ML20150F749
| ML20150F749 | |
| Person / Time | |
|---|---|
| Issue date: | 02/26/1988 |
| From: | NRC COMMISSION (OCM) |
| To: | |
| References | |
| NUDOCS 8804050396 | |
| Download: ML20150F749 (65) | |
Text
_ _ -_. -._._ - -- __ --...._.--_.. _-. - --- _._.....-.._.. - -.--......
i r
l l
UNITED STATES g
NUCLEAR REGULATORY COMMISSION i
In the Matter of:
l BWR MARK I CONTAINMENT PERFORMANCE INFORMATION EXCHANGE NORKSHOP
)
l
\\
CORRECTED COPY I
l Pages:
466 through 516 Place:
Baltimore, Maryland J
Date:
February 26, 1989
.-..u===------..---====---------.--==----------=.---...
I i
i i
l HERITAGE REPORTING CORPORATION osse w a m on m 1224 L Stnet N.W., Suke 640 WasMngton, D.C. 20005 8804050396 000PP (242) 62M PDR TOPRP EMt 46 C
, _.. - ~ _ _ _ _ _ _ _
466 NUCLSAR REGULATORY COMMISSION O1 2
)
3 In the Matter of
)
)
4 BWR MARK I CONTAINMENT
)
PERFORMANCE WORKSHOP
)
5
)
6 12th Floor Conference Room The Belvedere Hotel 7
1 Chase Street Baltimore, Maryland 8
- Friday, 9
February 26, 1988 10 The above-entitled matter came on for hearing, 11 pursuent to notice, at 8:30 a.m.
12 APPEARANCES:
13 On behalf of the Nuclear Regulatory Commission:
14 DR. THEMIS SPEIS 15 DR. WAYNE HOUSTON 16 DR. BRIAN SHERON 17 MR. JERRY HULMAN, Moderator l
18 19 20 21 22 23 S24 (G
25 CONTENTS Heritage Ro Wrting Corporation (202) 628-4838
467 91 CONTENTS STATEMENT OF:
PAGE:
2 TREVOR PRATT, BNL 468 3
GREGORY KRIEGER, PECO 470 4
HERSHEL SPECTOR, NYPA 475 5
6 7
8 9
.10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 lieritage Re[2 )70rting Corporation (20 628-4888
468 1
PROCEEDI NGS 2
DR. SPEIS:
Good morning, ladies and gentlemen.
I think we're ready to start.
3 4
It's going to be a brief morning.
We have one more 5
session.
Hopefully, we'll be over in an hour. After that, we 6
plan to present a kind of a short overview or summary and maybe 7
a few people on the audience might want to say something.
8 I don't think there is a need to have the so-called 9
Issue Focusing Discussions.
We had enough of that the last two 10 days.
So, the plan is that we might be able to get out of here 11 somewhere between 10:30 and 11.
So, I hope that is good news.
12 The first -- this last session deals with Potential 13 Automatic Depressurization Improvements.
The first speaker is-14 Trevor Pratt from BNL.
15 Trevor.
16 STATEMENT OF TREVOR PRATT, BNL 17 MR. PRATT:
Okay.
We'll again follow the same format 18 in terms of these discussions that we've had in the previous 19 presentations.
20 Firstly, I'd like to try to indicate the problem 21 description and then indicate some of the potential 22 improvements that are being suggested, 23 In this case, core melt down with the primary system
,a (v')
24 at high pressure has been identified as an area of significant 25 uncertainty.
In particular, it could result in the potential lieritage Rewrting Corporation (202) 628-4888
469 1
for early containment failure.
2 It's been suggested that an improved automatic 3
depressurization can potentially eliminate the potential for 4
high pressure core melt down.
5 Some of the suggested improvements are on this 6
viewgraph.
The feeling is that improved air and nitrogen 7
supply has been addressed in the TMI action plan, that long-8 term loss of AC power batteries may be depleted rendering the 9
ADS system inoperable. So, it's been suggested that alternative 10 dedicated back-up power supplies, batteries, a small generator-11 driven power supply with appropriate connections, could 12 increase the availability of the ADS system.
13 In addition, it's also been suggested that the 14 cabling for the ADS may need to be insulated or shielded to 15 assure operation during the long-term station black-out 16 sequence.
17 These are the questions that we would like to address 18 this morning, if possible.
Are the types of improvements that 19 have been described sufficient to improve the reliability of 20 the ADS system?
What other recommendations would you suggest?
21 What are the benefits associated with this improved 22 availability?
And, in fact, will it reduce the risk from, for 23 example, direct containment heating?
p()
24 What is the potential adverse interactions between 25 any of the proposed improvements and the existing safety Heritage Hemrting Corporation (202) 628-4888 1
470 1
systems?
Finally, are there any other improvements that also 2
may be feasible for the Mark I Containments to allow them to 3
depressurize under these circumstances?
4 I think that's really that all I had to say.
5 Perhaps I could ask for the next speaker to come up 6
then.
7 STATEMENT OF GREGORY KRIEGER, PECO 8
MR. KRIEGER:
My name is Greg Krieger from 9
Philadelphia Electric, and I'd like to talk about accident 10 sequence definitions and their impact on some of the 11 phenomenological issues we've been looking at.
12 In relating some of the new information coming out of 9
13 NUREG 1150, I just want to touch on TBUX, which is the short-14 term station black-out, and TB, which is the long-term station 15 black-out.
16 TBUX is defined as the loss of all off-site power and 17 simultaneous loss of all station batteries.
This causes 18 failure of all injection systems and failure of diesels, 19 failure to flash the field on the diesels.
20 This was calculated as everyone is aware as a 21 dominant sequence in the draft PRA front-end for 1150.
It is 22 currently used as we have seen in a number of phenomenological 23 calculations, and a re-analysis of 4550 indicates, as I q
(
)
24 mentioned the other day, even more than an order of magnitude 25 de rease in-probability of around a factor of fifty.
Heritage Re p rting Corporation (202) 628-4888 I
v,
471 1
Just to touch on a few of the reasons for this 2
decrease, the use of additional plant in f o rma tion, such as 3
maintenance records at Peach Bottom, are testing methods.
More 4
detail was received on the actual configuration and how the 5
batteries relate to each other.
6 Items that were not taken into account in the draft, 7
the first draft, were Conowingo Dam, which is a hydro-electric 8
facility, approximately seven miles downstream from Peach 9
Bottom, that has black start capability for our grid, grid 10 specific initiatore instead of generic, and our loss of grid 11 procedures in the drills we partake in in restoring power.
12 Next slide, please.
Jast to touch on common mode f^,t t
s_J 13 battery failure for a minute, in the draft, the probability of 14 a second battery failing, given the first, was given a 15 probacility of four percent.
The remaining six batteries were 16 assumed to fail with a hundred percent probability, and just a 17 simple calculation shows what the probability of failing any 18 five batteries are.
19 The re-analysis adjusts some of those comnon mode 20 factors on the bottom, and as you can see, there's a factor of 21 three difference just from looking at that.
A factor of five 22 or six is due to the specific five batteries we must feil to 23 fail the diesele and our ADS system.
(
)
24 The battery configuration at Peach Bottom, just to go a
25 through it a little bit, Unit 2 has four batteries, Unit 3 has Heritage Rep rting Corporation (202) 628-4888
472 1
four batteries.
Two of the batteries in Unit 2 flash fields 2
for the diesels and two on Unit 3 flash fields for the diesels.
3 So, there is an intimate interaction between Units 2 4
and 3, although the PRA was specifically done.for Unit 2, and 5
how does tnis impact our perception of ADS reliability.
One, I 6
should mention that ADS, the way I use ADS is in a collective 7
term.
8 ADS is not only the automatic system that is used, 9
but it.is also the manual depressurization on the vessel also, 10 since the EPGs do indicate in many instances for the operator 11 to manually depressurize.
ADS is qualified for design basis 12 conditions. This means a hundred percent humidity and
(
\\
'J 13 appro'ximately 340 degrees F.
14 As I mentioned before, manual depressurization is 15 required by the procedures, and it all comes down to the 16 reliability of ADS is really questioned because of some 17 unrealistic battery assumptions that have gone into draft 1150.
18 To turn to long-term station black-out, this is the 19 loss of all AC, off-site and on-site.
We do have long-term 20 injection and ADS available for an assumed six hours. Re-21 analysis indJ. cates for 1150 that this is the dominant sequence.
22 The six-hour battery life is assumed, and it does 23 have an impact. Realistically, plant operations would run HIPC
,3 5
)
24 and RCIC in a sequential manner.
Also, ADS could be available 25 even though HIPC and RCIC may not.
lieritage Reporting Corporation (202) 628-4888
473 1
HIPC and RCIC are 250 volt DC systems.
ADS is a 125 2
volt DC system. Degradation of certain batteries may lose --
3 you may Jose control of HIPC and RCIC but still have ADS.
