ML20132C477
| ML20132C477 | |
| Person / Time | |
|---|---|
| Issue date: | 09/24/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1448, NUDOCS 8509270062 | |
| Download: ML20132C477 (315) | |
Text
ORIGINAL esror
. ~ ~
/ 'I UNITED STATES OF AMERICA V NUCLEAR REGULATORY COMMISSION In the matter of:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Combined Meeting of Subcommittees on Structural Engineering, Seismic Design of Piping, and Metal Components Docket No.
Location: Washington, D. C. Pages: 328 - 522 Date: Tuesday, September 24, 1935
. .* 4M3,
~
KCHSDFRCECOPYx
- ouemergageme 1625 I St., N.W.
Suite 921 pJ Washington, D.C. 20006 t?n?) 793-3950 fr g 6\\ 8509270062 850924 448 PDR
i .,
1
^
k-328 1
UNITED STATES OF, AMERICA.
t 1 C 2 '
NUCLEAR REGULATORY COMMISSION
. ( >
- 1. 3 - --
~
4 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS e
5 COMBINED MEETING OF SUBCOMMITTEES ON STRUCTURAL ENGINEERING, SEISMIC DEdlGU OF_ PIPING, AND METAL COMPONENTS
~
6 } , 3
, 7 - - -
, 8 PUBLIC MEETING \
9 Room 1046
% : u 1G 1717 H Street, N' W .
11 Washington, D.C.
~ 5 1
12J -
h a
, 13' Tuesday,, September 24, -
1985 b , 14 The Subcommittees met, pursuant to notice, at i
15 8:30 a . rs .' '
, s .
16 ACRS MEMBERS PRESENT:
- 17 C. Siess, Chairman '
i 18 P. Shewmon, Chairman
/ 's ,
19 J. Ebersole, Member
'( .
20 .
J.C. Mark, Member ' t 4
21 E. Rodabaugh,-Consultant
\ 7, 2 ?, M. Bender, Consultant 5
23 FEDERAL DES 1'GNATED EMPLOYEE:
24 A. Igne ;
4 s s
25 y
s t
's \ .a,.._..__.___..-- __ -
, . - - ~ . _ - . . . . ~ - . - - - - - . . - . . . . - - . . . - - . - . - . - . - - . . -~.. - - . -_ .. . _ . . ..
i-j :, .
4 i
[ 329
} ..
- j. 1' SPEAKERSi 2 J. Richardson i
! 3 J. Costello h 4 H. A'shar 5 J. Burns i
6 R. Cloud 7 G. Kurtz I
8 B. Romer j '
9 B. Bosnak i.
{ 10 i
11 12 1 13 **
- I i
1 14 i
- i. 15
.1 4
i 16 i
i 17 18'
- i 19 i !
l 20 j l
l' 21 1 !
l .
l 22 l i
! 23 j 24 i 25 ;
l
J 330 1 P ROCEED I NG S 2 MR. SIESS: This i s the second day of the Combined 3 Meeting of the ACRS Subcommittees on Structural Engineering, 4 Seismic Design of Piping and Metal Components.
5 The two major items on the agenda for today, one is 6 the containment integrity program, and the other is really not 7 a research area, is it, Paul? It is an application of the 8 results of research on a leak before break situation. That 9 portion of the meeting may start before noon, and we can do a
10 that with the people here because it is a continuing meeting, a
11 and that will be chaired by Paul Shewmon as Chairman of the 12 Metal Components Subcommittee.
13 Who is going to start?
14 MR. RICHARDSON: I am just going to say a few words, 15 Chet. I would like to have you refer back to the front of tro j 16 package that we handed out yesterday, and this morning we are 17 going to talk about our Containment Integrity Research 18 Program.
19 The third sheet in is the budget sheet for this 20 program, and if you will notice, in 1985 our budget, including q 21 containment buckling, which is not a real integral part of i 22 this program but it is in the containment area, is about $3.9 23 million. In 1986 we are going to $3.7 million and staying ,
24 level in 1987 for $3.7 million.
, 5 I
25 The program has three major elements: experiments on
?
4
- - - , - - - - ,,,-r - , - -
-,,,y --,,-y.w..,,,-w--m,---,..___,-,,-_,_-vv - - ,
--.._-..-,oyy,,,.,..m-,,,w,,,3yy.-w,,.-,,,,y,,,,-.,www-~-w.,--
e 331 1 containment models, containment integrity, which is really the
) 2 analytical portion of the program, and the third element is 3 integrity of containment penetrations.
4 MR. SIESS: As I said, I don't want to get too 5 involved in budget figures, but how firm is your 1986?
6 MR. RICHARDSON: I have no idea. The budget is as 7 firm as Congress will make it firm.
8 MR. SIESS: Let's see. 1986 is at what stage in 9 Congress. Have you got an appropriation?
10 MR. RICHARDSON: No. We have a Senate mark and a 11 House mark, and they are vastly different.
12 MR. SIESS: Okay. The House mark was a whopping 13 cut.
14 MR. RICHARDSON: A big cut. And the Senate version 15 was the OMB mark plus $8 million to cover the demanded salary 16 cut that never came about.
17 MR. SIESS: Okay.
18 MR. RICHARDSON: So the Senate is full budget and 19 the high version i .s --
20 MR. SIESS: Well, where do you stand internally?
21 MR. RICHARDSON: Internally, with the House budget, 22 if the House budget were to prevail, the contingency plans, 23 then, would be to substantially cut severe accident research.
I 24 That would be one of the plan areas for cut. If that were to 25 come about, then, our Containment Research Program would be
a 332 1 also severely cut, probably at least in half.
) 2 MR. SIESS: All right. Now, if this is cut in half, 3 what are you going to do, just postpone that concrete 4 containment?
5 MR. RICHARDSON: We will probably postpone the 6 concrete containment. It will probably bring it to a halt.
7 We put a hiatus on it this summer when we heard the House 8 budget number. We put a hiatus on the construction. However, 9 about the 1st of August we released that and allowed the 10 contractor to proceed because it was becoming -- we were at a 11 cost impact point, whereas if we did not proceed, it would 12 start costing us a great deal of money.
13 The judgment of Bob Minogue and others was that the 14 program would probably survive. It was not a made gamble to go 15 ahead.
16 MR. SIESS: Personally, I would agree with that 17 priority in terms of the research needs, not in terms of the 18 dollars because I can understand that it may cost you more to 19 defer something than it does to do it. But of all the things 20 we have heard about, I think that one is least likely to 21 affect regulation in the forseeable future.
22 That is just a personal opinion, for what it's 23 worth.
24 MR. MARK: What does this mean? We -- the i
\/ 25 committee, that is, have been making noises for a little while l l
J
e 333 1 about the need of a containment performance term in the safety
) 2 goals. Is that directly connected with what we are just 3 talking of here?
4 MR. SIESS: Can I put an intermediate question 5 in? Do you know what we mean by --
6 MR. MARK: You can rephrase my question, if you 7 want?
8 MR. SIESS: No, I just want to ask them if they know 9 what we mean by containment performance criteria.
10 MR. RICHARDSON: No.
11 MR. BENDER: Is a consultant allowed to express an 12 opinion?
13 MR. SIESS: Well, I don't know what we mean by it 14 either, and I'm hoping somebody can explain it to me.
15 MR. RICHARDSON: I turn to my colleagues if they 16 have any --
17 MR. COSTELLO: Jim Costello from the NRC 18 Staff. Since I have been just made a member of a task force to 19 come up with a plan for possibly implementing or examining the 20 options for implementing a containment performance part in the 21 safety goal, the best I could offer is the guidance that we 22 have in the internal staff exercise, which is to examine 23 options which would add or permit the addition of some 24 implementable requirement to the safety goal prescription 25 which would reflect a defense in depth concept.
e 334 1 MR. SIESS: Okay. Well, there have been figures in
) 2 one of our letters that said -- maybe it's a letter that never 3 got sent out, I'm not sure -- that there should be at least a 4 one in ten probability that the containment will function as 5 assumed, on top of the 10 to the minus 5 core melt 6 probability, you see.
7 Now, that I understand, but the function of the 8 containment varies tremendously, depending on what containment 9 you are talking about and what accident sequence you are 10 talking about. It might be a basemat melt-through, it might 11 be a steam explosion, and now we are still talking about a 12 conditional probability of one in ten on the containment 13 performance, which is quite a different thing. If you are 14 talking about a steam explosion, then you are talking about 15 flow overpressure; right?
16 MR. RICHARDSON: Oh, yeah.
17 MR. SIESS: Now, I think we know that if you have 18 flow overpressure from basemat melt-through or sitting there 19 and boiling off water, the containment will fail unless it 20 leaks enough or you open a hole in it. Okay? If you are 21 going to open a hole, deliberately vent, there is some 22 advantage in knowing at what point it would fail, although how 23 precisely you need to know that, I don #t have the slightest i l
24 idea.
25 You know, when are you going to vent? You don't ,
l l
e i
335 1 j
1 want to vent too early because everybody is going to be i 2 uncomfortable. This is, presumably, one reason we are doing 3 ultimate load capacity tests on containments. Now, you look 4 at the steel containment and what do you say you learn? You 5 want a fairly good, sure point that it won't burst before you 6 start venting. I think we know how to do it for steel ones.
7 Yield. General yield, if you want. First yield isn't 8 positive, but if it is going to leak, none of this has any 9 application.
10 So even if I look at that, I would put this as a 11 bottom priority because -- you have heard me say it -- that 12 test is going to ask questions, not answer them, and I think
N 13 we have got enough questions now that we don't need to spend (G 92 million to ask more if we haven't got the $2 14 15 million. That's why I just said I would agree with your 16 priorities.
17 MR. BENDER: Well, I think you said most of it, but 18 it seems to me that for the existing containments, which is 19 really the severe accident policy that you are talking about 20 here, a few kinds of containments that are more vulnerable 21 than others, and I think Chet made the point properly that it 22 is largely the steel containments that are the issues in the 23 safety goal policy if you are going to implement it. And 24 knowing the modes of failure concerning those, the ones that 25 have most influence on what the overall consequences might be.
a 336 1 Everything I have seen shows that the concrete
( 2 containments, reinforced or prostressed either one, have so 3 much extra reserve strength in them that you don't have to 4 worry about them getting up too high.
5 MR. EBERSOLE: May I ask a question? I see a piece 6 of money in here for buckling, which recalls an old issue of 7 about 19 years ago. We found that in steel containments, 8 fairly thin, that buckling, it's autocatalytic, it's 9 unstable. You know, if it buckles, it bursts real quick. The 10 issue 3came up, should you test the hypothetical buckling 11 pressure with some margin over that since structural 12 catastrophe would result if you really buckle?
13 We had cold water we were spraying containments
\
14 with, and you really cooled it down. As a matter of fact, 15 some of our plants had to have the water warmed to keep the 16 containment from buckling. So the question I would ask you 17 here 19 years later is do you think you should require a proof 18 test of buckling strongth?
19 We did it much against the objection of the 20 hieararchy.
21 MR. SIESS: You can't do it on the containment, 22 Jesse.
23 MR. EBERSOLE: Yes, we did. We pulled a vacuum on 24 it.
25 MR. SIESS: You mean build one and test it --
e 337 1 MR. EBERSOLE: No, no, no. I'm talking about in 2 situ.
3 MR. SIESS: Well, you can test a containment up to 4 115 percent of LOCA load, but there is no way of testing a 5 containment up to severe accident load.
6 MR. EBERSOLE: I'm talking about negative pressure.
7 MR. SIESS: Oh, negative pressure. Oh, 8 MR. EBERSOLE: Buckling. You see, that thin ice 9 condenser containment is like a paper bag.
10 MR. SIESS: Oh, yeah. No, they are not looking at 11 negative pressure at all.
12 MR. kICHARDSON: Buckling is from overpressure in 13 the knuckle region.
14 MR. EBERSOLE: Well, what about the negative 15 pressure aspect? You can certainly wreck a plant if you spray 16 too cold water in it.
17 MR. SIESS: I assume somebody is looking at it 18 because in the severe accident, if your sprays come on at the 19 wrong time, you could develop a negative pressure.
20 MR. RICHARDSON: Yes. We are not working at it, but
( 21 Jack, why don *t you --
22 MR. BURNS: Jack Burns from the Staff. I think you l
23 might say about the only area we are looking at the negative 24 pressure in steel containments was the reinforcements around f 25 the cutouts to see what kind of knockdown factors and so on t
L
e 338 1 - would be required to at least bring the cutouts up to the C
( 2 normal pressure strength of the cylinder itself.
3 MR. SIESS: But it seems to me I have heard recently 4 in the severe accident the same kind of thing you got in the 5 BWR. When you turn the sprays on, you can condense a lot of a
6 steam in there and pull a negative pressure. And I thought I 7 heard some talk about vacuum breaker valves or something.
8 MR. EBERSOLE: That includes whether they perform or 9 not. Sure.
10 MR. BURNS: No, we had not looked at that in our 11 program. We had nothing giving that literally as a need from 12 our user office.
13 MR. SIESS: Well, can you follow that up and see if t
14 the Containment Performance Working Group -- who was on thatt 15 You, Jim?
16 MR. COSTELLO: I am not really on the working group, 17 but I was on one of the advisory committees.
16 MR. SIESS: Well, what they are looking here is with 19 the nonhemispherical heads.
20 MR. EBERSOLE: 1 understand that. I just jumped to ,
21 that.
l l
22 MR. SIESS: Remember we went through it on OESSAR. '
23 MR. EBERSOLE: Yes, I understand, but this other l
24 thing has -- you know, if you get a sticky vacuum release, it 25 can get messy. 1
i e
339 1 MR. SIESS: Well, somewhere you have looked to 2 buckling due to earthquake loads, haven't you?
3 MR. BURNS: Yes, that is part of the program now.
4 At the moment, or coming on the heading buckling problem, that 5 was not an underpressure. That was really an overpressure 6 condition we were looking at for the shallow heads. If you 7 pressurise, you can develop a shear knuckle buckling.
8 +-
MR. SIESS: Yes, that was a toroidal.
9 MR. EBERSOLE: Of course, this was the first time we 10 ever saw a vessel that had substantial catastrophic potential 11 if it buckled in the inward direction and busted the pipes and 12 things.
13 MR. SIESS: Well, previously the only --
}
14 containments have been looked at for something like 2 or 3 psi 15 negative. From what -- ! can't remember. What is the source 16 of that?
17 MR. BURNS: Spring of cold water. Probably a spray 18 condition.
19 MR. SIESS: In the LOCA condition?
20 MR. EBERSOLE: Anything that steams it up, and 21 especially the --
22 MR. SIESS: I'm talking about what they have done in 23 the design, not what we are looking at now for severe 24 accident.
\'-
25 MR. RICHARDSON: Are there any steel containments
e 340 1 that maintain a subatmospheric --
b 2 MR. SIESS: No, no. No, that's all the concrete.
(
3 MR. EBERSOLE: There is another thing. If you have 4 lost the component atmosphere which is in there and you have, 5 say, 100 percent steam --
6 MR. SIESS: Oh, yes, I know the circumstances. I'm 7 trying to talk about what has been the design basis because 1 8 have been -- Hans, what is the design basis now for 9 containment for negative pressur'e?
10 MR. ASHAR: Generally for the vacuum due to the 11 tornado loading, some steel containment had been designed for 12 3 psi or so.
13 MR. SIESS: Tornado loading produces a negative 14 pressure outside.
15 MR. ASHAR: Outside, yes, correct. Tornado pressure 16 outside of 3 or 4 psi and then there are vacuum breakers.
17 MR. SIESS: But what would -- something that will 18 produce an external pressure.
19 MR. ASHAR: But they are only in the range of 3 psi 20 or 4 psi, not more than that.
21 MR. SIESS: If a containment is designed for 50 or 22 60 pst gauge inside, you mean they have just added 2 or 3 pst 23 to that from the vacuum?
24 MR. ASHAR: Oh, no. I'm talking about negative O 25 pressure. ,
.-_ _~ _ __ _ _ _ _ _ . _
341 1 MR, SIESS: How do you get a negative pressure O)
( 2 inside from a tornado? A tornado produces a pressure drop 3 outside, and if you are atmospheric inside, that is a positive 4 pressure inside gauge, j S MR ASHAR: Chet, there is the pressure from outside 1 .
6 and the pressure from inside. On the other side there is --
7 MR. SIESS: I still don't understand.
8 MR. EBERSOLE: I don't understand how you can get a 9 positive pressure from a tornado unless you vent it.
10 MR. BENDER: You can get some negative pressure on 11 the downstream side of the wind.
12 MR. BURNS: That is just an unbalanced pressure 1
v
} 13 around the cylinder.
14 MR. SIESS: But there is a design basis. I know it 15 becausse I have seen it in FSARs for 2 or 3 psi negative, but 16 that's r et --
17 MR. EBERSOLE: Let me set the stage for the real 16 hairy thing. You have depressurization from someplace. You 1
19 have the purge valves open. You blow out a substantial 20 fraction of the atmosphere. I am making it bad. Then they 21 succeed in closing, and then you hit it with real oold water, 22 and then, if you want to compound it, the vacuum breakers 23 don't all work.
~. 24 You have available 15 pai. Hot containment How 25 much will you get and when will you collapse, and what is the
a 342 1 margin?
2 MR. SIESS: This has to be looked at in the severe 3 aooident thing.
4 MR. BURNS: Keep in mind, though, that a buckle 5 itself may not be a failure of the containment. It has to be 6 !arge enough to impact on something inside or to crack the 7 containment in one form or another through a buckle or through 8 distortions of the penetrations.
9 MR. EBERSOLE: Right, but if it starts to buckle, as 10 I understand it, it's unstable and it will rapidly get worse.
11 MR. SIESS: It depends on whether the load is --
12 MR, EBERSOLE: Yes, It may just implode.
/ 13 MR. MARK: Chet, when we were at Sandia, they were b} 14 doing some very small-scale looks at buckling features of this 15 dome.
16 MR. SIESS: They were looking at two things. One is 17 they were looking at the code reinforcement formulas to see if 18 that adequately took care of buckling due to overturning 19 vertical compressions in the shell due to seismio loads 20 overturning, and then they were looking at buckling in the 21 knuckle region of those tortsphertoal heads. I believe with 22 the hemispherical heads, you don't run into the problem.
23 MR. BURNS: That was our conclusion from the test, 24 including a large test at Chicago Bridge and tran.
25 MR. SIESS: Those were the two things they were
e 343 1 looking at, and one was due to ovt: turning from
( ) 2 earthquake-type loadings, and the other was due to internal 1 3 pressure. If you remember, GESSAR went through this 4 complicated analysis because they have got the tortspherical 5 head and -- but the internal vacuum pressure -- I think we 6 ought to have A1 look at it with the NRR people and try to 7 find out what the situation is, and you can follow it up and 8 see if there is a problem there and whether somebody is 9 thinking about it.
10 MR. EBERSOLE: Chet, I want to ask you this. I 11 never have personally been able to rationalize where there 12 isn't a very conservative rettei valve on containments.
l
/~} 13 MR. SIESS: Well, there are a number of us looking 14 at -- we see figures now saying that if you have a 10 square 15 inch opening, you can never overpressure the containment 16 unless you have something like a steam spike.
17 MR. EBERSOLE: You can back it up with a valve that 18 closes by motor.
19 MR, SIESS: The question of deliberately venting, 20 which the BWRs are proposing, hasn't been seriously considered
, 21 -- well, it has for the PWRs. I know EPRI looked at it in 1
! 22 their studies. But what people have come up with is that the I
I 23 slow overpressure that you could relieve with a modestly small 24 valve takes a week to burst the containment, and by then the 25 fission products are bound to a low level You could have
--- ~ - _ . - , . _ _ - - - _ _ - . - - - _ , . . . _ - , - . - - _ - , - _ _ , - , - - - , _ _ - - , _ -
344 1 evacuated people. And the consequences become negligible, q_) 2 MR. RICHARDSON: Unless you get a catastrophic 3 failure of containment where you might reintroduce aerosols 4 that have plated out.
5 MR. SIESS: The first thing, it isn't a nice idea, 6 let's face it, but neither is a deliberate release.
7 Now, the near-term operating license requirement is 8 that they provide an opening in the containment, equivalent 9 3-foot diameter hole for possible vented filtered 10 containment. Three foot diameter.
11 MR. EBERSOLE: With no intention of really using the 12 hole.
13 MR. SIESS: Well, the idea someday that they might
\
14 run a pipe through there to 20,000 tons of gravel; not to a 15 lake, to gravel and all of this stuff. The vented filtered 16 containment has been looked at. Even what they call a simple 17 cheap filter. But now the boilers are proposing to vent into 18 cooler water and take a terrific benefit from scrubbing, and l
19 if you don't want to overpressure this thing, it looks now 20 from the scenarios that have developed that a 3 or 4-inch hole 1
21 will take it out.
22 Now, somebody has looked at this for me and said I
23 that some of the containments with the double containment in l
l l 24 the annulus and a filter system on the annulus can take a fair 25 discharge right there. If you could discharge it under water i
1
T e
345 1
into your cooling tower basin, ing1,your ultimate heat sink
\
("'/
\
(, 2 and take some credit for scrubbing --
3 MR. EBERSOLE: vou can scrub the annulus if you want 4 to.
5 MR. SIESS: What about the spent fuel pool?
6 MR. ESERSOLE: There are all sorts of ways of doing 7 this.
l l 8 MR. SIESS: But on Indian Point -- and I talked to
- 9 Herschel about it -- they say, you know, that is so far down 10 the line you are preventing something where the consequences 11 are negligible, and if you have had the core melt, you have 12 already done the public relations damage. You haven't hurt 13 anybody yet, you know.
14 MR. BENDER: Well, just so it doesn't get too 15 distorted, a lot depends on when the venting occurs. The 16 Indian Point concept is based on the integrity of the 17 containment being fully assured up to some pressure point, but 18 it doesn't allow for the potential of a leak early in the 19 event.
20 MR. SIESS: Oh, yes. They have made 21 analyses. Studies have been made as to what the doses are if 22 you have a leak early. Spector sent me a copy of a paper --
23 1 don't know whether anybody else has gotten it. Los Alamos, 24 I think, made a parametric study of early leakage and 25 mentioned the Containment Performance Group.
e 346 1 It's not simple, obviously, but if you do have the
[ leakage early, you won't overpressure but you will release
( 2 3 early, and now you have got the trade-offs. If the late 4 release due to rupture of the containment doesn't hurt anybody 5 because everything is settled out and it doesn't resuspend, 6 then early venting or early inadvertent venting or leakage may l
7 not be good.
8 MR. BENDER: They may have been able to make that 9 case. The analysis may be correct, but they had a very 10 prescriptive way of defining the accident You have to accept 11 their prescription.
12 MR. SIESS: You are talking about Indian Point?
13 MR. BENDER: Yes, I'm talking about Indian Point.
14 The other areas that have to be thought about, this )
15 big, 3-foot diameter opening that was put in way back when, ! ;
16 think there is no defensible filter system that goes with that 17 particular scheme. l i
18 MR, SIESS: Well, somebody had one in mind. !
19 MR. BENDER: They just don't have enough information 20 about it. BJt the smaller systems where you vent with a j I
21 10-inch pipe or thereabouts, they can be bubbled through water ;
22 or anything and they will do most of the decontamination that i 23 is needed.
24 MR. S!ESS: Yes, but the BWRs just haven't looked at 25 that seriously. As I said, EPRI looked at it and their l I
i I
l I
347 i
1 cost-effectiveness is poor. Because the dose reduction was l
i O(_,/ 2 negligible, there wasn't any dose there to speak of, so they 3 couldn't afford to spend more than 650,000 to prevent rupture 4 of the containment seven days down the line. That was the 5 kind of calculation that was -- I said EPRI IDCOR. Excuse 6 me.
l l
l 7 And if you have an early steam spike or a hydrogen l
l 8 detonation, that is not going to help you anyway.
l l 9 Jim? i l
l 10 MR. COSTELLO: As far as I know in the matter of l
11 practice overseas, the Swedes are, I am virtually certain, 12 going to do something with a PWR that is similar to their BWR 13 venting scheme. I don't know what that is yet.
14 MR. SIESS: Oh, but it involves a huge gravel 15 filter.
16 MR. COSTELLO: That is the BWR. Are they going to 17 use the same system on the PWRt i 18 MR. SIESS: Yes. But the Swedes came to that 19 solution before they had the benefit of the severe accident 20 sequence analyses. You know, I think their decision to go to l
l 21 vented filters was almost a political one without the 1
1 22 engineering basis that we might now have for it, but the RDA 1
l 23 studies on vented filters, you have seen those, and they come
)
24 up with one that is fairly cheap looking. It's a separate l
O 25 butiding sitting over here that is hardly passive. It has got
\
l l
l
_________________-______-____-_____-___-___________-_______________-_____\
a 348 1 fans. One of them even uses ice condensers. They just go on 2 condenser steam. I thought that was interesting.
3 But I just can't see why the PWR doesn't fall 4 somewhere in the same ballpark as the BWR, The BWR is not 5 going to vent through a 3-foot hole.
6 Now, the other point, though, is, whether or not you 7 deliberately vent, there may be enough leakage; but again, so 6 far, looking at what we have looked at, you are not going to 9 get an awful lot of leakage until you get up fairly high in 10 pressure compared to the design pressure; right? That's what 11 everything seems to be telling us so far.
12 And if it takes three days to get up to where you 13 have significant leakage and you start looking at dose 14 calculations, you don't see too much benefit. If you look at 15 defense in depth, you see something else, I think.
16 MR. COSTELLO: I could also offer from my 17 understanding of what I have heard about the French thinking.
18 For one thing, they feel -- double containment is unlined, so 19 they feel they can handle leakage through the inner 20 containment into the anular region and clean it up. Or so 21 they say.
22 MR. BIESS: Well, how much can they handle?
l 23 MR. COSTELLO: I have never seen a clear number on I
24 that, but they say enough.
(~ As far as the PWR, their 900 megawatt PWR with the 25 l
l i
l
e 349 1 similar containment and the liner, I have heard talk over
) 2 there about venting schemes, gravel pit, that kind of stuff, i
I 3 although I have never seen any indications they are 4 implementing that.
l 5 MR. SHEWMON: The last time the group was over here 6 a year or two ago, they said they had decided to go to a 7 vented, filtered containment. That was a decision. The l
8 building people didn't like it.
9 MR. SIESS: One problem I have is I can't tie 10 numbers together. Some people are talking about leakage and 11 the terms we use for leak testing, percent of the volume per 12 day. Then I see filtered systems described as so many cubic 13 feet per minute capacity, and then somebody else tells me from 14 the severe acoident study that a whole 10 square inches -- and 15 I can't relate 10 square inches to cubic feet per minute to 16 percent per day and get any feeling as to whether what goes 17 out of a hole 10 square inches can be handled by the annulus 18 filter system or not. You seet 19 Hans, did you do some looking up on this for met 20 Was it yout 21 MR. ASHAR: This is Hans Ashar, NRC Staff. Some l
22 work has been done, and the way we are going, for testing i
23 purposes, is by cubic centimeter or centimeter or length,
- o 24 25 parameter of penetration, four second, that kind of a parameter, because we are going to integrate a lot of leakages l
- \
l 350 l I
l 1 from a lot of penetrations, but as far as the 10-inch diameter '
l
, 2 or 6-inch diameter, in my own opinion I think it all depends 3 upon the flow rate between the outside and the inside, and !
I 4 depending upon what kind of flow you assume.
l
! S That's the reason it's not very definite that this 1
j 6 diameter hole is going to depressurine containment to the !
l 7 extent that it won't pressurize and rupture the 8 containment. This is an uncertainty that is always there, i
9 MR. SIESS: Yes, but then the containment I
10 performance working group has got curves in their report ;
f 11 showing various size openings, and it depends on a nurab e r of 1 12 things, I will admit; but the simple question -- take a 13 specific plant with an annulus, a 6-foot annulus with a filter 14 system in it, and tell me -- take the filter capacity and feed 15 it back into one of the accident sequences and tell me what it 16 does.
17 This is a filtered vented release through an r
18 existing sytem, which ! think is Category I sf necessary, 19 MR. ESERSOLE: Chet, may I ask a question?
20 MR. SIESS: And you can't get the answers to that, i 21 MR. ESERSOLE: I would like to know, and I think j 22 some pressure should be put on the industry to determine what 23 the real leakage rate is under design conditions. My 24 impression is the leakage tests are done like the old n\/ 25 general's visit to the Army post You knew he was coming ten
,-_ _..u_____-
- . . _ ~ _ _ - - _ . __ _ _ _ _ . _ _ _
e 351 l
1 days in advance and you fixed it.
2 MR. RICHARDSON: Well, with the modification that we i 3 are in the process of rulemaking right now to modify Appendix 4 J, the integrated leak rate test procedure, and in fact, we 5 are trying to plug that loophole. The Staff has never 6 considered that the applicant had or the utility had the 7 privilege of fixing up before the test. However, some 8 utilities disagree with us. We are going to make that very i
9 clear in the revision to Appendix J, that they can't do that.
10 MR. SIESS: But you have been getting some reaction 11 from CROR on that.
12 MR. RICHARDSON: Oh, yes.
13 MR. SIESS: This is under review.
14 MR. RICHARDSON: I might say the CROR is now -- we 15 have now gotten resolution of CROR and it is ready to go now.
16 MR. EBERSOLE: Oreat. What does this mean, that 17 there will be a SWAT team show up at your front door some day 18 and it won't impact on their running condition?
19 MR. RICHARDSON: No. It just says that any time you 20 make repairs to valves, you have to count that in the tally of 21 leak rates, i
22 MR. EBERSOLE: Oh, well that's not what I mean. I'm talking about a black hat team shows up at 2 a.m. in the I 23 i l
24 morning and says: We are going to do a leak test. l l
[h
\- 25 MR. SIESS: You can't do it, Jesse, i
l I
f
352 1 MR. RICHARDSON: No, because that involves shutdown.
i ( 2 MR. EBERSOLE: No, I am talking about a careful l
l 3 program that would not impact on generation.
l 4 MR. BENDER: What do you want to know, whether it 5 leaks at .1 percent --
6 MR. EBERSOLE: I want to know whether it leaks l
7 before they fix it. I just want to find out what it leaks, 8 Mike.
9 MR. 81ESS: Well, their procedure would not tell you 10 at any p ra i n t in time what the leak was because a valve could 11 he open for 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> before somebody --
12 MR. EBERSOLE: Well, I'm talking about the automatic 13 closure of the system by punching a button. That's the way it 14 works.
15 MR. SIESS: I don't know if it can be done the way 16 you say without --
17 MR. EBERSOLE: Well, how does it work at all?
18 MR. SIESS: Well, give me a couple hours sometime or 19 the next time we look at Appendix J. But there are other ways 20 the Staff is looking at it. You are going to hear a lot from B
21 Jim Costello about what they are doing on seals and i
22 penetrations and so forth, but what they are not doing in the !
23 research program is 1 coking at what they refer to as 24 pre-existing leakage. They have said that can be handled l
25 administratively. Well, obviously you can't do anything in
. . . . _ _ _ . ~ _ _ - . , . ..
e 353 1 terms of research on it, and there are a couple of 2 approaches. One of them, Conn Yankee, for example, operates, 3 or at least one of the plants operates under 2 or 3 psi.
4 Now, with their recording that they have, it takes 5 them about a week to detect one-tenth of one percent leakage 6 and two-tenths of one percent leakage, whatever their tech 7 spec leakage is. But we are not worried about one or
! 8 two-tenths percent a day leakage. When you get up to one 9 percent a day or something like that, I'm sure they could 10 detect it a lot earlier, 11 The subatmospheric plants, which have to pull about 12 6 psi vacuum -- that's about what it is -- they operate at 9 13 or 10, don't they? Well, somewhere in that range. They pull 14 a few pst vacuum. Now, that is leakage in the other 15 direction, but it would catch the open valve. There they are 16 giving them credit for pretty good control because the leakage 17 can't be greater than the capacity of their pumps that are 18 keeping the air pressure down. So there are ways of looking 19 at that.
20 There is another approach, called a blunder test, 21 which is every time you have had an outage and you close 22 things back up, you run a test to be sure that somebody didn't 23 leave something open. But that is the pre-existing leakage, 24 and you won't hear anything from Research on that because it's 25 not a physical research-type problem, and I believe that
e 354 1 administrative controls are being handled in NRR somewhere.
2 Don't you have Pacific Northwest or somebody looking 3 at--- ,
5 .MR. SIESS: Oh, they've got a technical assistance i
6 program.
7 MR. COSTELLO: But I guess I could say a'lso on the 8 question of the inerted BWRs, there is also a monitoring 9 there, because nitrogen costs money, and the utilities find 10 leakage.
11 MR. SIESS: And inerted BWRs have also got a 12 control, but you are going to hear from Jim about leakage A) 13 studies, and there are two aspects about leakage.
14 One is, if it leaks enough, it won't overpressure, 15 which is good if you don't want the containment to fail. And 16 what it does to doses is something else, and if it leaks 17 early, you get doses.
18 Now our whole thinking about leakage has been LOCA, 4
19 and we don't want leakage more than 1/10th of 1 percent, i
20 because combining that with the source term and the 21 meteorology exceeds 25 rem. You know, that rigamarole we have 22 gone through.
' So the research is trying to get some numbers, but 23 24 the. thinking about what to do about it is supposed to be 6
( 25 coming out of the severe accident program. And I haven't seen
- \
i l
355 1 that much thought being given to it, frankly.
2 MR. RICHARDSON: I have just a couple of words for 3 those that may not be intimately familiar with our containment 1
4 integrity program.
l 5 The overall objective, of course, is to study, I
6 evaluate and assess the structural and leak behavior of j l
l 7 containments up to and including severe accidents, and we are 8 doing this experimentally to validate, improve analytical 9 models and predictions, so that we can then take this 10 methodology into all the various containments that exist.
11 The program consists of three basic elements. That 12 is, the large model testing of steel and concrete 13 containments, and the only concrete that we are testing is a i
14 reinforced concrete. We have made a conscious decision not to 15 test prestressed concrete.
16 MR. SIESS: On the basis that it no longer exists?
17 The British are going to test one.
18 MR. RICHARDSON: And it still looks like the British 19 are going to test one. That's the word we art getting from 20 NII.
21 The CEBG doesn't want to do it, but it looks like 22 NII is going to force them to do it. And, of course, our 23 penetrations program, which is a separate effects program, 24 looking at the seal and gasket materials and testing a few N- 25 penetrations by themselves, using the data from the large
,'5 ,
356 L
1 model tests as boundary conditions, and then an analytical
) 2 effort that puts it all together and' develops an analytical 3 process that can be used to assess containments.
4 That, coupled with-some supporting programs in the s
5 research areas that are feeding into this program, such as the 6 electrical penetrations, ~are being evaluated under the 7 electrical equipment qualification program and the containment
\
8 isolation valv s being evaluated in the mechanical equipment 9 qualification program.
10 MR. EBERSOLE: HowJeome these programs are mixed up
. ?
11 like this? How come the programs are mixed up like this? Who i
12 really runs the show?' We do.
m 13 Well, let me call out then, as late as,,on GESSAR II, 14 I saw a line that says electrical penetrations not carrying 1E 15 loads won't need s 1E overcurrent protection. That's a flat 16 wrong statement, because these penetrations carry fault loads 17 under accident conditions. If you don't clear them, more 18 , often than not, they are the fuses in the circuit and they'll 19 blow up.
20 MR. RICHARDSON: Well, maybe I answered too quickly 21 who has 'c on t r o l . The control of the research program of l 22 containment integrity rests with this branch, and we guide the 23 research in the electrical penetrations area, in that we I I
l 24 specify what we want. .
25 MR. SIESS: But,does the research on electrical i
(
- 6 , m ,- - - ~ - - - - , ~ , - , - , - , , , - - -g. , , , -n- n-- - - -
a 357 1 penetrations look at electrical overloads as a way of rm k) 2 disintegrating --
3 MR. RICHARDSON: No. It is looking at pressure, 4 pressure, temperature, radiation.
5 MR. EBERSOLE: Well, you know the conductors that go 6 through those, that's the pinch point.
7 MR. COSTELLO: They are measuring fault currents and 8 things like that in these.
9 MR. EBERSOLE: And what clears them?
10 MR. COSTELLO: That I couldn't tell you.
11 MR. EBERSOLE: That's what keeps them from being 12 sustained and melting.
[~' 13 MR. RICHARDSON: Perhaps Hans can --
14 MR. ASHAR: The last electrical penetration they 15 tested, what they saw was under pressure and temperature, 16 after aging for so many days -- I don't remember how many days 17 -- the radiation and temperature, they subjected to 18 temperature and pressure, simulating an accident, and what 19 they observed was there was no leakage from inboard or from 20 outboard.
21 MR. EBERSOLE: That's not the point.
22 MR. ASHAR: But let me say -- go further. But there 23 was a -- to the conductors, and the conclusion -- not 24 conclusion, some kind of a decision was that some of the 25 instrumentation depending upon penetration would not be
358 1 functioning, because of the heavy damage to the insulation.
2 But --
3 MR. SIESS: I'm sorry, but the point Jesse is making 4 is that electrical overload could damage that penetration and 5 reduce its ability to resist pressure or temperature in the 6 subsequent stages of the accident.
7 MR. EBERSOLE: No. It could just fully blow it out.
8 MR. SIESS: Well, I don't think that's a part of the 9 research program. I think it should be looked at.
10 MR. COSTELLO: Okay.
11 MR. RICHARDSON: You have a good point.
12 MR. EBERSOLE: Well, it's an old, old issue, you
[ ) 13 know. For instance, the main coolant pumps, they have been V
14 typically protected by a single circuit breaker out in the 15 yard and on seismic 1E or whatever, with a battery that has no 16 pedigree, and so when you have a LOCA, these big pumps go to 17 fault current conditions, and the circuit design almost always 18 causes the penetration to be the pinch point of 19 current-carrying ability, so it will simply blow like a big 20 fuse. Physically. I am talking about explosive. Unless you 21 clear it in micro-milliseconds. That has to be done by a t i. g 22 breaker.
23 MR. SIESS: There are some penetrations that may not 1
24 be protected from overload because the circuits are not 25 needed, but if the penetration fails --
1
359 1 MR. RICHARDSON: Yes, you may develop a leak.
N
- 2 MR. SIESS: Let me supplement what you said. Two 3 things, Jim:
4 One thing that's very important to these large 5 containment tests, the one that they have tested and the one 6 ihat they propose to test, we all realise that there are 7 probably 50 or 60 different containment designs out there, if 8 not 90 or 120, when you really get to looking at them, and 9 there is no hope of ever testing all of them, or even testing 10 parts of all of them.
11 So the whole process here is we think we can analyse 12 containments for their ultimate pressure-carrying capacity, 13 and we are going to test a couple and test out our analysis, 14 and then try to set limits on the analysis, that we can use 15 analysis to handle all these different varieties.
16 Now that is an extremely important assumption, and 17 the results are still going to have " uncertainties."
18 The other point I want to make is t h 'a t even those 19 tests, failure of a containment is defined now not as 20 structural failure, but as leakage beyond some appropriate 21 limit. The function of a containment is to contain, and when 22 it no longer contains, it has failed.
23 So the emphasis is not on structural failure, but 24 they are integral tests, they are separate effect tests, and 25 they are analyses of the integral tests intended to verify or
360 1 validate the analyses and the separate effects tests are 2 intended to provide input, too.
3 I think this is very important, because there's a 4 very basic philosophy behind this, and it's not one that 5 guarantees an answer, by any means.
6 MR. EBERSOLE: Chet, I want to make one comment.
7 You are sort of the chairman of the containment integrity 8 program -- this is an old issue that goes way back again, at 9 least 15 years, and a lot of patch work and compromises have 10 been done on it, but I for one don't know how it was really 11 wrapped up, especially with regard to old plants, because all 12 of them were found to be deficient in this matter of trip
[~h s
13 integrity. And so it might be worth an overview.
14 MR. RICHARDSON: We will certainly look into it.
15 Well, that's all I have, and I am going to turn it 16 over to our two project managers, Hans Ashar and Jim Costello, 17 who will describe the program.
18 [ Slide.]
19 MR. COSTELLO: I am James Costello from the NRC 20 Staff.
21 To get you to the right page in the handout, I have 22 put up here the cover page.
23 As has been discussed previously, we have a large 24 effort concentrated on three areas. As Prof. Siess pointed 25 out, the underlying hypothesis behind these efforts is that
361 1 the state-of-the-art analytical methods, intelligently
()
n 2 applied, will give answers that are good enough to make
- 3 inferences about containment leakage and behaviour near 4 failure.
5 Of course, the question of how good the analyses 6 really turn out to be has to be checked experimentally and, 7 indeed, we will find out, as we get into the history of the 8 steel model test, the iterative process of pre-test 9 prediction, post-test analyses and an assessment of how good 10 the state of the art is.
11 The three activities listed up there are followed 12 by, in your package, a bunch of format pre '85 activity, '85 13 activity, '86 and '87 activity.
14 MR. SIESS: This is all Sandia, right?
15 MR. COSTELLO: All three of these activities are 16 being performed at Sandia Laboratories.
17 We have in '85 sliced things up into three FINS, 18 mainly for programmatic and management reasons. We have found 19 in the past that leaving the analysis efforts in the same FIN 20 with the model construction led to insufficient attention i
21 being devoted to the analytical methods. l l
22 I mean it is a simple fact of life, that getting a 23 large experiment designed is a very ticklish undertaking, and 24 attention gets devoted to the most difficult task.
\- 25 So we have attempted to break things out this way.
e 362 1 It somehow makes it difficult when we come for a presentation,
) 2 when we compartmentalize things this way, but we will do our 3 best to give you presentations on the highlights. And since 4 we are doing it all in the same morning, I trust you can see 5 how they hang together.
6 MR. EBERSOLE: May I ask a question in sort of an i
7 overview? I would have thought this test would include severe 8 accident potential capability of the containment, which goes 9 whole-hog down the road.
