program.
g MR. SIESS: It's interesting because steel has a 17 fairly clear nor-linearity in terms of stress, and concrete 18 starts being non-linear as soon as it cracks.
39 MR..ARNDT: Well, we have rated the priorities.
1 20 MR. SIESS: And cyclic loading and reverse loading 21 in the containment, a lot of work has been done on reverse
- 22 loading and so forth.
23 MR. SHEWMON: What does limit state mean?
24 MR. CHOKSHI: Limit state is the definition of a 25 functional failure or structural failure. It's t
(_-
l l
l
218 P ) matt.ematically described limits.
(
2 MR. SHEWMON: This is'in strain or cycles?
3 'MR. ARNDT: It's a probability of exceeding a 4 given strain or stress level, depending on what you decide 5 is critical. And they are using a 40 year design life of 6 the plant. It's a probabilistic value of exceeding that 7 level.
8 MR. SIESS: Well, that's one definition. There 9 must be --
10 MR. ARNDT: It's a probability-based definition.
11 MR. SIESS: Well, limit state is used in other 12 ways. Limit state is just that, it's any state you decide
., 13 as being the-state I want to work at. It could be two
( \
( ,) 14 inches deflection; it could be ending up in two pieces; it 15 could be excessive leakage. You have to define a limit 16 state and then they can work to that.
17 MR. SHEWMON: But usually -- well, invariably the 18 steel would work harden whereas shear-walls do not, as I 19 understand it.
20 MR. SIESS: Shear-wall would work harden if you go 21 far enough. Rebars work harden.
22 NR. SHEWMON: They certainly work sof ten as you 23 start opening up cracks.
24 MR. SIESS: Oh, yes. But reinforcement -- I mean, 25 the reinforcement work hardens, and that's what's carrying
l l
l 219 f) v i the load once you get the cracks, half of it, anyway.
2 MR. ARNDT: We prioritized the kinds of structures 3 to look at under this program that the most important and 4 the best pay-off is to look at the concrete structures.
5 The second thing would be to look.-- including 6
c ncrete containments.
7 The second priority would be steel containments, 8
and the third priority would be remaining steel structures.
9 Under the work that we are kind of wrapping up in jo fiscal '85, we deferred the steel containments and we have ij dropped the other steel structures. We feel that you 12 improve the design on steel frame structures on a 33 probabilistic basis and try to go back and design them more 34 efficiently on that approach.
15 There isn't that much in atauclear power plant 16 that is cost effective to retrain the ALES.
17 MR. SIESS: But now, Gunther, what you are looking is at is a distribution of whatever your parameter is you are j9 concerned with. If'it's a containment, it's not just a 20 method of calculating a limit state which might be a one-21 inch crack or leakage, line of failure or whatever. You 22 want a distribution. That is the object of this stochastic 23 thing, isn't it7 i
24 MR. CHOKSHI: Yes.
i 25 MR. SIESS: Because if you don't want a O
V l
l
e 1
220 7
7' i distribution, its deterministic calculation isn't all that
\_/
2 great.
3 MR. CHOKSHI: No, sir.
4 MR. SIESS: You can do it with a finite element 5 program. But if you want' dynamics and if you want load 6 reversals, are they going to consider load reversals in 7 this thing?
8 MR. ARNDT: I think so, but I'm not sure.
9 MR. SIESS: You know, if you are going to make to PRAs and containment failures important -- which I'm not 11 sure it is -- then this will give you a distribution to 12 stick into a PRA.
13 If you asked me off-hand, I wouldn't give it a yi
? 14 very high priority -- on two grounds. First, I'm not sure 15 there is a tremendous need for it to evaluate safety, 16 health of the public.
17 Second, I think the prospects of coming up with 18 something that's generally-useful and credible to enough 19 people are not all that great.
20 MR. ARNDT: We started out to level out the safety 21 margin.that's designed into the structures.
22 MR. SIESS: But now we are not designing, we are 23 evaluating.
24 MR. ARNDT: ,Right. Now we are looking at this in 25 the next couple of years in an evaluation phase and trying
~
i
(>
l l
!' - t- , _ _ ._ _ . _ _ . - _ _, _ _ _ _ _ . _ _ _ _ _ _ _ . _ . . _ . _ _.
~
i l
221 j l
) i to contribute to, again, the general pool of knowledge of 2 how good is a structure that's already built and is being 3 hit with a larger load, and what kind of response will it 4 give.
5 MR. SIESS: And keep in mind that some of those 6 structures are pretty good because when they were-designed, 7 they were designed very conservatively and elastically to 8
1 ads that are a lot less than we are now looking at.
9 (Laughter) 4 jo MR. ARNDT: Right, which balances out.
11 MR. SIESS: You know, we don't want too sharp a 12 pencil unless we know the loads with a lot more certainty j3 than we do.
(_)
s, ja MR. ARNDT: One of the things we are doing is, we 15 are working with the laboratory at this point to establish i
i 16 the relationship between the work proposed here and all the I
17 other work that they are doing both for Research and NRR. .
is MR. SIESS: Now, is this Morris Reich in WANG 7 19 MR. ARNDT: Yes, it is. We want to establish the 20 relationship there and you can see in Objective 3 on the
(
21 second slide there is another contract with NRR where we 22 want to do some tie-in between the results of this research 23 and the results of what they are doing there.
l 24 We had figured that it wuuld be very useful for us 25 to come to the committee at some point with this research l
e 222 y \
(
} I and have both Brookhaven and us sit down and discuss it 2 with you.
3 We had not reached a point in development pr'ior to 4 this that we felt merited doing so. But we would like very 5 much your input on this before we are too far down the road 6 with it.
7 MR. SIESS: It ain't going to be easy to get a
because the end use -- and I think the first question, do 9 we need to know probabilistically the ultimate pressure 10 capacity of a containment?
11 I thought they did some of that in the other work, 12 that was local loads. So that was essentially what they 13 did, and that was in elastic; wasn't it?
(
(M,,) 14 MR. ARNDT -All I'm familiar with is what they 15 have done in the elastic range so far. They have recently, 16 in the last three moths, come out with a report on the 17 concrete containment subjected to elastic loads and come up 18 with load factors for that -- the dead life and earthquake 19 load.
20 MR. SIESS: Jack, can you add to that?
21 MR. RICHARDSON: I'm not aware of any elastic --
22 MR. SIESS: I guess I'm thinking of, they were 23 talking about failure. They just didn't define failure as 24 going in elastic, I think.
25 In the load combination stuff they've got to look s
s_-
223 m
f at a failure of some sort, don't they?
V) i 2 MR. ARNDT: The limit states reflect here in shear 3
4 MR. SIESS: Okay.
l $ MR. ARNDT: -- in concrete containments and 6 concrete shear-walls.
7 MR. SIESS: Yes.
8 MR. ARNDT: They al so are ' just coming out --
9 MR. SIESS: But their limit state was first to jo yield or something?
u MR. ARNDT: Pardon me?
12 MR. SIESS: Their limit state was first to yield?
13 MR. ARNDT: It's cracking in the concrete and 34 yield in the rebar.
15 MR. SIESS: Okay.
16 MR. ARNDT: There is also tangential shear in i7 concrete containments they are just putting out a draft 18 report on.
39 MR. SIESS: You see, the limit state for a 20 containment for severe accident, not for earthquake now, 21 has got to be measured in terms of a leakage -- either a 1
22 rupture of the liner or rupture in the penetration, or a 23 large opening; right? That's the way you have to define 24 y ur limit state.
i 25 If y u are looking at a limit state for a concrete i
N
e 224 containment under seismic loading, you might -- that is
) I 2 certainly one because you assume the earthquake is going to 3 eventually cause a release. You want to know that the 4 containment is there to hold it.
5 You could start thinking of some others, enough 6 motion to drop the crane on something, I guess, or 7 attachments to it.
8 But I don't see - you know, I find it difficult 9 to define a limit state for a containment in terms of to anything other than leakage.
11 MR. ARNDT: They haven't, so far, established a 12 limit state based upon leakage. I think they are going to
__ i3 leave it to us to worry about how to apply some of that.
14 What they have done so far is, they have said, the big 15 question is how big an SSE is going to be applied to that --
16 two times SSE, three times SSE.
17 They have run several different limit states using 18 different multiples of an SSE.
19 MR. SIESS: It seems to me you do here what you 20 are doing back in the seismic margins, how much of an SSE 21 you can take before you have reached the limit stato and 22 let somebody else worry about how big the earthquake is 23 going to be.
2i MR. ARNDT: They are doing it to some degree.
25 They are saying, well, if you use a certain level, here is J
e 225
() i what you get. You multiply it by three instead of by two, here~is a different limit, different probability that you 2
I 3 will get. But they are not saying which one we should use.
4 MR. SIESS: Okay.
5 MR. ARNDT: They are just giving us different 6 steps.
7 MR. SIESS: All right. But they are not defining 8
a limit state in terms of leakage. What is the limit state 9 they are working?
jo -MR. ARNDT: It 's the material 's behavior in the ij concrete structure,.and it's the compressive capability of 12 the concrete and the yield and the rebar, as I recall.
j3 MR. SIESS: That's not a limit state for o j4 containment.
j$ MR. ARNDT: I think they used the same thing in 16 both containment and shear-wall.
37 MR. SIESS: The liner can probably take twice as 18 much strain as the rebar. I don't know whether the
- 9 equipment hatch can. But, you know, if you separate those -
~
20 21 MR. CHOKSHI: I think you will see in the 22 objectives,'the third item, fragility of concrete 23 structures. Those are the types of limit states which you 24 spoke.about we are trying to do,'one relating to the 25 leakage -- strains which will lead to the -- performance of l ,
\>
- 226 r
a i openings and penetrations.
)
2 MR. SIESS: Where is that? I
, 3 MR. CHOKSHI: That's in the far right under 4 objective.
5 MR. SIESS: Under which task? Oh, under ,
6 " objectives." Yes.
7 MR..CHOKSHI: That's right.
8 MR. SIESS: Yes. Well, they've got the right 4
9 objectives, but I don't see that they are going the right --
10
[ 11 MR. CHOKSHI: They are just starting that work.
12 MR. SIESS: Yes.
,,, . i3 MR. CHOKSHI: Their objective is to look at the
/
(_j 1 14 more' realistic limit states.
15 MR. SIESS: Okay, I think we ought to keep this in 16 mind and look at it when it gets a little farther along.
17 It certainly has a relation to something ACRS has talked 18 about frequently without defining. You may have seen the j 19 phrase. " Containment Performance Objectives" in some of our 20 letters.
l 21 I think I could relate this to that. I'm not 22 quite sure how because I'm not quite sure what they are .
j 23 doing and I'm not quite.sure what we mean by performance i
24 objectives for containment.
25 MR. BURNS: Yes, the work we have done up to date I
k
(
227
[s j on the load combinations, as was described, was based on 2 linear conditions. As far as getting into this --
J 3 arrangement, the definition of " load states" is concerned, 4 this program is just being generated and we really haven't 5
defined the state that we are really working to as far as 6 containment.
7 Obviously, it has to be related back to the 8
leakage conditions, which may be through a distortion-type 9 of condition rather than an actual failure event.
10 MR. SIESS: I think it will.
11 MR. BURNS: Yes.
12 MR. SIESS: On the other hand, we may find out i3 that first yield is so far above what we need that on
) ;4 existing reactors we don't have to worry any more.
j$ MR. BURNS: This could be true.
i 16 MR. SIESS: And I would not be happy designing 37 containments on leakage limit states with the general state 18 of the knowledge. I wouldn't be happy if we had designed j9 our containments originally on that basis for LOCAs and 20 then come along with severe accidents and have four times 21 the pressure. We wouldn't have any margin.
22 MR. ARNDT: Well, we have generated a great deal 23 f information over the last two years under this program, 24 and we are currently trying to tie this all up in a summary 25 report that will give the user a good idea of what has been l l 1
l
228
! D I done, how to use it, how to take all these NUREG reports 1
- 2 that have been issued.
t 3 MR. SIESS: That will help. Is E11ingwood going 4 to be involved in this thing at all, or is he out of i t?
5 MR. ARNDT: That's still an open item. We are 6 still discussing what should be done and matching the i 7 research program to NRR's needs. We have not gotten to 4
8 that level.
. 9 MR. SIESS: When you get to a point where you i
j 10 would like to talk to us, let us know and I'll try to find i
i 11 at least one consultant that can understand the language 12 because, I'll admit I'm not expert in the kind of i 13 probabilistic stuff Morris is doing.
14 MR. ARNDT: Right.
I 15 MR. SIESS: I'll start looking for an expert.
16 MR. ARNDT Okay, a translator, right.
17 (Laughter)
I 18 MR. ARNDT: I'm going to go on to the next set of
- 19 slides in just the second or two that I have to present to 20 you. It's not on your agenda list but it's in the package 21 that you have, and it's called " Structural Degradation."
22 This is something that we are not doing this year.
l 23 This is something we are not proposing to do in fiscal I 24 '86. That is something we are proposing and' considering 25 starting in fiscal 1987.
r (w
. - , , . m ,. -, - - . _.. - . - _ . ., ~m _~ , - , _- .
..y.,. = . _ . _ , _ _ . _ _ _ _ . , , _ , , - - _ , - _ _ - , , v,- .. , ,#. ,..-,_,--.r.
e 229 ,
What it is, is an area of study that we would (o} j 2 admit is not a high priority concern at this point in time, 3
but it's one of those items that it seems prudent to start 4 1 coking at before it becomes a high priority concern.
i 5
There are cropping up all over the place 6
discussions about extension of plant life, that the 40-year 7
design life of the nuclear power plants will inevitably be ,
8 extended for at least parts of that plant. And one could 9 assume that for the containment building and other similar 9
jo rugged structures the licensees will start looking for ij additional years of operation.
12 As a matter of fact, this has already happened on y3 Calvert Cliffs where they have gotten, I believe, a five-1 (N
(_,) 34 year extension on plant life. But that, as I recall, is 15 not due to any change in conditions but simply when they 16 g t their license for date of commercial operation --
p MR. SIESS: Since the containment is one of the T
18 first things they built, it will end up being older.
39 MR. ARNDT: Also, the utility industries such as 20 Potomac Electric Power locally has announced plans for 21 their fossil plants to review plant life extension.
22 What this does is, we can with certainty say that '
23 the NRC staff will be faced with a licensee coming in one 24 of these days -- or one of these years, rather -- saying i
25 that they want to extend their licente for "x" period of m
230 U\ t years.
2 The staff will then be faced with, here are these 3 apparently rugged structures which we are going to tack on 4 another 25, 30, 40 percent for their life-time. What 5 should we be concerned about?
6 It isn't just for concrete, it's for concrete, for 7 steel, for pre-stressing tendons, for materials that are .
8 e.mbedded, such as poly-vinyl chloride water stops for 9 insulation that is attached. There are chemical reactions 1
to that take place in the concrete over a long period of time 11 that are known to the industry but within the usual life 12 span of a commercial structure are not a concern.
13 We would want to be able to identify these, if
, 14 they exist or not. There are questions of this type.
15 And there is an additional point of interest which 16 is not the primary one but contributes. That is the 17 experience from Three Mile Island where we found that we 18 had a very hostile environment within the. containment 19 building where everything got sprayed with chlorated water, 20 where we had a hydrogen burn take place. Where we had 21 radiation present. We had very high humidity present, and 22 we had, as usual, a very warm interior, all of which are 23 very conducive to a attacks on materials.
24 At the same time, if anything took place you 25 couldn't go in and do any repairs or adjustments to
/
O
e 1
231
() i 2
anything. And this post-LOCA environment within the containment and its effect on supports to components, to 3
various structural things that are within there, is
)
4 something we think might merit a little closer look because 5
nobody can go in during that however many months it was at ,
6 Three Mile Island to do anything.
7 Some of the supports and supporting systems could 8
be subject to attack under that environment.
9 MR. EBERSOLE: That won't go so far as to consider jo core melt and vessel failure, will it? .
11 .MR. ARNDT: No because I would say if everything 12 has gone to pot and you have blown the containment apart, 1
13 you are going to worry about it.
() ja MR. EDERSOLE: Well,.there is lots of hypotheses 15 that say you can melt the vessel and metl the core and 16 still live through it.
37 MR. SHEWMON: His concern is whether you are going is to re-license it.
39 MR. ARNDT: There are several things thrown into 20 the bag here, and re-licensing is one of them. Whether we 21 have looked sufficiently long at the hostile environments
~
22 that can exist in a severe accident or post-LOCA condition 23 is an ther question.
24 MR. SIESS: That's not a very fruitful line of 25 endeavor. For the one that might occur, you know, there i
/"N i
l k_
1 232 7
1 might be a 50 percent chance of another one occurring in 2 the life of these plants.
3 But looking at ordinary deterioration from 4 ordinary sources -- of course, a lot of these areas, you 5 realize, there is a lot known about this.
6 MR. ARNDT: That's true.
7 MR. SIESS: If you fi'nd somebody that can find out 8 who knows it, you are going to get a lot of answers.
9 MR. ARNDT: But the main thing is that we 10 identified for the NRC staff what should we look at when 11 somebody comes in and says, "We want another 15 years for 12 the containment building, 20 years."
13 Do we just say, "Oh, yes, it's a concrete 14 building, it's fine. You've got it."
15 MR. EBERSOLE: Let me ask a hypothetical 16 question. Let's say TMI 2 in fact did succeed in melting 17 the primary vessel and it fell on the floor. Would you 18 pump water in?
19 MR. ARNDT: Would I pump water in?
20 MR. EBERSOLE: Yes. Somebody is going to have to
, 21 say yes or no. I'd like to ask a concrete man.
22 MR. SIESS: What do you mean? -
23 MR. EBERSOLE: Well, I'm going to pump water in, 24 so something better than melting concrete might occur.
25 Shoot that out of the saddle.
j d
l 0 _ - _ _ . .
.- = _ _ - . _
233
() i 2
MR. SIESS: You pump water in, you get a lot of steam and bust the containment.
3 MR. EBERSOLE: How do you know?
4 MR. SIESS: If water turns to steam -- that core 5 is a lot hotter than that.
6 MR. EBERSOLE: There may be condensing systems.
7 It's perfectly capable of coping with that.
8 MR. SIESS: Not from what I read.
9 MR. EBERSOLE: No. I'm just saying, we may be io confronted with that question like, you know, who knows it when?
12 MR. SIESS: There is enough water in the concrete 33 to make enough steam to bus the concrete, just melt it
' s ,/ ja through the basemat.
i3 MR. EBERSOLE: I'll ask the question as it 16 stands. What do you do?
17 MR. SIESS: I don't think you will get the answer is out of a structural engineer.
19 (Laughter) 20 MR. EBERSOLE: I don't know who to get it from.
21 MR. ARNDT I would not pass myself off as having 22 the answer to that question.
23 MR. SHEWMON: .Let me go back to a more fruitful, 24 m re immediate question --
25 (Laughter)
\
e 234
) 1 MR. SHEWMON: -- what does happen to the ten-year 2 extension? Do you think you could go out and learn enough 3 from people what is known~before? Are there special kinds 4 of tests?
5 What you presented is very open-ended. To what 6 extent is there field knowledge on the existence of these 7 and are there indeed degradation processes that you know 8 of?
9 MR. ARNDT: Yes, there are degradation processes 10 we know of. If a licensee came in and asked for a plant 11 ex tensi on , we would have to look at the structures, the 12 design of the structures and see what would perhaps be 13 vulnerable to these processes.
i
~~'
(
(_) 14 For example, it would make a difference if it were 35 prestressed containment or reinforced containment. We 16 might look at --
17 MR. SIESS: Relaxation.
is MR. ARNDT: -- prestressing which we have an in-19 service inspection program for. We have also had questions 20 come up in some of the bridge research models, concrete 21 reactor vessels. The thermal input on the organic -- that 22 is used to protect the prestressing tendons might produce 23 s me by-products that attack tendons and fail them.
24 MR. SIESS: That's what you got a surveillance 25 program for.
!Q b
l l
l
235
() MR. ARNDT: True. Within the concrete itself, you have chemical interactions; aggregate silica reaction that 2
3 can be very low-grade types of reactions, but start to 4 become more important as time goes by -- 10, 12, 15 years 5 later or 20 years later.
6 MR. SIESS: But not 407 7
MR. ARNDT: I don't know.
8 There was one case in Canada, Ontario Hydro, where 9 they suspected they had that, or at least the beginnings of 10 it. They sealed the surface of the containment, took the i ji water out and stopped the process.
12 MR. SIESS: You might tell them to replace all the 13 prestressing tendons, start over.
O. j4 MR. RICHARDSON: Well, there is a philosophical 33 statement, I think, that should be inserted here. That is g the objective of any research would not be to solve g7 industry's problems for them-for life extension, but to 18 Provide the licensing staff with the right questions to ask 39 industry on what critical areas they should provide 20 substantive --
21 MR. SIESS: Well, that's one way of putting it, 22 Jim. But we just went through a pretty good list of 23 questions you could ask right now. But what you want to do 24 is narrow it down to the important questions to ask --
25 MR. RICHARDSON: That's right.
O
, O .
l l
a 236 l
1 1 MR. SIESS: -- because I think that's very (V7_a}
2 important. I think this is deserving of some l ow-l evel 3 worrying'and getting maybe somebody to look at it and tell 4 you some things you don't need to worry about.
~
5 MR. RICHARDSON: Exactly. .
6 _MR. SIESS: You know, a lot of this is pretty well 7 known stuff.
8 MR. RICHARDSON: We certainly don't see this 9 developing into any big, long-term research program.
10 MR. SIESS: This is your share of aging?
11 (Laughter) 12 MR. RICHARDSON: Yes.
13 MR. ARNDT: It's a little bit more definitive.
/ 14 Aging is a very broad term and maybe nothing happens. But 15 here we are looking at various materials and how they 16 deteriorate with time, and what mechanisms are present.
17 MR. SIESS: We have had some concrete around for la 2,000 years.
19 MR. ARNDT: We have had some concrete fail very 20 rapidly too.
21 MR. SIESS: That's not around any more.
22 MR. ARNDT: Right.
23 MR. SIESS: So we just have to figure out what's 24 good about that 2,000 year concrete and about the stuff 25 that isn't.
h n
a 237
- i MR. ARNDT
- And, as I say, the term " structures" 2 or " structural degradation" is not strictly limited to 3 concrete.
4 MR .' SIESS: Steel rusts.
5 MR. ARNDT: Poly-vinyl chloride insulation on 6 liners. We may be looking at the liners because the PVC is 7
known to deteriorate and off-gas HCL which produces 8
hydrochloric acid can be particularly harmful to stainless 9 steel penetrations, may be going through the liner.
10 MR. SIESS: How often do seals get replaced, is penetrations? Do they get replaced?
12 MR. ARNDT: They do get replaced. I'll give you a 13 theoretical answer and that is based upon the leakage y,) 14 testing program --
15 (Laughter) 16 MR. ARNDT: -- they should establish the need for 17 replacement at what frequency. The same thing applies in 18 electrical cable insulation, how rapidly that deteriorates.
j9 MR. SIESS: When they open an-equipment hatch --
20 which isn't very often, I assume -- do they replace the 21 58815?
l 22 MR. ARNDT: The'only ones I am familiar with, I 23 don't think have seals there, are seal-welded on the 24 equipment hatch.
25 MR. SIESS: All of the studies that have been done O
l
e l
l 238 on the leak work -- drops seals and all of that?
FmV) 1 2 MR. ARNDT: On personnel hatches. I imagine some 3 equipment hatches do have seals. But I don't know at what 4 frequency --
5 MR. RICHARDSON: If they are all welded, we are 6 doing something wrong on our containment tests because they 7 do have seals.
8 MR. BENDER: May I ask a standard question? Is 9 there a list of materials that you are investigating?
10 MR. ARNDT: No, we haven't reached that stage.
11 This was simply a --
12 MR. RICHARDSON: This is the 1987 program.
_ 13 MR. BENDER: Today is the time to make the h
14 materials list. You don't have to have a research program 15 to decide what problems are out'there. I just don't 16 understand why you don't do it.
17 MR. ARNDT: We do have candidate materials. There is is concrete, steel, poly-vinyl chloride --
19 MR. SIESS: There is a list.of them there.
20 MR. ARNDT: -- prestressing tendons, the organic 21 compounds --
22 MR. SIESS: Anchor heads, please.
23 MR. BENDER: Is that really "the" list or is that 24 "a" list?
25 MR. ARNDT: I would say that is "a" list.
(vD l
l
e 239
() i MR. SIESS: If they had "the" list, they wouldn't need the research program, Mike.
2 3 MR. BENDER: Well, I understand what you are 4 saying, but the point I want to emphasize is, you make a 5
laundry list and then you go down through it and say, "Here 6
are the problems and here is where they are." And you.also 7
ask, "Well, what is the industry doing to worry about these 8
things? Why am I worrying about them myself?"
9 It seems to me that's part of the advance research in planning that ought to be done even though they are not 11 planning on doing the research and come back in '86.
12 MR. SIESS: .Anything else for Gunther?
- 13 (No response) s, 14 MR. SIESS: Is that all your presentation?
15 MR. ARNDT: Those are all.
g MR. SIESS: Okay. The farther in advance the j7 project is, the more paper we get.on it, it 'looks like.
18 (Laughter) 19 MR. ARNDT: The better we try to justify it.
20 MR. SIESS: Okay, let's take a short break.
21 (Whereupon, at 3:10 p.m., a recess was taken until 22
.3:28 p.m.)
23 MR.-SIESS: Okay, one more item under the heading 24 of what, seismic; right?
! 25 MR. RICHARDSON: .Yes.
I (~h t
,]
e 240
';m~
MR. SIESS: I just couldn't wait for that break,
(% )) 1 2 though.
3 MR. RICHARDSON: Herman Graves is our project 4 manager on this.
5 MR. SIESS: The first thing we are going to hear 6 about, computers.
7 MR. GRAVES: Benchmarking.
8 As Jim said, my name is Herman Graves and I am the 9 project manager for the program on standard' problems for 10 structural computer codes.
11 (Slide presentation) 12 MR. GRAVES: This is beginning with the slides 13 after the structural degradation that Gunther Ar.ndt was
(,,/ 14 just talking about.
15 The contractor is Brookhaven National Laboratory, 16 and you see the budgets there on-the page. We are spending 17 about 2OOK in FY '86.
18 The issue here is that various analytical 19 approaches with different degrees of approximation have 20 been developed and used to determine structural response 21 and source-structure interaction for nuclear-structures 22 subjected to seismic and dynamic loads.
23 These approximations used may or may not represent 24 the physical behavior of the structure. Thus, we see a 25 need-to establish benchmark solutions to determine validity
(~
\
e 241 er i and accuracy as well as the confidence and analytical i
)
2 methods that people use.
3 Thus the objective is to determine or to establish a problems with experimentally known solutions, " Benchmarks" 5 for use by the licensing staff to validate major plots of 6
licensee methods used to calculate structural response and 7
soil-structure interaction of safety-related buildings, 8
systems, and components. -
d 9 Now, these solutions will be compared or will be 10 based partly on data that has been collected on it earthquakes.
12 The program is integrated as follows:
! 13 The findings can be used in efforts to determine l
( 14 seismic safety margins; experimental results from various 15 studies currently being performed for the NRC at Sandia 16 National Labs with containment integrities will require 37 consultants or. engineering characterization of ground 18 motion. Anu Los Alamos National Labs has one structures j9 program.