4 What impact does this have?
Just in looking at some 5
phenomenological issues on timing, the water level, of course, 6
is a big difference.
The six-hour battery life gains 7
approximately ten and a half hours before the uncovery of the 8
core.
9 As you can see, relocation of the core debris vessel 10 bottom head failure are a little longer. These numbers are 11 coming from an Oak Ridge analysis, from the 1150 expert review 12 process.
{\\ >}
13 I should mention that the dry well pressure, looking 14 at the dry well pressure, at vessel failure, the long-term 15 station black-out is quite high.
This would indicate upon 16 vessel failure, there is a possibility that the containment 17 would fail on over-pressure, which is a totally different 18 sequence than failing with dry well liner and burn through.
19 Just to outline some of the major differences, the 20 manual depressurization gains a significant time to core 21 recovery. I think we have heard that yesterday from Don Kneckt.
22 It reduces the temperature inside the vessel and, therefore, if 23 you would assume to lose ADS and repressurize, there is much
()
24 more time for core recovery.
25 The suppression pool is saturated in a long-term fleritage Reporting Corporation (202) 628-4888
P
. t 2
t 474 j
1 station black-out sequence, where it is not in the short-term.
2 This has an impact on secondary containment.
There is possible i
3 steam inerting in secondary containment instead of possible j
f 4
hydrogen deflagrations.
l 5
The containment failure as analyzed in the CB&I study j
9 i
6 on over-pressure most likely being the Torus air space, fission i
l 7
product scrubbing would then be available, and just in general, 8
because of timing of the sequences, there's much longer time 1
i 9
for debris to reach the dry well shell.
A 10 So, just in conclusion, some of the perceived I
11 improvements from a number of the items we have seen in the 3
1 12 past few days are based on a limited set of sequences, accident f
i 13 sequence, maybe just one accident sequence.
They.are very low 14 probability non-dominant events, and we have seen that one 4
.i 15 assumption on common mode battery failure has had a lot of t
l 16 impact on analyses we have seen in the past few days.
t 17 These assumptions, estimations and the front-end j
18 analysis have to be looked at more carefully before we even go 19 into the phenomenological issues.
20 I shor' mention that a lot of this, as has been said
)
j 21 in the past few days, is plant specific.
Other plants may not
}
j 22 be similar to Peach Bottom, and just in closing, I wou'ld like l
23 to offer Philadelphia Electric assistance to any rational 1
{
24 laboratory doing studies on Peach Bottom.
l 25 We like to see the correct information go out and l
3 j
Heritage Reprting Corporation (202) 628-4888 i
475 1
used in the right fashion, such that correct -- the correct 2
benefits and assumptions are realized.
3 DR. SPEIS:
Thank you.
4 Our next speaker is Hershel Spector from NYPA.
I see 5
you're always teamed with Peter Bioniarz.
6 MR. SPECTOR:
Good morning.
7 Can you hear me?
8 Actually, we covered this material yesterday as part 9
of the over-pressure /over-temperature session, but Jerry Hulman 10 asked me to repeat some of the highlights.
11 This is the analysis that we did where we assumed the 12 ADS was operative in the station black-out, and we did it for 9
13 two cases, one where the batteries were lost right at T equal 14 zero, and the other one where the batteries were lost after six 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />.
16 And, in general, what we want to remind you of this 17 morning is that what we found was that there was significant 18 cooling of the inner surfaces in the reactor vessel when you 19 continued to operate the ADS, and this, in turn, promoted the 20 trapping of radionuclides.
21 The second thing that we learned is that since you 22 had transferred a certain amount of energy from inside the 23 reactor vessel system over to the pool, there was less energy
\\)
24 available at the time of vessel failure and that, in turn, had 25 a lot of positive effects, one of which, now that you had a. low Heritage Hemrting Corporation (202) 628-4888
i l
476 1
pressure melt rather than a high pressure melt, it reduced the i
1 42 probability of having a concern about direct heating in the 3
i I
3 containment.
)
i 4
It also -- we had calculated earlier that, in fact, 5
it exceeded the assumed failure pressure when you didn't have 6
ADS.
We got a little higher value than was just reported, l
7 whereas in the case where you did transfer energy via the ADS,
~
l 8
when the vessel finally did fall, they didn't have sufficient 9
pressure in the containment to reach the failure point and that 10 was reached considerably later.
l 11 And, lastly, we found that there was a significant t
12 reduction in the source term, roughly a factor of ten, below P
13 the numbers which we had mentioned earlier, which, in 14 themselves, are already relatively small in the one or two or t
15 three percent range.
f 16 So, you come up often with source terms in the 17 fractions of a percent.
So, there appears to be some interest 18
- here, j
19 I want to add a few cautions, though, this morning.
i 20 I've come to understand that running the ADS under these 21 circumstances may require a little closer scrutiny. I think I
22 there have been two different analyses.
Ours, which assumed l
23 that ADS was put into operation when the water is at the top of 24 the active fuel, and another analysis, which we didn't perform, l
25 but I understand to be rather similar, but the initiation point l
5 Heritage Rep rting Corporation (202) 628-4888 l
L
477 1
was when the water was perhaps closer to the bottom of the 2
fuel, maybe sixty inches or so off the bottom.
3 My limited understanding of these two differences-is 4
that it may be significant differences in the hydrogen 5
production where you -- at what water level you initiate ADS.
6 You may get considerably less hydrogen if you initiate ADS at 7
the top of active fuel than later on.
It's something worth 8
looking into.
9 On the other hand, I've compared cases where the --
j 10 ob.viously when you initiate the ADS, you are transferring mass 11 over to the pool and, therefore, the time between removing the 12 water from the top of active fuel to fuel melt may actually be 13 shortened.
This ia, in a senso, reducing somewhat your ability 14 to prevent a core damage if, by some manner, you were able to 15 recover coolability during the time you were boiling off the 16
- water, 17 Obviously, if you're transferring mass, you're going 3
18 to get to the point of core melt a little bit quicker once l
19 electric power is lost.
So, there are some things that one has l
20 to be a bit cautiouc about and they're not fully investigated.
21 These things, some of tnem, have just more or less surfaced 22 during the hallway conversations, and I thought I would pass t
23 them along to the group here.
h 24 Finally, I'd like to add that it wouldn't surprise me n
25 that as people begin to look at their irdividual plants, that Heritage Reporting Corporation (202) 628-4888
~.
4 478 1
this is only one of several schemes that become apparent.
2 Perhaps there are other ways, if proven to be useful, that you 3
can accomplish the same thing, namely reconnecting the reactor 4
to the pool or changing high pressure sequences into low 5
pressure sequences.
6 So, I would not wich to say this is the be all and 7
end all fix for everything, but, rather, that this should be an 8
invitation for people to look at this in the general manner.
9 You may do a lot better when you look at your own plants.
10 Thank ;ou.
l 11 DR. SPEIS:
I guess you were the last speaker.
12 Let's spend a few minutes or as many as necessary to O
13 see if anyone has any questions on the last three i
14 presentations.
15 Any questions?
16 (No response.)
17 DR. SPEIS:
We just saved another half an hour.
18 Thank you, Hershel.
19 Let's see.
I understand that copies of all the 20 presentations will be available at around 10:30, is that right?
21 10:30 a.m.
Up here some place?
Okay.
22 I'm looking for Jerry Hulman.
He's supposed to be 23 writing kind of a brief summary. I am looking for Jerry Hulman.
24 I don't see him any place.
25 So, maybe we'll take an early break for about five or Heri'tage Rep rting Corporation (202) 628-4888
_ ~ _. _ _ _. _ _
a i
i 479 1
ten minutes and then we'll reassemble here.
Okay?
3 i
2 (Recess.)
l l
3 MR. HULMANt I'm going to summarize very generally j
i
]
4 what's happened in the last couple of days, and then we're l
1 i
l 5
going to call on three people to see if they can summarize what l
1 6
they believe might be the effects of water.
Then, we're going 7
to throw it open for discussion.
l i
8 We've talked about several different challenges to t
q 9
containment in the BWR Mark I.
We've talked about early 10 challenges. We've talked about late challenges.
We've talked 11 about bypass. We've talked about steam spikes or explosions, j
i 7
[
12 and we've talked about liner failure.
j l
13 The record has been filled out and we've gotten some i
~
14 information at this meeting on the challenges that we didn't 15 have before.
We've gotten information from not only NRC 16 contractors, but from industry.
17 We've talked about the improvements.
Post of the i
t 18 information that's been presented at this meeting has come from
]
19 NRC contractors, but there has been some contribution of new 20 information on hydrogen controls, sprays, core debris, venting, 21 ADS improvements, and there was discussion on training and
[
22 procedures.
4 j
23 On experiments, we've heard from NRC contractors on 1
l 24 their conclusions, the bases for which have not all been
)
25 documented.
The same thing for industry.
We've heard j
Heritage rg poration
_.4
480 1
conclusions, for' example, from experiments conducted by Bob 2
Henry, documentation is not yet complete.