10 MR. COSTELLO: It does, yes, sir.
11 MR. EBERSOLE: Well, then, for instance --
12 MR. SIESS: Except temperature.
N 13 MR. EBERSOLE: Well, there I go.
14 MR. SIESS: I know. Temperature doesn't have any 15 effect on the steel itself. It's mainly on the penetration.
16 MR. EBERSOLE: Well, let's say the hypothesis is we 17 melt the core and the vessel fails and it falls down into 18 something, and I think the rationale has to be it is 19 concluding now whatever the heck we're going to do under those 20 circumstances, we will never fail to try to drive water to the 21 core, we'll do it anyway. We don't know what else to do. l i
i 22 So this says if I can't drive water, I'll surely be 23 in trouble. If I can, I might be. And this automatically 24 leads to the thesis that a containment might be optimized by 1
\
]
25 having a pre-existing pool of water into which the residual of
363 1 the vessel and fuel would progressively descend.
2 I'm talking about a hypothetical model, not -- we've 3 got, as Chet says, maybe a hundred models, but I'm talking 4 about one that you hope would work.
5 What is being done about any sort of thing like 6 this?
7 MR. COSTELLO: This program? Literally nothing.
8 The focus of this program is developing reliable predictions 9 of the pressure and temperature conditions.
10 MR. EBERSOLE: In the field of containments we have?
11 MR. COSTELLO: Yes. Under which leakage or i
12 significant leakage will begin.
I 13 MR. EBERSOLE: There's no attempt then to really 14 devolve a containment --
15 MR. COSTELLO: No, sir, this is just establishing 16 facts.
17 MR. BENDER: It just seems to me that based on the 18 discretion here, that you would be way ahead if you just wrote 19 down a few scenarios and related the experiment to the few 20 scenarios that you are using. They exist somewhere, and it 21 seems to me that if they were just tied together a little bit 22 --
23 MR. SIESS: Mike, what do you mean?
24 MR. BENDER: For the purpose of showing what these
(
25 containment tests tell you. You have to have.some reference t
-._-_m. _ . - - . --
e 364 1 accident condition that you are relating to, and I think it is f 2 still quite abstract.
3 MR. SIESS: You mean the rate of pressure increase?
4 MR. BENDER: No, I am talking about the conditions 5 including rate of pressure increase that have to be 6 considered.
7 MR. SIESS: In the penetration studies, as you will 8 hear, the pressure temperature from accident scenarios is 9 being involved, or being considered. In the model containment 10 test, there is no way they could get the temperatures on 11 there, and they would have no effect on the basis shell. That 12 couldn't be predicted. And the idea is to verify or validate 13 analysis, so they are simply applying a pressure.
14 Now the time rate would be the only variable there, 15 and they can't really do anything about it, because they have 16 got to stop and take readings and so forth.
17 MR. BENDER: I am not trying to argue that the 18 experiment should simulate the event. I am trying to get --
19 MR. RICHARDSON: What is it going to address?
20 MR. BENDER: Yes, what event? Not what event, but 21 what events?
22 MR. SIESS: Well, that's covered. Containment loads 23 working group has come up with scenarios and they are 24 summarized in 956, or whatever it is, in the appendix, and 25 they are all in their report, and they have taken various
365 1 scenarios, Mike, at various leak rates and various kinds of
'4 O 2 centainments.
3 MR. BENDER: I don't disagree with the point you're 4 making, Chet. All I'm saying is that if it is put into 5 context, it will make it a lot easier to understand why the 6 experiments have meaning.
7 MR. SIESS: Well, the thing is, it's going to be 8 hard for them to put it into context here when we have spent 9 several days in meetings on the Class 9 accidents, looking at 10 the scenarios.
11 I think there is a problem here for the consultants 12 who haven't been sitting in those meetings.
( 13 MR. RICHARDSON: It probably would have improved the
(
14 presentation if we would have shown some bounding curves, 15 showing the range of accidents that are being considered and 16 how this fits into what we are addressing.
17 MR. SIESS: Yes. Well, Mike and Rodabaugh are at a 18 disadvantage there. I have seen them, I think Carlton has 19 seen them from Class 9 meetings, and Jesse, and this is not 20 being done in a vacuum, I can assure you.
21 MR. RICHARDSON: We confess that we were probably 22 under -- having worked with Chet pretty closely at this, we 23 made the assumption that we weren't going to give much 24 background. But your point is well taken.
25 MR. BENDER: I'm not all that uncomfortable with
r 4
- 366 i
)
i 1 that myself.
1 i 2 MR. SIESS: Mike, what we should get to you is that i
3 discussion -- I forget what chapter it is of 9.56 that came 4 out of the containment loads working group, where they take I 5 the five or six plants and take the various scenarios and take 6 the various containment failure possibilities, and they build 7 up the scenarios very well in that. It is very well written.
I 8 And it will give you exactly the picture you want. And what i 9 is in 9.56 is the executive summary I think from that report, i
t
- 10 and it is very good and we'll get a copy of that for you 11 because I think it will help you a lot.
12 MR. COSTELLO: But your point is well taken, and it 13 might not hurt for us to start off the first 10 minutes with a j _-
14 perspective of that type. We will make note of that. Thank 15 you.
16 But as has been pointed out, in the large model e
17 tests we are proceeding -- oh, and by the way, I suppose the 1
4 1,
18 order in which the FIN activities are listed here are in
- 19 increasing order of expenditure. The analytical efforts are 20 relatively modest; the penetration experiments which are done 21 in separate effects are more costly, and the construction of 1
i 22 large models is most costly.
23 Going from the bottom up, the large model tests, it 24 became apparent that (1) pressure effects are the predominant j 25 factor leading to leakage with temperature effects being 1
367 1 clearly secondary as the conventional wisdom has it for steel f
2 containment and likely to be more important than steel but not 3 likely to be dominant in concrete containments.
4 So we have focused the large model tests on behavior 5 under increasing pressure up to total failure, and it attempts 6 to correlate how well state of the art predictive methods 7 could predict what is going to happen.
8 The penetration tests are focused on possibilities i
9 for generating local leakage, and here temperature, time of 10 temperature are likely to be very important. And 11 consequently, accident scenario temperature and pressure 12 histories are used in the penetration experiments.
13 MR. EBERSOLE: Let me ask you about the penetration 14 experience and then the findings that you might make. You 15 come to a conclusion pretty quick that the penetration ought 16 to have a very robust margin in current-carrying capacity l 17 relative to the conductors that enter it and leave it. That 18 when you started out, you should have designed it as not the 19 weak point but the strong point in the circuit. l 20 But you're going to find real quickly -- this is 21 just a random combination of this, but to go back to what it 1 22 ought to be might be worth considering, which says 23 electrically speaking, from the time / temperature standpoint 24 carrying overloads and current, the penetration ought to be
(
25 the best thing in the whole circuit.
a \
l 368 1 MR. COSTELLO: That's something we will convey to
)
)a 2 the electrical angineering research people when we talk to 3 them. But we are also focusing on the mechanical penetrations 4 which have great potential for leakage -- the larger diameter I
I 5 ones being the most likely. l l
6 MR. SIESS: Jim, you didn't mention that the large i 1
l 7 containment tests emphasized deformation measurements that j 8 would be input to the penetration tests.
9 MR. COSTELLO: That's correct.
10 MR. SIES3: Not just the fa' lure. l 11 MR. EBERSOLE: In that latter thing, mechanical 12 penetrations, I hear that flued fittings are getting out of 13 style now because you can't look at them. You know, the 14 double-walled flued fittings that fire the broken pipe 15 contents back into the containment? Is that true, that we 16 won't be using anymore flued fittings?
17 MR. COSTELLO: I can't speak to that.
18 MR. SIESS: Jesse, this whole thing is directed at 19 existing plants.
20 MR. EBERSOLE: Oh, they have flued fittings.
21 MR. BENDER: I just want to get a point clear in my 22 mind. The tests that you're doing are directed not to the 23 organic materials that provide the sealing devices for the
~, 24 electrical fittings, are they? They are directed to the 25 structural -- to the main structure?
369 i NR. COSTELLO: No. The effort on electrical 2 penetrations actually are full-sized tests. We're not talking 3 about these here today, but the electrical penetration tests 4 are actual full-sized electrical penetrations subjected to 5 these constructed severe accidents -- the histories.
6 MR. BENDER: Well, that's all right but I guess I am 7 pre-supposing something that may not be correct. But the 8 verification that the seals themselves have the capability to 9 withstand the structural loadings is done in some other way, 10 then, in these tests. That would be just incidental if they 11 failed here due to something, I would guess.
12 MR. COSTELLO: In the separate effects tests? No, 13 the pressure and temperature histories.
14 MR. BENDER: No, the organic materials that are used 15 to make sure that --
16 MR. SIESS: Are aged and subject to pressure and 17 temperature.
18 MR. BENDER: Are they pre-conditioned to establish 19 their historical --
20 MR. COSTELLO: Yes, sir. Temperature and radiation 21 aging, yes. And also, in some of the separate seal and gasket l l
22 materials tests that are being performed in this program, my 23 colleague, Hans Ashar will talk about it later on the agenda 24 but we will give you a little rundown on what's done.
25 MR. SIESS: We're not going to cover the electrical
a 370 1 penetrations?
(-S 2 MR. COSTELLO: No, sir, not today.
3 MR. SHEWMON: Well, the penetrations tests then 4 consisted of heating it up, putting pressure on it and seeing S if it will move; is that it?
6 MR. COSTELLO: The electrical penetrations test, 7 yes.
8 MR. SHEWMON: Is that what penetration means here?
9 MR. COSTELLO: No. Here we will be talking about 10 the efforts looking at the mechanical penetrations and 11 potential leakage.
12 MR. SIESS: These guys are not doing the electrical 13 penetration tests.
14 MR. SHEWMON: Well, they are doing structural 15 integrity of the containment and I'm trying to find out what 16 structures they're talking about. So there's a lot of 17 penetrations, and you're talking about doors and hatches more?
18 MR. SIESS: I think it might help if we would let 19 them tell us.
20 MR. SHEWMON: Well, we won't find it on what was 21 given out as a handout, so maybe he has a separate slide.
22 MR. SIESS: Well, he will explain it.
23 MR. COSTELLO: Again, corresponding to the agenda, 24 we will run through the activities which are currently going 25 on and will be going on in 1987, and give some historical
e 371 1 background on what went on in the past to try and put a 1
( 2 picture together.
(
3 MR. RICHARDSON: Can I interrupt because I think it 4 is such a simple thing. The containment, of course, is made 5 up of the shell and many types of penetrations. We have the 6 large mechanical penetrations such as airlocks and equipment 7 hatches and large pipe penetrations with bellows. There are 8 valves on those pipes that represent penetrations and 9 electrical penetrations. Those are the four classes of 10 penetrations of the containment.
11 Within this program, we are evaluating the large 12 equipment hatches, airlocks and pipe bellows. In two other 13 programs, in the equipment qualification area we are looking 14 at the electrical penetration and the valves. So what you 15 will not hear today is the leak integrity investigation of 16 electrical penetrations and valves. What you will hear today 17 are hatches and pipe penetrations.
I 18 MR. SHEWMON: Thank you.
19 MR. COSTELLO: And in the order of the presentations 20 that we have today we will first talk about the planned 21 concrete model tests that my colleague, Hans Ashar, will now 22 speak to. And then I will be back to talk to you a bit about 23 what went on in the large steel model series, and summarize
_s 24 that and what we learned in the way of pre-test analysis, s
25 post-test evaluation.
l
e 372 1 Then I guess -- and the problem of making models.
1
(~'sj 2 And Bob Kassawara is not here and I would leave it up to y o u_, i 3 Mr. Siess, to decide how would you best like us to speak to, 4 or when you would like us to speak to EPRI's undertakings.
5 MR. SIESS: I think, Jim, about all we can do there 6 is to have -- is for you to assume that I know what they are 7 doing and not take the time to try to explain to everybody 8 else because I think that's too much of a burden to put on 9 you. And to take a few minutes to tell me what you think it 10 will help you do. Okay?
11 MR. COSTELLO: And I will speak to that when I come 12 back to talk to you again about pre- and post-test analyses,
(~N 13 in the next go-around.
i i 14 Then we will pass over the EPRI part of the agenda, 4
15 and then my colleague, Hans Ashar, will be back to talk to you 16 about what is ongoing and planned in the efforts on 17 penetration experiments, and then I will finish up by talking 13 about what we sense as the likely scope of the effort on the 19 seismic question and ask for your advice on how -- any input 20 you may to help in proceeding in 1987 and beyond on the 21 seismic question.
22 MR. SIESS: Jim, this is an area where I am 23 particularly unhappy that EPRI is not here because one thing 24 that concerns me is where their containment program, which is 25 all concrete containments, fits in, in the event that your
. . - - - - - _ = -. . . ._ _ __ _
373 1 model test has to be deferred because of budgetary reasons.
) 2 Because that then becomes the only work on concrete 3 containments.
4 MR. COSTELLO: Well, if I can defer it I will speak 5 to that when we get into the pre and post-tests. Okay.
6 MR. SIESS: Well, I think that was particularly i
7 important because that may be all that gets done on concrete 8 containments for a couple of years if the budget gets whacked.
9 MR. COSTELLO: Thank you. I guess with that in 10 mind, then, I guess I will have Hans Ashar come and talk 11 about the concrete model experiments for which the initial 12 construction has begun on the model right now.
/' 13 MR. ASHAR: The first part of this presentation will
(
14 be just a boilerplate format in which I will talk about the
/
15 overview and budget requirements in a very generic way; what 16 is the issue, objective and integration, et cetera.
17 In the second part I will show you some of the 18 slides on what is going on in the reinforced concrete model 19 testing.
20 [ Slide.]
21 The contract is with Sandia National Lab and the 22 budget requirement in 1985 going on, there was $2,060,000.
- 23 For FY86 we have earmarked about $2 million. For 1987 we are 24 estimating about $2,400,000, and in that one there will be 25 completion of reinforced concrete model as well as some of the
-- - - ._- = .-. . - - - - -
374 1 needed separate effects test, such as we are thinking it all 2 depends on how the evaluation by Sandia turns out as to the 3 temperature effect on concrete containment. If needed, we 4 might perform some separate effects tests for temperature on 5 liner-concrete interaction.
6 And another thing that is an issue is: is there a 7 different behavior of pre-stressed versus the reinforced.
8 This is the one which we'll be talking of in the near future 9 and work out if we need any separate effects to get the idea 10 about the behavior of pte-stressed, if it is different from 11 the reinforced concrete containment for overall pressure i
12 capacity.
13 [ Slide.)
\
l 14 MR. RODABAUGH: Are you going to be using air as the 1
15 test pressure medium? Are you going to be using for test 16 pressure?
17 MR. ASHAR: For test pressure?
18 MR. COSTELLO: No. Nitrogen.
19 MR. ASHAR: Nitrogen will be the medium.
20 MR. SIESS: Why nitrogen instead of air?
21 MR. COSTELLO: The principal reason is it turns out 22 to be cheaper. Considerable effort was done in scoping it out 4
23 back when the initial steel model tests were done. The 24 , question of the size of pump you needed and the power 25 requirements when compared against the easy availability of
e 375 1 nitrogen with oil field trucks and the like pointed to 2 nitrogen.
3 MR. SIESS: And the leakage through e crack would be 4 or would cost you about the same?
- 5 MR. COSTELLO
- There was an old scoping calculation I 6 and it was felt that with the level of accuracy we're going to 1
7 get, that would not be a problem.
I 8 MR. ASHAR: Okay. The issue is well known to you.
9 You have discussed it here for about 45 minutes, which is to i 10 have some kind of idea as to the analytical prediction of 11 containment failure modes by experimental means. We can find 12 out as to what would be the failure capacities of various
'~} 13 types of containments. That is the aim. For steel 14 containment we know we learned a lot, and it will be 15 transferred.
16 The objective -- perform experiments on containment 17 models that will prcduce failure data under extreme loading 18 conditions and permit an evaluation of the capability of 19 state-of-the-art calculational methods.
20 As it was discussed, the pressure that we are 21 putting in is a steady pressure at a particular rate of 22 pressurization. It may not completely simulate pressure l
23 during the severe accident, but analyses and evaluations as to 24 the effect on the containment itself show that it won't make i \ss# 25 any difference at what rate we pressurize the containment.
l l
, _ . . . _ . . - . . _ . . _ . . _ _ . _ _ _ _ . _ _ _ - . _ . _ . _ ___,,_.__________._,_.__-,.c._- _ . _ . , _ . _ _ _ , _ _ , _ _ _ . . _ , _ - , _ _ _ _ _ _.
376 1 The temperature effect will be separately evaluated for steel
( 2 as well as concrete containment, if there is any temperature 3 effect.
4 MR. RODABAUGH: Do you anticipate building the 5 pressure up and then dropping it to make inspections?
6 MR. ASHAR: For reinforced concrete containment the 1
7 way the pressurization will be done is first it will be done
, 8 at peak calculated pressure. That means, containment design 9 pressure. It will be stopped and the readings will be taken 10 for the deformations, et cetera. Then it will be stopped 11 again at structural integrated test pressure, which is about 12 1 15 the design pressure.
13 MR. RODABAUGH: My question was are you ever going 14 to drop the pressure before you get to your final pressure 15 stage?
16 MR. COSTELLO: Not intentionally.
17 MR. SIESS: Hans, what did you say about 18 temperature?
l 19 MR. ASHAR: For reinforced concrete containments? l 20 What I said was you might be aware of the discussion on 21 including the temperature into the model, but what has l
22 happened is that there are more problems with including 23 temperature than deriving any benefit we can get out of it.
24 So we have decided is that Sandia is going to make 25 an evaluation based on the expert advice on temperature
]
377 1 effects on concrete containment. They are going to come out
) 2 with a recommendation as to do we need any separate effects 3 test to see that we can superimpose some of the effects in the 4 discontinuity areas.
5 MR. SIESS: The effect you're considering is that 6 the increased temperature puts the liner in compression.
7 MR. ASHAR: That's correct. But in the 8 discontinuity area that effect might be a little different 9 than in some areas. This is what Sandia is evaluating, and 10 they have a letter report to us --
11 MR. SIESS: I don't care what they come up with; 12 you're going to need separate effects test, I think.
13 MR. ASHAR: Probably you are right.
b 14 MR. SIESS: It will depend somewhat on how it fails, 15 you know.
t 16 MR. RODABAUGH: When I run tests with nitrogen 17 coming out of compressed nitrogen bottles I ended up with 18 quite a cooling effect. Have you worried any about -- that 19 you may be running your tests at lower temperature than you 20 want tot 21 MR. COSTELLO: Well, on the steel tests that was 22 compensated for to some extent by the fact that we actually 23 had heaters in the containment, primari,1y to try and keep 24 constant temperature during the diurnal cycle. And that had 25 the effect also of --
e 378 1 MR. RODABAUGH: Okay, very good.
(*)
(, 2 MR. ASHAR: The integration. As pointed out by Jim, 3 this project has been integrated together with the containment 4 penetration as well as the analytical pretest analysis and 5 post-stress analysis, as well as with the wall testing program 6 and the electrical penetration program.
7 In 1985 what we have done is the facility, as you 8 all might have seen, the facility at Sandia National Lab, 9 probably, tests the failure of three small steam models, and 10 in '85 the large steel model also was tested, as you must be 11 aware. Prior to 1985 was the site selection and three small 12 steel model testing, and in '85 we completed the large steel
[ 13 model, one-eighth scale steel model testing, and completed 14 design and started construction of the concrete model.
15 In '86 what we have planned is the complete 16 construction and instrumentation of the concrete model and 17 initiating testing. In 1987, complete testing to failure of 18 concrete model, performing necessary supplementary tasks, what 19 I was talking about today, the separate effects test; initiate I
20 experiments in seismic performance, that is, after the 21 planning stages are over in that area. That is what Jim will 22 talk about a little more later.
23 Regulatory use. Confirmatory assessment of severe 24 accident policy statement that the containments are safe, n
v 25 probably, and the basis for possible addition of containment
i e
379 l l
1 performance requirements for safety goal implementation. )
2 Now, the practical usage has been made in quite a 3 number of areas, like in preparing NUREG-1037 for containment 4 performance working group. They have taken the analyses done 5 by Argonne National Lab on penetration, and they are finding 6 out if their penetrations are the weak points at Sequoyah or 7 at Brown's Ferry or at Peach Bottom, whichever containments 8 they were looking into.
9 So they made extensive use of the survey and of the 10 analyses of Argonne and Sandia National Lab. Also in the j 11 regulatory use, in addition to this very generic regulatory 12 use, there will be code development as part of the computer l i
13 code developments. These will be useful to a lot of people in 14 conforming their nonlinear codes or to make sure that they 15 modify their codes to suit certain data being available in ,
i 16 this particular test.
17 MR. SIESS: Hans, are the analyses, the pretest ,
18 analyses, far enough along on this to give you a prediction of 19 what will happen? ,
i 20 MR. COSTELLO: I guess I could speak to that. There l l
21 has been a modest effort at Sandia to do at least an 22 axisymmetric modeling of the specimen to help in deciding j I
23 where to put instrumentation, and the axisymmetric modeling 24 again points to the basemat-wall intersection as an area of I
25 exceptional interest.
e 380 1 MR. SIESS: All right. That's where the concrete O)
( 2 will be most highly stressed. Is it predicted that the liner 3 will fail at that location?
4 MR. COSTELLO: What happens in this axisymmetrio 5 modeling at the level of detail that's done is the same that 6 has happened in other models of the same ilk, which is that 7 the local loss of stiffness leads to numerical instability in 8 the calculation, and the argument is whether that is a 9 numerical concrete modeling problem or it actually is 10 reflective of a large deformation in concrete liners.
11 The answer is the level of analysis done so far is 12 indeterminant on that point.
13 MR. SIESS: Has anybody ever considered getting away 14 from a finite element analysis and trying to make some 15 calculations using statics?
16 MR. COSTELLO: Hand calculations? I believe we may 17 have actually --
18 MR. SIESS: Nobody knows how anymore.
19 MR. COSTELLO: We may have actually come to that.
20 MR. SIESS: Well, once I get to a certain point. I 21 could model this as a little angle here and then I am pushing 22 it apart.
23 MR. RICHARDSON: Do they teach those simple 1
24 techniques in college anymore?
25 MR. SIESS: I don't know whether they teach anybody l
381 1 any more. We do because we don't use computers that much. We y
\
) 2 think people ought to know how to do things that computers are 3 doing.
4 MR. ASHAR: Now I will concentrate on some of the 5 details of the reinforced concrete model testing. First, let 6 me give you the status of where we are in this particular 7 area. As Jim Richardson pointed out before, we have stopped i
the work in between and restarted again after Mr. Minogue gave 8
i 9 us the go-ahead. So there was a slight delay, but it appears 10 that there won't be a schedule delay because of that l
11 Now, at this time they have already broken the 12 ground. They are all ready to put the mud mat for the basemat
[~* 13 and are ready to make the fabrication for reinforcing for the 14 mat That's where they are. They expect to start putting the 15 reinforcing for the mat sometime in October. That's what l
16 I heard last from Sandia. So that's where we are.
17 But the one-sixth scale reinforced concrete 18 containment, the contract was given to United Engineers in 19 design and construction of the model United will work with 20 two subcontractors, Chicago Bridge and Iron for the liner 21 fabrication and supplying the penetrations, and Wiss, Janney &
22 Elstner will be doing some modeling and testing of the l
23 materials.
i 24 MR. RICHARDSON: Is there any problem with pouring i - O 25 concrete in the winter out in that desert?
a 382 1 MR. SIESS: None that they couldn't overcome. They 2 could keep it heated.
3 MR. ASHAR: Well, there might be some, but I think 4 there are enough precautions taken by Wiss, Janney in that 5 area to make cold weather concreting a possibility. I don't 6 think there should be a problem. Maybe they may have to delay 7 some work once in a week or something like that, but there 8 won't be a break, as far as I understand it.
9 MR. SHEWMON: A different question. I am confused, 1
10 maybe, by the use of your word "model." Now, do you recall --
11 I take it you use the word "model" to describe the one-sixth 12 scale reinforced concrete containment. The word "modeling" is
/~N 13 often used to try to write a computer program or do
(% ))
14 something. So what this Janney & Elstner did was to design 15 the one-sixth scale structure; is that right?
16 MR. ASHAR: Yes. Modeling includes quite a number of 17 things, like making an assessment as to the size of the liner.
18 MR. SIESS: Excuse me. I think scaling would be a 19 better word. And this was designed as a small containment. It 20 was not scaled down from a full-scale design. They were told 21 to design a containment of this size using the code. They i 22 didn't try to design something with No. 18 bars and scale 23 everything down, but to keep things about right, an'd they used 24 Wiss, Janney for scaling questions.
(
25 M F. . BURNS: In particular, it does not represent any
e 383 1 single containment in existence.
f 2 MR. ASHAR: But it is a containment as such.
3 MR. SIESS: It is typical and will be analyzed.
4 MR. COSTELLO: I can't let the question about the 5 weather in the winter go. The answer is fundamentally it 6 depends an awful lot on the winter as to how much time might 7 be lost.
8 MR. ASHAR: The model is designed to the Code, 9 Section III, Division 1 and 2, and there won't be a 10 requirement for an N-Stamp on this one.
11 MR. MARK: How many pounds extra would you expect if 12 you did have an N-Stamp?
13 MR. ASHAR: Well, as a matter of fact, at this time 14 NRC does not require an N-Stamp on regular containments even.
15 MR. SIESS: No, this thing would not be N-stamped.
16 [ Slide.] l 17 MR. ASHAR: The work with United Engineers has been 18 divided into two phases. The first phase includes development 19 of quailty assurance program, plans for support tests for 20 concrete using reinforced steel and splices, et cetera, 21 conduct the support tests, design the model, prepare 22 fabrication drawings, report the results of the prefabrication 23 test, and summarize the design work to Sandia in design review 24 meetings.
25 The first one is completed and all the drawings are i
l l
a j '
384 1 available now. W e a r'e in Phase II right now, revise the
) 2 design drawings and. supporting specifications. I understand 3 that has been done. Sandia has reviewed the comments. Site t
4 preparation has been done. Fabricate the model -- that will be 5 done. Perform the post-fabrication support tests -- that will 6 be done in fiscal year 1986.
3 7 [ Slide.]
B[ These are some of the features of the model 9 testing. Its design accident pressure will be 46 psig. There 10 will be two equipment hatches with three doors -- I think I 11 have a slide to explain the three doors -- two personnel 12 airlock representations, and two constrained penetrations. I
/"N '13 will be explaining that to you later.
14 , 1/16th inch and 1/12th inch steel liner with i t
[ 15 studs. The 1/16th is for the cylinder part of the containment 16 model, and the 1/12th inch will'be'for the dome area. That is
/ 17 1 why there are two sizes of steel, liner. Piping penetrations 18 will be there, simulated'p'iping penetrations, and layers of 19 major reinforcing just to simulate what would happen to
\
20 various layers in the cracking because of the layering.
21 MR. RODABAUGH: What kind of material will that *
[ 22 1/16th inch steel lining he? ,
23 MR. ASHAR: I think I have something on it later 24 that I will come to.
25 MR. R O D A B A'J G H : Thank you.
' 385 1 [ Slide.3
( 2 MR. ASHAR: The equipment hatches are two 40-inch 3 diameter hatches. One is centered in the cylinder wall. It 4 has two hatch covers. One is inside and one is outside. In 5 the initial testing, up to a certain level, line 1,15, the 6 design pressure, it will be used -- the inside cover will not 7 be in. The outside will be in. So that we can know that 3
8 unseating of the hatch cover, what will be done, will it leak, 9 will anything happen to the seals and gaskets and that kind of 10 thing.
11 MR. SIESS: The outside door will open out?
12 MR. ASHAR: We will try to open it. Whether it will 13 open or not, we don't know. It is a pressure unseating.
14 MR. SIESS: Which is not the way we designed them.
15 MR. ASHAR: That's correct, yes. This is just to 16 understand the behavior.
17 MR. SIESS: You are looking at seals, though, 18 drywell seals, to see if drywell seals do open up. Drywell 19 hatches, if I'm not mistaken, do open up.
20 MR. ASHAR: Unseating hatches uses a double-tongue 21 and groove seal geometry. Pressure seating hatch uses a 22 double gumdrop seal geometry. These are used in regular 23 nuclear power plants. The other hatch projects out from the 24 containment has one pressure seating cover and uses double
\~' 25 gumdrop seal geometry.
e 386 4
1 MR. SIESS: These are not intended to be typical 2 equipment hatches; they are intended to test doors.
3 MR. ASHAR: They test doors. Airlocks, all 4 equipment hatches are not typical, but the equipment hatch is 5 more or less simulated.
j 6 [ Slide.]
3 f
i 7 The steel liner. It will be A414 grade D that will 8 be used. Generally what is being used on concrete j 9 containments is A516 steel, but A516 cannot be available in
! 10 these sizes, so they had to make some tests on the A414 grade, l 11 and they did testing to see if the properties of A516 and A414 12 are approximately the same as far as the use strength and ,
13 ultimate strength. They have been reviewed by the expert 1
14 committees, and it looks to be a promising one. It almost 15 simulates the steel liner --
l i 16 MR, SIESS: Does it have a flat yield?
17 MR. ASHAR: No, it is quite a curving yield. It 18 isn't flat.
19 MR. SIESS: What is the ultimate strength?
20 MR. ASHAR: The ultimate strength is in the area --
21 guaranteed ultimate strength is about 60 ksi 22 MR. SIESS: No, strain.
- 23 MR. ASHAR
- Oh, strain? Twenty percent.
J 24 MR. SIESS: Sc that compares pretty well with the 25 liners -- l t l t
1
-.m.e -- ,, , - -r-y -------,%.%%,m-w. ,- m-_,--y m w .-- - ,v % www, .gr _ , , , y m % y ,w.m , , e ym-wwww,w,m--.--,--r--,.-ve-ww,%=
387 1 MR. ASHAR: Yes, that's correct O() 2 MR. SIESS: And the typical liners don't have a flat 3 yield either?
4 MR. ASHAR: No.
5 MR. SHEWMON: A curvy yield means there is a yield 6 drop?
7 MR. ASHAR: It is a variation. It is some kind of 8 variable plateau, and you weigh the yield strength at that .1 9 percent or .05 percent, what ever is defined as the yield 10 point. 1/16th inch thick in the cylinder region, 1/12th inch 11 thick in the hemispherical dome, thickened in discontinuity 12 regions as required, and 3/4 inch and 1/2 inch long studs will 13 be attached to the liner as designed in free field, one to one 14 replacement near the penetrations.
15 Now, on start, I think there is a linear deviation 16 from the scaling. In the general shell area, the spacing of 17 the studs, attachment into the concrete liner, attachment to n
18 the concrete, is more spaced than what would be in the typical 19 containment, but in the area of discontinuity like hatches and 20 piping, it has been one to one scaling ratio.
21 There has been quite a long discussion on this as to 22 whether we should spend money on buying hundreds of studs. It 23 was decided by the peer review committee and all the other 24 experts involved that it is not necessary to have one to one A 25 stud pacing for model testing, because we are doing it for
a 388 1 pressure and not for seismic or anything.
( ) 2 MR. SIESS: And you don't have a temperature?
3 MR. ASHAR: We don't have temperature in this.
4 MR. SIESS: The original stud spacing was based on 5 buckling between studs?
6 MR. ASHAR: Between the studs, and under pressure 7 only, there is no question of buckling as such, but in the 8 area of penetrations there is some need for it, so you have 9 the one to one ratio in that area.
10 MR. SIESS: And the one to one has nothing to de 11 with the studs?
12 MR. ASHAR: No.
/h 13 MR. SIESS: It's just confusing there.
's 14 MR. ASHAR: Yes. It means that it is simulated 15 correctly in that area.
16 [ Slide.]
17 Reinforcing steel ASTM, A615, Grade 60. That is 18 typically used in concrete containments. No. 4's are the i
19 major reinforcing which are available. Four layers of hoop 20 steel, two layers of meridonial steel, two layers of diagonal 21 steel. Diagonal steel will be for the shear reinforcing.
22 Simulator seismic reinforcing, really. Nothing more than that.
23 MR. RODABAUGH: What kind of a space do you have 24 between the liner and the concrete?
\ 25 MR. SIESS: None.
l l
389 1 MR. RODABAUGH: None at all.
2 MR. ASHAR: They are attached to each other by 3 studs. Studs are connected to the liner.
4 MR. SIESS: There might be a space, but it isn't 5 deliberate.
6 MR. RODABAUGH: Okay, thank you.
7 MR. SIESS- It is uniform.
8 MR. ASHAR: Additional layers of discontinuity 9 regions, 6 millimeter, No. 3, No. 5 and No. 6 bars will be 10 used, reinforcing bar splices, 4500 cold swaged bar splices, 11 and 50 modified cadwelds. Now, there are some problems l 12 because what was happening was, generally in the typical i [~'\ 13 containment, the cadwelds are No. 18 bars. No. 18 bar l
V 14 cadwelds slipped about yield strength or something, and No. 4 i 15 bar cadwelds don't slip. This was being discussed at length I
16 in our peer review meeting, and they have tried to define a l
17 cadweld that would slip for No. 4 bars. That is another way 18 of getting around it.
19 MR. SIESS: Lots of luck. I have tried it with 20 bigger ones.
21 MR. ASHAR: And the last I heard on this one was 22 that they could slip it but then the ultimate capacity was 23 reduced, and that we cannot tolerate. So probably we might 24 use a few cadwelds for No. 4 bars, but without simulating the O 25 slip for the No. 18 bars.
I
a 390 1 MR. SIESS: I think you could do that with separate 2 effects tests if you had to.
3 MR. ASHAR: That has been done. That has been done.
4 MR. SIESS: Are the small bars likely to end up a 5 little stronDer than the No. 8, and that you could take care 6 of in your analysis?
7 MR. ASHAR: Yes, that is the intent, correct.
8 [ Slide.)
9 Now, personnel airlocks. There are two of them.
10 Both have a 1 foot, 8 inch diameter. One is used as an escape 11 lock, which is not representative of an actual airlock because 12 of the bulkhead door openings and the seals and everything. It 13 is not simulating the actual airlock. The other has two
\
14 bulkheads with a typical stiffener pattern.
15 As far as the shell penetration interaction, it 16 simulates very well with the actual airlocks, but as far as 17 the detail of the envelope itself, of the airlock itself, it 18 still is not the right one because of the diameter.
19 MR. SIESS: As far as the structure is concerned, it 20 is seeing a typical airlock, but as far as the doors are 21 concerned, you are testing doors.
l 22 MR. ASHAR: That's correct.
23 Now, there are no operable airlock doors on that 24 one, so we are going to have a separate test on airlocks, 25 which I will talk about in the penetration presentation.
- - - . . _ - - . = . . . . . . - - _ _ - . .- . _ _ _ . - - _ - _ . _ _ _ _ _ __ _ - - _ _
l i
391 1 [ Slide.]
2 Constrained penetrations. In our testing of the I 3 steel containment, the one-inch steel containment model, we I
- 4 also had one constrained penetration, and the results of that 1
5 particular testing were quite interesting. We are going to do 4-6 the same kind of testing, a little different than what we had 7 in the steel containment, but the purpose of this one was to 4 8 represent the constraint placed on the containment by major 4
9 piping penetrations.
10 Basically, the supports for the main steamline and 11 feedwater lines, et cetera are sometimes very near to the 12 containment face, and in order to simulate how much restraint i
! 13 will be given by them -- this is only one of the parameters.
14 There might be much more. But this is to simulate that kind 15 of a constraint to the piping penetration as to will it yield '
i 16 faster, will it punch through. This is the thing we are 17 trying to figure out.
18 MR. SIESS: Now, in the steel test, that didn't have
)
j 19 any effect, did it?
l 20 MR. ASHAR: As far as we know, there was no effect 21 because of the constraint placed, but the results around that 22 penetration are very interesting. That is why the strained 23 configuration.
1 1 4
24 MR. SIESS: Did the tie yield? )
< l 1
) 25 MR. ASHAR: No. It was not supposed to yield and it
392 i
k i didn't yield.
2 MR. SIESS: It wasn't a weak spot in the 3 containment, ,
4 [ Slide.3 l
5 MR. ASHAR: The final drawing shows you how the 6 model looks. The concrete model will have --
7 MR. SIESS: Where is the ground surface?
l l
l 8 MR. ASHAR: The ground surface is here [ indicating).
9 MR. SIESS: I just wondered if you had any seismic 10 instrumentation on this in case they had an earthquake.
11 [ Laughter.3 12 MR. ASHAR: No, there is no provision for that, but j
(
13 one foot diameter is not correct now because after the 14 preliminary analysis that was done between the wall and the I
( 15 mat, there was absolutely no convergence of the finite element 16 analyses they are performing.
17 MR. SIESS: And you have got an interior concrete on 18 top of the base? i I
i 19 MR. ASHAR: Yes. There is going to be fill concrete l i
l 20 on top of the base, so this dimension is going to be 40 inches 21 now, 3 feet 4 inches, am I correct, Jim?
22 MR. COSTELLO: Yes.
23 MR. SIESS: Does that line knuckle the same all the 24 way around?
25 MR. ASHAR. You mean this one here? Yes.
t
e
- 393 1 MR. SIESS
- Is any thought given to changing that
( j 2 around the periphery? They are different details if I'm not 3 mistaken. Am I right?
4 MR. ASHAR: Well, there has been a code detail used 5 on the sloping bottom part, but that is not dissimilar. All 6 the recent reinforced concrete containments have a straight 7 face, so --
8 MR. SIESS: I'm talking about the details of the 9 steel liner. If that is the weak spot in the concrete and the 10 liner sees large deformation, that might be where the liner 11 ruptures if it does. And if it is, it is going to be a 12 function of a detail, I would think.
13 MR. RODABAUGH: I would think so, too, and I wonder 14 if there is a weld there.
15 MR. SIESS: There is a weld there, but that could be 16 had by separate effects test. But I was thinking would it have 17 been possible to take three arcs around and put a different i l
18 detail in each one.
19 MR. COSTELLO: Frankly, we haven't thought of that.
20 MR. SIESS: At this stage I'm not about to suggest 21 it.
22 MR. COSTELLO: We hadn't thought about that.
23 MR. ASHAR: What happened was this detail was a very 24 typical one being used for the recent reinforced concrete
,0 25 containment, like Seabrook has the same detail
e 394 l l
l 1 MR. SIESS: After what happened to the steel one, I
( 2 wouldn't use the word " typical" on anything.
3 MR. ASHAR: That is true.
4 MR. RODABAUGH: This is 1/16th inch steel liner, a 5 16th-inch thick?
6 MR. ASHAR: Yes. 1/16th inch.
7 MR. RODABAUGH: You are going to have the problem of 8 welding 1/16th inch steel in a corner weld? Is that just a 9 corner weld at that critical point here?
10 MR. ASHAR: Yes 11 MR. SIESS: Is it just this?
12 MR. ASHAR: The way we are planning it is that up to
/ 13 some point here, this part will be welded outside -- on the 14 fill, and then it will be placed on the site. This part will 15 also be done in the same way, so that there is no grouting.
16 It will be --
17 MR. SHEWMON: But the question is you have to match 18 this up and it is very easy to burn through.
19 MR. ASHAR: That's correct, and CBI has developed a 20 special procedure for the T liner welding. We hope it does 21 simulate the way it should be done.
22 MR. SIESS: Those steel models, how thick were the 23 small-scale steel models?
24 MR. COSTELLO: The small-scale steels were .056. )
i
~~ 25 MR. SIESS: And they welded those.
l l
1
_ _ _ ____ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ . _ _ ______..______________J
a 395 1 MR. RICHARDSON: Is this just a fillet weld in the 2 corner?
3 MR. ASHAR: No, it will be a groove weld.
4 MR. SIESS: This may be better than the typical ones 5 by the time they get through.
6 MR. ASHAR: This is a diameter of 32 feet and a 7 height up to the spring line of 32 feet, 3 inches, and it is a 8 hemispherical dome.
9 MR. SIESS: Can I ask you a question about material 10 properties? Do you know the NDT of the liner material? Is 11 that a test that will be made?
12 MR. ASHAR: Yes. The test was going to be conducted 13 on it to make sure that it is as per the ASME Code.
\
14 MR. SIESS: Do you know the NDT of the rebar?
15 MR. ASHAR: No, we don't know the NDT of the rebar, 16 either in this model or in actual prototype containment.
17 MR. SIESS: I know what it is in ordinary rebar.
18 It's about the temperature in here.
19 MR. ASHAR: Right.
20 MR. SHEWMON: You don't do a Sharpy bar on 1/16th 21 inch steel.
22 MR. ASHAR: They are going to do some energy tests.
23 Whether it's Sharpy or drop rate or what, I have no idea at 24 this time, but I will let you know if you are interested.
/ 25 MR. SIESS: I would save some robar just in case you w -
e a 396 1 wanted to know what the NDT is later, (m) 2 MR. SHEWMON: I rebar you could do it, but none of 3 those tests are done on 16th-inch sheet. They are all thick 4 plate.
5 MR. ASHAR: That's right. They are at least 4
6 half-inch specimens, yes. They plan to do -- I have not heard 7 about the results and how they have done it, but I will find 8 out for you.
9 MR SHEWMON: Chet, the other thing is there is no 10 Sharpy test on stuff thinner than half-inch or something 1 11 because you can't get the thickness to make it relevant. It 12 is all shear lift.
{}
\/
13 MR. ASHAR: Well, this is about the end of my 14 presentation. If you have any question on details, maybe I 15 have some slides I can show you.
16 MR. SIESS: And right now they have poured the mud 17 mat.
18 MR. ASHAR: Yes, they are ready to pour the mud mat.
I 19 MR. RICHARDSON: How many channels of 20 instrumentation do you anticipate?
21 MR. ASHAR: About 1500. There are more than in the 22 steel model. The steel model had about 1100 or so.
23 MR. COSTELLO: About 7. We do have a graph where we 24 have some of the --
%- 25 MR. RICHARDSON: I was just after a rough order of
e i 397 1 magnitude.