20 These studies will, we feel, compliment this 21 Program.
22 Also, we have quoted in some work sponsored by 23 EPRI on the Simquake experiments.
24 And the last part of integration we have gained 25 through a cooperative agreement data from Fukushima I
h1 G
l l
e 242
, r;"~s (U ) 1 earthquake..
- 2 Scope and accomplishments:
3 For FY '85, the work was concentrated primarily in 4 the source structure interaction area, specifically with 5 determining the effects of foundation lift-off, high 6 groundwater table, and soil layering on SSI results.
7 In FY '86, we plan to complete our studies on high 8 groundwater table, We are going to initiate a structural 9 data survey, and start to evaluate the data that is to collected.
11 Finally in FY '86, we hope to sponsor a workshop 12 on source structural interaction.
__, 13 MR. RICHARDSON: Herman, could you expand a little i
' (N,,/\ bit on the structural data survey, what type of data you 14 13 would be gathering?
16 MR. GRAVES: Hopefully, we want to get some data 17 the predicts structural response not only in the building 18 forces that a structure might see, but what is.the capacity 19 of that structure strength wise. That's basically the type 20 , of data that we hope to get.
21 We are going to be primarily concentrating on 22 reinforced concrete structures.
23 MR. EBERSOLE: May.I ask a question about this 24 f undation lift-off and higher groundwater table as it 25 relates to the maximum possible floods that we have to look a
.e l
243 at at these plants?
) i 2 I was fascinated to find out that the thesis at 3 Browns Ferry of providing eight-foot high dikes and 4 protective walls was invalidated because they never did 5
know how badly the foundation leaked even when they had a 6
modest rise in groundwater.
7 Are they due to deterioration of some of the 8
seals, which would have meant that the logic was false in 9 the first place. They simply would have been drowned had a jo flood occurred. They couldn't even hold high groundwater ij without some sump pumps they didn't have.
12 So how does this relate to the fact that once in a
. 13 blue moon you are going to have water that is going to come ja way up above the ground in some of these plants?
15 MR. GRAVES: Well, I'm not that familiar with 16 Browns Ferry. But there is'a plant in Louisiana that's, I j7 believe, below sea level.
18 MR. EBERSOLE: Grand Gulf?
j9 MR. GRAVES: Not Grand Gulf.
20 MR. EBERSOLE: Riverbend? Oh, wait a minute, the 21 ne at New Orleans, Waterford.
22 Waterford, yes.
23 MR. GRAVES: Yes, Waterford. It's not below, it's 24 just slightly above.
25 MR. EBERSOLE: It's below river level.
O v
i
1 l
244 i% i i MR. SIESS: Below river level.
2 MR. EBERSOLE: Yes.
I 3 MR. SHEWMON: The river doesn't fall too fast down 4 there.
5 (Laughter) 6 MR. GRAVES: Well, below river level.
7 MR. EBERSOLE: Are you talking about including 8
' consideration of these major high water levels?
9 MR. GRAVES: What we are talking about here is to -
10 - we want to look at situations such as Waterford to 11 determine what effects that high groundwater would have on 12 the source-structure interaction.
13 In other words, determine the effect of neglecting u high groundwater situations.
15 MR. EBERSOLE: Soil-structure interaction. And 16 you are talking about it in the context of seismic?
17 MR. GRAVES: Yes.
18 MR. EBERSOLE: Oh, but you are talking about high 19 groundwater coincident with the seismic event.
20 MR. SIESS: Now the plants, the more recent plants 21 are all designed for lateral earth pressures'due to -- or 22 lateral pressures due to fairly high groundwater levels; 23 right?
24 I know in the SEP plants, that was one of the l 25 deficiencies on every plant. The fix was either to l
i v
i f
_ . - = . .
a
. 245
'l
() i determine where the groundwater.couldn't get above or to assume that it was at the surface.
2 f
- l. 3 But then you are saying that they haven't always 4
J 4 considered this in a source-structure interaction case.
5 MR. GRAVES: Well, I'm not saying that they ,
6 haven't always considered it. But what we are looking at 7 here, just like we are looking at the methods to put some
! kind of confidence level on the analytical techniques that 8
~
9 are used to predict the effect of a high groundwater table.
jo MR. SIESS: Now, have we got some tests that'would j ij tell us we could use to benchmark that? I mean, the 12 essence is, we have to have tests somewhere, experiments.
13 MR. GRAVES: We don't have any experimental data, I
- ij or we haven't used any data dealing with groundwater tables i
15
.in this program. We hope to get from our studies or our 16 Cooperative agreement with EPRI from the Taiwan experiments
} 17 1
j- 18 MR. SIESS: Simquake must have had a low water i
ig table.
20 MR. COSTELLO: Jim Costello --
l 21 .MR. SIESS: It was'in.a desert somewhere, wasn't 22 it?
I 23 MR. COSTELLO: -- with the NRC staff.- -
24 That is correct that the Simquake was at the i
25 McCormack Ranch in Albuquerque, and that's about as dry as l0
.2
.~ . . _ . .. - . _. _ , . . _ _ , , _ _ _ _ _ _ ,,_.,2._____. _ _ , . . . . _ _ . ~ _ _ .. .
4 e
246 I l
I
[e~\ 1 you are going to get.
l 2 The Lotong site in Taiwan has got a very high 3 water table.
4 MR. SIESS: Okay.
5 MR. COSTELLO: The question that arises has to do 6 with the analytical modeling that is traditionally done.
7 It does not, for example, look at the mixed fluid soli'd 8 problem ALABIO, and the question is, would that give you a 9 .significant difference in your answer?
10 MR. SIESS: Has anybody got a source-structure 11 interaction model of a PECA 2 Phase material?
12 MR. COSTELLO: Yes, sir. There have been a 13 couple. One of them Professor Costantino has developed --
/%
k,) 14 MR. SIESS: But they don't use them.
15 MR. COSTELLO: Not usually, no.
16 MR. SIESS: Okay.
17 MR. MARK: There must have been some consideration 18 of a high water table in connection with any seismic 19 analysis of the floating nuclear plant.
20 (Laughter) 21 MR. SIESS: Well, that happens to be a single-22 phase system, though. I think it's the two-phase system i
23 that bothers you.
24 MR. BENDER: Somehow I get the feeling that 25 benchmark is not quite the right term for this.
\J.
e 247 I)
V i MR. GRAVES: Yes, that's why we call it pretty 2 much standard problems for --
3 MR. SIESS: Benchmark has been used so long, let's 4 don't change it.
5 MR. BENDER: I guess I'm too old. I have some 6 recollection of what " benchmarks" used to be.
7 MR. SIESS: Yes. You could call it validation, 8
except that has another meaning. And it's not really a 9 standard problem because that's what you usually think of jo as you try several methods on a standard problem and most 11 of these are only going to be one method.
12 We are just trying to find if the best one we've 13 got is good enough.
34 MR. GRAVES: Well, actually, Professor Si ess , on 15 this effort for the last couple of years the Brookhaven 16 subcontractors have actually tried to look'at the elements 17 that go into both fine element models and lump spring is models.
39 MR. SIESS: We are still on soil structure.
20 MR. COSTELLO: Yes, sir. And to see to what 21 extent or how well these elements captured certain pieces 22 in the Simquake data and in the Fukushima data.
23 That's the substance of the report that Mr. Graves 24 has in his last slide.
25 MR. SIESS: By spring model, you mean simple O
248 spring --
) I 2 MR. GRAVES: Yes,'that's lump parameter.
3 MR. SIESS: Lump parameter.
4 MR. MARK: Fukushima.
5 MR. SIESS: I remember Mr..Chan of the NRC staff 6 used to prefer that because it was a lot easier to see what 7 was happening and what the variables did with the lump 8 parameter model than with finite elements. It was probably 9 easier to run several cases and find out how things varied.
10 MR. MARK: Fukushima, I believe you said, had a 11 high water table.
12 MR. COSTELLO: No, sir, it was the Lotong site in i3 Taiwan where the EPRI experiments are going to be run.
I MR. MARK: So there will be shock-induced defects?
14 15 MR. COSTELLO: We would think so, yes.
16 MR. SIESS: This is the instrumented system --
17 MR. SHEWMON: When was the Fukushima earthquake?
18 MR. COSTELLO: There were two that were reported 19 in.the literature. One, I think, was in the late '60s and 20 one was '79.
21 MR. SIESS: The proper question was, when was the 22 last Fukushima earthquake. Then you can say when is the 23 next one.
24 MR. COSTELLO: It's known as the Sindai 25 earthquake.
/"T
e 249
(% i MR. GRAVES: This Fukushima earthquake was 7.4 5
N 2 magnitude.
3 MR. SIESS: But not at the site.
4 MR. RICHARDSON: No, not at the site.
5 MR. SIESS: What kinds of "g" values did they see 6 at the site?
7 MR. COSTELLO: The peak horizontal recording, I 8 think, was around .3 or .2.
9 MR. GRAVES: .2, .3.
10 MR. COSTELLO: I forget. We have a limited subset ij of records that Tanaka made available and those are what is 12 used.
13 MR. SIESS: Are the Humble Bay figures useful for q,) ja the benchmarking?
15 MR. GRAVES: We had taken a look at that data.
g MR. SIESS: I mean, Harry Seeps has already used 17 them.
ig MR. GRAVES: Yes. I think the contractor 39 determined that there wasn't enough information there or it 20 wasn't instrumented properly to really take a look at and 21 evaluate that. .
l l 22 MR. SIESS: Don't 1et them get their standards 23 too high. Let's take whatever data we can get.
24 MR. GRAVES: Yes. Thus far, we got the Fukushima l 25 and the Simquake.
I
^
%d
e 250 In FY '85, we hope to develop SSI benchmarks and
) 1 2 structural concrete benchmarks, and finally evaluate any 3 new data that becomes available.
4 MR. RICHARDSON: You said '85.
5 MR. GRAVES: '87.
6 MR. SIESS: Structural concrete is the CAT-1 7 structure data.
8 MR. GRAVES: Yes, that would be.
\ 9 MR. SIESS: I mean, that's shear-well type stuff.
10 MR. GRAVES: Shear-wall type stuff, yes. And 11 other data that we can find.
12 MR. SIESS: You might even have a containment by
- 13 then.
. 14 MR., RODABAUGH: Well, getting back to this 15 structural data survey, what are you going to be trying to 16 gather there, just SSI information?
17 MR. GRAVES: No, this would be data on concrete 18 structures, the structural response due to an earthquake.
19 MR. RODABAUGHz I guess my problem is, I seem to
- 20 see a lot of overlap in the various programs with respect 21 to the data gathering phase.
22 MR. GRAVES: Well, in this program we are 23 basically looking at the analytical techniques.
24 MR. RODABAUGH: Yes, but.I'm talking about the 25 data gathering.
l
(>
\-
i l
l l
., .- - - ,n --- , .
e 251 MR. GRAVES: Right, right. The other programs are
( i 2 on the experiments and generating the data, and then we are 3 taking-the data and massaging it, taking a look at it and 4 comparing the equations to actual data.
5 MR. SIESS: They are going to gather data from 6 other people, other projects.
7 MR. RODABAUGH: Yes, but I'm trying to see the 8
difference between, for example, the designs margin people 4
9 were doing the same thing, were they not?
ja MR. SIESS: No, no.
11 MR. RODABAUGH: For structures?
12 MR. SIESS: No, they are getting data on 13 survivability-type stuff, and this was mainly predicting
) i4 behavior to deformation and loads, and stresses for design.
15 This is trying to find out how good the codes are.
16 MR. RODABAUGH: Yes, I understand that.
17 MR. SIESS: And the seismic margin really isn't is looking at how good they are, they are using what they've i
i 19 got.
20 MR. RODABAUGHz Yes. But why, then, would this 21 program do a separate effort to accumulate data?
22 MR. RICHARDSON: Well, in many cases it . won 't be 23 separate, they will be using a common data base.
24 MR. SIESS: Oh, I'm sorry.
25 MR. RODABAUGH All right.
O V
I
e 252 MR. RICHARDCON: Wherever possible, we use common
) 1 2 data.
3 MR. RODABAUGH: That's why I was asking what you 4 are going to do in initiation of structural data. It may -
5 be that it's just a supplement.
6 MR. RICHARDSON: Oh, yes, yes. And wherever 7 possible, it will be.
l 8 MR. RODABAUGH: My curiosity was also prompted 9 because earlier you talked about, you got a look at these 10 computer programs for SSI -- as I understand -- systems and 11 components. What are you thinking of in a component in the 12 context of this program, what is a component?
,_, 13 I'm looking at the first page of the handout.
, 14 MR. SIES'S: Oh, the response of systems and 15 components.
16 MR. RODABAUGH: Right, response of systems and --
17 MR. SIESS: Now, by response of components you are
, 18 talking about things like floor spectra?
19 MR. GRAVES: Right, floor response spectra.
4 20 MR. RODABAUGH: Okay, you are not talking about 21 response of the pump sitting on the floor.
22 MR. GRAVES: No.
23 MR. SIESS: We need some better words.
24 MR. GRAVES: .The regulatory use for this program 25 has been that preliminary efforts so far have been used by
>~
b 1
e 3
253
() i 2
NRR and the Structural Geotechnical Branch to resolve SSI issues on near-term operating licensings, such as Byron, 3
Beaver Valley-2, and Diablo Canyon nuclear plants.
4 Benchmarks will be used to assess the accuracy of 5
the licensees' methods to predict the behavior of safety-6 related structures under seismic loads.
7 Finally, benchmarks and other data developed will 8
be used to evaluate, revise, and where necessary establish 9 new seismic design criteria, for example, revised sections jo of the SRP or revised regulatory guides that assist in this ji area.
12 MR. SIESS: I don't really think that what will come out iof this will affect design criteria. It might 13 ja affect how you tell people to analyze something. But 15 design criteria to me is more related to limit states, and 16 this is strictly related to the codes you used to analyze.
17 Some people don't separate them, but I do. Who is ,
jg doing this at Brookhaven, the same people that are doing i9 this other stuff? No, this is not probabilistic.
20 MR. GRAVES: Professors Miller and Constatino from 21 City College.
22 MR. SIESS: Okay. I'm glad to see something on 23 s urce-structure interaction because a couple of years ago-24 benchmarking was comparing one code with another code and 25 if they both gave the same answer, people were happy.
\
l J
i l
I
e -
254 Any other questions on this?
V(K) 1 2' (No response) 3 MR. SIESS: Okay, thank you.
4 MR. BURNS: Chet, just one ccmment there. In 5 1986, we are going to have a workshop on source-structure 6 interactions. We.will notify the ACRS when that comes on.
7 MR. SIESS: Yes. After the meeting next week, two 8 weeks from now, we'll have our schedule for '86.
9 In case you don't get it, it's on the ACRS to bulletin board system. Dial up and ask for 4-3147, or 11 whatever it is, and the ACRS schedule is on there. We are 12 going to post it on the bulletin board out here, I think.
__ 13 So far, they have hit two meetings.
m 14 There is no charge for the ACRS bulletin board.
15 (Laughter) 16 MR. SIESS: All right, we are changing the subject 17 but we are still in my subcommittee, right? I'm looking 18 for an opportunity to turn the chair over to Dr. Shewmon.
19 Okay, we'll give you the floor.
20 MR. GUZI: I'll be talking about the piping that 21 our branch will be sponsoring in the next few years. I 22 guess before I start, a little overview --
23 MR. SIESS: That's an optimistic statement if I 24 ever heard one.
25 (Laughter) t
(
l l
e 255 fm MR. GUZI: Basically, of all the piping
( )
2 restructure I'll be talking about today, all of these are 3 essentially coming out of recommendations made by the NRC 4 Piping Review Committee.
5 The break of what we do and what the Materials J
6 Branch do, you can take it of the NUREG-1061 that the 7
Piping Review Committee put their recommendations in. We 8
are addressing those research needs in Volume 2 and Volume 9 4, which are the volumes of the Task Group on Seismic jo Design and the Volume from the Task Group on the Evaluation it of Other Dynamic Loads and Load Combinations.
12 So, the research I will be talking about will not 13 address pipe crack, pipe break, inspection to the main ms/ 34 extent, I guecs. It addresses itself to dynamic response 1
15 and failure modes.
g (Slide presentation) 17 MR. GUZI: Now, in my presentation I have broken is the research down to about three groups. Of the Volumes 2-39 and 4, there are about 12 ticket items that the Piping 20 Review Committee recommended, and I have grouped them here, l
73 for this presentation, in categories of piping dynamic load 22 capacity, damping, and other research.
. 23 The research as a whole, I guess, has been 24 prioritized by the Piping Review Committee. We have had a 25 research plan which has been circulated and has been
?
(7 1
256 tr( l 1 reviewed by NRR. That addresses the recommendations in
} l 2 Volume 3 and 4, and Mark Hart: man has been involved in that 3 review.
4 So, I guess with that I'll start in with the first 5 category which is piping -- and the title, the first title 1
6 is NRC/ERPI Programs on Piping Dynamic Load Capacity.
7 In there, there are about two on-going programs 8 and about three items in later years that will be covered 9 under this one titic.
10 The main program in this is a cooperative effort 11 between the EPRI and NRC, which is the piping and fitting 1
12 dynamic reliability program. This is a program consisting 13 of, initially, of tests, of piping at the component level --
P~
()$ ' t .t which I mean elbows and "ts."
13 The specimen level, which is in this case f atigue 16 ratcheting specimens, and the systems level. And once the 17 tests are done, there will be analyses and recommendations
.i j is for code rule changes. .
19 Another program that is' covered in this general 20 category is the ETEC demonstration seismic fragility pipe 21 test. This is separate from the main EPRI/NRC program but 22 it's being performed in conjunction with it, and it's part 23 of cur, the NRC's research agreement to essenti ally sponsor 24 this.
25 It will also serve -- hopefully, it will serve as f
(
l l
I e - . . _ _ - . . _ . .
257 i a precursor to the systems tests that will be done later on (S}
2 in the main EPRI/NRC program.
3 I mention three other items here that will not 4 begin until probably next year and will probably be fully 5 underway in 'S7. This is the independent assessment of 6
pr p sed code rule changes, and I'll get into that a little 7 later.
8 MR. SIESS: Proposed by whom?
9 MR. GUZI: The EPRI/NRC program will -- the main 10 contractor is General Electric, and under the eye of NRC ij and EPRI and also a group of. consultants, GE will develop 12 recommendations for code rule changes.
13 So, essentially it will be GE --
14 MR. SIESS: Will a review committee be set up?
15 MR. GUZI: Well, no. This will be a change to 16 design rules. The piping --
37 MR. SIESS: Oh, stress.
18 MR. GUZI: For stress, for stress rules. The 19 Piping Review Committee looked at possible reclarification 20 of seismic inertial stresses. They discussed them but they 21 didn't reach any conclusions to make a recommendation.
They did in lieu of that recommend this type of research.
{ 22 23 The-next item would be some sort of synthesis or 24 integration with the degraded piping program. The main 25 program now, the EPRI program, is looking at piping as l
O
e 258
) 1 built, which'is uncracked or not cracked more than a new 2 piping system.
3 But at the same time we realize that the strength 4 of piping is affected by the cracks that grow, and we have 5 this other program under the Materials Branch for a 6 degraded piping program.
7 Well, there are areas of overlap and areas where 8 the two.could be sort of pullad together to give a 9 realistic information.about failure levels and failure 10 modes of piping that's in service. I'll get into that if 11 you want me to later on. But this is something we see 12 being done after our cooperative effort is done and after
- 13 the degraded piping program is farther along.
\ 14 The last item has to do with using this 15 inf ormation f rom the programs f or a better defis.a tion or 16 validation of piping fragilities that are.used in the PRA.
17 MR. SIESS: Dan, I think I missed something. Your 18 activity connects when the Piping Review Committee is 19 completed, then.
20 - MR. GUZI: The activity -- this is a research 21 activity that was recommended by the Piping Review 22 Committee.
,23 MR.1SIESS: Now, there were no questions that came 24 up about shifting and --
25 MR. GUZI: Yes. Those will be addressed under my
~/
e 259 m
( ) i third grouping, our second grouping, of --
2 MR. SIESS: I want to interrupt you completely and 3 say something to Jim for a minute. When we were talking 4 the seismic margin thing, what was not addressed was what-5 is being done about the specific research needs recommended 6 by the expert panel.
7 MR. RICHARDSON: Yes.. Oh, what I think Dan 8
identified, what the committee had recommended, some of 9 those fall outside of this branch's area and fall into the 10 DRAO, such as human factors.
11 MR. GUZI: But one of them was piping fragility, 12 which this will, this program will partially -- I mean 13 component fragility --
rO MR. SIESS: All of those that fall under yours are (ms/ 9 15 covered somewhere.
16 MR. RICHARDSON: No. I have to confess design and 17 construction errors.
18 MR. SIESS: I don't know who that belongs to.
j9 MR. RICHARDSON: Well, we are still studying a way 20 of -- we right now don't know enough to really define a 21 good resec,ch program to do it.
22 MR. SIESS: Fragility, you think --
23 MR. RICHARDSON: We are handing that through the 24 c mp nent fragilities.
25 MR. SIESS: Seismic hazard clearly isn't.
A.
\ I l x.)
(
l l
l l
e 260 MR. RICHARDSON: Yes.
) 1 2 ;iR. SIESS: System and operator interaction --
3 MR. RICHARDSON: That's DRAO.
4 MR. SIESS: Design and construction errors, you 5 are not sure. ,
6 MR. RICHARDSON: If we can ever define how to do 7 it,'we'll do it.
8 MR. SIESS: And modeling and analysis?
9 MR. RICHARDSON: Division of Risk Assessment.
10 MR. SIESS: Modeling and analysis?
11 MR. RICHARDSON: I beg your pardon?
12 MR.~SIESS: Modeling and analysis?
13 MR. RICHARDSON: Modeling and analysis?
\
(g_,1 14 MR. SIESS: Yes. Non-linear behavior of 15 . ductility.
16 MR. RICHARDSON: Yes.
17 . MR . SIESS: That comes under --
18 MR. RICHARDSON: That's under validation and some 19 of the work that Gunther talked about.
20 MR. SIESS: Yes, okay.
21 MR. SHEWMON: Would you tell me what you are going 22 to do on this program? I.have heard you say it and I can 23 read it. I still don't know what the issue is.- If I go 24 back and read the issue, ASME code design rules for piping
, 25 -of under -- I'm not sure, "of"-shouldn't be there maybe?
V(~$
~ ' , ~
261 r MR. GUZI: I think "under" shouldn't be there.
i (v) 2 Piping of dynamic loads.
3 MR. SHEWMON: Okay. What aspect do you feel is 4 unrealistic, or what are you trying to change?
5 MR. GUZI: It's the definition of inertial loads 6 and if you are familiar with the code, it's Equation 9.
7 Right now, the way we design piping for inertial loads we 8
c nsider to be a primary stress. It's treated the same as 9 .a static load, a pressure load or static load.
10 MR. SIESS: It's an allowable stress question.
11 MR. GUZI:- Yes. It's more than just the stress 12 value, it's the way you classify it and the way you treat 13 it.
ja MR. SIESS: Yes. But I mean, it makes a 15 difference in how thick it has to be --
16 MR. GUZI: Well, where supports have to be and 17 what your support spacing is on it. On your FUD analysis 18 you compare it to allowable stress, j9 MR. SIESS: Yes.
20 MR. GUZI: Essentially, if your allowable is 21 higher, you need less supports.
22 MR. SHEWMON: And for already built plants this 23 w uld allow them to take out supports --
24 MR. GUZI: -Yes.
25 MR. SHEWMON: -- after they had done if for some O)
( ,_-
l
262 1 reason they wanted to to ease inspection.
2 MR. SIESS: Or it might show higher margins.
3 MR. RICHARDSON: You might use that as a way out 4 of not modifying the pipe system in light of a higher 5 earthquake being imposed.
6 MR. GUZI: We haven't reached a recommendation 7 yet, but I.think if inertial stresses can be reclassified 8 as secondary stresses or something similar, this should 9 have a bigger impact on piping design than damping has now, 10 which is our Number 1 candidate as far as getting rid of 11 snubbers and restraints.
~
12 So this goes along with snubber minimization,
__, 13 getting rid of, you know, unnecessary supports et cetera.
f )N
.\ 14 MR. SHEWMON: Now, is the fatigue ratcheting one 15 word like "Damyankee," or something, or are these two 16 different things?
17 (Laughter) 18 MR. GUZI: It'sJa combination of both. It is a i 19 hyphenated word, perhaps. It's a phenomenon of ratcheting 20 that occurs in failure by fatigue. The ratcheting being
[ 21 seismic inertial loads in combination with either pressure-I 22 loads or static loads, and the failure mode which has i
23 analytically been thought to be_the primary mode.
24 It would be -- under static, the thing would, the 25 piping system, would gradually go down until you got~large
(T b
e 263
() i strains and you got fatigue failure.
MR. SHEWMON: Are you saying it sags?
2 3 MR. GUZI: Yes.
/
4 MR. SIESS: Without fatigue failure?
5 MR. GUZI: Well, I think if you get the strains 6 large enough, you get essentially --
7 MR. SIESS: Cyclic.
8 MR. GUZI: Cyclic-type failure, okay. Also, with 9 the pressure, some of the tests which have been run --
10 limited tests -- show essentially a ballooning out in an ij elbow or something and then cracks growing in that.
12 MR. SHEWMON: So in that sense there is a cyclic --
13 as y u cycle the stress in the system, the stuff g,) ja macroscopically buckles or distorts.
15 MR. GUZI: Distorts and essentially if you have 16 large strains and you have cracks, the cracks grow.
37 MR. RICHARDSON: Every time that your strain 18 exceeds the yield, the constant pressure force there allows i9 it-to expand and then, when the load is relaxed because you 20 had plastic deformation, it doesn't come back and it just 21 keeps ratcheting on up.
22 MR. GUZI: It's a function of your seismic load, 23 the number of cycles 4n. the other load, which is the
- 24 pressure in the static.
25-MR. SIESS: How many cycles does it take to get O
t >
v
e
, 264 m}
( 1 you into trouble?
2 MR. GUZI: Well, we are talking about big 3 earthquakes. We are trying to find when this thing fails.
4 .It's a question of how you measure margin, whether-the 5 cycles are in the loads.
6 MR. SIESS: Will two do it?
7 MR. GUZI: What's that?
8 MR. SIESS: Will two do it?
9 MR. GUZI: Two cycles? No, I think it would be to the number of cycles you would have in a large earthquake.
11 Essentially, we are modeling or we are trying to test at a 12 high enough e rthquake' level with a realistic number of
__ i3 cycles that.would go along with it, say 10 or 20, you know -
Ib
(_/ 14 - probably about ten -- to represent that earthquake, and 15 to see the failure mode at the super earthquake.
16 MR. SIESS: Are you defining a large earthquake as 17 one with a lot of cycles, or also a lot of cycles of high 18 "gs?
t 19 MR. GUZI: No. I think the thinking now is to tie 20 it to one with a big "g" level and scaling up from, say, an 21 SSE type.