3 There's concern over all of the experimental work 4
with respect to whether it's adequately scaled and whether it's 5
considered the range of boundary conditions that may be 6
important.
l 7
I encourage all of the people workir.g on the 8
experiments to document their work so that we can consider the 9
scaling effects and the boundary conditions.
10 On the analytical side, we've heard discussions of a 11 probabilistic approach to assessing the problem. We've also 12 heard from Hershel Spector on the importance of source terms.
13 The documentation of these comments is important.
14 We've heard conclusions, but not all of the information is 15 fully documented, and I want to encourage the people at the 16 ineotlag to document their work.
17 Finally, the schedule.
I want to reiterate what I l
18 started out with at the beginning of the meeting, saying we've i
19 made a commitment to the Commission to give them an interim j
20 report on Mark I Containment Performance in April, 21 We've told them we're going to give them a final 22 report by the end of Auo%t.
We also told them that we're 23 going to give them reports on the other containment types by l
FN sg 24 the end of August of next year.
25 I think'it's important for the staff and I also i
tieritage Her>orting Corporation (202) 628-4888 j
~_ _
)
i 481 i
i believe it may be important for industry to close out this i
2 issue of containment performance, and we intend to try and meet i
3 those dates.
j 4
I heard a lot of comments about the importance of the i
j j
5 IPE letter.
We agree.
We have been trying to get it out, and I
i 6
at thic point in tirse, it's being considered by the highest 7
levels of managemont and the staff.
We hope to have it.out in 1
8 April.
2 I
9 You also heard that we think research on accident 10 management is important.
That's going forward.
What you may l
11 not have heard but what the staff believes to be important is 12 watching the operations of the individual reactors and making l
Q 13 sure that the operations are good ones.
Sometimes referred to j
i 14 as improved operations, f
l 15 Lastly, be aware that there's going to be continuing l
16 research on NRC's part, on a.'sas that we think are important,
]!
17 areas of uncertainty, areas on phenomenon.
Thet research will 18 go forward also.
l l
19 That's a brief general summary.
I wanted to lay the 20 foundation for discussions with a brief summary.
l 21 Hith that, I'd like to call on Dana Powers to see if I
j 22 he could give us an overview and a summary of the role of water l
23 as he sees it.
l l
24 MR. POWERS:
I've been asked to give some summary i
lj 25 comments on how water affects the progression of accidents in j
Heritage Reporting Corporation (202) 628-4388 j
l
482 0
1 the Mark I BWR.
2 I think it goes without saying that if we can get 3
water any place into a Mark I BWR during an accident, we would 4
like to get it into the vessel and arrest accident progression, 5
keep the core intact within the vessel itself.
6 If you look at the Mark I BWR, there are multiple 7
sources of water to the vessel.
Some of them are operative 8
only at low pressures, several of them can be optimized to 9
operate at high pressures.
10 Unfortunately, under accident conditions, it's easy 11 to hypothesize that these water sources may not be availdble 12 and the accident will progress to the point that core debris
')
13 does come ex vessel.
'w 14 So, the question turns on how does the core debris 15 behave once it's escaped the reactor versel and passed into the 16 dry well and can that behavior be altered by providing water 17 into the dry well probably by the dry well sprays.
18 In the NRC Research Task Force, we examined the i
19 question of how water might affect core debris behavier, and we l
20 looked at it in two questions. When could we alter the debrid 1
21 flow attack on the concrete and possible attack on the liner, s
22 and could we alter the fission product source term.
23 Wo looked at the available data base, and we found O')
24 that there was not data to prove that water would prevent core 25 debris from coming into the dry well, that it would prevent Heritage Reporting Corporation (207) 628-4888
483 1
core debris from attacking the concreto, or that it 5.ould 2
prevent the attack on the liner.
3 br. Henry presented some interesting results in which 4
he showed the water could have a tremendous effect on the S
initial flow, initial freezing behavior of core debrie.
We 6
have to remember that thone experiments only addr40s the 7
initial transients.
The longer-term interaction, by longer-8 term, I mean by several minutes, was not addressed by those 9
experiments.
10 When we've examined the data base that might be 11 applicable to those longer-term, that several minutes up to a 12 halt an hour, we still don't see data that conclusively 13 demonstrates you can pr' vent liner melt through by the presence e
14 of water.
15 On the other hand, we looked at how water might 16 affect the source term, and there we found tremendous evidence 17 that water could have a great impact on the magnitude of the 18 source term that could arise during a reactor accident.
19 We found two mechanisms by which water might preven *.
20 excessively large source terms, first, the sprays would provide 21 pools over-lying the core debris.
These pools act much the 22 same way as the suppression pool does to trap fission products.
23 We found experimental data that showed that there g
(j 24 were an order of magnitude reduction in the source term coming 25 from melt concrete interactions with possible by-water pools lloritage Reporting Corporation (202) 628-4888
484 5
1 over-lying the core debris.
0 t
2 The second mechanism was the sprays themselves could f
3 trap aerosols, and there we found an existing technology that l
t 4
seemed to suggest not just order of magnitude reductions in the i
5 source term, but several order of magnitude reductions in the 6
source term'were possible if the atmosphere could be cleansed f
i 7
by the action of the sprays, 8
It is my personal conclusion *that if I were looking f
i 9
for accident mitigation methods for very severe accidents that t
f 10 had progressed to the ex vescel stage, I would.look not at 11 trying to prevent liner failure, but to lock to water as a 12 means of reducing the source term to the point that liner i
O 13 t
failure would be a largely inconsequential event in the reactor j
14 accident.
I 15 MR. HULMAN:
Dr. Theofanous.
16 DR. THEOFANOUS:
Well, I think I will take a position f
i 17 that maybe some of you might fjed it somewhat extremist and l
i 18 maybe not easily supportable, but I would claim that we don't i
19 have at this point correct experiments that will tell us one i
20 way or the other what is the effect of water in relation to the i
21 dry well melt through question.
i 22 I would like to disagree with my friend, Dana Powers, 23 that the experiments presented by Henry give some indication
(~))
24 that the initial transient may be the effect of what is i
25 important. I will think that even for the initial transient, I Heritage Re prting Corporation
]
(202) 628-4888
i i
485 1
have very significant questions as to whether these experiments 2
tell us that the water can actually prevent something happening f
3 to the liner.
4 The reason I'm saying that is because i think-the 5
significant scale of effect and those experiments do just the i
6 opposite of what I was askino George Greene the other day.
In 7
George Greene's experiment, he's_using a lab.
It's a small 8
scale.
He under-estimates his heat losses and, therefore, the l
9 whole thing is just like'a flow of water.
i 10 In Henry's experiments, he over-e'stimates his heat i
t 11 losses because it is a small scale experiment, surface to 12 volume mass is so large, that actually per unit volume of melt
[
T 13 that is trying to spread in, you have tremendous heat loss.
{
14 You put water in the tub and route it to an FCI unit tub,
[
15 there's no doubt that you're going to freeze the damn thing 16 quite closely before it gets to the liner.
l 17 I would like to take the position that if you really i
18 want to know what is the effect of water and, in fact, how the I
19 melt might come and spread out into the pedestal and dry well 20 area, I think, first, people need to do, including myself, l
21 their homawork.
We need to do first a water analysis to figure i
.2 out what are the important phenomena, do some scaline, find out 23 what the different' time scales are, and only after this 24 analysis is done exhaustively and after it's been discussed 25 among the different people, because, as you know, everyone can-i Heritago Reprting Corporation (202) 628-4888
+
l 486 1
take different approaches in apprc)ching those kinds of 2
technical issues, and after there is a thorough discussion and i
3 kind of wrap-up between people, only then, I think, one can 4
actually know what kind of experiment is an appropriate 5
experiment to do.
6 It might actually turn out that one can not do an 7
integral simulation experiment in scale, but, rather, maybe one 8
can do some kind of side experiments that will shed light on 9
some aspects of this problem that before doing that kind of 10 thing, one could then put it into some analysis and try to make 11 a reasonably technical prediction as to whether so many tons of 12 melt coming out into the pedestal will go how far.
b
/
13 All ri-So, that's the first point I want to make 14 and to summarize tuat because I think I rambled a little bit, 15 is that we don't have the right kind of experiments now one way 16 or the other.
We could very well find, I will make allowance 17 for that, we could very well tind that water has very 18 significant impact, all right, but this still has to be 19 de:aons trated, and I would like - o urge everybody to do a water 20 analysis,.a lot of thinking about how to simulate or what 21 aspects even can be simulated correctly.
It's small scale 22 because that's all we can do, of course, and then go ahead and 23 do the experiment and we go back and complete the analysis.
24 So, that is the first point.
25 The second point that I want to make is that as far
)
Heritage Reporting Corporation j
(202) 628-4888 j
i m
e 487 1
as the liner question, the e f fect of water, I think, and I'm 2
speculating at this point, I feel that from what thought I have 3
given to this problem, that ninety-nine percent of the answer 4
as far as the liner lies on the in vessel portion of the 5
accident.