2 MR. COSTELLO: Around 1100, I think.
3 MR. SIESS: Let me ask a question. The reason for 4 using a gas to test this -- of course, there are two reasons.
5 One is you are interested in leakage and it is leakage of a 6 gas, and the other is that the mode of ultimate rupture will e
7 be, presumably, different with a gas than with a liquid. But 8 is it conceivable that you could get far enough in a gas test 9 that the thing -- if it leaked, you would have called it a 10 leakage failure and you could back off and make a final test 11 with a liquid where you could observe things more easily.
12 MR. COSTELLO: I think that's possible if that is 13 the outcome. If we do get a small stabilized tear in the liner 14 which leads to depressurisation and inability to add the 15 pressure in the gas, then we are in mode 2 of what to do 16 next.
17 MR. SIESS: Would it be possible? I mean are your i
18 gauges protected inside well enough that you could fill it 19 with water?
26 MR. COSTELLO: I think with a bladder.
21 MR. SIESS: That would be pretty expensive. You 22 know, a Canadian test was made with water and a bladder, and 1
23 they learned quite a bit from that, but I'm not sure they i 24 couldn't have learned it with air, but they couldn't have used i
25 air at their site. You know, the pieces would be a lot easier I
i l
l l
I 398 1 to find.
) 2 MR. COSTELLO: But I think, you know, we haven't 3 gotten down to Phase II yet, but our thought in the steel 4 model was to bring it up to pressure failure and then see
- 5 where we were about the possibility of maybe finding the next 6 failure mode up by repairing and raising pressure again One 7 of the thoughts there was -- you know, what you might find out 8 depends an awful lot on how far you are out in strain.
9 MR. SIESS: Well, it is just a thought.
10 MR. EBERSOLE: May I ask a question? Chet, you said 11 the Canadians did this with a bladder?
12 MR. SIESS: They build a model of prestressed A
13 concrete, a can-do type containment. They don't line their J
14 containments with steel. They had a rubber liner in it.
15 MR. EBERSOLE: Well, can't you do a test like this 16 with a mixed environment inside which is mostly filled with 17 water or with x percent in a pneumatio ball and diminish the 18 terrible consequences but still get --
19 MR. SIESS: The tests have to be with gas because 20 they are interested in leakage. It's a gas that is leaking.
21 I was just trying to find some way to avoid the explosive type 22 -- where you can't see what happens and you just hope your i
l
( 23 cameras and your strain gauges catch enough information.
I i>
24 This tends to be a one-shot test You spend a lot lO
! 25 of money on it and you would like to milk it for all it's
_ _ - . - _ _ - . - . - . - . _ _ . - . _ . _ _ _ _ _ _ _ _ _ . - _ _ _ _ _ ~ _ _ . . . _ _ _ _ _ . _ _ _ . , _ _ . _ . - ~ _ . - - - _ . _ - -
399 1 worth.
2 MR. EBERSOLE: Well, for a strength test, you could 3 fill it 99 percent water and have the rest pneumatic.
4 MR. RODABAUGH: The problem with water, of course, 5 is any inside instrumentation like strain gauges --
6 MR. SIESS: Yes, if they are not waterproofed, and 7 that is quite a job. At that stage you may not be interested 8 in strains anymore.
9 MR. RODABAUGH: It's very possible, but it might be 10 worth keeping in mind.
11 MR. SIESS: I mean you expect it to do more with the 12 steel 13 MR. COSTELLO: Another possible desirable use for 14 this model is we do get a localized tear in the liner and a 15 leakage through the concrete would be the same as you might 16 have used for the steel model as a vehicle for some nice 17 thermalhydraulics test in a large volume.
18 MR. BURNS: Actually, I would hope this would not be 19 a catastrophic failure but that we could use it for a test 20 vehicle for testing other things, like other forms of 21 penetration and such.
22 MR. EBERSOLE: What is the time interval for a 23 catastrophic failure? How many milliseconds are we talking 24 about? ,
1
' I have a little something 25 MR. COSTELLO: Very few.
l
e 400 1 to show you here later.
) 2 MR. EBERSOLE: Not time enough to detonate a member, 3 though, or anything like that.
4 MR. SIESS: I don't think this one can go the same 5 way, but then --
6 MR. RODABAUGH: Chet, for my information, what 7 happened to the steel model?
8 MR. SIESS: They are going to show you. I think that 9 is waht we have here up on the film.
10 MR. COSTELLO: We have a little summary --
11 MR. SIESS: Is that the failure test?
12 MR. COSTELLO: Yes, sir.
13 MR. SIESS: Okay. We will wait. They couldn't get 14 all the pieces in one photograph, I will tell you that.
15 MR. RODABAUGH: It failed under gas pressure. How 16 big was it? About the same size as the concrete one?
17 MR. COSTELLO: Just a little bit smaller, about a 18 one-eighth size.
19 MR. RODABAUGH: I would guess you have got pieces 20 around a radius of a mile, more or less. !
21 MR. SIESS: No, not quite.
22 MR. COSTELLO: No, thank God.
23 MR. SIESS: No, they weren't that far away from it.
24 A mile away would have put them over in the -- what is that 25 mountain? ,
1 l
l l
e 401 4
1 MR. RODABAUGH: There is a fair number of industrial
( 2 accidents where a tank had a gas pressure break off and could
! 3 go a mile.
4 MR. BURNS: If the pieces went a mile, I doubt if 5 Sandia would allow us to test this particular model at this i
6 site. l 7 MR. SIESS: This was at an isolated site, but not 1
8 that isolated.
9 MR. RODABAUGH: Battelle had a site they thought was 10 isolated. They broke windows within three miles.
11 MR. SIESS: Gentlemen, this is the time for a break, 12 and then what are we going to hear, the steel one?
13 MR. COSTELLO: After the break, I will talk about 14 pretest and post-test analyses, primarily with emphasis on the 15 steel model series, and then some lead-in into the concrete.
16 (Recess.3 17 MR. SIESS: All right.
18 MR. COSTELLO: James Costello from NRC Staff again.
19 I would like to start talking about the pretest 20 predictions and post-test analyses and attempts to make sense i 21 of it all, which activity we have more or less concentrated ,
1 1
22 into a single FIN, and having the other two FINS that Hans 23 Ashar was talking about concentrating on trying to get the 1
24 experiments built and done, which is the major undertaking 25 that we have.
e 402 1 1 would like to give some indication about some of
( 2 the cooperating organizations involved in pre and post-test 3 analyses. We have had with XFX in Germany some collaboration 4 on the steel model. We had somewhat hoped that they might be 5 inclined to try to predict the large steel model tests bearing 6 in mind that the containments that they have been looking at 7 in their research program are spherical.
8 However, they thought about it and decided that they 9 are better off putting their analytical money into spherical 10 models. However, they are cooperating with us to the extent 11 of testing, doing biaxial testing in a machine that they have 12 of some of the materials that we had in the steel models.
13 We also expect that we may well have cooperation
[V}
14 down the road in the seismio question in Germany. That is, we 15 have not yet gotten to that, and I will talk some more about 16 the seismio question later.
17 As far as the concrete experiment, we do have l 18 agreements with EPRI to provide protest predictions for the 19 model. We also have agreements with the Commisariat A 20 L'Energie Atomique, CA, Enssole and Fontanoeurs to perform 21 protest predictions. They now have copies of the drawings and 22 are at work. The same with the Nuclear Installations 23 Inspector in the United Kingdom. They are having a contractor 24 doing a pretest prediction. I will talk a little more of that O-
\' 25 later.
i
e 403
! 1 MR. SIESS: What are they predicting, leakage or 1
2 failure?
i 3 MR. COSTELLO: Deformation.
4 MR. RICHARDSON: I think there is an interesting S contrast between these two organizations. CEA is going to do 6 a very complex, three-dimensional, sophisticated analysis, 7 kind of representing the people that are enamored with 8 computers and 3D finite element approaches, where N!! is going l
l 9 to take a very simple approach, not quite back of the l 10 envelope, but a very simple approach, and I think it will give 11 us a chance to see the sensitivity of complex versus simple 12 methods and models.
\ 13 MR. COSTELLO: Yes. The whole idea here is to try b[~
14 to get some handle on how well state of the art techniques can 15 work and what do you have to model and how well to get 16 predictions that are viable.
17 (Slide.3 18 1 think at this time I would like to give a quick 19 summary of where we were. In fact, in talking about the 20 pretest predictions on the steel model series, I have to say 21 that is one place where I bet on the wrong horse. I had the l
22 mistaken belief that self-interest, enlightened self-interest 23 would have A&Es and/or computational mechanios firms want to 24 show off how well their calculational techniques could predict
('
l -
25 the outcome of these experiments.
-.~ _ . . - _ _ . . -- .. -.-. --..--- - - -- . - - - . _ - - . . .
e 404 '
1 1 It all went along fairly well. We did invest some l 1 I g 2 money in mailings and winnowing lists, asking people would 3 they be interested. The answers of a great number were yes, i
, 4 and apparently when it required only a 20 cent stamp, they are 5 willing to say yes. We sent out some drawings and responses j f
f 6 began to flag, and in fact, no A&E firm or computational 1
4 7 mechanics outfit actually went through the in-house expense of I
, 8 doing the calculation. Perhaps they just didn't see the j 9 market for it in the aftermath, but that is one place where I
- 10 --
I 11 MR. SIESS: Are you assuming that they knew how? .I 12 mean it's a lot different calculating when something will fail f / 13 and calculating when something will stand up, which is what i \s - ;
- 14 design is. ,
l i 15 MR. COSTELLO: I believe that all these firms had 16 the same tools that were available to the people who did the i I 17 calculations at Sandia. {
I 18 MR. SIESS: Inelastic?
i
{
19 MR. COSTELLO: Yes, sir. !
20 MR. SIESS: Yes, but how much experience did they I 21 have using them?
i i 22 MR. COSTELLO: More, I think, than some of the ,
i
! 23 people at Sandia who actually did the calculation in some 24 firms.
j 25 MR. SIESS: You mean the AEs had made inelastic l
i 1
e 405 1 calculations of containment behavior?
2 MR. COSTELLO: These have not, to my knowledge, made 3 much more than the capability prediction kind of calculation.
4 MR. SIESS: Some of those just say calculate the 5 yield strength and that's it. That's not really inelastic.
6 MR. COSTELLO: Or one percent --
7 MR. SIESS: That's still elastio.
8 MR. COSTELLO: But I think the people are there.
9 Certainly the A&E firms have garnered a large number of the 10 graduates of major universities over the last 10 to 15 years, 11 and all of these fellows, many of them have been engaged in 12 those kinds of calculations. Certainly the computational
/ 13 mechantos firms have that as their stock in trade, but I guess
(
14 they simply didn't see the market and didn't want to make the 15 in-house investment.
16 The activities currently going on I will talk about 17 on protest predictions of the model tests at the end.
18 I would like to go down to 1986 and 1987 a bit with 19 you and then go back.
20 As Jim mentioned before, we made a consolous 21 decision some years ago after discussions with our peer review 22 panel and other members of the NRC Staff that a single model 23 test of a reinforced concrete model, given that it was about 24 all we could afford, was also the most likely vehicle to give O
\m / 25 us enough understanding to try and predict prestress behavior.
e 406 1 We had as background the fact that there had been
) 2 the Canadian model test, albeit on line, at 1/14th size of a 3 prestressed containment. There had been in the past, albeit 4 it at much greater wall thicknesses, PCRV test to failure, and 5 the prospect that there might be a prestressed model test of 6 something representing the SNPPS type containment undertaken 7 in the U.K.
8 That's an off and on -- as Jim reports from last 9 August, it looks more on than off, but things might change by 10 October. Nonetheless, we had that for background, and we also 11 had the belief that, given that the oracking phenomena are the 12 hardest to model and the effects of the oracking phenomena on 13 the liner are the hardest to model, the reinforced concrete 14 model would be a harder test of the ability to model behavior 15 near fatture, 16 Further, or taking the belief that you could l
17 reliably predict failure in a reinforced concrete model, 18 extensions to prestressed where you have much less oracking 19 along the path would be easier done. We still believe that.
20 We now have to prove it. We will begin that in '86, 21 We also have to look seriously at the question of 22 what do we need to do, if anything, to look at the question of 23 degradation of containment capability under large seismic 24 loading. Is it possible that you might lose containment 25 capacity in a containment in an accident to follow)
e 407 1 MR. EBER80LE: When you do this, are you b)
(j 2 hypothesising that a coincident loss of coolant accident or a 3 secondary failure will occur?
4 MR. COSTELLO: No, sir. I will get to that later, 5 at the end. At the end I have to say a few words about where 4 we are gonig to go there.
7 We will then in 1987 he going back to try and figure 8 out what happened in the concrete model test, and I, frankly, 9 am not as optimistic about that as I was about the outcome in 10 the steel experiments. In the steel experiments, I stood here 11 cheerfully four years ago and said it will all come out, not 12 to worry, we will figure out rather easily what happened. I 13 think we did and I will talk about that a little later.
(
14 In the concrete model, the state of the art of large 15 deformation modeling is n6t anywhere near as vast and tried 18 and true as in other applications in steel, and I expect wo 17 will have, in Professor Stess' phrase, a good deal more 18 questions after the experiment than we had in the steel.
, 19 With that, I would like to spend about 10 minutes or 20 so on the videotape machine giving you a little precis of what l
21 went on in the steel experiments. Now, this is not a very 22 elaborate production. It is not, shall we say, of the caliber 23 of those produced by KFK on the HTM experiments. It is simply 24 a splicing together of plotures, graphs and the !!ke collected O 25 over time in the series of steel model experiments, with l
l I
= - _ = - - _ _ --__ - _ . . -- . - .. .- _ _ _ _ - - _ - - --
i 408 l
l I
1 admittedly a little bit of music and some file footage taken
/~N
()
l 2 in, but I think it does rather neatly summarise what went on 3 in the steel experiments area.
4 [A videotape was shown.3 S MR. SIESS: Jim, have you got an explanation for why 6 those cracks did not run in the first two steel models?
7 MR. COSTELLO: We have a pretty strong hypothesis 8 which is being checked now in the closeout. The answer is 9 most likely extreme thinning, and we have measurements on that 10 done by the grinding wheel at --
11 MR. SIESS: Yes, but why didn't it run?
12 MR. COSTELLO: The primary suspicion is that the 13 material was so thin that even --
14 MR. SIESS: The stresses outside of --
15 MR. COSYELLO: Yes, it was confined by the material.
16 MR. SIESS: In both cases?
17 MR. COSTELLO: In both cases. And that is one of 18 the thinga that we are going to check out. The fundamental 19 thing, in the words of Dr. Von Reisman, his observation was 20 that grinding wheels don't scale very well.
21 MR. SIESS: You know, I'm going to make a 22 semi-facetious comment that maybe that is the way we ought to 23 butid them. You know, just put a thin spot in there so that 24 when they do fall, it will just be a nice, good-sized leak and 25 they won't end up in pieces around the countryside.
1 1
1
409 1 MR. EBERSOLE: That notion of programmed failure, 2 Chet, that's what you are talking about, isn't it, programmed 3 failure?
4 MR. SIESS: This is leak before break.
5 MR. EBERSOLE: Yes. I never really had heard that 6 treated as a design thesis. I don't know why.
7 MR. SIESS: I mean it's just like a rupture disk.
8 MR. RICHARDSON: What would be the difference 9 between that and a relief valve?
10 MR. SIESS: Well, somebody has got to decide to open 11 that relief valve.
12 MR. RICHARDSON: Well, it could be set at some 13 pressure.
b 14 MR. SIESS: Well, they will debate that for five 15 years.
16 MR. SHEWMON: I was going to ask the question a 17 different way, and that is, how do you think the NRC is going 18 to approach what they will have for the limit on this. There 19 is a decision of, okay, when do you recommend opening it so it 20 doesn't blow?
21 MR. EBERSOLE: Well, it certainly doesn't seem i l
22 rational to build it like the wonderful one-horse shay. l 23 MR. SIESS: Paul, your question is great because, if 24 that question could be answered now, we would probably end up n\/ 25 saying that we want to be sure that it doesn't rupture, and we l
1 l
l 410 1 want to open a valve o have a rupture disk or something in l
O)
(, 2 it. We would probably choose a pressure that could be l
3 predicted with a great deal of coniidence, and we wouldn't 4 need these kind of tests. But nobody has answered that 5 question. In fact, they haven't even asked it properly. So we l
6 3 go ahead and try to get all the answers. You know, whatever 7 anybody decides to do, we will be able to tell them how this 8 thing will behave.
9 MR, COSTELLO: I think that, you know, an outcome of l
10 a series of steel experiments was to pretty strongly reinforce l 11 the suggestion that if you picked -- and some other 12 penetration experiments that Hans is going to talk about next 13 -- is to suggest that if you picked something like general 14 membrane yielding, you would be quite sure that you were well 15 below onset leakage from penetrations and that you are well 16 below likely gross failure.
17 Ma. SHEWMON: What was the pressure for general l
18 membrane strain on this one?
l 19 MR. COSTELLO: I don't have the Vu-graph -- well, it 20 was around 160, 21 MR. SIESS: Your post-test analysis, you can 22 probably explain why that thing failed, the big one that 23 failed when it did.
24 MR. COSTELLO: Yes, sir.
\
25 MR. RICHARDSON: But that doesn't mean you could i
.,, _ _ _ _ _ - , . , _ _ _ _ , . . - _ _ , , . . . , . , ...__,.,,om_,,__,,_,_,__m, , . , , , _ _ _ . , ,-% , _ _ . _ _ , . , , . y
.,.,__.,y,.... .-,m. 7,m,
e 411 1 predict it because it was sort of a localized effect of the
) 2 people going back looking at had they accounted for the 3 eccentricity Sure. And they debated it at the time when they 4 made their pretest predictions. They debated whether they 5 should model that eccentricity or not. They erroneously 6 concluded that it was a never-mind.
7 MR, SIESS: But how many places would they be in a 8 typical containment where somebody would have to make that 9 decision, you see.
10 MR. RICHARDSON: Probably a lot.
11 MR, SIESS: And yet, to make a decision to take a 12 general membrane yielding or smeared -- you know, you have 13 established you can smear the stiffness.
14 MR. SHEWMON: That is four times design. You know, 15 one shouldn't be greedy.
16 MR. SIESS: Yes, but it's four times LOCA design, 17 but it's not four times what the severe accident predicts at 18 certain stages, and you get peaks at early stages. In some 19 containments those peaks are within the uncertainty range on 20 the ultimate capacity.
21 MR. COSTELLO: Yes. TMLB sequences and things like l
i 22 that.
23 MR. BCNDER: In these tests that you have run, 24 somehow or other they have to be put in relation to the 25 leakage characteristics of the closures that some people are
e 412 1 arguing will occur before the failure of the vessel itself ID happens.
( ) 2 Do we have any feeling yet for what the relative 3 capability of the sealed closures is?
4 MR. COSTELLO: Hans is going to talk to that in his 5 summary of the penetration work, but it simply comes down on 6 the' steel model test to this. The ovalization of the 7 equipment hatch, which was suspected to be the most likely 8 failure mode of leakage through that equipment hatch, was 9 predicted very accurately.
10 MR. BENDER: That is not quite what I'm asking, but 11 it may be around it. If the burst precsure of this vessel is 12 four times the design pressure or whatever it is, and the
'"N 13 leakage capability of the gaskets is 2.5 times the pressure --
14 I don't know really what it is. I'm just picking a couple os 15 numbers.
16 MR. COSTELLO: It,was substantia 11 more than 4.5 17 because the leakage did not occur.
18 MR. SIESS: This one did not leak.
19 MR. BENDER: Well, this one did not leak, but I 20 don'i know whether this is representative of the gaskets --
21 MR. COSTELLO: Let me speak to that question. One )
22 of the exercises we went through in the post-test analysis, l 23 which I have not reported here, is to ask the "What '
t 24 if?" question, to say, what if, instead of this equipment b
l
\ss/ 25 hatch reinforcement configuration, another one or two which is . - .
413 1 are typical of the population were put in?
w 2 It was easy to examine that question because one had 3 great confidence about the finite element model's ability to 4 predict the ovalization having been benchmarked against the 5 test data. I think five were examined that spanned the 4
6 population. The outcome was that two of them would surely not 7 have leaked in this experiment. The one that was in there 8 didn't, and two others would almost surely have leaked 9 earlier.
10 MR. BENDER: Well, there are a couple of ways to 11 deal with Paul's point. One obviously is to put in a rupture 12 disk. Really. Another is to have some weakened structural 13 capability, and the third is to select the kind of sealing for 14 the closure that would assure that leakage does occur in 15 sufficient quantities so that the bursting never occurs.
16 Now, all those remedies are out there and they are 17 all within the limits of the present regulatory system. The 18 only one that is not in there, maybe, is that nobody wants to 19 be in the position of saying, well, I'm going to be the guy 20 that is going to press the button. That's not exactly 21 rational, but that is the way it is.
J' 22 MR. COSTELLO: Again, as Professor Siess pointed 23 out, there is that question of how much more is really there I
24 that might be of benefit to you. If, indeed, you believed b
G 25 that the pressure in an accident was going to peak at 165 or
, /
e 414 1 170, you would not have wanted to push the button at 160 given
(_) 2 that this thing was going to ride out until nearly 200.
3 MR. BENDER: Well, if that is the game you want to 4 play. But if you really would prefer to be in the position of 5 saying I want to have an assured capability up to some level, 6 and when that level occurs, I want to direct whatever comes 7 out in a certain way, you would be far better off to select 8 some mechanism and say, I'm going to work on the reliability 9 of that failure mechanism happening.
10 Mn. SIESS: Mike, what you are bringing up, it would 11 be very nice to know at this point just how somebody is going 12 to use these results, and that is what I was stressing. In 13 contrast to some of the stuff we heard yesterday, the severe 14 accident business has not been carried far enough yet that 15 anybody knows how they are going to use it.
16 Now, we have had that trouble with the whole severe 17 accident program, that people could not tell us what they were
, 18 going to do with it when they got through; they were just i
19 going to learn everything about everything. The severe 20 accident program hasn't really addressed the possible uses of 21 this. Until some of this work was brought in, the people that 22 were calculating containment loads, containment pressures were 23 assuming it wasn't going to leak at all They were drawing f- 24 curves that just went on up. You know, the idea of U 25 containment leak never entered their minds.
a 415 1 And then their attention was gotten and they started 2 calculating containment loading as if it might leak, and they 3 are just beginning to think about these things.
4 MR. EBERSOLE: Well, Chet, it must surely be true 5 that nobody contemplates that we will ever allow a containment 6 to approach a burst pressure.
7 MR. SIESS: They haven't even thought about it yet, 8 Jesse.
9 MR. EBERSOLE: Well, it has been around for 20 or 30 10 years.
11 MR. SIESS: I know, but in addressing a policy on d
12 severe accidents, this has not been addressed except to talk 13 about vented filtered containment.
14 MR. EBERSOLE: I would rather have a big cannon out j 15 there on the site'to open it up if I had to.
16 MR. SHEWMON: Well, that is part of why you are 17 here, you know.
18 MR. SIESS: We are not so sure that the burst of 19 this thing seven days down the line makes that much 20 difference. ;
i 21 MR. EBERSOLE: Well, if it bursts violently and 22 re-entrainment occurs, that is something else.
4 23 MR. SIESS: That is something they are just 24 beginning to think about. As far as I can see --
4
\' 25 MR. EBERSOLE: 1 can't see by any stretch of the
l
. 416 1 imagination that we would ever approach a burst pressure; that
( 2 we would take violent means to preclude that.
3 MR. MARK: Well, this is at leat telling you that 4 you have a good chance to know where that burst pressure is.
5 MR. EBERSOLE: Right. That is all I see here, is 6 that you are finding the margin at which point you would 7 deliberately open it.
8 MR. SIESS: Well, if somebody would say that they 9 wanted a highly reliable, conservative 11'-established burst 10 pressure below which they could vent, I don't think we would 11 need that test.
12 MR. EBERSOLE: That is the point.
13 MR. RICHARDSON: Except you may need it to establish
[V'}
14 the validity of your models that predict that burst pressure.
l 15 MR. EBERSOLE: But you could put margins in the j l
16 strength.
17 MR. SIESS: But I think on the steel ones -- maybe 18 it took the big one. I'm not sure. I said to predict a 19 highly reliable lower bound, and if somebody says, okay, we 20 will accept yield as that, we don't really need a lot of model i
l 21 tests.
22 MR. RICHARDSON: But that may be so low that it 23 would --
k 24 MR. SIESS. It may be too low, but you see, you have l \
[ 25 got to think about it.
- - - - . ._ ,, . - , . - . . _ . . - . - _ ~ . _ . _ _ , _ _ _ _ . , , - . , _ - . . _ - . , _ . _ . _ _ . . - . . - _ _ . - _ _ . ,
e 417 1 MR. SHEWMON: We are talking about concrete now.
(~N) g 2 MR. SIESS: I'm talking about either one.
3 MR. SHEWMON: Well, but the steel, if that was four 4 times the design pressure --
5 MR. SIESS: I know, but it's not four times the 6 accident pressure. Design pressure doesn't mean anything when 7 you are talking severe accident.
8 MR. BENDER: It only sets an upper bound on what the 9 pressure is --
10 MR. SHEWMON: When you say a highly conservative 11 lower bound, the design pressure is that, and that is much too 12 low.
13 MR. SIESS: Well, no, 115 percent of design pressure 14 is a highly conservative lower bound, but the higher you go, 15 the greater the uncertainty is, obviously.
16 MR. EBERSOLE: Chet, let me ask you. With the 17 concrete vessel, I understand that at least for the primary 18 concrete vessels for gas, they have an intrins'ic phenomenon 19 that the reinforcement will stretch, the concrete will crack, 20 and it will become its own permeable relief system.
21 Will that apply to this containment?
32 MR. SIESS: The liner has to fail. The concrete is 23 not leak-resistant material. The liner is.
24 MR. SHEWMON: But the concrete limits the strain on
)
25 the liner so that you may well have -- well, you certainly
418 1 will have a limited rupture.
2 MR. SIESS: Yes, that's right, but the cracks in the 3 concrete can get pretty large before the liner fails. We 4 don't know. The reason for testing the concrete containment, 5 in my mind, is to see if there are some questions we haven't 6 thought about.
7 MR. EBERSOLE: If the liner failed and it 8 pressurizes the concrete inner surface, then the concrete is 9 pressurized, and will it not then act as its own relievable 10 structural --
11 MR. SIESS: Oh, yes. Unlined concrete, you can 12 probably get enough cracks so if the stuff would leak out, it
% 13 would go out through cracks, and then you have the question of
[b 14 how much deposition there is and aerosols in the cracks. You 15 know, the devious path and so forth. But again, the unlined 16 concrete one would not work for a LOCA, which has always 17 governed our design.
18 Now, in a prestressed concrete containment, they are 19 designed not to crack under LOCA load, and the Canadian 20 prestressed containments, the French prestressed containments 21 -- no, the French have liners, I believe, although they have 22 looked -- well, anyway, the Canadian one, I know, is not 23 lined. It has an epoxy paint that will take care of --
24 MR. EBERSOLE: It will stretch, b
25 MR. SIESS: But it is prestressed, so at LOCA loads
a 419 l
1 it is not cracked, and nobody has looked beyond LOCA loads k 2 because that is what they were designed for. But a reinforced 3 concrete one would not stand the LOCA. You could not get 4 one-tenth percent leakage at LOCA loads on a reinforced one 5 because you would have enough stress to get some cracks and 6 you would get some leakage. I don't know whether anybody has 7 even tried to calculate it.
8 But now we are talking about accident loads. At 9 Indian Point, for example, TMLB' gets you up within a few psi 10 of the predicted yield capacity, I think, in a coupe of hours, 11 if I am not mistaken, and then it drops off a little bit and 12 then a slow overpressurization comes and it kesps on going.
/' 13 But at that first stage where there is a lot of fission 14 products in containment, it is within a range where you might 15 like to know it better. In some others, you would be a mile 16 off.
17 You see, you remember the trouble we went to on 18 Sequoyah trying to get the capacity -- you know, we had four 19 different answers and we sat around a table one day. You take 20 the same design, essentially, for McGuire, where they didn't 21 reduce the thickness of the liner as they went up. Remember, 22 Sequoyah had the liner getting thinner and thinner as it went 23 up; McGuire didn't. So McGuire had something like, what, 50 24 or 60 percent more capacity. And this was just the designer's 25 choice of trying to save steel as he went up on this thing
e 420 1 because of vertical load. Remember that? So you have to look
( 2 at those things. It makes a difference.
3 But if somebody would tell us what they wanted --
4 they were really going to vent these things and they wanted, 5 you know, a 99 percent probability that they would vent below I
6 --
I 7 MR. EBERSOLE: Well, who is that somebody and why 8 don't we find him? i l
9 MR. BENDER: It seems to me if we were doing it in a i
10 rational way, you would lay out all the strategies, and if you 11 had a good probabilistic risk analyst, he would make the 12 arguments for the various data --
13 MR. SHEWMON: We would submerge him in severe C\
14 accident codes and he would never be seen again.
15 MR. BENDER: Well, that's probably right.
l 16 MR. SIESS: Well, if you believe the severe accident 17 codes, which, you know, takes a certain amount of faith, you 18 could be at that point right now. But as I pointed out, IDCOR 19 looked at this venting thing; but on a cost-benefit type 20 thing, venting to prevent rupture of the containment seven 21 days down the line isn't worth anything in terms of dose 22 release.
23 Now, they didn't talk about the resuspension.
l-24 MR. RICHARDSON: Yes, with catastrophic failure.
25 MR. EBERSOLE: That is what would be a key effect.
e 421 1 MR. BENDER: You won't have it if you will accept j 2 the argument of gasket leak --
3 MR. SIESS: But Jesse, the severe accident research 4 has placed -- I am prejudiced -- has placed 99 percent of its 5 emphasis on source term, and what is in the containment, all 6 the scenarios and how it gets there and so forth, and on 7 containment loading phenomena, and have not really been 8 thinking about the things t'e are asking questions about. I 9 have tried to get their attention on it.
10 MR. EBERSOLE: The cart is always before the horse, 11 then.
12 MR. BENDER: No, that's not quite right.
[}
V 13 MR. SIESS: The people that were doing the work got 14 absolutely fascinated with this suite of codes, and they are 15 going to get answers to what people thought were unanswerable 16 questions, and now we are doing it, you know, and we are doing 17 research on core-concrete interaction.
18 MR. EBERSOLE: The war is being won by the patrols.
19 MR. SIESS: Yes, it's the old question, Jesse. The 20 ACRS has been asking it ever since they started. What are you 21 going to do with this? What are you going to use it for?
22 MR. EEERSOLE: Yes. If you had already done it, 23 what would it be worth?
24 MR. SIESS: If I told you this, what would you do O 25 with it?
1
e 422 1.
1 MR. EBERSOLE: Yes. Well, I think every research 2 program has all this. Suppose you are already done? What 3 have you got?
4 MR. SIESS: Yes. What they did on severe accident 5 -- I said it and I will say it again. Their approach was we 6 are going to learn everything about everything, and when we 7 get through, we will know what to do about it. And they are 8 still trying to learn everything about everything. And here 9 we tried to learn everything about everything because nobody 10 has told us what they might do about it. And you can't ask 11 the structural engineers what to do about it.
12 MR. COSTELLO: Well, I can't say that we are so 13 immodest as to try to learn everything about everything.
14 MR. SIESS: No, I didn't say you are doing that, but 15 again, you may be learning more than you have any use for if 16 somebody could tell you what they wanted to do with it.
17 MR. COSTELLO: Yes, but one has to look at the 18 prospect that a high degree of confidence of when and where
- 19 it's going to fail is something in the end game that will be 20 important.
21 MR. SIESS: You heard me state one of my famous 22 generalities that one of my old professors taught me, and it 23 has been borne out by my own experience. It says tests on 24 full-sized structures never answer questions, they only ask 7s 25 questions. And this is about as close to a full-sized
423 1 structure as you are going to get.
O)
\s 2 Take the steel one. Did it ask any questions?
3 MR. COSTELLO: I would say it asked three questions, 4 two of which were easy to answer, one of which was a little 5 harder, and I think we have got the answers to them. And the 6 net outcome I say is this: that the capability exists within 7 the A&E firms and/or some of their consultants on behalf of 8 utilities to go back and look at the steel containments they 9 designed and built, use available technology and come up with 10 very good estimates of when and where the containment is going 11 to fail if they choose or the utility chooses to put the 12 money into it.
13 MR. BENDER: I wonder if I could make a point that
(
14 develops something Dr. Siess said a moment ago. When the 1
15 Sequoyah vessel was looked at, there were four methods of i i
16 analyzing them and they gave four different pressures.
17 MR. COSTELLO: I think it was closer to tpn .
18 MR. BENDER: Well, I'm not sure how many there are, 19 but four is enough for my argument. I am not clear right now 20 which of the four was near to the right answer based on the 21 experimental program that you did here, but it might be useful l
l 22 to be able -- to make that comparison just for the purpose of 23 showing --
- 24 MR. SIESS: Well, one thing they established clearly 25 in the small scale was that smearing works. You can smear the
424 1 stiffness, and that was an issue in the Sequoyah thing.
) 2 -Somebody assumed it, What's-His-Name with Westinghouse, with 3 Ops, made the finite element type thing and supported it, and 4 that came out very clear.
5 MR. BENDER: But I guess ths point I am really 6 trying to make is that, yes, all the architect engineers can 7 predict these capabilities. The question is which of the 8 several methods that is out there is the one that you want to 9 give preference to for this purpose. That's the most useful 10 part of this program.
11 MR. COSTELLO: I will give two answers to the 12 question. The answer to Problem 1 is, as I tended to call it
/' 13 over the years, that is a not unreasonable lower bound O) 14 estimate. I believe there is sufficient evidence for steel 15 containments that membrane yield quite suffices, and the 16 outcome of the attempts to get leakage around these 17 penetrations and some of the other experiments that Hans will 18 talk about, if I ever give him a chance, at temperature will 19 support that hypothesis.
20 The second one is Problem 2. Do you really have to 21 know when and where it is going to fail? I say that the 22 outcome of these experiments has been reasonably conclusive 23 that the state of the art will do it if someone wants to model i
,s 24 it in the kind of detail that's required.
U 25 MR. SIESS: I think Mike may not be aware of how I
e 425 ,
1 that steel containment failed and why.
D 2 [ Slide.]
3 MR. COSTELLO: Oh, yes. May I wrap up?
4 Here we go with the picture. Up here you have a 5 formed stiffener, a formed horizontal. There are eight such 6 details like this.
7 [ Slide.3 8 The design had what amounted to a half-inch of 9 eccentricity. That is in the lower picture. Why? In order 10 to make the weld, I guess. The belief by the fabricator was 11 that that is the common thing to do. The judgment was 12 made, after some agonizing early on in the pretest prediction, 13 to model it as if there was no eccentricity.
14 MR. SIESS: Now, here you mean mathematical model 15 and analyze that?
16 MR. COSTELLO: Yes. The analyzing pretest 17 analysis. Let's not try and bring in a little bit of bending 18 in here because tradition says that's not really 19 important. Again, in yielding, things will -- et cetera. So 20 they went with this model [ indicating 3.
21 The outcome of the experiment was that at around 165 22 psi, a crack was noticed i t: the formed stiffener. The crack 23 was spotted by an operator. One continued to raise the l
24 pressure in the programmed fashion. On the way from 192.5 to 7-25 195, there was the outcome.
e 426 1 This, as I say, was the third question, the hard
() 2 question that was harder to explain, but it turned out to be 3 not that hard. They went back -- you know, no curve fitting, 4 no derivation of constants. Just went back in and did a 5 finite element model, which included that eccentricity.
6 [ Slide.]
7 MR. SIESS: Would the model predict the crack 8 running?
9 MR. COSTELLO: Yes.
10 MR. SHEWMON: No, you don't really get into fracture 11 mechanics in your model, do you?
12 MR. SIESS: No, but you get into propagation. Was 13 the crack a brittle crack?
14 MR. COSTELLO: Oh, no. It's ductile.
15 What happened at predicted strain in the cylinder 16 right next to that juncture in the pretest analysis, the 17 membrane yielding roughly at 160, things beginning to happen, 18 but nothing that would lead you to believe that you would have 19 a catastrophic failure. And the predictions went out 20 somewhere up here, around 220.
21 The other predictions for deformations around the 22 hatch in ovalization suggested that you really would have lost 23 metal-to-metal contact around 210. You would expect leakage, 24 but it never got there.
O 25 MR. SIESS: Jim, the knee in the curve at 165 could
e 427 1 be predicted without a finite element analysis?
s_ 2 MR. COSTELLO: Yes.
3 MR. SIESS: That could be just membrane-type.
4 MR. BENDER: Well, the point that I was trying to 5 make, and I will try to make it again, is if I can establish 6 that 165 psi, or 160, whatever it is, is predictable by the 7 methods of analysis that are available to me right now, and I 8 don't have to worry about the sophisticated details that come 9 with having to predict the rest of the behavioral pattern, I 10 probably know enough to say if I can make the case for that 11 160 pounds being an acceptable upper limit and I can design 12 the vessel so it relieves before that point, I have got a 13 better case for accident control than I had with any of these
[d}
14 things based on ultimate failure.
15 MR. SIESS: Right.
16 MR. COSTELLO: That is certainly a more robust 17 argument if that is the one you want to make.
18 MR. BENDER: That's the one I ought to be trying to 19 make, and that's the one you ought to be trying to make.
20 MR. SIESS: I think that's the one they are about 21 ready to make on the steel containment. I think they are 22 satisfied now that if somebody will look at the details, which 33 conceivably could affect you even lower, but if they would i I
24 look at the details, that you could predict that level with a O 25 high degree of reliability and you could probably predict some
a 428 1 other level with somewhat less reliability because there is (D j) 2 always the question of completeness; right? And they don't 3 propose to do any more tests on steel containment.
4 MR. COSTELLO: No. As far as we are concerned, the 5 ballgame is over from a research point of view.
6 MR. SIESS: Now for concrete containments, what you 7 would hope to get out of the test is some similar surprises or 8 details or -- I hate to use the word " outliers" or something 9 of that sort. But that can be overcome by a suitable analysis 10 and end up on the same state with the concrete containment.
11 MR. BENDER: Let me go back to the other point that 12 I was trying to make. Sequoyah had an analysis. If I were to
[\ 13 take one of these analytical methods and relate it to 14 Sequoyah, what answer would I get?
15 MR. SIESS: Oh, you could predict 165 --
16 MR. BENDER: For that particular vessel, in 4
17 comparison to what they did get. Do you understand what I'm 18 trying to say?
19 MR. COSTELLO: Sure.
20 MR. BENDER: Let's get a real case and compare it 21 with a model so we have something that exists, that we can 22 make a case with.
23 MR. COSTELLO: Well, let me tell you how we are 4
-s 24 going to close this thing out. We are going to have -- it is V 25 part of this question. We are going to examine the question
429 1 of how different details in the population might lead you to qj 2 something like this. Just as we examined previously how 3 different penetration designs -- as part of doing that 4 examination, we are going to be looking at what kinds of 5 modeling would have to be done for real penetration, and real 6 details in reinforcement. And those kinds of calculations 7 will be done on a sampling basis.
8 The real answer is, to my thinking, that the persons 9 most capable of giving a good analysis of the containment are 10 the people who designed them.
11 MR. SIESS: But, Jim, you take the Sequoyah analysis 12 and use the smeared stiffeners, ignored the penetrations, 13 except for looking at the capacity for buckling and so forth.
14 The prediction that that analysis made of the general yield --
15 if you tested the model of Sequoyah, if you tested Sequoyah, 16 don't you think there's a very high probability that it would 17 reach that pressure --
18 MR. COSTELLO: I would say virtually certain, yes, l
19 sir. And in fact Gryman has done that calculation already.
20 MR. RICHARDSON: It seems to me Mike is asking, 21 look, a few years ago in an atmosphere of a lot of ignorance, 22 they made some predictions. Doesn't it make sense now to go 23 back and revisit how well they did, using today's technology?
24 MR. SIESS: I don't think you have to go back, O 25 because all they did was look at free field, smeared the
430 1 stiffeners and --
) 2 MR. RICHARDSON: It seems to me, though, it may not 3 be a big deal to do, but an interesting --
4 MR. BENDER: My reason for wanting to do is 5 pragmatic. It says, look, here is a case we had out here, 6 here is how the research would have influenced that action.
7 If you did what Jim did later and looked at all the variations 8 that have to be thought about, I think that is a constructive 9 thing to do.
10 MR. SIESS: But, Mike, I think it has already been 11 done.
12 MR. BENDER: I know, but they haven't put it out 13 there where people can see it.
14 MR. RICHARDSON: But it would lend credibility to 15 what we have done.
16 MR, SIESS: This shall reach general yielding, which 17 is all the Sequoyah analysis'was.
18 MR. BENDER: I think it would show that one of those 19 methods was the right one to use.
20 MR. SIESS: To predict general yielding. That it 21 wouldn't fail --
22 MR. BENDER: That was enough for them at the time.
I 23 MR. SHEWMON: Mr. Chairman, are we at 9:30 or 10:30?
24 MR. SIESS: Well, we are closer than you think.
O 25 I have suggested to Jim Richardson that we will skip l
a 431 1 seismic performance of containments, which is clearly in the
' ("')
( f 2 future, and pick it up at some other time. But I think we 3 should have you -- can we go on now to penetrations, and then 4 we will take a few minutes and see what time it is and talk a 5 little bit about the EPRI stuff.
6 MR. COSTELLO: I guess if I could say I'd like to 7 get the last word in on the steel models. Basically we are 8 not prepared to declare total victory yet, but we are having 9 this one set of closeout calculations in which someone else 10 will try different but equivalent code to see if they 11 reproduce those effects, and also look at the question of, if 12 you look at something like Sequoyah -- which in fact this 13 , person has looked at before, and look at some of the details, 14 would you come to the conclusion that you would get the
, 15 membrane yielding, or is it something you would have to model?