22 MR. SIESS: Well, a big "g" level.doesn't 23 automatically mean a lot of. cycles.
24 MR. GUZI: Okay. Well, a design-type of 25 earthquake with a lot of, or a moderate --
- i I
\
265
'O t j j MR. SIESS: Do the seismologists say this is
%.s 2 reasonable?
3 MR. GUZI: Well, people that are involved in 4 designs such as Kennedy are supplying us our target 5 earthquake. ,
6 Maybe I'll just mention who the consultants are on j this program. Ed Rodabaugh is one of the consultants; 8
Robert Cloud, Don Landers, Robert Kennedy, and Bob Bosnaeck 9 will serve as the NRC consultants. So, we have people that jo have been involved with piping.
ij But through Kennedy we have somebody who is 12 familiar with the load definition part of it.
j3 MR. SHEWMON: But this plastic deformation, then, ja is not from operating cycles but from seismic extreme 15 cycles.
g MR. GUZI: Right. It's much bigger than the j7 design level seismic load that's placed upon the system as 18 it's in operation.
i9 MR. SIESS: A progressive deterioration on a few 20 large --
21 MR. GUZI: Right. The intent here is not to over-22 cycle something to have it fail if it's not a realistic 23 number of cycles for the earthquake. You get to show a low 24 probability but realistic -- if there is such a thing --
25 earthquake that would fail this thing and have it fail in a tJ
e 266 f~hj realistic mode under those circumstances.
( 1 2 MR. SIESS: Did anybody ever test it?
3 MR. GUZI: Nobody has tested. dynamically piping 4 systems at this level.
5 MR. RODABAUGH: Dan, I wouldn't say that. The 6 ANCO tests to me are pretty good --
7 MR. GUZI: But they didn't fail the pipe.
8 MR. SIESS: Well, that's all right. If they put 9 enough load on it and it didn't fail --
10 MR. RODABAUGH: I think that's kind of an 11 explanation for the questions you are asking. Right now, 12 the way the code is set up, what you limited your thinking 13 about is the part starts swaying and sways so much that it
( )
s_/ 14 hits something, or it goes so far that it actually kinks.
is That's the present code.
16 That means that the earthquake has no limit to 17 that displacement. Okay,~ANCO runs these tests and other-18 people, and they put on a level'of earthquake type of 19 loading but much higher than ordinary earthquake.
20 They show that the pipe really doesn't do that, it 21 just sways back and forth within narrow limits. So that 22 _means that the code has to be changed from putting a limit 23 on that guards against collapse and change it over to a 24 fatigue limit.
25 In doing'that, it will remove many of these. If (R
i
(_-
e 267 i you look through the South Texas plants, you will see a
[G*t -
2 snubber about every ten steps.
3 MR. SIESS: We are a long way from Texas.
4 MR. RO9ABAUGH: This will go a long ways towards 5 taking out some of those earthquake restraints.
6 t1R . SIESS: I understand that, but I don't get the 7
relation between the two thoughte here.
8 MR. GUZI: As far as I know, nobody has tested 9 dynamically a piping system.
10 MR.-SIESS: Why would it have to be tested 11 dynamically to get the cyclic effect? Put it under 12 pressure and reverse the loads on it, wouldn't you get the 13 same thing?
kj j4 MR. GUZI: Well, essentially the Japanese have 15 done that.
16 MR. SIESS: The dynamics really doesn't give you a 17 loading rate high enough to get into load rate effects; 18 does it?
i9 MR. RICHARDSON: Oh, no.
20 MR. SIESS: -So why does it have to be dynamic?
21 What you are concerned about is reversed modes.
22 MR. SHEWMON: He has been trying to answer your 23 last two questions.
24 MR. GdZis Everett may need to help me with this a 25 little bi.t. There are two types of testing -- well, three r~N-s
)
e e 268 I 1 types but the two types I'm talking about now is the J
2 component test and the systems test.
3 I think in a systems test to a-large extent there 4 needs to be some demonstration. We were talking about 5 inertial loads here and we were talking about, you know, 6 dynamic effects. I think we need to show a piping system 7 failing under those types of facts, and I think if you look 8 at these statically, you may not capture that.
9 I think there is a certain confidence that we need to from doing an actual test rather than trying to have a 11 pseudo dynamic test.
12 The second thing,-as far as the component level, I 13 think it's actually cheaper to test.these things with
, 14 essentially masses on these things rather than to do a lot 15 f -- I don't think'it's that expensive to do'the component i
16 tests as a dynamic test rather than a static test.
17 MR. SIESS: I find that literally unbelievable.
18 It seems to me there are two things here that I don't 19 understand at all.
20 One of them is analyzing the response of a system, 21 and the other is determining the strength of it. Am I 22 mixing -- am I not understanding what you are doing, or are i
23 you doing both?
24 MR. GUZI: I think we are doing both, but we are 25 concentrating -- at least right here -- in the strength, t
! (j
269 r
i capacity of a system.
(\
2 MR. SIESS: Okay.
3 MR. HARTZMAN: I'm Mark Hart: man from the staff, 4 Mechanical Engineering Branch.
5 The reason we want to do or think it would be 6
appropriate to do dynamic tests is that the material 7
properties on their actual dynamic loading conditions are 8
different from those on the static conditions.
9 In other words, there are effects, strain rate 10 effects, which tend to -- excuse me?
ij MR. SHEWMON: If you read the staff's piping, the 12 seismic part of the staff's piping study group, they go 13 through an analysis of that and conclude that the effect is
/^
( }f ja insignificant.
15 MR. HARTZMAN: Well, that is to be verified by 16 tests.
j7 MR. SHEWMON: That has been verified repeatedly at 18 much higher ranges for. that exact same material. It's a i9 fairly standard test to do strain rate effects. And what 20 they showed in that report'was, that since the frequency l
you have is the et-der of one a second, the strain rates l 21 l
22 that you have just are insignificant 1y different from a l
tensile test.
) 23 24 If you went up and did orders of magnitude 25 difference in these materials, it gets to be some b
v
l
. I i 270 4
'~
1 different.
2 MR. GUZI: The strain rate, as far as my way of
. 3 thinking, is not something we try to capture. I think, as 4 I said, the component tests I believe will be done 5 economically as dynamic tests. .
6 The systems tests, I think, we will try to capture i
7= phenomenon'that you can't pre-predict or predict with 8 analysis.
9 MR. SIESS: That's response used.
f l~
10 .MR. GUZI: With response in the failure range, 11 which we haven't --
4 12 MR. SIESS: You are trying to predict a non-linear i -
esponse of piping.
~ ut 14 'MR. GUZI: It only responds in failure of the-15 iping.
16 17 i
, 18 i 19 4
20 i_
22 l
j 23 24 25 J
\
.J
271
) 1 I mean a piping system is pretty complex with the
- 2 elbows behaving differently than you would analyze l 3 analytically. There are secondary effects from the 4 pressurization. If you had a collapsed mechanism, you have l 5 multiple hinges forming, which apparently won't happen.
! 6 But I think we need some tests, dynamic tests to 7 capture any phenomenon that we can't capture analytically.
8 I don't think we have that much confidence in our 9 analytical abilities right now to capture at tailure and, 10 secondly, I think it adds to our confidence to have a real
- 11 dynamic type test to fail the piping and see what happens.
12 Everett, do you have anything more to say about 13 that?
14 MR. RODABAUGH: It isn't a simple subject by any l
15 stretch of the imagination, and nothing we are talking 16 about is.
~
17 If you took an elbow and just put a force on the is end of it, the static test, you increase the force and 19 eventually the. elbow will just fold over and collapse.
20 One point of.what I am doing in that test is I 21 have no limit on the amount of energy I am inputting into 22 the piping assembly test. I am just putting in whatever 23 energy it takes to collapse it in the static test.
24 MR. SIESS: Use a control strain rather than 25 control load. That is the way the Japanese tested full-O
1 272
() i scale buildings.
MR. RODABAUGH: Here is another part of our 2
3 problem. In nuclear power plants we are talking about 4
4 evaluating something when you talk about all the systems of 5 100 to 200 piping systems. That means each system might i
~
6 have 15 or 20 elbows in it, reducers, valves bodies sitting 7 in it and different supports. So we are talking about a a great deal of analytical effort.
9 So far we have to just physically get that work io done, we have always used an elastic analysis. Now there it is the flaw in what we are doing. We are using.an elastic 12 analysis to predict behaviors. Where we are getting non-i3 linear behavior, plus an earthquake, we are getting
( i4 behaviors. See, when we do a response spectra analysis, we is don't really talk about what is the total energy that the 16 earthquake can put in the piping system. We are missing i7 that point, and we are missing the energy content point and is we are missing any non-linear effects.
19 So we have got this thing that we have got lots of I 20 analyses to do. We have been sticking for that reason to 21 an elastic analysis and generally a response spectra l
i 22 analysis. So we are missing some very major aspects of 23 what an' earthquake really does to piping system. Now this 24 shows up very clear when you run a test. We don't get 25 anywheres near as bad a result as our elastic response
- O i
i i
i .. . . - - . _ - - . _
l 273 f) i spectrum analysis indicated.
I 2 And so a simple way or what seems like the best 3 way to do that is run a few tests and correct our analysis
- 4 method somewhat empirically.
i 5 MR. SIESS: That is nice, but you just said there 6 are a lot of very Complex systems, and I don't think you 7 are going to be able to correct all of those systems 8 empirically with a few tests.
' 9 MR. GUZY: Well, the testing is at three levels,
. io as I said, and hopefully we can tie these three levels and 11 get back to defining and being able to predict the basic 4
12 phenomenon of failure.
l 33 The system test is sort of an overall proof and to
! 14 make sure that we haven't missed anything. But the i
is component tests, where we have 40 component tests planned i 16 of different types of components, T's. and elbows and 17 pressures and whatever, the intent is to try to reset the i
- 18 spectra of variations in these components and get some hard 19 data on how they fail.
20 But then from there we have these specimen tests 21 to try to capture the basic phenomena of fatigue 22 ratcheting, if that is a failure mode, and we can tie ---
23 MR. SIESS: What will be tested?
l l 24 MR. GUZY: In the fatigue ratcheting test?
25 MR. SIESS: In the first one I gathered you were C
= 1 274
() i going to have multiple lengths of piping and you were going 2
to shake them, and in the second you were going to have
)
3 shorter pieces of pipe elbow and you were going to shake 4 them, and what is the third one now? l 5 MR. GUZY: The fatigue ratcheting specimens, which 6 it will be a ---
7 MR. SIESS: What are you going to test?
g MR. GUZY: There will be specimens under a static 9 load.
io MR. SIESS: I didn't say how you were going to 11 test. You are talking about a specimen. That is just a 12 word.
33 MR. GUZY: You will test under ---
i4 MR. SIESS: What does the specimen look like? Is i
- i3 it a .505 inch diameter that you test?
16 MR. GUZY: I am not sure exactly ---
17 MR. RODABAUGH: It is a notched bar specimen.
is MR. GUZY: I think there is more than one specimen i9 they are considering now.
20 MR. SIESS: It is a coupon test.
21 MR. GUZY: If you turn a few pages forward, the 22 ~first picture you have, this essentially is the ETEK piping 23 system, which is a six-inch piping system with one smnller 24 branch line with a valve on it.
25 This is essentially what we are testing under that i
a --
m u A 275 i demonstration test, but it should also be somewhat
)
2 representative of what the final configuration is in the 3 systems tests which have not been awarded yet. This should 4 be similar to the ANCO tests which have been run before.
5 MR. SIESS: Now can you predict the deformations 6 of the various points on this piping system in the elastic 7 range?
8 MR. GUZY: In the elastic range we could predict 9 it.
10 MR. SIESS: With what kind of accuracy?
11 MR. GUZY: I think on the tests we had before if 12 we were within 50 percent we felt good.
13 MR. SIESS: Did you get-the right mode shapes?
14 MR. RICHARDSON: The frequencies of mode shapes we 15 do a pretty good job. Where we seem to miss is the 16 amplitude, the response amplitude, and that seems to be i
17 depending upon things like gaps and supports, the geometric is non-linearities.
l l n The next slide shows a representative ---
20 MR. SIESS: Are you going to come up with a 21 dynamic inelastic analysis that will do as well or better?
22 MR. RICHARDSON: I would be very happy if it would 23 do as well.
i 24 MR. GUZY: I guess the point of this program is 25 not so much to develop analytic prediction techniques as it 0
\
q
- 276 i
is to identify failure modes that occur and change the c
-accordingly.
In other words, the primary objective is not 2 l tical to develop a better computer code to analysis ana y 3
4 response. It is to ---
Now taking this piping system right MR. SIESS:
5 here, what would be the failure mode that wouldf be analysis 6
7 predicted by the first curve, criteria, efforts o a
and design?
The present criteria, the code criteria MR. GUZY:
9 d would be you would look at the high stress point an high 10 h n
essentially limit your design because of whatever t e stress point, probably an elbow in this case.
12 MR. SIESS: What would be the failure mode that e 13 f l n would be predicted in the present state of the art be or b i4 your research is carried out?
You said that is what your 15 So I am asking you now what concern is, the failure mode.
16 failure mode would you be able to predict of thing thing 17 MR. GUZY: The failure mode that the code is bas is So if you were using this same philosophy on is collapse.
i9 as the code, you would predict a multiple collapse.
20 MR. SIESS:
We are using it.
21 MR. GUZY:
So you would predict sort of a 22 collapsing.
23 So now you would design this thing of MR. SIESS:
24 25 i
I O
m
277 1 auch dimensions and such supports that some portion of it 2 would not collapse where the water couldn't go through it 3 or the steam. i 4 MR. GUZY: Well, you would design it so you 5 wouldn't get that maximum stress. You don't really design 1
6 for failure. You design for maximum stress and keep that 7 below a certain level.
8 MR. SIESS: Stress never hurt anybody until ---
9 MR. SHEWMON: The code still works with the 10 stress.
11 MR. SIESS: I know, but stress is a means of 12 controlling something. What I am trying to explore'is you i3 said to prove knowledge of failure modes. So the first
() 14 design is postulated o collapse and some stress.
15 ~ MR. GUZY: Some. maximum stress, maximum stress i 16 calculated primarily above linear calculations.
, 17 MR. SIESS: And you think the research would show is that it doesn't fail that way or it doesn't-fail by
} 19 collapsing at that stress?
20 MR. GUZY: It doesn't fail that way. Primarily it 21 is to show a cyclic type of failure.
22 MR. RICHARDSON: It results in a change in failure l
23 level, but by showing that it doesn't fail by collapse, but 24 rather it fails by fatigue ratcheting, you will demonstrate 25 that the ---
i ~S I
r
- - - + - e 4--- -ur-g'v--- w +- 9 irt--*p-ee m -aa- --
---t- g-a-**-- -v-rt*9w'w-+-ew--en --"'-"P-9-c?'----- -T- ~ * - - - - -
278
() i 2
MR. SIESS: But I would rather have it collapse and then I wouldn't have a pipe break.
i 3 MR. RICHARDSON: But that will occur at a much 4 higher level than what what we would predict as collapse 5 load.
6 MR. SIESS: Why is taa ecce so far off on 7 predicting collapse?
MR. RICHARDSON: I don't think the code predicts 8
9 collapse. It basis its allowable on the assumption that io you get collapse.
11 MR. SIESS: That to me says that if I exceed that 12 allowable it will collapse. That is the way I write codes.
i3 MR. RODABAUGH: The code criteria is based on f()
i4 static loads. If you look at this page after the piping is system, you see a very simple elbow test assembly, and this 16 is going to be set up ---
17 MR. SIESS: You are now into the component test, is right?
19 MR. RODABAUGH: Yes. I wanted to say how come the 20 code is so far off. This is set up for a dynamic test 21 where the base is going to go through an earthquake 22 motion.
23 The code had available a lot of static tests on 24 this elbow where you can imagine somebody just pulling on 25 the top end of this assembly, or the cable and continuing O
279 1 to pull until the elbow collapsed, and it will reach a 2 maximum load and become unstable.
3 Okay, that is the basis of the code equations, 4 static load, unlimited energy. That is not what happens in s an earthquake. There is an energy limit.
t 6 MR. SIESS: I can't argue. I agree with that and 7 that is 100 percent obvious. But now do we have to go a through something to prove that the obvious is true?
9 MR. RICHARDSON: Well, you are not going to ---
10 MR. SIESS: We certainly know that an earthquake 11 puts in a limited amount of energy, and the notion that it 12 was based on infinite amount of energy is wrong.
1
__, 13 MR. RICHARDSON: Well, certainly I think it is
- 14 recognized that these are demonstration tests to the i
15 consensus bodies that write these codes, and we have to have evidence to convince them to change the codes.
l 16 17 MR. SIESS: Will there also be enough information is to tell them what to do in place of what they do?
19 MR. EBERSOLE:
20 MR. RODABAUGHz Yes. Chet, let me go through 21 this ----
22 MR. SIESS: I think we have gone far enough. I am 23 not convinced that you have got to do it dynamically, but I 24 am not paying for it right now, and I think we have taken l 25 enough time trying to explain it. So let's go ahead.
0
J 280 i MR. GUZY: Do you want me to move on to the next 2 topic?
3 MR. SIESS: You just move to wherever you wanted 4 to be before I started asking you questions.
5 MR. MARK: Excuse me, Chet. Does it make ,an awful 6 lot of difference that this pipe is at 500 degrees 7 Fahrenheit and 2000 psi and has a seam weld in it?
8 MR. GUZY: The pressure makes a difference at 9 least in this context. The temperature is more ---
io MR. SIESS: The temperature is small, isn't it?
11 MR.'GUZY: Well, the temperature is in the 12 material properties essentially in this case, but the 13 pressure as far as this ratcheting mode is important.
14 MR. SIESS: It is well above NDT. I can tell you 15 that.
16 MR. SHEWMON: It is not brittle and the strength 17 levels don't change appreciable until you get up to is something like 500 C instead of 500 F.
! 19 MR. MARK: Right.
20 MR. SIESS: It is well above NDT.
l 21 MR. MARK: So you don't need to heat this up and 22 Pressurize it to make the test significant?
I 23 MR. GUZY: Excuse me?
24 MR. MARK: You don't have to have this hot and 25 Pressurized for the test to be significant?
- O
281 1 MR. SHEWMON: Well, not hot, but pressurized and 2 how that interacts _with the other applied stress seems to 3 be part of what they are saying.
4
~
MR. GUZY: I think we covered the first two slides l 5 pretty well. I will just maybe skip this scope and 6 accomplishments.
7 In this fiscal year we finally reached a a cooperative agreement with EPRI and the NRC, and the 9 program is essentially underway. There is some planning 10 that has to be done in mid-course as far as a systems 11 test. And, by the way, the systems ---
12 MR. SHEWMON: Can you tell me where you are i3 reading from?
14 MR. GUZY: Under scope and accomplishments, PY-85.
is MR. SIESS: It is the next page. Three bullets,
! 16 three years.
i j 17 MR. RICHARDSON: Not, it is before that.
18 MR. SHEWMON: Okay, thank you.
19 MR. RICHARDSON: I might add at this point that we 20 are getting a pretty good bang for our research dollar 21 here, and the NRC is contributing something in the order of 22 $800,000 and EPRI is putting about $2 million in it.
23 MR. GUZY: Basically the program began essentially 24 in June of this last year, and GE is the main contractor.
25 As far as the testing, it has not quite O
- . - _ _ = _ = _ . . . . - _ - .. . . -- - - . - . -
282 l
()
4 1 i begun. At ANCO the test setup has been made and there is l 2 some problems in getting the money out to them, but we 3 should begin any minute now on the first of the component
- j. 4 tests.
I 5 There will be a total of 40 tests, and the first 6 phase will include 12 tests, selected tests.
7 The ETEK demonstration test, which is a separate a program, has also been initiated as part of our agreement i
j 9 with EPRI. The testing on that has not quite begun yet.
j to They have got their shakers being I guess repaired now. In 11 their shake-down part they had some problems, but they will i 12 be, testing shortly.
i3 In '86 we will continue with more of the component
() i4 tests, and then in '86 this is when the systems tests will is begin. The systems tests I note here for the seismic will I
! 16 begin in '86, but as part of the overall program there will i
! 17 be tests to simulate mid-frequency and hydrodynamic type is loads, and there will be a separate water hammer type 19 test. So the systems tests will cover all types of dynamic '
! 20 loadings. The NRC is primarily sponsoring the seismic part 21 of all the tests.
22 MR. EBERSOLE:
May I ask a question. 'It is true, 23 isn't it, that all these tests that are done on pipes for 24 which the physical properties have been carefully measured 25 and you know exactly what you are doing, and isn't the O :
l
4 283 4
) 1 major problem the the statistics of how uniform piping is 2 that goes into the plants and how you can count on having 3 this beautiful uniform qualities that you have in this
! 4 test?
5 MR. GUZY: I think these are specimens and we will i
6 know what the properties are.
7 MR. EBERSOLE: This is a laboratory university i
! a experiment. What is in real life out there in the pipes?
i 9 MR. GUEY: I think these pip'es will represent to pipes as configured and we will just know more about them 11 than other pipes.
12 MR. EBERSOLE: Yes, but you know what I am trying
,,,, i3 to.get at, the uniformity of qualities.
(s- 14 MR. SIESS: Are you thinking of flaws, Jesse?
f 15 MR. EBERSOLE: Flaws or statistical measurements
! 16 being inadequate or batches getting mixed or, you know, the 17 whole sloppy business that we find in reactors.
18 MR. SIESS: Explain more about the normal l
19 variations and properties of this kind of pipe, thickness l 20 and. material properties.
21 MR. GU2Y: These pipes are supposedly typical of 22 PiP ing that is made. We.will know more about them because I 23 'we have looked at them.
i 24 MR. SIESS: Well, you won't with this small a i
25 sample, but has anybody ever looked -- there must be some l
284
() i tests made during manufacture, and it may be only-heat tests, right? They don't cut coupons out of pipes or cut 2
3 Pi eces out like we do on rebar?
4 MR. RODABAUGH Yes ---
5 MR. SIESS: For example, I could show you, tests on 6 hundreds of reinforcing bars and on the distribution of 7 yield strength ultimate strength and maximum strain. Do we a have any kind of information like that on supergrade 9 stainless steel piping?
io MR. RODABAUGH: If I may comment a bit on'that.
11 Yes, I have written a report which shows normal 12 distributions of nuclear piping. ,
, i3 MR. SIESS: And things like thickness and C) 14 strength?
MR. RODABAUGHs Thickness -- well, I don't think I 15 16 have seen any significant variation of material properties 37 with thickness. There is ---
is MR. SIESS: I mean statistical data variation and i9 thickness itself, or is this stuff ---
20 MR. RODABAUGH: That is pretty well controlled by 21 the specifications for the pipe. There is another
! 22 tolerance of either ten thousandths of an inch or 12.5
! 23 percent of the thickness, and there is an overweight 24 tolerance. Between those two you could probably depend on 25 saying that the pipe thickness will be within nominal plus
(.)
e 285 1 or minus 12 percent.
2 MR. EBERSOLE: WJll, you are talking about primary 3 piping, aren't you? ,
4 MR. RODABAUGH: No, all piping, all pressure 5 piping.
6 MR. EBERSOLE: If I go after the service water 7 system, which carries the lowest grade piping that we have a in ---
9 MR. RODABAUGH: Yes, the same rules.
10 MR. EBERSOLE: The same rules and I can count on 11 the statistics of brittleness or ductility and measurement?
12 MR. RODABAUGH: Well, no, wait a minute. You are 13 off on another subject.
( 14 MR. EBERSOLE: Well, I am talking about overall is quality.
16 MR. SIESS: It is a different material.
17 MR. RODABAUGH It is a different material and ---
is MR. EBERSOLE: I know, and that is why I called it 19 up.
20 MR. RODABAUGHz That is another story. The 21 ductility and the toughness is ---
22 MR. EBERSOLE: I have to have ductility if you are 23 going to do these kinds of analyses.
24 MR. BENDER: We haven't said enough about what we 25 are going.to apply these tests to, and there is enough
286
() 2 i confusion to think that maybe we ought to know.
P l anning to apply these tests to every fluid system where Are we 3 P ii P ng is used?
4 MR. SIESS: If they change the code they will. I 5 MR. BENDER: I am trying to find out what the ---
6 MR. GUZY: We are changing the design rules for, 7 or the objective is to change the design rules for all a Class 1, 2 and 3 piping, and the design rules specifically 9 will be changed in the area of how we now define allowable io st'resses for our inertial loads. ,
11 MR. BENDER: Now what is it likely to do? Right 12 now we have static loading limits on the piping and we i3 apply sometimes stress concentration factors or something
'() i4 like that to the fittings. Are you thinking about the same
- is kind of thing in this case?
16 MR. GUZY: I think the impact will be that it if n it is fatigue ratcheting that essentially we will be is looking at inertial stresses which are essentially the i, controlling stresses now in your piping design as part of 20 your cyclic stresses. So you will be either in a class one 21 some sort of fatigue type analysis and class two and three 22 will be other considerations which will be built-in 23 systems.
24 MR. BENDER: I am listening to what you are l
25 saying, but I am trying to put it in some kind of I
287
) 1 application sense. If these cyclic stresses don't go on 2 for very long in an earthquake, and we are thinking about 3 cumulative strain in this case created by stress reversals 4 I believe. Are we trying to argue that we will take so 5 much elastic strain and so much inelastic strain developed 6 by an analysis of this sort as a basis for the new code?
7 What is it we are trying to do?
8 MR. RICHARDSON: Can I try it? I think the bottom 9 line is that the present day code, the equations that you
.io must plus in treat the inertial stresses as primary il stresses, basic P over A plus minus MC over I where the M 12 has a component earthquake and it says that that must be 13 less than a certain allowable. And that is based on the
(__/)
s- 14 theory that the pipe will fail by collapse. We don't 15 believe that and we are going to use these tests to prove
~
16 that point.
17 If in fact we can prove that point, then the is inertial stresses from the earthquake would be removed from 19 the primary stress category and be put into either a 20 secondary stress category or a fatigue category so you 21 would have to check it against it fatigue life, and that in 22 effect would be changing the allowable equations and allow 23 you higher stresses to exist in the pipe, which in effect 24 would allow you then to remove supports and allow the 25 P ii P ng to be more flexible in earthquakes. I think that is
a l l
288
() i the practical application of how it is going to end up.
2 MR. BENDER: We could take forever to talk this
- 3 over. To put the thing in some kind of context, 4 particularly with respect to Mr. Ebersole's questions, one i 5 of the questions you have to ask is whether the kinds of 6
flaws that normally exist in pipe would have any influence 7 on these answers. And the answer you are giving me is no, a they won't because you have examined and have some 9 statistical information.
io I think you have to tell me whether it relates to i
.n the particular piping systems that we are talking about 12 MR. RICHARDSON: Well, I think the test specimens I
that were chosen, the fittings, were chosen to represent a i3
, i4 wide population of fittings.that are used in different size piP es and different types of pipes. So that the results l
15 16 they have would be a generic conclusion that it is not 37 peculiar .tx> a particular size or type of fitting.
18 MR. BENDER: Well, that will establish the range l
i, of fittings in your test, but what about the properties of 20 the fittings? Do they represent the ranges of flaws and 21 variations in materials that you have to consider, or do 22 you care?