6
'm pretty much convinced myself through the simple
~
7 analysis that I showed you the first day that the melt is going 8
to come out of the boiling water reactor essentially in a very, 9
very small quantity and through some kind of a last failure of 10 the lower head, and what leads me to that conclusion is because 1.1 I feel it's an indisputable fact and that you don't need to 12 have a very detailed model with one hundred parameters to turn.
13 we just do a simple energy balance and you find out that no, 14 matter how you take that melt or that core to slam into the 15 lower head of the boiling water reactor, because of the 16 tremendous amount of heat things that you have there, there's 17 just no two ways about it, the stuff is going to be quenched, 18 Even if there is no water there at all, I claim the 19 stuff is going to be quenched.
Just imagine putting fuel that 20 in molten through pipes that are about an inch and a half in 21 diameter and four meters long.
Just imagine that and just do 22 that.
Any temperature.
I'd like you to pick even 300 degrees 23 super-heat and find out how far it's going to go and find out 24 after it stays there for one minute or two minutes whether it's 25 going to be quenched or not.
Heritage Reprting Corporation (202) 628-4888
488 1
That's really what exactly the boiling water reactor 2
lower head looks like.
All right. So, I am convinced that the 3
stuff is going to be quenched when it gets there.
Because it 4
is quenched, then I'm convinced that the lower head is going to 5
give well before everything becomes molten again, and I think 6
the answer to the liner problem lies there and what I suggest 7
there is that pecole actua.11y -- more people do this kind of 8
analysis and then we play a game where everybody draws a i
9 probability distribution with respect to how m h of the core 10 they expect to be molten at the time that the vessel breaches.
11 And if we find that people agree, then I think we 12 shall have the liner problem right there and then.
You don't O
13 need to know anything more.
14 Now, if we find that there is some good basis and l
15 there is some good analysis that show that there is a 1F significant fraction of the core that is molten at the time of 17 vessel breach, then that takes us to the next step.
18 All right.
Now, so, we go down to the next step and i
19 the next step is on the question of spreading and the effect of l
20 water now in the pedestal and in outside of the pedestal in the 21 spreading process.
22 I think it's very critical, very critical for us to 23 recognize two time regimes.
There is this initial time regime j
24 that I will characterize as the initial whatever is molten just 25 comes out, is something less than a minute, of course, you Heritage Re wrting Corporation (202) 628-4888
-g--
.n.
489 1
will take it, if it's molten, and then you've got a big-hole i
~
2 that's going to come out and the short time, so it will splash 3
and it will spread.
Okay.
That's the short-time regime.
4 Short-time element.
5 In this part, I have some doubts as to what the 6
effects of water would be.
I think it's possible that the 7
water may create some water FCI if the quantities are not too 8
great. If we're talking some of the order of even thirty tons 9
or forty tons of melt, I think the water can have a very 10 significant effect in slowing you down, but if you also are 11 talking about thirty to forty tons anyway, I don't think that 12 you are worried about the liner because it isn't that much to Os P
13 go all the way out there anyway, to submerge it.
14 All right.
Now, if you talk about, on the other 15 hand, about the hundred tons of melt coming out all at once, I 16 think you're going to have a hard time to show that even if the 17 water hits,,some water interactions, scme local high heat flux, 18 I think that you can visualize various things that you can 19 never test where there may be some local interaction in the 20 water way and the rest of the melt goes over to the lag.
21 I think you begin to get into a no-man's land at that 22 point when you get to deal with such large quantities of melt.
23 But I think it's going to turn out that this is not
/~T
()
24 to be the case.
So, we are okay in that respect, unless 25 somebody comes tomorrow and proves me wrong.
All right.
Heritage ReW rting Corporation (202) 628-4888
-r.,
g
..------~,,-,.-a
,w-
490 1
Then, the -- however, on the long-term, which is the 2
regime now of this long gradual addition of melt by melting 3
through whatever is left in the lower plenum and in the core or 4
if it's come out in a large quantity slowly, what melts into 5
the pedestal area.
That now I call the long-time regime.
5 Obviously, if you visualize all the stuff going 7
inside the pedestal, it would be something like three feet, 8
four feet deep, obviously that's going to melt again, all 9
right,.and it will want to spread.
10 Now, I believe in that kind of a physical 11 configuration, the effect of water is going to be paramount, is 12 going to be critical.
I see the water, because of the
/~T i
w) 13 treme'ndous amount of heat loss that the water can provide 14 because it's being right next to where the melt is, not 15 necessarily from the point of view of maybe making coolable the 16 whole thing, you couldn't do that because the amount of 17 dimensional effect, okay, but by mainly from the point of view 18 of affecting these gradual spreading, some material melts and 19 melts and now it actually tries to work itself against the 20 crusts.
21 So, that is, I think, where I am going to be doing 22 some work, already I have decided that, and I think I will i
23 encourage other people to do work because I think it's critical p) x 24 to know how those crusts now, as the material is melting inside 25 and gas is produced with the presence of the water, how the Heritage Re w rting Corporation
'(202) 628-4888
491 1
thing is going to be,.how the configuration is going to lead to 2
that spreading.
3 I will not be surprised if, after this work is 4
completed, we find that there is significant non-uniformity 5
even in what I call th6 final after everything is said and 6
done.
Kind of steady state ultimate disposition of the melt.
7 I wouldn't be surprised if there was very significant non-8 uniformity in heights, where a lot of melt actually still 9
resides in with the pedestal area, and maybe even eating its 10 way around a little bit on the concrete on the pedestal wall, 11 while only a small amount actually has gotten as far as even up 12 to the liner.
i 13 That's something that has to be seen, has to be 14 examined, and I feel that the water will have a very profound j
a 15 effect on that long-term kind of a spreading situation.
16 And, then, at the end, of course, that is 17 indisputable fact.
I like to agree with Dana that, indeed, the 18 water has a very important beneficial effect in the short-term, 19 and, so, from any side I look at it, I like to encourage 20 everybody to think about making sure there is water in the Mark 21 I's.
22 Thank you.
23 MR. HULMAN:
Bob Henry.
24 MR. HENRY:
Water is obviously a very powerful source 25 of cooling the core,of tho reactor system.
That's what it's i
fleritage Reporting Corporation (202) 628-4888
492 1
designed for.
And as Tag eloquently described for us 2
yesterday, if the plant is having a bad day, water is a very 3
important way that we try to deal with that in an emergency set 4
of sequences, and if we actually end up in core damage, water 5
added to the vessel is the way that we can keep it in vessel 6
and any time we look at these systems, that's the way that the 7
attention should be directed.
8 If, for some reason, we can't provide water to the 9
reactor system, we can provide it to the containment, it has a 10 tremendously beneficial r' ole in the containment as well, and I 11 think it has that beneficial role both in terms of the short-12 term of handling high temperature material which may be coming 13 out of the vessel and may be coming out slowly or rapidly.
In 14 either case, it has a beneficial effect, and it has a 15 beneficial effect long-term, both in terms of perhaps taking 16 fission products out of the atmosphere and long-term cooling of 17 the material outside of the vessel region.
It has all these 18 beneficial effects.
19 with respect to the experiments, do they represent, 20 are they prototypic of all the reactor conditions under these 21 kinds of anticipated releases?
No.
Were they intended to be 22 representative of all the reactor conditions under these kinds 23 of releases?
No.
24 What they were intended to do was to demonstrate that 25 if high temperature material comes out and sees water, that the 11eritage Reprting Corporation (202) 628-4888
493 1
dynamic interaction between the two promotes very rapid energy 2
transfer between the melt and the water, and something which
~
3 has not been modeled in anybody's analysis to date, so I would 4
agree with Theo and Dana, let's do some analyses and let's look 5
at the dynamic interacticns of water in a realistic way and 6
that's what the experiments were directed at.
7 They demonstrate that this interaction of heat loss 8
is with water and that is the thing that was coming out very 9
rapidly into the quench tank and it does definitely demonstrate 10 that. So, that is the rapid heat loss and that would be 11 prototypic of the reactor system in terms of the time scale, 12 and I think the last thing that one gets from that is the fact O
13 that it's a mistake if we don't carry out experiments in the l
14 future of this nature with the presence of water so that we can 15 understand the true benefits that it has.
16 I guess we're in violent agreement, if you will, that 17 from the perspective of the plant, the water has tremendously 18 beneficial aspects, not only in the reactor system but some 19 reason, as Tag said, you're having a tremendously bad day in 20 the containment as well.
21 MR. HULMAN:
Questions?
Comments?
Discussion?
22 George Greene.
Would you come up to the microphone 23 in the center?
Make sure the mike is on.
(
24 MR. GREENE:
In my discussion yesterday, I talked 25 about the immediate need to continue ahead energetically now i
Heritage ReW rting Corporation (202) 628-4888
494 1
for more analyses on this Mark I steel shell problem as opposed 2
to instead at this time pursuing experimentally attempts to 3
narrow the ranges of uncertainty in the dominant physical 4
phenomenon.