16 And I think at the end of that we can claim that the 17 game will be over and the good guys have won.
18 MR. SIESS: But in the meantime, would you try to 19 convince some of these people that they ought to decide what 20 they are going to do with this?
21 Okay, let's go on to Hans, and he will describe what 22 Sandia is doing on large hatches, small hatches, and pipe 23 penetrations.
i I
24 MR. EBERSOLE: Chet, while he is getting ready, it 4
\- 25 always bothers me that if a failure occurs prior to core
e 432 1 damage, which can be a sequence of events, if the failure 2 occurs in the regressive context, that it blows into the aux 3 buildings of that region where the penetr,ations generally are, 4 that will virtually guarantee core failure and disaster.
5 If it can be made to fail some place into open 6 atmosphere, you have still got a chance.
7 MR. SIESS: Well, we could probably arrange that.
8 MR. ASHAR: Before I start on penetrations, let me 9 make a little remark on Sequoyah.
10 Sequoyah has been looked at quite closely in NUREG 11 1037 on leak rate estimates in which, as far as I remember, at 12 a peak pressure or peak acceleration pressure for the worst 13 case scenario, they are predicting the pressures which does 14 not prove the leak-before-break hypothesis.
15 The leakage is so small for penetrations in all the 16 other areas that they have examined, they figured that if the 17 pressure was any higher than what they had predicted, it would 18 rupture the containment rather than leakage.
19 MR. RICHARDSON: Hans will discuss this later. But 20 Mike's point is, it is in essence a PR -- an excellent one. :
1 i
21 MR. SIESS: NUREG 1037 chose Sequoyah, and you would 22 have gotten a different answer if you had chose McGuire, which 23 is the same design of plant but a different design of 24 containment.
25 [ Slide.3
e i
433 1 MR. ASHAR: I am going to show first the boiler N
i 2 plate slides on the project, and then I will show you some of 3 the work that has been done in the past, and what will be done 4 in the future, and then some of the sketches of the pressure 5 chambers.
6 This is with Sandia National Lab, and the budget for 7 '85 was 900K. For '86, it has been earmarked for 900K. For 8 '87, it has been marked down to 600K.
9 We wish to complete most of the seal and gasket 10 testing and probably most of the needed tests for the bellows 11 for the fixed penetrations.
12 [ Slide.]
13 The issue, I think it has been pointed out quite a 14 number of times by the ACRS members, that leak-before-break is 15 the one that we should examine more closely than anything 16 else, and this is the attempt to do that. As to whether the 17 containment will rupture first, or there will be leakage 18 enough through the various penetrations in the containment 19 that that will start depressurizing the containment.
20 So the issue is basically the analytical prediction 21 of the containment failure modes -- that is what will be the 22 failure mode, will there be leak-before-break or break of the 23 containment.
24 The objective is to determine the major 25 characteristics of operable and fixed containment penetrations 1
, . . , . ,e ------,a-.,---- ----y , - - - , , - - , - - , - - -n - - - - , - - - - - - . . , - - - , - - - - - - r--m.----- .--.w. .-- -
a 434 1 that could contribute to leakage during severe accidents.
/ 2 The effect of pressure, temperature and deformation 3 must be studied.
4 Now let me define for you, i' t i. o r e are some who are 5 not aware of it, what our operable and fixed penetrations 6 are. There are obvious terms, but still operable penetrations 7 which are likely to be opened and closed during the operating 8 years.
9 MR. SIESS: Other than valves?
10 MR. ASHAR: Other than valves. Valves are operable, 11 but they are not included in this project.
12 What comes out are equipment hatches, air locks, 13 escape hatches, drywell of the BWR head, and maybe in some 14 cases fuel transfer tubes which are blind flanged, but they 15 are operable, they can be opened and closed.
16 So these are the operable penetrations. The others 17 are piping penetrations which are considered fixed l 18 penetrations.
19 MR. RODABAUGH: When you use the term 20 leak-before-break, are you thinking of gasket leakage or 21 mechanical?
22 MR. ASHAR: We are thinking of gasket and seal 23 leakage.
24 Integration. It has been pointed out before that
('
25 the three projects are closely integrated. In addition, the
435 1 valve leakage program, as well as the electrical penetration
[T s ,/ 2 and assembly program, they are all integrated to assess the j
3 containment capacity.
4 MR. SIESS: The term leak-before-break here is 5 different. It means anything that happens before the 6 containment ruptures.
7 MR. BENDER: Yes, it is a different kind of leakage.
8 MR. SIESS: Yes. But they just adopted the term 9 here.
10 MR. BENDER: They ought to fight not to use it, 11 because it confuses people.
12 [ Slide.]
/~') 13 MR. ASHAR: This project was initiated in 1984. The 14 first thing that has been done was to survey penetrations. We 15 really didn't know what type of penetrations exist, what type 16 of seals and gaskets are used in penetrations, what are their 17 configurations and sizes, their types, and the materials, et 18 cetera.
19 So the first thing that was done was a survey of the 20 existing operating plants to see as to what are the parameters 21 that we ought to look into as far as testing was concerned. !
22 I will go more into detail along with the survey l
23 results.
24 That was done prior to FY '85.
)
N/ 25 In '85, basically what has been done is testing of
e 436 i seals and gaskets. Seal and gasket tests have been done.
) 2 Some of them have been done at Sandia National Lab in the 1
3 pressure chamber. Some of the tests have been done at Argonne 4 National Lab, and by a subcontractor.
5 In '86, we plan to complete the seal and gasket 6 tests and complete the air lock tests.
7 I pointed out during the presentation of containment 8 model tests that air lock in the model does not represent the 9 real air lock. It doesn't have an opening in the bulkheads 10 required, et cetera. That's why we are going to test one air 11 lock which we have procured at a bargain price. I will show 12 you the schedule later.
13 Procurement of bellows and development of test 14 plans. This will be done in 1986 so we can test the bellows i 15 in 1987.
16 [ Slide.)
17 In '87, the evaluation of the air lock test and j 18 performance of bellows test. The regulatory use in a generic 19 way, it is the same as before. Confirmatory assessment of 20 severe accident policy statement, and the basis for possible i
21 addition of containment performance requirement for safety 22 goal implementation.
23' But in detail, there are quite a number of uses.
i j 24 The first thing, it has been used in preparing NUREO 1037,
! 25 which is going to be a part of the severe accident reporting.
- , . - - - - , , _ , - - - . _ _ _ - - . _ ,,,,_,,y., y -c,_, ,,n.___ ,_.,m. _,_..~,._..,--.._.,,..-__,,,,.,.,_,y..._,--,,_e..,_ _
e 437 1 Also, we are making a move to have some of the 2 results used for the standards committees, such as the 3 replacement of gaskets and seals. Right now there are no set 4 requirements that people have to replace certain gaskets.
5 We will try to push that kind of idea into the ASME 6 Section 11 committees, and there are a lot of details which 7 have been learned from this testing which we will be 8 transferring to the code committees which would be of use for 9 modifying the codes and standards, 10 MR. SIESS: Hans, who has the job of looking at 11 pre-existing leakage? NRR?
12 MR. ASHAR: That particular project is under John 13 Wang in Containment Systems, NRR. BNL has looked at it. And 14 he has some of the results.
15 MR. SIESS: And they don't call that research?
16 MR. ASHAR: They call it technical assistance, 17 because it helps them in assessing as to what is available.
18 [ Slide.3 19 Now I am going to give some details of what has been 20 done. A survey of mechanical penetrations. The survey was 21 conducted by Argonne National Lab under subcontract from 22 Sandia National Lab.
23 A preliminary analysis of some of the structural 24 analyses was done by Argonne and later on by Sandia National C% 25 Lab.
I
438 1 Tests on penetrations are being performed by Sandia, 2 and Idaho National Lab.
3 An investigation of electrical penetration 4 assemblies is being conducted in another program, which I 5 pointed out before.
6 [ Slide.]
7 MR. SIESS: That is in another program because it 8 started out under qualifications.
9 MR. ASHAR: That is correct.
1 10 [ Slide.]
l 11 Program steps are: survey of penetration types and 12 designs performed by ANL on 48 U.S. plants. 48 plants, I fa 13 think, is a good sample. It's about half the available 14 sampling, and attempts we re made to make sure that there is 15 enough PWRs, enough BWRs, enough steel in the prestress and 16 reinforced concrete containment, because the inside is 17 different in each of those cases.
18 MR. SIESS: There is a report out on that.
19 MR. ASHAR: There is a report out on it, and it is 20 available, yes. If anybody is interested, I can send it to 21 them.
22 Qualitative figures of merit developed. Figures of 23 merit is a matter of finding out which are the parameters in g 24 the operable penetration assembly. As to how the sleeve 25 length,.how the diameter, how the thickness of the collar or
. _ - . _ . . _ . _ , - . - , _ _ . , _ _ . . _ - - . - _ . . , - . . .- .. _ .- ,._ . - ... _ ., . - - - _ , - ~
e 439 1 thickness of the reinforcing plate around the sleeve would O(,,/ 2 affect the potential performance.
3 These are figures of merit according to what is 4 considered as more important than the other. So that we can 5 figure out what we should do and what we should keep in our 6 minds. And that also is in that particular report. If any of 7 you are interested, I can send these reports.
8 MR. SIESS: I think what you omitted was those 48 9 plants yielded about several hundred different types of 10 penetrations.
11 , MR. ASHAR: Oh, absolutely, yes, they did.
12 Preliminary analyses were conducted based on the T 13 survey. They came out with some typical design for analysis 14 purposes as to what a typical -- really, it's not typical, but 15 just for the purpose of analyses, to get an idea about what is 16 predominating. The preliminary analysis -- there is a report 17 on those analyses, too. So if anybody is interested, I will 18 be able to transmit the reports to them.
19 Program plan developed and conduct the program. I 20 will be getting more details on those.
21 [ Slide.)
22 Experiments and analyses of penetrations. The shell 23 structure interaction, equipment hatch and personnel air 24 lock. They were all tested on a 1/32nd scale, 1/8th scale
,v 25 steel model, as well as they will be presented in a 1/6th
t a
440 1 scale concrete model.
2 The equipment hatch with seal, 1-to-8 scale steel 3 model, had one type of seal there. The equipment hatches, two 4 pressure seating and one unsestinD in 1/6th scale reinforced 5 concrete model, I showed you before in the concrete model 4
6 [ Slide.]
7 MR, SIESS: Excuse me. Is it clear to everybody 8 that what we are looking for there is deformation, such that 9 the two metal surfaces separate far enough that the 10 . elastomeric seals are'no longer sealing?
11 MR. RODABAUGH: I was about to ask a question on 12 that. l l
/ 13 MR. SIESS: There are two issues. One is the (
\ )) l l
14 deformation that occurs, and the other is the elasticity of 15 the integrity of the seal material which might degrade with ,
16 temperature.
17 MR. RODABAUGH: CAF v: i scale -- I was wondering 18 about that 1/8th and even .m a a i . e- scales. Assuming the 19 elastomer seals.
/
20 MR. SIESS: Those were the tests you just saw. And 21 all they used those for were to get the effect ci the shell 22 behavior in the deformations. The other tests are not being -
23 scaled.
24 MR. RODABAUGH: Okay. The ones that I was just O 25 looking at are not --
i 441 l
1 MR. ASHAR: They were performed separately, and will ,
l g 2 be integrated together when the final reports come out. So as 3 Dr. Siess pointed out, the deformation is how the structure in 4 general will affect the deformation on the sleeves, on the 5 penetration.
6 MR. SIESS: They have seals in there, but they don't 7 expect to get definitive answers with those particular seals.
8 It's just illustrative.
9 MR. ASHAR: A full sized personnel air lock will be 10 tested in '86.
11 Expansion of bellows in '86, '87, and drywell head 12 for BWR Mark I and Mark II analyis did not indicate a need for
[~]
%s' 13 experiment.
14 I will have a little bit on the drywell head later.
15 MR. SIESS: That was interesting. It started out 16 being pretty bad.
17 MR. ASHAR: Yes, it did.
18 [ Slide.3 19 MR. SIESS: Can you get to pressure without the 20 temperature?
l 21 MR. ASHAR: Well, in the Limerick study done on 22 NUREG 1037, they predict leakage at very low pressure, around 23 86 psig or something for Mark II containment, only considering <
l l
24 the pressure, and if the temperature is included, then it just ;
O N# 25 so happens that it sinks back, and there is no leakage until
a 442 1 up to 130 psi. I will be elaborating on that a little later.
2 Pressure seating equipment hatches. Models for 3 predicting the deformation and leakage for pressure seating i
4 hatches in the steel containment were verified by the 1/8th 5 scale steel model test. Jim showed some of the slides and 6 they have been discussed before.
7 Significant interaction between the cylinder and the 8 penetration sleeve occurs. That is quite a legitimate 9 concern.
10 Average membrane strength from 2.3 percent up to 5 11 percent must arrive before the sleeve deformation is 12 sufficient to cause a mismatch that results in leakage. We 13 are up to 2.3 to 5 percent, and they do occur in the C
14 discontinuity areas in some places.
15 MR. BENDER: When you talk about membrane strains in 16 this case, what are you talking about? What part of the 17 membrane?
18 MR. COSTELLO: Free field. That's one of the back 19 calculations I referred to. The range was a factor of two, 20 looking at existing typical penetration geometries, and in 21 that experiment, some of the low end would surely have leaked 22 before we had the rupture, and some of the high end surely 23 would not have.
24 MR. BENDER: Can you generalize, tell me that is 25 typical of all shells?
_ .-_ . ~ - . _ . _ _ _ _ _ _ . - , _ , _ _ _ _ _ _ _ . . _ _._ _ _ _ _ _ .___._ _ __
443 l
1 MR. COSTELLO: Of steel models, it certainly -- you.
N 2 can generalize that one of the population would leak before 3 they break, and some would break before they leak.
4 MR. BENDER: Until I get the 2.3 percent strain, 5 none of them would leak?
6 MR. COSTELLO: I would be hesitant to do that l
7 without looking at the penetration design itself. But of the l l
8 population they found, the one that had the thinnest or i 9 skimplest reinforcing still took that kind of free field 10 strain before --
11 MR. SIESS: Jim, you had a couple of small scale 12 models that ballooned. What kind of strains did they reach?
13 MR. COSTELLO: I believe it was around 15 or 16 14 percent.
15 MR. SIESS: Now you could generalize, I think -- or 16 could you generalize and say that a real containment with 17 hatches in it could not reach 15 percent without leaking?
18 MR. COSTELLO: Absolutely. That one I will go for. l 19 MR. SIESS: Okay. Every little bit helps.
20 MR. BENDER: I've got even more of an axe to grind.
21 I would like to be selective in the seals that I want to leak, 22 and if I am selective, I can make the leak go where I want it 23 to go, and when you're doing something like this, that's part 24 of what you want to be able to display.
O 25 MR. COSTELLO: I think it is fair to generalize that
444 1 if your choice is the large diameter equipment hatch and you
_,/ 2 are willing to take the leakage there, then that should be 3 rather easy to arrange.
4 MR. BENDER: But that's not where I want it. That 5 thing is too big, and I can't --
6 MR. SIESS: Suppose you used a metal O ring on the 7 equipment hatch. Would it keep the local leakage down if you 8 still have regulations on that? And about what membrane 9 strain would that start to leak?
10 MR. COSTELLO: I'm afraid that's too complex a 11 question for me to try and wing off the top of my head.
12 MR. ASHAR: Let me address that question. If PWR 13 conditions, as I will show you later on, probably you would
[V~')
, 14 not need to go that far to use the metal rings --
15 MR. SIESS: I'm talking about an equipment hatch.
16 MR. ASHAR: Even for an equipment hatch in a PWR, it 17 won't be necessary. I will show you by the seal and gasket 18 test we have performed. For a PWR, that is one of the 19 possibilities that we should look into.
20 MR. SIESS: You said the BWR drywe,. head wouldn't?
21 MR. ASHAR: Oh, yes, BWR is a different thing we l
22 should think about, I'm saying.
23 MR. SIESS: All right. Now I was talking about the l l
24 equipment hatch for day containment. 14 to 20 foot hatch. I 25 MR. ASHAR: Yes, okay. For that, I don't think it I
a 445 1 will be necessary to go to that type of failure break. It is 2 an elaborate system, but I'll show you that it is not of 3 concern.
4 [ Slide.]
5 Now back to the drywell head analyses. The CPWG, 6 which is essentially NUREG 1037, shows that there will be 7 unseating of drywell head due to internal pressurization at, 8 as I pointed out, approximate 86 psig or something. And then 9 substantial amount of leakage in the Mark I and Mark II 10 containments.
11 However, the conclusion is based on analysis that 12 did not take into account the effect of elevated temperatures 13 which are also present in the accident.
14 Argonne National Lab and Sandia have conducted 15 independent analyses to show that these elevated temperatures 16 inside the containment are considered in the analysis.
17 Metal-to-metal contact will be maintained between the sealing 18 surfaces, thereby precluding leakage.
19 Differential thermal expansion between the drywell 20 head flanges, which house the seal, and the bolts which attach 21 the drywell head to the lower part of the containment, have 22 the effect of additional preload on the bolts, thereby l 23 substantially increasing the pressure needed to unseat drywell l
_s 24 heads.
25 I will show you some of the sketches on this --
- j 446 l 1 MR. SIESS: I don't think you need to go into that.
s 2 MR. EBERSOLE: Is this due to the thermal gradient 3 across the metal?
4 MR. ASHAR: Yes. Basically this is due to the 5 difference between the flange of the drywell and the bolt 6 outside.
7 MR. EBERSOLE: It's like a fin?
8 MR. ASHAR: Yes. There is a fin going out there, 9 and then there is a bolting on top of it. The bolt tension 10 keeps it compressed because comparatively it is in 11 compression.
12 MR. SIESS: Are there any scenarios which would give
/~ 13 you the pressure without the temperature?
k, 14 MR. ASHAR: That also I will look into.
15 MR. EBERSOLE: Well, surely ATWS does.
16 MR. SIESS: No, you've got to have a coremelt to get 17 anything in there.
18 MR. ASHAR: There is one scenario in a BWR that I 19 recall.
20 MR. SIESS: Inside the vessel, the drywell head, not 21 the vessel head.
22 MR. EBERSOLE: I'm offbase, yeah.
23 MR. SIESS: This drywell head.
24 [ Slide.)
O- 25 MR. SIESS. The drywell equipment hatch is pressure
447 1 unseating?
/
r~%
)
(,,/ 2 MR. ASHAR: You mean the BWR equipment hatch? No, 3 they are pressure seating, most of them.
4 MR. SIESS: On the inside? .
5 MR. ASHAR: From the inside. Now a little more in 6 drywell head analyses. Argonne National Lab used a 7 sophisticated finite element approach, whereas Sandia derived 8 closed form solutions based on the strength of materials 9 approach.
10 Sandia analyzed the response of drywell heads in 11 eight containments. There were significant differences in the 12 drywell head diameter, flange thickness, number of bolts, bolt 13 diameter and bolt preload, which is quite important, within
%_/
14 this group.
15 Only one of the eight unseated at the maximum 16 accident pressure of 120 psig, and the gap was less than 2 17 mils, which would most likely be closed by the seal material.
18 There are significant conservatisms in Sandia's 19 analysis, and the actual separation pressure is probably 20 higher.
21 Based on the analysis, the test of the drywell head 22 was not felt to be necessary, so we have been able to 23 eliminate that elaborate test.
24 [ Slide.]
25 Pressure unseating equipment hatches. A pressure
e 4@
1 unseating equipment hatch will be tested in the 1-to-6 scale
( ,/ 2 reinforced concrete model I have shown you there will be one 3 equipment hatch model in the reinforced concrete model which 4 will have two, an inner cover and an outer cover.
5 The outer cover will be tested first as a pressure 6 unseating type of cover. The same methodology used to analyze 7 drywell heads will be used to analyze the pressure unseating 8 hatch.
9 Data from the test, gap measurements, compression in 10 the flanges, will be used to assess the method.
11 [ Slide.]
12 Personnel air locks. Tests will be conducted in FY 13 '86. The air lock has been procured. The RFQ will be going a
)
14 out to the public somewhere next month or the month after, as 15 I understand, for test facility, where the testing will be 16 conducted.
17 The preliminary analyses suggest that no significant 18 interaction occurs between the shell and sleeve, and the door 19 and bulkhead will undergo locked deformations associated with 20 bending.
21 That is the basic reason for conducting a separate 22 test for the air locks.
23 The first thing, there is no significant interaction 24 between the shell and the sleeve of the air lock, because the 25 air lock sleeve is quite a bit protruding out of the shell
e 449 1 liner or in the steel containment and steel face.
2 MR. SIESS: It seems to me there has to be an 3 interaction between the shell and the sleeve, but the effect 4 on the sleeve doesn't carry out to where the doors are.
5 MR. ASHAR: That is what it is, yes.
6 Finite element analyses will be performed with the 7 actual material properties and a leakage criterion will be 8 proposed prior to the test.
9 MR. SIESS: So it is just the bending of the door, 10 MR. ASHAR: And the bulkheads, yes.
11 [ Slide.]
12 MR. SIESS: What about the doors that have -- I
[~ 13 don't know what you call it, a mechanical lock where it k.
14 actually rotates and slides in and locks it tight? Are those 15 considered?
16 MR.-ASHAR: In the personnel air lock?
17 MR. SIESS: Yes, in the personnel locks.
18 MR. ASHAR: All these personnel locks have a hinge 19 on one side and they have been bolted on the other side by 20 bolts.
21 MR. SIESS: Yes, and what holds them shut? What 22 squeezes that seal?
23 MR. ASHAR: I will show you the sketch of the air i
24 lock later.
O 25 MR. SIESS: Okay.
l
450 1 MR. ASHAR: Test on seals' materials. As a result
(
b)\ 2 of the survey that was done on the 48 containments, these are l
d 3 the four materials of importance, but the most used materials 4 are the silicon rubber and the EPI /EPDM material, which is --
4 5 I don't remember the whole name. It's a polymer, some kind of i 6 a-polymer material, with specific material properties on it.
4 7 Viton has been used on a few plants in some of the i
8 penetrations, and unspecified rubber or neoprene has been used l 9 on some of the older plants.
]
l 10 We will be doing some tests, trying to get samples l 11 from those plants.
12 MR. SIESS: These are for both kinds of locks?
13 MR. ASHAR: All types of operable penetrations.
14 MR. SIESS: I thought some had inflatable seals 15 MR. ASHAR: Yes, that I have not shown, but that is 16 a different configuration. But the material will still be in 17 the same range.
18 Cross section geometry. There are O rings, dog ears 19 and gumdrop and tongue-in-groove type.
20 I might show some of them. I might have some slides 21 to show them. j
! l 22 MR. BENDER: Some of these seal materials are more ,
i 23 temperature-sensitive than others? l 24 MR. ASHAR: That is correct. I will show you some 25 slides on how bad they are.
-. ~. _ _
l 451 l
- l. 1 [ Slide.3 O 2 Tests on seals. Environment. The seals will be
( ,/
l l 3 aged and unaged. Both will be tested. Aging will be thermal 4 and radiation exposure before test.
i j 5 Now on the tests that have been performed at Sandia, 6 they will undergo both types of aging, radiation and thermal, 7 but the tests which are being performed at Argonne National 8 Lab will only undergo thermal aging, because they don't have 9 the facility to do the radiation. But we will be able to get l
10 a simulated comparison as to what the effect of various aging 11 is.
12 Pressures will be up to 160 psia. I will be showing l
I 13 you the sketches on the pressure and temperature later on.
14 [ Slide.:
15 The test set-up pressure chamber. I will be showing 16 you the pressure chamber, 17 Saturated or superheated steam, heated air or 18 nitrogen, 19 Sandia in all cases will be superheated steam.
20 Sometimes they might use saturated steam just to get the 21 comparison as how it will influence the material properties.
l 22 At Argonne National Lab, they are only using l
l
l l
24 The mating surfaces, metal-to-metal contact, O)
\'"' 25 prescribed gp, rotational surface, machine-inclined.
'. .-___.._-.._..,_,___._..,r. ., . . _ _ _ _ . _ _ . _ , _ . , . - _ . . _ _ , . _ _ . - , _ _ , . . _ . . . . . - - - _ _ _ _ . . _ . . _ . . . - . , . . . , _ _ _ _ _ _ -
e
' 452 1 Metal-to-metal contact, there will be another prescribed gap, 2 but the gap will depend upon 48 mil or 10 mil, and the 3 rotational one-surface.
4 This is to simulate the potential rotation that that 5 particular mating surface would undergo because of the shell 6 deformation.
7 There will be three rotations that we will be 8 examining.
9 MR. SIESS: On the door deformation? Not just the 10 shell? On equipment hatches it is door deformation, isn't it?
11 MR. ASHAR: Yes. Yes, that is correct.
12 [ Slide.3
(}
V 13 Future work. Complete tests on seals, investigate 14 effect of aerosols on leakage, test personnel air lock, 15 equipment hatch, test reinforced concrete model These are 16 the things I already mentioned.
17 [ Slide.]
18 Let me show you some of the sketches which might be 19 of interest to you.
20 This is the test chamber at Sandia National Lab.
21 This is the whole chamber, and these are the three lengths of 22 fixtures in the seal stage. I will show you a separate sketch 23 of this particular assembly, but this will be exposed to the 24 severe accident pressures and temperatures inside, like a 25 pressure seating kind of condition.
e 453 1 MR. SIESS: And leakage will be measured?
/
( 2 MR. ASHAR: The leakages will be measured.
3 [ Slide.]
4 Let me show you the profiles under which the seal 5 tests are conducted. This is a PWR accident scenario. The 6 pressures will be going from here to here (indicating) at 155 7 psia, will be PWR conditions, temperature will start from 293 8 or something and 361 degrees will be the highest temperature 9 maintained for PWR conditions. And these are enveloped --
10 MR. BENDER: With regard to the air locks, somebody 11 pointed out once that the way these air locks are designed, 12 the outer air lock -- there is a chamber between the inner air
/~' 13 look and the outer air lock which reaches a temperature 14 barrier. Does that influence what is up there?
15 MR. ASHAR: These temperature-pressure profiles are 16 containment air profiles. They are not even at the site of 17 the containment, which is lower than what they are shown here.
18 MR. SIESS: That's what the inside door will see, 19 and what the outside door will see if the inside door failed.
30 MR. BENDER: Well, if it failed -- even if you get a 21 leak there, if it's stagnant, they'll have a big insulator 22 there. I am just asking because it may have an influence --
23 MR. SIESS: You could get the pressure on the <
l t
24 outside door, but you are not likely to get temperature.
25 MR. EBERSOLE: I would like to raise a question J
454 1 about these air locks, including both the equipment as well as
/
k 2 personnel. I think they are an anachronism.
3 MR. SIESS: They are there, Jesse.
4 MR. EBERSOLE: I know, but they look like submarine 5 air locks, and we have long since departed from the thesis a
6 that we don't have a moment in time to exposure ourselves to 7 an open containment. As a matter of fact, containments for 8 years ran with open purge valves that you couldn't even shut.
9 32 inch butterflies. They couldn't even shut them against an 10 inflow from a LOCA. There had to be revolution that had to be 11 imposed on NRC to get those things fixed.
12 So we have run, therefore, in the presence of wide 13 open containments blowing air through them that didn't have a 14 ghost of a chance of shutting up for a LOCA. In the meantime, 15 we go to the ridiculous opposite extreme of having to double 16 ourselves through this lock, thus saying I can't expose myself 17 to an open containment for five minutes while I go through.
18 MR. SHEWMON: When the Westinghouse advanced reactor 19 subcommittee chairman gets through with that, I hope they have 20 something in it about better containments that don't have 21 these problems. i I
22 MR. EBERSOLE: Well, I just want to recognize the l 23 curious contradictions in our logic.
24 MR. SIESS: Well, there's a proposal from RDA for an N_J 25 unpressurized containment. It just vents out through a
l 455 1 system. You would like that one.
2 [ Slide.]
3 MR. ASHAR: This is a BWR profile for pressure, 4 which the maximum is 135 degrees.
5 [ Slide.]
6 This is a BWR profile for temperature, which some 7 people say is not conservative enough. Some of the studies 8 have shown the temperatures as high as 1000 degrees, but the 9 consensus showed that 10 MR. BENDER: Did you say that's not conservative l 11 enough?
12 MR. ASHAR: Yes. We heard from the IDCOR study ,
I 13 using some code, I don't remember, and Oak Ridge did some 14 studies which showed that the temperatures could be much 15- higher in certain severe accident sequences than 700 degrees.
16 MR. SIESS: Is this hydrogen burning?
17 MR. ASHAR: No, 'there's no hydrogen burning; just )
l' 18 core-concrete interaction giving very high amount of energies 19 and temperatures.
1 20 MR. SIESS: Because there is an issue, isn't there, 21 about hydrogen coming out of the wetwell and burning as it 22 escapes, in the MARK III?
23 MR. ASHAR: Yes. I don't know the exact reasons, 24 but some people do think the temperatures will be higher in 25 certain BWR severe accident scenarios.
l l
e 456 1 But we had modified this particular chamber from 500 l'"
2 degrees to 700, and we are not ready to do 700 to 1200 degrees 3 because it is quite a job to change the chamber from one to 4 another. And everyone said that this would give us enough 5 idea about the material properties. And we will see from the 6 results that it does.
7 [ Slide.]
8 MR. RICHARDSON: But Hans, how will you eventually 9 answer the question of these people that are claiming 1000 10 degrees? That would not be an acceptable answer, that our 11 test fixture wasn't high enough to get it.
12 MR. SIESS: Have to look at the probabilities, Jim.
13 MR. ASHAR: You will see in some of the results on 14 seals and gasket that by 700 degrees, those things deteriorate 15 to such an extent, it doesn't matter 800 degrees or 1000 16 degrees.
17 MR. SIESS: Yes, but at 1000 degrees I'd start 18 worrying about the steel.
19 MR. ASHAR: That's correct. Yes.
20 [ Slide.]
21 These are the assemblies for the three parameters.
22 We wanted to see the scale effect as to the parameter is 23 larger, is the leak rate larger. That's why we've got three 24 separate diameters here. And these are the channels through 25 which the leakage will be munitored. It gets into one
E e
457 !
1 particular place and --
k 2 MR. SIESS: Where is the test specimen? Where are 3 the seals you're testing.
4 MR. ASHAR: The seals are here in the top set.
5 MR. RODABAUGH: How long are you going to run these 6 tests? How many hours? The length of time?
7 MR. SIESS: You saw the curves.
8 MR. RODABAUGH: But I didn't make a note of it.
9 Twelve hours or something like that?
10 MR. ASHAR: They will be going for around 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> 11 or so at the highest temperature. First, at the PWR pressure 12 and t erap e r a t u r e , they will be maintained at 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, and then 13 it will be raised to BWR temperatures.
)
14 MR. RODABAUGH: Okay. So at 700F in 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> you 15 wouldn't expect, for example, neoprene to be anything but a 16 piece of liquid.
I 17 MR. ASHAR': That's correct.
18 MR. SIESS: Incidentally, there arer test data 19 available on seal materials, independent of these tests that 20 they're using I think to decide on those lengths. Am 1 21 right? I remember seeing great big sets of curves on seal 22 materials.
23 MR. COSTELLO: Oh, yes. They are basic supplier 24 material tests.
25 MR. SIESS: So you know this is long enough.
4
- - - - - - . + ~ - , , , , ,.w - - -
.,-.-.-n.- , --,..-,.-,..-c-w - ,, --,n, ,e , , -. ,,_,_nn-.---._n.,- _-, ,.,- ,, n m - - . , , , . . . - , --n.__,
458 1 Mr. Rodabaugh says it's probably much longer than it needs to
' b..
(,
3 MR. RODABAUGH: Yes. I was just wondering what you 4 are going to get. I was trying to imagine. If you've got 5 something like neoprene in there at 700F for 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, that 6 would be a liquid.
7 MR. COSTELLO: The real question here is, are there, 8 in the whole panoply of geometries and material combinations, 9 the first question is are there any surprises about the seal 10 behavior beyond handbook kind of stuff.
11 The second question is, can you account for size 12 effect in the different diameters and different perimeters.
13 That's the second question.
G' 14 MR. RODABAUGH: Okay. In that test setup, were you 15 able to measure leakage in between -- it looked like you had 16 four rights on that top set.
17 [ Slide.3 18 MR. ASHAR: There are two rings in each of the 19 sections.
20 MR. RODABAUGH: Okay. And on what pair -- are you 21 going to measure leakage across pair rings?
22 MR. ASHAR: They will represent something like a 23 double dog ear, or --
24 MR. SIESS: Where are you going to measure leakage?
\
25 You've got two rings. Are you going to measure both --
l
e 459 1 MR. ASHAR: In between channels; after the first
,nm 2 seal, there is a channel, After the second seal there is 3 another channel, so both will be measured to see what goes 4 from first to second, and then how it transpires.
5 MR. RODABAUGH: What I'm thinking is some of those 6 materials at 700F the gas will be pushing through this liquid 7 gasket --
8 MR. ASHAR: I'll be telling you some of the 9 results. I don't have the slide for some of the results we 10 have seen so far.
11 MR. EBERSOLE: May I ask, can the outboard door be 12 justified on the basis it will be out in a cooler
/" 13 environment? That's the only reason I could justify it.
14 MR. ASHAR: For BWR you are thinking about?
4 15 MR. EBERSOLE: Well, any containment.
16 MR. SIESS: Two-door personnel hatch.
17 MR. ASHAR: Yes, okay.
18 MR. SIESS: The outside door will not see the same 19 temperature. It's the same problem you went through with the 20 electrical penetrations.
21 MR. ASHAR: Absolutely, that's correct.
22 MR. EBERSOLE: Well, in that case it is useful.
23 Otherwise, it isn't.
24 MR. SIESS: It just may turn out to be.
\- 25 MR. BENDER: It's a way of getting some use out of
_ . ~ . . _ . , _ . . _ _ -
i-460 1 it.
f~
( 2 MR. ASHAR: Now let me give you some of the results 3 just from what I remember. For silicone rubber material --
4 MR. SIESS: Let's don't get into results. The main 5 purpose here is to talk about what you're doing and why.
6 We'll explore the results at another meeting.
7 MR. ASHAR: Okay, very good.
8 MR. SIESS: What about the bellows, the pipe 9 penetrations?
10 MR. ASHAR: They are being studied at this time.
11 MR. SIESS: And the mode of failure that you're 12 looking for?
13 MR. ASHAR: In the bellows? Is the excessive N,
14 compression or twisting.
15 MR. EBERSOLE: What about the big purge valves?
16 Those are big things.
17 MR. SIESS: That's another program.
18 MR. ASHAR: There's another program where purge 19 valves are being tested.
20 MR. SIESS: And we ought to look at that sometime l
21 but that's not in Research and it's under the cognizanoe of, I 22 guess, the Containment Performance Working Group, isn't it?
23 MR. ASHAR: Purge valves are under the research.
24 MR. RICHARDSON: That is under our Equipment s 25 Qualification program. They are being tested at -- l l
l
461 1 MR. SIESS: I'm talking about the whole problem of
( 2 pre-existing leakage. Are the big valves a part of that? If 3 they open, they open.
4 MR. RICHARDSON: Yes, right.
5 MR. SIESS: And how much they leak under pressure is 6 something else. But purge valves are the prime suspect 7 because they open at both ends, don't they? Containment 8 atmosphere at one end and --
9 MR. RICHARDSON: You mean normally open?
10 MR. SIESS: No. The other so-called isolation 11 valves that don't open at containment and don't open into the 12 atsmophere. Those valves open at containment atmosphere at 13 one end and the aux building at the other --
Os 14 MR. EBERSOLE: I hope it isn't the aux building.
15 MR. SHEWMON: Who is responsible for that program?
16 MR. RICHARDSON: For the valve penetration? That is 17 under our branch and it is in our Equipment Qualification 18 Program. But we're also going up to severe accident 19 conditions.
20 MR. EBERSOLE: By the way, what Chet said is 21 certainly not true, is it? Those purge valves don't ever open 22 into the aux building, do they?
23 MR. SIESS: I don't know where they open to.
24 MR. RICHARDSON: I don't know.
- 25 MR. EBERSOLE: If they do, that ought to be hammered
a 462 1 out right now and fixed. I hope they open into the annulus or 2 the outdoors or something. If they open into the aux building 3 --
4 MR. SIESS: Don't they open into some filtered area, 5 or do they purge directly --
6 MR. EBERSOLE: I think they open into the filtered 7 -- the secondary system. Secondary filtrated, you know, 8 secondary containment. But if they open into the aux 9 building, Chet, we have got a regressive consequence 10 instantly.
11 MR. SIESS: Jim, could we talk about the EPRI work 12 in a very short length of time? Let me introduce it and tell
/~N 13 the other people what I think they ought to know about it, ts -)
14 And correct me if I leave something out, 15 EPRI is having some tests made at the Construction 16 Technology Laboratory, the Portland Cement Association. They 17 built a testing rig that can suggest or subject a slab of 18 concrete of full containment thickness to biarial in-plane 19 tensions like it would see under pressure, and to a transverse 20 load that it could see on a penetration, and can measure 21 leakage.
22 In other words, it can stretch it in two directions 23 with reinforcement in it with a liner on it, and it could be 24 full scale. So they are going to be making tests on sections 25 of a containment wall to look at what might happen through the
a o 463 1 wall or through the penetration; they can imbed penetrations.
(
l Now, is that reasonably correct?
2 3 MR. COSTELLO: I think that's a fair estimate', yes.
4 MR. SIESS: And there's no modeling. They have done i 5 some work on smaller-scale models in another area and some in 6 this area, but this is their potential.
4 7 Now, are you involved? You are following this, I'm 8 sure.
9 MR. COSTELLO: Yes, we follow it. We are also 10 involved to the extent that usually, Dr. Tang from EPRI or i
11 Dr. Wahl from EPRI does appear at our meetings of our peer
! 12 review panel i
13 MR. EBERSOLE: Is this the Dr. Tang that used to be
\,
14 up here?
15 MR. COSTELLO: No, it's H.T. Tang, used to be G.E.,
16 I believe. But EPRI does come and chat with our peer review 17 panel and lets them know of progress and that way, fill in 18 ourselves and the Sandia staff.
19 We have a rough idea of their test plan and in 20 the last iteration they have been taking the benefit of 3
21 comment from our peer review panels to help them think about 22 what they're going to do in the outyear.
l 23 But their basic. approach is to focus on the 24 possibility of premature liner failure, the so-called, as I 25 call it, the Rashid hypothesis.
464 1 MR. SIESS: Okay. That's really what they're
, 2 looking at; to see when the liner will fail in relation to the 3 concrete behavior. .
4 MR. COSTELLO: Yes. With the suspicion that --
5 MR. SIESS: So that if you developed a premature 6 liner failure in your concrete model, that work could be used 7 to expand it into looking at how the anchors might have 4 8 affected it and so forth.
9 MR. COSTELLO: Yes, sir.
10 MR. SIESS: If you don't develop anything in your 11 concrete model and they do in their full scale model --
12 MR. COSTELLO: It should be interesting, yes.
13 ,
MR. SIESS: Okay. And if neither one of you s
14 develops it, we quit.
15 MR. COSTELLO: Maybe.
i 16 Now the other thing that they are attempting to do 17 and it's rather difficult and it's not clear yet that they 18 will pull it off, is to do a modeling of the basemat-wall 19 intersection that we talked about as being so interesting 20 before.
l 21 MR. SIESS: And the same kind of answers go.
22 MR. COSTELLO: With something less than a complete 23 wall with a segment. They have had difficulty -- and with 24 loading by a bladder. They have had difficulty in getting the 25 loading on in a couple of tries, but that should give us some I
a 465 1 insight --
2 MR. SIESS: [ Inaudible.]
3 MR. COSTELLO: That's something I haven't explored 4 with them.
5 MR. SIESS: There was something said about possibly 6 testing a 360 degree ring of something. What's that supposed 7 to show?
8 MR. COSTELLO: I am not sure that that's an EPRI 9 proposal or a PCA proposal.
10 MR. SIESS: I saw it in a report that EPRI 11 distributed.
12 MR. COSTELLO: We have certain difficulty in getting 13 EPRI plans in the same frame as we are. You see, we're 14 talking about the 1987 budget. As of now, EPRI is just l 15 firming up their 1986 budget, so we have a timescale 16 difference. But we hope that they are able to bring off this 3 17 basemat wall test and get something out of there that will be 4
18 beneficial to us in the model.
19 MR. SIESS: You had questions about scaling of your 20 shear connectors, and that is not unrelated, I guess, to 21 Rashid's problem and what they are looking at there. Will 22 they do anything, varying the shear connector spacing?
23 MR. COSTELLO: I believe the series of tests they 24 have do have different shear connector spacings and different
\
25 imbedments, both studs and channels.
1 l
{
e 466 1 MR. SIESS: Okay.
( 2 MR. BENDER: Jim, there is another variable of this 3 thing that probably they should look at if they haven't. And 4 that is that the conformance of that shell basemat connection l 5 to the concrete backup or whatever it is they've got there 6 varies all over the map, depending upon who has designed and 7 who built it and when it was built.
8 And I wonder if just one model without some 9 understanding of what the performance is going to be of the 10 steel shell and the concrete will wind up being all that 11 meaningful 12 MR. COSTELLO: I think in many cases, the liner was 13 used as one of the forms, right?
14 MR. BENDERt In some cases. But even when it was 15 I think you will find that there is some uncertainty about 16 whether the concrete really was up a2ainst the steel liner.