23 MR. RICHARDSON: In the beginning of the program ,
i 24 it is going to represent essentially new material, unflawed 25 or at least flawed no more than the extent that the code O
. - _ _ - - - - . - - . - - . ~ - - , - - . - . ~ . - . . - . . . - _ . - - - . - - - . . . .
289
() i would allow.
- _ '2 MR. SIESS
- Will you know whether it is flawed?
3 MR. RICHARDSON: It will be examined.
4 MR. GUZY: They will examine it like they will-5 examine, you know, essentially new piping under your Class 6 1, 2 and 3 ---
7 MR. SIESS: They will certainly look at your a failures.
9 MR. BENDER: Well, I think you are half missing to the point. The real question is that it is going to be l
11 pipe out of the hopper somewhere, but is pipe out of the 12 hopper likely to cover the range of uncertainties that are 1
13 associated with the materials?
14 MR. SHEWMON: When they make a-code case they have I is to have a range of properties and they set their design 16 lines down below that and that is in a sense the answer to 17 an earlier question. They will inspect the -- I don't know is whether these things are cast or rolled and welded. Many t
19 of the fittings are cast, and there may be a seam weld in 20 these, but these things will be inspected by radiography if i
21 it is cast stainless steel and by UT and radiography if it 22 is the other ---
23 MR. BENDER: Well, let's start with one. If it is 24 cast stainless steel, it is not the piping system, Jesse, 25 that ---
e 290
() i this.
MR. EBERSOLE: Let me go back to what started 2
3 MR. SHEWMON: Many of the fittings are cast 4 stainless steel.
5 MR. EBERSOLE: I am talking about almost 20 years 6 ago starting in about '67 or '68. We found out that 7 everybody was using Virtually any kind of pipe for the a
service water systems, low pressure, no temperature, and 9 yet that was the life blood of the plant, what to do? And to I think the decision was we will use A-106B. What is the 11 range of properties of that? Well, it went all over the 12 map, and there was even talk about batching the quality of A-106B at the plant to establish groupings of good and bad,
()
13 u and here we are talking about a degree of uniformity is infinitely better than that.
16 MR. RICHARDSON: But I think we are missing a 17 basic point here. Our objective is not to address the is uncertainties in the load or the material properties 39 themselves, but a more phenomenology related question on 20 what is the failure mechanism of these pipes, and that is 21 not material property variation dependent.
22 What we.are talking about is the basic mechanism 23 for failing these pipes collapse or is it ---
24 MR. EBERSOLE: You are working on an academic 25 Problem now.
291 1 MR. RICHARDSON: No, it is not academic because it .
2 has a real impact on how the code equations are i 3 constructed. They are constructed today on the theory that 4 the pipe will fail by collapse, and we believe that is 5 wrong and that it will fail by fatigue ratcheting. And if 6 that is true, that is going to Change that basic design 7 equation.
8 MR. EBERSOLE: As I recall, we might have argued 9 that it might fail by simply breaking because a lot.of this 10 stuff was brittle.
4 11 MR. SIESS: Jim, let me ask you something. People 12 have been out looking at pipe and plants that have been i3 subjected to earthquakes and they have never found a pipe
) 14 that failed during an earthquake.
15 MR. RICHARDSON: Other than by severe corrosion or 16 differential movement.
17 MR. SIESS: Would the present code have predicted is a failure by collapse under those conditions?
19 MR. RICHARDSON: I don't think that is a relevant 20 question because the code is not a failure predictive 21 code. It is a design ---
22 MR. SIESS: Well, here is a plant that one 23 designed against an earthquake, and I could go in and now 24 apply .25 G to it and do an analysis like I would do for a 25 new plant, and would that analysis show that that pipe was O
i 5 - - -- --- + - - - - -- -
292
() i well above the code stresses allowed?
would, wouldn't it?
I would think it 2
3 MR. RODABAUGH Well, Chet, to do that, first you 4 would know about this plant in an earthquake. You would 5 know the earthquake G levels in the field around the plant.
6 MR. SIESS: That is what people have been doing.
7 MR. RODABAUGH: Okay. Now you have got to a
translate that to building motion with your soil structure 9 interaction, and then you have got to translate it up to io where the supports of the piping are. So those are the 11 steps you would have to go through.
12 MR. SIESS: I am doing a research project and I am 33 spending, you know, several hundred thousand dollars a year
{ u on it.
l is MR. RODABAUGHz It is a point that when I first 16 saw some of the published papers on what happened to piping i
17 in earthquakes, I suggested that that would be a good thing
- 18 to do, is to go through our routine procedure and see what
! 19 the code would indicate.
i 20 MR. SIESS: We have done this on buildings, and 21 that is how we have learned a lot about what we learned 22 about how buildings behave, because it seems to me that 23 I what you are saying is that the code assumes a particular 24 type of failure and you design against it.
.25 MR. RICHARDSON: I understand your question now.
i O
293
() 1 MR. SIESS: And things don't fail that way, and 2 the fact that they don't fail that way is at least partly 3 an explanation for why we don't see failure.
4 MR. RICHARDSON: In my judgment, if you were to 5 apply the assumption of collapses of failure mechanism in 6 using the equation 9 of the code, it would have predicted 7 failures in earthquakes that in fact did not happen.
8 MR. SIESS: Okay. So now you look at the code and I 9 it says the code won't predict a failure, you know, there f to is something wrong with the code. I have got a pipe out 11 here that was stressed 50 percent over the code level and 12 it didn't fail. So there is something wrong with the 1 13 code. If it is going to fail, it is going to fail a 14 different way at a higher level, and you are trying to is demonstrate that.
16 MR. RICHARDSON: Yes.
17 MR. SIESS: Now the question of flaws I think is 18 different. One question you would want to ask is would a 19 flaw have a greater deteriorating effect for the motor 20 failure you are postulating as being a real one than it
! 21 would have for a mode of failure that somebody was 22 postulating when they wrote the code? But the pipe that is 23 out there in the field, some of it is flawed and some of it 24 ain't, you know. \nd if it hadn't failed, it hadn't 25 failed.
O
294
) i Now we can come up'with a design approach that is 2 more realistic in terms of telling when something will fail 3 or won't fail. Too-often we are designing on the basis 4 that it is okay if it meets this stress level, but it is j
5 not okay if it is higher than that. But what we are really 6 trying to do is design something that won't fail under the 7 loadings we assume.
8 And to say that a design method is not predicting 9 failure is to say it is not really doing what it is 10 supposed to do. That is the question that the code keeps 1
11 asking.. What is the control on that deformation.
12 Now we think the code is wrong, but we want to be
-- 13 sure that when we change the code we are not getting into
- 14 hot water in some other way.
15 MR. SIESS: You can collapse a pipe if you take it 16 and just push it and push it until it collapses. The
' 17 question is is it possible for an earthquake to make it 18 move that far and can an earthquake put enough energy in to i
19 make it .asve, and the answer is probably ---
20 MR. BENDER: But really what you are investigating 21 here is whether you have to use that piece of the margin 22 that is there primarily for pressure loading to accommodate i
23 the seismic loading. And you want to put it in the second 1
24 stress strain category so it isn't added in the same way, 25 as I understand it.
l0
295
() 2 1 MR. SIESS:
an imposed deformation and not an imposed load.
It is a secondary stress because it is 3 MR. RICHARDSON: Yes, but the code right now 4 demands that primary load, i 5 MR. RODABAUGH: That is correct. You can collapse 6 a piece of pipe and people have done it by dynamic type 7 loadings. What is the deformation limit? You have got a 8 long piece of pipe, and what is the deformation that is 9 being controlled? That is the question that the code keeps 10 asking. What is the control on that deformation.
11 Now we think the code is wrong, but we want to be 12 sure that when we change the code we are not getting.into 13 hot water in some other way.
14 MR. SIESS: You can collapse a pipe if you'take it is and just push it and push it until it collapses. The 16 question is is it possible for an earthquake to make it 17 move that far and can an earthquake put enough energy in to 18 make it move, and the answer is probably no.
19 MR. RODABAUGH I think we have some tests that I
20 strongly suggest that the experience shows that.
1 21 MR. GUZY: Well, just to finish up on this slide, 22 in '87 essentially the testing will be finished for both 23 the component tests, the other systems tests and the 24 fatigue ratcheting tests.
i- 25 In '87 the phase will come into development of i
l O
I
296
) I recommended code changes, and the intent is to get 2 something specific to replace equation 9 with, and 3 hopefully something that is design oriented and simple that 4 you can use and still based on elastic analysis and not 5 requiring a lot of fatigue evaluation for Class 2 and 3.
i 6 MR. SIESS: And all of the tests you are talking 7 about, the dynamic tests, are to go in the inelastic range?
8 MR. GUZY: Yes, and hopefully they will go in the 9 failure range.
j 10 MR. SIESS: Yes. It would be nice if you couldn't 11 get them'there, but you have tried. Just try hard enough 12 to get them to fail, and if they don't, then they don't.
13 MR. GUZY: Maybe as an aside, when the NRC first 14 help sponsor the ANCO tests that EPRI and NRC both funded, 15 our intent there was to fail the piping and to get some 16 idea of what the failure level was. The piping wasn't 17 failed due to the level. Essentially it was a failure due j
18 to overfatigue of testing.
19 But those tests are rererenced now even though 20 they didn't fail, and they are referenced now as to the 1
21 justification for a lot of these other proposed changes we 22 are looking at. We want to demonstrate a failure 23 mechanism.
24 MR. SIESS: It is easy if you make static tests.
25 MR. GUZY: Well, it is easier, but can you O
l
297
() I convince everybody you are right.
2 Anyway, so there will be proposed code rule 3 changes and there.will be an assessment of these ---
4 MR. SIESS: I am talking about on the components 5 only, but go ahead.
6 MR. GUZY: And then also, as we discussed before, 7 there is a need to bring in the findings of the degraded a piping program, and I think this has been mentioned around 9 here. Basically we are 1 coking at piping as designed. So 10 whatever flaws are the ones that escaped your initial 11 examination.
12 There is also a look at what is realistic cracking 13 in piping and what is the dynamic resistance of these which 14 will come out of the degraded piping program. They will 15 look at the two types of program as a whole as far as a way 16 of predicting ---
17 MR. RODABAUGHz Well, doesn't that aspect of it 18 address to some extent Mr. Ebersole's concern?
19 MR. GUZY: Yes. I think the degraded piping 20 program is going to pick up -- if you know what the flaw is 21 in your piping, then you need to have some mechanisms to 22 evaluate whether that is going to withstand a load. Now i
23 the degraded piping program, as far as I know, is looking 24 more at design levels, and saying you have got a flaw this i
25 big and your margin is maybe based on a crack size rather
298
) I than the earth break level.
2 Ours is sort of looking at an uncracked pipe and ,
I 3 the margin is based on load level, but we are trying to l 4 combine them together to have some more integrated approach 5 to it.
6 MR. EBERSOLE: Is internal corrosion accounted for
! 7 in this? I recall that extraordinary problem they had at 8 Watts Bar. They were developing globules of in essence 9 rust, cancerous growths inside the pipe, and there main 10 problem was the Delta P in the pumping systems.
l 11 Nevertheless, they were chewing away at the l
12 structure of the pipe.
13 MR. SHEWMON: That was not in the primary. That
' 14 was some dump water ---
15 MR. SBERSOLE: Service water, critical service 16 water right out of the river.
17 MR. SIESS: There were complications of internal 18 erosion, too.
19 MR. EBEPSOLE: Yes, in a lot of theirs. They 20 thought the solution would be to go to stainless, Paul.
21 MR. SHEWMON: Pardon?
22 MR. EBERSOLE: They have got a lot of stainless 23 now.
l 24 MR. SHEWMON: Yes, but is pumping out of the i
25 river, and it is at low pressure. That is certainly not j
O
l 299
( 1 class one, and probably not class two, and what class three 2 is, I don't know.
3 MR. EBERSOLE: It is out there in the darkness, '
4 but nevertheless that carries the vital fluid for shutdown.
5 MR. SIESS: You didn't explain this shotgun 6 diagram.
7 MR. GUZY: That shotgun diagram was put there out a of place. That is the damping values.
9 (Laughter.)
10 MR. SIESS: I was looking at the plant 11 designations on top and wondering ---
12 MR. GUZY: I think you can decode them because you 13 are clever enough.
14 MR. SIESS: It looked like it was computer 15 variables.
16 (Laughter.)
17 MR. GUZY: As mentioned, the regulatory use is to is essentially endorse, as we routinely do I guess, the NRC's 19 endorsement of the ASME Code rules changes. The overall 20 program will provide another means for getting rid of 21 snubbers, and essentially this is EPRI's main objective is 22 to use this as a way of eliminating the number of 23 restraints you use.
24 Then the third thing will be to feed into the 75 fragilities margins validation effort I guess.
O
300
) 1 MR. BENDER: Mostly this is going to be useful in 2 the secondary steam systems and some related service ---
3 MR. GUZY: It should apply to all Class 1, 2 and 3 4 piping. The main loop has been taken care of, but as far as 5 your other lines, your surge lines ---
6 MR. BENDER: I am talking about the secondary' 7 lines, which are the places where the seismic design 8 features are becoming so horrendous that you can't-trust 9 them.
i j
10 Okay, if we are through this this, I would like to
! 11 move on to the damping and discuss that next, which is out ,
12 of order here. It will be about five viewgraphs in from 13 the pictures, and the title would be parameters influencing 14 damping in piping systems.
.j$f, MR. SIESS: I hope somebody will learn to put
! 16 numbers on their slides. -
1 l 17 MR. GUZY: I was out of' town last week and I had la to leave these off and this is the way they came out.
, 19 MR. RICHARDSON: We were working to the last day 20 to get these out and didn't have time to paginate.
21 MR. GUZY: This essentially is one program. It is 4
22 a continuation of the program we have had at INEL, and some 23 of the information that has come out has already been used 24 by the piping review committee in supporting the PVRC 25 damping values, which is now in the code as Code Case N-
-y d
. =
301
() 1 2
- 11. But some of that work is ongoing as far as the testing.
3 Under the INEL program they have essentially 4 assembled this data base. Part of that was used for the 5 PVRC. Since then they have updated the data base and they 6 show that basically the initial curves were pretty good.
7 With twice as much data, as I understand, or 100 percent a more data they show that you still have five percent at the 1
9 lower frequencies and a drop-off as you get to higher to frequencies was reasonable and seems to fit pretty well.
11 MR. SHEWMON: I understand they have a better 12 curve than what you show here.
l 13 MR. GUZY: Well, I will explain that curve when I 14 get to it.
15 MR. SIESS: 'I hope it is not what I think it is.
16 MR. GUZY: I will get back to that curve.
17 The issues here again are overly stiff piping 18 systems and trying to get rid of snubbers and pipe whip j 19 restraints and those other evils.
20 Another new issue that came out of the piping i
21 review committee was that, as recommended now, the PVRC 22 damping values in Code Case N-411 limits itself to response i 23 in the zero to 33 hertz range, essentially in the seismic 24 range.
25 The Reg. Guide used to be limited to the seismic O
L
i 302 i
) I range, but people could easily use that past and out into 3
l 2 the mid-frequency range because you essentially used one I
3 value and went out with it. It is not so obvious with the l
4 PVRC damping curves what you would use.
I 5 It was requested' that we get more guidance on or 6 data on the damping in the high frequency range, and here 7 we are talking about 33 hertz to 100 hertz, essentially I 8 something that would be representative of hydrodynamic 4 9 loads.
10 So this recommendation was being made, the other ,
11 program in INEL was going on, the experimental stage of l 12 it. So we have adjusted the program to include high 13 frequency inputs to get measures of the high frequency
} __,
i 14 response and subsequently get data for the mid-frequency 15 damping. -
1 16 MR. GIESS: Is damping that important for water 17 hammer type cases?
- 18 MR. GUZY: I am not sure about water hammer, but 19 it is for the hydrodynamic loads. It is sort of an 20 artificial thing.
j 21 MR. SIESS: What is an example of-an hydrodynamic i 22 load? >
l 23 MR. GUZY: Blow-down- loads and chugging loads and
{ 24 that kind of stuff.
25 MR. SIESS: Blowing down in the torus?
I
%s Y l
303
() 1 MR. GUZY: Right.
- 2 MR. SIESS: Cyclic loads. >
3 MR. GUZY: Basically if you look at the.PVRC' 4 curves now, I think it~goes down two percent at the 33 5 hertz level, and we don't know where it goes beyond that.
6 If you interpret the old. reg. guide, Reg. Guide 161 values, 7 you could either' Classify your hydrodynamic loads as OBE as j 8 SSE, depending on how you classify them, and OBE is one 1
i 9 percent.
10 MR. SIESS: That reg. guide was developed for 11 ' seismic, too.
l 12 MR. GUZY: But it is being used because it is the 13 only thing available.
14 I guess we can move on to the objectives.
is One is to compile and evaluate existing data.
16 INEL has collected an extensive amount of test data, in l 17 situ test data and experimental test data for piping systems. Some of this, as I mentioned, has been used j 18 19 already by the PVRC.
20 MR. SIESS: Is this all elastic now.
21 MR. GUZY: Yes, this is all elastic and it is i 22 different levels of exitation. Some of it is low level and 23 some of it is approaching the design levels.
24 Also, to get this more high-level data and to be 25 able to control some parameters, they have set up some
() l l
S 304 1 simple test systems.
~
2 In '85, other than what I naid about establishing 3 a world data base on damping, they have -- I guess last
, 4 year they completed their testing of simple straightline 5 piping systems of a. couple of different diameters.. Now 6 they testing a three-dimensional system.
7 If you flip back to your pictures again, the 8 picture before the shotgun pattern is essentially the 9 picture of the 3-D damping system that INEL is testing. As 10 you see on that, there is a No. 1 and 2, and those are l
11 support locations.
12 Now one of the things they have looked at it, 1
, 13 changing supports around to see how supports affect i 14 damping.
15 This year they have tested low-strain amplitudes 16 and different support configurations, meaning at those 17 locations having supports there or not there.
7 They have looked at pressurized and unpressurized 19 systems, and they have looked at different forcing 20 functions. They have had snap-back, hammers and side-sweep 21 and random inputs into the systems.
22 Now I guess this next step would have been to 23 insulate the piping to test that and then to take the 24 piping off and to test it at high strain loads, which will i 25 be done, but we have combined the high frequency stuff. So Os !
\
_ . _. -, _ . ~ . ., , __. _._ _
305 m
I we are going to take the system as mentioned this year for 2 the seismic and do the same kind of tests for high-3 frequency inputs. Essentially it will be low strain 4 amplitude tests at high frequencies and different support 5 configurations, et cetera, to get data on the damping 6 behavior in this high frequency range.
7 Next year on the piping system it will be
~
8 insulated, and the effect of insulation will be studied 9 both with the high-frequency and seistaic frequency inputs.
10 Then the insulation will be taken off and there will be 11 high strain seismic loads. Now that will probably be the 12 last type of loads on the system, figuring there will be i 13 either some either damage or non-linear -- well, 14 essentially some damage done to the.line.
15 MR. SIESS: What kind of insulation will be used?
16 MR. GUZY: They,are using your wrap-around I think 17 it was a ceramic type, or whatever it is.
18 MR. SIESS: That kind of stuff that we don't like 19 because it clogs sumps and ---
20 MR. GUZY: Yes. It is not metal insulation.
21 MR. SIESS: Metal vauldn't have any energy
- 22 absorbing capacity probably, right?
23 MR. GUZY: I don't think it has much for our type l 24 of piping.
25 MR. SIESS: What is the justification for making
306
) I tests using a type of insulation that we are trying to get 2 rid of?
3 MR. GUZY: Because it is representative I guess of 4 What is out there.
t 5 MR. SIESS: We don't want to encourage people.when 6 they have to replace it to replace it with this same 3
7 stuff. That is what we heard last week.
8 MR. GUZY: Also under the high-frequency stuff, 9 they will go back and retest some of these one-dimensional 10 systems, which means straightline piping and with different 11 support configurations.
12 So that will give us hopefully enough data,
___ 13 experimental data to help fill out what we know about the 14 high-frequency stuff.
15 Now if you go back to the drawing that you picked 16 out earlier, that drawing essentially represents the data 17 as it exists now for plants having damping data above 33 l 18 hertz.
I 19 MR. SIESS: It looks to me like it is above five 20 hertz.
21 MR. GUZY: Well, it depends on where you draw your 22 line. One thing you have to note here is that this t
represents only those plants that had damping in the high
~
23 24 frequency range, and there is a lot more essentially in the 25 seismic range. So this thing would probably be black if
307 ls 1 .you had all of them.
2 MR. SIESS: These are measured peak values? -
3 MR. GUZY: Yes, measured.
4 MR. SIESSt. Low excitation --- [
j 5 MR. GUZY: No. These are essentially I think i '
6 hammer tests. They go and hit piping with hammers. l 7 MR. SIESS: How can you justify anything higher a than zero? ,
i 9 MR. GUZY: Well, first of all, what you had before i 10 is you looked at responses in higher ranges, and then i
j 11 essentially fit a curve through a regression analysis. So 12 this may represent the lower spectrum of amplitudes.
- 13 MR. SIESS
- But if I am analyzing this pipe, how
! 14 do I justify ~ assuming more than zero?
i
, 15 MR. GUEY: You mean because there are some at the
! 16 zero level?
l 17 MR. SIESS: There are a lot of them at the zero 18 level.
i 19 MR. GUEY: Well, it looks like a lot at the zero f
20 level, but the intent of the new piping curves was to be l 21 median values and not to be conservative values.
f a 22 MR. SIESS: That means only half the pipes will j l i l 23 fail. l r
i 24 MR. GUEY: No, that means only half the piping
! 25 will have damping levels at design levels. .One thing that '
i lO
)
i
308 I comea in here is that there is a conservatism that hasn't 2 been quantified because we don't have the data, that if you 3 estaolish a damping level for an SSE and OBE analysis, the 4 feeling is that damping is going to go way up before it 5 fails. So even though you had a median value at SSE, you 6 still have margin there. It is matter of how accurate you
! 7 are predicting ---
] 8 MR. SIESS: If you can uncouple the OBE from the 9 SSE, it will be even better.
10 MR. GUZY: They have done that in the damping I 18 area. Essentially the new PVRC curves and the code case l 12 curves are the same for OBE and SSE.
i 13 MR. RICHARDSON: But low damping values aren't l
14 necessarily conservative. That makes your piping stiffer is and that may not be necessarily a conservative way to go. l j 16 The feeling is that we ought to be using more like a median
! 11 type damping rather than a lower bound.
i l
18 MR. GUZY: So moving along here, that was the 19 extent of the data that exists now in the world data base.
]
20 So we are going to enhance that with the testing that is
- 21 being done next year, and our intent is to wrap this up say 22 by next summer, the regulatory use would be providing 23
- 24 I
I 1
i!
I 309
() i the basis for endorsing Code Case No. 411 and eventually to 2 revise Reg. Guide. 161 and parts SVDR that reference it.
3 MR. MARK: If you can assuming a higher damping 4 value, does that mean that things are safe and designs are 5 more relaxed? ,
6 MR. RICHARDSON: What we are finding out is that t 7 more flexible pipe systems are better.
1 i
8 MR. MARK: Now does that go with a high damping 9 value or a low damping value?
10 MR. RICHARDSON: That goes with a higher damping 11 value.
l 12 MR. MARK: I don't see how you can justify
) 13 anything above zero from the scatter diagram.
I 14 MR. RICHARDSON: The point being is that you
- is should not necessarily use lower bound data for damping.
16 You ought to use more like a median value for damping 17 because lower damping isn't necessarily conservative.]
18 If I use a lower damping value, that forces me to 19 design a very stiff system to accommodate that low damping, f 20 and that requires me to put more snubbers in and intuition, J
21 judgement and experience shows that more flexible piping i
22 systems behave better in earthquakes.
4 23 MR. MARK: I sort of qualitatively appreciate 24 that. I am just looking at this diagram and am wondering 25 what support it gives for anything bigger than zero.
l
e i
310 l
) 1 MR. RICHARDSON: Well, No. 1, I would say ---
2 MR. SIESS: What is the source of~the variation in 3 this diagram?
4 MR. RICHARDSON: Well, I think there are a number 5 of things that go into that large amount of variation. As 6 Dan pointed out, a lot of that data is based on hammer i 7 tests, which are extremely low vibration amplitudes, and 8 essentially all you are getting is internal damping of the 9 material at very, very low stress levels.
10 Once you starting getting that at the higher 11 working stresses, the pipe starts banging against its 12 supports and you start absorbing energy in a lot more la mechanisms than just internal molecular friction within the 14 material.
15 MR. SHEWHON: One of the things that has always 16 offended me greatly is if you do a low cycle fatigue 17 experiment, the damping is easily 100 percent for a lot of 18 cycles.
l 19 MR. RICHARDSON: Oh, yes.
20 MR. SHEWMON: And any time you get the stress up
{
i 21 high enough and start doing plastic deformation on this 22 system, as you must before it fails, there is a lot of 23 energy absorbing capabilities that no self-respecting 24 owner will let you shake his plant up to ---
25 MR. RICHARDSON: That is exactly right. ,
l WI
. O i
- - . _ - - _ - _ _ _ _ . - _ . . - _ . - . . _ . . . _ , . . _ . , . . _ _ - , - . _ _ - _ _ _ _ _ . _ , _ _ _ . _ - -m, __ _ - , _ . _ . . _ , .
311
() 1 MR. SIESS: These are just different ,
2 configurations and supports ---
3 MR. RICHARDSON: Different configurations and 4 different supports, and some at extremely low stress-i 5 levels, and to me it is not necessarily a practical value 6 of damping at all.
7 MR. SIESS: Now if I just took a piece of pipe and a sat it up here on two supports and hit it with something, I 9 would get fairly consistent values at some level.
i 10 MR. RICHARDSON: Yes. There would be some 11 variation, but not nearly like the variation you are seeing 12 here.
13 MR. SIESS: And that would be variations in
, 14 dimensions of the pipe and wall thickness and probably that is is all. Where would that be expected to fall?
16 MR. RICHARDSON: That would be very low. That is t 17 were you get your very low damping values.
18 MR. SIESS: So we are getting damping here due to 19 restraint conditions, support conditions, added masses, et 20 cetera, et cetera.
i
! 21 MR. RICHARDSON: And the magnitude of the stress.
22 Once you start into the inelastic range you are going to 23 get ---
24 MR. SIESS: Some of these must have gotten 25 inelastic or something to get up to 20 percent.
- O
312
( 1 MR. GUZY: The picture you have, these are all low-2 level tests. I showed it just to show what data we have in 3 the high-frequency range. But the other tests have been ---
4 MR. SIESS: The 20 percent on there.
i 5 MR. RICHARDSON: Well, probably what is giving you i
6 your 20 percent are more geometic non-linearities and 7 things like snubbers that might be attached to the piping.
8 MR. SIESS: So all they did was go out and bang on 9 pipe and measure decay?
10 MR. RICHARDSON: Yes.
11 MR. GUZY: That is another parameter, how you i
12 measure the damping, log decrement or ---
l 13 MR. RICHARDSON: Usually you measure the log 14 decrement of succeeding amplitudes.
15 MR. SIESS: I don't know how else you can do it
. 16 with a hammer. You certainly can't pick off the points ---
17 MR. RICHARDSON: Oh, the half power point?
! 18 MR. SICSS: The half power point.