5 There seems to be a strong perception in this meeting 6
that there is a dearth or, in fact, some people would suggest, 7
a total absence of experimental data with which to apply to the r
t 8
resolution of the Mark I issue.
I do not share that opinion.
9 One of my slides on Wednesday afternoon, prior to 10 most of the other discussions and presentations, represented 11 information on a comparative assessment of the original 12 Brookhaven analysis.
O 13 I presented some information on,the comparative 14 assessment of the Brookhaven analysis and the IDCOR analysis 15 and FAI Report 86-1, and I made it perfectly clear in my 16 perception of where the uncertainties laid and the 17 phenomenological issues were in melt spreading, crusting 18 behavior of melts, and the boiling behavior of water.
19 I'm gratified that almost every discussion since then 20 has hit those three points.
We know where the differences lie.
21 They support the perceived range of uncertainty in these 22 phenomenon.
23 I went on at that point to describe the fact that we
\\)
24 were pursuing this experimentally and I suggested that there 25 were six experiments at Brookhaven that we were doing currently Heritage Rew rting Corporation (202) 628-4888 j
i
. ~.,. _ _ _, _
. 1
495 1
with respect to the phenomenological issues in the absence of a 2
full-blown probabilistic assessment now to guide those 3
experiments, and those were the following:
bubbling heat 4
transfer to vertical boundaries and inclined boundaries; melt 5
spreading; the effects of water on melt spreading; boiling of 6
water over core debris or simulated core debris; response of 7
steel structures in contact with some simulated core debris; 8
and the modes of freezing of simulated core d'ebris.
~
9 Every presentation that's gone on with respect to the 10 Mark I steel shell, based solely on analyses, has demonstrated 11 each of these effects is important.
We know where the 12 uncertainties lie.
O 13 Anyone who suggests that there's no experimental data 14 base at this point that's appropriate for issue resolution will 15 probably never be convinced. Those are the experiments at 16 Brookhaven Laboratory.
17 In addition to those experiments, the data base 18 that's come from the Sandia National Laboratories in melt 19 concrete interactions and the effects of water on aerosols and 20 the effect of water on MCCIs is staggering.
Simply, they have 21 the burn series, the CC series of tests, the Swiss testo, the 22 Turk tests, the Cert tests, and the list goes on.
23 If you're not familiar with the'se tests, I'd suggest i
24 that you do a literature search.
25 There seems to be a strong feeling that these test Heritage Rewrting Corporation (202) 628-4838
t i
496 1
series are inadequate to resolve the issue. They may not be 2
appropriate to resolve the issue completely right now, but the i
3 data base is, simply put, staggering.
We know a considerable i
4 amount about the phe'nomenon. Our experiments at Brookhaven and f
.I 5
Sandia will continue.
They will continue to augment the data i
6 base and our intent is to cooperate with the industry I
7 representatives to come to some sub-set of physical phenomena l
8 that are important, narrow the range of uncertainty in those 9
phenomena and resolve the technical issue in a timely fashion 10 experimentally.
l I
11 MR. HULMAN:
Do I understand, George, that that's an 12 invitation for critique and comment from industry on the 13 program as well as the data?
l l
14 MR. GREENE:
Absolutely.
15 MR. HULMAN:
Okay.
Any other comments?
Discussion t
16 on anything that's happened at the meeting, on the summaries?
17 Discussion?
j 18 (No response.)
i 19 MR. HULMAN:
Mel, did you want the opportunity?
Mel 20 Silverberg.
21 MR. SILVERBERG:
Thank you, Jerry.
22 I just wanted to add a few comments and perspectives 23 on some of the observations I've made over the past -- better 24 part of the past two days, which I think are worth noting.
25 I think the proceeding on some of the comments that Heritage Rew rting Corporation (202) 628-4888 wm,-%
~
-,e
-4
--,w,,
ev
1 497 1
were made this morning by George Greene and comments yesterday 2
by Bill Rasin that I think are quite appropriate, that the 3
differences that one is getting from the various calculations-4 of liner behavior are clearly driven by the differences in 5
input assumptions, and one can continue with, you know, 6
calculational exercises, but the cottom line lies in those 7
differences and I think some of the points made about using the 8
data that's available as well as focusing those to got at the 9
differences.
10 I think, again, it's a question of the uncertainties 11 and calculational exercises per se don't necessarily, won't 12 necessarily solve the problem or resolve them.
13 The -- I noticed that there's a_ tendency for people 14 discussing code predictions and I've heard a number of claims 15 on results and so forth, but without again really having an 16 attention to or an exposition of the uncertainties-that are 17 associated with analyses, someone says, well, I think I've got 18 a low source term, okay, but what's the up and the low side of 19 the -- the up side and the low side of that, and I think Mohsen i
20 Khatib-Rahbar from Brookhaven made that point quite clearly, 21 that you've got to go out, it's tough, but you've got to go out 22 and actually look at the uncertainties.
The up side and the 23 down side of your uncertainties, and point values in 24' themselves, while encouraged and. interesting and good to help 25 focus, given the uncertainties we have, one should not dismiss lieritage Re wrting Corporation (202) 628-4888 1
r
~v.,
i I
l 498 1
a good comprehensive uncertainty analysis.
2 I also detected and was concerned that what appears 3
to be a trend towards someone using a code, any code, another 4
code, and if they get an answer that they like or that is 5
favorable, then they, in effect, accept the analysis or the 6
code _or the methodology out of hand without a critical review, 7
whereas whenever there's a concern when, if you will, the 8
answer is not favorable,'one immediately says, well, gee, we've 9
got to go out and get a review.
10 I think the criteria should be similar'for all.
In 11 other words, given an analysis method, given a new methccdology 12 or an old methodology, whether -- regardless of what the 13 answers are, they need to be critically reviewed, and which 14 brings me to my next point.
15 When the various reviewers of the NRC program have 16 commented on the program, and I might say that the long list of 17 critical reviews that the NRC work has had has been quite, I 18 think, quite impressive and has involved a considerable effort 19 on the part of many people.
The -- I'm talking about the APS, 20 National Academy of Science review, they refer to the 21 importance of having the work and you refer to a lot of it, 22 Jerry, in terms of documenting results and things, it's
~
~
23 Jmportant to have, if you will, peer review, open peer review,
()
24 publications in referee journals, peer review journals.
25 But somehow there seems to be a fee.1ing that it's Heritagn Reporting Corporation (202) 628-4888 r
_ ~.. -
499 1
only the NRC work that needs the peer review, the open broad 2
peer review, and that others don't necessarily fall into that 3
category, and the reason why I make that observation, it kind 4
of goes back to the codes that have been bandled about, 5
starting with the source term code package,-which, of course, l
6 has had, as you know, a rather thorough open review, and then 7
we get the MAAP code and we get to the codes that the people 8
are using at NYPA, and perhaps there should be peer review in 9
their own community, but I don't think they have been subjected 10 to the broad critical open peer review that the NRC codes have,-
11 and my botto., line is that until one has a full and open 12 critical review of all the methodology, one will not make the
(~)
\\-V 13 kind of scientific and technical progress that one needs in the 14 complex technology that we're dealing with in severe accidents.
15 And, so, I think that I want to continue to encourage 16 others to participate in continuing the need for this open J
17 review and I think in the long run, I think we all will benefit 18 by it technically and also it will'also, I think, help speed
)
19 things up quite a bit.
20 Thank you.
21 MR. HULMAN:
Mr. Spector.
22 MR. SPECTOR:
I'd like to respond to Dr. Silverberg's 23 statement, speaking now just for the Power Authority as some
()
24 group which has, in fact, contributed to source term code work.
1 25 Let me first talk about the peer review process.
The Heritage Re wrting Corporation (202) 628-4888
)
i
_.m.,
500 1
Power Authority is in full agreement that any piece of 2
scientific work should go through the peer review process and 3
we reached that decision many years ago, and in implementing 4
that, we have come before the ACRS several times, made 5
presentations on our source term work, starting with the Indian 6
Point 3 work, and also the work that we did on Fitzpetrick.
7 That was several years ago.
We have come before the 8
NRC staff similarly.
We have come before the American Physical 9
Society, made presentations out at the meeting at Snowberg.
We 10 have published documents both nationally and internationally.
11 We have recently come before the American Nuclear 12 Society's source term 1150 group, just this February. We have, 13 in fact, when we submitted our comments on NUREG 1150, 14 submitted a pile of paper which is maybe this high, maybe a 15 little bit large, but it went down to the absolute minutia of 16 how we did our analyses with codes, equations, references, and 17 most of all, we have always made ourselves available to the 18 staff at any time and, in fact, in just discussions this week, 19 Jerry has very kindly extended an offer to look once again at 20 our most recent analyses and we have responded that we will 21 submit everything within a week, actually a lot of it is 1
22 already at the Commission and has been there since October 1st, 23 but we will resubmit a lot of it and we stand ready to meet
()
24 with the Commission at any time to satisfy any residual 25 questions they may have.