17 MR. COSTELLO: That's one of the things we're going 18 to be looking at in our tests.
19 MR. BENDER: They may have a very good setup for 20 their tests but the real skills may be a lot different, 21 MR. EBERSOLE: I thought that had to have about an 22
- inch or so of some sort of flexible membrane because of the i
23 differential expansion suddenly occurring, the ,
l 24 suddenly-occurring thermal profile when you swell up the liner 1
\- 25 against the concrete, and you had to leave room for it or
e 467 1 otherwise it will surely buckle before the concrete heats up. i 6
2 MR. SIESS: No, the stud spacing is based on 3 buckling.
4 MR. EBERSOLE: Isn't there an insulation pack in 5 there, though?
6 ME. BENDER: There are all sorts of variations in 7 that connection.
8 MR. SIESS: Has anybody looked at, you know -- has 9 anybody looked at the variation of details in those corners?
10 MR. COSTELLO: I don't know for sure. I do know 11 that the design was chosen based on being representative, but 12 I think we would be well advised to go back and look. i l
13 MR. SICSS: Yes, I think somebody ought to look.
14 MR. EBERSOLE: I know one containment has insulation 15 and I'm not so sure that it isn't Sequoyah.
16 MR. SIESS: That is a steel shell.
17 MR. EBERSOLE: Well, it's not that then, But I 18 remember one, they had trouble with it, Chet, this insulation.
19 MR. SIESS: That's another problem, you see. We 20 tested a steel shell with a steel bottom to it and the steel 21 shell sitting out there anchored, down to a concrete base and 22 that's quite a different story.
23 MR. EBERSOLE: Well, there's sore containment that 24 has this flexible, thick liner.
((() 25 MR. SIESS: Well, some of the steel ones do because l
e i 468 1 they had to allow --
l 2 MR. EBERSOLE: There was a lot of trouble about it 3 not always being, you know, fully cleared of concrete, too.
4 MR. SIESS: Well, steel shells have got some s
S flexibility, they will come away from the concrete but that's 6 not the concrete liner; that is the steel shell. And that 7 concerned me because I thought the steel one was going to fail 8 at the base.
9 Your steel one didn't have a bottom head, it had a 10 flat bottom, didn't it?
11 MR. COSTELLO: No, it had an elliptical head.
12 MR. SIESS: The steel ones vary quite a bit in that 13 detail where the steel and the concrete connect. There's a O%
14 concrete floor in there and ? figured that was going to be the 15 weak spot and then we were going to have another problem.
16 But you know, I don't see why, in the reinforced 17 concrete containment, the phenomenon is a local bending in 18 that joint. I don't think it is three-dimensional even. And 19 if you are going to explore it, I don't know why they just 20 can't do that -- they can't look for leakage but they could 21 look for the mechanics of how the thing fails. At some point, 22 you've got to quit worrying about leakage and just look for 23 cracks in the steel and assume if it cracks it'sigoing to leak g 24 through it.
25 MR. EBERSOLE: I recall a detail, Chet, where at one
l
\
l 469 2 l
1 point the steel actually was poured against the concrete but l
f'
( ,}f 2 as it turned vertical is where the insulation tore. So there 3 was a joint where a transition took place.
4 MR. SIESS: I think you're thinking about what some 5 people call the hybrids; the steel containment on a concrete 6 base. It's got a steel bottom to it and that is for 7 leaktightness, and it is not a code vessel. And those are the 8 ones that have a problem.
9 Gentlemen, I think it is time for lunch. I will let 10 the Chairman after lunch tell you what time to be back.
11 MR. SHEWMON: How about 50 minutes. 1:30, 12 [Whereupon, at 12:43 p.m. a lunch recess was taken,
(~'\ 13 the meeting to resume at 1:30 p.m. the same day.]
14 AFTERNOON SESSION 15 MR. SHEWMON: This afternoon we will hear about 16 containments of concrete, steel and piping. As I understand 17 it, gentlemen, this will deal with the extension of the leak 18 before break ideas into other parts of the primary system.
19 Without any other introduction I will turn it over to Gene 20 Kurts to talk to us.
21 MR. KURTZ: Thank you. I would like to thank the 22 subcommittee for giving us this opportunity to speak this 23 afternoon. We would like to keep the subcommittee abreast of 24 an effort that we have ongoing right now with NRR staff, which
~
O* 25 has currently been received quite well by the staff and
e 470 1 management of NRR. I am referring to the Whipjet program 2 which we are pursuing which we believe will improve plant 3 safety while reducing plant costs not only during the 4 construction phase but also continue to reduce costs 5 throughout the plant life. And-to explain the technical 6 aspects of it, I'd like to let Bob Cloud discuss the issues.
7 MR. SHEWMON: What is the name of the program?
8 MR. KURTZ: Whipjet is the name of the program that 9 Bob Cloud has blessed the endeavor with.
10 MR. CLOUD: Thanks, Gene. I was doing some work at 11 the Beaver Valley Power Plant last winter in a review mode 12 looking at the methods and the approaches being taken by Stone 13 & Webster in the conduct of the engineering qualification of 14 the piping systems.
! 15 As part of that, I realized following the trends 16 that were going on in the industry, I realized that Beaver 17 Valley would be in the situation in a year or 18 months where 1
18 they would be installing their whip restraints just about the 19 time that the industry decided that they weren't necessary at 20 all, which struck me as an unhealthy situation.
21 MR. MARK: Are you implying that Beaver Valley has 22 never had them?
23 MR. CLOUD: No, they have always had them, but they 24 are building them, they are welding them in.
25 MR. SIESS: Is this Beaver Valley 2?
I 471 1 MR. CLOUD: Two, I'm sorry. I should have said 2 Beaver Valley 2. It's a plant under construction.
3 MR. MARK: Oh, I was just wandering in my mind.
4 MR. CLOUD: Well, that's quite all right.
5 So to that end, we thought that it should be 6 possible to put together an engineering approach to the 7 problem of pipe break protection that would hopefully be an 8 improvement on the way it was planned to be done. And that's 9 what we have done and that's what I would like to explain to 10 you today.
11 [ Slide.]
12 Our program -- this is a contraction of the terms
(' 13 whip restraints and jet shields. And we have labled it that 14 to give us a reference point.
15 Fundamentally, what we are proposing, we are 16 proposing an alternative approach to the engineering for the 17 pipe break protection. And very specifically, we wanted to do 18 two things. The first is -- in the first place, I should 19 mention that the leak before break -- the fracture mechanics 20 technology -- has advanced in the last ten years from the 21 elastic fracture mechanics or the so-called linear elastic 22 fracture mechanics to the point where now it works very well, 23 quite adequately in the elastic plastic regime. And I think 24 it is non-controversial to say that that is accepted on a O 25 technical basis, that the method works when it's applied in
e 472 1 applicable situations.
2 Our approach is to through an engineering program, a 3 program of analysis and testing, to determine those places 4 where the leak before break approach is appropriate for the 5 balance of plant systems. When I say balance of plant, I am 6 talking about those systems other than the primary system.
7 Now also I should mention that Beaver Valley 2 has 8 requested, as have many others, relief from the so-called 9 arbitrary intermediate breaks, and I believe that this is in 10 the works and we'll talk about those s ocie more in just a 11 moment.
12 In those places where the leak before break approach 13 can be shown to be appropriate, then we will implement it.
14 That is to say we will perform those analyses and that testing 15 that's required to prove that the pipes will leak before they 16 will break and that they will leak in a detectable manner, and i
17 then we will review the leak detection capabilities as well.
1 18 An extremely important aspect of our program and our l l
19 approach is that in those areas where the leak before break 20 doesn't work or for some reason it's inapplicable or 21 inappropriate, we'll do the work in the conventional way. In 22 those cases where -- let's see, I guess I should mention that 23 most of those areas would be related to reasons that are f_
24 delineated in the recent Reg Guide 1061; they are laid out in 25 Volume 3.
e 473 1 Further, in the second --
2 MR. SHEWMON: Is that a NUREG or a Reg Guide?
3 MR. CLOUD: I'm sorry, NUREG. My mistake.
4 Another very important aspect of our work is that we 5 are attacking the issue of the dynamic effects on the pipes 6 and protecting the plant from that phenomenon alone, and all 7 those other aspects of the plant design that are related to or 8 based upon the assumptions of the double-ended break we are 9 leaving unaltered. In no way, for example, are we 10 re-examining any of the equipment qualification, nor 11 compartment pressures, nor containment pressure, nor ECCS 12 assumptions, nor any of those other aspects of plant design T 13 related to or based upon the assumptions of the double-ended 14 breaks.
15 I wanted to go over this very slowly and very 16 carefully because we talked with the staff about this two 17 weeks ago and we got a lot of questions that suggested to us 18 that we went over this slide perhaps a little too fast.
19 MR. E B E ftS O L E : Is this to say that the only thing 20 you intend to do physically then is to take off some pipe whip 21 restraints?
22 MR. CLOUD: I'm sorry?
23 MR. EBERSOLE. Take off some pipe whip restraints, 3
24 is that the objective?
25 MR. CLOUD: Our objective is to show that the leak
1 e
474 1 before break analysis will work in the balance of plant.
2 Assume pipe breaks, then apply that approach and do not 3 install whip restraints wherever the leak before break 4 approach works. Yes.
5 MR. EBERSOLE: But is it just in the context of 6 removing the pipe restraints?
7 MR. CLOUD: That's exactly correct.
8 MR. EBERSOLE: If you squirt water across 15 feet 9 and hit an electrical box you will retain that thesis? Is 10 that a leak?
11 MR. CLOUD: We have two kinds of break protection 12 that we are addressing here. The whip restraints which are
[ \ 13 supposed to physically restrain the pipe, and the second means 14 of protection are the so-called jet shields which are 15 protective covers placed over presumably vulnerable 16 equipment. Those two features are being addressed in our 17 p'rogram.
18 MR. EBERSOLE: And is this for just primary piping?
19 MR. CLOUD: No; for everything except primary 20 piping. Primary piping is already being addressed --
21 MR. EBERSOLE: Oh, this is for everything except 22 primary?
23 MR. CLOUD: That's exactly correct; for the balance 24 of plant systems. Everywhere else in the plant where we have 25 high energy lines and whip restraints related to them.
e 475 1 MR. BENDER: Bob, I think you ought to reiterate 2 again that it's where you can show that this concept is 3 justifiable that you are arguing for the elimination. Not 4 proposing universal elimination at this stage.
5 MR. CLOUD: Yes, exactly. I want to emphasize 6 that. I want to reiterate it. I want it absolutely clear 7 that we are not asking for any favors, we're not asking for 8 any relief, we are not asking for any exemptions. But instead 9 what we are proposing to do is to do a serious job of looking 10 at the engineering involved in the postulated breaks, and when 11 it's possible, to apply the leak before break which we think 12 is a technologically improved means. Then we want to change h 13 away from the old-fashioned, well, blindly installing whip 14 restraints in the hopes that it will do something.
15 MR. RODABAUGH: Bob, I would like to follow up on 16 Jesse's question. I think it's a key point here. Your 17 analysis will say that it's going to leak before it breaks.
18 It could be a very big leak. Now you don't have the shields 19 anymore. Just how big a leak is a leak?
20 MR. CLOUD: Yes, I understand that. Please know I
21 that we have not completed our program; we a r e ,.l u s t talking l 22 about doing it. We are here talking with you today, and we 23 are not proposing to develop a situation that will expose 24 sensitive equipment and then increase the risk of hazard.
(A}
N/ 25 So if we have that situation, then that would
476 1 definitely fall in the category of a jet shield that we know b 2 does some good, in which case we would leave it.
3 MR. RODABAUGH: Well, let me phrase it in terms of a 4 regulatory guide or SRP's. Right now, you have in a sense 5 postulated that you are going to get a split everywhere, for 6 the purpose of equipment qualification, spraying on cabinets 7 and so forth. Is your program -- or do you anticipate that l
8 you will change that part of it?
9 MR. CLOUD: I tried to say it before and I'm glad 10 you asked the question because I want it to be crystal clear.
11 Any equipment qualification that is required to be done before 12 our program will be done after our program. Our program in no 13 way, manner, shape or form affects the qualification of 14 equipment, whether it does or not.
15 MR. RODABAUGH: So you would assume in the case of l 16 Jesse's question that yes, you will have a leak, that it will 17 spray on an electrical cabinet nearby.
18 MR. CLOUD- If that cabinet is presumed to be 19 sprayed on today, or before our program, it will be presumed 20 to be sprayed on after o sr program, whether it will be or not.
21 MR. RODABAUGH: Okay, thank you.
22 MR. EBERSOLE: That's whether it's got a shield or 23 not.
24 MR. CLOUD: Yes.
25 MR. SHEWMON: Why don't we assume you have covered
__ . =.
. l 477 l 1 that slide, and we may return to it but let's go on. Maybe
) 2 you can show us what you're doing.
3 MR. CLOUD: Okay, splended.
4 The other thing is this. I want everybody to be 5 aware of a problem that we have and we are talking with the 6 NRR about it, and Mr. Bosnak is here and he can elaborate on 7 it. But we do have a bit of a licensing issue in the sense 8 that Beaver Valley 2 will come up for their license at what, 9 early 1987, and in order for us to do this program -- we can 10 do it, we can begin now, we can do it in the next 18 months or 11 16 months, 14 months, whatever. But then we will come into a 12 situation where it will be time for Beaver Valley to have
/~} 13 their license awarded and they will be -- there will be some 14 whip restraints that may not be installed because we are not 15 going to know which ones need to be installed until our 16 program is done.
17 We think, as a matter of actual fact, that there 18 will be very, very few that will be required. But still and 19 all, the situation will be that there will be time to load 20 fuel and operate the plant and there will be some whip 21 restraints not installed, and it will be necrssary for the l
22 power company to get some assurance that it will be okay to i
23 install those in the refueling outages that follow. l
,_s 24 We feel that that is a perfectly justified -- the t \
\,,) l 25 In fact, we think it is a preferable '
position to be in.
- l l
478 1 position to be in than the alternative. The alternative being h
2 to close our eyes and blindly install the whip restraints. We 3 feel that there will be very few whip restraints required.
4 In the second place, we feel that there will be very 5 low risk of operating in the first cycle or two or three early 6 in life with whip restraints absent because the initiation and 7 growth of cracks does, in fact, require time.
8 Secondly, any construction errors we think will have 9 been uncovered during the shakeout; that is to say, in the 10 testing before the plant goes online. And in the third place, 11 I think events have shown pretty well that in general, whip 12 restraints don't work anyway.
13 And so that's basically where we stand. Excuse my
[V) 14 bluntness on this point.
15 CSlide.3 16 MR. MARK: You said whip restraints don't work. Is 17 that what you really mean, or is that you mean they are not 18 necessary or they have nothing to do?
19 MR. CLOUD: Well, I meant what I said but I think 20 what you said is also true.
21 MR. MARK: What? We have been putting these in and 22 they just don't hang onto a pipe, or what?
23 MR. CLOUD: Well, I'm thinkang, for example, 24 specifically, we had an instance in a fossile plant of a major 25 pipe rupture which, by the way, we wouldn't have had in a
a 479 1 nuclear plant, at the Mojave Station out in California three 2 or four months ago. A number of us went down and looked at 3 that. Mr. Bedoin from my group went down to see it and 4 Mr. Bosnak went down to see it.
5 And if that line had had whip restraints on it at 6 both ends, it wouldn't have helped any. All those people that 7 were injured would have been injured. The line would have 8 broken and everything would have been the same as in fact 9 actually happened. Whip restraints wouldn't have done 10 anything for us.
11 And I think in general what happens is if we are 12 going to get a crack, if we're going to get a leak, it's going 13 to happen in some way that we didn't think of anyway, in some 14 location that we didn't think of, for reasons that we didn't 15 think of.
16 MR. BENDER: Bob, in that instance it was not a 17 double-ended guillotine break?
18 MR. CLOUD: That one was a split, 19 MR. BENDER: It was a split, and so it didn't follow 20 the same mechanism that pipe whip restraints would be expected 21 to resist.
22 MR. CLOUD: That's right. That's exactly right 23 because the pipe didn't break the way it was supposed to.
1 24 That's right.
a 25 What I'll be talking about today is this is the work
480 l
1 that we're going to be doing, and first I'll begin with the '
2 scope of the problem.
3 [ Slide.]
4 At Beaver Valley 2 --
5 MR. EBERSOLE: I can't remember, what is Beaver 6 Valley? Is it a PWR?
7 MR. CLOUD: Beaver Valley 2 is a pressurized water 8 reactor, Westinghouse primary system, three-loop plant, and 9 there are two units. One of them is now operating and one is 10 presumed or hopefully will start in a couple of years.
11 On the whip restraints and jet shields, where the 12 NRC has said that they are not needed on the primary system,
( 13 we have about nine. We have also said that we need no l o,n g e r
\_e 14 consider the arbitrary intermediate breaks; there's about 15 127. In the balance of plant, those which we are talking 16 about today, there's about 136, and it is a sheer coincidence 17 that these turn out to be exactly half. So these are the ones 18 we're talking about [ indicating 3.
19 MR. RODABAUGH: Bob, by primary system, does that 20 include the surge line to the pressurizer?
21 MR. CLOUD: No. The main coolant system --
22 MR. RODABAUGH: Not the whole primary system.
23 MR. CLOUD: Yes. The main coolant loop, right.
24 [ Slide.]
25 Our first jobs are to demonstrate the applicability
e 481 1 of the technical approach, and there are four major tasks s
2 involved in doing that. We will talk about each of them in 3 detail.
4 Secondly, after it is shown that it is applicable, l
i 5 then the performance of the work itself to show that it works. i 6 MR. EBERSOLE: In respect to the size range of 7 pipes, do you go clear down right on to the instrument lines?
8 MR. CLOUD: That's a very good question. At the 9 moment, and a priori, we are making no distinction as to 10 size. We know that there has been some discussion in learned 11 circles that perhaps bigger pipes -- that the approach works 12 better, and it doesn't work so well on small pipes. And we
,/ ~N 13 understand that, and we take that as a caveat.
14 We feel that it's wrong for us to tie our hands 15 before our back before we ever get started with the program.
16 We noticed that a high energy line -- there has been an 17 instance very recently of a high energy line rupture in a 18 nuclear plant of a one-inch line at Rancho Seco, which is 19 well known to you people, I'm sure. That was a one-inch line 20 and I think it was detected very, very early and very easily 21 and very readily.
l 22 MR. EBERSOLE: Well, there have been others. They 23 blow out at the welder let. I've heard of others where they 24 blow out at the welder let. You know, where they weld it in.
O 25 MR. CLOUD: Well, you're ahead of me on that one. I
I e
482 i 1 don't know which one you are referring to.
O 2 MR. EBERSOLE: There was one on a turbine, I think 3 at Sequoyah. I'm not sure. It came out where it was attached i 4 to the primary pipes. ,
5 MR. CLOUD: But specifically to answer your 6 question, we propose to study the entire range of sizes 7 involved, and if we find that we can't show that a detectable 8 leak will occur in the smaller lines, then we will continue in ;
9 the old-fashioned way.
10 [ Slide.]
11 Our first task is the stress corrosion review. Of 12 course, it's a major problem in some plants, the business of 13 the stress corrosion cracking, and it is much less of a
(
14 problem in the pressurized water plants, although we feel that 15 in some instances it could be a problem. And our task 16 involves a specific review on a system-hy-system basis where 17 we are assessing the causative factors for stress corrosion 18 cracking including stress levels, material, the environment, 19 and putting the whole picture together. l l
1 20 If we see either from the point of view of history 21 or for other reasons, we think that we have a place where we 22 can have stress corrosion cracking in a higher energy line, i
23 then that would be a cause for us perhaps to have second 24 thoughts about taking the new approach.
25 MR. BENDER: Bob, I guess just speaking personally,
a 483 1 I don't know that much about the materials that are involved 2 in the balance of plant to know whether stress corrosion 3 cracking is a big problem or a little problem or one that 4 hasn't been thought about much. Is there some set of l
5 materials or references that are around that you are using as 1 i
6 the basis for judgment, or are you going to create these? I 7 MR. CLOUD: Fundamentally, Mike, we have three --
8 MR. BENDER: If I'm getting ahead of you I'll wait.
9 MR. CLOUD: No, it's quite all right, we can discuss I i
10 it. Basically, we're talking about three materials. We are 11 talking about carbon steel, A106, Grade B, and perhaps there 12 may be some Grade C. The carbon steels, as you know, are 13 essentially immune to stress corrosion cracking except in
(
14 severe caustic environments. So we will look for that.
15 The second class of materials is the austenitic 16 stainless 304 and the stabilized grade, 316. Those materials, 17 as is well known, in the right circumstances are liable to 18 stress corrosion cracking. The water has to be hot, there has 19 to be a chloride or other causative agent and there has to be I 20 enough temperature and there has to be some stress. So we are 21 going to watch for those situations.
22 And further, we will review the history of stress 23 corrosion cracking in plants of this nature. We think that --
l
24 in fact, if my memory is correct, there have been instances of l a 25 stress corrosion cracking in pressurized water plants. If my .
1 l
e 484 1 memory is correct, they were not in high energy systems; they 2 were, in fact, in stagnant, low energy systems.
3 But as best ! can, that is the stress corrosion 4 picture.
5 MR. BENDER: What is the third material?
6 MR. CLOUD: 304, 316 and 106.
7 MR. BENDER: The high pressure steam lines are what?
8 MR. CLOUD: They are 106 carbon steel.
9 MR. SHEWMON: Bob, there has been work on stress 10 corrosion cracking of ferritic steels in BWR environments, and 11 there was a paper given at a conference in Monterey a week or 12 two ago by a man whose name I'll think of in a minute who used 13 to work for a Swiss turbine manufacturer, Brown-Bravera, and 14 he started off talking about what happens in rotor steels but 15 then had also done work in waters for other carbon steels and 16 found that indeed, he could get stress corrosion cracking 17 above the threshold in water, again with about the same 18 kinetics as the other, as the turbine steel 19 So I agree with you, it's not normally looked upon 20 as a problem but there are well-documented cases of it showing 21 up.
22 MR. CLOUD: Yes. Now, I want to be clear on what 23 you said. Did you say boiling water reactors or pressurized 24 water reactors?
O 25 MR. SHEWMON: I said boiling water reactors because
a 485 1 they have a higher oxygen content and that's what it got
) 2 blamed on. But he, I don't think, found that it varied much 3 or at all with the variation of water composition, but he had 4 not gone to extremely low oxygen, as I recall. It was not a 5 highly caustic solution, which is what you mentioned earlier 6 as a well-documented example.
7 MR. CLOUD: Right. Ey the way, I should mention 8 that I have Bob Romer from Stone & Webster here tod.y, as well 9 as Ron Bedoin from my group, and hopefully they are taking 10 notes on these questions.
11 MR. SHEWMON. Marcus Fidel was the author. He used 12 to be in charge of materials for Brown-Bravera in the research
/T 13 area and is now a professor in a university in Switzerland, U
14 MR. CLOUD: I think one of our people was at the 15 conference, so hopefully we will have that information.
16 [ Slide.]
17 Our second task, from a technical viewpoint, is the 18 assessment of water hammer analyses. What we are doing here 19 basically is following the guidance in NUREG-1061, which lists 20 the series of caveats that the writers of those reports were l
i 21 concerned about from the point of view of applying leak before 22 break. And water hammer or lines that are susceptible to 23 water hammer was one of them.
24 So what we will be doing is making a specific review
\
25 of the transients, also the history of water hammer and assess l
l
I l
l 486 l l
l 1 its potential for water hammer on a system by system basis, '
2 and we will provide the technical justification to show that 3 we don't have a water hammer problem in the high energy lines 4 that we are addressing ourselves to.
5 If we can't show that, then we will, as I said, 6 follow the old-fashioned way.
7 [ Slide.]
8 MR, SHEWMON: Let me go back to stress corrosion 9 cracking again. There has been this work of finding of cracks 10 in the joints in steam generators. Indian Point and, I guess, 11 Surry was the other, some Vepco plant. These may well have 12 had cracks in them, but that is the secondary side. Are you 13 talking primarily in the steam lines or water lines' 14 MR. CLOUD: We're talking +6out -- first I would say 15 this. Your concern about the stress corrosion cracking in the i
16 steam generator is a legitimate concern, it's one of our
!? concerns. We're talking about the high energy lines in the 18 balance of plant. This does include the main steam lines and 19 it does include a number of other lines pressurized with high 20 pressure water. And certainly, the steam generator experience
- 21. is applicable experience that I feel we will have to address 22 head on.
23 [ Slide.)
24 Another concern that the framers of the NUREG had O 25 was the concern about lines which might have an overly high
}
e 487 1 fatigue usage. It's an important question for the lines that (D ,) 2 we are working with because normally, these high energy lines 3 do not have -- except the Class 1 lines -- do not receive a 4 full-fledged fatigue analysis.
5 So what we will have the job of doing is to 6 characterize the fatigue loading and the fatigue resistance
' and the fatigue capability of our lines, which we will do when 8 we work out an approach to do that. And we will do 9 straightforward ASME fatigue analysis on a certain 10 representative number of them.
11 MR. SHEWMON: Does the fatigue come only from power 12- cycles, or do some of these lines attach to pumps?
13 MR. CLOUD: Undoubtedly, there will be lines 14 attached to pumps, and undoubtedly we will have to consider 15 that form of loading.
16 MR. SHEWMON: I remember some outfit, it wasn't 17 Beaver Valley, that the LER or whatever it was said this is 18 the third time in a year that that line off a pump had 19 broken. One wonders what kind of a clown was applying what 20 kind of codes to keep doing the damn thing back again -- but 21 that's enough. Go ahead.
22 MR. BENDER: It's the feedwater lines primarily 23 that you're going to be looking at, I guess.
_ 24 MR. CLOUD: Right.
25 MR. MARK: You have spoken of high energy lines;
488 1 those are the only lines I suppose which have whip restraints.
) 2 MR. CLOUD: That's right. That's what we're talking 3 about.
4 MR. MARK: And there are lines in pipes which are 5 not high energy lines, 6 MR. CLOUD: That's correct.
7 MR. MARK: And what is the rate point between the 8 two?
9 MR. CLOUD: 200 degrees F and 270 pounds per square 10 inch, either/or.
11 CSlide.)
12 MR. CLOUD: Another concern that has been raised --
13 and this is mainly much more of a concern in the primary 14 system than in the balance of plant, but nevertheless we are 15 going to consider it -- is the potential for the failure of a 16 high energy line as the secondary or as an ancillary event to 17 the failure of the supports for major equipment.
18 I think the original concern grew out of the 19 potential or the concern that the steam generator supports in 20 some plants might be vulnerable to a brittle failure, and 21 might precipitate the failure of the primary coolant system, 22 and so this is sort of a carryover concern, if you will, into 23 the balance of plant.
g 24 Nevertheless, we will make a systematic adjustment 25 of the potential for this problem to occur.
e 489 1 [ Slide.]
2 What I have just described to you is the work that 3 will be performed to show that the fracture mechanics analysis 4 and the leak-before-break methods will be applicable to the 5 lines in question.
6 If any of the systems that we are talking about or 7 any of the postulated breaks fail the test that they will be 8 subjected to as a result of this work then, as I mentioned, 9 the whip restraints will be installed or at least earmarked 10 for installation.
11 MR. MARK: You have had it on the slide, but you 12 didn't say much about it. It is necessary if it leaks before
[~N 13 break that you are in a position to detect the leak. There 14 must be situations when there's a little difficult to assure 15 yourself.
16 MR. CLOUD: That's correct. Yes, indeed, that is 17 exactly correct. And, in fact, that is the part that comes 18 into when we implement the program itself. We start assessing 19 detection capability so we will be talking about that.
20 MR. SHEWMON: That is one reason why small lines get 21 harder to be detected, too.
22 MR. CLOUD: That is exactly right.
23 [ Slide.]
24 I will go fairly quickly through the specific O 25 technical analyses that are earmarked for the
490 1 leak-before-break analyses.
0\
\ )
s 2 Basically we have three components. First, some of 3 these materials, we have to be sure of their properties, so we 4 are going to get left-over pieces of pipe from the plant and 5 do the materials testing of the actual materials of the plant.
\
6 And the second are the two kinds of analyses that 7 are very well and exhaustively described in the NUREG, the 8 stability analysis for a throughwall crack and the crack 9 growth rate analysis of a postulated flaw consistent with the 10 tenets of Section 11 11 MR. RODABAUGH: Before you leave the materials 12 testing, how about all these hundreds and hundreds of welds, 13 and what are you going to do about -- l 14 MR. CLOUD: We are going to do'some testing. In 15 fact, I am going to talk about that right now. Why don't I 16 just --
17 [ Slide.] ,
18 We hope -- it's a good question. We hope to, as I 19 mentioned, get actual material and we anticipate the need to 20 actually construct some welds, using the weld procedures that 21 are in force at the plant and the weld materials, the welding 22 rods in place at the plant, and characterize the properties of 23 those welds.
24 MR. BENDER: Bob, every now and then you hear people
)
25 with stories about whether the right materials are installed I
1 l
e 491 l
1 in the right lines, and while A106 pipe is very good material l 2 and it is highly ductile and probably the right stuff to make 3 this argument for, how are you going to make an argument for 4 being sure that the right material is there?
5 MR. CLOUD: Well, I might ask for some help in 6 answering that question, but I think what we would do -- and 7 certainly my right-out-of-the-barrel answer would be that we 8 do feel that -- and I'm sure that Stone & Webster feels very 9 strongly about their quality assurance and their quality 10 control programs, but we believe, and particularly for the 11 high energy lines in safety-related systems, that those 12 programs are adequate to ensure that we have the right
('N 13 material in the right place.
\,%
i 14 MR. BENDER: Well, the points that need to be l 15 considered, I think, are first of all that you are dealing 16 with a chemistry range that you have to worry about, and that 17 you have got a legitimate sample of the materials, of the 18 various types that are used in the plant, and I don't know who 19 is going to vouch for that, but that is often difficult.
20 MR. CLOUD: Yes. l 1
21 Bob, do you want to add anything to what I said?
22 MR. ROMER: Robert Romer, Stone & Webster.
23 What we had talked about doing -- and we are 24 interested in exactly the same concern -- was doing a detailed k
25 review of the CMTRs, material characteristics that came in for l I
.- , . , - . - . - c.. , - , ,--e,, ,- - - - - , - - . , - - - .
e 492 l
1 those three materials and determine that we do have a 2 representative sample, and work back to particular material 1
3 samples that will undergo the fracture mechanics test. l 4 MR. BENDER: Well, that helps some. Do you have any 5 residual pieces off the pipe that's actually in there?
6 MR. ROMER: Yes, and that was the intent, to take 7 actual residual marked pieces of piping from the plant and use 8 them in the actual testing program.
9 MR. BENDER: Thank you.
10 MR. EBERSOLE: May I ask a question about this pipe?
11 Is there such a phenomenon as tight cracks or cracks 12 that have been filled by corrosion products or crud? You
[~' 13 know, that conceal the fact that you have lost structural 14 strength until you lose too much of it?
15 MR. CLOUD: Excuse me, I'm not sure I heard you 16 right. I tried to understand you, but I'm not sure I heard 17 you.
18 MR. EBERSOLE: All right. I'll give you what I gave l
19 them yesterday, the old corn meal treatment to the Model T. I 20 can fill up cracks with crud so you never know they are 21 there. I don't know whether such materials are generated in 22 these loops you are talking about or not, to fill up a crack 23 and obscure the fact that you have lost structural strength,
,,s 24 but it is not leaking.
'" 25 MR. CLOUD: We used to do it with oatmeal, as I
l' 493 i recall. l
) 2 MR. EBERSOLE: Well, call it oatmeal 3 MR. CLOUD: In the transmissions, yes.
4 Well, our program does not depend upon a pipe crack 5 inspection. Our work is based upon the presumption that there j 6 will be a priori the largest crack that would be acceptable 7 under the ASME Section 11, on the one hand; and on the other 8 hand, the analysis that we are talking about requires the 9 postulation of a throughwall crack in the pipe to prove that 10 the pipe will remain stable in the presence of that crack.
11 So whereas it is conceivable that we might have 12 cracks in the piping that would be concealed by the presence 13 of the corn meal, that would in no way either affect our 14 conclusions or our program. ;
i 15 MR. EBERSOLE: Plain old rust makes a pretty good 16 crack filler.
17 MR. CLOUD: That is correct.
18 MR. EBERSOLE: It swells up. So how do I know you 19 haven't got a cracked pipe that is waiting for that next 20 impulse to break it?
21 MR. CLOUD: You don't, but -- you do not know.
22 However, what we will prove to you, if we continue with our 23 program, we will prove that the result that concealed crack 24 which just is getting ripe and springs open will produce a l 25 leak that will be detected, rather than a catastrophic break.
l i
l
e 494 1 MR. EBERSOLE: You mean if I hit it with a hammer or N
2 a jolt or water hammer or whatever, it will then spring into a 3 leak rather than a break?
4 MR. CLOUD: That is correct.
5 MR. RODABAUGH: Supposing you have a crack in a --
6 well, let's take a steam pipe 24 inches diameter by 1 inch 7 thick, and following up Jesse's point, there happens to be a 8 longitudinal seam in this pipe, and that is the point where
^ the crack is developing.
'9 10 It now, under Jesse's postulant, might grow to be 11 seven feet long. Now when that pops through, that is an i.
12 unstable crack. You are not going to show by fracture
[ 13 mechanics, unless we are talking of a different definition of 14 break. To.me, that would break. It will be an unstable 15 failure.
16 MR. CLOUD: I think what you are saying is that you 17 are visualizing that there will be some circumstances where it 18 will not be possible to show that there will,be a leak before i
19 break; is-that what you are telling me?
20 MR. RODABAUGH: Depending a bit on just what you 21 mean by a break, Circumferential breaks are one thing I 22 think you could almost rule out, and longitudinal braks are 23 something else.
-w 24 Yes, there may be longitudinal breaks where you are a 25 going to have a hard time showing that it can occur.
4 s
e 495 ,
1 MR. CLOUD: That's right. And we said we
(_j/ 2 anticipated that circumstance, and we have said at the very 3 outset -- and I am glad, in fact, that you brought it up, 4 because I don't want to lose the opportunity to become 5 repetitive on this point.
6 [ Slide.]
7 But I want to bring back my first slide again, to 8 say that we are going to retain the rupture restraint in the 9 areas where the leak-before-break is not applicable.
10 So I am going to keep this one handy, and I want to 11 say that we definitely have what we hope -- u e hope that we 12 have got -- for lack of a better term, I will say a more 13 intelligent approach to the pipe break protection by using --
}
14 by making use of the program that we have, which envisions 15 several different. ways.
16 MR. RODABAUGH: Is it not true that-a longitudinal 17 split requires much less in the way of pipe whip restraints?
18 MR. CLOUD: As far as I am concerned, a longitudinal 19 split -- a pipe whip restraint doesn't do any good on the 20 longitudinal split, anyway. And so probably they wouldn't be l
21 there. But it might be that there would be some jet shields l 22 nearby because of that.
23 MR. RODABAUGH: Yes.
24 MR. BENDER: Bob, you stated that you are going to 25 make your argument around the Section 11 inspection limit, and
1 a
496 i
1 I guess I am not all that comfortable with the assurance we O)
( 2 have that inspection process under Section 11 for this raft of 3 piping we have got out there necessarily will give you the 4 confidence that you have exposed all the cracks that are 5 within the limits.
6 1 think in a statistical sense they find most of 7 them, but how much does the argument hinge on Section 11 8 inspections?
9 MR. CLOUD: Well, not really at all It really 10 doesn't hinge on the Section 11 inspections at all, because 11 part of our methodology is to assume that --
12 [ Slide.)
13 -- this is the testing, this is the crack stability 14 analysis, and this is the crack growth rate analysfs. This is 15 an entire class of analyses which is based on beginning with 16 the assumption that you have a throughwall orack, a complete 17 throughwall crack of a certain size that of and by itself
/
18 would be significantly larger than Section 11 would ever 19 permit.
20 And then we need to prove that pipe will remain 21 stable in the presence of that crack.
22 The Section 11 comes in where we are basically 23 following the methodol-ogy or, I guess, the practice, really,-
24 that has been formulated by NRR'and others who have gone i
j 25 before me in saying okay, well, what kinds of crack growth l
l 1
l i
e 497 1 rates am I to anticipate in this circumstance, and to assume 2 that we will have the largest crack that will pass through 3 Section 11, and see how fast they would grow.
4 But we still need to show that a very large orack, 5 much larger than Section 11 cracks, would remain stable and 6 with detectable leaks.
7 MR. EBERSOLE: If I can take as an example the main 8 steam lines from the steam generators to wherever, you know, 9 the turbine, would you allow these to lay on top of the 10 control room roof if it was say a four inch concrete slab?
11 MR. CLOUD: I don't know. I think -- I intended to 12 say of course not.
13 MR. SHEWMON: What is your point, Jesse?
)
14 MR. EBERSOLE: The point is if he says it is not 15 going to break before it leaks, that would be perfectly all 16 right.
17 MR. CLOUD: Yes, If the leakage is acceptable in 18 that location --
J 19 MR. EBERSOLE: Yes, but a big split would not.
20 MR. CLOUD: And then we get into the situation that
- 21 would in fact we be vulnerable to a big split or not. And the 22 important thing, I think, today is that we are in a position 23 with the improvements in our knowledge, the analysis f- 24 procedures, we are in a position to say whether or not we are s
25 vulnerable or we are not vulnerable.
e 498 1 CSlide.]
2 Just to briefly wrap it up, these are the -- filling 3 in the blanks on the types and kinds of analyses. Once this 4 is done --
5 MR. MARK: Excuse me. Are materials testing -- you 6 have mentioned several kinds of material, like 106, whatever 7 it is, steel. You only have to test a sample of that from one 8 of these residual pipe ends. You don't have to go around and 9 sample the metal at which whip restraint location, do you?
10 MR. CLOUD: No, sir, I don't believe we do. I 11 believe that if we could characterize -- one of the things we 12 have to do is see how many heats of this material we are
[ } 13 dealing with.
NJ 14 MR. MARK: So you might go with taking a sample of 15 each heat?
16 MR. CLOUD: That's the kind of thing that we are 17 presently engaging in.
18 MR. MARK: Or each weld type, perhaps?
19 MR. CLOUD: Yes. And there are certain numbers of 20 types of welds that would be permitted, and weld materials 21 that would be permitted, and procedures that would be 22 permitted. So we would sort through all of that to make sure 23 that we have characterized the welding adequately, and 24 similarly with the materials themselves.
O' 25 [ Slide.]
a 499 1 Now we have got a situation where we have come this O
() 2 far, we have gotten our postulated breaks, and we have seen 3 that at these break locations we can be sure that we will get 4 a significant leak prior to the catastrophic failure.
5 So now we have to go through an assessment of our ,
l 6 leak detection capability to ensure that we have the -- that I
7 we have the ability to detect the leakage, and the most I
8 important thing is that we believe that there are several l 9 different ways to detect leakage, and we feel that a good and 10 a sound leak detection program will incorporate several 11 techniques.
12 In particular, it will have the instruments that are 13 presently available. We will do a good study, a thorough f'N 14 study of the plant instrumentation in the sense that we 15 believe that there is a tremendous amount of leakage that l
16 could be detected by a study of the pressure drops and liquid )
l 17 levels and other instrumentation that we have in place at the 18 moment.
19 And the outcome of this study would result in 20 instructions to the plant operators and alarms and things of !
l 21 that nature.
22 In addition, we have physical inspections that we 23 could impose upon the plant operators and we believe that 24 there are various other features. So we visualize a 1x 25 comprehensive leak detection program specifically for each of l
500 1 the -- and we will show the path for leak detection for each
(_j 2 of the postulated breaks that has the whip restraint 3 eliminated.
4 [ Slide.]
5 MR. BENDER: This leak detection concept is a a
6 two-edged sword. I think you might be able to make the case.
7 The constraints put on the operation may have to be kept in 8 mind, just as has been the case in the primary system. If you 9 specify leaks requiring action that are too small, you may 10 find the plant being shut down more than you would like, 11 because you can't be sure you know where the leak is coming 12 from.
13 It seems to me that side of it must be included when 14 you make your major case. .
15 MR. CLOUD: Yes, I guess I couldn't agree with you 16 more, Mike. We definitely don't want to get into a situation 17 where we are shutting the plant down for every valve leak.
18 That's exactly right.
19 MR. SHEWMON: The current tech specs talk in terms 20 of five gallons a minute for Beaver Valley 1? ]
l 21 MR. KURTZ: I believe it's one gallon per minute and l 22 10 gallons per minute; one being unidentified and 10'being ;
a 23 identified.
24 MR. RODABAUGH: To get back to Jesse's example, 25 which I think is a good one since we're talking about balance
e 501 1 of plant, is the steam line. You have a fairly healthy steam 2 leakage from the valve stems. What leakage method, other than 3 going outside of the containment -- you could go look at the 4 steam line. But inside the containment, would you depend then 5 on your existing in-containment leakage system?
6 MR. CLOUD: Well, you're asking me questions that 7 I'll be better able to answer when I'm done with the work. I i
8 believe that there are humidity indicators inside containment.
9 MR. RODABAUGH: Well, I was thinking of the valve 10 house in the steam line in particular, which is moist, it's 11 got steam.
12 MR. CLOUD: Yes.
13 MR. SHEWMON: How much lower than this one gallon a 14 minute do you think you'll have to get, or are you aiming 15 for? Are you talking about factors of 10 percent or 2 or an 16 order of magnitude, or do you know at this point?
17 MR. CLOUD: No, sir, I don't know at this point.
i 18 I will point out to you, though, that the Rancho l
19 Seco line was a one-inch.line, it was a very small line, and i 20 it developed a crack. Behaved, by the way, as we anticipated 21 it would have behaved even though it was improperly 22 supported. And I believe that that crack produced a leak of 23 some 30 gallons a minute. So that the high energy lines leak 24 a lot. Once they start leaking, well, it comes pretty good.
25 [ Slide.]
e 502 1 I'd like to talk briefly about what we visualize 2 some of the benefits. I'm going to talk about the benefits 3 from a technical point of view, and the benefits from a cost 4 point of view.