19 MR. RICHARDSON: No.
20 MR. GUZY: My point for showing this graph is this 21 is what we have now that has been recorded in other tests.
22 This will essentially be enhanced ---
23 MR. MARK: I question the usefulness of this graph 24 in trying to make the point which I believe you made 25 verbally without looking at the graph.
7 m 4
g-h
l 4
313
( 1 MR. SIESS: I thought this graph related to some 2 other tests; is that right? -
3 MR. GUZY: This test has already been made, not 4 the test that the INEL is doing now.
5 MR. SIESS: This is part of the discussion we just 6 finished cut the parameters influencing them?
7 MR. GUZY: Yes.. Part of that program is look at 8 the existing data, and this is the existing data that was 9 Collected by them but not produced by INEL.
10 MR. SHEWMON: The PVRC report did come up with 11 this frequency dependent increased damping. Now you 12 referred to a code case something ---
13 MR. GUZY: Code Case N-411.
f s- 14 MR. SHEWMON: Now has that been accepted by the 15 Commission or the code people or what is the status of it?
16 MR. GUZY: Essentially it has been approved by the 4 17 ASME. The NRC has not officially endorsed it yet.
- 18 Essentially we have a Reg. Guide 184 which routinely l
19 reviews and endorses code cases. It is now currently in 20 that cycle. So that it is out for comment.
. 21 MR. SIESS: NRR has never accepted it?
22 MR. GUZY: NRR accepted it on a case-by-case 23 basis, but they haven't generically endorsed the code case.
i 24 MR. SIESS: NRR has accepted it on a case-by-case l 25 basis.
O
314
) 1 MR. GUZY: And I am certain that there is high 2 confidence that they will endorse it generically with ,
3 certain caveats, such as being limited to the low-frequency 4 design and for spectral analysis and not time history 5 analysis. .
6 MR. SIESS: Why the latter?
7 MR. GUZY: What is that?
8 MR. SIESS: Why.
9 MR. GUZY: Why not time history analysis?
10 MR. SIESS: Yes.
Il MR. GUZY: From my perspective, part of the 12 acceptance of the code case damping values is based on 13 looking at spectral analysis in comparison with time 14 history analysis. And it is showing that even though this is is median centered damping curves, there is some 16 conservatism in the design process if you use spectral 17 analysis.
18 MR. SIESS: You are offsetting what might be not 19 quite conservative with the conservatism from a completely 20 different source.
21 MR. GUZY: I will get into this. Nobody really 22 looked in detail the analysis of what the effect of this 23 would be on the time. history analysis. It has left a void 24 in -- -
25 MR. SIESS: The bottom line then. Do you know O
315
) i that using the higher values with the spectral analysis 2 that you still get conservative results, and you don't know 3 that you would get conservative results with time history, 4 but they might be correct results.
5 MR. GUZY: Correct.
6 MR. SIESS: They wouldn't necessarily be 7 incorrect, but they just wouldn't have the known a conservatisms.
9 MR. GUZY: Right. And essentially NRC didn't look to at them. The studies they looked at were structural 11 analysis.
12 That moves into one of my topics in the next title
___ 13 if we are through with damping now I guess.
%. 14 Go back to that drawing we beat to death, and the is next page after that is something I called other NRC piping 16 research.
17 And in here I have lumped a lot of different is items, but they have one commonality, I guess, which is 19 that they are analysis oriented. The other two groups had 20 some testing in it. Plus a' lot of these were of a lower 21 Priority, and Volume V of 1061 they prioritized research.
22 Most of these fall in lower priority items, except for the 23 nozzle design, which was one of the higher priorities.
24 The ones that I have lumped together here are 25 nozzle design guidance. This is an ongoing part of our i l
O
316 ;
() i technical systems program at Oak Ridge.
The intent here is to look at nozzle flexibility and nozzle design rules and l
2 3 come up with recommended code changes to nozzle designs for 4 both nozzles and for branch connections.
5 In reality I think the nozzles that will have the f
6 9featest benefit are simply thinner walled nozzles like 7
tanks and stuff, and then also in the definition of what we a used for allowable criteria for branch connection in
? P ii P ng.
10 MR. SHEWMON: Is buckling what limits that or n fatigue or what?
12 MR. GUZY: Everett also knows about this since he 13 has been somewhat involved with it. As far as I understand i, it, it is a stress related calculation that didn't account is properly before the effect of the nozzle deformation itself i
16 on relaxing stresses in the. tanks.
17 The failure mode is the same as what you would is predict before, and in this case we are looking at static i 19 as well as dynamic type failure modes. So it is a maximum l
20 stress type. It is not buckling, but probably a Collapsed l
21 mechanism, although a lot of this is based on shell 22 theory. So I guess ---
23 MR. SIESS: Based on what?
24 MR. GUZY: Shell theory. To some extent it is an i
25 extension of what the PVRC has done in I guess it is the l
. \
317 j i containment area, but it has to do with nozzle definition 2 and has an update in WRC Bulletin 107, which is our basis 3 for a lot of the nozzle design now.
4 The next item is combinational procedures for 5 piping response spectral analysis.
6 There are a number of small objectives under this 7 program. The basis for most of them have come out of a Volume 4, and they were in relationship to the use of the 9 independent support motion method, which is the multiple 10 independent support motion method that the NRC has 11 recommended to be allowed for spectral type analysis.
12 Basically through Volume IV and through the 13 research at BNL various groupings and combinational 14 procedures were looked at in using this new method for 15 accounting for independent support motion.
16 And a method was adopted, but there were certain 17 areas identified by the piping review Committee where more la research was needed, and this is intended to take care of 19 that.
20 The next item is non-linear piping response 21 Predictions. This was another ticket item out of the 22 piping review committee, and the objective here was to look 23 at simple methods for accounting for non-linear effects and 24 essentially simply response methods.
25 And in particular looking at the inelastic 1
l l
l
- 318
() i spectral method, Kennedy's, SMA's, Campbell's, whoever it is, dynamic load factoring I guess is the method and what 2
! 3 they call it. Plus look at some simple non-linear codes, 4 and to use these and to compare against each other, and I 5 also the ETEK demonstration tests will serve as a benchmark 6 for this.
7 There will be pre-predictions made and some a
assessment will be made of how well these methods predict 9 the- response, j io Now HED shows because they have done similar work l 11 for' DOE on a one-inch line and they have had some 12 experience in doing this kind of comparison.
13 This will end -- essentially it is not funded past i4 this fiscal year, but since the ETEK tests have been
! 15 delayed, this will not be completed until after those
! 16 tests, but that is all we are going to put into that one I 17 guess.
is MR.' SHEWMON: As I recall the report, they wanted ig to get something that was simple enough that people would 20 use it instead of just opting out of the elastic.
21 MR. GUZY: That is right. They want something 22 that is essentially based on linear analysis, but a better 23 way of accounting for non-linear behavior.
! 24 The next item is actually two items that came out 25 of the piping review committee, assessment and improvement
319
() i of spectra broadening procedures. And in NUREG 1061 the 2 piping review committee recommended that at least from a 3 piping person's perspective we quantify the uncertainties '
4 involved with the plus or minus 15 percent peak broadening 5 in the reg. guide's specification for input spectra, and if 6 possible to develop the rationale for perhaps a 7 probabilistic base, lower and flatter spectre. So you a won't be so peak sensitive I guess in your piping design.
9 I guess the scope of this more or less is to look to at the state of the art now, including the CAT-I program
) 11 and what data we have out of that program so far, plus some
! 12 work that has been done by the ASCE, the SSMRP and its I
__ i3 precursor I guess, and to look at the state of the art as
- 14 far as what does plus or minus 15 percent give you in terms i
! 15 of the building input to the piping.
16 The piping review committee also specifically 17 asked us to look at the uncertainty in the-piping response is itself to see how the. combination of the two, the l
19 uncertainty in the building spectra and the uncertainty in l
20 the piping response meshed together and to make some sort 21 of assessment of the appropriateness of plus or minus 15 l
22 percent. That is the phase one we funded this' year. If l
l 23 succe'ssful, and if a good plan of attack is worked out as 24 far as the development of a criteria for' broadening the 25 spectra, then we will go ahead with that next year. This l
0
320
() i is not a big effort, but the goal is to find some probabilistic means of perhaps getting a much flatter 2
3 spectra, and in the extreme it could be a perfectly flat 4 spectra, which means you can get away with basically a 5 simpler way of analyzing piping, which a lot of people have 6 been asking for. So this is our research type program, and 7 it is a low-level thing.
a The next item is time history pipe damping, and it 9 was mentioned before that right now we don't have any' io criteria. We have the code case, except for the reg.
11 guide. The code case limits itself to spectral analysis, 12 and in this program it is thought that somehow we have to 33 account for the conservatisms that would be in a design 14 type of time history analysis, whatever that is.
j3 Again, this may be looking at the conservatism in 16 the input, and possibly could be looking at the i7 conservatisms going from the damping in SSE levels to the is highet levels. But I think right now our approach is that i9 we are asking for proposals on this and discussing it. I 20 think it is to look at more or less, and somehow 21 quantifying and showing to the piping engineers that you 22 have got conservatisms in your input spectra.
23 The last item is something called combined 24 effects. I guess the piping review committee 25 recommendations were more or less one of a kind, and one-by-O
- ,_ c--- .-_ . . -
l 321 l
() 1 one recommendations. Somewhere along the line if you find one or more things that are significant, the combined 2
3 effects have to be looked at.
4 In one case we are doing research on this now that 5 has turned out to be important, at least in terms of the 6 licensing people, and that was the use of the PVRC damping 7 in combination with the independent support motion method, a essentially Brookhaven's work with the PVRC.
9 The initial at Brookhaven was based on uniform to damping and time histories compared to the spectral 11 analysis, and I guess there was some fear that you are 12 taking away too many conservatisms with this median based i3 damping, and we were asked to look at what happens when you 14 combine the PVRC damping in combination with these 15 recommended grouping procedures.
16 That program is ongoing now, but the initial 17 results indicate that perhaps there is a loss in is conservatism and there may be some concern about endorsing 19 these things to be used in combination.
20 MR. SIESS: Why don't we just take out all the 21 conservatisms, get a real best estimate and then put in a 22 factor of safety and then we know what we have got. Why 23 are people so uncomfortable with that?
24 MR. GUZY: I can't argue with your approach.
25 MR. RICHARDSON: Because we have never done it.
C
322
() i MR. SIESS: The NRC has never done it, but there are a heck of a lot of other people out there that have.
2 3 Nobody ever thought about earthquakes until we started 4 building nuclear plants. If we had spent the money we are 5 spending now on piping research 20 years ago or 10 years 6 ago, we would have saved.how many times that?
7 MR. EBERSOLE: That is an old proposition.
i 8 MR. SIESS: How much are we spending on piping 9 research over a three or four-year period, $10 million?
10 MR. GUZY: Not quite that much.
11 MR. SHEWMON: Too soon old and too late wise.
MR. SIESS: Well, at least we are getting wise is f 12 i3 some comfort, but it is like, I will tell you.
14 MR. GUZY: I think I discussed most of the issues i
j is in the scope and objectives we went through.
1
- 16 Do you have any questions on what is coming out of i
17 these things we have lumped as other piping research?
is -(No response.) ;
j j9 MR. SIESS: I don't, and I don't see any around 20 the table.
- 21 Well, one question. In talking about the peak l
22 broadening, I thought we were getting away from peak
. 23 broadening to peak shifting.
I MR. GUZY: Well, we shift over a broadened peak.
! 24 25 So essentially we have to know what the broadened peak is i
'O
1 .
1 323
) i we are shifting over, if you understand what I am saying.
2 MR. SIESS: Now are you going to be able to tell 3 us anything about EPRI?
4 MR. GUZY: I know what they are doing, and I feel l
5 reluctant to represent them, but I can ---
6 MR. SIESS: Tell us what you know.
7 MR. GUZY: Okay. Well, the major part-of EPRI's 4
a piping research is the joint program I discussed before 9 this. They are contributing a lot more than we are to this 10 program. They have a few other research items that I am
' 11 aware of, and maybe I.can just go through those.
12 One of them that is being conducted I guess by the 13 EQE, Peter Yanoff and Company, is essentially doing to 14 piping what they did to SQUG, as far as I understand. From 15 what I understand, it is a more detailed, or at least more f '
16 expensive review of earthquake experience in piping, and.
17 this would go beyond essentially what the NRC sponsored !
- 18 with John Stevenson's report, which is the addendum to 19 Volume II of NUREG 1061 to give a little more systematic 20 SQUG type treatment of piping experience.
I 21 And I believe they are going to include all the 22 earthquakes that happened in SQAG and all the plants, plus
- 23 the Chilean earthquake.
24 So that is one item that is going on.
! 25 MR. SIESS: Did you say they are going to add i
324
() i Mexico City?
I am not sure if they are going to add 2 MR. GUZY:
3 it. They were going to add the Chilean on the last time.
4 EPRI is sponsoring a couple of I guess design 5 approach projects in the supports area. One that,I guess 6 we have been involved in the past, and apparently EPRI has 7
been involved and they becoming involved with again is the a use of energy absorbing restraints. They are simple plate 9 type restraints that go with this. The supports themselves go non-linear and add damping to the system, but they are
~
to 11 treated in a quasi-linear approach.
12 Essentially in the design process you have i3 something in the system that accounts for a lot of i4 damping. It creates a lot of damping. It creates a lot of is damping and it is a fairly simply type of support.
16 There is a code case now that addresses these 17 types of supports, and my understanding is that EPRI is' 18 doing more ---
19 MR. SHE4 MON: How many cycles is this good for?
20 MR. GUZY:- Apparently it is supposed to be good 21 enough for at least an earthquake and more.
22 MR. RICHARDSON: Apparently one of the properties 23 of this is that as long as you can get through one 24 earthquake and its associated aftershocks, they can go
- 25 through and measure the change in the magnetic properties
- O l
, - - - - . , .,,---.--,_---,....,,,,-,~---.-.~n-. . - - - , , _ - - , - . - - _ _ . . . . . . - - - - - - . - - - , . - . - . . - -
' 325
/\
of this material and determine its residual fatigue life
() 1 2 a,pd replace it if necessary.
3 MR. SHEWMON: This is the stainless steel then 4 that transforms the deformation?
5 MR. RICHARDSON: Yes. It was the program that was 6 started with the University of California at Berkeley and 7 Spence Bush and I was involved in it.
8 MR. SHEWMON: Now I have also seen things where 9 people just had a six-inch section of pipe and collapse
, 10 that. Now that is a different ---
11 MR. RICHARDSON: That would be more for impact 12 energy absorption like a pipe whip restraint or ---
__ , 13 MR. SHEWMON: And this is some unit that will
/
(,)T 14 actually collapse ---
15 MR. RICHARDSON: Historetic damping.
16 MR. GUZY: I am not so sure who is doing this, but 17 I think it is U.C. Berkeley, and Bechtel is essentially is marketing this. I don't know exactly what the relationship 19 is, but I understand that EPRI is sponsoring more work in 20 this area.
21 Another I guess similar type of support 22 configuration research is being done by Bob Cloud 23 Associates, and this is based on I guess sort of a box 24 around maybe an industrial -- what you mentioned, a frame 25 around a piping system with a fairly large gap so you don't
,. - , . ~ - - -_ - . , , - _ - , . . - , - - _ _ - _ _ . - - -
- . . - . _ - _ = - - - - . . .. .-- .-- -- - - . . - _ . . _ _
326
.l i worry about tolerances and it bumps around and absorbs 2 energy. I think it was published in Mechanical Engineering l 3 Magazine sometime this summer. It is a concept which I 4 understand is using some industrial plants, and it is a 5 fairly simple design and the impact is supposed to absorb
- 6 energy.
7 MR. SIESS: That thing at Berkeley, is that the 8 one that Spence Bush was spearheading? l
, 9 MR. RICHARDSON: Yes. I am sure the concept has (
1 ,
I io been refined..
l 11 MR. SIESS: Some of Freeliano's pictures showed ;
l
'I 12 showed angles and stuff around pipe and nice bends in them
, 13 and it absorbed a fair amount of energy.
14 MR. GUZY: I guess this would be like some of the ,
1 i
15 latter type of -- this is designed to essentially take 16 impacts, and it is designed to allow you some tolerances on l
I 17 your gaps.
- is MR. SIESS: You don't want it so strong that it i
i9 will bust the pipe.
l 20 MR. SHEWMON: Have we come to the end? We are 21 getting off here.
l !
22 MR. GUZY: One other research item at EPRI that I l 23 am aware of that I can think of now is that they are also
! 24 doing something similar to what we are doing in the non- "
i l 25
- linear, the simple non-linear prediction techniques. I }
l O
. 1 327 1
() i think we complement each other in that one of the methods we are looking at is a method that was developed 2
3 essentially at ETEK Rocketdyne. They are not the same 4 company, but they are right next to each other, you might s say.
6 This is one of the ones we are going to 7 benchmark. EPRI's program, as far as I understand, is a going to take this method and develop it further in terms 9 of a more usable predictive mechanism for accounting for to non-linear effects, but using it based on an elastic 11 analysis, and that has just started or is going to start.
4 12 MR. SIESS: Okay. I think that will conclude the
, 13 presentations for today, 14 We will reconvene tomorrow morning at 8:30.
15 (Whereupon, at 5:35 p.m., the subcommittee 16 recessed, to reconvene at 8:30 a.m., Tuesday, September 24, 17 1985.)
18 l
19 20 21 22 23 24 25 l /~'n
(-)
e
\, 1 CERTIFICATE OF CFFICIAL REPORTER 2
3 4
5 This is to certify that the attached proceedings 6 before the United States Nuclear Regulatory Commission in the 7 matter of' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS e
9 Name of Proceeding: Combined Meeting of the Subcommittees on Structural Engineering, Seismic Design of 10 Piping, and Metal Components 11 Docket No.
] () 12 Place- Washington, D. C.
- 13 Date
- Monday, September 23, 1985 14 f
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.
(Signature) g %
(Typed Name of Rep 6rter) Mary C? Simons 20 21 22 23 Ann Riley & Associates. Ltd,
, 0 24 l 25 l
l
- % sh6 w a$
Y23
, wa. OE O
EXPERT PANEL ON QUANTIFICATION'0F SEISMIC MARGINS h
R. BUDNITZ (CHAIR)
P. AMIC0 A. CORNELL W. HALL .
R. KENNEDY J. REED I M. SHIN 0ZUKA LLNL CONTRIBUTORS:
~
R. MURRAY O G. CunMiNGS P. PRASSINOS J. SAVY .
CONTRACTOR SUPPORT:
J. JOHNSON (NTS/SMA)
M. RAVINDRA (NTS/SMA)
_NRC LIAISON:
J. RICHARDSON J. KNIGHT -
D. GUZY O """~"~~"~'~"~'- -
M Center strustreusse 4teinerkeley, CA 94704t416-626.st t t
,. y _
. . - , , . . - , - ~ - - _ . _ _ . . _ _ . .--
! . 2 f
- ' REGULATORY'NEED' IDENTIFIED BY THE NRC' STAFF
- l 1
O :
HOW MUCH " SEISMIC MARGIN" Is sulLT INTO THE NUCLEAR REACTOR PLANTS NOW OPERATING ?
OR, IN PLAIN ENGLIsH ....
HOW LARGE AN EARTHQUAKE CAN EACH PLANT ACTUALLY WITHSTAND ?
I n \ EN TS y p.
O i
)
)
I 1
0 -
, , - - - - - , .,,.e -_
-~-,.,,,-_,_,,,_,_p.-_,-.w--,-u - , ,- -, . _-, -, - , , , - ,,y, , --n----- , . , , _ - , , , , - , - <
1 .3 .
HOW CAN EARTHQUAKE CAPACITY-BE DETERMINED ?
( f 2A d%
D ORDERLY METHODOLOGY HAS BEEN DEVELOPED :
(1) DEFINE EARTHQUAKE SIZE PARAMETRICALLY (2) DETERMINE CAPACITY OR FRAGILITY FOR EACH COMPONENT (3) DETERMINE LOCAL MOTION AND RESPONSE (4) PuT (2) AND (3) TOGETHER:
DETERMINE WHICH COMPONENTS FAIL VS. E0 SIZE (5) ANALYZE SYSTEM DEPENDENCIES SUPPORT SYSTEM DEPENDENCIES PHYSICAL PROXIMITY PROBLEMS ELECTRICAL, MECHANICAL DEPENDENCIES (6) DEVELOP FAULT TREES AND EVENT TREES FOR SEQUENCES (7) DETERMINE ACCIDENT FREQUENCIES:
CORE-MELT ACCIDENTS OFF-SITE RELEASES.(CONTAINMENT BREACH) d E ** hw A un u wm u y ws ,
% i a e el (PG9 0 .
_pttrF-a%3 x lS ew e. y trum P R &s -w \s6%c) 3. aN k* N *-
oy % w a
. 4
'O THE PANEL'S APPROACH i USE: SYSTEMS-ANALYSIS EXPERIENCE FROM PRAS USE: FRAGILITIES ANALYSIS EXPERIENCE O FROM PRAS O FROM TEST DATA O FROM ACTUAL EARTHQUAKE DATA DEVELOP:
AMETHODToREVIEWASPECIFICPLANTkGAINST A PRE-SELECTED EO ACCELERATION LEVEL.
4 O j ( THE PANEL CALLS THIS REVIEW A
" SEISMIC MARGINS REVIEW" )
I l
l l
l O
4
A. SCREENFUNCTIONS, SYSTEMS, COMPONENTS, STRUCTURES 5 f1) FAILURES IN A FEW FUNCTIONS CON MOST TO CORE-MELT FREQUENCY O THEsE ARE: REACTOR SUB-CRITICALITY LESSON FROM PRA EARLY ECES (EMFRGENCY.' COOL (P'4RS ONLY l} A INVENTORYMAINTENANC 2)
OTHER SAFETY FUNCTIONS DON T CONTRIBUT MUCH TO CORE-MELT FREQUENCY
. u.,n%f -koi 6 dwt a etw B.
SCREEN COMPONENT FRAGILITIES':
H_IGH C_0NFIDENCE OF LOW PROBABILITY O C=-
5% BEsT .
95 %
ESTIMATE 0.4 -
3 E .
u.
0.2 - .
I n Y ACCELERATION , '
Lff >
O V#" '
IHE PANEL'S MAIN CONCLUSION: 6
[~
t_-
) GIVEN TWAT:
AN EARTHQUAKE LEVEL IS CHOSEN FOR " MARGINS REVIEW THEN THE PANEL BELIEVES:
1)
IT IS FEASIBLE TO IDENTIFY THOSE (FEW) SAFETY FUNCTIONS WHOSE FAILURE CONTRIBUTES MOST TO CORE-MELT FREQUENCY (FOR PWRS), AND THE SYSTEMS AND COMPONENTS SUPPORTING THOSE FUNCTIONS.
2)
IT IS FEASIBLE TO DETERMINE THE HCLPF ACCELERATION VAL FOR EACH SUCH COMPONENT, AND TO DETERMINE IF IT IS ABOVE OR BELOW THE EQ LEVEL CHOSEN FOR REVIEW.
3)
IT IS FEASIBLE TO DETERMINE WHICH ACCIDENT SEQUENC RESULT FROM THOSE COMPONENTS WITH HCLPFS ABOVE THE EQ REVIEW LEVEL.
)
THE PANEL'S APPROACH WILL PROVIDE:
A) IF NO ACCIDENT SEQUENCES EXIST BELOW THE E0 REVIEW LEVEL, THERE WILL BE "HIGH CONFIDENCE OF LOW PROBABILITY OF FAILURE" AT THAT REVIEW LEVEL.
B)
IF SOME SEQUENCES DO CONTRIBUTE BELOW THE E0 REVIEW LEVEL THESE WILL BE IDENTIFIED, AND THEIR APPROXIMATE IMPORTANCE UNDERSTOOD.
f' \,
\ ,
I
1 7
PANEL'S PROGRESS TO DATE V !
FIRST REPORT -------- AucusT 1, 1985 o ESTABLISHES APPROACH o PROVIDES GUIDANCE ON MARGIN REVIEW PROCESS NEXT REPORT --------
OCTOBER - NOVEMBER 1985 o DETAILED GUIDANCE FOR DOING TRIAL REVIEW o
DISCUSSION OF STEPS AND IHEIR RELATIONSHIPS 7- 0 INTERACTIONS OF SYSTEMS ANALYSTS AND U FRAGILITY ANALYSTS LATER PANEL WORK --------
1986 o
IRIAL PLANT MARGIN REVIEWS (1 OR 2) ---- EARLY 1986 o
MODIFIED REVIEW Gu! DANCE ---------------- LATE 1986 h [dt tjuth JS nw O2"h - #
k' 7
O_Mo kNA
~~
6o Analyst task
, Fragility Analyst ' task i
Select a review earthquake level i
J Get infomation h
on systems. sorts Y % I A and B functions N Use infomation N Bather generic infomation on the f,"b 23 d plant (layout. drawings,ect.).
% Q Get infomation summary table 2.2from ref.1 and Confim screening Y% First plant walkdown:
[ Teams of + and
- concentrate on identification of problems without
\/ component specific. Emphasize system interaction.
9 levision of systems V
-f-
- elationships established
'n Step 3. c Q
V
/N % Second plant walkdown:
Team of
- primarily (and possible 6 + for checks). Emphasize actual 981 sic i size #8 c"erecter4st4cs
\[ of components requiring detailed analysis.
U itemine minimal cut sets
>r end points core melt, .}.
- Finalize HCLPF for all components in the final cut sets (components not sreened out).
1
O O O ,
BOLT l :
n i N j !
\ .
x i 1 TWO STEEL PLATES l d APPROX.17,800 lb EACH i NE '
Erl l 12 0 ,,
N ' s'
- 4 ,
' i 106 % y' i
\ 4 I I
imi i 90 ,- l t i .
/
4A
=
I
% / f s
mi R i- y '
I N I
I ALL FOUR INCH l '
% & s i WALLS HAVE No. 3 4
g \ W \ '
REBARS ON 4.9 INCH
\y <
,s lli '
s CENTERS '
N .
's s, N
j' 1
~~- p,3 :
s'
<' 120 s l 90 ,'
l DIMENSIONS IN INCHES 9
M Np6
O O O VALIDATION EFFORTS (FY 86 - 87)
TOPIC (S) FACILITY C011ABORATOR NRC CONTRACTOR PIPING..RE.SPONSE,. HDR KFK ANL i
STRUCTURAL RESPONSE DAMPING, VARIABILITY
, S0ll-STRUCTURE LOTUNG EPRI ANL 4
INTERACTION, STRUCTURAL RESPONSE
- TENTATIVE
! RESPONSE OF SHEAR CERL LANL l WALL STRUCTURES PIPE CAPACITY TESTS ANC0/ETEC EPRI ANC0/ETEC l
- -iw- ,- - + -w-% 4_. A . -wma.-aam . -. _ um-4 A ,mu_ __ -----e.m-4 e' 4+-w= ^ *'d -
O O O i
i i
I i MECHANICAL-STRUCTURAL ENGINEERING BRANCH j PRESENTATION TO ACRS SUBCOMMITTEE ON i
STRUCTURAL ENGINEERING SEISMIC DESIGN OF PIPING I METAL COMPONENTS SEISMIC, STRUCTURES, AND CONTAINMENT INTEGRITY RESEARCH PROGRAM l
1 l
. l SEP1 DEER 23 Alt 24,1985 e k
MSEB BUDGET FOR STRUCTURES AND SEISMIC RESEARCH BUDGET FY 1985 FY 1986 FY 1987.