Heritage Rom rting Corporation (202) 628-4888
i l
501 1
In addition, a large portion of the codes that we use 2
are codes that other people have developed, including the 3
Commission itself.
We've extended them, but we've corrected 4
them. So, they also benefit from the peer reviews that are 5
inherent in these sub-packages.
6 So, speaking at least for the Power Authority, we 7
have pursued peer review.
We have gotten it and we are more 8
than willing to continue that process because we think it 9
serves us all, and I think if I may generalize a bit, I think 10 this you find to be true throughout the nuclear industry and 11 sometimes the -- being quite candid, it has been difficult to l
12 find the time in the NRC staff to devote to the line by line
(
13 analysis that some of us frankly would enjor.
14 Thank you.
l 15 MR. HULMAN:
Mr. Fuller.
16 MR. FULLER:
I feel compelled to respond to Mel's 17 statement, also, t
18 I need not remind people like Mel or Themis about the 19 very extensive interaction process that took place between 20 IDCOR and the NRC when the technical information that was 21 produced by IDCOR was presented, including the MAAP code 22 results.
1 23 We had at least a half a dozen or so meetings.
They O(_j 24 were very well attended, including by a lot of people who are 25 sitting in this room today.
I Heritage Reprting Corporation (202) 628-4888 y
.y-m,r r
e
-c.-+e
1 502
(-]/
['
1 During those meetings, there was a very free exchange 2
of information, and the exchange was prefaced by the fact that 3
prior to the meetings, the IDCOR documentation had been sent to 4
4 the NRC, including some MAAP user manuals, so that the various 5
members of the NRC technical contractor community, national 6
community, did have time to review the material.
7 And I must say that the meetings that took place were 8
extremely lively, very productive, a lo't of information was 9
exchanged on all parts.
To me, that constitutes a pretty good 10 peer review, just for starters.
11 Since then, we in the industry have continued to --
12 and also I should add that on the basis of those meetings, as 13 everybody knows, certain issues were brought up as yet to be 14 resolved, to be worked on, and the industry has continued to 15 work on those to resolve the issues.
16 I'm not the person to talk about how those were 17 resolved because I got out of the IDCOR program during that 18 portion, but suffice it to way that if somebody wants to back 19 me up, he can.
Suffice it to say that they were worked on very 20 actively and are still being worked on.
21 Meanwhile, we at EPRI are continuing to foster the 22 continued development and acceptability of the MAAP code.
We 23 have undergone extensive verification and validation of the
()
24 code.
We have right now underway a very rigorous design review 25 of the code, which is designed -- which is -- has as its lieritage Re wrting Corporation (202) 628-4888
t 503 1
objective to get it into a form'that will be suitable for the 2
NRC to review in the context of 10 CFR 50 Appendix B.
3 We are very serious about this. So, we feel we're 4
continuing to work on our part to get adequate and acceptable 5
review.
t 6
The NRC has been asked to review the code and, 7
frankly, the rate of' progress on that has been less than 1
8 speedy.
I don't know why, but it has been less than speedy.
i 9
I think that, though, the time has come to heal 10 whatever wounds are opening up because we're all in this t
11 together.
We do need to understand analytically as well as 12 experimentally what we might expect would happen if we ever got
[
l 13 into one of these severe accidents, j
14 And, so, I call upon everybody to work in that 15 spirit.
I 16 MR. HULMAN:
Unte Cheh.
17 MR. CHEH:
Nrc, Unte Cheh.
I 18 This is a continuation of my remarks about the core 19 development, also tha analysis. When I see the comments made by 20 the national laboratory and industry, there's no mention of the j
21 problems they face in the steam spikes.
This is one of the l
22 technical areas one can easily resolve by reducing the time
~
23 interbar in that the abrupt change -- however, that's got to be
(
24 done so'that we will know that -- why that steam spike in terms 25 of how we can handle that thing and instead of shut down that Heritage Re p rting Corporation
[
(202) 628-4888 t
504 1
steam spike.
They've got to get it all the way up by using the 2
low scale.
Through my experience when I worked with the national 3
4 laboratory and the industry, we can handle this problem very 5
easily.
So, that is one thing I would like to point out.
I 6
saw some base cases.
Base cases selected in parameter studies 7
or sensitivity studies, whatever you call, when we selected the 8
base case, we've got to make some sense of when base case comes 9
in.
Everything's got to be compared with the reasonable base 10 case and why these things happen, whv is it that high, that 11 low, PCT or source term.
To eliminate all those parameters, 12 most parameters not important.
13 That's what base case is used. So, we've got to use 14 these base case based on reasonable parameters and then set up 15 the bench mark to compare other things against that bench mark.
i 16 The other thing is that again when we make the 17 idealization to set up the model and also the nodalization, 18 we've got to nodalize, otherwise if one test shows just that i
19 average of temperature only, but someone talking about the 20 temperature or pressure, what temperature, what pressure we 21 talking about?
It's a good question because when I left only a 22 few years ago this nuclear industry.
So, I cannot understand 23 what they're trying to do and one model they have used, so it's
(
24 like I get lost.
25 One of the presentations was here that degassing Ildritage Rem 7rting Corporation (202) 628-4888
505 I
takes place above the ablation temperature.
I talk informally
,(~
2 with the representative of that national laboratory, that 3
doesn't make any sense or something must be wrong.
Okay.
Here 4
it is.
That's right.
That was developed at another national-5 laboratory.
Okay.
There is something wrong. You ask them to 6
fix it.
He say yes.
Okay.
7 Then, what did they do.
One, I have been waiting.
8 Just one hour before the plane took off, I got it, so it be 9
fixed and be done and be present source again.
10 So, we have here a capability to review that thing.
11 If I feel uncomfortable, I can tell you right away that 12 something is terribly wrong here in computer code or anything.
I f
13 So, let's do work on this one.
My study goes rather to the,
14 people working in national laboratory.
It takes awhile for 15 them to graduate from schools. Come to national laboratory, 16 produce something sensible making sense.
It usually takes two 17 years or so and maybe some manpower prob)eias, but someone got 18 to keep some guidance to this national laboratory people in 19 this regard.
20 This comes from experience, though, rather training 21 and the experience.
That's what I feel, and I discussed this 22 matter somehow with my management.
23 Thank you.
24 MP. HULMAN:
Dr. Theofanous.
25 DR. THEOFANOUS:
Since the subject of peer review Heritage RON rting Corporation (202) 628-4888
506 j
1 came up, I want to make a small contribution.
I think the i
2 matter of peer review -- can you hear me now? No.
3 I said since the subject of peer review came up, I'd 4
like to make a small contribution.
I have found that we are 5
lacking the culture in this industry, the culture that is going 6
to encourage people to really be subjected to a peer review 7
process the way it is understood, this word, in'ternationally.
8 We seem to think that the peer review is when there 9
is a problem comes up, we go through a crash meeting and we i
10 call this a peer review. Of course, that is a form of peer 11 review, but the real benefit the peer review is when people are 12 continually throughout the exercise of their professional 13 duties, they are exposed to the opinion of their peers.
i 14 That is normally and for many, many years now, is 15 done in the framework of sending papers to the literature and 16 being reviewed anonymously, where actually under the anonymity, 17 the referee can actually express fully, without any bias and 18 without any hindrance, their opinion about the technical work.
19 Now, I think the way that this happens actually is 20 not the fault of the people that are doing the work, either the 21 national laboratory or some of us at the universities. Of 22 course, we have to do it for other reasons.
We have to publish 23 as part of our work, but people in national laboratories and 24 people in industry, the reason is mostly because the 1
25 management,-I have found, of these places does not encourage I
Heritage Reporting Corporation (202) 628-4888 w w
.m w-p
,7--
x, o
507 3
1 adequately or they presumably don't think that's a good way to 6
)
i 2
spend somebody's time to write papers.
3 Now, what we do at the universities actually, we many j
4 times stay up all night writing papers ^
We do it after hours.
{
5 I don't know if that's something that the national laboratory
[
l 6
people can do, but I would really like to encourage the j
r h
7 management of NRC and the management in industry and the j
8 utilities.
It is so extremely important to think of this r
9 review in these broader contexts of encouraging people to do j
1 10 technical work, to write their work, submit it to journals, and
?
11 be subjected to that on a continuing basis, and only then the i
12 standard can actually be elevated to where it's consistent with i
13 the standards of the other societies.
i 14 We tend to be very often too inbred and if you call t
i 15 what ANS does once in awhile with summary, I think it's a i
16 euphemism if there ever was one, this summary review in the 17 five minutes you look at a paper and it is ninety words long, 1
i 18 all right.
1 19 So, this inbreeding is very, very detrimental, I t
20 think, to us because then mechanical engineers, chemical I
21 engineers, analytical engineers,, people from outside, they look i
22 to our work many times from the outside.