5 We think that just getting rid of the whip 6 restraints, of and by itself, will be of benefit to the plant 7 by improvin,g the accessibility, minimizing restrictions on the 8 inservice inspection, and we feel that there is additional 9 radiation exposure that is incurred as a result of having a 10 congested plant.
11 By focusing the plant operation on the detection of 12 leakage, we feel that we are running the plant in a more 13 intelligent way. It is known that it is possible to get 14 binding with the whip restraints whenever there's a problem, 15 so we visualise that there are unanticipated events that will 16 be eliminated.
17 We are clearly, clearly, clearly going to have a 18 much better understanding of the behavior of our high energy 19 piping than we do using the current technology. The current 20 technology doesn't do anything; very little engineering on the 21 high energy piping. So we're talking about very serious study 22 of exactly how that piping behaves, and we ' f e e l that that, of 23 and by itself, will be a major contributor to improved safety 24 of the plant.
25 MR. EBERSOLE: How does this interface with the
503 1 other hoped-for advantage in getting rid of most of the f
( 2 snubbers?
3 MR. CLOUD: I'm sorry?
4 MR EBERSOLE: How do you interface this effort for 5 getting rid of the pipe restraints with the other ongoing 6 effort to get rid of the snubbers?
7 MR. CLOUD: Well, getting rid of the snubbers is --
8 those are essentially independent efforts. We have no 9 snubbers, for example, that serve as whip restraints. And the 4
10 snubbers are a part of the pipe support system, whereas the 11 whip restraints --
12 MR. EBERSOLE: They are? I didn't understand that.
13 I thought pipe supports were independent from snubbers.
[/}
14 MR. CLOUD: Well, yes and no. It depends on what 15 you consider to be independent. But the snubbers --
16 MR. SIESS: They are unbroken pipe supports as 17 opposed to broken pipe restraints.
i 18 MR. CLOUD: In any event, we believe that by 19 addressing ourselves and doing the engineering work on the 20 high energy lines, that we have got to end up with a better 21 plant.
22 MR. RODABAUGH: Bob, before you leave that slide, I 23 think Jesse and I keep running into the same confusion with 24 yesterday's program. An improved understanding of piping f
O 25 failure modes. Yesterday, we were discussing the difference
504 1 between collapsed failure mode and fatigue failure mode.
2 Now ordinarily, a pipe whip restraint, if properly J
3 designed, never touches the pipe. ,
l 4 MR. CLOUD: That's right.
5 MR. RODABAUGH: So in your program, how are you 6 improving the understanding of piping failure modes?
7 MR. CLOUD: Okay. We are improving the 8 understanding of the high energy piping failure modes because 9 of the fact that we will know, for example, about the 10 susceptibility of our high energy lines to stress corrosion 11 cracking.
12 MR. RODABAUGH: Okay. You're using the words
, /~T 13 " failure mode" in a different term than it was being used 14 yesterday.
15 MR. CLOUD: I'm using it in the general sense.
16 MR. RODABAUGH: All right, fine.
17 [ Slide.]
18 MR. CLOUD: We're talking about significant-amounts 19 of cost, too, which is what got me starte.. thinking about it 20 in the first place. I want to talk about two kinds of costs.
21 What we are going to call the quantitative costs.
22 Quantitative costs are costs that we can estimate reasonably 23 well, and we have two kinds of costs that w e. are going to be 24 talking about today.
25 The first are these quantitative costs that we can
e l
505 l
1 estimate reasonably well, and the second is a whole lot of A
k ,) 2 other costs that we know about, we know we are going to incur, 3 but we cannot estimate them nearly so well as these 4 quantitative costs. In the quantitative costs we have the 5 analysis, design, fabrication, installation and the indirect 6 costs of designing and installing the whip restraints. We're 7 looking at about $8 million for Beaver Valley, considering the 8 work that has already been done.
9 [ S l i,d e . ]
10 The more difficult to estimate costs are the 11 construction sequence frequently gets seriously screwed up 12 because of the presence of congestion. We have a major gap 13 adjustment verification program that must be conducted just b}
/
after all the whip restraints are in and just before 14 before --
15 the plant goes into operation. And this is a big program and 16 it varies a lot.
17 Structural changes that we know will be required 18 which will be due to the rupture restraint loads, -- big loads 19 get calculated out of these whip restraints.
20 Another congestion cost [ indicating 3.
21 And subsequently, in our outages we feel that we are 22 going to save, or make major savings, once again due to 23 congestion. And then there is the heat loss. The fact that i
24 we have the insulation off of the pipe at the rupture 25 restraints; we will be able to insulate the whole pipe. So we
a 506 4
1 lose a lot of heat in the operation of the plant.
v ) 2 [ Slide.]
3 Taken all together for this particular plant, we 4 have costs that will save somewhere in the range of $12 to $16 5 million.
6 [ Slide.3 7 And we felt that those are savings that are very 8 definitely worth going after, particularly in the present 9 situation where we have the leak before break technology that 10 we have developed over the years.
11 Secondly, from the questionable validity or the 12 questionable practicability of the rupture restraints anyway.
[\ 13 Thirdly, we feel that by doing this additional
\~sE 14 engineering work that we de facto, on the face of it, will end 15 up with a safer, better plant.
16 [ Slide.]
17 So taken all together, what we are doing is we are 18 proposing to provide for the pipe break protection using the 4
19 state-of-the-art today. We are not asking for any exemptions, 20 we are proposing to provide for the protection in a different 21 way. We want to use the ability to successfully detect pipe 22 leaks. We believe, as I mentioned, that this will be a major i 23 improvement in plant safety and design, both during operation I
24 and during the outages.
25 Second, it's a strong industry initiative. We are 4
_ ,.n. _ . - - , - . _ , _ _
- .,,.,,n..,. n, , . , _ . . , _ . . . _ _ _ _ _ _ , , , , . _ , - - , . - . , _ - , ,
a 507 1 here offering, I think, to do more than -- we are giving for
) 2 what we get on this one.
3 That is the conclusion of my presentation. Thank 4 you very much for taking the time to hear it through. And 5 understand the complexities of the licensing process which 6 will require some assurance be given Duquesne Light that if 7 this approach is followed, that there won't be a glitch at the 8 licensing -- at the time for licensing, and that the license 9 of the plant will not be jeopardized.
10 Thanks again, and I'll be glad to answer any further 11 questions if there are any.
12 MR. RODABAUGH: A couple of details on what you, at 13 this stage, think you may be addressing. There is a flange 14 joint, several flange joints, for example, off the safety 15 relief valve. Are you going to address in any way flange 16 joints?
17 MR. CLOUD: We have to consider the geometry of the 18 pipe as it is, so definitely we will consider the flange 19 joints.
20 MR. RODABAUGH: Well, I asked the question because 21 nobody that I'm aware of has done any leak before break on 22 flange joints -- analysis and tests.
23 I guess my second curiosity is the leak before break
~_ 24 technology is fairly well advanced for straight pipe and
(
\ 25 welds in straight pipe, but I've seen essentially nothing on
e 508 1 application of the leak before break technology to elbows, (O) 2 reducers, branch connections, which you run into a lot of.
3 MR. CLOUD: You're talking about -- when you say --
4 in the question you're asking, are you referring to the 5 calculation of the growth of cracks and the stability of 6 cracks?
7 MR. RODABAUGH: The stability, primarily. What is 8 the stable crack size, an approximation?
9 MR. CLOUD: Oh, yes. Well, we feel that whereas in 10 the case of straight pipes, for example, where we might have 11 very exact solutions for the various different kinds of 12 cracks, nevertheless, in the fundamental mechanics of the
[~' 13 behavior of a crack it's the same regardless of anything i
L 14 else. It is only complicated by the geometry of the material 15 and the geometry of the crack.
16 And we believe that, like any other analysis 17 problem, we are entirely confident that it will be possible 18 for those cases where we do not have exact solutions to have 19 approximate conservative solutions like we do in so many other 20 analysis areas.
21 MR. RODABAUGH: Well, I certainly want to encourage 22 work along these lines.
23 MR. CLOUD: You are in a position to do so, I might 24 point out i i N' 25 MR. RODABAUGH: Well, I'm not sure if I am or not,
e l
509 ,
l 1 but I think it's a very worthwhile thing to attempt to do.
[
\ 2 MR. CLOUD: I couldn't agree with you more.
, 3 MR. SHEWMON: Let me ask a simple question. A 4 branch is a T-joint of dissimilar diameter points?
5 MR. RODABAUGH: Yes.
6 MR. BENDER: Like Ev, I am a supporter of this 7 idea. I think it is constructive for a number of reasons, not 8 the least of which is that you get rid of pipe whip 9 restraints.
4 10 But there are some things that I think ought to be 11 done as a preliminary, and that the regulatory staff will have 12 to do in order to give you any assurance that what you are
[h 13 proposing will be acceptable after you do all the work. I
'O 14 think really you have to be somewhat more quantitative than 15 you have been about the size breaks, or size of crack that 16 you are going to have as the initiator, from which you are i
17 going to extrapolate growth rates. And I'm not at all sure 18 that I want to accept the Section 11 limits as being the l l
l 19 starting point.
20 But you probably don't have to start that small. I i
21 think you probably could start with something somewhat bigger 22 because the A106 pipe in particular has good fracture 23 properties, and you probably have a long way to go.
24 The second thing I want to say again, I think you k i l
25 may be letting yourselves in for pore than you know when you l
l l
l
---,.---cem---,,,,r , ,-m .--,---,-._,.w,w,-wr,,,,,vr- ----,crw--.,,-e<y ,,r-
I e
510 1 commit yourself to do something if you have a leak of a b
( j 2 certain size, recognizing the fact that finding out where the 3 leak is is sometimes a difficult job in a primary steam 4 system. And that that may influence, to some degree, what 5 parts of the system you want to defend.
6 The third point is the one which Ev has I think 7 articulated well enough, and that is that the fittings are a 8 separate case.
9 Now, the regulatory staff had difficulty dealing 10 with the fitting question for the primary system, as you know, 11 and stayed with the straight pipe. Whether they will get the 12 fittings in the picture or not later, I don't know but I think 13 that in view of that precedent, you ought to think about what 14 your case for fittings is going to be, and whether it fits 15 with the other arguments.
16 That's all I have.
. 17 MR. CLOUD: I would like to reply to those points 18 becasue they are all three important ones. The first thing, l
19 the most obvious and most important aspect of the reply is 20 that everything you say in every aspect of your remarks is in 21 fact true. Everything that you pointed out in every aspect of 22 the things in our program that you discussed are things that 23 are completely ignored with the conventional approach.
24 So that just by virtue of the fact that we are in 25 and poking around trying to understand how the systems behave 1
a 511 1 to protect against the potential breaks in an intelligent
) 2 way is almost, de facto, of and by itself, I think an advance.
3 And I agree, we certainly don't know all of our j l
4 answers in advance, but we believe that we can develop them as 5 we go along, and we believe that the state-of-the-art is 6 advanced to the point where we can develop workable, 7 satisfactorily conservative answers as we go through.
8 Another thing I'd like to point out with respect to 9 the discussion related to the primary system is that the 10 primary system piping in the pressurized water plants, that is
- 11 all cast material. That is the least tough material in the 12 plant, and that's the part that we've accepted the leak before
'\ 13 break on.
14 We have all the rod material in the balance of plant 15 is much tougher, much more fracture resistant. Mr. Mark?
16 MR. MARK: I am certainly favorably impressed with 17 the direction of things here. You have mentioned that the 18 state-of-the-art is advanced, and that was particularly with 19 respect to fracture mechanics and crack growth calculations.
20 How is it with respect to leak calculations?
21 MR. CLOUD: That is not my field and I'm not sure.
22 [ Laughter.3 23 I'm not in the fluid mechanics business, but I 24 believe we can calculate the amount of leak once we have a
'# 25 given --
e 512 1 MR. MARK: Well, once you have a through-wall crack 2 and it's big enough that the pressure will probably pry things 3 apart slightly, then I could understand some leaking must go 4 on but I'm not sure that that's the way things always happen.
5 Nor am I sure that you are necessarily so good at calculating 6 the opening of the crack as you may be at the progress of the 7 crack.
8 MR. SHEWMON: The opening should be elastic, and if 9 you assume it's elastic then it gets down to a question of, 10 you know, how rough it is. And they can calculate smooth 11 cracks -- smooth-walled cracks, I understand it, or you can do 12 tests on that. And they can talk about, with a given
["'j 13 pressure, given two surfaces, if we now make a cut through b
14 here, how much will it open up.
15 My impression is it gets worse as you start talking 16 about how much --
17 MR. MARK: Zig-sag, yes.
18 MR. RODABAUGH: There is a good deal of test data, 19 Carson, on leak rates and small cracks.
20 MR. MARK: It was a question; i t was not an 21 assertion of doubt.
22 MR. RODABAUGH: Well, your doubt is well taken, and 23 I think Jesse's point enters here, too. Crud stoppage of j l
24 cracks. But the research work is mostly on a piece of ;
\
25 straight pipe. I know of very little on leakage rates in a
,- .-n.,,_....-..--n._-,- , , - - - - - . , - - . . , - , . _ - ----- -. - ,,,,_-.---- _ - en. .,,..~--,.s ---,.e -----_..,.--s-,, ,.
l l
513 )
l 1 crack in branch connections, for example. So your question is i
k'D,) ' a good one.
3 MR. EBERSOLE: Paul, I was just going to renew that 4 question I discussed with you about the wide range of material 5 properties you can get within the context of A106, B. Are you 6 going to sharpen that up? I recall having --
7 MR. SHEWMON: He said that they knew what heats were 8 in the plant and thought they could get samples from that.
9 MR. EBERSOLE: Right. I'd like to know what that 10 range of property is which is under the cover of A106-B. I've 11 heard stories ranging from almost brittle at 40 degrees to 12 stuff that you could forage around and bend into duplex bends.
(J) 13 MR. '; LOUD : Well I'll tell you what. Give us the 14 go-ahead on this program and we will tell you what the range 15 is.
16 MR. EBERSOLE: 106B, I think that's something that 17 needs --
18 MR. SHEWMON: And also hot pipes at 200 or 270.
19 MR. EBERSOLE: Well, I'm interested in the cold 20 pipes, too, that service water.
21 MR. CLOUD: That's an excellent question, 22 Mr. Ebersole, and I'd like to reply in the same way I did to 23 Mike and that is that the questions you are asking you will 24 get the answer to if we could go ahead with the program. But
~
25 you're not going to know nothing if you don't, if we don't.
e 514 1 MR. SHEWMON: Let me ask one more simpleminded f Would you list the systems that you will be looking
( 2 question.
3 at -- steam systems, certainly. Feedwater.
4 MR. CLOUD: Those systms are listed and well 5 defined, and they are the so-called high energy line systems.
6 I don't have it right now, but it's the main steam lines, the 7 feedwater lines, and all the other lines that are higher than 8 275.
9 MR. EBERSOLE: Service water is about maybe 200 or 10 300 pounds at 40 degrees.
11 MR. SHEWMON: I see a violent shaking of the head 12 back there.
13 MR. CLOUD: Bob, can you help me ote t informally with s
14 anymore of the lines, anymore of the hydrogen systems?
15 MR. ROMER: Steam generator blowdown lines, low 16 pressure injection, other injection lines. Of course, we 17 mentioned feedwater, main steam, aux steam, charging. It 18 definitely doesn't cover service water.
19 MR. EBERSOLE: Component cooling? It doesn't cover 20 that?
21 MR. ROMER: I don't believe so. In some cases we 22 have postulated breaks in lines and the effects are virtually 23 non-existent.
-- 24 MR. EBERSOLE: So really you're looking at just the d 25 elimination of pipe whip.
a 5 05 1 MR. CLOUD: In general, that's correct.
2 MR. EBERSOLE: You're not looking at the loss of 3 service function due to breaks.
4 MR. CLOUD: That's exactly right.
5 MR. SHEWMON: So these are steam lines and feed 6 lines. Okay.
7 Any other questions for Bob before we go on tc the j
8 other Bob?
9 [No response.)
10 MR. SHEWMON: Okay. Thank you very much.
11 MR. BOSNAK: The staff has had several informal 12 meetings with the people from Beaver Valley, Duquesne Light,
[~->\ 13 and I would characterise the staff right now as being 14 sympathetic particularly to the removal of these items of 15 hardware that are in there to protect again'st postulated pipe 16 breaks. And we are also sympathetic to the approach that is 17 being proposed.
18 We are also I think -- in preface, we are pleased 19 that Duquesne has chosen to do these kinds of analyses and 20 testing. I want to stress that testing is an important part 21 of this thing because we believe that the broad scope rule 22 change which is in preparation now through ODC-4, I think will 23 be aided. It has to be helped by this program.
24 Now, what are some of the problems that we see? We
25 expect that within a couple of weeks we will get a formal
I 516 1 transmission with all of the meat on the skeleton that 2 Duquesne is able to come up with at the present time. But 3 certainly, identification of systems to which fracture 4 mechanics techniques may not apply -- that is going to be 5 quite important.
6 I think Bob said that the method works when applied 7 in appliable situations, and we quite agree. So we don't want 8 to be down at the last minute arguing about whether or not 9 this system is subject to stress corrosion cracking, or this 10 one has large dynamic loads which we don't know how to 11 analyse. So we want to be sure that they go through, at least 12 now, a preliminary screening and eliminate or kick out the f}
V 13 systems which we will be arguing about. So in particular, Bob 14 did mention the large dynamic loads, stress corrosion cracking 15 and the fatigue situations.
16 From a legal and administrative point of view, we 17 have talked several times with the staff counsel, and I 18 think originally they felt that the broad scope rule had to be 19 in place before they were able to go ahead with this. That 20 has changed now.
21 I think the caveats that they have are that the 22 broad scope rule is proceeding on track. In other words, 23 there have been no hangups identified. And of course, we and 24 they would be hopeful that there wouldn't be too many systems 25 for which these analyses and testing techniques cannot be
e 517 1 applied.
) 2 Now what they see happening and what we would see 3 happening is that there may be one or two systems at the end 4 -- and this is at the point at which the license has to be 5 issued, and I guess we're talking about the end of 1986 or 6 early 87 -- if that is the case, then we would handle that --
7 and this is what's being proposed now -- by a schedular 8 exemption that would be similar to what we are doing with the 9 narrow scope rule for the main loop.
10 By that I mean that they have gotten to the point 11 now where they are ready for the license and there is this one 12 system which we believe we need to install the hardware that 13 we have been used to. And if that is the case, an exemption
~
14 schedule or an exemption, if you will, will be granted for 15 perhaps one or two refueling outages, and after that period 16 the hardware would have to be installed.
17 So that's the way we look at it. Obviously, if 18 there are many, many such systems when we get down to the end, 19 that will be bad news, and it would have to delay the 20 licensing. We hope that there won't be such a situation. And 21 that is why right now we believe that preliminary work, going i
22 through and screening the number of systems, is very, very l
23 important.
l 24 MR. SIESS: Bob, if Beaver Valley 2 were allowed to 25 operate without all these pipe whip restraints in place, would
518 1 they be the only plant operating without pipe whip restraints?
[D,
( / 2 MR. BOSNAK: Other than the primary loop which we 3 have been talking about, they would be the only plant recently 4 licensed --
5 MR. SIESS: That wasn't my question. Could you say 6 yes_ or no and then elaborate?
7 MR. BOSNAK: If you go back to the very old plants 8 -- and there are those plants that were never designed for 9 these kinds of situations.
10 MR. SIESS: There are plants -- what do you call 11 very old? Dresden?
12 MR. BOSNAK: I'm not sure whether Dresden -. Maine 13 Yankee -- well Connecticut Yankee perhaps. Maine Yankee may 14 have --
16 MR. BOSNAK: Essentially, the SEP plants, but again, 17 those have been looked at from a consequence point of view and 18 during the recent SEP exercises. So even though they may not 19 have the hardware in there, they have been reviewed from a 20 systems and consequence point of view.
21 MR. SIESS: Okay. I was just trying to get some 22 perspective in terms of risk, of this plant operating, say, 23 for two years without these restraints versus the sum total of 24 risk. You know. Not a regulatory approach; strictly a safety 25 view of it, without consulting a lawyer.
e 519 1 MR. BOSNAK: Staff, at least in the Division of 2 Engineering, feels that these, particularly the pipe whip 3 restraints, really don't contribute to safety and may actually 4 detract from safety because you can have an interaction 5 between a whip restraint and a system in normal operation, and 6 over the years these things can settle and change position.
7 And that may, in fact, under certain situations, lead to a bad 8 situation.
9 MR. SIESS: So your problem was simply a legal one 10 of some way to defer this until the whole issue is settled by 11 the rulemaking. .
12 MR. BOSNAK: Well, even if --
/~' 13 MR. SIESS: You don't have a real safety concern.
14 MR. BOSNAK: That's right, because we are not 15 changing -- I think what Bob pointed out, even though some of 16 the letters that have come in on the narrow scope rule have 17 indicated we should go further, we are not making any changes 18 with respect to equipment qualification 19 When you are talking about balance of plant, that is 20 the important thing. That is not being changed.
21 MR. MARK: On one of the slides we saw there were 22 136 pipe whip restraints in the balance of plant. I 23 MR. BOSNAK: Items of hardware. Perhaps pipe whip 24 restraints, jet shields.
- 25 MR. MARK: Yes. Is that the total number that are i
i t
)
...,-.-,--.,-_,.------.,.,,_,,-__,__,,-,-._,.n-...._.,,.,-____,,,,,,,__,.,_.,,nn_,.-,_,,.
i, e
520 i 1 out there, some of which might later be decided were not -.
( j 2 MR. BOSNAK: I think that comes from about 400 or so 3 breaks and those are the only things for which they have 4 necessity to have this kind of break protection. In other 5 words, for the other breaks they expect that they don't need 6 these kinds of hardware.
7 MR. CLOUD: It is the total number. The numbers I 8 presented on the pie chart were the total numbers.
9 MR. MARX: So some of those probably won't be 10 relieved in this way, since you will find that they are 11 subject to corrosion or fatigue or something.
12 MR. CLOUD: That's exactly correct 13 MR. RODABAUGH: Bob, in terms of your MEB 36.2, 14 right now you have this requirement for longitudinal pipe 15 breaks which are more of a vigorous crack. What are your 16 plans regarding that part? Is that involved at all here?
17 MR. BOSNAK: Well, if you go into the fracture 18 mechanics approach, both the longitudinal and circumferential 19 breaks go away.
20 MR. RODABAUGH: I guess 36.2 requires the 21 longitudinal breaks even in a moderate energy piping.
22 MR. BOSNAK: Well, certain locations might have 23 longitudinal. It depends on the stress ratio.
24 MR. RODABAUGH: Maybe I'm confusing this with 25 through wall leakage cracks.
e 521 1 MR. BOSNAK: Through wall leakage cracks are there 2 in moderate energy situations, and you know if you don't have 3 a postulated pipe break in those areas, then you go into the 4 through wall leakage crack which --
5 MR. RODABAUGH: I am trying to explore this 6 equipment qualification bit -- maybe it wouldn't make any 7 difference.
8 MR. BOSNAX: Right now from what the equipment 9 people tell me, the difficult area is not the large break; 10 it's the smaller break in which they can get some superheated 11 steam.
12 MR. RODABAUGH: That's exactly the point I'm trying l
13 to follow. Would this impact how you evaluate --
14 MR. BOSNAK: It wouldn't impact that at all. Not at )
. l 15 the present time at least. But we do have all of these 16 suggestions that say we should go further but we are not 17 intending to do that now.
18 MR. SHEWMON: Any other questions? I think the 19 sense of the consultants and the members here is positive and 20 encouraging. Now, do you want anything other than warm l
21 feelings out of us at this time?
22 MR. BOSNAX: I think that's what we were looking 23 for. That.you don't see reasons that this program should not 24 proceed. That's what we wanted to know; if you had any doubts 25 at all. But if you have warm feelings --
. ~.-- -- . . . - . ,, _
e 522 1 MR. SHEWMON: I have a note from Jesse. He says, 4
2 " Sounds 1.ike a really worthwhile program," and I think that 3' tends to be the sense of the meeting. So I think on that, s
't 4 unless I am mi'squoting anybody, -- any other questions or 5 discussion? If not, the meeting is adjourned.,
6 (Whereupon, at 3:03 p.m., the s ub c osami t t e e meeting 7 was adjourned.3 8
1 9
10 11 12 's 4
13 14 15
- 16 17
, 18 19 4
20' 21 '4 22 23 24 1
25 l 1
l l
l l
l
("
'sms 1 CERTIFICATE OF OFFICIAL 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 Subcommittees on Structural Engineering, Seismic Design 10 of Piping, and Metal Components 11 Docket No.
12 Place: Washington, D. C.
(v~'\
13 'Date: Tuesday, September 24, 1985 14 15 were held as herein appears and that this is the original 16 transcript thereof for the file of the United States Nuclear 17 Regulatory Commission.
18 !
(Signature)
(TypedName'ofRegrter) [Su'zage B. G ng-20 21 22 23 Ann Riley & Associates. Ltd.
24 25
' I
-c
~
w F) i - e o O
O k.ewu
.asn cD MECHANICAL-STRUCTURAL ENGINEERING BRANCH PRESENTATION TO ACRS SUBCOMMITTEE ON STRUCTURAL ENGINEERING SEISMIC DESIGN OF PIPING METAL COMPONENTS SEISMIC, STRUCTURES, AND CONTAINMENT INTEGRITY RESEARCH PROGRAM
- SEFIDEER 23 AND 24,1985 i
O O MSEB BUDGET FOR STRUCTURES AND SEISMIC RESEARCH BUDGET TITLE FY 1985 FY 1986 'FY 1987 SEISMIC MARGINS STUDIES $650K $600K $900K l VALIDATION OF SEISMIC RESPONSE PREDICTIONS 1100 1750 1900 COMPONENT FRAGILITY DATA ACQUISITION AND 240 1100 1000 EVALUATION SEISMIC CATEGORY I STRUCTURES 820 1200 1500 RELIABILITY ANALYSIS OF NONLINEAR 0 200 - 400 -
BEHAVIOR OF CONCRETE STRUCTURES (But) '
I STRUCTURAL DEGRADATION 0 0 300 STRUCTURAL COMPUTER CODE BENCHMARKS 300 200 200
, NRC/EPRI PROGRAMS ON PIPING DYNAMIC 200 700 900 LOAD CAPACITY OTHER NRC PIPING RESEARCH 333 375 600 PARAMETERS INFLUENCING DAMPING IN 230 200 0 PIPING SYSTEMS
! $3873K $6325K $7700K
! PROGRAMS THAT TERMINATE IN FY 1985 1,418
$5291K j Su S % 0 Egg a u Wg 1
MSEB BUDGET FOR CONTAINMENT INTEGRITY RESEARCH
^
BUDGET l
! TITLE FY 1985 FY 1986 FY 1987
$2,060K $2,000K $2,400K '
EXPERIMENTS ON CONTAINMENT MODELS l UNDER EXTREME LOADING CONDITIONS 740 600 700 i CONTAINMENT INTEGRITY UNDER EXTREME LOADS 900 900 600 INTEGRITY OF CONTAINMENT PENETRATION UNDER SEVERE ACCIDENT CONDITIONS 1
250 200 0
! CONTAINMENT BUCKLING
$3950K $3700K $3700K i
I -
]
i 3
1 . .
EFFECTS OF EARTHQUAKES " "*ERATING PLANTS ISSUE SOME OPERATING PLANTS MAY BE REQUIRED TO WITHSTAND EARTHQUAKES LARGER THAN THEIR ORIGINAL DESIGN BASIS BECAUSE OF NEW SEISMOLOGICAL INFORMATION. IT IS IMPORTANT TO ESTIMATE AVAILABLE DESIGN MARGINS OF PLANTS TO WITHSTAND THESE LARGER EARTHQUAKES TO PROVIDE THE BASIS FOR REGULATORY DECISIONS AND AVOID UNNECESSARY SHUTDOWNS. %YL 4 A & tvata u ut 4m (O Ms .
-e 4Mnw W A-d % io, P b N*
OBJECTIVE (p BG % y 9-)
TO EXPERIMENTALLY VALIDATE COMPLEX ME_THOD_S USED TO PREDICT SEISMIC BEHAVIOR OF THE PLANT; IMPROVE ~
AND REDUCE UNCERTAINTIES IN COMPONENT FRAGILITY DATA BASE USED TO ESTIMATE MARGINS; DEVELOP PROCEDURES FOR ESTIMATING SEISMIC DESIGN MARGINS; AND ELIMINATE EXCESSIVE USE OF SNUBBERS TO PROVIDE BETTER BALANCE OF SAFETY.
ELEMENTS .
o CONDUCT EXPERIMENTS USING LARGE SCALE INTEGRATED FACILITIES TO VALIDATE COMPLEX PREDICTIVE _
METHODS
- 3DR NN
- b"N *
- ha. p . ska.6> - qcit t a_re s w - go i og . P g \mcy */7 _ W s en_(1 S.
o ASSEMBLE EXISTING COMPONENT FRAGILITY DATA AND CONDUCT FRAGILIT_Y TESTS TO IMPROVE THE DATA BASE.
- b o DEVELOP AND DEMONSTRATE PROCEDURES FOR ESTIMATING SEISMIC DESIGN MARGINS AND PROVIDE BASIS FOR REDUCING THE NUMBER OF SNUBBER _S USED IN SEISMIC DESIGN.
4... si INTEGRATION Wgod5 Q..
PROGRAM INVOLVES COOPERATION AND COORDINATION WITH EPRI, INDUSTRY, GERMANY AND JAPAN.
Mm,'d',io FEGULATORY USE
^*
METHODS, DATA BASE AND REVIEW PROCEDURES WILL BE USED BY THE REGULATORY STAFF TO EVALUATE LICENSEE SUBMITTALS, PROBABILISTIC RISK ANALYSES, AND ESTIMATE SEISMIC MARGINS. w/
%wa w u a
f
4 e
9 em o . W
= w (T
\* 5 R 5
- e eE "
- a $ 0 ~
E & w =
w a E
< < u E w u
\
m z d C
- 8 <u W M 5 t C d
m $d 2 g
un W G E M W 5 E = g f ; "= ow s=s 5mEu a=8 5"o ee 5 -
w m w w E e a E 8 w
9 E e 8 E E a"E 8 5 m b w w e "# *
- S S :
5 w
6ga
<-,o M N 2 aE
=
o W
- m 2, sE "
o M "
d - k m m E m
=
w M$w E eE w
w =
- g 2 - m p ~
W W5 **
E E E - o -
C eeS ~ E
- 8 0 s .= E ep= 0
= #
E I" E
e z
3 5 mE
- o =
goa5 e E2 & w 4 t
=
2 W e E W n <e m a ae*4 E
~
0
= e W 8 5 ~
= 5 o w S 8w 5 g W
m m*5 -
5 a- ameaas " ~
e5 w
b 5 g"*R
" ~
5 5 W " " S $ 5 < S o G m =
g g 5 g w E o - = 4 o W eW=
e -
E E a
u aa o
e m
so *u S m w W
o J
w w
=
e
=
=
y "
gm
- E 8 E w E g 5- E o w ra=m w <
o 5 - aE w =
R <nm gsr= p o
w
.Wo -
8 8
=
=
S
" 8
- gm m
E W w
- ~
- o .<"
o 5 *
= b w 5 m < o m M
M MM
- w
=
w a m w o a u
- W 5 W " o g
~
E e S 0
u a=g2 5 o- w gw
= ms"5 eseia M ew
=
g=
g E ee gE ew o 8m "
agaa s e= "= a s- "5w w2 5 -E 5 -
e=me 4 o = m E 5 E E E C 8 E G u
w v
a=
=
w g E y8C = 0 o
- e5 " $
" " E E B w 5 M "5 2 W w a w w m w w e a m
- =
a <.m W w
b- b 3 o
= a w
h o o k =
z - =
w eM eE b E W E 6 3 - m w -
" E s=m e: o a$= = 5 8 m B c "
e g= <
- w=
o
=
.easo e m e=5o -
m
- U o % *= o b w 5 m
a m 5 =m
~
W E e 5 g eE gW M M " M E 5 8 a"W E 5 W *
=
=
w wm asm se 8
a W
eE E=
o
=
=
E o
5
=
m E
w E
w e5E E
a -
w E " S
- = w m
Sw W
E w w
=
W W W
"
- g ~ 5 X
" E D $~uE d 2 w
m E
s<-=
- g m
E
=
o 5
u
=
w
=
o w
- - =
o o r
o
- w -w
=
g < w 2 = o -
= E E E 5EE G -
4 4 4 E
O- C b w
=
M
- o
EFFECTS OF EARTHQUAKES ON 0; ING PLANTS (CONTINUED)
PROG 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 VALIDATE SEISMIC CODES. ALSO COOPERATE WITH EPRI IN A SOIL-STRUCTURE INTERACTION EXPERIENT BEING BUILT IN AN ACTIVE EARTHQUAKE REGION IN TAIWAN (FY 1986-1988). PARTICIPATE IN FY 1987 IN VIBRATION EXPERIMENTS IN COOPERATION WITH THE JAPANESE ON LARGE INTEGRATED PIPING SYSTEM AT TADOTSU FACILITY (WORLD'S LARGEST SHAKER TABLE) TO VALIDATE NON-LINEAR PIPING COMPUTER CODES. VALIDATION EFFORT WILL BE COMPLETED IN FY 1987.
- 2. IMPROVED FRAGILITY DATA BASE - IN FY 1986. WILL COMPLETE GATHERING OF COMPONENT FRAGILITY (FAILURE) DATA IN COOPERATION WITH EPRI AND INDUSTRY (SEISMIC QUALIFICATION USERS GROUP) AND IDENTIFYING WHAT SPECIFIC FRAGILITY DATA IS NEEDED. IN FY 1987, WILL COMPLETE FRAGILITY TESTING OF SELECTED COMPONENTS. CONTINUE IN FY 1986 AND COMPLETE IN FY 1988 EXPERIMENTS ON SMALL SCALE CONCRETE SHEAR WALL STRUCTURES TO INVESTIGATE THEIR BEHAVIOR UNDER SEISMIC LOADS EXCEEDING THE DESIGN BASIS TO DETERMINE THE MODE AND LEVEL OF FAILURE.
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 IMPROVE THE PRO-CEDURES. IN FY 1987, SIMILAR PROCEDURES AND PLANT REVIEW WILL BE COMPLETED FOR A BWR AND WILL COMPLETE THIS EFFORT. ,
INTEGRATION ,
- 1. COOPERATING WITH THE FRG AT HDR FACILITY TO VALIDATE SEISMIC METHODS. (FRG-$10M,NRC-$500K)
- 2. NEGOTIATING WITH JAPANESE (MITI) TO COOPERATE ON TADOTSU SHAKER TABLE TO VALIDATE NUCLEAR PIPE CODES.
(MITI-$12M,NRC-$700K)
- 3. COOPERATING WITH EPRI ON SOIL-STRUCTURE INTERACTION EXPERIMENTS IN TAIWAN. (EPRI-$1M,NRC-$200K)
- 4. COOPERATING WITH EPRI AND INDUSTRY (SEISMIC QUALIFICATION UTILITY GROUP) TO GATHER FRAGILITY DATA.
(INDUSTRY-$50M,NRC-$500K)
S. COOPERATING WITH EPRI ON SEISMIC DESTGN MARGINS. (EPRI-$1M,NRC-$1.2M)
NRC RESEARCH IS FOCUSED ON DEVELOPING THE TECHNICAL BASIS FOR LICENSING DECISIONS AND ESTABLISHING REGULATORY CRITERIA USING RESULTS OF INDUSTRY RESEARCH WHERE POSSIBLE.
REGULATORY USE
- 1. VALIDATED SIMPLIFIED METHODS WILL BE USED TO EVALUATE LICENSEE SU3MITTALS.
- 2. IMPROVED FRAGILITY DATA BASE WILL BE USED TO EVALUATE AVAILABLE SEISMIC MARGINS AND PROVIDE IMPROVED PROBABILISTIC RISK ASSESSMENTS.
- 3. SEISMIC DESIGN MARGIN PROCEDURES WILL BE USED TO ASSESS CAPACITY OF NUCLEAR PLANTS TO WITHSTAND LARGER EARTHQUAKES AND MAKE SOUND REGULATORY DECISIONS WITHOUT UNNECESSARY MODIFICATIONS OR SHUTDOWNS.
- 4. ASME CODE, RECIATORY GUIDES, AND STANDARD REVIEW PLAN MODIFIED TO PERMIT REMOVAL OF SNUBBERS TO MAKE PIPE SYSTEMS MORE FLEXIBLE AND PROVIDE IMPROVED BALANCE OF OVERALL SAFETY.
o o O , ,
I $
l FOREIGN AND 00NESTIC l NRC SEISMIC 5AFETY llESEARCH PROGRAM j RELATED PROGRAMS I SElsatt HAZARO l
j SEl5M C MAZAa0 j
I o EPRI and Industry Probahtlistic Setssic Hazard Program ; EAllTH SCIENCES ! e Charlesten Earthquake l
j e Setsmetectonic Program l e h 8 Wck hWe o USGS Ground Metten Stedies l !
o mattenal Ear he Hazard Re- l e Sell Response Project !, e Site 1pecific Spectre l
4 ductten Prog. USGS. II5F. FDIA.
! N05) ;
O Corps of Engineers Setssic Sells Sett1 ment Program j ETH005 OEVELOPENT AIIO VAL 10AT1011 APPLICATIONS !
1 SEI5MIC RI_5E
- ~
j VALIGATI0li 0F SEISMIC E TH005 l e Setssic Safety Margins e Setssic Design !'
- * "" "8 "S**" !.
l 0 Germasqr - EfE MOR Program l I e Validatten 'of Setssic e Lead Combinations ! e Integrated Safety l Assessment Program e EPRI - Taiwas $51 Esperiments Calculatten Methods Project for Structural l l Des @ Project j e Japan - MITI/IIUPEC Setssic l /\ , e Rfst Methodology l Integretten 8
- i IAperiments j * ""
'VI !
FRAGILITIES 4110 RESP 0lISE !
I FilAGILITIES ABIO llESPONSE !
o EPRI Piping Program ;
j j
- I "C%""8I N'8P**S' P'*3'CI ! SEISMIC MulGINS i
o EPRI Equipment (lualification l Program l e Campenent Fragt11ttes Project j e NTOL Seismic Issues O uRC puent Qualtftcatten ! e Piping Reliablitty Projects fe Pipe Denytag (R.S. 1.61) ,
p ,
e Containn.nt suctiin. Project .I * * **a("5'*
e mRC Aetas Prae =
- l; ! e ASME Code-Piping 8 i o Setssic Qualtftcatten Utt11ty Structures Group Data ;
{
I : e E9=tmt 0=a11ficattea l :
SEisniC n4asins :
l : : e structural Damping e EPRI 5etsete Margins Research ;
l (R.G.1.61)
Program j .
4
. e Independent Pipe Supporti i e ladestry Probabilistic Risk ;
8*"
Analyses Figure 1 Project Interrelationship 5 l i l*
l and Regulatory Applications ,
l O
e W W
> CE:
E CE:
aC aC aC I
$ $ Z W W W I > > $ $ $
E E o.
at at E
> H O U U Z Z U '
3
} m e a Y
W Y
W a
Z Z Z Z Z - -
gE
< - - - Ze * -
O O
O M
O O O e O O O O O O 7 7 W 7 LaJ O 7 7 7 7 7 7 J
CE: 7 U U 7 7 6- >
e Z
W .
k & c' w Cz:
E .
at Cz: CL. W
- 7 6 c.
- -e e s m LaJ Z 3M J W Z *
. at CTw O at C Z C 3 C LaJ Cz: aC U w s m g, E 3 La. c. E D C'Y > sD p e aC > CY aC f C at m H m DC aC H
- 7 D.
t/i
' . Z .t;; '
E - Z E -
W
- aC M cc
~m-Cz: O La ee Ex:
W - aC E M E Ct: (D Cu: CD CL 2 Cz: CL *-e 2 Cz:
Q. E aC c. &
LaJ CL W
CL LaJEe H CL CX:
H O. act att Q. ed s LaJ A W A 7 LaJ s 7 D-- LaJ LaJ O LaJ J
LaJ Z .l Z J J I LaJ aC Z Oe CC Z Z Z L CD e
CD E ~
>= C H -
5b<
W H
Cr H Z J
O U
Z W
Z J
Z J
Z CD J O U
Z W
Z J
E LaJ Q
M:
_J Z
3 J
Z Z m aC M CD aC *-* LaJ w O . . . %
U . .
J *
- J J J J U J Jf 8 J Z Z Z T J J LaJ Z Z Z W J J O m J Z Z Z LaJ O Z Z I
Z Z J aC J aC W LaJ Z _J aC J J J W Z aC e=* CD aC CD J CE E ED J J W Z e=* J J J
.Z U.
e- O Cz:
E m
- e
\ Z m e-e Q Z 6-O O aC w E I" M
. LaJ U 6-Z W E O H
~ G O
- 2 W Z C - m m O Z e m aC =
Z - <Cz: O E W CD LaJ e-o Q H w Cz:
ee U m LaJ D
M i= CL D aC >
H aC LaJ H U U M Q H W U E LD Z c., a.aJ at Cz:
LaJ Z M J aC y Q aC alC aC E % Z E F O LaJ U W 0- LaJ O LaJ Z Z U Z LaJ m H Z to WE > alC J
H LaJ M E U Z W Z Z CL l--
Z LaJ m aC M
Q Cz:
3 O .-e M E O *-e M &
M CE: E C
C Z
LaJ m - : J CL E c. 6. c E
2 M
J J
- t. ,
O D
Q. *=*
LaJ C2: Z E E M *-e 4.0 m CD e-o en aC LaJ M C aC K J O O O Q D aC LaJ M Z D LaJ W sk: E w 4A CL LaJ H *.* CD L E L6J M Cr Ex: E > M E
E O M aC LaJ J s m O W La.I e=e Q 6 OC Ck: w w m w > s 6-- J Z E 6 6-LaJ J > J O 3 aC LaJ E aC H 6-Z A s-.e c s t--
Z LD w G G w CD 6 CE: O Z Z Z e-e CL c. l--
LaJ atC W s.aJ O D Z LaJ E D Z Z > #
C C LaJ LaJ E T Z CD - E m O O t3 LaJ SD M M alC O O CD CL E CL Z e o E LaJ _J J :)
CD e O O Z Z Z Z Z - O J s e-e O LaJ LaJ C3 H e-o e-e > Z J I J W M e-o M M Q. Z D D D M e-e e-e > O G e-* e-* CD CL & CL D U LaJ LaJ Z
e-o M
Q in D
CD D
CD C M
CL M
O.