TITL_E
$650K $600K $900K SEISMIC MARGINS STUDIES.
1100 1750 1900 VALIDATION OF SEISMIC RESPONSE PREDICTIONS ,
240 1100 1000 COMPONENTFRAGILITYDATAACQUISITIONAND EVALUATION 820 1200 1500
'SEISMICCdTEGORYISTRUCTURES ,
0 200 400 RELIABILITY ANALYSIS OF NONLINEAR BEHAVIOR OF CONCRETE STRUCTURES
- 0 0 300 STRUCTURAL DEGRADATION ,
l 200 300 200 STRUCTURAL COMPUTER CODE BENCHMARKS 200 700 900 NRC/EPRIPROGRAMSONPIPINGbYNAMIC LOAD CAPACITY 333 375 600 OTHERNRCPIPINbRESEARCH ,
230 200 0 i PARAMETERS INFLUENCING DAMPING IN PIPING SYSTEMS
$3873K $6325K $7700K 1,418 PROGRAMS THAT TERMINATE IN FY 1985 l $5291K
MSEB BUDGET FOR CONTAINMENT INTEGRITY RESEARCH BUDGET FY~1985 FY 1986 FY 1987_
TITLE
$2,060K $2,000K $2,400K EXPERIMENTS ON CONTAINMENT MODELS UNDER EXTREME LOADING CONDITIONS 740 600 700 CONTAINMENT INTEGRITY UNDER EXTREME LOADS 900 900 600 INTEGRITY OF CONTAINMENT PENETRATION UNDER SEVERE ACCIDENT CONDITIONS
- 250 200 0 CONTAINMENT BUCKLING $3700K
$3950K $3700K t
/
i I
= - - . _ - . - - . - . - . _ -
EFFECTS OF EARTHQUAKES ON/ TRATING PLANTS v O ISSUE _
SOME OPERATING PLANTS MAY BE REQUIRED TO WITHSTAND EARTHQUAKE IT IS IMPORTANT TO ESTIMATE AVAILABLE DESIGN BASIS BECAUSE OF NEW SEISM 0 LOGICAL INFORMATION.
l MARGINS OF PLANTS TO WITHSTAND THESE LARGER EARTHQUAKES TO PROVI 4 DECISIONS AND AVOID UNNECESSARY SHUTDOWNS.
OBJECTIVE l
TO EXPERIMENTALLY VALIDATE COMPLEX METHODS USED TO PREDICT SEISMIC BEHAVIOR AND REDUCE UNCERTAINTIES IN COMPONENT FRAGILITY DATA BASE USED TO ESTIMATE MA
~
l PROCEDURES FOR ESTIMATING SEISMIC DESIGN MARGINS; AND ELIMINATE EXCESSIVE USE O PROVIDE BETTER BALANCE OF SAFETY.
ELEMENTS '
! o CONDUCT EXPERIMENTS USING LARGE SCALE INTEGRATED FACILITIES TO METHODS.
o ASSEMBLE EXISTING COMPONENT FRAGILITY DATA AND CONDUCT FRAG BASE.
o DEVELOP AND DEMONSTRATE PROCEDURES FOR ESTIMATING SEISMIC DE FOR REDUCING THE NUMBER OF SNUBBERS USED IN SEISMIC DESIGN.
INTEGRATION PROGRAM INVOLVES COOPERATION AND COORDINATION WITH EPRI, INDUSTRY, GERMANY AND JA REGULATORY USE l
METHODS. DATA BASE AND REVIEW PROCEDURES WILL BE USED BY THE i SUBMITTALS, PROBABILISTIC RISK ANALYSES, AND ESTIMATE SEISMIC MARGINS.
4 4
4 .,u iL. - ._.a _ , - -. s % - . _ * - . - - - - - -
O w
o . W S
w
5
- E i E
\ 5 ~
E E w =
E E = $ $ u E w u
- w a -
m*i C d %
" d c5 o mE
. 5 - m E = -
m z
- 5 ? d = z um .
=
E mW w m . s - e
=
~ g a
~
w 4 S S
$0 z .
b E E E 5 S =
a 0
= g m" 5 w E m m f
s v
E E
o v
a c8 E W5 W W 5 E =
$5 m w
M E
8
=
W r
E o m
o =
M e E b 5 b h m o M
- 5 =
1 g !& -
m o
E m
m 5
w 2
=
0 m
5 5 g =
B m
5 5 - G E E G * =
o w m 5 M w
w z 1 = -
m < ~
< g ~ " C 5 E 8 E
- e b sE E
m z
=
e M o 4
- o = <
=
=
w z 2 = w < m - =
N g E
$oo " S 8w < g m 5 -
w
c=m8 M sw e-
=
=
i 3 m 5
=*== gSo ao$ = E b
5 t" 5 <- E
v o
o D
sa - w
= v 8 G -E 8 O
$ $ 5 O .
E E - a D <W = m m w W E 5 E e o g E ** 8 E s0 y p
a = = s= s *
=* 5 "= c w eo w
8
- 58 8 a o m-5 M! = m o
1 4 ' 5 5 s .m e : s - =I w R w m u - =
a g . B U = g .
p W $ 0 5 E W *
.J -" g < o" eW 5 W " o g 4
e f f 8 9 i
e r
n5 85 o w
E
$w W 4 s"5 S s52 Em0 M " =
w 8 I" E $ w o 8w W g"w 3 gE
$ S x
=E w 5 s
- o= m=m ga=a
- = -
- w a 2 w v E -
w .
w
=
E 5
O
gCmw 5 E 8
0 w
E D
- E E
5ma E IEw Ua$w. w I
m - = m m w w .
=
s -
w z - =
E J K = w - J
- =
w W - - h '
a eW
- 5d 9 m
~
- w n & M E d
~
53E$$$ l5
=
~
=
W 5 e
.sasm 5
5" b
d $l$$bE 8
5 WE m w G o =&
E o 5 mw o
W E 5 W b w 5 l
l gW G E E = 5 E E w m . se 8 .M eE i o- = E z 5
=
ae E
rw-m E "85
$ 5"s S S en==**E
= w E" - r 1 s. * *
- 2 "dI 0 -
M "e "= = c *e cw *E E a-5
- = -
m- o E u
z
-- =
o w o r -=
m o o r
o r
w
=
x w
m - = g 3 o- -
g
< w 2 = o ,
A A 4
= 5 5 5 5 9 1 Q A c
G b 8
m 2
- o
EFFECTS OF EARTHQUAKES ON OPE .J PLANTS (CONTINUED)
PROGRAM ELEMENTS:
- 1. VALIDATION OF METHODS - IN FY 1986 AND FY 1987, PARTICIPATE IN HIGH LEVEL VIBRATION EXPERIMENTS AT THE HEISSDAMPFREAKTOR (HDR) IN COOPERATION WITH THE FEDERAL REPUBLIC OF GERMANY TO PARTIALLY SEISMIC CODES. ' ALSO COOPERATE WITH EPRI IN A S0IL-STRUCTURE INTERACTION EXPERIENT BEING B ACTIVE EARTHQUAKE REGION IN TA!WAN (FY 1986-1988). PARTICIPATE IN FY 1987 IN VIBRATION EXPER COOPERATION WITH THE JAPANESE ON LARGE INTEGRATED PIPING SYSTEM AT TADOTSU FACILITY (WORL VALIDATION EFFORT WILL BE COMPLETED IN SHAKER TABLE) TO VALIDATE NON-LINEAR PIPING COMPUTER CODES.
FY 1987.
- 2. IMPROVED FRAGILITY DATA BASE - IN FY 1986 WILL COMPLETE GATHERING OF COMPDNENT FRAGILITY (
IN COOPERATION WITH EPRI AND INDUSTRY (SEISMIC QUALIFICATION USERS GROUP) ANO IDENTIFYING W FRAGILITY DATA IS NEEDED. IN FY 1987, WILL COMPLETE FRAGILITY TESTING OF SELECTED COMPONENTS. CONTINUE i IN FY 1986 AND COMPLETE IN FY 1988. EXPERIMENTS ON SMALL SCALE CONCRETE SHEAR WALL STRUCTUR THEIR BEHAVIOR UNDER SEISMIC LOADS EXCEEDING THE DESIGN BASIS TO DETERMINE THE MODE AND LEVEL OF THE COMPONENT AND STRUCTURAL FRAGILITY EFFORT WILL BE COMPLETED IN FY 1988.
! 3. SEISMIC DESIGN MARGINS - IN FY 1986, COMPLETE DEVELOPMENT OF PROCEDURES AND REVIEW GUIDELINES TO ESTIMATE THE SEISMIC MARGINS FOR PWR PLANTS AND CONDUCT A TRIAL PLANT REVIEW TO DEMONSTRATE AND IMP CEDURES. IN FY 1987, SIMILAR PROCEDURES AND PLANT REVIEW WILL BE COMPLETED FOR A BWR AND WILL COMPLETE
- THIS EFFORT. .
e INTEGRATION ,
(FRG-$10M,NRC-$500K)
- 1. COOPERATING WITH THE FRG AT HDR FACILITY TO VALIDATE SEISMIC METHODS. -
i 2. NEGOTIATING WITH. JAPANESE (MITI) TO COOPERATE ON TA00TSU SHAKER TABLE TO VALIDATE NUC (MITI-$12M,NRC-$700K) (EPRI-$1M,NRC-$200K)
- 3. COOPERATING WITH EPRI ON S0IL-STRUCTURE INTERACTION EXPERIMENTS IN TAIWAN.
- 4. COOPERATING WITH EPRI AND INDUSTRY (SEISMIC QUALIFICATION UTILITY GROUP) TO GATHER FRAGI (INDUSTRY-$50M,NRC-$500K)
(EPRI-$1M,NRC-$1.2M) i 5. COOPERATING WITH EPRI ON SEISMIC DESIGN MARGINS.
i NRC RESEARCH IS FOCUSED ON DEVELOPING THE TECHNICAL BASIS F0R LICENSING DECISIONS AND E i CRITERIA USING RESULTS OF INDUSTRY RESEARCH WHERE POSSIBLE.
REGULATORY USE l 1. VALIDATED SIMPLIFIED METHODS WILL BE USED TO EVALUATE LICENSEE SUBMITTALS. l
- 2. IMPROVED FRAGILITY DATA BASE WILL BE USED TO EVALUATE AVAILABLE SEISMIC MARGINS AND PROVIDE PROBABILISTIC RISK ASSESSMENTS.
- 3. SEISMIC DESIGN MARGIN PROCEDURES WILL BE USED TO ASSESS CAPACITY OF NUCLEAR PLANTS T EARTHQUAKES AND MAKE SOUND REGULATORY DECISIONS WITHOUT UNNECESSARY MODIFICATIONS OR
- 4. ASME CODE, REGULATORY GUIDES, AND STANDARD REVIEW PLAN M00!FIED TO PERMIT REMOVAL OF SNUBBERS TO MAK PIPE SYSTEMS MORE FLEXIBLE AND PROVIDE IMPROVED BALANCE OF OVERALL SAFETY.
i
i l l NRC SEl5MIC 5AFETT RESEARCH PROGRAM !
l _F1 REIGN ANO 00NESTIC ;
RELATED PROGAAM5 ;
i SEISMIC HAZARD { {'
j SEI5MIC MAZARD
' o EPRI and Industry Prehabilistic o Charlesten Earthquete Setsatc Hazard Program EARTH SCIENCES !
l .
! 'o Setsmetectente Program o m Swick hwe o USGS Ground Metten Studies l e mattenal Earthquake Hazard Re- e Seti Response Project ! o $145pecific %tre l ;
duction Preg. (USG1. NSF. FEMR. :
um) o Corps of E inoers Setsate Setis f ,!
Settlement as : APPLICATIONS.
- ETunat DEVELOPENT AND VALIDATION l SEI5MIC Al5K I
VALIDATION OF SEISMIC ETH005 : e setssic Safety Mergins e Setssic Desten !
I ResearchProgram(55 NIP) ) letrgins Project :
o Licensee PRAs o Gerensy - KfK Hot Progres ,
e Integrated Safety I e Valfdetten of Setssic e Lead Coetnations !
I
- for Structural ; Assessment Program o EPRI - Ta w m W im ats Calculatten Methods Project I Design Project
i
- o Japan - MIT!/NUPEC 5elsmic A ;
e Risk MethodologP lategration 4
! Emperteents ! Evaluetten Program
,Y ;
j FRASILITIES AND RESPONSE l
I FRAGILITIES AND RESPG15E ;
o EPRI PtPing Program !
2 e Strectural Response Project j HISMICMANEINS f o EPRI Egelpment Qualtftcatten Program e Component Fragt11ttes Project j e RTOL Selsetc Issues 1 !
o NaC tenant Qualification e Piping Rollahtitty Projects e pgpe Desping (R.G. 1.41)
~ s - < i.i>>
- o - 1 , Pro.a. i e -ta,.ent c.itog ,rsje. ! o A9E Code-Piptog 8
,1 l o Selselc Qualtficatten Utillty !
5tructures Group Osta :
l o Eguipuent Ous11fIcatten i
i SEISMIC ItANGINS 'l 1 : j e Structural Damping i
o EPRI 5etselc Margins Research ; (R.G. 1.61) py,,,,, * .
j e Independent Pipe Support <
o Industry Probattlistic Risk j "'"
j :
Analyses Figure 1 Project Interrelationships and Regulatory Applications ,
i i
- 1 E y E E <
W W W W W W $
5 0 t
E E $ $
$ a v
- g" E E o a
a W W -
o m m W W W k W W
- " W W W W W E W -
y E, 5, 5, 5, 5, y E, d E, "$ $ 5, 5, 8 5,
= u u WE E g":i
& E"g- -
m E 80 E J 4 .
E g"u.
O s E E f M g 5 m t
~
8 5 .
n? m
= .= = . . .
E
. E g 2 EEM E W E- 2 E E E W 6 E S m 4 S SEW 6 E S 4 5 S S E d
=
N fE 5 W g J E d.
d 8 -d
=
E J
5 a
y#g d a
~ = = g 5
s .
y d s d 6
o m a 4 4 5
4 8 8 dd 5 0 M *d dd 5d d
5 E<y a
. d 5 d i W E5
=
m h
- I =
~
$ =
E 2 ;:
l b
- & C I - E g -
2 W i
E
m m
E. -s s s e g 5 W a
5gr
=,r g
m <a =
e g g=e a
m m
n 3 8 -
. E-g.
o a s
=
5 mm m,
=sl e ssgmssa=g-Ee=m sm =s s 8! E _m-es=--=s!=E=s 8 m 3 n
d a
E eae.
e3e,,e=m=eaamam w ee egms-- -
w?ss5isas-EEE 8ee ?sE5 *gBw l ms e s
,eeewerra==.rrrrga&Ega=Reo s _ _ m a
' e 2 8 d C m
E
=
m >
I i
S =
d 5 S g
E $ $* $E
-=
)
l
" s a ws 5
2 ab E M gI a "
" S E - 5 g l E a 5 a * . - a 5 " ~
EeE E eE
" " xE 2 E
- # $ = g 5 "" w e*8 S m E $ o w e EW E O w e g o b *
- G a W - E a G e*
S 5 - g "5 gES J
? 8 agc
~
t -
u o 8 - < r o
5 w w aE o E E aE
" EW~ um= 0 8 5 b.
e
. E g 5 m b r 8
i E" 8 W S e i " 5 m E 5 S
- C W E
- W W E
4 "
g w p EO & b E
$ ! Eh E
- E N Nb E E ~N w S E d $w "
e 5 m ss e c
=* b = m
- eE Oz5 2u S
w d
=
o 5 E 8 W E 8 5G m E 5 S E
5 C
r o Es p
82 " N J
- y=E 1 G E" E 5 8 5 W w =
s 5
- 8 E E D W w 5 *2 E E o E 8 5 5 3*
E 5 s
S W
8 5 "*5 g5- k
" 5 - = a; m gg=
- - W
- 5 d k 4 o
E o m
et m
- m = w a o a w m- " 5 E W S e
8 8
E[S8 w
m m
e5 W m W W E
w ge*
m b a p 2
w
=
W *8 W
u
- * ~
u mg m b y W 8
d a E t
- W -
m w o "
E E EW W
5 W 5
$ls
,em a eg
$ $gbEE E
eag m
E W
a w EE8 Eg=
a - a w
a*m w -
ev -
5 3 s e*a 5 $G b 8 3sd i -
E g
W g
=a-s
=
"a oz m g
=ae c l = -
W8 a
e w
w c e
a w
m
!$ m a g
l E a a w l
e I
W 5 g J u h$ - -
!e : =C E
e aE E o
5
- d W
w sW E
m E "
gg
=y 5 w p W & b p R o u -
0 2 M R p 5 5 5 S 0 $
ew=e E a n-e b b "
E w = ~
- y5* 5 5 8 8 E Y' ' E @
C 5e m
g M -
esm. -
aBamo
- 3 2 I - r w
g5 E g E" g - a *e Ew
- - = -
e w 5 0 W W
- 5 gE EE E 8a W 9G w yE e
! s -
eaEW g <4 = =d aW ls 5 55 =
w aw Em"E es- g =;g e e. a nw =< ww a
a S
w 55 " W
=
eWC W
0 g5 m 5 g E =
o w e i
( E E
- g < E g a 8 %g8 W, 9
- E < 8 g g W EgW 4 =g E g. E aw 8aje e s
s- es
=-
ae=ma
- . s =e *8 m *se ts v. m mEO b " "g g~ 5 Eeg
E a g' E5eg
m
= w
- s. W s5
= < - <-
s E
m 5
w W
y w
o ggg -
a 9 m ,g = rs8 -
= E E
s egma *E B s <w ee
= c
-w s o
4 .
w <
gw w
5 W 5 W"5 E
w w E E a
- w a"W 4 w o n5 g *E " W$e8 y E
- Es am l" e-a M w
m o
- w g gm < g 1
e c
= ~ s Y - =s d *$ 5
=3 =1 5 E d d S s=w ea ggs-ees<m t
g -
em w a
.=s eBee<-n w
w W
E E
a
. JWM E W <
Q w
^ Z HOJ W m O m
- HZJ M
W J C s a > e Jww W H < J s O M O M Q3 s> H w U g O a M E
[
i W ZW OM M w
Q 3 m Z
Z O N
a Z W
Q
<<E C
%*/ EM MH U
E H H
<
- M
- w e
O M w OW ZEO (J WW w M W
J
< M C C J U O M O D mQ H
> O O W 4 HM&
- d M Z w Z O N U < < A
~
JD C M N M w J
@W g a wg Z
C w
w Z
Q w
M e W w
W N
Q MM M JMM kJ 20 - W Q E w w 8
D + 3 W CH og H W H M w M > >U H e> H e a U a H Q W W
>E Mg Z W Z Ze M UW Ze W M M w
- W 40 v ECE
<E QC O W W M E a O O Dw mm w H w MU M k>M Og H U e J W M e
Z K N w NA WW (M UM J Z M W W X M MO Q4 EH <Z W H E a W < MHM
<M Hw Q m O H W w M A
~WM M W E A J WWW
@ZH HZ A4 O
>g WP a 3
- M W M W
W Q
JM MM M
M ME W ww MW M < E W M O W aE EH Z = MQ m M W J w 8 MM C
- WW Z EW M QC 8 M WMZ Q mcZ QM Wa U OO J W C6 wM WH Q H w kmw MW O MJ J J O W w H wmd > O < W H mv ZZ >
w EMH CZ J J H Q W HW O W Z Zw < m W 3 Q E W Q OC *
>>Q w U w M W U M
- OM EMM Z
QC Mw < < O E M ZU A Z C LU JD Z H Z AO U H DH JE 4 W W O O OU
<ZJ. J <H * <C M U w< >< JAM J ImW UC wu . UZ M M
w U HQ W d4H M W ZZ KW O MH HK H & E E MJ mm M&M 3MO H mm WJ W
>< Z U Q Q W M M H *M' C HW w JW W Z U W >W HM Z% MOO Z
O EM m .J
< E O EM HEU Z WWW Ow Z Q M ZW W U M M w< Wm E EM m KZH ZD 40 4 O O D WM
< ZW4 mm O Z
O E W & W O CO ZZ EZ MED C o Z HZ MC N (ZJ JM MW - A4 W M & H <~ wWW Hmq ww MH H QJ J E HM EUQ W DZ <W 4 O
< ZW Ww Q Z<> < M O E J "O Z CHW WH CJ H
Z>
OE 3 O wW OO X U UZ
- W wW O W & W WU<
CO HZM C W H WW E JUH ZQQ Z H> E < HC AO E W W C U W * < M W D <g QA ZZW
< J WM wwwH MHW (m WO JC C HZ>Z a
M DW m MWM WM HD CUE HD W
MJ Ow MH
<Z
>D EU ON W
D W
<H MW Z HWW MQ CZ M ZOW J <w UZ D O MO <E H>EU M W W ZQ OZ> Zw JO TM w > H WZ WZ Wm WU 1 CO UOM oz Z FH OE a J
< W >D Z UH M ZQ MZ d <w WCQ4
> wZ Mw M W U HM QU 3 . N > M O AWZ M MW W O E J MDW w < WW HWM wmc H
< Z m E hw wE HW M4 -u H M H U WH>
M Q CW UJE HA E M M U (H> ZJA ZM WQ wH W o m M OA MEW k w OM WM JM JZ w M
Z w Q < MC H WW Q OH U3W &W W Q Q H W H > ZJE Wu Z A U U M U E Z (E WJZZ OMH MM < H M MQ EMwo U MDQ MM W
W W W DW O
U W
UA o
M W
b w O
w Z34 M HC
< H WWH w a WE W > w W W Z
< <MZZ D ZWE NW mW mE ZE O E M WJ HOOw O WO4 OE< > Q MO W m W EW T ZOUM W EQ kwo <@ Q E MW W Q wvW EMM >m- JW Jw WE & H NH m OEm U
< E MWJ JM J> u WM O WJE wm< ~ M< rm H M
W a A WE WOW AW MMW MM>
W W
D M A
32 MQ W 3Mw m <=
U M W W MZ WO MQ D
WO JE=U W
QWOJ WMH &Z H M Q Mwo HZ w l
c W4
>O J *Z H< Z- $Z W
D W W MW N W a -
GM ZH M D
W>A M M W m W O -Z W Z i
mW M
-Z M
w
- Eu mo N O O E HC MU MO 4 WO mE AG Z W w & Uw
-D> @ Q EO MM MO O > DM 6 mM M N>w WW W> Z m O J QZ >
- M
>MO -WH LH L WM HH w EM MMMM -
= <ZO WJ WW < D o m -
WO Q M 3 UO A M
M ZAO
- W MZ o
A E .m .O - Za QZ a J J M mH We W
Z H
J o
Z Om U O Hu H um J w w Q M OU W OED m
- W M Zm D E Z E We mo .m W .m 5 mW ..
MW J
3 U, u- Um o <W W W -W -W ~4 -> Z WC 3 w W AM EW > NW>
J Mm W m E O w w w U E 6>H
>E hw M. > > -
s W JW 24J >D-Q m
mu J m-W m
O H ->
uO u Z, Z O =
Q H MH-r rW==
=g $ ggeg wJ 4 mm a M =Z H 1
Ow ZHU Z- Zg ZZ we g W. W D m m<
g Um3 -Me M3 -
> Ue m
8 Wu W MW$B a . . Z . . . . .
= ma-M
= . .
n
< M - - N n < M Z-a - N
O O O CONTAIMENT INTEGRITY RESEARCH PROGRAM i
OTHER RES PROGRAM RELATIONSHIP _
SOURCES CONTRACTOR / SUS-CONTRACTOR l ELEENT_ ,
i EPRI, FRG COMPLEMENTARY SE/ UNITED ENGINEERS EXPERIMENTAL EVALUATION OF SNL/ CHICAGO BRIDGE & IRON ,
l OF CONTAIMENT MODELS UNDER ,
EXTREME LOADING EPRI, FRANCE, JOINT SNL CONTAIMENT INTEGRITY UNDER UK EXTREME LOADING
- EPRI COMPLEMENTARY SNL/INEL INTEGRITY OF CONTAINMENT PENETRATIONS UNDER SEVERE ACCIDENT LOADS PVRC COMPLEMENTARY LANL CONTAIMENT MARGINS TO FAILURE -
CONTAINMENT BUCKLING l - !
i 1
l I
i'
O O O ,
TITLE:
i SEISMIC MARGINS STUDIES i FIN; A 0398 i
- CONTRACTOR:
r LLNL EXPERT PANEL: R. BUDNITZ, P. AMICO, A. CORNELL, W. HALL, R. KENNEDY,
! J. REED, M. SHIN 0ZUKA
! BUDGET:
1 FY 1985 FY 1986 FY 1987 .
- $650K .$600K $1100K l t i i
O O O
~
ISSUE:
NEED TO DETERMINE THE CAPABILITY OF NUCLEAR PLANTS TO WITHSTAND EARTH-
! QUAKES LARGER THAN THE DESIGN LEVEL, OBJECTIVE:
l DEVELOP THE BASES FOR EVALUATING NUCLEAR PLANT SEISMIC MARGINS, 0 ASSESS EXISTING MARGINS INFORMATION (EXPERIENCE DATA, TESTS, PRA'S) l i
0 DEVELOP MARGINS REVIEW PROCEDURES 0 IDENTIFY INFORMATION NEEDS i
O O O SCOPE / ACCOMPLISHMENTS:
FY 1985 0 NRC SEISMIC DESIGN MARGINS PROGRAM PLAN EXPERIMENTS, O EXPERT PANEL REPORT ON "AN APPROACH TO THE QUANTIFICA
' MARGINS IN NUCLEAR POWER PLANTS" (NUREG/CR-4334).
- ASSESSMENT OF EXISTING MARGINS INFORMATION
- DEVELOPMENT OF THE HCLPF CONCEPT AND SCREENING APPROACH l' - OUTLINE OF REVIEW PROCEDURES
- 0 EXPERT PANEL REPORT ON GUIDELINES FOR SEISMIC MARGINS REVIE i
0 FY 1986 ,
- SEISMIC MARGINS REVIEW GUIDELINES REVIEWED BY NRC WORKIN AND FINALIZED AS A NUREG.
t
- TRIAL SEISMIC MARGINS REVIEWS OF A PWR.
- MULTILEVEL EARTHOUAKE " TARGETS".
- EXPERT PANEL TO MONITOR TRIAL REVIEWS.
1 i
g
4 i
O O O .
i !
4 0 FY 1987 l l
- REVIEW GUIDELINES FOR PWRS UPDATED BASED ON REVIEW EXPERIENCE AND 1
NEW RESEARCH.
i j -
BWR REVIEW GUIDELINES DEVELOPED.
l TRIAL SEISMIC MARGINS REVIEW (S) 0F A BWR. t l
l
- PROCEDURE FOR POST-0BE EVALUATION DEVELOPED. i
< s k
1 4
i i i
i t
4 -
i i
I .
1 ,
l l '
I I
l i -
- 4 O .