That is totally I
23 inconsistent with our standards.
24 MR. HULMAN:
Mr. Hill.
j 2$
MR. HILL:
I would like to reflect on what I think is 1
i
(
Heritage Re p rting Corporation (202) 628-4888
-a.-.,
508
[~')
1 the point of view of my particular utility on these matters.
(J l
2 First of all, I'd like to say that the objective that 3
PP&L has in doing risk assessment of any kind is not to 4
calculate the f rs,1ency of damage to our plant or to calculate 5
the frequency of severe or unacceptable off-site consequences.
6 It's rather to determine what kind of procedures we 7
should have and what kind of training we should have for our 8
operators, and to determine that these procedures and this i
9 training minimizes the risk or the frequency of damage or i
10 public consequence.
11 The perspective that you have from that point of 12 view, I think, is frequently quite different and sometimes 13 diametrically opposed to the perspective that you might have if 14 you're trying to do a conservative risk assessment, and I think 15 that that 3hould be kept very clearly in mind.
16 The second thing that I would like to point out is 17 that once again from the utility perspective, I believe we have i
18 a pretty good understanding of the equipment in our plant and 19 how it functions. We have a pretty good understanding of what 20 kind of people we have operating the plant.
i 21 We know pretty well how they perform.
At PP&L, we 22 make measurements on them, and as a result, we feel that the 23 front-end of the risk assessment process has relatively low 24 uncertainties.
It's true there are uncertainties, and it's 25 true we don't know everything that represents a risk to the Heritage Rew rting Corporation (202) 628-4d88 i
l l
t j
509 1
plant. But we do study the plant as we know it in considerable 2
detail and we take all of these tactors into account.
[
3 The result, we believe, is a relatively low
(
f 4
uncertainty picture of the front-end part of the risk
[
5 assessment process.
We'do not have the same level of 6
confidence in the back-end part of this analysis process.
{
7 Out to the point of where the reactor vessel fails, 8
up to that point, we feel reasonably comfortable, although we l
t 9
still have some questions as to the core damage progression 10 models having to do with relocation of material into the lower l
11 plenum of the vessel.
i i
12 As the vessel fails and material relocates on to the td 13 dry well floor, our lack of confidence in the models is even l
14 g rea te r -.
15 Our response to dealing with this situation is that 16 we believe we should make every attempt to keep the damage i
17 inside v;he vessel, if we can, and we believe that that should
(
i 4
18 be givea a great deal of attention.
19 Falling that, then we would like to know what to do j
i 20 in terms of our procedures and our training of our people to do i
21 to stop the progression once material is on the dry well floor.
l 1
l 22 We believe that if a proper front-end assessment is 23 made and the actual capability of the plant, its equipment and l
24 its operators are taken into account, the importance of some of 25 these back-end unc'ertainties is greatly reduced, and somehow or i
i IleritageRek$rtin2 Corporation (20 628 4888 t
510 1
other I don't get the feeling that others share this 2
perspective.
3 I would really like to recommend that much closer 4
attention be paid to defining the front-end accident sequences 5
that lead to the point of core damage or a threat to the plant 6
of some other form.
7 MR. HULMAN:
Any other comments or discussion?
8 Ken Bergeron.
9 MR. BERGERON:
We've seen a lot of analyses of the 10 Mark I corium spread and shell attack question, and I wanted to 11 make some comments tiat perhaps may be obvious, may be obvious I
12 to some here, may not be obvious to some about the nature of Y
13 analysis is and the kind of problems you can get into, the kind 14 of apparent uncertainty can be present when people take strong 15 positions.
l 16 It's the nature of the severe accident business that 17 you get into apparent controversy because people take -- people 18 fall in love with the scenarios that they have invented for the 19 starting point of their analysis and then defend them as if 20 they are their children.
21 That's an exaggeration, but there is a tendency to l
22 over-state confidence, I think, in positions.
23 I see accident sequence analysis consisting of 24 several stages.
There's first the scenario selection, model 25 selection, implementation of the model, calculation of the 1
Heritage Reporting Corporation (202) 628-4888
- _=
511 1
model, and then interpretation of it.
82 Now, differences in results between different
- 3 analysts can occur because of differences in any one of those 4
stages. In some areas of accident analysis, and aerosol 5
research is an example, you find differences because of
.edel 6
differences.
7 What I'm saying in this business, it's scenario 8
differences in many cases.
Different choices of initial 9
conditions.
10 Let me give you three examples of scenarios that 11 we've seen.
The calculations that I presented on shell attack 12 were based on what you might call pancake picture of the 13 corium.
That's the starting point for the calculation.
Yo'u 14 have a flat layer of corium in contact with the liner.
That's 15 the starting point.
That's the scenario.
I didn't calculate 16 that.
I assumed it.
17 Mike Corradini did an analysis which talked about the 18 approach to the pancake. One of the things that he did is that 19 he found that there are cases in which the zirconium could be 20 burned out before che pancake reached the edge of the dry well.
21 That makes a very big difference.
22 Theo just described the possibility of large non-23 uniformities in depth of the materia) that's deposited. It 1
24 wouldn't take very much non-uniformity to run out of max before 25 you go to the edge or to have a very thin contact region.
lieritage Re w rting Corporation (202) 628-4888
512 1
These are all possible scenarios.
I think every one 2
of them can be described as plausible.
3 I think, however, that we must face the fact that 4
they will remain uncertain.
The scale of these phenomena is so 5
large that experiments will never address whether Theo is right 6
or whether the pancake picture is right, in my opinion, not 7
never but at least not by August.
8 And, so, my point is for August closure, the authors 9
of this report are going to have to veight menarios.
They're 10 going to have to say this'is possible and that's possible, too, 11 and they're going to have to live with that uncertainty, and to 12 weight the pancake, the hot pancake, picture with zero weight 13 would be very unfortunate given the enormous amount of energy 14 available in the zircon oxidation, and the fact that that 15 amount of energy can reduce the material to a liquid.
16 Even a small weight on that scenario is probably 17 unwise.
However, I don't want to lobby for the picture. I just 18 w :nt
'o say all the scenarios have to be weighted. The research 19 that's been discussed in terms of experiments between now and 20 August will not sharpen everyone to a common picture, and, 21 therefore, the actions that are taken, the recommendations that 22 are made in these reports must act on the basis of these 23 weighted scenarios, and I think that if all of us accept that, 24 we can do that.
The work is for the NRC and for the industry 25 in a spirit of cooperation and shared purpose.
Heritage Reporting Corporation (202) 628-4888
513 1
MR. HULMAN:
Any other comments?
82 Please come up to the microphone and identify 3
yourself.
4 MR. MAINGI:
I'm Stan Maingi from Pennsylvania Bureau 5
of Radiation Protection.
6 I'm glad that I attended this session and I find that 7
the issue is really twofold.
One is that public health and 8
safety with regard to this issue.
We are finding a lot o'f 9
discussion that once the core gets out of the vessel and is on 10 the floor, the uncertainties of that are trying to be coded and 11 I can understand that and that is a good exercise because 12 before the final solution is made, one should have all those 13 things available.
14 However, it is also apparent there's a margin to 15 safety can be certainly improved without even calculating that 16 uncertainty.
That is, we don't expect, and to us at least, to 17 us at state, getting the core melt to the point that it is out 18 of the vessel, is absolutely unsatisfactory.
19 There are things which can be done to mitigate that 20 situation, so that we do not even come tt a point wb.ere the 21 vessel has melted and the core is on the floor.
Those are as i
22 you all know, improvement with the ADS systems, ability to pipe 23 in fire water rather easily, and in the last case, even if you 24 do get a certain amount of damage,to the vessel, release 25 certain amount of -- or in the case of even beyond this loca, lleritage Reporting Corporation (202) 628-4888
514 1
even if you get some large amount of core on the floor, do 82 relieve that through the suppression pool even in Mark I 3
Containments.
4 We are glad that now you have at least pinpointed 5
your schedule, that in April you have an interim report, and in 6
August, you will have a final report, but in the mean time, I 7
hope this discussion will not delay that any further because 8
that is -- we're compromising the margin of safety which is 9
well within the NRC hand to implement.
10 That's all.
11 Thank you.
12 MR. HULMAN:
Thank you.
8 13 Any other -- okay.
Dr. Speis and then I'll close the 14 meeting.
15 DR. SPEIS:
I'm not going to say anything technical.
16 Enough has been said. I just want to thank all of you for 17 coming here, for being subjected to the abuse that has taken 18 place the last two days, the verbal abuse, and I'm sure I have 19 contributed to that to some extent, but I think some of the 29 issues have been clarified.
I think all of us will be better 21 off.
I think it has been extremely useful.
22 So, from myself and my organization, The Office of 23 Research, I want to thank you for coming here and I want to 24 publicly thank here Jerry Hulman, the spirit behind this 25 effort, and hopefully we're not going to meet in Baltimore for Heritbge Re W rting Corporation (202) 628-4888
515 1
awhile because I have been here myself three times within the 82 last six weeks.