M CL M
CL LaJ E
CY LaJ CY LaJ CY LaJ lfaO- O m CD LaJ ~
E f O O O E ee n.
m ee m M
J D J M CL aC LaJ D M CY LaJ > m CD CL 1
l
l 4
m d 5 2
= b E
. W m 5 E =
E g 5 2 evE 8 E *
=
s = d e 5
~
" g o o u 5 E e w
E E E . W E g 5 E e$ g *E
- E g m
E
=
u u
5 s* -
5 g
=
E g.
w a o 5 5 E m a g i,
W *E E
E 8
e W E o
i gg E E Ww W m . $ d c W O E E W E O 8 E E E b o 8 ~
E x 5 w ~5 E E o 3 E aE em = c "
ee . eg b ~ ~ 8 5 2 5m E 8 W 2 W k 5 8m E E E S E k g E e EC E b E e g "
w C EE
! 5 W
w E
S g
=
l d o w E
b
~
5 E E S W "2 E ew ".
W d g - b E eE (m)
( gE g "
$ 5 s <g E G2 E_ em Ee m a
w 8 m Eg S5 =5 E W *8 W E' g 5 5 g2t w W m 5 - E w E" g g E mE =
e-
=
m aE 8
W
==
M
=
5 wa E W 1
5 a5
>2 YE a g = a E u 9 "
8 g
- E ew 5 8 8 5 5-
- S* w w # 5 E m " E E W 2 5"
E E[28 m
m W
W
=
e w
gE b E*
C 5
w
=
W *8 W
- - ~
5 g
E g - r d =
E 2 W W - m w . 8 E _W W E M 5
- gy "
= s- W bls m-a es E
W EBR s*m 8 eg E
W 5
SE8 Eg= E
- 8 m w w a -
E W E W - E W
- E E W g b 8 5ad W 5 I - 8 $0 g
w ==s
- =
e a" z m g=e-
.m c 2 n =
=
=
W=ae
=
w e w e g E a a 5 =
S
-,-.--,-re,, ,-,4----------m---,--------e, -,-,-e-,,.,,,--,.w...w,---..ew.-..-,w.--.-w,-,. , - , .-,---,-,--,-:-ev, _ _ - = -
N Y
4
- W W c
Es$ I E 5 \ r eaw a - = .
" d eD O I E E
< w
< m5
=w J R W
-am w=!e W
- b E $ E "
W
= 8sa -
- s eea aer o e w
"' =' es m i E - O W w = g w w w 8 E - *
- 5 E 8 M r 0 5e Ega'm W 5
g, esmo E5 ag ao 5
" hE = aeE Eo -
w .
e e-a W
- 5 gB w e 5 0 m m 5 E E E W E G w l 5 eaw a e gW=0s = QG#
- l. 5
- E
ss
e w
= = s5m-r-=g; a e ga=w < m e-es W 5 w
55
=
e=
-w W
E O P g=3 o
s u-b 5 m g
=
w 4 -
g E 5 "
- Ep Eg == x r
5 "2
=
=
< o e
s E_ dalge s =8
=aw 8 -
s e = -
- 5 =e ee=ma m e,a p e
- s" - g- 5 5 0 b gg gE m $ e E'
- e e =
m . -
5 EE$
- og=,EM 8 E E . i. n, wI E e y D 5 m g W E E8 " E B e g *r = c-s#e s
- _ _m n a s =w $ee a o
< w <
E w
5 E d 5 W" E "
E a
- w J w
r 5wE E < = = = *= a w g 5 *# E5
- a 5 E 8 2 5 w W i o - E E e 5 % $ g Egm W - 8 o
eoO "= E SS
. sE E S 8W t I-
" W W d W W W mwg d E WaW . ss=sa=
o
=
E< O E $w r 3 2 -
~
- g. m N
- o i .
. _ . . _ . - . . . . _ - . _ . _ _ , . . _ ~ . _ _ _ , _ _ - - _
O O, JWm W - E
? W o
65d w
E W
W F <
J J
O.
O.
U u w
O m
ted Jaw ZJ EW g H - U E O E F uC3
[ W Om m w < 3 E E o a E W
EE
<<E Em u E - O E N WH E O H H
- m O O Cd y 4J WW w m J <
- M e m - CWS \
< M < U Em8 Wm D WO E
H w
D W
E H
E E O O
8 N
W U
J
< U w
Hmm ZW JD M M w J <ZE HE J Z w W N w W CO ME O E Z e J E wwm AJ 20 m W O w o e M w 04 H E - - D
- 3 W JmM H *> H W H E w m > <wZ m< Z - a U & H Q W >UM
>Z UW EZ W m X < W Z Zn m W (E OW Q 3 W E w
- W CO w EQE D- wm w O W M E e o O OE U e J W H + H w mu m L>E WW <H m um J Z E m Z E HE - EA CA EH <Z W H w H W W UE m MC
<M HM Q m Z E a W < -HM W W Ww C H W w X A
- m <H W
- ZW EH W Z g E E W 6M JJ ZWM WH WC & w A w W WO Z mDD HZH AC >E w m W E a J Q C Cm HE C WH A D = m W M W m - WWW M O W WM MW M < A W O M m ZEA W *E EH Z
- MO m X O w e W H E
Q w@Z Om EW M QC e W J E XM Q e WW w Amw WW WA O O J W 46 m WMZ
> c O J J J Q W mM WH CCH H wm4 < W O H mw ZZ ME w EMH CZ EJ J W MW >
O W Z Zw D4 E WW 3 O O
E E W Q CO +
Z >>O M U w EHM C
hu OO mw < E <M O E W W H + Ow U H DH ME JD Z H E gm EU a LO E
CZJ J %H
- CC < W W O wO OU a
O ZwW UC wu
- UZ H E U w< ><
W JAM W H W ZZ EW C -H m W U HO mm J<H
-Hw D F wm WJ
WHW CM EZ < E O EM Z4 Z W mZ EZH ZD o m ZW W W E m w< Wm HZU ZW< mm H E 4C J O O D E EM
<O WO O E W E 6 O OO ZZ WmH U wED Q o Z HZ wc EZZ Ew <ZJ J. MW w A4 MWW W HMC ~~ MH H OJ J A m H H dw EUO C> Z<> DZ <W 4 J Q W JE HM EHw OMW mW CJ E E> < m o E <<
wW EW Oc WMw mau UE
- H CE 3 O T x U
@ JUM ZOO E Z H> W T W WW wo E. WUC E
W W O 40 W * < H F E <H HO
,W 4
WJO EOZ Em QE W- W D 42 06 ZZW
- MW MC E W J E4 O Em
.Z HWW ZWF JW Z
< Wm MWWF EM WWm wHW <m WO JC U I W
J C Z>E w
DH Wm m HD <UE HD W
mJ CM EH
<Z
>D ZU OE
<H D
W J
WE mW
<H MW Z HWW MC OZ ZOW J <w UZ <H J mo w WW E
ZO CO OZ>
UOw Zw O2 JC Z
ZM HH E
- D J
O b WZ CZ WZ H>EU WW WU bZ mwZ UH w ZO wZ OE < W >D Jw Z WOO 4 MH W U OU 3W N > M 4 hw w< WWW WWw H( H m E mW W O Z JH CW AWZ MDW 6W NCE Wmc KZ CH QE E e H O WZ ZZ E W MW WZ WM OM O M WW 40 OD E (Z M UJE WE HW He wE F M H HE U M mW mO OW WH Ha E U m2 E %WH>
U (H> w EJE Zw WO QA MEW b Ow WE Jw JZ MW m O - w wO H wm C ZOH W3W mW a0 C ww Z w Q Z4 OWU Z h EU U w O H W OH > EJE WJZZ QMF C4 W w U M Z (E
> mw MM DW U UA m . W w A mo ZO Emmo U mDO WH W W U Q HU Ww 234 4
3 H
EWE W UZW EWH OE4 O
w
- EW WE mW W
mE O
>E ZE E w O
W E
H m
E WJ m>
w
<mEZ HC O WQE MC W A W EW ZZ HOOM W Zo 6-Q (m > Q o wW OW EOUM Q -UW ZEm >mM JW Jw E H W EH m OZ
% E WJF Z MWJ Wa< Z JM w
J>
~<
WE HA
& H m E WM o WE UHA<Z WO>
AW EMD O> 3M 3Z 4 H W & E JE U MMW W 6 W W EU m M XZ MO W w W WO 3 W Em WmW W O m MC Em (E o UQ OWQJ WC J *E HE HE <
EW EZ W
H W
E
& M Cm OW Omwo
O E
MZ HC MU D
WHZZ mm< @
MZ w
- EC Ew E
40 Z W w a UM wo J WO MD>
m O
WW W> Z MO O > DM A 4 mw N>w WM HH w O J CZ H U
- m
>mo MWH w
AM MW A4 MZ ZH U > E J Zw Em MMMM WWH O Qw C<w
.. E <EC WJ WW HMw DH O A w UQ m
H EAQ w W
- wZ O .E A
m *w OD Zm E
m J 3
Z A
mW -
W WCWE zZJ O
@ U 20 NM mb OZ J < M4 WE HH E
W Q QM m
- mu GM UM J w W w o m =
E WED mOH m .m 5 mW . J 3 U m Zm D EW<OU O W W W MWE MW MJ M> E MW w a Mw UA Z
H W Z O wow 3 w W mm EW - EW J WEm HW w U < E 6>>H W JW >UO > >E >M w M.> W> Q MF JJ mDM mJ mm m H w> w Z, Z J mE mHwH E wJ EU J W m < EC ag E
M E
Ce EZ WH <H HUEE E &
g Um3 Om ZHU ww<
EM w2 ZZ
-H ZZ w< g Wm W D m m< Eg g gggg C
O .
H Z . . . . .
Ua EZ W mWHB WEZ E
m M W
m C
e w - W n C m =w E Emmm
~ '
O O O CONTAINMENT INTEGRITY RESEARCH PROGRAM RES PROGRAM OTHER ,
CONTRACTOR /SUB-CONTRACTOR SOURCES RELATIONSHIP ELEMENT _
SM./ UNITED ENGINEERS EPRI, FRG COMPLEMENTARY .
EXPERIMENTAL EVALUATION OF OF CONTAINMENT MODELS UNDER SNL/ CHICAGO BRIDGE & IRON EXTREME LOADING SNL EPRI, FRANCE, JOINT CONTAINMENT INTEGRITY UNDER UK EXTREME LOADING '
SNL/INEL EPRI COMPLEMENTARY INTEGRITY OF CONTAINMENT PENETRATIONS UNDER SEVERE ACCIDENT LOADS LANL PVRC COMPLEMENTARY CONTAlf01ENT MARGINS TO FAILURE -
CONTAINMENT BUCKLING
/
4
/'
O O &w (9
TITLE:
! SEISMIC MARGINS STUDIES j FIN; -
l' A 0398 i
j CONTRACTOR:
I LLNL i EXPERT PANEL: R. BUDNITZ, P. AMICO, A. CORNELL, W. HALL, R. KENNEDY, I J. REED, M. SHIN 0ZUKA .
i 4
BUDGET:
l FY 1985 FY 1986 FY 1987
$650K $600K' $1100K i
i j
I i
/3 i
O O O ISSUE:
NEED TO DETERMINE THE CAPABILITY OF NUCLEAR PLANTS TO WITHSTAND EARTH-QUAKES LARGER THAN THE DESIGN LEVEL. '
OBJECTIVE:
DEVELOP THE BASES FOR EVALUATING NUCLEAR PLANT SEISMIC MARGINS.
O ASSESS EXISTING MARGINS INFORMATION (EXPERIENCE DATA, TESTS, PRA'S1 0 DEVELOP MARGINS REVIEW PROCEDURES O IDENTIFY INFORMATION NEEDS e
/
h Eq@ % N
%a. a.3 c- .g. ^
f w m.e wh yuh ile
%'W/S t
~ '
O O~ O SCOPE / ACCOMPLISHMENTS:
FY 1985 0 NRC SEISMIC DESIGN MARGINS PROGRAM PLAN EXPERIMENTS.
0 EXPERT PANEL REPORT ON "AN APPROACH TO THE QUANTIFICATION OF SEISMIC MARGINS IN NUCLEAR POWER PLANTS" (NUREG/CR-4334).
ASSESSMENT OF EXISTING MARGINS INFORMATION DEVELOPMENT OF THE HCLP_F CONCEPT AND SCREENING APPROACH OUTLINE OF REVIEW PROCEDURES .
C we t96u 6 gut,adg p u ,
0 EXPERT PANEL REPORT ON GUIDELINES FOR SEISMIC MARGINS REVIEWS.
O FY 1986 SEISMIC MARGINS REVIEW GUIDELINES REVIEWED BY NRC WORKING GROUP AND FINALIZED AS A NUREG.
- TRIAL SEISMIC MARGINS REVIEWS OF A PWR, MULTILEVEL EARTHQUAKE " TARGETS". % AM')
EXPERT PANEL TO MONITOR TRIAL REVIEWS.
( g-
o
- e O ~
~
o, O FY 1987 REVIEW GUIDELINES FOR PWRS UPDATED BASED ON REVIEW EXPERIENCE AND NEW RESEARCH,
- BWR REVIEW GUIDELINES DEVELOPED,
- TRIAL SEISMIC MARGINS REVIEW (S) 0F A BWR,
- PROCEDURE FOR POST-0BE EVALUATION DEVELOPED, 4
e M
A
J O O O REGULATORY USE:
) 0 SEISMIC MARGINS REVIEW GUIDELINES WILL PROVIDE AN EFFICIENT AND COST EFFECTIVE WAY TO ADDRESS THE SIGNIFICANCE OF PERCEIVED CHANGES IN
! SEISMIC HAZARD, l.'
0 SEISMIC MARGINS REVIEW GUIDELINES CAN BE USED TO IDENTIFY PLANT-SPECIFIC " WEAK-LINKS" AND THE NEED TO UPGRADE SEISMIC DESIGN PROCEDURES.
i i
l l
l
> TITLE:
5 '"'
VALIDATION OF SEISMIC RESPONSE PREDICTIONS ( " P
, A pooc. we sv.
l FIN; Y" -
g 443 ,,, q A 2251, B 5702, D 1603 -
N"" '
l l CONTRACTOR:
i .
ARGONNE NATIONAL LABORATORY 1
BROOKHAVEN NATIONAL LABORATORY i NATIONAL BUREAU OF STANDARDS COOPERATING ORGANIZATIONS:
/ ELECTRIC POWER RESEARCH INSTITUTE-KERNFORSCHUNGSZENTRUM KARLSRUHE (FRG) mtn MINISTRY FOR INTERNATIONAL TRADE AND INDUSTRY,(JAPAN) l BUDGET:
l FY 1985 FY 1986 FY 1987 i $1100K $1750K $1900K i
/f
O O O ISSUE:
)
ANALYTICAL METHODS THAT ARE, ESSENTIALLY, LINEAR ARE USED TO PREDICT SEISMIC RESPONSE OF STRUCTURES AND COMP 0NENTS, THERE IS LITTLE DATA h wev6m M m:
TO INDICATE HOW WELL THESE METHODS CAN BE USED TO PREDICT BEHAVIOR ,9 g i NEAR FAILURE, NOR IS THERE A CLEAR PICTURE OF HOW GOOD PREDICTIONS h i I U" HAVE TO BE FOR SOUND DECISIONS ABOUT PLANT SAFETY. ~ ** T
) . OBJECTIVE:
l OBTAIN INFORMATION THAT CAN BE USED TO DEVELOP ACCEPTANCE CRITERIA FOR f
j PREDICTIONS OF THE BEHAVIOR OF NUCLEAR POWER PLANTS SUBJECTED TO LARGE l EARTHQUAKES.
i INTEGRATION:
l l 0 THE QUESTION OF ADEQUACY OF PREDICTIVE METHODS HAS A STRONG CONNECTION
! WITH EFFORTS TO DETERMINE SEISMIC SAFETY MARGINS, 0 ONG0ING NRC-SPONSORED EXPERIMENTAL PROGRAMS ON CONTAINMENTS AND OTHER CATEGORY I STRUCTURES ARE AN INTEGRAL PART OF THE VALIDATION EFFORT, l
l
/9
O O O SCOPE / ACCOMPLISHMENTS:
0 FY 1985 DEFINE SCOPE OF COLLABORATION IN HDR SEISMIC TESTS.
LOW LEVEL EXCITATION OF EPRI MODEL IN TAIWAN.
EXPLORATION OF INFORMATION EXCHANGE POSSIBILITIES WITH MITI.
SCOPING CALCULATIONS FOR EXTENDING JAPANESE TESTS AT TAD 0TSU FACILITY INTO THE INELASTIC RANGE.
O FY 1986 i
PARTICIPATION IN LARGE SHAKER TESTS AT HDR.
PRE-EARTHQUAKE PREDICTIONS OF RESPONSE OF EPRI MODEL. Oainw}
IMPLEMENTATION OF INFORMATION EXCHANGE AGREEMENT WITH MITI.
O FY 1987 PARTICIPATION IN LOCAL-PLASTIFICATION TESTS AT HDR.
CONTINUATION OF INFORMATION EXCHANGE AGREEMENT WITH MITI.
ASSESSMENT OF PREDICTIVE MODELS FOR EPRI MODEL RESPONSE. (uiura.w)
- a. aRs, dah wih.tartN,%.
2 tq. 6 -TAtdiw's a faiD. - -ruudL u% . w. ~
4
'i AOonu w - 2s,wo ,/ t m _t 4 u pa .
O O O REGULATORY USE:
0 ASSESSMENT OF SEISMIC MARGINS AT' EXISTING PLANTS, 0 JUDGEMENT ON THE RELIABILITY OF PROBABILISTIC RISK ANALYSES FOR i EARTHOUAKES. ]
l
3O&
~ ' ~
~
O O TITLE:
COMPONENT FRAGILITY DATA ACQUISITION AND EVALUATION FIN:
A 3278 -
CONTRACTOR: .
BROOKHAVEN MATIONAL LABORATORY (BNL) -
PROJECT MANAGER: C. H. HOFMAYER BUDGET: ,
FY 1985 FY 1986_ FY 1987_
$21:0K $1100K $1000K l
>1
O O O ISSUE:
MnRE RELIABLE AND REALISTIC COMPONENT FRAGILITY IT IS - ud INFORMAT d>Jnt J FOR SEISMIC MARGIN STUDIES AND PROBABILISTIC RISK ASSESSMENTS *C NOW BELIEVED THAT PESSIMISTIC ESTIMATES OF THE SEISMIC - h i eTHREAT
.Jt a NUCLEAR POWER PLANTS HAVE DEVELOPED BECAUSE OF '*) 9%CON UNDUE THE INPUT COMPONENT FRAGILITIES. %
- % taqt a %%
OBJECTIVES:
- a .'\.a. 6 %
, g gg FY 1985 0 TO INITIATE COOPERATION WITH DOMESTIC AND FOREIGN INSTITUTI a TAIN ALREADY EXISTING COMPONENT FRAGILITY DATA.
~
tahwweq,,% W pru - u a n <
TO ASSEMBLE, ANALYZE AND INTERPRET EXISTING COMP 0N$NT FRAGILITY DATA 0
AND COMPARE WITH INFORMATION CURRENTLY BEING USED.
NA "%* -
l 0 TO DEVELOP A SCHEME FOR PRIORITIZING AND GROUPING COMPONE l
i GILITY TESTING AND TO DEMONSTRATE A PROCEDURE FOR PERFORM l
FRAGILITY TESTS WHICH WILL PROVIDE A BASIS FOR COMPREHENSIV (LLNL OBJECTIVE
! IN FUTURE YEARS AND USEFUL FRAGILITY DATA IN FY 1985.
UNDER A DIFFERENT FIN).
l .
13 i
i Q
I
! OBJECTIVES: (CONTINUED) i 4 ,
! FY 1986 AND 1987 '
1 l
0 TO REDUCE BY ADDITIONAL DATA ACQUISITION AND SOME LI TESTING THE INADEQUACIES OF THE PRESENT COMPONEN l
l 0 TO IMPROVE SEISMIC PRAs AND OBTAIN BETTER ESTIMATE '
) BY USING MORE REALISTIC TEST BASED COMPONENT FRAGILITI i CHARACTERIZE ACTUAL FAILURE MODES. .
l SCOPE:
SCOPE INCLUDES ELECTRICAL AND MECHANICAL COMPONEN WITH EMPHASIS ON ELECTRICAL COMPONENTS SINCE THEY A (MECHANICAL COMPONENTS ARE INHERENTLY TORS ACCORDING TO PRESENT VIEWS, RUGGED SINCE THEY MUST, IN GENERAL, BE DESIGNED FOR TEMPERATU '
ONLY ACTIVE MECHANICAL COMPONENTS TREATED, BUT BOTH ACTIVE SURE LOADS). PIPING IS EXCLUDED FROM f'
AND PASSIVE ELECTRICAL COMPONENTS ARE CONSIDERED.
THIS PROGRAM.
l i
[
l O o O ,
I l
SCOPE: (CONTINUED)
! NO EMPHASIS IS ON SEISMIC LOADS, BUT CONCURRENT OPERATING LOADS IN ALSO SEVERE ENVIRON-l j OTHER SEVERE ENVIRONMENTAL AGING, SCALING AND ORANCHORAGE ACCIDENT LOADS TREATED VARIABILITY MENTAL EFFECTS EXCLUDED. HOWEVER, l
I INCLUDED. NOT NECESSARILY DEV0TED TO OBTAINING FRAGILITY LEVEL
- TESTING OR DATA ACQUISITION MUST DEMONSTRATE THA
- THAN PRESENTLY ASSUMED BEFORE BEING DISCONTINUED.
ACCOMPLISHMENTS: (SINCE NOVEMBER 198ti) 0 OBTAINED FRAGILITY INFORMATION FROM GE, WPPSS, BROWN-B0 VERI, -
EVALUATION OF TEST SDRC, FARWELL a HENDRICKS, ANCO AND OTHER SOURCES.
RESULTS UNDERWAY.
0 ESTABLISHED COOPERATIVE AGREEMENT FOR INFORMAT '
j TWO-WAY INFORMATION -
TRANSFER IN PROGRESS.
]
l 0 HOSTED THREE DAY WORKSHOP AT BNL WHERE CRITICA (SEVENTY-FIVE ATTENDEES AND TWEN-j AND TEST RESULTS WERE REVEALED. "'^ d-
' TY-TWO PRESENTATIONS), c 00rtg VSE 09 '
0 DEVELOPED PRIORITIZATION SCHEME AND IMPORTA I PONENTS.
\
I
- >c
o O O ACCOMLISHMENTS: (CONTINUED)
EXECUTION 0 DEVELOPED STANDARDIZED COMPONENT FRAGILITY TEST PROCEDURES OF TEST PROCEDURES BEGINS ON SEPTEMBER 24, 1985, FOR APPROXIMATELY TWO MONTHS USING A THREE BAY MCC LOADED WITH ELECTRICAL COMPONE TESTED FOR BOTH STRUCTURAL AND FUNCTIONAL FRAGILITY, FUTURE PRODUCTS:
FY 1986 0 RECOMMENDED LIST OF COMPONENT FOR TESTING BASED ON PRIORITIZ GROUPING.
O TEST PLANS AND TEST RESULTS FOR FRAGILITY TESTS ON ELECTRICA NENTS MOUNTED IN A THREE BAY MOTOR CONTROL CENTER OF VARYI TURAL STIFFNESS AND VARYING ANCHORAGE.
0
SUMMARY
OF CURRENT UNDERSTANDING OF COMPONENT FRAGILITY IN N POWER PLANTS,
) C;
O -
O FUTURE PRODUCTS: (CONTINUED)
FY 1987 0 RECOMENDATIONS FOR COMPONENT FRAGILITY LEVEL COMPONENTS FOR SEISMIC MARGIN AND PRA STUDIES, REGULATORY _ GUIDES OR STANDARD REVIEW PLAN SECTIO NONE; HOWEVER, COULD LEAD TO THE ESTABLISHMENT OF A C l
TY DATA BANK AT ONE OF THE NATIONAL LABORATORIES.
REGULATORY USE:
RESPONDSTOQUESTIONSRfLATINGTOEASTCOASTSEISMICITY,PART INTENDED TO PRO-THE CHARLESTON AND NEW BRUNSNICK EARTHQUAKE ISSUES.
VIDE A RATIONAL BASIS FOR DECISION SHOULD THE CENSING REQUIREMENTS ON EAST COAST OPERATORS,
~
O O 9"r:$
TITLE:
l .
SEISMIC CATEGORY I STRUCTURES PROGRAM
- u mpw e ; stru a. %T towk&M ;
i FIN-l
! A 7221 CONTRACTOR:
j LOS ALAMOS NATIONAL LABORATORY i BUDGET:
l FY 1985 FY 1986 FY 1987
! $820K $1200K $1500K 1
)
I i
1 4
i
\ . .
O O O ISSUE:
ABILITY OF CURRENT LIGHT WATER REACTOR PLANTS TO ACCOMODATE LARGER EARTHQUAKES THAN THOSE CONSIDERED IN THEIR INITIAL DESIGN.
OBJECTIVE:
REDUCE UNCERTAINTIES IN METHODS USED TO PREDICT HOW THE INCREASED Il EARTHQUAKE LOADS ARE TRANSMITTED BY THE PLANT BUILDINGS 10 THE SAFETY SYSTEMS AND COMPONENTS NEEDED TO OPERATE AND SHUTDOWN THE PLANT.
4 0 EXPERIMENTAL DATA ON THE SENSITIVITY OF STRUCTURAL BEHAVIOR TO
- VARIATIONS IN CONFIGURATION AND EARTHQUAKE LOAD.
0 CHANGES IN FLOOR RESPONSE SPECTRA RESULTING FROM INCREASED EARTH-l QUAKE LOADS. ,
! O EXPERIMENTAL' DATA TO VALIDATE COMPUTER PROGRAMS USED TO PREDICT l: STRUCTURAL BEHAVIOR IN THE-ELASTIC AND INELASTIC RANGES.
i i
l if
o
~
O O
~
INTEGRATION:
0 STRUCTURAL BEHAVIOR INFORMATION WILL BE PROVIDED T0:
VALIDATION OF SEISMIC CALCULATIONAL METHODS PROJECT SEISMIC DESIGN MARGINS PROJECT 0 FLOOR RESPONSE SPECTRA INFORMATION WILL BE PROVIDED T0:
PIP _I_NS_RELIARILITY PROJECTS
- COMPONENT _fjQGILIlY PROJECT e
a 36
--.G-O O O SCOPE / ACCOMPLISHMENTS:
0 PRIOR TO FY 1985 STATIC OR SEISMIC TESTING 0F 23 DIFFERENT MODELS R PRESENTING TWO TYPES OF STRUCTURES. .g , f C
MODELS BUILT TO TWO DIFFERENT SCALES (0NE INCH AND THREE INCH WALL THICKNESS). '
s on l>Ug gg ,% g NUMBER OF FLOORS IN THE MODELS VARIED FROM ONE TO THREE. 4,jup i. 2.
STIFFNESS OF THE MODEL WAS LOWER THAN THE COMPUTER UNCRACKED CROSS-SECTION VALUES BY A FACTOR OF 4.
S_CALEABILITY OF RESULTS ILLUSTRATED IN THE ELASTIC AND INELASTIC RANGE, HOWEVER, ALL MODELS WERE FABRICATED USING MICR0 CONCRETE.
/
1 1
k j V
O O
~
O SCOPE / ACCOMPLISHMENTS (CONTINUED) 0 FY 1985 WITH THE AID OF THE PROGRAM TECHNICAL REVIEW GROUP A MODEL WAS DESIGNED TO HELP RESOLVE THE STIFFNESS DIFFERENCE ISSUE ( . 4,j AND PROVIDE CREDIBILITY FOR SCALE MODEL EXPERIMENTAL DATA. unw h sk!.i w%ad ~~ p %
TEST TWO 1/4 SCALE MODELS OF THE TRG STRUCTURE. THESE MODELS -
WERE FABRICATED WITH MICR0 CONCRETE AND SIMULATED REBAR.
START FABRICATION OF THE FIRST TRG STRUCTURE. THIS MODEL WILL BE 7a ,a 1 441) v w-; =>
FABRICATEDWITHPROTOTYPICAL[0NCRETEANDREBAR. ,
4000 p w g
/
O
~
O O O SCOPE / ACCOMPLISHMENTS (CONTINUED) 0 FY 1986 TEST TWO TRG STRUCTURES; THE FIRST ONE WILL BE SUBJECTED TO LOW Uh LEVEL STATIC THAN INCREASED SEISMIC EXCITATION TO FAILURE (NOVEMBER 1985); THE SECOND ONE WILL BE TESTED QUASISTATICALLY TO W
j FAILURE (MARCH 1986).
tas.
PT 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 NU,MBER OF EXPERIMENTS TO MEET PROGRAM OBJECTIVES AND AID IN BENCHMARKING THE ANALYTICAL MODEL DEVELOPMENT.
'A' 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.
na 4 tmp.-b a%bp) 3h
' ~
~
O O O
~
SCOPE / ACCOMPLISHMENTS (CONTINUED) 0 FY 1987 i
ENHANCED LNALITICAL EFFORT TO REDUCE THE NUMBER OF TESTS AND l BENCHMARK ANALYTICAL METHODS, i
SEISMIC TESTS ON 1 FLOOR MODELS WITH 1 INCH THICK WALLS (IF FY 1986 l RESULTS ALLOW) TO DETERMINE THE AFFECTS OF WALL ARRANGEMENTS AND EARTHOUAKE INPUT.
SEISMIC TESTS ON 2 FLOOR MODELS WITH DIFFERENT WALL
! CONFIGURATIONS ON THE FIRST AND SECOND FLOORS (3 OR 4 INCH WALL l THICKNESS). (*. d W quae) '
I i
]
i Y
~ '
O O ,
O REGULATORY USE:
0 PROVIDE EXPERIMENTAL DATA AND ANALYTICAL METHODS TO DETERMINE THE BEHAVIOR OF REINFORCED CONCRETE BUILDINGS SUBJECTED TO EARTHOUAKES LARGER THAN THE DESIGN BASIS.
! O RECOMMEND IMPROVEMENTS AND REDUCE UNCERTAINTIES IN REGULATORY GUIDES I
(E.G., 1.61, 1.122) AND STANDARD REVIEW PLANS (E.G., SECTIONS 3.7, i
3.8, 3.9 AND 3.10).
i 0 ESTABLISH NEW REGULATORY CRITERIA CONSIDERING NONLINEAR RESPONSE j EFFECTS.
i j 0 WHEN COUPLED WITH COMPONENT AND PIPING RESEARCH HELP DETERMINE SEISMIC l DESIGN MARGINS IN OPERATING PLANTS.
) -
1 l
I j
l 3
i
~
] 8 -
N: S f .I, TITLE:
RELIABILITY ANALYSIS OF NONLINEAR BEHAVIOR OF CONCRETE STRUCTURES FIN; A 3292 CONTRACTOR:
BROOKHAVEN NATIONAL LABORATORY BUDGET-
- E'*< '
Wa. u LG.. m' _.., wh o \ g u.y . s . . }'. . . mu.at ovG , 6 md fM.
3 FY1985;FY1986 FY 1987
- Q Q & , y'n e..g A su a ,a
, $0 K i $200K $400K uoL@[M . -
t 0M reert ,
i
O O O OBJECTIVES:
0 DEVELOP STOCHASTIC, PROBABILISTIC, NONLINEAR ANALYSIS METHOD FOR MORE RIG 0ROUS 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: Fi % '
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 ACCELEf(ATION, 32
- O O O
, TASK 3. INITIATE EXTENSION OF THE RELIABILITY ANALYSIS METHOD TO INCLUDE NONLINEAR STRUCTURAL RESPONSE.
i
! m/w4=
W* w.,1. . u,- 1, W i, h (
j TASK 11. COMPLETEREPORTONSTOCbSTICEQUIVALENTLINEARIZAIION
- nt - ph beh j TECHNIQUE FOR CATEGORY I STRUCTURES. **
- n <v -
~
l I
17 OM senut u 14 i T Y W *-
i i
--o @ sam hjjlAN ( d KPs Pabo) fMMte
,,, 7 e (de>NL) h b t>dw-
- W' O-en.,g d.h; M 3 6- + %
... a ,
. . 4 ~d 4e> pu A c z_s c u m .
i
- 3 i
i
- REGULATORY USE:
l 0 DEVELOP MORE RIGOR _0US STRUCTURAL FRAGILITY ESTIMATES FOR USE IN PRAs l AND SEISMIC MARGIN STUDIES.
l 0 EVALUATION OF EXISTING STRUCTURAL CAPACITIES AND PRAs FOR OPERATING i REACTORS.
0 EXTEND BNL-DEVELOPED PROBABILITY-BASED LOAD COMBINATION METHODOLOGY TO
- INCLUDE NONLINEAR BEHAVIOR.
i i
i i
i 3
o
~
o co TITLE:
STRUCTURAL DEGRADATION FIN; TO BE ASSIGNED CONTRACTOR:
TO BE SELECTED BUDGET.
I FY 1985 FY 1986 .
FY 1987
$0 K $0 K $300K M _5
' i r; koi W pJeig Je4 [ '-
l
OBJECTIVE:
PREPARE NRC STAFF TO BE ABLE TO IDENTIFY:
O STRUCTURAL ADEQUACY UNDER EXTENSIONS OF PLANT DESIGN LIFE.
O POST-LOCA ENVIRONMENTAL EFFECTS ON STRUCTURAL RELIABILITY.
i 0 HOW Tl) ADJUST STRUCTURAL CAPACITIES AFFECTED BY DEGRADATION WHEN PER-l FORMING SEVERE ACCIDENT ANALYSES.
{ 0 WHAT SECONDARY EFFECTS RESULT FROM DETERIORATION OF MATERIALS UNDER j OPERATING OR POST-LOCA CONDITIONS.
a l SCOPE:
0 BEGIN INVESTIGATIONS INTO LONG-TERM CHEMICAL REACTIVITY BETWEEN SPE-CIFIC TYPES OF CONCRETE AGGREGATES, ADDITIVES (SUCH AS FLYASH), AND
] CEMENTS OF DIFFERENT ALKALI-NITIES.
l 0 MORE ADVANCED UNDERSTANDING 0F HOW DIFFERENT CONCRETES BEHAVE UNDER i U U LONG-TERM CONDITIONS OF LOCALLY HOT (150 F - 300 F) TEMPERATURES SUCH AS THOSE AT PENETRATIONS.
\ 4/
4
~
O O O i
j 0 EXAMINE AND DEFINE UNIQUE ENVIRONMENTAL CONDITIONS THAT EXIST l POST-LOCA IN THE CONTAINMENT AND THEIR EFFECTS ON STRUCTURAL MATERIALS 3
0F 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 -
i
! O LIMITS ON HOSTILE OPERATING ENVIRONMENTAL EXPOSURES 0 REDUCTION OF LICENSING RELIANCE PURELY ON INSPECTION AND SURVEILLANCE -
l 0 IMPROVEMENTS IN DAMAGE INSPECTION AND INCORPORATION INTO l SRP-REFERENCED NATIONAL DESIGN STANDARDS ,
WILL ENABLE NRC STAFF T0:
0 BETTER ACCEPT MATERIAL SELECTIONS i
i 0 REQUIRE APPROPRIATE DESIGN PROVISIONS PERMIT CONTINUED OPERATION NEAR, AT, OR BEYOND 140-YR DESIGN LIFE
) 0 i
\ .
O O O O DETERMINE POST-LOCA CONTAINMENT ENVIRONMENT (THERMAL, RADIATION, CHEM-ICAL) SHORT AND LONG TERM IMPACTS ON STRUCTURAL MATERIALS 0 ASK LICENSEES THE RIGHT QUESTIONS WHEN REVIEWING REQUESTS FOR EX-TENSIONS OF PLANT DESIGN LIFE EXAMPLES:
0 MICR0/ SHRINKAGE CRACKING STIFFNESS DEGRADATION IN LANL SHEAR WALL TESTS 0 USE AT SEABROOK OF AGGREGATE SHOWING ALKALI-SILICA REACTION AT PEASE AFB RUNWAY 0 CHEMICALREACTIVITYINONThRIO-HYDR 0CONTAINMENTBLDG 0 TENDON FAILURES IN ORNL MODEL PCRV, PERHAPS DUE TO THERMAL DECOMPOSI-TION OF TENDON GREASE-0 W. R. GRACE & CO. EVIDENCE THAT HALOGENS, ESP. PVC, DECOMPOSE LINEARLY VS. AT SOME THRESHOLD LEVEL OF ENERGY INPUT ffU#)
l 0 LONG TMI-2 CONTAINMENT INACCESSIBILITY, W/ MOISTURE, RADIATION, BORON, .
H 2 BURN EFFECTS ON RELIABILITY 0 WOLSUNG 1 (KOREA) CONCRETE BASEMAT SWELLING, DUE TO ALKALINE CEMENT REACTION W/ VOLCANIC AGGREGATE O DEGRADATION OF SONGS-1 EQUIPMENT SUPPORTS DUE TO PLANT ENVIRONMENT 0 9 MILE PT.-1 REACTOR BLDG WALL CRACK, POSSIBLY DUE TO THERMAL GRADIENT
- h ) ( g T*
6 - L u p m ,(
At&% : eqe b Asg Wgga 40 wM p im _ p u A fi phoo u. w ,.;g, Me: Mw dibt
S Y Q
O TITLE:
STANDARD PROBLEMS FOR STRUCTURAL COMPUTER CODES FINS A 3242 CONTRACTOR:
BROOKHAVEN NATIONAL LABORATORY (BNL)
BUDGET:
FY 1985 FY 1986 FY 1987
$300K $200K $200K M
ISSUE:
VARIOUS ANALYTICAL APPROACHES WITH DIFFERENT DEGREES OF APPROXIMATION HAVE BEEN DEVELOPED AND USED TO DETERMINE STRUCTURAL RESPONSE AND S0ll-STRUCTURE INTERACTION OF NUCLEAR STRUCTURES TO SEISMIC AND DYNAM-IC LOADS. APPROXIMATIONS 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.
e
o O
~
INTEGRATION:
0 THE FINDINGS OF THIS PROGRAM CAN BE USED IN EFFORTS TO DETERM MIC SAFETY MARGINS, O EXPERIMENTALRESyLTS.FROJj[ARIOUSSTUDIESCURRENTLYBEINGPERFORMED FORTHENRCATSNLIf0NTAINMENTS),WCC(E.0, CHARACTERIZATION)AND LANL CATEGORY I STRUCTURES) WILL COMPLEMENT THIS PROGRAM, ALSO WORK FUNDED BY EPRI (SIMQUAKE) HAS BEEN USED, O DATA HAS ALSO BEEN OBTAINED THROUGH COOPERATIVE %up% AGREEMENTS 9
M 4
~ ~
O O -
O SCOPE / ACCOMPLISHMENTS:
i 0 FY 1985 I - WORK WAS CONCENTRATED IN THE.SSI.. AREA; SPECIFICALLY, WITH DE-TERMINING THE EFFECTS OF FOUNDATION LIFT 0FF, HIGH GROUND WATER TABLE, AND SOIL LAYERING ON SSI ANALYTICAL METHODS, i
0 FY 1986 g
) -
COMPLETE FY 85 EFFORTS (WATER TABLE) i -
INITIATION OF A STRUCTURAL DATA SURVEY AND EVALUATION OF DATA
~ ~ ~ ~ 1% nyte - ^"* \* u * @ U~3'
) -
WORKSHOP DN SSI j 0 FY 1987 b ds 4 b e ws .
4 j -
DEVELOP SSI BENCHMARK PROBLEMSg56 3 44 M 4J DEVELOP STRUCTURAL (CONCRETE) BENCHMARK PROBLEMS EVALUATE AVAILABLE NEW DATA WITH THE AIM 0F IMPROVING BENCHMARKS i
i ers- we A6na g Lom u ...ma m,yL sa_,aa_
O 4
~ ~
O O O V'b REGULATORY USE: ,7 0 PRELIMINARY FINDINGS HAVE BEEN USED BY BNL TO HELP NRR (SGEB) RESOLVE 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. 4 p % e awhi eJ os2O .
O BENCHMARKSANDOTHERDATADEVELOPEDWILLBEUSEDTO) EVALUATE / REVISE AND WHERE NECESSARY ESTABLISH NEM SEISMIC DESIGN CRITERIA (I.E., SRP, REG. GUIDES). ,1 49
~ ~
\ O -
O O i
j REPORTS:
i
. S. SHARMA, M. REICH, AND TY CHANG, BROOKHAVEN NATIONAL LABORATORY,
" REVIEW 0F CURRENT ANALYSIS METHODOLOGY FOR REINFORCED CONCRETE
- STRUCTURAL EVALUATIONS," USNRC REPORT NUREG/CR-3284, APRIL 1983.
i C. A. MILLR, C. J. COSTANTINO, A. J. PHILIPPACOPOULOS, AND M. REICH, i
BROOKHAVEN NATIONAL LABORATORY," VERIFICATION OF SOIL-STRUCTURE INTERACTION METHODS," USNRC REPORT NUREG/CR-4128, MAY 1985.
i i
i I
j -
i 4 .
t 4
i
[
o o A:
l _% - s o r. t uy
- w 2;q tRe $#~
- TITLE: % t iew
{ L. des s i NRC/EPRI_ PROGRAMS ON PIPING DYNAMIC LOAD CAPACITY l 0 EPRI/NRC COOPERATIVE PROGRAM " PIPING AND FITTING DYNAMIC RE-i 5P'" 3
! LIABILITY" (FIN D1157) - * '"' M ' l'
! O ETEC DEMONSTRATION SEISMIC FRAGILITY PIPE TEST (FIN B3052)
- *? F 0 INDEPENDENT ASSESSMENT OF PROPOSED ASME PIPING RULE CHANGES *g9 PY' 0 INTEGRATION WITH NRC DEGRADED PIPING PROGRAM l 0 PIPING FRAGILITY REDEFINITION FOR PRA USE CONTRACTOR
l GE (THROUGH EPRI), ETEC, ANCO, OTHERS i
MSEB BUDGET:
LM.