T
_ S O N CN G I I D S
_ NS E AE D G
TN C NA I EH YM I C FS C I I I D TE FE NS FV E EI DE E I D NC A OR AR
_ EP E TG P
D DU I F E VO SO O UT RE PC ED N BE LA E O LC I I WF I SN NN A
_ EG E NI ND I S I N L LA EE E '
DH D" I T I S
_ U UK GS GN
_ S I WE WL -
ER E -
_ I D I K VA VD EA EE R RW O "
ST . S
_ N D NC I YR I I
_ GAA GF
_ RWZ RI .
- A A ACS
_ E MEH MEE S V PR U CI C CSU
_ R SES SNC O I FI I AO T EFE ELR A SES SPP
_ L U
G
_ E O R O O
1 : 4 ;; ;!I! .I l I; ! l!l: t !
O ,
)
N A
P A
J
(
_ Y R
T S
S U
=
N D O N
- I I o T C
I D
E R
E(
T
)
GD RN FA E
7 8
9 1
K 0
0 9
P UED Y 1 Y THA F $
E R I UR S OS TRT N YTD SS O RAR NLL P ORA I RA S TOD AN E ABN HKO R C I 6 3 RAA RMT 8 K C 0 OLT S AUA 9 0 I 6 B S ERN 1 5 AL N 7 M 1 LAF O STR Y 1 S
D NO I ENE F $
I LO T RET E , A I U A ZN S NTA Z RSI 2 EG F 0 OAE I N WNR O 7 I NR A OUO 5 T U A NB G PHF N NE R C O B O CSY I VL I RR 5 T , EAA G ROT 8 K A 1 : NHN N TFS 9 0 D 5 R NKO I CNI 1 0 I 2 O OOI T ERN 1 L 2 T GOT A LEI : Y 1 A C RRA R EKM T F $
- V A A ABN E E E R P G L ; T O D T N N O U I I O C B T F C O
i: ~ ;j: i ,l '! ! , , ;!,! I : ;I:
O O O ISSUE:
ANALYTICAL METHODS THAT ARE, ESSENTIALLY, LINEAR ARE USED TO PREDICT SEISMIC RESPONSE OF STRUCTURES AND COMPONENTS. THERE IS LITTLE DATA TO INDICATE HOW WELL THESE METHODS CAN BE USED TO PREDICT BEHAVIOR NEAR FAILURE. NOR IS THERE A CLEAR PICTURE OF HOW GOOD PREDICTIONS HAVE TO BE FOR SOUND DECISIONS ABOUT PLANT SAFETY 2 OBJECTIVE:
OBTAIN INFORMATION THAT CAN BE USED TO DEVELOP ACCEPTANCE CRITERIA FOR PREDICTIONS OF THE BEHAVIOR OF NUCLEAR POWER PLANTS SUBJECTED TO LARGE
INTEGRATION:
0 THE QUESTION OF ADEQUACY OF PREDICTIVE METHODS HAS A STRONG CONNECTION WITH EFFORTS TO DETERMINE SEISMIC SAFETY' MARGINS, i
/
0 ONG0ING NRC-SPONSORED EXPERIMtNTAL PROGRAMS ON CONTAINMENTS AND OTHER CATEGORY I STRUCTURES ARE AN INTEGRAL PART OF THE VALIDATION EFFORT.
O
. U ,
I S TT I I 0 T ,
MD I I TE A M HT I S T ,H MN I T L T O WA EI .HP S DW RTS T SS O DI E S ET MT HWR E I S N T TE I E TTL
.I T RM ANE CNL PE E D I A I E ,
EE SMO MWBS R R TEM SI I E D FG SE I ASN HOA ERI ETSA TGR S OP TEE AP NPA AS G N E RI J NN OE D E SOA I G R HLGG . TP H TN O ENNE SS C AA F O NDAI G I
N OHDN MCNA XER EEX TR E RFN CH I C S FXL I EE OI ET E OO T D
. I R XC K I SN O TPNEI AST A OM AEO T HNA L I R I RS S OM PTE
_ OFTOA I R - AV BOAFL ETO LMI A M E GCF ART LNRSN R I N COC L OONI ADI OFI OI FO L E L ND CTNI E RF .
I E AI TH NPO N R FT AT I I FP OI FL EN O COUO NKO N F
- EX CT OAI ONO S PENL N I UT I O T O OAI TQA TI T N CLI C AHT ATN E SET Y PTN PA E M VAGT I RE I UM H EERNI CAM CN S S NLOI L I EE I I S I I LPI T - L TTE L FWP OC R EP RNS
_ P 5 E OXC A 6 A R M 7 A OS M 8 D LESF 8 PPI 8 PC A O 9 9 9
_ C 1 1 1 C
A Y Y Y
_ / F - - - - F - - - F - - -
E P
O C
S 0 O 0 O
l
l O O O 3
4 .
REGULATORY USE: '
1 0 ASSESSMENT 0F SEISMIC MARGINS AT EXISTING PLANTS.
0 JUDGEMENT ON THE RELIABILITY OF PROBABILISTIC RISK ANALYSES FOR j EARTHQUAKES.
2 i
i 1
.i 3
t 1
1 s -
l 6
- i
) i i
l
.l 4
1 O~ o O i
! TITLE:
i I COMPONENT FRAGILITY DATA ACQUISITION AND EVALUATION I
FIN:
i
! A 3278 ,
1 I CONTRACTOR: - .\
BROOKHAVEN NATIONAL LABORATORY (BNL) -
PROJECT MANAGER: C. H. HOFMAYER l
I BUDGET: ,
I FY 1986 FY 1987 FY 1985 l $240K $1100K $1000K ;
I l
1 i
-- -- -- --_. __ ~ ~ " '
' i ,! . r . , : ! .
O _
T E
- N V DS 8 A A EGI EI N 0 T R NNT _
D I A FOI C ET O D PTE EI M RMSJ N OS T Y OOEB TI S T FCTO S . T N I I STA O L SGEL
. TA V I I TNVN NI I L NNER T G EMSL .
OERE MHS U A R NRN(
I T OOE TSTN F
A S O I
T PFH -
MECC S T MRE .
RSI N N OER5 OSME I E CPP8 M9 FASU I N ,
GRO1 N D N OD NOC I KEN G PE I F Y SSU I , MS P RF YI EA OU UEO TREF HO RT C G ORFN I
LCT OA GN RU I I I E FD NI GDS GTFS DY I E EI A ASOU NT TB DCST RI A A I S NOAA FLSC L I Y ARBD EE XL P O
I CI G AY TBTB I G ET N NA T NAA TA EI EBMD SR TE I ZEDL NOI E EF ER ORTP RR I TI I .
PPSO ELS . M PU TAVG M OT RC I ROA EE DN RTRR OD E E OSPF CNCVI HN TN' I N AI ET TO NO ROLL C TDI I P I I I SS L WMO T PMLU TEI EI DA EI F RDWE ,
- SI MVG NC NM O S)
- LUSHR I G O TC I ATS F EF ESES PTT TN ZN E I DF AI MDHN R RT YI N NN ES L ENWAT HA N DI TA E PI AH SSE NGALN OX NT C SGTRR ARHPO P OE AI NSAE AT C W AI TTYFEEF EM RM Y ,
L DEO ED EE PS I BCEWC TA LR BA OEYED
_. AI VO AE LTTR I MEPT U I R MP E I UA LSI TL EM
.- EI LRP 5 8 I A SO VYLT ETI UR REEAN S 9 N SC DILAGFE SBEI E 1 I N A D .
E L I D OI RNN RRWCE V OA ON OOOUH I Y TA TGFI U F TT
._ : MFNNT T C
E E U J S B S 0 0 O O_
O I
l !., l ,I l . ;
l ' l
l ,
O M
S -
S T - - O E
N NU SER I A B YEYF YS G LI RA OB R A
P TT RLR! PTCE D T
RT CH W CRN EAHL O A I C O KH T CX BD MI P SNEO A SH RB E LY T I W RI I U
DI EL E
TI FI G I G MA OT I U JASP E ,N I R N LF SL T MTMDINEEP EI TAAG AEETTIAP RN RE ET OE N MR SAOPOR .
SNO I F T
ST NYMFT D T
DP NM EN E EEONHCDSE ET R AO C OT LNN E N
RN EO iP MEAGED P
OCCISOS NI OT I N l M CNINEPN TE EO LSADMO I
ASE ES A
CSCEC I TEB NNM ,A A H
LR OC E
- Q T AD AE( LCAI S CE AH BBO HN T O P CO .RNT N T EP A
TF DSE NER MMSEAOWNH E AO D S ATU L DCEECN N I GEO P LE SCI ALV A N M AI A LCTI EO C AI N C RTF AI N ,V OA PS I L I I U CLA CRESTT I T CAL
) TI E Q I AU RCES T ERUA C
U 8 DA AN I A EE OYLT N 9 EE L I
T 1
Y I
SRE OZ ENTENCHOE O L N D BEH EMI R S GT E M
O C
N A ET V OGE ESN DI I EC) ,EA R
( C UG RNT USDLARNSV 6
- 9 1
EI RT S I U RA NFCSASOOSR I F C AP E
V Y OE OYH T B C E
EADLE P S TT E I
T F P O
PHSGERGRDI OT OUUNH C
E C S
CI SWT RSAT J C B 0 O
O O O (CONTINUED) NO SCOPE:
OPERATING LOADS INCLUDED.
ALSO SEVERE ENVIRON-EFFHASIS IS ON SEISMIC LOADS, DS TREATED: BUT CONCURRENT OTHER SEVERE ENVIRONENTAL OR ACCIDENT LOAAGING, ENTAL INCLUDED.
EFFECTS EXCLUDED.NOT NECESSARILY D INUED.
TESTING OR DATA ACQUISITION (SINCE NOVEPRER 1984)
ACC0FFLISHENTS: SS, BROWN-BOVERI, GOULD, .
RCES. EVALUATION OF TEST 0 OBTAINED FRAGILITY INFORMATION F RESULTS UNDERWAY. TION EXCHANGE WITH EPRI.
O ESTABLISHED COOPERATIVE AGREEMENT FO TWO-WAY INFORMATION TRANSFER IN -
0 HOSTED THREE DAY WORKSHOP AT BNL WHERE(
AIN) TEST RESULTS WERE REVEALED.
TY-TWO PRESENTATIONS). E CLASSIFICATION FOR COM-DEVELOPED PRIORITIZATION SCHEME AN O
PONENTS.
i m w
O -
O 0-(CONTINUED)
ACC0fR.ISHMENTS: ECUTION 0 DEVELOPED STANDARDIZED COMPONENT 288, 1985, FOR APPROXIMATELY NENTS TO BE TWO FRAG OF TEST PROCEDURES BEGINS ON SE TESTED FOR BOTH STRUCTURAL AND FUNCTIONAL >
I FUTURE PRODUCTS:
I FY 1986 ATION AND
'l 0 REC 08 TENDED LIST OF COMPONENT FOR T GROUPING, COMPO-0 TEST PLANS AND TEST RESULTS G STRUC-FOR FRAG NENTS MOUNTED IN A THREE BAY MOTOR CO TURAL STIFFNESS AND VARYING ANCHORAGE.
UCLEAR 0 SUFTIARY OF CURRENT UNDERSTANDING OF POWER PLANTS.
l l
- O O O i
l l
l.
i i
I
! (CONTINUED)
I FUTUNE PRODUCTS:
Ii I FY 1987 S OF SELECTED l ,
O RECON O DATIONS FOR COMPONENT FRAGIL COMPONENTS FOR SEISMIC ECTED: MARGIN AND PRA ST f '
I l
REGIAAiuRY GUIDES OR ST* " M REVIEW MPONENT FRAGILI- PLAN '
1 NONE: HOWEVER, COULD LEAD TO THE ESTABLISHME l TY DATA BANK AT ONE OF THE NATIONAL LABO .
1 .
REGULATORY USE: ITY, PARTICULARLY I
RESPONDS TO QUESTIONS RfLATING TO EAS INTENDED TO PRO-ES IMPOSE NEW LI-l THE CHARLESTON AND NEW B 1
CENSING REQUIREENTS ON EAST COAST OPERA i
4 l
_ _ _ _ _ - - , . _ . . - _ _ m-. - , , , _ . . _ - . _ . _ , . , - . _ . , . . _ . . . _ . _ . _ _ _ _ _ _ _ _
O o ,
4 i
4 i
i i ' TITLE:
SEISMIC CATEGORY I STRUCTURES PROGRAM- ~
1 f
i j FIN:
I t
i A 7221 l
CONTRACTOR:
li LOS ALAMOS NATIONAL LABORATORY -l Et;DGET:
1 FY 1986 FY 1987 j FY 1985
$1200K $1500K
$820K j ~
1 ,
t >
O 'c) L)
ISSUE:
ABILITY OF CURRENT LIGHT WATER REACTOR PLANT EARTHOUAKES THAN THOSE CONSIDERED IN THEIR INITIAL DESIGN, OBJECTIVE:
REDUCE UNCERTAINTIES IN METHODS USED TO PREDI EARTHOUAKE LOADS ARE TRANSMITTED BY THE PL SYSTEMS AND COMPONENTS NEEDED TO OPERATE AND 0 EXPERIMENTAL DATA ON THE SENSITIVITY OF STRUCT VARIATIONS IN CONFIGURATION AND EARTHOUAKE LOAD.
0 CHANGES IN FLOOR RESPONSE SPECTRA RESULTING F QUAKE LOADS.
0 EXPERIMENTAL DATA TO VALIDATE COMPUTER PROGRA STRUCTURAL BEHAVIOR IN THE-ELASTIC AND INELASTIC R
O O O ,
_ INTEGRATION:
0 STRUCTURAL BEHAVIOR,INFORMATION-WILL BE PROVIDED T0:
- VALIDATION OF SEISMIC CALCULATIONAL METHO
- SEISMIC DESIGN MARGINS PROJECT 0 FLOOR RESPONSE SPECTRA INFORMATION WILL B
- PIPING RELIABILITY PROJECTS t
- COMPONENT FRAGILITY PROJECT.
e l
b
O O O SCOPE / ACCOMPLISHMENTS:
0 PRIOR TO FY 1985
- STATIC OR SEISMIC TESTING OF 23 DIFFERENT MODE TYPES OF STRUCTURES.
- MODELS BUILT TO TWO DIFFERENT SCALES (ONE INCH WALL THICKNESS).
- NUMBER OF FLOORS IN THE MODELS VARIED FROM ONE T -
- STIFFNESSOFTHEM0DNLWASLOWERTHANTHECOMPUTERUNC CROSS-SECTION VALUES BY A FACTOR OF 4.
- SCALEABILITY OF RESULTS ILLUSTRATED IN THE ELA RANGE, HOWEVER, ALL MODELS WERE FABRICATED USING MICR0 C I
e
- . - _ , , _ _ ,_a _ __ __,
O O O SCOPE / ACCOMPLISHMENTS (CONTINUED) 0 FY 1985
- WITH THE AID OF THE PROGRAM TECHNICAL REVIEW GROUP A MOD DESIGNED TO HELP RESOLVE THE STIFFNESS DIFFERENCE ISSUE
> AND PROVIDE CREDIBILITY FOR SCALE MODEL EXPERIMENTAL DAT THESE MODELS TEST TWO 1/4 SCALE MODELS OF THE TRG STRUCTURE.
WERE FABRICATED WITH MICR0 CONCRETE AND SIMULATED REBAR.
THIS MODEL WILL BE
- START FABRICATION OF THE FIRST TRG STRUCTURE. .
FABRICATED WITH PROTOTYPICAL CONCRETE AND REBAR.
l l
O O O SCOPE / ACCOMPLISHMENTS (CONTINUED) 0 FY 1986
- TEST TWO TRG STRUCTURES; THE FIRST ONE WILL BE SUBJECTED TO LOW LEVEL STATIC THAN INCREASED SEISMIC EXCITATION TO FAILURE (NOVEMBER 1985); THE SECOND ONE WILL BE TESTED QUASISTATICALLY TO FAILURE (MARCH 1986).
CONTINGENCY PLANS HAVE BEEN DEVELOPED FOR SCENARIOS IDENTIFIED AS LIKELY POSSIBLE OUTCOMES OF THE TRG EXPERIMENT.
PENDING RESOLUTION OF THE STIFFNESS AND SCALEABILITY ISSUES INITIATE A LIMITED NUMBER OF EXPERIMENTS TO MEET PROGRAM OBJECTIVES AND AID IN BENCHMARKING THE ANALYTICAL MODEL DEVELOPMENT. ,
- A STATISTICIAN, KNOWLEDGEABLE IN EXPERIMENT DESIGN, WILL BE USED
- TO INSURE THAT CONTROLLED AND UNCONTROLLED VARIABLES ARE INCORPORATED INTO A COST-EFFECTIVE TEST MATRIX TO MEET PROGRAM OBJECTIVES.
O O O .
SCOPE / ACCOMPLISHMENTS (CONTINUED) 1 0 FY 1987 ,
ENHANCED ANALYTICAL EFFORT TO REDUCE THE NUMBER OF TESTS AND BENCHMARK ANALYTICAL METHODS.
' SEISMIC TESTS ON 1 FLOOR MODELS WITH 1 INCH THICK WALLS (IF FY 1986 RESULTS ALLOW) TO DETERMINE THE AFFECTS OF WALL ARRANGEMENTS AND EARTHQUAKE INPUT.
SEISMIC TESTS ON 2 FLOOR MODELS WITH DIFFERENT WALL CONFIGURATIONS ON THE FIRST AND SECOND FLOORS (3 OR 4 INCH WALL THICKNESS). ,!
4 l
I i
i
- O O ,
O REGULATORY USE:
i O PROVIDE EXPERIMENTAL DATA AND ANALYTICAL METHODS TO DETERMINE THE BEHAVIOR OF REINFORCED CONCRETE BUILDINGS SUBJECTED TO EARTHQUAKES LARGER THAN THE DESIGN BASIS. ,
0 RECOMMEND IMPROVEMENTS AND REDUCE UNCERTAINTIES IN REGULATORY GUIDES (E.G., 1.61, 1.122) AND STANDARD REVIEW PLANS (E.G., SECTIONS 3.7, 3.8, 3.9 AND 3.10).
0 ESTABLISH NEW REGULATORY CRITERIA CONSIDERING~ NONLINEAR RESPONSE EFFECTS.
0 WHENCOUPLEDWITHCOMP0NENfANDPIPINGRESEARCHHELPDETERMINESEISMIC DESIGN MARGINS IN OPERATING PLANTS.
O#
e
O o O l
TITLE:
RELIABILITY ANALYSIS OF NONLINEAR BEHAVIOR OF CONCRETE STRUC ;
FIN; A 3292 i
CONTRACTOR:
BROOKHAVEN NATIONAL LABORATORY BUDGET:
FY 1985 FY 1986 FY 1987
$0 K- $200K $fiOOK '
M g - N
l O O O OBJECTIVES:
O DEVELOP STOCHASTIC, PROBABILISTIC, NONLINEAR ANALYSIS METHOD FOR MORE l RIGOROUS STRUCTURAL CAPACITY ESTIMATES FOR USE IN PROBABILISTIC RE-LIABILITY ANALYSES AND STRUCTURAL SEISMIC MARGIN STUDIES.
O DEFINE WHEN APPROPRIATE TO USE THIS NONLINEAR METHOD OR SIMPLIFIED METHOD BY SAMPLE PROBLEM COMPARISONS.
0 VERIFY THE APPR0XIMATE NONLINEAR LIMIT STATES DEVELOPED FOR CONCRETE STRUCTURES UNDER FIN A3802, " FRAGILITIES OF CONCRETE STRUCTURES".
SCOPE:
TASK 1. DEVELOP STOCHASTIC, PROBABILISTIC, NONLINEAR STRUCTURAL ANALYSIS METHODS FOR CATEGORY I STRUCTURES. DEVELOP A STOCHASTIC EQUIVALENT LINEARIZATION METHOD IDEALIZED IN TERMS OF STICK MODELS, COMPARE RESULTS OBTAINED WITH RE-SULTS FROM MONTE-CARLO SIMULATION TECHNIQUES.
TASK 2. VERIFY ANALYTICAL RESULTS BY INITIATING PSEUD 0 DYNAMIC TESTS ON STICK MODELS OF IDEALIZED STRUCTURES. REPRODUCE NONLINEAR DYNAMIC BEHAVIOR UNDER SEISMIC LOADING ON A DECELERATED TIME SCALE, AND ESTABLISH A RELATIONSHIP BETWEEN THE MAXIMUM NONLINEAR RESPONSE AND PEAK GROUND ACCELERATION,
O O O 6
i TASK 3. INITIATE EXTENSION OF THE RELIABILITY ANALYSIS METHOD TO
! INCLUDE NONLINEAR STRUCTURAL RESPONSE.
i TASK 4. COMPLETE REPORT ON STOCHASTIC EQUIVALENT LINEARIZATION
]
i.
TECHNIQUE FOR CATEGORY I STRUCTURES.
1 i
i e*
b t
t I
I
(
O O o .
i REGULATORY USE:
0 DEVELOP MORE RIG 0ROUS STRUCTURAL FRAGILITY ESTIMATES FOR USE IN PRAs AND SEISMIC MARGIN STUDIES, l 0 EVALUATION OF EXISTING STRUCTURAL CAPACITIES AND PRAs FOR OPERATING REACTORS, ;
j 0 EXTEND BNL-DEVELOPED PROBABILITY-BASED LOAD COMBINATION METHODOLOGY TO ,
t
! INCLUDE NONLINEAR BEHAVIOR, 1
i i
j .
l 1
3 l
O o o i
TITLE:
I STRUCTURAL DEGRADATION J 1 .
i FINi 1
4 TO BE ASSIGNED -
1 i
l CONTRACTOR:
A i
l
- TO BE SELECTED I
i BUDGET: ,
l ,
t i FY 1985 FY 1986 FY 1987
$0 K $0 K $300K .
l ,
i i
i-l
! l I
{
i l
,_ . , - _ ..--. -- _y -, ,, _ . . _ _ . . . -, _ - , _ _ _ _. ___
, . ? , !
O
- H -
R - C .
E R ED RU -
P E PN ES .
- D SA D .
N N NS E U N , UE H E) R W S EH EU VT w
L WS N A TA AA
. . O I EY HR E Y I R BL EE F T T E F BP I I A T Y M m
L L D A TS SE I A M I A ET R V T
- e. N B I H E)
G A G F RF m I I E O TC S L D CU C0 E E N AS N0
._ D R Y O E( O0 R C3 B I T L T S N A D A LE T -
A R E R A V N L U T O CI E F P T C I I T RU C E R MI E 0 -
F U F E ED F 5 -
O R F T HD F1 A E C A I (
O T .
- S S D D -
Y N S M , T F O N E M . RS . WO I I O I . OS EES OH a
T S TS RN TTE H N N S I E FO - AI Y E E T CS I GGT FL m
D T C AY TT NEI 0L I X E PL LI ORN A
~
E F AA UD LGI GC m
O F CN S N GL NO T R E A EO OAA I L .
E L RC T K D S E D L AT NEL NFN L N A RN SA I TA AOO B U T UE TC E T I A N TD CO SRT SST Y E CI EL NCN RNA E C M UC F - ONE EOR B A N R C FT I OR DI T m
U O TA ES TCE NTE O Q R S O A F UI N T E I E YP GFF DE
-_ D V TR R I OI DNP F A N SE AR T D EO F E UV DO SS CCT A L JE N EEF N A T A A DS OG VPO AM
'S R C A CN NY VRE U O G EI I TS DES C T L ON ST T ATO
- R C - TI A NCN - H EGT E N U T M TR I I E V R S WR AE GFM RN I E T O OO HP EI E OOS T R S P HF WO : BCC MLA C A E E P P J E O B R C O P 0 O O 0 S O 0 O
~
! : l ,!
O O O 0 EXAMINE AND DEFINE UNIQUE ENVIRONMENTAL CONDITIONS THAT EXIST POST-LOCA IN THE CONTAINMENT AND THEIR EFFECTS ON STRUCTURAL MATERIALS OF UNFORESEEN STRESS OR CORROSION EFFECTS, AT A TIME WHEN FUNCTIONAL RELIABILITY IS ESSENTIAL BUT ACCESS FOR INSPECTION AND MAINTENANCE IS NOT PRACTICAL, REGULATORY USES:
0 IMPROVED PREDICTIONS OF LONG TERM MATERIAL DETERIORATION O LIMITS ON HOSTILE OPERATING ENVIRONMENTAL EXPOSURES 0 REDUCTION OF LICENSING RELIANCE PURELY ON INSPECTION AND SURVEILLANCE 0 IMPROVEMENTS IN DAMAGE INSPECTION AND INCORPORATION INTO SRP-REFERENCED NATIONAL DESIGN STANDARDS WILL ENABLE NRC STAFF T0:
0 BETTER ACCEPT MATERIAL SELECTIONS 0 REQUIRE APPROPRIATE DESIGN PROVISIONS 0 PERMIT CONTINUED OPERATION NEAR, AT, OR BEYOND 40-YR DESIGN LIFE d
l !
O Y
- - L M E I R E S S A H A O E C L E P N
- M I
, X L P A O L N E C O W T A E E I R D S T O R O A F A N E O L P I
H I A M DS S T M O A L T S R C RA S C E E L A E I
N E G H D
, R U R D T LE Q A ,
E L L AT A B O C MA R T V RM N C I T P E G I L N E HL N E U .
TA I N I P
( R W O S M D
- N S U E I I S ET TT I T I P U NC V A L A ,P E D A T A EU A K N H SN MR R R NI NT R L O E N G A C E OS P GY R E E O I N VO N
ES H
W S
D S
S G
N I
W 0
R D
Y V
R C
OG LR A E HN E
T N E O H O N H - P T TC F S O AF NA I I L HO EP TE F E R E T MM SF I A D - L NI EI T T EE UL S A T O I
G N M CV AM Q G E O NE TR N L EL NE TG N R HI I G N N ~ D OT K G I RE I D C GS A OS VL G I E C A EO A N RD A Y H C O R F T NE C 0 I I R .S
- D TA E I EN H TN L G O T aT SP A O C RO S A K R A UD O E E LN EE PT EF N B CM R S O I AY R I E N R EA AT A O EO H SW L F RS NH ES N A F G I S CN S T M I O / TU C NO 0S AR I O .A R) LI M DN R EL S : RT KN S CS EB E NO .
TA I E SF H EI .S EC SE E C TT WV DI AT L MT UA P
M A
X 0 0 0 0 E 0 0 0 O
lJ l , , ; ,
o o
~ ~
- O 4
O LONG TMI-2 CONTAINMENT INACCESSIBILITY, W/M0ISTURE, RADIATION, BORON, H
2 BURN EFFECTS ON RELIABILITY 0 WOLSUNG 1 (KOREA) CONCRETE BASEMAT SWELLING, DUE TO ALKALINE CEMENT REACTION W/ VOLCANIC AGGREGATE 0 DEGRADATION OF SONGS-1 EQUIPMENT SUPPORTS DUE TO PLANT ENVIRONMENT 0 9 MILE PT,-1 REACTOR BLDG WALL CRACK, POSSIBLY DUE TO THERMAL GRADIENT i
l I
l i
i I
i
o o O
i J ;
TITLE: I STANDARD PROBLEMS FOR STRUCTURAL COMPUTER CODES !