3 MR. HULMAN:
That was an energetic disassembly.
4 Housekeeping.
First, the copies of the material have 5
not yet arrived.
If you are still in the hotel by noon, stop 6
by.
Otherwise, on your way out, leave your name and address 7
with Ms. Kondulis or John Lane that's standing next to her, and 8
we'll mail them to you.
9 The record will not be clear in some places, as I 10 understand it.
Some of the material was not picked up or if 11 it's gotten picked up by the court reporter, it may come out a 12 little jumbled.
In order to make sure that the record is 8
13 clear, we may call you for clarification. We won't ask you to 14 add to anything, only to clarify what was said if it's garbled.
15 I again encourage you and, in fact, I'm going to try 16 and issue a challenge, to not only the researchers that we pay 17 for, but to industry, if you can contribute anything more to 18 the issues that were raised here in the way of documentation, 19 in the way of additional assessments, I'd like to hear from 20 you.
21 Realizing that we've made a commitment to the 22 Commission, we're going to try and fulfill it if we can.
23 Also, I want to reiterate the fact that there will be 24 continuing research.
Mr. Hill's comments impressed me the 25 most.
I think what we want to do, and I think everybody would lieritage Re m rting Corporation (202) 628-4888
516 1
agree, is prevent an accident f rom happeriing, and if an 82 accident occurs, we want to mcke sure that it's well managed 3
and if it et all possible prevent any public health and safety 4
i.mpac ts.
I think that's what we're all here for.
5 Last, I'm very impressed with the ability of the 6
technical people in this room.
I think this room represents 7
probably the best brains in the entire world on this subject 8
matter.
We left some at home, there may be a few othera, but 9
principally this constitutes the best people in the world, and 10 my understanding is that many foreign countries look to us fer 11 better knowledge on what goes on in reactors and how best to 12 manage them.
13 with that, I want to close and thank you all very 14 much for coming and making a contribution.
15 (Whereupon, at 10:26 a.m.,
the hearing was 16 concluded.)
17 18 19 20 21 22 23 111 25 lieritage Reporting Corporation (202) 628-4888
l CERTIFICATE S
2
.it it to certify that the attached proceedings before the United States Nuclear Regulatory Commission in the matter ofi
~
1:ame s BWR MARK I CONTAINMENT PERFORMANCE INFORMATION EXCHANGE WORKSHOP 7
Docket Number:
8 Place:
Baltimore, Maryland S
Date:
Februe.ry 25, 1988 10 were held as herein appears, and that this is the original 11 transcript thereof for the file of the United States Nuclear lj Regulatory Commission taken stenographically by me and, 13 thereafter reduced to typewriting by me or under the direction
'A of the court reporting company, and that the transcript is a 1
trueandaccuraterecordofthefp;egoigproceedings.
_i
/S/
b[.
'p " Ec1.h/dA
~.
7
' Signature typed): K.C.
Sekander 18 Official Reporter 19 Heritage Reporting Corporation 20 i
21 22 23 2<
25 8
Heritage P.cporting Corporation (202) 628-4888
I 8
1 e
POTENTIAL AUTOMATIC DEPRESSURIZATION IMPROVEMENTS 1
W. T. PRATT DEPARTMENT OF NUCLEAP. ENERGY BROOKHAVEN NATIONAL LABORATORY UPTON, NEW YORK S
~
PRESENTED AT WORKSHOP DN BWR/ MARK I CONTAINMENT ISSUES FEBRUARY 24-26, 1988 BELVEDERE HOTEL BALTIMORE, MARYLA.99 i
BROOKHAVEN NAll0NAL LABORATORY l} lj l A5500ATED UNIVERSITIES, INC(Illl
l 8
PROBLEM DESCRIPTION AN AREA 0F SIGNIFICANT PHENOMEN0 LOGICAL UNCERTAINTY RELATES TO CORE MELTDOWN WITH THE PRIMARY SYSTEM AT HIGH PRESSURE:
POTENTIAL FOR EARLY CONTAINMENT FAILURE AUTOMATIC DEPRESSURIZATION SYSTEM ( ADS) CAN POTENTIALLY ELIMINATE HIGH PRESSURE CORE NELTDOWN 8
8 A5500ATED UNIVER$1 TIES, ItK. (I til BROOKHAVEN Nail 0NAL LABORATORY l} g3 l
)
POTENTIAL IMPROVEMENTS t
I ACTllATION OF ADS DEPENDS ON ADEQUATE AIR OR N SUPPLY AND I
CONTROL POWER TO THE SOLEN 0 IDS:
ADDRESSED IN NUREG-0737 IN LONG-TERM LOSS OF AC POWER ACCIDENT BATTERIES MAY BE DEPLETED RENDERING ADS INOPERABLE:
ALTERNATE DEDICATED BACKUP POWER (BATTERIES OR SMALL GENERATOR DRIVEN POWER SUPPLY AND APPROPRI ATE CONNECTIONS)
FOR ADS COULD INCREASE AVAILABILITY l
CABLING FOR ADS MAY NEED THERMAL INSULATION OR SHIELDING.TO ASSURE OPERABILITY IN HOT DRYWELL ATMOSPHERE DURING LONG-TERM i
STATION BLACK 0UT BROOKHAVEN Nail 0NAL LABORATORY l} l)l ASSOCIATED UNIVERSITIES, INC. ll Ell
l.
S QUESTIONS TO BE ADDRESSED ARE THE TYPES OF IMPROVEMENTS CESCRIBED SUFFICIENT TO IMPROVE ADS RELIABILITY?
WHAT ARE OTHER RECOMMENDATIONS?
WHAT ARE THE BENEFITS?
FOR EXAMPLE, IS INCREASED AVAILABILITY OF THE ADS SYSTEM NECESSARY, AND WILL IT ELIMINATE THE RISK FROM DIRECT CONTAINMENT HEATING?
i WHAT ARE POTENTIAL ADVERSE INTERACTIONS BETWEEN ANY OF THE PROPOSED IMPROVEMENTS AND EXISTING SAFETY SYSTEMS?
ARE THERE OTHER IMPROVEMENTS THAT MIGHT ALSO BE TECHNICALLY FEASIBLE TO FRESERVE THE MARK I CONTAINMENT FUNCTION IN THE EVENT OF LOW PROBABILITY CHALLENGES?
8
The Effect of Sequence Definition h
.on Accident Progression (Station Blackout)
Gregory A. Krueger Philadelphia Electric Company 8
1 BWR Mark l Workshop Baltimore, MD Feb. 24-26,1988 8
TBUX-Loss of Offsite Power and h
Simultaneous Loss of Station Batteries
- Was Calculated as the Dominant Sequence for Peach Bottom in Draft NUREG/CR-4550
' is Used As A Basis For Current Calculations g
- Reanalysis in NUREG/CR-4550 Indicates an Order of Magnitude Decrease in Probability 8
i Common Cause Battery Failure 8
Draft NUREG/CR-4550 Assumed a 4% Failure Probability for the Second Battery Falling Given the First and a 100% Failure Probability for the Remaining Fal!ure of 5 Batteries was Cals:ulated as (1.3E-3)(.04)(1)(1)(1) = 5.2E-5 Reanalysis uses Readjusted Numbers Based Upon Actual Plant Maintainance Practices 8
1 Failure is Now Approximately Calculated as I
(1.3E-3)(.04)(.52)(.76)(.88) = 1.8E-S 8
r
ADS Reliability
- ADS is Qualified for Design Basis Conditions
- Manual Depressurization is Required By Procedures
- Reliability is Questioned due to Unrealistic Battery Assumptions S
8
i TB - Loss of Offsite and Onsite AC Power 8
- Long Term injection and ADS Available i
- Reanalysis indicates it is the Current Dominant Sequence l
1 i
8 1
2
)
8
Significant Timing and Containment Response Exists Between the Two Sequences EVENT (Time in Minutes)
TBUX TB Water Level at the Top of Core 27 630 i
Relocation of Core Debris Begins 76 738 l
Vessel Bottom Head Penetration Failure
- 194 902 Drywell Pressure at Vessel Failure (psia) 55 100 i
l
- Vessel Failure Occurs at High Pressure in Soth Sequences
Differences g
U Manual Depressurization Gains Significant Time to Core Uncovery Suppression Pool is Saturated - Possible Steam inerting in Secondary Containment i
i Possible Containment Failure in Torus Airspace Due to Overpressure h
Much Longer Time for Debris to Reach Shell from Gnset of Accident l
l l
- The Perceived Benefit of Improvements Could Be Misleading and Are Based Upon:
j
- A Limited Set of Accident Sequences
- Low Probability, Non-Dominant Events i
l
- Assumptions / Estimations in the ' Front End' Analyses S
i t
- -, - -,,, - -,, - - - - - - -, -,