3 FY 1985 FY 1986 FY 1987 N A%
$200K $700K ' $900K Lao
)
% s__I-een_
usg 2 no' b uo l
MNM lN ampAaa i
O i -
~
~
O O O
~
j ISSUES:
- O ASME CODE DESIGN RULES FOR PIPING OF UNDER DYNAMIC LOADS MAY HAVE AN UNREALISTIC (0VERALLY CONSERVATIVE) BASIS.--tg S i 0 PRA PIPING FRAGILITY ESTIMATES NEED VALIDATION.
OBJECTIVES:
0 EXPERIMENTALLY DETERMINE DYNAMIC FAILURE LEVELS AND MECHANISMS (FA-
) I TIGUE/RATCHETTIN.G,0R COLLAPSE) FOR PIPING SYSTEMS AND COMPONENTS.
! u v.i i
- 0 DEVELOP NEW DESIGN RULES FOR ASME CLASS 1, 2, AND 3 PIPING.
i
! O ASSESS THE IMPACT OF THESE PROPOSED CHANGES.
i 0 IMPROVE PRA PIPING FRAGILITY ESTIMATES.
l 0 BENCHMARK NONLINEAR BEHAVIOR IN THE PIPE FAILURE RANGE.
k vas.* h8 & whau itsqu.y n d w )') &
l wu a w,gp.
i i
~
\ .
SCOPE / ACCOMPLISHMENTS:
0 FY 1985 NRC/EPRI COOPERATIVE AGREEMENT ESTABLISHED PROGRAM PLANNING COMPLETED ANC0 COMPONENT TESTS INITIATED (FIRST 12 0F 40 TESTS)
ETEC DEMONSTRATION TEST INITIATED 0 FY 1986 COMP 0NENT TESTING CONTINUES SEISMIC PIPING SYSTEMS TESTS (SEISMIC) ---" "Y'd* *
- t^> -
FATIGUE /RATCHETTING S,PECIMEN 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 f.3
~
O O O
~
REGULATORY USE:
0 PROVIDES THE BASES TO ENDORSE CHANGES TO ASME CODE RULES (SECTION III, SUBSECTIONS NB/NC/ND 3600).
O PROVIDE FURTHER JUSTIFICATION FOR SNUBBER MINIMIZATION.
0 ESTABLISH PIPING FAILURE MECHANISMS AND MARGINS.
e O
/
yQ ad .
i @ 'e U U - TEST ARTICLE PIPING SYSTEM l
' k Y
' 4 6 ,r
/N x x . I i
N.
\
d'(* PIPE MATL: ASTM A-106 GRADE 9
,> ~
(/
6"sca 40
'0.' #
NOTES-l (1) EXISTING DUMMY VALVE j
{2) ALL WILDS TO BE MADE RADIOGRAPHED AND INSPECTED PER SECTIONS III AND N h IX OF ASME CODE (3) ELB0tlS ARE 90' LONG RADIUS
/
[N-#
agt) D *-h.
(4) FLANGES ARE 600 LB CLASS yk ,
PROTOTYP!C CLASS 1 '
PIPE ATTACMENT (l,g*
. X I .
s[.6 3" ses.fo - .
2'.4'i .
f$$
/ 97*
v
'%f q b'
~dC -
O
@s t;,p "
e)
%p d4 s.s Q a
. - ar su u.4, u 1,.
- SW%N
7 0
g .
2
- M g
pl, g .s i
'd .
% A - N . +
I t ,_ ,
r m'
\ -.
l '
/ l i
- l
/ 3i ,
j <__=
- >r *---_
/ .
e E[ {l /, , j e l i llI e
, i< ,
...l I l H M*4 [f'I '-sis ~ ~' N ~Dz '
I q n vj v b;e 3
=. a
\:
E ]p E 12
!bh f- -
I .
+
o o
J
,- , i s
c I *(-( 8 . O O
00 o
o **
l l
- l
[
l l-f }
e-- G 1
! I~ .
~
I ..
3-2 g
l i
i O -
n a
f 4
l ,:
Q 1
rr T2 -
~ j "gg !
w - .. - .
Figure 1. Typical 5-in. pipe arrangement.
I 1
- . - l 5-in. 3-in.
System (Fig. 1) System (Fig. 2)
End supports only no supports no supports
~
Hydraulic snubber 1 Mechanical snubber 1 1 Rigid strut 1 2 Combination strut 1,2 -
1,2 l and snubber j
. Strut combination 1,2
. -- . - --. _ .._ _ . -- _ __ __.___---.- - ._..= -. ._ --.- - --__--
i O I l i
O
+
E 8
- ++.
+$ -
1
- 3 I + ~
i + ++++ _o 3
- 8 3
+ +,% $ "
l 4 +*+7 +
mj + -
+# +
{ **
l- Q w* +
i
+- * * +_O*%*
- + $ *++$+ '
+
l hk , +$+*$
z .; +
! + + [
\ O]
- +1 +
++ _q d
- 4. +
++
! M + +b+
+
i Uk. = +
- +
i 4 + ++ + +$t + t i '
++ + +g * '
o
,' + + n
' ++ +
.g. +++
- g 4++*,44g
+ + + gH -
i
+
++ ++
! + +
i4 iiii4 i ii4 4 4 1 4 4 4 4 y**% % N~
ff ffff ,
O ~ " ~ ' = '~>' ~ "
w I g
TITLE:
OTHER NRC PIPING RESEARCH ( h h h chau, u u[A m)$o - )
N0ZZLE DESIGN GUIDANCE (ORNL FIN B0474)
COMBINATIONAL PROCEDURES FOR PIPING RESPONSE SPECTRA ANALYSIS (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) - Tm' + kbg M _
wpp s pud, "T A BUDGET:
' (vn opudi FY 1985 FY 1986 FY 1987
$333K $375K ,
$600K
O O O ISSUES:
O 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-l SIGN ISM METHOD IN COMBINATION WITH PVRC DAMPING RECLASSIFICATION OF SEISMIC STRESSES i -
SUPPORT DESIGN SCOPE / ACCOMPLISHMENTS:
O FY 1985 AND 1986 l 0 IMPROVE N0ZZLE DESIGN GUIDANCE i
0 EVALUATE ISM METHOD IN COMBINATION WITH PVRC DAMPING i 0 IMPROVE MODEL COMBINATION METHODS ISM WITH KNOWN PHASE CORRELATION l
CLOSELY SPACED MODES j -
HIGH FREQUENCY MODES
! 6#
i l
1
o o
~
O
, O EVALUATION (AND VALIDATION) 0F SIMPLE NONLINEAR RESPONSE PREDICTION TECHNIQUES 0 EVALUATION OF APPROPRIATENESS OF R.G. 1.122 PEAK BROADENING RANGE l
i 0 DEVELOPMENT OF BASES FOR " FLATTENED AND BROADENED" SPECTRA
- 0 RECOMMENDATIONS FOR DAMPING VALUES .
TO BE USED IN TIME-HISTORY ANALYSIS l 0 FY 1987 0 INVESTIGATION OF COMBINED EFFECTS OF NEW PIPING DESIGN CRITERIA i
I -
DAMPING NEW ASME DESIGN RULES
- I -
NEW N0ZZLE DESIGN CRITERIA PIPING SUPPORT DESIGN RULES I
i i 6f i
i
__ _ ~ _ __._____1 ____ ~
o O O TITLE:
PARAMETERS INFLUENCING DAMPING IN PIPING SYSTEMS ( *^* "j) 10f Gjtet.a.i )
FIN A 6316 CONTRACTOR:
INEL BUDGET:
FY 1985 FY 1986 FY 1987
$230K $200K $0 K
. 3 l
O O O ISSUES:
0 OVERALLY STIFF PIPING DESIGNS HAVE RESULTED DAMPING VALUES.
O DATA AND GUIDANCE IS LACKING FOR HIGH FRE BY NONSEISMIC LOADS.
OBJECTIVES:
0 COMPILE AND EVALUATE EXISTING DAMPING DATA 0
TEST SIMPLE SYSTEMS, VARYING RESPONSE AND DESIGN PARAMETE .
_ SCOPE / ACCOMPLISHMENTS:
O FY 1985 0 ESTABLISHED WORLD DATA BASE ON PIPE DAMPING 4
43
-- - - ....,,.r.2.... 7.,c.. .
, y ,, g j..
O
~
O O O TESTING OF 3-DIMENSIONAL SYSTEM LOW STRAIN AMPLITUDES DIFFERENT SUPPORT CONFIGURATIONS PRESSURIZED AND UNPRESSURIZED DIFFERENT FORCING FUNCTIONS 0 REVIEW 0F EXISTING HIGH FREQUENCY DAMPING DATA 0 FY 1986 0 CONTINUED TESTING OF 3-DIMENSIONAL SYSTEM HIGH FREQUENCY LOADS INSULATION HIGH STRAIN SEISMIC LOADS 0 TESTING OF 1-DIMENSIONAL SYSTEM AT HIGH FREQUENCY 0 FINAL REPORT
' ~
O
~
- O O
, fvae %d n> - 4 ) \
REGULATORY USE: '^ g '
l'
. Y b n O PROVIDE BASIS FOR ENDORSING ASME CODE CASE 6 "Y N fill )
- O PROVIDE BASIS FOR REVISING R.G. 1.61 AND SRP i .
i t i i
i .
l 1
i i
I l
a
/
b$
o o o I CONTAINMENT INTEGRITY PROGRAM ACTIVITIES
) FIN A-1401 ANALYSIS AND EVALUATION j -
PLANNING OF EXPERIMENTS PRETEST PREDICTIONS
{
l -
POST TEST EVALUATIONS f
j -
DEVELOPMENT OF PREDICTIVE METHODS INTERACTION WITH PEER REVIEW PANEL
! FIN A-1375 PENETRATION TESTS i
j -
PROCUREMENT OF SPECIMENS ,
j.
DETAILED DESIGN OF TEST FIXTURES AND INSTRUMENTATION i -
CONDUCT OF EXPERIMENTS COMPILATION OF TEST REPORTS .
l FIN A-1249 LARGE MODEL TESTS j -
DESIGN AND CONSTRUCTION OF MODELS i -
INSTALLATION OF INSTRUMENTATION EXPERIMENTAL PLAN AND SAFETY ANALYSES i -
CONDUCT OF EXPERIMENTS 1
COMPILATION OF TEST REPORTS l
! 46
o O O TITLE:
CONTAINMENT INTEGRITY UNDER EXTREME LOADS FIN; A 1401 CONTRACTOR:
SANDIA NATIONAL LABORATORY COOPERATING ORGANIZATIONS .
KERNFORSCHUNGSZENTRUM KARLSRUHE (FRG)
ELECTRIC POWER RESEARCH INSTITUTE (EPRI)
COMMISARIAT A L'ENERGIE ATOMIQUE (FRANCE)
NUCLEAR INSTALLATIONS INSPECTORATE (UK)
BUDGET:
FY 1985 FY 1986 FY 1987_
$740K $600K $700K e
O O O
~
ISSUE:
THE CONSEQUENCES OF A SEVERE ACCIDENT WOULD BE SIGNIFICANTLY ANALYTICAL PREDICTIONS OF BY Tile PERFORMANCE OF THE CONTAINMENT.
CONTAINMENT FAILURE MODES REQUIRE EXPERIMENTAL VALIDATION.
OBJECTIVE:
DEVELOP A METil0DOLOGY THAT CAN BE USED TO PREDICT Tile PERFORMAN CONTAINMENT BUILDINGS DURING SEVERE ACCIDENTS.
INTEGRATION:
l 0 LARGE MODEL TESTS PERFORMED UNDER FIN A1249 PROVIDE THE DATA FOR BENCHMARKING ANALYTICAL PREDICTIONS OF STRUCTURAL DEFORMATIONS.
O PENETRATION TESTS PERFORMED UNDER FIN A1375 PROVIDE DATA FOR ANALYTICAL PREDICTIONS 0F LEAKAGE.
t5 l
o O O SCOPE / ACCOMPLISHMENTS:
.; O PRIOR TO FY 1985 I
i -
PRE TEST PREDICTIONS AND POST TEST ANALYSES FOR S SERIES.
I O FY 1985
- POST TEST ANALYSIS OF LARGE STEEL MODEL TEST.
- PRE TEST PREDICTIONS FOR CONCRETE MODEL TEST.
l
! O FY 1986 -
r
- DEVELOP A BASIS FOR EXTENDING CONCRETE MODEL TEST INS
! PRESTRESSED CONTAINMENTS.
l
- DETERMINE EXPERIMENTS NECESSARY TO VERIFY PREDICTIONS O
- PERFORMANCE UNDER. SEISMIC LOADS.
l 0 FY 1987 i
- POST TEST ANALYSES FOR CONCRETE MODEL.
- PRE TEST PREDICTIONS FOR SEISMIC EXPERIMENTS.
1
- DEVELOPMENT OF A LEAKAGE METHODOLOGY BASED ON PENETRATIO
! EXPERIMENTS.
i 41 i ,
f . . _ _ . . - - - - - . . -
REGULATORY USE: ,
0 CONFIRMATORY ASSESSMENT OF SEVERE ACCIDENT POLICY STATEMENT.
O BASIS FOR POSSIBLE ADDITION OF CONTAINMENT PERFORMANCE REQU SAFETY GOAL IMPLEMENTATION, e
/
70
O O O
~
! TITLE:
l INTEGRITY OF CONTAINMENT PENETRATIONS UNDER SEVERE ACCIDENT CONDITIONS
' F_IL l
A 1375 l
CONTRACTOR:
1 SANDIA NATIONAL LABORATORY BUDGET:
- FY 1985 FY 1986 FY 1987
$900K $900K $600K j -
N i
i l
1
\ 7/
i --- - - --
e
. O o n H u)
O ZH
- Z W LLJ Z C ~
H EWW
- v Z C3 -
.-= - c3
>= 3d CD m
w Lt. J M3 y,,, O 'CC U C3 44 "C N u HUH CC "
at U)
E Z a:C M LD *:C O -J Z
>= O U LLJ LLJ J ~Z
- 3- cc st3 ~
HH .
Q LLJ H .E Q *-* ~ -
Z ar U. O Z id X LL >J U)D M "C Lt-C E
T*
vctd .= O -.J Id i-g H u. U2 =E C '- ~
1,L. cc cc t.D M LLJ Q a--
- Q. C3 C3 Z aC CE . E Z .
Z Z.-= O DD ,
J J og cc U)
(D. ag og DH Z.-= U) C3 . .
u) -U '> (d Q a:t LL- LLJ W ' --
-J E - CC E td H- ac J cc 14 cc CE u) Q. M Tr'.:
O go >J H et to O E a:C LL.
LLJ H ca Z x u-5 wxw OJat LLJ CC Zw Z Z LLJ o:
CL. sC C:2 DE - C3 LLJ m aC E OW J C3 Z Q. f O
- g cg Z Q% 3 M t.l
'.T L.L. LLJ LLJ U> l~~
H 1-- CL.
. td OO< Q. C- ZM ,;
H CX ZZX u) Ld J LLJ O id
-H t. e-
% t,LJ LLJ u,J O
-Hm LLJ m Id H p C.:.s. EZ id >Z -
y -. cc HDH C 14 C *C . l
% I.LJ E u) Q~ t-- Q r
y aC - **CQ ~ *CC H a:C H 3 CE E LLJ C WO ;
ZQ w l-- CL. W Q =% ZE f.
td O t.u H ZE l-- LLJ Q. LLJ \ ;
E U C2: U O LM CW CD ,
LLJ w ca ac v l- HZ CE E i M id $
>J LL Z t.LJ at O = u) Q. LLJ aC H -.- I u) H c3 2 J td C3 at t.L. E O v mO3cc O LLJ ZZ LL) '
E LLJ O g u u) l-- *C Z u) ~
Lt. LLJ O l-J LL l--
O LLJ C2:
vs 7HLM EZ C3 O
- C a:C cc O LLJ H u) Z.
LLJ at -*
J E ZF- C. Id t.D Z r
U E aC t.L. >
a-a N l.LJ E~
LLJ Z c e Lt. I u) O D *C H ZCH I--
U U) J Q t LLJ m t.L. O >-- ~ *C SM OEZ 1.LJ - U Q LLJ LL.
O >4
- =
WU3 tdC. (E d Z HcCLLJ u =* ~ l.LJ E E Z Z' CE u) CD Z EEu O Q. td LLJ O cc H LLJ UZa14 C LD l--
LLJ 14 LLJ CE v) >--
>-- l-- t.Lt Z LLJ 1- Z *C LLJ~Z Z ::> t-- et
- LLJ
-O 1d LM Z CC H J F-*
2 C2 Q. H
.. HC>Q U 6- -
U C
Id w' LLJ H m 7 Z O 4 u) CC ~l C M Oi
~j O
I
- - ~ ~ . . -
o
~
O O SCOPE / ACCOMPLISHMENTS:
0 PRIOR TO FY 1985
- SURVEY OF PENETRATION TYPES, GE0METRIES, AND MATERIALS TO IDENTIFY MOST LIKELY CANDIDATES FOR LEAKAGE TESTS.
- DESIGN OF TEST MATRIX FOR SEAL AND GASKET GE0METRIES AND MATERIALS.
O FY 1985
- INITIATION OF SEAL AND GASKET TESTS.
- PROCUREMENT OF A FULL SCALE AIR LOCK AND SELECTION OF A TEST' FACILITY.
- DESIGN DETAILS COMPLE.TED FOR AIR LOCK AND EQUIPMENT HATCHES CONCRETE MODEL.
O FY 1986 .
- COMPLETTION OF SEAL AND GASKET TESTS, COMPLETION OF AIR LOCK TEST.
PROCUREMENT OF BELLOWS AND DEVELOPEMENT.OF TEST PLAN. <
73
O O O
~
O FY 1987 EVALUATION OF AIR LOCK TEST.
PERFORMANCE OF BELLOWS TEST.
REGULATORY USE:
0 CONFIRMATORY ASSESSMENT OF SEVERE ACCIDENT POLICY STATEMENT.
O BASIS FOR POSSIBLE ADDITION OF CONTAINMENT PERFORMANCE REQURIEMENT FOR SAFETY G0AL IMPLEMENTATION.
e W
l
o O O TITLE:
N EXPERIMENTS ON CONTAINMENT MODELS UNDER EXTREME LOADING CONDITI6.
FIN; A 1249 CONTRACTOR:
SANDIA NATIONAL LABORATORY BUDGET: .
FY 1985 FY 1986 FY 1987
$2060 $2000K $2400K
/
s
/5
o
~
O O ISSUE:
THE CONSEQUENCES OF A SEVERE ACCIDENT WOULD BE SIGNIFICANTLY A ANALYTICAL PREDICTIONS OF BY THE PERFORMANCE OF THE CONTAINMENT.
CONTAINMENT FAILURE MODES REQUIRE EXPERIMENTAL VALIDATION.
OBJECTIVE:
PERFORM EXPERIMENTS ON CONTAINMENT MODELS THAT WILL PRODUCE FAILUR -
DATA UNDER EXTREME LOADING CONDITIONS AND PERMIT AN EVALUATION O CAPABILITY OF STATE-0F-THE-ART CALCULATIONAL METHODS TO PREDICT CONTAINMENT FAILURE UNDER SEVERE ACCIDENT CONDITIONS.
INTEGRATION:
0 THE MODEL EXPERIMENTS ARE COORDINATED WITH THE ANALYTICAL PREDI EFFORT PERFORMED UNDER FIN A1401.
O THELARGEDIAMETERPdNETRATIONSINCLUDEDINTHEMODELSPROVIDE DEFORMATION DATA USED IN T'HE PENETRATION EXPERIMENTS PERFORMED UND FIN A1375.
4 74
- .- - - - -- - . - _ - - - =_. _ -_. - . .-. - _ - _ _ __ .-- __
O O O I SCOPE / ACCOMPLISHMENTS:
O PRIOR TO FY 1985 l
DEVELOPMENT OF A REMOTE TESTING FACILITY FOR HIGH PRESSURE l EXPERIMENTS. .
i TESTS TO FAILURE OF A SERIES OF THREE SMALL STEEL MODELS TO VERIFY THE ABILITY OF LARGE DEFORMATION COMPUTER CODES TO MODEL STIFFENER BEHAVIOR AND BEHAVIOR AT PENETRATION DISCONTIVITIES. ,
! - DESIGN, CONSTRUCTION, AND INSTRUMENTATION OF A LARGE STEEL MODEL.
, 0 FY 1985 i .
COMPLETED TESTING TO FAILURE OF A LARGE STEEL MODEL.
COMPLETED DESIGN AND STARTED CONSTRUCTION OF CONCRETE MODEL.
l i
l 0 FY 1986 i .
COMPLETE CONSTRUCTION AND INSTRUMENTATION OF CONCRETE MODEL AND l
i INITIATE TESTING.
i 0 FY 1987 l -
COMPLETE TESTING TO FAILURE OF CONCRETE MODEL PERFORM ANY NECESSARY SUPPLEMENTARY TESTS INITIATE EXPERIMENTS ON SEISMIC PERFORMANCE
\ 77 i
O O O
- REGULATORY USE
j 0 CONFIRNATORY ASSESSMENT OF SEVERE ACCIDENT POLICY STATEMENT,
) 0 BASIS FOR POSSIBLE ADDITION OF CONTAINMENT PERFORMANCE REQURIEMENT FOR I
SAFETY G0AL IMPLEMENTATION, k
}
i 5
3 .
l 1
i 1
i i
k l
4 4
5a . / /c
! O O O i
1
[ POST-TEST AXALYSIS OF FAIIERE
! The failure mechanism and capacity of the model were not '
' aredicted correctly. Metallurgical investigations showed ihat the failum was ductile, and no degradation or defects l
j in the materials or welds were detected.
A crack in the formed stiffener around EH1 was observed
! during analysisthe wastest (first reported conducted to investigate: at 165 p(sig). A post-test 1) wha1 initial crack in the formed stiffener, and (2) how a crack in the stiffener would affect the response of the contain-l ment wall.
The model was similar to that used for the pretest analyses, except for an " eccentricity" at the junctures between the formed stiffeners and cimumfemntial stiffeners.
i i
O O O 1
POST-TEST ANALYSIS - RESULTS At 165 psig, the max. principal strain was 16% at the outer
! fibers of the formed' stiffener where it intersected the l cimumfemntial stiffener. This is consistent with the observed crack. With the intmduction of the eccentricity, l equilibrium at the junctum could not be satisfied with l
membrane fomes only, thereby resulting in bending moments l and high shear.
1 i A strain concentration occurred in the cylinder near the i stiffenefs juncture, with the maximum principal strain j (primarily hoop) exceeding 14% at 195 psig. Again, the calculated strain in conjunction with the ruptum criterion is consistent with the experiment. This strain
, concentration is a direct result of the redistribution of j load fmm the stiffener (after it has cracked) into the i shell wall.
l j;
i
---,_._________________.___________a - -__
O /
x "
6
- RENEFFITS
)YRTEMOEG LAUTCA(
j TSET-TSOP i
"50 I
RENEFFITS NEMECROFNIER RENEFFITS HE
) DEMROF -
+
O TSET-ERP O .) ~
>:: 6
lll s
o S
- - - - - - - - I 7 '
ER RE ' 4%
UN 1 N
A NFI R
UT ' ' 2T
_ JS 1
S EH ' L A
HT TI ' '
0 PI R W 1
C AR ' N I
EE T R
NN Ss Ee P
' S R E Tsy
_ O E DA NS F
F T
TL SA ON X
A M
__ I Pa ' S L E YD C
CE '
A M / F NR I
e, 4R #
O U NF I '
S AE TS SI T RH E S ' 2X T
S T TYL '
E F EA O RN '
PA -
- ,, J - - -
O 1
0 0 o . 0 o o 0 0 0 O 0 8 s .
2 o s 8 4 2 2 1 i 1 i
. ." aEm* aR n
?
O 11lll(l l'
y y- r +
O O .
i O
IPROGRAM ACTIVITIES j SURVEY OF MECHANICAL PENETRATIONS (ARGONNE NATIONAL LABORATORY)
PRELIMINARY ANALYSIS (ANL, SANDIA) .
TESTS ON PENETRATIONS (SANDIA)
TESTS ON SEALS (SANDIA, IDAHO NATIONAL ENGINEERING LABORATORY)
~
NOTE - lNVESTIGATION OF ELECTRICAL PENETRATION l
ASSEMBLIES IS BEING CONDUCTED IN ANOTHER PROGRAM
o o
~
O
.I PROGRAM STEPS .
. SURVEY OF PENETRATION TYPES AND DESIGNS PERFORMED BY ANL ON 48 U.S. PLANTS .
. QUALITATIVE FIGURES OF MERIT DEVELOPED
. PRELIMINARY ANALYSES CONDUCTED
. PROGRAM PLAN DEVELOPED
. CONDUCT PROGRAM .
l
i j O o O EXPERIMENTS /AN ALYSIS OF PENETR ATIONS i
l, PERSONNEL AIR LOCK 1986 (FULL SIZE)
EXPANSION BELLOWS 1986-37 DRYWELL HEAD FOR BWR ANALYSES DO NOT INDICATE I MK l OR MK ll NEED FOR EXPERIMENT
l c
O O O s
l PRFSSERE SEATI3G EQUIPMENT HKUTFS i ~
Models for pmdicting the deformation and leakage for .
pmssure seating hatches in a steel containment wem verified by the 1:8-scale steel model test Significant interaction between the cylinder (shell) and the penetration sleeve occurs.
't; Average membrane strains from 2.3% up to 5% must arise '
before the sleeve deformation is sufficient to cause a mismatch with the cover that results in leakage.
Two additional pressum seating equipment hatches will be
- i tested in the 1
- 6-scale minforced concrete model.
l! '
1 i
~~
O O d l
DRYWELL HEAD ANALYSES CPWG cites unseating of drywell head due to internal amssurization as a major contributor to leakage in Mk I and Mk II containments in a severe accident. However, the conclusion is based on analysis that did not take into account the effect of elevated temperatures, which are also
! pmsent in the accident.
I I
I ANL and Sandia have conducted independent analyses that -
! show that if elevated . temperatures inside the containment am considered in the analysis, metal to metal contact will be maintained between the sealing surfaces, thereby pmeluding leakage.
i Differential thermal expansion between the drywell head flanges, which house the seal, and the bolts, which attach l
the drywell head to the lower part of the containment, has i the effect of additional pmload on the bolts, themby '
substantially increasing the pressure needed to unseat the
! drywell head.
l
O O O i
l DRYWEIL HEAD ANALYSES ANL used a sophisticated finite element ap muach, whereas Sandia derived closed form solutions basec. on a strength of materials approach.
Sandia analyzed the response of drywell heads in eight
' containments. There were significant differences in the drywell head diameter, flange thickness, number of bolts, bolt diameter, and bolt preload within this group. Only one ,
of the eight unseated at the maximum accident pressure (120 l
psig), and the gap was less than 2 mils, which would most l
likely be closed by the seal material There are significant conservatisms in Sandia's analyses, and the actual separation pressure is probably higher. .
3 l
Based on the analyses, a test of a drywell head was not felt '
to be necessary.
i l
I
O o d PRESSERE ESSEATISG EQUIPEST HATCHFS t
A pressum imseating equipment hatch will be tested in the 12-scale reinfomed conemte model The same methodology used to analyze drywell heads will be used to analyze the pmssure unseating hatch.
Data fmm the test (gap measumments, compression in the flanges) will be used to assess the method. .
i O O O i
I PERS0 DEL' AIRLOCKS
'Ibst will be conducted in FY 86 Preliminary analyses suggest that- (1) no significant interaction occurs between the shell and sleeve, and (2) the i door and bulkhead will undergo large deformations associated i
with bending.
Finite element analyses will be performed with the actual ,
material properties, and a leakage criterion will be proposed y , r to the test.
i f
i I
i
! O O O 1
l l
! TESTS ON SEALS i
l -
1 SILICONE RUBBER f MATERIALS -
EPM /EPDM VITON (FKM)
NEOPRENE cross-SECTION "O" RING GEOMETRY - DOG EAR l GUM DROP ,
TONGUE-AND-GROOVE
)
f 1 - - - - - - - - _ - _ _ ________ ____ __ __ _ __ _
O O O TESTS ON SEALS
.i ENVIRONMENT - SEALS AGED AND UNAGED l
! AGING (THERMAL AND RADIATION EXPOSURE BEFORE TEST) e PRESSURE - UP TO 160 PSIA (1100 KPA) ,
TEMPERATURE - UP TO 700 F (370 C) i -
o
! O O j
i i
1 l
l TESTS ON SEALS TEST SET-UP - PRESSURE CHAMBER (SATURATED OR SUPERHEATED STEAM, HEATED AIR, OR NITROGEN)
MATING SURFACES - METAL-TO-METAL CONTACT PRESCRIBED GAP ROTATION OF ONE SURFACE (MACHINED INCLINE) 9 I
I
I O O d i
i l
EX P E RIM E N TS/A N A LYS ES .
OF PEN ETR ATIONS i SHELL-STRUCTURE INTERACTION TESTS ON 1:32 SCALE EQUIPMENT HATCH STEEL,1:8 SCALE STEEL, &
PERSONNEL AIR LOCK 1:6 SCALE CONCRETE l -
MODELS EQUIPMENT HATCH WITH SEAL 1:8 SCALE STEEL MODEL EQUIPMENT HATCHES 1:6 SCALE REINFORCED 2 PRESSURE SEATING CONCRETE MODEL,1987 1 PRESSURE UNSEATING i
O -
O O FUTURE W.O R K
. COMPLETE TESTS ON SEALS
. INVESTIGATE $FFECT OF AEROSOLS ON LEAKAGE (PLUGGING)
. TEST PERSONNEL AIR LOCK (THERMAL AND STRUCTURAL)
. EQUIPMENT HATCH TEST IN REINFORCED CONCRETE MODEL
I O
' # ' T
- ( T3,RW "". )J n
'"L 4W%
'y . , *
\
1 I
f1 , j ' "
i 9 - N f ?, , .
s .- .
Ei !. ._
I \ /
l EI$ F%e "a d (s eA: p "O sk o
3 ATTACHING LUGS ~--- -
- . O' RING GROOVES &
l '
'O' RINGS ENVIRONMENTAL '
- O' RING TEST FIXTURE CHAMBER PLATES
\ N _. _ ,._
, /
O SEALS AND GASKET TEST ASSEMBLY
4r 4
+ ,
- k er x x e as e e e -x
,' + * -
o *
- Nt :: +---- -g r 4 0 -
S
['i e l e
\j k 4 x x x ,
d' - ,
~
v Test "O" Rings i
Bolts for Pre Loadshwsawl nmw 4
. I m mxw '
M+.en6-l YAMM i 7/M&;Mi
% A w a kN% % w E9e h % h h N h M W Pressure ~
Monitoring Ports' sectio = a a O FIG. 6.0: FLAT 1GE ASSEMBLY FOR DOUBLE "O" RINGS TEST
, y j, i
gg, h d $ Ti -
~g * $
e Wh@
Z u
. ~Z d.
c e. .
J g d 1
/
g
-mmr -
)ac: , C 1
q w a
! G 2 k -- / s v3 y
,r,;/
s '
r ;d '
o g -
e Lt s.
_WW 1_ i T, 1 w" .
j __..
M.i
/i q h -: >
a -
i g
% . h_ 4 i
t p 1 G i g -H LiL n. .
- a
, g
" f ~
j
-j u 3-
-ay g
.d t u? ~
2(56%
I l c k.4 '7 .
y 1
-; o , )- - >@ m i [%) 1 &m:
Y 1 m t -
w .I L l
{2
! g
!- g!
F A
'[Om\
d g
j i
u Wq l
. I . :I F- m
'3 l ,
s gGv os;
,)I .
(
i
) '
! i
- 5 .i.
i
'n i i . -
l .
l
- e e .e _
1
. A/ .i
- L i. D L [
/c/ ,/Q
( l l
I l
t i
I DUQUESNE LIGHT COMPANY
. BEAVER VALLEY POWER STATION -
UNIT 2 !
( i i
i i
+
{
t 1
l 4
)
l me f I
l 1
e
(-
MHIEEEI EBOGBAM AN ALTERNATIVE ENGINEERING METHOD FOR PIPE BREAK PROTECTION PERFORM TEST & ANALYSIS TO SHOW LBB APPROACH IS APPROPRIATE FOR B.O_P. SYSTEMS IMPLEMENT THE LBB APPROACH o DEMONSTRATE LEAKS OCCUR FIRST o DEVELOP LEAK DETECTION PROGRAM RETAIN RUPTURE RESTRAINTS IN AREAS LBB IS NOT APPLICABLE OR ECONOMICAL RETAIN DEGB ASSUMPTION FOR ECCS, CONTAINMENT DESIGN, ETC. I l
k.
O (
PROGRAM o SCOPE l
o ACTIVITIES j STRESS CORROSION REVIEW WATER HAMMER ANALYSES FATIGUE EVALUATION EQUIPMENT SUPPORT EVALUATION LBB ANALYSES LEAK DETECTION PROGRAM COST BENEFIT ANALYSIS l
o
SUMMARY
AND CONCLUSIONS I
.m e
o e
O SCOPE
(
RUPTURE MITIGATION HARDWARE I
i ARBITRARY BALANCE O Fr l INTERMEDIATE PLANT l BREAKS (WHIPJET) -
127 136 PRIMARY SYSTEM 9
i THE NUMBERS REFrER TO BVPS-2 O RESTRAINTS AND JET SHIELDS
(
ACTIVITIES DEMONSTRATE APPLICABILITY OF LBB o STRESS CORROSION REVIEW o WATER HAMMER ANALYSES '
o FATIGUE EVALUATION o EQUIPMENT SUPPORT EVALUATION 1
1 IMPLEMENTATION OF LBB PROGRAM l
o LBB ANALYSES o LEAK DETECTION PROGRAM o COST BENEFIT ANALYSIS t
i O
9 o .
1 l
I l
(
l 1
l l
l l
1 STRESS CORROSION REVIEW i
l o CONDUCT A SPECIFIC STRESS CORROSION REVIEW ON A SYSTEM-BY-SYSTEM BASIS o ASSESS COMBINATION OF STRESS LEVELS, MATERIAL, AND CORROSIVE FLUID ENVIRONMENT FOR EACH CLASS l OF PIPING l o ASSESS THE POTENTIAL FOR STRESS CORROSION CRACKING AND FORMULATE A CONCLUSION FOR B.O.P. SYSTEMS l
( '
O
(
WATER HAMMER ANALYSES o REVIEW TRANSIENTS CONSIDERED APPLICABLE TO BVPS-2 WATER HAMMER ANALYSES o ASSESS THE POTENTIAL FOR HARMFUL WATER HAMMER ON A SYSTEM-BY-SYSTEN BASIS o PROVIDE TECHNICAL JUSTIFICATION TO SHOW WATER HAMMER IS ADEQUATELY CONTROLLED l
l l
o DOCUMENT TREATMENT OF FLOW l TRANSIENTS FOUND APPLICABLE TO BVPS-2
, l I
1 I
O (
l f=ATIGUE EVALUATION o THERMAL FATIGUE EFi=ECTS PREDICTED BY USING=
NUMBER OF THERMAL CYCLES w (s ALLOWABLE STRESS LEVELS TO
-OBTAIN CUMULATIVE USAGE FACTORS CCUF>
c P E RFr O R M MORE f=ATIGUE EVALUATION Fr O R LINES WITH CUF GREATER THAN O1
d O c i
l i
EQUIPMENT SUPPORT EVALUATION o ESTABLISH CRITERIA TO EXAMINE EQUIPMENT SUPPORTS AFFECTING ,
B.O.P. PIPING O (
o SUPPORT FAILURE POTENTIAL MUST BE EXAMINED TO ENSURE PLANT SAFETY IS NOT COMPROMISED
h' O
O (
LDD -
ANALYSES O ,
ATERIAL PIPE FRACTURE PIPE CRACK TESTING ANALYSIS ANALYSIS l
f
.)
'n,
I 4
,\ -
i t
}
t i i i
l LBB ANALYSES l
t }
i !
1 ,
i i
i l
MATERIAL l
- TESTING I l
BVPS-2 MATL i
i i
STRESS-STRAIN ?
J-R CURVES
- I f
- PLANT j SPECIFIC
, i P
( h 5
- i
- i I
i I
I n j I
- - - - -- r -
w - @b m --- '=w "w'tr v - N' - - - -
4 4
LBB ANALYSES l
1
.I PIPE FRACTURE ANALYSIS e
I
- CONSERVATIVE i
CIRCUMFERENTIAL
- , CRACKS I
- CRITICAL CRACK l LENGTH
- LEAK RATES
\
l l
O I (
Y U >
UW
(" ( F 0 W EW [ lt U> g q ,
J g gz yI W( q s jp aZ ga m3 .
"4 F ([ (0 1 I O( F[
0 JE to 0 l
l 1 l l l
l i
G W
W m>
mJ O JE Z
( I l
O
S 9
i j Les
! ANALYBCS i
a MATERIAL eIeE FRACTURE eire. CRACK
{
TESTING ANALYSIS ANALYSIS i
O i
1 i
- - sves-2 mATL - CCNsERvATIve -
o-T AreRoACs J
- STRESS-STRAIN - CIRCUMFERENTIAL - LOAD SAFETY l CRACKS MARGIN J-R CURVES CRITICAL CRACK - STABLE f - eLANT LENGTH GROWTH
} SPECIFIC l -
LEAK RATES i .
, l 1
)
J J
l l
R. R 1
4 l
'a O <
LEAK DETECTION o PROVIDE CAPABILITV TO DETERMINE, LOCATE, AND QUANTIFY PIPE CRACK LEAKAGE o OPTIMUM LEAK DETECTION TECHNIQUE IS A COMBINATION OF MORE THAN ONE METHOD (
o EXAMINE LEAK DETECTION RATES OF EXISTING AND/OR ADDITIONAL INSTRUMENTATION o ESTABLISH OVERALL LEAK DETECTION STRATEGY O -
1
O (
l 1aO f
Q %. - ,#
~
. :w=@#.
M$j j b eb , ,
I
"%;f i t 4
Dae I s ,
I vg
_@ %r,es
~ g_,
Q
O (
BENEfrITS o IMPROVE ACCESSIBILITV TO MINIMIZE RESTRICTIONS TO ISI AND REDUCE PERSONNEL RADIATION EXPOSURE o TIMELY AWARENESS OF LEAKAGE Oo MINIMIZE UNANTICIPATED THERMAL EXPANSION STRESS DUE TO INTERFERENCE o IMPROVE UNDERSTANDING OF PIPING FAILURE MODES o REDUCE PLANT COST
QUANTITATIVE COST SAVINGS (NET) i HAZARDS ANALYSIS $1,420,000 DESIGN -
330,000 hABRICATION 2,980,000 i l
INSTALLATION 1,890,000 INDIRECT COSTS 1,400,000 TOTAL QUANTIFIABLE COST SAVINGS (NET) $8,O20,000 l
l nv
O (
QUALITATIVE COST SAVINGS o INCREASED FLEXIBILITY IN THE CONSTRUCTION SEQUENCE o ELIMINATION OF GAP ADJUSTMENT VERIFICATIONS DURING START-UP AND POWER ASCENSION WALKDOWNS I
o ELIMINATION OF STRUCTURAL CHANGES WHICH ARE DUE TO PIPE 4 RUPTURE RESTRAINT LOADS o REDUCTION OF ENGINEERING AND CONSTRUCTION COSTS WHICH ARE DUE TO DESIGN AND CONSTRUCTION IN CONGESTED AREAS o DECREASED OUTAGE COSTS DUE TO ISI DIFFICULTIES AND CONGESTION o REDUCTION IN HEAT LOSS FROM PIPING TO ADJACENT PIPE RUPTURE RESTRAINTS. l t
.--,..--,...---,_,.__.,.--,__.,,.---,,,,n.-,.,---a.,---._,-__,
, . n--,,,n._,,-.,. . . , _ , , , , , , ,.,.-n.,--.-., . , , , . -
a O (
l COST BENEFIT DIRECT COST SAVINGS IN DESIGN, ANALYSIS, FABRICATION, AND s INSTALLATION BM O ,
INDIRECT COST SAVINGS DUE TO BETTER SCHEDULE AND AN IMPROVED PHYSICAL PLANT 4M TO BM OTAL COST SAVINGS 12M TO 16M e
O
6 O <
SUMMARY
AND CONCLUSIONS o PROVIDE PIPE BREAK PROTECTION USING STATE-OF-THE-ART ENGINEERING TECHNOLOGY o USE THE ABILITY TO SUCCESSFULLY DETECT PIPE LEAKS ;
1 o IMPROVE PLANT SAFETY AND DESIGN DURING OPERATIONAL PERIODS AND OUTAGES o SHOWS REASONABLE, AGGRESSIVE INDUSTRY INITIATIVE TO IMPROVE SAFETY AND LOWER COSTS
. 1 l
l SCHEDULE I
NRC GRANTS SCHEDULAR E X Elv1P T I O N FOR B.,O P Wi-1I P RESTRAINTS SEPT- 1985 DLC COMPLETES WHIPJET PROGRAM DEC 198G INSTALL WHIA RESTRAINTS, I fr ANY REfrWELING ARE REQUIRED OUTAGES O
'a