FIN;
}
l A 3242 I
CONTRACTOR:
l BROOKHAVEN NATIONAL LABORATORY (BNL) l .;
BUDGET:
i FY 1985 FY 1986 FY 1987 i
$300K $200K $200K l
/
4 :
i i
l l
I ,
l i I
O O O ISSUE:
VARIOUS ANALYTICAL APPROACHES WITH DIFFERENT DEGREES OF APW:0XIMATION HAVE BEEN DEVELOPED AND USED TO DETERMINE STRUCTURAL RESPONSE AND S0ll-STRUCTURE INTERACTION OF NUCLEAR STRUCTURE'S TO SEISMIC AND DYNAM-IC LOADS, APPR0XIMATIONS USED MAY OR MAY NOT REPRESENT THE PHYSICAL BEHAVIOR OF THE STRUCTURE, THUS, THERE IS A NEED TO ESTABLISH BENCHMARKS SOLUTIONS TO DETERMINE THE VALIDITY AND ACCURACY OF ANALYT-ICAL METHODS PRESENTLY USED, OBJECTIVE:
TO ESTABLISH PROBLEMS WITH EXPERIMENTALLY KNOWN SOLUTIONS (BENCHMARKS)
FOR USE BY THE LICENSING STAFF TO VALIDATE MAJOR PARTS OF LICENSEE METHODS USED TO CALCULATE STRUCTURAL RESPONSE AND SSI 0F SAFE-TY-RELATED BUILDINGS, SYSTEMS AND COMPONENTS, .
l i
1 0
O O O INTEGRATION:
0 THE FINDINGS OF THIS PROGRAM CAN BE USED IN EFFORTS TO DETERMINE SEIS-MIC SAFETY MARGINS.
0 EXPERIMENTAL RESULTS FROM VARIOUS STUDIES CURRENTLY BEING PERFORMED FOR THE NRC AT SNL (CONTAINMENTS), WCC (E,0. CHARACTERIZATION) AND LANL CATEGORY I STRUCTURES) WILL COMPLEMENT THIS PROGRAM, ALSO WORK FUNDED BY EPRI (SIMQUAKE) HAS BEEN USED, 0 DATA HAS ALSO BEEN OBTAINED THROUGH COOPERATIVE AGREEMENTS (FUKUSHIMA)
O W
l
- l ,
i i . i o
S K
R R
- E A A ET T M DA A H W D C H N TD F E I N O B WU N G O N
,R O I YG , I V
L S T LHD A O AGO U SR CI H L MP I HT A EM F E V LI I ,
M E B CF OF EFL D R0 P0A N P STC A M FI KI
TA RA ED T S M) W EE BT I EU _
HON T LTA TFO AL BET WA ORA NFG ( R RCD I ON U PN I ST OW DSR TC KCE ETE RU R( N TCY OR A AEA FT ' MLE RFL FS HAL TF E CRB _
NEl AI NUA E l 5 S ETL
- CE0 8FS BCI _
. : NHS O UA S OT Y N I RV T C D FNO STA N GN O SS E SNA EI P E M AI TTO PPT H WN , EAH OOA S I E L I S LLU I KML PTK EEL L RRB MI R VVA P 5 OEA 6 ONO 7 EEV
. M 8 WTT 8 CI W 8 DDE O 9 9 9 C 1 1 1 C
A Y Y Y
/ F - F - - - F - - -
E P
O C
S 0 0 0 O
- , , ; i ~ : f! ;il i I .
O O O REGULATORY USE:
O PRELIMINARY FINDINGS HAVE BEEN USED BY BNL TO HELP NRR (SGEB) RESOLVE ;
l SSI ISSUES ON BYRON, BEAVER VALLEY 2, AND DIABLO CANYON NUCLEAR
! PLANTS. .
O BENCHMARKS WILL BE USED TO ASSESS THE ACCURACY OF LICENSEE METHODS USED TO PREDICT THE BEHAVIOR OF SAFETY RELATED STRUCTURES UNDER SEIS MIC LOADS.
I .
O BENCHMARKS AND OTHER DATA DEVELOPED-WILL BE USED TO EVALUATE / REVI AND WHERE NECESSARY ESTABLISH NEW SEISMIC DESIGN CRITERIA (I.E., SRP, REG. GUIDES),
i F
i I
O '
O O REPORTS:
S. SHARMA, M. REICH, AND TY CHANG, BROOKHAVEN NATIONAL LABORATORY, l
" REVIEW 0F CURRENT ANALYSIS METHODOLOGY FOR REINFORCED STRUCTURAL EVALUATIONS," USNRC REPORT NUREG/CR-3284, APRIL 1983.
1 C. A. MILLR, C. J. COSTANTINO, A. J. PHILIPPACOP0ULOS, AND M. REICH, BROOKHAVEN NATIONAL LABORATORY," VERIFICATION OF S0Il-STRUCTURE INTERACTION METHODS," USNRC REPORT NUREG/CR-4128, MAY 1985.
I i
, 'l j -
?
l 4
l .
O O O TITLE:
NRC/EPRI PROGRAMS ON PIPING DYNAMIC LOAD CAPACITY I O EPRI/NRC COOPERATIVE PROGRAM " PIPING AND FITTING DYNAMIC RE-LIABILITY" (FIN D1157) 0 ETEC DEMONSTRATION SEISMIC FRAGILITY PIPE TEST (FIN B3052) 0 INDEPENDENT ASSESSMENT OF PROPOSED ASME PIPING RULE CHANGES 0 INTEGRATION WITH NRC DEGRADED PIPING PROGRAM 0 PIPING FRAGILITY REDEFINITION FOR PRA USE CONTRACTOR:
GE (THROUGH EPRI), ETEC, ANCO, OTHERS MSEB BUDGET:
FY 1985 FY 1986 FY 1987 1
$200K $700K ' $900K 4
i
_ _ - _ _ a
^
O O O ISSUES:
0 ASME CODE DESIGN RULES FOR PIPING 0F UNDER DYNAMIC LOADS MAY HAVE AN UNREALISTIC (0VERALLY CONSERVATIVE) BASIS, O PRA PIPING FRAGILITY ESTIMATES NEED VALIDATION.
OBJECTIVES:
0 EXPERIMENTALLY DETERMINE DYNAMIC FAILURE LEVELS AND MECHANISMS (FA-i TIGUE/RATCHETTING OR COLLAPSE) FOR PIPING SYSTEMS AND COMPONENTS.
0 DEVELOP NEW DESIGN RULES FOR ASME CLASS 1, 2, AND 3 PIPING. .
O ASSESS THE IMPACT OF THESE PROPOSED CHANGES.
O IMPROVE PRA PIPING FRAGILITY ESTIMATES.
0 BENCHMARK NONLINEAR BEHAVIOR IN THE PIPE FAILURE RANGE.
4
~
O O O SCOPE / ACCOMPLISHMENTS:
0 FY 1985 i -
NRC/EPRI COOPERATIVE AGREEMENT ESTABLISHED PROGRAM PLANNING COMPLETED t ANC0 COMPONENT TESTS INITIATED (FIRST 12 0F 40 TESTS)
ETEC DEMONSTRATION TEST INITIATED O FY 1986 COMPONENT TESTING CONTINUES SEISMIC PIPING SYSTEMS TESTS (SEISMIC)
FATIGUE /RATCHETTING SPECIMEN TESTS 0 FY 1987 COMPONENT TESTS COMPLETED HYDRODYNAMIC LOAD a WATERHAMMER PIPING SYSTEMS TESTS FATIGUE /RATCHETTING TESTS COMPLETED DEVELOPMENT OF PROPOSED CODE CHANGES ASSESSMENT OF THESE CODE CHANGES RESULTS OF TESTS INTEGRATED WITH RESULTS OF THE DEGRADED PIPING PROGRAM PRA FRAGILITY MODELS VALIDATED / IMPROVED 4
l j REGULATORY USE:
i 0 PROVIDES THE BASES TO ENDORSE CHANGES TO ASME CODE RULES (SECTION III, !
SUBSECTIONS NB/NC/ND 3600).
- 0 PROVIDE FURTHER JUSTIFICATION FOR SNUBBER MINIMIZATION, e
i l
! i
- t
! l
. l i
I i
[
[
i i
TEST ARTICLE PIPING SYSTEM Y"
f s
'$,f f
~
i
/N x r
x g.a.
N
\
d'g. PIPE MATL: ASTM A-106 GRADE B
/
6"sca 40 s*
g.0 #
NOTES.
(1) EXISTING DUMMY VALVE (2) ALL WELDS TO BE MADE RADIOGRAPHED s AND INSPECTED PER SECTIONS III AND N ,#-h.
'/ genM D IX OF ASME CODE (3) ELBOWS ARE 90' LONG RADIUS V4 ,k, (4) FLANGES ARE 600 LB CLASS
$2 '#-
PROTOTYPIC CLASS 1 PIPE ATTACMENT Mg* X 5
~p 3"scs.4o - f.4'4."
,998 f T Y
/ .
7* T.Nf
' g.C 3 '- c' Q @
5
w 1
1
\
o i go I
e s
w .
- g N i
( ' -%
\ ,
$j..
I /
,/ q
--> D . s. p <-
i*
2 ; - . . . ,
E
$r u.
- E i i 111 s
jg '
-w L
>c ;;
I i
5 0((.r\-i lv I l -
v ' i li. . . .
1 5
' i
- o I.T /A _.
s w
7p"$
3*
I O *4 , N5 a-4 3
.E a
\
W J
a o, , .. -
w .
..-) \
1 I .(..
w (. -
o e
o 7
,.i
. - - -ps g y
- --0 (,
5
-t t .
~
E -
3-2 1 -_ _ - - . _.
~
O .
/
l,;
- e e, .
rr Ta o I
99
! e_
h
=
i l
l :
l Figure 1. Typical 5-in. pipe arrangement.
l 5-in. 3-in.
System (Fig. 1) System (Fig. 2)
End supports only no supports no supports
~
i Hydraulic snubber 1 Mechanical snubber 1 1 l Rigid strut 1 2 Combination strut 1,2 - 1,2 and snubber Strut combination 1,2
O O Lis:ek Reactors v.
(Q d 1, Q d2,Drs2,Drs3,Frma,Zmr1,Kuo,Lal f)
? 20 , V h 19 -
18 -
17 -
+
16 - 4 a 16 -
6 14 0 + + ;
+
$ 13 - 4 k 12 - + 4+
+
+ +
- y fr-
+ +
+
10 - +
g' + .*+ * ++ +
9~ 4+
an 8- + % s Q4 +* g +
C *+ + +
fe 7-s-
5-
+
,,+%
- k #f* 4 t++
++++ ,
- ++
+
+. ++
4
~
+
+
s .
l C5 +
4-
+ ,+ +++ 4 l +
- ~ + +% +
[ +++ 4 4; 2-i
,, + 4,+, +,.
+ + 4+ + 4,
),,4+ +
f-0
- i i i 80 i .
100 L
40 s0 0 20 Freguancy (Es>
i I
I a l ,w ltAl O *Y 4
e
~
O O O TITLE:
OTHER NRC PIPING RESEARCH N0ZZLE DESIGN GUIDANCE (ORNL FIN B0474)
COMBINATIONAL PROCEDURES FOR PIPING RESPONSE. SPECTRA ANALY (BNL FIN A3287)
NONLINEAR PIPING RESPONSE PREDICTIONS (HEDL FIN D1611)
ASSESSMENT AND IMPROVEMENT OF SPECTRUM-BROADENING PROCEDURE (LLNL FIN A0453)
TIME HISTORY PIPE DAMPING (UNDESIGNATED)
ADDITIONAL STUDIES OF COMBINED EFFECTS (UNDESIGNATED) .
BUDGEI:
FY 1985 FY 1986 FY 1987
$333K $375K $600K 1
JSSUES:
0 NEED TO IMPROVE PIPING RESPONSE , PREDICTION TE AS ITEMIZED IN NUREG 1061 (PIPING REVIEW ,COMMITTEE . _ _ _ _ _ _ RECOMMENDA
~~~j--
O O O 3
ISSUES _:
0 NEED TO IMPROVE PIPING RESPONSE PREDICTION TECHNIQUES, AS ITEMIZED IN NUREG 1061 (PIPING REVIEW COMMITTEE RECOMMENDATIONS) 0 NEED TO ADDRESS COMBINED EFFECTS OF PRC RECOMMENDATIONS ON PIPING DE-SIGN ISM METHOD IN COMBINATION WITH PVRC DAMPING RECLASSIFICATION OF SEISMIC STRESSES SUPPORT DESIGN SCOPE / ACCOMPLISHMENTS: ,
- 0 FY 1985 AND 1986 0 IMPROVE N0ZZLE DESIGN GUIDANCE O EVALUATE ISM METHOD IN COMBINATION WITH PVRC DAMPING 0 IMPROVE MODEL COMBINATION METHODS
- ISM WITH KNOWN PHASE CORRELATION CLOSELY SPACED MODES HIGH FREQUENCY MODES
~
O O O TITLE:
OTHER NRC PIPING RESEARCH N0ZZLE DESIGN GUIDANCE (0RNL FIN B0474)
COMBINATIONAL PROCEDURES FOR PIPING RESPONSE SPECTRA ANALYSI (BNL FIN A3287)
NONLINEAR PIPING RESPONSE PREDICTIONS (HEDL FIN D1611)
ASSESSMENT AND IMPROVEMENT OF SPECTRUM-BROADENING PROCEDURES (LLNL FIN A0453)
TIME HISTORY PIPE DAMPING (UNDESIGNATED)
- ADDITIONAL STUDIES OF COMBINED EFFECTS (UNDESIGNATED)
BUDGET:
FY 1985 FY 1986 FY 1987-
$333K $375K $600K l
O O O .
0 -EVALUATION (AND VALIDATION) 0F SIMPLE NONLINEAR RESPONSE PRE ,
TECHNIQUES 1.122 PEAK BROADENING RANGE 0 EVALUATION OF APPROPRIATENESS OF R.G.
O DEVELOPMENT 0F BASES FOR " FLATTENED AND BROADENED" SPECTRA
! O RECOMMENDATIONS FOR DAMPING VALUES TO BE USED IN TIME-HISTO r
0 FY 1987 0 INVESTIGATION OF COMBINED EFFECTS OF NEW PIPING DESIGN CRITERI '
l DAMPING r NEW ASME DESIGN RULES ,
- NEW N0ZZLE DESIGN CRITERIA
- PIPING SUPPORT DESIGN RULES i
i J
4 i
i
i O O O 5
TITLE:
i PARAMETERS INFLUENCING DAMPING IN PIPING SYSTEMS FIN; A 6316 .
CONTRACTOR:
! INEL i BUDGET:
i j FY 1985 FY 1986 FY 1987
$230K $200K $0 K l
l i
i j
j i i
4 t I
1 __ __ _- _ _ _ _ m .
- .r. _
L i
! ISSUES:
1 1 0 OVERALLY STIFF PIPING DESIGNS HAVE RESULTED FROM USE OF R.
j DAMPING VALUES.
f 0 DATA AND GUIDANCE IS LACKING FOR HIGH FREQUENCY MODAL BY NONSEISMIC LOADS.
i ~
i OBJECTIVES:
1
- O COMPILE AND EVALUATE EXISTING DAMPING DATA I
0 TEST SIMPLE SYSTEMS, VARYING RESPONSE AND DESIGN PARAMETERS .
l SCOPE / ACCOMPLISHMENTS:
i l
l 0 FY 1985 !
ESTABLISHED WORLD DATA BASE ON PIPE DAMPING 0
i' l
l l
1 ., . - ,
o
~
O O i
0 TESTING OF 3-DIMENSIONAL SYSTEM 1
LOW STRAIN AMPLITUDES
- DIFFERENT SUPPORT CONFIGURATIONS .
PRESSURIZED AND UNPRESSURIZED DIFFERENT FORCING FUNCTIONS 0 REVIEW 0F EXISTING HIGH FREQUENCY DAMPING DATA O FY 1986 0 CONTINUED TESTING OF 3-DIMENSIONAL SYSTEM HIGH FREQUENCY LOADS INSULATION HIGH STRAIN SEISMIC LOADS -
0 TESTING OF 1-DIMENSIONAL SYSTEM AT HIGH FREQUENCY 0 FINAL REPORT i
l l
I
~~ ' '
i i
O O O i
l REGULATORY USE:
O PROVIDE BASIS FOR ENDORSING ASME CODE CASE N-411 O PROVIDE BASIS FOR REVISING R.G. 1,61 AND SRP f
{
l.
1 .
1 I
f i
o
~
i.
O O CONTAINMENT INTEGRITY PROGRAM ACTIVITIES FIN A-1401 ANALYSIS AND EVALUATION PLANNING OF EXPERIMENTS 4
PRETEST PREDICTIONS POST TEST EVALUATIONS DEVELOPMENT OF PREDICTIVE METHODS l
INTERACTION WITH PEER REVIEW PANEL FIN A-1375 PENETRATION TESTS j
PROCUREMENT OF SPECIMENS .
DETAILED DESIGN OF TEST FIXTURES AND INSTRUMENTATION CONDUCT OF EXPERIMENTS COMPILATION OF TEST REPORTS FIN A-1249 LARGE MODEL TESTS-DESIGN AND CONSTRUCTION OF MODELS INSTALLATION OF INSTRUMENTATION EXPERIMENTAL PLAN AND SAFETY ANALYSES CONDUCT OF EXPERIMENTS COMPILATION OF TEST REPORTS
O O O
't TITLE:
CONTAINMENT INTEGRITY UNDER EXTREME LOADS FIN; A 1401 CONTRACTOR:
i SANDIA NATIONAL LABORATORY COOPERATING ORGANIZATIONS KERNFORSCHUNGSZENTRUM KARLSRUHE (FRG)
ELECTRIC POWER RESEARCH INSTITUTE (EPRI) l COMMISARIAT A L'ENERGIE ATOMIQUE (FRANCE) l 1
NUCLEAR INSTALLATIONS INSPECTORATE (UK)
I BUDGEI:
l FY 1986 FY 1987 FY 1985 l $600K $700K
! $740K 1
ISSUE:
THE CONSEQUENCES OF A ANALYTICAL SEVEREPREDICTIONS ACCIDENT OF WOULD BE S BY THE PERFORMANCE OF THE CONTAINMENT.
CONTAINMENT. FAILURE MODES REQUIRE EXPERIMENTAL VALIDA OBJECTIVE:
DEVELOP A METHODOLOGY THAT CAN BE USED TO PREDICT CONTAINMENT BUILDINGS.DURING SEVERE ACCIDENTS. ,
INTEGRATION:
0 LARGE MODEL TESTS PERFORMED UNDER FIN A1249 PROVID 1 BENCHMARKING ANALYTICAL PREDICTIONS OF STRUCTURAL D 0 PENETRATION TESTS PERFORMED UNDE'R FIN A1375 PROVIDE ANALYTICAL PREDICTIONS 0F LEAKAGE.
1
_ _ _ _ _ _____t __ __ __
O ,
L E
9 0 O _
N T L S N E T O E H F I -
T G O T -
S I A S S R L N N T L I O E A I N M T T E S S C P E I R T D .N O . .
E SO F L R T TT E P ND S SS D EE E EE O Y MS S TT M F I A I
Y LL .RB L E R L E A EE T E EPY N DD E V OO R D XG A MM C O OEO L N T M T LE . CO S ET O EI D E C YS TMO O T T
S EE TR SC N
G N
I RD AA SO SL ESH RI T CEE NS M O EO D E GC .CC O P R N ESEI CRE AR OG D
N LO TTNM XN S RFA K
A F EESI ' O F MTE FS NE A S OS N RNNS SOL N SO OI E .
EI O I I FAMR TI E STA I
T ST SNRD YC C Y C OEN LI F LI I
SCPU X A DO I
AD NE .
D E NE A BDEE ARTS PNT 5 R AR E C EN S P ASEN T T 8 P T S NA S TME N 9 T ST PEI M ESPM E 1 S . ES ORMR TEOI TL R M
H Y ES TE T L TRO T EE S F TE I T ESEF SEVP SE VETR EREE 7 OREX PPDE I
L O ER OR 6 8 P T RE 5 8 PP 8 DPDP 9 M PS 9 1 O R 9 1 1
C O Y C I Y Y - F - - -
A R F - - F -
/ P -
E P
O O C 0 O S O O
- l 'ii . I
, ;l ! l ' i : ! ! :IIti:: !.!0 .L ;iI,i e!l!:L ri O .
R O
F _
T N _
_ E -
M E
I -
, R U -
T N
E M
Q E
R E -
T E
_ A C T N .
S A
_ M
_ Y R
_ C 0 .
I F .
_ L R O E _
P P -
_. T T
_ N N E E M
_ D o
I N
_ C I C A A T .
N E O -
_. R C
_ E V F
_ E O S
N . - _
F O
ON I O TI T I T
_ N E
DA DT
_ M AN S E S EM E LE S BL
_ S I P A SM
. Y S I O
R PL
- O A -
._ E S
T A
R0 OG
_ U M F R Y
_ Y I ST -
R F I E
_ O N SF T O A A A C BS L
U G
E
_ R O 0 O .
O o o i TITLE: -
3 l'
i INTEGRITY OF CONTAINENT PENETRATIONS UNDER SEVERE ACCIDENT CO l
ELNi
)
4 l A 1375 2
1 CONTRACTOR:
i i SANDIA NATIONAL LABORATORY 4 .
BUDGET:
FY 1985 FY 1986 FY 1987 j
! .$900K $900K $600K ;
)
h -
1 :
l
, l j .
I O O O ISSUE:
THE CONSEQUENCES OF A ANALYTICAL SEVEREPREDICTIONS ACCIDENT OF WOULD BE BY THE PERFORMANCE OF THE CONTAINMENT.
CONTAINMENT FAILURE MODES REQUIRE EXPERIMENTAL VALIDA OBJECTIVE:
DETERMINE THE MAJOR CHARACTERISTICS OF OPERABL PENETRATIONS THAT COULD CONTRIBUTE TO LEAKAGE D i
THE EFFECT OF PRESSURE, TEMPERATURE AND DEFORMATION MUST BE S INTEGRATION:
0 THE PREDICTIVE METHODS TO BE DEVELOPED UNDER FIN LARGE MEASURE ON THE PENETRATION EXPERIMENTS.
O TESTS OF LARGE DIAMETER PENETRATIONS UNDER PRESSURE INTEGRATED INTO THE LARGE MODEL TESTS PERFORMED UN i
I -
o
- O O -
i
< SCOPE / ACCOMPLISHMENTS:
0 PRIOR TO FY 1985 i
- SURVEY OF PENETRATION TYPES, GE0METRIES, AND MATERIALS T0 IDENTIFY MOST LIKELY CANDIDATES FOR LEAKAGE TESTS.
DESIGN OF TEST MATRIX FOR SEAL AND GASKET GE0METRIES AND .
MATERIALS.
,! O FY 1985 f
INITIATION OF SEAL AND GASKET TESTS.
PROCUREMENT OF A FULL SCALE AIR LOCK AND SELECTION OF A TEST .
FACILITY.
DESIGN DETAILS COMPLETED FOR AIR LOCK AND EQUIPMENT HATCHES IN CONCRETE MODEL.
O FY 1986 .
I - COMPLETTION OF SEAL AND GASKET TESTS, j -
COMPLETION OF AIR LOCK TEST.
PROCUREMENT OF. BELLOWS AND DEVELOPEMENT OF TEST PLAN.
4l 1
1 O O O l
! O FY 1987 i.
EVALUATION OF AIR LOCK TEST.
l -
PERFORMANCE OF BELLOWS TEST.
' REGULATORY USE:
1 0 CONFIRMATORY ASSESSMENT OF SEVERE ACCIDENT POLICY STATEMENT.
O BASIS FOR POSSIBLE ADDITION OF CONTAINMENT PERFORMANCE RE SAFETY G0AL IMPLEMENTATION, l
- i 1
4 l
S i
j i
1 l
4 O o O l
I TITLE:
i EXPERIMENTS ON CONTAINMENT MODELS UNDER EXTREME LO FIN; A 1249 i
!, CONTRACTOR:
i 1
! SANDIA NATIONAL LABORATORY ,
! BUDGET _: .:
I FY 1986 FY 198Z
- FY 1985 ,
$2000K $2400K 1 $2060 l i i ,
i 4
1
! a l
1 .
(
d l
_ . - y , ,e- -, _ _ , -e -
O O O i ISSUE:
! THE CONSEQUENCES OF A SEVERE ACCIDENT WOULD BE SIGNIFICANTLY AFFEC ANALYTICAL PREDICTIONS OF BY Tile PERFORMANCE OF THE CONTAINMENT.
l CONTAINMENT FAILURE MODES REQUIRE EXPERIMENTAL. VALIDATION.
OBJECTIVE:
PERFORM EXPERIMENTS ON CONTAINMENT MODELS THAT WILL PRODUCE FAILURE l'
DATA UNDER EXTREME LOADING CONDITIONS AND PERMIT AN EVALUATION OF THE CAPABILITY OF STATE-0F-THE-ART CALCULATIONAL METHODS TO PREDICT CONTAINMENT FAILURE UNDER SEVERE ACCIDENT CONDITIONS.
l INTEGRATION:
0 THE MODEL EXPERIMENTS ARE C0ORDINATED WITH THE ANALYTICAL PREDICTIO -
EFFORT PERFORMED UNDER FIN A1401.
0 THE LARGE DIAMETER PENETRATIONS INCLUDED IN THE MODELS PROVIDE DEFORMATION DATA USED IN THE PENETRATION EXPERIMENTS PERFORMED UNDER FIN A1375.
I
O o O SCOPE / ACCOMPLISHMENTS: '
\
i 0 PRIOR TO FY 1985 I
- DEVELOPMENT OF A REMOTE TESTING FACILITY FOR HIGH PR l EXPERIMENTS.
l - TESTS TO FAILURE OF A SERIES OF THREE SMALL STEEL MOD l VERIFY THE ABILITY OF LARGE DEFORMATION COMPUTER C ,
l STIFFENER BEHAVIOR AND BEHAVIOR AT PENETRATION DISCONTIVITIES, l -
DESIGN,' CONSTRUCTION, AND INSTRUMENTATION OF A LARGE STEEL MOD
! O FY 1985 .
COMPLETED TESTING TO FAILURE OF.A LARGE STEEL MODEL, l -
COMPLETED DESIGN AND STARTED CONSTRUCTION OF CONCRETE MODEL,
/
l b 0 FY 1986
- COMPLETE CONSTRUCTION '
AND INSTRUMENTATION OF CONCRE
' INITIATE TESTING.
0 FY 1987 3
1 j
- ' COMPLETE TESTING TO FAILURE OF CONCRETE MODEL
- PERFORM ANY NECESSARY SUPPLEMENTARY TESTS i - INITIATE EXPERIMENTS ON SEISMIC PERFORMANCE 4
)
j - -
i O O O 1
REGULATORY USE:
0 CONFIRMATORY ASSESSMENT OF SEVERE ACCIDENT POLICY STATEMENT, l
- O BASIS FOR POSSIBLE ADDITION OF CONTAINMENT PERFORMANCE REQURIEMENT FOR SAFETY GOAL IMPLEMENTATION.
i i t i .
i ;
i I
l i l i
i-i i j
1 I
__ _ _ _ - __ _ _ _ _ _ _ _ _ _ _ _